Establishment of Galleria mellonella as a model for Achromobacter xylosoxidans infection

TL;DR

This paper shows that wax moths can be used as a model to study Achromobacter xylosoxidans infection and immune responses, similar to mouse models.

Contribution

The study establishes Galleria mellonella as a valid infection model for Achromobacter xylosoxidans, mirroring mouse outcomes and immune interactions.

Findings

Mortality in Galleria mellonella after Ax infection mirrors survival in infected mice.

Ax is internalized into Galleria hemocytes, similar to vertebrate immune responses.

The Galleria model can assess antimicrobial activity against Ax.

Abstract

Achromobacter xylosoxidans (Ax) is an emerging pathogen with a strong capacity to adapt to different niches, but its pathogenesis is poorly understood. To investigate the virulence of this versatile bacterium, alternative infection models are valuable.

Galleria mellonella wax moth presents cost and ethical advantages as an in vivo infection model. Here, we investigate the utility of Galleria as a model of Ax infection.

We demonstrate that mortality following Ax infection in Galleria recapitulates survival outcomes observed in infected mice. We further show that the Galleria infection model can be used to examine antimicrobial activity against Ax. Visualization of hemocytes suggested that Ax was internalized into immune cells, similar to what is observed in vertebrate models.

Overall, our work establishes Galleria mellonella as a model of Ax infection that mirrors disease severity and innate immune cell interactions in murine models.

Figures

Survival of mice and Galleria mellonella larvae infected with Achromobacter xylosoxidans GN050 and GN008. Eight-week-old mice were infected with different doses of GN050 (A) and GN008 (B) by intratracheal instillation; n = 8. Galleria mellonella larvae were infected with different doses of GN050 (C) and GN008 (D); n = 25. Survival was monitored up to 96 h post infection (hpi). p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. *p < 0.001, p < 0.0001.

Survival of mice and Galleria mellonella larvae infected with Achromobacter xylosoxidans NIH-018 series. Galleria mellonella larvae were infected with different doses of NIH-018-1 (A), NIH-018-2 (B), and NIH-018-3 (C); two biological replicates, n = 25/replicate. Eight-week-old mice were infected with different doses of NIH-018-1 (D), NIH-018-2 (E), and NIH-018-3 (F) by intratracheal instillation; n = 4. Survival was monitored up to 96 hpi. p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. *p < 0.001, p < 0.0001.

Survival of infected Galleria mellonella larvae under imipenem (IMI) treatment. Galleria larvae were infected with 102 CFU (A) or 104 CFU (B) of GN050, then treated with different doses of IMI at 4 hpi. Survival was monitored up to 96 hpi. (C) PBS control was injected to Galleria followed by injection of IMI at 4 hpi for toxicity assessment. p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. p < 0.0001.

Growth and association of Achromobacter xylosoxidans in Galleria. Galleria larvae were infected with GN050 (A) and GN008 (B) at 102 CFU and 103 CFU, respectively. Bacterial growth (CFU) in hemolymph was monitored every 24 h until 96 hpi; n = 5. All GN050 infected larvae were either dead or euthanized by 72 hpi. Galleria larvae were infected with different Ax isolates, including GN050, GN008, NIH-018-1, NIH-018-2, and NIH-018-3 at 103 CFU. Ax attachment to hemocytes (C,D) and internalization into hemocytes (E,F) were monitored at 24 and 48 hpi; n = 8. p-values were determined by nonparametric one-way ANOVA with Dunn’s multiple comparisons test, comparing other time points to time 0 in (A,B) or comparing other isolates to GN050 in (C–F). *p < 0.05, p < 0.01, *p < 0.001, p < 0.0001.

Uptake of Ax by hemocytes. Galleria larvae were infected with GN050-GFP at 102 CFU. At 24 hpi, hemocytes were collected, stained, and visualized using fluorescence microscopy. Representative images of identified hemocyte populations were divided into Cell type 1: oenocytoids-like cells, Cell Type 2: granulocytes-like cells, and Cell Type 3: plasmatocytes-like cells. DAPI: cell nucleus, TRITC: GN050-GFP, Cy5: CellMask Deep Red plasma membrane stain. Membrane outlines and merged images are shown.

Full text

Introduction

Achromobacter xylosoxidans (Ax) is a gram-negative emerging opportunistic pathogen that is highly associated with immunocompromised patients (Giacoia, 1990; Ronin et al., 2021; Bayless et al., 2025). In addition to lung infections, Ax can cause skin and wound infections, urinary tract infections (UTIs), and systemic infections (Tena et al., 2014; Isler et al., 2022). These infections can cause severe organ damage as a result of bacteremia and dissemination to immune-privileged sites, such as the central nervous system (Manckoundia et al., 2011; Imani et al., 2021). Ax can cause recurrent and prolonged infections and is highly resistant to antibiotic treatment (Weitkamp et al., 2000; Acharya et al., 2025a; Isler et al., 2020). However, Ax has been largely understudied (Billiot et al., 2023; Turton et al., 2023).

To investigate Ax infection, in vitro cell based assays and in vivo mouse assays were used to model infection (Billiot et al., 2023; Wills et al., 2023). While in vitro models are useful for screening and characterizing virulence factors, they can fail to capture the complexity of in vivo infection. Indeed, infections are known to elicit different effects in vivo than those observed in vitro (Stratton, 2006; Hoogeterp et al., 1989). Thus, identification of an alternative model that recapitulates the in vivo murine model while reducing cost and ethical concerns would be beneficial.

Galleria mellonella is an invertebrate species that has gained increasing attention as an alternative to mammalian models of infection (Mai et al., 2023; Axline et al., 2025; Prakoso et al., 2022; Menard et al., 2021). Although Galleria lacks an adaptive immune response, its innate immune response is strikingly similar to that of mammals (Gallorini et al., 2024; Campbell et al., 2024). Therefore, Galleria has been employed to assess infection-associated mortality and treatment efficacy for related pathogens, including Pseudomonas aeruginosa and Bordetella bronchiseptica (Sun et al., 2022; White et al., 2025; Michetti et al., 2025; Szymczak et al., 2020). Galleria has also become a widely used model for studying infection caused by a wide range of agents, from bacteria to fungi (Smith et al., 2025; Primm et al., 2025). Development of a Galleria mellonella infection model would provide a cost-effective, time-efficient, and ethically improved alternative to murine models. Here, we describe the application of Galleria mellonella as an in vivo infection model by comparing infection outcomes in Galleria to those observed in a murine intratracheal instillation model and by demonstrating the use of this model in examining antibiotic treatment of Ax infection. We establish Galleria as an intermediary infection model for Ax that bridges in vitro assays and murine models.

Methods

Bacterial culture

Achromobacter xylosoxidans isolates (GN050, GN008, NIH-018-1, NIH-018-2, and NIH-018-3) were obtained from previous work (Pickrum et al., 2020; Ray et al., 2025). GN050-gfp was obtained as described below. All Ax bacterial strains used in this study were first streaked onto Luria-Bertani Agar (LBA) (10 g/L tryptone (RPI CAS #91079-40-2), 5 g/L Yeast extract (RPI Cat. No. 8013-01-2), 10 g/L NaCl (RPI Cat. No. 7647-14-5), 15 g/L Bacteriological Grade Agar (RPI Cat. No. 9002-18-0)) containing no antibiotics and grown at 37 °C overnight. Strains were then incubated at room temperature for 24 h before infection.

Preparation of Galleria moth larva

Galleria mellonella larvae were acquired from Waxworms (waxworms.com). After delivery, larvae were portioned out into 100 mm Petri dishes, with 25–50 larvae per dish. Larvae weighing between 180 and 250 mg per larva and exhibiting ideal coloration (i.e., light tan with minimal or no spots or markings) were selected. This weight range correspond to 6th instar stage, which has been used in previous studies (Lange et al., 2019; Asai et al., 2019; Ten et al., 2023). Larvae were then stored at 37 °C overnight to ensure they were starved for at least 24 h prior to infection.

Galleria larva infection

Following 2 days of incubation as described above, bacteria were collected from the plates, suspended in phosphate-buffered saline (PBS, Gibco Cat. No. 14190-144) and optical density at 600 nm (OD600) was measured using a spectrophotometer (Eppendorf BioPhotometer 6131). The resulting suspension was diluted in PBS to the OD600 of 0.5, corresponding to 10^9^ CFU/mL. The suspension was then serially diluted in PBS to achieve concentrations of 10^8^, 10^6^, 10^4^ CFU/mL, resulting in final inocula of 10^6^, 10^4^, and 10^2^ CFU per 10 μL injection volume. These inocula correspond to 10^6^, 10^4^, and 10^2^ CFU per larva, or approximately 5,000, 50, and 0.5 CFU per mg of larval body weight, respectively. We selected these inocula based on our pilot experiments to identify a dose range that resulted in minimal killing in GN008 and significant killing in GN050, consistent with observations in murine model (Wills et al., 2023).

Following dilution, 100 μL of the bacterial suspension was plated out onto LBA without antibiotics to validate the actual inoculum used. Plates were incubated at 37 °C overnight, and colony-forming units (CFUs) were counted after 24–48 h of incubation, depending on growth rate. Using the same inoculum, larvae were infected with 10 μL of the prepared inoculum into the left side of the last ventral proleg using a 27-gage needle attached to a 1 mL syringe. The injection was done using a KD Scientific syringe pump set to 7.20 mL/h to deliver a total volume of 0.01 mL. Following inoculation, the larvae were moved to 37 °C and monitored every 12 h to assess mortality. Dead larvae were removed from the population. For antibiotic treatment, after a 4-h incubation at 37 °C, 10 μL of imipenem (Goldbio Cat. No. I-600-1) suspended in PBS was injected into the right ventral proleg adjacent to the last proleg of the larva. Imipenem concentrations were performed at 5 ng, 1.25 ng, or 0.3125 ng/mL, corresponding to approximately 0.2 μg, 0.05 μg, and 0.0125 μg per larva, respectively. Larvae were then monitored at 8 h post-treatment and every 12 h thereafter.

Measurement of bacterial CFU inside hemolymph

Hemocyte collection and fluorescence microscopy were performed using previously described methodologies, incorporating established polymyxin B treatment procedures (Pickrum et al., 2020; Acharya et al., 2025b; Admella and Torrents, 2022). Following the Galleria infection, five Galleria larvae were collected from the live larvae every 24 h starting from 0 to 96 h post-infection (hpi). The larvae were transferred individually to a − 20 °C freezer for 3 min or until they became inactive (anesthetized). The larvae were then submerged in prechilled 70% ethanol for 20 s to achieve surface sterilization and anesthesia, and briefly dried on a sterile towel. The base of the larva was removed using sterile surgical scissors, and the hemolymph was gently squeezed out into a 1.5 mL tube containing 100 μL hemolymph anticoagulant solution (26 mM sodium citrate (Fisher cat BP327-1), 30 mM citric acid (Thermoscientific Cat. No. 036664.36), 100 mM glucose (Thermoscientific Cat. No. 41095-5000), 140 mM NaCl (RPI Cat. No. 7647-14-5), pH 4.11) (Admella and Torrents, 2022) using blunt-ended forceps. The tube was spun down at 200 × g for 10 min (Eppendorf Centrifuge 5420), resuspended in 100 μL anticoagulant solution, and plated to determine the total CFU. To assess adherent and internalized bacteria per hemocyte, pelleted cells were resuspended in 200 μL anticoagulant and the wash step was repeated 2 times. The cell suspension was split into two aliquots: one aliquot contained no antibiotics to measure total cell-associated bacteria, i.e., adherent and internalized bacteria, and the other was treated with 50 μg/mL polymyxin B (RPI Cat. No. 1405-20-5) for 1 h at room temperature to kill off bacteria outside hemolymph and quantify internalized bacteria. The number of hemocytes per mL was calculated using a hemocytometer and trypan blue staining. Each aliquot was then plated on LBA containing 20 μg/mL chloramphenicol (RPI Cat. No. 56-75-7).

Tn7 fluorescent labeling of Achromobacter xylosoxidans

Fluorescent labeling of Ax strains was conducted by Tn7 mutagenesis (Choi and Schweizer, 2006). The strain GN050 was first grown overnight at 37 °C in 3 mL LB without antibiotics. Concurrently, two strains of DH5α S17λpir E. coli containing either pGPTn7: GFP (NovoPro) or pUXBF13 (Addgene) were grown overnight in 3 mL LB supplemented with 100 μg/mL carbenicillin (Fisher Cat. No. 4800-94-6). Overnight cultures were spun down at 6000 rpm for 5 min, and the supernatant was removed. Cell pellets were resuspended in 1 mL LB, mixed at a 1:1:1 ratio, and spun down at 6000 rpm for 5 min. The resulting pellet was resuspended in 100 μL LB, spot-plated onto LBA and incubated for 48 h at 37 °C. Single colonies of Ax were streaked onto LBA supplemented with kanamycin (20 μg/mL) (Fisher Cat. No. 25389-94-0) and chloramphenicol (100 μg/mL) (RPI Cat. No. C61000). Insertion of gfp genewas confirmed by PCR using the screening primers OHL706 (GCAGGAAAGAAACGTCGCGGGT) and OHL707 (ATTTCACATCTTTCTTTCCG). The presence of GFP-producing Ax colonies was confirmed by visualizing the bacteria under a fluorescent microscope and GFP^+^ colony was chosen for downstream fluorescence microscopy.

Fluorescence microscopy of hemolymph

Prior to hemolymph collection, coverslips were sterilized in 70% ethanol and allowed to air dry. Sterilized coverslips were placed into a 6-well plate and treated with 1 mL of 0.01% poly-L-lysine for 10 min at room temperature. The coverslips were carefully removed using sharp-ended forceps and excess poly-L-lysine solution was removed. They were washed 3 times with 1 mL PBS and transferred to a covered container at 4 °C overnight to dry. Galleria hemolymph was collected as noted above, except that the wash solution was supplemented with 1 μL/mL CellMask Deep Red (Invitrogen Cat. No. C10046) and 0.0734 μL/mL Hoechst 33432 (Thermoscientific Cat. No. 62249). Following the initial spin at 200 × g, the pelleted cells were washed twice with 500 μL anticoagulant solution. The resulting cell suspension was applied to the prepared coverslips, spun onto the coverslip at 200 × g for 10 min and allowed to settle for 10 min at room temperature. Excess material was removed, and the coverslips were washed 3 times with 2 mL PBS (Thermo scientific Cat. No. 28908). Cells were fixed with 1 mL 4% paraformaldehyde (PFA, Fisherscientific Cat. No. 50-980-494) for 10 min at room temperature, washed 3 times with PBS, and incubated with 1 mL 125 mM glycine (G-biosciences Cat. No. RC-055) for 10 min. Excess glycine was removed, and coverslips were mounted on slides using Prolong glass antifade mountant (Life Technologies Corperation Cat. No. P36984). Cells were imaged using a NIKON AXR fluorescent confocal microscope at 60X magnification using DAPI, TRITC, and Cy5 channels. Images were acquired as Z-stacks (0.2 μm slices, 1,024 × 1,024 pixels). Final images were generated using maximum-intensity Z-axis projections for each channel. Image processing was done using ImageJ. For all images, maximum and minimum brightness levels were adjusted to the same threshold to remove background prior to generating composite images. Cell outlines were collected by softening the cell membrane channel three times, followed by image thresholding, after which the resulting outline was obtained.

Mouse infections

To assess mouse survival following Ax infection, we employed an intratracheal instillation model as previously described (Wills et al., 2023) using 8-week-old C57BL/6 J male mice. Eight mice per group were used for infections with GN050 and GN008, and four mice per group were used for infection with NIH-018-1, NIH-018-2, and NIH-018-3. Three inocula were examined, including 5 × 10^7^ CFU, 1 × 10^7^ CFU, and 5 × 10^6^ CFU. Mice were anesthetized by exposure to isoflurane (VetOne Fluriso, NDC 13985-528-60) at a flow rate of 500 mL/min at 5% using a SomnoFlo electronic vaporizer (Kent Scientific, #SF-01). A total of 50 μL of inoculum was then administered to each mouse via the intratracheal instillation method. Immediately following infection, mice were weighed to determine initial weights. Initial weights ranged between 22 and 28 g per mouse. Mice were monitored every 8 h for weight loss, activity level, and physical appearance. Physical Appearance: 0 = Normal, BAR (bright, alert, responsive); 1 = BAR, mildly hunched and ruffled, abnormal stance, mild ocular or nasal discharge; 2 = QAR (quiet, alert, responsive), hunched, squinted eyes, ears pinned, ruffled fur, heavy ocular or nasal discharge, increased breathing; 3 = Unresponsive, recumbent, eyes closed, severely hunched, labored breathing. Activity: 0 = BAR, normal; 1 = BAR, abnormal gait, mildly lethargic; 2 = QAR, reluctant to move, lethargic; 3 = Moribund/inability to move. Weight Change (loss and/or body condition score): 0 = <10% and/or body condition score (BCS) 3–5; 1 = 10–15% weight loss and/or BCS 2 (wet food will be provided); 2 = 16–20% weight loss and/or BCS 2 (wet food will be provided); 3 = >20% weight loss and/or BCS ≤ 1. Once mice returned to a clinical score of ≤2, the observation period was extended to every 12 h. Mice with a total clinical score of ≥5 or a score of 3 in any single category were considered moribund and euthanized. Euthanasia was conducted by 30–70% CO_2_ displacement followed by secondary cardiac exsanguination.

Statistical analysis

Statistical assessment of infection data was performed using Kaplan–Meier survival analysis with multiple-comparison correction done using the Benjamini-Hochberg test to control the false discovery rate. Statistical analyses of other data were performed using one-way ANOVA with Dunnett’s multiple comparisons test, unless otherwise noted. All statistical analyses were performed using GraphPad Prism software version 10.

Results

Survival of Galleria mellonella larvae infected with Achromobacter xylosoxidans recapitulates infection outcomes observed in mouse models

To investigate whether Galleria mellonella is an effective model for assessing disease severity of Ax infection, we compared Galleria larval survival with murine survival using the clinical isolates GN050 and GN008. GN050 and GN008 were previously characterized as cytotoxic and non-cytotoxic isolates, respectively (Pickrum et al., 2022), and have been used in murine infection models to investigate pathogenesis and immune responses during respiratory infection (Turton et al., 2023; Wills et al., 2023). We therefore used these strains for the initial validation of Ax infection in Galleria.

We performed an intratracheal infection of C57BL/6 J mice with Ax. 8-week-old mice were inoculated with either 5 × 10^7^ CFU, 1 × 10^7^ CFU, or 5 × 10^6^ CFU and monitored for up to 96 h post infection (hpi). Mice infected with GN050 at 5 × 10^7^ CFU showed extremely severe symptoms, requiring euthanasia at the highest inoculum by 30 hpi (Figure 1A; Supplementary Figures S1A–C,P). Infection with either 1 × 10^7^ CFU or 5 × 10^6^ CFU of GN050 resulted in ≥50% mortality by 60 hpi. GN008, by contrast, showed substantially less severe disease and required the highest dose 5 × 10^7^ CFU to significantly reduce mouse survival, with the majority of mice requiring euthanasia due to weight loss rather than elevated overall clinical scores (Figure 1B; Supplementary Figures S1D–F,Q).

*Survival of mice and Galleria mellonella larvae infected with Achromobacter xylosoxidans GN050 and GN008. Eight-week-old mice were infected with different doses of GN050 (A) and GN008 (B) by intratracheal instillation; n = 8. Galleria mellonella larvae were infected with different doses of GN050 (C) and GN008 (D); n = 25. Survival was monitored up to 96 h post infection (hpi). p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. ***p < 0.001, ***p < 0.0001.

We then infected Galleria mellonella larvae with GN050 and GN008. The procedure was outlined in Supplementary Figure S2. Galleria larvae infected with GN050 required only approximately 1 CFU to achieve about 60% mortality by 96 hpi (Figure 1C). By contrast, GN008 did not result in substantial mortality until the highest dosage tested (1 × 10^6^ CFU) which resulted in 100% mortality by 48 hpi (Figure 1D). These results suggest that survival outcomes in the Galleria infection model followed the same trend as those observed in murine models.

The Galleria infection model reveals a disconnect between in vivo virulence and in vitro cytotoxicity

We have access to NIH clinical isolates from previously published work (Acharya et al., 2025a; Ray et al., 2025). Among them, the NIH-018 series was of particular interest. NIH-018-3, NIH-018-2, and NIH-018-1 were collected roughly 1 year apart, with NIH-018-3 collected first and NIH-018-1 collected last. In vitro characterization revealed a high degree of cytotoxicity among these isolates, which appeared to increase over time, and exceeded what was observed in GN050 (Acharya et al., 2025a). Thus, we were interested in further investigating their in vivo virulence. Following the same methodology as with GN050 and GN008, we infected Galleria larvae with the NIH-018 series clinical isolates (Figures 2A–C). Surprisingly, the most cytotoxic isolate, NIH-018-1, was by far the least virulent of the strains tested and showed similar mortality to GN008. The second biological replicate exhibited the same results. These findings suggest that in vivo virulence can diverge from in vitro cytotoxic phenotypes.

*Survival of mice and Galleria mellonella larvae infected with Achromobacter xylosoxidans NIH-018 series. Galleria mellonella larvae were infected with different doses of NIH-018-1 (A), NIH-018-2 (B), and NIH-018-3 (C); two biological replicates, n = 25/replicate. Eight-week-old mice were infected with different doses of NIH-018-1 (D), NIH-018-2 (E), and NIH-018-3 (F) by intratracheal instillation; n = 4. Survival was monitored up to 96 hpi. p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. ***p < 0.001, ***p < 0.0001.

To further validate whether the survival trend observed in Galleria reflected outcomes in murine models, we infected mice with the NIH-018 series. The results showed that infection of mice with the NIH-018 series exhibited survival patterns similar to those observed in Galleria (Figures 2D–F; Supplementary Figures S1G–O,R–T). All mice infected with NIH-018 series isolates at the 5 × 10^7^ CFU inoculum showed extremely severe symptoms by 16 hpi, necessitating euthanasia as a result of morbidity (Supplementary Figures S1G–O,R–T). NIH-018-1 infection resulted in no survival at 5 × 10^7^ CFU inoculum, but no morbidity resulting from 1 × 10^7^ CFU or 5 × 10^6^ CFU inoculum (Figure 2D). NIH-018-2 demonstrated an intermediate level of morbidity in mice with no survival at 5 × 10^7^ CFU inoculum, and no statistically significant reduction in survival at the 10^7^ CFU inoculum (Figure 2E). NIH-018-3, by contrast, showed the most severe infection with significant morbidity following infection at 1×10^7^ and 5 × 10^7^ CFU (Figure 2F). Assessment of the clinical scores and body weight indicated greater disease severity at the onset of infection in our NIH-018 strains compared to GN050 or GN008 (Supplementary Figure S2). Overall, infection outcomes in mice infected with Ax isolates mirrored those observed in Galleria, while both differed markedly from trends observed in in vitro assays.

Antibiotic treatment reduces Achromobacter-induced mortality in Galleria mellonella

Ax is known to be highly resistant to many antibiotics, posing a significant challenge for patient treatment (Acharya et al., 2025a; Hu et al., 2015) and highlighting the urgent need for novel antimicrobial discovery and development. Galleria has been employed as a model to evaluate antibiotic efficacy against other pathogens, such as Acinetobacter and Pseudomonas (White et al., 2025; Maslova et al., 2023; Tsai et al., 2016; Barton et al., 2024). Antibiotics are known to exhibit different effects in vivo compared to in vitro (de Araujo et al., 2011). Thus, to facilitate the development of novel antimicrobial treatment strategies, we examined whether Galleria could be used to evaluate the in vivo efficacy of antibiotics against Ax.

Galleria larvae were infected with the strain GN050 at inocula of 1 × 10^2^ CFU and 1 × 10^4^ CFU. At 4 hpi, larvae were treated with imipenem (IMI), a broad-spectrum carbapenem used clinically to treat Ax infections (Acharya et al., 2025a). IMI-treated infected larvae showed a substantial reduction in mortality, with survival extending to at least 4 days post-infection at both GN050 inocula when treated with imipenem at 12.5 μg/mL or higher (Figures 3A,B). As a control, uninfected larvae treated with IMI showed no larval death, indicating that the antibiotic concentrations tested were not toxic to the larvae (Figure 3C). These results suggest that Galleria is a promising model for evaluating the in vivo effectiveness of antimicrobial agents against Ax.

*Survival of infected Galleria mellonella larvae under imipenem (IMI) treatment. Galleria larvae were infected with 102 CFU (A) or 104 CFU (B) of GN050, then treated with different doses of IMI at 4 hpi. Survival was monitored up to 96 hpi. (C) PBS control was injected to Galleria followed by injection of IMI at 4 hpi for toxicity assessment. p-values were calculated using Kaplan–Meier survival analysis with correction for multiple comparisons by controlling the false discovery rate. ***p < 0.0001.

Growth of Achromobacter xylosoxidans and its association with hemocytes in Galleria mellonella

We monitored the total CFUs in larval hemolymph over time to evaluate how effectively Galleria larvae clear infection (Figures 4A,B). We observed an initial expansion of GN008 at 24 hpi, corresponding to a 10-fold increase and over a 1,000-fold increase for GN050. Subsequent reduction in CFUs in GN008-infected larvae indicated clearance of bacteria from the hemolymph. GN050 showed a slower reduction with CFUs remaining 100-fold above the inoculum at 72 hpi.

*Growth and association of Achromobacter xylosoxidans in Galleria. Galleria larvae were infected with GN050 (A) and GN008 (B) at 102 CFU and 103 CFU, respectively. Bacterial growth (CFU) in hemolymph was monitored every 24 h until 96 hpi; n = 5. All GN050 infected larvae were either dead or euthanized by 72 hpi. Galleria larvae were infected with different Ax isolates, including GN050, GN008, NIH-018-1, NIH-018-2, and NIH-018-3 at 103 CFU. Ax attachment to hemocytes (C,D) and internalization into hemocytes (E,F) were monitored at 24 and 48 hpi; n = 8. p-values were determined by nonparametric one-way ANOVA with Dunn’s multiple comparisons test, comparing other time points to time 0 in (A,B) or comparing other isolates to GN050 in (C–F). *p < 0.05, **p < 0.01, ***p < 0.001, ***p < 0.0001.

Since Ax has been shown to adhere to and be internalized by macrophages in vitro (Acharya et al., 2025a; Turton et al., 2023), we investigated whether Ax can similarly adhere to and become internalized by Galleria hemocytes. In addition, it was also shown that different Ax isolates can behave differently (Acharya et al., 2025a; Turton et al., 2023). Thus, Galleria larvae were infected with different Ax isolates, GN050, NIH-018-2, and NIH-018-3 at 1 × 10^2^ CFU and GN008 and NIH-018-1 at 1 × 10^4^ CFU. These doses were chosen to ensure sufficient survival for monitoring of Galleria at 24 and 48 hpi, starting from a population of 30 larvae per condition (Figures 4C–F). Measurement of bacterial attachment to hemocytes revealed an overall low percentage of bacteria associated with hemocytes, with the NIH-018 series showing lower attachment than GN050 at 48 hpi (Figures 4C,D), but comparable levels of uptake, with the exception of NIH-018-1 (Figures 4E,F). NIH-018-1 exhibited neither attachment nor internalization. GN008 was also not internalized by hemocytes. These results suggest that internalization into hemocytes is associated with overall disease severity in vivo.

Localization of Ax during Galleria infection

In order to verify the presence of Ax in Galleria hemocytes, fluorescent imaging was performed on hemocytes collected at 24 hpi from larvae infected with a Tn7-GFP-labeled Achromobacter GN050. Among the collected hemocytes, very few cells appeared to contain bacteria, consistent with our observations when measuring CFUs per hemocyte (Figure 4). Among the hemocytes observed to harbor bacteria, the majority appeared to be morphologically similar to oenocytoids (Figure 5, Cell Type 1), which are known to be phagocytic (Wu et al., 2016). Additional cell morphologies consistent with granulocyte-like (Figure 5, Cell Type 2) and plasmatocyte-like hemocytes (Figure 5, Cell Type 3) were also identified. Granulocytes, which function through lysis or degranulation to promote plasmatocyte binding, did not appear to contain Ax bacteria. In contrast, plasmatocytes were observed to contain bacteria, a finding consistent with previous studies (Wu et al., 2016; Browne et al., 2013). These results suggest that Ax is internalized by phagocytic innate immune cells in Galleria, analogous to what occurs during vertebrate infection.

Uptake of Ax by hemocytes. Galleria larvae were infected with GN050-GFP at 102 CFU. At 24 hpi, hemocytes were collected, stained, and visualized using fluorescence microscopy. Representative images of identified hemocyte populations were divided into Cell type 1: oenocytoids-like cells, Cell Type 2: granulocytes-like cells, and Cell Type 3: plasmatocytes-like cells. DAPI: cell nucleus, TRITC: GN050-GFP, Cy5: CellMask Deep Red plasma membrane stain. Membrane outlines and merged images are shown.

Discussion

Investigation of the pathogenesis and virulence of novel pathogens can be expensive and ethically demanding due to the use of mice or other vertebrate models of infection. Here, we present an alternative model for studying Ax infection and antimicrobial discovery. Galleria larvae are substantially less expensive and raise fewer ethical considerations than vertebrate models, making them an excellent alternative in vivo model. Our Galleria infection model recapitulated the same survival trends observed in murine models (Figure 1), yet these in vivo trends diverged from the in vitro cytotoxicity data (Figure 2; Acharya et al., 2025a). Previous studies have suggested the similar trend that different clinical isolates may have disparities in infection dynamics and severity that are not correlated with in vitro assays (Acharya et al., 2025a; Wills et al., 2023). Therefore, using Galleria as an in vivo model in virulence studies can mitigate the substantial costs and ethical burden associated with mouse models.

Achromobacter sp. is known for highly antibiotic-resistant phenotypes (Hu et al., 2015; Bador et al., 2013). Galleria larvae have been utilized as a model for assessing the effectiveness of antibiotics in vivo against infections caused by other pathogens (White et al., 2025; Barton et al., 2024). Here, we validated that Galleria can serve as an effective in vivo model for evaluating antibiotic treatment against Ax (Figure 3). Beyond antibiotic testing, Galleria has also been utilized to characterize the in vivo effectiveness of virulence inhibitors (Sun et al., 2022; Song et al., 2024). The utility of such a model becomes apparent in the context of Ax virulence factors, such as ArtA, an RTX adhesin important for adherence and cell invasion (Wills et al., 2023). The design of small molecules to inhibit adhesins and other virulence factors to reduce disease severity may be a promising area of research, in which the Galleria model can be used to assess in vivo efficacy. In short, the Galleria model can serve as an initial in vivo model to test the effectiveness of antimicrobials, facilitating novel drug discovery and development.

Cytotoxic Ax isolates are known to kill innate immune cells such as macrophages, while non-cytotoxic isolates are less capable of causing cell death (Acharya et al., 2025a; Pickrum et al., 2020). In the Galleria infection model, GN050 (a cytotoxic isolate) was able to multiply rapidly by 24 hpi and maintain a high number of bacteria inside Galleria at later time points, although the CFU counts dropped after 24 hpi. The reduction in GN050 growth could be due either to immune-mediated bacterial killing or to nutrient limitation resulting from the high bacterial density inside Galleria. Given the ~1,000-fold increase in GN050 bacterial numbers, the latter explanation may be more likely. In contrast, GN008 showed limited replication and a significant reduction in bacterial counts inside Galleria by 72 hpi, suggesting rapid and effective clearance of GN008 by Galleria. Since Galleria lacks an adaptive immune system, these findings suggest that innate immunity plays an important role in controlling Ax infection; however, more cytotoxic and virulent isolates may be able to evade or counteract the innate immune system, leading to persistence or proliferation, indicating that adaptive immunity may contribute to host defense against Ax to avoid chronic infection.

Assessment of Ax adhesion to and internalization by Galleria hemocytes revealed an overall low level of bacterial association. GN050 attached to approximately 10% of hemocytes, while other isolates attached to less than 1% of the cells at 48 hpi (Figure 4). Uptake of GN050, NIH-018-2, and NIH-018-3 was comparable and occurred in around 0.1% of hemocytes while uptake of GN008 and NIH-018-1 was significantly lower than that of the other isolates. In agreement with bacterial uptake trends, disease severity in mice and Galleria was much higher for GN050, NIH-018-2 and NIH-018-3 compared to the other two isolates. These observations suggest that uptake of Ax by phagocytic cells is associated with disease severity. Moreover, in our previous in vitro studies (Acharya et al., 2025a), attachment and uptake among NIH-018 isolates were similar; however, they differed in the Galleria infection model, with NIH-018-1 exhibiting the least uptake. Moreover, NIH-018-1 is the most cytotoxic isolate in vitro, but causes the least severe infection among the NIH-018 isolates. These results highlight the complexity of host-pathogen interactions in vivo, which could be missed in in vitro experiments.

To visualize internalized bacteria in hemocytes, we performed fluorescent imaging of GN050-GFP infecting Galleria larvae. Hemocytes were collected at 24 hpi and visualized. Oenocytoids, granulocytes, and plasmatocytes are the main phagocytic cells in Galleria, and granulocytes can take up E. coli at the fastest rate (Wu et al., 2016). From our in vivo samples, Ax was found inside oenocytoids and plasmatocytes, but not granulocytes (Figure 5). This discrepancy may reflect enhanced susceptibility of granulocytes to GN050-mediated killing or a pathogen-specific immune response. Further studies are needed to investigate which specific immune cells are critical for responding to Ax, which is beyond the scope of this study.

Galleria model offers several advantages over other non-murine models. Beyond the lungs, Ax can infect other tissues including skin, bone, and, in extreme cases, the central nervous system (e.g., meningitis). Ex vivo models (organoids, human skin and human bone) exist for investigating these infections (Fan et al., 2022; Kuehling et al., 2022; Wurbs et al., 2024) in other bacteria. Ex vivo models are derived from human cells and preserve tissue architecture, making them more physiologically relevant than invertebrate models. However, they are generally high-cost, with limited throughput and lack both adaptive immunity and systemic infection capacity. A few non-murine models have been developed for Ax, including Caenorhabditis elegans (nematode), Drosophila melanogaster (fruit fly), and the non-mammalian vertebrate Zebrafish (Danio rerio) (Yilmaz et al., 2022; Besse et al., 2025; Aryal et al., 2017). C. elegans provides large sample sizes for study but is extremely difficult to infect via injection. Similarly, Drosophila enables high-throughput experimentation and can provide substantial insights into genetic factors of disease, however the ability of the flies to escape in a BSL-2 environment poses substantial hazards. The zebrafish model allows for the visualization of organs during infection but requires specialized housing and is challenging to infect. In contrast, although lacking adaptive immunity, Galleria has a functional innate immune system, provides a readily manipulable organism that is easily housed, abundant in number for high throughput virulence screening, and poses minimal hazard to laboratory personnel. In fact, Galleria larvae have been widely employed to assess pathogen-mediated killing by numerous human pathogens (Kaito et al., 2020).

While our results strongly support the use of Galleria as an in vivo infection model, this study has limitations. Our work utilized a small number of mice for the exploratory infections with the NIH-018 series. However, this limitation itself underlines the value of Galleria mellonella as a preliminary in vivo model, given the cost and ethical implications of mouse work.

In summary, we have demonstrated that the Galleria infection model enables researchers to obtain important insights into disease severity and interactions between bacteria and the innate immune system. It can be a useful tool for the identification of novel virulence factors in Ax clinical isolates. In addition, the Galleria model may serve as an effective system for the preliminary screening and testing of novel antimicrobials.