Twenty-five Male Sprague-Dawley rats (initial age 8 weeks, initial weight 200–250 g, Animal Experiment Center, Southern Medical University, Guangzhou, China) were used in this study. Rats were housed 5 per cage under a standard 12 h/12 h light/dark cycle with food and water available ad libitum. Before any experimental procedures, rats were acclimated for 7 days. All experiments followed the recommendations of the International Association of Studies on Pain, and were approved by the Laboratory Animal Care and Use Committee of Nanfang hospital, Southern Medical University (NFYY-2014-86).

The rats were randomly divided into five groups (five rats per group): AOPPs-Rat serum albumin (AOPPs–RSA) group, PBS group, Rat serum albumin (RSA) group, apocynin group, and AOPPs–RSA+apocynin group. In our study, we injected AOPPs–RSA into healthy rats, according to plasma AOPPs concentration in the CFA model and our previous study (Ding et al., 2016). AOPPs–RSA (50 mg/kg), RSA (50 mg/kg), and PBS (pH 7.4) were intravenously injected each day of the study, Antioxidant apocynin (50 mg/kg, Sigma–Aldrich Corp, MO, United States) was dosed by intragastric administration each day of the study duration (Ding et al., 2016). The basic threshold of each rat was tested 3 days before drugs were injected. After treatment, rats were tested on Days 0, 3, 6, 9, ....., 54, and 57. The time interval was 2 days. DRG tissues were obtained after deeply anesthetized with pentobarbital sodium (150 mg/kg) on Day 57.

As previously described, AOPPs-Rat Serum Albumin (RSA) was prepared in vitro by incubating RSA (Sigma–Aldrich Corp, MO, United States) with hypochlorous acid (HOCl) as described in a previous study (Ding et al., 2016). The samples were dialyzed for 24 h against PBS at 4°C to remove the free HOCl. Control incubation was performed in native RSA dissolved in PBS alone. All the preparations were passed through a Detoxi-Gel column (Thermo Fisher Scientific, MA, United States) to remove any contaminated endotoxin. To quantify AOPPs concentration, SpectraMax M5 Multifunctional Microplate Reader (Molecular Devices, CA, United States) was used to measure the absorbance at 340 nm immediately.

This study was performed at the same time of day and by an experimenter unaware of the treatment applied. Each rat was placed in a clear plastic chamber on an elevated wire grid and acclimated for at least 15 min to the testing environment prior to the experiment. Withdrawal responses to mechanical stimuli were determined using an Electronic von Frey Anesthesiometer system (IITC Life Science, CA, United States). The electronic von Frey polypropylene tip was applied perpendicularly to the mid-plantar surface of the selected hind paw and the intensity of the stimulus was automatically recorded when the paw was flexed reflexively followed by a clear flinch response after paw withdrawal. All rats were tested 5 times, with an inter-test period of 15 min (Shibasaki et al., 2010).

Immunofluorescence staining of DRG slides was performed as previously described (Morell et al., 2014). Briefly, rats were anesthetized with isoflurane and pentobarbital sodium (150 mg/kg) at Day 57. The lumber DRGs (L4–6) were removed and postfixed in 4% paraformaldehyde overnight at 4°C and then were transferred into 30% sucrose (in PBS) at 4°C for at least 24 h. Samples were cryostat sectioned at a thickness of 14–16 μm and stored at –80 °C until use. Sections were permeabilized, blocked in 5% bovine serum albumin in PBS for 1 h at room temperature and then incubated overnight at 4°C with the following antibodies: rabbit anti-Nox4 (1:500) and chicken anti-beta III Tubulin antibody (1:1000) from Abcam (MA, United States). After rinsing in PBS, slides were incubated with secondary antibodies conjugated with FITC (1:500) or Cy3 (1:200) for 2 h at room temperature. Then, reacted slides were mounted and cover slipped. Images were captured with an Olympus FluoView FV10i self-contained confocal laser scanning microscope system (Olympus America Inc., PA, United States).

To determine whether cell apoptosis occurs in the primary afferent neurons during AOPPs-induced pain hypersensitivity rats, we used terminal deoxynucleotidyl transferase-mediated dUTP nick end-labeling (TUNEL) assay to evaluate programmed cell death in the dorsal root ganglia (location of the cell bodies of the primary afferent neurons) in each group of rats. Rats were deeply anaesthetized with sodium pentobarbital (150 mg/kg) and the lumber DRGs (L4–6) were removed and postfixed in 4 % paraformaldehyde overnight at 4°C. Cryosections were evaluated for TUNEL staining using the In situ Cell Death Detection Kit, Fluorescein (Roche) according to the manufacturer’s instructions (Ben-Zvi et al., 2008).

Dorsal root ganglion from thoracic and lumbar spinal cord of rats were minced in cold PBS and incubated for 60–90 min at 37°C in DMEM (4.5 g/L glucose, Gibco) containing (in mg/mL): 0.5 trypsin, 1 collagenase type IA, and 0.1 DNase type IV. Fetal bovine serum was added to neutralize trypsin. Neurons were pelleted, suspended in DMEM containing 10% fetal bovine serum, 100 U/mL penicillin, 0.1 mg/mL streptomycin, 2 mM glutamine. Cells were plated on polylysine-coated wells. Cytosine arabinoside (1 μM) was added to eliminate mitotic cells, which including Schwann cells and fibroblasts. Cultures were maintained at 37°C in a water-saturated atmosphere with 5% CO₂ for 3 d prior to start of the experiment (Amadesi et al., 2004).

Dorsal root ganglion neurons were cultured at 75% confluence and treated with AOPPs–RSA at increasing concentrations for 24 h or 100 μg/mL AOPPs–RSA for increasing time durations. After treatment, the cells were resuspended in 500 μL of binding buffer containing 5 μL Annexin V–fluorescein isothiocyanate (FITC) and 3 μL of propidium iodide (PI) according to the manufacturer’s instructions (BD Biosciences, CA, United States), and incubated for 10 min in the dark. Quantification of Annexin V–FITC and PI staining was performed by FACSCanto II (BD Biosciences) using channels FL-1 (Annexin V–FITC) and FL-3 (PI).

To introduce siRNA into rat DRG neurons, the cells were plated on 6-well plates at 90% confluence before transfection. Individual siRNAs (at 5 nM), lipofectamine 3000 and Opti-MEM were mixed and incubated at room temperature for 20 min. siRNA– lipofectamine 3000 complexes were added to cells for 24 h and the medium was replaced by fresh serum DMEM medium after transfection. Nox4 siRNA (190243) were purchased from Thermo Fisher Scientific (MA, United States). Experiments were performed 72 h after transfection. Knockdown of Nox4 was assessed by Western blot analysis which detected the protein level of Nox4 (Liu et al., 2016).

The level of intracellular ROS was assessed by fluorescence microplate reader and confocal laser scanning microscope system (Olympus America Inc., PA, United States) with the probe 2′,7′-dichlorofluorescein diacetate (DCFH-DA), which oxidizes to fluorescent dichlorofluorescein (DCF) in the presence of ROS, as described previously (Ding et al., 2016). Briefly, DRG neurons were suspended in DMEM at a given concentration of 10⁸/L. Each sample was incubated in 10 μM DCFH-DA (Sigma–Aldrich Corp, MO, United States) for 30 min in darkness. The excitation and emission wavelengths were 488 and 525 nm, respectively. The resulting data were normalized using the control values.

The change in MMP caused by AOPPs was assayed using a membrane-permeant JC-1 dye (5,50,6,60-tetrachloro-1,10,3,30-tetraethylbenzimidazolylcarbocyanine iodide; Keygen Biotech, Nanjing, China). The JC-1 dye accumulates in mitochondria in a potential-dependent manner. Consequently, mitochondrial depolarization is indicated by a decrease in the red/green fluorescence intensity ratio. The potential-sensitive color shift is due to concentration-dependent formation of red fluorescent J-aggregates (Sun et al., 2016). After 24 h of AOPPs–RSA treatment, 5 μg/mL of JC-1 dye was added to the cells with incubation at 37°C for 20 min, and the cells were washed with incubation buffer. The alteration in DRG neurons was observed using the Olympus FluoView FV10i self-contained confocal laser scanning microscope system.

Samples were homogenized in ice-cold RIPA buffer with 1 mM PMSF, protease, and phosphatase inhibitors (Sigma–Aldrich Corp., MO, United States) and cleared by centrifugation (12 000 rpm, 4°C, 10 min). Equivalent amounts of extracted proteins (50 μg) were separated by 10 or 12% SDS–polyacrylamide gel electrophoresis, then electro-blotted onto polyvinylidene difluoride membranes (Millipore, MA, United States). The membranes were blocked using 5% blocking buffer for 1 h at room temperature. Blots were incubated overnight at 4°C with the following primary antibodies: rabbit anti-Nox4 (1:1000) from Abcam (MA, United States), rabbit anti-(p-)MAPK family (1:1000), rabbit anti-Bax (1:1000), rabbit anti-Bcl2 (1:1000), rabbit anti-Cytochrome c (1:1000), mouse anti-caspase 9 (1:1000), rabbit anti-cleaved-caspase 9 (1:1000), rabbit anti-caspase 3 (1:1000), rabbit anti-cleaved-caspase 3 (1:1000), rabbit anti-PARP (1:1000), and rabbit anti-cleaved-PARP (1:1000) from Cell Signaling Technology (CST, MA, United States). The membrane was washed 3 times with TBST for 10 min and followed by incubation with appropriate secondary antibody for 2 h, and then washed again three times with TBST. Relative levels of immunoreactivity were quantified using the Kodak In vivo Imaging System (Kodak, NY, United States). Rabbit anti-β-actin (1:2000, Abcam) was used as an internal control for the concentration of proteins loaded.

Analyses were performed with SPSS 20.0 (IBM, NY, United States) and GraphPad Prism 5 (GraphPad Software, CA, United States) software. All experiments were repeated at least three times. Continuous variables were expressed as mean ± standard deviation (SD). Comparisons were analyzed with student t-test (2 groups only) or ANOVA followed by the Bonferroni post hoc tests analysis. A value of P < 0.05 was considered statistically significant.

To explore whether AOPPs could induce mechanical hypersensitivity in rats, the electronic von Frey was used to measure the mechanical threshold of rats. As shown in Figure 1A, after giving different treatments, the paw mechanical threshold of each rat was tested every 3 days. After intravenous injection of AOPPs–RSA, the mechanical threshold significantly declined from Day 3 (23.91 ± 0.51 g, P < 0.05) to Day 30 (5.22 ± 0.33 g, P < 0.05) and remained a low level until Day 57 (5.22 ± 0.13 g, P < 0.05). In addition, we found that intragastric administration of apocynin could notably improve mechanical hypersensitivity induced by AOPPs. The weight of rats in each group was recorded in Figure 1B, and there was no significant difference between each group (the range was from 225.00 ± 4.21 g to 341.80 ± 8.46 g, P > 0.05).

Immunoreactivity of anti-Nox4 was observed in sections of rat DRG (L4-L6) neurons that were co-stained with the neuronal marker, β-tubulin III. Figure 1C revealed that Nox4 was markedly increased in AOPPs–RSA group compared with that in PBS group, and apocynin partly inhibited AOPPs-induced Nox4 expression. To confirm our results, we used western blot assays to study the expression level of Nox4 in DRG tissues (L4-L6). Figure 1D indicated that the expression of Nox4 increased ∼2-fold in AOPPs–RSA group compared with PBS group (P < 0.05), and apocynin could reduce the expression level by almost 75% (P < 0.05).

We performed TUNEL assay in order to evaluate the occurrence of apoptosis in DRG neurons of AOPPs-induced rats. The results from this study showed the evidence of programmed cell death or of necrotic cell death in the DRG of AOPPs-induced rats (Figure 1E). These studies indicate that cell apoptosis was occurred in the DRG neurons. Taken collectively, our data indicate that AOPPs could activate Nox4 expression, induce apoptosis of DRG neurons, finally lead to causing pain hypersensitivity in rats.

The MTT assay (Figure 2A) showed that incubation in 0–200 μg/mL AOPPs–RSA had minimal effects on cell fatality (cell viability was above 90%, P > 0.05). To eliminate the fatal effects of AOPPs, DRG neurons were subjected to increasing concentrations of AOPPs–RSA (50, 100, and 200 μg/mL) for 24 h. Quantitative fluorescence-activated cell sorting analysis of FITC–annexin V/PI staining indicated that AOPPs–RSA induced DRG neurons apoptosis (Figure 2B, P < 0.05). Figure 2C showed that 100 μg/mL AOPPs–RSA for 12 and 24 h could significantly induce apoptosis of DRG neurons (P < 0.05).

The caspase family is a key section in cell apoptosis, and activation of caspase cascade and PARP-1 results in substantial damage in molecule level. To demonstrate whether caspase cascade and PARP-1 participate in AOPPs-induced DRG neurons apoptosis, their expression was detected by western blot analysis. The expression of pro-caspase9 decreased in DRG neurons under the condition of 50, 100, or 200 μg/mL AOPPs–RSA in 24 h, while, cleaved-caspase9 increased significantly (Figure 3A, P < 0.05). Next, we used 100 μg/mL AOPPs–RSA to treat cells in indicated periods, the results indicated that the levels of pro-caspase9 was significantly down-regulated at 12 and 24 h, whereas, cleaved-caspase9 was up-regulated (Figure 3B, P < 0.05). As for cleaved-caspase3, the expression level was similar to cleaved-caspase9. However, the level of pro-caspase3 had no significant changes in indicated concentrations or periods (Figures 3C,D, P > 0.05). PARP-1 and cleaved-PARP-1 increased in AOPPs-treated cells (Figure 3E, P < 0.05), and the level of protein increased from 2 h after AOPPs–RSA treatment and reached the highest level at 24 h (6.57-fold and 5.76-fold compared with t = 0, P < 0.05, Figure 3F).

Loss of mitochondrial membrane potential (MMP) was considered as a form of mitochondrial dysfunction. MMP loss was assessed through JC-1 staining. In healthy cells with a high MMP, JC-1 accumulates in the mitochondria as J-aggregates with red fluorescence. As the MMP declines in apoptotic cells, JC-1 stays in the cytosol as monomers with green fluorescence. The results suggested a dose-dependent reduction of red fluorescence and concomitantly enhanced green fluorescence, which indicated the existence of MMP loss in AOPPs-treated DRG neurons (Figure 4A).

The intrinsic apoptosis pathway is regulated primarily by Bcl-2 family proteins, notably the proapoptotic Bax and antiapoptotic Bcl-2 proteins. Figure 4B showed that the expression of Bax was activated in a concentration-dependent manner, while Bcl-2 was down-regulated (Figure 4D), and the highest Bax/Bcl-2 ratio was 2.93 ± 0.15 at the concentration of 200 μg/mL AOPPs–RSA (Figure 4F, P < 0.05). As shown in Figures 4C,E, the expression of proapoptotic Bax significantly increased from 12 h, with a significantly decrease in antiapoptotic Bcl-2 at the same time, and the highest Bax/Bcl-2 ratio was 1.45 ± 0.12 at 24 h (Figure 4G, P < 0.05).

Cytochrome c increased in cytoplasm when mitochondrial membrane potential decreased, and it can induce apoptosis of cells. As shown in Figures 4H,I, cytochrome c increased significantly after AOPPs–RSA treated DRG neurons in indicated concentrations or periods.

To determine whether AOPPs, as oxidative stress products, can cause abundant ROS accumulation inside DRG neurons, we examined intracellular ROS levels in AOPPs-treated DRG neurons. We subjected DRG neurons cultures to increasing concentrations of AOPPs–RSA or for different time durations. ROS production was increased in DRG neurons cultured with AOPPs–RSA in a dose- and time-dependent manner (Figures 5A,B). To elucidate the function of Nox4 in ROS generation, DRG neurons were, respectively, pre-treated with Nox4 siRNA (Supplementary Figure S1), Nox inhibitor DPI (10 μM, 2 h) and ROS scavenger NAC (2 mM, 2 h). After these treatments, DRG neurons were incubated with 200 μg/mL AOPPs–RSA for 360 min. The level of ROS generation was significantly decreased in DRG neurons that were pretreated with Nox4 siRNA, DPI or NAC separately (Figure 5C). For further confirmation, ROS generation in DRG neurons cultured with control medium, RSA, AOPPs–RSA, AOPPs+siNox4, AOPPs+DPI, AOPPs+NAC was observed under a confocal laser scanning microscope system using DCFH-DA (Figure 5D).

To investigate the effect of AOPPs–RSA on Nox4 expression in rat DRG neurons, the cells were stimulated with different concentrations of AOPPs–RSA (0, 50, 100, and 200 μg/mL) or for different time durations (0, 2, 6, 12, and 24 h). The protein level of Nox4 was analyzed by western blot from AOPPs-stimulated rat DRG neurons. The Nox4 protein level gradually increased after AOPPs stimulation for 100–200 μg/mL (Figure 5E). AOPPs–RSA stimulation also time-dependently induced Nox4 protein expression (Figure 5F).

Previous studies had demonstrated that ROS could activate MAPK family, thereby, lead to cell apoptosis. In this study, c-Jun N-terminal kinase (JNK), and p-JNK consecutively increased from 2 to 24 h (Figure 5G). Their activation was also mediated by AOPPs in a dose-dependent manner (Figure 5H). However, p38/p-p38 and p44/p42 had no significant changes in the process of AOPPs-induced DRG neurons apoptosis (Supplementary Figure S2).

To further elucidate the roles of Nox4, ROS, JNK, caspase cascade, and PARP in AOPPs-induced apoptosis, DRG neurons were incubated with siRNA Nox4, DPI, NAC, a JNK special inhibitor (SP600125), and the broad-spectrum caspase inhibitor Z-VAD-FMK before AOPPs–RSA treatment. DRG neurons apoptosis was found to be significantly abolished under the protection of these inhibitors (Figure 6A) and PARP-1 activation was suppressed by these inhibitors (Figure 6B).