Inhaled NO prevents hyperoxia-induced white matter damage in neonatal rats

Experimental Neurology 252 (2014) 114–123

Contents lists available at ScienceDirect

Experimental Neurology journal homepage: www.elsevier.com/locate/yexnr

Inhaled NO prevents hyperoxia-induced white matter damage in neonatal rats Hoa Pham a,b,c, Gaelle Vottier a,b,c, Julien Pansiot a,b,c, Sy Duong-Quy d, Bieke Bollen a,b,c,e, Jérémie Dalous a,b,c, Jorge Gallego a,b,c, Jean-Christophe Mercier b,f, Anh Tuan Dinh-Xuan d, Philippe Bonnin b,g,h, Christiane Charriaut-Marlangue a,b,c, Olivier Baud a,b,c,i,⁎ a

INSERM, UMR 676, 75019 Paris, France Université Paris Diderot, UFR de médecine Denis Diderot, Sorbonne Paris Cité, 75010 Paris, France c PremUP foundation, 75014 Paris, France d Assistance Publique-Hôpitaux de Paris, Université Paris Descartes, Hôpital Cochin, Service de Physiologie, 75014 Paris, France e University of Leuven, Laboratory of Biological Psychology, Leuven, Belgium f Assistance Publique-Hôpitaux de Paris, Université Paris Diderot, Sorbonne Paris Cité, Hôpital Robert Debré, Pediatric emergency department, 75019 Paris, France g INSERM, UMR 965, 75010 Paris, France h Assistance Publique-Hôpitaux de Paris, Université Paris Diderot, Sorbonne Paris Cité, Hôpital Lariboisière, Physiologie Clinique–Explorations Fonctionnelles, 75010 Paris, France i Assistance Publique-Hôpitaux de Paris, Université Paris Diderot, Sorbonne Paris Cité, Hôpital Robert Debré, Neonatal intensive care unit, 75019 Paris, France b

a r t i c l e

i n f o

Article history: Received 9 August 2013 Revised 23 November 2013 Accepted 26 November 2013 Available online 7 December 2013 Keywords: Nitric oxide White matter damage Hyperoxia Developing brain Neuroprotection

a b s t r a c t White matter damage (WMD) and bronchopulmonary dysplasia (BPD) are the two main complications occurring in very preterm infants. Inhaled nitric oxide (iNO) has been proposed to promote alveolarization in the developing lung, and we have reported that iNO promotes myelination and induces neuroprotection in neonatal rats with excitotoxic brain damage. Our hypothesis is that, in addition to its pulmonary effects, iNO may be neuroprotective in rat pups exposed to hyperoxia. To test this hypothesis, we exposed rat pups to hyperoxia, and we assessed the impact of iNO on WMD and BPD. Rat pups were exposed to either hyperoxia (80% FiO2) or to normoxia for 8 days. Both groups received iNO (5 ppm) or air. We assessed the neurological and pulmonary effects of iNO in hyperoxia-injured rat pups using histological, molecular and behavioral approaches. iNO significantly attenuated the severity of hyperoxia-induced WMD induced in neonatal rats. Specifically, iNO decreased white matter inflammation, cell death, and enhanced the density of proliferating oligodendrocytes and oligodendroglial maturation. Furthermore, iNO triggered an early upregulation of P27kip1 and brainderived growth factor (BDNF). Whereas hyperoxia disrupted early associative abilities, iNO treatment maintained learning scores to a level similar to that of control pups. In contrast to its marked neuroprotective effects, iNO induced only small and transient improvements of BPD. These findings suggest that iNO exposure at low doses is specifically neuroprotective in an animal model combining injuries of the developing lung and brain that mimicked BPD and WMD in preterm infants. © 2013 Elsevier Inc. All rights reserved.

Introduction Bronchopulmonary dysplasia (BPD) and white matter damage (WMD) are the two most common complications of preterm birth. Their incidence reaches 20 to 40% of infants born below 28 weeks of gestation (Jobe, 2011; Volpe, 2009). Whereas advances in neonatal intensive care have resulted in a marked decrease in neonatal mortality,

Abbreviations: BPD, bronchopulmonary dysplasia; WMD, white matter damage; iNO, inhaled NO; RAC, radial alveolar count; MLI, mean linear intercept; CC3, cleaved caspase 3. ⁎ Corresponding author at: INSERM U676, Hôpital Robert Debré, 48 blvd Sérurier, F75019 Paris, France. Fax: +33 1 40 03 19 95. E-mail address: [email protected] (O. Baud). 0014-4886/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.expneurol.2013.11.025

the incidence of both morbidities in these infants remains high. Several common risk factors, such as low gestational age, perinatal inflammation and excessive oxidative stress make more likely the devastating co-occurrence of BPD and WMD (Dammann et al., 2004). In addition, BPD increases the risk for cerebral palsy (CP), which is a common consequence of WMD in infants born prematurely. Among these common risk factors, an excessive release of free radicals induced by oxygen therapy in preterm neonates with respiratory distress was involved in the pathogenesis of WMD as well as BPD. Furthermore, recent experimental studies have supported that hyperoxia causes oxidative stress and triggers maturation-dependent cell death, maturation arrest of developing oligodendrocytes, and disruption of axon–oligodendrocyte integrity, all key features of WMD (Back et al., 2007; Gerstner et al., 2008; Ritter et al.,

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

2013; Schmitz et al., 2011; Vottier et al., 2011). For these reasons, the design of therapeutic strategies that take into account both BPD and WMD is a highly relevant, to date poorly explored challenge. In the present study, we focused on the therapeutic use of nitric oxide (NO) in neonates. NO is widely recognized as an important messenger and effector molecule in a variety of acute and chronic inflammation systems, and also as a mediator of vascular tone and tolerance to damage (Moncada et al., 1991; Toda and Okamura, 2003; Vaucher et al., 2000). Interestingly, NO is both a physiological mediator of the central nervous system and a key factor for lung angiogenesis and alveolarization, two developmental phenomena involved in BPD pathophysiology (Jobe, 2011; Yun et al., 1996). On the other hand, we have recently reported that inhaled NO (iNO) has several critical properties in the developing brain by promoting myelination and inducing neuroprotection against excitotoxic-induced brain injury and neonatal stroke (Charriaut-Marlangue et al., 2012, 2013; Olivier et al., 2010; Pansiot et al., 2010). Interestingly, similar results were reported in preclinical model of adult stroke (Terpolilli et al., 2012, 2013). Thus, iNO appears to be a promising candidate for the prevention and/or the clinical management of WMD and BPD, but remains to be evaluated in an animal model mimicking simultaneously these two developmental diseases. Here, we used a rat model of prolonged postnatal hyperoxia that induced BPD and WMD. We made the hypothesis that iNO, in addition to its potential effects on lung injury, may be neuroprotective in rat pups exposed to hyperoxia. The present results showed that iNO induces neuroprotection with a significant effect at both histopathological and behavioral levels, whereas it had only transient and mild effects on BPD. Material and methods Experimental protocol and gas exposure This study was approved by the National Institute of Health and Medical Research and complied with the instructions of the Institutional Animal Care and Use Committees INSERM 676—Paris. The day before delivery, pregnant rats (Sprague–Dawley, Janvier S.A.S., Le Genest-St-Isle, France) were placed in a transparent Plexiglas chamber supplied with a gas mixture that either induced hyperoxia (FiO2 = 80 +/− 0.5%) until postnatal day (P)7 or maintained normoxia (FiO2 = 21 +/− 0.5%). Hyperoxia exposure of pregnant rats began the day before delivery to make sure that all rat pups were placed under hyperoxic condition immediately after birth. Adult rats were switched every 24 h between O2 exposed and room air-exposed litters. Oxygen concentration was monitored using a Proox (Biopherix, USA). CO2 concentration was consistently kept under 0.1% using soda lime (Intersurgical, France). To investigate the impact of exogenous NO on the developing brain and lung, iNO at low concentration (5 ppm) was introduced in chambers from embryonic day (E)21 to P7 and monitored using iNOvent system (INOTherapeutics, Clinton, NJ). NO2 concentration was kept under 1 ppm. Low concentration of NO (5 ppm) was used according to current use in neonatal intensive care units and because a higher dose (20–40 ppm) would not be feasible for a protracted exposure. From P7, rat pups and their mothers in all experimental groups were kept in room air. Animals were housed under controlled temperature (22 ± 1 °C) and light conditions (12 h day/night cycle) with food and water ad libitum. Neonatal mortality was checked daily. Blood gas analysis Blood gas was analyzed in hyperoxic and control rat pups using a clinical blood gas analyzer (ABL 80, Radiometer). Pups were decapitated and blood samples collected at various time points (1 h, 12 h, 3 days and 7 days of life) from the neck in heparinized capillary tubes, and gases measured immediately.

115

Ultrasound imaging Thermoregulated rat pups (at P1 and/or P6, n = 5 per group) were subjected to ultrasound measurements under isoflurane [0.5% inhalation via a facemask in (O2/N2O) (1:3)] anesthesia using an echograph (Voluson i, GE Heatlthcare, Aulnay-sous-bois, France) equipped with a 12-MHz linear transducer (Bonnin et al., 2011). Heart rate and spatialaveraged-time-averaged mean blood-flow velocities (mBFV) were measured in the intracranial carotid arteries (ICA) and the basilar trunk (BT, when detectable) 1 day (P1) and 6 days (P6) after air, iNO (5 ppm), hyperoxia, and hyperoxia in combination with iNO exposure. Heart rates reflected changes in cardiac output, as the left ventricular ejection volume is quite invariable in newborns. Tissue preparation Pups (n = 7–9/group for each staining protocol) were sacrificed at P3, P10 or P21. The left lung was washed in physiological serum, fixed with 4% paraformaldehyde in PBS, pH 7.4, for 24 h, and paraffinembedded. Symmetrical 5 μm-thick sections were cut from the hilum to the pleural surface. For brain histology studies, two distinct fixation protocols were used: (1) brains were directly removed and immersed in 4% formaldehyde and embedded in paraffin, (2) pups were perfused transcardially with 4% paraformaldehyde in phosphate buffer (PB 0.24 M, pH 7.4). Brains were equilibrated with 10% sucrose in PB for 2–4 days, frozen in liquid nitrogencooled isopentane, stored at −80 °C and cut coronally into serial 10 μm-thick sections. Lung alveolarization assessment All sections were stained with hematoxylin and eosin, and alveolarization was assessed by performing radial alveolar counts (RAC) and median linear intercepts (MLI) as previously described (Dunnill, 1962; Emery and Mithal, 1960; Thurlbeck et al., 1970). Images of each section were captured with a magnified digital camera through a Leica microscope and were saved as PICT.jpg files. At least ten counts were performed per animal and 6–8 animals were used for each experiment. Immunohistochemistry Primary antibodies used in this study are listed in Supplemental Table S1. All quantification of immuno-reactive cells was carried out by investigators blind to the experimental groups. Lung angiogenesis was evaluated by measuring pulmonary vessel number and pulmonary vascular volume density (Vv), using factor VIII [von Willebrand Factor (vWF)] as an endothelial marker. Pulmonary vessel number was determined by counting microvessels (20–80 μm) stained with vWF in each high power field (100X magnification). Vv was measured by superimposing a grid of 100 points onto color photomicrographs (400× magnification) of ten random noncontiguous fields per animal. Vv was calculated as the ratio of the number of points coinciding with vWFpositive sites to the number of points on lung parenchyma (excluding large vessels and airways). For brain immunohistochemistry, coronal sections (+ 1.44 to − 0.48 mm from bregma) were selected and processed as previously described (Olivier et al., 2005). In each experimental group, we studied 7–8 pups in three separate experiments. Immunolabeling was visualized using the streptavidin–biotin–peroxydase method. Doublelabeling was performed with secondary antibodies coupled to the green fluorescent marker Fluoroprobe S488 (Interchim, Montluçon,

116

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

France) or the red fluorescent marker cyanine 3 (Jackson Immunoresearch laboratories, West Grove, PA). Optical density of myelinated and axonal fibers The optical density (OD) of MBP-stained and axonal fibers was measured in 2 different regions (cingulum and lateral corpus callosum) as previously described (Fontaine et al., 2008), using a computerized image-analysis system (Image J 1.41o, NIH, USA) that reads optical density as gray levels. Four sections per brain were examined for each animal at P10. Non-specific background densities were measured at each brain level in a region devoid of MBP and axonal fibers immunostaining and were subtracted from values of the region of interest. TUNEL staining DNA fragmentation in the white matter at P3 was detected using TUNEL staining (In Situ Cell Death Detection Kit, Roche, Meylan, France) according to the manufacturer's instructions. Labeled cell nuclei were counted in 4 sections for each hemisphere on 40× images. Western blotting Snap-frozen brain cortices including white matter were subjected to different protocols to separate nuclear proteins, cytosolic proteins and mitochondrial proteins. Proteins (40 μg) were separated on 10% polyacrylamide gels, using Ponceau-S staining and β-actin as the loading standard for cytosolic proteins, and histone-3 as the loading standard for nuclear proteins. Bands were transferred to membranes, which were incubated in primary antibodies diluted in blocking buffer (rabbit polyclonal anti-AIF, rabbit polyclonal anti-cleaved caspase-3 or mouse monoclonal anti-cytochrome C) overnight at 4 °C. They were then incubated with a secondary antibody (HRP-linked anti rabbit or anti mouse: 1/5000) 1.5 h at room temperature, and positive signals visualized using enhanced chemiluminescence (Bio-Rad). Western blots were performed in triplicate using 4–6 animals per group. RT-PCR experiment Total mRNA was isolated using the RNeasy Lipid Tissue kit (Qiagen, France) according to the manufacturer's instructions and as previously reported (Fontaine et al., 2008). The cDNA was produced from 600 ng of total mRNA by reverse transcription with the Iscript™ kit (Bio-Rad). Sequences for the forward and reverse oligonucleotide primers, designed with the Primer Express software package (Applied Biosystems), are listed in Supplemental Table S2. Each experiment was run twice with 8 animals per group, and in both cases measurements were carried out in triplicate.

Statistical analysis Values are expressed as the mean ± SEM and/or mean ± SD (for mBFVs). Statistical comparisons between normoxic control and exposed groups were performed using one-way ANOVA test followed by the Dunnett's multiple comparison tests. Pairwise comparisons between groups with or without iNO were done using Mann–Whitney non parametric test. Statistical tests were run on Prism version 5.01 (Graphpad Software, San Diego, CA) and Statview version 5.0 for Mac Os (Abacus Concepts, Berkeley, CA). Results Body, lung and brain phenotypes induced by postnatal hyperoxia Postnatal hyperoxia resulted in a pronounced decrease in body weight at both P3 and P10 (Supplemental Table S3), as well as increased postnatal mortality rates overall (15.5% vs. 1% in controls, p b 0.001). Blood gas analyses performed showed that blood concentrations of pO2 and fraction of oxyhemoglobin (fO2Hb) were significantly elevated no longer than 1 h after the onset of hyperoxia exposure (Supplemental Fig. S1). Hyperoxic condition also induced a significant but transient increased in pCO2 blood concentration at P3. No difference was reported in body temperature between experimental groups at P3 and P10 (data not shown). Hyperoxia was associated with significant deleterious effects on the developing brain recapitulating WMD pathological features observed in human preterm neonates and summarized in Fig. 1. At early time-point, hyperoxic condition induced microglial activation (ED1+ cells) and increased cell death (cleaved caspase 3 (CC3), TUNEL+ cells), decreased cell proliferation (Ki67+ cells) in P3 and P10 rat pups, and decreased density of pre-oligodendrocytes within the cingular white matter was observed in P3 animals. Astrogliosis (GFAP + cells), and myelination delay (MBP+ fiber density) together with a decrease in mature oligodendrocyte density (APC + cells) were observed in P10 animals. All these hyperoxia-induced anomalies appeared to be transient as they were no longer detectable at P21. To characterize the hyperoxia-induced lung injury in P3 and P10 animals, both median linear intercepts (MLI), radial alveolar count (RAC), which assesses alveolar number across terminal respiratory units, and vascular volume density (Vv) were calculated as described in the Material and methods. Hyperoxia was associated with a significant increase in MLI in hematoxylin–eosin-stained lung tissues in both P3 and P10 (Figs. 2A, B). Similarly, RAC was significantly decreased in hyperoxia-challenged pups both at P3 and P10 (Figs. 2C, D). Vascular volume density was consistently reduced after hyperoxia (Figs. 2E, F). Together, the reduced alveolarization and lung angiogenesis detected in our rat model reproduced the main pathological features of BPD in humans. Thus postnatal hyperoxia induced both lung and brain disease in rat pups, closely mimicking BPD and WMD.

Cognitive assessment Impact of iNO on BPD Associative abilities were assessed at P3 and P10 using an odor preference conditioning test. In this test, the pups learned to associate two artificial odors, with either warm (34 °C) or cold (26 °C) temperature (the CS+ (Conditioned Stimulus) and the CS− odor, respectively). The pups were randomly assigned to the peppermint CS+ or to the banana CS+ group. Immediately after acquisition phase, the pups underwent a two-odor choice test (either peppermint, 0.5 ml 97% menthyl acetate or banana 97% butyl propionate, Aldrich, Steinheim, Germany) in cold conditions using a setup previously described (Bouslama et al., 2005). A preference score was calculated as: 100 × (Time on CS+ − Time on CS−) / (Time on CS+ + Time on CS−). Effective learning was considered to occur in a given group if the mean learning score was significantly above zero.

We explored the impact of iNO on survival, body growth and the main pathological features of BPD. Inhaled NO (5 ppm) prevented body weight reduction in pups subjected to hyperoxia (Supplemental Table S3, p b 0.001). Furthermore, normal pups given iNO had significantly higher body weights than controls (p b 0.001). iNO exposure induced significant changes in MLI and RAC in hyperoxia-challenged pups at P3 but not at P10 (Figs. 2A–D). Similarly, iNO did only prevent the deleterious effects of hyperoxia on vascular volume density in P3 pups but not in P10 pups (Figs. 2E, F). Therefore, at low concentration, iNO only limited the extent of deleterious effect of postnatal hyperoxia on P3 rat pups but failed to prevent lung injury after a 8-day exposure.

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

117

Fig. 1. Impact of postnatal hyperoxia (FiO2 = 80%) on the developing brain. Graph depicted at three separate stages (P3, P10 and P21) the effects of postnatal hyperoxia on microglial activation (ED1), hemispheric cell death (CC3, TUNEL), proliferating cells (Ki67), developing oligodendrocytes (O4, NG2), mature astrocytes (GFAP), mature oligodendrocytes (APC) cell density and myelin density (MBP) within cingulate white matter compared to controls. For all, *p b 0.05, **p b 0.01 compared to normalized controls using Mann–Whitney test.

Fig. 2. Impact of iNO on BPD pathological features in the developing lung during and after postnatal hyperoxia. A–D: Quantitative analysis of the effect of iNO exposure on alveolarization assessed by MLI (A, B) and radial alveolar count (RAC; C, D) in P3 and P10 animals. Low dose of iNO (5 ppm) had only significant effects on hyperoxia-induced lung damage in P3 animals compared to airexposed controls. E, F: Quantitative analysis of the effect of iNO exposure on pulmonary vascular volume density (Vv) in P3 and P10 rat pups. For all, *p b 0.05, **p b 0.01, ***p b 0.001 compared to controls using one-way ANOVA with the Dunnett's correction. #p b 0.05 using Mann–Whitney test for comparisons between air- and iNO-exposed groups. ns: not significant.

118

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

iNO attenuated microglial activation, astrogliosis and cell death following postnatal hyperoxia In addition to a mild and transient impact on lung histology in hyperoxia-challenged rat pups, iNO attenuated microglial activation induced by the hyperoxia in the developing brain (Fig. 3A). Similarly, GFAP + cell density within the white matter was reduced in iNOexposed P10 pups when compared to air-exposed animals (Fig. 3B). To test the hypothesis that iNO prevents cell death in brain subjected to postnatal hyperoxia, we performed TUNEL assays in P3 pups. The density of TUNEL-positive nuclei significantly increased in the hemispheric white matter of hyperoxia-exposed pups when compared to controls. Conversely, pups exposed to iNO had similar TUNEL-positive nuclear densities as controls (Fig. 3C). Similarly, the density of CC3positive cells was significantly increased by hyperoxia in the cingular white matter (Fig. 3D). In contrast, pups exposed to iNO displayed CC3-positive cell densities similar to those of controls. These results

were confirmed by western blotting (Fig. 3E). Cytosolic levels of cytochrome c, which is released from the mitochondria during apoptosis, reflected the changes observed with CC3 (Fig. 3E). Caspaseindependent cell death is associated with the translocation of Apoptosis-Inducing Factor (AIF) from the mitochondria to the nucleus. Hyperoxia exposure dramatically reduced mitochondrial AIF and increased nuclear AIF, while iNO exposure effectively reversed this nuclear translocation in P3 rat pups (Fig. 3F). iNO is thus capable of reducing hyperoxia-induced DNA fragmentation through both caspasedependent and -independent pathways. Impact of iNO on myelination of hyperoxia-injured white matter We next asked whether iNO could modify oligodendroglial developmental maturation in hyperoxia-challenged pups. At P10, the density of myelinated (MBP-positive) fibers in hyperoxia-challenged pups exposed to iNO was significantly increased compared to air-exposed

Fig. 3. iNO reduced inflammation and cell death occurring in the developing brain following postnatal hyperoxia. A, B: Quantitative analysis of the microglial activation (A: ED1+ cells on P3) and astrogliosis (B: GFAP+ cells on P10) in the cingulate WM in rat pups subjected to hyperoxia compared to controls (Ctl), with or without iNO exposure. C, D: Quantitative analysis of TUNEL-positive cells within hemispheric WM (C) and the cleaved caspase 3 (CC3) positive cells within the cingulated WM (D) in P3 rat pups subjected postnatal hyperoxia compared to controls (Ctl), with or without iNO exposure. E: Western blot analysis showing an increase in both cleaved caspase 3 (CC3) and cytochrome c (cyt c) cytosolic content in P3 pups subjected to postnatal hyperoxic insult compared to controls (Ctl) (n = 6). In contrast, iNO exposure fully prevented these alterations compared to controls. F: Western blot analysis of Apototic Inducing Factor (AIF) translocation from mitochondria to nucleus (histon 3 was used as nuclear internal control) (n = 6). Again, iNO exposure prevented AIF translocation induced by postnatal hyperoxia in P3 pups. For all, *p b 0.05, **p b 0.01, ***p b 0.001 compared to controls using one-way ANOVA with the Dunnett's correction. #p b 0.05, ##p b 0.01 using Mann–Whitney test for comparisons between air- and iNO-exposed groups.

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

119

Fig. 4. Impact of iNO on oligodendroglial cells maturation and myelination in the developing brain following postnatal hyperoxia. A–C: Quantitative analysis of the MBP-positive fibers optical density in the cingulated (A, B) and lateral (C) white matter in P10 rat pups subjected to hyperoxia, with or without iNO exposure. These data demonstrated that myelin content was defective in pups subjected to hyperoxia compared to controls (Ctl). In contrast, low dose (5 ppm) iNO exposure restored a MBP + fibers optic density similar to controls. Bar = 200 μm. D: Quantitative analysis of the APC positive cells in the cingulated WM in P10 rat pups subjected to hyperoxia compared to controls (Ctl), with or without exposure iNO. E: Quantitative analysis p27kip1 gene expression in P3 rat pups subjected to hyperoxia compared to controls (Ctl), with or without exposure to iNO. For all, *p b 0.05, **p b 0.01, ***p b 0.001 compared to controls using one-way ANOVA with the Dunnett's correction. #p b 0.05, ##p b 0.01 using Mann–Whitney test for comparisons between air- and iNO-exposed groups.

injured pups in both the cingulate and lateral white matter (Figs. 4A–C). In addition, the density of mature myelinating oligodendrocytes, as indicated by APC immunoreactivity in the lateral corpus callosum, was significantly decreased in pups subjected to a hyperoxic insult at P10 but restored to normal levels by exposure to iNO (Fig. 4D). Furthermore, iNO exposure enhanced the density of APC positive cells in control P10 pups, as previously reported (Olivier et al., 2010). Consistently, using quantitative PCR analysis, we found that iNO significantly increased the gene expression level of P27kip1, a pro-maturation factor in hyperoxia-challenged rat pups (Fig. 4E). In contrast to myelin content and mature oligodendrocyte density, axons did not seem to be injured by postnatal hyperoxia in P10 rat pups; iNO had no effect on axonal optic density both in cingular and lateral white matter (Supplemental Fig. S2). All these findings were gender-independent. Impact of iNO on developing oligodendrocytes following postnatal hyperoxia We next asked whether iNO could modify pre-oligodendrocyte cell density, proliferation and survival in challenged animals. Postnatal

hyperoxia induced a significant decrease in the density of O4- and NG2-immunolabeled immature oligodendrocytes in the white matter at P3 (Fig. 5A). Postnatal exposure to iNO remarkably attenuated this deleterious effect. NG2 positive cells appeared healthier and slightly more mature in iNO-exposed animals compared to the controls (Figs. 5B, C). We found that iNO prevented the hyperoxia-induced decrease in pre-oligodendroglial cells through a significant improvement of proliferating oligodendrocytes (NG2/Ki67 double positive cells) and a significant reduction in developing oligodendrocyte cells death (O4/CC3 double positive cells) in P3 rat pups (Figs. 5D–F). Impact of iNO on neurotrophic factor expression Since white matter injury appears to be transient in our model, we explored next an intrinsic repair mechanism that could be potentiated by iNO exposure. Specifically, we studied the expression of several neurotrophic factors potentially involved in these mechanisms, including BDNF, VEGF-A and its receptor, VEGFR-2. In P3 pups, the expression level of BDNF was significantly decreased in P3 rats subjected to

120

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

Fig. 5. Impact of iNO on developing oligodendrocytes in rat pups subjected to postnatal hyperoxia. A: Quantitative analysis of the NG2 and O4 positive cells in the cingulated WM in P3 rat pups subjected to hyperoxia compared to controls (Ctl), with or without exposure iNO. B, C: Photomicrographs showing NG2-positive oligodendrocytes (arrows, higher magnification in insets) in P3 rat pups subjected to postnatal hyperoxia, without (air: B) and with (iNO: C) exposure to iNO. Bar = 50 μm. D: Quantitative analysis of the NG2-Ki67 double positive dividing pre-oligodendrocytes in the cingulated WM in P3 rat pups subjected to hyperoxia compared to controls (Ctl), with or without exposure iNO. E: Photomicrograph showing NG2/Ki67 double immunostaining (arrows: double positive proliferating oligodendrocytes; arrowheads: non proliferating NG2+ oligodendrocytes) within the lateral corpus callosum in coronal sections from P3 rat pups. F: Quantitative analysis of the O4-cleaved caspase 3 double positive dying pre-oligodendrocytes in the cingulated WM in P3 rat pups subjected to hyperoxia compared to controls (Ctl), with or without exposure iNO. For all, *p b 0.05, **p b 0.01 and ***p b 0.001 compared to controls using one-way ANOVA with the Dunnett's correction. #p b 0.05, ##p b 0.01 using Mann–Whitney test for comparisons between air- and iNO-exposed groups.

hyperoxia but those of VEGF-A and VEGFR-2 were found unchanged (Supplemental Fig. S3). iNO exposure completely prevented this decrease. In P10 pups, we failed to demonstrate any changes in BDNF, VEGF-A, and VEGFR-2 gene expression in response to hyperoxia. Impact of iNO on early behavior of rat pups Finally, we asked whether postnatal hyperoxia could have a functional impact on learning abilities. Olfactory conditioning experiments,

which allowed detection of early dysfunction in associative memory were performed on P3 and P10 (Figs. 6A, B). These experiments showed that postnatal hyperoxia resulted in an important learning deficit, irrespective of age. iNO exposure did not modify behavioral performances in control pups. In contrast, iNO exposure did significantly attenuate the loss in learning abilities and did improve the score of pups exposed to postnatal hyperoxia to a level similar to that of non-exposed pups (Fig. 6). Thus, iNO proved efficient to protect associative abilities in pups exposed to postnatal hyperoxia.

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

121

Discussion

Fig. 6. Odor preference conditioning test performed at P3 (panel A) and P10 (panel B). *p b 0.05 and **p b 0.01 mean that learning scores were significantly above zero after analysis using unilateral t-test. Postnatal exposure to oxygen abolished learning abilities (hyperoxia by treatment interaction, p = 0.0496, and effect of hyperoxia: p = 0.0104). However, iNO treatment improved learning scores in hyperoxia-exposed pups to a level similar to that of non-exposed pups (effect of iNO: p = 0.0133). Age had no significant effect, neither as a main between-factor nor in interaction with other factors analyzed. N = 24 for each group. Values are means ± SEM.

Impact of iNO on hemodynamic responses A significant increase in mean blood-flow velocities (mBFV) was evidenced in the internal carotid arteries in animals subjected to iNO (5 ppm) after 6 days of exposure (Fig. 7B); this effect was not detected 1 day after exposure (Fig. 7A). Hyperoxia very significantly reduced mBFV at both 1 and 6 days of exposure; in contrast, iNO was able to reverse this effect and mBFV returned to values measured under air at the 2 developmental stages investigated (Figs. 7A, B). Heart rates (HR) were not significantly different between animals exposed to air (360 ± 23 bpm), iNO (376 ± 20 bpm) and hyperoxia with iNO (362 ± 36 bpm). In contrast, animals under hyperoxia alone displayed reduced HR (194 ± 18 bpm, p = 0.0012).

In this study, we developed an animal model that combines the two main complications associated with prematurity, BPD and WMD, which share several risk factors including low gestational age, inflammation and oxidative stress. Using this model, we demonstrated that iNO exposure during the first postnatal week significantly attenuated the main features of hyperoxia-induced WMD in neonatal rats and the devastating effects of hyperoxia on learning abilities. Our animal model simultaneously displayed both WMD and BPD. Previous reports have only explored one of these entities at a time (Baud et al., 2004; Shaffer et al., 1987), and do not reflect the integrated clinical course of premature infants with both brain and lung injuries. We previously published a model that simultaneously mimicked WMD and BPD by subjecting rat to postnatal hyperoxia (80% +/− 0.5 O2) during the first seven days of life following gestational hypoxia (10% +/− 0.5 O2). In this model, we demonstrated that iNO reduced brain injury, but mortality rate in this model was high and mechanisms triggering lesions quite complicated (Pham et al., 2012). In the rat model reported here, mortality rate was strongly reduced (10–15%) and developmental programs not altered by antenatal hypoxia. Thus, it could be a better experimental tool to clarify the complex relationship between the development of BPD and the occurrence of WMD (Dammann et al., 2004). Elevated pO2 induced by hyperoxia exposure only lasted few hours probably due to rapid lung disease. Brain injury and disruption of developmental programs in the white matter were therefore likely related to either early neonatal hyperoxia or BPD itself. Since respiratory disease is associated with WMD independently of the degree of prematurity (Anjari et al., 2009), we could speculate that the effect of iNO on WMD is a consequence of its pulmonary benefits. iNO is known to markedly improve cell survival and lung histology, growth and alveolarization by inhibiting inflammation and reducing capillary-alveolar alterations in rat pups with BPD (Lin et al., 2005; Rieger-Fackeldey and Hentschel, 2008; Tang et al., 2004; Tourneux et al., 2009). In our model, the low doses of iNO used (5 ppm) prevented the reduction in body weight and the increase in mortality of pups subjected to postnatal hyperoxia, but failed to induce protracted improvements in lung histology/angiogenesis in challenged animals. Similarly, low dose iNO did not alter the biological, histological and physiological consequences of BPD induced by the combination of antenatal protracted hypoxia and postnatal hyperoxia (Pham et al., 2012). Therefore, the neuroprotective effect of iNO in our model is probably due less to its modest pulmonary impact than to its specific impact on the developing brain. The remarkable effect of iNO reported here is consistent with our previous work demonstrating that both endogenous and exogenous (inhaled) NO affect myelination and the developmental program of

Fig. 7. Impact of iNO on internal carotid blood flow in normoxic and hyperoxic conditions. A–B: Mean blood-flow velocities were measured in the internal carotid arteries (ICA) in animals subjected to hyperoxia exposure at 1 (A) and 6 (B) day(s) compared to controls (Ctl), with or without exposure iNO (n = 5 in each group). Note that iNO was able to significantly reverse the detrimental effect of hyperoxia. ***p b 0.001 versus Ctl; #p b 0.05; ##p b 0.01; ###p b 0.001 air versus iNO.

122

H. Pham et al. / Experimental Neurology 252 (2014) 114–123

white matter maturation in normal rodent pups (Olivier et al., 2010). We have also demonstrated that iNO exposure: - increases cortical NO concentration, as demonstrated by voltammetric detection in rodents (Charriaut-Marlangue et al., 2012), - is neuroprotective following excitotoxic brain damage in P5 rat pups (Pansiot et al., 2010), - induces collateral recruitment following brain ischemia (Bonnin et al., 2012), - induces significant neuroprotection when given at low doses during ischemia (Charriaut-Marlangue et al., 2012, 2013). Interestingly, Terpolilli et al. (2012, 2013) recently reported similar results in adult ischemic stroke and traumatic brain injury. Taken together, these findings indicate that low doses of iNO not only act locally on the pulmonary vasculature but also have remote effects on the developing brain, both under basal and challenged conditions. The idea that exogenous NO could be neuroprotective is fairly controversial. Indeed, numerous studies have demonstrated the deleterious effects of reactive nitrogen species accumulation following ischemic– reperfusion brain injury through energy failure, lipid peroxidation, protein nitrosylation, DNA alterations and increased permeability of the blood–brain barrier (Iadecola, 1997). Excitotoxicity and hypoxic– ischemic insults result in inflammation, and high concentrations of NO and peroxynitrite produced locally by activated microglia could become toxic to neurons and immature oligodendrocytes both in vitro and in vivo (Haynes et al., 2005). The immature brain is especially vulnerable to damage by free radicals due to its high concentrations of unsaturated fatty acids, low antioxidant levels and increased availability of free redox-active iron (Ikonomidou and Kaindl, 2011). Rosenberg et al. have, however, reported that NO could have positive or negative effects depending on the intracellular redox state (Rosenberg et al., 1999). Here, we used very low doses of iNO, ie. 5 ppm, which could attenuate the deleterious consequences of postnatal elevated PaO2 and related inflammation by scavenging oxygenderived free radicals. In addition, several reports suggest that NO could have neuroprotective effects under specific experimental conditions. Thus, NO release by endothelial cells regulating local blood flow appears to protect the brain after middle cerebral artery occlusion, possibly by improving cerebral circulation (Chiueh, 1999; Huang et al., 1994). Increased brain infarct volumes have also been reported following decreased NO production due to NOS inhibitors (Yamamoto et al., 1992). In a murine stroke model Li et al. showed that 60 ppm iNO could improved recovery from subarachnoid hemorrhage and reduced the inflammatory response accompanying ischemic stroke (Li et al., 2013). In another model of P9 mice subjected to unilateral hypoxia–ischemia (HI), Zhu et al. reported that intraischemic 50 ppm iNO was able to reduce brain injury in males but not in females (Zhu et al., 2013). Here we did not confirm that iNO provides neuroprotection in a gender-related manner. Other possible neuroprotective mechanisms could involve bloodborne delivery of NO or NO-related metabolites such as hemoglobinderived S-nitrosothiol or nitrites/nitrates, with distant vasodilatory effects (Pawloski et al., 2001), but the role of these metabolites/ NO-carriers remains to be elucidated in the setting of iNO administration. We and others recently reported that iNO was delivered to the brain and this effect was associated with increased arterial blood-flow velocities (Charriaut-Marlangue et al., 2012; Li et al., 2013; Terpolilli et al., 2012). In the present study, iNO was also able to reverse the adverse effect of hyperoxia by increasing mBFV, highly suggesting recruitment and vasodilation of small vessels in the microcirculation as previously reported (Bonnin et al., 2011; Terpolilli et al., 2012). In addition to its potent pulmonary arteriolar vasodilator properties, iNO has recently been proposed for the treatment of BPD in premature

infants because it promotes pulmonary vascular growth (Jobe, 2011). However, several therapeutic trials using iNO have yielded disappointing results, especially at low doses, and its subsequent neurodevelopmental effects are poorly understood (Mercier et al., 2010). iNO must therefore be used with caution while the results of longterm follow-up studies are still awaited, as stated by a recent NIH Consensus Development Conference (Cole et al., 2011). The current challenge is to identify premature babies capable of benefiting from iNO exposure, and the optimal dosage of iNO specific to each endpoint (BPD and/or WMD). The need for further preclinical evaluation appears crucial and urgent, and our experimental study contributes to this quest by demonstrating the neuroprotective potential of iNO at low concentrations. Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.expneurol.2013.11.025. Role of funding source This study was financially supported by the Institut National Sur la Recherche Médicale (INSERM), the INSERM AVENIR Program and the Mairie de Paris. Ikaria Inc. partially supports the experimental research program of Prof. Olivier Baud and co-workers with respect to inhaled nitric oxide. However, Ikaria Inc. has taken part neither in the study design nor in the data analysis. Acknowledgments We are grateful to Mathias Schroijen for his help in cognitive assessment. This study has been supported by the Prime Brain project and de Spoelberch Foundation. References Anjari, M., Counsell, S.J., Srinivasan, L., Allsop, J.M., Hajnal, J.V., Rutherford, M.A., Edwards, A.D., 2009. The association of lung disease with cerebral white matter abnormalities in preterm infants. Pediatrics 124, 268–276. Back, S.A., Riddle, A., McClure, M.M., 2007. Maturation-dependent vulnerability of perinatal white matter in premature birth. Stroke 38, 724–730. Baud, O., Daire, J.L., Dalmaz, Y., Fontaine, R.H., Krueger, R.C., Sebag, G., Evrard, P., Gressens, P., Verney, C., 2004. Gestational hypoxia induces white matter damage in neonatal rats: a new model of periventricular leukomalacia. Brain Pathol. 14, 1–10. Bonnin, P., Leger, P.L., Deroide, N., Fau, S., Baud, O., Pocard, M., Charriaut-Marlangue, C., Renolleau, S., 2011 May 15. Impact of intracranial blood-flow redistribution on stroke size during ischemia-reperfusion in 7-day-old rats. J Neurosci Methods 198 (1), 1039. Bonnin, P., Leger, P.L., Villapol, S., Deroide, N., Gressens, P., Pocard, M., Renolleau, S., Baud, O., Charriaut-Marlangue, C., 2012. Dual action of NO synthases on blood flow and infarct volume consecutive to neonatal focal cerebral ischemia. Exp. Neurol. 236, 50–57. Bouslama, M., Durand, E., Chauvière, L., Van den Bergh, O., Gallego, J., 2005. Olfactory classical conditioning in newborn mice. Behav. Brain Res. 161, 102–106. Charriaut-Marlangue, C., Bonnin, P., Gharib, A., Leger, P.L., Villapol, S., Pocard, M., Gressens, P., Renolleau, S., Baud, O., 2012. Inhaled nitric oxide reduces brain damage by collateral recruitment in a neonatal stroke model. Stroke 43, 3078–3084. Charriaut-Marlangue, C., Bonnin, P., Pham, H., Loron, G., Leger, P.L., Gressens, P., Renolleau, S., Baud, O., 2013. Nitric oxide signaling in the brain: A new target for inhaled nitric oxide? Ann. Neurol. 73, 442–448. Chiueh, C.C., 1999. Neuroprotective properties of nitric oxide. Ann. N. Y. Acad. Sci. 890, 301–311. Cole, F.S., Alleyne, C., Barks, J.D., Boyle, R.J., Carroll, J.L., Dokken, D., Edwards, W.H., Georgieff, M., Gregory, K., Johnston, M.V., Kramer, M., Mitchell, C., Neu, J., Pursley, D.M., Robinson, W.M., Rowitch, D.H., 2011. NIH Consensus Development Conference statement: inhaled nitric-oxide therapy for premature infants. Pediatrics 127, 363–369. Dammann, O., Leviton, A., Bartels, D.B., Dammann, C.E., 2004. Lung and brain damage in preterm newborns. Are they related? How? Why? Biol. Neonate 85, 305–313. Dunnill, M.S., 1962. Quantitative methods in the study of pulmonary pathology. Thorax 17, 320–328. Emery, J.L., Mithal, A., 1960. The number of alveoli in the terminal respiratory unit of man during late intrauterine life and childhood. Arch. Dis. Child. 35, 544–547. Fontaine, R.H., Olivier, P., Massonneau, V., Leroux, P., Degos, V., Lebon, S., El Ghouzzi, V., Lelièvre, V., Gressens, P., Baud, O., 2008. Vulnerability of white matter towards antenatal hypoxia is linked to a species-dependent regulation of glutamate receptor subunits. Proc. Natl. Acad. Sci. U. S. A. 105, 16779–16784. Gerstner, B., DeSilva, T.M., Genz, K., Armstrong, A., Brehmer, F., Neve, R.L., FelderhoffMueser, U., Volpe, J.J., Rosenberg, P.A., 2008. Hyperoxia causes maturationdependent cell death in the developing white matter. J. Neurosci. 28, 1236–1245.

H. Pham et al. / Experimental Neurology 252 (2014) 114–123 Haynes, R.L., Baud, O., Li, J., Kinney, H.C., Volpe, J.J., Folkerth, D.R., 2005. Oxidative and nitrative injury in periventricular leukomalacia: a review. Brain Pathol. 15, 225–233. Huang, Z., Huang, P.L., Panahian, N., Dalkara, T., Fishman, M.C., Moskowitz, M.A., 1994. Effects of cerebral ischemia in mice deficient in neuronal nitric oxide synthase. Science 265, 1883–1885. Iadecola, C., 1997. Bright and dark sides of nitric oxide in ischemic brain injury. Trends Neurosci. 20, 132–139. Ikonomidou, C., Kaindl, A.M., 2011. Neuronal death and oxidative stress in the developing brain. Antioxid. Redox Signal. 14, 1535–1550. Jobe, A.H., 2011. The new bronchopulmonary dysplasia. Curr. Opin. Pediatr. 23, 167–172. Li, Y.S., Shemmer, B., Stone, E., A. Nardi, M., Jonas, S., Quartermain, D., 2013. Neuroprotection by inhaled nitric oxide in a murine stroke model is concentration and duration dependent. Brain Res. 1507, 134–145. Lin, Y.J., Markham, N.E., Balasubramaniam, V., Tang, A., Maxey, J.R., Kinsella, J.P., Abman, S.H., 2005. Inhaled nitric oxide enhances distal lung growth after exposure to hyperoxia in neonatal rats. Pediatr. Res. 58, 22–29. Mercier, J.C., Hummler, H., Durrmeyer, X., Sanchez-Luna, M., Carnielli, V., Field, D., Greenough, A., Van Overmeire, B., Jonsson, B., Hallman, M., Baldassarre, J., EUNO Study Group, 2010. Inhaled nitric oxide for prevention of bronchopulmonary dysplasia in premature babies (EUNO): a randomised controlled trial. Lancet 376, 346–354. Moncada, S., Palmer, R.M., Higgs, E.A., 1991. Nitric oxide: physiology, pathophysiology, and pharmacology. Pharmacol. Rev. 43, 109–142. Olivier, P., Baud, O., Evrard, P., Gressens, P., Verney, C., 2005. Prenatal ischemia and white matter damage in rats. J. Neuropathol. Exp. Neurol. 64, 998–1006. Olivier, P., Loron, G., Fontaine, R.H., Pansiot, J., Dalous, J., Thi, H.P., Charriaut-Marlangue, C., Thomas, J.L., Mercier, J.C., Gressens, P., Baud, O., 2010. Nitric oxide plays a key role in myelination in the developing brain. J. Neuropathol. Exp. Neurol. 69, 828–837. Pansiot, J., Loron, G., Olivier, P., Fontaine, R., Charriaut-Marlangue, C., Mercier, J.C., Gressens, P., Baud, O., 2010. Neuroprotective effect of inhaled nitric oxide on excitotoxic-induced brain damage in neonatal rat. PLoS One 5, e10916. Pawloski, J.R., Hess, D.T., Stamler, J.S., 2001. Export by red blood cells of nitric oxide bioactivity. Nature 409, 622–626. Pham, H., Vottier, G., Pansiot, J., Dalous, J., Gallego, J., Gressens, P., Duong-Quy, S., DinhXuan, A.T., Mercier, J.C., Biran, V., Charriaut-Marlangue, C., Baud, O., 2012. Inhaled NO protects cerebral white matter in neonatal rats with combined brain and lung injury. Am. J. Respir. Crit. Care Med. 185, 897–899. Rieger-Fackeldey, E., Hentschel, R., 2008. Bronchopulmonary dysplasia and early prophylactic inhaled nitric oxide in preterm infants: current concepts and future research strategies in animal models. J. Perinat. Med. 36, 442–447. Ritter, J., Schmitz, T., Chew, L.J., Bührer, C., Möbius, W., Zonouzi, M., Gallo, V., 2013. Neonatal hyperoxia exposure disrupts axon-oligodendrocyte integrity in the subcortical white matter. J. Neurosci. 33, 8990–9002.

123

Rosenberg, P.A., Li, Y., Ali, S., Altiok, N., Back, S.A., Volpe, J.J., 1999. Intracellular redox state determines whether nitric oxide is toxic or protective to rat oligodendrocytes in culture. J. Neurochem. 73, 476–484. Schmitz, T., Ritter, J., Mueller, S., Felderhoff-Mueser, U., Chew, L.J., Gallo, V., 2011. Cellular changes underlying hyperoxia-induced delay of white matter development. J. Neurosci. 31, 4327–4344. Shaffer, S.G., O'Neill, D., Bradt, S.K., Thibeault, D.W., 1987. Chronic vascular pulmonary dysplasia associated with neonatal hyperoxia exposure in the rat. Pediatr. Res. 21, 14–20. Tang, J.R., Markham, N.E., Lin, Y.J., McMurtry, I.F., Maxey, A., Kinsella, J.P., Abman, S.H., 2004. Inhaled nitric oxide attenuates pulmonary hypertension and improves lung growth in infant rats after neonatal treatment with a VEGF receptor inhibitor. Am. J. Physiol. Lung Cell. Mol. Physiol. 287, L344–L351. Terpolilli, N.A., Kim, S.W., Thal, S.C., Kataoka, H., Zeisig, V., Nitzsche, B., Klaesner, B., Zhu, C., Schwarzmaier, S., Meissner, L., Mamrak, U., Engel, D.C., Drzezga, A., Patel, R.P., Blomgren, K., Barthel, H., Boltze, J., Kuebler, W.M., Plesnila, N., 2012. Inhalation of nitric oxide prevents ischemic brain damage in experimental stroke by selective dilatation of collateral arterioles. Circ. Res. 110, 727–738. Terpolilli, N.A., Kim, S.W., Thal, S.C., Kuebler, W.M., Plesnila, N., 2013. Inhaled nitric oxide reduces secondary brain damage after traumatic brain injury in mice. J. Cereb. Blood Flow Metab. 33, 311–318. Thurlbeck, W.M., Dunnill, M.S., Hartung, W., Heard, B.E., Heppleston, A.G., Ryder, R.C., 1970. A comparison of three methods of measuring emphysema. Hum. Pathol. 1, 215–226. Toda, N., Okamura, T., 2003. The pharmacology of nitric oxide in the peripheral nervous system of blood vessels. Pharmacol. Rev. 55, 271–324. Tourneux, P., Markham, N., Seedorf, G., Balasubramaniam, V., Abman, S.H., 2009. Inhaled nitric oxide improves lung structure and pulmonary hypertension in a model of bleomycin-induced bronchopulmonary dysplasia in neonatal rats. Am. J. Physiol. Lung Cell. Mol. Physiol. 297, L1103–L1111. Vaucher, E., Tong, X.K., Cholet, N., Lantin, S., Hamel, E., 2000. GABA neurons provide a rich input to microvessels but not nitric oxide neurons in the rat cerebral cortex: a means for direct regulation of local cerebral blood flow. J. Comp. Neurol. 421, 161–171. Volpe, J.J., 2009. Brain injury in premature infants: a complex amalgam of destructive and developmental disturbances. Lancet Neurol. 8, 110–124. Vottier, G., Pham, H., Pansiot, J., Biran, V., Gressens, P., Charriaut-Marlangue, C., Baud, O., 2011. Deleterious effect of hyperoxia at birth on white matter damage in the newborn rat. Dev. Neurosci. 33, 261–269. Yamamoto, S., Golanov, E.V., Berger, S.B., Reis, D.J., 1992. Inhibition of nitric oxide synthesis increases focal ischemic infarction in rat. J. Cereb. Blood Flow Metab. 12, 717–726. Yun, H.Y., Dawson, V.L., Dawson, T.M., 1996. Neurobiology of nitric oxide. Crit. Rev. Neurobiol. 10, 291–316. Zhu, C., Sun, Y., Gao, J., Wang, X., Plesnila, N., Blomgren, K., 2013. Inhaled nitric oxide protects males but not females from neonatal mouse hypoxia–ischemia brain injury. Transl. Stroke Res. 4, 201–207.