Salmonella enterica spp. enterica serovar Typhimurium N-15 (S. Typhimurium N-15) was isolated in 2007 from an infected person in Switzerland and obtained from the National Center for Enteropathogenic Bacteria (NENT, Luzern, Switzerland). It was routinely cultivated in tryptic soy broth (TSB, Difco, Basel Switzerland) at 37°C for 18 h.
E. coli L1000 wt, producing microcin B17 , was kindly provided by Hans-Dieter Grimmecke (Laves-Arzneimittel GmbH, Schötz, Switzerland). A mutant strain lacking microcin B17-phenotype (E. coli L1000 MccB17-) was also used . B. thermophilum RBL67, initially isolated from baby feces , was obtained from our culture collection.
Intestinal in vitro colonic fermentations
Intestinal colonic fermentations were performed as previously reported . In brief, two three-stage continuous in vitro fermentation models (F1 and F2) inoculated with the same immobilized child fecal microbiota were infected with S. Typhimurium N-15. These models were operated in parallel for 65 days to test and compare the effects of treatments with probiotic E. coli L1000 wt and MccB17-, followed by B. thermophilum RBL67, and prebiotic inulin, on gut microbiota composition, activity, probiotic growth and Salmonella colonization . Specific retention times (RT) and pH were applied to the three reactors of each model corresponding to the physiological conditions in child proximal (R1), transverse (R2) and distal (R3) colons: RT = 5 h and pH 5.7 for R1, RT = 10 h and pH 6.2 for R2, and RT = 10 h and pH 6.6 for R3, respectively [43, 44].
Continuous fermentations were divided into six consecutive experimental periods illustrated in Figure 1 and presented in detail by Zihler et al. . Briefly, the first model F1 used to test E. coli L1000 wt, included the following conditions: (1) system stabilization [Stab, 10 days], (2) S. Typhimurium N-15 beads addition to R1 to induce Salmonella infection [Sal, 9 days], (3) first E. coli L1000 wt beads addition to R1 [Ecol I, 14 days], (4) second E. coli L1000 wt beads addition to R3 [Ecol II, 8 days], (5) first B. thermophilum RBL67 beads addition to R1 [Bif, 11 days], and (6) second B. thermophilum RBL67 beads addition to R1 [Bif II, 10 days]. In the second model F2 E. coli L1000 wt was replaced by E. coli L1000 MccB17- to assess the effect of microcin B17 phenotype. Similar periods as F1 were tested except for the last period (6) during which prebiotic inulin was tested [Inulin, 10 days].
Effluents (13 ml) were collected daily from each reactor of the two models and processed within 1 h for the enumeration of S. Typhimurium N-15 (selective plating), quantification of main bacterial populations (real-time qPCR analyses), and metabolic analysis . Fresh effluents were also directly applied on intestinal HT29-MTX cells.
Salmonella enumeration by plate counts
Salmonella viable cell counts were measured during the last 3 days of each experimental period corresponding to pseudo-steady-state conditions. Effluent samples were serially diluted 10-fold in peptone water (0.1%, pH 7.0) and plated in duplicate on CHROMAgar™Salmonella (Becton Dickinson AG, Allschwil, Switzerland). Plates were incubated at 37°C for 48 h.
E. coli L1000 and B. thermophilum RBL67 enumeration by real-time qPCR analysis
E. coli L1000 and B. thermophilum RBL67 concentrations in reactor effluents were estimated by real-time qPCR analysis as described before . Mean copy numbers (MCN/ml) were calculated for the last 3 days of each experimental period of F1 and F2.
Short-chain fatty acids [SCFA: acetate (A), propionate (P) and butyrate (B)] concentrations in effluent samples were determined in duplicate by high-performance liquid chromatography (HPLC) analysis .
The human mucus-secreting intestinal colon cancer cell line HT29-MTX , obtained after long-term treatment of human carcinoma HT-29 cells with the anti-cancer drug methotrexate , was kindly provided by Dr. Thécla Lesuffleur (INSERM, Lille, France). Cells were routinely maintained at 37°C in a humidified incubator (10% CO2) in complete Dulbecco's Modified Eagle medium Glutamax (DMEM; Invitrogen AG, Basel, Switzerland) supplemented with 10% (V/V) fetal bovine serum (FBS; Invitrogen AG) and 1% (V/V) antibiotics (10'000 U/ml penicillin + 10'000 μg/ml streptomycin; Invitrogen AG). For invasion assays, cells were seeded in 24-well tissue culture plates (2 cm2 well-1; Bioswisstec AG, Schaffhausen, Switzerland) at a concentration of 4 × 104 cells per well and cultivated for 21 days to reach complete confluence and differentiation. The medium was replaced every 2 days and cell viability was determined by tryptan blue staining (0.1% (V/V) in 10 mM phosphate buffered saline (PBS), pH 7.3). DMEM without antibiotics was used for the last medium change before using the cells for invasion assays.
For transepithelial electrical resistance (TER) measurements, HT29-MTX cells were seeded in cell culture inserts with a 0.45 μm filter membrane and a 0.7 cm2 surface area (24-well culture plate, Millipore AG, Zug, Switzerland) at a concentration of 2.3 × 105 cells per insert and cultivated as described above.
A gentamicin-based assay, as described by Steele-Mortimer et al. (2008) but with some modifications, was performed to determine the capacity of Salmonella present in reactor effluents to invade HT29-MTX cells. Briefly, 1 ml effluents obtained during the last 3 days of each fermentation period from proximal (R1), transverse (R2) and distal (R3) colon reactors were applied directly in duplicate on cell layers of three consecutive passages and incubated at 37°C for 90 min. To kill non-invading bacteria, cell layers were washed twice with 250 μl PBS before adding 250 μl DMEM supplemented with 150 μg/ml gentamicin (Sigma-Aldrich Chemie GmbH, Buchs, Switzerland) per well followed by an additional incubation period for 60 min at 37°C. After a further washing step with PBS, 250 μl Trypsin-EDTA (1X, Invitrogen) were added followed by another incubation for 10 min. Finally, cells were disrupted by adding 250 μl 0.1% (V/V) Triton X-100 (Sigma) per well and incubating for 10 min before samples were collected for enumeration of invaded Salmonella. The same protocol but without gentamicin treatments was used for the determination of cell-associated Salmonella (accounting for both invasive and adherent bacteria). The number of adhered Salmonella was then calculated from the difference of cell-associated to invaded bacteria. Adhesion and invasion ratios were expressed as the percentage of adhered and invaded bacteria, respectively, related to the total number of Salmonella present in effluents. Invasion efficiency measured during different probiotic and prebiotic treatments was expressed as the percentage of invaded bacteria related to the number of cell-associated Salmonella.
The same protocol was used to measure the invasion efficiency of S. Typhimurium N-15 in pure culture when applied in artificial DMEM medium. Therefore, the pellet of an overnight culture of Salmonella obtained by centrifugation (8000 g, 5 min) was diluted in DMEM to reach a concentration of 1.0 × 107 cfu/ml. 125 μl of this bacterial suspension was added in duplicate to cell monolayers that corresponded to a Salmonella concentration (1.3 × 106 cfu/ml) measured in effluents from the two models during Sal periods.
Transepithelial electrical resistance (TER) measurements
TER measurements were performed to estimate the degree of cell monolayer's integrity loss that occurs during Salmonella infection due to disruption of tight junctions . To measure the epithelial integrity of HT29-MTX cells, 400 μl of effluent was applied directly to the apical compartment of PBS-washed HT29-MTX cell culture inserts that were prepared as previously described. TER measurements were performed before effluent application and after 1, 2, 3 and 24 h of incubation at 37°C. The resistance of cell layers was calculated by subtracting the intrinsic resistance of the filter insert alone from the total measured resistance (filter insert plus cell layer and effluents) and expressed as Ω per cm2 surface area. The same protocol was used to measure the influence of S. Typhimurium N-15 on TER of HT29-MTX cells in artificial DMEM medium as presented before.
Microscopic analysis of tight junctions
To visualize the effects of Salmonella infection on cell monolayer integrity before and during probiotic treatments, tight junctions and the nucleus of confluent HT29-MTX cells were fluorescently stained according to previous studies [35, 47].
Briefly, HT29-MTX cells were seeded at 9.6 × 104 cells/ml on a coverslip in a 6-well tissue culture plate and cultured to confluence before incubation with 1 ml of distal colon reactor (R3) effluents from the last day of different treatment periods of F1. DMEM-high glucose without Phenol red (Invitrogen AG, Basel, Switzerland) supplemented with 10% (V/V) fetal bovine serum (FBS; Invitrogen AG) and without antibiotics was used for the last medium change before invasion assays. After incubation of 1 ml effluent for 90 min, cells were washed thrice with PBS and fixed overnight in 1 ml per well of a chilled 4% (V/V) formaldehyde (Sigma-Aldrich Chemie GmbH, Buchs, Switzerland) in PBS solution. After a second washing step (3 times with PBS), cells were permeabilized by treating them with 200 μl of 0.1% Triton X-100 in PBS for 3 min at room temperature. After a third washing step (3 times with PBS), cells were treated with 1 ml of 3% (V/V) albumin bovine serum (BSA, Sigma-Aldrich Chemie GmbH) in PBS to prevent non-specific binding of fluorescent dyes. Tight junctions were stained for 40 min with 1 ml of a 1:200 PBS-diluted stock solution (0.1 mg/ml) of phalloidin-tetramethylrhodamine B isothiocyanate (phalloidin-TRITC, Sigma-Aldrich Chemie GmbH) in methanol, while nuclei were stained for 3 min with 1 ml of a 1:100 PBS-diluted stock solution (5 mg/ml) of 4', 6-diamidino-2-phenylindole (DAPI, Sigma-Aldrich Chemie GmbH) in ultrapure water. After a last washing step, coverslips were mounted inverted on a coverglas by applying one drop of the embedding media Glycergel (DakoCytomation; Glostrup, Denmark). Microscopic analyses were performed with a confocal laser scanning microscope (SP 2, Leica Microsystems, Mannheim, Germany). Different series of images were obtained and stacked by using the Imaris 7 software (Bitplane AG, Zürich, Switzerland).
All statistical analyses were performed using JMP 8.0 for Windows (SAS Institute Inc., Cary, NC, USA). Bacterial counts as well as adhesion and invasion data were log10-transformed to stabilize the variance and normalize residuals values for variance homogeneity.
A one-way analysis of variance (ANOVA) was performed to compare the effects of two consecutive treatments on mean Salmonella counts, adhesion and invasion capacities, as well as percentage changes in invasion and adhesion ratios, invasion efficiencies and transepithelial electrical resistance (TER). Measurements during the last 3 days of each fermentation period corresponding to a pseudo-steady-state were used as repetition. Salmonella counts, invasion and adhesion ratios, as well as invasion efficiency and TER measured during the last 3 days of each experimental period were not significantly different for F1 and F2, which were inoculated with the same child fecal microbiota immobilized in beads. Therefore, data obtained during system stabilization (Stab), Salmonella colonization (Sal) as well as E. coli L1000 (Ecol) and B. thermophilum RBL67 (Bif) treatment periods of F1 and F2 were used as independent replicates. TER data measured after 1, 2 and 3 h of incubation were not significantly different (P > 0.05). Therefore, mean TER values for the three incubation times were reported. Treatment means were compared using the Tukey-Kramer-HSD test with probability levels of P < 0.05 and P < 0.01.