• Crohn's disease;
  • ulcerative colitis;
  • intestinal microbiota;
  • Escherichia coli;
  • adherent-invasive strains


  1. Top of page
  2. Abstract


Crohn's disease (CD) and ulcerative colitis (UC), known as inflammatory bowel diseases (IBD), are characterized by an abnormal immunological response to commensal bacteria colonizing intestinal lumen and mucosa. Among the latter, strains of adherent-invasive Escherichia coli (AIEC), capable of adhering to and invading epithelium, and to replicate in macrophages, have been described in CD adults. We aimed at identifying and characterizing AIEC strains in pediatric IBD.


In all, 24 CD children, 10 UC, and 23 controls were investigated. Mucosal biopsies, taken during colonoscopy, were analyzed for the presence of AIEC strains by an adhesive-invasive test. Protein expression of the specific AIEC receptor, the carcinoembryonic antigen-related cell adhesion molecule 6 (CEACAM6), was evaluated by western blot and immunohistochemistry, while tumor necrosis factor alpha (TNF-α) and interleukin (IL)-8 mRNA expression was detected by real-time polymerase chain reaction (PCR), after bacterial infection. Transmission electron microscopy and trans-epithelial electric resistance assays were performed on biopsies to assess bacteria-induced morphological and functional epithelial alterations.


Two bacterial strains, EC15 and EC10, were found to adhere and invade the Caco2 cell line, similar to the well-known AIEC strain LF82 (positive control): they upregulated CEACAM6, TNF-α, and IL-8 gene/protein expression, in vitro and in cultured intestinal mucosa; they could also survive inside macrophages and damage the epithelial barrier integrity. Lesions in the inflamed tissues were associated with bacterial infection.


This is the first study showing the presence of adhesive-invasive bacteria strains in the inflamed tissues of children with IBD. Collective features of these strains indicate that they belong to the AIEC spectrum, suggesting their possible role in disease pathogenesis. (Inflamm Bowel Dis 2011)

Inflammatory bowel diseases (IBD), including Crohn's disease (CD) and ulcerative colitis (UC), are relatively common chronic disorders, thought to result from inappropriate and continual activation of intestinal mucosa immune system due to a complex interplay between genetic, immunological, and microbial factors.1, 2 Clinical and experimental studies indicate that intestinal bacteria are involved in the initiation and amplification of IBD: indeed, colitis does not occur in most gene knockout models of IBD when animals are kept in a germfree environment,3, 4 and both animal models of colitis and patients with IBD respond to antibiotics and probiotics.5 In addition, most genes associated with susceptibility to IBD, including NOD2/CARD15, ATG16L1, and IRGM, encode proteins involved in host–microbial interactions.6 Immunological studies have also shown that the majority of patients with IBD develop serological and T-cell response to their own enteric flora.7 The role of bacteria in the pathogenesis of IBD is further supported by the observation that surgical diversion of fecal stream or treatment with bowel rest and total parenteral nutrition have successfully been used to control enteric inflammation.8 Recent advances in molecular techniques in IBD patients have shown a reduction in beneficial bacteria groups such as members of the phyla Firmicutes and Bacteroidetes and an increase in potentially pathogenic bacteria such as Proteobacteriae, in particular Escherichia coli.9, 10 An association between particular adherent invasive E. coli (AIEC) strains and IBD has recently been hypothesized.11–13 These strains are able to adhere and to invade cultured intestinal epithelial cells and can survive extensively within macrophages, inducing secretion of high levels of tumor necrosis factor alpha (TNF-α).14 Adhesion of these strains depends on expression of type 1 pili in the bacterial surface as well as of cell adhesion molecule 6 related to carcinoembryonic antigen (CEACAM6) in the apical surface of ileal epithelial cells.15, 16 CEACAM6 acts as a receptor for AIEC and has been shown to be abnormally expressed in the ileal epithelial cells of adult patients with CD.17 It is still unclear if the increase in certain types of E. coli in IBD is a consequence of inflammation or the cause.

Due to the important role played by these bacteria in the pathogenesis of IBD, as previously suggested by other studies carried out in adults, the aim of the present study was to identify and characterize adhesive-invasive bacterial strains in pediatric IBD. Moreover, tissue damage, potentially associated with the presence of bacteria, was investigated. We show for the first time the presence of adherent-invasive bacteria, thought to belong to the AIEC spectrum, in the inflamed intestinal tissue of pediatric IBD patients.


  1. Top of page
  2. Abstract


In all, 34 IBD patients, 24 CD (seven female; median age 14.0 years, range 9–18 years), and 10 UC (five female; median age 13,0 years, range 10–17 years), were recruited between 2008 and 2010 at the Pediatric Gastroenterology and Liver Unit of the Sapienza University of Rome.

The diagnosis of CD and UC was based on widely agreed endoscopic and histological criteria as well as on the exclusion of infectious diseases, systemic disorders, food allergies, and malabsorption syndromes.18 Clinical disease activity was measured using well-established parameters of the pediatric CD and UC activity indexes (PCDAI and PUCAI, respectively).19, 20 All children underwent ileocolonoscopy performed by the same endoscopist with a pediatric videocolonoscope (Pediatric Videocolonoscope Olympus, PCF 160 AL) under general anesthesia or conscious sedation with intravenous pethidine (1–2 mg/kg) and midazolan (0.1 mg/kg).

Table 1 reports patient demographic and clinical characteristics as well as therapy at the time of endoscopy. When included in the study, three CD and two UC patients were not receiving any treatment, whereas the others were treated with different drugs, including immunomodulators (azathioprine), mesalazine, or oral corticosteroids at low doses; three patients were treated with biological drugs. No patient received antibiotics for 6 months before colonoscopy. The control population consisted of 23 children (five female; mean age 12.4 years, range 10–17 years), investigated for symptoms and signs of functional gastrointestinal disorders, without organic or inflammatory disease, as documented by normal endoscopy and histology.

Table 1. Demographic and Clinical Characteristics of the IBD Study Population
SubjectSexAge (Years)Disease (Years)Therapy at Time of EndoscopyPCDAILocationa
CD2M152Azathioprine Budesonide40L3
CD8F151 45L1
CD11M134Azathioprine Mesalamine30L3
CD12M142Mesalamine Azathioprine10L1
CD13F173Mesalamine Azathioprine30L1
CD17M150 10L1
CD19F141Adalimumab L3
CD21F131 15L2
SubjectsSexAge (Years)Disease (Years)Therapy at Time of EndoscopyPUCAILocationb
  • a

    Montreal Classification for Crohn's disease. L1: ileal; L2: colonic; L3: ileocolonic.

  • b

    Montreal Classification for ulcerative colitis. E1: ulcerative proctitis; E2: left-sided UC (distal UC); E3: extensive UC (pancolitis).

UC2F130 25E2
UC4M125Methotrexate, Mesalamine10E2
UC5M163Mesalamine Beclometasone5E2
UC7F130 25E3
UC9F103Mesalamine Beclometasone45E3

The study was approved by the Ethics Committee of the University Hospital Umberto I where patients were admitted. For each patient informed consent from parents was obtained; in addition, children older than 12 years gave an informed consent through an appropriate-for-age form.

Biopsy Treatment

In addition to specimens for routine histology, three biopsies were taken from each subject enrolled in the study. For bacterial isolation, biopsies were immediately incubated in Dulbecco's Modified Eagle's Medium (DMEM; Invitrogen, Carlsbad, CA) supplemented with 100 μg/mL gentamicin (Invitrogen) for 1 hour, washed three times in phosphate-buffered saline (PBS), and lysed in 1% Triton-X-100 in PBS for 15 minutes. Aliquots were cultured on MacConkey agar plates at 37°C overnight and lactose-fermenting colonies were enumerated and propagated in Luria Bertani (LB) broth for 4 hours at 37°C under aerobic conditions. Individual clones were stored in 50% glycerol at −80°C until further use. For protein analysis, samples were immediately snap-frozen in liquid nitrogen. For organ cultures, specimens were immediately placed in Trowell T8 medium and treated as described below. For immunohistochemistry or ultrastructural examinations, tissue specimens were immediately placed in fixative and processed as described below.

Cell Cultures

HEp2 (human epidermoid carcinoma, larynx), Caco2 (human colorectal adenocarcinoma), and RAW 264.7 (mouse leukemic monocyte macrophage) cell lines were obtained from the American Type Culture Collection (ATCC, Rockville, MD). HEp2 cells were grown at 37°C in a humidified atmosphere with 5% CO2 in Minimal Essential Medium (MEM, Gibco, Grand Island, NY) containing 1.2 g/L NaHCO3 and supplemented with 5% inactivated fetal calf serum (FCS, Euroclone) and 2 mM L-glutamine.

Caco2 cells were grown at 37°C in a humidified atmosphere with 5% CO2 in DMEM (Gibco) supplemented with 10% fetal calf serum (FCS) and 2 mM L-glutamine.

RAW 264.7 cells were grown at 37°C in a humidified atmosphere with 5% CO2 in RPMI 1640 medium (Gibco), supplemented with 10% FCS and 2 mM L-glutamine.

Bacterial Adhesion and Invasion Assay

The ability of the E. coli strains to adhere to HEp2 cells was assessed as previously described.21 Cells were grown on 24-well plates to confluence and infected with each E. coli strain at a multiplicity of infection (MOI) of ≈10:1 for 3 hours at 37°C. To quantify the adherence of E. coli strains, infected cells were washed twice in PBS and lysed for 10 minutes with 200 μL of 0.1% Triton X-100 in PBS buffer. Adherent bacteria were recovered and plated on LB agar plates. The latter were incubated at 37°C overnight and then colonies were counted for statistical analysis. To obtain an accurate count of adherent bacteria, experiments on bacterial invasivity were performed simultaneously as those on bacterial adhesivity and, after, the number of the former was subtracted from the number of the latter.

To assess the invasiveness of E. coli strains, HEp2 cells were infected as above. After 3 hours incubation at 37°C and 5% CO2, cells were washed twice in sterile PBS and then incubated in MEM with 100 μg/mL gentamicin for 1 hour to kill extracellular bacteria. Cells were washed twice in sterile PBS, lysed in 0.1% Triton X-100 in H2O, plated on LB agar plates, and incubated at 37°C. Colonies were counted the following day as above.

Bacterial Strains

The adherent-invasive E. coli strains utilized in this study were LF82 (ileal Crohn's strain, kindly provided by Arlette Darfeuille-Michaud, Clermont Université, Université d'Auvergne, Clermont-Ferrand, France), EC10, isolated from an ileal mucosal specimen of a CD patient, and EC15, isolated from a colonic mucosal specimen of a UC patient. Bacteria were cultured on Tryptone Soya Broth (TSB, Oxoid) at 37°C. Before infecting Caco2 cells, bacteria were washed and resuspended in DMEM at a concentration of 2 × 107 CFU/mL.

Bacterial Survival and Replication in Macrophages

Bacterial uptake, survival, and replication were measured by the gentamicin protection assay according to Glasser et al.14 Before infection, RAW 264.7 cells were washed twice in PBS and the medium was replaced with 1 mL of RPMI 1640 supplemented with 10% heat-inactivated FCS. Cells were infected at an MOI of 10 bacteria per macrophage. After 2 hours of incubation at 37°C in 5% CO2, infected macrophages were washed twice in PBS, and fresh cell culture medium, containing 100 μg of gentamicin/mL, was added to kill extracellular bacteria. After 1 hour the medium was removed and fresh medium, containing 20 μg of gentamicin/mL, was added for longer postinfection periods. To determine the number of internalized bacteria, cells were washed in PBS and 0.5 mL of 1% Triton X-100 (Sigma Chemical Company, St. Louis, MO); deionized water was placed in each well for 5 minutes to lyse the eukaryotic cells. Triton X-100 at the concentration used did not affect bacterial viability for at least 30 minutes. Samples were then removed, diluted, and plated onto Mueller-Hinton agar plates to determine the number of CFU recovered from lysed monolayers. The number of gentamicin-surviving bacteria was determined after antibiotic treatment for 1, 4, and, 24 hours.

Bacterial Characterization

Api-20E Test

All bacterial strains thought to be invasive were cultured in Api-20E test strips (Biomerieux, Durham, NC) under aerobic and anaerobic conditions according to the manufacturer's specifications.

Phylogenetic PCR Grouping

Phylogenetic groups of E. coli strains were identified using multiplex polymerase chain reaction (PCR) as described by Clermont et al.22

Identification of Pathogenicity and Adhesion-associated Genes by PCR

Invasive bacteria identified as E. coli were screened by PCR for three pathogenicity genes associated with virulent E. coli strains: intimin,23 enteroaggregative factor and verotoxin,24 and for a fimbrial adhesin gene. The DNA to be amplified was extracted from whole-cell suspensions using Qiagen DNA extraction kits (Hilden, Germany). Appropriate bacteria, encoding each virulence gene, were included as PCR-positive controls. Amplification products underwent gel electrophoresis in 2% agarose, followed by ethidium bromide staining.

Real-time PCR

Expression of CEACAM6, IL-8, and TNF-α was detected by real-time PCR. Primers were designed to nonredundant sequences using Primer Express v. 3.0 (Applied Biosystems, Foster City, CA). Total RNA (1 μg) was reverse-transcribed to cDNA by a High Capacity cDNA Reverse Transcription Kit (Applied Biosystems). Real-time PCR amplification was done with an ABI PRISM 7300 Sequence Detection System using the SYBR Green kit (Applied Biosystems). Primers were: CEACAM6: forward 5′-GAAATACAGAACCCAGCGAGTGC-3′; reverse 5′-CAGTGATGTTGGGGATAAAGAGC-3′; TNF-α: forward 5′ TCTGGCCCAGGCAGTCAGATC-3′; reverse 5′-CAGTGATGTTGGGGATAAAGAGC-3′; IL-8: forward 5′-ATG ACT TCC AAG CTG GCC GTG GCT-3′; reverse 5′-TCT CAG CCC TCT TCA AAA ACT TCT C-3′. Relative transcript levels were determined using β-actin as endogenous control gene; primers used: 5′-CCTGGCACCCAGCACAAT-3′ and 5′-GCCGATCCACACGGAGTACT-3′.


Immunohistochemical analysis of inflamed and normal epithelial tissue samples was performed on paraffin sections (4-μm thick). Before incubation in 0.3% H2O2 in methanol for 30 minutes, sections were dewaxed, rehydrated, and pretreated by incubation in sodium citrate buffer (pH 6.0) for 8–10 minutes at 37°C. The antihuman CEACAM6 monoclonal antibody 9A6 (1:25 dilution; Genovac, Freiburg, Germany) was used. Sections were incubated overnight at 4°C. Detection was carried out with the Vectastain Elite ABC Kit (Vector Laboratories, Burlingame, CA) by using mouse biotinylated secondary antibodies for 1 hour at room temperature. After incubation with avidin-biotin immunoperoxidase, immunohistochemical staining was visualized with 3-amino-9-ethylcarbazole (Vector Novared kit; Vector Laboratories) and slides were counterstained with hematoxylin.

Protein Detection

Snap-frozen biopsy specimens were homogenized in ice-cold lysis buffer. Equal amounts of proteins were run on sodium dodecyl sulfate-polyacrylamide gel electrophoresis under reducing conditions. Proteins were transferred to a polyvinylidene difluoride (PVDF) membrane (Amersham, Little Chalfont, UK). Antihuman CEACAM6 polyclonal antibodies (Abcam, Cambridge, MA) was used as primary antibody; goat antirabbit antibody conjugated to horseradish peroxidase (HRP) (Santa Cruz Biotechnology, Santa Cruz, CA) as secondary antibody. Specific signals were detected using ECL reagents (Amersham, Biosciences Europe, Freiburg, Germany) for chemiluminescence. The anti-β-actin monoclonal antibody (Sigma) was used to control the equivalence of protein loading for total extracts. Densitometric analysis of the blots was performed with a GS-700 densitometer (Bio-Rad, Hercules, CA) using the software Quantity One (Bio-Rad).

Cytokine Induction

Proinflammatory cytokine induction was performed by treating Caco2 cells at 70% of confluence with 10 ng/mL of TNF-α and 1000 U/mL of interferon gamma (INF-γ) (Peprotech, Inalco, Milan, Italy) for different times.

Organ Cultures

Mucosal biopsy specimens were placed on iron grids with the mucosal face upward in the central well of an organ culture dish (Falcon, Becton Dickinson, Lincoln Park, NJ) containing Trowell T8 medium and NCTC-135 medium (ratio 3:1) (Biowest, Miami, FL), supplemented with L-glutamine (2 mM), penicillin (100 U/mL), streptomycin (100 mg/mL), and gentamicin (100 μg/mL) (Life Technologies-Gibco BRL, Milan, Italy).

The inflammatory process trigger was performed by adding to the medium 40 ng/mL of TNF-α (Sigma) and 1 U/mL of INF-γ (Sigma). Infections were performed by adding 106 CFU/mL of bacteria. Dishes were then placed in a modular incubator chamber (MP Biomedicals, Aurora, OH) at high oxygen saturation (95%) and incubated at 37°C. After 6 and 24 hours, biopsies were collected and total RNA was extracted for reverse-transcription PCR (RT-PCR) analysis.

Transepithelial Electric Resistance (TEER) Assay

Caco2 cells were grown at 37°C in 5% CO2 on polyethylene terephthalate membrane inserts, pore size 0.4 mm (Falcon, Becton Dickinson) in DMEM supplemented with 10% FCS. TEER values were measured using a Millicell-ERS voltohmmeter (Millipore, Billerica, MA) according to the manufacturer's instructions. Electrodes were sterilized by immersion in 70% ethanol, then they were washed in sterile PBS prior to use. Calculations for Ωcm2 were made by subtracting values of blank inserts from all samples and multiplying by the area seeded with cells. Subtracting filter resistance, epithelial monolayers were grown to an average TEER of about 500 Ωcm2 tissue monolayer. Cells were infected with EC10 or EC15 or the reference AIEC strain LF82, applied to the apical transwell compartment, at MOI 10:1 at 37°C in 5% CO2. Changes in TEER over time were determined every 2 hours and triplicate TEER readings from three independent experiments were normalized, averaged, and expressed as percent change.

Transmission Electron Microscopy (TEM)

Bioptic specimens were gathered from inflamed as well as noninflamed segments and immediately fixed in 2.5% glutaraldehyde, in 0.1 M sodium cacodylate buffer, pH 7.2. Samples were washed in cacodylate buffer and postfixed with 1% osmium tetroxide in the same buffer for 1 hour at room temperature. Fixed specimens were washed, dehydrated through a graded series of ethanol solutions 30%–100% ethanol, and embedded in Agar 100 (Agar Scientific, UK). Ultrathin sections obtained by a MT-2B Ultramicrotome (LKB, Pharmacia) were stained with uranyl acetate and lead citrate and examined with a Philips 208S electron microscope. For bacteria–cell interaction visualization, Caco2 cells, seeded on Transwell filters, were infected with E. coli strains as described above and processed for TEM.

Statistical Analysis

All experiments were repeated three times. Data are given as mean ± SD. The Kolmogorov–Smirnov test showed significant departures from the normal distribution for some of the analyzed parameters. Therefore, comparison of the groups was performed with a Mann–Whitney U-test (significance taken as P < 0.05).


  1. Top of page
  2. Abstract

Epithelium Morphology Is Altered in the Inflamed tissue of IBD Children

To investigate the impact of bacteria on the inflamed intestinal epithelium, an ultrastructural analysis by TEM was performed on biopsies from controls and IBD pediatric patients. In the latter, specimens were collected close to the macroscopically involved areas of the ileum and colon. Results showed microvilli fused and altered in size, shape, and orientation as well as changes in tight junction (TJ) distance in the inflamed tissue; UC patients also showed an absence of microvilli in some areas and a reduced number of goblet cells. A reduction of TJ strands and a higher electrodense cytoplasm in enterocytes were also seen, as compared with healthy tissue of controls (Fig. 1A–H). The altered morphology of the inflamed tissue is in agreement with an increased microbial antigen entry and a diminished bacterial clearance in the intestinal mucosa.

thumbnail image

Figure 1. Morphological alterations of intestinal inflamed tissue from IBD pediatric patients as shown by TEM. Healthy intestinal epithelium from a control subject. Both microvilli (MV) and tight junctions are normal (white arrows) in thickness and height. GC: goblet cell (A,B). Dilated intercellular spaces (asterisks) of TJ complexes (white arrows), microvilli altered in size and shape and frequently fused together (black arrows) in CD ileal specimens (C–E). Extensive microvillar and TJ structural alterations (white arrows), deep cytoplasmic derangement in UC colonic specimens (F–H).

Download figure to PowerPoint

EC10 and EC15 Are Adhesive-invasive Bacteria Isolated from a CD and a UC Patient

The gentamicin protection assay was performed in bioptic specimens from 27 IBD patients (10 UC and 17 CD) and 23 controls to identify and isolate invasive bacterial strains. We analyzed 50 biopsies (10 UC, 17 CD, and 23 controls) and observed an average low colony density. Bacterial growth, following overnight incubation, was detected in tissue specimens from 7 out of 17 CD and 7 out of 10 UC patients, plus from 19 out of 23 controls. Eighty-five lactose-fermenting isolates were obtained following propagation under aerobic conditions overnight. Among these we identified two strains showing a significant invasion capability in HEp2, named EC10 and EC15, from a CD and a UC patient, respectively. The invasivity indexes were 1.4% (EC10) and 1.0% (EC15), whereas that of the prototype LF82 (positive control) was 1.29%. Both strains were identified as belonging to the E. coli group. They were negative for the principal virulence genes associated with pathogenic E. coli, such as intimin, verotoxin (shiga-like-toxin), and enteroaggregative factor gene, while positive for the adhesive fimbrial factor H gene (FimH). In the same bioptic repertoire we isolated three other E. coli strains, named EC18 (from a CD patient), EC23 (from an UC patient), and EC8 (from a control), showing adhesivity but not invasivity to cells, and used as controls.

EC10 and EC15 Can Survive Inside Macrophages

A main feature of AIEC strains is known to be the capability to survive and replicate extensively within murine macrophages.14 To verify whether EC10 and EC15 strains were able to resist the phagocytosis, the gentamicin protection assay was performed in RAW 264.7 macrophage cell line infected with each strain. LF82 and the adhesive noninvasive EC8 strain were used as positive and negative controls, respectively. Results showed that both strains could survive and replicate within macrophages (Fig. 2). In addition, EC10 and EC15 did not induce any cell necrosis or apoptosis at 24 hours postinfection: this is a feature characterizing AIEC bacteria and distinguishing them from other invasive pathogens.25

thumbnail image

Figure 2. EC10 and EC15 E. coli strains survival and replication within macrophages. Bacterial survival and replication are expressed as intracellular bacteria counts after gentamicin treatment. AIEC LF82 is used as a positive control, while adhesive bacterial strain, EC8, as a negative control. P ≤ 0.05 (*) considered statistically significant. Data are expressed as the mean of three experiments ± SD.

Download figure to PowerPoint

CEACAM6 Protein Expression Is Increased in the Ileum of IBD Patients

The ileal mucosa of CD patients is supposed to be abnormally colonized by bacterial strains able to adhere and invade intestinal epithelial cells. These bacteria specifically recognize the CEACAM6 surface receptor. Therefore, by western blot we analyzed CEACAM6 protein expression in the ileal bioptic specimens of 24 CD and 10 UC pediatric patients and compared results with those of 23 healthy controls. There was a significant (P < 0.05) increase of CEACAM6 expression in the inflamed ileum of CD patients as compared with controls, as well as in uninflamed areas of the ileum of CD and UC patients (P < 0.05) (Fig. 3A,B). Even if invasive bacteria are known to be typically more associated with the ileal mucosa, we also investigated CEACAM6 expression level in the colon of patients. The result did not show any difference between inflamed colonic mucosa of CD and UC children as compared with controls (data not shown).

thumbnail image

Figure 3. CEACAM6 protein expression in the ileum of IBD patients and healthy controls. Bioptic specimens were obtained from the inflamed and uninflamed terminal ileum of CD patients and from the uninvolved ileum of UC patients. CEACAM6 protein expression evaluated by western blot; expression level is normalized against the housekeeping gene, β-actin (A). Densitometric analysis of the bands (B). P ≤ 0.05 (*) considered statistically significant. Data are expressed as the mean of three experiments ± SD. Apical localization of CEACAM6 in the inflamed ileal mucosa of patients (C) as compared with the healthy mucosa of controls (D), as evaluated by immunohistochemistry. [Color figure can be viewed in the online issue, which is available at]

Download figure to PowerPoint

CEACAM6 Is Highly Concentrated at the Surface of the Intestinal Inflamed Epithelium

To assess whether CEACAM6 is localized on the surface of inflamed epithelial cells or dispersed throughout, immunohistochemistry analysis on inflamed and healthy tissues was performed. Thus, intestinal mucosal sections were set up from the inflamed bioptic tissue of a CD patient as well as from the healthy tissue of a control, in which an inflammatory insult was given through the incubation with proinflammatory cytokines (TNF-α and INF-γ). The results showed that the protein was highly concentrated at the surface of the inflamed epithelium, in agreement with the hypothesis that it works as a surface bacterial receptor (Fig. 3C,D).

CEACAM6, TNF-α, and IL-8 mRNA Expression Is Upregulated in Presence of EC10 and EC15 Strains

To assess whether EC10 or EC15 strains were able to alter CEACAM6 and proinflammatory cytokine mRNA expression, a confluent monolayer of Caco2 cells, as a model of intestinal epithelium, was infected for 3 hours with each strain. LF82 was used as a positive control. The results showed that both EC10 and EC15 were able to induce a significant increase of CEACAM6, TNF-α, and IL-8 gene expression (P < 0.01 and P < 0.05, respectively), similar to that induced by LF82 (Fig. 4A–C). Moreover, to evaluate a possible synergistic effect between proinflammatory cytokines and bacteria in upregulating CEACAM6 mRNA expression, cells were treated first with INF-γ for 48 hours and then with EC10 or EC15 for 3 hours. The results were compared with those obtained from cells treated with INF-γ, EC10, or EC15, separately. A significant increase (P < 0.01) was seen in CEACAM6 mRNA expression in cells incubated with the mixed treatment (Fig. 4A).

thumbnail image

Figure 4. CEACAM6, TNF-α, and IL-8 mRNA expression in vitro and in cultured mucosal explants after EC10 and EC15 infection. Caco2 cells were infected with invasive strains, EC10 and EC15, or with a mixed treatment (INF-γ + each bacterial strain); CEACAM6 mRNA level was evaluated by real-time PCR and compared with uninfected cells or with cells incubated with INF-γ alone. LF82 was used as a positive control (A). TNF-α (B) and IL-8 (C) mRNA levels were evaluated after infection with EC10, EC15, or LF82. Caco2 cells were infected with adhesive strains, EC8, EC18, and EC23. CEACAM6 (D), TNF-α (E), and IL-8 (F) mRNA expression was evaluated. Ileal mucosa explants of two healthy controls were cultured and infected for 24 hours with EC10 or EC15. CEACAM6, TNF-α, and IL-8 mRNA expression was analyzed. LF82 was used as a positive control, while the untreated tissue was used as a negative control (G). P ≤ 0.05 (*) considered statistically significant. Bars represent the mean of three experiments ± SD. Fold of activation is calculated as the ratio between expression level in treated and untreated cells.

Download figure to PowerPoint

In order to verify whether the increased gene expression levels were related to the invasiveness of bacterial strains, cells were incubated in the presence of three noninvasive adhesive bacterial strains, EC8, EC18, and EC23. The results showed a significant increase of TNF-α and IL-8 (P < 0.05), but not of CEACAM6 gene expression (Fig. 4D–F).

Finally, the altered CEACAM6, TNF-α, and IL-8 mRNA expression induced by EC10 and EC15 was also analyzed in the ileal mucosa explants taken from two control subjects after 24 hours of incubation. LF82 was used as a positive control and untreated tissues as negative controls. Similar to in vitro experiments, the results showed a significant upregulation of all genes (P < 0.05) (Fig. 4G).

CEACAM6 mRNA Expression Is Inducible by TNF-α and INF-γ

To assess whether CEACAM6 gene expression was inducible by inflammatory stimuli, Caco2 cells were incubated for 6, 24, and 48 hours in the presence of the proinflammatory cytokines, TNF-α and INF-γ. We showed that mRNA expression significantly increased (P < 0.01) after 6 and 24 hours of treatment with TNF-α and INF-γ, respectively (Fig. 5A). Furthermore, to better emulate the in vivo conditions we repeated the same experiment using bioptic specimens, taken from two control subjects, which were cultured for 6 and 24 hours as above. A significant upregulation of CEACAM6 mRNA (P < 0.01) was observed after 6 hours of incubation, confirming the inducibility of CEACAM6 following inflammation (Fig. 5B).

thumbnail image

Figure 5. CEACAM6 is inducible in vitro and in cultured mucosa by proinflammatory cytokines. Caco2 cells were treated with TNF-α and INF-γ for 6, 24, and 48 hours; CEACAM6 mRNA expression was analyzed by real-time PCR (A). The same experiment was repeated in mucosa explants cultured for 6 and 24 hours (B). Fold of activation is calculated as the ratio between CEACAM6 expression level in treated and untreated cells (A). Data are expressed as the mean of three experiments ± SD. P ≤ 0.01 (**) considered statistically significant.

Download figure to PowerPoint

Epithelial Barrier Function Is Altered in Caco2 Cells Infected with EC10 and EC15 Strains

To investigate the effect of EC10 and EC15 on the intestinal epithelial barrier function, Caco2 cells were grown to an average TEER of 485.55 Ω/cm2 ± 67.11 and infected at MOI 10:1 for 24 hours with LF82, EC10, or EC15 strains. The results showed that EC15 infection caused a 55% reduction in TEER, as compared with uninfected cells, while EC10 and LF82 caused a complete TEER suppression (Fig. 6A).

thumbnail image

Figure 6. EC10 and EC15 reduce the epithelial barrier function and cause morphological alterations. Caco2 cells were grown to an average TEER of 485.55 Ω/cm2 ± 67.11 and infected with EC10, EC15, or LF82, as a positive control. Data from three independent experiments were normalized and expressed as percent change over baseline (A). TEM analysis (B) was performed in uninfected Caco2 cells (A) or in the presence of EC10 strain (B–D).

Download figure to PowerPoint

To confirm TEER data and to deeply investigate the bacteria–host interaction, a TEM analysis was performed on Caco2 cells, grown on membrane inserts and incubated for 24 hours with EC10 strain (Fig. 6B). Figure 6A shows uninfected cells, while Figure 6B–D show cells in the presence of bacteria. It was shown that EC10 adhered well to cells (Fig. 6B,C), attaching to microvilli by fimbriae (Fig. 6C, insert). Epithelial tight junctions of infected cells were strongly damaged by bacteria (Fig. 6C,D, white arrows), as compared with those of uninfected (Fig. 6A). This behavior is similar to that reported in the literature for LF82.


  1. Top of page
  2. Abstract

Although very little is known about the identification of bacterial species and their interaction with specific immunology cell repertoire in patients with IBD, it seems that E. coli, with specific adherent, invasive, and survival features may play a predominant role in the pathogenesis of the disease. Indeed, while E. coli bacteria are commonly found in the lumen of the gut and are mostly considered nonharmful, nevertheless, some strains of them, while not containing virulence factors, have a capacity to adhere to and invade epithelial cells in vitro and, therefore, should be considered a separate category of E. coli, which may cause intestinal diseases in humans.24, 26–31

Involvement of luminal bacteria in the pathogenesis of IBD is supported by a plethora of studies, most of them carried out in adult disease, while very few in pediatric disease.32–35 On the other hand, the distinction between adult and pediatric disease is a highly topical subject, given the recently reported rise in the incidence of IBD in pediatric age36 and the currently held view that children with IBD are a distinctive population with specificities that justify a highly skilled and specialized approach and a special management.

To contribute to the improvement of knowledge in the field of pediatric IBD, we aimed at identifying and characterizing invasive E. coli strains in a population of children with CD and UC and in an age-matched control population. For this purpose we analyzed a total of 50 bioptic specimens from 17 CD and 10 UC patients plus 23 controls and identified two adhesive-invasive E. coli strains, EC10 and EC15, from 85 lactose-fermenting isolates. The two strains were found in one CD patient and the other in a UC patient, while none in the control group. At first glance, the contribution of invasive strains on the total microbiota seems to be lower in children than in adults, where it is reported to be around 29%–36% in CD patients, 12%–19% in UC patients and 3%–9% in controls.12, 24, 27 However, how relevant the impact of AIEC on the total microbiota or how often they affect people with IBD is still controversial. Raso et al26 reported the identification of AIEC strains in only three out of eight CD patients investigated, while no patient with UC nor any control.

As reported by Darfeuille-Michaud et al,13 AIEC bacteria must be able to colonize the intestinal mucosa by adhering and invading intestinal epithelial cells, survive and replicate intracellularly without inducing host cell death, and induce the release of high amounts of TNF-α. These virulence properties characterize them as possible pathogens with the potential ability to induce persistent intestinal inflammation.

In our study we analyzed our strains to see how much they adhered to features previously described. Indeed, EC10 and EC15 were able to adhere and invade epithelial cells at levels comparable to that of the well-known AIEC strain, LF82, and survive in cultured macrophages, without inducing any cell necrosis or apoptosis at 24 hours postinfection. In addition, they were negative for virulence factors, but positive for the fimH gene, which, encoding the terminal subunit of type 1 pili (fimbriae), is a marker of cell adhesivity. It is also accepted that AIEC adhesion is dependent on CEACAM6 expression, which is situated on the apical surface of ileal epithelial cells,16, 30 and that this expression in intestinal epithelial cells is upregulated by AIEC infection.16 Therefore, we first analyzed CEACAM6 protein expression in inflamed and uninflamed areas of ileum and colon of patients and controls. We show that CEACAM6 is strongly upregulated in both inflamed and uninflamed ileum of CD and UC, but not in the colon. This was expected, as the receptor is known to be more typically associated with ileal than colonic mucosa. We also observed, by immunohistochemistry, that a large amount of the protein was localized on the surface of ileal tissue, in agreement with the notion that it works as a surface receptor. Interestingly, CEACAM6 expression was also increased in the uninflamed ileal mucosa of patients: as suggested by other authors,37 this may indicate that some individuals are genetically predisposed to express high levels of the protein and could also differently modulate (as compared with people expressing low levels) their bacterial colonization, favoring more harmful species and, thus, the development of the disease.

We also showed in vitro and in cultured organ explants that CEACAM6 is inducible by TNF-α and INF-γ. Moreover, we showed that the presence of EC10 and EC15 could upregulate in vitro and ex vivo CEACAM6 as well as TNF-α and IL-8 gene expression and that this effect was emphasized by incubating cells simultaneously with bacterial strains and INF-γ; this seems to suggest a synergistic action between bacteria and proinflammatory cytokines in inducing inflammation. To verify that the observed effect was due to the bacterial invasive properties, rather than adhesive ones, experiments were repeated by incubating cells in the presence of the three noninvasive adhesive bacterial strains, EC8, EC18, and EC23, that we had previously isolated and characterized: it was shown that only TNF-α and IL-8, but not CEACAM6, expression was increased. These findings confirm, in agreement with other studies,38 the mechanism whereby, in inflamed mucosa, CEACAM6 expression is increased by proinflammatory cytokines: this facilitates the entrance of AIEC strains, which, in turn, upregulate the TNF-α production, creating a loop of colonization and inflammation leading to chronic disease with the secretion of other proinflammatory cytokines.

Given the impact of EC10- and EC15-induced effects on the gut mucosal immune system, we also wished to test their role in the derangement of the intestinal epithelial barrier. It is known that increased intestinal permeability is a feature of IBD patients; moreover, it was reported that CD-isolated LF82 bacteria were able to disrupt apical junctional complexes in differentiated enterocytes.29, 31 In this study we showed, by TEM analysis, that the increased microbial antigen entry and the diminished bacterial clearance caused morphological and structural alterations in the inflamed intestinal epithelium of IBD children. These alterations mainly consisted of altered microvilli shape and size and changes in TJ distance. In addition, we demonstrated that EC10 could affect microvilli and TJ morphology. In detail, bacterial fimbriae were seen to contact microvilli, showing a behavior typical of AIEC strains, which use long polar fimbriae (LPF) to better interact with susceptible cells.39, 40 Finally, in vitro TEER analysis revealed that both strains, EC10 and EC15, could strongly reduce the epithelial barrier function to levels comparable to those of LF82.

In conclusion, this is the first study, to our knowledge, to show the presence of adhesive-invasive bacterial strains in a pediatric population with IBD. Due to their behavior in terms of adhesivity and invasiveness criteria previously described,12 they are thought to belong to the AIEC spectrum, suggesting a role in the pathogenesis of the pediatric disease. This is more intriguing if these bacteria are viewed as target organisms whose elimination could reduce the intestinal inflammatory process and disease progression.


  1. Top of page
  2. Abstract