ELISA experiments showed that TNF-α was not secreted by mock-infected cells or HB101-treated cells (Fig. 7E). These results were expected, because tnf-α mRNA was not detected by RT-PCR. However, E2348/69 infection activated TNF-α secretion at a high value (252 ± 8 ng/ml) at 2 h of infection, which decreased at 4 h post-infection (151 ± 13 ng/ml). In E22-infected cells, there was no decrease and TNF-α secretion was similar at 2 h (252 ± 8 ng/ml) and 4 h (247 ± 13 ng/ml) post-infection (Fig. 7E).

Therefore, as with E2348/69, E22 infection activates TNF-α synthesis and secretion. E22Δeae infection caused IWR-1 molecular weight contrary effects on TNF-α secretion depending on the infection time. At 2 h of E22Δeae infection, TNF-α secretion was of 282 ± 8 ng/ml, while at 4 h of infection, cells secreted 50% less TNF-α (126 ± 13 ng/ml) than cells infected with E22 WT (247 ± 13 ng/ml). TNF-α secretion in cells infected

with E22ΔescN ABT263 was not reduced (236 ± 8 ng/ml) at 2 h and decreased at 4 h (192 ± 13 ng/ml), whereas the secretion at 2 h in cells infected with E22ΔespA was 191 ± 8 ng/ml and at 4 h of 116 ± 13 ng/ml. Thus, T3SS is involved in the activation of TNF-α release. E22ΔfliC infection caused a reduced secretion of TNF-α at 2 h (201 ± 8 ng/ml), and at 4 h TNF-α was completely absent from the supernatants (Fig. 7F). Evidently, flagellin is a factor which is necessary to activate TNF-α secretion, and it is essential to maintain this cytokine in the supernatants of infected cells (strikingly similar is the effect of flagellin in IL-8 release). Inflammation induced by EPEC results from

the balance of positive and negative factors [39]. Here, we analysed the role of the EPEC virulence factors T3SS, EspA, intimin and flagellin, on the epithelial inflammatory response. Sirolimus chemical structure The evaluation comprised TLR5 signalling activated by EPEC flagellin [25], activation of ERK1/2 [28] and NF-κB [27] pathways and transcription of proinflammatory cytokine genes [33, 39]. EPEC-induced cell signalling was reproduced and unified in an in vitro epithelial cell infection model, which consisted in using HT-29 cells infected with the prototype strain E2348/69 and the strain E22, which is a strain pathogenic for rabbits, which contains LEE but no BFP [40], and can be considered an atypical EPEC. The role of the virulence factors was studied using isogenic E22 mutants, to be able to corroborate the results in vivo through the experimental rabbit infection model [33]. The significance of EPEC flagellin in the activation of proinflammatory response is well established [25]. However, TLR5 expression, localization and functionality in intestinal epithelial cells have all been unclear [38], and previous studies have been focused on TLR5 distribution in polarized cells [41, 42].