Mechanisms of Neuroprotection from Hypoxia-ischemia (HI) Brain Injury by Up-regulation of Cytoglobin (Cygb) in a Neonatal Rat Model

Results: Up-regulation of Cygb reduces HI injury and cognitive

cognitive and/or sensory deficits, mental retardation, learning disabilities, and epilepsy (3,4). These impairments significantly impact life experience and social welfare. To date, there are no effective therapies for the treatment of these neurological disorders. Thus, this limitation has provided an incentive to search for new, more effective therapeutic interventions which may lead to better outcomes.
The globins have striking functions, which initially attracted our research interest. Cytoglobin (Cygb), the fourth member of the vertebrate globin family of hemoproteins, is detectable in various organs (5). Hemoglobin (Hgb) and myoglobin (Mb) are abundant hemoproteins that have been extensively studied. Neuroglobin(Ngb), the third heme protein (6), can be induced by neuronal hypoxia and cerebral ischemia, and can protect neurons from the effects of hypoxia in vitro (7,8) and in vivo (8)(9)(10)(11). The most recently discovered androglobin (Adgb), the fifth member of globin family, is preferentially expressed in the testis and is insensitive to experimental hypoxia (12). The physiological function of Cygb remains unknown, but it is believed to have various roles due to its family homology, such as storing O 2 , facilitating O 2 diffusion, detoxifying reactive oxygen species, acting as an O 2 sensor, and functioning as an NO dioxygenase (5,(13)(14)(15)(16). Thus, Cygb may play a cytoprotective role under hypoxic and/or ischemic conditions. Increasing evidence suggests that Cygb may be up-regulated in the liver, heart, brain, muscle, and kidney (16)(17)(18)(19) under hypoxic and/or ischemic conditions, suggesting that Cygb might be a protective factor in these organs. Previous studies have also indicated that Cygb acts as a stress-responsive hemoprotein expressed both in the developing and adult brain (20) and is protective under oxidative stress in cell lines (21,22). More recently, Cygb was found to act as a tumor suppressor gene (23,24), and protect kidney fibroblasts under ischemic conditions (25), which are related to oxidative stress (25-27).
Our hypothesis was that Cygb might protect neonatal rats from the damage associated with HI brain injury. The aims of the present study were to investigate the potential role of Cygb in neonatal HI and clarify the underlying mechanisms, as well as to provide a new therapeutic target for the treatment of neonatal HI brain injury. For the first time, we identified the expression profiles of Cygb mRNA and protein induced by HI using a widely applied animal model. We also provided the first evidence that HI outcomes could be significantly affected under the condition of over-expression and knock-down of Cygb using adenovirus transfection systems in vivo. MDA and SOD assays were also carried out in order to demonstrate the antioxidant ability of Cygb in vivo. In addition, we identified the molecular mechanisms associated with Cygb under HI injury. These findings suggest that Cygb plays an important role in protecting the developing brain against HI injury.
of each group (animals used for studying the endogenous expression profiles of Cygb and confirming the efficiency of ADV transfection reagents were considered as sham) in the first 36 h after treatment were listed in Fig 1C. Log-rank test were used to compare survival curves among groups at 28 days when they were ready for Morris Water Maze (MWM) test.
Reagents-In situ cell death detection kits were purchased from Roche (Roche Applied Science, Mannheim, Germany). The PCR primers (as listed in Table 1), cDNA, and shRNA for Cygb were synthesized by Sangon Biological Engineering Technology & Services Co., Ltd (Shanghai, China). The RT-PCR kits were purchased from Takara Biotechnology Co., Ltd (Dalian, China). Rabbit polyclonal antibody for Cygb , as well as goat polyclonal antibody for CD31 was purchased from Santa Cruz Biotechnology, Inc. (Santa Cruz, CA, USA). Primary monoclonal antibody for β-actin, cleaved Caspase-2 and Caspase-3, and horse radish peroxidase-conjugated secondary antibody were from Cell Signaling Technology, Co., Ltd (USA). The SuperSignal western blotting detection kits was obtained from Pierce Biotech. Co., Ltd (Rockford, IL, USA). Cy3 conjugated anti-rabbit secondary antibody and 4', 6-diamidino-2-phenylindole (DAPI) for immunofluorecence were from Sigma Aldrich (Shanghai, China). Rabbit anti-Vegf and relative Streptavidin-Peroxidase immunohistochemical staining kits were purchased from Biosynthesis Biotechnology Co., Ltd (Beijing, China). All surgical materials used were acquired from Yuehua Medical Products Ltd (Shantou, Guangdong, China). Other chemicals and reagents were of molecular biology grade and were purchased from local commercial stores. Establishment of Animal Model-HI brain injury was induced in P7 SD rats based on the classical "Levine method" (28) with modifications by snipping the left common carotid artery (CCA) between double ligations and prolonging hypoxia duration from 90 min to 120 min. Briefly, each rat was deeply anesthetized by inhalation of isoflurane. The CCA was exposed through a midline cervical incision, permanently double ligated with 5-0 silk sutures, and severed. The total time for surgery in each animal lasted approximately 3 min. After surgery, animals were given 1~2 h to recover from anesthesia. Following recovery, animals were placed in an airtight container partially submerged in a 37 °C water bath to maintain a constant thermal environment. A 120 min continuous hypoxia (8% O 2 , 92% N 2 ) was used to induce systemic hypoxia. Sham animals received anesthesia and exposure of the left CCA but did not receive HI treatment. After another 30 min recovery, all surviving rats were returned to their cages and kept in a standard environment as previously described in the animal housing section.

Construction of Adenovirus-mediated Transfection Systems-shRNA and cDNA of
Cygb sequences were designed based on a protocol used in a previous study (23). We obtained plasmids PDC316-mCMV-Zsgreen-cDNA-Cygb and PDC316-ZsGreen-shRNA-Cygb, with null vectors transfected as negative controls (29-32). The plasmids were cloned to adenovirus vectors and the viral null vectors were propagated in human embryonic kidney 293 (HEK 293) cells. Viral titer was determined by using standard plaque assays on HEK293 cells. The resulting titers for Cygb-shRNA-ADV and Cygb-cDNA-ADV were 3×10 9 PFU/ml and 2×10 9 PFU/ml, respectively.

Intracerebroventricular
Injection of Adenovirus-Adenoviral vectors delivery were conducted on P3 SD rats based on previously described protocols (29,33). Briefly, the rats were anesthetized with isoflurane and mounted onto a SR-6N stereotactic frame (Narishige Scientific Instrument Laboratory, Tokyo, Japan). 4 μL of adenoviral vectors or saline (NaCl solution) were injected into the lateral ventricle (0.8 mm posterior and 1.0 mm lateral in relation to lambda and at a depth of 3 mm from the skin surface of the brain) (33) using a Hamilton syringe with a 10-gauge needle. The vector or saline was injected over a period of 2 min and the needle was left in the place for another 2 min, and then removed slowly over 2 min. In order to avoid the potential effect of tissue damage caused by the injection procedure on the assessment of HI injury, all injections were made into the cerebral ventricle contralateral to the injury hemisphere.
To confirm adenovirus-mediated expression in vivo, ADV vectors were tagged with green fluorescence protein (GFP) and expression was detected using a Zeiss Axio Imager Z1 inverted microscope (Massachusetts, USA) 4 days post-injection. To evaluate temporal change of the expression of Cygb protein after transfection, rats were euthanized at 1, 3, 5, and 7 days after injection of the vectors (n=3), and the cortex and hippocampus were harvested for Western blot analysis.
Immunofluorescence-Neonatal rats were sacrificed 4 days after saline or ADV vectors injection. Three rats were used for each group, and three sham rats were used as control. Targeted brain tissues (3-4 mm) were obtained as indicated in Fig.1 B and then fixed for 18 h in 4% paraformaldehyde in PBS at 4 °C, followed by incubation in 30% sucrose in PBS overnight at 4 °C. Then, the fresh brain tissues were embedded in O.C.T and cut into 25 um-thick sections using a cryostat, and the sections were mounted onto glass slides and stored at -80 °C. The brain sections were fixed with pre-cooled acetone for 15 min at 4 °C and rinsed with PBS for 5 min. The sections were then incubated in PBS buffer containing 0.1% Triton X-100 for 15 min and then blocked with 5% BSA in PBS. Blocked sections were then treated overnight with primary antibody at 4 °C. The primary antibody used was rabbit anti-Cygb diluted (1:100) in PBS containing 1% BSA. The sections were rinsed with washing buffer and then incubated in secondary antibody for 1 h at room temperature in the dark. The secondary antibody used was Cy3 conjugated goat anti-rabbit IgG (1:1000). Samples were washed and stained by DAPI for 10 min at room temperature in the dark and then washed again, mounted with glycerol/PBS, and observed using a Zeiss Axio Imager Z1 inverted microscope (Massachusetts, USA).

Reverse Transcription and Polymerase
Chain Reaction (RT-PCR) -Total RNA was prepared from samples of injured tissues (Fig.1B) using Trizol reagent (Invitrogene, Carlsbad, CA, USA) following the manufacturer's instructions. After establishing HI for 0, 6, 12, 18, 24 and 36 h time points, rats from different groups were euthanized. RT-PCR experiments and results analysis were performed as described previously (34) to evaluate the mRNA expression pattern of Cygb under HI conditions. Then, we conducted HI treatment within specific time windows (the time for reaching the peak of endogenous Cygb mRNA up-regulation) to rats which had been given saline, ADV only, shRNA, or cDNA treatment to reveal expression of the potential candidates, including Hif-1α, Vegf, Caspase-2, and Caspase-3. The sham group served as control. The PCR primers used are shown in Table 1. β-actin was used in each assay to normalize the amount of mRNA.
Western Blotting-To evaluate changes of endogenous expression of Cygb after HI injury, rats were euthanized at 0, 6, 12, 18, 24 and 36 h and the injured hemisphere of the brain (approximately 100 mg, indicated in Fig1 B) was collected. In addition, as mentioned previously, rats were euthanized at 1, 3, 5, and 7 days after receiving injection of the Cygb-shRNA or Cygb-cDNA vectors. P7 rats (4 days after injection) injected with the ADV vector only or with saline served as the ADV vector-only group and control group. Brain tissues were homogenized in cold RIPA lysis buffer (Beyotime, Jiangsu, China) and the protein concentrations were determined by BCA assay kits (Beyotime, Jiangsu, China). Equal amounts of protein (40 μg) were separated on 12% SDS-PAGE and transferred onto polyvinylidene fluoride (PVDF) membrane (pore size, 0.22 μm; Millipore). The membrane was blocked by a 1 h incubation at room temperature in a Tris-buffered saline solution (TBS-T: 20 mM Tris, pH 7.6, 135 mM NaCl, and 0.05% Tween) containing 5% nonfat dry milk, and then incubated with different primary antibodies including anti-Cygb (1:400, Santa Cruz), anti-β-actin(1:1000,CST) overnight at 4 °C. After washing the membrane 3x with TBS-T, Neuroprotective effect of Cygb in a neonatal rat model with HI injury the secondary antibody HRP-labeled goat-anti-rabbit IgG was then added to the membrane according to the vendor's recommendation (1:8000 dilution; CST) and incubated for 1 h at room temperature and then washed again as described previously. The bound antibodies were detected by using SuperSignal western blotting kits. Quantity One Software (v4.5.2, Bio-Rad, Hercules, CA, USA) was used to perform densitometric analysis of western blots. The expression of β-actin was used as a control for equal protein loading.
To reveal the association between caspase activation and Cygb expression, we conducted HI treatment at specific time windows (the time for reaching the peak of endogenous Cygb protein up-regulation) to rats which had been given saline, ADV only, shRNA, or cDNA treatment. The sham group served as control. Samples were harvested at the time point that demonstrated the highest transfection efficiency, which was 24 h post-HI in this study. The activation levels of both cleaved-caspase-2 and cleaved-caspase-3 were studied following the steps described above. The conditions of primary antibodies used were: anti-cleaved-caspase-2 (1:1000), anti-cleaved-caspase-3 (1:1000).
TTC Staining-As previously described (35), 24h after HI induction, 2,3,5-triphenyltetrazolium chloride (TTC) staining was performed (n = 8) to measure the infarct volume. The animals were perfused transcardially with cold saline under deep anesthesia and brains were quickly removed and subsequently imbedded in brain matrix and frozen for 3 min at -80 °C. Each brain was sliced coronally at 2-mm intervals with the matrix. Four sliced sections were then subsequently stained with 1% TTC (w/v) at 37°C for 12 min and then fixed in 4% (w/v) formaldehyde in PBS for 24 h at 4°C. Finally, the brain slice images were captured using a digital camera (SONY DSC-W150, Japan), and the areas of unstained tissue (the infarct areas) were delineated manually using the Image Pro Plus 6.0 software by a person blinded to the treatment groups. The corrected infarct area was calculated as that total area of the infarct area -the edema area (calculated as the ischemic hemisphere area -the contralateral hemisphere area). Corrected infarct volume was then calculated by the sum of the corrected infarct area on each section multiplied by slice thickness. Total corrected infarct volume was expressed as percentage of the contralateral hemisphere area. As this kind of data is usually binomial distributed, an arcsine transformation which can be calculated by using the formula " " should be done in the statistical software first before further analysis.

Nissl
Staining-The targeted brain specimens (Fig. 1B) obtained 24 h after HI were fixed in 4% (w/v) formaldehyde in PBS, paraffin embedded, and sliced in sections of 3 μm thickness. For Nissl staining (n = 8), after dewaxing in xylene and rehydration through graded ethanol, the sections were hydrated in 1% (w/v) toluidine blue at 37 °C for 20 min. After rinsing with double distilled water, they were dehydrated and mounted with permount. 8 slices per brain were used for cell counting. Six fields of each slice in the injured side both of the cerebral cortex and hippocampus (CA1) were chosen randomly at 400 × magnifications to count staining cells. Imaging-Pro-Plus software 6.0 was used to perform quantitative analysis of cell number counts. The mean number of intact neurons in the six views was used for cell counts in each section. The final average number of the eight sections from each sample was used for analysis.
TUNEL Staining-TUNEL staining (n = 8) was performed on paraffin-embedded sections by using the in situ cell death detection kit, according to the manufacturer's instructions. Briefly, sections were deparaffinized in xylene, rehydrated through graded ethanol, rinsed in 3% hydrogen peroxide and treated with proteinase K (20 mg/ml) for 25 min at room temperature. Subsequently, the sections were incubated with the TUNEL reaction mixture for 1 h at 37 . ℃ After washing with PBS, the sections were incubated with coverter-POD for 30 min and then visualized with DAB. Sections were then counterstained with hematoxylin for 3 min and Neuroprotective effect of Cygb in a neonatal rat model with HI injury rinsed under running water. After dehydration in graded ethanol series and transparent in xylene, the brain sections were mounted onto gelatin-coated slides. Apoptotic cell counting was performed in hippocampus (CA1) and cerebral cortex in the hemisphere that was ipsilateral to the injured hemisphere. Cells with yellow-brown granules in the nucleus were considered to be apoptotic cells. In evaluating numeric density, total TUNEL positive stained neurons were calculated in six views under the light microscope with 400 × magnifications. The mean number of apoptotic neurons was used for cell counts in each section. The final average number of apoptotic neurons of eight sections from each sample was used for analysis and the severity of brain damage was evaluated by apoptotic index, defined as the average number of TUNEL-positive neurons.

Histology and Immunohistochemistry-4
paraffin sections from each sample (n = 8) from sham, HI, shRNA, and cDNA group were processed for Hematoxylin and Eosin (HE) for the examination of brain tissue under light microscopy. For Vegf and CD31 staining, briefly, after deparaffinization and redehydration, non-specific endogenous peroxidase activity was blocked by treating sections with 3% hydrogen peroxide in methanol for 25 min. The antigen was recovered by boiling the sections for 15 min in 10mM citrate buffer (pH 6.0). Nonspecific binding was blocked with 1% non-immune serum in PBS for 30 min. The sections were then incubated with anti-Vegf (1:300) or anti-CD31 (to stain endothelial cells) (1:60) overnight at 4 . They were then washed with ℃ PBS, incubated with a biotinylated goat anti-rabbit IgG (1:300) for 1 h at 37 °C, washed and incubated with an avidin peroxidase conjugate solution (1:100) for 1 h. Finally, the sections were developed with diaminobenzidine for 5 min. Negative controls were similarly processed without the primary antibody. Five slides from each brain, with each slide containing 6 fields were digitized under 100 × magnifications by two independent observers who were blinded to the experimental conditions. The density and size of the vasculature were analyzed in a blinded manner in digital images using the Imaging-Pro-Plus software 6.0.
MDA and SOD Assays-Brain samples were taken from the injured hemispheres of rats at 24 h after hypoxia-ischemia induction (n = 8). The MDA level and SOD activity were measured according to the manufacturer's specifications (Beyotime, Jiangsu, China). Briefly, the tissues were homogenated and centrifuged at 2000×g for 10 min at 4 °C. The supernatant was collected to measure the MDA level and SOD activity. The protein content of the supernatant was measured as described in the Western blot section. The MDA level and SOD activity were measured using the thiobarbituric acid (TBA) and xanthine oxidase methods. The absorbance levels of the MDA and SOD test samples were measured using a spectrophotometer (Thermo) at 532 nm and 550 nm. The MDA levels and SOD activities were expressed as nmol/mg protein and U/mg protein.

Morris Water Maze Test-Spatial learning
and memory abilities were evaluated using the Morris Water Maze (MWM) test P28-P33 rats (n = 8; 36). A circular pool (160 cm diameter × 50 cm high) divided into four quadrants was filled with water (22 ± 1°C), and an 12 × 12 cm platform was positioned 1 cm below the water surface in the center of one of the quadrants. Four points on the perimeter of the pool were designated and room lights illuminated the pool. The swimming path of the rats was recorded using a video camera mounted above the center of the pool and analyzed using a video tracking and analysis system (Institute of Medica, Chinese Academy of Medical Sciences). On each training day (P28-P32), the rats received eight consecutive training trials during which the hidden platform was kept in a constant location. A different start location was used on each trial, which consisted of a search followed by a 20 s platform sit. The time to reach the platform was recorded. If a rat could not find the platform within 90 s, it was led to the platform by the experimenter for a 20 s rest. On the fifth day (P33), memory retention was evaluated during a 90 s probe trial carried out 24 h after the last training session in the absence of the escape platform. The test parameters included time to reach the platform, swimming track, swimming distance of the platform space, and traversing times of the platform.
Statistical Analysis-Each experiment was performed at least three times. Statistical analysis was performed using GraphPad Prism software, Version 5.01 (San Diego, CA, USA). Continuous data were expressed as mean ± standard deviation (SD). One-way analysis of variance (ANOVA) followed by Student-Newman-Keuls's test was used to compare the differences among multiple groups. Log-rank test were done to compare the survival curves among groups. Data from Morris Water Maze tests were analyzed using the general linear models repeated-measures analysis of variance. A probability (P) value less than 0.05 was regarded as statistically significant.

RESULTS
Survival-In this study, 535 neonatal rats were used and the total survival rate of animals was 87.4%. Pups' survival rate appeared unaffected by pretreatment before HI, indicating that vehicle injection did not significantly affect pups' survival. The survival rates among HI, saline-HI, and ADV-HI groups were not significantly different (Fig.1, C) 36 h after HI induction. However, the survival rate of the Cygb-shRNA-ADV-HI group was 70.4% while the Cygb-cDNA-ADV-HI group was 89.8%. This indicated that over-expression of Cygb increased the survival rate of animals whereas inhibition of Cygb expression decreased survival. In addition, the 28-day survival curves also showed consistent result (Fig. 1 D). Over-expression of Cygb improved the survival of rats suffering from HI.

Cygb is Induced by HI in vivo-
We evaluated the changes of endogenous Cygb expression in the brain after HI in the P7 rat pups (Fig. 2). Cygb mRNA levels were measured by RT-PCR at 0, 6, 12, 18, 24 and 36 h time points after HI. As shown, the transcriptional level of Cygb increased in a time-dependent manner 18 h after HI (Fig. 2, A  and B). The protein level of Cygb was detected by Western blotting at 0, 6, 12, 18, 24 and 36 h after HI. Endogenous Cygb expression in the ipsilateral hemisphere significantly increased from 6 h time point post-HI and gradually peaked at 24 h after HI (Fig. 2, C and D). Cygb was elevated nearly 3-fold as compared with sham group at the 24 h time point. Therefore, Cygb may be involved in the pathogenesis process of HI.

Adenovirus-mediated Transfection Alters
Expression of Cygb in vivo-ADV vectors were tagged with green fluorescence protein (GFP) and expression was measured in the periventricular area (Fig.3E). Cygb protein level was detected on day 1, 3, 5, and 7 after injection in P3 rats (Fig. 3A-D). Expression of Cygb peaked at 5 days post-injection (Fig. 3, B and D). In order to further assess whether the transfection systems were down-regulated or up-regulated, Cygb expression were examined in region of interest (ROI) using immunofluorescence co-localization analysis 4 days post saline or ADV vector injection (Fig.  3F). Although the subcellular distribution of Cygb was not thoroughly assessed, we found that Cygb was localized in both the cytoplasm and nuclei of hippocampus neurons (Fig. 3F), a finding that is consistent with the distribution pattern previously reported (15). The results showed that Cygb-shRNA-ADV and Cygb-cDNA-ADV exhibited significant effects in regulating Cygb expression of brain, which is consistent with the western blotting data (Fig. 3, A-D). Results from saline or ADV only group indicated that there was no effect on Cygb expression. Together with these results and the expression profiles of Cygb mRNA and protein shown above, animal HI model was induced at day 4th post injection for subsequent assays in this study.

Adenovirus-mediated Transfection Alters the Expression of Cygb in the Developing
Brain 24 h Post-HI-To confirm how different levels of Cygb expression affects the outcomes of neonatal HI brain damage, we assessed the Cygb mRNA and protein levels in developing brain tissues from different groups 18 h or 24 h post-HI. Western blotting revealed that Cygb protein expression was significantly lower in the Cygb-shRNA-ADV group as compared to the sham, ADV only, and saline groups, and expression levels were higher in the Cygb-cDNA-ADV group as compared to sham group (Fig. 4, B and D). In addition, RT-PCR analysis also exhibited that Cygb-shRNA-ADV inhibited Cygb mRNA expression while Cygb-cDNA-ADV enhanced its expression as compared to ADV only or saline groups 18 h post-HI (Fig. 4, A and C). As compared to the sham group, expression levels of Cygb in HI-ADV and HI-saline groups were significantly higher, indicating that HI could induce Cygb expression (Fig. 4, A-D). Together, these data show that Cygb expression was mediated by exogenous Cygb-cDNA-ADV and Cygb-shRNA-ADV despite endogenous up-regulation in the neonatal HI model.

Cygb Reduces Injury in the Neonatal HI
Model-Rats from different groups were sacrificed 24 h post-HI and morphological injury was assessed by TTC staining, Nissl, and TUNEL staining. Fig. 5 shows images of neonatal rat brain.
Quantitative assessment of TTC-stained sections was used to assess brain infarct volume in different groups (Fig. 5A). Our findings showed that injection with Cygb-shRNA-ADV dramatically increased the infarct volume at 24 h after HI injury (p < 0.01) (Fig. 5B). Injection with Cygb-cDNA-ADV resulted in a significant decrease in brain infarct volume as compared with the ADV only group or the saline group (p < 0.01). There was no difference between saline, ADV only pretreatments, and untreated HI group, yet all HI groups with or without pre-injection of different reagents were significantly different as compared to the sham control group (p < 0.05).
Neuronal cell loss in both the ipsilateral cortex and hippocampus were identified by Nissl staining 24h post-HI injury (Fig. 5C). As expected, HI significantly reduced the number of cells and resulted in smaller and irregularly arranged neurons in the brain 24h after HI as compared with the contralateral part or the sham control group (Fig. 5E). Furthermore, inhibition of Cygb significantly reduced the number of neuronal cells under HI as compared with saline or ADV only pretreated group (p < 0.01). However, more Nissl-stained cells were observed in Cygb over-expression group than in the other HI group (p < 0.01).
TUNEL staining of tissue sections was carried out to determine if Cygb could prevent apoptosis at an acute stage of neonatal HI injury. Significantly more apoptotic cells were observed in the Cygb-shRNA-ADV group compared with control HI groups 24 h post-HI (p < 0.01; Fig. 5, D and F); however, the number of cells was significantly lower in the Cygb-cDNA-ADV treatment group (p < 0.01). There was no significant difference in the number of TUNEL-positive cells in the cortex and hippocampus across three control HI groups while HI induced greater number of apoptotic neurons as compared with sham group (p < 0.05).

Cygb Exhibits Anti-oxidant Action in the HI Model-
To evaluate the anti-oxidant activity of Cygb, we measured the MDA level and SOD activity 24 h after HI insult (n = 8).
These findings indicate that inhibition of Cygb significantly increased MDA levels and decreased SOD activity whereas Cygb over-expression significantly decreased MDA levels and increased SOD activity as compared to the HI group (p < 0.01; Fig. 6A). There was no difference between HI, HI-saline, and HI-ADV groups, although all three groups demonstrated increased MDA levels and decreased SOD activity as compared with the sham group (p < 0.01).

Cygb Plays a Critical Role in Antiapoptosis
and Promotes Angiogenesis-Transcriptional levels of Caspase-2, Caspase-3, Hif-1α, and Vegf in HI brain tissue (n = 3) were measured by RT-PCR 18 h after HI in three different groups (HI, HI-Cygb-shRNA-ADV, and HI-Cygb-cDNA-ADV; Fig. 6, B and C). Cygb was inhibited by Cygb-shRNA-ADV but enhanced by Cygb-cDNA-ADV (Fig. 6B). Hif-1α and Vegf are protective factors in HI, whereas Caspase-2 and Caspase-3 play an important role in apoptotic pathways. In the present study, we found that Cygb-cDNA-ADV could induce Vegf mRNA expression but inhibit Caspase-2 and Caspase-3 mRNA levels as compared to the HI groups (p < 0.01; Fig. 6C). These results are consistent with previous studies (37)(38)(39) and suggest that these three factors are critical for understanding the mechanism underlying the neuroprotective effect of Cygb in neonatal HI.
Furthermore, to confirm whether the effect of Cygb overexpression/knockdown on Vegf, Caspase-2, and Caspase-3 is specific to HI conditions, we conducted a RT-PCR assay. RNA was extracted from samples harvested five days after Cygb-shRNA-ADV and Cygb-cDNA-ADV injection (Fig. 6D). The results indicated that Cygb expression did not significantly affect the expression of these three genes under basal conditions (Fig. 6E).
HE staining revealed that cell morphology and microvessels were disrupted (Fig. 6F), and immunostaining indicated that higher levels of Vegf protein were expressed in the Cygb over-expression group (Fig. 6, G and H). Immunohistochemical staining of CD31 revealed that a smaller number of microvessels were distributed in the shRNA group, whereas more microvessels were distributed in the cDNA group compared to the HI group (Fig. 6, I and J), similar to the number of microvessels with a diameter larger than 15 µm per section (Fig. 6, I and K).
Western blotting showed that HI induced both cleaved-Caspase-2 and cleaved-Caspase-3 were declined by Cygb over-expression whereas they were increased in the Cygb knockdown (Fig. 7, A-C).
There was no significant difference between the Cygb-shRNA-ADV and Cygb-cDNA-ADV groups in Hif-1α mRNA expression levels (Fig.  6C).
Thus, we propose that Hif-1α may be the upstream factor promoting Cygb expression, similar to neuroglobin (40), and Vegf, Caspase-2, and Caspase-3 are downstream target genes of Cygb.

Cygb Improves Long-term Leaning and
Memory in the HI Model-The Morris water maze test (n = 8) was performed at 3 weeks post-HI (P28 rats) to evaluate long-term spatial learning and memory abilities dependent upon the function of the hippocampus and cortex (41). The escape latency (EL) in training days (P28-32) is shown in Fig, 8A. The results showed longer EL in the HI group as compared to sham group (p < 0.01). EL in the Cygb-shRNA-ADV group was significantly longer compared with HI group (p < 0.01), whereas Cygb-cDNA-ADV groups had significantly shorter EL (p < 0.01). There was no significant difference in EL between the HI, saline, and ADV only groups. The space probe trial was conducted after spatial maze training. Platform crossing times and the time spent in the target quadrant were also recorded (Fig.8, B and C). Animals in the sham group had approximately 10 times crossing frequency, while the HI, saline, and ADV group had about 3 times crossing frequency, and there was no significant difference among these three groups (Fig. 8B). The crossing frequency was reduced to twice in the Cygb-shRNA-ADV group. In contrast, Cygb-cDNA-ADV significantly increased the crossing frequency. HI rats spent less time in the target quadrant compared with sham rats (p <0.01), and Cygb inhibition significantly reduced time spent in the target quadrant compared with HI rats (p < 0.01; Fig.  8C). Cygb-cDNA-ADV significantly increased the percentage of time spent in quadrant as compared to the HI group (p < 0.01; Fig. 8C). These results demonstrated a long-term neuroprotective effect of Cygb on HI brain injury.

DISCUSSION
Perinatal hypoxia-ischemia (HI) brain injury is a major cause of morbidity and mortality in infants and children (1,2), and an increasing number of studies are being conducted to investigate the pathogenesis of the disease. In present study, we show for the first time that Cygb is neuroprotective in neonatal rats exposed to HI through possible anti-oxidative and anti-apoptotic mechanisms and by promoting angiogenesis. First, we identified the endogenous up-regulation patterns of Cygb mRNA and protein in the developing brain under the condition of HI brain injury. Moreover, we found that changes in Cygb expression result in corresponding changes in the severity of histological and functional deficits after HI, in a manner consistent with an endogenous neuroprotective action of Cygb. In summary, this study suggests that Cygb plays an important role in the signaling pathway underlying oxidative stress, angiogenesis and neuronal apoptosis in the neonatal HI rat model.
Though several previous studies have demonstrated that Cygb is affected by hypoxia in vitro (16,21,42) and in vivo (17,18,20,42,43), the detailed expression patterns of Cygb mRNA and protein in neonatal rats brain suffered from HI have not been report until now. The present study is the first to undertake a comprehensive analysis of Cygb expression in response to neonatal rat brain HI injury at acute stage. Our study demonstrates that Cygb is significantly up-regulated at both transcriptional and translational levels in a time-dependent manner. However, it must be noted that two previous studies reported that the expression of Cygb was not significantly up-regulated in the ischemic model of adult rats (44,45). A possible explanation for the paradoxical findings is that there were some limitations in these two studies, such that no detailed data were presented to illustrate the changes in Cygb expression during the first 36 hours after ischemia, and different animal models that only suffered ischemic treatment were used. Moreover, it is possible that the age of the rats also contributed to the differences in results. Since we found that the expression of Cygb protein peaks at 24 h post-HI (Fig.2), it appears that Cygb is up-regulated during the acute stage of HI. Based on previous data of Cygb expression change under hypoxic conditions in vivo and in vitro (18,19,42,43), another possible explanation for the discrepancy is that hypoxia may play a more important role than ischemia in up-regulating Cygb expression. Since there was significantly higher Cygb mRNA and protein levels in neonatal HI injury rats compared to sham rats, which point to an important role of Cygb in hypoxia-ischemia adaptation, RNA interfering technology was used and a Cygb knock-down model was successfully established using the Cygb-shRNA-ADV system 4 days before HI induction (Fig. 3). Significantly aggravated brain injuries were observed at 24h after HI (Fig.5). Cognitive functions were also impaired severely under conditions of HI in the Cygb-shRNA-ADV group as demonstrated by the performance of MWM test (Fig. 8). These data indicated that Cygb is a novel endogenous neuroprotective factor in the neonatal HI model.
Many investigators have explored the development of therapeutic methods for neonatal HI disease. Although hypothermia, the most widely applied clinical intervention in asphyxiated babies, improves outcome, the efficiency is limited because it is only mildly affective for children born at term (46,47). Our study focused on the effect of Cygb over-expression in HI injury and was initially stimulated by the application of gene therapy to various brain diseases. The results showed that Cygb over-expression attenuated cerebral infarction and neuron cell apoptosis caused by HI in neonatal brain (Fig. 5). The protective effect improved long-term neurological function 3 weeks after HI insult (Fig. 8); thus, we propose that Cygb functions as an endogenous neuroprotective protein, providing a new therapeutic target for neonatal HI disease.
Previous findings have shown that Cygb is expressed in distinct regions of the mouse brain as compared with neuroglobin (Ngb), and these regions (hippocampus, thalamus, and hypothalamus) play an important role in protecting from oxidative stress (20). As compared with the adult brain, the immature brain is highly susceptible to oxidative stress because of its high concentration of unsaturated fatty acids, rate of oxygen consumption, and availability of redox active iron, but poorly developed scavenging systems (48). Therefore, oxidative stress is thought to be one of the major factors that induce neuronal cell death in the immature brain (49). In this study, MDA and SOD activity were measured to confirm the anti-oxidant role of Cygb in vivo. The results indicated that Cygb could increase neonatal tolerance to HI partly due to its anti-oxidative effect at the early stage of HI injury. The anti-oxidant effects of Cygb have recently been investigated, and emerging evidence has demonstrated a protective role of Cygb against oxidative stress, especially in models of fibrosis that involve hypoxic reperfusion and subsequent oxidative injury (19,25,(50)(51)(52)(53). It was also reported that Cygb expression protected neuronal cells from oxidative damage in vitro (21,22,54,55); however, it is unclear whether this is a direct anti-oxidative effect of Cygb under HI condition. Downstream signaling in the Cygb pathway may result in these anti-oxidative effects. Due to its NO dioxygenase activity (56)(57)(58), Cygb might reduce intracellular NO concentration, which in turn may prevent accumulation of peroxynitrite, a strong oxidant formed from the reaction between superoxidaxide and NO. It has been speculated that Cygb, like other hexacoordinated globins, may eliminate reactive oxygen species (ROS) utilizing haem and thiol residues (59,60), and it has been found that Cygb may reduce the induction of intracellular ROS formation (55,61). The processes of lipid-induced transformation of Cygb from hexa-co-ordinate to penta-co-ordinate may allow the cell to up-regulate antioxidant defenses before extensive oxidative damage occurs (62). Nevertheless, additional studies are required to confirm the pathways and mechanisms of the protein in protecting brain HI damage from oxidative stress.
In our study, we show that Cygb mediates a neuroprotective effect through reducing cerebral infractions and apoptosis caused by oxidative stress in vivo. Furthermore, the molecular mechanisms and signal pathways which are responsible for the neuroprotective property of Cygb are also of great interest. Fordel et al. reported that the up-regulation of Cygb, as well as Vegf gene, was abrogated in brain tissue of Hif-1 knockout mice upon hypoxia (42,63), and found that the mechanism of Cygb induction is Hif-1α-dependent (63). Guo et al. proposed that Hif-1 could regulate Cygb expression by binding to hypoxia responsive elements (HREs) of the protein (64) which was confirmed by Singh et al. in 2009 (19). Our results show that Hif-1α was not affected by Cygb, supporting the assumption that Hif-1α might be the upstream gene of Cygb in vivo under HI condition. In present study, we found that ischemic infarction could be affected by different expression levels of Cygb (Fig.5, A  and B), and both the Vegf mRNA and protein expression levels were regulated by Cygb (Fig.6) which is consistent with previous findings that Cygb may induce the induction and synthesis of Vegf in vitro (16). We also show that Cygb over-expression increased both the density and diameter of microvessels in brain tissue under HI conditions, due at least in part to the Vegf up-regulation. Vegf was found to increase phosphorylation of protein kinase B (Akt) and extracellular-signal regulated kinase 1/2 (ERK1/2) in the cortex in the neonatal rat HI model (65). Jones and Bergeron have found that activation of ERK1/2 contributes to HI-tolerance in the neonatal brain in part by preserving vascular and white matter integrity (66). A role for ERK1/2 in neurons has been demonstrated in neuronal gene expression regulation (CREB, Elk-1, and c-Myc) and long-term potentiation and memory formation (67). Dash et al. demonstrated ERK activation in the hippocampus and cortex during MWM training (68,69), which is consistent with our results that Cygb can promote long-term spatial learning and memory abilities. These findings are strengthened by data suggesting that Vegf can reduce infarct size, improve neurological performance, markedly enhance angiogenesis in the ischemic brain, and reduce neurological deficits during stroke recovery (39,70). Thus, we believe that "Hif-1α-Cygb-Vegf" signaling pathway might be an important part in the Cygb antioxidant mechanism to protect neurons against HI injury in neonatal brain and reduce infarct. An in vitro model confirming this "Hif-1α-Cygb-Vegf" signaling pathway is needed for future study. In addition, it remains to be tested whether other signal molecules may also act as upstream mediators of Cygb in vivo, such as erythropoietin (EPO), which has been confirmed to have anti-oxidative effects by regulating several downstream genes including Cygb (64).
Since previous studies of Cygb mainly focused on nitric oxide dioxygenase and lipid peroxidase activities, the role of Cygb in process of cell apoptosis remains unknown. Apoptosis is one of many key factors contributing to the injury of developmental brain induced by HI or other insults. Caspase inhibitors have been reported to provide neuroprotection in neonatal rat models (71). Caspase-2 is a developmentally regulated initiator caspase, and recent data has demonstrated that Caspase-2 mediates HI brain injury (38). In our study, we performed a TUNEL apoptosis assay and found a decreased number of apoptotic cells in both the cortex and hippocampus of Cygb over-expression in neonatal rats subjected to HI when compared with Cygb knockdown rats. Our findings demonstrate that Cygb over-expression suppresses the activation of Caspase-2 and Caspase-3. These findings indicate that the Caspase family members are downstream targets in the anti-apoptotic effects of Cygb. In the present study, we show that the anti-apoptotic function of Cygb in neonatal brain was achieved in part through the negative regulation of expression of Caspase-2 and Caspase-3. Previous studies have reported that genetic inhibition of Caspase-2 could reduce neonatal HI brain injury, and that Caspase-3 activity was also attenuated by inhibition of Caspase-2, indicating that Caspase-3 is the downstream gene of Caspase-2 (38). Caspase-3 has also been demonstrated to own a role in regulating ischemic neuronal injury (72), mediated by Vegf, in the regulation of neuronal death under hypoxia (73). Several studies have also suggested that Caspase-2 induced mitochondrial outer membrane permeabilization (MOMP) which led to the release of pro-apoptotic molecules from mitochondria, such as the release of Cytochrome C (74-76) by direct processing of full-length Bid to activated truncated tBid (77). Interestingly, ERK was reported to protect hypoxic cortical neurons via phosphorylation of Bid (78). Thus, it is possible that Cygb plays a neuroprotective role against HI injury partially due to reduction of oxidative stress-mediated caspase-dependent apoptosis.
In conclusion, Cygb, is up-regulated by HI in the neonatal brain. Using gene transfection technique, we demonstrated for the first time that Cygb displayed significant functions of neuronal protection and repair in the HI injured neonatal brain. Also, for the first time it is proposed that Cygb might exert its protective activity by molecular mechanisms of antioxidant and anti-apoptosis, and that the effects of reducing infarct might occur through regulating Vegf expression, thus affecting angiogenesis. Nevertheless, further investigation is required to explain the mechanism by which Cygb contributes to the molecular pathogenesis of neonatal brain HI disease and to pursue the utilization of Cygb in the clinical management of such disorders.    The results show endogenous up-regulation of Cygb in rats exposed to HI, and both mRNA and protein levels of Cygb were inhibited by shRNA but enhanced by cDNA at 18 h and 24 h, respectively (B, D). The ADV-null vector did not affect the expression of Cygb. a, p < 0.01 as compared with the sham group; b, indicates there is no significant difference as compared with the saline group; c, p < 0.01 as compared with the saline group. data are expressed as mean ± SD (n = 8). a, p < 0.01 as compared with the sham group; b, indicates there is no significant difference; c, p < 0.01 as compared with the HI group.

FIGURE 6. Cygb exerts its neuroprotective activity via anti-oxidant and anti-apoptotic mechanisms but also promotes angiogenesis.
A, the comparison of MDA level and SOD activity among different groups (n = 8). SOD activity was higher in the cDNA treatment group but lower in the shRNA group as compared with HI group. A representative image of Cygb related gene is shown (B).
RT-PCR showed that over-expression of Cygb leads to up-regulation of transcriptional level of Vegf, which plays a crucial role in promoting angiogenesis but down-regulates, apoptosis-related genes, Caspase-2 and Caspase-3 (C). β-actin was used as an internal control. The expression of Vegf, Caspase-2, and Caspase-3 were not affected by Cygb transfection under normal conditions (D, E). The data are expressed as mean ± SD in triplicate independent experiments (n = 3). Disrupted microvessels and cell morphology in both HE and anti-CD31 staining images of the ischemic border areas from sham, HI, shRNA+HI and cDNA+HI groups were shown, arrow head: microvessels, scale bar: 200 μm (F, I). Representative images of Vegf immunostaining from the corresponding groups show that the expression of Vegf protein is Cygb-dependent under HI conditions, ↓: Vegf positive cells (G). Quantification of Vegf positive cells/field is also shown (H). Immunohistochemical staining of CD31 of the number and diameter of microvessels in different groups are expressed as mean ± SD in triplicate independent experiments (n = 3) (J, K). a, p < 0.01 as compared with the sham group. b and "ns", indicates there is no significant difference; c, p < 0.01 as compared with the HI group.

FIGURE 7. The activation of both Caspase-2 and Caspase-3 induced by HI were suppressed by
Cygb over-expression. Representative electrophoretic images for both the activation of Caspase-2 (upper) and Caspase-3 (lower) from different groups are depicted (A). The results showed that the active levels of Caspase-2 and Caspase-3 protein were suppressed by Cygb (B, C). Results are presented as mean ± SD in triplicate independent experiments (n = 3). a, p < 0.01 as compared with HI group; ns, not significant.

FIGURE 8. Results from the Morris Water Maze tests (n=8).
Repeated measures ANOVA revealed statistically significant differences (p < 0.01) (day*group interaction) between the HI, shRNA and cDNA groups, but did not reveal differences between the saline or ADV-null vectors injection groups (A). ▼, Cygb-shRNA-ADV injection + HI group; •, HI group; □, saline injection + HI group; ■, ADV vectors only + HI group; ▲, Cygb-cDNA-ADV injection + HI group; ○, sham group. On the first day, the escape latency (EL) values of these groups were narrow in range. On the fourth day, the EL of the shRNA group were higher as compared with the HI group; however, EL values in Cygb-cDNA injection group were much lower than those in HI group (p < 0.01). On the fifth day, the sham group had a short swim path to the former platform location, while the rats of the other groups had longer swim paths. The platform crossing (swimming into the area where the former platform was located within 90 seconds after removing the platform) was lower in HI group, saline, and ADV only groups compared with the sham group (p < 0.01). (B). The number of platform crossings in the shRNA injection group was lower compared with HI model group (p < 0.05), also significantly less compared with the cDNA injection group (p < 0.01) (B). The comparison of the percentage of time spent in the target quadrant among different groups (C). a, p < 0.01 as compared with Sham group. b, indicates there is no significant difference; c, p < 0.01 as compared with HI group. which could promote angiogenesis, may be up-regulated by Cygb over-expression. Cygb may exert its anti-apoptotic function via negative regulation of Caspase-2 and Caspase-3 expression, potentially leading to reversal of impairments in neonatal rats that suffered from HI. Table 1 Gene Name (ID) Sequences (Forward/Reverse) Product Size Neuroprotective effect of Cygb in a neonatal rat model with HI injury