Yogurt starter strains ameliorate intestinal barrier dysfunction via activating AMPK in Caco-2 cells

ABSTRACT

Lactic acid bacteria (LAB) are commonly used probiotics that improve human health in various aspects. We previously reported that yogurt starter strains, Lactobacillus delbrueckii subsp. bulgaricus 2038 and Streptococcus thermophilus 1131, potentially enhance the intestinal epithelial barrier function by inducing the expression of antimicrobial peptides in the small intestine. However, their effects on physical barrier functions remain unknown. In this study, we found that both strains ameliorated the decreased trans-epithelial resistance and the increased permeability of fluorescein isothiocyanate-dextran induced by tumor necrosis factor (TNF)-α and interferon (IFN)-γ in Caco-2 cells. We also demonstrated that LAB prevented a decrease in the expression and disassembly of tight junctions (TJs) induced by TNF-α and IFN-γ. To assess the repair activity of TJs, a calcium switch assay was performed. Both strains were found to promote the reassembly of TJs, and their activity was canceled by the inhibitor of AMP-activated protein kinase (AMPK). Moreover, these strains showed increased AMPK phosphorylation. These observations suggest that the strains ameliorated physical barrier dysfunction via the activation of AMPK. The activities preventing barrier destruction induced by TNF-α and IFN-γ were strain-dependent. Several strains containing L. bulgaricus 2038 and S. thermophilus 1131 significantly suppressed the barrier impairment, and L. bulgaricus 2038 showed the strongest activity among them. Our findings suggest that the intake of L. bulgaricus 2038 and S. thermophilus 1131 is a potential strategy for the prevention and repair of leaky gut.

KEYWORDS:

• intestinal barrier function
• tight junction
• AMPK
• calcium switch assay
• Lactobacillus delbrueckii subsp. bulgaricus
• Streptococcus thermophilus

1. Introduction

The intestine is one of the most important organs for living organisms to absorb water and nutrients to maintain life activities. Intestinal barrier functions are necessary to eliminate harmful substances, such as toxins, allergens, and pathogens. Increased intestinal permeability by barrier dysfunction, termed as “leaky gut,” results in the influx of harmful substances containing endotoxins, such as lipopolysaccharides (LPS), into the body and its translocation to various organs via blood vessels.¹ Moreover, pathogens carried to organs through the blood lead to the spread of infection. This causes chronic low-grade inflammation and various diseases, such as inflammatory bowel disease (IBD),² irritable bowel syndrome (IBS),³ diabetes,⁴ nonalcoholic fatty liver disease⁵ and depression.⁶ Therefore, maintaining and enhancing the intestinal barrier function is important for human health.

Tight junctions (TJs) play one of crucial roles to form the intestinal barriers that divide the internal and external domains of the intestine. Claudin (CLDN),⁷ occludin (OCLN)⁸ and junctional adhesion molecule (JAM)⁹ are well-known transmembrane proteins of TJs, and the zonula occludens (ZO) family anchors them to the actin cytoskeleton.^10,11 Intestinal TJ integrity is regulated by several factors, including dietary components, prebiotics, probiotics, and microbiota.¹² The expression levels of TJs are closely related to IBD and IBS.¹³ Hence, the regulation of TJ integrity is important for the treatment and prevention of leaky gut that induces these diseases.

Leaky gut is caused by various factors. In food ingredients, gliadin, a protein composed of gluten, decreases trans-epithelial electrical resistance (TEER) value, increases permeability, and alters the TJ protein expression.¹⁴ High-fat diet induces the invasion of pro-inflammatory macrophages into intestinal epithelial cells, increasing the intestinal permeability.¹⁵ Ethanol causes displacement of ZO-1 protein by the activation of myosin light-chain kinase.¹⁶ Endurance exercise is associated with intestinal barrier dysfunction.¹⁷ Moreover, intestinal dysbiosis caused by genetic or environmental factors leads to leaky gut by inducing a pro-inflammatory response that destroys TJ integrity.¹⁸ Therefore, we are continuously exposed to factors that induce leaky gut in our daily life. Aging is also known to result in intestinal barrier dysfunction.¹⁹ Small intestine biopsies from the elderly showed lower TEER values compared to young controls, while CLDN-2, a leaky type TJ protein,²⁰ had high expression levels in the elderly.²¹ With progressive aging, increase in the incidence of leaky gut must be arrested to prevent the outbreak of various diseases resulting from leaky-gut-originated low-glade inflammation.

Lactic acid bacteria (LAB) have been widely used as starters for fermented foods, such as yogurt. Some of them promote human health as probiotics. We previously reported that the long-term intake of a yogurt fermented with Lactobacillus delbrueckii subsp. bulgaricus 2038 and Streptococcus thermophilus 1131 in mice increases the gene expression of antimicrobial peptides in the small intestine.²² We further showed that L. bulgaricus 2038 and S. thermophilus 1131 solely induced gene expression of the regenerating family member 3 family in the small intestine of mice via the stimulation of dendritic cells and type 3 innate lymphoid cells.²³ These studies suggest that L. bulgaricus 2038 and S. thermophilus 1131 enhance the intestinal chemical barrier function, but their effects on the physical barriers, such as TJs, remain unknown.

Toll-like receptor (TLR)-2 stimulation of Caco-2 cells regulates the epithelial barrier function by preserving its integrity and polarization.²⁴ L. bulgaricus 2038 and S. thermophilus 1131 are gram-positive bacteria whose cell walls contain peptidoglycans (PGs) and lipoteichoic acids (LTAs) that are recognized by TLR2. Our previous reports showed that both strains induced the production of interleukin (IL)-23 from dendritic cells.²³ As the production of IL-23 is induced by the TLR2 ligand,²⁵ it is suggested that both strains induce IL-23 expression via TLR2 activation. Therefore, we hypothesized that L. bulgaricus 2038 and S. thermophilus 1131 may enhance the intestinal physical barrier function. In this study, we evaluated the effects of L. bulgaricus 2038 and S. thermophilus 1131 on the intestinal physical barrier function and examined the underlying molecular mechanism.

2. Materials and methods

2.1. Bacteria culture

Bacteria were cultured anaerobically with AnaeroPouch-Anaero (Mitsubishi Gas Chemical, Tokyo, Japan) at 37°C for 18 h in the medium described in Table 1. Bacteria were washed twice with phosphate-buffered saline (PBS; pH 7.4) and suspended in PBS so that the optical density at 600 nm was 10.0 using a U-2810 spectrophotometer (Hitachi, Tokyo, Japan). Heat-treated bacteria were obtained by incubating at 75°C for 1 h.

Table 1. Bacteria and medium used in this study.

Strain	Medium
L. bulgaricus 2038	MRS*
S. thermophilus 1131	LM17**
L. bulgaricus JCM 1002^T	MRS
L. bulgaricus ME-876	MRS
S. thermophilus NCIMB 8510^T	LM17
Lactobacillus paragasseri ME-879	MRS
Lactobacillus paragasseri ME-880	MRS
Lacticaseibacillus paracasei ME-881	MRS
Lacticaseibacillus paracasei ME-882	MRS
Lactococcus lactis subsp. lactis ME-883	MRS
Bifidobacterium bifidum ME-884	GAM***
Bifidobacterium longum ME-885	GAM
Propionibacterium freudenreichii ME-886	GAM

*de Man Rogosa Sharpe broth (Becton Dickinson, Cockeysville, MD, USA).

**M17 broth (Becton Dickinson) supplemented with 1% lactose.

***Gifu anaerobic medium (Nissui Pharmaceutical, Tokyo, Japan).

L. bulgaricus JCM 1002^T and S. thermophilus NCIMB 8510^T were purchased from RIKEN BRC (Ibaraki, Japan) and NCIMB, Ltd. (Aberdeen, Scotland, UK), respectively. The other bacteria used in this study were obtained from Meiji Co. Ltd (Tokyo, Japan).

2.2. Cell culture

Caco-2 cells, which are widely used as small intestinal epithelial models, were purchased from the European Collection of Authenticated Cell Cultures and cultured in a minimum essential medium (MEM; Gibco, Thermo Fisher Scientific, Rochester, NY, USA) supplemented with 10% (v/v) fetal bovine serum (Biowest, Nuaillé, France), 1% (v/v) MEM non-essential amino acid solution (Thermo Fisher Scientific), 100 U/mL penicillin, and 100 µg/mL streptomycin (Thermo Fisher Scientific). The culture was maintained at 37°C and 5% CO₂. Medium without penicillin – streptomycin was used during co-culture with LAB and Caco-2 cells.

To assess the intestinal barrier function, cells were seeded at a density of 8.9 × 10⁴ cells/cm² on the membranes in Transwell inserts (0.4 µm pore size; Costar, Cambridge, MA, USA) and cultivated for 14 d to form a monolayer. The medium was changed every two or three days. TEER value was measured using a MILLICELL-ERS voltohmmeter system (Millipore, Burlington, MA, USA).

2.3. Paracellular permeability assay

The permeability of Caco-2 cell monolayers was assessed by adding fluorescein isothiocyanate-dextran with an average molecular weight of 4,000 (FD-4). FD-4 (1 mg/mL; Sigma-Aldrich, St Louis, MO, USA) was added to the apical side. The medium on the basolateral side was collected after 1 h, and the fluorescence emission at 535 nm was measured after excitation at 480 nm using a Synergy H1 microplate reader (BioTek, Santa Clara, CA, USA).

The apparent permeability coefficient (Papp) was calculated according to the following equation: Papp = (ΔQ/Δt)/(C₀ × S), where ΔQ/Δt is the amount of FD-4 permeated per second, C₀ is the initial FD-4 concentration on the apical side, and S is the surface area of the membrane.

2.4. Barrier destruction by TNF-α and IFN-γ

TNF-α and IFN-γ are reported to increase barrier permeability of Caco-2 cells.^26,27 TNF-α (10 ng/mL; R&D Systems, Minneapolis, MN, USA) and IFN-γ (10 ng/mL; R&D Systems) were added to the basolateral side for 18 h. To assess the protective activities of bacteria on barrier functions, live or heat-treated bacteria were simultaneously added to the apical side. After incubation, TEER value was measured and a paracellular permeability assay was conducted. Cell viability was evaluated using a WST-1 cell proliferation assay system (Takara Bio, Shiga, Japan).

2.5. Calcium switch assay

Caco-2 cells were cultured on membranes in Transwell inserts for 14 d. The cells were washed twice with calcium-free MEM (S-MEM; Gibco) and incubated with S-MEM for 16 h. After removing S-MEM, cells were washed with serum-free MEM and incubated with serum-free MEM for the indicated time.

To assess the promotion of TJ assembly by LAB, LAB were added to serum-free medium. To determine the involvement of AMP-activated protein kinase (AMPK), 0.5 mM 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR; Tokyo Chemical Industry, Tokyo, Japan), an AMPK activator, and 10 μM dorsomorphin (Wako, Osaka, Japan), an AMPK inhibitor, were added.

2.6. RNA isolation and gene expression analysis

Total RNA from Caco-2 cells was prepared using the Maxwell RSC simpleRNA Cells Kit (Promega, Madison, WI, USA), according to the manufacturer’s protocols. RNA was quantified using a Nanodrop 8000 spectrophotometer (Thermo Fisher Scientific). Complementary DNA was synthesized from 1 μg of total RNA using PrimeScript RT Master Mix (TaKaRa Bio), and real-time polymerase-chain reaction (PCR) was performed using the QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific) and SYBR Premix Ex Taq II (TaKaRa Bio), according to the manufacturer’s protocol. The primers used were: CLDN4 forward 5′-GGCTGCTTTGCTGCAACTGTC-3′ and reverse 5′-GAGCCGTGGCACCTTACACG-3′, OCLN forward 5′-TCAGGGAATATCCACCTATCACTTCAG-3′ and reverse 5′-CATCAGCAGCAGCCATGTACTCTTCAC-3′, F11R forward 5′-GGTCAAGGTCAAGCTCAT-3′ and reverse 5′-CTGAGTAAGGCAAATGCAG-3′, TJP1 forward 5′-CGGTCCTCTGAGCCTGTAAG-3′ and reverse 5′-GGATCTACATGCGACGACAA-3′, and glyceraldehyde 3-phosphate dehydrogenase (GAPDH) forward 5′-GCACCGTCAAGGCTGAGAAC-3′ and reverse 5′-TGGTGAAGACGCCAGTGGA-3′. Quantitative comparisons were performed using the ΔΔC_T method. Data were normalized to the values of GAPDH, and the results were expressed as fold-changes to threshold cycle values relative to controls.

2.7. Immunofluorescence microscopy

Caco-2 cells cultured in the inserts of Transwell plate were washed twice with PBS. Cells were fixed with 4% paraformaldehyde (Wako) for 10 min at room temperature (RT). After rinsing with PBS, PBS containing 0.2% Triton X-100 was added and the cells were incubated for 5 min at RT. One drop of Image-iT FX Signal Enhancer (Thermo Fisher Scientific) was added and incubated for 30 min at RT. The cells were blocked with 10% normal goat serum for 20 min at RT and labeled with anti-CLDN-4 (MAB42191-SP; R&D Systems), anti-OCLN (GTX114949; Genetex, Irvine, CA, USA), anti-JAM-A (NBP2–71876; Novus Biologicals, Littleton, CO, USA), or anti-ZO-1 (GTX108627; Genetex) antibodies for 2 h, followed by incubation for 1 h with fluorescein-conjugated goat anti-IgG (ab150087 and ab150117; Abcam, Cambridge, MA, USA). The membranes were cut with a scalpel and placed on glass slides. The cells were visualized using a confocal laser-scanning microscope (LSM880; Zeiss, Oberkochen, Germany). The fluorescence intensities of ZO-1 in the calcium switch assay were measured using ImageJ Fiji, as previously described.²⁸ The fluorescence intensity was quantified using the following equation: fluorescence intensity = mean fluorescence intensity (MFI) in a field – MFI of the background. The fluorescence intensities of three randomly selected fields were then averaged.

2.8. Western blotting

Caco-2 cells were lysed in the radioimmunoprecipitation assay buffer (Nacalai Tesque, Kyoto, Japan) containing 1 mM ethylene glycol tetraacetic acid, 1 mM ethylenediaminetetraacetic acid, protease inhibitor cocktail (Sigma-Aldrich), and phosphatase inhibitor cocktail (Sigma-Aldrich). Cell lysates were sonicated with a Bioruptor (Sonicbio, Kanagawa, Japan) for 5 × 5 s and centrifuged for 5 min at 15,000 ×g at 4°C. The supernatants were collected to measure the total protein concentration using a BCA protein assay kit (Thermo Fisher Scientific). Ten micrograms of the extracted proteins was separated by sodium dodecyl sulfate-polyacrylamide gel electrophoresis using Mini-Protean 4–20% gels (Bio-Rad, Hercules, CA, USA) and transferred to a polyvinylidene difluoride membrane (Bio-Rad). The membranes were blocked using EveryBlot Blocking Buffer (Bio-Rad) for 5 min. The membranes were probed with primary antibodies for 1 h at RT. The primary antibodies used were anti-total AMPK (1:1000; Cell Signaling Technologies, Danvers, MA, USA), anti-phosphorylated AMPK (pAMPK; 1:1000; Cell Signaling Technologies), and β-actin (1:1000; Cell Signaling Technologies). The membranes were incubated with anti-rabbit (Cell Signaling Technologies) or anti-mouse (Cell Signaling Technologies) secondary antibodies for 1 h at RT. Bands were detected with a ChemiDoc XRS+ system (Bio-Rad), and the intensity was quantified using the Image Lab software version 6.1.0 (Bio-Rad).

2.9. Statistical analysis

All of the experiments were repeated at least 2 times producing similar results, and one of the representative data were shown. Data are presented as the mean ± standard error. Statistical analysis was performed using the Tukey–Kramer test. For comparative study of the activity on the intestinal barrier function, student’s t-test was performed. Statistical significance was set at P < 0.05.

3. Results

3.1. L. bulgaricus 2038 and S. thermophilus 1131 ameliorate the impairment of barrier function of Caco-2 cells by TNF-α and IFN-γ

We examined whether L. bulgaricus 2038 and S. thermophilus 1131 ameliorated the impairment of barrier function caused by TNF-α and IFN-γ. Eighteen hours of treatment with TNF-α and IFN-γ significantly decreased the TEER value and induced FD-4 permeability (Figure 1A, B). Live L. bulgaricus 2038 and S. thermophilus 1131 strains significantly suppressed these effects (Figure 1A, B). Heat-treated strains tended to suppress the decrease in TEER value and significantly reduced FD-4 permeability (Figure 1C, D). These results suggest that both strains improve the impaired physical barrier functions in the live and heat-treated states.

Figure 1. Lactobacillus delbrueckii subsp. bulgaricus 2038 and Streptococcus thermophilus 1131 improved the intestinal barrier dysfunction by tumor necrosis factor (TNF)-α and interferon (IFN)-γ.

L. bulgaricus 2038, S. thermophilus 1131, TNF-α, and IFN-γ were simultaneously added to the Caco-2 monolayers. (A) Trans-epithelial electrical resistance (TEER) percentage of initial value and (B) the permeability of fluorescein isothiocyanate-dextran with an average molecular weight of 4,000 (FD-4) after 18 h in the experiments using live lactic acid bacteria (LAB) are shown (n = 5). (C) TEER value (n = 5–6) and (D) FD-4 (n = 4–7) permeability in the experiments using heat-treated LAB are shown. Data are expressed as the mean ± standard error (SE). *P < 0.05, †P < 0.1.

To determine whether metabolites, such as lactic acid produced by live LAB, influenced the viability of Caco-2 cells, a cell viability assay was conducted. The results showed no significant differences among the experimental groups (Figure S1).

Hereafter, we conducted further experiment with live LAB only.

3.2. L. bulgaricus 2038 and S. thermophilus 1131 rescue the decrease in TJ gene expression using TNF-α and IFN-γ

To demonstrate the improvement of intestinal barrier dysfunction by live L. bulgaricus 2038 and S. thermophilus 1131, we measured the gene expression of TJ proteins, CLDN-4, OCLN, JAM-A (F11R), and ZO-1 (TJP1), which are essential for TJ structure. TNF-α and IFN-γ significantly decreased OCLN and TJP1 levels, and L. bulgaricus 2038 and S. thermophilus 1131 significantly suppressed these effects (Figure 2). There were no significant changes in CLDN4 and F11R levels (data not shown).

Figure 2. L. bulgaricus 2038 and S. thermophilus 1131 suppressed the decrease in the gene expression levels of tight junctions (TJs) by TNF-α and IFN-γ.

Live L. bulgaricus 2038, S. thermophilus 1131, TNF-α, and IFN-γ were simultaneously added to the Caco-2 monolayers. Caco-2 cells were collected after 18 h of incubation. Gene expression levels of (A) occludin (OCLN) and (B) zonula occludens-1 (ZO-1/TJP1) were evaluated via real-time polymerase-chain reaction (PCR). Data were normalized to glyceraldehyde 3-phosphate dehydrogenase (GAPDH) expression and are shown as relative expression levels. Data are expressed as the mean ± SE (n = 8–9). *P < 0.05.

3.3. L. bulgaricus 2038 and S. thermophilus 1131 contribute to the maintenance of TJ protein structure destroyed by TNF-α and IFN-γ

We conducted immunofluorescence assay to analyze whether live L. bulgaricus 2038 and S. thermophilus 1131 contributed to maintaining the TJ structure. The immunofluorescent images showed that TNF-α and IFN-γ caused abnormal distribution of the TJ proteins, CLDN-4, OCLN, JAM-A, and ZO-1, because the internalization of TJ proteins was observed, and the cell outline became more obscure after treatment (Figure 3). However, treatment with L. bulgaricus 2038 and S. thermophilus 1131 maintained the morphology of TJ proteins (Figure 3).

Figure 3. L. bulgaricus 2038 and S. thermophilus 1131 suppressed the destruction of TJ proteins by TNF-α and IFN-γ.

Live L. bulgaricus 2038, S. thermophilus 1131, TNF-α, and IFN-γ were simultaneously added to the Caco-2 monolayers. After 18 h of incubation, TJ structures (claudin [CLDN]-4, OCLN, junctional adhesion molecule [JAM]-A, and ZO-1) were visualized and analyzed via immunofluorescence microscopy (scale bar = 20 μm). White arrows show the internalization of TJ proteins.

3.4. L. bulgaricus 2038 and S. thermophilus 1131 promote the reassembly of TJ proteins

To evaluate the activity of live L. bulgaricus 2038 and S. thermophilus 1131 on the reassembly of TJ proteins, we conducted a calcium switch assay, which has been widely used to assess the reassembly of epithelial TJ proteins.^29–31 As extracellular calcium ions are required for cell – cell adhesion, depletion of calcium causes the translocation of TJ proteins to the cytoplasm, and the addition of calcium induces TJ reassembly to the cell membrane. Caco-2 monolayers on Transwell inserts were cultured in S-MEM for 16 h and changed to calcium-containing medium (calcium switch). Simultaneously, L. bulgaricus 2038 and S. thermophilus 1131 were added to the medium to evaluate the reassembly activities by monitoring the shift in TEER values, paracellular permeability, and immunofluorescence. As AMPK plays an important role in TJ assembly, AICAR, an AMPK activator, and dorsomorphin, an AMPK inhibitor, were used as the positive and negative controls, respectively.

Cultivation of the Caco-2 monolayer with S-MEM for 16 h decreased the TEER value to approximately 5% of the initial TEER value before changing to S-MEM (Figure 4A). The TEER values gradually increased in all groups after the calcium switch and reached almost 100% at 8 h (Figure 4A). AICAR and L. bulgaricus 2038 significantly increased the TEER value compared to the control group at 8 h (Figure 4A). Furthermore, dorsomorphin significantly suppressed the increase in TEER value induced by L. bulgaricus 2038 (Figure 4A), suggesting that L. bulgaricus 2038 promoted TJ reassembly via AMPK activation. S. thermophilus 1131 also increased the TEER value, but did not show a significant difference compared with the control group (Figure 4A).

Figure 4. L. bulgaricus 2038 and S. thermophilus 1131 promoted the assembly of TJ proteins in a calcium switch assay.

Caco-2 monolayers were cultured in S-MEM for 16 h. Then, S-MEM was replaced with normal medium, and L. bulgaricus 2038, S. thermophilus 1131, 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), and dorsomorphin were added simultaneously. (A) TEER values were measured every hour (n = 6–8). TEER value was calculated before changing to S-MEM at 100%. (B) Permeability of FD-4 after 8 h was evaluated (n = 6–8). (C) Immunofluorescence staining of ZO-1 at calcium switch (-Ca²⁺) and 8 h was analyzed via immunofluorescence microscopy (n = 3–4; scale bar = 20 μm). White arrows show clear formation of cell shape. (D) Fluorescence intensity of ZO-1 was quantified (n = 3–4). (E) Caco-2 cells were collected at 8 h and the expression of phosphorylated AMP-activated protein kinase (p-AMPK) was evaluated via western blotting. Expression of p-AMPK was normalized to total AMPK expression, and the relative expression levels are shown (n = 4). Data are expressed as the mean ± SE. *P < 0.05, †P < 0.1.

FD-4 permeability was evaluated 8 h after the calcium switch. AICAR and L. bulgaricus 2038 significantly decreased the FD-4 permeability compared to the control, and dorsomorphin tended to suppress the activity of L. bulgaricus 2038, similar to the TEER value (Figure 4B).

Next, we performed immunofluorescence staining to visually evaluate the activities of the reassembly of ZO-1 on behalf of TJ proteins at 8 h. Cultivation in S-MEM caused abnormal distribution of the ZO-1 structure (Figure 4C). After calcium switch, AICAR, L. bulgaricus 2038, and S. thermophilus 1131 promoted the reassembly of ZO-1 protein, because the cell shapes observed clearer than control group (Figure 4C). The intensity of ZO-1 in the AICAR, L. bulgaricus 2038, and S. thermophilus 1131 groups was significantly greater than that in the control group (Figure 4D). Furthermore, dorsomorphin canceled the activities of L. bulgaricus 2038 and S. thermophilus 1131 (Figure 4D).

To clarify whether AMPK was activated by L. bulgaricus 2038 and S. thermophilus 1131, the levels of AMPK phosphorylation during the calcium switch assay were measured. AMPK phosphorylation 8 h after calcium switch was promoted by both strains (Figure 4E).

3.5. L. bulgaricus 2038 and S. thermophilus 1131 promote AMPK phosphorylation

The calcium switch assay revealed that both L. bulgaricus 2038 and S. thermophilus 1131 promoted the reassembly of TJ proteins via AMPK activation. We evaluated whether L. bulgaricus 2038 and S. thermophilus 1131 also promoted AMPK activation in normal state. L. bulgaricus 2038 and S. thermophilus 1131 were added to Caco-2 monolayers cultured on Transwell inserts, and the expression of p-AMPK was measured via western blotting. Stimulation with L. bulgaricus 2038 and S. thermophilus 1131 for 9 h significantly increased the p-AMPK expression (Figure 5).

Figure 5. L. bulgaricus 2038 and S. thermophilus 1131 increased the expression of p-AMPK.

Caco-2 monolayers were stimulated with L. bulgaricus 2038 or S. thermophilus 1131 for 9 h. Caco-2 cells were collected and the expression of p-AMPK was evaluated via western blotting. Expression of p-AMPK was normalized to total AMPK expression, and the relative expression levels are shown. Data are expressed as the mean ± SE (n = 6–7). *P < 0.05.

3.6. The activity against barrier destruction by TNF-α and IFN-γ is strain-dependent

To assess whether the protective activity against barrier destruction by TNF-α and IFN-γ was specific for L. bulgaricus 2038 and S. thermophilus 1131, the activities of other strains of various species were evaluated. The results showed that the activities were strain-dependent (Figure 6). L. bulgaricus 2038, S. thermophilus 1131, B. longum ME-885 and L. bulgaricus ME-876 significantly decreased FD-4 permeability compared to TNF-α and IFN-γ group, while several other strains such as L. paracasei ME-881, P. freudenreichii ME-886, and L. lactis ME-883 showed no protective activity. Among them, L. bulgaricus 2038 showed the strongest activity.

Figure 6. Protective effect on FD-4 permeability increased by TNF-α and IFN-γ was strain-dependent.

Bacteria, TNF-α, and IFN-γ were simultaneously added to Caco-2 monolayers. Permeability of FD-4 after 18 h is shown. Data are expressed as the mean ± SE (n = 3). *P < 0.05 compared to TNF-α and IFN-γ group. Dashed line shows the values for the control group. L.b, L. bulgaricus; S.t, S. thermophilus; L.c, Lacticaseibacillus paracasei; P.f, Propionibacterium freudenreichii; L.l, Lactococcus lactis subsp. lactis; L.g, Lactobacillus paragasseri; B.b. Bifidobacterium bifidum; B.l., Bifidobacterium longum.

4. Discussion

In the present study, we used a barrier destruction model with Caco-2 cells induced by TNF-α and IFN-γ. TNF-α and IFN-γ levels are elevated in patients with IBD and increase the intestinal barrier permeability.^32,33 Moreover, TNF-α and IFN-γ synergistically inhibit the barrier function.^26,27 Therefore, this model may be appropriate as an in vitro leaky gut model. We used both live and heat-treated bacteria in this study to clarify whether the cell components are responsible for the activity, or the metabolite or some unknown mechanism due to the live status of the bacteria are required for the efficacy. The results showed that live L. bulgaricus 2038 and S. thermophilus 1131 significantly ameliorated the impairment of the intestinal barrier function induced by TNF-α and IFN-γ. In contrast, heat-treated strains significantly suppressed the increase in FD-4 permeability by TNF-α and IFN-γ, but did not significantly ameliorate the reduced TEER value. Although the TEER value is an indicator of barrier function, changes in the expression of leaky-type TJ proteins, such as CLDN-2, influence the apparent value, because leaky-type TJ proteins are involved in paracellular transportation of small ions, but not large substances that induce inflammation such as LPS. Therefore, it is conceivable that FD-4 permeability reflects actual gut leakage. This observation suggests that the intake of both strains prevents the permeability of harmful substances through the small intestine, even non-viable status. Furthermore, treatment with both strains maintained the structure of TJ proteins, despite the addition of TNF-α and IFN-γ. However, these experiments could not determine whether both strains repaired the damaged TJ structure or suppressed the inflammatory response, leading to barrier dysfunction. To evaluate the former possibility, we conducted a calcium switch assay and found that both strains promoted the assembly of ZO-1. This result suggests that L. bulgaricus 2038 and S. thermophilus 1131 can repair the damaged intestinal barrier.

TJ integrity is necessary to maintain intestinal barrier function. AMPK activation promotes TJ expression^34,35 and assembly by phosphorylating TJ proteins.^{29,30,36–38}. Akkermansia muciniphila ameliorates LPS-induced intestinal barrier dysfunction by activating AMPK.³⁹ This report also showed that a TLR2 inhibitor canceled AMPK activation by A. muciniphila. Therefore, we evaluated whether L. bulgaricus 2038 and S. thermophilus 1131 reinforced the intestinal barrier function by activating AMPK. We found that L. bulgaricus 2038 and S. thermophilus 1131 ameliorated the decrease in TJ mRNA expression induced by TNF-α and IFN-γ, and promoted the expression of p-AMPK. Moreover, in the calcium switch assay, dorsomorphin suppressed the induction of TJ protein assembly in both strains. These results suggest that AMPK activation is the main factor affecting the activities of L. bulgaricus 2038 and S. thermophilus 1131. In human fetal small intestine, TLR2 is expressed on the surface of the villus epithelium.⁴⁰ Therefore, ingested L. bulgaricus 2038 and S. thermophilus 1131 are expected to exert their activities in the lumen of the human small intestine. In a previous report,³⁹ the signaling pathway responsible for the recognition by TLR2 and AMPK activation also remained unclear. Liver kinase B,⁴¹ Ca²⁺/calmodulin-dependent protein kinase beta,⁴² and transforming growth factor beta-activated kinase 1 (TAK1)^43,44 can phosphorylate AMPK. TAK1 is activated by multiple TLRs.⁴⁵ Therefore, we deduced that L. bulgaricus 2038 and S. thermophilus 1131 promote the intestinal barrier function via the TLR2–TAK1–AMPK axis. Further studies using inhibitors of TLR2 and TAK1 or gene engineering are needed to clarify the signaling pathway responsible for this effect.

The protective activity against FD-4 permeability induced by TNF-α and IFN-γ was found to be strain-dependent. PG and LTA structures differed according to the species or strain.^46,47 It is possible that these different structures influence recognition by TLR2, although the mechanism underlying this difference is yet to be elucidated. L. bulgaricus 2038 and S. thermophilus 1131 significantly decreased FD-4 permeability. Moreover, among the bacteria used in the experiment, L. bulgaricus 2038 showed the strongest activity. Therefore, the intake of a yogurt fermented with L. bulgaricus 2038 and S. thermophilus 1131 may be a potential strategy for the prevention and repair of leaky gut. We aim to further investigate whether a yogurt intake promotes human intestinal barrier function in a future clinical study.

In conclusion, we have shown that both live and heat-treated L. bulgaricus 2038 and S. thermophilus 1131 ameliorate intestinal barrier dysfunction by modulating TJ integrity in Caco-2 cells. Moreover, AMPK activation was responsible for the activities of L. bulgaricus 2038 and S. thermophilus 1131. These observations suggest that the intake of L. bulgaricus 2038 and S. thermophilus 1131 is a potential strategy for the prevention and treatment of leaky gut, which results in various diseases. However, the mechanism of recognition of L. bulgaricus 2038 and S. thermophilus 1131 by Caco-2 cells and their activities in humans need to be investigated further in future studies.

Author contributions

Conceptualization: K.K. and T.S. Experimental method design: K.K. Experiments with Caco-2 cells: J.M. Western blotting experiments: J.M. and F.Y. Data analysis: K.K. Manuscript writing: K.K. Manuscript editing: T.S.

Acknowledgments

The authors would like to thank Editage (www.editage.jp) for English language editing.

Disclosure statement

All authors are employees of Meiji Co. Ltd.

Supplemental data

Supplemental data for this article can be accessed online at https://doi.org/10.1080/21688370.2023.2184157

Additional information

Funding

This study received no external funding.