Parecoxib, A Selective Cyclooxygenase Inhibitor, Attenuates C-Jun N-Terminal Kinase Activation in Experimental Subarachnoid Hemorrhage Induced Early Brain Injury

Background: Increased evidence has disclosed early brain injury (EBI) might determine the outcome of patients suffering from aneurismal subarachnoid hemorrhage (SAH). This study is of interest to examine the efficacy of parecoxib, a cyclooxygenase-2(COX-2) inhibitor, on SAH-induced EBI. Methods: The rodent SAH model was employed. 500/1000/2000ug/kg/day parecoxib was administered via osmotic mini-pump. CSF samples were collected to examine IL-1β, IL-6, IL-8 and TNF-α. Cerebral cortex was harvested for C-Jun N-terminal kinase (cJNK), NMDARs (western blot), B-cell lymphoma 2 (Bcl-2), caspases (rtPCR) measurement. Results: Parecoxib dose-dependently reduced the bio-expression of cJNK when compared with the SAH groups. No significant differences were found in the levels of IL-8 among the experimental groups. The levels of IL-1β, IL-6 and TNF-α were increased in animals subjected to SAH, compared with healthy controls, but not in the mediumand high-dose parecoxib treatment group. Moreover, the levels of TNF-α in the SAH-only and SAH-plus vehicle groups were significantly elevated, and treatment with parecoxib (1000ug/kg) reduced the level of healthy controls. Cleaved caspase-3 and -9a were both significantly reduced in 2000ug/kg parecoxib treatment groups. Likewise, NMDAR-2a was significantly suppressed in parecoxib treatment SAH groups (p <0.01). Conclusion: Treatment with parecoxib exerts its neuroprotective effect through the dual effect of inhibiting cJNKactivated neuro-inflammation and reduced mitochondrion-related caspase-9a expression. Besides, parecoxib decreased CSF levels of TNF-α and IL-1β, which contributes to the anti-delayed vasoconstriction effect. This study lends credence to support the COX-2 inhibitor could attenuate SAH-induced EBI.


Introduction
Delayed ischemic neurological deficit, and acute cerebral ischemia subordinate to subarachnoid hemorrhage (SAH) has become a major disability in patients following a ruptured aneurysm [1][2][3][4][5]. Despite more than fifty years of studies focusing on the correction and prevention of delayed vasospasm, the efforts, in a multifaceted way, trying to reverse this condition, have proven to illustrate poor outcome. Increased evidence shows early brain injury (EBI) [6][7][8][9][10] enrolled cortical spreading depression, early cortical depolarization waves, and impaired neurovascular coupling, play key roles in neurological deterioration after SAH. Connolly stated the mortality rate in SAH patients remained 27% to 44%, and more than half of all SAH patients survive with serious cognitive and functional impairment [11]. These results stimulate efforts to dissect the cellular and molecular basis of EBI accompanying SAH to establish rational therapeutic targets.
Blood clots, existing in the subarachnoid space, are able to induce acute arteries and arteriolar constriction, passive venous obliteration and delayed arterial spasm [1,[11][12]. Previous studies have focused on the vasoconstriction of arteries; however, the precise mechanism and pathogenesis of arteriolar constriction and venous obliteration associated EBI remain unclear. A growing body of both direct and indirect evidence shows possible mechanisms to produce EBI include oxidative stress [13], nitric oxide(NO)/nitric oxide synthase(NOS) uncoupling [14], matrix metalloproteinase -9 (MMP-9) induced BBB breakdown [15], nuclear factor erythroid related factor 2 (Nrf2) pathway related oxidation [16], activation of c-Jun N-terminal kinase pathway, [14,17,18] increased proinflammatory cytokines [19] vascular endothelial growth factor, and mitogen-activation protein kinase [20]. Energy over-loaded after hemoglobin degrading has been shown to induce EBI. Owing to lack of effective therapies to achieve a better outcome in SAH patients, studies have become focused on the pathogenesis of EBI.
Parecoxib, a intravenous administration form of a cyclooxygenase-2 (COX-2) selective nonsteroidal anti-inflammatory drug (named celecoxib), is used to treat osteoarthritis, rheumatoid arthritis, ankylosing spondylitis, acute pain, and juvenile rheumatoid arthritis [7]. COX (a prostaglandin H (PGH) synthase) is a key regulatory enzyme that catalyzes the conversion of arachidonic acid and associated prostaglandin (PG) G2 and PGH2 to a variety of eicosanoids, such as PGE2, PGD2, PGF2a, PGI2, and thromboxane (TX) A2 [21]. Recently, COX-2 inhibitor has been reported to have a neuroprotective effect in the experimental intracerebral hemorrhage model as well as induce apoptosis in Huh7 and HepG2 cell line cultures [22]. In the study of SAH rabbits, intramuscular administration of parecoxib was shown to attenuate SAH induced vasospasm and basilar artery wall thickness [23]. Owing to the concerning side effects of major cardiovascular events, increased risk could be avoided by shortening the duration of administered COX-2 inhibitor and reducing the therapeutic dosage in patients who have pre-existing cardiovascular risk factors or disease. In this study, parecoxib was administered via an osmotic mini-pump at a minimum dosage and within a short duration, devoid the mentioned side effect of cardiovascular and renal problems. 0.9% saline was used as a vehicle.
Taking these findings together, we propose that COX-2 inhibitors, with its unique property in PGs expression, may be effective in SAHinduced EBI. Given the importance of arterial lesion formation, associated pro-inflammatory cytokine stimulation on leukocyte and endothelial dysfunction in the rat SAH model was used to test the hypothesis that COX-2 inhibitor significantly attenuates caspaserelated apoptosis and C-Jun N-terminal kinase (cJNK)-associated inflammation following experimental SAH.

Induction of experimental SAH
Fifty-four male Sprague-Dawley rats, weighing between 320-450 g (BioLasco Taiwan Co., Ltd., authorized by Charles River Lab), were used in this study. All the experimental protocols were authorized and superintended by the University of Kaohsiung Medicine Animal Research Committee. The rats received anesthesia by an intraperitoneal injection of a mixture of 0.9 mg/100 gm xylazine and 5.0 mg/100 gm KetaVed. 0.1 ml fresh arterial blood was withdrawn and injected into the cisterna magna using a stereotactic apparatus (Stoelting, Wood Dale, IL, 60191, US) [24,25]. After the induction, animals were placed in ventral recumbent position for 30 minutes to let ventral blood clot formation. After monitoring for respiratory distress and giving mechanical ventilation if necessary, the animals were returned to the vivarium until becoming fully awake. A habitat with a 12 hr light-dark cycle and access to food and water ad lib was offered.

General design of experiments and treatment groups
The animals were subdivided into the following groups (9 rats/ group): 1) sham operated (no SAH); 2) SAH only; 3) SAH plus vehicle; 4) SAH plus parecoxib (500 ug/kg/day); 5) SAH plus treatment with parecoxib (1000 ug/kg/day); and 6) treatment with 2000 ug/kg parecoxib in SAH rats. The dosage was adjusted according to our pilot study, devoid of hepato-renal toxicity and G-I bleeding. The first injection was given at 1hr after induction of SAH by using an osmotic mini-pump (Alzet corp, Palo Alto, CA 94301, USA). After reanesthesia, CSF sampling was obtained through a 30-gauge needle into the cranial-cervical junction by using stereotactic apparatus. The animals were sacrificed by perfusion-fixation 72hr after SAH. Cortical tissue were sampled by means of placing a 22-gauge needle inserted 3 mm in depth into the skull bone (N=5) through a burr hole craniectomy (2 mm apart from the bregma) at a 24 hr interval.

Perfusion-Fixation
By the end of the study, the animals were re-anesthetized by administration of 7 mg/kg Zoletil 50 (a combination of tiletamine hypochloride and zolazepam hypochloride. VIRBAC, L.I.D., Carros 06516, France). The femoral artery was catheterized to monitor blood pressure and obtain blood samples to determine arterial blood gas, BUN, Cr, GOT and GPT levels. Perfusion-fixation was performed as opening the thorax, and the left ventricle was canalled with a NO18 catheter with clamping the descending aorta, and opening the right atrium. 120 mL of 70 mm Hg of 0.01M phosphate buffer (pH 7.4) was used for perfusion, followed by fixation with 120 mL 2% paraformaldehyde in the PBS solution at 36°C under a perfusion pressure of 80 mm Hg. The harvested brain was immersed in a fixative at 4°C overnight. Visual inspection made sure that formed blood clots covered the basilar artery (BA) in all SAH animals.

Neurological assessment
A modified limb-placing tests (MLPT), [26] and a modified Voetsch neuro-score test [27] were performed before and at 24hr after the induction of SAH by an investigator blinded to the experiment set. The behavioral examination comprised two parts: ambulation and placing/stepping reflex examinations were performed before and after animals were subjected to SAH. The modified limb-placing tests (MLPT) was composed of two limb-placing tasks to examine the sensorimotor integration of the forelimb and the hindlimb through tactile and proprioceptive stimulation. The ambulation task was performed by suspending the rat 10 cm over a table and evaluated the stretch of the forelimbs towards the table: normal stretch, 0 points; abnormal flexion, 1 point and then, by positioning the animal along the edge of the table, and let its forelimbs placed out of the edge and allowed to move freely. The same performance of hindlimb was checked as normal performance, 0 point; performance with a delay (2 s) and/or incomplete, 1 point. A motor deficit index (MDI) was calculated for each rat at a set time interval. The final index was the sum of the scores of walking with lower extremities and placing/ stepping reflex. Animals with MDI score more than three were defined as paraplegic, whereas animals with MDI score less than three were regarded as neurological deficit. The modified Voetsch neuro-score was to evaluate the head movement, confrontation, reflex stimulated by ear pinch, proprioception and four limbs movement. The modified Voetsch neuroscore is a vertebrobasilar scale score of sensorimotor ability. To examine the spontaneous movement of head: moves in all dimensions, 3 points; prefers one side, 2 points; only movement to 1 side, 1 point; flexed to unilateral side, 0 point. Confrontation task was defined as: fully responsive, 3 points; moderately responsive, 2 points; minimally responsive, 1 point and coma, 0 point. The hearing test by auditory startling: finger rubbing, 3 points; snap of fingers, 2 points, loud clapping, 1 point and no startling, 0 point. Pain reflex after ear pinch: brisk and symmetrical reaction, 3 points; slightly diminished or asymmetrical reaction, 2 points; severely diminished and asymmetrical reaction, 1 point; no reaction, 0 point. To check the sensation of the neck in a cage: reached 2 walls, 3 points; touched 1 wall, 2 points; moved along the base, 1 point and no movement, 0 point. To check craniocaudal movement, turned bilaterally, 3 points; preferred circling 1 side, 2 points; only turned circling to 1 side, 1 point; fallen to 1 side, 0 point. To test the axial sensation, brisk and symmetrical reaction to stimuli, 3 points; slightly diminished or asymmetrical reaction, 2 points; greatly diminished and asymmetrical movement, 1 point and no reaction, 0 point. To evaluate forelimb and hindlimb movement: equal and bilateral movement, 3 points; slight asymmetry, 2 points; prominent asymmetry, 1 point; paresis, 0 point. To check the ability to climbing, the rat is able to climb to top, 3 points; impaired climbing, 2 points; stationary gripping, 1 point and fell immediately, 0 point. The final stage is to examine movement to the stimuli of beam, explored both ends, 3 points and some movements, 2 points. The sum of neuroscores ranged from 0 to 30, which stood for normal performance to comatose status and paraplegia.

Immuno-staining with MCP-1, NeuN polyclonal and GFAP monoclonal antibody and brain edema studies
The basilar arteries (BA) were frozen and cut into 25-μm-thick sections with the assistance of ultra-microtome E (Leica EM UC7, Union Optical & Instruments and Nanotechnology. New Taipei City 22101, Taiwan.) Five BA cross-sections of each animal (the middle third) were analyzed by two investigators blinded to the experiment set. A MCP-1 recruited monocytes, memory T cells, and dendritic cells were detected using a video-assisted microscope (x400). Briefly, coronal sections of the isolated rat basilar arteries were stored overnight on slides at -80˚С. Rabbit monoclonal anti-rat MCP-1 antibody, polyclonal anti-rat NeuN antibody, glial fibrillary acidic protein (GFAP) monoclonal antibody were used at a dilation of 1:40, 1: 20 and 1:40, respectively, and immunostaining was performed for 40 min at 25˚С followed by drying overnight as described in the mouse monoclonal alkaline phosphatase anti-alkaline phosphatase (APAAP) technique [28]. A modified wet-dry weight method was used for determining the water content of the brain at 48 hr, when the weight lost was robust and reached a peak for 2 days after the induction of SAH [28]. The brain cortex was harvested from the skull bone, pondered on a precision microbalance (Precision Weighing Balances, Bradford, MA 01835, USA), and then let dried in an oven at 60°C for 48 hr. The dry specimen was re-weighed, and brain edema index was calculated as (wet cortex weight / dry cortex weight).

Quantification of mRNA expression of IL-1β, IL-6, IL-8, and MCP-1
ABI PRISM 7900 System (Applied Biosystems, Foster City, CA, 94404, USA) was used to examined the expression of cytokines mRNA. As indicated by the manufacturer's instructions, the mRNAs for IL-1β, IL-6, IL-8, and MCP-1 expression were estimated, while 18S was assigned as a housekeeping gene owing to its stable expression in the condition of SAH stimuli. Each sample was launched into a TaqMan ® Human Cytokine Card that enclosed probes and primers for specific targets as well as the 18S ribosomal RNA as a standard control. This procedure was performed through an affixed filling reservoir and a vacuum loading process via the ABI PRISM ® Card Filling Station.
Target mRNAs were standardized according to the reference gene (18S), and final data were expressed as a relative fold from baseline values. Comparative mRNA expression was set by the Livak and Schmittgen CT method [30]. The results were analyzed once a 5-fold increase was observed in the mRNA levels compared with baseline to let data consistent.

Estimation of Bcl-2, cleaved caspase -3 and -9a mRNA by rt-PCR
The levels of activated caspase-3 and caspase-9a mRNA in the cortical homogenates were determined by TriPure RT-PCR Reagent reverse: 5'-CCAGTGAGTTTCCCGTTC-3'. After incubation with Avian Myeloblastosis Virus Reverse Transcriptase (AMV RT) (Promega, WI 53711, USA), the cleaved caspase -3, -9a and GAPDH cDNA were amplified. The amplified cDNA fragments were detected via agarose gel electrophoresis swirled with 1 µL of ethidium bromide. The intensity of the gene bands was measured via a comet assay method. Owing to its stable expression under subarachnoid hemorrhage stimuli, GAPDH mRNA was used as a housekeeping gene for multiplexing. The tissue of every five animals was used for rt-PCR and three measurements for each animal specimen performed to obtain a mean value. The pro-caspase-9 was catalyzed and cleaved into caspase -9a (majority) and -9b. We chose cleaved caspase-9a to represent the activated caspase-9.

Statistical analysis
Data are expressed as the means ± standard deviation. For group comparisons, one-way analysis of variance was used for analyses (ANOVA) followed by the Newman-Keuls post-hoc test and the Student t test statistically. Difference, at a probability value less than 0.01, was considered significant.

General observation
Through all the course of this experiment, all animals showed no significant differences in the physiological parameters recorded, including glutamate oxaloacetate transaminase (GOT), glutamate pyruvate transaminase (GPT), blood urea nitrogen (BUN), creatinine, pH, blood pressure and arterial blood gas analysis among all groups. It proved that continued parecoxib pumping in the selected dosage has a number of pleiotropic effects, but devoid of hepatic and renal toxicity and bowel bleeding ( Table 1). According to the brain edema after the induction of SAH, there is no significant difference among all the experimental groups (Figure 1, lower panel). Increased MCP-1(+) cells in the adventitia of vascular wall, NeuN (+) neurons with vacuolar formation in the cortex and GFAP(+) astrocytes were observed in the SAH groups, when compared with the parecoxib treatment SAH group, which correspond to the observation of MCP-1 mRNA ( Figure  1 upper and middle panels).

Neurological deficit after SAH
A summary score of four limbs ambulation was obtained to assess the sensorimotor incorporation of the forelimb and hindlimb, and placing/stepping reflex as a reflective response to tactile and proprioceptive stimuli in the SAH groups was significantly higher than the healthy controls and parecoxib groups. The sum of these two tests was considered as motor deficit index (MDI). The values of MDI in the SAH and SAH+vehicle groups were 2.92 ± 0.46 and 2.87 ± 0.40 respectively, compared with a score of 0 and 1.20 ± 0.46 in the healthy controls and 2000ug/kg parecoxib respectively. Treatment with parecoxib significantly improved the MDI in the SAH groups ( Table  2). The modified Voetsch neuro-scores were dose-dependently improved in the parecoxib + SAH group and the healthy controls when compared with the SAH groups (Table 3).

mRNA expression of IL-1β, IL-6, IL-8, and MCP-1
After SAH, the CSF levels of IL-1β, IL-6, IL-8, and MCP-1 were found to increase 1000-and 3000-fold at 72hr, when compared with the sham operated group. Administration of 2000 ug/kg parecoxib reduced cytokine levels by 7%, 25%, 42% and 21% for IL-1β, IL-6, IL-8, and MCP-1 respectively, 24hr after SAH ( Figure. 2). Level of IL-1β was reduced in the parecoxib treatment groups in a dose-dependent mechanism in this study. 2000 ug/kg/day parecoxib significantly reduced the levels of IL-1β, IL-6 and MCP-1 in animals at the time points of 24hr and 72hr after the induction of SAH. Meanwhile, treatment with parecoxib tended to reduce the level of proinflammatory cytokines; the IL-8 level (at 72hr) failed to achieve statistical difference from the SAH group ( Figure 2).

cJNK (p55/p46) expression
The cJNKs were demonstrated to play a pivotal role in the early cascades of pro-apoptotic signaling. cJNK was chosen as an example of modulating the program for cellar survival and differentiation, which has emerged as a key molecule in regulating the pathogenesis of SAHinduced EBI. In this study, parecoxib dose-dependently reduced the expression of cJNK (p55) in SAH animals. (Figure 3, p<0.01) The expression of NMDAR-1, NMDAR-2a, and NMDAR-2b protein induced in the SAH groups (SAH only, treatment with vehicle) when compared with the sham-operated groups (p<0.01). Parecoxib (at 1000 and 2000 ug/kg/day) decreased NMDAR-2a expression when compared with the SAH groups (P<0.01). Parecoxib dose-dependently reduced the expression of NMDAR-1 (P<0.01). Increased NMDAR-2b expression was noted in all the SAH groups (Figure 4, P >0.01).

The expression of activated Caspase-3 and -9a mRNA
Cleaved caspase-3 and -9a (the major functional variant) were used to examine the neuronal apoptosis. The activated caspase-9a was observed to increase in the SAH groups when compared with the sham-operated groups (p<0.01). 1000 ug/kg and 2000 ug/kg parecoxib decrease cleaved caspase-3 level when compared with the SAH groups respectively (Figure 5, p<0.01). Only high-dose parecoxib could attenuate apoptosis-related caspase-9a and caspase-3 when compared with the SAH groups.

ApoPrimer for Bcl-2 mRNA
To assess the regulation of apoptosis through mitochondrial cytochrome c release, Bcl-2 mRNA was evaluated. The Bcl-2 mRNA was induced in the treatment of 2000 ug/kg/day parecoxib in SAH rats, when compared with the SAH group ( Figure 6, p<0.01).

Discussion
In this study, we have shown that parecoxib, a selected cyclooxygenase-2 inhibitor, ameliorates neurologic deficits in the behavior studies and brain damage, including neuronal death, and brain edema in the MCP-1 and NeuN immunostaining in rats subjected to SAH. EBI may promote delayed neuronal death by SAHinduced inflammation, associated increased BBB permeability, cerebral edema and intracranial pressure, and resulting neuronal dysfunction [8]. The family of cyclooxygenases (COX) includes two isoforms: COX-1, which functions in cellular homeostasis, and COX-2, which is responsible for up-regulating inflammatory cytokines and mitogens [21,30]. Induced COX-2 has profound effects on the brain via the complex effects of prostanoids that are abundant in the central nerve system, which modulate glutamate release, cerebral vasoconstriction, induction of reactive oxygen species (ROS) release, and make alterations on neuroendocrine function [7]. Besides, it is possible that COX-2 up-regulation in injured neurons could contribute to neuronal death following brain injury. Inhibition of COX-2 activity has been demonstrated to protect against ischemic damage, kainate-induced seizures, and neuroinflammation-induced neurotoxicity. [21] In this study, parecoxib was observed to be able to reduce early COX2 expression via inhibiting cJNK (p55) expression (dose-dependently), reduce IL-1β, IL-6 and MCP-1( at 2000 ug/kg/day at the time point of 24hr and 72hr after the induction of SAH), IL-8 (of the dosage of 2000 ug/kg/day at 24hr after the induction of SAH), and also reduce cleaved caspase-1 (at 1000 and 2000 ug/kg/day) and Caspase-9a (at 2000 ug/kg/day).
Disrupted energy metabolism within the central nerve system was believed to cause deadly damage to neurons in both acute and chronic neurological disorders [32]. Overstimulation of glutamate receptors, especially NMDARs related to intracellular glutamate release, results in acute and delayed neuron injury, which included chronic diseases: Alzheimer's disease, Parkinson's disease, Huntington's disease, HIV associated dementia, multiple sclerosis, amyotrophic lateral sclerosis (ALS), and acute disorders: stroke, central nervous system (CNS) trauma and seizures [33]. Both might be initiated by distinctive mechanisms but might share an ultimate similar pathway to neuronal injury. Excitotoxicity, owing to over-activation of NMDA-sensitive glutamate receptors, allows excessive Ca2+ influx through the NMDAR-associated ionotropic channel. Increased intracellular Ca2+ resulted in free peroxide radicals production and excessive proteolytic activity, which finally led to cellular damage and death [34]. Unwarranted glutamate released from the necrotic cells leads to a cascade of auto-destructive events and progressive cellular death, which can last for hours or even days after the primary injury. In the present study, parecoxib was found to reduce the NMDAR-1, and NMDAR-2a (at 1000 and 2000 ug/kg) hours after the induction of SAH. Unlike pro-inflammatory cytokines observed in CSF and released from microglia or immigrating from the breakout of BBB in cerebral ischemia, activated NMDARs were short-lived and believed primary to the activation of other neuro-toxic events as early cyclooxygenase-2 activation.
There are three major pathways observed in the caspase-related apoptosis of mitochondria existing in mammals: the cell-surface (extrinsic) receptor-related pathway, the apoptosome (intrinsic) pathway and the cytotoxic lymphocyte-initiated granzyme B pathway [35][36][37][38][39][40][41]. Shi stated there are 14 caspases identified in mammals, subdivided into three groups: one is the initiator caspases, such as caspase-8 and -9, one is the downstream effecter caspases, like caspase-3, -6, and -7, and the other include pro-inflammatory enzymeenrolled caspases-1, -4, -5, -11, -12, and -13. Caspase-3 is believed to be the major effecter caspase in neuronal apoptosis [42]. It has been observed that caspases act on upstream intrinsic apoptosis, and on downstream of cytochrome c release [43]. Caspase-9 behaves as an initiator and has been shown able to uncouple the mitochondria and increase ROS production [13,40]. In this study, parecoxib was observed able to reduce the caspase-9a (2000 ug/kg/day) and decrease caspase-3 (of a dosage of 1000 and 2000 ug/kg/day). It is believed parecoxib, at a selected dosage, has a neuroprotective effect by blocking both initiator caspase-9a and effector caspase-3 activation.
Four conserved Bcl-2 homology (BH) domains were found in Bcl-2 family members, designated BH1, BH2, BH3, and BH4 [43,44]. An anti-apoptotic characteristic was conserved in a carboxy-terminal hydrophobic domain, through which BCL-2 is essential for its targeting to membranes such as the mitochondrial outer membrane [44]. The anti-apoptotic molecules, such as Bcl-2 or Bcl-XL, can inhibit the activation of BAX following a death signal. In contrast to inactive Bax, [43] Bcl-2 is an integral membrane protein heavily localized to mitochondria, exerts an anti-apoptotic effect. 2000 ug/kg/day parecoxib enhanced the cytoplasm Bcl-2 in the SAH study. The mechanism that parecoxib exerts anti-apoptotic effect via inducing cytoplasm Bcl-2 expression merits a further study lateral.
Increased adhesion molecules and pro-inflammatory cytokines were observed increased in cerebrospinal fluid (CSF) after aneurismal SAH [5,24]. However, the relationships among the development of inflammatory response, early brain injury, and delayed cerebral ischemia in the brain after SAH need to be clarified. The cumulative findings indicate that the up-regulation of pro-inflammatory cytokines is antecedent to radiographic vasospasm (peak at 4th to 14th days after SAH) in human, and attenuation of cytokines tends to minimize vascular constriction and reduced early brain ischemia in animals. [6,44] Monocyte chemoattractant protein-1 (MCP-1) is able to predict outcome and vasospasm following aneurismal SAH and is highly expressed to reduce the tight junctions of vascular endothelium [19,46]. Elevation in the mRNA levels for IL-1β, IL-6, IL-8, TNF-αand adhesion molecules has been observed in the pilot study. In the study, parecoxib was observed to be able to reduce IL-1βand MCP-1 (at the dose of 1000 and 2000 ug/kg/day) and tended to reduce the IL-8 level. This result indicated parecoxib selectively reduced the inflammatory cytokines.
cJNK, belonging to the family of mitogen-activated protein kinase (MAPK), plays a role in T cell differentiation and cellular apoptosis pathway [14,39]. Though mediated by regulated-upon-activation normal T cell expression and presumably secreted (RANTES) and colony stimulating factor 2 (CSF2), cJNK is believed to be able to modulate cellular apoptosis, neuro-degeneration, proliferation and differentiation, cytokine production and subsequent inflammatory reactions [17]. In Jadhav et al's study, through modulating cJNK/ MAPK-related molecules, microRNA was able to suppressed microglia-elicited neuro-inflammation [18]. In the study of Hela cell culture, cleaved caspase-3 was demonstrated to disassemble the JIP1/JNK complex and set off the initiation of apoptosis, which corresponds to our findings [17]. Through suppression of cleaved caspase-3 (at a dosage of 1000 and 2000 ug/kg/day), parecoxib was able to reduce cJNK, and attenuated associated cytokine activation.
In summary, the results of this study show that continued administration of parecoxib, at a selected dosage, is safe and efficacious in the prevention of EBI in this experimental model and is meritorious of further investigation. Decreased levels of NMDAR-1 as well as caspase-3 may contribute to the anti-apoptotic effect of this compound. Besides, parecoxib, by reducing cJNK activation, exerts a dual effect on the MAEK-related apoptotic pathway and IL-1β related inflammatory cascade.

Conclusion
The outcome of SAH patients is persistently devastating, and has changed little after decades of research and treatment focusing on cerebral vasospasm. These accumulated results arouse interest to consider the pathogenesis of SAH-induced EBI and its effects dictating the patient's outcome. The ultra-short activated cyclooxygenase-2 following SAH leads a cascade of neuro-inflammatory reaction. The breakout of BBB accompanying SAH may be another complicated pathway underlying the development and maintenance of EBI. This study shows that administration of parecoxib diminishes SAHinduced cerebral edema and subsequent glutamate-related excitotoxicity in a rodent model of SAH. This study suggests that parecoxib, an early cyclooxygenase-2 inhibitor, could prove clinically useful in treating SAH-induced early brain injury.