RESUMO
BACKGROUND: Experimental animal models are indispensable components of preclinical sepsis research. Reproducible results highly rely on defined and invariant baseline conditions. Our hypothesis was that the murine gut microbiota varies among different distributors of laboratory animals and that these variations influence the phenotype of abdominal sepsis derived from a bacterial inoculum model (intraperitoneal stool injection). MATERIALS AND METHODS: Male C57BL/6 mice (8-wk old) purchased from Charles River (CR), Janvier (J), and Harlan (H) were sacrificed, and the bacterial composition of feces was analyzed using CHROMagar orientation medium. Stool was injected intraperitoneally into CR mice, followed by clinical observation and gene expression analysis. Experiments were repeated 16 mo later under the same conditions. RESULTS: Stool analysis revealed profound intervendor differences in bacterial composition, mainly regarding Staphylococcus aureus and Bacillus licheniformis. Mice challenged with CR as well as H feces developed significantly higher severity of disease and died within the observation period, whereas stool from J mice did not induce any of these symptoms. Real-time polymerase chain reaction revealed corresponding results with significant upregulation of proinflammatory cytokines and vascular leakage-related mediators in CR and H injected animals. Sixteen months later, the bacterial fecal composition had significantly shifted. The differences in clinical phenotype of sepsis after intraperitoneal stool injection had vanished. CONCLUSIONS: We are the first to demonstrate vendor and time effects on the murine fecal microbiota influencing sepsis models of intraabdominal stool contamination. The intestinal microbiota must be defined and standardized when designing and interpreting past and future studies using murine abdominal sepsis models.