Macrophage and <it>Galleria mellonella </it>infection models reflect the virulence of naturally occurring isolates of <it>B. pseudomallei</it>, <it>B. thailandensis </it>and <it>B. oklahomensis</it>

Abstract

Abstract Background Burkholderia pseudomallei is the causative agent of melioidosis, a tropical disease of humans with a variable and often fatal outcome. In murine models of infection, different strains exhibit varying degrees of virulence. In contrast, two related species, B. thailandensis and B. oklahomensis, are highly attenuated in mice. Our aim was to determine whether virulence in mice is reflected in macrophage or wax moth larvae (Galleria mellonella) infection models. Results B. pseudomallei strains 576 and K96243, which have low median lethal dose (MLD) values in mice, were able to replicate and induce cellular damage in macrophages and caused rapid death of G. mellonella. In contrast, B. pseudomallei strain 708a, which is attenuated in mice, showed reduced replication in macrophages, negligible cellular damage and was avirulent in G. mellonella larvae. B. thailandensis isolates were less virulent than B. pseudomallei in all of the models tested. However, we did record strain dependent differences. B. oklahomensis isolates were the least virulent isolates. They showed minimal ability to replicate in macrophages, were unable to evoke actin-based motility or to form multinucleated giant cells and were markedly attenuated in G. mellonella compared to B. thailandensis. Conclusions We have shown that the alternative infection models tested here, namely macrophages and Galleria mellonella, are able to distinguish between strains of B. pseudomallei, B. thailandensis and B. oklahomensis and that these differences reflect the observed virulence in murine infection models. Our results indicate that B. oklahomensis is the least pathogenic of the species investigated. They also show a correlation between isolates of B. thailandensis associated with human infection and virulence in macrophage and Galleria infection models.