A novel danshensu derivative ameliorates experimental colitis by modulating NADPH oxidase 4‐dependent NLRP3 inflammasome activation

Abstract We have previously reported a novel compound [4‐(2‐acetoxy‐3‐((R)‐3‐(benzylthio)‐1‐methoxy‐1‐oxopropan‐2‐ylamino)‐3‐oxopropyl)‐1,2‐phenylene diacetate (DSC)], derived from danshensu, exhibits cytoprotective activities in vitro. Here, we investigated the effects and underlying mechanisms of DSC on dextran sodium sulphate (DSS)‐induced experimental colitis. We found that DSC treatment afforded significant protection against the development of colitis, evidencing by suppressed inflammatory responses and enhanced barrier integrity. Intriguingly, DSC specifically down‐regulated DSS‐induced colonic NADPH oxidase 4 (Nox4) expression, accompanied by a balanced redox status, suppressed nuclear factor‐κB (NF‐κB) and NLRP3 inflammasome activation and up‐regulated nuclear factor (erythroid‐derived 2)‐like 2 and haeme oxygenase‐1 expression. In vitro study also demonstrated DSC also markedly decreased Nox4 expression and activity associated with inhibiting reactive oxygen species generation, NF‐κB activation and NLRP3 inflammasome activation in bone marrow‐derived macrophages. Either lentiviral Nox4 shRNA‐mediated Nox4 knockdown or Nox4‐specific small‐interfering RNA mimicked effects of DSC by suppressing NLPR3 inflammasome activation to alleviate experimental colitis or inflammatory macrophage response. Collectively, our results provide the first evidence that DSC ameliorates experimental colitis partly through modulating Nox4‐mediated NLRP3 inflammasome activation.


| INTRODUC TI ON
Inflammatory bowel disease (IBD), including ulcerative colitis and Crohn's disease, is a common relapsing inflammatory disorder of the gastrointestinal tract and is clinically characterized by recurrent and long-lasting episodes of diarrhoea and abdominal pain. 1,2 Mucosal barrier damage, inflammation homeostasis disruption and inflammatory cell infiltration are associated with decay process of the illness. 3 Current therapies for IBD depend mainly on anti-inflammatory drugs, antibiotics or biologics that are either not all effective, of high cost or with adverse effects. Therefore, the development of novel nutraceuticals or therapies for IBD remains imminently needed. 4 Accumulating evidence suggests that IBD arises from aberrant innate and/or adaptive immune responses due to destroy intestinal homeostasis. 5 Inflammatory conditions in general have been linked to the overproduction of reactive oxygen species (ROS). 6 Overproduction of ROS is a hallmark of inflammation and also elevated in clinic and experimental colitis. [7][8][9] Excessive ROS also further induces inflammatory cytokine (such as interleukin-6 (IL-6)) release and enzyme (such as cyclooxygenase-2 (COX-2) and inducible nitric oxide synthase (iNOS)) expression in colon. Mounting evidence indicates nicotinamide adenine dinucleotide phosphate (NADPH) oxidase family, a group of membrane enzymes producing ROS, is responsible for pathological progression of colitis. [9][10][11][12][13] Increased NADPH oxidase expression and activity have also been correlated with the severity of IBD in humans and in animal models. 14 Thus far, seven NADPH oxidase isoforms (Nox1-5 and Duox1 and 2) have been identified, which form the basis of distinct NADPH oxidases and have varying requirements for other protein subunits. 15 Distinct from other NADPH oxidase members, Nox4 is not expressed constitutively in normal tissues but can be rapidly induced in various cell types and tissues, including the colon. [16][17][18] In addition, Nox4 is constitutively active and does not require cytosolic factors for its activation. 19 Meanwhile, inhibition or knockdown of Nox4 attenuates inflammatory responses in macrophage and ischaemic myocardium. [16][17][18]20 Ours as well as other previous studies have demonstrated that enhanced Nox4 appears important in various diseases, including colitis. 13,16,[21][22][23][24] Importantly, it is still unclear to what extent Nox4 is important in mediating the process of colitis.
Salvia miltiorrhiza Bunge (S miltiorrhiza) is a traditional Chinese herbal medicine of the Labiatae family and has been implicated in cardiovascular disorders, blood circulation diseases, neurodegenerative disorders and gastrointestinal inflammation. 25 S miltiorrhiza contains broadly two types of bioactive constituents, hydrophilic phenolic (danshensu, salvianolic acid B, etc) and lipophilic quinines (tanshinone I, tanshinone IIA, dihydrotanshinone I, etc). 26 Danshensu (3-(3',4'-dihydroxyphenyl)-(2R)-lactic acid) has been demonstrated with various therapeutic effects. 27 However, its chemical instability of the phenolic hydroxyl groups and relative low abundance in S miltiorrhiza limit its application. Our groups have earlier synthesized danshensu derivatives asymmetrically to improve the chemical stability while improving the desired biological functions. 21 Importantly, we have synthesized and demonstrated its L-cysteine derivative S-propargyl-cysteine with various salubrious biological activities, including anti-inflammatory, anti-oxidative and anti-apoptotic activities. 28 To further obtain a polyvalent pharmaceutical or nutraceutical candidate, we chemically synthesized a series of novel amide and thioester conjugates of danshensu-cysteine derivatives by joining two drugs through an appropriate linker or bond in the light of the guidance of medicinal chemical hybridization. 26 Among them, a novel conjugate [4-(2-acetoxy-3-((R)-3-(benzylthio)-1-methoxy-1oxopropan-2-ylamino)-3-oxopropyl)-1,2-phenylene diacetate] (DSC, Figure 1A) has been demonstrated with eminent anti-inflammatory and anti-oxidative properties in vitro and in vivo. 26,29,30 In the present study, we explored the therapeutic potential of DSC in dextran sulphate sodium (DSS)-induced experimental colitis with special focus on finding novel intracellular molecular target of it.

| Animals and ethics
Seven-to eight-week-old male C57BL/6J mice (18-20 g) were purchased from JOINN Laboratories, Inc (Suzhou, China). They were acclimated for 1 week with tap water and basal diet under the conventional housing conditions of humidity (50 ± 10%), temperature (23 ± 2°C) and light (12/12 hours of light/dark cycle). All animalrelated experimental protocols were approved by the Animal Ethics  In short, all mice were anaesthetized by sodium pentobarbital (50 mg/kg, ip) and given an intrarectal enema of 50% ethanol (vol/ vol in distilled H 2 O) before instillation. Pretreatment with ethanol enemas has been shown to increase intestinal transduction with other vectors. Two hours after the enema, 0.1 mL of the concentrated viral suspension with a titre of 1 × 10 9 TU/mL or vehicle was instilled intrarectally. The mice were inverted for 30 seconds after administration of intrarectal products to prevent leakage.

| Induction of colitis and treatment
Intrarectal instillation was carried out 3 days before 2.5% DSS stimulation. Mice were randomly assigned to control group, DSStreated group, lentiviral control shRNA-treated group (DSS-treated mice also administered lentiviral control shRNA), lentiviral Nox4 shRNA-treated group (DSS-treated mice also administered lentiviral Nox4 shRNA) (n = 8). At the end of the treatment, mice were killed and the colonic tissues were taken out. The tissues were separated into several parts for extraction of protein and mRNA for subsequent assays. The distal part was fixed in 4% buffered formalin for examination, and other parts were frozen in liquid nitrogen immediately and kept at −80°C until use.

| Plasmid construction and lentiviral particle production
Lentivirus generation was performed as our previously described. 21 The sequences of small hairpin RNA (shRNA) targeting mouse Nox4

| Scoring of disease activity index
The disease activity index (DAI) was calculated by assigning well-established and validated scores. Scores were evaluated based on weight loss, stool type and bleeding as described. 33 All scores were assessed by an independent observer blinded to the treatment.

| Macroscopic assessment and histologic analysis
After the mice were killed, colon and spleen tissues were taken out and length of the colon and the wet weight of the spleen were measured. Colonic shortening was determined by measuring the length between the ileocecal junction and the proximal rectum. An independent observer who was blinded to treatment status determined the macroscopic characteristics. Briefly, the macroscopic scores were 0 (no damage), 1 (hyperaemia without ulcers), 2 (hyperaemia and wall thickening without ulcers), 3 (one ulceration site without wall thickening), 4 (two or more ulceration sites), 5 (0.5-cm extension of inflammation or major damage) and 6-10 (1-cm extension of inflammation or severe damage. 34

| Western blot analysis
Western blot analysis was carried out as our previously described. 36 Equal amounts of protein were separated by electrophoresis in SDS-PAGE and then transferred onto polyvinylidene difluoride membranes (Millipore, USA). The membranes were blocked with 5% nonfat dried milk and then incubated with respective primary antibodies overnight at 4°C. After being washed with Tris-buffered saline with Tween, the membranes were incubated with horseradish peroxidaseconjugated anti-rabbit or antimouse secondary antibodies (1:5000, Thermo Fisher Scientifics) for 2 hours at room temperature. The reaction was followed by enhanced chemiluminescence reaction (Thermo Fisher Scientifics), and the blots were quantified using Alpha Imager

| Cytokine measurement
The levels of IL-1β and IL-6 in serum and the levels of IL-1β in culture supernatant of BMDM were measured by ELISA kits (R&D Systems, Minneapolis, MN, USA) following the standard procedure of the manufacturer. Absorbance was measured at 450 nm with a microplate reader (M1000, TECAN, Austria GmbH, Grödig, Austria).

| Real-time quantitative PCR (qPCR)
Total RNA was isolated from colonic tissues by using Trizol (TaKaRa Biotechnology, Dalian, China) according to manufacturer's protocol.
RNA (1 μg) of each sample was reverse transcribed using the reverse transcription system of Takara (TaKaRa) following the manufacturer's instructions. An equal volume of cDNA was used as a polymerase chain reaction (PCR) template for determining the mRNA expression level using SYBR-Green Quantitative PCR kit (TaKaRa) by iCycler iQ system (Bio-Rad, Hercules, CA, USA). The relative mRNA levels were normalized to mRNA levels of β-actin (housekeeping control), and calculations for fold change of each mRNA were made on comparative cycle threshold method (2 −ΔΔCt ). The primers used in this study are provided in Supporting Information Table S1.

| Tissue and mitochondrial ROS production measurement
For analysing tissue ROS generation, the redox-sensitive, cell-permeable superoxide indicator dihydroethidium (DHE) was used to detect the cellular production of superoxide in vivo. 37

| Glutathione (GSH) and glutathione disulphide (GSSG) assay
The tissue GSH and GSSG levels were measured using GSH and GSSG assay kit (S0053, Beyotime Institute of Biotechnology, Shanghai, China) according to the manufacturer's instructions. Briefly, protein removal reagent mol/L solution (70 μL) was added to homogenates of colonic tissues (10 mg). Tissue homogenates were allowed to stand at 4°C for 10 minutes and then centrifuged for 10 000 g at 4°C for 10 minutes. The samples were centrifuged, and the supernatant was subjected to GSH and GSSG assay. The total glutathione level was measured by the 5,5′-dithiobis (2-nitrobenzoic acid)-GSSG recycling assay. The absorbance was measured at 412 nm by a microplate reader (M1000, TECAN, Austria GmbH, Grödig, Austria). The amount of total glutathione in samples was calculated according to the standard curve. After samples and standards were treated with 2-vinylpyridine to block GSH, the GSSG level was quantified similarly to the total glutathione. The amount of GSH was obtained by subtracting the amount of GSSG from the total glutathione. Protein concentration of the supernatants was determined by bicinchoninic acid assay. Values were normalized to protein concentration and expressed as GSH/GSSG ratio.

| Cell transfection
To introduce siRNA into BMDM, the cells were plated in 6-well or 24-

| Statistical analysis
Data are expressed as means ± SEM. Statistical analysis was performed by one-way analysis of variance (ANOVA) followed by Tukey's post hoc test for multiple comparisons and Student's twotailed t test was used for comparing two groups using Stat View of GraphPad Prism (version 6; GraphPad Software Inc, San Francisco, CA, USA). The values of P < .05 were considered to indicate a statistically significant difference.

| DSC attenuates DSS-induced experimental colitis
The optimal dose of DSC for treating mice with DSS-induced co-  Figure 1D). Macroscopic analysis revealed that DSC treatment improved the colonic histological score in DSS-treated mice ( Figure 1E).
Collectively, these results indicate that DSC attenuates DSS-induced colonic damage in mice.  (Figure 2A,B). Similarly, DSC also induced muc-2 mRNA expression in experimental colitis (Figure 2A). Inflammatory responses and leucocytes infiltration underpin the development of both experimental and clinical IBD, leading to the mucosal damage. 5 As shown in Figure 2C

| DSC alleviates colonic Nox4-mediated redox imbalance
NADPH oxidases have been implicated in the pathogenesis of colitis. [9][10][11] To understand the underlying mechanisms of DSC in suppressing experimental colitis, expression of Nox1, Nox2 and Nox4 in colon  Figure 3I).
In line with our previous study, 21 we also found that Nox4

| DSC inhibits the Nox4-mediated activation of NLRP3 inflammasome
It is reported that activation of NLRP3 inflammasome plays a critical role in DSS-induced colitis. 38 Next, we investigated the effect of DSC on NLRP3 inflammasome in vivo and in vitro. NLRP3 inflammasome was activated after DSS administration, evidencing by up-regulated NLRP3, ASC caspase-1 p20, cleaved IL-1β and cleaved IL-18 expression in the colon, which was markedly attenuated by DSC treatment ( Figure 5A). Up-regulation of Nox4 promotes NLRP3 inflammasome activation in vitro and in vivo. 20 As expected, lentiviral Nox4 shRNA, but not control shRNA, mimicked the effects of DSC on alleviating DSS-mediated NLRP3 inflammasome activation ( Figure 5B). To further determine whether Nox4 was involved in the inhibitory effects of DSC on NLRP3 inflammasome activation, we induced activation of NLRP3 inflammasome with nigericin in BMDM primed with LPS. As shown in Figure 5C, DSC (50 μmol/L) alone had no effect on NLRP3 inflammasome activation. Intriguingly, DSC treatment, F I G U R E 4 DSC inhibits Nox4-mediated redox imbalance in BMDM. BMDM was treated with or without DSC (50 μmol/L) or Nox4 siRNA as described in Materials and Methods. A, Representative bands and quantitative analysis of Nox4. B, Representative images and quantitative analysis of mitochondrial ROS production. C, Quantitative analysis of intracellular H 2 O 2 generation. D, Quantitative analysis of NADPH oxidase activity. Data shown are means ± SEM of n = 8 in each group. *P < 0.05 vs Control cell (CTL), # P < 0.05 vs LPS-stimulated BMDM F I G U R E 5 DSC inhibits the activation of NLRP3 inflammasome in vivo and in vitro. A, Colitis was induced as described in Materials and Methods and treated with or without DSC (50 mg/kg). Representative bands and quantitative analyses of NLRP3, caspase-1 p20, cleaved IL-1β and cleaved IL-18 in colonic tissues. The β-actin was used as loading control (n = 8 in each group). B, Colitis was induced as described in Materials and Methods and treated with or without lentiviral Nox4 shRNA. Representative bands and quantitative analyses of NLRP3, caspase-1 p20, cleaved IL-1β and cleaved IL-18 in colonic tissues. The β-actin was used as loading control (n = 8 in each group). C, NLRP3 inflammasome was induced in BMDM as described in Materials and Methods and treated with DSC (50 μmol/L). Representative bands of NLRP3, ASC and caspase-1 in BMDM and caspase-1 p20 in supernatant (SN) by Western blot and quantitative analyses of IL-1β in supernatant by ELISA. The β-actin was used as loading control. Data shown are means ± SEM of n = 8 in each group. *P < 0.05 vs Control cell or mice (CTL), # P < 0.05 vs DSS-treated mice or LPS + nigericin-stimulated cells mimicking Nox4 knockdown, dramatically reduced NLRP3 inflammasome activation and IL-1β secretion in BMDM. Taken together, these results suggest that DSC suppresses NLRP3 inflammasome activation by modulating Nox4.

| DSC inhibits NF-κB activation and induces Nrf2-mediated signalling pathways
NF-κB is a transcription factor regulating expression of inflammatory mediators (including NLRP3, COX-2 and iNOS) and various proinflammatory cytokines, and the effect of DSC on the activation of NF-κB was evaluated. Phosphorylation of p65 at serine 536 appeared to play a critical role in NF-κB DNA binding activity. 39 As shown in Figure 6A, DSC treatment dramatically reduced DSS-mediated the phosphorylation of NF-κB p65 (serine 536) in colonic tissues, which was also mimicked by lentiviral Nox4 shRNA ( Figure 6B). Similarly, both DSC and silencing Nox4 remarkably suppressed NF-κB p65 phosphorylation (serine 536) and NF-κB p65 DNA binding activity in LPS-stimulated BMDM ( Figure 6C,D).
Our previous study reported that induction of Nrf2/HO-1 might be, at least in part, responsible for the cytoprotective property of DSC. 26 In the present study, the nuclear accumulation of Nrf2 and  Figure 7D). Together, our data indicate that DSC restore redox homeostasis by down-regulating Nox4 expression, contributing to suppressed inflammatory response.

| D ISCUSS I ON
Danshensu, a hydrophilic bioactive constitute of Danshen, attracts considerable interest due to its salubrious biological activities. We have earlier reported prominent anti-oxidative and anti-inflammatory activities with a novel danshensu derivative DSC in vitro and in vivo. 26,29,30 In the present study, we clearly demonstrated that DSC suppressed DSS-induced experimental colitis in vivo.
Mechanistically, the anti-inflammatory effects of DSC on inhibition of NF-κB-NLRP3 inflammasome activation were mediated via modulating Nox4-mediated ROS signalling pathway. Our study suggests DSC mitigating Nox4 signalling as a novel therapeutic strategy for IBD. F I G U R E 6 DSC inhibits NF-κB activation. Colitis was induced as described in Materials and Methods and treated with or without DSC (50 mg/kg). A, Representative bands and quantitative analysis of p-NF-κB p65 in colonic tissues, NF-κB p65 was used as loading control. B, Colitis was induced as described in Materials and Methods and treated with or without lentiviral Nox4 shRNA. Representative bands and quantitative analysis of p-NF-κB p65 in colonic tissues, NF-κB p65 was used as loading control. Overproduction of ROS may stimulate NF-κB activation, thereby leading to activation of NLRP3 inflammasome and release of IL-1β and other inflammatory mediators including iNOS and COX-2, which are associated with IBD. 14,43 Aberrant over-activation of NF-κB plays a critical role in the pathological process of IBD. 44 In addition, Nox4derived ROS generation is one of common mechanisms of NLRP3 inflammasome activation, which may additionally contribute to accelerated inflammatory responses. 20,23 Phosphorylation of NF-κB p65 (serine 536) controls the kinetics of NF-κB nuclear translocation. 39,45 Meanwhile, phosphorylation of p65 at serine 536, but not other sites, appears to play a pivotal role in p65 DNA binding activity and transcriptional activity. 46 In support of this, our results F I G U R E 7 DSC restores redox balance in colonic tissues. A, Colitis was induced as described in Materials and Methods and treated with or without DSC (50 mg/kg). Representative bands and densitometry analysis of Nrf2 nuclear translocation and HO-1 expression in colonic tissues. Histone H3 and β-actin were used as loading control, respectively. B, Colitis was induced as described in Materials and Methods and treated with lentiviral Nox4 shRNA or lentiviral scrambled shRNA. Representative bands and densitometry analysis of Nrf2 nuclear translocation and HO-1 expression in colonic tissues. Histone H3 and β-actin were used as loading control, respectively. BMDM was stimulated with or without DSC (50 μmol/L) as described in Materials and Methods. C, Representative bands and densitometry analysis of Nox4. β-actin was used as loading control. D, Representative bands and densitometry analysis of Nrf2 nuclear translocation and cytoplasmic HO-1 expression in BMDM. β-actin or histone H3 was used as loading control. Data shown are means ± SEM of n = 8 in each group. *P < 0.05 vs Control cell or mice (CTL), # P < 0.05 vs DSS-treated mice or LPS-stimulated cells demonstrated that both DSC treatment and genetic Nox4 knockdown interfered with DSS-induced Nox4 expression and NF-κB activation in experimental colitis in mice. Additionally, we also observed that Nox4 silencing or DSC treatment significantly attenuated LPSmediated ROS generation and NF-κB activation in inflammatory macrophages. NF-κB is one key mechanism to mediate NLRP3 inflammasome activation, which may additionally contributes to accelerated inflammatory responses. 47 Emerging evidence suggests the pivotal role of NLRP3 inflammasome in the development and pathogenesis of IBD and NLRP3 -/mice were protected in experimental colitis. 48 Consistent with these findings, we showed that DSC dramatically prevented NF-κB activation and subsequently suppressed NLRP3 inflammasome activation (NLRP3, ASC, cleaved caspase-1, cleaved IL-1β and IL-18) in vivo and in vitro. Taken together, these data supported that DSC alleviated experimental colitis at least in part through inhibiting Nox4-mediated NF-κB and NLRP3 inflammasome activation.
Loss of redox homeostasis plays a pivotal role in the development of IBD. 49 The transcription factor Nrf2 is a master regulator of redox homeostasis. 50 Nrf2 regulates the expression of a battery of cytoprotective genes, including HO-1. 51 We found that DSC significantly induced nuclear translocation of Nrf2 and increased HO-1 expression in colonic tissues, associated with reduced ROS generation and NF-κB activation. Our findings were consistent with previous reports that showed the protective effects of Nrf2 on the colitis are mediated by mechanisms involving ROS scavenging and/or inhibition of the NF-κB pathway. 52,53 Aberrant Nox4 up-regulation and the impairment to induce the Nrf2-mediated antioxidant response indicated the involvement of oxidative stress. 54 On the other hand, inducing Nrf2 activation is shown to inhibit Nox4 expression. 54 Consistent with the previous report, we found that DSC activated Nrf2 with the suppression of Nox4 expression in vivo and in vitro. In LPS-stimulated BMDM, the effects were abolished by Nrf2 siRNA. In addition, increasing evidence has suggested that ROS produced by NADPH oxidases can activate Nrf2 and induce HO-1. 55,56 In LPS-mediated inflamed BMDM, we showed that an aberrant up-regulation of Nox4 was coupled with a Nrf2 induction, resulting in a sustained redox imbalance, which promoted inflammatory response. Interestingly, Nox4 siRNA knockdown markedly decreased ROS generation and subsequently down-regulated Nrf2 nuclear translocation. The present data are consistent with the earlier report that Nox4 can activate the Nrf2-regulated pathway through oxidative post-translational modification. 57 Taken together, Nrf2 and Nox4 reciprocally regulate the inflammatory response in DSS-mediated colitis and LPS-stimulated BMDM, and therefore, it is logical to believe that DSC restores redox homeostasis by regulation of the balance between Nrf2 and Nox4.

| CON CLUS IONS
In summary, our study demonstrated that DSC alleviated DSSinduced colitis by inhibiting Nox4-mediated NF-κB and NLRP3 inflammasome activation. Although the protective mechanism by which DSC modulates various signalling pathways remains to be fully understood, our findings suggest that DSC has promising therapeutic potential for IBD.

ACK N OWLED G EM ENTS
The work was supported by funds from the National Natural

CO N FLI C T O F I NTE R E S T
The authors have declared no conflict of interest.

DATA AVA I L A B I L I T Y S TAT E M E N T
The authors declare that the main data supporting the findings of this study are available within the article and its Supplementary Information. Extra data are available from the corresponding author upon request.