Identification and Characterization of Cri1, a Locus Controlling Mortality During Citrobacter Rodentium Infection in Mice
Total Page:16
File Type:pdf, Size:1020Kb
Genes and Immunity (2011) 12, 280–290 & 2011 Macmillan Publishers Limited All rights reserved 1466-4879/11 www.nature.com/gene ORIGINAL ARTICLE Identification and characterization of Cri1, a locus controlling mortality during Citrobacter rodentium infection in mice E Diez1,2, L Zhu1,2, SA Teatero1,2, M Paquet3, M-F Roy4, JC Loredo-Osti5, D Malo2,6 and S Gruenheid1,2 1Department of Microbiology and Immunology, McGill University, Montreal, Quebec, Canada; 2Complex Traits Program, McGill University, Montreal, Quebec, Canada; 3Veterinary Comparative Pathology Services, Comparative Medicine and Animal Resources Centre, McGill University, Montreal, Quebec, Canada; 4Department of Veterinary Clinical and Diagnostic Sciences, Faculty of Veterinary Medicine, University of Calgary, Calgary, Alberta, Canada; 5Department of Mathematics and Statistics, Memorial University, St John’s, Newfoundland and Labrador, Canada and 6Departments of Medicine and Human Genetics, McGill University, Montreal, Quebec, Canada Infection of inbred mouse strains with Citrobacter rodentium represents an ideal model to reveal the genetic factors controlling host resistance to noninvasive enteric bacterial pathogens. We have chosen a positional cloning approach to identify putative gene(s) that control the known difference in survival between resistant C57BL/6J and susceptible C3H/HeJ and C3H/HeOuJ mice. Our work has identified one major locus within proximal chromosome 15 that is responsible for the marked susceptibility of both C3H strains, and we formally exclude Tlr4 from control of survival to this pathogen. We have named this new host resistance locus Cri1 (Citrobacter rodentium infection 1). The Cri1 genetic interval currently spans B16 Mb and it confers survival to the infection in a recessive manner. Transfer of the Cri1 locus from the surviving B6 mice into a congenic mouse with a C3Ou genetic background confirms its overall chromosomal localization and its highly significant effect on host survival. The C3Ou.B6-Cri1 mice thus produced have also enabled us to dissociate the control of mouse survival from the control of bacterial load early in the infection as well as from control of colonic hyperplasia. Genes and Immunity (2011) 12, 280–290; doi:10.1038/gene.2010.76; published online 17 February 2011 Keywords: QTL; Citrobacter rodentium; mouse; survival; Cri1 Introduction increase.10 These observations are all the more worri- some as there are currently no specific therapeutic Citrobacter rodentium infection of mice was first described regimens for EHEC infections; antibiotic use is not in the 1960s and 1970s as the suspected causative agent recommended as it exacerbates disease and increases of outbreaks of diarrhea, colitis and colonic hyperplasia the rate of development of complications.11 in mouse colonies in the United States and Japan.1–5 More The parallels between C. rodentium infection of mice recently, infection of mice with this intestinal pathogen and EHEC and EPEC infection of humans are striking. has become recognized as an excellent model for These Enterobacteriaceae colonize the gastrointestinal infection with the human-specific diarrheal pathogens tract of their hosts by the formation of characteristic enterohemorrhagic and enteropathogenic Escherichia coli attaching/effacing lesions, leading to the development of (EHEC and EPEC; reviewed in Mundy et al.6). EPEC kills diarrhea, colitis and hyperplasia of the intestinal epithe- several hundred thousand children each year in devel- lium. Using a type III secretion system that is conserved oping countries,7 whereas EHEC is commonly associated between all three bacterial species, they inject bacterial with foodborne diarrheal outbreaks in the developed virulence proteins directly into host cells. Recent geno- world.8 Over 110 000 cases of EHEC infection and 61 mic and proteomic characterization of the secreted EHEC-associated deaths are estimated to occur each year protein repertoires in EHEC, EPEC and C. rodentium in the United States alone.9 Furthermore, recent epide- indicates that the vast majority of injected virulence miological data indicate that both the frequency of EHEC proteins are common to all three bacterial species.12,13 outbreaks and the severity of disease suffered are on the Consequently, it is not surprising that the C. rodentium model is now widely used to investigate the role of bacterial virulence proteins and host defense mechan- Correspondence: Dr S Gruenheid, Department of Microbiology and isms in the attaching/effacing pathogenesis. Immunology, McGill University, 3649 Promenade Sir William Osler, Studies in humans indicate that the disease severity Room 365, Montreal, Quebec, Canada H3G 0B1. E-mail: [email protected] and clinical outcome of EHEC and EPEC infection can Received 16 March 2010; revised 14 July 2010; accepted 24 August vary greatly between individuals.14–17 Whether these 2010; published online 17 February 2011 differences have a genetic basis is undetermined. On the Cri1 and susceptibility to C. rodentium infection E Diez et al 281 other hand, differences in susceptibility to C. rodentium 100 infection in mice were first described in 1977, when an abnormally high mortality rate was noted in the 80 C3H/HeJ strain.18 Later studies revealed that C3H/HeJ mice are highly susceptible to Salmonella infection and are hyporesponsive to bacterial lipopolysaccharide be- 60 cause of a mutation in Toll-like receptor 4 (Tlr4), raising the question of whether the Tlr4 defect was also responsible 40 for C. rodentium susceptibility. However, the finding Percent survival that the closely related but Tlr4-wild-type mouse strain B6 C3H/HeOuJ is also susceptible to C. rodentium infection 20 C3Ou suggested that C. rodentium susceptibility was indepen- C3 19 dent of Tlr4. Additional support for this was provided 0 by side-by-side phenotypic studies of Tlr4-deficient and 0 5 10 15 20 25 30 20 Tlr4-wild-type mice. In this study we apply a systema- Days post infection tic genetic approach to the C. rodentium infection model to investigate the genetic basis for high mortality in B6 10 C3H/HeJ and C3H/HeOuJ mice. C3Ou C3 8 Results 6 Phenotypic differences between C57BL/6J and C3H inbred mice during C. rodentium infection 4 The inbred mouse strains C3H and C57BL/6J (B6) differ in their susceptibility to C. rodentium infection.18,19 The 2 most striking difference being that mice of the C3H/HeJ (C3) and C3H/HeOuJ (C3Ou) strains succumb to the infection within 12 days following orogastric infection, Bacterial shedding (log cfu / g of feces) 0 0 5 10 15 20 whereas B6 mice not only survive but fully recover and clear the infection (cumulative survival curves, Days post infection Figure 1a). Besides the obvious difference in survival Figure 1 Phenotyping of inbred mouse strains following Citrobac- between the B6 and both C3H strains, the C3Ou strain ter rodentium infection. (a) Cumulative survival in the resistant displays a reproducible and statistically significant C57BL/6J and the susceptible C3H/HeJ (C3) and C3H/HeOuJ (C3Ou) inbred mouse strains following gastric gavage with 2.5 Â 108 increase in survival time when compared with the C3 CFUs of C. rodentium. A total of 10 animals from each strain (5 males strain (Po0.0001 with a log-rank statistical test). Com- and 5 females) were infected. No significant differences in survival pared with B6 mice, C3 and C3Ou mice also display a were noted between the sexes. Besides the obvious difference in much more rapid rate of colonization of the colon by survival between the C57BL/6J and both C3H strains, the C3Ou C. rodentium, leading to significantly greater bacterial strain displays a reproducible and statistically significant increase in loads early on in infection.19 Figure 1b shows a time survival time when compared with the C3 strain (Po0.0001 log- rank statistics). (b) Bacterial shedding in fecal pellets from infected course of C. rodentium shedding in the surviving B6 inbred mice. Bacterial shedding is measured as C. rodentium CFU g–1 mouse strain and in the susceptible C3 and C3Ou strains. of homogenized fecal matter. Five mice from each inbred strain Two important observations can be made from the data were infected and transferred into bedding-free isolation cages at obtained. First, the colony-forming unit (CFU) counts the specified time points to collect six fresh fecal pellets from each. increase significantly faster in both C3H strains com- Bacterial shedding was measured every 24 h after infection. Data pared with B6 mice (Po0.05 at days 3–5 post infection), collection was stopped for the C3H strains at day 7, when signs of 19 morbidity/mortality appeared. For the B6 mice, CFU counts were and that is consistent with previously published data. further compiled at days 10, 14 and 21 post infection to show Second, B6 mice reach similar high levels of C. rodentium bacterial clearance, as described previously.33 Mean log-transformed shedding in fecal pellets before they start clearing the bacterial counts are shown in the graph and error bars depict s.d. infection after day 10 post infection. Although it has been Nonlinear curve fitting was performed using fourth-order hypothesized that the greater bacterial loads early in polynomial functions. infection may lead to mortality in C3H strains,19 the link between these two phenotypic differences remains an open question. pellets. Colonic hyperplasia and hemorrhage are evident In order to gain more insight into the phenotypes as well, even at the macroscopic level. In contrast, B6 under study, the parental mouse strains were examined tissues display low levels of hyperplasia and only in detail at day 7 post infection. This is the time point slightly increased water content in their fecal pellets. closest to the peak of infection in C57BL/6J mice where Although B6 mice are undeniably sick at that point, it is there are still some surviving C3 mice (Figure 1a). Mice nowhere near the phenotype observed in C3H mice. We from each of the infected inbred strains were killed and hypothesized that the observed phenotypic differences their cecum and colon dissected.