Synergistic Interactions between Cytokines and AVP at the Blood-CSF Barrier Result in Increased Chemokine Production and Augmented Influx of Leukocytes after Brain Injury

Several lines of evidence indicate that the blood-cerebrospinal fluid barrier (BCSFB), which primarily resides in the choroid plexus (CP), plays a significant pathophysiological role not only in neuroinflammatory diseases, such as multiple sclerosis, but also in traumatic brain injury (TBI). Here we investigated how arginine vasopressin (AVP) regulates function of the BCSFB in the context of post-traumatic neuroinflammation. It has previously been shown that AVP exacerbates various forms of brain injury, but the mechanisms underlying this AVP action are poorly understood. Type 1A AVP receptor is highly expressed on the CP epithelium and the CP synthesizes AVP. Using the controlled cortical impact model of TBI, we demonstrated decreased post-traumatic production of proinflammatory mediators by the CP and reduced influx of inflammatory cells across the BCSFB in AVP-deficient Brattleboro rats when compared with Long-Evans rats, a parental strain for Brattleboro rats. Arginine vasopressin was also found to play an important role in post-traumatic activation of c-Jun N-terminal kinase (JNK) in the CP. In the CP epithelial cell cultures, AVP augmented the tumor necrosis factor-α– and interleukin-1β–dependent increase in synthesis of proinflammatory mediators, including neutrophil chemoattractants, an action largely dependent on the JNK signaling pathway. Under in vivo conditions, a selective JNK inhibitor decreased the post-traumatic production of neutrophil chemoattractants by the CP and reduced the influx of neutrophils across the BCSFB. These results provide evidence for the synergistic interactions between proinflammatory cytokines and AVP, a ligand for G protein-coupled receptors, and support a pathophysiological role of AVP in post-traumatic neuroinflammation.


Introduction
The choroid plexus (CP) is best known for its ability to produce cerebrospinal fluid (CSF), but this highly vascularized tissue located in all four cerebral ventricles is also a major site for the blood-CSF barrier (BCSFB). The BCSFB is formed by a single layer of cuboidal epithelial cells enclosing the leaky choroidal microvessels [1]. Several lines of evidence indicate that the BCSFB plays an important pathophysiological role in neuroinflammatory diseases, such as multiple sclerosis [2,3]. We have also demonstrated that in traumatic brain injury (TBI), which is accompanied by neuroinflammation, the BCSFB plays a significant role in synthesis of proinflammatory cytokines and CXC and CC chemokines, and in post-traumatic invasion of neutrophils and monocytes [4,5]. Post-traumatic influx of leukocytes represents an integral part of secondary injury processes, and laboratory studies of TBI have shown that invading leukocytes significantly augment the formation of cerebral edema, the loss of neural tissue, and functional deficits resulting from injury [6][7][8][9][10]. Since posttraumatic neuroinflammation progresses at a relatively slow pace, there is an extended window of opportunity for anti-inflammatory intervention in TBI [11].
In the present study, we investigated the effect of arginine vasopressin (AVP) on function of the BCSFB. It has been demonstrated that AVP plays an important pathophysiological role in various forms of brain injury, including TBI, cerebral ischemia, and intracerebral hemorrhage [12][13][14][15][16], but the underlying molecular and cellular mechanisms are not well understood. These investigations are supported by observations that the type 1A AVP receptor (AVPR1A) is highly expressed on the choroidal epithelium [17] and that the CP has the ability to produce AVP [18]. Here we tested the hypothesis that AVP acts synergistically with proinflammatory cytokines tumor necrosis factor-a (TNF-a) and interleukin-1b (IL-1b) to amplify the brain inflammatory response to injury. Using AVP-deficient Brattleboro rats and their parental Long-Evans strain, we showed that AVP increases the post-traumatic production of proinflammatory mediators by the CP and significantly enhances the leukocyte influx across the BCSFB. We also showed that AVP has a substantial effect on posttraumatic activation of c-Jun N-terminal kinase (JNK) in the choroidal tissue. In the CP epithelial cell cultures, AVP amplified the TNF-a-and IL-1b-dependent synthesis of neutrophil chemoattractants. This AVP action was primarily mediated by the JNK signaling cascade. Wild-type rats treated with a selective JNK inhibitor showed a decrease in post-traumatic production of CXC chemokines in the CP and a reduction in the magnitude of neutrophil influx across the BCSFB.

Rats
Adult male Long-Evans and homozygous AVP-deficient Brattleboro rats (Avp di/di ) weighing 200-250 g (Harlan, Indianapolis, IN) were used. Brattleboro rats have a single-nucleotide deletion in the Avp gene [19]. The rats were kept at 22uC with a 12-h light cycle and maintained on standard pelleted rat chow and water ad libitum.

The rat model of TBI
The surgical and animal care procedures used in this study were approved by the Animal Care and Use Committee of Rhode Island Hospital and conformed to international guidelines on the ethical use of animals. Six to nine rats per group/time point were used. The controlled cortical impact model of TBI was employed as previously described [4,5]. In brief, rats were anesthetized with intraperitoneal pentobarbital sodium (60 mg/kg) and a 4-mm craniotomy was performed on the right side of the skull to expose the dura, with the center of the opening located 3 mm posterior to the bregma and 2.5 mm lateral to the midline. The velocity of impact was 5 m/sec and the duration of impact was 50 msec. The diameter of the impactor's tip was 2.5 mm and the depth of brain deformation was set at 3 mm. At selected time points after injury, the rats were perfused transcardially with ice-cold 0.9% NaCl and the samples of the lateral ventricle CP, both ipsilateral and contralateral to injury, were collected separately and pooled into three subgroups (2-3 rats per subgroup). To inhibit the JNK activity in vivo, a selective JNK inhibitor SP600125 was injected i.p. at 4 and 8 h post-TBI at a dose of 30 mg/kg for each injection. SP600125 was dissolved in DMSO at a concentration of 12.5 mg/ ml.

CP epithelial cell cultures
Immortalized rat CP epithelial cells Z310 [20] were kindly provided by Dr. Wei Zheng (Purdue University, West Lafayette, IN). These cells have previously been characterized [21]. Lowglucose Dulbecco's modified Eagle medium supplemented with penicillin (100 U/ml), streptomycin (100 mg/ml), gentamicin (40 mg/ml), and 10% fetal bovine serum was used to culture Z310 cells. They were grown to confluence in six-well cell culture plates at 37uC in a humidified atmosphere of 5% CO 2 /95% air. The cells were serum starved in serum-free medium containing 0.1% BSA for 24 h prior to experimentation. On the day of experiment, the cells were exposed to TNF-a or AVP alone, or to the combination of TNF-a and AVP. A separate series of experiments was also performed in which IL-1b was used instead of TNF-a. To inhibit JNK or p38 activity, or to interfere with extracellular signal-regulated kinase (ERK) or NF-kB signaling, the Z310 cells were pre-incubated for 1 h with respective inhibitors and these inhibitors were also added to the culture media when the cells were exposed to TNF-a and AVP. The concentrations of CXCL1-3 and lipocalin 2 (LCN2) in the culture media were determined by ELISA using the DouSet kits from R&D Systems.

Real-time RT-PCR
Total RNA was isolated using NucleoSpin RNA II kit (Macherey-Nagel, Düren, Germany). First-strand cDNAs were synthesized using oligo(dT) 20 primer (50 pmol) and 15 U of ThermoScript RNase Hreverse transcriptase. The 20-ml reverse transcription reactions also contained 40 U of RNase inhibitor RNaseOut. For each reaction, 0.5 or 1 mg of total RNA was used and the reactions were carried out for 1 h at 50uC. Real-time PCR was performed using TaqMan chemistry. The sequences of primers and TaqMan probes for TNF-a, IL-1b, CXCL1-3, and CCL2 were previously described [4,5]. The following primers and TaqMan probe were used for LCN2: 59-CTGTACGGAAGAAC-CAAGGG-39 (forward primer), 59-CATTGGTCGGTGG-GAACA-39 (reverse primer), 59-CTGTCCGATGAACTGAAG-GAGCGAT-39 (probe). Cyclophilin A was used for the normalization of the data obtained [4,5]. The 25-ml PCR reaction mixtures contained 0.2 mM mixed dNTPs, 0.2 mM each primer, 0.1 mM TaqMan probe, 5 mM MgCl 2 , 1 U of HotStart Taq DNA polymerase, and 1/20 of the reverse transcription reaction product. For cyclophilin A, 1/1000 of the reverse transcription reaction product was used. The reaction mixtures were heated to 95uC for 15 min and were then subjected to 45 cycles of denaturation (15 sec) at 94uC and annealing/extension (60uC, 60 sec).

Western blotting
Proteins from the CP or Z310 cells were extracted using isotonic lysis buffer (150 mM NaCl, 50 mM Tris-HCl, pH 7.4, 2 mM EDTA, 1% Triton X-100), containing protease inhibitors (1 mM benzamidine, 100 U/ml aprotinin, 20 mg/ml antipain, 20 mg/ml leupeptin, 1 mg/ml pepstatin A, 1 mM PMSF) and phosphatase inhibitors (10 mM sodium pyrophosphate, 1 mM sodium orthovanadate, 1 mM sodium fluoride, 1 mM b-glycerophosphate). Proteins were resolved via SDS-polyacrylamide gel (4-12%) electrophoresis under reducing conditions and were transferred onto 0.2-mm nitrocellulose membranes (Invitrogen). After blocking with 5% ECL Advance blocking agent (GE Healthcare, Little Chalfont, UK) for 1 h at room temperature, the membranes were incubated with primary antibodies overnight at 4uC. Membranes were subsequently incubated with horseradish peroxidase-conjugated anti-rabbit or anti-mouse antibody for 1 h at room temperature. For detection, Lumigen TMA-6 (Lumigen, Southfield, MI) or SuperSignal West Dura extended duration (Pierce, Rockford, IL) chemiluminescence substrate and the Bio Imaging System Chemi Genius2 (Syngene, Frederick, MD) were used. In the analysis of the optical density of the bands on immunoblots, the levels of phosphorylated JNK, c-Jun, and ATF2 were normalized to the levels of total phosphorylated and nonphosphorylated forms of these proteins. This analysis was performed using ImageJ software (http://rsb.info.nih.gov/ij/).

JNK activity assay
Non-radioactive JNK activity assays were performed using recombinant c-Jun as a JNK substrate. The assays were completed in kinase buffer (25 mM Tris-HCl, pH 7.4, 10 mM MgCl 2 , 2 mM DTT), containing 0.1 mM sodium orthovanadate and 5 mM bglycerophosphate. The 50-ml reaction mixtures contained 10 mg of total protein (extracted either from the CP or Z310 cells), 200 ng of c-Jun fusion protein (Cell Signaling), and 200 mM ATP. The reactions were carried out for 30 min at 30uC. The reactions were terminated with 25 ml of 3X SDS sample buffer and the reaction mixtures were boiled for 5 min. Ten microliters of each reaction mixture were loaded on SDS-polyacrylamide gel, and after electrophoresis, proteins were transferred onto nitrocellulose membranes. The membranes were incubated with anti-p-c-Jun antibody (40 ng/ml). The amount of total c-Jun was assessed using anti-c-Jun antibody diluted 1:5000.

Myeloperoxidase (MPO) activity assay
To assess the number of neutrophils accumulating in the ipsilateral CP after trauma, the MPO activity assays were performed as previously described [4]. The activity of MPO was assayed spectrophotometrically at 24 h after TBI. One unit of MPO activity was defined as the degradation of 1 mmol of H 2 O 2 per min. The results are expressed as units of MPO activity per gram of tissue.

Statistical analysis
For statistical evaluation of data, ANOVA was used, followed by the Newman-Keuls test for multiple comparisons among means. The results are presented as mean values 6 SEM. p,0.05 was considered statistically significant.

Results
Vasopressin upregulates the post-traumatic production of proinflammatory mediators by the CP Using the same rodent model of TBI, which produces unilateral injury to the cerebral cortex and subcortical structures, we have previously shown that TBI results in a rapid and substantial increase in choroidal production of proinflammatory cytokines TNF-a and IL-1b, and neutrophil and monocyte chemoattractants [4,5]. To investigate the possible role of AVP in augmenting the post-traumatic production of proinflammatory mediators by the CP, we used the homozygous AVP-deficient Brattleboro rats. When compared with their parental wild-type Long-Evans strain, Avp di/di rats had reduced levels (by 50-59%) of mRNA for CXC and CC chemokines in the ipsilateral CP at 6 h post-TBI, a time point after injury at which a peak in chemokine synthesis was observed in these two rat strains (Fig. 1). Since TNF-a and IL-1b are well-known inducers of chemokine synthesis, we also assessed the differences in post-traumatic production of these two cytokines in Avp di/di versus wild-type rats. The levels of mRNA for TNF-a and IL-1b in the ipsilateral CP at 6 h post-TBI were lower by 57% and 37%, respectively, in Avp di/di versus wild-type rats; however, these differences did not attain statistical significance because of variations in the levels of cytokine synthesis among individual animals. A significant difference between Avp di/di and wild-type rats was observed in post-traumatic production of LCN2, a proinflammatory mediator whose synthesis in the CP epithelium is highly upregulated in response to inflammatory stimuli [22]. Unlike the chemokine synthesis, the post-traumatic production of LCN2 in the ipsilateral CP of wild-type rats increased gradually over the period of 48 h post-TBI, with the levels of LCN2 mRNA in wild-type rats exceeding those observed in AVP-deficient rats by 19-fold.

Vasopressin acts synergistically with TNF-a and IL-b to increase the production of proinflammatory mediators by the choroidal epithelium
We next investigated the ability of AVP to stimulate the chemokine production in the CP epithelial cell cultures. For these experiments, the CP epithelial cell line Z310 was employed. Using RT-PCR, we confirmed the presence of the AVPR1A in the Z310 cells. The AVPR1B or AVPR2 was not expressed in these cells. The effect of AVP on CXCL1 synthesis was examined, with the AVP concentration ranging between 10 -10 and 10 -6 M. After 1-h incubation, the maximum effect was observed at a peptide concentration of 10 -7 M, but AVP produced only a moderate (6-fold) increase in CXCL1 mRNA. Therefore, we tested the hypothesis that AVP acts synergistically with proinflammatory cytokines to amplify the cytokine-dependent production of chemokines. A significantly larger increase in CXCL1 mRNA was observed after the Z310 cells were exposed to a combination of TNF-a and AVP compared to the increase in CXCL1 synthesis seen in these cells after exposure to TNF-a alone ( Fig. 2A). Similarly, AVP augmented the IL-1b-dependent increase in expression of CXCL1, but the levels of CXCL1 mRNA were substantially higher than those found in response to TNF-a or a combination of TNF-a and AVP (Fig. 2B).
Despite the higher levels of CXCL1 mRNA observed in response to IL-1b or a combination of IL-1b and AVP, the amplification of the TNF-a-induced CXCL1 synthesis by AVP was more robust than that of the IL-1b-induced production of this chemokine (10 times for TNF-a at 5 ng/ml versus 3 times for IL-1b at 0.5 ng/ml with AVP at 10 -7 M). Accordingly, in subsequent experiments, the synergistic interactions between TNF-a (5 ng/ml) and AVP (10 -9 M) were investigated. Vasopressin not only augmented the TNF-a-dependent increase in mRNA for CXCL1 but also amplified the choroidal expression of other neutrophil chemoattractants, such as CXCL2 and CXCL3, and LCN2 (Fig. 3A). Compared to other CXC chemokines studied, the production of CXCL1 was the most affected by AVP (augmented 5X by AVP), but this AVP effect was also short lasting compared to that exerted by AVP on expression of other proinflammatory mediators. The synergistic interactions between TNF-a and AVP also resulted in increased release of proinflammatory mediators into the culture media (Fig. 3B). Among the CXC chemokines studied, CXCL3 was produced at the highest rate. However, this chemokine appears to be a weaker neutrophil chemoattractant compared to CXCL2 [23], which was produced at the lowest rate.
Vasopressin amplifies the cytokine-dependent production of proinflammatory mediators by augmenting the activation of JNK To identify the signal transduction pathways mediating AVP action on the choroidal epithelium, we used selective inhibitors of JNK, MEK1, and p38. In the cultures of the Z310 cells, the inhibition of JNK completely abolished an increase in CXCL1  1 mM) activity, or to interfere with ERK (MEK1 inhibitor PD-98059; 10 mM) or NF-kB (SN50; 100 mg/ml) signaling, the Z3110 cells were pre-incubated for 1 h with respective inhibitors. These inhibitors were also added to the culture media when the cells were exposed to TNF-a (5 ng/ml) or a combination of TNF-a (5 ng/ml) and AVP (1 nM) for an additional hour. Note that the selective JNK inhibitor SP600125 completely abolished the induction of CXCL1 synthesis observed in response to TNF-a alone or to a combination of TNF-a and AVP. A selective MEK1 inhibitor PD-98059 partially (by .50%) inhibited CXCL1 synthesis, whereas p38 inhibitor SB203580 had no effect. In comparison, the inhibition of NF-kB signaling resulted in the amplification of chemokine synthesis. *p,0.05, **p,0.01 for treatment versus control. { p,0.05, {{ p,0.01 for the incubation with inhibitor versus the incubation without inhibitor (n = 3 per group). doi:10.1371/journal.pone.0079328.g004 expression observed in response to either TNF-a or a combination of TNF-a and AVP (Fig. 4). The inhibition of ERK signaling only partially reduced the production of CXCL1, whereas the inhibition of p38 had no effect on chemokine synthesis. Interestingly, the inhibition of NF-kB signaling actually increased the synthesis of CXCL1 by the choroidal epithelium.
To confirm the results obtained with the pharmacological inhibition of JNK, we assessed changes in activation of JNK and its target transcription factors c-Jun and ATF2 in response to TNF-a or a combination of TNF-a and AVP using the JNK activity assays and Western blotting (Fig. 5). When added to the culture media at a concentration of 10 -9 M, AVP by itself had no effect on JNK activity. However, it significantly augmented the TNF-a-dependent activation of this kinase. Similarly, AVP significantly increased the activation of c-Jun and ATF2 occurring in response to TNF-a, while on its own it had no effect on the level of phosphorylation of these transcription factors.
Vasopressin plays an important role in augmenting the post-traumatic activation of JNK in the CP In wild-type rats, TBI resulted in a rapid (within 6 h post-injury) activation of JNK in the ipsilateral CP, which was determined by Western blotting with phospho-specific anti-JNK antibodies and the JNK activity assays (Fig. 6). The extent of post-traumatic activation of JNK in the ipsilateral CP was considerably attenuated in Avp di/di rats, suggesting the important role of this neuropeptide in the activation of this signaling pathway after injury.
In vivo inhibition of JNK decreases the post-traumatic production of neutrophil chemoattractants by the CP and reduces the influx of neutrophils across the BCSFB To inhibit the activity of JNK in wild-type rats sustaining injury, a selective JNK inhibitor SP600125 was administered at 4 and 8 h post-TBI. The efficacy of this pharmacological inhibition of JNK was judged by the extent of reduction in post-traumatic activation of ATF2 in the ipsilateral CP. Another target transcription factor for JNK, c-Jun, was undetectable in the CP. The administration of SP600125 resulted in a significant reduction in the level of posttraumatic activation of ATF2 in the ipsilateral CP when assessed at 6 h after injury (Fig. 7A). Consequently, a decrease in posttraumatic production of neutrophil chemoattractants in the ipsilateral CP was also observed (Fig. 7B). To determine the effect of JNK inhibition on the magnitude of post-traumatic influx of neutrophils across the BCSFB, we employed the MPO assays. The influx of neutrophils was assessed at 24 h post-TBI, a time point after injury at which the maximum accumulation of these inflammatory cells in the ipsilateral CP was seen [4]. Consistent with the reduction in post-traumatic production of neutrophil chemoattractants observed after the pharmacological inhibition of JNK, we also found a reduction (by 79%) in the magnitude of neutrophil influx across the BCSFB in animals treated with SP600125 versus those receiving vehicle (Fig. 7C). When we compared wild-type rats with AVP-deficient animals, we also observed a significant reduction (by 62%) in the magnitude of post-traumatic influx of neutrophils across the BCSFB (Fig. 7C). These observations support an important role of AVP in augmenting the post-traumatic invasion of neutrophils across the BCSFB.

Discussion
In this study, we have demonstrated the synergistic interactions between proinflammatory cytokines and AVP, a ligand for G protein-coupled receptors. By interacting with TNF-a and IL-1b, AVP amplifies the cytokine-dependent production of proinflammatory mediators by the CP, consequently exacerbating the brain inflammatory response to injury. This AVP action is mediated by its cognate AVPR1A receptors. We have confirmed by RT-PCR that the AVPR1B or AVPR2 is not expressed in the CP. Our observations challenge the traditional view of AVP as a pressor/ antidiuretic hormone and support a new role for AVP as a regulator of post-traumatic neuroinflammation. Clinical studies have shown that TBI increases the levels of circulating AVP [24]. There is also evidence based on studies of experimental TBI that brain injury results in a rapid increase in AVP concentration in the CSF [25]. We have previously shown that the AVPR1A is localized to the apical (CSF-facing) surface of the choroidal epithelium [17]. This suggests that AVP released into the CSF after injury augments the post-traumatic production of proinflammatory mediators by the CP.
When compared with AVP-deficient rats, wild-type animals had increased synthesis of neutrophil chemoattractants in the ipsilateral CP after neurotrauma. This was associated with the larger magnitude of post-traumatic influx of neutrophils across the BCSFB in wild-type rats compared to Avp di/di rats. In both in vivo and in vitro studies, a short-lasting (for up to 6 h) increase in mRNA for neutrophil chemoattractants was only observed. However, studies in the Z310 cells demonstrated a prolonged (lasting for at least 24 h) increase in production of chemokine proteins, with AVP having the most significant effect on TNF-a-dependent chemokine synthesis at later time points during the 24-h experimental period. This may explain why a peak in the magnitude of post-traumatic influx of neutrophils across the BCSFB is observed at 24 h after injury, and a very limited accumulation of these inflammatory cells in the ipsilateral CP is seen at 6-8 h post-TBI [4].
The synergistic interactions between proinflammatory cytokines and AVP are predominantly mediated by the JNK signaling pathway. In vivo and in vitro studies indicated that AVP by itself has little effect on JNK activity in the CP. However, AVP could effectively enhance JNK activation induced by cytokines and, possibly, other factors involved in the initiation and propagation of inflammatory cascade in the injured brain. Previous studies have demonstrated that the JNK signaling pathway plays a significant part in regulating the synthesis of neutrophil and monocyte chemoattractants [26][27][28], which supports the mediatory role of JNK in AVP action. Interestingly, we have found that NF-kB plays a negative regulatory role in AVP-dependent amplification of choroidal synthesis of chemokines. When the selective NF-kB inhibitor was added to the Z310 cell culture media, a larger increase in chemokine synthesis in response to TNF-a or a combination of TNF-a and AVP was observed. Our findings are consistent with previous observations that NF-kB deficiency prolongs the TNF-a-dependent activation of JNK [29]. These previous studies have also demonstrated that TNF-a induces the production of reactive oxygen species (ROS) that inhibit JNKinactivating phosphatases, and NF-kB acts to reduce the TNF-adependent production of ROS. There is evidence that AVP has also the ability to induce the production of ROS [30]. It is thus possible that AVP augments the cytokine-dependent synthesis of proinflammatory mediators by inhibiting JNK-inactivating phosphatases.