Effects of Brain-Derived Neurotrophic Factor on Local Inflammation in Experimental Stroke of Rat

This study was aimed to investigate whether brain-derived neurotrophic factor (BDNF) can modulate local cerebral inflammation in ischemic stroke. Rats were subjected to ischemia by occluding the right middle cerebral artery (MCAO) for 2 hours. Rats were randomized as control, BDNF, and antibody groups. The local inflammation was evaluated on cellular, cytokine, and transcription factor levels with immunofluorescence, enzyme-linked immunosorbent assay, real-time qPCR, and electrophoretic mobility shift assay, respectively. Exogenous BDNF significantly improved motor-sensory, sensorimotor function, and vestibulomotor function, while BDNF did not decrease the infarct volume. Exogenous BDNF increased the number of both activated and phagocytotic microglia in brain. BDNF upregulated interleukin10 and its mRNA expression, while downregulated tumor necrosis factor α and its mRNA expression. BDNF also increased DNA-binding activity of nuclear factor-kappa B. BDNF antibody, which blocked the activity of endogenous BDNF, showed the opposite effect of exogenous BDNF. Our data indicated that BDNF may modulate local inflammation in ischemic brain tissues on the cellular, cytokine, and transcription factor levels.


Introduction
Stroke is a major cause of death and long-term disability worldwide [1,2]. Brain-derived neurotrophic factor (BDNF) can decrease infarct volume and improve neurological outcome either by exogenously supplied or overexpression in vivo using genetic methods in experimental stroke. Inhibition of BDNF exaggerates damage of ischemia. BDNF exerts neuron protection against ischemic injury through binding to two membrane receptors, p75 neurotrophin receptor and tyrosine kinase receptor B (trkB) [3]. 7,8-dihydroxyflavone as a bioactive high-affinity TrkB agonist also protects neurons from apoptosis and decreases infarct volumes in animal model of stroke [4].
Inflammation plays an essential role in the pathogenesis of ischemic stroke [5,6]. Rapid activation of resident inflammatory cells (mostly microglia), productions of inflammatory cytokines such as interlerukin10 (IL-10) and tumor necrosis factor α (TNF-α) and translocation of intercellular transcription factors such as nuclear factor-kappa B (NF-κB) are characters of local inflammatory responses to ischemia in brain [7][8][9][10][11]. Different responses may have different functions in the pathogenesis of stroke. Activated microglia could excrete neurotrophic effects such as BDNF to alleviate ischemic injury and exhibit phagocytic activity disposing of degenerating elements [7,11]. There are several cytokines involved in inflammatory process. TNF-α, as an important proinflammatory cytokine, appears to exacerbate cerebral injury of ischemia [8] while IL-10, an anti-inflammatory cytokine, ameliorates ischemic insult of brain [9]. Activation of NF-κB, an important transcription factor, could mediate translation of many downstream genes and promote survival of neurons [10].
BDNF promotes cell proliferation, increases phagocytic activity and inhibits apoptosis of microglia in brain [12]. BDNF downregulates the expression of TNF-α and upregulates the expression of IL10 in the model of multiple sclerosis [13]. NF-κB activated by BDNF protects cells from damages, such as the serum starvation and glutamate toxicity [14]. However, whether BDNF modulates inflammatory processes in ischemic stroke is unclear. In present study, we evaluated the effect of BDNF on local cerebral inflammatory process on 2 Mediators of Inflammation  Food and water were provided ad libitum. All efforts were made to minimize the number of animals used and their sufferings.

Experimental Groups.
The rats used in the study were randomized into control (n = 18), only vehicle was given, BDNF group (n = 18), BDNF was given, and antibody group (n = 18), BDNF antibody was induced. The procedure was shown in Table 1.

Administration of Drugs.
Rats were anesthetized with pentobarbital sodium (40 mg/kg i.p., Sigma-Aldrich, USA) and placed in a stereotaxic frame. Stereotaxic injections were made by Hamilton syringe using the following coordinates: 0.5 mm rostral to bregma, 3.5 mm lateral to midline, and 5.5 mm ventral to the skull surface. Rats in control group were given 10 μL phosphate buffered saline (PBS, pH 7.4); rats in antibody group were given 5 μg BDNF antibody (PeproTech, USA) diluted in 10 μL PBS (pH 7.4); rats in BDNF group were given 10 μg BDNF (PeproTech, USA) diluted in 10 μL PBS (pH 7.4). At the end of injection, the needle was left in place for 5 minutes before being slowly withdrawn.

Transient Middle Cerebral Artery Occlusion (MCAO)
and Reperfusion. The MCAO procedure has been described before [15]. In brief, rats were anesthetized with pentobarbital sodium (40 mg/kg i.p.). Carotid artery was exposed and a 4-0 silicone-coated nylon filament was gently advanced from external carotid artery into the lumen of internal carotid artery until the rounded tip blocked the origin of the middle cerebral artery. A laser Doppler flow meter (LDF; Perimed PF5000, Stockholm, Sweden) was used to confirm the decrease of the middle cerebral artery blood flow immediately after the occlusion to about 20% of the basic cerebral blood flow. After 2 hours, rats were briefly reanesthetized and the filament withdrawn. Rectal temperature was maintained at 37 • C using a rectal probe and heating pad during the surgery. Physiological parameters were monitored pre-and during ischemia (

Behavioral Testing.
Following recovery from anesthesia, behavioral neurologic deficits were assessed 24 h after MCAO. The battery consisted of four tests to systematically evaluate motor, sensory, and vestibulomotor deficits. The following behavioral tests were performed. Postural reflex and hemiparesis test as described before [17]: (0) no observable neurologic deficits, (1) left forepaw flexion, (2) decreased resistance to lateral push and forepaw flexion without circling, (3) decreased resistance to lateral push and forepaw flexion with circling, and (4) cannot walk spontaneously. Forepaw placing test was done as described previously [18]. For each test, limb placing scores were as follows: (0) immediate and complete placing, (1) delayed and/or incomplete (>2 s), and (2) no placing.
Modified beam balance test was done as described [19]: The scale was as follows: (1) steady posture with paws on top of the beam, (2) paws on side of the beam or wavering, (3) one or two lim(s) Slip off, (4) three limbs slip off, (5) attempts with paws on the beam, but falls, and (6) drapes over the beam, then falls or falls with no attempt.
2.9. Enzyme-Linked Immunosorbent Assay (ELISA). At 6 h and 24 h of reperfusion, 6 rats of each group were sacrificed and brain homogenates were obtained from the ischemic hemisphere. The concentrations of BDNF, IL-10 and TNFα in brain homogenates were measured using specific ELISA kits according to the manufacturers' instructions (R&D system, USA).

Real-Time Quantitative PCR.
Total RNA was isolated from frozen brain tissues using the TRIzol reagent (Invitrogen, USA) according to the manufacturer's recommendation, and subjected to DNase (Promega, USA) treatment. Reverse transcription (RT) reaction was carried out using the Firststrand cDNA synthesis kit (Takara, Japan) according to the manufacturer's instructions. Obtained cDNA were amplified using the following primers: for TNF-α, 5 -GCATGA-TCCGAGATGTGGAA-3 and 5 -AGACACCGCCTGGAG-TTCTG-3 , for IL-10, 5 -CCTTACTGCAGGACTTTAAGG-GTTA-3 and 5 -CTGGGCCATGGTTCTCT-3 , and for βactin, 5 -GACAGGATGCAGAAGGAGATTACT-3 and 5 -TGATCCACATCTGCTGGAAGGT-3 . The amplification and data acquisition were run on a real time PCR system (Bio-Rad, USA) using SYBR green PCR Master Mix (Takara, Japan). The conditions were predenaturation at 95 • C for 10 minutes, followed by 40 cycles at 95 • C for 15 seconds and 55 • C for 1 minute. All samples were analysed in triplicates in three independent experiments. Reactions without cDNA were used as no template control and no RT controls were also set up to rule out genomic DNA contamination. Gene expression levels were calculated using the 2 −ddCt method [21].

Electrophoretic Mobility Shift Assay (EMSA).
Nuclear extracts were prepared by hypotonic lysis followed by high salt extraction. EMSA was performed using a kit (Gel Shift Assay System, Promega, USA) to assay NF-κB DNA-binding activities. The NF-κB oligonucleotide probe, 5 -AGTTGAGGGGACTTTCCCAGGC-3 , was end-labeled with [γ-32 P]. Protein-DNA binding assays were performed with 50 μg of nuclear protein. The binding medium contained 4% glycerol, 1% NP40, 1 mM MgCl 2 , 50 mM NaCl, 0.5 mM EDTA, 0.5 mM DTT, and 10 mM Tris/HCl (pH 7.5). In each reaction, 20000cpm of a radiolabeled probe was included. Samples were incubated at room temperature for 15 minutes, and nuclear protein with 32 P-labeled oligonucleotide complex was separated from free 32 P-labeled oligonucleotide by electrophoresis through a 4% native polyacrylamide gel in 0.5 × TBE. After separation was achieved,   The concentration of BDNF in brain one hour before ischemia, 5 μg of BDNF antibody in 10 μL PBS was injected to the brain of rats. 10 μg BDNF in 10 μL PBS was given immediately after ischemia. At 6 h and 24 h of reperfusion, rats were sacrificed and brains were removed. The concentration of BDNF was measured after reperfusion using ELISA kits. The level of BDNF was significantly increased in BDNF group both 6 h and 24 h after reperfusion. Bars represent mean ± SD (n = 6). * P < .05 versus control group, # P < .05 versus BDNF group. the gel was dried (80 • C, 30 minutes) and exposed to X-ray film (Fuji Hyperfilm) at −80 • C with an intensifying screed.

Cell
Counting. Images from brain sections with immunofluorescence were captured and analyzed by using Image-Pro Plus 6.0. Cells were counted in the ischemic cortex. 5 slides, 2 mm interval from bregma, were used for counting in each rat. 6 random areas of each slide were involved and the average number of OX-42 or ED1-positive cells/mm 2 was calculated.

Statistical
Analysis. The data were presented as mean ± SD. Differences between groups were compared using analysis of variance (ANOVA) followed by post hoc ttest with SPSS13.0 software (USA). P < .05 was considered to be statistically significant.

BDNF Level was Increased in Brain.
To assess the concentration of BDNF in brain tissues, the brain homogenates were obtained 6 h and 24 h after reperfusion and the concentration of BDNF was measured using ELISA kits. BDNF level was significantly increased by 19.7 fold in BDNF group while administration of BDNF antibody did not influence BDNF level in brain (1.00 ± 0.05 versus 0.98 ± 0.13 ng/g, n = 6, P > .05, Figure 1). The concentration of BDNF was decreased at 24 h of reperfusion than that at 6 h of reperfusion; however, it was significantly increased compared with control group (11.36 ± 0.91 versus 1.42 ± 0.08 ng/g, n = 6, P < .05, Figure 1). There was no significant difference of BDNF level between antibody group and control group (1.38 ± 0.14 versus 1.42 ± 0.08 ng/g, n = 6, P > .05, Figure 1).

BDNF Promoted Anti-Inflammatory Cytokine Expression.
IL10 is a well-known anti-inflammatory cytokine. IL-10 and its mRNA expression were tested by ELISA kits and real time qPCR, respectively. When BDNF antibody blocked activity of endogenous BDNF in brain, IL10 was markedly decreased compared to control group (10.85 ± 0.48 versus 14.28 ± 0.82 ng/g, n = 6, P < .05, Figure 4(a)). Application of BDNF antibody also prevented upregulation of IL10 mRNA 6 h after reperfusion (0.50 ± 0.02 versus 1.01 ± 0.09, n = 6, P < .05, Figure 5(a)). However, there was no significant difference of IL10 and its mRNA expression between BDNF group and control group 6 h after reperfusion. Exogenous BDNF markedly increased local IL10 level in brain tissues 24 h after reperfusion (19.80 ± 0.83 versus 13.31 ± 0.36 ng/g, n = 6, P < .05, Figure 4(a)) and increased mRNA expression by 1.93 fold compared with control group. However, BDNF antibody did not influence the level of IL10 and its mRNA 24 h after reperfusion.

BDNF Increased DNA-Binding Activity of NF-κB after
Stroke. The DNA-binding activity of NF-κB was activated by many stresses such as ischemia, the activity was measured using EMSA and expressed as arbitrary densitometric units (AU). Exogenous BDNF increased DNA binding activity of NF-κB by 10.7% compared with control group 6 h after reperfusion, and when endogenous BDNF was suppressed by BDNF antibody, the activity was significantly decreased (35.92 ± 0.99 versus 39.97 ± 0.70, n = 6, P < .05, Figure 6). At 24 h of reperfusion, exogenous BDNF could significantly increase DNA bind activity compared with control group (45.38 ± 0.86 versus 42.46 ± 0.27, n = 6, P < .05, Figure 6).  Figure 5: BDNF modulating local cytokine in brain after stroke Interleukin10 (IL10) and tumor necrosis factor α (TNF-α) in rat brain after stroke were measured using ELISA kits. (a) Exogenous BDNF upregulated IL-10 24 h after reperfusion while BDNF antibody decreased IL-10 6 h after reperfusion. (b) Exogenous BDNF decreased TNF-α at 6 h and 24 h of reperfusion than control group and BDNF antibody overwhelmed the effect. Bars represent mean ± SD (n = 6). * P < .05 versus control group, # P < .05 versus BDNF group.

Discussion
Our data suggested that BDNF could alleviate cellular injury of ischemic insult, reduce the neurologic deficits and modulate local inflammation on cellular, cytokine and nuclear factor levels in brain after stroke. Exogenous BDNF could increase the number of activated and phagocytotic microglia, upregulate IL-10, downregulate TNF-α and increase the DNA-binding activity of NF-κB while BDNF antibody blocked these effects of BDNF in ischemic brain tissues.
Firstly, we showed that introducing BDNF directly to brain increased the concentration of BDNF. We also found that BDNF antibody may not change the expression of BDNF in brain tissues after stroke. Our data confirm previous reports that injection of BDNF directly to brain could raise concentration of BDNF.
Secondly, exogenous BDNF could protect brain from ischemic injury and reduce the neurologic deficits. We found exogenous BDNF significantly decreased the number of TUNEL-positive cells. We also found that BDNF improved the motor-sensory function, sensorimotor function, and vestibulomotor function. Our results were consistent with previous reports [22,23]. However, BDNF may not reduce the infarct volume which was different from some previous researches [3,24]. Schäbitz et al. [22] also found that BDNF may not change the infarct volume.
Thirdly, we provided evidences that BDNF modulated local inflammation in ischemic brain tissues on cellular, cytokine and transcript levels. On cellular level of local inflammation, we found that exogenous BDNF could increase the number of both activated and phagocytotic microglia as early as 6 h after reperfusion and lasted at least for 24 h. When the activity of endogenous BDNF was blocked by BDNF antibody, the number of activated and phagocytotic microglia was decreased. Previous reports showed that BDNF could promote microglial proliferation and phagocytic activity in vitro and in vivo and inhibit microglial apoptosis [11,12]. These results were consistent with our data. The mechanism BDNF activates microglia might be that BDNF could sustain elevation of intracellular Ca 2+ of microglia [25,26]. It is well accepted that activated microglia protects the brain against ischemic and excitotoxic injury [27]. Once microglia activated, it could excrete neurotrophins such as BDNF. BDNF could protect neurons against ischemia. More BDNF could activate more microglia and it is an autoloop. Phagocytotic microglia could engulf damaged cells and other inflammatory cells which may initiate more damage to brain in ischemic stroke. These may partly explain the protection of BDNF against ischemic insult.
On cytokine level of brain local inflammation in ischemic stroke, we found that BDNF could decrease brain TNFα in ischemic stroke. Exogenous administration of BDNF decreased brain TNF-α and inhibited the mRNA expression of TNF-α. When the activity of BDNF was blocked by its antibody, the protein level and mRNA expression of local TNF-α in brain were decreased. These results were consistent with previous research, which shows that exogenous BDNF inhibits the expression of TNF-α in mouse brain of model of multiple sclerosis [13]. TNF-α, an important proinflammatory cytokine, plays a vital role in the pathophysiology of ischemia stroke [6]. Exogenous administration of TNF-α exacerbates ischemic brain injury while inhibition of TNFα could reduce brain damage [8,28]. BDNF could provide protection of brain in ischemic stroke via decreasing local TNF-α.
BDNF not only decreased local proinflammatory cytokine, it also increased local anti-inflammatory cytokine. IL10 is an important anti-inflammatory cytokine. Our data showed BDNF increased the level of IL10 and upregulated the mRNA expression of IL10 at 24 h of reperfusion. Once activity of endogenous BDNF was blocked by BDNF antibody, local IL-10 and its mRNA in brain were increased. These results were consistent with other studies [13]. Previous publications showed that exogenous pre/postischemic administration of IL10 can provide neuroprotection following MCAO [9,29]. Over-expression of IL10 in vivo markedly protected cortical tissue against cerebral ischemia using the IL10 transgene mice [30]. Our data suggested that BDNF might protect brain from ischemia through upregulating local IL10 in brain.
On the transcription level of local inflammation in brain in ischemic stroke of rats, we found that exogenous BDNF increased the DNA-binding activity of NF-κB. We also provided evidence that the DNA-binding activity of NF-κB was inhibited when the effect of BDNF was suppressed. These results indicated that BDNF could modulate NF-κB's activity. The way BDNF activates NF-κB in different cells is through the TrkB-PI3-kinase-Akt pathway [31]. Inhibition of TrkB, the receptor of BDNF in the cell, decreases the activation of NF-κB [32]. The activation of NF-κB induced by BDNF protects cells from a variety of damages, including the serum starvation, glutamate toxicity and ischemia [14]. Once the activity of NF-κB was inhibited by different inhibitors, the protections of BDNF against different damages were lessened [31]. Our data suggested that NF-κB played an important role in neuron protection of BDNF.
In our study, effect of BDNF on local inflammation in brain showed no significant difference between antibody group and control group 24 h after reperfusion. This may because that BDNF antibody only blocked the activity of BDNF and may not suppressed the expression of local BDNF in brain after stroke. Our data showed that BDNF antibody did not change BDNF level (Figure 1). 24 h after reperfusion (25 h after BDNF antibody was given), the expression of new BDNF may replace the antibody-conjuncted BDNF, so the effect of BDNF antibody might be removed.

Conclusions
In summary, our data suggested that BDNF may alleviate cellular injury of ischemic insult, reduce the neurologic  Figure 7: BDNF enhancing DNA-binding activity of NF-κB The DNA-binding activity of NF-κB in ischemic cortex in three groups was measured using Electrophoretic mobility shift assay (EMSA) 6 h and 24 h after reperfusion. Levels of NFκB DNA binding activity were expressed as arbitrary densitometric units (AU.) using Image J. Exogenous BDNF increased the activity of NF-κB 6 h and 24 h after reperfusion. Inhibition of endogenous BDNF by its antibody significantly decreased the activity 6 h after reperfusion. Bars represent mean ± SD (n = 6). * P < .05 versus control group, # P < .05 versus BDNF group.
deficits and modulate local inflammation on cellular level, cytokine level, and transcription factor level in ischemic stroke.