Comparison of Galleria mellonella , Epithelial Cell Cytotoxicity, and Mouse Model of Bacteremia to Measure Pseudomonas aeruginosa Virulence

Abstract Considerable effort has focused on identifying alternatives to mouse models in research studies. In the field of bacterial pathogenesis, Galleria mellonella and epithelial cell lines have been widely used for this purpose, but the concordance of these models with mice remains unclear. To begin to address this knowledge gap, we used 105 clinical isolates of Pseudomonas aeruginosa for which the virulence had been previously determined in a mouse bacteremia model. A semistrong correlation was observed between G. mellonella median time to 50% mortality and mouse 50% pre-lethal dose (LD50) values (Spearman’s rank correlation coefficient [ρ] = 0.75), whereas percent A549 epithelial-like cell lysis during co-culture showed a weak correlation to mouse LD50 values (ρ = -0.47). Given the stronger correlation between G. mellonella and mouse virulence, we next examined whether G. mellonella could substitute for mice when asking questions about the virulence of large numbers of P. aeruginosa isolates. Results from mice indicated that isolates with resistance to more antibiotics were significantly less virulent, and the use of G. mellonella identified the same inverse correlation. Furthermore, both models found no evidence for the existence of hypervirulent clonal lineages. In particular, isolates belonging to sequence types defined as high-risk clones were not consistently more virulent than other isolates, despite the known association of high-risk clones with poor clinical outcomes. These findings suggest that G. mellonella can serve as an adequate substitute for mice when addressing specific population-based virulence questions, although conclusions should be confirmed in mice. Author Summary We found that virulence measurements in a G. mellonella infection model showed a semistrong correlation with those from a mouse bacteremia model and that this insect larval model adequately detected population-level trends similarly to mice. In contrast, A549 epithelial-like cell lysis during bacterial co-culture correlated less well with mouse virulence. Together, these results support the use of G. mellonella as a scalable, low-cost, and humane first-line model for assessing P. aeruginosa virulence but also indicate that conclusions should be validated in mice. Competing Interest Statement The authors have declared no competing interest.