Chemoprevention of dietary digitoflavone on colitis-associated colon tumorigenesis through inducing Nrf2 signaling pathway and inhibition of inflammation

A DNA fragment containing 8 copies of the ARE sequence (GTGACAAAGCACCC) were subcloned into the pGL3 vector. After transient transfection with the expression plasmid, different concentrations of digitoflavone were added to the cell culture and incubated for 8 hours and 24 hours respectively. Parallel cell viability assays revealed no obviously cytotoxic effects (>95% viability) for the digitoflavone treatment when the concentration of digitoflavone is lower than 10 μM in Caco-2, HepG2, HEK-293 cells and 5 μM in HT-29 cells (Figure 1F). 10 μM digitoflavone induced the highest level of luciferase activity after 8 hours exposure, about 5-fold increases of control (Figure 1B). Another human epithelial colorectal adenocarcinoma cell line HT-29 also showed that low concentrations (5 μM) of digitoflavone can increase the ARE-luciferase activity with no obviously cytotoxic effects (Figure 1C). To evaluate the ARE-driven luciferase activity of digitoflavone in other cell lines, HepG2 (Figure 1D) and HEK-293 (Figure 1E) cell lines were transient transfected with the pGL3-ARE-luciferase plasmid respectively and tested with 1–20 μM digitoflavone for 8 hours. All tested cell lines showed over 2-fold increases of the luciferase activity at 1–10 μM concentrations of digitoflavone. These result suggested that digitoflavone, at low concentrations (<10 μM), is a potent activator of the Nrf2/ARE antioxidant pathway.

To verity whether activation of luciferase activity by digitoflavone in Caco-2 cells reflected the expression of the endogenous ARE-driven genes, the mRNA levels of GR, TR, HO-1, γ-GCSc, γ-GCSm, NQO1, and MRP2 were examined in the presence or absence of digitoflavone. In Caco-2 cells treated with 10 μM digitoflavone for 8 hours, the mRNA levels of GR, TR, HO-1, γ-GCSc, γ-GCSm, UGT1A1 and UGT1A10 increased 1.2-, 6.0-, 1.5, 1.7-, 1.8-, 1.5, 1.8-fold, respectively (Figure 2A). Similarly, evaluation of the Nrf2-mediated antioxidant enzymes, such as γ-GCSc and TR by Western blotting showed that exposure of Caco-2 cells to 1–15 μM digitoflavone strongly induced γ-GCSc, γ-GCSm and TR protein expression in a dose and time-dependent manner (Figure 2).

Previous studies described that under normal conditions, Keap1 sequestered Nrf2 in the cytoplasm and that translocation of Nrf2 into the nucleus is essential for the transactivation of various targeted genes [26, 27]. Therefore, to further investigate effects of digitoflavone one the Nrf2/ARE activation, we examined the protein expression and subcellur location of Nrf2 in Caco-2 cells after digitoflavone treatment. As show in the Figure 2B, Western blot analysis demonstrated a significant increase of Nrf2 protein expression after digitoflavone treatment in dose- and time-dependent manner. Western blot analysis of the nuclear fraction (Figure 2C) and Immunofluorescence analyses (Figure 2D) showed Nrf2 accumulation in the nucleus of Caco-2 cells after digitoflavone treatment.

To confirm the requirement of Nrf2 in the digitoflavone-induced antioxidant activities, we transfected the Caco-2 cells with Nrf2 target siRNA before digitoflavone treatment. As show in Figure 2E, silencing Nrf2 expression significantly inhibited the digitoflavone-induced γ-GCSc, γ-GCSm and TR up-regulation, suggesting that digitoflavone induced antioxidant activities in an Nrf2/ARE-dependent manner.

We also investigated changes in GSH content in Caco-2 cells after incubation in varying concentrations of digitoflavone for 8 h. Digitoflavone increased GSH content and decreased the level of GSSG in a dose-dependent manner, which resulted in a dose-dependent increase in the ratio of GSH/GSSG (Figure 2F-H). This result is consistent with increased levels of γ-GCSc and γ-GCSm mRNAs, which encode the rate-limiting enzymes in GSH synthesis, in Caco-2 cells.

Nrf2 is a key component in protection against carcinogenesis and oxidative stress [26, 28, 29]. Previous reports have suggested that oxidative stress plays an important role in tumor promotion [30, 31]. H₂O₂ may induce self-generation of free radicals known as the ROS-induced ROS release at the mitochondrial level [32], which has been widely used as a model of exogenous oxidative stress. In this study, we validated if antioxidant activities induced by digitoflavone can actually protect against H₂O₂-induced damage in Caco-2 cells. The protective effects of digitoflavone against the H₂O₂-induced cytotoxicity were detected by MTT assay. As show in Figure 3A and B, pretreatment of digitoflavone for 4 h exhibited dose-dependent protective effects in the H₂O₂-damage model and the Nrf2 target siRNA transfection group (Figure 3C), while the GSH synthesis inhibitor BSO partially abolished the digitoflavone-induced protective effect (Figure 3D).

Intracellular ROS levels affect cell viability and high ROS levels can cause cellular damage. Using flow cytometry analysis, we examined the effects of digitoflavone on intracellular ROS levels. As shown in Figure 3E and F, H₂O₂ treatment led to a significant increase in ROS levels. Statistical analysis showed that digitoflavone reduced the H₂O₂-induced intracellular ROS level in a dose-dependent manner. We further confirmed the anti-apoptotic effects of digitoflavone through the quantitative analysis of FITC-Annexin V/PI staining by flow cytometry. In the normal control group, the percentage of apoptotic cells was 8.7%. The percentage of apoptotic cells increased up to 33.9% in the H₂O₂ model group (Figure 3G and H). The protective effects of digitoflavone against cell apoptosis was concentration dependent (apoptosis rate 29.7 ± 0.88% and 21.2 ± 1.18% for 1 and 10 μM digitoflavone treatment respectively).

Under normal conditions, the interaction of Nrf2 with the Kelch-like ECH-associated protein 1 (Keap1) traps Nrf2 in the cytosol, leading to a rapid degradation of the cytosolic Nrf2 by the 26S proteasome, via the Cullin3-based E3-ligase ubiquitination complex [26]. A number of studies have shown that several signaling pathways, including PI3K [33, 34], MAPK [35, 36], and PKC [37, 38], are involved in the induction of Nrf2/ARE-driven gene expression. To elucidate the signal transduction pathways leading to the activation of Nrf2 and the induction of antioxidants expression in the digitoflavone-treated cells, we examined the effects of digitoflavone on the expression of Keap1 and the phosphorylation of PKC, AKT, ERK1/2, and p38 MAPK. Upon digitoflavone treatment, time-dependent increases in the phosphorylation of AKT, ERK1/2, and p38 MAPK were observed (Figure 4A). To determine whether such activations of AKT, ERK1/2, and p38 MAPK contribute to the digitoflavone-induced Nrf2 activation, several kinase inhibitors, including wortmannin (for PI3K), PD98059 (for MEK1/2), and SB202190 (for p-p38 MAPK), were employed. As show in Figure 4B-D, inhibition of the phosphorylation of AKT and ERK1/2 did not decrease the digitoflavone-induced Nrf2 activation. However, the p38 MAPK inhibitor SB202190 significantly inhibited the digitoflavone-induced Nrf2 activation (Figure 5A) and nuclear accumulation (Figure 5B). To determine whether such activation of p38 MAPK contribute to the digitoflavone-mediated protections against the cytotoxic effects of H₂O₂, the Caco-2 cells were pre-incubated with SB202190 for 2 hours before the 4 hours digitoflavone treatment, Cells were then challenged with 500 μM H₂O₂ for additional 24 h for MTT assay, 4 h for ROS detection, and 6 h for apoptosis detection, respectively. As show in Figure 5C, SB202190 eliminated the protective effects of digitoflavone. SB202190 also reversed the digitoflavone antioxidant activity (Figure 5D and E). Further, the anti-apoptosis ability of digitoflavone was also abolished by SB202190 (Figure 5F and G).

We further explored chemopreventive effects of digitoflavone on tumor progression by administering it to mice from week 2 to day 13, after the AOM and 3 cycles of DSS treatments (Figure 6A). Compared with the AMO group, digitoflavone treatment reduced the numbers and size of macroscopical tumors remarkably and the shorted colon length was resvered by digitoflavone when compared with AOM group, also less loss of crypts was observed in mice with digitoflavone treatment (Figure 6B-D). Next, activation of Nrf2 and its downstream targets were assessed to demonstrate that the beneficial effect of digitoflavone against tumor progression is attributed to activation of the Nrf2 pathway. Protein expression of Nrf2 and Nrf2 downstream targets TR, γ-GCSc and γ-GCSm and mRNA expression of GR, TR, HO-1, γ-GCSc, γ-GCSm, NQO-1, UGT1A1 and UGT1A10 were slightly changed in AOM group, indicating induction of the Nrf2 pathway by colon oxidative stress. As expected, treatment with digitoflavone markedly increased the protein levels of Nrf2, TR, γ-GCSc, γ-GCSm and HO-1 and mRNA levels of GR, TR, HO-1, γ-GCSc, γ-GCSm, NQO-1, UGT1A1 and UGT1A10 (Figure 6E and F). The mRNA Levels of colonic inflammatory cytokines TNF-a, IL-1β and IL-6 were increased in AOM group, and digitoflavone reduced TNF-a, IL-1β and IL-6 mRNA Levels when compared with AOM group (Figure 6G).

AOM, DSS (MW 36,000–50,000), digitoflavone, SB202190, DCFH-DA [2′, 7′-Dichlorofluorescin diacetate], Trypsin, MTT [3-(4, 5-dimethylthiazol-2-yl)-2, 5-diphenyltetrazoliμM bromide], BSO [L-Buthionine-sulfoximine ], DNase-free RNase and SB202190 were obtained from Sigma-aldrich, USA. Digitoflavone was dissolved in dimethyl sulfoxide (DMSO) and was used in all experiments. Maxima® SYBR Green/ROX qPCR Master Mix (2×) and Maxima® First Strand cDNA Synthesis Kit were purchased from Fermentas life science (Fermentas, MBI). PD98059, Wortmannin, Lysis buffer was purchased from Beyotime, China. Primary antibodies (Nrf2 for mouse, γ-GCSc, γ-GCSm, TR, HO-1, Keap1, PKC, Lamin B) were obtained from Santa Cruz Biotechnology, CA, USA. Rabbit anti-Nrf2 (for human) was purchased from Abcam, USA. Primary antibodies (p-PKC, ERK1/2, p-ERK1/2, p-p38, p38, AKT and p-AKT) were purchased from Cell Signaling Technology, MA, USA. Goat anti-rabbit IgG and goat anti-mouse IgG antibodies were purchased from LI-COR, Lincoln, NE, USA. Rabbit anti-Goat IgG was obtained from KPL, Gaithersbhrg, MD, USA. Monoclonal mouse anti-glyceraldehyde-3-phosphate dehydrogease (GAPDH) was obtained from KangChen, China.

Human epithelial colorectal adenocarcinoma cell line Caco-2, Human colon adenocarcinoma grade II cell line HT-29, human liver carcinoma cell line HepG2 ,Human Embryonic Kidney 293 cell HEK-293 were purchased from Cell Bank of Shanghai Institute of Biochemistry and Cell Biology. Cells were cultured in DMEM-medium (for HepG2, HEK-293 cells), MEM-medium (for Caco-2 cells), McCOY’s 5A(for HT-29) supplemented with 10% fetal bovine serum (FBS), 100 U/ml penicillin and 100 μg/ml streptomycin (all available from Invitrogen, Grand Island, NY, USA). All cultures were maintained in a humidified environment with 5% CO² at 37°C.

We used the luciferase reporter assay to investigate the Nrf2-mediated transcriptional activity of Nrf2. Firstly, 8 copies of antioxidant-responsive element (ARE)–luciferase reporter plasmids were generated using the pGL3 promoter vector (Promega UK). After the plasmids were generated, the DNA sequence of the inserts was verified. The Dual-Luciferase Reporter Assay System (Promega, UK) was used to determine reporter gene activity in transiently transfected cells. Transient transfection was performed in 96-well plates at a cell density of 50% ~ 70% confluence per well. Then the 8 × ARE pGL3 plasmid were co-transfected with the pRL-TK plasmid(transfection ratio of 8 × ARE pGL3:pRL-TK is 10:1), encoding Renilla luciferase as an internal control for transfection efficiency for 24 h using Lipofectamine 2000 (Invitrogen, Carlsbad, CA, USA) according to the manufacturer’s instructions. After transfection, cells were treated with test samples for indicated time, and then cell lysates were prepared for assessment of luciferase activity. Fire fly and Renilla luciferase activities were measured using a luminometer (Centro XS3 LB960, Berthold, Germany) according to the manufacturer’s instructions. Relative fire fly luciferase activity was normalized to Renilla luciferase activity and activity was expressed as fold induction after treatment with compounds compared with vehicle control (DMSO).

Cell viability was determined using the MTT assay. Briefly, cells in logarithmic phase were seeded at the density of 70 ~ 80% confluence per well in 96-well plates at 37°C with 5% CO² for overnight incubation and treated with appropriate concentrations of test samples for the indicated times. After treatment, 10 μl of 5 mg/ml MTT was added and the cells were incubated for 4 h at 37°C. The supernatant was discarded and 100 μl of DMSO was added to each well. The mixture was shaken on a mini shaker at room temperature for 10 min and the spectrophotometric absorbance was measured by Multiskan Spectrum Microplate Reader (Thermo, USA) at 570 nm and 630 nm (absorbance 570 nm, reference 630 nm). Triplicate experiments were performed in a parallel manner for each concentration point and the results were presented as mean ± SD. The net A_570nm-A_630nm was taken as the index of cell viability. The net absorbance from the wells of cells cultured with DMSO was taken as the 100% viability value. The percent viability of the treated cells was calculated by the formula: % viability = (A_570nm-A_630nm) _treated/(A_570nm-A_630nm) _control × 100%.

Caco-2 cells were cultured in MEM and then treated with test samples for indicated time. Proteins were isolated by lysis buffer (Beyotime, China) and measured using the Nanodrop 1000 Spectrophotometer (Thermo, USA). Protein samples were separated on 10% SDS-polyacrylamide gels (SDS-PAGE) and transferred onto the PVDF membranes (Millipore, USA). After blocked with 1% BSA in TBST for 2 h, membranes were incubated with primary antibodies overnight at 4°C. Blots were washed and incubated with secondary antibodies for 1 h at room temperature. Membranes were again washed three times with TBST and were scanned with an Odyssey infrared fluorescent scanner (LI-COR) and analyzed with Odyssey software version 3.

Caco-2 cells were treated with various concentrations of digitoflavone (1–15 μM) or vehicle control (DMSO). After 8 h incubation, the cellular GSH and GSSG were quantified using GSH/GSSG-Glo Assay kit (Promega) according to the manufacturer’s protocol. GSH and GSSG levels were normalized to protein concentrations and the GSH/GSSG ratio was calculated.

Cells in logarithmic phase were seeded in logarithmic phase were seeded at the density of 70 ~ 80% confluence per well into 24-well chamber slides. After treatment with test samples for the indicated times, cells were fixed with cold 4% (w/v) paraformaldehyde for 20 min, rehydrated in PBS for 15 min, and permeabilized in 0.1% (w/v) TritonX-100 at room temperature for 10 min. After being washed with PBS, the cells were blocked unspecific fluorescence with 3%BSA for 1 hour and then incubated with primary antibody at 4°C overnight followed by Texas Red-conjugated secondary antibody for 1 h at room temperature. The images of Nrf2 with Texas Red staining were captured using a fluorescence microscope.

Nuclear extract protein was prepared according manufactory’s instruction (Thermo, USA). Briefly, after treatment with digitoflavone for indicated times, Caco-2 cells were harvested, washed with PBS, centrifuged, and resuspended in ice-cold buffer CERI. After 10 min of incubation on ice, cells were added with ice-cold CERII and centrifuged again, the supernatant (cytoplasmic extract) was immediately transferred to a clean pre-chilled tube. The insoluble (pellet) fraction was resuspended with NER, and vortex for 15 seconds every 10 min for a total 40 min. The tube was centrifuged and the supernatant (nuclear extract) was immediately transferred to a clean pre-chilled tube. The cytoplasmic and nuclear extract protein was stored at −80°C until use. For Western blot analysis, LaminB and GAPDH were used as internal controls for nuclear and cytoplasmic extracts, respectively.

Caco-2 cells were treated with different concentrations of digitoflavone for indicated times, then treated cells were washed with PBS, total RNA was extracted from the treated cells using trizol reagent (Invitrogen, Carlsbad, CA) and then RNA was converted to cDNA by reverse transcriptase (Thermo, USA) according to the manufacturer’s instruction. Primers used for the reactions were purchased from Genscript and the sequences were listed in Table 1. Real-time qPCR analysis for mRNA expression was performed using SYBR Green probes and an ABI 7500. ALL genes’ mRNA expression was normalized against GAPDH expression.

The production of cellular ROS, mainly H₂O₂, was detected using the DCFH-DA fluorescence assay. Briefly, cells were seeded in 24-well plates at the density of 70 ~ 80% confluence per well for overnight incubation. After treatment with appropriate concentrations of test samples, cells were harvested, placed into 1.5 mL round-bottom polystyrene tubes, and washed with PBS twice. Subsequently, the cells were centrifuged for 5 min at 400 × g at room temperature, and the supernate was discarded. The cells were resuspended in 500 μL ROS detection solution, stained in the dark at 37°C for 30 min, and analyzed by FACScan laser flow cytometer (Guava easycyteHT, Millipore, CA).

Caco-2 cells in logarithmic phase at were treated with test samples for indicated time. Then they were harvested, washed and resuspended with PBS. Apoptotic cells were determined with an FITC Annexin V Apoptosis Detection Kit (BD Biosciences, USA) according to the manufacturer’s protocol. Briefly, the cells were washed and subsequently incubated for 15 min at room temperature in the dark in 100 ul of 1 × binding buffer containing 5ul of Annexin V-FITC and 5 ul of PI. Afterward, apoptosis was analyzed by FACScan laser flow cytometer (Guava easycyteHT, Millipore, CA).

Nrf2-specific short interfering RNA (siRNA) and scramble control siRNA were obtained from RIBOBIO (GuangZhou, China). Transfection was performed using LipofectAMINE 2000, according to the manufacturer’s protocol, with Nrf2-specific siRNA SMARTpool L-003755-00-0050; human NFE2L2 (NCBI Accession No. NM006164); target sequences including TCCCGTTTGTAGATGACAA, GAGAAAGAATTGCCTGTAA and GCAACAGGACATTGAGCAA. Briefly, cells were transfected with 10 nmol/L siRNAs directed against Nrf2 and non-targeting scramble control siRNA for 48 h, followed by treatment with the test samples for the indicated times. The cells were harvested and the protein status, MTT test and flow cytometry analysis.

All animal experiments were conducted in accordance with the NIH Guidelines for the Care and Use of Laboratory Animals. Pathogen-free 8- to 12-week old C57BL/6 male mice were housed in a temperature-controlled room (22°C) with a controlled 12-h light/dark cycle. The mice were given free access to diet and water during the course of experiments. They were allowed to adapt to the Experimental Animal Laboratory for 1 week before beginning the experiment. Mice were injected intraperitoneally with 10 mg/kg body weight of AOM dissolved in physiological saline. One week later, 2% DSS was given in the drinking water over 7 days, followed by 14 days of regular water. This cycle was repeated a total of 3 times. Body weight was measured every week, and the animals were sacrificed at week 13 for macroscopical inspection, histological analysis, and total RNA and protein extraction. In digitoflavone group, digitoflavone at 50 mg/kg dose suspended in 0.5% carboxymethyl cellulose (CMC) was given as gavage to mice and mice of control group and AOM group were given 0.2 mL 0.5% CMC solution every day from week 2 to week 13 (Figure 6A).

Results are expressed as mean ± SD. Statistical tests were performed using SPSS 15.0. Unpaired Student t tests were used to compare the means of two groups. For multiple comparisons between groups, a one-way ANOVA was performed to detect statistical differences. Differences within the ANOVA were determined using a Tukey’s post-hoc test. P value of less than 0.05 was considered to be statistically significant.