It was generally accepted that TLRs were not expressed or expressed at low levels on the AP surface of IEC in order to avoid constitutive intestinal inflammation caused by commensal flora and their products. Based on a study of polarized T-84 IEC it was determined that the flagellin receptor, TLR5, was exclusively location on the BL surface of IEC 
allowing only Gram-negative pathogens capable of translocating through IEC access to the BL membrane. However, the location of TLR5 on polarized IEC and ability of luminal flagellin to induce innate immune responses is controversial as other investigators have detected TLR5 expression on their luminal surface 
and shown that flagellin interacts with the AP surface of cultured human IEC to induce pro-inflammatory responses 
. Our current study concurs with these findings as we report herein that the AP application of flagellin from virulent or avirulent bacteria as well as the flagellin protein constructs of TLR5-binding regions alone stimulated healthy polarized Caco-2BBe and T-84 cells to induce IL-8 secretion and migration of neutrophils and DCs. However, we extend these findings and are the first to show that flagellin is internalized by IEC as directed by apically present TLR5 followed by the co-localization of flagellin with endosomal and lysosomal compartments and subsequently found not to translocate to the BL surface of IEC.
Using a transwell filter system to establish polarized IEC allowed us to expose the AP or BL surface of cells to full-length flagellin from intestinal pathogens or commensal bacteria. Our findings showed that IL-8 was secreted from both the AP and BL surfaces of T-84 and Caco-2BBe cells regardless of which surface was stimulated. However, significantly higher IL-8 secretion was detected in the BL supernatant of both IEC stimulated with full-length flagellins from virulent or avirulent bacteria. These results correlate with the migration of DC and PMN across the IEC barrier. Interestingly, polarized IEC also secreted IL-8 subsequently resulting in the induction of immune cell migration following AP exposure to truncated flagellin protein (ND1/2ECHCD2/1), which contains only the amino and carboxyl conserved flagellin regions. But the hypervariable D3 flagellin region stimulated only minimal to no IL-8 from both cell lines. These findings suggest that flagellin is binding to TLR5, as the conserved regions of this protein are the only domains recognized by TLR5 resulting in the stimulation of host inflammatory responses 
. In light of these findings, it is not surprising that flagellin from avirulent E. coli
strain K12 and flagellin from an E. coli
strain isolated from a Crohn’s disease lesion induced IL-8 secretion and immune cell migration when exposed to the AP surface of IEC. Previous studies have shown that flagellin from a commensal E. coli
strain induced NF-κB activation via TLR5 binding and the recruitment adaptor protein myeloid differentiation factor-88 (MyD88) in vitro and ex vivo 
To determine the role of TLR5 in IEC uptake of flagellin, we suppressed expression of TLR5 by siRNA or blocked TLR5 with an antibody on Caco-2BBe cells then exposed the AP surface to purified flagellin. Both approaches resulted in the detection of AP flagellin but not BL flagellin. Moreover, exposing the siTLR5 cells to purified flagellin significantly decreased IL-8 secretion regardless of AP or BL applied flagellin suggesting that TLR5 is expressed on both surfaces of IEC. These results are in line with other reports that directly detected AP expression of TLR5 on cultured IEC 
and on human intestinal biopsies 
. Under certain circumstances live, intact bacteria can influence TLR expression on host cells 
. However, we did not find that the virulent intestinal pathogen, S. Typhimurium
and/or an avirulent bacteria, E. coli
K12, altered TLR5 expression on IEC.
Our findings, as well as others 
, suggest that the intestinal tract is arranged in such a manner that makes it difficult for resident microflora and their products to contact the AP surface of IEC under steady-state conditions and also has mechanisms in place so that the normal gut does not mount a full-scale immune assault against residential bacteria and their products 
. Two key factors that limit exposure of the IEC AP surface to commensal bacteria and participate in maintaining intestinal homeostasis is IgA and mucous 
. Our recent data indicated that intestinal IgA blocked host systemic immune responses to commensal bacteria 
. However, pathogenic organisms of the gut have evolved mechanisms (e.g. type III secretion system, exotoxins, endotoxins, use of host proteins as receptors, bacterial effector proteins that manipulate the host cell) that allow them to circumvent host defenses, move through the mucus layer and gain access to IEC 
. By breaching these initial defenses, pathogens and their products, such as flagellin, gain access to the AP surface of IEC where specific TLRs provide IEC with the opportunity to detect and respond to pathogens in a timely manner 
. Sierro et al
showed that Salmonella
flagellin apically stimulated secretion of CCL20 from IEC resulting in the subsequent migration of DCs. Substantial damage to the luminal surface inflicted by the invading pathogen may also provide an opportunity for commensal flora to access the AP IEC surface 
. Our studies demonstrate that AP applied flagellin from the commensal E. coli
K12 strain stimulated IEC leading to IL-8 secretion and immune cell migration. In line with these results, it has been demonstrated that Salmonella 
and Enteroaggregative E. coli 
flagellin as well as flagellin from commensal E. coli 
is capable of inducing inflammatory responses in human IEC.
The intestine has been shown to internalize bacterial products, such as LPS 
and peptidoglycan 
. Majority of apically applied purified LPS 
and/or LPS naturally released from bacteria 
was internalized by IEC in vitro and in vivo. Our findings showed that a predominant antigen from commensal flora that is found in experimental colitis models known as CBir1 flagellin 
was internalized by IEC, which could represent potential implications for the contribution of this specific flagellin to Crohn’s disease. Likewise, Caco-2BBe cells took up Salmonella
flagellin, which was found to co-localized with an early endosomal surface marker and within 1 h flagellin co-localized with a lysosomal membrane surface marker as analyzed by confocal microscopy and cell fractionation. Similarly, it was previously reported that the trafficking of LPS revealed that, following IEC internalization, the highest percentage of LPS co-localized with late endosomes and lysosomes 
. Our data showed that in regards to cell fractionation of flagellin-exposed cells, the long residence time in the recycling endosome was unexpected. It should be noted that transcytotic traffic was not followed beyond the recycling endosome but the recycling endosome was shared by traffic from both AP and BL surfaces. Therefore, it is possible that residence in the recycling endosome does not preclude traffic across the cell (in either direction).
The biological relevance and significance of our findings demonstrates that luminal recognition of flagellin serves multiple purposes. First, we showed herein that in normal human intestinal epithelial cells, luminal application of flagellin induced innate immune responses that did not disrupt the integrity of the intestinal barrier. Subsequently, IEC internalized flagellin, in a TLR5-dependent manner, which co-localized to early endosomal and lysosomal compartments where, without the addition virulence factors from the live pathogen, it was likely degraded, as it was undetectable on the basolateral side of IEC (i.e. not reaching the immune cell rich lamina propria). These data expand our knowledge of how the normal intestinal epithelial host defense recognition of commensal bacterial products does not elicit a full-scale, unrestrained immune response contributing to intestinal homeostasis. The importance of our findings also lends creditability to emerging evidence that flagellin can be administered under controlled conditions without evoking severe inflammation thus reducing harmful side effects, making it a promising vaccine adjuvant 
Second, our data also demonstrate that normal IEC do not distinguish between flagellin monomers of different bacterial species as the amino and carboxyl regions of this protein are conserved among bacterial strains and species. Thus recognition of flagellin, from pathogens such as S. Typhimurium and E. coli O83:H1 by apically localized TLR5 likely serves as the initial warning against bacterial invaders. Virulence factors possessed by pathogenic bacteria, but absent in commensal flora, are additional contributors to the full-scale, and at times unrestrained, host immune responses. Similar to the lungs, the intestine is constantly exposed to indigenous and environmental bacteria, thus is susceptible to infection. Therefore, it is reasonable to conclude that the host evolved an elaborate, complex immune system to detect dangerous invaders, and their products, at the earliest possible opportunity in order to fight against the infecting pathogens. Mounting evidence shows that the luminal surface of IEC is at the forefront of detecting factors that are injurious to the host, allowing the fight against these harmful factors to begin before they gain access to the lamina propria.