COX-2/sEH Dual Inhibitor PTUPB Attenuates Epithelial-Mesenchymal Transformation of Alveolar Epithelial Cells via Nrf2-Mediated Inhibition of TGF-β1/Smad Signaling

Background Arachidonic acid (ARA) metabolites are involved in the pathogenesis of epithelial-mesenchymal transformation (EMT). However, the role of ARA metabolism in the progression of EMT during pulmonary fibrosis (PF) has not been fully elucidated. The purpose of this study was to investigate the role of cytochrome P450 oxidase (CYP)/soluble epoxide hydrolase (sEH) and cyclooxygenase-2 (COX-2) metabolic disorders of ARA in EMT during PF. Methods A signal intratracheal injection of bleomycin (BLM) was given to induce PF in C57BL/6 J mice. A COX-2/sEH dual inhibitor PTUPB was used to establish the function of CYPs/COX-2 dysregulation to EMT in PF mice. In vitro experiments, murine alveolar epithelial cells (MLE12) and human alveolar epithelial cells (A549) were used to explore the roles and mechanisms of PTUPB on transforming growth factor (TGF)-β1-induced EMT. Results PTUPB treatment reversed the increase of mesenchymal marker molecule α-smooth muscle actin (α-SMA) and the loss of epithelial marker molecule E-cadherin in lung tissue of PF mice. In vitro, COX-2 and sEH protein levels were increased in TGF-β1-treated alveolar epithelial cells (AECs). PTUPB decreased the expression of α-SMA and restored the expression of E-cadherin in TGF-β1-treated AECs, accompanied by reduced migration and collagen synthesis. Moreover, PTUPB attenuated TGF-β1-Smad2/3 pathway activation in AECs via Nrf2 antioxidant cascade. Conclusion PTUPB inhibits EMT in AECs via Nrf2-mediated inhibition of the TGF-β1-Smad2/3 pathway, which holds great promise for the clinical treatment of PF.


Introduction
Pulmonary fibrosis (PF) is a prototype of chronic, progressive, and fibrotic lung disease. An altered extracellular matrix replaces healthy tissue, and alveolar architecture is destroyed, which leads to decreased lung compliance, disrupted gas exchange, and ultimately respiratory failure and death [1]. Although pirfenidone and nintedanib have been authorized by the Food and Drug Administration [2], they only slow down lung function decline in patients with the mild and moderate disease [3]. Therefore, it is urgent to develop an effective treatment for PF.
Epithelial-mesenchymal transition (EMT) is a reversible process in which epithelial cells lose their cellular polarity and obtain migration characteristics through downregulation of E-cadherin-mediated cell adhesion [4]. EMT is involved in wound healing, fibrosis, embryonic development, and cancer metastasis [5]. Most investigators concur that alveolar type II epithelial cells undergo EMT during PF development [6,7]. Studies have shown that pulmonary fibroblasts are derived from various routes, of which about one-third are derived from alveolar type II epithelial cells via EMT [8]. Transforming growth factor (TGF)-β1 is the most studied and is a key EMT inducer [9]. TGF-β1 activates its downstream Smad signaling pathway and plays an important role in fibrosis [10]. TGF-β1 binds to its receptor to trigger intracellular signaling and phosphorylates Smad2 and Smad3. Phosphorylated Smad2 and Smad3 are transported to the nucleus and regulate the transcription of target genes [11]. Consequently, blocking the EMT of alveolar epithelial cells (AECs) might be a promising strategy for the treatment of PF.
Oxidative stress accelerates TGF-β1-mediated fiber formation by increasing hydrogen peroxide levels, protein damage, DNA degradation, and lipid peroxidation [12]. The transcription factor nuclear factor erythroid 2-related factor-2 (Nrf2) plays an important role in intracellular antioxidant responses. Activated Nrf2 is transported to the nucleus to promote the transcription of antioxidant enzymes [13]. Nrf2 balances not only oxidative stress but also has negative effects on TGF-β1-mediated profibrotic signal transduction [14,15]. Previous studies have shown that Nrf2 plays an important role in preventing lung inflammation and fibrosis [16,17]. These results indicate that strategies targeting Nrf2 have antipulmonary fibrosis potential.
Our previous study suggested that the expressions of sEH and COX-2 are significantly increased in the lungs of PF mice induced by bleomycin (BLM) [30]. A compound that concurrently inhibits both COX-2 and sEH is called 4-(5-phenyl-3-{3-[3-(4-trifluoromethylphenyl)-ureido]-pro-pyl}-pyrazol-1-yl)-benzenesulfonamide (PTUPB), which prevents the release of PGs and increases the blood levels of EETs [31]. PTUPB is more potent in suppressing inflammatory pain and tumor growth than celecoxib, t-AUCB (an inhibitor of sEH), or the combination of celecoxib and t-AUCB [31,32]. We have shown that PTUPB can alleviate acute lung injury [33], nonalcoholic fatty liver disease [34], and sepsis [35] in mice. What is more, we have found that PTUPB significantly attenuates BLM-induced PF in mice [30]. However, it is not clear whether PTUPB can inhibit EMT. Therefore, the present study aimed to investigate the effects of PTUPB on TGF-β1-mediated pulmonary EMT.

Materials and Methods
2.1. Animal. C57BL/6J mice (adult male, 6-8 weeks) were obtained from Hunan SJA Laboratory Animal Co., Ltd. (Hunan, China). Mice were placed in specific pathogenfree conditions for a 12 h day-night cycle. Mice have free access to food and water.

Pulmonary Histopathology
Analysis. The left lung tissue was placed in a tube filled with 4% paraformaldehyde (Servicebio, Wuhan, China, G1101), followed by conventional paraffin embedding. Paraffin-embedded sections were made. Hematoxylin-eosin staining (HE) was used to observe the morphological changes in lung tissue of mice, and Masson staining was used to observe the collagen deposition.   . The deposition of α-SMA and E-cadherin was detected by immunofluorescence ((g) bar =100 μm). Data are expressed as the mean ± SD. Differences among multiple groups were performed using ANOVA. Tukey's test was used as a post hoc test to make pairwise comparisons. * P < 0:05, * * P < 0:01, and * * * P < 0:001.   Data are expressed as the mean ± SD. Differences between the two groups were determined by an unpaired t-test. * P < 0:05.   was reverse transcribed using PrimeScript RT reagent Kit (Takara). Real-time PCR was carried out to detect mRNA expression levels as described in our previous study [37]. Relative expression of genes was computed by the 2 -ΔΔCT method according to our previous study [38]. The sequence of primers used in this study is shown in Table 1  . The data shown are from a representative experiment with biological triplicates. Data are expressed as the mean ± SD. Differences among multiple groups were performed using ANOVA. Tukey's test was used as a post hoc test to make pairwise comparisons. * P < 0:05, * * P < 0:01, and * * * P < 0:001. 7 Oxidative Medicine and Cellular Longevity 2.11. Statistical Analyses. All data were presented as means ± standard deviation. Statistical analysis was performed using GraphPad Prism 7 (GraphPad Software, Inc., San Diego, CA, USA). Multiple group comparisons were made using a one-way analysis of variance. Tukey's test was used as a post hoc test to make pairwise comparisons. Differences between the two groups were determined by an unpaired ttest. All experiments were independently repeated three times. P < 0:05 was considered statistically significant.

Results
3.1. PTUPB Reduces PF in Mice Induced by BLM. In this study, a COX-2/sEH dual inhibitor PTUPB (5 mg/kg, s.c. once a day) was employed on the 7th day after BLM admin-istration. HE and Masson staining results showed that PTUPB treatment for 14 days also significantly reduced BLM-induced lung histological changes and collagen deposition in the lungs (Figure 1(a)). PTUPB significantly decreased Collagen I protein (Figures 1(b) and 1(c)) and the expression of tissue inhibitors of metalloproteinase 1 (Timp1) mRNA (Figure 1(d)). At the same time, we found that PTUPB significantly reduced α-SMA expression and restored E-cadherin expression in the lungs (Figures 1(e)-1  (g)). These results suggest that the reduction of PF by PTUPB is related to the reduction of EMT.  . The data shown are from a representative experiment with biological triplicates. Data are expressed as the mean ± SD. Differences among multiple groups were performed using ANOVA. Tukey's test was used as a post hoc test to make pairwise comparisons. * P < 0:05, * * P < 0:01, and * * * P < 0:001.

Oxidative Medicine and Cellular Longevity
We found that both COX-2 and sEH protein levels were increased in TGF-β1-treated A549 (Figures 2(a)-2(c)) and MLE-12 cells (Figures 2(d)-2(f)), indicating that dysregulation of ARA metabolism participates in the development of EMT. These results suggest an important role of COX-2 and sEH dysregulation in the development of EMT.

Prophylactic Treatment of PTUPB Suppresses the TGF-
β1-Induced EMT in AECs. Then, we wondered whether PTUPB suppressed the EMT induced by TGF-β1 in vitro. We observed that PTUPB alone did not affect the EMT of A549 cells ( Figure S1). Further, we found that the treatment with TGF-β1 (10 ng/mL) for 12 h significantly increased the mRNA expression of actin alpha 2 (ACTA2) (encoding α-SMA) and Vimentin, indicating the occurrence of EMT, which was effectively suppressed by the pretreatment with PTUPB in A549 cells (Figures 3(a)  and 3(b)). We found that PTUPB (1 μM) was the most effective inhibition concentration. In addition, western blotting results showed that the pretreatment with PTUPB (1 μM) reduced α-SMA protein expression and restored Ecadherin protein expression induced by TGF-β1 (10 ng/ mL) (Figures 3(c)-3(h)). Collectively, these results provide strong evidence that PTUPB directly suppresses the EMT induced by TGF-β1 in AECs.

Prophylactic Treatment of PTUPB Inhibits the Migration
Induced by TGF-β1 in A549 Cells. We further investigated the effect of PTUPB on TGF-β1-induced cell migration with the scratch wound-healing assay. The results showed that TGF-β1 treatment (10 ng/mL) for 48 h significantly pro-moted the migration of A549 cells. PTUPB could significantly reduce this effect (Figures 4(a) and 4(b)). In order to confirm that PTUPB inhibits cell migration but not cell proliferation, we further evaluated proliferation with CCK-8. Results showed that this effect did not attribute to the alteration of cell proliferation (Figure 4(c)). Taken together, these results indicate that PTUPB suppresses cell migration by inhibiting EMT in AECs.

Prophylactic Treatment of PTUPB Inhibits the Collagen
Synthesis Induced by TGF-β1 in AECs. The collagen synthesis can directly reflect the severity of PF. We found that the gene expression of COL1A1 and fibronectin (FN) was significantly increased in A549 cells stimulated by TGF-β1, which was effectively suppressed by the pretreatment with PTUPB (Figures 5(a) and 5(b)). TGF-β1 treatment also induced the increase of protein expression of Collagen I in A549 cells and MLE-12 cells (Figures 5(c)-5(f)). Pretreatment with PTUPB restored these changes induced by TGF-β1. Altogether, these results indicate that PTUPB inhibits the TGF-β1-induced collagen synthesis in AECs.

Prophylactic Treatment of PTUPB Disrupts the TGF-β1-Smad2/3 Signaling Pathway in AECs.
To elucidate the mechanism of PTUPB on EMT, we focused on the downstream signaling pathways of TGF-β1, including Smad, MAPK, and PI3K signaling pathways. We found that PTUPB had no significant effect on MAPK and PI3K signaling pathways after TGF-β1 activation ( Figure S2). However, after PTUPB pretreatment, TGF-β1-induced phosphorylation of Smad2 and Smad3 in A549 cells was significantly reduced . The data shown are from a representative experiment with biological triplicates. Data are expressed as the mean ± SD. Differences among multiple groups were performed using ANOVA. Tukey's test was used as a post hoc test to make pairwise comparisons. * P < 0:05, * * P < 0:01, and * * * P < 0:001.

12
Oxidative Medicine and Cellular Longevity  Figure S3). Furthermore, immunofluorescence was used to observe that PTUPB reduced nuclear translocation of Smad2 and Smad3 in TGF-β1-stimulated MLE12 cells (Figure 6(g)). Then, we found that treatment with PTUPB suppressed the gene expression of the downstream targets of TGF-β1-Smad2/3 signaling, including ZEB1 and SNAIL1 (Figures 6(h) and 6(i)). These data indicate that PTUPB blocks the TGF-β1 signaling pathway through the inhibition of TGF-β1-Smad2/3 activation in AECs.

16
Oxidative Medicine and Cellular Longevity activation of Nrf2 plays an important role in the regulatory effects of PTUPB on the TGF-β1/Smad axis.

Discussion
The transition of AECs into mesenchymal cells has been reported to cause and/or aggravate PF [6]. In this study, the direct effects of PTUPB on the TGF-β1-induced EMT were investigated. We found that PTUPB restored the phenotype changes, reduced the migration ability, and inhibited the collagen synthesis of TGF-β1-stimulated AECs by disrupting the TGF-β1-Smad2/3 pathway. We demonstrate for the first time that PTUPB blocks TGF-β1-induced EMT in AECs by inhibiting the TGF-β1-Smad2/3 signaling pathway. We found that PTUPB restored phenotypic changes, reduced migration ability, and inhibited collagen synthesis of TGF-β1-stimulated AECs. We demonstrated for the first time that PTUPB blocks EMT of AECs by upregulating Nrf2 and inhibiting the TGF-β1-Smad2/3 signaling pathway. ARA is one of the most abundant polyunsaturated fatty acids in the body [39]. ARA is involved in a variety of biological processes, such as angiogenesis, cell migration, and apoptosis [40]. It has been found that inhibiting sEH could increase endogenous EETs content and reduce the EMT process [29,41]. 14,15-EET and its synthetic analog EET-A could decrease the expression of the EMT inducer factors, ZEB1 and Snail1, prevent the decrease of E-cadherin, and reduce the expression of mesenchymal/myofibroblast markers in the UUO model [29]. However, another ARA pathway, COX-2 metabolism, promotes EMT. COX-2 inhibitor-induced EMT reversal with restored E-cadherin expression has been observed in several cancer cells [42,43]. The COX-2 metabolite PGJ2 induces EMT by upregulating the expression of snails [44]. It can be seen that different metabolites of ARA play different roles in the process of EMT. We found that the protein expression of sEH and COX-2 increased significantly during the TGF-β1induced EMT process, manifested by the CYP/COX-2 metabolism disorder in ARA.
Studies have found a common phenomenon in the three metabolic pathways of ARA: inhibition of any one of these pathways may shunt ARA to the other pathway, thereby reducing efficacy and causing adverse reactions [45][46][47]. For example, NSAIDs may have anti-inflammatory effects by inhibiting COX, but their side effects may increase the risk of stroke and kidney failure [48]. At the same time, selective inhibition of COX-2 reduces the levels of endothelin PGI2 and the platelet aggregator TXA2, which increases the risk of cardiovascular disease [46]. Therefore, the development of bimolecular inhibitors targeting ARA metabolism has become increasingly important. It has long been found that drugs targeting a single molecule can produce other toxicity and drug resistance, while drugs targeting multiple molecules are less likely to develop resistance and have better therapeutic effects [49]. PTUPB is a novel COX-2 and sEH dual inhibitor [31], and we demonstrated that PTUPB could suppress PF [30], acute lung injury [33], nonalcoholic fatty liver disease [34], and sepsis [35]. However, the direct effects of PTUPB on TGF-β1-induced EMT in AECs are unknown. In the present study, PTUPB significantly improved E-cadherin expression, decreased α-SMA expression, and reduced excessive extracellular matrix deposition in BLM-treated mice. TIMPs serve an important role in controlling tissue organization and fibrosis following injury [50]. We found that PTUPB decreased the expression of Timp1 mRNA in BLM-treated PF mice lung tissue, which may be one of the reasons for decreased collagen synthesis. Further, in vitro EMT models of MLE-12 and A549 cells were induced by exogenous TGF-β1. We found that PTUPB attenuated TGF-β1-induced the acquisition of mesenchymal markers (such as α-SMA), prevented TGF-β1-induced the loss of epithelial markers (such as E-cadherin), decreased TGF-β1-induced the enhancement of migration ability, and reduced TGF-β1-induced the accumulation of collagen synthesis. These results suggest that regulating COX-2/CYP metabolism in AECs alleviates TGF-β1-induced EMT. Our results support the hypothesis that inhibition of COX-2/ sEH by PTUPB potently inhibits the progression of EMT. In short, our findings indicate that a COX-2 and sEH dual inhibitor show a pivotal therapeutic potential for EMT.
ROS plays an important role in the process of EMT, and TGF-β1-induced EMT can be inhibited by interfering with related upstream molecular events or by treating cells with antioxidants to block ROS production [51,52]. These studies indicate that ROS production is an important signal for EMT initiation. It has been found that restoring intracellular antioxidant signaling pathways can reduce TGF-β1-induced EMT. For example, piperine enhances the Nrf2 antioxidant cascade, reduces TGF-β1-induced ROS accumulation, and eliminates EMT in AML-12 hepatocytes [14]. Our data show that PTUPB restored Nrf2 protein expression and nuclear translocation in TGF-β1-stimulated MLE12 cells, while reducing TGF-β1-induced intracellular ROS levels. In addition, we unveiled that inhibition of Nrf2 abrogated the protective activity of PTUPB against TGF-β1. Thus, it is reasonable to speculate that targeted activation of Nrf2 is a pivotal contributor to the lung-protective activity of PTUPB.
TGF-β1-activated Smads play an important role in the process of EMT [53]. The combination of activated Smad2 or Smad3 and Smad4 can transcriptionally regulate EMT, while blocking the expression of Smad2 or Smad3 can reduce TGF-β1-induced EMT [54]. TGF-β1 activates TβRI by acting on the receptor complex and directly phosphorylates the C-terminal of Smad2 and Smad3. After phosphorylation, Smad2, Smad3, and Smad4 form trimers, which are transported to the nucleus, bind to DNA-binding transcription factors, and cooperatively regulate the transcription of target genes [53]. Our study found that PTUPB significantly reduced TGF-β1-induced phosphorylation of Smad2 and Smad3 in A549. Meanwhile, PTUPB also reduced the phosphorylation level of Smad3 induced by TGF-β1 in MLE12 and tended to decrease the phosphorylation level of Smad2 induced by TGF-β1 in MLE12. From the multiple of Smad2/3 phosphorylation change, we believe that PTUPB mainly inhibited the phosphorylation level of Smad3 in AECs. It was further found that PTUPB decreased the expression of ZEB1 mRNA and SNAIL1 mRNA downstream of the TGF-β1-Smad signaling pathway. These data indicate that PTUPB could inhibit activation of the TGF-β1-Smad2/ 3 pathway, therefore suppressing TGF-β1-induced EMT.
Moreover, we also unveiled that in MLE12 cells, inhibition of Nrf2 crippled the regulatory effects of PTUPB on TGF-β1/Smad signaling. This finding suggests that activation of Nrf2 is an important upstream event that explains PTUPBmediated modulation of intracellular TGF-β1/Smad pathways.

Data Availability
The datasets generated during and/or analyzed during the current study are available from the corresponding author on reasonable request. Figure S1. Effects of different concentrations of PTUPB on mRNA expression of TIMP1, MMP9, and CDH1 in A549 cells. Figure S2. Prophylactic treatments of PTUPB have no effect on the PI3K and MAPK signaling pathways of AECs. Figure S3. Prophylactic treatment of PTUPB does not affect total Smads protein in AECs. (Supplementary Materials)