All surgical procedures and methods were approved by Colorado State University Animal Care and Use Committee (protocol numbers: 09-1351A and 11-2663A) and adhered to the Guide for the Care and Use of Agricultural Animals in Agricultural Research and Teaching. Thirteen male adult Walker hounds (25 – 30 kg), 4 years of age, were housed in individual stalls (Veterinary Teaching Hospital of Colorado State University, Fort Collins, Colorado, USA) , provided water and food ad libitum in 5000 feet of altitude, were adapted to these conditions for at least 1 week before being used in the experiments, and randomly assigned to one of three groups: (1) control (CON, n = 3), cerebral normoxia with systemic normoxia (SaO2 ~ 90 %); (2) cerebral hypoxia (CH, n = 6), cerebral hypoxia (SaO2 ~ 55 %) with systemic normoxia (SaO2 ~ 90 %); (3) global hypoxia (GH, n = 4), systemic hypoxia (SaO2 ~ 60 %). All efforts were made to minimize the suffering and the number of animals. This work is written according to the ARRIVE guidelines (http://www.nc3rs.org.uk/ARRIVE/).
Anesthetic and surgical procedures were adapted from those previously described methods  and reflecting a modification to the original model developed by Moss and Stein . Following an overnight fast, the dogs were sedated with fentanyl (10 μg/kg SQ), atropine (0.04 mg/kg SQ) and midazolam (0.2 mg/kg SQ) 30 min prior to anesthetic induction. Anesthesia was induced by propofol (2.5 mg/kg IV) and maintained by fentanyl (0.01 mg/kg IV). Once a surgical plane of anesthesia was confirmed, animals in the GH group were intubated and placed on a volume ventilator (10 breaths/minute) with 100 % oxygen and isoflurane (1.0-2.0 %) during surgical isolation of the jugular veins, carotid, and femoral arteries. A catheter was placed underneath the meninges to estimate intracranial pressure (ICP). ICP was recorded every 15 min on a pressure monitor (Marquet 7000, Fridley, MN). Cerebral perfusion pressure (CPP) (mmHg) was calculated by subtracting the ICP (mmHg) from the mean arterial pressure (MAP) (mmHg). Cardiac output (CO) was measured by thermodilution method  with 10 mL of iced saline. Pulmonary arterial pressure (PAP) and capillary wedge pressure (Pwedge) were assessed using a Swan-Ganz catheter inserted into the pulmonary artery. A catheter was placed in a dorsal pedal artery to measure MAP. Stroke volume (SV) and systemic vascular resistance (SVR) were calculated from the quotients of CO (L/min) divided by HR (bpm), and MAP divided by CO respectively. Pulmonary vascular resistance (PVR) was calculated by the subtraction of Pwedge from PAP divided by CO ((PAP-Pwedge)/CO) and is presented as the product of dyne and time relative to cm5 (dyn*s/cm5). A 2 mm flow probe (Transonic, Transonic Systems, Inc., Ithaca, NY) was placed around each carotid artery to measure blood flow. Data were collected at baseline and then every 30 min during the experiment.
Induction of cerebral or global hypoxia
Dogs assigned to the cerebral hypoxia cohort had their left femoral vein cannulated with a 14 Fr cannula and connected to the inlet port of a cardiopulmonary bypass pump (Roller pump, Cobe, Lakewood, CO, USA). Both carotid arteries were cannulated with an 8 Fr cannula and connected to the pump outlet with a “Y” connector. When venous and arterial isolations were achieved, the animals were weaned to 21 % O2 and ventilation rate adjusted to 8 - 10 breaths/minute to maintain normal systemic partial pressure of CO2 and oxygen saturation (SaO2 ~ 95 %). To maintain control over cerebral hypoxic blood flow and insure oxygenated arterial blood was not entering the brain via the vertebral arteries, both vertebral arteries were isolated at the level of 7th cervical vertebrae and occluded by tourniquets during the time course of cerebral hypoxia.
After normal carotid blood flow was achieved from the cardiopulmonary bypass pump, carotid arteries were perfused with femoral venous blood desaturated to 60 % for the CH group or saturated to ~95 % for the CON group. Carotid arterial blood flow of the CH cohort was matched to the blood flow of the GH cohort. Carotid arterial blood flow in the CON group was clamped at baseline values obtained prior to cannulation of carotid arteries. All animals underwent 2 hours of hypoxia (or normoxia in the CON group).
As described above, once a surgical plane of anesthesia with isoflurane (1.0-2.0 %) was confirmed, animals in the GH group were intubated and placed on a volume ventilator (10 breaths/minute). Dogs received 10 % O2 and when necessary room air was introduced to maintain 60 %.
Blood and tissue acquisition
Blood (6 mL) was drawn from the dorso-pedal arterial and jugular venous and internal carotid catheters at baseline, 30, 60, 90 and 120 min of hypoxia or normoxia. Samples were transferred into chilled vials, one set containing ethylenediaminetetraacetic acid (EDTA; 1.8 mg K3 EDTA per 1 mL of blood) and the other set containing EDTA, 0.3 M ethylene glycol tetraacetic acid (EGTA), 0.3 M glutathione. Plasma was separated by centrifugation (4 °C, 14,000 x g; 10 min), and stored at -80 °C until assayed. At the end of 2 hours of hypoxia/normoxia, the animals were euthanized with sodium pentobarbital (10 mL IV) and the heart and lungs were removed by median sternotomy. The left caudal lung lobe was removed, weighed, and oven dried (65 °C for 96 hours until stable weight was achieved) for lung wet weight to lung dry weight (LWW/LDW) ratios as indices of pulmonary edema. The right caudal lobe was frozen in liquid nitrogen and stored at -80 °C for further analysis. Remaining lung lobes were fixed in 10 % formalin for 24 hours, placed into 70 % ethanol, paraffin embedded and cut into 4 µm sections for histological analyses. Brainstem, cerebellum, and left half of the cerebral cortex, were removed, snap frozen in liquid nitrogen, and stored at -80 °C for later analysis. A coronal section through the cortex around the prefrontal lobe was prepared for histological analyses in a manner identical to that used for lung tissue.
Western blot analyses
Approximately 50 mg of frozen brain cortex and lung were homogenized (Next Advance Inc, Averill Park, NY, USA) in 1 mL of ice-cold buffer (40 mM Tris HCl, 10 mM Tris Base, 5 mM MgCl2, 100 mM NaCl, 1 % TritonX-100, 1 mM EDTA, pH 7.4) with protease and phosphatase inhibitors (Halt, Thermo Fisher, Rockford, IL, USA). Samples were centrifuged (4 °C, 10,000 x g, 10 min) and protein concentration of the supernatant was determined using a bicinchoninic acid assay (Thermo Fisher, Rockford, IL, USA). Samples were heat denatured in Laemmli buffer, separated using 10 % SDS-PAGE, transferred to nitrocellulose paper, and incubated in 5 % milk in TBST (Tris-buffered saline with tween) for 1 hour prior to immunoblotting. To prepare the nuclear extracts, 40 mg of cerebral cortex and lung tissue were homogenized in 200 volumes of Thermo Scientific NE-PER Nuclear and Cytoplasmic Extraction Kit buffer (Thermo Fisher, Rockford, IL, USA). 20 µg of the nuclear protein (lung and brain) was then prepared for immunoblotting as described above. Antibodies were purchased from Abcam (Cambridge, MA, USA; HO-1 # Ab13248, NQO1 # Ab2346) and Santa Cruz Biotechnology (Santa Cruz, CA, USA; Nrf2 #SC-1302, SOD1# SC-8637). Blots were incubated overnight at 4 °C with primary antibodies diluted 1:200 in TBST, washed in TBST, and incubated with HRP-conjugated secondary antibody diluted 1:1000 in 5 % milk in TBST for 1 hour at room temperature (anti-rabbit for Nrf2, anti-goat for SOD1 and NQO1 and anti-mouse for HO-1) followed by chemiluminescence detection (West Dura; Pierce, Rockford, IL, USA). Images were captured and densitometry conducted using a UVP Bioimaging system (Upland, CA, USA). Equal loading was verified using ponceau staining as well as actin antibodies (SC-8432; Santa Cruz Biotechnology, Santa Cruz, CA, USA).
Aforementioned 4 μm thick paraffin sections from brain cortex were mounted on poly-l-lysine slides. Slides were dewaxed and sections rehydrated by immersion in ethanol (100 %, 95 %, and 70 %) and then distilled water. After washing, sections were preincubated in PBS supplemented with 0.5 % BSA and 10 % normal horse serum (Santa Cruz Biotechnology, Santa Cruz, CA, USA) for 1 hour, then incubated overnight with mouse monoclonal anti-APPA4 antibody diluted 1:100 in PBS containing 0.5 % BSA and 15 % normal horse serum. The sections were then incubated for 1 hour with 1:1000 biotin-labeled anti-mouse secondary antibodies (Santa Cruz, CA, USA), followed by streptavidin-biotin-horseradish peroxidase solution containing 3′, 3′-diaminobenzidine (DAB) tetrahydrochloride dihydrate (Dakocytomation, CA, USA) and hydrogen peroxide. Finally, the sections were counterstained with hematoxylin. Signal density was quantified using Image J software (NIH, USA).
Assessment of oxidative stress
Carbonylated proteins of lung and brain tissues were detected and analyzed following derivatization of protein carbonyl groups with 2, 4-dinitrophenylhydrazine, using the OxyBlot kit (Millipore, Billerica, MA, USA). Immunodetection was performed using 15 μg of protein per lane (3µg/µl) and primary antibody directed against dinitrophenylhydrazone (Millipore, Billerica, MA, USA). To measure malondialdehyde (MDA) concentrations, 25 mg of lung and brain tissue was quantitated using a TBAR assay kit (Cayman Chemical, Ann Arbor, MI, USA) per manufacturer recommendations. Plasma catecholamines: Plasma obtained from the blood that was collected into tubes with EGTA/glutathione was used to quantitate circulating norepinephrine and epinephrine with enzyme-linked immunosorbent assay (ELISA) (Rocky Mountain Diagnostics Inc., CO, USA).
Hemodynamic data from all 13 subjects were included in analyses (CH=6, GH=4, CON=3). Unless otherwise noted, the remaining analyses were conducted using data from 12 dogs (tissues were not collected from the first CH procedure). Statistical analyses were performed with the SAS (version 9) statistical package (SAS, Cary, NC) and SPSS (IBM, version 19). Inter-subject variability in baseline hemodynamic measurements was controlled for by considering baseline measurements as covariates. Data were analyzed by one-way repeated measures of analysis of variance (ANOVA). Western blot, IHC, and catecholamine data were analyzed by one-way ANOVA. Significance was established a priori at p < 0.1 due to the necessary small sample size using this large animal model.