Journal of Cerebral Blood Flow & Metabolism (2013) 33, 1008–1014 & 2013 ISCBFM All rights reserved 0271-678X/13 $32.00 www.jcbfm.com

ORIGINAL ARTICLE

Genetic elimination of eNOS reduces secondary complications of experimental subarachnoid hemorrhage Mohammed Sabri1,2,3, Jinglu Ai1,2,3, Elliot Lass1,2,3, Josephine D’abbondanza1,2,3 and R Loch Macdonald1,2,3 Delayed complications of subarachnoid hemorrhage (SAH) such as angiographic vasospasm, cortical spreading ischemia, microcirculatory dysfunction, and microthrombosis are reported in both patients and animal models of SAH. We demonstrated previously that SAH is associated with increased oxidative stress in the brain parenchyma, and that this correlates with dysfunction of endothelial nitric oxide synthase (eNOS) (homodimeric uncoupling). Uncoupling of eNOS exacerbated oxidative stress and enhanced nitric oxide (NO) depletion, and was associated with multiple secondary complications such as microthrombosis, neuronal apoptosis, and release of reactive oxygen species. Thus, we hypothesized that genetic abbrogation of eNOS would confer a beneficial effect on the brain after SAH. Using a prechiasmatic injection model of SAH, we show here that eNOS knockout (KO) significantly alleviates vasospasm of the middle cerebral artery and reduces superoxide production. Endothelial nitric oxide synthase KO also affected other nitric oxide synthase isoforms. It significantly increases neuron nitric oxide synthase expression but has no effect on inducible nitric oxide synthase. Endothelial nitric oxide synthase KO decreases Zn2 þ release after SAH, reduces microthrombi formation, and prevent neuronal degeneration. This work is consistent with our findings where, after SAH, increased oxidative stress can uncouple eNOS via Zn2 þ thiolate oxidation, or theoretically by depletion or oxidation of tetrahydrobiopterin, resulting in a paradoxical release of superoxide anion radical, further exacerbating oxidative stress and microvascular damage. Journal of Cerebral Blood Flow & Metabolism (2013) 33, 1008–1014; doi:10.1038/jcbfm.2013.49; published online 3 April 2013 Keywords: eNOS knockout; mice; subarachnoid hemorrhage; vasospasm

INTRODUCTION Delayed complications of subarachnoid hemorrhage (SAH) such as angiographic vasospasm, cortical spreading ischemia, microcirculatory dysfunction, and microthrombosis are reported in some animal models of SAH.1–9 They are postulated to contribute to delayed cerebral ischemia in humans and thus, contribute to death and disability in patients with SAH.3 One theory behind how these secondary complications cause cerebral ischemia and infarction revolves around the depletion of or excess generation of the potent vasodilator nitric oxide (NO).3,6,10 Nitric oxide is synthesized enzymatically by three main nitric oxide synthase (NOS) isoforms, endothelial, neuronal, and inducible NOS, or in a nonenzymatic fashion via a series of nitrate– nitrite reduction–oxidation reactions.11 Nitric oxide produced under physiologic conditions confers a number of beneficial effects such as vasodilation of the microcirculation and maintenance of normal vascular tone, antithrombotic effects, prevention of excess platelet adhesion and aggregation, inhibition of endothelial apoptosis, and vascular smooth muscle cell hyperplasia.12 We demonstrated previously that SAH in mice is associated with oxidative stress in the brain parenchyma, and that this correlates with dysfunction of eNOS (homodimeric uncoupling). Uncoupling of eNOS exacerbated oxidative stress and enhanced NO depletion, and was associated with multiple secondary complications such as microthrombosis, neuronal apoptosis, and release of reactive oxygen species (ROS).4,10

Therefore, in the current study, we hypothesized that genetic abrogation of eNOS would confer a beneficial effect on the brain after SAH. Additionally, we investigated the mechanisms behind eNOS dysfunction, the potential role of other isoforms, and their potential contributions to secondary complications after SAH, if any. We show here that eNOS knockout (KO) significantly alleviates vasospasm of the middle cerebral artery (MCA) and reduces superoxide production, increases nitric oxide synthase (nNOS) expression but has no effect on inducible nitric oxide synthase (iNOS). Endothelial nitric oxide synthase KO decreases Zn2 þ release after SAH, reduces microthrombi formation, and prevents neuronal degeneration. MATERIALS AND METHODS Animals and Subarachnoid Hemorrhage Model Experiments were approved by the Animal Care Committee of St Michael’s Hospital, University of Toronto and complied with regulations of the Canadian Council on Animal Care and guildlines of ARRIVE (http:// www.nc3rs.org/ARRIVE). The eNOS KO mice were from Jackson Laboratory (Stock #007073, Bar Harbor, ME, USA). The KO mice were created by targeted mutation with a 1.2-kb neomycin cassette replaced with 129 bp of exon 12 of the gene. This replacement disrupts the calmodulin-binding site essential to eNOS function and introduces a premature translation stop codon into the eNOS transcripts.13 Mice homozygous for the eNOStargeted mutation were viable and fertile. All animals were bred in our animal care facility.

1 Division of Neurosurgery, St Michael’s Hospital, University of Toronto, Toronto, Ontario, Canada; 2Labatt Family Centre of Excellence in Brain Injury and Trauma Research, Keenan Research Centre, Li Ka Shing Knowledge Institute of St Michael’s Hospital, Toronto, Ontario, Canada and 3Department of Surgery, Institute of Medical Science, University of Toronto, Toronto, Ontario, Canada. Correspondence: RL Macdonald, Division of Neurosurgery, St Michael’s Hospital, University of Toronto, 30 Bond Street, Toronto, ON M5B 1W8, Canada. E-mail: [email protected] This study is supported by grants from The Heart and Stroke Foundation of Canada, and The Brain Aneurysm Foundation. Received 8 January 2013; revised 7 March 2013; accepted 11 March 2013; published online 3 April 2013

eNOS knockout improves SAH M Sabri et al

1009 Subarachnoid hemorrhage was created (by MS) by injection of autologous blood into the prechiasmatic cistern as previously reported.14 The head was fixed in a stereotactic frame equipped with a mouse adaptor (Stoelting Company, Wood Dale, IL, USA). Relative cerebral blood flow (CBF) was measured using a laser Doppler flow probe (BLT21, Transonics systems, New York, NY, USA). Body temperature was maintained at 37±0.5 1C with a homeothermic heating pad (Harvard apparatus, Holliston, MA, USA) and monitored with a rectal probe. A 0.9-mm hole was drilled in the midline of the skull 4.5 mm anterior to the bregma. The drill was angled 401 caudally, and a 27-gauge spinal needle was advanced at the same angle through the burr hole to the base of the skull. For the SAH group, 80–100 ml autologous blood (nonheparinized) was withdrawn from the ventral tail artery using a 25-gauge needle and 60 ml was injected through the spinal needle over 15 seconds. For the control group, the same volume of saline was injected. Animals in the sham groups underwent insertion of the needle but no injection. Allocation to groups was sequential with an SAH and a control animal operated each day. Mice were killed 48 hours after surgery. For histologic studies, mice were perfused through the left cardiac ventricle with NaCl, 0.9%, 10 mL, followed by 150 mL, 4% paraformaldehyde in phosphate-buffered saline. Brains were removed and postfixed in 4% paraformaldehyde for 48 hours. Sections of brain were processed and embedded in paraffin, and 7-mm sections were cut using a microtome (Leica, Wetzlar, Germany). For western blot and NO/ O2 detection, brains were isolated without perfusion. They were used fresh or immediately frozen at  80 1C for later analysis.

Hematoxylin and Eosin Staining and Vasospasm Measurement Paraffin sections were incubated in xylene for deparaffinization and rehydrated through a decreasing gradient of ethanol solutions. Slides were stained with hematoxylin and eosin. After dehydration through an increasing gradient of ethanol solutions and three changes of xylene, slides were coverslipped with xylene-based mounting medium (Permount, Sigma, Oakville, Ontario, Canada) and viewed under a light microscope. The lumen perimeter of the MCA was quantified by a masked observer (by EL) at  200 magnification using Image J (National Institutes of Health (NIH), Bethesda, MD, USA).

Zinc Assay Free zinc (Zn2 þ ) content in brain homogenates was measured using 4-(2-pyridylazo)-resorcinol disodium salt (PAR) assay.15 Absorbance at 500 nm was measured using a spectrofluorometer (SpectraMAX-Gemini, Molecular Devices, Sunnyvale, CA, USA). PAR in the absence of Zn2 þ is associated with low basal absorbance. However, when it is bound with Zn2 þ , the PAR2Zn2 þ complex has absorbance at 500 nm. The PAR assay specificity was tested by adding 5 ml of samples into a cuvette containing 150 mM PAR in Chelex buffer prepared with Chelex 100-treated 50 mM Tris, 100 mM NaCl, pH 7.8. Chelex was added to eliminate background cations and to test for the specificity of PAR assay, and was not utilized in subsequent buffers for brain homogenates. Assays were run at 231C in rapidly stirred cuvettes, and the total assay volume was 1.5 mL. Toward the end of the assay, 1 mM N,N-bis(carboxymethyl) glycine (NTA) was added. Under these assay conditions, NTA selectively chelates Zn2 þ from the PAR2Zn2 þ , resulting in a gradual reduction in absorbance at 500 nm, allowing for the calculation of the amount of Zn2 þ released from the protein.

Western Blots for Endothelial Nitric Oxide Synthase, Neuron Nitric Oxide Synthase and Inducible Nitric Oxide Synthase Brain tissue was homogenized in 300 ml 1% RIPA buffer with 0.1% protease inhibitor, and centrifuged at 13,000 r.p.m. for 12 minutes at 4 1C. Protein was quantified using the Bradford method, with RIPA buffer used as the blank standard. Thirty micrograms protein was loaded and separated by electrophoresis on 8% sodium dodecyl sulfate—polyacrylamide gels and transferred onto nitrocellulose membrane. We used Ponceau S and Gel Code to stain the membrane and gel, respectively. Blots were incubated with 5% milk for 60 minutes, followed by incubation with primary monoclonal antibodies (1:1000 dilution) against phosphorylated S1177eNOS (BD Biosciences, San Jose, CA, USA), nNOS (1:500 Abcam, Cambridge, MA, USA) and iNOS (1:500 Abcam ). After washing in PBS, membranes were incubated in horseradish peroxidase-conjugated antigoat polyclonal antibody (Abcam) at a dilution of 1:1000 for 50 minutes at room temperature. Reactions were developed with ECL reagent mix (Amersham Biosciences, Amersham, UK). Protein intensities were quantified by densitometric analysis utilizing Image J software (NIH). Values are expressed as relative unit or arbitrary unit after normalization to betatubulin controls. For eNOS monomer and dimer detection, we used low-temperature sodium dodecyl sulfate—polyacrylamide.16,17 Samples were subjected to sodium dodecyl sulfate—polyacrylamide on 8% gels that were kept in an ice bath at 4 1C. The gels were then blotted onto nitrocellulose membranes and blocked. The membranes were incubated with primary antibodies against eNOS (Cell Signalling, Danvers, MA, USA) and the remainder of the procedure was identical to the standard western blot steps.

Immunohistological Staining of Fibrinogen for Microthrombi Coronal brain sections were dereparaffinized and rehydrated. Antigen was retrieved by heating the sections for 20 minutes in 0.01 mmol/L sodium citrate (pH 6.0) at 96 1C. We permeabilized sections with 0.3% Triton X-100 for 10 minutes and then incubated them with 10% normal goat serum, which was diluted in 1% bovine serum albumin and 0.1% sodium azide, for 60 minutes. Slides were incubated with primary antibody (rabbit polyclonal antifibrinogen 1:200, Abcam) and secondary antibody (Alexa Fluor 568conjugated antirabbit for fibrinogen (1:1000, Invitrogen, Burlington, Ontario, Canada). Slides were washed and coverslipped with antifading mounting medium and sealed with nail polish. Slides were viewed in a confocal microscope and images were captured using consistent parameters (pinhole size, exposure time, and laser intensity).

Fluoro-Jade Staining Fluoro-jade B (Histo-Chem, Jefferson, AR, USA) staining was performed according to a previously published protocol.14 After deparaffinization and rehydration, the slides were incubated in 0.06% potassium permanganate (VWR International, Strasbourg, France) for 15 minutes. Slides were then rinsed in deionized water and immersed in 0.001% Fluoro-jade B in 0.1% acetic acid for 30 minutes. They were washed and dried at 60 1C for 15 minutes. Sections were cleared in xylene and coverslipped with a nonaqueous, low-fluorescence, styrene-based mounting medium (Sigma). Slides were viewed under a confocal microscope and images taken using constant parameters (laser power, exposure time, and pinhole size).

NO Superoxide Anion Radical Detection Superoxide anion radical (O2 ) and NO were detected in homogenized fresh or frozen brain tissue using spectrophotometric methods. The cellpermeable fluorophore 4,5-diaminofluorescein-2-diacetate (DAF-2DA, Alexis Biochemicals, Gruenberg, Germany) was used to detect NO, and a chemiluminescence probe, 2-methyl-6-(p-methoxyphenyl)-3,7-dihydroimidazo[1,2-]-pyrazin-3-one (MCLA), was used to detect O2 . Homogenized brain tissues were incubated with either 10 mmol/L 5-diaminofluorescein-2diacetate for 30 minutes or 4 mmol/L MCLA in transparent 96-well plates (Fisher, Ottawa, ON, Canada) at room temperature in the dark. DAF-2DA was excited at 495 nm and emission read at 515 nm in a spectrofluorometer (SpectraMAX-Gemini, Molecular Devices). The luminescence signal of MCLA was read directly at 495 nm. Reactions were followed for 10 minutes for 5-diaminofluorescein-2-diacetate and 6 minutes for MCLA and repeated three times. To test the specificity of MCLA, increasing concentrations of superoxide dismutase were used to reveal a concentration-dependent MCLA luminescence signal (data not shown). Since MCLA crosses cell membranes, the O2 detected is intra- and extracellular. & 2013 ISCBFM

Statistical Analysis and Data Quantification All data are presented as means±standard deviation (SD), and were compared between groups by analysis of variance (ANOVA). If significant variance was found, a posthoc test was performed using the Holm-Sidak method. Student t-test was used for continuous variables. Po0.05 was considered significant. To keep the quantification on staining consistent, we preselected five fixed symmetrical areas for cortex and three for hippocampus on a proper coronal section of mouse brain.18 For fibrinogen staining, we took one image from each fixed area (10 images from cortex, six images from hippocampus) and counted all the microthromboemboli in the section. To determine Fluoro-jade B staining in the cortex, we used higher magnification (400  ), and took three images from each of the preselected areas (30 images for both sides of the cortex). For the hippocampus, we counted all positive cells in all regions of the hippocampus (dentate gyrus, CA3, and CA1). All counting was done by researchers (by EL and JD) who were masked to the mouse groups. Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

eNOS knockout improves SAH M Sabri et al

1010 RESULTS All assessment was done 48 hours after creation of SAH or control surgery.

Endothelial Nitric Oxide Synthase Knockout Does Not Affect Cerebral Blood Flow Acutely after Subarachnoid Hemorrhage The CBF changes are similar to that in our previous reports.4,6,14 In sham-operated mice, there were no changes in CBF after the insertion of the needle (Figure 1A). Subarachnoid hemorrhage animals showed a steep drop in CBF seconds after the injection. Raw CBF data of tissue perfusion units were standardized to baseline, and expressed as percent change from an averaged baseline. At the time of injection or needle insertion, the average CBF decrease was to 6.5%±2.2% for wild-type SAH, 5.7%±3.4% for eNOS KO mice, 55%±4.4% for saline-injected, and 95.2%±1.9% for sham controls (Po0.001 all groups compared with sham controls, one-way analysis of variance, n ¼ 7 for all groups). There was no significant difference between the wildtype and eNOS KO SAH groups. It took B10 minutes for the CBF to recover and stabilize at 75% to 85% range in both SAH groups. In contrast, the CBF in saline-injected controls recovered to the baseline level within 30 minutes after injection.

Endothelial Nitric Oxide Synthase Knockout Decreases Vasospasm The eNOS KO mice exhibited less vasospasm after SAH compared with wild-type mice (Figure 1B). The perimeter of the MCA was 93±14 mm for eNOS KO SAH mice and 53±4 mm for wild-type SAH mice (mean±s.d., Po0.001, Student t-test, n ¼ 7 for both groups, Figure 1B). Endothelial Nitric Oxide Synthase Knockout Prevents Nitric Oxide Reduction after Subarachnoid Hemorrhage We previously reported that SAH is associated with decreased NO production.4 To determine whether reduction of vasospasm by eNOS KO is related to NO production, we measured NO in brains from mice with or without SAH (Figure 1C). Nitric oxide value is expressed as relative fluorescence units (RFU). The eNOS KO mice without SAH induction had the same NO production as wild-type mice (NO was 875±340 RFU for naive eNOS KOas compared with 813±339 RFU for wild-type naive animals, P40.05). Subarachnoid hemorrhage significantly reduced NO production in wild-type animals (294±143 RFU for SAH, 813±339 RFU for naive animals, Po0.001, Figure 1C) but not in the eNOS KO mice (570±410 RFU for SAH, 875±340 RFU for naive, P40.05, Figure 1C). Thus, eNOS KO partially prevents SAH-induced reduction of NO as compared with wild-type animals. Endothelial Nitric Oxide Synthase Knockout Reduces Superoxide Anion Radical after Subarachnoid Hemorrhage We previously showed that SAH resulted in upregulated but uncoupled eNOS, which produced O2 instead of NO (O2 values are expressed as relative luminescence unit (RLU)).4 Similarly, in this study, SAH produced a significantly higher amount of O2 than naive controls in the wild-type mice (66±12 RLU for SAH, 27±5 RLU for naive wild-type mice, Po0.001, Figure 1D). Endothelial nitric oxide synthase KO alone without SAH did not change O2 production (29±4.5 RLU for naive eNOS KO mice, 27±5 RLU for naive wild-type mice, P40.05). However, eNOS KO did reduce O2 production after SAH induction (44±8 RLU for eNOS KO SAH mice, 66±12 RLU for wild-type SAH mice, Po0.01). Endothelial nitric oxide synthase KO did not completely eliminate O2 production after SAH, suggesting that eNOS is only partially responsible for O2 production in SAH animals (Po0.001 SAH versus naive in eNOS KO mice, Figure 1D).

Figure 1. Endothelial nitric oxide synthase (eNOS) knockout (KO) reduces vasospasm, nitric oxide (NO) depletion, and O2 production with no effect on cerebral blood flow (CBF). (A) Cerebral blood flow measurement showing that CBF in both wild-type (WT) and eNOS (KO) mice were similarly reduced after blood injection, and recovered and remained around 75–85% of baseline level throughout the whole experiment. (B) Bar graph showing a significantly larger perimeter in the middle cerebral artery (MCA) from eNOS KO subarachnoid hemorrhage (SAH) animals compared with WT SAH animals (n ¼ 7 for both groups), representative images of MCA from both groups are shown. (C) There is a trend of higher NO production in eNOS KO mice as compared with WT SAH, in which NO is significantly reduced as compared with naive animals of both WT and eNOS KO. (D) O2 production is significantly higher in both SAH groups, with eNOS KO animals showing a smaller increase. Values are mean±s.d., n ¼ 7 for all groups. RFU, relative fluorescence unit; RLU, relative luminescence unit. Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

Endothelial Nitric Oxide Synthase Knockout Increases Neuron Nitric Oxide Synthase Expression but has no Effect on Inducible Nitric Oxide Synthase In wild-type animals, there was significantly more eNOS monomer than dimer after SAH as compared with that in eNOS KO mice, where both eNOS monomer and dimer were expressed at a similar, almost undetectable level (monomer/dimer ratio for wild type is13.6±0.9, for KO is 1.4±0.03, Po0.001, t-test, Figure 2). There was a significant increase of nNOS expression in eNOS KO mice after SAH as compared with wild-type mice (nNOS expression was 167±13 AU for eNOS KO mice and 129±4 AU for wild-type mice, Po0.05, t-test, Figure 3). Compared with naive controls, SAH significantly increased iNOS protein in both wildtype and eNOS KO mice (6.4±0.4 AU for naive, 34.2±5.7 AU for SAH in eNOS KO and 37.3±4.6 AU for SAH in wild-type mice, Po0.001, one-way analysis of variance, Figure 3). However, there was no significant difference between wild-type and eNOS KO mice with SAH (P40.05), showing that eNOS KO did not prevent iNOS induction by SAH. Endothelial Nitric Oxide Synthase Knockout Decreases Zinc Release after Subarachnoid Hemorrhage A hallmark of eNOS uncoupling is Zn2 þ release from its thiolate complex. To investigate further the effect of SAH and eNOS KO on & 2013 ISCBFM

eNOS knockout improves SAH M Sabri et al

1011

Figure 2. Endothelial nitric oxide synthase (eNOS) expression. (A) Analysis of eNOS monomer and dimer expression and monomer/ dimer ratio demonstrating that as compared with wild-type (WT) animals, eNOS knockout (KO) mice only show trace amount of expression of eNOS protein in both monomer and dimer form. (B) Representative western blots demonstrating the absence of eNOS protein in eNOS KO mice. A substantially increased eNOS monomer expression is detected in WT animals with subarachnoid hemorrhage. Values are mean±s.d., n ¼ 4 for all groups. RU, relative unit.

NOS uncoupling, we measured Zn2 þ content in brain homogenates from different groups of animals. Zn2 þ value is expressed as relative absorbance (RA) at 500 nm. Figure 4 shows that SAHinduced Zn2 þ release was significantly decreased in eNOS KO animals as compared with wild-type animals (90±10 RA for eNOS KO, 161±25 RA for wild type, Po0.001). In contrast, in shamoperated or saline-injected mice, eNOS KO animals had a significant increase in Zn2 þ release compared with wild-type animals. In sham animals, Zn2 þ release was 51±7 RA for eNOS KO mice and 34±6 RA for wild-type mice. In saline-injected animals, Zn2 þ release was 49±10 RA for eNOS KOs and 36±7 RA for wildtype mice (Po0.01 for both comparisons, n ¼ 5 for all groups). Interestingly, SAH induced significant Zn2 þ release as compared with sham-operated or saline-injected controls for both eNOS KO or wild-type animals (Po0.001, one-way analysis of variance, n ¼ 5 for all groups). Endothelial nitric oxide synthase Knockout Reduces Microthrombi and Neuronal Degeneration Having observed that mice with no eNOS had less vasospasm, a trend towards less NO production and a significantly higher O2 as compared with wild-type SAH mice, we investigated whether these changes were accompanied by prevention of microthrombi that have been postulated to contribute to brain injury after SAH.2,3 Antifibrinogen staining of cortical and hippocampal sections of saline-injected and SAH mice showed that mice with SAH had many fibrinogen-positive stained microvessels scattered throughout multiple layers of the cerebral cortex and hippocampus (Figure 5B). In comparison, sections from eNOS KO mice showed fewer fibrinogen-stained microvessels in the brain (Figure 5B). The number of microthrombi in wild-type SAH animals was significantly increased in both cortical and hippocampal sections in comparison to that in eNOS KO SAH animals (Figure 5A). In hippocampus, the counts are 79±16 for wild-type SAH and 29±16 for eNOS KO SAH, (Po0.001, t-test, n ¼ 6 for KO, 4 for wild type). In cerebral cortex, the counts were 124±29 for wild type and 48±26 for eNOS KO SAH (Po0.01, t-test, n ¼ 6 for KO, 4 for wild type, Figure 5A). & 2013 ISCBFM

Figure 3. Neuron nitric oxide synthase (nNOS) and inducible nitric oxide synthase (iNOS) expression. (A) Bar graph shows significantly increased nNOS expression in eNOS knockout (KO) mice as compared with wild-type (WT) controls after subarachnoid hemorrhage (SAH). Inducible nitric oxide synthase expression is notably increased in both WT and eNOS KO mice after SAH as compared with naive controls. There is no significant difference between the two SAH groups. (B) Representative western blots of nNOS and iNOS. Values are means±s.d., n ¼ 4 for all groups. AU, arbitrary unit.

Figure 4. Zn2 þ release determined by 4-(2-pyridylazo)-resorcinol disodium salt assay. Bar graph shows that subarachnoid hemorrahge -induced Zn2 þ release was significantly decreased in endothelial nitric oxide synthase (eNOS) knockout (KO) animals as compared with wild-type animals (WT). In contrast, in shamoperated or saline-injected control groups, eNOS KO animals have a significant increase of Zn2 þ release than that in WT animals. Values are means±s.d., n ¼ 5 for all groups. RA, release assay. Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

eNOS knockout improves SAH M Sabri et al

1012

Figure 5. Quantification of microthrombi. (A) Bar graph showing significantly more microthrombi in the wild-type (WT) mice subarachnoid hemorrhage (SAH) group as compared with endothelial nitric oxide synthase (eNOS) knockout (KO) mice in both hippocampus and cerebral cortex. Values are means±s.d., n ¼ 6 for eNOS KO mice, n ¼ 4 for WT mice. (B) Representative images from immunohistochemical staining of fibrinogen, demonstrating the presence of microthrombi in the two brain regions. More microthrombi could be seen in WT animals with SAH as compared with the eNOS KO mice.

Fluoro-jade staining to assess neuronal degeneration gave similar results to those of microthrombi counts (Figure 6). Subarachnoid hemorrhage was associated with degenerating neuronal cells in cerebral cortex and hippocampus in the wildtype mice (Figure 6B). In comparison, eNOS KO reduced the number of degenerating neurons in both regions. In hippocampus, the counts of degenerating cells were 86±19 for wild-type SAH and 19±8 for eNOS KO SAH (Po0.001 t-test, n ¼ 5 for KO, 3 for wild type). In cerebral cortex, the counts were 125±14 for wild type and 60±8 for eNOS KO SAH, (Po0.001, t-test, n ¼ 5 for KO, 3 for wild type Figure 6A). DISCUSSION Subarachnoid hemorrhage in various animal models and in humans was associated with oxidative stress, likely due to clotderived ROS.19–23 We demonstrated that this oxidative stress was associated with eNOS uncoupling in the brain parenchyma and that this contributed to nitrosative stress, resulting in neuronal apoptosis, microvascular dysfunction, and large artery vasospasm.10,24 The eNOS uncoupling could be reversed by drugs like simvastatin, which has antioxidant and other effects Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

Figure 6. Quantification of degenerating neuronal cells. (A) Bar graph showing significantly higher number of Fluoro-jade-positive cells in the wild-type (WT) mice subarachnoid hemorrahge (SAH) group as compared with endothelial nitric oxide synthase (eNOS) knockout (KO) mice in both hippocampus and cerebral cortex. Values are means±s.d., n ¼ 5 for eNOS KO mice, n ¼ 3 for WT mice. (B) Representative images from Fluoro-jade staining showing the presence of degenerating neuronal cells in the two brain regions. More degenerating neuronal cells could be seen in the images from WT animals with SAH as compared to the eNOS KO mice.

that may preserve eNOS function.4,24 However, simvastatin has multiple potential mechanisms of action that preclude definition of the exact mechanisms of its beneficial effects. A recent randomized, double-blind, placebo-controlled trial of simvastatin does not support a beneficial effect of simvastatin in patients with SAH, which suggests that translation from bench-to-bedside could be difficult.25 A larger phase-3 trial (SimvasTatin in Aneurysmal Subarachnoid Hemorrhage (STASH)) is in progress (http://urlm.co/ www.stashtrial.com). In the current experiments, eNOS KO mice were used to overcome these limitations and gain more insight into the role of eNOS in brain injury after SAH. It should be emphasized that at baseline, naive eNOS KO mice are similar on all the endpoint measurements as compared with wild-type controls. They differed, however, in the response to SAH. The eNOS KO mice exhibited less vasospasm, microthrombi, and brain injury after SAH. Furthermore, they had reduced oxidative stress as measured by reduced O2 content. Another new finding of this study is that other isoforms of NOS changed differently in eNOS KO animals in comparison to wild-type mice. Subarachnoid hemorrhage increased iNOS in both wild-type and eNOS KO mice, which is consistent with other studies in wild-type animals.26,27 Increased & 2013 ISCBFM

eNOS knockout improves SAH M Sabri et al

1013 iNOS activity could produce more ROS and aggravate brain injury and large artery vasospasm.28–30 Yatsushige et al,31 however, found that inhibiting the increase in iNOS after SAH in rats did not reduce blood–brain barrier breakdown, brain edema, neuron death, or mortality. Neuronal NOS has been less studied after SAH. After SAH in humans, nNOS messenger ribonucleic acid was unchanged in brain tissue except in patients in poor clinical condition in whom it was increased.32 We found that nNOS was significantly elevated after SAH in eNOS KO but not wild-type mice, suggesting that this is a compensatory response to the lack of eNOS. Prior studies have suggested that nNOS decreases in brain33 and cerebral arteries34 after SAH, which is similar to the current findings. The hypothesis supported by the current data is that eNOS is uncoupled and causes detrimental effects after SAH.12 In the current study, eNOS monomers were increased after SAH in wildtype mice, and we further demonstrate that SAH in wild-type mice was associated with Zn2 þ release in comparison to animals injected with saline or in sham-operated mice. The eNOS KO mice with SAH also demonstrated a similar increase in Zn2 þ , however, there was a much smaller increase in comparison to wild-type mice. These data are consistent with previously published work showing peroxynitrite-induced oxidation of Zn2 þ from the Zn2 þ thiolate cluster of eNOS, which results in the formation of disulfide bonds between the monomers. This could explain why the dimer can be separated in SDS-reducing gel conditions.35 It has been documented that oxidative stress and increased peroxynitrite target the thiolate complex in eNOS rather than free thiols.35 Experiments demonstrated that peroxide-oxidized thiols had no added effect on eNOS dimer or Zn2 þ content. Recent studies demonstrated that Zn2 þ homeostasis is very important for normal function of neurons and glia in the central nervous system. Zn2 þ accumulation may contribute to selective and delayed degeneration of hippocampal pyramidal neurons in the CA1 subregion after focal and transient global ischemia.36 Intramitochondrial Zn2 þ accumulation leads to a loss of mitochondrial membrane potential and increased generation of ROS that exacerbate oxidative stress and eventually lead to cell death.36 This suggests that Zn2 þ accumulation from eNOS uncoupling after SAH may represent a novel mechanism and potential target for therapy of observed neuronal cell death in the disease. Interestingly, even in the absence of eNOS, there was some Zn2 þ release after SAH, possibly due to release from other NOS isoforms (nNOS and iNOS). Neuronal NOS can also uncouple and release Zn2 þ under some conditions.35,37–39 There are multiple possible mechanisms underlying the uncoupling of eNOS after SAH, such as depletion of tetrahydrobiopterin (BH4) or oxidization to dihydrobiopterin (BH2), processes that have been demonstrated to result in eNOS dysfunction in a number of studies. For example, Santhanam, et al,40 demonstrated that reduced bioavailability of tetrahydrobiopterin in cerebral microvessels of hyperphenylalaninemic mice resulted in increased superoxide production and reduced NO release. The role of eNOS was also supported by prevention of these findings with L-N(G)nitro arginine-methyl ester, a NOS inhibitor. This work is consistent with our findings where after SAH, increased oxidative stress can uncouple eNOS via Zn2 þ thiolate oxidation, or theoretically by tetrahydrobiopterindepletion or oxidation, resulting in a paradoxical release of superoxide anion radical, further exacerbating oxidative stress and microvascular damage. Overall, these findings have implications for treatment of brain injury after SAH. DISCLOSURE/CONFLICT OF INTEREST RL Macdonald is a consultant for Actelion Pharmaceuticals and Chief Scientific Officer of Edge Therapeutics. The remaining authors declare no conflict of interest.

& 2013 ISCBFM

ACKNOWLEDGEMENTS RL Macdonald receives grant support from the Physicians Services Incorporated Foundation, Brain Aneurysm Foundation, Canadian Stroke Network, Canadian Institutes of Health Research and the Heart and Stroke Foundation of Ontario.

REFERENCES 1 Friedrich B, Muller F, Feiler S, Scholler K, Plesnila N 2012Experimental subarachnoid hemorrhage causes early and long-lasting microarterial constriction and microthrombosis: an in-vivo microscopy study. J Cereb Blood Flow Metab 32: 447–455. 2 Macdonald RL, Pluta RM, Zhang JH 2007Cerebral vasospasm after subarachnoid hemorrhage: the emerging revolution. Nat Clin Pract Neurol 3: 256–263. 3 Pluta RM, Hansen-Schwartz J, Dreier J, Vajkoczy P, Macdonald RL, Nishizawa S et al. 2009Cerebral vasospasm following subarachnoid hemorrhage: time for a new world of thought. Neurol Res 31: 151–158. 4 Sabri M, Ai J, Marsden PA, Macdonald RL 2011Simvastatin re-couples dysfunctional endothelial nitric oxide synthase in experimental subarachnoid hemorrhage. PLoS One 6: e17062. 5 Sabri M, Ai J, Macdonald RL 2011Dissociation of vasospasm and secondary effects of experimental subarachnoid hemorrhage by clazosentan. Stroke 42: 1454–1460. 6 Sabri M, Ai J, Macdonald RL 2011Nitric oxide related pathophysiological changes following subarachnoid haemorrhage. Acta Neurochir Suppl 110: 105–109. 7 Vergouwen MD, Vermeulen M, Coert BA, Stroes ES, Roos YB 2008Microthrombosis after aneurysmal subarachnoid hemorrhage: an additional explanation for delayed cerebral ischemia. J Cereb Blood Flow Metab 28: 1761–1770. 8 Dreier JP, Korner K, Ebert N, Gorner A, Rubin I, Back T et al. 1998Nitric oxide scavenging by hemoglobin or nitric oxide synthase inhibition by N-nitro-L-arginine induces cortical spreading ischemia when K þ is increased in the subarachnoid space. J Cereb Blood Flow Metab 18: 978–990. 9 Dreier JP, Ebert N, Priller J, Megow D, Lindauer U, Klee R et al. 2000Products of hemolysis in the subarachnoid space inducing spreading ischemia in the cortex and focal necrosis in rats: a model for delayed ischemic neurological deficits after subarachnoid hemorrhage? J Neurosurg 93: 658–666. 10 Sabri M, Ai J, Knight B, Tariq A, Jeon H, Shang X et al. 2011Uncoupling of endothelial nitric oxide synthase after experimental subarachnoid hemorrhage. J Cereb Blood Flow Metab 31: 190–199. 11 van Faassen EE, Bahrami S, Feelisch M, Hogg N, Kelm M, Kim-Shapiro DB et al. 2009Nitrite as regulator of hypoxic signaling in mammalian physiology. Med Res Rev 29: 683–741. 12 Forstermann U, Munzel T 2006Endothelial nitric oxide synthase in vascular disease: from marvel to menace. Circulation 113: 1708–1714. 13 Shesely EG, Maeda N, Kim HS, Desai KM, Krege JH, Laubach VE et al. 1996Elevated blood pressures in mice lacking endothelial nitric oxide synthase. Proc Natl Acad Sci USA 93: 13176–13181. 14 Sabri M, Jeon H, Ai J, Tariq A, Shang X, Chen G et al. 2009Anterior circulation mouse model of subarachnoid hemorrhage. Brain Res 1295: 179–185. 15 Crow JP, Sampson JB, Zhuang Y, Thompson JA, Beckman JS 1997Decreased zinc affinity of amyotrophic lateral sclerosis-associated superoxide dismutase mutants leads to enhanced catalysis of tyrosine nitration by peroxynitrite. J Neurochem 69: 1936–1944. 16 Hemmens B, Goessler W, Schmidt K, Mayer B 2000Role of bound zinc in dimer stabilization but not enzyme activity of neuronal nitric-oxide synthase. J Biol Chem 275: 35786–35791. 17 Leber A, Hemmens B, Klosch B, Goessler W, Raber G, Mayer B et al. 1999Characterization of recombinant human endothelial nitric-oxide synthase purified from the yeast Pichia pastoris. J Biol Chem 274: 37658–37664. 18 Paxinos G, Franklin KBJ, 2004The Mouse Brain in Stereotaxic Coordinates. 2nd ed., London, UK: Academic Press. 19 Loftspring MC 2010Iron and early brain injury after subarachnoid hemorrhage. J Cereb Blood Flow Metab 30: 1791–1792. 20 Clark JF, Sharp FR 2006Bilirubin oxidation products (BOXes) and their role in cerebral vasospasm after subarachnoid hemorrhage. J Cereb Blood Flow Metab 26: 1223–1233. 21 Barry C, Turner RJ, Corrigan F, Vink R 2012New therapeutic approaches to subarachnoid hemorrhage. Expert Opin Investig Drugs 21: 845–859. 22 Macdonald RL, Weir BK 1994Cerebral vasospasm and free radicals. Free Radic Biol Med 16: 633–643. 23 Sakaki S, Kuwabara H, Ohta S 1986Biological defence mechanism in the pathogenesis of prolonged cerebral vasospasm in the patients with ruptured intracranial aneurysms. Stroke 17: 196–202. 24 Sabri M, Kawashima A, Ai J, Macdonald RL 2008Neuronal and astrocytic apoptosis after subarachnoid hemorrhage: a possible cause for poor prognosis. Brain Res 1238: 163–171.

Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

eNOS knockout improves SAH M Sabri et al

1014 25 Vergouwen MD, Meijers JC, Geskus RB, Coert BA, Horn J, Stroes ES et al. 2009Biologic effects of simvastatin in patients with aneurysmal subarachnoid hemorrhage: a double-blind, placebo-controlled randomized trial. J Cereb Blood Flow Metab 29: 1444–1453. 26 Zheng B, Zheng T, Wang L, Chen X, Shi C, Zhao S 2010Aminoguanidine inhibition of iNOS activity ameliorates cerebral vasospasm after subarachnoid hemorrhage in rabbits via restoration of dysfunctional endothelial cells. J Neurol Sci 295: 97–103. 27 Maddahi A, Ansar S, Chen Q, Edvinsson L 2011Blockade of the MEK/ERK pathway with a raf inhibitor prevents activation of pro-inflammatory mediators in cerebral arteries and reduction in cerebral blood flow after subarachnoid hemorrhage in a rat model. J Cereb Blood Flow Metab 31: 144–154. 28 Sayama T, Suzuki S, Fukui M 1999Role of inducible nitric oxide synthase in the cerebral vasospasm after subarachnoid hemorrhage in rats. Neurol Res 21: 293–298. 29 Widenka DC, Medele RJ, Stummer W, Bise K, Steiger HJ 1999Inducible nitric oxide synthase: a possible key factor in the pathogenesis of chronic vasospasm after experimental subarachnoid hemorrhage. J Neurosurg 90: 1098–1104. 30 Ayer RE, Zhang JH 2008Oxidative stress in subarachnoid haemorrhage: significance in acute brain injury and vasospasm. Acta Neurochir Suppl 104: 33–41. 31 Yatsushige H, Calvert JW, Cahill J, Zhang JH 2006Limited role of inducible nitric oxide synthase in blood-brain barrier function after experimental subarachnoid hemorrhage. J Neurotrauma 23: 1874–1882.

Journal of Cerebral Blood Flow & Metabolism (2013), 1008 – 1014

32 Berra LV, Carcereri DePA, Suzuki H, Pasqualin A 2007The role of constitutive and inducible nitric oxide synthase in the human brain after subarachnoid hemorrhage. J Neurosurg Sci 51: 1–9. 33 Sehba FA, Chereshnev I, Maayani S, Friedrich Jr. V, Bederson JB 2004Nitric oxide synthase in acute alteration of nitric oxide levels after subarachnoid hemorrhage. Neurosurgery 55: 671–677. 34 Pluta RM, Thompson BG, Dawson TM, Snyder SH, Boock RJ, Oldfield EH 1996Loss of nitric oxide synthase immunoreactivity in cerebral vasospasm. J Neurosurg 84: 648–654. 35 Zou MH, Shi C, Cohen RA 2002Oxidation of the zinc-thiolate complex and uncoupling of endothelial nitric oxide synthase by peroxynitrite. J Clin Invest 109: 817–826. 36 Sensi SL, Paoletti P, Bush AI, Sekler I 2009Zinc in the physiology and pathology of the CNS. Nat Rev Neurosci 10: 780–791. 37 Xia Y, Zweier JL 1997Superoxide and peroxynitrite generation from inducible nitric oxide synthase in macrophages. Proc Natl Acad Sci USA 94: 6954–6958. 38 Huhmer AF, Nishida CR, Ortiz de Montellano PR, Schoneich C 1997Inactivation of the inducible nitric oxide synthase by peroxynitrite. Chem Res Toxicol 10: 618–626. 39 Huhmer AF, Gerber NC, de Montellano PR, Schoneich C 1996Peroxynitrite reduction of calmodulin stimulation of neuronal nitric oxide synthase. Chem Res Toxicol 9: 484–491. 40 Santhanam AV, d’Uscio LV, Smith LA, Katusic ZS 2012Uncoupling of eNOS causes superoxide anion production and impairs NO signaling in the cerebral microvessels of hph-1 mice. J Neurochem 122: 1211–1218.

& 2013 ISCBFM