The use of chicken and insect infection models to assess the virulence of African Salmonella Typhimurium ST313

Over recent decades, Salmonella infection research has predominantly relied on murine infection models. However, in many cases the infection phenotypes of Salmonella pathovars in mice do not recapitulate human disease. For example, Salmonella Typhimurium ST313 is associated with enhanced invasive infection of immunocompromised people in Africa, but infection of mice and other animal models with ST313 have not consistently reproduced this invasive phenotype. The introduction of alternative infection models could help to improve the quality and reproducibility of pathogenesis research by facilitating larger-scale experiments. To investigate the virulence of S. Typhimurium ST313 in comparison with ST19, a combination of avian and insect disease models were used. We performed experimental infections in five lines of inbred and one line of outbred chickens, as well as in the alternative chick embryo and Galleria mellonella wax moth larvae models. This extensive set of experiments identified broadly similar patterns of disease caused by the African and global pathovariants of Salmonella Typhimurium in the chicken, the chicken embryo and insect models. A comprehensive analysis of all the chicken infection experiments revealed that the African ST313 isolate D23580 had a subtle phenotype of reduced levels of organ colonisation in inbred chickens, relative to ST19 strain 4/74. ST313 isolate D23580 also caused reduced mortality in chicken embryos and insect larvae, when compared with ST19 4/74. We conclude that these three infection models do not reproduce the characteristics of the systemic disease caused by S. Typhimurium ST313 in humans. Author summary Salmonella Typhimurium ST313 is associated with systemic infection in human populations in sub-Saharan Africa, and contrasts with the related pathovariant ST19 which causes gastrointestinal disease worldwide. Although the systemic pathology associated with ST313 infection in humans has been comprehensively documented in clinical and epidemiological studies, such pathology has been inconsistently reproduced in animal models of infection. Animal models that reliably recapitulate ST313 infection in humans are needed to study the biological mechanisms underpinning the systemic disease caused by ST313. In this study we performed extensive infection experiments, using several defined and alternative animal infection models to look for robust phenotypes that differentiate infection by S. Typhimurium ST313 from ST19. Large sample sizes and multivariate statistical analysis of infection data for inbred chicken lines allowed us to detect a small but consistent differences between the strains. Overall, S. Typhimurium ST313 was associated with a reduced infection burden and pathology relative to ST19. This subtle phenotype may reflect the limitation of animal models to accurately represent infection by pathogens that have adapted to specific host phenotypes, for example, immunodeficiency in humans. Our study demonstrates the challenge of using animal models to differentiate closely-related bacterial pathovariants, and shows that inter-pathovar differences detected in animal models of infection often do not reflect clinical differences in humans at the level of disease mechanism.