Crohn disease–associated adherent-invasive E. coli bacteria target mouse and human Peyer’s patches via long polar fimbriae

Crohn disease (CD) is a multifactorial disease in which an abnormal immune response in the gastrointestinal (GI) tract leads to chronic inflammation. The small intestine, particularly the ileum, of patients with CD is colonized by adherent-invasive E. coli (AIEC) — a pathogenic group of E. coli able to adhere to and invade intestinal epithelial cells. As the earliest inflammatory lesions are microscopic erosions of the epithelium lining the Peyer’s patches (PPs), we investigated the ability of AIEC bacteria to interact with PPs and the virulence factors involved. We found that AIEC bacteria could interact with mouse and human PPs via long polar fimbriae (LPF). An LPF-negative AIEC mutant was highly impaired in its ability to interact with mouse and human PPs and to translocate across monolayers of M cells, specialized epithelial cells at the surface of PPs. The prevalence of AIEC strains harboring the lpf operon was markedly higher in CD patients compared with controls. In addition, increased numbers of AIEC, but not LPF-deficient AIEC, bacteria were found interacting with PPs from Nod2^–/– mice compared with WT mice. In conclusion, we have identified LPF as a key factor for AIEC to target PPs. This could be the missing link between AIEC colonization and the presence of early lesions in the PPs of CD patients.

Crohn disease (CD) and ulcerative colitis (UC) are multifactorial diseases, occurring in individuals with genetic predisposition in whom an environmental or infectious trigger causes an abnormal immune response (1–3). Several lines of evidence suggest that bacteria play a role in the onset and perpetuation of IBD (4). E. coli bacteria have been assigned a putative role in CD. They are abnormally predominant in early and chronic ileal lesions of CD, and most E. coli strains isolated from the ileal mucosa of CD patients adhere to intestinal epithelial cells (IECs) (5–7). In addition to their ability to adhere, these E. coli bacteria are able to invade IECs and belong to a new pathogenic group of E. coli, designated adherent-invasive E. coli (AIEC) (8). Several independent studies have reported the abnormal presence of AIEC associated with ileal mucosa of CD patients (7, 9–13) owing to increased ileal expression of CEACAM6, which acts as a receptor for AIEC binding via type 1 pili to the intestinal mucosa (14, 15). The adhesion and invasion process of reference AIEC strain LF82 involves, in addition to type 1 pili, flagella, outer membrane proteins, and outer membrane vesicles (OMVs) (16–18). In particular, the LF82 invasion process occurs via the interaction between the ER-localized stress response chaperone Gp96 and the outer membrane protein OmpA expressed at the surface of OMVs, allowing OMVs to fuse with IECs (19). In addition, analysis of the genome sequence of AIEC strain LF82 revealed the presence of several known virulence genes and 4 putative pathogenic islands carrying virulence-related genes (20).

Clinical observations suggest that the sites of initial inflammation in ileal CD are the lymphoid follicles (21), since microscopic erosions are observable at the specialized follicle-associated epithelium (FAE), which lines Peyer’s patches (PPs) (22). The most prominent feature of FAE is the presence of specialized membranous/microfold cells (M cells), which are optimized for antigen adherence and transport and for immunological sampling of microorganisms (23). Several microorganisms, particularly invasive bacteria, take advantage of the transcytotic characteristics of M cells to use them to cross the intestinal barrier. For example, Yersinia enterocolitica and Yersinia pseudotuberculosis cross the intestinal epithelial barrier through M cells of the FAE (24), as does rabbit enteroadherent E. coli (25). The mechanism responsible for Y. pseudotuberculosis targeting M cells is essentially mediated by invasion (26). Salmonella Typhimurium bacteria selectively interact ex vivo with PPs (27), and a recent study reported that glycoprotein 2 (GP2), specifically expressed on the apical plasma membrane of M cells among enterocytes (28), is recognized by FimH, the adhesin of type 1 pili on bacteria (29). These observations demonstrate that the entry of type 1 piliated bacteria, such as E. coli and Salmonella Typhimurium, into M cells could occur by means of the GP2-FimH recognition. In addition, the long polar fimbriae (LPF), encoded by the lpf operon, were equally responsible for Salmonella Typhimurium specific adherence to M cells of the murine FAE (30). In the case of Shigella flexneri, bacteria entry occurs essentially through the M cell monolayer of the FAE, as shown in a rabbit ligated intestinal loop assay of infection, with the involvement of IpaB and IpaC effectors (31).

As the earliest observable lesions of CD are microscopic erosions of the FAE lining the PPs and as AIEC bacteria colonize the ileal mucosa of CD patients, we aimed in the present study to search for the presence of virulence factors in AIEC that could be involved in the targeting of PPs by these invasive bacteria. We report here, in addition of the involvement of type 1 pili, the presence of a functional lpf operon, encoding LPF in AIEC, that allows AIEC bacteria to interact with murine and human PPs and to translocate across M cells.

Analysis of the AIEC bacteria interaction with PPs. With the aim of identifying AIEC bacterial factors putatively involved in the targeting of PPs, we studied the interaction between AIEC LF82 bacteria and isolated PPs in Ussing chambers. As E. coli bacteria were previously reported to interact with PPs via the recognition of GP2 specifically expressed on apical plasma membrane of M cells by type 1 pili (28, 29), we studied the interaction of LF82 bacteria with murine PPs in the presence of antibodies raised against GP2. The addition of anti-GP2 antibodies during AIEC LF82 infection statistically decreased the number of LF82 bacteria interacting with PPs, from 31.89 × 10⁶ ± 8.73 × 10⁶ CFU in the absence of antibodies to 6.75 × 10⁶ ± 2.38 × 10⁶ CFU in the presence of anti-GP2 (Figure 1A). Immunostaining experiments using anti-GP2 antibodies to visualize M cells and anti-O83 antibodies to visualize LF82 bacteria confirmed that numerous LF82 wild-type bacteria colocalized with GP2 (Figure 1B). We also observed decreased numbers of AIEC bacteria interacting with murine PPs for the type 1 pili–negative mutant LF82-ΔfimA. However, we still observed type 1 pili–negative LF82 mutant associated with GP2-positive cells. In addition, we also still observed LF82 bacteria interacting with PPs when 0.5% methyl α-d-mannopyranoside was added in order to block the binding of type 1 pili to GP2. Together these results indicated that LF82 should express additional factor(s) targeting PPs in a mannose-independent manner.

Figure 1

Interaction of AIEC bacteria with PPs. (A) Interaction of bacteria with murine PPs placed in Ussing chambers after a 4-hour infection period, with or without anti-GP2 antibody (1 μg/ml) and with or without 0.5% methyl α-d-mannopyranoside (mann). When needed, PPs were preincubated with anti-GP2 antibody. (B) Confocal analysis of PP sections after labeling of AIEC LF82 with LPS O83 antibody (green), of M cells with anti-GP2 antibody (red), and DNA with Hoechst (blue). Scale bars: 20 μm. Arrowheads show clear colocalization of bacteria and M cells. (C) Interaction of wild-type LF82 and lpf mutant with murine PPs placed in Ussing chambers after a 4-hour infection period. (D) HES staining and confocal analysis of indicated areas after Cy3-EUB228 FISH staining to detect bacteria (red) and Hoechst to identify DNA (blue). Scale bars: 100 μm for HES staining and 20 μm for confocal analysis. Images in the bottom row correspond to the boxed regions in the top row. (E) In vivo interaction of wild-type LF82 and lpf mutant with murine PPs using ileal loop assay after a 4-hour infection period in the presence of 0.5% methyl α-d-mannopyranoside. (F) Interaction of wild-type LF82 and lpf mutant with murine small intestine mucosa without PPs after a 4-hour infection period in Ussing chambers. All results are expressed as numbers of mucosa-associated bacteria; each value is the mean ± SEM of at least 5 separate experiments. (G) Interaction of wild-type LF82 bacteria with murine PPs placed in Ussing chambers and coincubated with nonpathogenic MG1655 E. coli K-12 strain expressing or not expressing LPF_LF82. *P < 0.05, **P < 0.01, ***P < 0.001.

Identification of lpf operon in CD-associated E.coli strains. We analyzed the genome sequence of AIEC strain LF82 (Genoscope, ColiScope Project, NC_011993 890) with the aim of identifying AIEC putative virulence factors involved in PP targeting other than type 1 pili expressed in the entire family of Enterobacteriaceae and by about 60% of fecal isolates (32). This screening indicated the presence of a putative functional lpf operon encoding LPF (Supplemental Figure 1; supplemental material available online with this article; doi: 10.1172/JCI44632DS1).

The AIEC LF82-ΔlpfA mutant was impaired in its ability to interact with murine and human isolated PPs. To identify the involvement of LPF in the ability of AIEC strain LF82 to target PPs, we generated the LF82-ΔlpfA isogenic mutant. After 3 hours infection of murine PPs in the presence of 0.5% methyl α-d-mannopyranoside to better assess the role of LPF, a significant (P < 0.01), 6.7-fold decrease in the number of associated AIEC LF82 bacteria was observed for the LF82-ΔlpfA isogenic mutant compared with wild-type AIEC strain LF82 (Figure 1C). This was not due to differences in bacterial growth, since growth curves for the wild-type strain LF82 and the LF82-ΔlpfA mutant were similar (Supplemental Figure 2). Nor was this due to lack of expression of type 1 pili, since the LPF-negative mutant still expressed these pili at a level similar to that in LF82 wild-type bacteria (Supplemental Figure 3). The transcomplementation of the mutant with the cloned lpf operon restored the interaction with murine PPs. FISH experiments using probe EUB338 to visualize bacteria indicated numerous bacteria inside PPs for the wild-type strain LF82 (Figure 1D), and very few PP-associated bacteria for the LF82-ΔlpfA mutant. The number of PP-associated bacteria was fully restored when the LF82-ΔlpfA mutant was transcomplemented with cloned lpf operon, indicating that LPF are essential for AIEC LF82 bacteria to interact with murine PPs. To confirm these ex vivo data with murine intestinal tissues, we performed in vivo experiments using mouse ileal loop containing one PP. We observed a significant (P < 0.01), 6.5-fold decrease in the number of PP-associated bacteria for the LF82-ΔlpfA isogenic mutant compared with wild-type AIEC strain LF82, and the wild-type phenotype was fully restored for the LPF-negative mutant expressing cloned LPF (Figure 1E). To identify whether LPF also mediate AIEC adhesion to IECs, we performed additional experiments in Ussing chambers using small intestine without PPs. Results interestingly showed that the ability of the LF82-ΔlpfA isogenic mutant to adhere to intestinal epithelium was similar to that of the wild-type strain LF82 (Figure 1F). Thus, this indicated that the difference in the ability of LF82 bacteria and LPF-negative mutant to bind PPs was related to a specific adhesion of LPF to FAE. In addition, the interaction of bacteria with PPs was studied in competition experiments between LF82 and nonpathogenic E. coli MG1655 expressing or not expressing LPF. A slight, 1.6-fold-decreased adhesion of LF82 to PPs was observed during coinfection with the nonpathogenic MG1655 E. coli strain, probably due to competition of type 1 pili binding to the glycoprotein GP2; and a 4.5-fold-decreased LF82 adhesion to PPs was observed during coinfection with MG1655 strain expressing LPF, confirming that the adhesion of LF82 bacteria to PPs is related to the presence of LPF (Figure 1G).

To analyze the involvement of LPF expressed by AIEC bacteria to target PPs in humans, we performed similar experiments using human isolated PPs. Confocal microscopy analyzing the uptake of bacteria across the FAE and into the subepithelial dome (SED) of human PPs using Ussing chamber indicated that, compared with wild-type strain LF82, LF82-ΔlpfA mutant showed 4.9-fold-decreased transepithelial uptake (P < 0.01 vs. LF82) (Figure 2, A and B). Transcomplementation of the mutant with the cloned lpf operon restored its ability to interact with human PPs, indicating that LPF are also essential for AIEC LF82 bacteria to interact with human PPs.

Figure 2

Involvement of LPF to promote interaction between AIEC LF82 bacteria and PPs. (A) Quantification of intra-human PP-associated bacteria by confocal microscopy. Each value is the mean ± SEM of 2–4 separate experiments, with 3–5 sections studied for each experiment. **P < 0.01. (B) Representative confocal photomicrographs of uptake of bacteria across human FAE. Scale bars: 100 μm for HES staining and 50 μm for confocal analysis. Images in the bottom row correspond to the boxed regions in the top row. See the Figure 1D legend for staining.

The AIEC LF82-ΔlpfA isogenic mutant was impaired in its ability to translocate across M-like cell monolayer and to interact with murine M cells in vivo. We used the in vitro model of the specialized M cells, in which polarized enterocyte-like Caco-2-cl1 cells were cocultured with the Raji B cell line (33). To analyze whether these M-like cells express only a few developed microvilli as previously described (34), we performed F-actin phalloidin-TRITC labeling, and results were consistent with the presence of cells displaying morphological characteristics of M cells (Supplemental Figure 4A).

We investigated the interaction of AIEC strain LF82 with Caco-2-cl1 monocultures compared with M-like cells in the presence of 0.5% methyl α-d-mannopyranoside. Confocal analysis showed numerous AIEC LF82 bacteria interacting with the monolayer of M-like cells compared with the Caco-2-cl1 monolayer (Figure 3, A and B). This was also observed in the absence of methyl α-d-mannopyranoside and related to an absence of CEACAM6 expression by Caco-2-cl1 monolayers (Supplemental Figure 4B). Analysis of translocation across these monolayers indicated that only low numbers of LF82 bacteria translocated across Caco-2-cl1 monolayers even after 4 hours of infection, whereas high numbers of LF82 bacteria translocated across M cell monolayers (Figure 3D). As transepithelial electrical resistance (TEER) stayed constant during the 4 hours of infection (data not shown), this indicated that the large increase in the number of translocated bacteria was not the result of a loss of monolayer integrity.

Figure 3

LPF are involved in the ability of AIEC strain LF82 to target M cells. Interaction of AIEC bacteria LF82 (A and B) and LF82-ΔlpfA isogenic mutant (C) with Caco-2-cl1 (A) or M-like cell (B and C) monolayers in the presence of 0.5% methyl α-d-mannopyranoside. Phalloidin-TRITC labeling of F-actin (red), anti-O83 antibody labeling of LF82 bacteria (green), and Hoechst labeling of DNA (blue) were used. Confocal photomicrographs of interaction of bacteria with in vitro M cells are representative of 3 separate experiments. Scale bars: 25 μm. (D) Translocation across Caco-2-cl1 or M-like cell monolayers. Results are expressed as CFU of translocated bacteria. Each value is the mean ± SEM of at least 5 separate experiments. (E) Translocation of LF82 bacteria and LF82-ΔlpfA isogenic mutant across M-like cell monolayers after 4 hours infection. Results are expressed as translocated bacteria relative to those obtained for strain LF82, taken as 1. (F) Confocal analysis of murine PP sections after labeling of AIEC LF82 with LPS O83 antibody (green), of M cell with UEA-1 TRITC (red), and DNA with Hoechst (blue). Scale bars: 20 μm. Arrowheads indicate UEA-1–positive cells. (G) Quantification of murine M cell–associated bacteria by confocal microscopy analysis. Bars represent the mean. **P < 0.01, ***P < 0.001.

Decreased interaction of the LF82-ΔlpfA mutant compared with the wild-type strain was observed with M-like cells (Figure 3, B and C), and after 4 hours of infection, significantly decreased translocation was observed for the LF82-ΔlpfA mutant compared with the wild-type strain (Figure 3E). Transcomplementation of LF82-ΔlpfA mutant with cloned lpf fully restored the phenotype, indicating that LPF play a key role in targeting M cells. The observation that LPF expression is crucial for AIEC interaction with M cells was confirmed ex vivo using PPs isolated from mice and confocal analysis to search for colocalization of bacteria with M cells stained with TRITC-labeled Ulexeuropaeus agglutinin 1 at the surface of whole PPs. Results showed that the number of bacteria associated with M cells was significantly higher (P < 0.001) for wild-type strain LF82 than for the LF82-ΔlpfA mutant (Figure 3, F and G).

Prevalence of lpf-positive adherent-invasive E. coli strains in CD patients and controls. Various lpf operons are observed among bacteria belonging to the Enterobacteriaceae family, including Shigella spp., Salmonella spp., Citrobacter rodentium, enterohemorrhagic E. coli (EHEC), and rabbit enteropathogenic E. coli (REPEC). BLAST analysis ( http://blast.ncbi.nlm.nih.gov/Blast.cgi) indicated that LpfA protein from AIEC strain LF82 (CU651637) was most closely similar to that of Salmonella Typhimurium strain LT2 (NC_003197), with 86% similarity, and to that of EHEC strain EDL-933 encoded by lpf1 (NC_002655), with 82% similarity (Supplemental Figure 1 and Supplemental Figure 5A). DNA sequencing of the LpfA-encoding genes of all the CD-associated E. coli strains positive in hybridization showed that, among ileal-associated E. coli strains isolated from 55 CD patients, none harbored the lpfA gene of Salmonella Typhimurium. lpfA_LF82 -positive E. coli strains were found in 8 (14.5%) patients, and lpfA_Shigella-positive E. coli strains were found in 23 (41.8%) patients (Supplemental Figure 5). We only observed E. coli strains positive for lpfA1_EHEC or lpfA2_EHEC in 4 patients, in 2 of whom the E. coli strains associated with the ileal mucosa also harbored the lpfA_Shigella gene.

As a high prevalence of putatively functional lpfA_Shigella-positive AIEC strains was observed in CD patients, we compared LPF_Shigella and LPF_LF82 in their ability to mediate bacterial interaction with PPs. Thus, experiments were performed with CD-associated AIEC strain LF110 harboring an lpfA gene similar to that of S. flexneri 2a strain 301 (with 99% similarity, Supplemental Figure 5B) and the corresponding lpfA-negative mutant. The comparison of the ability of AIEC strain LF110 and strain LF82 to interact with murine PPs revealed similar numbers of PP-associated bacteria for the two AIEC strains (Figure 4). Like the LF82-ΔlpfA mutant, the LF110-ΔlpfA mutant showed a significantly decreased ability to interact with murine PPs compared with the wild-type strain. Therefore, the analysis of the presence of any lpf operon in ileal-associated E. coli strains indicated a statistically significant difference (P = 0.006) in the prevalence of lpfA-positive E. coli strains associated with ileal specimens between CD patients and controls (Table 1). Of the 55 CD patients, 26 (47.3%) harbored lpfA-positive E. coli strains, in contrast to only 5 (17.2%) of the 29 controls without IBD. When we analyzed the presence of the lpfA gene in AIEC and non-AIEC strains, we observed that 12 CD patients harbored lpfA-positive AIEC strains. Fourteen CD patients harbored lpfA-positive E. coli strains that were not initially classified as AIEC on the basis of the expression of AIEC-specific virulence properties. Even without considering a possible underestimation of the AIEC phenotype, the prevalence of subjects harboring lpf-positive AIEC strains was significantly (P = 0.027) higher in CD patients than in controls (21.8% versus 3.5%). Of note, most of the E. coli strains associated with the ileal mucosa of controls were non-AIEC and did not harbor the lpfA gene, which was highly significant compared with CD patients (P = 0.00006).

Figure 4

Comparison of ability of LPFShigella and LPF_LF82 bacteria to interact with murine PPs. (A) Quantification of murine PP-associated bacteria for AIEC strains LF82 and LF110 and corresponding ΔlpfA mutants. See the legend to Figure 1C. Each value is the mean ± SEM of at least 5 separate experiments. (B) Visualization of bacteria interacting with murine isolated PPs. Scale bars: 100 μm for HES staining and 50 μm for confocal analysis. Images in the right column correspond to the boxed regions in the left row. See the legend to Figure 1D. *P < 0.05, **P < 0.01.

Table 1

Prevalence of adherent-invasive E. coli positive for lpfA gene in ileal specimens from CD patients and controls

LPF-dependent increased interaction of AIEC LF82 bacteria with Nod2^–/– PPs. Confocal quantification of M cells on PPs using TRITC-labeled UEA1 indicated a 2.9-fold-higher number in Nod2^–/– mice compared with wild-type mice (Figure 5, A–C), confirming previous findings by Barreau et al. (35). In addition, quantification of GP2-positive cells using anti-GP2 antibody indicated a 2.8-fold-higher number of GP2-positive cells in Nod2^–/– mice compared with wild-type mice (Figure 5, D–F). As GP2 expression is restricted to M cells in FAE (29), this confirmed an increased number of M cells in PPs of Nod2^–/– mice compared with wild-type mice.

Figure 5

LPF-dependent increased interaction of AIEC LF82 bacteria with Nod2^–/– PPs. Visualization (A, B, D, and E) and quantification (C and F) of UEA-1 TRITC–stained M cells (red) (A–C) and GP2-positive cells (red) (D–F) present in FAE of wild-type (A and D) and Nod2^–/– (B and E) mice after cryostat section. Scale bars: 20 μm. Arrowheads indicate UEA-1– or GP2-positive cells. In C and F, results are expressed as number of M cells present in 100 μm of FAE. Each value is the mean ± SEM of 6 separate experiments, with 5–6 sections studied for each experiment. (G) Ability of wild-type LF82 bacteria and LPF-negative mutant to interact with murine PPs from wild-type and Nod2^–/– mice. See the legend to Figure 1C. Each value is the mean ± SEM of at least 5 separate experiments. *P < 0.05, **P < 0.01, ***P < 0.001.

We investigated the ability of AIEC LF82 bacteria and LF82-ΔlpfA mutant to interact with murine PPs isolated from wild-type and Nod2^–/– mice. Results presented in Figure 5G show that the number of PP-associated LF82 bacteria was significantly higher for PPs from Nod2^–/– mice than from wild-type mice (P < 0.05). In contrast, the number of LPF-negative LF82 mutant interacting with PPs was similar regardless of the Nod2 genotype, indicating that the increased number of LF82 bacteria observed with PPs from Nod2^–/– mice was linked to the expression of LPF.

Clinical observations suggest that the sites of initial inflammation in ileal CD are the lymphoid follicles (21, 22). A correlation exists between peak of CD case frequency (between 15 and 25 years old) and peak representing the number of PPs according to age, suggesting that CD develops as an inflammatory process specifically targeting these important lymphoid structures (36). Moreover, recent studies reported a defect in the intestinal barrier to control bacterial translocation across PPs of CD (37, 38). It is well documented that several enteroinvasive pathogens, such as Yersinia, Shigella, and Salmonella, invade the intestinal mucosa via the M cells of the FAE (24, 27, 31). The molecular mechanism of the specific interaction between the invasive pathogens and M cells involve the Inv invasin for Yersinia (26), the IpaB and IpaC effectors for Shigella (31), and LPF for Salmonella (30). It is also known that enteropathogenic E. coli (EPEC) are able to translocate across an in vitro M cell model and that this translocation is regulated by the type III secretion system, but the bacterial effector involved is not yet known (34).

Here we show that AIEC bacteria are able to interact with PPs involving at least one virulence factor in addition to type 1 pili. The analysis of the genome sequence of AIEC reference strain LF82 indicates the presence of an lpf operon encoding LPF. Such lpf operons were reported in Salmonella Typhimurium, Shigella boydii, and S. flexneri and in EHEC EDL933 (30, 39), but this is the first report to our knowledge of the presence of an lpf operon in CD-associated AIEC bacteria. In Salmonella Typhimurium, LPF promote bacterial interaction to murine PPs (30), and ex vivo assays performed with a murine intestinal organ culture model showed that their interaction with PPs leads to full expression of LPF (40). No role has been yet reported for LPF in Shigella. For EHEC, experiments in pigs and sheep with O157:H7 strain 86-24 indicated that LPF contribute to intestinal colonization (41). To study the involvement of LPF in AIEC LF82 interaction with PPs, we performed in vitro experiments with M cells, and ex vivo assays with murine and human PPs. We observed that LPF have a crucial and type 1 pili–independent role in mediating the interaction between AIEC LF82 bacteria and PPs. Numerous AIEC LF82 bacteria were observed within PPs, as shown by ex vivo assays using murine and human PPs for wild-type strain, but not for the LPF-negative LF82 mutant.

M cells can serve as an entry portal for many enteroinvasive pathogens. We observed here that AIEC LF82 bacteria translocated at a very high level through M cell monolayers, but not as the result of the loss of monolayer integrity, since we observed that the TEER of M cell monolayers was not modified after infection with strain LF82. A decreased ability to translocate was observed for the LPF-negative mutant, and confocal analysis showed that AIEC LF82 bacteria interacted with M cells at the surface of murine PPs only when LPF were expressed. Together, these experiments demonstrate a functional role for the LPF of AIEC strain LF82 in targeting M cells on the surface of PPs. Of note, the enteric pathogens crossing the FAE need to survive phagocytosis and bacterial killing by resident and recruited macrophages in the dome of the lymphoid follicle. This is not really a problem for AIEC bacteria, since one of the characteristics of AIEC is their ability to survive and replicate at a high rate in phagolysosomes within macrophage (42, 43). Thus, the expression of LPF in AIEC bacteria is not deleterious for these invasive bacteria crossing the FAE.

Mutations in the CARD15/NOD2 gene are associated with abnormal development and function of PPs, since Nod2-knockout mice exhibit an increased number of PPs also characterized by an excess of M cells and CD4⁺ T cells (35). In the present study, probably owing to the increased numbers of M cells, we observed LPF-dependent increases in numbers of AIEC LF82 bacteria interacting with PPs from Nod2^–/– compared with WT mice. Interestingly, AIEC bacteria are more frequently isolated during early recurrence of ileal CD (9), and Nod2 mutations are mostly associated with ileal involvement, structuring complications, and a slightly earlier age of onset (44). In this context, the analysis of a large cohort of CD patients is necessary to search for a possible correlation between NOD2 polymorphisms and the presence of adherent-invasive E. coli expressing LPF. Of note, NOD2-deficient mice do not develop macroscopically or clinically evident colitis in response to their own microbiota or when challenged with AIEC bacteria (N. Barnich, unpublished observations), indicating that some other genetic or environmental factors are required to create the necessary conditions for increasing host susceptibility to invasion by LPF-expressing bacteria. Perhaps the generation of humanized transgenic mice expressing CEACAM6 in the small bowel using a villin promoter and knockout for Nod2 could indicate whether ileal bacterial colonization involving abnormal CEACAM6 expression is a missing host susceptibility factor.

Two major lpf operons were identified in CD-associated E. coli strains, similar to those found in strain LF82 and Shigella. Interestingly, we show here that the roles of the LPFShigella and LPF_LF82 in their ability to interact with PPs are similar, and that there was a statistically significant increase in the prevalence of lpfA-positive E. coli strains, whatever the lpfA origin, in ileal specimens from CD patients compared with controls. We therefore suggest that lpfA probes are very useful and sensitive molecular tools for investigating the presence of fully pathogenic AIEC in CD patients, since we observed some lpfA-positive CD-associated E. coli strains that were not initially classified as AIEC on the basis of all AIEC-specific virulence properties.

In conclusion, our findings, in addition to the suspected major role of PPs in CD as observed by histological and functional examination of the gut mucosa, show that CD-associated AIEC bacteria by expressing LPF can use PPs as an open gate to induce early stages of the disease. This observation could be very relevant, since there is a potential role of gastrointestinal infections in the onset of IBD as reported by two studies, including a recent Danish population-based cohort long-term follow-up study that showed that gastroenteritis caused by nontyphoid Salmonella or thermophilic Campylobacter increases the risk of developing IBD (45, 46). LPF are expressed by Salmonella, giving weight to the idea that PPs are major players in IBD. Further analyses are therefore needed to investigate the presence of LPF-expressing pathogens in CD tissues. In particular, such studies should address whether the Gram-negative intramucosal bacteria observed in 73% of CD patients express LPF, since 58% of these intramucosal bacteria were not identified as E. coli (7). Taken together, our data reinforce the idea that underlying genetic components in hosts affected by IBD might create the necessary conditions for increasing host susceptibility to invasion by LPF-expressing bacteria, which aggravates the disease. This is another example of co-evolution between pathogens and the host, as we previously reported that AIEC bacteria express type 1 pili variant to better bind to abnormally expressed CEACAM6 by ileal enterocytes of CD patients (14, 15). From the data presented here, the co-evolution could involve, on the host side, increased numbers of PPs and M cells in patients and, on the pathogen side, the presence of virulence genes encoding LPF. We previously reported that CEACAM6 expression by IECs is positively regulated by AIEC adhesion and also by proinflammatory cytokine stimulation, and so we can speculate that more bacteria will reach the FAE underlying macrophages, and that more proinflammatory cytokines will be synthesized, leading to an amplification loop of colonization and inflammation.

View Supplemental data

This study was supported by the Ministère de la Recherche et de la Technologie (JE2526), INRA (USC-2018), and by grants from the Association F. Aupetit (AFA), the European Commission through the FP7 IBDase project, and the Swedish Research Council (VR-M 2005-7013). Studies in B.J. Campbell’s laboratory were supported by the National Institute of Health Research Specialist Biomedical Research Centre in Microbial Disease (01CD1), the Wellcome Trust UK (074949/Z/04/Z), and the National Association for Colitis and Crohn’s Disease UK (M/08/1). We thank Lothar Beutin (Department of Biological Safety, Robert Koch Institut, Berlin, Germany) for anti-O83 antibodies. We also thank Pierre Sauvanet for help with tissue inclusions and sections, Adeline Sivignon for help with hybridization experiments, and Nicolas Barnich for help with Ussing chambers and animal experiments. We thank Christelle Blavignac, Claire Sczepaniak, and Lorraine Gameiro (CICS, Université Auvergne) for technical assistance with electron microscopy and the CICS platform for confocal microscopy.

Address correspondence to: Arlette Darfeuille-Michaud, Pathogénie Bactérienne Intestinale, CBRV, 28 place Henri Dunant, 63001 Clermont-Ferrand, France. Phone: 33.4.73.17.79.97; Fax: 33.4.73.17.83.71; E-mail: arlette.darfeuille-michaud@u-clermont1.fr.

Conflict of interest: The authors have declared that no conflict of interest exists.

Reference information: J Clin Invest. 2011;121(3):966–975. doi:10.1172/JCI44632.

See the related article at Adherent-invasive E. coli in Crohn disease: bacterial “agent provocateur”.

1. Kaser A, Zeissig S, Blumberg RS. Inflammatory bowel disease. Annu Rev Immunol. 2010;28:573–621.
   View this article via: PubMed CrossRef Google Scholar
2. Strober W, Fuss I, Mannon P. The fundamental basis of inflammatory bowel disease. J Clin Invest. 2007;117(3):514–521.
   View this article via: JCI PubMed CrossRef Google Scholar
3. Xavier RJ, Podolsky DK. Unravelling the pathogenesis of inflammatory bowel disease. Nature. 2007;448(7152):427–434.
   View this article via: PubMed CrossRef Google Scholar
4. Sartor RB. Microbial influences in inflammatory bowel diseases. Gastroenterology. 2008;134(2):577–594.
   View this article via: PubMed CrossRef Google Scholar
5. Darfeuille-Michaud A, et al. Presence of adherent Escherichia coli strains in ileal mucosa of patients with Crohn’s disease. Gastroenterology. 1998;115(6):1405–1413.
   View this article via: PubMed CrossRef Google Scholar
6. Swidsinski A, et al. Mucosal flora in inflammatory bowel disease. Gastroenterology. 2002;122(1):44–54.
   View this article via: PubMed CrossRef Google Scholar
7. Martin HM, et al. Enhanced Escherichia coli adherence and invasion in Crohn’s disease and colon cancer. Gastroenterology. 2004;127(1):80–93.
   View this article via: PubMed CrossRef Google Scholar
8. Boudeau J, Glasser AL, Masseret E, Joly B, Darfeuille-Michaud A. Invasive ability of an Escherichia coli strain isolated from the ileal mucosa of a patient with Crohn’s disease. Infect Immun. 1999;67(9):4499–4509.
   View this article via: PubMed Google Scholar
9. Darfeuille-Michaud A, et al. High prevalence of adherent-invasive Escherichia coli associated with ileal mucosa in Crohn’s disease. Gastroenterology. 2004;127(2):412–421.
   View this article via: PubMed CrossRef Google Scholar
10. Martinez-Medina M, et al. Molecular diversity of Escherichia coli in the human gut: new ecological evidence supporting the role of adherent-invasive E. coli (AIEC) in Crohn’s disease. Inflamm Bowel Dis. 2009;15(6):872–882.
    View this article via: PubMed CrossRef Google Scholar
11. Baumgart M, et al. Culture independent analysis of ileal mucosa reveals a selective increase in invasive Escherichia coli of novel phylogeny relative to depletion of Clostridiales in Crohn’s disease involving the ileum. Isme J. 2007;1(5):403–418.
    View this article via: PubMed CrossRef Google Scholar
12. Sasaki M, et al. Invasive Escherichia coli are a feature of Crohn’s disease. Lab Invest. 2007;87(10):1042–1054.
    View this article via: PubMed CrossRef Google Scholar
13. Eaves-Pyles T, et al. Escherichia coli isolated from a Crohn’s disease patient adheres, invades, and induces inflammatory responses in polarized intestinal epithelial cells. Int J Med Microbiol. 2008;298(5–6):397–409.
    View this article via: PubMed CrossRef Google Scholar
14. Carvalho FA, et al. Crohn’s disease adherent-invasive Escherichia coli colonize and induce strong gut inflammation in transgenic mice expressing human CEACAM. J Exp Med. 2009;206(10):2179–2189.
    View this article via: PubMed CrossRef Google Scholar
15. Barnich N, et al. CEACAM6 acts as a receptor for adherent-invasive E. coli, supporting ileal mucosa colonization in Crohn disease. J Clin Invest. 2007;117(6):1566–1574.
    View this article via: JCI PubMed CrossRef Google Scholar
16. Barnich N, Boudeau J, Claret L, Darfeuille-Michaud A. Regulatory and functional co-operation of flagella and type 1 pili in adhesive and invasive abilities of AIEC strain LF82 isolated from a patient with Crohn’s disease. Mol Microbiol. 2003;48(3):781–794.
    View this article via: PubMed CrossRef Google Scholar
17. Boudeau J, Barnich N, Darfeuille-Michaud A. Type 1 pili-mediated adherence of Escherichia coli strain LF82 isolated from Crohn’s disease is involved in bacterial invasion of intestinal epithelial cells. Mol Microbiol. 2001;39(5):1272–1284.
    View this article via: PubMed CrossRef Google Scholar
18. Rolhion N, Carvalho FA, Darfeuille-Michaud A. OmpC and the sigma(E) regulatory pathway are involved in adhesion and invasion of the Crohn’s disease-associated Escherichia coli strain LF82. Mol Microbiol. 2007;63(6):1684–1700.
    View this article via: PubMed CrossRef Google Scholar
19. Rolhion N, et al. Abnormally expressed ER stress response chaperone Gp96 in CD favours adherent-invasive Escherichia coli invasion. Gut. 2010;59(10):1355–1362.
    View this article via: PubMed Google Scholar
20. Miquel S, et al. Complete genome sequence of Crohn’s disease–associated adherent–invasive E. coli strain LF82. PLoS One. 2010;5(9) ). :pii: e12714.
21. Morson BC. The early histological lesion of Crohn’s disease. Proc R Soc Med. 1972;65(1):71–72.
    View this article via: PubMed Google Scholar
22. Gullberg E, Soderholm JD. Peyer’s patches and M cells as potential sites of the inflammatory onset in Crohn’s disease. Ann N Y Acad Sci. 2006;1072:218–232.
    View this article via: PubMed CrossRef Google Scholar
23. Kraehenbuhl JP, Neutra MR. Epithelial M cells: differentiation and function. Annu Rev Cell Dev Biol. 2000;16:301–332.
    View this article via: PubMed CrossRef Google Scholar
24. Marra A, Isberg RR. Invasin-dependent and invasin-independent pathways for translocation of Yersiniapseudotuberculosis across the Peyer’s patch intestinal epithelium. Infect Immun. 1997;65(8):3412–3421.
    View this article via: PubMed Google Scholar
25. Inman LR, Cantey JR. Specific adherence of Escherichiacoli (strain RDEC-1) to membranous (M) cells of the Peyer’s patch in Escherichiacoli diarrhea in the rabbit. J Clin Invest. 1983;71(1):1–8.
    View this article via: JCI PubMed CrossRef Google Scholar
26. Isberg RR, Leong JM. Multiple beta 1 chain integrins are receptors for invasin, a protein that promotes bacterial penetration into mammalian cells. Cell. 1990;60(5):861–871.
    View this article via: PubMed CrossRef Google Scholar
27. Jones BD, Ghori N, Falkow S. Salmonellatyphimurium initiates murine infection by penetrating and destroying the specialized epithelial M cells of the Peyer’s patches. J Exp Med. 1994;180(1):15–23.
    View this article via: PubMed CrossRef Google Scholar
28. Terahara K, et al. Comprehensive gene expression profiling of Peyer’s patch M cells, villous M-like cells, and intestinal epithelial cells. J Immunol. 2008;180(12):7840–7846.
    View this article via: PubMed Google Scholar
29. Hase K, et al. Uptake through glycoprotein 2 of FimH(+) bacteria by M cells initiates mucosal immune response. Nature. 2009;462(7270):226–230.
    View this article via: PubMed CrossRef Google Scholar
30. Baumler AJ, Tsolis RM, Heffron F. The lpf fimbrial operon mediates adhesion of Salmonellatyphimurium to murine Peyer’s patches. Proc Natl Acad Sci U S A. 1996;93(1):279–283.
    View this article via: PubMed CrossRef Google Scholar
31. Sansonetti PJ, Arondel J, Cantey JR, Prevost MC, Huerre M. Infection of rabbit Peyer’s patches by Shigellaflexneri: effect of adhesive or invasive bacterial phenotypes on follicle-associated epithelium. Infect Immun. 1996;64(7):2752–2764.
    View this article via: PubMed Google Scholar
32. Johnson JR. Virulence factors in Escherichiacoli urinary tract infection. Clin Microbiol Rev. 1991;4(1):80–128.
    View this article via: PubMed Google Scholar
33. Kerneis S, Bogdanova A, Kraehenbuhl JP, Pringault E. Conversion by Peyer’s patch lymphocytes of human enterocytes into M cells that transport bacteria. Science. 1997;277(5328):949–952.
    View this article via: PubMed CrossRef Google Scholar
34. Martinez-Argudo I, Sands C, Jepson MA. Translocation of enteropathogenic Escherichiacoli across an in vitro M cell model is regulated by its type III secretion system. Cell Microbiol. 2007;9(6):1538–1546.
    View this article via: PubMed CrossRef Google Scholar
35. Barreau F, et al. CARD15/NOD2 is required for Peyer’s patches homeostasis in mice. PLoS ONE. 2007;2(6):e523.
    View this article via: PubMed CrossRef Google Scholar
36. Van Kruiningen HJ, Ganley LM, Freda BJ. The role of Peyer’s patches in the age-related incidence of Crohn’s disease. J Clin Gastroenterol. 1997;25(2):470–475.
    View this article via: PubMed CrossRef Google Scholar
37. Salim SY, et al. CD83+CCR7- dendritic cells accumulate in the subepithelial dome and internalize translocated Escherichiacoli HB101 in the Peyer’s patches of ileal Crohn’s disease. Am J Pathol. 2009;174(1):82–90.
    View this article via: PubMed CrossRef Google Scholar
38. Keita AV, et al. Increased uptake of non-pathogenic E. coli via the follicle-associated epithelium in longstanding ileal Crohn’s disease. J Pathol. 2008;215(2):135–144.
    View this article via: PubMed CrossRef Google Scholar
39. Torres AG, et al. Identification and characterization of lpfABCC’DE, a fimbrial operon of enterohemorrhagic Escherichiacoli O157:H7. Infect Immun. 2002;70(10):5416–5427.
    View this article via: PubMed CrossRef Google Scholar
40. Norris TL, Kingsley RA, Bumler AJ. Expression and transcriptional control of the Salmonellatyphimuriumlpf fimbrial operon by phase variation. Mol Microbiol. 1998;29(1):311–320.
    View this article via: PubMed CrossRef Google Scholar
41. Jordan DM, Cornick N, Torres AG, Dean-Nystrom EA, Kaper JB, Moon HW. Long polar fimbriae contribute to colonization by Escherichiacoli O157:H7 in vivo. Infect Immun. 2004;72(10):6168–6171.
    View this article via: PubMed CrossRef Google Scholar
42. Bringer MA, Barnich N, Glasser AL, Bardot O, Darfeuille-Michaud A. HtrA stress protein is involved in intramacrophagic replication of adherent and invasive Escherichiacoli strain LF82 isolated from a patient with Crohn’s disease. Infect Immun. 2005;73(2):712–721.
    View this article via: PubMed Google Scholar
43. Glasser AL, Boudeau J, Barnich N, Perruchot MH, Colombel JF, Darfeuille-Michaud A. Adherent invasive Escherichiacoli strains from patients with Crohn’s disease survive and replicate within macrophages without inducing host cell death. Infect Immun. 2001;69(9):5529–5537.
    View this article via: PubMed CrossRef Google Scholar
44. Economou M, Trikalinos TA, Loizou KT, Tsianos EV, Ioannidis JP. Differential effects of NOD2 variants on Crohn’s disease risk and phenotype in diverse populations: a metaanalysis. Am J Gastroenterol. 2004;99(12):2393–2404.
    View this article via: PubMed CrossRef Google Scholar
45. Gradel KO, Nielsen HL, Schonheyder HC, Ejlertsen T, Kristensen B, Nielsen H. Increased short- and long-term risk of inflammatory bowel disease after Salmonella or Campylobacter gastroenteritis. Gastroenterology. 2009;137(2):495–501.
    View this article via: PubMed CrossRef Google Scholar
46. Porter CK, Tribble DR, Aliaga PA, Halvorson HA, Riddle MS. Infectious gastroenteritis and risk of developing inflammatory bowel disease. Gastroenterology. 2008;135(3):781–786.
    View this article via: PubMed CrossRef Google Scholar
47. Datsenko KA, Wanner BL. One-step inactivation of chromosomal genes in Escherichiacoli K-12 using PCR products. Proc Natl Acad Sci U S A. 2000;97(12):6640–6645.
    View this article via: PubMed Google Scholar
48. Guzman LM, Belin D, Carson MJ, Beckwith J. Tight regulation, modulation, and high-level expression by vectors containing the arabinose PBAD promoter. J Bacteriol. 1995;177(14):4121–4130.
    View this article via: PubMed Google Scholar
49. Hogardt M, Trebesius K, Geiger AM, Hornef M, Rosenecker J, Heesemann J. Specific and rapid detection by fluorescent in situ hybridization of bacteria in clinical samples obtained from cystic fibrosis patients. J Clin Microbiol. 2000;38(2):818–825.
    View this article via: PubMed Google Scholar
50. Hitotsubashi S, Fujii Y, Yamanaka H, Okamoto K. Some properties of purified Escherichiacoli heat-stable enterotoxin II. Infect Immun. 1992;60(11):4468–4474.
    View this article via: PubMed Google Scholar
51. Lo D, et al. Cell culture modeling of specialized tissue: identification of genes expressed specifically by follicle-associated epithelium of Peyer’s patch by expression profiling of Caco-2/Raji co-cultures. Int Immunol. 2004;16(1):91–99.
    View this article via: PubMed CrossRef Google Scholar
52. Gullberg E, et al. Expression of specific markers and particle transport in a new human intestinal M-cell model. Biochem Biophys Res Commun. 2000;279(3):808–813.
    View this article via: PubMed Google Scholar