Schisandrin A protects intestinal epithelial cells from deoxynivalenol-induced cytotoxicity, oxidative damage and inflammation

Extensive research has revealed the association of continued oxidative stress with chronic inflammation, which could subsequently affect many different chronic diseases. The mycotoxin deoxynivalenol (DON) frequently contaminates cereals crops worldwide, and are a public health concern since DON ingestion may result in persistent intestinal inflammation. There has also been considerable attention over the potential of DON to provoke oxidative stress. In this study, the cytoprotective effect of Schisandrin A (Sch A), one of the most abundant active dibenzocyclooctadiene lignans in the fruit of Schisandra chinensis (Turcz.) Baill (also known as Chinese magnolia-vine), was investigated in HT-29 cells against DON-induced cytotoxicity, oxidative stress and inflammation. Sch A appeared to protect against DON-induced cytotoxicity in HT-29 cells, and significantly lessened the DON-stimulated intracellular reactive oxygen species and nitrogen oxidative species production. Furthermore, Sch A lowered DON-induced catalase, superoxide dismutase and glutathione peroxidase antioxidant enzyme activities but maintains glutathione S transferase activity and glutathione levels. Mechanistic studies suggest that Sch A reduced DON-induced oxidative stress by down-regulating heme oxygenase-1 expression via nuclear factor (erythroid-derived 2)-like 2 signalling pathway. In addition, Sch A decreased the DON-induced cyclooxygenase-2 expression and prostaglandin E2 production and pro-inflammatory cytokine interleukin 8 expression and secretion. This may be mediated by preventing DON-induced translocation of nuclear factor-κB, as well as activation of mitogen-activated protein kinases pathways. In the light of these findings, we concluded that Sch A exerted a cytoprotective role in DON-induced toxicity in vitro, and it would be valuable to examine in vivo effects.

Sch A prevents Don-induced cell cycle arrest and, to a lesser extent, apoptosis. Since the induction of apoptosis may be mediated through the regulation of cell cycle, the impact of different treatments on cell cycle perturbations was analysed by PI staining (Fig. 2). DON alone showed significant decrease in the percentage distribution of cells in G0-G1 phase ( Fig. 2A). Percentages of cells in S phase (Fig. 2B) and G2-M phase (Fig. 2C) were significantly increased following DON treatment alone. These data indicate that DON alone induced S-phase and G2-M phase cell cycle arrest. Pre-treatment with Sch A, tends to decrease the percentages of cells in S phases, but significant differences were not detected. No significant changes in percentages of cells in G0-G1, S-and G2-M phases occurred when cells were treated with any of Sch A concentrations alone (Fig. 2).
For quantification of apoptosis, cells were stained with Annexin V-FITC and PI. Both early and late apoptotic cells were considered apoptotic cells. In Fig. 3, apoptosis induced by DON was not significantly higher compared to control cells, whereas Sch A treatment alone resulted in significant reduction in apoptosis at all the concentrations tested. This may be due to the relatively low concentration of DON used in this study, which may be insufficient to induce apoptosis. However, Sch A pre-treatment at 5 and 10 µM showed significant reductions in the proportion of apoptotic cells compared to DON-treated cells.
Sch A inhibits DON-induced ROS production, nitrite production but not lipid peroxidation. To measure changes in the cellular redox status in response to DON with or without Sch A pre-treatment, cells were first exposed to 2.5-10 µM Sch A followed by DON exposure at 1 µM from 0.5 to 24 hours (Fig. 4A). Treatment with DON significantly increased ROS from 0.5 to 6 hours (p < 0.05) but not at 24 hours. Pre-treatment with Sch A at 2.5, 5 and 10 µM followed by DON caused a significant decline in ROS release (p < 0.05). www.nature.com/scientificreports www.nature.com/scientificreports/ Thiobarbituric acid reactive substance (TBARS) assay was carried out to measure MDA production in HT29 cells (Fig. 4B). Results obtained demonstrated that DON increased MDA level significantly when compared to control (p < 0.05), while Sch A alone did not result in significant change. MDA production in samples pre-treated Sch A remained comparable to that of DON alone.
NO assay was performed to measure effects of DON and Sch A on nitrite production (Fig. 4C). Treatment with DON significantly reduced nitrite concentration (p < 0.05) but pre-treatment with 5 µM and 10 µM Sch A HT-29 cells were treated with 1 µM of DON for 24 hours in the presence and absence of Sch A pre-treatment (2.5, 5 and 10 µM) for 24 hours. Control received appropriate carriers. Percentage distribution of cells in (A) G0-G1 phases, (B) G2-M phases and (C) S phases were shown as mean of ±SEM (n ≥ 6), which are at least six separate experiments. Different letters indicate significant differences at p < 0.05. G0, resting phases; G1, growth 1 phase; G2, growth 2 phase; M phase, mitosis; S phase, synthesis phase.
www.nature.com/scientificreports www.nature.com/scientificreports/ increased nitrite production slightly, though not statistically significant compared to samples treated with DON only. Samples treated with Sch A only did not have a significantly different nitrite concentration than control.

Sch A lowers DON-induced CAT, SOD and GPx antioxidant enzyme activities but maintains
GSt activity and GSH levels. CAT, SOD, GPx activity assays were performed to examine how Sch A and DON affected antioxidant enzyme activities ( Fig. 5A-C). CAT, SOD and GPx are primary enzymes participated in repairing damage caused by oxidative stress. DON alone increased the activities of these enzymes significantly (p < 0.05). But Sch A pre-treatment showed no significant changes in CAT activity and SOD activity, while the decrease in GPx activity was statistically significant (p < 0.05). Sch A alone slightly increased CAT activity but did not affect other enzyme activities significantly.
GST are detoxification enzymes that act through conjugation of xenobiotics and GSH, an antioxidant that protects cells from ROS. DON significantly reduced GST activity and GSH level (p < 0.05), but pre-treatment with 10 µM Sch A increased GST activity and GSH level significantly (p < 0.05) when compared with samples treated with DON alone. It is noteworthy that Sch A alone significantly (p < 0.05) lowered GSH levels (Fig. D,E).
Sch A suppresses DON-induced HO-1 expression through modulation of Nrf2 signalling pathway. The HO-1 mRNA and protein expression in HT-29 cells were displayed in Fig. 6. DON significantly up-regulated HO-1 mRNA and protein expression level. With Sch A pre-treatment, both HO-1 mRNA and protein were suppressed. Sch A alone caused significant up-regulation of HO-1 protein levels but not mRNA levels.
To examine the effect of DON and Sch A on the activation of Nrf2, Nrf2 was labelled with a green fluorescent tag. As revealed in Fig. 7, nuclear staining of Nrf2 was increased after three hours when DON was added. Pre-treatment with Sch A also increased nuclear Nrf2, indicating an activation of Nrf2 signalling pathway.
Sch A suppresses DON-induced COX-2 mRNA and protein expression and PGE2 production. Cyclooxygenase (COX)-2 is regarded as an inflammation marker. mRNA and protein expression of COX-2 in DON treated cells without or with Sch A pre-treatment were measured. Figure 8A indicated that DON treatment significantly increased mRNA expression. Pre-treatment of Sch A down-regulated the COX-2 mRNA significantly. Sch A alone did not cause any changes in the COX-2 mRNA expression. Previous reports indicate that COX-2 is not solely regulated at the transcription level but also via post-transcriptional mechanisms in human intestinal cells. Moreover, COX-2 protein was degraded through ubiquitin proteolysis, and its half-life was ∼3.5-8 h 41 . Therefore, protein expression of COX-2 was measured in cells after 2, 6 and 24 hours of exposure to DON. Western blot analyses (Fig. 8B) showed that DON significantly up-regulated COX-2 protein expression after 6 and 24 hours of DON exposure. Sch A pre-treatment significantly down-regulated COX-2 protein expression after 6 and 24 hours of exposure to DON.
PGE2 is one of the important mediators produced at the inflammatory sites by the COX-2 enzyme. Thus, the impact of Sch A on DON-induced production of PGE2 in HT-29 cells was also studied by quantifying its levels in the cellular supernatant using ELISA. Figure 8C indicates that stimulation of HT-29 cells with DON alone did not increase PGE2 concentration in the culture medium; however, with Sch A pre-treatment, PGE2 production was significantly reduced. Sch A alone significantly decreased PGE2 production compared to control.
Sch A prevents Don-induced nf-κB expression and nuclear localization of nf-κB. Western blot analyses were performed at various time intervals of 1, 3 and 6 hours for NF-κB expression (Fig. 9). The results indicate that not significant changes in NF-κB protein expression were observed in the whole cell extracts of DON-treated cells. We then examined the NF-κB protein expression in both cytoplasmic and nuclear fractions. Significant up-regulation of NF-κB protein expression was observed in the cytoplasmic fractions of DON-treated www.nature.com/scientificreports www.nature.com/scientificreports/ cells at all time points. However, pre-treatment of Sch A did not suppress NF-κB protein expression. In nuclear fractions, DON significantly up-regulated NF-κB protein expression at all time points; only pre-treatment with Sch A at 10 µM significantly down-regulated NF-κB protein expression. The results indicate that Sch A affected DON-induced NF-κB expression by preventing nuclear localization of NF-κB.
Sch A inhibits DON-induced production of pro-inflammatory cytokine IL8. We then assessed if the Sch A effects on NF-κB activation were followed by an effect on IL8 mRNA expression and secretion (Fig. 10). DON treatment resulted in significant up-regulation of IL8 mRNA and amount of IL8 released in the medium as quantified by qPCR and ELISA, respectively. Sch A pre-treatment down-regulated IL8 mRNA and secretion in a concentration-dependent manner. It is also interesting to note that Sch A alone at 5 and 10 µM significantly reduced the amount of IL8 released but down-regulation of IL8 mRNA was not observed with the same treatments.

Sch A inhibits DON-induced activation of MAPK pathways. Previous studies have established that
the MAPK signalling pathways participated in DON-induced inflammation in intestinal epithelial cells 20,42 . Thus, the potential involvement of these signalling pathways in Sch A-mediated inhibition of DON-induced inflammation was investigated in this study. Western blot data showed that DON treatment markedly promoted the For Sch A pre-treatment, cells were exposed to Sch A (2.5, 5 and 10 µM) for 24 hours, then exposed to DCF-DA for 30 minutes followed by 1 µM of DON for 0.5-6 hours. For cells with 24 hours of DON treatment, DCF-DA was added at the end of the incubation. Results were shown as mean of ±SEM (n ≥ 6), which are at least six separate experiments performed in triplicates. #, ## p < 0.05 and 0.01 significantly different from control, * , ** , ***p < 0.05, 0.01 and 0.001 significantly different from DON treated cells. (B) Nitrite generation measured using Griess reagent. (C) Malondialdehyde (MDA) prodiction as measured by thiobarbituric acid reactive substances (TBARS) assay. Cells were treated with 1 µM of DON for 24 hours in the presence and absence of Sch A pre-treatment (2.5, 5 and 10 µM) for 24 hours. Results were shown as mean of ±SEM (n ≥ 6), which are at least six separate experiments. Different letters indicate significant differences at p < 0.05. www.nature.com/scientificreports www.nature.com/scientificreports/ phosphorylation of three MAPKs, JNK, p38 MAPK and extracellular signal-regulated kinase (ERK) (Fig. 11); yet, the phosphorylation of p38 and ERK but not JNK were abrogated by pre-treatment with Sch A. Thus, it is likely that Sch A inhibited the inflammatory response by supressing p38 and ERK signalling pathways in DON-stimulated HT-29 intestinal cells.

Discussion
DON is a ribotoxic mycotoxin that exerts its toxicity by provoking oxidative stress and inflammatory responses 27 . Oxidative stress causes many pathological changes, including inflammation and cancer. In principle, oxidative stress can result from increased ROS/reactive nitrogen species (RNS) production and lowered antioxidant defence. To combat oxidative stress and the molecular alterations associated with that, it is suggested that www.nature.com/scientificreports www.nature.com/scientificreports/ nutraceuticals or antioxidants may be supplemented via diet. The cytotoxic effects of DON are well documented, however, studies to alleviate the toxicity are sparse. Recently, Sch A was shown to suppress inflammation and oxidative stress in LPS-treated RAW 264.7 macrophages 40 . This study was the first to investigate how Sch A affected DON induced cytotoxicity, reactive oxidative and nitrogen species production and inflammation on a human HT-29 intestinal cells.
The results clearly showed that Sch A can act as a cytoprotective agent against DON-induced cytotoxicity in HT-29 cells. CCK assay revealed that Sch A at 10 μM offered cytoprotection by protecting the cells against DON-induced cytotoxicity. The cytoprotective effect of Sch A against DON-induced toxicity was further evaluated by examining cell cycle distribution and apoptosis. Our results suggest that DON induced G2-M and S phase arrest without induction of apoptosis, possibly attributed to the relatively low but physiologically relevant concentration of DON employed in the current study 43 . Sch A seems to inhibit DON-induced S phase arrest and cause a remarkable decrease in the proportion of apoptotic cells, implying the cytoprotection effect was possibly mediated via cell cycle regulation, to a lesser extent, affecting cell apoptosis.
It was believed that cytoprotective effect of Sch A against DON-induced toxicity may be attributable to the capacity of Sch A to act as a powerful antioxidant and free radical scavenger 44 . In the present study, Sch A can reduce DON-induced ROS at a concentration of 10 μM, implying that Sch A's antioxidant and free radical scavenging activity contributes an important action in cytoprotection. Also, ROS production in HT-29 cells exposed to DON increased from 0.5 to 6 hours, but a negligible increase in ROS production was found at 24 hours. This was in accordance with previous studies, which showed significant release of ROS following DON treatment at shorter exposure time (e.g. 15-60 minutes) 27,28 but not for longer exposure time (e.g. 24 hours) 26 . All these studies indicated that differences in exposure times could affect ROS generation.
In addition to ROS production, NO production was also measured. It was demonstrated that DON-induced NO production was partially inhibited by Sch A. Excessive production of ROS/RNS could induce oxidative stress and cause damage to lipids, proteins or DNA, thus affecting the normal cellular functions 45 . Consistent with previous studies, DON induced lipid peroxidation on intestinal cells as measured by MDA formation 25,27 . This may alter membrane integrity, cellular redox signalling and antioxidant status of the cells 46 . However, Sch A pre-treatment did not decrease the MDA production caused by DON in our study, suggesting that the protective mechanism of action of Sch A does not seem to be related to lipid peroxidation. Cellular response to oxidative stress is dependent of enzymatic and non-enzymatic anti-oxidant defences in the cell 27,28 . To fight against oxidative stress, a balance between of these factors needs to be attained. In the current study, both enzymatic (SOD, CAT, GPx and GST) and non-enzymatic antioxidants (reduced GSH) were measured. SOD detoxifies a superoxide anion to H 2 O 2 , CAT and GPx then convert H 2 O 2 to water and oxygen. GST is a secondary enzyme which plays a function in glutathione metabolism 47 . Extensive evidence in literature has shown that host cells respond to oxidative stress by increasing these antioxidant enzyme activities 48 . Our results are in agreement with earlier reports which demonstrated that DON resulted in elevated antioxidant enzymes SOD, CAT and GPx in response to oxidative stress 27,28 . Although oxidized glutathione (GSSG) was not measured in this study, a concomitant drop in total GSH levels was seen after DON treatment, which may be explained by the fact that GSH is used as an antioxidant in order to cease oxidative reactions triggered by DON by scavenging toxic free radicals and thus contribute to the anti-oxidative effects 49 . In fact, aberrant GSH levels can undesirably disturb various cellular processes, such as mitochondrial function, homeostasis and death 50 . Pre-treatment with Sch A, as an antioxidant, defended the cells from ROS, by preserving GSH levels. It is interesting to note that Sch A alone diminished cellular GSH level, which was in agreement with another study using macrophages/ www.nature.com/scientificreports www.nature.com/scientificreports/ lymphocytes in mice 39 . The anti-inflammation effect produced by Sch A may be linked to their ability to decrease cellular GSH, which may be partly associated with the modification of a redox-sensitive regulatory cysteine residue in NF-κB under the reduced cellular GSH content environments 51 .   www.nature.com/scientificreports www.nature.com/scientificreports/ To further characterize the mechanisms responsible for Sch A-mediated anti-oxidative functions, our study assessed whether Sch A influenced the DON-induced activation of Nrf2 pathway that modulates oxidative stress and inflammatory responses by controlling crucial antioxidant and detoxification enzyme genes via the antioxidant response element. DON treatment induced cellular oxidative stress, Nrf2 then dissociated from Keap-1 followed by its translocation to the nucleus. Different antioxidant enzymes (such as HO-1, GPx, etc.) increased to keep cells from oxidative injury. In the current study, DON significantly up-regulated HO-1 expression, a cytoprotective enzyme synthesized by Nrf2 activation, which was consistent with earlier reports 21,52 . Upon the Nrf2 nuclear translocation, it up-regulated HO-1 expression that helped to protect cells against oxidative stress. Furthermore, HO-1 level was also significantly increased by Sch A alone, which may imply that Sch A itself plays an crucial cytoprotective effect by removing ROS to conquer oxidative damage, inflammation and apoptosis 53,54 . It is surprising to note that pre-treatment with Sch A at 5 and 10 μM followed by DON exposure significantly decreased HO-1 expression. It suggested that activation of HO-1 can cause an initial rise in the cellular antioxidant status, but its continuous activation would reduce it. The HO-1 protective effect might be restricted to a relatively limited threshold of overexpression, and excessive HO-1 may even sensitize the cell to oxidative stress by releasing reactive iron 55 . Our results showed that DON aggravated the oxidative stress response that resulted in the unwanted up-regulation of HO-1 expression; pre-treatment with Sch A may exert a protective effect on HT-29 cells by preventing HO-1 overexpression, which may be even more detrimental to the cells.
NF-κB transcription factor plays a critical part in regulating immune and inflammatory responses, as well as controlling cell proliferation and cell death 56 . Furthermore, elevated ROS levels could activate NF-κB signalling through nuclear translocation 57 . NF-κB if it is activated excessively or improperly is undesirable to the host. The ability of DON to activate the NF-κB pathway has been extensively documented in literature 19,21,27,28,52 . In this study, we report that DON activated NF-κB during inflammation. Our immunofluorescent data clearly showed the DON induced translocation of Nrf2 from cytoplasm to nucleus. This result appears to be strongly linked to the mycotoxins' capacity to trigger ROS production. Sch A reduced DON-induced NF-κB activation by preventing its nuclear translocation. Also, Sch A diminished intracellular ROS, thus affected NF-κB activation and the downstream inflammatory responses.
It is evident that DON activates NF-κB signalling which then triggered various signalling pathways, such as MAPKs that are essential for controlling inflammation by mediating the production of inflammatory factors 19,20 . Therefore, to determine whether the MAPK signalling pathways contributed to the Sch A-induced inhibition of the inflammation, the three MAPKs (JNK, p38 MAPK and ERK) were examined. Our data revealed that the DON-induced phosphorylation of p38 and ERK were markedly suppressed by Sch A pre-treatment. This suggested that Sch A reduced inflammation caused by DON, at least partially, by down-regulating the MAPK signalling pathways, which then blocked NF-κB inactivation.
COX-2 is an immediate early response gene, and is regarded as an inflammatory marker and linked to numerous inflammatory conditions as well as carcinomas 58 . PGE2 is one of the important mediators produced at the inflammatory sites by the COX-2 enzyme. Our results were consistent to previous findings where COX-2 expression was induced following DON exposure 27,28 . However, the failure of induction of PGE2 by DON may be due to the relatively low concentration of DON employed in the current study. Pre-treatment with Sch A considerably diminished COX-2. Sch A also inhibits PGE2 production by down-regulating COX-2 mRNA and protein levels. Sch A appears to down-regulate DON-triggered COX-2 expression partially by affecting the cellular redox status and NF-κB activation. Sch A may also control COX-2 expression by modulating the Peroxisome proliferator activated receptor γ/Retinoid X receptor (PPARγ/RXR) complex, as demonstrated by other investigators 59 . Additional mechanistic studies are necessary to elucidate these.
In our study, we also examined if the Sch A effects on IL8 mRNA expression and secretion. IL8 is secreted by phagocyte and mesenchymal cells following inflammation, infection, ischemia, trauma, etc 60 . It activates www.nature.com/scientificreports www.nature.com/scientificreports/ neutrophil chemotaxis and promotes inflammation at the site of infection 61 . Our results showed that DON significantly stimulated IL8 mRNA expression and secretion as in other studies 19,20,22 . It has been proposed that the augmented IL8 level following DON exposure was resulted from the activation of the MAPKs and NF-κB inflammatory pathways 20 . Pre-treatment with Sch A effectively repressed the mRNA and secretion of IL8.
Over the years, research on dietary antioxidants and their cytoprotective role in preventing DON toxicity has gained enormous popularity. As DON is a frequently occurring food and feed contaminant; it would be of great interest to discover any substances that could be used as potential therapeutic to decrease the toxic effects caused by DON in the human body. Our study suggested that Sch A exerts its cytoprotective effects against DON by protecting the intestinal epithelial cells from cytotoxicity, oxidative damage and inflammation. Such effects were perhaps mediated by NF-κB, MAPKs and Nrf2/HO-1 signalling pathways. A summary of the cytoprotective mechanisms of Sch A on DON-induced toxicity was depicted in Fig. 12. Several Fusarium species and toxins such as nivalenol, zearalenone and fumonisins co-occur in wheat and maize 62 , Sch A modulation may have a www.nature.com/scientificreports www.nature.com/scientificreports/ positive effect on these mycotoxins. Application of modern molecular biology techniques will also help to elucidate the mechanistic pathways regarding the dynamic interaction that occurs between mycotoxins, Sch A and IECs. Although more future studies of the exact mechanisms is necessary, our study has demonstrated that Sch A may be used as a natural bioactive compound with anti-oxidative and anti-inflammatory functions and may be a useful therapeutic approach for oxidative or inflammation-mediated diseases such as chronic intestinal inflammatory diseases like IBD. Despite there are several limitations in this cell-culture approach, these data provide invaluable information for future designs of more comprehensive of animal and human studies.  22,63,64 and are similar to the levels that could be encountered in the gastrointestinal tract of animals or human tissues after ingestion of DON-contaminated food or feed. Assuming that DON ingested in one meal is diluted in 1 liter of gastrointestinal fluid and is totally bioaccessible, the in vitro concentrations to be used in this study correspond to food contamination ranging from 150 ng/g to 295 ng/g of DON. Sch A was added to the cells at a final concentration of 2.5, 5 and 10 μM. These concentrations were selected on the basis that they did not affect the cell viability as determined by the CCK assay. Cytoprotective effect of Sch A was studied by exposing cells to Sch A for 24 hours followed by DON for 24 hours. The final concentration of DMSO in all the treatments did not exceed 0.5% v/v. At the end of the treatment, CCK-8 solution (10 μl) was added and incubated at 37 °C for one hour. The optical density (O.D.) was recorded on the Multiskan microplate reader (ThermoFisher Scientific, Waltham, MA, USA) at 450 nm. Cell viability is expressed as the percentage of the mean value normalized to the negative control (untreated cells with 0.5% v/v DMSO). cell cycle analysis. HT-29 cells after 24 hours of Sch A pretreatment were exposed to DON for another 24 hours. The cells were fixed in 70% (v/v) ice cold ethanol and then stained with propidium iodide (PI) solution estimation of intracellular reactive oxidative species (RoS). Cells (1 × 10 4 cells/well) were incubated with 2′,7′-dichlorodihydrofluorescein diacetate (DCF-DA) (0.05 mM) (Sigma) for 30 minutes, followed by incubation with 1 μM DON for 0.5-6 hours. In the Sch A pre-treatment samples, the cells were treated with 2.5-10 μM Sch A for 24 hours and then exposed to DCF-DA for 30 minutes followed by 1 μM DON for 0.5-6 hours. For cells with 24 hours of DON treatment, cells were incubated with DCF-DA for 30 minutes at the end of incubation. The samples were read in a fluorimeter, with excitation and emission wavelengths of 480 nm and 520 nm (Model VICTOR X3, PerkinElmer, Courtaboeuf, France). The values were expressed as % relative fluorescence as compared to the control.

Materials and
Thiobarbituric acid reactive substances (TBARS) assay. The malondialdehyde (MDA) content was quantitated as a marker of lipid peroxidation. This assay was based on the reaction between MDA and 2-thiobarbituric acid (TBA) as a thiobarbituric acid reactive substance (TBARS) to form a 1:2 MDA-TBA adduct. Hence, the amount of TBARS correlated with the MDA produced. To obtain the whole cell lysates, cells undergoing the same treatment as described above, were lysed in (radioimmunoprecipitation assay) RIPA buffer. The total protein concentration was determined by the DC protein assay (Biorad) using bovine serum albumin (BSA; GE Healthcare, Chicago, IL, USA) as the standard. 100 μl of whole cell extracts were then incubated with 100 μl of 10% (v/v) trichloroacetic acid (TCA; Sigma) and centrifuged. The resulting supernatants (100 μl) were mixed with 100 μl of 8% (v/v) SDS solution and 1 ml of 0.8% TBA solution in 10% acetic acid (Sigma) and boiled in 95 °C water bath for 1 hour. The mixtures were cooled and centrifuged. The absorbance was measured at 532 nm with the Multiskan microplate spectrophotometer (ThermoFisher Scientific). The concentration of TBARS was calculated using the MDA standard curve and is expressed as μM/mg of protein.
nitric oxide (no) assay. Following the same treatments as described above, the level of NO in the culture supernatants was measured by the NO assay using the Greiss reagent kit (ThermoFisher Scientific, Waltham, MA, USA), which measures the amount of nitrite in the culture supernatants. Antioxidant enzymes. Cells following the above-mentioned treatments, were lysed in RIPA buffer as described above to obtain whole cell lysates for the following antioxidant enzyme assays.
Glutathione S transferase (GST) was determined by the method of Habig et al., with slight modifications 65 . The assay mixture contained 0.1 M phosphate buffer, pH 6.5, 100 mM 1-chloro-2,4-dinitrobenzene (CDNB; Sigma), 200 mM GSH (Sigma) and 20 μl extracted protein in a final volume of 200 μl. The change in absorbance was measured at 340 nm for 5 minutes at 1-minute intervals in a 96-well plate (Corning, NY, USA) using the Multiskan microplate spectrophotometer (ThermoFisher Scientific). The enzyme activity was calculated based on extinction coefficient of GS-CDNB, E340 = 0.0096 μM −1 cm −1 and expressed as nmoles CDNB conjugated/minute/mg protein. Catalase (CAT) activity was measured using an Amplex Red Catalase Assay Kit (Molecular Probes, Eugene, OR, USA) according to the manufacturer's instruction. Intracellular superoxide dismutase (SOD) activity was measured using the SOD Assay Kit-WST (Dojindo). Glutathione (GSH) level was quantified with a GSH/GSSG ratio assay kit (Abnova, Walnut, CA, USA) according to the manufacturer's instruction. Glutathione peroxidase (GPx) activity was determined as described previously with some modifications 49 . The reaction mixure contained 0.1 M phosphate buffer pH 7.5, 1 mM EDTA and 2 mM sodium azide, 0.1% Triton X-100, 2 mM reduced GSH and 0.2 mM NADPH, 2.5 U freshly prepared glutathione reductase (GR) and 0.25 mM H 2 O 2 . Twenty microliter of cell extracts were added to 200 μl of reaction mixture. One unit of GPx will reduce 1 μmol of glutathione disulfide (GSSG) per minute at pH 7.5. Assays were conducted at 25 °C during 5 minutes. GPx enzymatic activity was calculated by using the molar absorptivity of NADPH (6.22 mM −1 cm −1 ) and expressed as μmol of NADPH oxidized/minute/mg of protein.

Measurement of interleukin 8 (IL8) and prostaglandin E2 (PGE2).
Following the same treatments as described above, the level of IL8 in the culture supernatants was determined using IL8 human ELISA kit (ThermoFisher Scientific); and the PGE2 level was determined using PGE2 monoclonal enzyme immunoassay (EIA) kit (Cayman Chemicals, Ann Arbor, MI, USA) according to the manufacturer's instructions.
Real-time quantitative polymerase chain reaction (qPCR). HT-29 cells were treated as described above. Total RNA was extraction using RNAiso Plus following the manufacturer's protocol. Total RNA (500 ng) was converted to cDNA using the HiScript TM RT SuperMix for qPCR. qPCR was perfomed on a StepOnePlus Real-Time PCR system (Applied Biosystems, Foster City, CA) using AceQ qPCR SYBR Green Master Mix to measure the mRNA expression levels of the downstream gene of Nrf2 signalling pathway, heme oxygenase (HO-1). Furthermore, COX-2, a key enzyme in production of PGE2, together with inflammatory cytokines, including