Absence of TLR11 in Mice Does Not Confer Susceptibility to Salmonella Typhi

Salmonella enterica serovar Typhi (S. Typhi) causes typhoid fever, a systemic disease of humans that is estimated to cause more than 200,000 annual deaths (Butler, 2011; Crump and Mintz, 2010 ; Parry et al., 2002). Unlike other Salmonella enterica serovars, which can infect a broad range of animals, S. Typhi can only infect humans, which has hampered the development of a convenient animal model for the study of typhoid fever. Recently, it was reported in Cell (Mathur et al., 2012) that mice lacking Toll-like receptor 11 (TLR11) could be lethally infected with S. Typhi after oral or systemic inoculation. It was also postulated that TLR11-mediated recognition of Salmonella flagellin prevents S. Typhi infection in wild-type (C57BL/6) mice and that the lack of functional TLR11 renders humans susceptible to the bacterial infection (Mathur et al., 2012). It was therefore proposed that TLR11-deficient mice could serve as a convenient animal model for typhoid fever (Mathur et al., 2012). However, we report here that infection studies conducted in four different laboratories have found that TLR11-deficient mice do not show enhanced susceptibility to S. Typhi regardless the route of inoculation. We also observed no binding of flagellin to TLR11 and found no differences in the response of wild-type and TLR11-deficient mice to the administration of bacterial flagellin, which are inconsistent with the proposed role of this Toll receptor in the detection of this bacterial protein (Mathur et al., 2012). We infected TLR11 / and wild-type isogenic control mice either orally or intraperitoneally with different strains of S. Typhi. In addition to the strain used in the original study (Ty2) (Mathur et al., 2012), we used other virulent S. Typhi strains, including ISP2825 (Galán andCurtiss, 1991), and A8/14353363984-6 and 3290A481, two clinical isolates obtained from theYale ClinicalMicrobiology laboratory. Mice were obtained from Dr. Sankar Ghosh, and their genotype was verified by standard genotyping techniques (see Supplemental Experimental Procedures). None of the mice showed any signs of disease, even when inoculated with doses up to 100-fold higher than those used by Mathur et al. (2012) (Figure 1 and Supplemental Information). Consistent with this finding, no S. Typhi colony forming units (CFU) were recovered from the spleens of animals (WT or TLR11 / ) orally or intraperitoneally inoculated with S. Typhi (Figures 1A–1G). Mice orally or intravenously inoculated with S. Typhi Ty2 showed no differences in CFU detected in fresh fecal pellets (Figure 1H), spleens (Figure 1I), or livers (Figure 1J). Survival studies independently conducted at four different geographical sites demonstrated that mice defective in TLR11 receptor do not show enhanced susceptibility to S. Typhi. Taken together, these results indicate that TLR11 / mice cannot serve as a model for the study of typhoid fever and S. Typhi pathogenesis as originally proposed (Mathur et al., 2012). Unlike the previous report (Mathur et al., 2012), we found no evidence of flagellin binding to TLR11. We injected WT and TLR11 / micewith recombinantSalmonella flagellin and measured serum levels of IL-12 and IL-6 2 hr post-injection. We found no differences in the levels of these cytokines in WT and TLR11 / mice (Figure S1). Similarly, we found no differences in the levels of IL-12, IL-6, and TNF produced by sort-purified WT and TLR11 / lamina propria mouse macrophages or splenic dendritic cells in response to flagellin (Figure S1). These results are inconsistent with the proposed recognitions of flagellin by TLR11 (Mathur et al., 2012). We cannot offer an explanation for the variance between our collective studies and those reported by Mathur et al. (2012). The differences in the results obtained are drastic in that TLR11 / mice not only did not succumb to S. Typhi after oral or systemic infection, but the animals cleared the infection. The strain of S. Typhi used in this study was the same as that used in the study of Mathur et al. (2012), and the same observations were