The effects of fermentation products of prebiotic fibres on gut barrier and immune functions in vitro

Faecal collection

Fresh faecal samples were obtained from three healthy volunteers (Donor 1: 26 years old, female, BMI 28.1; Donor 2: 25 years old, male, BMI 26.3; Donor 3: 22 years old, male, BMI 23.0) at University of Minnesota, Department of Food Science and Nutrition (Carlson et al., 2017). All donors were healthy subjects and consumed a regular Western diet. Exclusion criteria were variables known to affect the balance of the gut microbiota, including specific diets, any known gastrointestinal diseases, the use of antibiotic in the last year and the use of dietary supplements. The samples were collected according to the procedure described previously (Carlson et al., 2017). All data and samples collected were done in accordance with University of Minnesota policies and procedures, including ethical approval of the study (Carlson et al., 2017).

Batch fermentation

Oat bran contains 28% oat β-glucan, 24% insoluble fibres, 23% protein, 9% starch, 5% lipids, 6% water, 4% ash (DSM Nutritional Products, Ltd., Kaiseraugst, Switzerland). Dried chicory root contains 70–75% inulin, 6–8% pectin, and 4–6% hemi/cellulose (WholeFibre, Inc., Pennington, NJ, USA). β-glucan contains 94% high viscosity β-glucan from oat flour (Megazyme, Inc., Bray, Ireland). 70% corn-derived XOS (AIDP, Inc., City of Industry, CA, USA), 90% inulin (Cargill, Inc., Minneapolis, MN, USA) and a fully digestible maltodextrin (Sigma-Aldrich, Buchs, Switzerland) were also included in the study.

The faecal samples were collected anaerobically, immediately homogenized in phosphate buffer solution (1:6 w/v) using a vortex. Obtained faecal slurry was combined with prepared reducing solution (2.52 g cysteine hydrochloride, 16 mL 1 N NaOH, 2.56 g sodium sulfide nonanhydride, 380 mL DD H₂O) at a 2:15 ratio. Ten mL of the prepared faecal inoculum was inoculated into the fermentation flasks within 5 min post defecation (Carlson et al., 2017). All samples (1.0 g of oat β-glucan 28%, delivering approximately 0.28 g of oat β-glucan; and 0.5 g of all other fibres) were hydrated in 40 mL sterile trypticase peptone fermentation medium (McBurney & Thompson, 1987) for 12 h at 4 °C followed by 2 h at 37 °C. The fermentation medium contained (per litre): 2.49 g trypticase peptone, 1.00 g ammonium bicarbonate, 8.75 g sodium bicarbonate, 1.43 g anhydrous sodium phosphate, 1.55 g anhydrous potassium phosphate monobasic, 0.60 g magnesium sulphate, 0.12 mg resazurin, 1.12 mmol calcium chloride, 0.63 mmol manganous chloride, 0.15 mmol cobalt chloride, 0.04 mmol ferric chloride (McBurney & Thompson, 1987). Ten mL of the faecal slurry was inoculated to each serum bottle. The final concentration of oat β-glucan 28% and other fibres is 2% and 1% (w/v), respectively. Batch fermentation was carried out at 37 °C under anaerobic conditions as described previously (Carlson et al., 2017). Effluent samples were collected from each fermentation flask before (0 h fermentation) and after fermentation (24 h fermentation), sterilized by filtering through 0.22 μm filters and immediately frozen at −80 C for in vitro analysis in cell culture systems.

Cell culture

Co-cultures of Caco-2 and HT29-MTX-E12 cells were used to assess the protective effect of the fermentation supernatants.

CaCo-2 ECACC 86010202 and HT29-MTX-E12 ECACC 12040401 cells (both from European Collection of Cell Cultures, Salisbury, UK) were both cultured at 37 °C, in atmosphere of 5% CO₂ in Dulbecco’s Modified Eagle Medium (DMEM) medium supplemented with 4.5g/L D-Glucose, four mM L-Glutamine, one mM Sodium Pyruvate, 1% MEM Non-Essential Amino Acids, 50 μg/mL Gentamicin (Life Technologies Europe B.V., Zug, Switzerland) and 10% heat-inactivated fetal bovine serum (FBS) (Sigma-Aldrich, Buchs, Switzerland). Sub-confluent cells were trypsinized using 0.25% Trypsin/EDTA (Life Technologies Europe B.V., Zug, Switzerland) and passaged at a ratio of 1:5–1:10 twice a week. Cells between passages 55 and 85 were used for this study.

Cells were seeded at a density of 20,000 cells/well in a 7:3 ratio (Caco-2:HT29-MTX-E12) in Corning HTS Transwell-24 system PET membranes, 0.4 μM pore size, cell growth area 0.33 cm²/well (Corning BV, Amsterdam, Netherlands) and cultured as described above. Media was changed every second to third day.

Barrier integrity measurement in healthy gut model

After 13 days in culture, initial transepithelial electrical resistance (TEER) readings were obtained using an EVOM² Voltohmmeter (EVOM; World Precision Instruments, Berlin, Germany) equipped with STX100 Electrodes. TEER values correlate with the tightness of the confluent monolayer (Srinivasan et al., 2015).

The background TEER (insert) was subtracted from total TEER (cell monolayer plus insert) to yield the monolayer resistance and then normalized to surface area by multiplying with the area of the insert (Srinivasan et al., 2015). The integrity of the monolayer was determined at basal level by measuring TEER. A TEER of 300 Ωcm² was regarded as a tight monolayer (see Results section). Afterwards, the cells were treated with sterile filtered fermentation supernatants or fermentation medium as a control (diluted 1:20 in growth media) (Fig. 1). After 24 h incubation the resistance across each cell monolayer was measured again and the percentage change in TEER compared to initial TEER for each insert was calculated to represent a healthy gut model. All TEER measurements were done by placing the plates containing the inserts on a 37 °C warming plate and were performed in the same duration.

Barrier integrity measurement in leaky gut model

The integrity of cell monolayers after the stressor test was evaluated by both TEER values and by measuring the transport of Lucifer yellow from apical to basolateral compartments (leaky gut model). A prewarmed Hank’s Balanced Salt Solution (HBSS) transport buffer composed of HBSS pH 7.4 with Ca²⁺ and Mg²⁺, containing 5.5 mM D-(+)-glucose, Sodium Bicarbonate, and supplemented with four mM L-glutamine, 20 mM HEPES (Life Technologies Europe B.V., Zug, Switzerland) was used for all permeability experiments. The inserts were rinsed twice with HBSS using a reservoir tray, 100 μL HBSS were then added to the apical site of the inserts and subsequently moved to new wells with fresh HBSS and allowed to equilibrate at 37 °C for 60 min in the incubator followed by TEER measurement to obtain baseline resistance readings.

Lucifer Yellow CH dilithium salt (Sigma-Aldrich, Buchs, Switzerland) was dissolved in HBSS at a final concentration of 100 μg/mL. Basolateral stressor EtOH (Merck, KGaA, Darmstadt, Germany) was diluted to 5% in HBSS. Apical Stressor Rhamnolipids R90 (AGAE Technologies, Corvallis, OR, USA) was dissolved in HBSS containing Lucifer Yellow at a final concentration of 350 μg/mL. All solutions were prepared freshly each time.

The HBSS transport buffer from the basolateral chamber was aspirated first, then the HBSS from the apical chamber and replaced with Lucifer yellow solution. For the apical stressor test, Lucifer yellow solution containing 350 μg/mL Rhamnolipids or Lucifer yellow solution without stressor for the control wells was loaded into the apical compartment, blank HBSS was subsequently added to the basolateral chamber. For the basolateral stressor test, Lucifer yellow solution was added to the apical side of all wells, and either 5% EtOH in HBSS or blank HBSS for the control wells was placed in the basolateral compartment.

After 3 h incubation at 37 °C the permeability of the cell monolayer was evaluated by measuring TEER values. TEER data were expressed as the percentage of the initial values.

After TEER measurements, 100 μL aliquot samples were withdrawn from the basolateral sites in duplicates to quantify the amount of Lucifer yellow transported. The fluorescence level (excitation at 420 nm and emission at 530 nm) was measured in a 96-well fluorescent plate reader (Spectramax M5 Series Multi-Mode Microplate Reader, Molecular Devices, Sunnyvale, CA, USA).

Mucus staining with Alcian Blue

HT29-MTX-E12 cells were seeded in 96-well tissue-culture plates (Corning BV, Amsterdam, Netherlands) at an initial cell density of 40,000 cells/well in complete culture medium. Confluent (72 h post-seeding) cells were then incubated in growth media containing fermentation supernatants (1:40 diluted) or fermentation medium as a control.

After 4 day-incubation with fermentation supernatants, Alcian Blue staining was performed to detect acidic mucous substances in the monolayers (Behrens et al., 2001). In brief, cells were washed with Phosphate-buffered saline (PBS) and subsequently fixed with 4% paraformaldehyde (Thermo Scientific, Rockford, IL, USA) for 30 min at room temperature. Cells were washed with PBS and subsequently stained using the Alcian Blue pH 2.5 mucin stain kit (Abcam, Cambridge, UK) according to the manufacturer’s instructions. To this aim, the cells were pre-treated in Acetic Acid Solution for 3 min. Acidic Sulfated Mucosubstances and Sialomucins were stained with Alcian Blue for 30 min at room temperature. The plates were washed 2 min in running tap water, followed by two changes of distilled water. Absorbance at 600 nM in 100 μL distilled water was read with a Spectramax M5 Series Multi-Mode Microplate Reader (Molecular Devices, Sunnyvale, CA, USA).

Subsequently, nuclei were stained with 1μg/mL Hoechst 33342 solution (Life Technologies Europe, Bleiswijk, Netherlands) for 30 min at room temperature.

Automated analysis of the cell number and normalization

Fourty-nine adjacent images (fields) of each well were acquired with a 10× objective using an ArrayScan VTI high-content screening system (Thermo Fisher Scientific, Pittsburgh, PA, USA), resulting in a field width of 640 microns. Channel one (Ch1) is the focus channel in which objects (Hoechst-stained nuclei) were identified. The cell numbers were determined by counting the fluorescent staining (Hoechst dye) and the mucus production was quantified by colorimetric analysis of Alcian Blue staining. The Alcian Blue absorbance was normalized to OD/100,000 cells.

HT29 cell culture and quantification of immunological biomarkers

HT29 cells were obtained from American Type Culture Collection (Manassas, VA, USA). Cells were cultured in DMEM supplemented with 10% FBS, 50 units/mL penicillin, 50 μg/mL streptomycin, L-glutamine, and nonessential amino acids (Life Technologies Europe B.V., Zug, Switzerland). Cells were seeded into 12-well plates at 8 × 10⁵ cells per well and used after 2 days of preculture. They were starved in DMEM containing 0.25% FBS for 18 h before treatment. Cells were treated for 24 h with TNF-α at 100 ng/mL or fermentation supernatants which were sterile filtered and diluted 1:20 in DMEM containing 0.25% FBS. All treatments were done in triplicate. Consequently, cytokines, chemokines, and interleukins were quantified in HT29 supernatants using the Luminex Technology (LiquiChip Workstation IS 200, Qiagen, Hilden, Germany) with Bio-Plex Pro Human Cytokine Panel kits (Bio-Rad, Hercules, CA, USA) or with Luminex Screening Assay kits (R&D Systems, Inc., Minneapolis, MN, USA), following the manufacturer’s instructions. The data were acquired with the Luminex IS 2.3 software and evaluated with the LiquiChip Analyser software (Qiagen, Hilden, Germany).

Statistical analysis

Statistical analyses were carried out with IBM SPSS Statistics 22.0 (IBM SPSS Inc., Chicago, IL, USA). TEER data were calculated as percentage change after treatment with the tested fibres and expressed as mean ± SEM. Lucifer yellow permeability was expressed as mean relative fluorescence units ± SEM. In healthy gut model, TEER values after the treatment with fermentation supernatants without any stressor were tested for normal distribution using Shapiro–Wilk test, and means were compared pairwise using Student’s t-test for normally distributed data. A non-parametric Mann–Whitney test was performed when data were not normally distributed. In the leaky gut model, TEER and relative fluorescence units were compared pairwise using Student’s t-test. Mucus production data were expressed as normalized Alcian Blue absorbance ± SD and were compared using two tailed Student’s t-test. Cytokines and chemokines concentration were expressed as pg/mL ± SD and were compared using two tailed Student’s t-test. P-values < 0.05 were considered significant.

Effects of fibres on gut barrier integrity using healthy gut model

To study the impact of oat β-glucan 28%, oat β-glucan 94%, inulin 90%, dried chicory root containing 75% inulin, XOS 70% and maltodextrin on epithelial barrier function, TEER was measured in a co-culture of absorptive enterocytes (Caco-2 cells) and mucus-secreting goblet cells (HT29-MTX-E12 cells) (Figs. 2A–2F). TEER significantly increased (P < 0.001) in cells treated with supernatant from 24 h fermentation of oat β-glucan 28% compared to cells treated with supernatant collected before fermentation (0 h fermentation) (24 h: 123.0 ± 0.8, 0 h: 114.9 ± 1.2). Treatments with supernatant from 24 h fermentation of dried chicory root containing 75% inulin (118.1 ± 1.1 vs 105.5 ± 0.5), XOS 70% (122.9 ± 0.9 vs 114.9 ± 1.0), inulin 90% (119.3 ± 0.9 vs 110.3 ± 1.0), oat β-glucan 94% (116.3 ± 1.0 vs 110.7 ± 0.4) and maltodextrin (118.4 ± 0.76 vs 106.3 ± 0.41) significantly increased TEER compared to supernatant collected before fermentation (0 h fermentation).

Effects of fibres on gut barrier integrity using a basolateral induced leaky gut model

The impact of oat β-glucan 28%, oat β-glucan 94%, inulin 90%, XOS 70%, dried chicory root containing 75% inulin and maltodextrin fermentation products on tight junction development of co-cultures of Caco-2 and HT29-MTX-E12 were further studied via the basolateral induction of leaky gut by ethanol (Fig. 3). The treatment of the co-culture with ethanol as the basolateral stressor resulted in approximately 50% reduction of TEER compared to the untreated control (without stressor: 67.7 ± 3.4%, after stressor: 31.5 ± 1.9%). The supernatant of the in vitro fermentation of oat β-glucan 28% by the gut microbiota for 24 h resulted in a significant increase in TEER (P < 0.01) compared to the basolateral stressor control (24 h: 33.8 ± 0.5%, basolateral stressor: 29.0 ± 1.4%) (Figs. 3A–3F). There was no significant change in TEER in cells treated with 24 h fermentation samples compared to 0 h fermentation samples. TEER after treating the cells with supernatant from 24 h fermentation of XOS 70%, inulin 90%, oat β-glucan 94%, and dried chicory root containing 75% inulin was not significantly different compared to supernatant from unfermented samples (0 h fermentation) and ethanol control. In contrast, supernatant from the fermentation of maltodextrin for 24 h significantly decreased TEER compared to 0 h fermentation supernatant (24 h: 31.0 ± 0.2%, 0 h: 32.1 ± 0.3%).

The permeability of Lucifer yellow was measured to monitor the integrity of the tight junction of the co-cultures under basolateral stress (Figs. 3G–3L). We observed significant decrease of relative fluorescence units in cells treated with supernatant of 24 h fermentation of oat β-glucan 28%, oat β-glucan 94% and maltodextrin compared to cells treated with supernatant of unfermented samples (0 h fermentation), but not for inulin 90%, XOS 70%, and dried chicory root containing 75% inulin (oat β-glucan 28%: 193.5 ± 6.0% vs. 224.5 ± 7.4%; oat β-glucan 94%: 192.1 ± 3.7% vs. 230.7 ± 4.3%; maltodextrin: 223.6 ± 4.0% vs. 253.7 ± 5.5%). The supernatant of 24 h fermentation of all tested fibres resulted in a significant reduction in Lucifer yellow permeability compared to the basolateral stressor control (oat β-glucan 28%: 193.5 ± 6.0% vs. 257.5 ± 9.4%; XOS 70%: 167.0 ± 3.3% vs. 214.4 ± 2.5%; Inulin 90%: 184.7 ± 3.3% vs. 213.4 ± 5.5%; oat β-glucan 94%: 192.1 ± 3.7% vs. 242.7 ± 2.8%; dried chicory root containing 75% inulin: 166.3 ± 3.1% vs. 225.9 ± 2.9%; maltodextrin: 223.6 ± 4.0% vs. 268.4 ± 5.0%). We also observed a significant effect of the fermentation supernatant at time point 0 on reducing Lucifer yellow relative fluorescence units compared to the basolateral stressor control (P < 0.05) in the fermentations with XOS 70%, and dried chicory root containing 75% inulin.

Effects of fibres on gut barrier integrity using an apical induced leaky gut model

The impact of oat β-glucan 28%, oat β-glucan 94%, inulin 90%, XOS 70%, dried chicory root containing 75% inulin and maltodextrin on tight junction development of co-cultures of Caco-2 and HT29-MTX-E12 were further studied using apical induction of leaky gut by rhamnolipids (Fig. 4). Similar to the treatment with basolateral stressor, the concentration of the apical stressor was chosen such that the treatment of the co-culture resulted in approximately 50% reduction of TEER compared to the untreated control (without stressor: 72.1 ± 3.8%, after stressor: 31.6 ± 4.5%).

There was no significant impact on TEER values of cultures treated with 24 h fermentation samples of all tested fibres compared to cultures treated with samples before fermentation (0 h fermentation) and compared to cultures treated with the apical stressor control (Figs. 4A–4F).

We observed significant decrease of relative Lucifer yellow fluorescence in cultures treated with supernatant of 24 h fermentation of oat β-glucan 28% (218.8 ± 7.3% vs. 257.3 ± 6.8%) and dried chicory root containing 75% inulin (210.8 ± 7.1% vs. 265.0 ± 14.1%) compared to cells treated with unfermented supernatant (0 h fermentation) (Figs. 4G–4L). The treatment with 24 h fermentation samples (24 h fermentation) of inulin 90% and dried chicory root containing 75% inulin resulted in a significant reduction in Lucifer yellow permeability compared to the apical stressor control (Inulin 90%: 211.3 ± 6.3% vs. 252.8 ± 8.9%; dried chicory root containing 75% inulin: 210.8 ± 7.1% vs. 265.4 ± 6.8%) (Figs. 4G–4L).

Effects of fibres on mucus production of HT29-MTX E12 cells

To study the effects of fibres on mucus production, HT29-MTX-E12 cells were stained by Alcian Blue. We observed inter-individual variability in mucus production of cells treated with supernatant from fermentation of faecal samples from three donors (Figs. 5A–5C). Mucus production was significantly increased in cells treated with supernatant from 24 h XOS 70% fermentation using faecal samples from donor 2 (0.143 average 600 nm absorbance ± 0.001 vs. 0.133 ± 0.005) and donor 3 (0.139 ± 0.003 vs. 0.120 ± 0.002). Oat β-glucan 94% fermentation products significantly increased mucus production of cells treated with the supernatant from fermentation of faecal sample obtained from donor 1 (0.163 ± 0.010 vs. 0.146 ± 0.001). On the other hand, oat β-glucan 94% decreased mucus production of cells treated with 24 h fermentation supernatant from flasks inoculated with donor 2 and donor 3, albeit without statistical significance. The 24 h treatment of supernatant from maltodextrin fermentation of faecal samples from donor 1 (0.152 ± 0.001 vs. 0.147 ± 0.003) and donor 3 (0.136 ± 0.005 vs. 0.120 ± 0.004) resulted in a significant reduction in mucus production compared to time point 0.

Effects of fibres on immunological biomarkers

Similar changes in levels of all immunological biomarkers were seen between donor 1 and donor 2 after 24 h fermentation with the tested fibre supernatants. Levels of all immunological biomarkers in samples delivered from donor 3 were below the detection levels and therefore not reported in this study. We observed in effluent from 24 h fermentation of oat β-glucan 28% using faecal samples from donor 1 (Figs. 6A–6E) and donor 2 (Figs. 6F–6J) a significant increase in IFN-γ (Donor 1: 9443.3 ± 4539.8 vs. 1134.0 ± 196.5 pg/mL, Donor 2: 2203.3 ± 962.6 vs. 52.7 ± 2.8 pg/mL, respectively), IL-10 (Donor 1: 66.4 ± 37.2 vs. 6.1 ± 0.5 pg/mL, Donor 2: 11.6 ± 5.0 vs. 0.0 ± 0.0 pg/mL), IL-17 (Donor 1: 1427.7 ± 765.7 vs. 128.0 ± 15.6 pg/mL, Donor 2: 347.0 ± 174.1 vs. 4.7 ± 0.2 pg/mL), IL-2 (Donor 1: 292.3 ± 132.0 vs. 40.4 ± 5.7 pg/mL, Donor 2: 73.9 ± 33.3 vs. 0.0 ± 0.0 pg/mL), and IL-9 (Donor 1: 3086.7 ± 1682.0 vs. 237.7 ± 27.7 pg/mL, Donor 2: 669.7 ± 308.3 vs. 5.8 ± 1.2 pg/mL) compared to timepoint 0. Levels of all immunological biomarkers were higher in 24 h fermentation samples with oat β-glucan 94% compared to other tested fibres. Levels of IFN-γ (Donor 1: 20300.0 ± 4596.7 vs. 983.3 ± 197.0 pg/mL, Donor 2: 6763.3 ± 2422.9 vs. 44.5 ± 9.8 pg/mL), IL-10 (Donor 2: 39.8 ± 15.3 vs. 0.0 ± 0.0 pg/mL), IL-17 (Donor 1: 3430.0 ± 1088.1 vs. 122.0 ± 17.1 pg/mL, Donor 2: 1207.3 ± 508.2 vs. 3.5 ± 0.0 pg/mL), IL-2 (Donor 1: 687.0 ± 202.1 vs. 36.4 ± 5.3 pg/mL, Donor 2: 220.3 ± 75.1 vs. 0.0 ± 0.0 pg/mL), and IL-9 (Donor 1: 8243.3 ± 2743.2 vs. 216.0 ± 42.8 pg/mL, Donor 2: 2306.7 ± 938.3 vs. 6.5 ± 1.8 pg/mL) found in 24 h oat β-glucan 94% fermentation samples were significantly higher compared to time point 0 in both donors. We observed in effluent from 24 h fermentation of inulin 90% using faecal samples from both donors a significant increase in IFN-γ (Donor 1: 5430.0 ± 2208.7 vs. 678.7 ± 41.5 pg/mL, Donor 2: 2263.3 ± 868.4 vs. 62.8 ± 7.8 pg/mL, respectively), IL-10 (Donor 1: 35.8 ± 15.6 vs. 2.8 ± 0.4 pg/mL, Donor 2: 11.8 ± 5.6 vs. 0.0 ± 0.0 pg/mL), IL-17 (Donor 1: 571.3 ± 248.8 vs. 87.0 ± 4.7 pg/mL, Donor 2: 370.3 ± 142.6 vs. 6.1 ± 2.5 pg/mL), IL-2 (Donor 1: 157.2 ± 58.5 vs. 22.7 ± 1.8 pg/mL, Donor 2: 78.8 ± 29.7 vs. 0.0 ± 0.0 pg/mL), and IL-9 (Donor 1: 1463.0 ± 640.7 vs. 147.7 ± 11.0 pg/mL, Donor 2: 758.3 ± 296.2 vs. 6.1 ± 1.9 pg/mL) levels compared to time point 0.

Levels of IFN-γ, IL-17, IL-2, and IL-9 found in 24 h XOS 70% and maltodextrin fermentation samples were significantly higher compared to time point 0 in donor 2. Significant higher concentration of IL-2 were detected from 24 h fermentation samples of dried chicory root containing 75% inulin compared to time point 0 in donor 2.