Abstract
Introduction
The risk of BPD in these cohorts is inversely proportional to the gestational age at birth (25). Major pathologic features of BPD are failure in alveolarization leading to simplified air space and lowered alveolar density, inflammation, and respiratory distress. Lung injury in BPD is thought to result from early developmental arrest probably associated with prenatal exposure or genetic factors and interrupting alveolar growth as observed in extreme prematurity (“new” BPD), or from structural damage of relatively more developed saccular lungs characterized by surfactant deficiency (“old” BPD) that receive respiratory support with mechanical ventilation and prolonged oxygenation (3). BPD survivors often have clinically significant respiratory symptoms and functional abnormalities that persist into adolescence and early adulthood, indicating lifelong consequences of BPD (33).
Innovation
The current study is the first to use lung transcriptomic and pathway analyses to understand the role of Nrf2 in the molecular events during saccular-to-alveolar stage transition. Transcriptome analysis of lungs from Nrf2−/− and Nrf2+/+ mice also supported a functional role for Nrf2 and related downstream effector mechanisms (e.g., Gpx2 and Marco) in the pathogenesis of hyperoxia-induced lung injury in neonates. Results collectively provide insights into Nrf2-driven host defense mechanisms in developing lung, and suggest a therapeutic potential of specific Nrf2 activators in bronchopulmonary dysplasia and other neonatal diseases associated with oxidative stress (e.g., respiratory syncytial virus disease).
Hyperoxia-induced injury in underdeveloped lungs of newborn rodents has been investigated as a model for BPD (49). Angiogenesis proteins including vascular endothelial growth factor (VEGF) (27, 43), keratinocyte growth factor (20), and matrix metalloproteases (MMPs) including MMP-9 (29), are essential in lung development to protect against BPD pathogenesis. In contrast, cathepsin S (23), transforming growth factor beta (TGF-β) (46), or cytokines such as interleukin 1β (7) contribute to lung injury in experimental BPD.
Therapeutically administered hyperoxia to premature infants has been considered to be one contributing cause of BPD, and reactive oxygen species (ROS) are implicated in its pathogenesis (6, 17). NF-E2 related factor 2 (Nfe2l2, Nrf2), a transcription factor for antioxidant response element (ARE)-mediated antioxidant and defense gene expression (24), is essential in tissue protection (26). Using adult mice genetically deficient in Nrf2 (Nrf2−/− ), a protective role of Nrf2 and ARE-responsive effector genes has been shown in oxidant-mediated lung injury (9, 12, 13, 16, 36, 37, 44).
The current study was designed to identify Nrf2-mediated molecular events during the late-phase lung maturation, and determine the role of Nrf2 in the pathogenesis of hyperoxia-induced injury in neonatal mouse lungs. For this purpose, transcriptome analysis revealed lung gene expression profiles from saccular stage (postnatal days P1–P3) and more mature late saccular/early alveolar phase (P4) that differed between newborn Nrf2+/+ and Nrf2−/− mice. We also exposed Nrf2+/+ and Nrf2−/− mice to hyperoxia during early postnatal ages (P1–P4), and differential susceptibility to lung injury and abnormal alveolarization were found. Lung microarray gene profiling and computerized algorithmic screening for ARE characterized potential downstream mechanisms of Nrf2-mediated protection against development of hyperoxia-induced lung injury.
Results
Saccular lung maturation during P1–P4
Developmental lung gene expression profiles
Statistics-based analysis (Fig. 1A) and visual data mining (Fig. 1B) determined the greatest transcriptome differences between P1 and P4 during early postnatal period in Nrf2+/+
neonates. Among significantly varied genes (n=1674, p<0.01) between saccular stage (i.e., P1–P3) and late saccular/alveolar stage (i.e., P4), 324 transcripts were lower (by ≥50%) at P1–P3 (mostly P1 and/or P2) than at P4. They encoded genes for multiple centromere proteins, kinesin family members, cyclins, and DNA polymerases associated with cell cycle (DNA replication, recombination, and repair), cellular assembly, and epigenetic pathways (Fig. 1A–C and Supplementary Table S1A; Supplementary Data are available online at
Saccular-to-alveolar transition of postnatal lungs
Normal lung maturation in Nrf2+/+ newborns (Table 1) was characterized by saccular-to-alveolar transition from simple, poorly septated saccules at P1 (stage 1) to simple septation of saccules at P3 (stage 2) and the appearance of branched septa and multilobular alveoli at P4 (stage 3). Branched septi/alveoli were evident at P4 in the majority of neonates (67%). Lung maturation in Nrf2−/− newborns was comparable to Nrf2+/+ newborns with branched septi/multilobular alveoli present in 75% of mice by age P4 (Table 1).
Saccular-to-alveolar stage transition status of developing lungs.
Numbers indicate fraction of the corresponding animal tissue sections to the total animal tissue sections in the study group used for microscopic evaluation of histopathology (one tissue slide per mouse). Each microscopic slide includes proximal (G5) and distal (G11) lung sections stained with hematoxylin and eosin.
Airway lesions were evaluated by the severity of interstitial inflammation (II), alveolar edema (AE), and alveolar inflammation (AI). Score 1 is for minimal, 2 for mild, 3 for moderate, and 4 for marked injury.
Role of Nrf2 in developmental lung transcriptome
Nrf2 deficiency significantly affected expression of 9737 transcripts during P1–P4; these transcripts have roles mainly in tissue and organ development, cancer, cell death, and infectious disease and mechanism (Fig. 2A). Transcripts constitutively lower (≥50%) in Nrf2−/− than in Nrf2+/+ neonates (Fig. 2B and Supplementary Table S2A) were networked in DNA replication, recombination, and repair; tissue development; lipid metabolism; and redox cycle and stress response pathways. In contrast, several genes associated with immunity, lymphatic system development, and cell–cell interaction networks were expressed higher in Nrf2−/− than in Nrf2+/+ at P1–P4 (Fig. 2B and Supplementary Table S2B). Visual profile analysis also revealed distinct patterns of Nrf2-dependent transcript expression (Fig. 2C and Supplementary Table S3). Profile 1 included transcripts (e.g., Ifi44 and Tcf7) markedly upregulated at P2 in Nrf2−/− mice; profile 2 included transcripts (e.g., Myh and Tnn) overexpressed at P2–P3 in both genotypes, but more highly in Nrf2−/− mice; and transcripts in profile 3 (e.g., Aox1, Clstn2, and Sbf2) were totally attenuated in Nrf2−/− mice.
Role of Nrf2 in hyperoxia-induced lung injury phenotypes
Growth and mortality of neonates
Hyperoxia retarded growth as indicated by lower body weight than air exposure in both genotypes after 100% and 70% O2 (Fig. 3A). However, suppression of weight gain was significantly greater in Nrf2−/− than in Nrf2+/+ mice (Fig. 3A). Hyperoxia (100%) also caused greater mortality in Nrf2−/− mice relative to Nrf2+/+ mice (Fig. 3B). Seventy percent hyperoxia did not cause death of any neonatal mice.
Lung injury and inflammation
Minimal-to-mild inflammation was found in control mice of both genotypes exposed to room air (Table 1). Hyperoxia (100%, 3 days) caused significant pulmonary protein edema as assessed by increased protein concentration and cellular inflammation (neutrophils and monocytes) in bronchoalveolar lavage (BAL) fluids from Nrf2+/+ and Nrf2−/− neonates (Fig. 3C). Compared with Nrf2+/+ mice, more necrotic (lysis) and apoptotic (nuclear fragmentation) cells were found in BAL returns from Nrf2−/− mice (Fig. 3D, E, no deoxynucleotidyl transferase-mediated dUTP nick-end labeling [TUNEL]-positive cells detected in air controls). Protein exudation in air spaces, alveolar inflammation and disruption, and perivascular-peribronchiolar edema were more severe and frequent in Nrf2−/− neonates relative to Nrf2+/+ neonates after 3 days of 100% O2. Hyperoxia caused exudative-phase diffuse alveolar damage in 50% of Nrf2−/− lungs examined, while no lungs from Nrf2+/+ mice had this severe pathology (Fig. 3F and Table 1). Alveolar development in Nrf2+/+ mice after 3 days of 100% O2 was comparable to their air controls with branched septi and alveoli present in 67% of mice (stage 3, Table 1). However, relatively fewer Nrf2−/− neonates had developing multilobular alveoli and branched septi at 3 days after 100% O2 (stage 3: 2/8; stage 2: 6/8), indicating a delay in alveolar maturation (Table 1). Consistent with the histopathologic findings, radial alveolar count (RAC) determined that hyperoxia exposure caused significant alveolar simplification in both genotypes of mice, but RAC was significantly lower in Nrf2−/− neonates relative to Nrf2+/+ neonates after hyperoxia (Supplementary Fig. S1). Mild hyperplasic changes in airway epithelium of both strains and mild-to-moderate perivascular and peribronchiolar edema in Nrf2−/− neonates were found after 70% O2, while BAL parameters were not significantly altered by 70% O2 in either genotype (data not shown). Because 70% O2 did not cause significant lung injury in either of the genotypes, all further studies were performed with 100% O2.
TGF-β, VEGF, and ANGPT2 protein expression
Protein concentrations of TGF-β associated with neonatal hyperoxia-induced lung injury were greater in Nrf2−/− relative to Nrf2+/+ neonates at baseline and after hyperoxia (Fig. 3G). Basal levels of the angiogenesis factors VEGF and angiopoietin-2 (ANGPT2) were slightly lower in Nrf2−/− mice than in Nrf2+/+ mice. Hyperoxia-increased VEGF levels at 3 days were higher in Nrf2+/+ neonates relative to Nrf2−/− neonates (Fig. 3G). ANGPT2 was increased by O2 (1–3 days) in Nrf2+/+ but not in Nrf2−/− mice (Fig. 3G).
Pulmonary Nrf2 activation and oxidative stress after hyperoxia
Hyperoxia increased mRNA expression, nuclear translocation, and total ARE binding activity of pulmonary Nrf2 over the age-matched constitutive levels in Nrf2+/+ neonates after 2 and 3 days (Fig. 4A). Total glutathione (GSH) level was significantly lower in Nrf2−/− compared with Nrf2+/+ neonates at baseline and after 2–3 days of O2 exposure (Fig. 4B). Hyperoxia significantly increased GSH in Nrf2+/+ (1–3 days) and Nrf2−/− mice (1–2 days), but the induced GSH level was significantly lower in Nrf2−/− than in Nrf2+/+ mice throughout exposure (Fig. 4B). At P4, significantly higher levels of baseline oxidized lipid (malondialdehyde [MDA]) were found in the lungs from Nrf2−/− relative to Nrf2+/+ neonates (Fig. 4C). Hyperoxia (3 days) caused significantly higher increases of MDA in Nrf2−/− than in Nrf2+/+ mice (Fig. 4C). Basal level of oxidatively modified proteins was higher in Nrf2−/− than in Nrf2+/+ neonates at all times (Fig. 4D). Protein oxidation band intensity was elevated over baseline after 2 and 3 days of O2 in Nrf2−/− lungs, and enhanced protein oxidation levels were markedly higher than those in Nrf2+/+ lungs (Fig. 4D).
Transcripomics during the development of hyperoxia-induced lung injury and the role of Nrf2
Lung genes modulated by hyperoxia in Nrf2+/+ neonates
Transcripts upregulated by hyperoxia were enhanced mostly from 2 days and remained elevated, and they encoded antioxidant defense proteins, DNA damage/repair and apoptosis proteins, AP-1 family and other transcription factors, multiple chemokines and cytokines, cell adhesion/migration molecules, and organ development and angiogenesis factors (Supplementary Table S4A). Conversely, hyperoxia downregulated transcripts most predominantly at day 3. Affected genes encoded signal transducers of canonical pathways including cytochrome P450-mediated xenobiotic metabolism, lipid and hormone metabolism, molecule transport, and humoral immunity (Supplementary Table S4B). Overall, genes significantly modulated during the development of hyperoxia-induced injury (n=8529, p<0.01) have roles in cell cycle, tissue development, gene expression, lipid metabolism, cell–cell interaction, and immune response networks (Supplementary Table S4c).
Genes differentially modulated by hyperoxia in Nrf2+/+ and Nrf2−/− neonates
During hyperoxia (1–3 days), 437 transcripts were significantly (p<0.01) and differentially (≥2-fold) modulated between Nrf2+/+ and Nrf2−/− neonates. k-Means clustering analysis determined genes that are relatively suppressed in Nrf2−/− compared with Nrf2+/+ neonates after hyperoxia (Fig. 5A sets 2, 4, and 5; Supplementary Table S5). They encoded proteins for gene expression machinery and cell growth (DNA replication/repair, apoptosis, transcription, translation, and cell cycle) and redox homeostasis including multiple ARE-responsive antioxidant/defense genes. Conversely, gene transcripts with higher expression in Nrf2−/− lungs were involved in endocytosis, transport, and development (Fig. 5A sets 1 and 3; Supplementary Table S5). These Nrf2-dependent genes have functions in cancer, cell growth and development, lipid metabolism, and small molecule biochemistry during the development of hyperoxic lung injury (Fig. 5B). Canonical pathway (Supplementary Table S5) and functional network (Supplementary Fig. S2) analyses depicted time-dependent events modulated by Nrf2 during the pathogenesis, which were acute-phase organ injury (1 day), organ morphology (1–2 days), cell growth and proliferation (2 days), vasculature development (2 days), immune response (2–3 days), TGF-β signaling (3 days), hematological system development and function (3 days), and cell–cell interaction and signaling (3 days). Visual data mining (Fig. 5C and Supplementary Table S3) determined a unique profile of Nrf2-dependent genes suppressed basally and did not respond to hyperoxia in Nrf2−/− mice (profile 4; e.g., Gpx2, Txnrd1, and Creg1) while several genes (e.g., Mtr, Sbf2, and MHCII) were markedly overexpressed at P2–P3 and were relatively decreased by O2 only in Nrf2−/− mice (profile 5).
Because many genes associated with DNA replication/repair and redox homeostasis were relatively suppressed in Nrf2−/− neonates during hyperoxia, DNA damage and protein levels of a DNA replication and cell proliferation marker PCNA were compared in Nrf2−/− and Nrf2+/+ lungs. Significant increases in genomic (1 and 3 days) and mitochondrial (1 and 2 days) DNA base lesions were found only in Nrf2−/− neonates exposed to 100% O2 (Fig. 5D). In addition to postnatal age-dependent increase (Fig. 1D), hyperoxia-induced increase of nuclear PCNA was marked at 2 days in both genotypes, relative to corresponding air controls (Fig. 5D). However, the hyperoxia-induced level as well as basal abundance at 1–2 days (P2–P3) were relatively lower in Nrf2−/− neonates compared with Nrf2+/+ neonates.
Validation of microarray gene expression profiles and functional relevance
Nrf2-dependent lung expression profiles
Quantitative reverse transcription–polymerase chain reaction (qRT-PCR) analysis (Fig. 6A, left) and western blot analysis (Fig. 6A, right) confirmed microarray expression profiles of selected transcripts or their protein products, including histocompatibility 2 (H2-Q1, H2-D1, and H2-Ea), Inta4, Nqo1, Akr1b8, Hc, Aox1, Jag1, Rad51, and Egr2, which varied between Nrf2+/+ and Nrf2−/− neonates during P1–P4. Message profiles of selected hyperoxia-responsive, Nrf2-dependent genes Akr1b8, Clstn2, Nqo1, Hc, Ang3, Slc7a11, Gpx2, GCS (for Gclc), GST-μ (for Gstm), HO-1 (for Hmox1), and MARCO determined by qRT-PCR (Fig. 6B, left) or western blotting (Fig. 6B, right) were also consistent with those observed by microarray analysis.
Putative AREs in the promoter of potential Nrf2 effector genes
Putative ARE or ARE-like sequences were analyzed in 5-kb upstream sequences of selected genes regulated in an Nrf2-dependent manner to evaluate their potential as direct Nrf2 effectors using a position weight matrix (PWM) statistical model (48). A majority of the listed genes in Table 2 (a selected binding motif from each representative gene presented) had multiple ARE-like binding sequences, and their PWM scores and matrix similarity score were as high as those of the validated AREs in Nrf2 target genes (e.g., Txnrd1 and Ftl). Importantly, many of these target genes containing potential AREs were also identified as Nrf2 effectors by chromatin immunoprecipitation sequencing performed in human (8) and/or mouse cells (30), thus supporting their functional roles. Taken together, results strongly suggest a role for Nrf2 through binding to these genes in modulation of angiogenesis, cell cycle, tissue development, and cell-to-cell interaction in newborn lungs and during injury pathogenesis.
Uppercase indicates ARE core-like sequences (5′-RTKAYnnnGCR-3′; R=A or G, K=G or T, Y=C or T, n=A, C, G, or T), forward (F) or reverse complementary (R).
ND, not detected; ARE, antioxidant response element; TSS, transcription start site; PWM, position weight matrix; MS score, matrix similarity score; ChIP-Seq, chromatin immunoprecipitation-sequencing; M, Malhotra et al. (2010) (30); C, Campbell et al. (2010) (8).
Functional role of Nrf2 downstream effectors: glutathione peroxidase 2 and macrophage receptor with collagenous structure
ARE-bearing macrophage receptor with collagenous structure (Marco) and glutathione peroxidase 2 (Gpx2) were identified as key Nrf2 downstream effectors in the current model as they were distinctly suppressed in Nrf2−/− neonates at P1–P4, and were highly induced during exposure to O2. Lung injury responses to hyperoxia in Marco−/− and Gpx2−/− neonates were compared with their wild-type controls. Significantly greater numbers of BAL neutrophils were found in Marco−/− and Gpx2−/− mice compared with respective wild-type mice after 3 days of O2, and the number of BAL macrophages after O2 was also significantly greater in Marco−/− than in Marco+/+ neonates (Fig. 6C, D). Overall, data indicate functional roles for Marco and Gpx2 in hyperoxia-induced lung inflammation.
Discussion
While Nrf2 is a critical modulator for protection against a broad range of oxidative disorders in adults, its role in tissue development or the pathogenesis of neonatal or childhood disease has received little attention. In the current study, transcriptomic analysis indicated that Nrf2 is critical to processes/networks for cell cycle and DNA repair, immune function, morphogenesis and lung development, and antioxidant defense during postnatal normal lung maturation in mice. Importantly, we found a beneficial role for Nrf2 in hyperoxia-induced injury of undeveloped lung. Nrf2 is a susceptibility gene for protection against acute lung injury caused by 100% O2 in adult mice (13, 14, 38). Our results demonstrate Nrf2-dependent alleviation of hyperoxia-induced injury phenotypes in the saccular phase of lung, including arrest in alveolar development evidenced by lower RAC and reduced appearance of multilobular alveoli/branched septi as well as severe exudative-phase diffuse alveolar damage characterized by edema, leukocyte inflammation, and cell death in Nrf2−/− mice. Moreover, highly suppressed GSH pools as well as heightened pulmonary oxidation and DNA lesions in Nrf2−/− neonates indicated the critical roles for ROS and Nrf2-directed defense in the pathogenesis of hyperoxia-induced lung injury. The current study warrants further investigation of Nrf2 in other oxidant-associated lung disease models at early ages. Juvenile Nrf2−/− mice that were exposed to hyperoxia as neonates had more severely hindered resolution of lung damage relative to juvenile Nrf2+/+ mice that were similarly exposed as neonates (31). This observation supported an association of the severity of hyperoxia-induced injury in infancy with persisting or long-term pulmonary outcome. It also suggests the potential for exacerbation of oxidative pulmonary disease in adults or adolescents who had BPD in infancy.
The current study initially characterized complex gene expression networks in the saccular stage during postnatal lung maturation. Variation in lung gene expression was greatest at P1 relative to age P4, likely reflecting the influence of direct contact of airway cells to the extra-uterine environment. Importantly, Nrf2 significantly modulated genes involved not only in redox balance but also in tissue and organ development, cancer, cell death, and infectious disease during saccular-to-alveolar transition. The marked overexpression of multiple major histocompatibility complex, class II (MHCII), lymphatic system, and cell–cell interaction genes (e.g., Itga4 and Cxcl14) in naive lungs from Nrf2−/− neonates suggested their aberrant basal immunity as evidence by enhanced susceptibility of adult Nrf2−/− mice to asthma and allergy (28, 37).
Our microarray analysis also evaluated Nrf2-dependent antioxidant capacity under normoxic and hyperoxic conditions in the immature lungs. Direct antioxidants, including superoxide dismutases (SODs), are known to be highly activated in the lung shortly after birth (34), and we found that all the redox genes that varied during P1–P4 were higher at P1 relative to P2–P4. In utero expression of airway antioxidant enzymes is known to increase toward term gestation to prepare for birth into an O2-rich (from 3% to 21%) environment (39). Therefore, preterm infants with low birth weight are not only more sensitive to increased O2 concentrations compared with adults (10), but they also have diminished/compromised endogenous antioxidant activity relative to full-term infants (39), which contributes to the critical consequence of hyperoxic insult in BPD pathogenesis. However, overall clinical results from therapies with antioxidants (e.g., SODs, vitamins A and E, N-acetylcysteine, and metalloporphyrin) have remained inconclusive in preterm infants (1, 45). In the current study, we identified novel antioxidants (e.g., Akr1b8, Cbr2, Pgd, and Slc7a11) that were induced during the development of neonatal hyperoxic injury, but not in lungs of adult mice exposed to hyperoxia (15). These gene products have roles in redox balance through a broad spectrum of pathways, including metabolic process, small molecular biochemistry, and membrane transport. Importantly, Slc7a11 [solute carrier family 7 (cationic amino acid transporter, y+ system), member 11] encodes xCT that is a key component of high-affinity cysteine/glutamate exchange transporter system χc −, which mediates cellular cystine uptake for GSH synthesis (41). Identification of putative AREs in these Nrf2 effectors suggests their therapeutic potentials in preventing oxidant-induced injury in the neonate lung.
Neonatal pulmonary oxidative stress was obvious after hyperoxia exposure regardless of genotypes, while Nrf2 deficiency elevated oxidative proteins and lipid peroxidation at baseline as well as after hyperoxia. Although widely used, the amount of MDA as a lipid peroxidation marker is known to be affected by several variables (22). It would be worth validating the effect of Nrf2 deletion on oxidant tissue injury by measurement of a more reliable lipid peroxidation marker 8-iso-Prostaglandin F2α (32). Importantly, oxidative DNA damage is considered a causative factor in diverse pulmonary disorders, including neoplasia and acute lung injury. Previous studies have shown that hyperoxia caused base adduct formation (e.g., 7,8-dihydro-8-oxo-guanine) and DNA strand breakage in lungs of adult mice (4); DNA adduct formation was found in most lung cells after the exposure, while the most severe level of damage, phosphodiester backbone breakage, was found only in type 2 cells, which resulted in the cell death and was associated with lung injury. Consistent with Nrf2-dependent variation in pulmonary apoptotic cell death and DNA lesions, microarray analysis also identified Nrf2-dependent dysregulation of genes involved in the DNA damage/repair and methylation pathways under conditions of normoxia (e.g., Rad51 and Chek1) and hyperoxia (e.g., Rbm14 and Mtrr). In particular, significant suppression of mitochondrial superoxide dismutase (Sod2; Supplementary Table S4a) is likely to be a factor leading to increased mitochondrial damage in Nrf2−/− neonates. Although epigenetic effects of hyperoxia and Nrf2 dependency were beyond the scope of the current analysis, evidence indicates that hyperoxia causes hypermethylation in CpG islands of a lung gene in rats (50). Further, hyperoxia-induced DNA damage influenced global DNA methylation status in lung epithelial cells (35).
Impaired pulmonary vasculature development in ventilated preterm infants is thought to be caused by complex interactions of lung immaturity and postnatal factors including O2, which results in arrest of alveolar growth (42). Maturation of pulmonary vessel walls in BPD involves numerous growth components, including VEGF-A and ANGPTL2, and other factors, such as angiogenins and extracellular matrix proteins (42). In the current study, lung ANGPTL2 proteins and Ang3 transcripts were suppressed in Nrf2−/− mice relative to Nrf2+/+ mice constitutively and/or after hyperoxia, and 5′-flanking regions of Ang2 and Ang3 contained potential AREs. We speculate that although there was no effect on normal lung maturation, the differential constitutive levels of these proteins between two genotypes may predispose Nrf2−/− neonates to enhanced susceptibility to hyperoxia. Other multiple angiogenic or antiangiogenic genes (e.g., Agtrl1, Tie1, Eng, Il6, and Itga1) were also Nrf2 dependent after hyperoxia exposure. Overall, observations imply a potential adverse effect of Nrf2 deficiency on vessel development and endothelial differentiation in the saccular lung.
In conclusion, Nrf2 modulates genes involved in sustaining lung morphogenesis, cell growth machinery, and lymphocyte immunity during saccular lung maturation (Fig. 7). Nrf2 is also critical to protection of immature lungs from development of oxidative stress phenotypes caused by hyperoxia. Genetic loss of Nrf2 caused augmented oxidation, inflammation, DNA lesions, and aberrant alveolarization of saccular lungs (Fig. 7). Transcriptome analysis suggested that Nrf2 in the immature lung protected against O2 toxicity through regulation of DNA replication and cell cycle, various metabolism and small molecular process, and cell–cell interaction as well as redox homeostasis (Fig. 7). Functional bioinformatics elucidated downstream targets and in vivo validation of the role for Gpx2 and Marco indicated that Nrf2 has a dual role in lung maturation and protection against hyperoxia-induced lung injury. Results contribute to our understanding of the role of Nrf2 in molecular processes of alveolarization and in lung diseases of prematurity, and may suggest a potential therapeutic role for specific Nrf2 inducers (agonists) in protection against human BPD.
Materials and Methods
Mice
Breeding colonies of Nrf2−/− (24), Gpx2−/− (19), and Marco−/− (2) mice were maintained in the National Institute of Environmental Health Sciences (NIEHS) animal facility. ICR (Nrf2+/+ ) and C57BL/6J (Gpx2+/+ and Marco+/+ ) mice were purchased from Taconic and Jackson Laboratory, respectively. The mice and their neonate foster dams (Black Swiss or Swiss Webster; Taconic) were provided food (modified AIN-76A for Nrf2+/+ and Nrf2−/− , NIH-31 for others) and water ad libitum.
Inhalation exposure
Time-dated pregnant mice were cohabitated with appropriate time-pregnant foster dams on their delivery. Neonatal mice at P1 were placed in cages of a hyperoxia chamber, and exposed to 70% or 100% O2 (National Welders) with their foster dams. Air control mice were placed in cages in room air for the same exposure duration. All animal use was approved by the NIEHS Animal Care and Use Committee. Additional detail is provided in the Supplementary Data.
Bronchoalveolar lavage
Whole lung of each neonate was lavaged in situ with Hank's balanced salt solution. The pooled BAL fluid returns were analyzed for total protein content and cell differentials as described in the Supplementary Data.
Histopathology and morphometry
Left lung tissues from each mouse were processed for hematoxylin and eosin staining, immunohistological detection of PCNA and TUNEL staining. Morphometric analysis was done to quantify the TUNEL-positive apoptotic cells (21) and the number of alveoli in the terminal respiratory unit (radial alveolar count, RAC) as a parameter of alveolar simplification (11, 18). Details are in the Supplementary Data.
Protein analysis
Lung proteins (20–100 μg) were subjected to Western blotting using specific primary antibodies (Supplementary Data). ARE binding activity of nuclear protein (5 μg) was determined by gel shift analysis on [γ32P] ATP end-labeled consensus sequence (13).
Redox measurement
Total GSH levels were determined in lung homogenates (60 μg) by a colorimetric method (12). Oxidized protein amount was determined in lung protein aliquots (15 μg) by immunoblotting with anti-2,4-dinitrophenyl hydrazine (DNP) antibody after derivatization of carbonyl moieties using DNP (12). Amount of MDA was detected in BAL fluid (25 μl) for lipid peroxidation (OxiSelect TBARS Assay Kit; Cell Biolabs, Inc.). qPCR was performed to determine DNA base lesions in nuclear and mitochondrial genomes (40) as described in the Supplementary Data.
Microarray
Total lung RNA (n=3/group) was applied to Affymetrix mouse genome 430 V2.0 (Affymetrix, Inc.) in the NIEHS Microarray Core Facility. The array data were analyzed by GeneSpring (Agilent Technologies, Inc.), Ingenuity Pathway Analysis (IPA; Ingenuity Systems, Inc.), and Spotfire (TIBCO Software, Inc.) software. The microarray data are deposited in Gene Expression Omnibus (GEO, accession number: GSE29632) and in NIEHS Chemical Effects in Biological Systems (CEBS, accession number: 005-00003-0012-000-0). Greater details of microarray analysis are in the Supplementary Data.
Reverse transcription–polymerase chain reaction
qPCR (5) was performed on lung cDNA using 240 nM of gene-specific primers (Real Time Primers; LLC) in a 7700 Prism sequence detection system (Applied Biosystems). Semi-quantitative PCR was done for Nrf2 messages (12).
Bioinformatics for ARE
Potential ARE sequences were determined in the 5 kb promoter regions using a PWM statistical model (48). Details are in the Supplementary Data.
Statistics
SigmaStat 3.0 software program (SPSS Science, Inc.) analyzed statistics. Individual t-test was done on TUNEL data (p<0.05). Two-way (Nrf2 mRNA expression) or three-way (other data) analysis of variance was followed by Student-Newman-Keuls test for a posteriori comparisons (p<0.05). Data were expressed as mean±standard error of the mean (S.E.M).
Footnotes
Acknowledgments
This research was supported by the Intramural Research Program and Director's Challenge Program of the NIEHS, National Institutes of Health (NIH), Department of Health and Human Services. Hyperoxia exposures were conducted at the NIEHS. The authors thank Dr. Mary Grant and Mr. Norman Gage for installation of hyperoxia exposure chamber apparatus. The authors also thank Dr. Sue Edelstein of Image Associates for the professional artwork. We thank NIEHS Microarray Core personnel (Dr. Kevin Gerrish and Ms. Laura Wharey), and Ms. Laura Hall in the National Toxicology Program for submitting array data to GEO and NIEHS CEBS. Histopathologic evaluation of neonatal lung was performed by the pathologist Dr. Arlin Rogers (ILS, Inc., Research Triangle Park, NC) with support by Ms. Julie Foley of NIEHS Cellular and Molecular Pathology Branch. Drs. Pierre Buschel and Rick Paules of the NIEHS provided excellent critical review of the article.
Author Disclosure Statement
No competing financial interests exist.
Abbreviations Used
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
