Fifty-four Sprague-Dawley male rats (5–7 weeks old, weighing between 200 and 220 g) were purchased from the experimental animal center at Dalian Medical University [Certificate of Conformity: No. SYXK (Liao) 2013-0006]. The experimental protocol was approved by the Animal Care and Ethics Committee of Dalian Medical University on June 8, 2012. The animal protocol was designed to reduce pain and discomfort in the animals. The rats were acclimated to laboratory conditions (23°C, 12/12 h light/dark, 50% humidity, ad libitum access to food and water) for 2 weeks prior to the experiments. Rats were housed one per cage and they were deprived of food for 12 h before the experiments. All rats were euthanized by barbiturate overdose (intravenous injection, 150 mg/kg pentobarbital sodium) for intestinal tissue collection.

PFB (EC 3.4.22.33) was purified by us from crude proteins of pineapple by high-speed counter-current chromatography (17). Briefly, Matured Pineapple fruits were purchased from a local store (Dalian, China) and authenticated by Dr. Yunpeng Diao (Dalian Medical University, Dalian, China). Pineapple fruit was used to extract the juice, after the juice refined by centrifugation (10,000 × g, 30 min, 4°C), finely powdered ammonium sulfate was added into the juice gradually to obtain 50% saturation with continuous stirring for 1 h. Overnight aging at 4°C, some precipitate was separated out and recovered by centrifugation at 10,000 × g for 30 min at 4°C. Finally, 7.6 g dry protein was obtained, which was used for subsequently isolation. HSCCC coupled with a reverse micelle solvent system was successfully applied to separate fruit bromelain from fruit of pineapple, and the protein content of separated fraction was reached to 99%, the electrophoresis of obtained fraction purity was 100%, and the activity recovery was 95.5%. Stem bromelain complex (E.C. 3.4.22.32) and sulfasalazine (SASP) were purchased from Tianjin Kingyork Group Co. Ltd. (Tianjin, China). Antibodies to TNFR1 (ab90463), TNFR2 (ab109322), NF-κB (ab16502), MLCK (ab76092), occludin (ab167161), Bcl-2 (ab59348), and Bax (ab53154) were purchased from Abcam Ltd. (Hong Kong, China). Chemicals were obtained from Sigma-Aldrich (St. Louis, MO, USA), unless otherwise indicated.

Rat intestinal IEC-6 epithelial cells and human Caco-2 cells were obtained from the cell bank of the Shanghai Institute (Shanghai, China). The cells used in this study were evaluated before the experiments, and no significant interspecies variations in TNFR signaling were observed, which may have affected the results. Cells were maintained at 37°C in a 5% CO₂ environment. The culture medium consisted of DMEM with 4.5 mg/mL glucose, 50 U/mL penicillin, 50 U/mL streptomycin, 4 mM glutamine, 25 mM HEPES, and 10% fetal bovine serum. Both fetal bovine serum and DMEM were purchased from Invitrogen (Waltham, MA, USA).

Twenty-four of the 54 rats were used to test the toxicity of bromelain in vivo. The rats were divided randomly into four groups (n = 6) including normal control rats, 2.5, 20, and 160 mg/kg PFB-treated groups. The remaining 30 rats were divided randomly into five groups (n = 6). The rats were treated as follows: group I, sham-operated control with intracolonic administration of saline; group II, colitis group; group III, SASP (100 mg/kg body weight, intragastric, dissolved in saline); group III, low-dose PFB (10 mg/kg body weight, intragastric, dissolved in saline); group IV, high-dose PFB (80 mg/kg body weight, intragastric, dissolved in saline), 1 day after colitis induction. SASP and PFB were administered by gavage once daily for 14 consecutive days. The rat colitis model was induced as described previously (19). Briefly, rats were fasted for 24 h with free access to drinking water. A catheter was inserted through the anus to approximately the level of the splenic flexure (8 cm proximal to the anal verge) under urethane anesthesia. The colon was then infused with 1 mL of TNBS dissolved in ethanol (50% v/v) at a dose of 125 mg/kg. The rats were allowed to eat and drink ad libitum from 1 h after the operation. Distal colon samples from full-thickness intestinal walls were harvested for biochemical studies.

Animal body weight and total food intake for each group were measured daily. Macroscopic colon damage was scored on a scale of 0–10 (20). Colon preparations were stained with hematoxylin and eosin (HE), and the results were evaluated according to previously defined morphological criteria (21–23). Levels of myeloperoxidase (MPO) and pro-inflammatory cytokines were examined using enzyme-linked immunosorbent assay (ELISA) kits (R&D Systems, Minneapolis, MN, USA), according to the manufacturer’s instructions.

Intestinal epithelial barrier function was measured in vivo according to a previous study (19). Briefly, rats received gavage administration of 150 µL (80 mg/mL) fluorescein isothiocyanate-4 kDa dextran (FD-4) (Sigma-Aldrich, St. Louis, MO, USA), before which rats were fasted but free to water for 3 h. One and three hours later, serum was harvested and measured using a Synergy HT plate reader (BioTek, Winooski, VT, USA). Intestinal barrier function in vitro was represented by transepithelial electrical resistance (TER) using an epithelial voltohmmeter. Caco-2 cells (4 × 10⁵) were seeded in the upper chamber of a transwell filter. A barrier dysfunction cellular model was established in Caco-2 monolayers exposed to lipopolysaccharide (LPS).

Colon sections were isolated from rats in each group and immediately stored in liquid nitrogen. Total protein was isolated from epithelial layer of the colon section using a Total Protein Extraction Kit (KeyGen Biotech, Nanjing, China). Blots were transferred to nitrocellulose filters and probed with corresponding antibodies at 4°C with gentle shaking overnight. Bands were detected and quantified using a MultiSpectral Imaging system (UVP, Cambridge, UK).

IEC-6 cells were transfected using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) with TNFR1/TNFR2-targeted or control small interfering RNA (siRNA) oligos (Dharmacon, Lafayette, CO, USA), according to the manufacturer’s instructions [Takara Biotechnology (Dalian) Co., Ltd.]. The siRNA sequence for TNFR1/TNFR2 was produced by Genepharma, Ltd. (Shanghai, China). The efficiency of gene silencing was confirmed by western blotting.

The animal experiments, in vitro experiments, and data analyses were conducted according to a single-blind study design. Data were compared among three or more groups using a one-way ANOVA, and between two groups using Student’s t-tests. Data were expressed as the mean ± SD. Data were normally distributed and each group showed similar variances. Further evaluations were carried out using Kruskal–Wallis rank sum tests. All experiments were repeated at least six times and a P value < 0.05 was considered statistically significant.

The cytotoxicity of PFB in cell lines was examined in vitro to determine the effective concentration with no toxicity for subsequent experiments. Exposure to 2.5–160 µg/mL PFB for 36 h had no significant effect on the viability of IEC-6 (Figure 1A). Intragastric administration of PFB (2.5–160 mg/kg) for seven consecutive days had no significant effect on cytokine profiles [TNF-α, interleukin (IL)-1β, and IL-8] (Figures 1B–D) in colon tissue. Furthermore, intragastric administration of PFB had no significant effect on the expression of mRNA and protein of TNFRs (Figures 1E,F). Based on these preliminary experiments and preliminary results reported by other researchers (15, 24), we used <80 mg/kg PFB for in vivo and <80 µg/mL PFB for in vitro experiments in this study.

Colitis symptoms for rats in TNBS group (TNBS-colitis) are shown in Table 1. Rats in the TNBS-colitis group regained consciousness about 2 h after anesthesia. Symptoms included loose stools, increased stool frequency, outflow of red or dark red liquid from the anus, positive fecal occult blood test, and loss of appetite. On day 3, these symptoms peaked and lasted for about 7 days. Seven days following colitis induction, loose stools were still seen in the TNBS-colitis group. Deaths were recorded throughout the experiment, as outlined in Table 1. New rats were added to maintain six rats per group.

Successful establishment of colitis was confirmed with biochemical and macroscopic analysis. TNBS challenge provoked apparent colonic mucosal injuries, including serious hyperemia, edema, and ulcers, some of which were covered with necrosis on the surface of colonic mucosa. HE-staining revealed remarkable inflammatory cell infiltration even in muscle layers, irregular arrangement of glands, crypt abscesses, and thickened submucosal edema (Figure 2A). In the colitis group, rats had significantly lower body weight (Figure 2B) and lower food intake (Figure 2C) compared with the sham group, as well as more visible macroscopic damage (Figure 2D) and a higher colon weight-to-length ratio (Figure 2E). Colitis rats also had higher MPO activity (indicating neutrophil infiltration into the damaged tissue), higher pro-inflammatory cytokine levels, including TNF-α, IL-1β, and IL-8. PFB (10 and 80 mg/kg) and SASP reversed the pathological changes in the colitis model after 7 and 14 days of drug treatment, suggesting that PFB significantly ameliorated the colitis symptoms (Figures 2 and 3).

In vivo, intestinal barrier dysfunction leads to increase in serum recovery of FD-4. In single layer of Caco-2 cells, LPS leads to intestinal barrier dysfunction and significant TER reduction. In this study, the serum recovery of FD-4 was significantly increased in the colitis control group compared with the sham group. Gavage administration of PFB significantly reduced the serum recovery of FD-4. In in vitro studies, LPS induced decrease of TER, which was also significantly reversed by PFB (Figure 4). These results suggest that intestinal epithelial barrier dysfunction is recovered by PFB treatment.

In the TNBS group, expression levels of TNFR1, TNFR2, and NF-κB were significantly increased compared with the sham group, and all of these changes were reversed by treatment with PFB (Figure 5). The apoptosis-related protein Bax was significantly increased and Bcl-2 was significantly decreased in colitis rats, and these changes were also reversed by PFB. The epithelial tight junction dysfunction-related protein MLCK was significantly increased and the tight junction protein occludin was significantly decreased, and these changes were also reversed by PFB treatment. However, the increased expression level of TNFR mRNA was not significantly affected by PFB.

To confirm the role of PFB-induced TNFR inhibition in the treatment of colitis, we examined this effect in IEC-6 cells with or without siRNA targeting TNFR2. In normal cells, LPS induced significant increases in NF-κB and MLCK, which reflects the induction of inflammation and epithelial barrier dysfunction. These increases were significantly inhibited by both 10 and 80 mg/kg PFB (Figure 6). The PFB-induced inhibition of NF-κB and MLCK could be significantly abrogated by inhibition of TNFR2 by RNA interference (Figure 6). Moreover, the PFB-induced inhibition of NF-κB but not MLCK could be abrogated by inhibition of TNFR1 by RNA interference (Figure 7), suggesting that PFB alleviated inflammation and epithelial barrier dysfunction in a TNFR2-dependent manner. The results above suggest that TNFR-regulated inflammation, epithelial apoptosis, and tight junction barrier dysfunction in colitis may be blocked by PFB, which in turn may inhibit the release of cytokines, as well as decrease epithelial permeability.

The bromelain used in this study is PFB, but commercial bromelain is a complex natural mixture of proteolytic enzymes derived from pineapple stems (10). The effect of PFB, fruit bromelain complex, and commercial stem bromelain complex on the expression of TNFR1 and TNFR2 in IEC-6 cells stimulated by LPS was studied. Compared with fruit bromelain and stem bromelain complexes, the inhibition of TNFR2 induced by PFB was stronger than the inhibition of TNFR1. These results indicate that PFB showed a stronger selective inhibitory effect on TNFR2 than TNFR1 (Figure 8).