High carriage of adherent invasive E. coli in wildlife and healthy individuals

The ECOR collection contains 61 strains of fecal origin. Of these, 29 were from human and 32 from animals (Table 1). An invasion assay was performed on these strains using I-407 cells based on gentamicin protection, according to previously described methods [6, 15, 21, 22]. All of the strains were first tested for their susceptibility to gentamicin using disc diffusion method as only strains susceptible to gentamicin can be tested. All the strains were found to be susceptible to gentamicin, determined as a zone diameter superior to 17 mm, according to EUCAST recommendations.

The invasion levels of the noninvasive references E. coli strain K-12 C600 and E. coli Nissle were 0.0561 ± 0.0373% and 0.0254 ± 0.0265% respectively. Reference strain LF82, included in all of the assays as a positive invasive control, gave a mean invasion level of 1.0101 ± 0.9880%. Another positive invasive control E. coli NRG857c showed a mean invasion of 0.1683 ± 0.0368%. The invasion levels were done in triplicate.

Among strains belonging to the ECOR collection, 18 strains (30%) were classified as invasive based on this assay, with invasion rates ranging between 0.10 and 0.64%. Eight of these were of human origin (28%), four belonged to the phylogroup A (ECOR 1, 2, 9 and 15), one to phylogroup B2 (ECOR 63), two to phylogroup D (ECOR 35 and ECOR 36) and one to the phylogroup E (ECOR 43). Ten were of animal origin (31%), two belonging to the phylogroup A (ECOR 7 and 23), four to the phylogroup B1 (ECOR 45, 67, 69 and 70), two to the phylogroup B2 (ECOR 52 and 57) and two to the phylogroup D (ECOR 44 and 46) (Fig. 1).

During the invasion assay, for five of the strains, an unexpected pathogenic feature was discovered: the ability to detach intestinal cells from the microplate. All of these strains were of human origin, one belonging to the phylogroup A (ECOR 6) and the 4 others to the phylogroup B2 (ECOR 51, 53, 54 and 59) (Table 3). Invasion scores of these cell-detaching strains were established but were not interpretable due to their capacity to detach adherent cells from the microplate. The actual concentration of the intracellular bacteria was therefore skewed. Further investigations are necessary to define their propensity to invade.

The strains revealed as invasive were tested for their ability to adhere. All the invasive strains from the ECOR collection were adherent, with an adhesion index ranging between 67 and 334, all showed an adhesion index superior to 1 bacteria per cell (Table 3).

The 18 fecal strains previously classified as adherent and invasive were genotyped. The presence of 12 selected virulence genes associated with diarrheagenic E. coli pathotype was assessed. This included common virulence factors present in diarrheagenic E. coli pathotypes EHEC (stx-1, stx-2, eae and ehxA), EIEC (ipaC), ETEC (estA, elt and tia), EPEC (eae, ehxA, bfpA), EAEC (aggR and pCVD432) and DAEC (afaD) (Additional file 1: Table S1). Three strains were positive for the tia gene (ECOR 57) and tia and afaD genes (ECOR 63 and 70). These strains are not affiliated as AIEC pathovars and belong to ETEC/DAEC pathovars. The other 15 invasive strains (ECOR 1, 2, 9, 15, 35, 36, 43, 7, 23, 45, 67, 69, 52, 44, 46) were negative for the tested virulence genes excluding their affiliation to the previously described pathotypes (Table 2).

The cell-detaching strains were also tested for their expression of virulence genes associated of enteropathogenic strains. The ECOR 51 strain was shown to express the afaD virulence gene, showing that this strain is classified as a DAEC. The other cell-detaching strains were negative for the tested genes (Table 2).

The two other strains (ECOR 46 and 7) were not able to replicate inside macrophages with an index of 70.42 and 74.51% respectively (Table 3). These strains were thus not affiliated to the AIEC pathovar.

Bacteria from the ECOR collection were kindly provided by Laurent Debarbieux (Institut Pasteur Paris) (Table 1). E. coli referent strains belonging to different pathogenic groups respectively ETEC (ATCC 35401), EHEC (ATCC 43895), EPEC (ATCC 43887), EAEC (ATCC 33780), EIEC (ATCC 43893), DAEC (AfaA30, kindly provided by Pr Eric Oswald, IRSD, INSERM UMR 1220, Toulouse, France) were used as positive controls in PCR genotyping; while E. coli K12 was used as a negative control. E. coli ATCC 25922 was used as a positive control for the gentamicin susceptibility assay. Bacteria were routinely grown on Luria–Bertani (LB) agar plates or in LB broth at 37 °C.

All the strains were assessed for their susceptibility to gentamicin by applying the in vitro diffusion method. Agar plates were inoculated with a 0.5 McFarland bacterial suspension. Then, gentamicin discs (Oxoid, 10 μg) were placed in the inoculated media and incubated at 37 °C for 24 h. The susceptibility of the bacteria to gentamicin was determined by measurement of the diameters of the inhibition zones surrounding the discs in millimeters according EUCAST values. A control standard strain (E. coli ATCC 25922) for which the susceptibility to antimicrobial agents is known is included in the test to validate the assay [30].

Intestine-407 cell lines (I-407, ATCC accession number CCL-6™) were used to determine bacterial invasiveness. I-407 cells were maintained in Basal medium Eagle (BME, Thermofisher, France) with 10% heat-inactivated fetal calf serum (Euroclone Italy) supplemented with 1% Glutamax and 1% penicillin/streptomycin (Thermofisher, France), in 5% CO₂ atmosphere at 37 °C.

For AIEC identification, adhesion and invasion assays, based on gentamicin protection, were performed according to a previously described method [6, 15, 21, 22]. 24-well plates containing 4 × 10⁵ cells/well incubated for 20 h without antibiotics were infected with the E. coli strains at a multiplicity of infection of 10 during 3 h of incubation time. For the bacterial adhesion assays, the cell monolayers were washed with phosphate-buffered saline and then lysed with 1% Triton X-100. Adherent bacteria were quantified by plating them on LB agar. Plating was performed over a maximum period of 30 min in order to avoid bacterial lysis by Triton X-100. Adherence ability (I_ADH) was determined by calculating the mean number of bacteria per cell. For the bacterial invasion assays, cells were washed twice with phosphate buffer saline (PBS), and fresh cell culture medium containing 100 μg ml⁻¹ of gentamicin was added for 1 h to kill extracellular bacteria. Cells were then washed three times with PBS and 1 ml of 1% Triton X-100 was added in each well for 5 min to lyse the eukaryotic cells. After cell lysis, the number of intracellular bacteria was determined by plating. The invasive ability was expressed as the percentage of intracellular E. coli recovered from the initial inoculum, taken as 100%. An isolate was considered invasive when invasion was ≥ 0.1%. Each experiment was done in triplicate. The non-invasive strains of E. coli K12 and E. coli Nissle and the invasive AIEC LF82 and AIEC NRG857c strains were used as negative and positive control in each experiment.

The macrophage-like J774A.1 cell line (ATCC TIB-67) was used as a model in E. coli survival and replication assays. E. coli isolates with known adherence and invasion properties were checked for their ability to survive and replicate inside macrophages. Briefly, macrophages were seeded at 4 × 10⁵ cells per well in two 24-well plates and incubated for 20 h. After incubation, the medium was replaced with fresh medium and bacteria were seeded at a multiplicity of infection of 100. To promote internalization of bacteria by the macrophages, the samples were centrifuged at 1000 rpm for 10 min and incubated for an additional 10 min at 37 °C with 5% CO₂. Nonphagocytosed bacteria were killed with gentamicin. Intracellular bacteria were quantified in the same manner as described for the invasion assays after 1 and 24 h of infection. The results are expressed as the mean percentages of bacteria recovered at 1 and 24 h postinfection: I_REPL (%) = (CFU ml⁻¹ at 24 h/CFU ml⁻¹ at 1 h) × 100. Those strains with an I_INV of > 0.1 and an I_REPL of > 100% were classified as AIEC strains in the present study. All assays were performed in duplicate.

The presence of 12 virulence genes specifically associated with diarrheagenic E. coli pathotypes (EPEC, ETEC, EIEC, EHEC, DAEC, EAEC), and uropathogenic E. coli (UPEC) was analyzed (Additional file 1: Table S1) [8, 31]. E. coli reference strains were used as positive controls.

Adherent/invasive bacteria able to survive and to multiply within macrophage, without expressing virulence genes associated with diarrheagenic E. coli pathotypes and uropathogenic E. coli were considered as AIEC strains.