Bifidobacterium adolescentis Alleviates Liver Steatosis and Steatohepatitis by Increasing Fibroblast Growth Factor 21 Sensitivity

8-week-old male C57BL/6J mice purchased from the GemPharmatech Company (Nanjing, China) were housed in a 12-h light-dark cycle in specific pathogen-free (SPF) facilities. The mice were subjected to normal chow diet (Research Diets, containing 10% of total calories from fat) or choline-deficient high fat diet (CDHFD; Research Diets, containing 45% of total calories from fat) and water for 16 weeks. From the start of the 17^th week, the mice were divided into four groups as follows: a normal chow diet control group (Control), choline-deficient high fat diet group (CDHFD-water), CDHFD + live B. adolescentis group (CDHFD-B.a), and CDHFD + heat-killed B. adolescentis group (CDHFD-hk B.a). Drinking water supplemented with B. adolescentis (DSMZ 20086) was administered to mice in the CDHFD-B.a group, and the same amount of heat-killed B. adolescentis (121°C under 225-kPa pressure for 15 min) was added in water administered to mice in the CDHFD-hk B.a group. According to the preliminary experiments, effective dose of B. adolescentis was 4×10¹⁰ CFU/day and mice would drink 8 ml water per day. The concentration of B. adolescentis in drinking bottles was 5×10⁹CFU/mL to ensure a dose of 4×10¹⁰ CFU/day to each mouse. Bottles were changed once a day to assure the activity of bacteria as reported (10). Water was supplied to mice in bottles with leak prevention to accurately measure the consumption of water and bacteria. The treatment lasted for 8 weeks. At the end of the treatment period, all mice were sacrificed after 6 h of fasting. Heart blood and tissue samples were obtained. Part of the tissues were fixed for histological analysis, and the remaining tissues were immediately snap-frozen and stored at -80°C for further investigations. All animal experimental protocols were approved by the Animal Ethics Committee of the Shanghai Jiao Tong University Affiliated Sixth People’s Hospital.

Blood samples were allowed to stand for 1 h at room temperature and centrifuged at 4000 rpm for 20 min to separate the upper serum. Serum lipopolysaccharide (LPS) was measured by the Limulus Amebocyte Lysate assay (Hycult Biotech). Serum levels of triglyceride (TG), alanine aminotransferase (ALT), and aspartate aminotransferase (AST) were measured by commercial assay kits (Nanjing Jiancheng). All measurements were performed in accordance with the manufacturer’s instructions.

Hepatic TGs were quantified using Folch extraction (18). Liver samples from mice were sectioned to chloroform-methanol (2:1) to extract lipids. The organic extract was then dried and reconstituted in isopropanol. The hepatic TG levels were determined using the assay used for determining the serum TG levels. The concentration of hepatic FGF21 was measured using a Mouse FGF-21 ELISA Kit (Immunodiagnostics Limited). All measurements were performed in accordance with the manufacturer’s instructions.

Liver and ileum tissues were acquired when mice were sacrificed and using 4% paraformaldehyde to fix. After at least 48 h, these tissues were subsequently embedded in paraffin and then sectioned for further hematoxylin-eosin (HE) staining, Masson staining, immunofluorescent staining, or immunohistochemistry staining. For immunofluorescent staining, nonspecific protein binding was blocked by 5% normal goat serum. Ileum sections were incubated with anti–zona occuldens protein-1 (ZO-1, Abcam) or occludin (Abcam) antibodies. Then, samples were incubated for 1 hour with FITC-conjugated secondary antibodies (Life Technologies) and counterstained with 4,6-diamidino-2-phenylindole for visualization of cell nuclei. For immunohistochemistry staining, after the endogenous peroxidase activity had been inhibited by hydrogen peroxide (H₂O₂) for 20 min, sections were incubated overnight with anti-E. coli LPS antibody (Abcam), followed by staining with horseradish peroxidase–conjugated secondary antibody.

Three to six regions from each slide with immunofluorescent staining or immunohistochemistry staining were selected and scanned using the 10x objective by the Vectra imaging system (Perkin Elmer). The images acquired were analyzed with inForm software (v2.4.8, Akoya) for the calculation of histochemistry score (H-score).

NAFLD Activity Score (NAS) (19) of liver sections with HE staining were evaluated by two experienced pathologists independently, based on 4 semi-quantitatively histological features: steatosis (0-3), lobular inflammation (0-2), hepatocellular ballooning (0-2), and fibrosis (0-4). NAS from each pathologist was acquired by summing the steatosis, lobular inflammation, and hepatocellular ballooning scores unweighted. Final NAS of a mouse was acquired by averaging two scores from the pathologists.

Stool was lyophilized with a Freezemobile12XL instrument (VirTis) at −77 °C. Following this, the caloric content of stool samples was measured by an isoperibol calorimeter instrument (HWR-15E, Shanghai Institute of Measurement and Testing Technology).

The intestinal barrier function was measured using in vivo gut permeability assays as previously described (20). Mice were fasted for 6 h, then FITC-labeled dextran (DX-4000-FITC, Sigma-Aldrich) were administered by gavage with a dose of 500 mg/kg body weight. 1 h later, blood samples were collected through tail vein. The measurement of serum concentration of DX-4000-FITC was conducted by a fluorescence spectrophotometer (SpectraMax i3x, Molecular Devices) with an excitation wavelength of 485 nm and an emission wavelength of 535 nm.

As described previously (21, 22), the following steps were performed to collect the mucosa-adhered microorganisms: the gut (from jejunum to rectum) was first separated from mice, and the feces in lumina was removed by squeezing. Then, intestine was sectioned longitudinally and the material adhered to the superficial layer was obtained by scraping the superficial layer of the with a sterile cotton swab. The content was transferred to sterile microtubes with DNA stabilizer and then stored at −80°C before DNA extraction.

About 1 g of stool sample was collected in sterile tubes and frozen at −80°C before further experiments. DNA from the fecal samples or mucosa samples was isolated using the PSP Spin Stool DNA Kit (STRATEC Molecular) according to the manufacturer’s instructions. A microvolume spectrophotometer (NanoDrop, Thermo Fisher Scientific Life Sciences) was used for the measurement of nucleic acids concentration.

The hep1-6 cell line was used to measure whether LPS treatment would change the expression level of FGFR1 and KLB. The cells were cultured in DMEM medium with 10% FBS, 100 U/ml penicillin, and 100 mg/ml streptomycin at 37°C in a 5% CO₂ humidified incubator. The cells were plated in six-well tissue culture dishes. When cultured to 70–80% confluency, cells were then treated by different concentration of LPS (0ng/mL,10ng/mL, 20ng/mL,50ng/mL, and 100ng/mL) for 12h, and harvested for the subsequent measurement of the mRNA level of FGFR1 and KLB.

Total RNA was extracted with the TRIzol method using standard TRIzol RNA extraction protocol. TRIzol reagent (Thermo Fisher Scientific Life Sciences) was used to homogenized liver samples to isolated total RNA. A microvolume spectrophotometer (NanoDrop, Thermo Fisher Scientific Life Sciences) was used to determin RNA purity and concentration. RNA (1 μg) was reverse transcribed to generate cDNA using the PrimeScript RT reagent Kit (Takara) in accordance with the manufacturer’s protocol.

The gene transcript levels and bacterial DNA quantification were assessed using real-time polymerase chain reaction (RT-PCR). Sequence-specific primers used in the study were shown in Table 1. SYBR Green PCR Master Mix (Roche) was used for amplification reactions. RT-PCR were performed in a Light Cycler 480 system (Roche) following the manufacturer’s instructions. The cycle threshold (Ct) values of the target genes determined by the system were normalized to the Ct value of the endogenous control gene glyceraldehyde 3-phosphate dehydrogenase (GAPDH), and whereas those of B. adolescentis DNA were normalized to those of total bacterial DNA. Relative changes were calculated using the 2^−ΔΔCt method.

Liver tissue proteins were extracted using lysis buffer containing protease inhibitors (ST505, Beyotime Biotechnology) and phosphatase inhibitors (P1082, Beyotime Biotechnology). Protein was loaded onto sodium dodecyl sulfate polyacrylamide gel (SDS-PAGE) to separate by electrophoresis. After the target protein was separated, the protein would be transferred onto polyvinylidene fluoride (PVDF) membrane (Millipore) and blocking in 5% bovine serum albumin (BSA), followed by incubation with primary antibodies at 4°C overnight. On the next day, secondary antibodies would be incubated with the membrane. Immunoreactivity was detected using enhanced chemiluminescent autoradiography (Millipore). Chemiluminescence was determined using ChemiDoc Imageing System (BIO-RAD) or Optimax X-ray Film Processor (Protec GmbH & Co. KG). The Image J software (National Institutes of Health, USA) was used for band quantification. The primary antibodies used for western blotting were against Phospho-NF-κBp65 (Ser536) (3033, Cell Signaling Technology), NF-κBp65(8242, Cell Signaling Technology), KLB (AF2619, R&D), FGFR1 (12511, Cell Signaling Technology), extracellular signal−regulated protein kinase 1/2 (Erk1/2, 9102, Cell Signaling Technology), phosphorylated (p)- Erk1/2 (Thr202/Tyr204) (9101, Cell Signaling Technology), and glyceraldehyde 3-phosphate dehydrogenase (GAPDH,5174, Cell Signaling Technology).

After 8-week intervention, mice were conducted theFGF21 response test (Figure 4G). Under a general anesthetic, left lobe liver of mice was ligated and part of it were sectioned and flash frozen with liquid nitrogen. Then 2 mg/kg recombinant mouse FGF21 (rmFGF21) or vehicle was injected to mice through inferior vena cava. 15min later, part of liver from right lobe were collected. Protein from these liver samples would be used to detect the level of Erk1/2 phosphorylation.

All statistical analyses were performed using GraphPad Prism (GraphPad Software). Results were presented as mean ± S.E.M. (standard error of the mean). One-way analysis of variance with the consequent post hoc test of Fisher’s LSD was applied. A p-values of <0.05 was considered statistically significant.

Eight-week-old male C57BL/6J mice on control diet or CDHFD were allocated to treatment of drinking water (control; CDHFD-water), water with B. adolescentis (CDHFD-B.a), or water with heat-kill B. adolescentis for 8 weeks (Figure 1A), followed by further assessments. The supplementation of B. adolescentis in drinking water did not alter food intake nor body weight, as well as the water consumption (Supplementary Figures 1A, B; Figure 1B). And we measured the energy content in feces to investigate whether the absorption of the nutrients was affected. No significant difference was found among the four groups (Supplementary Figure 1C). Hepatic pathology showed that compared to control diet, CDHFD-induced massive accumulation of fat and mild fibrosis in mice (Figures 1C, D). Supplementation with living B. adolescentis attenuated these lesions and reduced the NAS (Figures 1C, D). In accordance with the histology findings, B. adolescentis abrogated the increase of circulating and hepatic levels of TGs induced by CDHFD (Figures 1E, F). The lower levels of AST and ALT in CDHFD-B.a group indicated a reversal of diet-induced liver damage (Figures 1G, H). The heat-killed B. adolescentis also show some benefits on NAS, hepatic TG, and serum AST. But it’s effect is weaker than live B. adolescentis (Figures 1D, E, G).

The B. adolescentis content in fecal and gut mucosa after treatment was measured by the real-time PCR analysis. 4.01*10¹⁰ CFU/day of B. adolescentis intake from drinking water increased its amount both in feces and in gut mucosa (Figures 2A, B). Another two prominent microbes were not changed (Supplementary Figures 1D, E). Changes in gut permeability have been associated with NAFLD development. In particular, the increase in gut permeability lead to the entry of gut-derived metabolites into circulation, such as ethanol and LPS, which affect NAFLD development (23). To investigate whether B. adolescentis administration could decrease intestinal permeability, an in vivo gut permeability assay was performed. 1 hour after gavage of fluorescent-labeled dextran (DX-4000-FITC), its serum level in CDHFD-water group was twice as high as that in control group, indicating an increase in intestinal permeability (Figure 2C). Such an increase in intestinal permeability was significantly abolished by supplementation with live B. adolescentis (Figure 2C). The gut permeability was largely regulated by the tight junctions between epithelial cells of intestine, which form a physical barrier defending the intrusion of pathogens and bacterial products. Therefore, we wonder whether B. adolescentis could affect the level of epithelial tight junction protein ZO-1 and occludin. As immunofluorescent staining of ileum sections showed (Figures 2D, E), mice on CDHFD have lower ZO-1 than their counterparts on control diet, whereas those with B. adolescentis have similar level with control mice. The level of occludin showed the same trend as ZO-1 although not statistically different (Figure 2F). In line with the change in gut permeability, serum LPS concentration was higher in mice on CDHFD, whereas B. adolescentis treatment could decreased the diet induced elevation of LPS (Figure 2G).

To investigate whether the change in circulation LPS would affect liver, we measured the localization of gut-derived LPS in liver and by quantifying the H-score of LPS we discovered a similar trend with serum LPS (Figures 3A, B). LPS infiltration in liver has been shown to increase liver damage through a TLR4-mediated pathway (24). As the main receptor of LPS, the levels of TLR4 and CD14 were upregulated in the CDHFD-water group compared to those of control group (Figures 3C, D). The activation of TLR4/NF-κB signaling pathway in CDHFD mice was confirmed by the increased phosphorylation level of a NF-κB subunit p65 (Figures 3E, F). However, supplementation with live B. adolescentis abolished the CDHFD-induced upregulation of the TLR4 and CD14 (Figures 3C, D). And the phosphorylation level of NF-κB was also decreased by B. adolescentis treatment (Figures 3E, F). Collectively, a suppression of LPS/TLR4/NF-κB pathway was discovered in mice supplemented with B. adolescentis.

LPS injection has been shown to suppress FGF21 signaling in mice (25). To investigate if circulating LPS can alter the signaling of the hepatic FGF21 pathway, we examined the expression of FGF21, as well as FGFR1 and KLB in the liver. In accordance with previous findings, compared to control diet, CDHFD induced a higher expression of FGF21(Figures 4A, B). Despite that CDHFD increased the level of FGF21, the expression of FGF21 receptors was downregulated (Figures 4C–G), which indicate a potent FGF21 resistant state in these mice. By supplementation with B. adolescentis, the CDHFD induced FGF21 elevation was reduced (Figures 4A, B), accompanied by the upregulation of receptors (Figures 4C–G). To measure the response to FGF21 in CDHFD-B.a group, we administered recombinant mouse FGF21 (rmFGF21) or vehicle to mice (Figure 4H), and detected the change of Erk1/2 phosphorylation. As shown in Supplementary Figures 1F, G, the technique itself could not produce the downstream phosphorylation of Erk1/2. And the level of Erk1/2 phosphorylation was significantly higher in CDHFD-B.a group than CDHFD-water group or CDHFD-hk B.a group in response to FGF21 (Figures 4I, J). It confirmed that treatment with live B. adolescentis increase the liver FGF21 sensitivity in mice. To testify that B. adolescentis increased liver FGF21 sensitivity by reducing LPS, we treated the hep1-6 cells with different concentration of LPS. Correspondently, the expression of FGFR1 and KLB was suppressed by LPS in a dose-dependent manner (Figures 4K, L).

The original contributions presented in the study are included in the article/. Further inquiries can be directed to the corresponding authors. 1