Plague is a disease with devastating effects on the host that are fatal if left untreated. These effects are the result of the ability that Y. pestis
displays to suppress host immune responses and to promote systemic dissemination at remarkably high rates. Numerous studies have described many virulence factors that are essential to suppress host immune responses [2
]. The direct contributions of these virulence factors to bacterial dissemination, however, are still unclear. The study of dissemination per se is a field that is lagging behind in plague research. BLI is a tool that allows for the visualization of a pathogen in a host during infection and a very promising alternative to better understand Y. pestis
dissemination. A recent report described the use of BLI in a subcutaneous (SC) model of bubonic plague [25
]. In this report, the pGEN-luxCDABE
plasmid was described to have no effect on the virulence of Y. pestis
and to be suitable for BLI as luminosity correlated with bacterial counts in vivo; our results confirmed and expanded upon these findings. Our goal was to determine whether BLI could be used to follow dissemination and colonization of Y. pestis
in mice after using different routes of inoculation that closely mimic bubonic and pneumonic plague. Moreover, we tested whether BLI could be used to detect mutants with defects in colonization or dissemination.
After inoculation with a strain of Y. pestis
that contains pGEN-luxCDABE
, we showed that animals can be imaged through the course of infection in such a way that bacterial spread could be followed over time for three different models of infection. Our results from the SC inoculation model support the previous notion that, during bubonic plague, Y. pestis
travels from the site of inoculation to the proximal lymph node prior to dissemination to deeper tissues [16
]. We observed that bacteria were maintained at the site of inoculation during the course of infection, as previously reported for ear intradermal (ID) infections [15
]. For both, the SC and ID models, the bacterial population at the site of inoculation appeared not only to be maintained, but also to expand. However, while we quantified signal from the site of infection in the SC-inoculated animals, we cannot conclude such signal comes from the skin alone. In our SC model, the patch of inoculated skin is located in an anatomical position on top of the superficial cervical LNs and thus, both, skin and LNs, are imaged as a single source of radiance. We could determine that signal was coming partly from the site of inoculation after removing the patch of skin and imaging it individually. This complication is minimized in the ID model, where the site of inoculation (ear pinna) is distant from the draining LN (superficial parotid LN). While an increase overtime in signal intensity from the ear was observed, we were not able to quantify the signal, as it was difficult to place the ears of all mice at the same position inside of the animal isolation chamber.
Images taken during the first hours following intranasal (IN) infections suggested that, in isolated cases, at least part of the inoculum can go to the stomach. The IN route requires delivering small drops of inoculum into one of the nostrils (total volume of 20 μL), and some of this inoculum could be swallowed rather than inhaled. Signal from the stomach never seemed to last beyond the 6 hpi time point, suggesting that gastric infections with Y. pestis
in these mice are cleared quickly. We also observed that the feces of half of the mice produced detectible signal, indicating that Y. pestis
was being shed. This was only observed at very early time points (6 hpi), indicating that bacteria were fully shed from the gastrointestinal tract by 24 hpi. In humans, it has been shown that transmission can occur after ingestion of contaminated food [32
]. While mice are coprophagous, it is not know whether a fecal-oral route could be a mechanism for Y. pestis
to disperse or infect other individuals. Detecting signal from the tip of the nose also opens the question whether bacteria could be transmitted to other individuals with whom food and water are shared. We do not know whether signal from the stomach or the tip of the nose would still be present after an aerosol infection, a route that pneumonic plague is assumed to be transmitted in nature. All mice, independent of the presence of signal from the stomach or feces, showed the same progression of infection with comparable levels of signal from the thorax. More importantly, all animals showed signs of disease and mortality at very similar times. This observation suggests that the fraction of the inoculum that may go to the gastrointestinal tract has no effect on the overall pneumonic infection.
The low number of mice used during BLI is one of its more important advantages. However, it can also be a disadvantage because of the variability in bacterial load for a specific organ from animal to animal and sudden death, both inherent aspects of plague infections. The differences in the levels of significance from time point to time point when comparing radiance values between the wild type and double mutant infected animals are due to this high variability of bacterial load and death. Despite these challenges, we found that BLI is a suitable method for studying dissemination/colonization of Y. pestis in three separate models of plague, and that significant differences in radiance could be detected between wild type and a mutant of modest attenuation using relatively few mice.