Murine Neonatal Oxidant Lung Injury: NRF2-Dependent Predisposition to Adulthood Respiratory Viral Infection and Protection by Maternal Antioxidant

NRF2 protects against oxidant-associated airway disorders via cytoprotective gene induction. To examine if NRF2 is an important determinant of respiratory syncytial virus (RSV) susceptibility after neonate lung injury, Nrf2-deficient (Nrf2−/−) and wild-type (Nrf2+/+) mice neonatally exposed to hyperoxia were infected with RSV. To investigate the prenatal antioxidant effect on neonatal oxidative lung injury, time-pregnant Nrf2−/− and Nrf2+/+ mice were given an oral NRF2 agonist (sulforaphane) on embryonic days 11.5–17.5, and offspring were exposed to hyperoxia. Bronchoalveolar lavage and histopathologic analyses determined lung injury. cDNA microarray analyses were performed on placenta and neonatal lungs. RSV-induced pulmonary inflammation, injury, oxidation, and virus load were heightened in hyperoxia-exposed mice, and injury was more severe in hyperoxia-susceptible Nrf2−/− mice than in Nrf2+/+ mice. Maternal sulforaphane significantly alleviated hyperoxic lung injury in both neonate genotypes with more marked attenuation of severe neutrophilia, edema, oxidation, and alveolarization arrest in Nrf2−/− mice. Prenatal sulforaphane altered different genes with similar defensive functions (e.g., inhibition of cell/perinatal death and inflammation, potentiation of angiogenesis/organ development) in both strains, indicating compensatory transcriptome changes in Nrf2−/− mice. Conclusively, oxidative injury in underdeveloped lungs NRF2-dependently predisposed RSV susceptibility. In utero sulforaphane intervention suggested NRF2-dependent and -independent pulmonary protection mechanisms against early-life oxidant injury.


Introduction
Lungs of preterm infants born at 24-36 weeks of gestational age are in the saccular phase of lung development. Extensive developmental changes are represented by widening of distal airways for subsequent formation of alveoli, differentiation of type 1 and 2 alveolar cells, and thinning of the air-blood barrier [1]. Alveolarization in fetal and new-born lungs is critical as the type 2 cells not only produce surfactants but are also involved in innate immunity [2]. In addition, type 1 cells differentiated from type 2 pneumocytes are responsible for barrier function and gas exchange [3]. Bronchopulmonary dysplasia (BPD) is a common outcome of the very low birth weight premature infants who required mechanical ventilation and oxygen therapy for acute respiratory distress. It is one of the most frequent causes of chronic respiratory morbidity in survivors of premature birth [4,5]. BPD is characterized by failure of alveolarization leading to alveolar simplification and dysmorphic pulmonary vascularization that results from an imbalance between lung injury and lung repair processes, which interrupts normal lung development and leads to diffuse lung fibrosis [2,6]. Persistent lung impairment later in life leads to long-term negative pulmonary outcomes and BPD survivors often have functional abnormalities and increased risk for adverse respiratory symptoms (e.g., asthma, bronchiolitis, infectious airway diseases, sleep disorders) as young adolescents and adults [7][8][9][10]. Knowledge about BPD pathogenesis has markedly accumulated and advances in neonatal care has improved survival rate in recent years. However, the pathogenesis and molecular mechanisms that lead to lung damage in BPD are not completely understood and there are limited therapeutic options available for prevention and treatment of BPD, which have been only partly satisfactory [8].
Hyperoxia exposure to newborn rodents whose lungs are in the saccular phase has been a model for BPD because their lungs are structurally simplified and functionally affected similar to BPD phenotypes [11,12]. Long-term consequence of neonatally exposed hyperoxia including significantly altered pulmonary function, pulmonary hypertension, and shortened life span were found in rodents [13]. Exposure to early postnatal hyperoxia in mice was also reported to exacerbate cigarette smoke-induced obstructive pulmonary disorder and influenza and rhinovirus infection later in life [14][15][16].
Nuclear factor, erythroid derived 2, like 2 (NFE2L2, NRF2) is a transcriptional inducer of antioxidant response element (ARE)-bearing host defense genes. Studies with mice genetically deficient in Nrf2 (Nrf2 −/− ) found that NRF2-mediated induction of antioxidants and various cytoprotective genes is essential in mitigation of inflammatory and oxidative airway diseases [17]. Lack of Nrf2 exacerbated hyperoxia-induced BPD-like arrest in alveolarization evidenced by lower radial alveolar count and reduced appearance of multilobular alveoli/branched septi as well as severe exudative-phase diffuse alveolar damage including edema and inflammation in saccular stage lung [18,19]. Sulforaphane (4-methylsulfinylbutyl isothiocyanate, SFN) is a bioactive phytochemical antioxidant and one of its mechanisms is through NRF2 stabilization by inhibition of the cytoplasmic suppressor KEAP1 [20,21]. SFN was determined as a potential drug target that can perturb hyperoxia-induced transcriptome changes in mouse neonates [22]. We recently reported that orally administered SFN lessened hyperoxia-induced lung injury in adult mice in a NRF2-dependent manner [23]. In addition, SFN attenuated airway inflammation and injury against bacterial infection following emphysema, respiratory syncytial virus (RSV), and inhaled arsenic [17,23]. However, overall clinical efficacy of SFN is undetermined [24,25]. Furthermore, little is known about prenatal effects of NRF2 modulation or SFN in postnatal respiratory health.
RSV is a ubiquitous seasonal pathogen that may cause acute lower respiratory tract infection in children and infant hospitalization [26,27]. More recent studies have demonstrated increased risk of severe RSV disease among infants with BPD compared to those with no BPD [28,29]. Gestational age is one of the critical risk factors for infant RSV hospitalization rates and appearance of respiratory outcomes [30]. Palivizumab, the humanized RSV monoclonal antibody, has been the only approved option for RSV prophylaxis and it is indicated to reduce RSV-caused serious lower airway infection in high-risk premature infants within six months of the anticipated RSV season [31,32]. NRF2 and antioxidants played a protective role in RSV-induced upper and lower airway disorders and SFN inhibited lung viral replication and neutrophilic inflammation in mice [33,34].
The current study was designed to investigate whether (1) hyperoxia-induced injury in the developmental murine lung affects the later life RSV susceptibility, and (2) maternal supplementation of SFN protects newborn mice against hyperoxic lung injury. We also examined the effect of Nrf2 deficiency in these two models using wild-type (Nrf2 +/+ ) and Nrf2 −/− mice.

Neonate Hyperoxia Exposure
Two days before scheduled delivery, time-pregnant Nrf2 +/+ and Nrf2 −/− mice were cohabitated with time-pregnant foster dams. Upon birth (PND0), neonatal mice from multiple litters with the same prenatal treatment or same genotype were pooled and randomly reassigned to exposure groups with a foster dam per group. The number of mice per cage was kept at a maximum of 12 and matched between comparison groups to control for the effects of the litter size on nutrition and growth. At 24 h after birth (PND1), neonatal mice were placed in cages an inserted into an inhalation chamber and exposed to hyperoxia (UHP grade, Min. purity 99.994% O 2 , National Welders, Durham, NC, USA) continuously for 2 or 3 days with their foster dams. During exposure, foster dams had free access to food (modified AIN76A; Harlan Teklad) and water. The temperature (72 ± 3 • F) and humidity (50 ± 15%) of the chamber were monitored, and mice were exposed to a 12-h light-dark cycle. The entire volume of gas in the chamber was flushed via total system flow every 3-4 min to maintain O 2 concentration consistent and to avoid CO 2 accumulation. Neonatal mice and their foster dams assigned to room air controls were placed in cages placed on a countertop with food and water provided ad libitum for the same exposure duration. The hyperoxia chamber was opened briefly after 24 h exposure to replace foster dams and to check animal health (morbidity and mortality). For neonate hyperoxia end points in the prenatal SFN study, Nrf2 +/+ and Nrf2 −/− pups with foster dams were removed from the chamber at the end of 3-day exposure to hyperoxia or room air and pups were euthanized by sodium pentobarbital overdose (0.02 mL per neonate). For the adulthood RSV study, Nrf2 +/+ and Nrf2 −/− pups and their foster dams were removed from the chamber and foster dams were replaced after the end of 2-day exposure to hyperoxia or room air. Nrf2 +/+ and Nrf2 −/− pups with their foster dam were kept in a virus-and antigen-free room. After weaning at PND14, Nrf2 +/+ and Nrf2 −/− mice were housed individually with water and mouse chow (modified AIN76A) until RSV infection at 5 wk of age.

Bronchoalveolar Lavage (BAL) Analyses
Whole lungs of each neonate from prenatal treatment groups were lavaged in situ, four consecutive times with Hank's balanced salt solution (HBSS, 0.08 mL/2.5 g body weight) using a 24-gauge BD Angiocath cannula (BD Biosciences, San Diego, CA) as previously described in detail [19]. The pooled BAL fluid returns were centrifuged (1000× g, 10 min at 4 • C). Aliquots of lavage supernatants (50 µL) were analyzed for total protein concentration (a marker of hyperpermeability/edema) in Bradford reagent (Bio-Rad, Hercules, PA, USA) following the manufacturer's procedure. Cell pellets from lavage returns were resuspended in 1 mL HBSS and 10 µL aliquots were counted for total cells using a hemocytometer. Aliquots of cell suspension (150 µL) were cytocentrifuged and stained with Wright-Giemsa for differential epithelial and inflammatory cell counts. For young adult mice in RSV study groups, the right lung of each adult mouse was lavaged in situ four consecutive times with HBSS (0.5 mL/25 g b.w.) as indicated elsewhere [37]. Briefly, The BAL fluid returns were centrifuged and an aliquot of the 1st BAL return supernatants (50 µL) and aliquots of pooled cell pellet suspension were analyzed for total protein concentration and cell differentials, respectively, as described above. Lactate dehydrogenase (LDH) content was evaluated in aliquots of pooled (neonates, 50 µL) or 1st (adults, 25 µL) BAL return fluids by a colorimetric assay (Sigma-Aldrich, St. Louis, MO, USA) as a cell lysis and cytotoxicity marker.

Lung Lipid Oxidation Measurement
Lung lipid oxidation levels were measured in BAL fluids by estimation of released malondialdehyde (MDA) which forms 1:2 adduct with added thiobarbituric acid (TBA) in acidic condition. Briefly, aliquots of BAL fluid (75 µL) were treated with acid reagent (75 µL) at room temperature for 15 min and centrifuged (14,000× g, 4 min). Supernatants were incubated with TBA reagent at 47 • C for 2 h. MDA-TBA adducts formation indicated by color changes was determined spectrophotometrically at 532 nm and pre-incubation absorbance at 532 nm was subtracted. The amount of MDA in BAL fluids was quantified from a standard curve (TBARS Parameter Assay; R&D Systems, Minneapolis, MN, USA) that was run in parallel.

Lung Histopathology
Whole lungs of neonate mice were inflated intratracheally in situ with 10% neutral buffered formalin (NBF) and the trachea was ligated. The inflated lung was removed from the mouse and fixed in a contained filled with 10% NBF for 2 days. The fixed left lung was trimmed into two equal sections and tissue blocks were embedded in paraffin and cut into 4 µm thickness for hematoxylin and eosin (H&E) staining. The left lung from each adult mouse was inflated intratracheally in situ with 10% NBF and fixed following the routine procedure described elsewhere [38]. The fixed lung lobe was cut at proximal (around generation 5) and distal (around generation 11) levels of the main axial airway [39], and paraffin-embedded tissue blocks were cut into 5 µm-thickness to be stained with H&E and Masson's trichrome.

Serum Immunoglobulin E (IgE) Measurement
An aliquot of serum (1 µL) from each mouse blood sample was collected by cardiac puncture and diluted in an assay diluent (1:100). IgE levels were colorimetically determined Antioxidants 2021, 10, 1874 5 of 28 using a mouse-specific sandwich ELISA kit (BD Opt EIA). Similarly prepared serial dilution of IgE standards (BD Biosciences) were used for IgE quantification.

SFN Metabolites Detection in Urine and Milk Bands
Non-metabolized (unconjugated) SFN and major conjugated metabolites, including SFN-GSH, SFN-cysteine, and SFN-NAC, in mouse urine and milk bands were analyzed by HPLC-mass spectrometry as previously described [23]. That is, an aliquot (10 µL) of pooled urine from neonates (n = 8-13/group) and foster dams (n = 2 for PBS/Air, n = 2 for SFN/Air) was diluted with 90 µL 5% methanol and 0.1% formic acid. The sample was centrifuged at 8 • C and 80 µL of supernatant was transferred to an autosampler vial containing a low volume insert. Milk bands were weighed and homogenized in 300 µL of 0.1% formic acid using a bead mill. A 30 mg of milk homogenate was diluted to 500 µL with 0.1% formic acid and centrifuged at 8 • C. The supernatant was loaded onto Oasis HLB solid phase extraction cartridges. Cartridges were then washed with 1 mL of 0.1% formic acid, dried with a stream of nitrogen, washed with 1 mL hexane to remove neutral lipids, and eluted with 2 mL of 9:1 acetonitrile:methanol containing 0.1% formic acid. Eluted milk band samples were dried under vacuum then reconstituted in 50 µL 5% methanol/0.1% formic acid. An aliquot (10 µL) of prepared urine or milk sample was injected onto an ACE Excel 3 CN-ES 2.1 × 100 mm HPLC column with gradient separation at a flow rate of 250 µL/min. Electrospray ionization and selected reaction monitoring mass spectrometry were used for analyte quantitation.

Protein Extraction and Detection by Western Blot Analysis
Right lung or placenta tissues (n = 3/group) were homogenized in RIPA buffer containing PMSF (10 µg/mL) and protease/phosphatase inhibitor cocktail (Sigma-Aldrich, St. Louis, MO, USA) and total proteins were isolated by centrifugation. Snap-frozen neonate lung tissues were pulverized in dry ice (two pooled sample/group, four lungs/sample) for nuclear proteins isolation using a kit following the manufacturer's direction (Active Motif, Carlsbad, CA, USA). Quantified proteins were aliquoted and stored at −80 • C until used. Lung nuclear (7 µg) or pooled total placenta (50 µg) proteins were subjected to Western blotting using antibodies against mouse c-Fos (Santa Cruz Biotechnology, Inc., Dallas, TX, USA), lamin B (Santa Cruz, Dallas, TX, USA), phosphor-p42/p44 mitogen-activated protein kinase/extracellular signal-regulated kinase (P-p42/p44 MAPK/P-ERK, Santa Cruz) and pan-actin (Santa Cruz). The assay was done in duplicate and scanned band images were quantitated by densitometry using Image J Gel Analysis software and relative protein band intensities normalize to PBS/air-Nrf2 +/+ (lung c-Fos) or PBS-Nrf2 +/+ (placenta P-ERK) were depicted.

Protein Oxidation Detection
Protein carbonyl groups introduced into protein side chains by oxidative reactions were quantified in lung total protein aliquots by a colorimetric analysis. Briefly, an aliquot of total lung protein (1 µg in 100 µL volume) and oxidized/reduced BSA standards were adsorbed onto a 96-well plate (OxiSelect Protein Carbonyl ELISA kit; Cell Biolabs, Inc., San Diego, CA, USA) overnight at 4 • C, and protein carbonyls were derivatized by 2.4dinitrophenyl hydrazine (DNPH). The DNPH-derivatized protein samples were then incubated with an anti-DNP antibody and an HRP-conjugated secondary antibody in turn following the manufacturer's instructions. The protein carbonyl contents indicating oxidated protein level were determined by colorimetric analysis at 450 nm using a reduced and oxidized BSA standard curve.

Nuclear DNA Binding Assay
Nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) P65-DNA binding activity was quantitated using a transcription factor binding assay kit (Abcam, Cambridge, MA, USA). Briefly, nuclear protein aliquots (2 µg, duplicates) were added to Antioxidants 2021, 10, 1874 6 of 28 ELISA plate wells coated with NF-κB consensus dsDNA sequence and incubated overnight at 4 • C. The wells were then washed and added with NF-κB p65 primary antibody for incubation. After consecutive incubation with secondary antibody and detection reagent, colorimetric analysis at 450 nm determined DNA-bound NF-κB p65 amount.

Lung DNA Damage Measurement
Quantitative polymerase chain reaction (PCR) was used to determine the presence of genomic or mitochondrial DNA lesions (including base modifications, AP sites, and strand breaks) which blocks the progression of DNA polymerase so that only undamaged templates can be amplified [40]. Briefly, total cellular DNA was isolated from frozen lungs using a kit (Qiagen Inc., Valencia, CA, USA), and a fragment of DNA from either nuclei (6.5 kb of DNA polymerase β) or mitochondria (117-bp fragment and 10-kb fragment) were amplified in ABI Prism 7700 Sequence Detection System (Applied Biosystems by Life Technologies, Foster City, CA, USA) using specific primer sets [19]. The amount of DNA amplification is inversely proportional to the amount of DNA damage such that increasing damage causes a decrease in amplification (1,10). Amplification (as determined by picogreen fluorescence) of hyperoxia-treated samples was compared with that of undamaged controls to calculate the relative amplification. Relative amplification values were then used to calculate the average number of lesions per 10 kb of the genomic or mitochondrial DNA, using a Poisson distribution. The amount of damage is presented as ln (mean of hyperoxia amplification/mean of the control samples). The amount of long mitochondrial amplification product (10 kb) was normalized to the mitochondrial DNA copy number using the small 117-bp fragment.

Lung RNA Isolation and Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)
Total RNA was isolated from left lung or placenta homogenates (RNeasy Mini Kit, Qiagen Inc., Dusseldorf, Germany). One µg of RNA was reverse transcribed into cDNAs using GeneAmp PCR System 9700 (Applied Biosystems) and an aliquot of cDNA (40 ng) was added to the reaction for semi-quantitative PCR (1 kb Nrf2 amplicon) as previously described [33] or quantitative PCR in 25 µL reaction containing 12.5 µL 2X Power SYBR Green Master Mix (Applied Biosystems, Foster City, CA, USA) and 240 nM of customdesigned [41] or commercially available (Real Time Primers, LLC, Elkins Park, PA, USA) primers for mouse Nrf2, glutamate-cysteine ligase catalytic subunit (Gclc), glutathione-stransferase P1 (Gstp1), Gstt1, NAD(P)H quinone dehydrogenase (quinone) 1 (Nqo1), ADAM like decysin 1 (Adamdec1), and ubiquitin C-terminal hydrolase L1 (Uchl1) by 10 min hold at 95 • C and up to 45 cycles of 95 • C (15 s)-60 • C (1 min) using an ABI Prism 7700 Sequence Detection System (Applied Biosystems) or CFX Connect Realtime System (Bio-Rad). The target gene expression was calculated using the comparative threshold cycle (C T ) method from the fluorescence detected C T difference between18s rRNA and target gene (∆C T ) in the same sample.

cDNA Microarray Analyses
Total lung RNA (100 ng, n = 3/group) from placenta (E18.5) or neonate lung (PND4) was applied to mouse 430 2.0 arrays (Affymetrix, Inc., Santa Clara, CA, USA) in the NIEHS Microarray Core Facility as described previously [19]. Array data were analyzed statistically using GeneSpring software (Agilent Technologies, Inc., Santa Clara, CA, USA). Briefly, array raw data were filtered by lower expression percentile (at least 1 sample had values within 20% cut-off rage) and the expression levels were normalized to the mean value of the experimental control (lung-PBS/Air/Nrf2 +/+ , placenta-PBS/Nrf2 +/+ ) for each gene by quantile algorithm. Student's t-test (PBS vs. SFN in each genotype for placenta or lung) or two-way ANOVA (PBS vs SFN and air vs. hyperoxia in each genotype lung) identified differentially expressed genes in Nrf2 +/+ or in Nrf2 −/− mice using GeneSpring (Agilent Technologies, Inc., Santa Clara, CA, USA). Overlapping or unique genes in both genotypes were determined by Venn Diagram analyses. Potential Antioxidants 2021, 10, 1874 7 of 28 molecular interactions, biological functions, downstream and upstream pathways were determined by Ingenuity Pathway Analysis (IPA, Qiagen). Microarray data are deposited in Gene Expression Omnibus (accession numbers: GSE164699 for the neonate lungs and GSE164700 for the placenta).

Statistics
Data are presented as the group mean ± standard error of the mean (S.E.M.). The effects of neonate inhalation exposure/adulthood intranasal treatment (air/vehicle, air/RSV, O 2 /vehicle, or O 2 /RSV) and genotype (Nrf2 +/+ , Nrf2 −/− ) in adulthood RSV study or the effects of prenatal treatment/neonate exposure (PBS/air, PBS/O 2 , SFN/air, SFN/O 2 ) and genotype (Nrf2 +/+ , Nrf2 −/− ) in gestational SFN study was evaluated by two-way ANOVA. Nrf2 mRNA assessment in Nrf2 +/+ mice was done by One-way ANOVA. Student-Newman-Keuls test was used for a posteriori comparisons of means (p < 0.05). SigmaPlot 13.0 or 14.0 package (Systat Software, San Jose, CA, USA) was used for all statistical analyses.

NRF2-Dependent Augmentation of Lung Inflammation and Injury
Significantly greater RSV-induced neutrophilic inflammation (1 d post-IN) was found in Nrf2 −/− mice neonatally exposed to O 2 than in those neonatally exposed to air, and the heightened neutrophilic infiltration was remained elevated in neonatal O 2 -exposed Nrf2 −/− mice at 5 d ( Figure 1A). Vehicle treatment also caused significant increase of neutrophils in neonatal air-or O 2 -exposed Nrf2 −/− mice ( Figure 1A). In Nrf2 +/+ mice, there was no significant additive effect of neonatal O 2 exposure in RSV-induced neutrophilic inflammation ( Figure 1A). RSV infection caused significant lymphocyte infiltration in Nrf2 −/− mice among mice pre-exposed to air at 1 d ( Figure 1B). Neonatal O 2 exposure significantly potentiated lymphocytic inflammation in both Nrf2 +/+ and Nrf2 −/− mice after either vehicle or RSV at 1 d ( Figure 1B). RSV-induced lymphocytic inflammation remained elevated by 5 d in Nrf2 −/− mice exposed to either air or O 2 , while there was no significant exacerbation of lymphocytic infiltration by previous O 2 exposure in either genotype of mice at this time ( Figure 1B). Significant RSV-induced epithelial cell sloughing was found in neonatal air-exposed Nrf2 +/+ (5 d) and Nrf2 −/− (1 and 5d) mice ( Figure 1C). BAL epithelial cell number was significantly abundant in Nrf2 −/− mice than in Nrf2 +/+ mice either exposed to prenatal air or O 2 and RSV-induced epithelial injury was exacerbated by neonatally exposed O 2 in Nrf2 +/+ mice at 1 d ( Figure 1C). Consistent with neutrophil numbers, neutrophil chemoattractant CXCL1 level at 1 d was significantly increased by neonate O 2 relative to neonate air in both genotypes while the augmentation by O 2 was more significant in Nrf2 −/− mice than in Nrf2 +/+ mice ( Figure 1D). Proinflammatory cytokine IL-1β contents in BAL were also significantly greater in O 2 /RSV-Nrf2 −/− mice compared to O 2 /RSV-Nrf2 +/+ mice at 1 d ( Figure 1D). BAL LDH levels reflected significantly enhanced lung tissue damage and epithelial cell death in neonatal O 2 -exposed, RSV-infected Nrf2 −/− mice ( Figure 1D).

NRF2-Dependent Potentiation of Serum IgE Increase
Serum IgE level detected as a marker for RSV-induced bronchial epithelial hyperplasia/smooth muscle thickening was significantly greater in Nrf2 −/− mice exposed to neonatal air compared with neonatal air/vehicle-Nrf2 −/− mice at 1 d and 5 d post-RSV ( Figure 3B). Neonatal O 2 exposure further heightened circulating IgE level in Nrf2 −/− mice at 1 d post-RSV, which was not evident in Nrf2 +/+ mice ( Figure 3B).

NRF2-Dependent Augmentation of Lung Oxidative Stress
Oxidation levels of endogenous macromolecules, proteins, and lipids, were measured to estimate the lung oxidative stress at 5 d post-IN. RSV infection significantly increased BAL oxidized lipid contents (determined as MDA levels) only in neonatal O 2 -exposed groups, and it was significantly more increased in O 2 /RSV-Nrf2 −/− mice than in O 2 /RSV-Nrf2 +/+ mice ( Figure 4A). The amount of oxidized protein carbonyl was significantly higher basally (air/vehicle) in Nrf2 −/− mice than in Nrf2 +/+ mice ( Figure 4B). Relative to O 2 /vehicle groups, oxidized protein level in O 2 /RSV groups was significantly greater in both Nrf2 −/− and Nrf2 +/+ mice and it was significantly more potentiated in O 2 /RSV-Nrf2 −/− mice than in O 2 /RSV-Nrf2 +/+ mice ( Figure 4B).

NRF2-Dependent Potentiation of Serum IgE Increase
Serum IgE level detected as a marker for RSV-induced bronchial epithelial hyperplasia/smooth muscle thickening was significantly greater in Nrf2 −/− mice exposed to neonatal air compared with neonatal air/vehicle-Nrf2 −/− mice at 1 d and 5 d post-RSV ( Figure  3B). Neonatal O2 exposure further heightened circulating IgE level in Nrf2 −/− mice at 1 d post-RSV, which was not evident in Nrf2 +/+ mice ( Figure 3B).

Lung Nrf2 and ARE-Bearing Antioxidant Enzyme Expression
Pulmonary Nrf2 mRNA abundance was significantly increased by RSV infection at 1 d and 5 d in air-exposed Nrf2 +/+ mice compared to time-matched air/vehicle-Nrf2 +/+ mice ( Figure 4C). Neonatal exposure to O 2 also significantly upregulated the lung Nrf2 mRNA expression compared to air/vehicle at both time points ( Figure 4C). Compared to air/RSV-Nrf2 +/+ mice, Nrf2 message was significantly heightened in O 2 /RSV-Nrf2 +/+ mice ( Figure 4C). RSV infection significantly induced pulmonary Gstp1 expression at 1 d (in air group) and 5 d (in O 2 group) in Nrf2 +/+ mice ( Figure 4D). Gstp1 and Gclc mRNA levels were significantly enhanced in O 2 /RSV-Nrf2 +/+ mice than in air/RSV-Nrf2 +/+ mice at 5 d ( Figure 4D). Abundance of these antioxidant enzyme genes was relatively or significantly lower in most experimental groups of Nrf2 −/− mice compared to time-and exposurematched Nrf2 +/+ mice ( Figure 4D).

Detection of Gestational SFN Metabolites in Urine and Milk Bands
HPLC-mass spectrometry detected SFN supplemented during gestation and its conjugated metabolites in milk bands (SFN-NAC, SFN) and urine (SFN-Cys, SFN-NAC, SFN) 8 d (PND4) after the last oral administration to foster dams (i.e., E17.5). Primary

Decreased Oxidative Stress and Antioxidant Expression in Neonates by Prenatal SFN
HPLC-mass spectrometry Nrf2 +/+ neonates had lower magnitude of neonatal hyperoxiaincreased protein and lipid oxidation than Nrf2 −/− neonates, and prenatal SFN treatment significantly suppressed the enhanced protein oxidation ( Figure 7A). Gestational SFN significantly suppressed both lung protein and lipid oxidation in hyperoxiaexposed Nrf2 −/− neonates ( Figure 7A). However, the SFN-reduced protein and lipid oxidation levels in hyperoxia-exposed Nrf2 −/− neonates were significantly higher than those in similarly exposed Nrf2 +/+ neonates ( Figure 7A). Lung Nrf2 mRNA expression was decreased in SFN/O 2 -Nrf2 +/+ neonates than in PBS/O 2 -Nrf2 +/+ neonates ( Figure 7B). As we reported previously [19], ARE-bearing antioxidants, Gpx2 and Nqo1, were transcriptionally induced by neonatal hyperoxia only in Nrf2 +/+ neonates ( Figure 7C). Prenatal SFN treatment lowered the hyperoxia-induced antioxidant expressions in Nrf2 +/+ neonates ( Figure 7C), indicating lowered oxidative stress by SFN. This corresponded to the lowered Nrf2 mRNA expression in SFN/O 2 -Nrf2 +/+ pups ( Figure 7B) and was consistent with the results of SFN pretreatment in adult Nrf2 +/+ mice [23,33]. Message levels of Sod2, a NRF2-independent antioxidant enzyme, increased by hyperoxia in both genotypes were also decreased by maternal SFN ( Figure  7C), indicating the role of reactive oxygen species-quenching direct antioxidants including SOD2 in reduced oxidation and inflammation of SFN/O 2 -Nrf2 −/− lungs. metabolites were SFN-Cys and SFN-NAC in neonate an conjugated SFN was minimally detected ( Figure 6A). In n jugated SFN and SFN-NAC were detected at PND4 (Fig  from dams or dams given prenatal PBS had no SFN or its c SFN-NAC) detected. Metabolite levels or composition w exposure.

Validation of Microarray and Pathway Analyses Results
Neonate lung. NF-κB and MAPK/AP-1 have been determined to be potential central molecules in the networks of prenatal SFN-modulated genes in Nrf2 +/+ and Nrf2 −/− lungs exposed to hyperoxia (see Figure 8C,D). Hyperoxia-increased lung nuclear NF-κB (p65)-DNA binding activity was significantly greater in PBS/O 2 -Nrf2 −/− neonates than in PBS/O 2 -Nrf2 +/+ neonates ( Figure 10A). Activity was decreased in SFN-treated lungs of both strains but significant suppression of NF-κB activity was more evident in SFN/O 2 -Nrf2 −/− neonates ( Figure 10A). The magnitude of NF-κB activity in all experimental groups paralleled expression patterns of many immune and inflammatory response genes (see Table 2, Tables S3 and S4). Redox responsive c-Fos, an AP-1 transcription factor, was increased markedly after neonatal hyperoxia exposure in both strains ( Figure 10B). Concurrent with the lung levels of SFN-reduced oxidative stress and direct antioxidant (Sod2) expression in hyperoxia-exposed Nrf2 +/+ and Nrf2 −/− lungs (see Figure 7A,C), the amount of c-Fos in both strains exposed to hyperoxia was lower in SFN-treated groups than in PBS-treated groups ( Figure 10B). In addition, consistent with upregulation of multiple genes encoding DNA damage repair proteins in SFN/O 2 -Nrf2 −/− mice than in PBS/O 2 -Nrf2 −/− mice, hyperoxiaenhanced nuclear DNA damage was significantly reduced in SFN/O 2 -Nrf2 −/− lungs than in PBS/O 2 -Nrf2 −/− lungs. Significantly increased mitochondrial DNA damage was found only in PBS/O 2 -Nrf2 −/− lungs, and prenatal SFN decreased mitochondrial DNA damage in Nrf2 −/− lungs ( Figure 10C). Genomic and mitochondrial DNA damages were significantly less frequent in Nrf2 +/+ mice compared to Nrf2 −/− mice, and they were not significantly altered by gestational SFN (Figure 10C).
Placenta. qRT-PCR confirmed significant induction of ubiquitin carboxy-terminal hydrolase L1 (Uchl1), one of the commonly altered placenta genes by SFN in both mice ( Figure 10D). Prenatal SFN markedly reduced expression of Gstt1 and Adamdec1 expression in Nrf2 +/+ placenta while they were barely expressed in PBS-treated Nrf2 −/− placenta ( Figure 10D). Activated (phosphorylated) protein level of placenta P-ERK1/2, a molecule proposed to be important in the networks of SFN-altered placenta transcriptomics in both genotypes (see Figure 9B,C) was heightened by SFN in both Nrf2 +/+ and Nrf2 −/− placenta ( Figure 10E). . Pathway analysis demonstrated that mitogen-activated protein kinase (MAPK) cascade including extracellular-signal-regulated kinase (ERK) may play central roles in SFN-altered cell survival and organ growth and development transcriptome in Nrf2 +/+ lungs. In Nrf2 −/− lungs, SFN mainly modulated genes facilitating cell morphology and cellular maintenance. (B) In neonate lungs exposed to hyperoxia (O2), transcriptome changes by maternal SFN in Nrf2 +/+ (n = 258) and Nrf2 −/− (n = 239) were different. Pathway analysis indicated that prenatal SFN effect in O2exposed Nrf2 +/+ neonate lung favors to inhibit cell death and potentiate a network of tissue development (C), while SFN in Nrf2 −/− neonate lungs may inhibit upstream regulators such as tumor necrosis factor ligand (e.g., TNF12) to suppress inflammatory response genes and activate cell-to-cell signaling and cellular maintenance network genes (D). In both strains of neonates exposed hyperoxia, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and/or MAPK/AP-1 were predicted to play central roles in crosstalk of SFN-affected lung genes and their influence on cell/tissue functions. Molecular color and its intensity indicate up (red) or down (green) regulation degree of the genes by prenatal SFN compared to prenatal PBS after air or O2 exposure in each genotype. Ingenuity Pathway Analysis and GeneSpring software were used for data analyses. . Pathway analysis demonstrated that mitogen-activated protein kinase (MAPK) cascade including extracellular-signal-regulated kinase (ERK) may play central roles in SFN-altered cell survival and organ growth and development transcriptome in Nrf2 +/+ lungs. In Nrf2 −/− lungs, SFN mainly modulated genes facilitating cell morphology and cellular maintenance. (B) In neonate lungs exposed to hyperoxia (O 2 ), transcriptome changes by maternal SFN in Nrf2 +/+ (n = 258) and Nrf2 −/− (n = 239) were different. Pathway analysis indicated that prenatal SFN effect in O 2 -exposed Nrf2 +/+ neonate lung favors to inhibit cell death and potentiate a network of tissue development (C), while SFN in Nrf2 −/− neonate lungs may inhibit upstream regulators such as tumor necrosis factor ligand (e.g., TNF12) to suppress inflammatory response genes and activate cell-to-cell signaling and cellular maintenance network genes (D). In both strains of neonates exposed hyperoxia, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) and/or MAPK/AP-1 were predicted to play central roles in crosstalk of SFN-affected lung genes and their influence on cell/tissue functions. Molecular color and its intensity indicate up (red) or down (green) regulation degree of the genes by prenatal SFN compared to prenatal PBS after air or O 2 exposure in each genotype. Ingenuity Pathway Analysis and GeneSpring software were used for data analyses. Table 2. Representative lung genes significantly changed by prenatal maternal sulforaphane (SFN) in hyperoxia-exposed wild-type (Nrf2 +/+ ) and Nrf2-deficient (Nrf2 −/− ) neonates.   Supplementary Table S3 (Prenatal SFN-altered genes in hyperoxia-exposed Nrf2 +/+ , 2-Way ANOVA with p < 0.01 n = 258) and Table S4 (Prenatal SFN altered genes in hyperoxia-exposed Nrf2 −/− , 2-way ANOVA with p < 0.01, n = 239). Commonly altered genes in both genotypes are solute carrier family 25 (mitochondrial carrier, dicarboxylate transporter), member 10 (Slc25a10), MIF4G domain containing (Mif4gd), and centrosomal protein 128 (Cep128). ERK was also predicted to play central roles in crosstalk of SFN-modulated Nrf2 −/− placenta genes involved in cellular assembly and organization and organ development. Molecular color and its intensity indicate increased (red) or decreased (green) regulation degree of the genes by maternal SFN compared to maternal PBS in each genotype. Ingenuity Pathway Analysis and GeneSpring software were used for data analyses. ERK was also predicted to play central roles in crosstalk of SFN-modulated Nrf2 −/− placenta genes involved in cellular assembly and organization and organ development. Molecular color and its intensity indicate increased (red) or decreased (green) regulation degree of the genes by maternal SFN compared to maternal PBS in each genotype. Ingenuity Pathway Analysis and GeneSpring software were used for data analyses.  Aliquots of nuclear proteins (7 μg) were subjected to colorimetric DNA-binding activity assay for nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB) p65 which was predicted to be involved in Nrf2 +/+ and Nrf2 −/− neonate lung transcriptomics altered by maternal SFN after hyperoxia (O2) exposure. Group mean ± S.E.M presented (n = 4/group). (B) Western blotting determined nuclear protein level of c-Fos transcription factor which was proposed to play roles in SFN-altered lung transcriptomics in Nrf2 +/+ and Nrf2 −/− neonates exposed to O2. Lamin B level was detected as a housekeeping control for nuclear proteins. Representative images from duplicate analyses presented. Scanned band images were quantitated by densitometry and relative protein band intensities normalize to PBS/air-Nrf2 +/+ were depicted (mean ± S.E.M.). kDa = kilodalton. (C) qPCR was performed to determine SFN-mediated changes in DNA base lesions caused by O2 in nuclear and mitochondrial genomes. Lesion frequencies in genomic (DNA polymerase β gene) and mitochondrial DNA were compared in Nrf2 +/+ and Nrf2 −/− neonate lungs after air or O2. All data were normalized to air-exposed Nrf2 +/+ and group mean ± S.E.M. presented (n = 4/group). Background noise level (dashed lines) is set at ± 0.15. *, significantly different from genotype-matched PBS/Air controls (p < 0.05). +, significantly different from prenatal treatment/neonatal exposure matched Nrf2 +/+ mice (p < 0.05). #, significantly different from genotype-matched PBS/hyperoxia (O2) mice (p < 0.05). (D) Placenta expressions of SFN-altered genes, ubiquitin carboxy-terminal hydrolase L1 (Uchl1), glutathione-stransferase, theta 1 (Gstt1) and ADAM-like, decysin 1 (Adamdec1) were determined by qRT-PCR. Group mean ± S.E.M. presented (n = 3/group). (E) Western blotting determined placenta protein level of activated (phosphor) p42/p44 mitogenactivated protein kinase/extracellular signal-regulated kinase (P-ERK) which was proposed to play roles in SFN-altered placenta transcriptomics of Nrf2 +/+ and Nrf2 −/− mice. Pan-actin level was determined as a housekeeping protein control for total proteins. Representative images from duplicate analyses presented. Scanned P-ERK band images were quantitated by densitometry and relative protein band intensities normalize to PBS-Nrf2 +/+ were depicted (mean ± S.E.M.). kDa = kilodalton. #, significantly different from genotype-matched PBS controls (p < 0.05). +, significantly different from prenatal treatment matched Nrf2 +/+ mice (p < 0.05). (A) Aliquots of nuclear proteins (7 µg) were subjected to colorimetric DNA-binding activity assay for nuclear factor kappa-light-chainenhancer of activated B cells (NF-κB) p65 which was predicted to be involved in Nrf2 +/+ and Nrf2 −/− neonate lung transcriptomics altered by maternal SFN after hyperoxia (O 2 ) exposure. Group mean ± S.E.M presented (n = 4/group). (B) Western blotting determined nuclear protein level of c-Fos transcription factor which was proposed to play roles in SFN-altered lung transcriptomics in Nrf2 +/+ and Nrf2 −/− neonates exposed to O 2 . Lamin B level was detected as a housekeeping control for nuclear proteins. Representative images from duplicate analyses presented. Scanned band images were quantitated by densitometry and relative protein band intensities normalize to PBS/air-Nrf2 +/+ were depicted (mean ± S.E.M.). kDa = kilodalton. (C) qPCR was performed to determine SFN-mediated changes in DNA base lesions caused by O 2 in nuclear and mitochondrial genomes. Lesion frequencies in genomic (DNA polymerase β gene) and mitochondrial DNA were compared in Nrf2 +/+ and Nrf2 −/− neonate lungs after air or O 2 . All data were normalized to air-exposed Nrf2 +/+ and group mean ± S.E.M. presented (n = 4/group

Discussion
The current study demonstrated that oxidative lung injury caused by hyperoxia during postnatal lung development increased pulmonary susceptibility to RSV-induced injury and inflammation in young adulthood, indicating long term respiratory consequences of BPD-like phenotypes. In addition, genetic deletion of murine Nrf2 further exacerbated RSV disease in later life. We also demonstrated that prenatal maternal administration of the phytochemical antioxidant SFN significantly reduced neonatal oxidative lung injury similar to BPD phenotypes in both Nrf2 +/+ and Nrf2 −/− mice. The protective effects of gestational SFN were greater in hyperoxia-susceptible Nrf2 −/− pups than in hyperoxiaresistant Nrf2 +/+ pups, indicating NRF2-dependent and -independent mechanisms of SFN actions. Supporting this outcome, comparative transcriptome analysis in Nrf2 +/+ and Nrf2 −/− lungs determined that maternal SFN modulated different sets of genes involved in similar host defense pathways to facilitate lung cell survival and organ development and to suppress oxidative inflammation and DNA damage responses. Similarly, differential placenta transcriptome changes in Nrf2 +/+ and Nrf2 −/− mice suggested a common beneficial role for prenatal SFN in utero signaling associated with inhibition of perinatal death and activation of angiogenesis.
Increasing evidence indicates that, compared to full-term infants, babies born prematurely are at significantly greater risk of adverse respiratory outcomes including BPD, respiratory failure, and infectious diseases [42][43][44]. Infants with BPD had increased risk of severe RSV disease relative to those without BPD [28,29]. Furthermore, low gestational age (<32 wk) and low birth weight (<1500 g) are critical risk factors for infant RSV hospitalization [30,[43][44][45]. Consistent with these clinical observations, we found that young adult mice that were exposed to hyperoxia at the saccular stage of neonatal lung development carried on persistent lung injury including impaired alveoli and focal fibrotic changes and predisposed to more severe RSV disease. In addition, these long-term outcomes and RSV susceptibility were NRF2-dependent as were the hyperoxia-caused BPD-like lung injury in neonate mice and RSV disease severity in adult mice [19,33]. McGrath-Morrow et al. [18] also demonstrated NRF2-dependent persistent BPD-like phenotypes, and neonatally hyperoxia-exposed juvenile Nrf2 −/− mice had greater delay in lung damage repair relative to similarly-exposed juvenile Nrf2 +/+ mice. Overall, our results indicate an essential role for NRF2-ARE-responsive defense and immune system in protection of prematurity with BPD against later life airway viral infection. Orally administered butylated hydroxyanisole, a NRF2 agonist, protected against RSV-induced lung oxidation, inflammation, viral replication, and airway hyperreactivity [34]. We also reported that RSV replication and lung neutrophilia was significantly decreased by SFN only in Nrf2 +/+ mice [33]. Moreover, SFN pretreatment significantly induced transcriptome of mitochondrial oxidative phosphorylation and energy metabolism and ameliorated hyperoxia-induced acute lung injury in adult mice [23]. The collective observations strongly suggest that therapeutic intervention of NRF2-ARE modulation may significantly impact RSV disease severity, post-BPD phenotypes, and perhaps other oxidative airway disorders in adulthood.
Recent research has highlighted the prenatal effects of functional foods or stimuli on postnatal offspring health to investigate the fetal determinants on human disorders. As we reported in the current study, beneficial effects of prenatal maternal SFN on fetal programming or postnatal health has been reported in model disorders. In spontaneously hypertensive stroke-prone rats, dietary supplementation of SFN precursor during pregnancy reduced oxidative stress and hypertension and offspring also had lower blood pressure and tissue inflammation tissue inflammation [46]. Prenatal SFN precursor diet also markedly decreased breast cancer growth in mice and the anti-carcinogenesis efficacy of prenatal diet regimen was greater than that of early postnatal or adulthood supplementation [47]. In mouse embryos, intraperitoneally given SFN significantly reduced maternal ethanol exposure-induced apoptosis [48]. In newborn mouse skin, maternal intraperitoneal SFN treatment prevented Krt14 mutation-related epidermis loses, and SFN was suggested to reprogram offspring keratin biosynthesis for skin integrity [35]. SFN nanoparticles in early chick culture alleviated heterocyclic aromatic amine compound-induced defects in central and peripheral nervous system and increased the embryo survival rate [49].
Unexpectedly, our results indicated that prenatal SFN-mediated postnatal protection against BPD-like phenotypes are not NRF2-dependent. Prenatal SFN markedly improved hyperoxia-caused severe BPD-like lung injury parameters in Nrf2 −/− pups while we observed relatively marginal protection by in utero SFN in hyperoxia-resistant Nrf2 +/+ pups. SFN is a strong NRF2 and ARE gene inducer for cytoprotection by NRF2 stabilization through inhibition of a cytoplasmic NRF2 inhibitor Kelch-like ECH-associated protein 1 (KEAP1) via its covalent binding to KEAP1 thiol groups [20,21]. However, SFN also acts through other mechanisms including NF-κB inhibition [50,51], MAPK activation [52,53], and histone deacetylase inhibition [54] for anti-inflammation, chemoprevention, apoptosis, and autophagy. From our microarray analyses, we found that gestational SFN supplementation affected different transcriptomes of postnatal lungs in Nrf2-sufficient and -deficient mice. However, pathway analyses indicated that similar biological functions were activated or inhibited by the SFN-altered genes in the two strains to prevent perinatal death and activate organogenesis. In addition, NF-κB, MAPK, and AP-1 were predicted to be common central components of crosstalk among the SFN-altered genes in both strains. These analyses indicated the role for these proteins in NRF2-independent protection by SFN. In support of these findings, we reported inhibition of hyperoxia-increased nuclear NF-κB activity and c-Fos level by SFN in postnatal lungs of two strains. SFN-induced decreased expressions of many NF-κB target inflammatory genes, particularly in Nrf2 −/− lungs (e.g., Ltb, S100a8, Il12a, Igll1, see Table 2 and Table S4), further supported SFN-mediated inhibition of NF-κB pathway in oxidative inflammation for neonate lung protection against hyperoxia. In addition to SFN, known NRF2 agonists include N-acetyl -L-cysteine, synthetic triterpenoids (e.g., CDDO-Im (1[2-Cyano-3,12-dioxooleana-1,9(11)-dien-28-oyl]imidazole), and resveratrol, curcumin, and oltipraz [22,55]. On the other hand, there are pipeline NRF2 antagonists (pharmacological inhibitors) as anti-cancer therapies because NRF2 provides growth advantage of tumor cells [56].
Studies have demonstrated that hyperoxia causes formation of DNA adducts such as 7,8-dihydro-8-oxo-guanine and DNA strand breakage in mouse lungs [57]. Relative to DNA adduct formation common in most lung cell types, phosphodiester backbone breakage is the most severe level of damage and was unique in type 2 cells, and it led to the cell death and lung injury [57]. Genetic loss of Nrf2 dysregulated redox homeostasis which promoted DNA lesions in neonatal lungs exposed to hyperoxia as seen previously [19]. Importantly, we determined in the present study that maternal SFN suppressed hyperoxia-induced nuclear and mitochondrial DNA damage in Nrf2 −/− neonate lungs, and it is postulated to be one of the NRF2-independent protective mechanisms of SFN. This was consistent with upregulation of multiple DNA repair/damage checkpoint genes (e.g., Uimc1, Nbn, Neil3, Smc4, Smc6, Tonsl, Wdr48, see Table S4) by SFN in hyperoxia-exposed Nrf2 −/− lungs. Genes involved in microtubule assembly, chromatin modification, histone acetylation, and genetic imprinting in Nrf2 −/− lungs (e.g., Cep128, Clip1, Dek, Arid4b, Tdrd3, Ehmt1) as well as in Nrf2 +/+ lungs (e.g., Stk10, Cep128, Drosha, Phf20) were also increased by maternal SFN, indicating association of DNA methylation events (see Tables S3 and S4). Evidence indicated that hyperoxia causes increased DNA methylation (hypermethylation) in rat lung genes [58]. Lung epithelial cells also had hyperoxia-induced DNA damage which affected global DNA methylation status [59]. Studies with maternal SFN supplementation also suggested epigenetic alteration which was thought to underlie the positive effects of maternal SFN. In neural crest cells of mouse embryos, ethanol-induced reduction in histone acetylation at the Bcl2 promoter was reversed by SFN, and it may be involved in anti-apoptosis function of SFN [48]. In utero SFN also downregulated histone deacetylase 1 and increased histone acetylation in the mouse mammary tumors for reactivation of tumor suppressor genes (e.g., p16 and p53), and it may have governed the transcriptome changes affecting BRCA1 signaling pathway for anti-carcinogenesis [47]. Investigation of SFN-reprogramed epigenomes in embryonic and extra-embryonic tissues as well as postnatal lungs are further warranted.
A recent study found that NRF2 is critical in placenta integrity for fetal growth and reduced total and labyrinth volume of placenta in Nrf2 −/− mice at E18.5 was predicted to compromise and abolish the adaptational increase of nutrient transfer capacity to meet fetal growth demands [60]. Investigation of the NRF2 pathway detected significant compositional changes of amniotic fluids during late gestation and upregulated ARE-bearing epidermal differentiation complex genes, small proline-rich protein 2 (Sprr2d, Sprr2h), which reverted the defected epidermal barrier formation in loricrin (Lor)-deficient mouse embryos [61]. In the current study, we compared maternal SFN-induced placenta transcriptome changes between Nrf2 +/+ and Nrf2 −/− mice at E18.5. SFN significantly induced multiple epidermal differentiation complex genes (Lor and Lce1 subfamily genes) in Nrf2 +/+ placenta, but not in Nrf2 −/− placenta. However, the Nrf2 −/− placenta differentially responded to SFN and modulated compensatory epidermal/mesenchymal development (e.g., Lect1, Irs2, Arhgap5, Thoc2, Tead1) and angiogenesis (e.g., Hbegf, Kdr, Apold1) genes to reprogram the feto-placenta barrier function. Interestingly, we found that SFN-altered placenta genes such as Adamdec1, H2-D1, Gstt1, and Igfbp1 in Nrf2 +/+ mice were also changed by maternal SFN in fetal lungs of the same gestation age. Pathway analyses also indicated similar functional networks (e.g., organ and system development, lipid metabolism, cell death and survival) affected by maternal SFN in placenta and fetal lungs.
In conclusion, we demonstrated that early-life oxidant-induced acute lung injury had significant consequences later in life on NRF2-dependent RSV susceptibility in mice. We also determined that increased antioxidant conditions in utero potentially contribute to a decreased risk of postnatal airway disease as we found that prenatal antioxidant SFN protected developing lungs from BPD-like oxidative pathogenesis in mice. Among the NRF2-independent protective mechanisms of SFN in Nrf2 −/− lungs were upregulation of DNA damage repair molecules and NF-κB inhibition. Our study provided new insights into the infant oxidant lung injury severity influence on persistence of pulmonary morbidity and the therapeutic intervention for NRF2 agonists. Our results also provided justification for further studies on feto-placental barrier crossing of SFN metabolites and SFN-triggered molecular and epigenetic aspects of maternal cues for barrier and fetal lung signaling.