Methylene Blue Administration During Cardio- Pulmonary Resuscitation and Early Reperfusion Pro- Tects Against Cortical Blood-Brain Barrier Disruption

Home , Methylene blue

Methylene blue administration during cardiopulmonary resuscitation and early reperfusion protects against cortical blood-brain barrier disruption

Miclescu Adriana, Sharma Hari Shanker, Cécile Martijn, Wiklund Lars

Department of Surgical Sciences/Anaesthesiology and Critical Care Medi- cine, Uppsala University Hospital, SE-751 85 Uppsala, Sweden

Objective: The present study was time. The immunohistochemistry designed to study the effects of car- analysis indicated less blood brain diac arrest and cardiopulmonary barrier disruption in the animals resuscitation (CPR) on blood-brain receiving MB (CA+MB group) barrier (BBB) permeability and sub- evidenced by decreased albumin sequent neurological injury. It also leakage (P<0.01), water content tests the cerebral effects of MB on (P=0.02), potassium (P=0.04), but the maintenance of BBB integrity, also decreased neuronal injury the production of nitric oxide (NO) (P<0.001) in this group in and regulation of nitric oxide syn- comparison with the group that was thases (NOS) in cerebral cortex. not treated with MB (CA-MB). Intervention: The control group (CA, Similarly, MB treatment reduced n=16) underwent 12 min cardiac nitrite/nitrate ratio (P=0.02), iNOS arrest without subsequent CPR, after expression (P<0.01), and nNOS which the brain of the animals was expression (P<0.01). removed immediately or after 15 and Conclusion: Cerebral edema, in- 30 min. The other two groups with 12 crease BBB permeability and neuro- min cardiac arrest and subsequent 8 logic injury are observed early in min CPR received either an infusion ischemia induced by cardiac arrest. of saline (CA-MB group, n=10) or an MB markedly reduced BBB disrup- infusion of saline with MB (CA+MB, tion and subsequent neurologic in- n= 12) started one minute after the jury. In addition with these cerebral start of CPR and continued 50 min morphologic effects, the exposure to after return of spontaneous circula- MB is associated with a decrease of tion (ROSC). In both the latter (CA- NO as measured by nitrate/nitrite MB and CA+MB) groups the brains content and partially inhibition of were removed for histological analy- NOS activity. sis at the following time points: 30, Key words: Cardiopulmonary resus- 60, 180 min after ROSC. citation (CPR); Blood-brain barrier Main Results: In all the groups an (BBB); Methylene Blue (MB); Nitric increase of necrotic neurons and oxide (NO); Nitrite/nitrate content; albumin immunoreactivity was Nitric oxide synthases (NOS) demonstrated with increasing duration of ischemia and reperfusion Introduction synthases (NOS) and consequently quantify the nitric oxide metabolite The role of the blood-brain barrier levels after global brain ischemia. We (BBB) in inducing neuronal damage previously demonstrated that the following cardiac arrest (CA) is still most serious effects on cerebral blood not well clarified. Suggested causes flow and oxygen metabolism oc- of disruption of the BBB in ischemia curred in cerebral cortex (11) that is and reperfusion are thought to be most vulnerable to ischemic insult. multifactorial and to involve effects Thus, this study focused on the of glutamate, calcium, zinc and nitric changes in BBB function and neu- oxide (NO) as well as other free radi- ronal injury in this particular cerebral cals (1). Upregulation of nitric oxide area. synthase (NOS) is associated with increased production of NO that Material and Methods induces breakdown of the BBB (2). It has been suggested that pretreatment The Uppsala Institutional Review with pharmacological agents that Board for Animal Experimentation reduce excess nitric oxide (3) or approved this prospective, rando- oxidative stress (4) might reduce mized, laboratory animal study. The disruption of BBB permeability (3, 4) piglets were handled according to the caused by ischemia/reperfusion in- guidelines of the Swedish National jury. Methylene blue (MB), a non- Board for Laboratory Animals and toxic dye and also a scavenger, re- the European Convention of Animal cently proved to be a possible aid in Care. Anesthesia was used in all resuscitation from cardiac arrest by surgical interventions. attenuating oxidative, inflammatory, myocardial and neurologic injury (5, Animals 6). The therapeutic effects of MB are ascribed to the presence of the nitric The following inclusion criteria were oxide/cyclic guanylyl mono- applied: No apparent pre-existing phosphate (NO/c-GMP) signaling disease, PaCO2 between 5-5.5 kPa, system and to its action on iron- PaO2 > 10 kPa (75 mm Hg) at base- containing enzymes. Thus, MB has line after stabilization. The animals direct inhibitory effects on nitric included in this study (n=45) were oxide synthases (7) and also blocks 12-14 weeks old and of so-called synthesis of cyclic guanosine mono- triple breed. Piglets were obtained phosphate (cGMP) by inhibiting the from a single provider source. All soluble guanylate cyclase enzyme piglets were kept fasting but had free (8). Due to xanthine oxidase inhibi- access to water during the night, and tion (9), MB also protects against the were taken to the laboratory on the toxic effects of some free oxygen morning of the experiment. radicals (10). The present study was designed to Anesthesia test the possible protective effects of Immediately upon arrival at the la- MB on maintenance of BBB integrity boratory, the animals received an after cardiac arrest and CPR. Fur- intramuscular injection of 6 mg·kg-1 ® thermore, the study tried to identify Zoletil (tiletamine and zolazepam, the potential regulation of nitric oxide Reading, France) combined with 2.2 2

mg·kg-1 Rompum® (xylazine, Bayer, Fluid administration -1 Germany) and atropine 0.04 mg·kg All animals received fluid replace- (Atropin , NM Pharma, Stockholm, ment with acetated Ringer’s solution Sweden). A peripheral venous cathe- (Ringer-acetat®, Frezenius Kabi, ter (18 Gauge) inserted in an ear vein Stockholm, Sweden) by means of an was used for induction and mainten- infusion pump (Life Care , Abbot ance of anesthesia, as well as for Shaw, USA), as follows: 30 mL·kg-1 fluid administration. All the piglets during the first hour of preparation in received an intravenous injection of order to obtain a pulmonary artery ® 20 mg morphine (Morfin Bioglan ). wedge pressure of 7-10 mm Hg, and An injection of 100 mg ketamine a continuous infusion of 10 mL·kg- ® (Ketaminol ,Veterinaria AG, Swit- 1·h-1 thereafter. zerland) was added as an intravenous bolus to achieve surgical anesthesia. Preparation and procedures The absence of motor response to An 18-G arterial catheter was ad- painful stimuli was considered indic- vanced into the aortic arch via a ative of an adequate level of anesthe- branch of the right carotid artery for sia. An intravenous infusion of 8 -1 -1 withdrawal of blood samples and mg·kg ·h sodium pentobarbital measurement of blood pressure. A (Apoteket, Sweden), morphine (Mor- 14-G saline-filled double lumen ca- phine®, Pharmacia, Uppsala, Swe- -1 -1 - theter was placed into the right den) 0.5 mg·kg ·h and 0.25 mg·kg atrium via a cutdown of the right 1·h-1 pancuronium bromide (Pavu- ® external jugular vein to measure right lon , Organon, Netherlands) dis- atrial pressure and for drug adminis- solved in 2.5% glucose solution was tration. This catheter was connected started and used for maintenance of to pressure transducers, which were anesthesia. The piglets were secured calibrated to ambient pressure at the in the supine position and were tra- level of the right atrium. cheotomized, which was done to secure a free airway during the expe- Measurements of hemody- riment. The animals were mechani- cally ventilated (Servo-i V3.1, Sie- namic variables mens Medical, Solna, Sweden) with 30% oxygen in air after preparation. During CPR and after ROSC, hemo- Volume-controlled mechanical venti- dynamic parameters including leads lation was instituted with an I: E ratio II and V5 of the ECG, heart rate of 1:2, a fixed frequency of 25/min (HR), systemic arterial blood pres- and a minute volume set to maintain sure, and right atrial pressure were arterial PaCO2 at an average of 5-5.5 continuously displayed (Solar 8000 kPa. A positive end-expiratory pres- monitor, Marquette Medical Systems, sure (PEEP) of 5 cm H2O was ap- Milwaukee, WI, USA) and recorded plied to prevent atelectasis. The cap- (Workbench 3.0, Strawberry Tree nogram and saturation level were Inc., Sunnyvale, CA, USA) until the displayed continuously until extuba- end of the experiment. tion (CO2SMO Plus-8100, Novame- Samples trix, Wallingford, CT, USA), as were Samples of arterial blood were taken leads II and V5 of the electrocardio- for blood gas analysis and acid-base gram (ECG). balance (ABL 300, Radiometer, Co- 3

penhagen, Denmark) at baseline, and J (Figure 1). If restoration of sponta- at 14, 16 and 19 min during CPR, and neous circulation (ROSC) was not after ROSC at 5, 15, 30, 60 and 120 accomplished, another two defibrilla- min, and before 180 min. Oxygen tory shocks (200 J, 360 J) and a bolus saturation and hemoglobin were injection of epinephrine 20 μg ·kg-1 determined simultaneously on an were administered. DC shocks were OSM3 Hemoximeter (Radiometer, then applied at the same energy level Copenhagen, Denmark) at the same of 360 J during a maximum period of time points. 5 min. CPR was discontinued if ROSC was not achieved during this Experimental protocol time. Restoration of spontaneous circulation was defined as return of After preparation, the piglets were co-ordinated electrical activity result- ventilated with 30% oxygen in air, ing in a systolic blood pressure great- stabilized for one hour, after which er than 60 mm Hg for at least 10 baseline measurements were obtained consecutive minutes. If the arterial before cardiac arrest. Thereafter, pH was less than 7.20 or the base ventricular fibrillation was induced deficit more than 10 mmol/L at 5 min by a 50-Hz, 40-60-V alternating after ROSC, acidosis was corrected with 1 mmol·kg-1 of a tris buffer transthoracic current applied via two ® subcutaneous needle electrodes. Car- mixture (Tribonat , Kabi Fresenius, diopulmonary arrest was defined as a Stockholm, Sweden) and by increas- decrease in aortic blood pressure to ing minute ventilation, aiming at a below 25 mm Hg and the presence of PaCO2 within the range of 5.0 to5.5 kPa. After the resuscitation phase and ventricular fibrillation on the ECG. ® Mechanical ventilation was stopped ROSC, dobutamine (Dobutrex , Eli at this point. After 12 min of un- Lilly Sweden, Solna, Sweden) was administered in a solution of 12.5 treated cardiac arrest, closed-chest -1 -1 CPR was performed with a pneumat- mg/ml starting at 5µg·kg · min to ically driven automatic piston device keep systolic blood pressure above 70 for CPR (Lucas®, Jolife AB, Lund, mm Hg. After completion of the Sweden), and mechanical ventilation study, all animals received an injec- with 100% oxygen was resumed with tion of 10 mL potassium chloride 20 the same ventilatory settings as be- mmol/mL and were sacrificed. Six- fore induction of cardiac arrest. One teen animals served as controls and minute after commencement of CPR received no further interventions all animals received 0.4 U·kg-1 of except anesthesia and placement of vasopressin (Arg8-vasopressin, Sigma catheters. Within 5 min after death, Chemicals Co, St Louis, MO, USA) the skull of all animals was opened as a bolus administered via the right with the animal in the prone position and the brain was rapidly removed atrial catheter. After 8 min of external chest compressions, a monophasic for histopathological analysis. countershock was delivered through To study disruptions of the BBB and defibrillation electrode pads (Med- the brain water and electrolyte con- tronic Physio-Control Corp. Seattle, tent, 38 pigs were divided into three WA, USA) at an energy level of 200 groups. The first group (CA, n=16)

Figure 1 Experimental protocol. After stabilization for 1 h, animals were subjected to 20 min of ventricular fibrillation (VF) including 12 min of cardiac arrest and 8 min of cardiopulmonary resuscitation (CPR). MB (methylene blue); SAL (saline); ROSC (restoration of spontaneous circulation); AC (alternating transthoracic shock); DC (defibril- latory shocks); KCL (potassium chloride).

that served as control group, under- saline (CA-MB group, n=10) or the went cardiac arrest without subse- same dose of saline infusion with 7.5 quent CPR. In some animals the brain mg·kg-1· h-1 methylene blue (Metyl- was removed immediately (n=5) after tioninklorid 10 mg·mL-1 equivalent to cardiac arrest, and in others it was 8.56 mg·mL-1 water free MB, Apote- removed at 15 min (n=5) or 30 min ket, Umeå, Sweden) (CA+MB (n=6) after cardiac arrest. The other group). After ROSC the normal sa- two groups underwent cardiac arrest line infusion was reduced to 16.5 and subsequent CPR and received mL·kg-1· h-1 and the methylene blue either an infusion of saline (CA-MB infusion was reduced to 2.25 mg·kg-1· group, n=10) or an infusion of saline h-1. In these latter groups the brains with MB (CA+MB, n= 12). One were removed for histological analy- minute after the start of CPR the sis in three animals (CA-MB) and animals in the last two groups were four animals (CA+MB), respectively, randomly assigned to receive an at each of the following time points: infusion of 55 mL·kg-1· h-1 normal 30, 60, 180 min after ROSC.

rum, followed by incubation with the Immunohistochemistry, primary antibodies (1: 400 for albu- brain water and electrolyte min). This was followed by incuba- content tion with biotinylated linking antibody and HRP Vector laboratories, Burlingame, Ca, USA), with brief a. Fixation rinses in PBS between incubations. One of the brain hemispheres was The reaction was visualized using immersed in 4% buffered formalin DAB 3,3’-Diaminobenzidine (Vector and stored at 4º C for 1 week. Small Laboratories, Burlingame, CA, tissue pieces (<3x5 mm) were cut USA). Reagent controls (omitting the from the cerebral cortex and primary antibody or substituting processed for histology or immuno- nonimmune serum for the primary histochemistry. The tissue pieces antibody in the staining protocol) on were dehydrated in a graded series of tissue sections were used to confirm alcohol, rinsed in xylene, and embed- the specificity of the primary antibo- ded in low-temperature paraffin (56- dies used (19, 20). The numbers of 58º C) according to a standard proto- albumin-positive cells were counted col (12). Multiple 3-5- μm-thick three times, in a blinded fashion, in sections (6 to 8) were cut from each one identical area of the cortex from tissue block and collected on glass each animal. The median value was slides. After deparaffinization, dupli- used for the final calculation. It is cate sections were stained with either often difficult to differentiate glia Nissl (Cresyl violet) or hematoxylin from neurons, particularly when and eosin using a commercial proto- edema and neural damage alter the col (13). For each animal the number shape of cells. Although most albu- of distorted neurons in one whole min that is present inside cells ap- section was counted at least three pears to be in neurons, a few glial times in a blinded fashion and the cells can also be positive. However, median values were recorded for data albumin leakage into the extracellular analysis. environment is also seen in these cases. Therefore, albumin-positive b. Albumin immunostaining cells appear to represent both damaged neuron and glial cells. Immunostaining of the endogenous albumin to detect leakage across the c. Immunohistochemistry of BBB is a well-standardized method in neuropathological laboratories and NOS has previously been used by several groups (14, 15, 16). Immunohisto- Immunostaining was performed on 3- chemical analysis for albumin was μm thick paraffin sections using a performed on paraffin embedded (3- monoclonal NOS antiserum as de- µm thick) sections using a rabbit scribed earlier (21, 22, 23). In brief, polyclonal anti-human albumin anti- the antibodies of nNOS were diluted body (DAKO, Glostrup, Denmark), 1:5000 and applied for 48 h with and the streptavidin-HRP-biotin continuous shaking at room tempera- technique as described previously ture (10). The immune reaction was (17, 18). Briefly, the endogenous developed using a peroxidase- peroxidase activity was blocked with antiperoxidase technique and visualized in the light microscope. No 3% H2O2 and 10% normal goat se- immunoreactivity was detected in cerebral cortex were cut and stained controls where the primary antibody with Haematoxylin and eosin or Nissl step was omitted. The number of for light microscopy to analyze cellu- nNOS positive cells in each group lar changes. For each animal, the was counted in a blind fashion (10). numbers of distorted neurons in one whole section were counted at least d. Western Blot three times, in a blinded fashion, and the median values were recorded for Protein homogenates of brain sam- data analysis. ples were prepared by rapid homoge- nization in Tissue Extraction Reagent f. Brain water and electrolyte II (Invitrogen Corporation) according content to the manufacturer’s instructions. The protein concentration was deter- Small tissue pieces of pig cerebral mined using the Bio-Rad RC DC kit. cortex were used to measure water Protein extracts (50 g) for each and electrolyte content according to a group were pooled from 3 animals, standard protocol (24). The water resolved by electrophoresis on 12% content was calculated from the dif- SDS-acrylamid gels and transferred ference between dry and wet weights to Hybond-P PVDF membrane (GE of the sample (12), and ion content + + – Healthcare). The membranes were (Na , K and Cl ) was measured from blocked with Phosphate-buffered the dry weight of the samples as saline (PBS) containing 0.2% Tween- described earlier (18, 25). These 20 and 5% non-fat dry milk, prior to measurements were made in slices incubation, overnight at 4oC, with from the same hemisphere of the either anti NOS1 (1:10,000, Euro- brain at the level of the parietal cere- Diagnostica AB), anti-NOS2 bral cortex. To measure water con- (1:10,000, abcam), anti-NOS3 tent, after obtaining the wet weight, (1:200, Santa Cruz Biotechnology) or the samples were placed in an oven, anti–actin (1:200, Sigma) in PBS- maintained at 90 °C, for 72 h or until Tween-20. After washing, mem- three dry weight measurements were branes were incubated (1 h at RT) constant. Water content was calcu- with anti-rabbit IgG-conjugated to lated according to the formula: wet horseradish peroxidase (1:10,000). weight-dry weight/wet weight × 100 After washing, immunoreactive bands (26, 27). To measure ions, after ob- were visualized by enhanced taining the dry weight, the samples chemiluminescence using Lumi- were heated at 550 °C for 24 h to Lightplus(Roche Diagnostics) and produce ash. The ash was dissolved protein band densities were digitally in 5 mL of 3 mmol nitric acid and quantified and normalized to the diluted with deionized water (1: 10). loading control (β actin) using Quan- Na+, K+ and Cl– were determined at tity One software (Biorad). 330, 404 and 512 nm, respectively, with an Atomic Absorption Spectro- e. Cerebral cortex tissue in- photometer (Packard Instruments, jury (evidenced histological- Downers, IL, USA) using an air- ly) acetylene flame (28, 29).

Three-micrometer paraffin sections from identical tissue blocks from the 7

Measurement of nitrite and Bowmann Spectrophotofluorometer nitrate in the cerebral cor- Series 2 (Thermo Elecron Corp, tex of pigs Madison, WI, USA) at excitation wave lengths of 360 nm and 450 nm. The NOx levels were quantified us- a. Sample preparation: ing external standard solutions of Immediately after removal, the brain sodium nitrite or sodium nitrate re- tissue was dissected out at 4° C, then duced with reductase. For nitrite frozen on dry ice and stored at -70 °C measurements, the samples were not until assay. treated with reductase. The brain NO levels were expressed as pmoles/mg b. Tissue protein (33). For sample preparation, the tissue was thawed and homogenized in 20 Statistical analysis mM Tris buffer, 10 mM EDTA (pH 7.4), and centrifuged at 4000 x g for Water content, ion analysis, hemody- 15 min, before the total protein con- namic data were analyzed using tent in cytosol was determined using ANOVA followed by Bonferroni’s Lowry method (30). test for multiple group comparison. The histological and immunohisto- c. Fluorometric Assay of ni- chemical results obtained for all trite and nitrate groups were compared using Kruskal Wallis non-parametrical analysis of variance and if a statistical difference Nitrite and nitrate levels were meas- was detected post-hoc analysis of ured using fluorescence of 2, 3- each group compared to the control diaminonapthotriazole as described group was performed by the Dunn`s earlier (31, 32). The cytosol was multiple comparisons test. filtered to remove hemoglobin Results through a 10 kDa cutoff filter at 15 000 g for 90 min and then processed Out of 45 animals, seven were not for nitrite or nitrate measurement resuscitated, thus leaving 38 to be (33). The NOx levels were measured studied after ROSC. There were no following reduction of nitrate to ni- differences between groups at base- trite with nitrate reductase and line, before cardiac arrest. NADPH regenerating system (G-6- P/G-6-PDH). The complete reaction a.Hemodynamic parameters was developed using nitrate reductase 30mU: NADPS 3 µM; G-6-P 750 During CPR, mean arterial pressure µM; G-6-PDH 48 mU in a total reac- (MAP) increased significantly, but it tion volume of 100 µl. The samples stayed below baseline values at all were incubated at room temperature points (P<0.001 versus baseline). for 90 min. At the end of incubation, Heart rate increased significantly 30 µl DAN reagent (50 µg/ml, 0.62 N after restoration of spontaneous cir- HCl) was added, and after 10 min, 30 culation and then gradually returned µl of 1.4 N NaOH was added to the to baseline values by the end of the incubation mixture. The fluorescence first hour after ROSC. After ROSC, was measured in an Aminco- there was a period of 5-10 min of spontaneous overshoot in arterial 8

blood pressures, in all groups (P the group untreated with MB <0.0001). After that a hypotensive (P<0.01) (Fig.2). period gradually followed in all b. Blood gases Compared to groups, with the lowest levels at 30 pre-arrest values, it was observed that minutes after ROSC (P<0.001).The during the 8 min of CPR the arterial decreased blood pressure lasted ap- blood pH gradually decreased, but proximately 1 h after ROSC, fol- acidemia was not profound in any of lowed by a gradually increased to the three groups. In contrast acidemia baseline levels. Differences between was profound (<7.2) 5 min after the two groups that underwent CPR ROSC and needed correction with were seen at 15 and 30 min after tris buffer mixture and increased ROSC with increased MAP in ventilation. CA+MB group in comparison with

200 CA-MB

CA + M B

** baseline

100 **

M A P(m m H g ) ROSC

CA CPR 5 15 30 60 120 180 m in

Tim e

Figure 2 Mean arterial pressure. The group treated with methylene blue (CA+MB) represented with black circles and the group without MB (CA-MB) denoted with white circles. Significant differences (P**<0.01) between groups were seen at 15 and 30 minutes after return of spontaneous circulation (ROSC). CA-cardiac arrest; CPR cardiopulmonary resuscitation. Albumin immunoreactivity, water, electrolyte content and neuronal injury During the experiment the brain wa- sium content in the CA group ter content increased progressively, (P<0.001), as well as a decrease in with a maximum in all the groups at cerebral potassium content in the the end of the experiment in compari- CA+MB group (P<0.01) at 180 min. son with controls (P<0.001). This The chloride ion simultaneously was accompanied by an increase in accumulated in the CA group. In all brain sodium content in all groups the groups an increase in necrotic (P<0.05) and an increase in potas- neurons and albumin 9

immunoreactivity was demonstrated and potassium, as well as neuronal with increasing duration of ischemia damage, were decreased at every and reperfusion time (P<0.001). The time point (30, 60, 180 min) in the immunohistochemical analysis CA+MB group compared to the CA- indicated less BBB disruption in the MB group. A significant difference animals receiving MB (CA+MB in chloride content between the group), as evidenced by a decreased CA+MB and CA-MB groups was albumin leakage (P=0.004), tissue observed only at 30 min (P<0.01). content of water (P=0.02), sodium The changes in immunohistochemical (P=0.06) and potassium (P=0.04), as staining were assessed using semi- well as decreased neuronal injury quantitative analysis, and the positive (P=0.008) in comparison with the albumin cells were counted in the group that was not treated with MB same region of the cortex (Fig. 4). (CA-MB) (Fig.3). The cerebral contents of albumin, water, sodium

Figure 3. Graphic representations of alterations in blood-brain barrier (BBB) immunoreactivity for albumin (A), brain edema (B), neuronal injury (C) and tissue electrolytes content [(sodium (D), potassium (E), chlor(F)] in cerebral cortex. Untreated animals (CA- MB group) exhibited profound increase in the permeability of BBB for albumin (p=0.004), increase in brain water (p=0.02), and potassium (p=0.04). Semiquantitative data in the same group evidenced significant increase in the number of distorted neurons as a measure of neuronal injury (p=0.008). The groups described were: CA group (cardiac arrest without CPR); CA-MB group (cardiac arrest with CPR without administration of MB) and CA+MB (cardiac arrest with MB administration).

The CA group was sacrificed imme- nerve cell damage was also present in diately after cardiac arrest. There was the subcortical regions, such as the a mild to moderate degree of nerve mid-thalamic nuclei and the brain cell damage in the cortical regions stem reticular formation. The magni- that increased over time, but that was tude of nerve cell damage was most mitigated by the administration of pronounced in brain stem (13). MB (P=0.008). Moderate to severe

b. Nitrate, nitrite and nitrite/nitrate content In general, there was a rapid increase Thus, methylene blue administration in the values of nitrate, nitrite and the reduced both nitrate and nitrite con- nitrite/nitrate content throughout the tent, with a more important reduction experiment.The increase over time in in nitrite resulting in a lower ni- nitrite levels was more pronounced trite/nitrate ratio in the CA+MB than that for nitrate levels. However, group (P=0.02) in comparison with the increase was mitigated in the the CA-MB group. group that was treated with MB (CA+MB) as compared to the group that did not receive MB (CA-MB). Figure 4. Representative examples of albumin immunostaining in the parietal cortex (left panel) and temporal cortex (right panel) of pig brain from control (CAwithout CPR) (a), CA-MB group at 30 min (b) and methylene blue treated animals (CA+MB) (30 min, c). Control group occasional- ly exhibited a few albumin positive cells in the cortex (a). CA-MB 30 min resulted in a massive increase in numbers of albumin positive cells in the cortical areas (b). Many albumin-containing neurons are showing deformed shape and perineuronal edema can be clearly seen in this group (arrows). The neuropil is swollen and sponginess is evident around the damaged neurons in this group (b). CA+MB 30 min showed a marked reduction in albumin positive cells (arrow-heads) on the cerebral cortex. Moreover the shape of neurons is also very similar to that seen in the control group indicating quite good neuroprotection in MB treated group following CA 30 min

(c). Sponginess and edema in the neuropil is much less11 apparent in MB treated CA group (c).

Figure 5. Representative example of neuronal (left panel), inducible (middle panel) and endothelial (right panel) nitric oxide synthase (NOS) expression in the parietal cerebral cortex of pig brains in control (a), cardiac arrest (CA-MB) 30 min (b) and following methylene blue (MB) treated CA group 30 min (CA+MB) (c). Several neurons express constitutive neuronal NOS (nNOS) in the control cerebral cortex (a), whereas only a few cells show inducible NOS (iNOS) expression (arrows). Some neurons in control group exhibited endothelial NOS (eNOS) expression (arrows) together with endothelial cells. CA 30 min markedly upregulated nNOS expression in several neurons in the cortex (b). The nNOS expression is clearly seen in deformed neurons and axons distributed in the neuropil (arrows). Much intense nNOS expression is seen in deformed and damaged nerve cells showing marked perineuronal edema. These neurons are present in areas showing sponginess and edema in the cortex. On the other hand, a distinct increase in iNOS is seen in the similar cortical areas, however, the number of iNOS positive cells is much less compared to nNOS positive cells in the same animal. The pattern of iNOS expression in neurons is also markedly different than nNOS expression (middle panel, arrows). In CA

group, pronounced increase in eNOS immunostaining is seen in several neurons located widespread in the cortical areas (b right panel) compared to control group (a). The eNOS containing cells are deformed and show intense immunoreaction product largely in the neuronal cytoplasm (arrows, right panel, b). MB-treatment profoundly reduced the number of nNOS and iNOS immunostaining in nerve cells in the cortical areas following CA 30 min (c, left and middle panel, respectively). Few neurons in this group show intense NOS immunoreaction product (arrows). The occurrence of perineuronal edema and sponginess of neuropil are much less apparent in MB-treated CA group (c). However, eNOS expression by MB treatment at 30 min is only slightly attenuated (c, right panel). Thus, eNOS expression is seen in many neurons although the intensities of immunoreaction product are much less intense compared to untreated CA group (b, right panel). Interestingly, perineuronal edema and sponginess is much less around the eNOS immunostained neurons in MB- treated CA group (c, right panel). c.NOS activity/expression (Figs. 5,6,7,8) In the control condition there was very little activity/expression of any of the NOS isoforms in the cerebral cortex indicated by immunohistochemistry. Experimental global ischemia from cardiac arrest rapidly led to the upregulation of the nNOS and iNOS isoforms. After only 5 min of untreated CA, significant nNOS expression was obvious. Thus, nNOS activation increased over time in all groups, resulting in an increased number of activated cells in comparison with iNOS (p<0.001). In contrast, there were no changes in eNOS expression over any of the time points studied. MB mitigated iNOS and nNOS upregulation at 30 min after ROSC, and thus less iNOS (p<0.01) and nNOS (p<0.01) activation was seen at this time point in the group that was treated with MB in comparison with the CA-MB group. However, MB appeared to attenuate nNOS and iNOS activity only during the first 30 min after ROSC, i.e. during the time of administration; no effects were observed later. In contrast to this, in the group with MB we observed by immunohistochemistry an activation of eNOS not only at 30 min (p=0.02), but also at all time points until the end of the experiment (P<0.01).

Figure 6. NOS activity in cerebral cortex. Panel A shows iNOS activity; panel B represents nNOS activity; panel C shows eNOS expressions. Data are expressed as means± SD. The gray bars represent the control group, where NOSs activation was determined at baseline, and at 5, 15, and 30 min after cardiac arrest. The white bars represent the CA-MB group and the black bars the CA+MB group. In these groups NOSs expressions were obtained at 30, 60, and 180 min after ROSC.

Ischemic neurons tend to pick up all kind of stainings. Accordingly, we have done a few of control experiments in order to fully appreciate the presented data. Therefore the immunohistochemical (IHC) stainings included a negative control where the primary antibody is omitted in order to exclude unspecific binding of the secondary antibody, and a positive control (Figure 7).

Figure 7. Negative and positive control images for albumin and NOS staining. In negative control (antibodies are omitted) a very clean image represent specific staining The positive control represents 5 min controls (cardiac arrest without CPR) showing some mild immuno for NOS and albumin but this is much negligible than experimental groups.

NOS activity was also investigated with Western blot analysis . In comparison with immunohistochemical analysis of NOS that were performed only from cerebral cortex, these Western blot analysis investigated NOS activity from different brain regions (Figure 8). Western blotting of superficial cerebral tissues pooled from all animals at respective time points revealed results which varied somewhat from the the immunohistochemical analysis for nNOS and iNOS but reasonably consistent with microscopical evaluation of immunostained sections of cerebral cortex in the 14

temporo-parietal region. No increase in eNOS protein expression could be shown using western blot technique.

30 min 60 min 180 min

50 Control CA-MB CA+MB CA-MB CA+MB CA-MB CA+MB 45 eNOS 40 35 30 CA β- 25 CA-MB actin 20 CA+MB 15 10

Adjusted normalizedAdjusted volume 5 0 control 30 min 60 min 180 min

50 45 30 60 180 min 40 min min 35 Control CA-MB CA+MB CA-MB CA+MB CA-MB CA+MB 30 CA nNOS 25 CA-MB 20 CA+MB β- 15 10

actin normalizedAdjusted volume 5 0 control 30 min 60 min 180 min

iNOS 30 60 180 min min min Control CA-MB CA+MB CA-MB CA+MB CA-MB 50 CA+MB 40 iNOS CA 30 CA-MB β- 20 CA+MB

actin 10

Normalized volumeadjusted 0 control 30 min 60 min 180 min

Figure 8. Expression of eNOS, nNOS and iNOS by western blot analysis. Protein extracts of superficial cerebral tissue from 3 animals were pooled in control, CA-MB (white bars) and CA+MB (black bars) groups at 30, 60 and 180 min after ROSC respectively. The expression of ß-actin was determined as an internal control (A). Relative density analysis showed in the bar graphs for eNOS, nNOS and iNOS respectively, was expressed as a ratio to ß-actin after background subtraction (B)

Discussion early as during the initial phase of an untreated as well as a treated cardiac It has long been accepted that global arrest and progressed throughout the cerebral anoxia induced by circulato- 3 h observation period. MB adminis- ry arrest in humans can result in neu- tration during CPR and the initial rological damage and neurological phase after ROSC reduced these dysfunction after only a 5-min car- harmful effects considerably, but did diac arrest and subsequent restoration not, on the other hand, reverse this of spontaneous circulation (34). The ongoing detrimental process. present results provide experimental We have previously provided histo- evidence for leakage of albumin, pathological proof in a porcine model water and electrolytes into the brain of BBB disruption as early as 5 min parenchyma after cardiac arrest and after untreated cardiac arrest (13). In ROSC. The BBB disruption began as the present investigation evidence of neurological injury was recorded 15 15

min after ROSC, and the magnitude HSD was administered. Therefore the of this injury increased with increas- neuroprotective effects must be as- ing time following the initial ischem- cribed to the MB, as the study proto- ic insult. cols in these studies were otherwise Evidence of widespread microvascu- identical. In our study, counting the lar and neuronal damage 4 h after distorted neurons in parietal-temporal cardiac arrest and ROSC has pre- cortex was used to compare the se- viously been documented in infant verity of post cardiac arrest induced piglets (35). In those studies it was brain damage. Findings by Radovsky considered that young animals might et al (43), however, indicate that be more prone to a delayed increase overall brain damage, rather than in BBB permeability after cardiac damage in any one specific region, is arrest and CPR (36, 37). Maximal of most significance for the continued disruption of BBB has been reported life of the experimental animal with at 3 hours after reperfusion, with cortical injury (43). Based on these normalization 24 h after ROSC (38). and previous findings it seems that However, some studies have indi- MB itself easily passes the BBB, cated that BBB postischemic distur- since after systemic administration bances are of a biphasic nature (39, this dye crosses the BBB and 40) and that disruption of the BBB selectively stains the brain tissues does not increase with extended pe- (45). Thus, in this situation MB has a riods of ischemia and recirculation much higher concentration in the (41). The discrepancy with the brain tissue than in the circulation present results might be due to differ- (44). This effect was demonstrated by ences between various animal mod- measuring the distribution of MB els, different methods for assessing with high-performance liquid chro- BBB function and the different time matography (44), as well as by the points of investigation. In any case, obvious blue tint of the animal’s there were no indications in the brain when removed 5 min after it present study that leakage through the was sacrificed (34). BBB was not a continuous process. Systemic hypertension that occurs This finding of an early leakage after immediately after successful CPR cardiac arrest is also supported by increases cerebral blood flow as studies using Evans blue (13). This demonstrated previously by our dye has been shown to represent group by means of the Laser Doppler vascular protein leakage macroscopi- technique (5, 6) and positron emis- cally (42). sion tomography (11). When BBB Previous research from our group has permeability is increased, it allows suggested that MB co-administered extravasation of albumin and other with hypertonic saline dextran may high and low molecular weight com- be an effective mechanism for atten- pounds (46) into the extracellular uation of oxidative, inflammatory and compartment of the brain, leading to neurological injuries (5, 6). MB may vasogenic edema formation and sub- contribute to those effects by inhibit- sequent cell injury . In contrast to this ing NO production as well as reduc- kind of reasoning the vasogenic ede- ing effects of excess NO, and/or by ma that results after the systemic scavenging of oxygen-derived free hypertensive response has been radicals (5, 6). In the present study no claimed to be a major contributor to 16

the BBB disruption (35, 47). In this an inhibitory effect on NO production context it may be noted that greater and its subsequent oxidation. blood pressure in the group receiving All three NOS isoforms were ex- MB in comparison with the group not pressed after global ischemia follow- receiving MB did not increase BBB ing cardiac arrest. However, both permeability. On the contrary, lower immunohistochemical and Western albumin leakage, neuronal injury and blot analysis confirmed previous water content was demonstrated in studies demonstrating iNOS tran- the group treated with MB. The lack scriptional induction occurs transcrip- of significant increase in water con- tionally long after activation of post- tent and albumin in the MB treated transcriptionally activated eNOS and group is consistent with the lack of an nNOS, suggesting that iNOS may not increase in BBB permeability. The actively contribute to early cerebral rationale for our finding would be injuries (51). Thus, it is generally that the magnitude of the blood pres- agreed that action by nNOS is re- sure increase after ROSC cannot be sponsible for the formation of NO associated with the BBB disruption that initially is harmful to the brain when MB had been administrated. (52). There are no previous data on The hyperemia that occurred after the relative inhibition of the NOS ROSC has been suggested to cause isoforms by MB. Earlier studies sug- disruption of the BBB which ulti- gest that the large NO tissue content mately may worsen neurological produced a few hours after ROSC by injury (3). These effects, and cerebral activation of iNOS is detrimental, hyperemia observed after global and that the considerably smaller ischemia have been suggested to be tissue content of NO derived from caused by high levels of NO pro- eNOS is beneficial (53) due to its duced during ischemia and early vasodilating effects after cerebral reperfusion (48). Thus, there is strong ischemia. Hence, it was also consi- evidence that NO is involved in the dered that the beneficial effects of mechanisms of neurotoxicity after non-selective inhibitors were limited, cerebral ischemia (49). Hence, it because they inhibit eNOS to a simi- seems logical to speculate that the lar extent (51) and may therefore mechanism behind less BBB disrup- aggravate the effects of brain ische- tion in the CA+MB group occurred mia. MB was considered a non- as a result of NOS and guanylate selective inhibitor. However, the cyclase inhibition by MB. It is note- interpretation of this data requires worthy that this effect occurred in caution in the light of other studies spite of the limited duration of the that describes dose-related neurotox- MB effect on nNOS and iNOS acti- icity of methylene blue on central vation. Nitrite and nitrate are stable nervous system (54) for higher doses metabolites of NO, accessible to (between 5 mg/kg and 50 mg/kg), quantitative analysis. Therefore, than the doses used in the present measurement of nitrite and nitrate study and also possible interspecies turned out to be the most suitable and differences (54). In this study the practical method for assessing NO neuroprotection has been assessed in synthesis (50). MB decreased the terms of BBB function and cerebral nitrite/nitrate content, demonstrating water content, together with histology. The limitation of the present 17

research is that the potential of neu- dismutase and catalase in piglets. roprotection concerning neurologic Stroke 1994;25:1830-4 function in terms of electrophysiolo- 5. Miclescu A, Basu S, Wiklund L: gy or in behavior could not be fully Methylene blue added to a hyper- assessed. tonic hyperoncotic solution increases short-term survival in ex- In conclusion: perimental cardiac arrest. Critical Care Medicine 2006;34:2806-13 The present results unequivocally 6. Miclescu A, Basu S, Wiklund L: demonstrate that MB markedly re- Cardio-cerebral and metabolic duced BBB disruption and subse- effects of methylene blue in quent neurological injury after hypertonic sodium lactate during ischemia/reperfusion following car- experimental cardiopulmonary diac arrest. Accordingly, this ap- resuscitation.Resuscitation 2007; proach is worth testing in human 75:88-97 cerebral ischemia and reperfusion 7. Mayer B, Brunner F, Schmidt K: after cardiac arrest. Inhibition of nitric oxide synthesis by methylene blue. Biochem Acknowledgements Pharmacol 1993;45:367-74 The authors are grateful to Anders 8. Mayer B, Brunner F, Schmidt K: Nordgren, Monika Hall, Elisabeth Novel actions of methylene blue Pettersson for excellent technical Eur Heart J 1993;14:22-6 assistance. 9. Salaris SC, Babs CF, Voorhees III WD: Methylene blue as an inhibitor of superoxide genera- References: tion by xanthine oxidase. A potential new drug for the attenua- tion of ischemia/reperfusion in- 1. Kaur C, Ling EA: Blood brain jury. Biochem Pharmacol 1991; barrier in hypoxic-ischemic con- 42:499-506 ditions. Curr Neurovasc Res 10. Kelner MJ, Bagnell R, Hale B, et 2008;5:71-81 al: Potential of methylene blue to 2. Kubes P, Granger DN: Nitric block oxygen radical generation oxide modulates microvascular in reperfusion injury. Basic Life permeability. Am J Physiol 1992; Sci 1988;49:895-8 262: H611-H615 11. Mörtberg E, Cumming P, Wik- 3. Schleien CL, Kuluz JW, Gelman lund L: A PET study of regional B: Hemodynamic effects of ni- blood flow after experimental tric oxide synthase inhibition be- cardiopulmonary resuscitation. fore and after cardiac arrest in in- Resuscitation 2007;75:98-104 fant piglets. Am J Physiol Heart 12. Sharma HS, Cervos-Navarro J: Circ Physiol 1998;274:H1378- Brain oedema and cellular H1385 changes induced by acute stress 4. Schleien CL, Eberle B, Shaffner in young rats. Acta Neurochir DH, et al: Reduced blood-brain Suppl (Wien) 1990;51:383-6 barrier permeability after cardiac 13. Wiklund L, Sharma HS, Basu S: arrest by conjugated superoxide Circulatory Arrest as a Model for Studies of Global Ischemic In- 18

jury and Neuroprotection. Ann 21. Sharma HS, Westman J, Alm P, NY Acad Sci 2005;1053:205–19 et al: Involvement of nitric oxide 14. Salahuddin TS, Kalimo H, Jo- in the pathophysiology of acute hansson BB, et al: Observations heat stress in the rat. Influence of on exudation of fibronectin, fi- a new antioxidant compound H- brinogen albumin in the brain af- 290/51. Ann NY Accad Sci1997; ter carotid infusion of hyperos- 813:581-90 molar solutions. An immunohis- 22. Sharma HS, Alm P, Westman J: tochemical study in the rat in Upregulation of hemeoxygenase- longlasting changes in the brain II in the rat spinal cord following microenvironment and multifoc- heat stress. In: Nielsen- Johans- al nerve cell injuries. Acta Neu- sen B, Nielsen R (eds) Thermal ropathol (Berlin) 1988;76:1–10 physiology 1997. The August 15. Pettersson CA, Salahuddin TS, Krogh Institute, Copenhagen Olsson Y: Presence of plasma 135-8 proteins in spinal nerve roots. An 23. Sharma HS, Alm P, Westman J: immunohistochemical study in Nitric oxide and carbon monox- the rat. Acta Neurol Scand 1990; ide in the pathophysiology of 82:21–7 brain functions in heat stress. In: 16. Sharma HS: Pathophysiology of Sharma HS, Westman J (eds) the blood–spinal cord barrier in Brain functions in heat environ- traumatic injury. In: HS Sharma ment Prog Brain Res 1998;115: and J. Westman, Editors, The 297-333 blood–spinal cord and brain 24. Kiyatkin EA, Brown PL, Sharma barriers in health and disease, HS: Brain edema and breakdown Elsevier, Academic Press, San of the blood-brain barrier during Diego 2004; pp. 437–518 methamphetamine intoxication: 17. Sharma HS, Kretzschmar R, critical role of brain hyperther- Cervos-Navarro J, et al: Age- mia. Eur J Neuroscience 2007; related pathophysiology of the 26 :1242–53 blood–brain barrier in heat 25. Joo F, Szucs A, Csanda E: Me- stress. Prog Brain Res 1992;91: tiamide treatment of brain øede- 189–196 ma in animals exposed to 90 yt- 18. Sharma HS: Pathophysiology of trium irradiation. J Pharm blood–spinal cord barrier in Pharmacol 1976;28:162–3 traumatic injury and repair. Curr 26. Rapoport SI: Blood–Brain Bar- Pharmacol Des 2005;11:1353-89 rier in Physiology and Medicine. 19. Gordh T, Chu H, Sharma HS: 1976; Raven Press, New York Spinal nerve lesion alters blood– 27. Sharma HS: Methods to produce spinal cord barrier function and brain hyperthermia. Curr Prot activates astrocytes in the rat. Toxicol 2005;14:1–24 Pain 2006; 124: 211–21 28. Dux E, Temesvari P, Szerdahelyi 20. Sharma HS, Ali SF: Alterations P, et al: Protective effect of anti- in blood–brain barrier function histamines on cerebral edema in- by morphine and amphetamine. duced by experimental pneumo- Ann N Y Acad Sci 2006;1074: thorax in newborn piglets. Neu- 198–224 roscience 1978;22:317–21

29. Sharma HS, Westman J: The 38. Betz AL, Dietrich WD: Blood- Blood-Spinal Cord and Brain brain dysfunction in cerebral Barriers in Health and Disease. ischemia. In Ginsberg MD, Bo- 2004 Academic Press, San Diego gousslavsky J: Cerebrovascular 30. Lowry OH, Rosebrough NJ, Farr disease: Pathophysiology, diag- AL, et al: Protein measurement nosis and management. Black- with the folin phenol reagent. J well Science 1998;1:359 Bio Chem 1951; 193:265-75 39. Mossakowski MJ, Lossinsky AS, 31. Sharma HS, Olsson Y, Dey PK: Pluta R, et al: Abnormalities of Early accumulation of serotonin the blood-brain barrier in global in rat spinal cord subjected to cerebral ischemia in rats due to traumatic injury. Relation to experimental cardiac arrest. Acta edema and blood flow changes. Neurochir Suppl (Wien) 1994; Neuroscience 1990;36:725-30 60:274-6 32. Misko TP, Schilling RJ, Salve- 40. Preston E, Sutherland G, Finsten mini D, et al: A fluorometric as- A: Three openings of the blood- say for the measurement of ni- brain barrier produced by fore- trite in biological samples. Anal brain ischemia in the rat. Neu- Biochem 1993;214:11-6 rosci Lett 1993;149:75-8 33. Rao AM, Dogan A, Hatcher JF, 41. Pluta R, Lossinsky AS, Wis- Dempsey RJ. Fluorometric assay niewski HM, et al: Early blood- of nitrite and nitrate in brain tis- brain barrier changes in the rat sue after traumatic brain injury following transient complete ce- and cerebral ischemia. Brain Res rebral ischemia induced by car- 793: 265-70 diac arrest. Brain Res 1994; 34. Wiklund L, Basu S, Miclescu A, 633:41-52 et al: Neuro- and cardioprotec- 42. Clasen RA, Pandolfi S, Hass tive effects of blockade of nitric GM: Vital staining, serum albu- oxide action by administration of min and the blood-brain barrier. methylene blue. Ann NY Acad J Neuropathol Exp Neurol 1970; Sci 2007;1122:231-44 29:266-84 35. Caceres MJ, Schleien CL, Kuluz 43. Radovsky A, Safar P, Sterz F: JW: Early endothelial damage Regional prevalence and distri- and leukocyte accumulation in bution of ischemic neurons in piglet brains following cardiac dog brains 96 hours after cardiac arrest. Acta Neuropathol (Berl) arrest of 0 to 20 minutes. Stroke 1995;90:582-91 1995;26:2127-34 36. Schleien CL, Koehler RC, 44. Peter C, Hongwan D, Kupfer A, Shaffner DH: Blood-brain bar- et al: Pharmacokinetiks and rier integrity during cardiopul- organ distribution of intravenous monary resuscitation in dogs. and oral methylene blue. Eur J Stroke 1990;21:1185-91 Clin Pharmacol 2000;56:247-50 37. Schleien CL, Koehler RC, 45. O`Leary JL, Petty J, Harris AB, Shaffner DH, et al: Blood-brain et al: Supravital staining of barrier disruption following CPR mammalian brain with intra- in piglets. Acta Neurochir Suppl arterial methylene blue followed (Wien) 1990;51:213-5 by pressurized oxygen. Stain Technol 1968;43:197-20 20

46. Nelson CW, Wei EP, Povlishock JT, et al: Oxygen radicals in cerebral ischemia. Am J Physiol 1992; 263: H1356-H1362 47. Mayhan WG, Heistad DD: Role of veins and cerebral venous pressure in disruption of the blood-brain barrier. Circ Res 1986;59:216-20 48. Gally JA, Montague PR, Reeke Jr, et al: The NO hypothesis: possible effects of a short-lived, rapidly diffusible signal in the development and function of the nervous system. Proc Nat Acad Sci USA 1990;87:3547-51 49. Moro MA, Cárdenas A, Hurtado O, et al: Role of nitric oxide after brain ischaemia. Cell Calcium 2004;36:265–75 50. Tsikas D: Methods of quantitative analysis of the nitric oxide metabolites nitrite and nitrate in human biological fluids. Free Radical Research 2005;39:97- 815 51. Murphy S, Gibson CI: Nitric oxide, ischaemia and brain in- flammation. Biochem Soc Trans- actions 2007;35:1133-7 52. Won SJ, Kim DY, Gwag BJ: Cellular and molecular pathways of ischemic neuronal death. J Bi- ochem Mol Biol 2002;35:67-86 53. Chan PH: Reactive oxygen radicals in signaling and damage in the ischemic brain. J Cerebral blood flow metabolism 2005;21:2-14 54. Vutskits L, Briner A, Klauser P, et al: Adverse effects of methylene blue on the central nervous system. Anesthesiology 2008;108:684-92