Mechanistic study of mtROS-JNK-SOD2 signaling in bupivacaine-induced neuron oxidative stress

Manganese superoxide dismutase (SOD2) is a key enzyme to scavenge free radical superoxide in the mitochondrion. SOD2 deficiency leads to oxidative injury in cells. Bupivacaine, a local anesthetic commonly used in clinic, could induce neurotoxic injury via oxidative stress. The role and the mechanism of SOD2 regulation in bupivacaine-induced oxidative stress remains unclear. Here, bupivacaine was used to treat Sprague-Dawley rats with intrathecal injection and culture human neuroblastoma cells for developing vivo injury model and vitro injury model. The results showed that bupivacaine caused the over-production of mitochondrial reactive oxygen species (mtROS), the activation of C-Jun N-terminal kinase (JNK), and the elevation of SOD2 transcription. Decrease of mtROS with N-acetyl-L-cysteine attenuated the activation of JNK and the increase of SOD2 transcription. Inhibition of JNK signaling with a small interfering RNA (siRNA) or with sp600125 down-regulated the increase of SOD2 transcription. SOD2 gene knock-down exacerbated bupivacaine-induced mtROS generation and neurotoxic injury but had no effect on JNK phosphorylation. Mito-TEMPO (a mitochondria-targeted antioxidant) could protect neuron against bupivacaine-induced toxic injury. Collectively, our results confirm that mtROS stimulates the transcription of SOD2 via activating JNK signaling in bupivacaine-induced oxidative stress. Enhancing antioxidant ability of SOD2 might be crucial in combating bupivacaine-induced neurotoxic injury.


INTRODUCTION
Reactive oxygen species (ROS), as a byproduct of oxidative phosphorylation, are mainly produced in the mitochondria [1,2]. Enough evidence has revealed that ability deficiency of scavenging ROS leads to the damage of mitochondrial lipid membrane, the release of mitochondrial cytochrome c and the activation of mitochondrial death pathway [3,4]. With reports of cauda equine syndrome and transient neurological symptoms following continuous spinal anesthesia or high concentration of local anesthetic application, more and more clinicians pay attention to the local anesthetic-induced neurotoxic injury [5,6]. Bupivacaine (BPV), an amide compound, is widely administrated for regional nerve block and analgesia [7]. It uncouples oxidative phosphorylation, inhibits ATP production, and collapses the mitochondrial membrane potential [8]. The decrease of ATP activates adenosine 5´-monophosphate (AMP)-activated protein kinase signaling which results in a marked increase of intracellular ROS. We have demonstrated that oxidative stress-mediated apoptosis is a crucial mechanism of BPV-induced neuron injury [9]. AGING Manganese superoxide dismutase (SOD2) is the essential mitochondrial antioxidant enzyme that detoxifies the free radical superoxide in mammalian cells [10,11]. It transfers highly reactive O 2 − into H 2 O 2 , which is reduced to H 2 O in the mitochondria [12]. The homeostasis balance in the activation of SOD2 and the production of free radical superoxide determine whether cells suffer from oxidative stress and apoptosis. SOD2 deficiency leads to mitochondrial oxidative stress and glycation of mitochondrial DNA, which plays a crucial role in mitochondria oxidative stress and neuron apoptosis [13,14]. The interaction between SOD2 transcription and ROS production is different in some pathological processes. SOD2 deficiency reportedly exacerbates the mitochondrial ROS (mtROS) overproduction and oxidative damage in Chagas disease [15]. At the same time, previous evidence demonstrates that ROS stimulates SOD2 expression through activation of p53 [16]. The mechanism is that ROS drives the nucleus translocation of extracellular regulated protein kinases, where it phosphorylated p53 at Ser15, leading to the activation of p53 and subsequent up-regulation of SOD2 transcription [17]. However, the role and transcription of SOD2 in BPVinduced oxidative stress remains unclear.
C-Jun N-terminal kinase (JNK) is a stress-inducible kinase in response to various extracellular and intracellular nociceptive stimulus, such as oxidative damage, UV light, chemicals and biological agents [18][19][20]. When phosphorylated, JNK signaling is activated to change stress related proteins transcription for modulating cell survival or death [21,22]. Whether or not SOD2 transcription is regulated through mtROSdriven activation of JNK signaling in oxidative stress has not been reported.
In this study, BPV was used to treat Sprague-Dawley rats with intrathecal injection and culture human neuroblastoma (SH-SY5Y) cells for developing vivo injury model and vitro injury model. This study may elucidate the mechanism of mtROS-JNK-SOD2 signaling in BPV-induced neuron oxidative stress and provide promising therapy for above neurotoxic injury.

BPV induced spinal reflex dysfunction and apoptotic injury in vivo
Spinal reflex function was assessed by paw withdrawal threshold (PWT, g) and thermal withdrawal latency (TWL, s). They were tested in different times (pre-drug, 6 th h, 12 th h and 24 th h) after rats with intrathecal injection of 2.5% BPV. In group BPV, PWT and TWL values were significantly elevated in 6 th h, 12 th h and 24 th h after injection (vs. pre-drug, P< 0.05). PWT and TWL values were also significantly elevated in 6 th h, 12 th h and 24 th h after intrathecal injection of BPV (group BPV vs. group Con, P< 0.05). (Figure 1A, 1B).
Spinal cord apoptotic injury was determined with cleaved caspase-3 expression. They were measured in different times (pre-drug, 6 th h, 12 th h and 24 th h) after rats with intrathecal injection of 2.5% BPV. Cleaved caspase-3 expression was significantly elevated in 6 th h, 12 th h and 24 th h after intrathecal injection of BPV (vs. group Con, P< 0.05, Figure 1C, 1D).

BPV caused oxidative injury, activated JNK signaling and elevated SOD2 transcription in vivo
After intrathecal injection of 2.5% BPV, JNK phosphorylation and SOD2 expression were elevated in 6 th h and 12 th h in spinal cord of rats (vs. pre-drug, P< 0.05, Figure 1C-1F). Spinal cord oxidative injury was measured with malondialdehyde (MDA) and 8hydroxydeoxyguanosine (8-OHdG) generation. MDA and 8-OHdG production significantly were increased in 6 th h, 12 th h and 24 th h after intrathecal injection of 2.5% BPV in spinal cord of rats (vs. group Con, P< 0.05, Figure 1G, 1H).

SOD2 transcription was up-regulated via mtROS-JNK signaling in BPV-induced vitro oxidative injury model
N-acetyl-L-cysteine (NAC, a ROS scavenger) was employed to determine the effect of mtROS on JNK-SOD2 signaling. The results showed that NAC significantly reduced mtROS production, the increase of JNK phosphorylation and SOD2 transcription in group AGING AGING NAC (vs. group Con, P< 0.05). It also significantly reduced BPV-induced mtROS generation, the increase of JNK phosphorylation and SOD2 transcription (group NAC+BPV vs. group BPV, P< 0.05); At the same time, NAC significantly attenuated BPV-induced neurotoxic injury (group NAC+BPV vs. group BPV, P< 0.05). (Figure 3) Next, small interfering RNA (siRNA) and sp600125 (an inhibitor of JNK signaling) were employed to confirm the mechanic of JNK signaling in SOD2 transcription. Down-regulation of JNK expression significantly decreased SOD2 transcription in cells (group siJNK vs. group Con, P< 0.05). Simultaneously, the same effect can be achieved at inhibiting activation of JNK signaling (group SP vs. group Con, P< 0.05). More importantly, SOD2 transcription was significantly inhibited in group siJNK+BPV or group SP+BPV (vs. group BPV, P< 0.05). However, JNK gene knock-down had no effect on BPV-induced mtROS production (group siJNK+BPV vs. group BPV, P> 0.05). (Figure 4)

SOD2 gene knock-down enhanced BPV-induced mtROS over-production, oxidative injury and apoptosis in vitro
Cells were transfected with SOD2 siRNA or negative control siRNA. SOD2 gene knock-down increased mtROS generation in cells (group siSOD2 vs. group Con, P< 0.05). Meanwhile, it also enhanced BPVinduced mtROS over-production (group siSOD2+BPV vs. group BPV, P< 0.05). (Figure 5A-5D) Further, the effect of SOD2 gene knock-down on JNK phosphorylation was determined. The results showed that SOD2 gene knock-down elevated JNK phosphorylation in group siSOD2 (vs. group Con, P< 0.05). However, it had no effect on JNK phosphorylation in group siSOD2+BPV (vs. group BPV, P> 0.05). ( Figure 5E and 5F) Mitochondrial depolarization was measured with JC-1 staining. Mitochondrial membrane potentials (MMP) was declined in group siSOD2 or group BPV (vs. group Con, P< 0.05). SOD2 gene knock-down significantly aggravated BPV-induced the decrease of MMP (group siSOD2+BPV vs. group BPV, P< 0.05). MDA and 8-OHdG production were increased in group siSOD2 or group BPV (vs. group Con, P< 0.05). This effect was further exacerbated in group siSOD2+BPV (vs. group BPV, P< 0.05). The cells apoptosis was detected by flow cytometry and cleaved caspase-9 expression. Cells apoptosis was elevated in group siSOD2 or group BPV (vs. group Con, P< 0.05). SOD2 gene knock-down significantly enhanced BPV-induced cells apoptosis (group siSOD2+BPV vs. group BPV, P< 0.05). As a control antioxidant in mitochondria, mito-TEMPO was used to protect cells against oxidative injury and 25 µM mito-TEMPO performed significant oxidation resistance in SH-SY5Y cells as previously described [23]. So, this concentration of mito-TEMPO was used to preculture cells and it could attenuate BPV-induced neurotoxic injury (group mito-TEMPO+BPV vs. group BPV, P< 0.05). (Figure 6)

DISCUSSION
The purpose of this study is to characterize SOD2 in BPV-induced neuron oxidative stress and find out its mechanistic pathway. There are three main findings in the present study. First, BPV caused neuron oxidative stress with concomitantly stimulating activation of JNK signaling and SOD2 transcription. Second, mtROS-JNK signaling was a co-regulator of SOD2 transcription in BPV-induced neurotoxic injury. Third, SOD2 deficiency enhanced BPV-induced neurotoxicity and mito-TEMPO protected cells against above injury.
Local anesthetics are widely used in regional anesthesia and analgesia, and have certain neurotoxic effects on neuron. Previous studies have confirmed that local anesthetics can trigger intracellular Ca 2+ homeostasis imbalance, mitochondrial and endoplasmic reticulum oxidative stress. A large amount of ROS production is a crucial factor in local anesthetics-induced toxic injury such as BPV, lidocaine and ropivacaine [9,24,25]. However, previous studies focus on cell oxidative injury and whether the results of vivo model are consistent with above changes is not clear. In this study, vivo injury model was developed and the results showed BPV caused spinal cord oxidative stress and apoptotic injury, and subsequent spinal reflex dysfunction. Next, BPV was used to culture SH-SY5Y cells and the results also demonstrated that it stimulated mtROS production in a concentration-and time-dependent manner with parallel cellular toxic injury. SOD2 resides predominantly in the mitochondrial matrix as a key antioxidant enzyme. It was found also in nucleoid complexes with mitochondrial DNA to protect them from oxidative injury and inactivation respectively. Amazingly, SOD2 appears to subject to inactivation in response to oxidative stress, and tyrosine nitration [26]. In contrast, we found that SOD2 transcription was upregulated in BPV-induced neuron oxidative stress. The mechanism governing the interaction of mtROS production and SOD2 transcription in above oxidative injury model is not clear.
Cells survival or death is closely related to the activation of JNK signaling which modulates stress related pathway in oxidative injury [27]. SOD2 transcription through mtROS-driven JNK signaling has AGING not been explored. In the present study, we investigated JNK phosphorylation and SOD2 transcription. The results showed that the activation of JNK and SOD2 transcription were increased in response to BPVinduced oxidative stress. So, we speculated that BPV could stimulate mtROS production, following activated JNK signaling which up-regulated SOD2 transcription. Antioxidant NAC was used to inhibit mtROS production for confirm its effect on the activation of JNK signaling and SOD2 transcription. The results  AGING showed that NAC effectively reduced mtROS production, antagonized the JNK phosphorylation and down-regulated SOD2 transcription. Further, JNK gene knock-down inhibited BPV-induced the up-regulation of SOD2 transcription. At same time, a JNK inhibitor (sp600125) also decreased JNK phosphorylation and down-regulated SOD2 transcription. Above evidence suggested that the activation of JNK signaling may be responsible for mtROS-stimulated SOD2 transcription in BPV-induced neuron oxidative stress.
Cells lose the capacity for being adequate to scavenge ROS contributing to mitochondria oxidative stress and apoptosis pathway [28]. The scavenging ability of SOD2 plays a critical role in maintaining the dynamic balance of ROS production and the integrity of mitochondria. SOD2 deficiency shows a variety of mitochondrial abnormalities such as reduced complex I and II activities [26,29], enhanced lipid peroxidation and increased oxidative stress [30]. Over-expression of SOD2 decreases lipid peroxidation and mtROS generation, subsequently inhibits oxidative stress and cells death [31]. Mitochondrial depolarization is a critical and relatively early event in mitochondrial oxidative injury process, which eventually leads to mitochondrial permeability transition and subsequent apoptosis [32]. For determining the role of SOD2 in BPV-induced neurotoxic injury, siRNA was used to knock down SOD2 gene. The results showed that SOD2 gene knocke-down enhanced BPV-induced mtROS production, accompanied with MMP declining and apoptotic injury. Further, mito-TEMPO, a mitochondrial antioxidant, could protect cells against BPV-induced neurotoxic injury. Above results suggested that Some limitations should be noted that BPV application in clinical is 0.5%-0.75%. For building spinal cord injury model, 2.5% BPV was used for intrathecal injection in rats as previously described [33], which was greater than clinical doses. In vitro injury model, BPV concentration was 2 mM, which was equal to 0.06% clinical concentration. According to previous report, almost all SH-SY5Y cells cultured with clinical concentration of BPV will be killed [34]. In clinic, the axons in the nerve roots of the cauda equina suffer from the brunt of BPV intrathecal injection which is greater than the cells culture concentration. So, BPV-induced neurotoxic injury can be confirmed in vivo and vitro models.
In summary, our study reveals that mtROS-JNK signaling is a co-regulator of SOD2 transcription in BPV-induced oxidative stress. Enhancing the antioxidant ability of SOD2 might be employed as a promising therapy in the prevention of BPV-induced neuron oxidative injury.

Surgical procedure and group assignment of rats
This study was approved by the Animal Research Center of Southern Medical University (protocol number: SYXK-2016-0167, Guangzhou, China), which follows the Guide for the Care and Use of Laboratory Animals (NRC1996). Animal experiments were conducted in male Sprague-Dawley rats (8-10 weeks old, 250-300 g) from Animal Research Center of Southern Medical University. All rats received intrathecal catheterization as described previously [33]. Rats were allowed at least 2 days to rest for recovery from the operation. Rats with tail movements or motor dysfunction in the hind limbs were not used in next experiments. For building BPV-induced rat spinal cord injury model, we used 2.5% BPV as previously described [33]. Under sevoflurane anesthesia (1.5%) with oxygen and room air via a nose cone, 2.5% BPV hydrochloride (dissolved in 0.9% saline) or 0.9% saline was injected intrathecally at the L 5 -L 6 intervertebral space using a 50gauge needle. The volume of the injectate was 0.2 µl/g body weight.

Spinal reflex function in rats
Spinal cord function was assessed by evaluating hind limb withdrawal reflex responses to mechanical and thermal stimuli with calculating the PWT (g) and TWL (s) in different times (before injection, 6 th h, 12 th h and 24 th h after BPV injection), with the experimenter blinded to initial treatment group as previously described [36].

Measurement of LDH
LDH (a maker of cell toxic injury) in culture medium was detected with Assay Kit (Beyotime Biotechnology, China) according to the manual instructions as described [37]. The amount of LDH from injury cells was quantified using absorbance captured at 490 nm.

Measurement of MDA and 8-OHdG
As described in previous study [38], rat spinal cord and cell DNA were extracted with DNeasy Blood and Tissue Kit (Qiagen, Germany) according to manufacturer AGING instructions. MDA and 8-OHdG production were determined using an ELISA kit (R&D Systems, MN, USA) according to kit instructions. The average MDA and 8-OHdG concentration per microgram of protein for each experimental group was calculated.

Western blot assay
Preparation of lysates, determination of protein concentrations, electrophoresis, and immunoblotting were conducted as previously described [9]. They were immunoblotted with anti-SAPK/JNK (1:500 The immunocomplexes were visualized using chemiluminescence. Band densities were measured using a densitometer and analyzed with Quantity One analysis software (Bio-Rad, Hercules, CA). Relative protein expression levels were normalized to corresponding βactin bands.

Detection of mtROS
Measurement of mtROS was performed using MitoSOX (Invitrogen, Carlsbad, CA). Briefly, after drugs treatment in 6-well plates, cells were incubated with 5 μM MitoSOX at 37 °C for 15 min, then collected and washed three times with phosphate buffer saline (PBS). The concentration of MitoSOX at 5 uM was chosen based on previous study conducted in SH-SY5Y cells [23], and our preliminary experiments which showed that the suitability to catch oxidative stress in our experimental settings. Measurement was performed using flow cytometry (BD FACS Calibur, BD Biosciences, USA) at excitation/emission wavelengths of 510/580 nm.

MMP assay
Mitochondrial depolarization was measured with JC-1 assay kit (Life Technologies Corporation, USA). After treatment, cells were incubated for 20 min with JC-1 staining solution (5 μg/ml) at 37 °C and rinsed twice with PBS. MMP was monitored by determining the relative amount of dual emissions from mitochondrial JC-1 monomers or aggregates using flow cytometry [39]. In injury cells with low MMP, JC-1 remains in the monomeric form and shows only green fluorescence. Mitochondrial depolarization is indicated by a decrease in the polymer/monomer fluorescence intensity ratio.

Apoptosis assay by flow cytometry
After treatment in 24-well plates, cells were rinsed with PBS, and resuspended in 500 μl binding buffer. Annexin V-FITC (5 μl) and propidium iodide (5 μl) (KeyGEN, Nanjing, China) were added in cells following suspension. After a 10 min incubation, cell apoptosis was determined by flow cytometry (BD FACS Calibur, BD Biosciences, NJ, USA). Apoptosis are Annexin V-FITC positive and PI-negative (statistics of upper right quadrant and lower right quadrant).

Statistical analysis
Data are presented as means ± standard error of the means (SEMs). Statistical differences of multiple groups were calculated by multiple comparisons with variance analysis, followed by Turkey's post hoc test. Differences between two groups were calculated by two tailed unpaired or paired Student´s t test. Statistical analysis was performed with SPSS soft-ware 13.0 (SPSS Inc., Chicago, IL). Significance was set at P< 0.05.

AUTHOR CONTRIBUTIONS
L.J. and X.Y. designed the study the experiments and wrote the paper. X.L., J.H. and Z.W. performed the experiments. X.Y. analyzed the data. X.R. revised the manuscript. All authors read and approved the final manuscript.