Lactobacillus improves the effects of prednisone on autoimmune hepatitis via gut microbiota-mediated follicular helper T cells

Autoimmune hepatitis (AIH) is a chronic, immune-mediated liver dysfunction. The gut microbiota and T follicular helper (Tfh) cells play critical roles in the immunopathogenesis and progression of AIH. We aimed to investigate the effect of gut microbiota combined with prednisone therapy on Tfh cell response in AIH. Samples from AIH patients and mouse model of experimental autoimmune hepatitis (EAH) were analyzed using real-time quantitative polymerase chain reaction, enzyme-linked immunosorbent assay, western blotting, flow cytometry, and hematoxylin–eosin staining to determine the role of gut microbiota on AIH. Lactobacillus significantly increased the levels of Bacteroides fragilis, Clostridium, Clostridium leptum, Bifidobacterium, and Lactobacillus and significantly enhanced the suppressive effects of prednisone on the levels of AIH clinical indexes in AIH patients. Lactobacillus exerts the same prptective effects as prednisone in EAH mice and enhanced the effects of prednisone. Lactobacillus also reinforced the inhibitory effects of prednisone on the levels of serum IL-21 and the proportions of Tfh cells in peripheral blood mononuclear cells. Mechanistically, prednisone and Lactobacillus regulated Tfh cell response in EAH mice in an MyD88/NF-κB pathway-dependent manner. Our results suggested a therapeutic potential of Lactobacillus in the prednisone-combined treatment of AIH.6BN4VU1A1WC6Nqr62iBRGo Video Abstract Video Abstract

immune responses during AIH progression. However, the underlying mechanism of Tfh cell dysregulation in AIH remains to be determined.
The gut-liver axis means the bidirectional relationship between the liver and the gut and its microbiota and is recently confirmed in the clinical settings of numerous diseases [12,13]. The role of gut microbiota in the regulation of AIH has received extensive attention [1,14]. Yiran et al. reported the altered composition and function of gut microbiome in AIH, potentially indicating gut microbiota as a noninvasive biomarker to assess disease activity [1]. Moreover, the data from humanized murine models showed that gut microbiota is engaged in the pathogenesis of AIH [15]. Gut microbiota can transform glucocorticoids [16] and regulate host's behaviors [17]. Various reports have elucidated the relationship between gut microbiota and Tfh cells. Gut microbiota has been reported to regulate arthritis, an autoimmune disease, through Tfh but not Th17 cells [18]. Teng et al. showed that Peyer's patch Tfh cells are essential for gut-residing segmented filamentous bacteria-induced systemic arthritis despite the production of auto-antibodies predominantly occurring in lymphoid tissues [19]. Given the vital role of Tfh cells in the pathogenesis of AIH, questions have emerged concerning the effect of gut microbiota on Tfh cells in the pathogenesis of AIH.
Probiotics are living microbiota, and their administration confers health benefits to the host by altering the gut microbiota composition [20]. VSL#3, a mixture of 8 probiotic bacteria, is able to alleviate dextran sulfate sodium-induced colitis by downregulating Tfh cells [21]. Arai et al. showed that oral administeration of probiotic Lactobacillus paracasei MCC1849 increases the proportion of IgA( +) B cells and Tfh cells in Peyer's patches in mice [22]. In this study, we investigated the effect of prednisone and prednisone + Lactobacillus on Tfh cell response in AIH. Lactobacillus improved the treatment effects of Prednisone on alleviating the symptoms of AIH patients and EAH mice by inhibiting Tfh cell response.

Participants and treatment
Fifty patients with active AIH were recruited at the First People's Hospital of Changzhou. All patients were diagnosed according to the AIH simplified diagnostic score system proposed by the International Autoimmune Hepatitis Group in 2008 and the AIH diagnosis and treatment guidelines updated by the American Association for the Study of Liver Diseases in 2010. All included patients were in the active phase of the AIH and were not under pharmacological intervention. The excluded patients are listed as follows: (1) patients who have taken antibiotics or probiotics that may affect the microbial composition of the intestine in the past 6 months; (2) patients with other combined liver diseases or extrahepatic diseases and those with poor compliance or in critical condition; (3) those who have consumed spicy and irritating foods in the past week. The experimental protocol was conformed to the 1975 Declaration of Helsinki (6th revision, 2008) and was approved by the Ethical Committee of the First People's Hospital of Changzhou and the Third Affiliated Hospital Soochow University. Informed consent was obtained from each participant.
All patients were randomly divided into the following groups: group I, prednisone, and group II, prednisone + Lactobacillus. On days 3 and 7 after treatment, fecal pellets and serum samples were collected and subjected to gut microbiota and clinical index analysis.

Preparation of liver antigen
Liver antigens were freshly prepared from female C57BL/6 mice after liver perfusion with phosphatebuffered saline (PBS). Livers were homogenized on ice. Nuclei and remaining intact cells were centrifuged at 150 g for 10 min. Next, the supernatants were centrifuged at 100,000 g for 1 h. The remaining supernatants, called S-100, is used for immunization.

Establishment of experimental autoimmune hepatitis (EAH) model
Female C57BL/6 mice (6-8 weeks old) were purchased from Nanjing Experimental Animal Center (Jiangsu, China). CD103-deficient (CD103 − / −) and MyD88deficient (MyD88 − / −) mice were obtained from The Jackson Laboratory (Bar Harbor, ME). All animal experiments were approved by the Ethical Committee of the First People's Hospital of Changzhou and the Third Affiliated Hospital of Soochow University.
The EAH model was established as described previously [11,23]. In brief, freshly prepared S-100 (0.5-2 mg/ mL) was dissolved in PBS, emulsified in complete Freund's adjuvant, and intraperitoneally injected into mice. A booster dose of S-100 was given seven days later. The mice were randomly divided into five groups with ten mice in each group. Mice in group I (sham) served as controls and were injected with normal saline. Mice in group II (EAH) were intraperitoneally injected with S-100. Mice in group III (EAH + prednisone) were intraperitoneally injected with S-100 followed by intragastrical administration of prednisone (0.5 g/kg/day; Sigma-Aldrich). Mice in group IV (EAH + Lactobacillus) were intraperitoneally injected with S-100 followed by intragastrical administration of Lactobacillus (0.66 g/day; LeTuoEr, France). Mice in group V (EAH + prednisone + Lactobacillus) were intraperitoneally injected with S-100 followed by cotreatment of prednisone (0.5 g/kg/day) and Lactobacillus (0.66 g/day) by gavage. The intervention was administered one day after the model establishment. Seven days after intervention treatment, the fecal pellets, blood, liver tissues, and colon tissues of mice in the five groups were collected. Blood was collected from mice by eyeball extraction. Approximate 0.6 mL blood was collected from each mouse.
The effects of DCs and MyDD88 in gut microbiotamediated Tfh cell response in EAH were determined in CD103 − / − or MyD88 − / − mice, which have been confirmed to be sensitive to S-100 stimulation (Additional file 1: Figure S1). CD103 − / − or MyD88 − / − mice were intraperitoneally injected with S-100, intragastrically administrated with prednisone (0.5 g/kg/day), or intragastrically administrated with prednisone (0.5 g/kg/day) and Lactobacillus (0.66 g/day). Normal C57BL/6 mice that received the same procedures served as the controls.

Real-time quantitative polymerase chain reaction (RT-qPCR) for 16S rRNA
RT-qPCR was conducted to reveal the levels of Bacteroides fragilis, Clostridium, Clostridium leptum, Bifidobacterium, and Lactobacillus, which are dominant microflora in the gut, greatly affecting the function of the whole flora and determining the physiological and pathological significance of the flora to the host. Total DNA from fecal pellets of patients and mice was isolated using the QIAamp Fast DNA stool mini kit (Qiagen) following the manufacturer's protocols. RT-qPCR primers utilized in the present experiment were as follows: Bacteroides fragilis, forward

Clinical index assay
The levels of alanine transaminase (ALT), aspartate aminotransferase (AST), total bilirubin (TBIL), anti-nuclear antibody (ANA), smooth muscle antibody (SMA), Immunoglobulin G (IgG), Immunoglobulin M (IgM), and Immunoglobulin A (IgA) in serum of patients were determined using an automatic biochemistry analyzer according to the manufacturers' instructions (Beckman Coulter). Peripheral blood was collected by eyeball extraction from mice, and serum ALT, AST, TBIL levels were measured using the same automatic biochemistry analyzer.

Dection of endotoxin (ET) and diamine oxidase (DAO)
A ToxinSensor ™ Chromogenic LAL Endotoxin Assay Kit (#L00350; GenScript) was used to detect serum ET levels in human and mice. Human and mouse (#CSB-E10137h and #CSB-E10090m; Beijing BioDee BioTech Corporation Ltd) diamine oxidase ELISA Kits were used to detect serum DAO activity.

Hematoxylin-eosin (H&E) staining
Liver and colon tissues were collected from the EAH model for H&E staining. H&E staining was performed according to our previous report [23]. In brief, liver and colon tissues were harvested, fixed with formalin, embedded in paraffin, and cut into 5 μm sections. After staining with H&E, pathological changes of the samples were observed using a light microscopy.

Cell clustering
Clustering of CD4 + cells was performed by Seurat (v1.4.0.8) [24] in a stepwise manner. The Seurat "Find-AllMarkers" function was used to find markers for the identified clusters. We initially performed low-resolution clustering and then annotated CD4 + CXCR5 + ICOS + clusters as Tfh cells.

RT-qPCR analysis
Total RNA from liver tissues was extracted using TRIzol reagent (TaKaRa) according to the manufacturers' instructions. A cDNA Reverse Transcription Kit (TaKaRa) and an SYBR PrimeScript RT-qPCR Kit (Takara) were used for mRNA expression assessment according to the manufacturer's instructions. The primers' sequences were listed in Table 1. GAPDH served as an internal reference.

Statistical analysis
The one-way analysis of variance and Tukey's post hoc test were used to assess statistical significance among different experimental groups. Data are expressed as the mean ± SEM. P values less than 0.05 were considered significant.

Lactobacillus capsules improves the clinical symptoms of prednisone-treated AIH patients
Patients were orally administrated with prednisone or prednisone + Lactobacillus capsules to determine the clinical effect of Lactobacillus on AIH. As shown in Fig. 1A, the levels of Bacteroides fragilis, Clostridium, Clostridium leptum, Bifidobacterium, and Lactobacillus levels were significantly higher in the fecal pellets of AIH patients who received prednisone + Lactobacillus treatment for three and seven days than that of AIH patients who received prednisone for three and seven days. Serological test results revealed that prednisone treatment significantly decreased the levels of ALT, AST, TBIL, SMA, ANA, IgG, IgM, and IgA. Cotreatment of prednisone + Lactobacillus capsule further reduced the relative levels of ALT, AST, TBIL, SMA, ANA, IgG, IgA, and IgM in the serum of AIH patients (Fig. 1B). Moreover, administration of prednisone + Lactobacillus markedly downregulated serum DAO and ET levels in AIH patients compared with patients who received prednisone treatment (Fig. 1C,  D).

Lactobacillus ameliorates EAH in mice and enhances effects of prednisone
We further used a murine model of EAH to demonstrate the effect of Lactobacillus treatment on AIH progression. The levels of Bacteroides fragilis, Clostridium, Clostridium leptum, Bifidobacterium, and Lactobacillus were significantly higher in the fecal pellets of mice of the EAH + Lactobacillus group than mice of the control, EAH, or EAH + prednisone groups. Levels of these probiotics showed no significant difference between EAH group and EAH + prednisone group ( Fig. 2A). There was a damage of normal intestinal structure in mice of the EAH group compared to normal mice. However, prednisone or Lactobacillus decreased intestinal tract lesions in EAH mice (Fig. 2B). Moreover, Lactobacillus treatment significantly reduced ALT, AST, TBIL, DAO, and ET levels and attenuated liver injury in EAH mice (Fig. 2C-E). There was no significant difference between the effects of Prednisone and Lactobacillus in clinical indexes and liver/intestine injury in EAH mice. We then used the combination of Prednisone + Lactobacillus and identified that Lactobacillus enhanced the effects of Prednisone in reducing DAO and ET levels in mice (Fig. 2E).

Lactobacillus promotes the suppressive effects of prednisone on Tfh response and regulates expression of relative cytokines in EAH mice
Our previous studies have indicated that the uncontrolled accumulation of Tfh cells plays an important role in the pathogenesis of AIH [11,23]. As shown in Fig. 3A, Lactobacillus treatment has markedly enhanced the suppressive effects of prednisone on serum IL-21 levels in EAH mice. The proportions of Tfh cells were significantly decreased in the blood of EAH mice after prednisone + Lactobacillus treatment as compared to prednisone treatment (Fig. 3B). Moreover, the mRNA expression levels of Tfh cell-related cytokines including IL-21, Bcl-6, and CXCR5 were decreased in the liver of EAH mice after prednisone + Lactobacillus treatment compared with prednisone treatment (Fig. 3C). Next, we evaluated the concentrations of six cytokines in the serum of mice. Serum TGF-β, KGF-1, KGF-2, and IL-10 concentrations were decreased, while IL-6 and TNF-α concentrations were increased in EAH mice compared to the control group. Prednisone rescued the EAHinduced changes on concentrations of these cytokines.

Prednisone and Lactobacillus regulate Tfh response in EAH mice through CD103
Previous studies showed that gut microbiota regulates T cell activation and differentiation in a DC dependent manner [26]. Herein, we hypothesized that Lactobacillus inhibits Tfh response in EAH mice by regulating DCs. CD103 is a surface marker for type I conventional DC in mice. The present study thus used CD103 − / − mice for establishment of the EAH model and treated mice with prednisone or prednisone + Lactobacillus. As shown in Fig. 5A, prednisone or combination of prednisone + Lactobacillus downregulated the levels of IL-21 in WT EAH mice but not CD103 − / − EAH mice, indicating that CD103 is essential for prednisone and Lactobacillusregulated IL-21 levels in EAH. More importantly, prednisone or combination of prednisone + Lactobacillus reduced the proportions of Tfh cells in PBMCs of WT mice but did not influence the proportions of Tfh cells in CD103 − / − mice after EAH induction (Fig. 5B, C). Prednisone or cotreatment of prednisone + Lactobacillus suppressed IL-6, increased IL-10, KGF-1, KGF-2 levels in WT EAH mice but caused no significant effects on these cytokines in CD103 − / − EAH mice (Fig. 5D). The effect of prednisone + Lactobacillus on Tfh response was more significant than prednisone alone.

Prednisone and Lactobacillus regulate Tfh response in EAH mice via the TLR4/MyD88/NF-κB pathway
Given that the TLR4/MyD88/NF-κB pathway is a key regulator for gut microbiota-mediated DC activation [27], we explored whether Lactobacillus regulates Tfh response in EAH mice by this pathway. As shown in Fig. 6A, the mRNA levels of TLR4 and MyD88 were significantly increased in EAH mice. However, prednisone treatment reduced TLR4 and MyD88 mRNA expression. Lactobacillus further enhanced the effects of prednisone on TLR4 and MyD88 mRNA expression (Fig. 6A). Furthermore, the inhibitory effects of prednisone on TLR4, MyD88, and p-p65 proteins in EAH mice were promoted by Lactobacillus (Fig. 6B).
Next, we used MyD88 − / − mice to further investigate the effect of MyD88 on gut microbiota-mediated Tfh cell response. As shown in Fig. 7A, prednisone or combination of prednisone + Lactobacillus treatment downregulated the levels of IL-21 in WT mice but did not affect IL-21 expression in MyD88 − / − mice after EAH induction (Fig. 7A). More importantly, prednisone or prednisone + Lactobacillus cotreatment reduced the proportions of Tfh cells in WT mice but did not influence the proportions of Tfh cells in MyD88 − / − mice after EAH induction (Fig. 7B, C). Moreover, prednisone or cotreatment of prednisone + Lactobacillus decreased IL-6 levels, increased IL-10, KGF-1, KGF-2 levels in WT EAH mice. Concentrations of these cytokines in MyD88 − / − EAH mice showed no significant response to prednisone nor prednisone + Lactobacillus (Fig. 7D). The effect of prednisone + Lactobacillus on Tfh response was more significant than prednisone alone.

Discussion
In the last decade, numerous studies indicated that gut microbiota is a key mediator of autoimmune diseases. In Albino Oxford rats, gut microbiota ameliorates the experimental autoimmune encephalomyelitis, a chronic inflammatory disease of the central nervous system, by ). E The levels of DAO and ET in the serum of EAH mice after prednisone, Lactobacillus treatment or prednisone + Lactobacillus cotreatment. The data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 vs sham, # P < 0.05, ## P < 0.01 vs EAH, & P < 0.05 vs EAH + prednisone Fig. 3 Lactobacillus promotes the suppressive effects of prednisone on Tfh response in EAH mice. A The level of IL-21 in the serum of EAH mice by treatment of prednisone or prednisone + Lactobacillus was assessed by ELISA. B The portion of CD4 + CXCR5 + ICOS + Tfh cells in the PBMCs of EAH mice by treatment of prednisone or prednisone + Lactobacillus was detected by flow cytometry analysis. C The mRNA levels of IL-21, Bcl-6, and CXCR5 in the liver of EAH mice by treatment of prednisone or prednisone + Lactobacillus were revealed by RT-qPCR. The data are presented as the mean ± SEM. ***P < 0.001 vs control, # P < 0.05 vs EAH, @ P < 0.05 vs EAH + prednisone Fig. 4 Prednisone and Lactobacillus regulate expression of Tfh cell-related cytokines in EAH mice. The protein levels of TGF-β, KGF-1, KGF-2, IL-10, IL-6, and TNF-α in the serum of EAH mice after treatment of prednisone or prednisone + Lactobacillus were revealed by ELISA. The data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 vs control, # P < 0.05 vs EAH, @ P < 0.05 vs EAH + prednisone regulating IFN-γ and IL-17 productions [28]. Gut microbiota and bacterial translocation promote autoimmune cholangitis by TLR2-independently increased gut permeability [29]. In AIH patients, the distinct microbial composition is associated with disease progression. Notably, the most strongly disease-associated taxa were positively correlated with the serum level of aspartate aminotransferase and liver inflammation [1]. Cai et al. used transgenic AIH mice carrying HLA-DR3 and showed a close relationship between microbiota and AIH [14]. In the current study, we demonstrated that Lactobacillus capsules increased the levels of Bacteroides fragilis, Clostridium, Clostridium leptum, Bifidobacterium, and Lactobacillus in AIH patients, and significantly decreased the levels of ALT, AST, TBIL, SMA, ANA, IgG, IgM, IgA, DAO, and ET in the serum of patients with AIH who received prednisone therapy. Both prednisone and Lactobacillus were found to alleviate AIH symptoms in EAH mice. Moreover, prednisone + Lactobacillus exerted more significant effects than prednisone on DAO and ET levels, suggesting that altering gut microbial composition by Lactobacillus may improve the therapeutic effect of prednisone in AIH.
Previous studies have reported the crucial roles of Tfh cells in autoimmune diseases [30]. Alcohol and its metabolite acetate promote autoimmune arthritis by altering the functional state of Tfh cells in vitro and in vivo [31]. Faliti et al. showed that P2X purinoceptor 7 receptor restrains pathogenic Tfh cell generation in systemic lupus erythematosus [32]. Our team has also previously reported the importance of Tfh cells in AIH patients and EAH mice [11,23]. Dysregulation between T follicular regulatory cells and Tfh cells causes excessive autoantibody production and immune homeostasis destruction, leading to the immunopathological process in AIH patients [11]. Moreover, we have explored the significance of Tfh cell-related molecules on C57BL/6 mice with EAH [23]. Gut microbiota has been reported to regulate the biological functions of Tfh cells in autoimmune diseases [18,19]. Therefore, we explored how gut It is well known that gut microbiota regulates T cell activation and differentiation in a DC dependent manner [26]. For example, gut microbiota modulates Mincle-Syk Axis in DCs, and regulates the productions of IL-17 and IL-22 by CD4 + T cells, promoting intestinal barrier integrity [33]. Herein, we hypothesized that Lactobacillus inhibits Tfh response in EAH mice by regulating DC cells. We used CD103 − / − mice and found that prednisone or prednisone + Lactobacillus decreased the levels of serum IL-21, the proportions of Tfh cells as well as Tfh cell-related cytokines in WT mice, but had no influence on the Tfh response in CD103 − / − mice after EAH induction. Moreover, effect of prednisone + Lactobacillus was more significant than prednisone. These results indicated that Lactobacillus enhanced the inhibitory effects of prednisone on Tfh response in EAH mice via CD103. However, depletion of CD103, a surface marker for type I conventional DC in mice, is not completely equal to depletion of DCs. More studies are required to directly reveal the role of DCs in Lactobacillus-mediated Tfh response in EAH mice.
Intestinal DCs recognize gut microbiota through TLRs, leading to the activation of the MyD88/NF-κB pathway and regulation of T cell activation and differentiation. Liang et al. showed that functionally specialized DCs promote inflammatory Th17 via MyD88 [34]. Therefore, we investigated if Lactobacillus regulates Tfh response in EAH mice via the MyD88/NF-κB pathway. Our results indicated that prednisone or prednisone + Lactobacillus treatment downregulated the levels of TLR4, MyD88, and p-p65 in EAH mice. More importantly, we found that prednisone or prednisone + Lactobacillus decreased serum IL-21, the proportions of Tfh cells, and Tfh cell-related cytokines in WT mice but had no influence on the Tfh response in MyD88 − / − mice after EAH induction. These results indicated that prednisone and Lactobacillus might inhibit Tfh response in EAH mice via the MyD88/ NF-κB pathway. However, our study possesses some limitations. We did not investigate other critical signaling pathways in DCs, such as the TRAM/TRIF pathway, and their roles in the Tfh response in AIH [35]. Further in-depth research is still needed. in EAH mice by treatment of prednisone + Lactobacillus were detected by RT-qPCR. B The protein levels of TLR4, MyD88, p65, and p-p65 in EAH mice by treatment of prednisone + Lactobacillus were revealed by western blotting analysis. The data are presented as the mean ± SEM. **P < 0.01, ***P < 0.001 vs control, # P < 0.05, ## P < 0.01, ### P < 0.001 vs EAH, @ P < 0.05 vs EAH + prednisone