Flagella are important virulence factors for bacteria, facilitating motility and the ability to adhere to and invade eukaryotic cells; they are possessed by a number of important human pathogens including Clostridium difficile, Escherichia coli, Shigella flexneri and Helicobacter pylori.
There are two defined cellular mechanisms for detection of flagellin by mammalian cells, which contribute to host defence against pathogens. Extracellular flagellin interacts with plasma membrane-localised Toll-Like Receptor 5 (TLR5), whereas internalised flagellin (the result of insertion by bacterial type 3 or type 4 secretion systems) can interact with the NAIP4/NCLR5 inflammasome.
In 2007, Vijay-Kumar and colleagues produced a TLR5KO mouse and showed that a quarter to a third of these mice demonstrated colitis by 8-12 weeks of age, which was accompanied by increased faecal bacterial content and a greater number of adherent bacteria associated with the colonic mucosa . Interestingly, the remaining mice that did not develop overt colitis went on to demonstrate typical features of metabolic syndrome including obesity, increased food intake, insulin resistance and elevated blood glucose .
The same research group went on to demonstrate the intrinsic instability in the microbiome of TLR5KO mice, with colitic mice demonstrating an increase in Proteobacteria phylum, particularly Enterobacteria such as E. coli . They also showed that the pathological changes observed in TLR5KO mice were driven by the microbiota in these mice, by establishment of a germ-free colony, which failed to show any of the features of the inflammatory phenotype demonstrated by this strain maintained using standard husbandry.
Subsequently, mouse strains with either intestinal epithelial cell-specific (TLR5ΔIEC) or dendritic cell-specific (TLRΔDC) deletion of TLR5 were generated . The TLR5ΔIEC strain demonstrated a similar phenotype to the original TLR5KO strain, with modest inflammatory changes (increased large bowel weight, inflammatory cell infiltrate) in the large bowel and metabolic syndrome; these changes were not evident in the TLRΔDC strain. Colonic inflammation in the TLR5ΔIEC mice was associated with increased faecal bacterial load, higher levels of faecal flagellin and LPS and greater number of adherent bacteria at the apical surface of the mucosal epithelium. TLR5ΔIEC mice were also more susceptible to induction of colitis, either by DSS or by deletion of IL‑10. These data support the role of TLR5 expression by intestinal epithelial cells in influencing the composition of the microbiota and thus controlling intestinal inflammation.
A key insight into the modulation of the microbiome induced by TLR5 insufficiency came from studies in which germ-free wild type or TLR5KO mice were infected with adherent-invasive E. coli. (AIEC) . Under germ-free conditions, TLR5KO mice showed neither colitis nor metabolic syndrome phenotypes. Following AEIC infection, wild type mice clear the majority of bacteria after six weeks (based on CFUs per gram of stool), and demonstrated no intestinal inflammation. TRL5KO mice showed similar levels of bacteria by six weeks (post-infection), however, the bacterial load in these mice was significantly elevated above control levels during the first five weeks post-infection. Although there was eventual equivalence in bacterial load by 120 days post-infection, TLR5KO mice demonstrated a failure to thrive relative to wild type mice, together with histopathological changes and increased inflammatory markers indicative of chronic colitis. If, after infection, AIEC-infected wild type and TLR5KO mice were maintained under standard SPF conditions, the AIEC are cleared more rapidly – within two weeks – by both strains . This was a consequence of colonisation by (and competition from) bacterial species endogenous to the vivarium SPF environment. Despite the rapid clearance of AIEC (under SPF conditions), TLR5KO mice still demonstrated a sustained elevation of inflammatory markers at 90 days post-infection, which was associated with changed species abundance and decreased species diversity within the microbiome.
The findings from these studies lead to the conclusion that there is a short window after birth that is critical for defining the gut microbiome and, as a consequence, the long-term health of the animal. This hypothesis has been investigated in a new publication from Fulde and colleagues . Analysis of Tlr5 expression by intestinal epithelial cells from neonatal mice revealed that levels were more than ten times greater during the first two weeks of life, than in older mice. In contrast, Tlr5 was not expressed at elevated levels during the early neonatal period by splenocytes, and expression showed no age-dependency; also, genes coding for NAIP4/NCLR5 inflammasome components did not show any age-related regulation. Given the short window of elevated Tlr5 expression by intestinal epithelial cells after birth, the authors examined whether a difference between the microbiome of wild type and TLR5KO mice could be observed during this time. For this experiment, wild type and TLR5KO pregnant dams were co-housed up to day E18, at which point they were caged separately. Bacterial samples from the small intestine of the pups were obtained and principal component analysis was performed on 16S rDNA sequence data, which demonstrated that a significant difference in the microbiome of the two genotypes could be observed just 8 days after birth. In contrast, no significant difference was seen between the microbiomes of co-housed litter mates of different genotypes (resulting from breeding of heterozygous Tlr5+/- pairs).
When either wild type- or TLR5KO-derived microbiota were transferred to neonatal, germ-free wild type recipients (via their dam at day E18), principal component analysis of 16s rDNA sequence data showed change and convergence of the two microbiomes over a four week period following birth. Repeating the same experiment over a longer time period, demonstrated that the similarity of the wild type- and TLR5KO-derived microbiomes originally transferred to the neonatal wild type animals was maintained for up to 21 weeks. The convergence was associated with a similarity in the level of flagellated bacteria between the two groups of mice. The clearance of flagellated bacteria following transfer of the TLR5KO-derived microbiota, was in part due to TLR5 signalling, via MyD88, upregulating the expression of the antimicrobial peptide Reg3γ. In contrast, when neonatal TLR5KO mice were the recipients for the microbiota transfer, there was no significant convergence of wild type- and TLR5KO-derived microbiomes in the mice.
When adult, germ-free, wild type recipient mice were used, although both the wild type- and TLR5KO-derived flora demonstrated change over the same period of time, they did not converge. The non-convergence of the two microbiomes was confirmed up to 12 weeks post-bacterial transfer (18 weeks of age), with higher levels of flagellated bacteria being present in the mice that received the TLR5KO-derived microbiota. In adult, germ-free TLR5KO recipients, however, there was evidence of some convergence of the different microbiomes from four weeks post-bacterial transfer.
This work has shown that TLR5 expression by the intestinal epithelium early in neonatal development is an important factor in preventing increased colonisation by flagellated bacteria, which can drive chronic pathologies such as metabolic syndrome and colitis, even if levels of the trigger organism(s) are subsequently “normalised”. The antimicrobial peptide, Reg3γ, was identified as one factor that played a role in controlling the level of flagellates in the neonatal microbiome. It was noted, however, that exposure to the microbiota of litter mates of different genotypes is also an important factor in shaping the development of the microbiome in neonatal animals.
Adult mice, in which levels of TLR5 expression are much lower, were found to be much less able (than neonates) in clearing flagellated bacteria when colonised by a TLR5KO-derived microbiome. But the authors suggested that reduced TLR5 expression in the adult intestine may be a good thing by limiting inflammatory responses; this observation was linked to the reported effects of Tlr5 polymorphisms in humans, where gain-of-function mutation is positively associated with Crohn’s disease and loss-of-function mutation shows negative association.
 Vijay-Kumar M, Sanders CJ, Taylor RT, Kumar A, Aitken JD, Sitaraman SV, Neish AS, Uematsu S, Akira S, Williams IR, Gewirtz AT. Deletion of TLR5 results in spontaneous colitis in mice. J Clin Invest. 117, 3909-3921: 2007. DOI: 10.1172/JCI33084
 Vijay-Kumar M, Aitken JD, Carvalho FA, Cullender TC, Mwangi S, Srinivasan S, Sitaraman SV, Knight R, Ley RE, and Gewirtz AT. Metabolic Syndrome and Altered Gut Microbiota in Mice Lacking Toll-Like Receptor 5. Science. 328, 228–231; 2010. DOI: 10.1126/science.1179721
 Carvalho FA, Koren O, Goodrich JK, Johansson ME, Nalbantoglu I, Aitken JD, Su Y, Chassaing B, Walters WA, González A, Clemente JC, Cullender TC, Barnich N, Darfeuille-Michaud A, Vijay-Kumar M, Knight R, Ley RE and Gewirtz AT. Transient inability to manage proteobacteria promotes chronic gut inflammation in TLR5-deficient mice. Cell Host Microbe. 12, 139-152: 2012. DOI: 10.1016/j.chom.2012.07.004
 Chassaing B, Ley RE, Gewirtz AT. Intestinal epithelial cell toll-like receptor 5 regulates the intestinal microbiota to prevent low-grade inflammation and metabolic syndrome in mice. Gastroenterology. 147, 1363-1377; 2014. DOI: 10.1053/j.gastro.2014.08.033
 Chassaing B, Koren O, Carvalho FA, Ley RE and Gewirtz AT. AIEC pathobiont instigates chronic colitis in susceptible hosts by altering microbiota composition. Gut. 63, 1069-1080: 2014. DOI: 10.1136/gutjnl-2013-304909
 Fulde M, Sommer F, Chassaing B, van Vorst K, Dupont A, Hensel M, Basic M, Klopfleisch R, Rosenstiel P, Bleich A, Bäckhed F, Gewirtz AT, Hornef MW. Neonatal selection by Toll-like receptor 5 influences long-term gut microbiota composition. Nature. 560, 489-493: 2018. DOI: 10.1038/s41586-018-0395-5