For generation of fPue, Pueraria L. was purchased from Omniherb Inc.; 400 g of that was crushed and boiled in 4 L of water at 121℃ for 30 minutes and the aqueous extract was filtered using Whatman filter paper 2 (GE Healthcare, Little Chalfont, United Kingdom). Dextrose (2%) and yeast extract (1%) were autoclaved and added to the Pue extract with 1% Lactobacillus brevis, aD110^T, and the Pue extract was incubated at 30℃ for 5 days. Following incubation, the Pue extract was centrifuged at 12,000 rpm for 15 minutes, and the supernatant was filtered. The filtered extract was concentrated at 40℃ and 80 rpm, using a rotary evaporator (Eyela, Irvine, CA).

Specific pathogen-free female ICR mice (5 weeks old) were purchased from Orient Inc., Gyeonggi-do, Korea and the mice were acclimatized for 1 week prior to the experiment. All mice were maintained under 12 hours light/dark cycles at 22±5℃ and 60±5% humidity and provided with food (Orient Inc.) and water ad libitum. The information about components and calorie of the food is supported by Orient Inc. All experiments with animals were approved by the Institutional Animal Care and Use Committee at Lee Gil Ya Cancer and Diabetes Institute (IACUC No., LCDI-2014-0010).

Raw264.7 and Caco-2 cell lines were purchased from Korea Cell Line Bank, Seoul, Korea. The cell lines were cultured in 10% FBS-supplemented DMEM (WelGene, Daegu, Korea) with streptomycin (100 µg/mL)/penicillin (100 units/mL) at 37℃ in a humidified chamber with 5% CO₂. Raw264.7 and Caco-2 cells were treated with Pue or fPue for western blot, RT-PCR and immunofluorescence analysis. For detection of NF-κB signaling by western blot, Raw264.7 cells were treated with Pue or fPue (0, 25, 100, 400 µg/mL) for 4 hours, followed by treatment with lipopolysccharide (LPS) for 1 hour. Total protein was isolated and western blot was performed. For determination of transcriptional level of cytokines in Raw264.7 cells, the cells were treated with Pue or fPue (0, 25, 100, 400 µg/mL) for 4 hours, followed by treatment with LPS for 3 hours. Total RNA was isolated using TriZol (Invitrogen, Waltham, MA) and RT-PCR was performed. Caco-2 cells were seeded in 12-well plates and cultured with or without a cover slip for immunofluorescence assay or western blot, respectively. The cells were incubated until they become 100% confluent, followed by incubation for an additional 7 days. The cells were treated with fPue (0, 25, 100, or 400 µg/mL) for 48 hours, followed by treatment with 2% DSS for 24 hours. Then, immunofluorescence assay and membrane fractionation for western blot were performed.

The mice were divided into 3 groups (7 mice in each group): (1) no treatment group; (2) 5% DSS-treated group; (3) 5% DSS with fPue (100 mg/kg)-treated group. DSS (MP Biomedical, Burlingame, CA) and fPue were dissolved in distilled water. Five percent DSS was administered to the mice in the drinking water for 7 days and normal distilled water was then supplied for 4 days for recovery (Figure 2A). During the recovery period, the mice were treated with fPue, or distilled water daily by gavage (Figure 2A). At the end of the recovery period, disease activity index (DAI) score was determined as described in the previous report [26] and necropsy was then performed.

The mice were sacrificed at the end of the recovery period. Following excision of the whole colon from anus to cecum, the length was measured. Approximately 0.5 cm of the distal colon was excised and frozen in liquid nitrogen for RT-PCR and histological analysis, and the remaining colon was Swiss-rolled and fixed in 10% formalin. The tissue samples were then paraffin-embedded and sectioned at 5 µm-thickness. The sections were stained with hematoxylin and eosin to evaluate the severity of colitis, according to three parameters described in the previous report [26]. The severity of colitis was scored blindly by 2 researchers.

Twenty micrograms of each protein sample was used for western blot analysis. The primary antibodies are as followed: anti-rabbit ZO-1, anti-mouse Occludin, 1:1,000, anti-mouse Claudin-1 (Invitrogen, Waltham, MA); anti-rabbit p-P65 (Cell Signaling, Danvers, MA), anti-rabbit P65, anti-goat Lamin B (Santa Cruz, CA) and anti-mouse GAPDH (Merck Millipore, Billerica, MA).

Using 3 µg of RNA, cDNA was synthesized using the PrimeScript RT reagent Kit (TaKaRa, Shiga, Japan) as described in the manufacturer's instructions. RT-PCR was performed, using the primers listed on Supplementary Table 1 (Table S1).

The 5 µm-thick frozen colon tissue and cells on the slides were fixed in an acetone-methanol (1:1) mixture at -20℃. The slides were incubated with primary antibodies (anti-rabbit ZO-1, anti-mouse Occludin, anti-mouse Claudin-1; Invitrogen, Waltham, MA) overnight at 4℃ and secondary antibodies (anti-rabbit IgG FITC, Santa Cruz; Alexa Fluor 488 goat anti-mouse IgG; Invitrogen) for 2 hours at room temperature. After mounting the slides with VectaShield Mounting Medium with DAPI (Vector Laboratories, Burlingame, CA), the slides were sealed with nail polish and stored at -20℃ until observation, using a confocal microscope (Nikon Instruments Inc., Melville, NY).

Numerical data for all graphs were expressed as mean±standard deviation. Statistical changes in all data were determined using one-way analysis of variance (ANOVA). Scheffe was used for the post hoc test. The significance level was limited to 5% (P<0.05). SPSS 22.0 was used as a tool for statistical analysis.

To determine whether fPue can enhance anti-inflammatory effect, compared to Pue, the activity of P65, a subunit of NF-κB, was examined in murine macrophage cell line Raw264.7. The cells were treated with Pue or fPue, followed by activation by LPS, which induces phosphorylation of P65 in Raw264.7 cells. Results of western blot showed that treatment with fPue significantly inhibited phosphorylation of P65 while Pue slightly inhibited that of P65 (Figure 1A). Activation of NF-κB leads to upregulated transcription of pro-inflammatory cytokines, including IL-1β, TNFα, IL-6, and IL-12. Treatment of LPS resulted in increased transcriptional levels of IL-1β, TNFα, IL-6 and IL-12 (Figure 1B). On the other hand, when pre-treated with fPue, the expression levels of the cytokines were decreased while treatment with Pue did not affect them (Figure 1B). These results suggest that fPue would improve the anti-inflammatory effect.

Because fPue exerted an anti-inflammatory effect in Raw264.7 cell line (Figure 1), we examined whether fPue could improve DSS-induced colitis. DSS-induced colitis is characterized by disruption of epithelial structure and infiltration of inflammatory cells and loss of goblet cells [27]. To evaluate the effect of fPue on DSS-induced colitis, DAI was assessed by fecal analysis, to examine how much water and blood the fecal pellet contains and histological analysis was performed by H&E staining on the paraffin block sections to assess the severity of colitis. In the control group, DAI score was nearly 0 and that of the DSS-treated group showed a significant increase (Figure 2B), indicating that colitis was induced by DSS. On the other hand, DAI score of the DSS+fPue-treated group was significantly decreased, suggesting that fPue treatment improved the colonic function (Figure 2B). Histological analysis demonstrated that DSS treatment induced severe inflammation in mucosa, edema in submucosa and epithelial cell loss while they were significantly lower in the DSS+fPue-treated group, compared with the DSS+Vehicle-treated group (Figure 2C, 2D). In addition, alcian blue staining showed that while goblet cells were lost in the DSS-treated group, they were restored by fPue treatment (Figure 2E). These results suggest that treatment with fPue resulted in amelioration of colonic inflammation and promoted recovery of epithelial structure in the colonic mucosa, following recovery from diarrhea.

In a DSS-induced colitis model, inflammatory cells, mainly macrophages, are infiltrated into submucosa and lamina propria of the inflamed colon [1628]. To determine whether the number of infiltrated macrophages is decreased by treatment with fPue, immunostaining was performed for F4/80, a marker for macrophages, on frozen sections of colon tissue. Immunofluorescence analysis revealed that while the number of F4/80-positive cells present in mucosa and submucosa of the inflamed colon was significantly increased, that of F4/80-positive cells was markedly reduced in the colon of the DSS+fPue-treated group (Figure 3A, B), indicating that macrophage infiltration was reduced by fPue treatment. In addition, decrease in transcriptional expression of pro-inflammatory mediators, IL-1β, TNFα and IL-6 was observed by performing RT-PCR (Figure 3C).

As previously mentioned, epithelial dysfunction is a major aspect in the pathogenesis of IBD. To determine whether the epithelial barrier, disrupted in DSS-induced colitis, is recovered by treatment with fPue, the expression levels and the arrangement of ZO-1, Claudin1 and Occludin, were examined by immunofluorescence assay, using cryosections of colon tissue. As shown in Figure 4A, all three TJ proteins were expressed along with the inner lining of the columnar epithelium of normal colon tissue. While the expression levels of TJ proteins were decreased and their architecture was disrupted in DSS-induced colitis, they were recovered as DSS-induced colitis was ameliorated by fPue treatment (Figure 4A). In addition, to investigate whether the physiological function of the epithelial barrier was recovered, bacterial translocalization was examined by blood agar assay, using cells from mesenteric lymph nodes. As shown in Figure 4B and C, bacterial translocalization was induced as the epithelial barrier was disrupted by DSS treatment. On the other hand, in the DSS+fPue-treated group a decreased number of bacteria was present in the lymph node (Figure 4B, C), indicating that bacterial translocalization was reduced as disrupted tight proteins were rescued by fPue treatment.

To determine whether DSS-induced disruption of TJ would be recovered by fPue treatment, an in vitro experiment was performed using colonic epithelial cell line Caco-2. Caco-2 cells form a monolayer and TJ between the cells, and the TJ is disrupted by DSS treatment [29]. Consistent with the previous study, treatment of Caco-2 cells with DSS resulted in a decrease in protein expression of Occludin, which was detected by western blot, while that of ZO-1 did not change (Figure 5A). Although ZO-1 expression level was not affected by DSS, its architecture was disrupted as well as that of Occludin (Figure 5B). However, fPue treatment prevented disruption of the architecture of ZO-1 and Occludin (Figure 5B). These data suggest that fPue could have a protective effect directly on intestinal TJ.