**1. Introduction**

Diabetic retinopathy (DR) is a very important microvascular complication of diabetes [1]. It is characterized by a progressive increase in vascular permeability, retinal ischemia and edema, and neovascularization, which results in visual impairment and legal blindness [2]. Retinal vascular endothelial cells play an important role in maintaining the blood-retinal barrier (BRB), which provides a physiological border for retinal homeostasis [3]. Previous studies demonstrated that hyperglycemia induces the activation of oxidative stress and generates reactive oxygen species (ROS) within retinal vascular endothelial cells [4,5]. The accumulation of ROS alters the homeostasis and enhances the migration of retinal vascular endothelial cells, triggers cellular apoptosis, and increases vascular

permeability and basement membrane leakage in the retina. These pathological changes may lead to the breakdown of the BRB and DR development [6,7]. Therefore, suppressing oxidative stress could inhibit apoptosis in retinal vascular endothelial cells and reduce the risk of DR progression.

Fenofibrate, a peroxisome proliferator-activated receptor type α (PPARα) agonist [8], is used clinically to treat hypertriglyceridemia and hyperlipidemia. However, evidence from two large randomized clinical trials, the Fenofibrate Intervention and Event Lowering in Diabetes (FIELD) and Action to Control Cardiovascular Risk in Diabetes (ACCORD), has shown that fenofibrate significantly prevents DR progression and reduces the use of laser treatment in DR [9,10]. In experimental diabetic models, the expression of PPAR α is significantly downregulated in retina [11]. In addition, high glucose medium downregulates PPAR α expression in retinal cells [11]. Moreover, over-expression of inflammatory factors, retinal vascular leakage, and more severe DR are found in diabetic PPAR α knockout mice [11]. Fenofibrate exerts anti-inflammatory and anti-oxidative e ffects. Treatment of retinal pigment epithelial cells with fenofibrate reduces high-glucose-induced ROS generation [12]. Fenofibrate downregulated NF-κB, significantly inhibited the expressions of inflammatory mediators and reduced the concentrations of oxidative products in a diabetic rat model [13]. Fenofibrate can significantly reduce lipopolysaccharide (LPS)-induced ROS and increase endothelial nitric oxide (eNOS) levels in human umbilical vein endothelial cells (HUVECs) [14]. Fenofibrate also decreases apoptosis in human retinal endothelial cells and pericytes by activating the AMP-activated protein kinase (AMPK) pathway and downregulating the NF-κB pathway, respectively [15,16]. PPARα over-expression or fenofibrate treatment has also been shown to attenuate retinal vascular permeability in diabetic animals [11,17] and may protect against BRB leakage through the down-regulation of basement membrane components [18]. Despite the persuasive results from these clinical and experimental studies, the mechanisms through which fenofibrate protects the eye from DR remain elusive, and further studies are still required to clarify the protective mechanisms of fenofibrate.

Thioredoxins (Trxs) are small proteins that are essential for embryonic development and could protect cells against oxidative stress [19]. There are two main forms of Trx. Trx-1 exists in cytosol, acts as a cofactor of peroxiredoxins (Prx), and plays a direct role in reducing oxidative stress [20,21]. Trx-2 exists in mitochondria and plays an important role in the mitochondrial cellular apoptosis pathway [22,23]. Trxs are essential for life, and Trx gene deficiency is embryonic lethal [24,25]. Trxs are involved in multiple redox-regulated signaling pathways. Trxs bind to apoptosis signal-regulated kinase-1 (Ask-1) in the cytosol and mitochondria, thereby blocking the initiation of the cellular apoptotic process and inhibiting c-Jun amino-terminal kinase (JNK/p38 mitogen-activated protein kinase [MAPK]) [26]. Furthermore, previous studies have shown that Trx is a PPARα target gene and that PPARα activation induces the translocation of Trx to the nucleus and modulates Trx expression [27]. PPARα activator significantly enhances the activation of the Trx promoter and increases Trx-1 expression in human macrophages [28]. However, the role of Trxs against oxidative stress in retinal vascular endothelial cells, as well as the mechanism through which fenofibrate modulates Trxs expression in DR have not been reported.

In this study, we hypothesized that fenofibrate can counteract oxidative stress and attenuate oxidative stress-induced cell apoptosis and death by modulating Trx expression in retinal vascular endothelial cells. Paraquat (PQ) is a common stimulator to induce oxidative stress in in vitro and in vivo studies about retinal degeneration [29–31]. Oxidative stress induced by PQ is thought to play an important role in type 2 diabetes through the impairment of insulin action [32–34]. Therefore, we used PQ as the inducer of oxidative stress to simulate the condition in DR. We used retinal/choroidal vascular endothelial cell (RF/6A) cells as the cell model system. RF/6A cell line is a monkey choroidal–retinal vascular endothelial cell line and has been widely used to study retinal vascular diseases and DR previously [35–44]. The study was performed in two parts. First, we investigated the role of oxidative stress in initiating apoptosis and evaluated the protective e ffects of fenofibrate. Second, we investigated the modulatory e ffect of fenofibrate on Trx expression and analyzed the related apoptosis and stress signaling pathways.

#### **2. Materials and Methods**

#### *2.1. Cell Culture and Fenofibrate Pretreatment*

The RF/6A cell line is a monkey choroidal–retinal vascular endothelial cell line. RF/6A cells were purchased from the American Type Culture Collection (Rockville, MD, USA). RF/6A cells were maintained in Dulbecco's modified Eagle's medium (DMEM) with 10% fetal bovine serum, 4.5 mg/mL glucose, 100 units/mL penicillin, and 100 μg/mL streptomycin (all from Thermo Fisher Scientific, Waltham, MA, USA) in a 5% CO2 atmosphere at 37 ◦C. The cells were pretreated with different concentrations of fenofibrate (CAS Number 49562-28-9, Sigma-Aldrich, St. Louis, MO, USA) before exposure to PQ (Sigma-Aldrich, St. Louis, MO, USA).

#### *2.2. Cell Viability Assay*

The RF/6A cells were seeded at a density of 1 × 10<sup>4</sup> cells per well onto 96-well plates and incubated at 37 ◦C. The cells were exposed to 0, 0.2, 0.4, 0.6, 0.8 and 1.0 mM PQ for 24 h. The cells in the fenofibrate treated group were pretreated with 25, 50, 75 or 100 μM fenofibrate for 1 h prior to 24-h exposure of 1.0 mM PQ. 5 mg/mL 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT, Chemicon, Millipore, Burlington, MA, USA) was added to each well for 4 h. Then we removed the culture medium supernatant, and formazan was dissolved with dimethyl sulfoxide (DMSO, Sigma-Aldrich, St. Louis, MO, USA) for 30 min at room temperature. The absorbance (570 nm) was measured with a microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

#### *2.3. Analysis of Apoptosis by Flow Cytometry*

The RF/6A cells were pretreated with 25, 50, 75 or 100 μM fenofibrate for 1 h prior to 1.0 mM PQ exposure. The proportion of apoptotic RF/6A cells was determined at 24 h by flow cytometry using a staining solution containing 5 μL of annexin-V-FITC and 5 μL of propidium iodide (PI) (Strong Biotech, Taipei, Taiwan) in 250 μL of binding buffer. Cells were washed with PBS and centrifuged at 200 g for 5 min. Then we resuspended the cell pellet in 100 μL of staining solution and incubated for 10 min at 20 ◦C. Finally, 900 μL of binding buffer was added to the samples, and the samples were analyzed on a FACScan cytometer (BD Bioscience, Franklin Lakes, NJ, USA).

#### *2.4. Detection of Intracellular ROS*

We measured intracellular ROS levels by 2-,7--dichlorodihydrofluorescein diacetate (2-,7--DCFDA, Sigma-Aldrich, St. Louis, MO, USA) oxidation. The RF/6A cells were pretreated with 25, 50, 75 or 100 μM fenofibrate for 1 h prior to 24-h exposure of 1.0 mM PQ. RF/6A cells were then exposed to 10 μM 2-,7--DCFDA for 10 min. The cells were analyzed by FACScan cytometer (BD Biosciences, Franklin Lakes, NJ, USA) using the FL-1 channel (515–545 nm).

#### *2.5. Quantitative Detection of ROS-Induced Cellular Oxidation*

The RF/6A cells were pretreated with 25, 50, 75 or 100 μM fenofibrate for 1 h prior to 1.0 mM PQ treatment. After 24-h PQ exposure, DNA oxidation, lipid peroxidation, and protein oxidation levels were determined using an 8-hydroxydeoxyguanosine (8-OHdG) Check Kit (JaICA, Shizuoka, Japan), a thiobarbituric acid reactive substances (TBARS) Assay Kit (Cayman Chemical, Ann Arbor, MN, USA), and a Protein Carbonyl Colorimetric Assay Kit (Cayman Chemical, Ann Arbor, MN, USA), respectively. Cellular DNA was extracted for 8-OHdG detection using a cellular genomic DNA Extraction Kit (T-Pro Biotechenology, New Taipei County, Taiwan). Cellular homogenates were prepared for TBARS or carbonyl colorimetric assays according to the manufacturer's instructions.

#### *2.6. Determination of Mitochondrial Dysfunction*

To detect the extent of mitochondrial dysfunction, we measured the mitochondrial membrane potential of cells with JC-1 stain (Cayman Chemical, Ann Arbor, MN, USA). The RF/6A cells were seeded at a density of 1 × 10<sup>4</sup> cells per well onto 96-well plates and incubated at 37 ◦C. We added different concentrations of fenofibrate (25, 50, 75, 100 μM) to the cells exposed to 1.0 mM PQ. After a 24-h incubation, 50 μL of JC-1 staining solution buffer was added to 1 mL of culture medium, and the plate was incubated at 37 ◦C for 15 min. The fluorescence signals for J-aggregates with Texas Red (healthy cells, excitation/emission = 560/595 nm) and JC-1 monomers with FITC (apoptotic or unhealthy cells, excitation/emission = 485/535 nm) were measured with a microplate reader (Bio-Rad Laboratories, Hercules, CA, USA).

#### *2.7. Preparation of RNA and cDNA*

The RF/6A cells were incubated with 10 μM GW6471 (a PPAR-α antagonist, R&D systems, Minneapolis, MN, USA) for 1 h. After removing GW6471, the cells were then pretreated with 50 or 100 μM fenofibrate for 1 h prior to 1.0 mM PQ treatment. After 24-h PQ exposure, we extracted RNA from RF/6A cells with TRIzol reagen<sup>t</sup> (Thermo Fisher Scientific, Waltham, MA, USA). 1 μg of total RNA was incubated with 300 ng of Oligo dT (Promega, Madison, WI, USA) for 5 min at 65 ◦C. Samples were then reverse transcribed into cDNA using Moloney murine leukemia virus reverse transcriptase (MMLV-RT; Thermo Fisher Scientific, Waltham, MA, USA) for 1 h at 37 ◦C. The reaction was terminated by heating the samples for 5 min at 90 ◦C.

#### *2.8. Analysis of mRNA Expression Levels*

The resultant cDNA product was subjected to PCR using Prx, Trx-1, Trx-2, B-cell lymphoma 2 (Bcl-2), Bcl-xl, B-cell lymphoma 2-associated X protein (Bax), and β-actin primers. The amplification was performed by thermocycler (MJ Research, Waltham, MA, USA). The 25 μL reaction mixture was composed of 5 μL of cDNA, 200 μM of each deoxynucleotide (DTT), 1 μL of sense and antisense primers, 1.25 U of GoTaq polymerase (Promega, Madison, WI, USA), and 5 μL of 10× Taq polymerase buffer. PCR was performed at an annealing temperature of 56 ◦C with GoTaq polymerase, cDNA, and the following primers: Prx: 5--CTTCAGGAAATGCAAAAATTGGGCAT-3- (forward), 5--GAGTTTCTTAAATTC TTCTGCTCTA-3- (reverse); Trx-1: 5--CCCTTCTTTCA TTCCCTCTGTG-3- (forward), 5--GAACTCCCCAACCTTTTGACC-3- (reverse); Trx-2: 5--CGTACAAT GCTGGTGGTCTAAC-3- (forward), 5--GTCTTGAAAGTCAGGTCCATCC-3- (reverse); Bcl-2: 5--CTGGTGGACAACATCGCTCTG-3- (forward), 5--GGTCTGCTGACCTCACTTGTG-3- (reverse); Bcl-xl: 5--CCCCAGAAGAAACTGAACCA-3- (forward), 5--AGTTTACCCCAT CCCGAAAG-3- (reverse); Bax: 5--TGGTTGCCCTTTTCTACTTTG-3- (forward), 5--GAAGTAGGAAAGGAGGCCA TC-3- (reverse); β-actin: 5-- CTGGAGAAGAGCTATGAGCTG-3- (forward), 5-- AATCTCCTTCTGCAT CCTGTC-3- (reverse). The DNA fragments were amplified for 25–30 cycles (30 s at 94 ◦C; 1 min at 50–52 ◦C; and 1 min at 72 ◦C), followed by a 7 min extension step at 72 ◦C. The products were then subjected to electrophoresis on a 1.5% agarose gel and analyzed by gel analyzer system. β-actin was used as the internal control.

#### *2.9. Protein Extractions and Western Blot Analysis*

The RF/6A cells were incubated with 10 μM GW6471 for 1 h. After removing GW6471, the cells were then pretreated with 50 or 100 μM fenofibrate for 1 h prior to 1.0 mM PQ exposure. After 24-h or 1-h (for phospho-Ask1 and phospho-JNK) PQ exposure, we extracted proteins from RF/6A cells with radioimmunoprecipitation assay (RIPA) lysis buffer, which contained 0.5 M Tris-HCl (pH 7.4), 2.5% deoxycholic acid, 10% NP-40, 1.5 M NaCl, 10 mM EDTA, and 10% protease inhibitors (Complete Mini; Roche Diagnostics, Indianapolis, IN, USA). Mitochondrial proteins and cytosolic proteins were isolated using a mitochondria isolation kit (Thermo Fisher Scientific, Waltham, MA, USA), following

the protocol description. For the western blot analysis, the protein samples were separated by a 10% sodium dodecyl sulfate (SDS)-polyacrylamide gel and then transferred to a polyvinylidene difluoride (PVDF) membrane (Immobilon-P; Millipore, Burlington, MA, USA). The primary antibodies used in the experiment were as follows: anti-PPARα (at a 1:500 dilution, Santa Cruz Biotechnology, Dallas, TX, USA); anti-Prx-1 (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-Trx-1 (at a 1:500 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-Ask-1 (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-phospho-Ask1 (at a 1:2000 dilution, Bioss, Woburn, MA, USA); anti-JNK (at a 1:3000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-phospho-JNK (at a 1:3000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-Bcl-2 (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-Bcl-xl (at a 1:500 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-Bax (at a 1:5000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-cytochrome c (at a 1:1000 dilution, Abcam, Hong Kong, China); anti-VDAC1 (at a 1:5000 dilution, Abcam, Hong Kong, China); anti-Trx-2 (at a 1:2000 dilution, R&D System, Minneapolis, MN, USA,); anti-apoptotic protease activating factor-1 (Apaf-1) (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-caspase-9 (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti-caspase-7 (at a 1:1000 dilution, Cell Signaling Technology, Danvers, MA, USA); anti- poly (ADP-ribose) polymerase-1 (PARP-1) (at a 1:1000 dilution, Abcam, Hong Kong, China); and anti-β-actin (at a 1:5000 dilution, Bioss, Woburn, MA, USA). Immunodetections were performed using enhanced chemiluminescence (Pierce Biotechnology, Waltham, MA, USA). Protein levels were determined using densitometry analysis of the protein bands. Protein levels were normalized to β-actin.

#### *2.10. Caspase-3 Activity Assay*

The RF/6A cells were incubated with 10 μM GW6471 for 1 h. After removing GW6471, the cells were then pretreated with 50 or 100 μM fenofibrate for 1 h prior to 1.0 mM PQ exposure. After 24 h PQ exposure, the caspase-3 activity of RF/6A cells was analyzed by the Caspase-3/CPP32 Colorimetric Assay Kit (BioVision, Milpitas, CA, USA). Assay procedures were performed following the manufacturer's instructions.
