Structural and transcriptomic response to antenatal corticosteroids in an Erk3-null mouse model of respiratory distress


Neonatal respiratory distress syndrome in preterm infants is a leading cause of neonatal death. Pulmonary insufficiency-related infant mortality rates have improved with antenatal glucocorticoid treatment and neonatal surfactant replacement. However, the mechanism of glucocorticoid-promoted fetal lung maturation is not understood fully, despite decades of clinical use. We previously have shown that genetic deletion of Erk3 in mice results in growth restriction, cyanosis, and early neonatal lethality because of pulmonary immaturity and respiratory distress. Recently, we demonstrated that the addition of postnatal surfactant administration to antenatal dexamethasone treatment resulted in enhanced survival of neonatal Erk3 -null mice.


To better understand the molecular underpinnings of corticosteroid-mediated lung maturation, we used high-throughput transcriptomic and high-resolution morphologic analysis of the murine fetal lung. We sought to examine the alterations in fetal lung structure and function that are associated with neonatal respiratory distress and antenatal glucocorticoid treatment.

Study Design

Dexamethasone (0.4 mg/kg) or saline solution was administered to pregnant dams on embryonic days 16.5 and 17.5. Fetal lungs were collected and analyzed by microCT and RNA-seq for differential gene expression and pathway interactions with genotype and treatment. Results from transcriptomic analysis guided further investigation of candidate genes with the use of immunostaining in murine and human fetal lung tissue.


Erk3 –/– mice exhibited atelectasis with decreased overall porosity and saccular space relative to wild type, which was ameliorated by glucocorticoid treatment. Of 596 differentially expressed genes ( q < 0.05) that were detected by RNA-seq, pathway analysis revealed 36 genes ( q < 0.05) interacting with dexamethasone, several with roles in lung development, which included corticotropin-releasing hormone and surfactant protein B. Corticotropin-releasing hormone protein was detected in wild-type and Erk3 –/– lungs at E14.5, with significantly temporally altered expression through embryonic day 18.5. Antenatal dexamethasone attenuated corticotropin-releasing hormone at embryonic day 18.5 in both wild-type and Erk3 –/– lungs (0.56-fold and 0.67-fold; P < .001). Wild type mice responded to glucocorticoid administration with increased pulmonary surfactant protein B ( P = .003). In contrast, dexamethasone treatment in Erk3 –/– mice resulted in decreased surfactant protein B ( P = .012). In human validation studies, we confirmed that corticotropin-releasing hormone protein is present in the fetal lung at 18 weeks of gestation and increases in expression with progression towards viability (22 weeks of gestation; P < .01).


Characterization of whole transcriptome gene expression revealed glucocorticoid-mediated regulation of corticotropin-releasing hormone and surfactant protein B via Erk3 -independent and -dependent mechanisms, respectively. We demonstrated for the first time the expression and temporal regulation of corticotropin-releasing hormone protein in midtrimester human fetal lung. This unique model allows the effects of corticosteroids on fetal pulmonary morphologic condition to be distinguished from functional gene pathway regulation. These findings implicate Erk3 as a potentially important molecular mediator of antenatal glucocorticoid action in promoting surfactant protein production in the preterm neonatal lung and expanding our understanding of key mechanisms of clinical therapy to improve neonatal survival.

Click Video under article title in Contents at

Since the initial observations of Liggins in 1969 that demonstrated dexamethasone-induced inflation of fetal lungs in premature lambs, antenatal glucocorticoids have come to serve as the mainstay of clinical management of preterm birth for the promotion of fetal lung maturation. This practice is now a standard of care in expected preterm birth at >34 weeks of gestation. Surprisingly, despite widespread use, documented clinical efficacy, and focused study, the molecular mechanism by which antenatal steroids function is not understood fully, which underscores our studies herein. Although guidelines on administration recommend a single dose of antenatal glucocorticoids, the most effective dosing regimen remains highly controversial. A better understanding of the molecular mechanism by which this prevalent and life-saving therapy manifests functionally is expected to inform clinical decisions regarding its optimal use.

To investigate the mechanisms behind pulmonary insufficiency, we used our previously described Erk3 ( Mapk6 ) knockout mouse, which are created by targeted disruption of the encoding gene. ERK3 is an atypical mitogen-activated protein kinase with relatively uncharacterized upstream regulators, which have been shown to accumulate during terminal cellular differentiation. In vivo analysis further supports a role for Erk3 in regulation of cellular differentiation in the lungs. Erk3 –/– mice display growth restriction and reduced organ weight without gross abnormalities in organ structure yet exhibit uniform neonatal lethality because of respiratory distress with the major hallmarks of lethal pulmonary immaturity. Histologic evaluation of Erk3 –/– fetal lungs revealed normal branching morphology with impaired saccular development, which suggests that, although overall organogenesis of the lungs is preserved, lung maturation is impaired. Ultrastructural analyses revealed that Erk3 –/– type II alveolar cells contain abundant glycogen granules and attenuated villi when compared with wild type controls, which is consistent with incomplete differentiation.

In our characterization of the model, we tested whether the neonatal mortality rate of Erk3 –/– mice was due to potentially clinically modifiable pulmonary immaturity through transplacental administration of the glucocorticoid dexamethasone. Treatment of pregnant mice at embryonic days 16.5 and 17.5 (during the early saccular stage of fetal lung development) significantly improved the immature histologic architecture of Erk3 –/– lungs, which was demonstrated by the restoration of pulmonary airspace and reduction of the number of immature glycogen-containing cells. Interestingly, despite antenatal dexamethasone treatment Erk3 –/– pups remained cyanotic, dying in the early neonatal interval. Further investigation, however, revealed that the administration of exogenous surfactant improved offspring survival, which suggests a deficiency in surfactant production similar to that seen in clinical neonatal respiratory distress syndrome (RDS). Taken together, the Erk3 –/– model recapitulates morbidity and death observed in the preterm infant and allows examination of the effects of dexamethasone on lung maturity distinct from increased pulmonary surfactant. Having recently demonstrated that surfactant replacement in combination with steroid treatment rescued the lethal postnatal pulmonary defect associated with Erk3 loss, we sought to expand our studies into the antenatal period to identify physiologically and clinically relevant molecular mechanisms by which glucocorticoids mediate fetal lung maturation.

Materials and Methods

Animal husbandry and dexamethasone administration

Animals were housed under pathogen-free conditions according to the procedures and protocols approved by the Institutional Animal Care and Use Committees of both Montreal University and Baylor College of Medicine (AN-4826). Mice were bred by intercrossing heterozygous Erk3 mice, where the presence of a vaginal plug indicated the beginning of gestation (E0.5). On embryonic days 14.5, 16.5 or 18.5 pregnant females were killed by CO 2 euthanasia, and fetuses were removed by cesarean delivery and processed for further analysis. Genotype was determined by polymerase chain reaction (PCR), as previously described. Dexamethasone (0.4 mg/kg; Sandoz, Princeton, NJ) or saline solution was administered intramuscularly to pregnant females on embryonic days 16.5 and 17.5.

RNA sequencing and analysis

High integrity (>8.5 Agilent Bioanalyzer; Agilent Technologies, Santa Clara, CA) coding 3’ polyadenylated messenger RNA (mRNA) was extracted from lungs of E18.5 mice (n = 7) and purified with the Dynabeads mRNA Purification Kit (Invitrogen, Carlsbad, CA). Ambion RNA Fragmentation Reagents (Ambion, Inc., Foster City, CA) were used to fragment mRNA. Double-stranded complementary DNA (cDNA) was made with the use of the Superscript Double-Stranded cDNA Synthesis Kit (Invitrogen, Carlsbad, CA) and random hexamer primers (Invitrogen; 50 ng/μL). DNA sequencing libraries were generated with cDNA according to manufacturer’s protocol. Cluster generation and sequencing was performed on the Illumina cBot station and Illumina Hiseq 2000 (Illumina, San Diego, CA) producing 114,868,849 60bp single end reads ( Supplemental Table 1 ; Supplemental Figure 1 ). Reads were aligned to the mouse genome assembly (NCBI37/UCSC mm9) with Tophat2, and differentially expressed genes were identified with the intersection of Cuffdiff and NOISeq. Programs were run with standard settings and corrected for multiple testing with false discovery rate. Gene pathway interactions were determined with Ingenuity Pathway Analysis (IPA) software (QIAGEN, Redwood City, CA) and DAVID Gene Ontology.

Quantitative real time-PCR (RT-PCR)

Total RNA was isolated from E18.5 lung tissue (n = 12) and purified with the RNeasy kit (Qiagen, Valencia, CA). RNA was reverse transcribed with the High Capacity cDNA Archive Kit with random primers (Applied Biosystems, Foster City, CA). RT-PCR was performed using commercially available sense and antisense oligonucleotides (Applied Biosystems), TaqMan PCR Universal Master Mix (Applied Biosystems), and gene specific TaqMan probes. The Applied Biosystems PRISM 7900HT Sequence Detection System (Applied Biosystems) was used to detect amplification level. Glyceraldehyde-3-phosphate dehydrogenase was used as endogenous control. Relative gene expression was quantitated by the 2 -ΔΔ CT method. The genes studied were corticotropin-releasing hormone ( Crh ), corticotropin-releasing hormone receptor 1 ( Crhr1 ), corticotropin-releasing hormone receptor 2 ( Crhr2 ), insulin-like growth factor 1 ( Igf1 ), insulin-like growth factor 2 ( Igf2 ), surfactant protein A1 ( Sftpa1 ), surfactant protein B ( Sftpb ), surfactant protein C ( Sftpc ), short stature homeobox 2 ( Shox2 ), thyrotropin-releasing hormone ( Trh ), and urocortin 2 ( Ucn2 ).

Histology, immunohistochemistry, and analysis

Human tissue sample collection and experimental protocol was reviewed by the Institutional Review Board (IRB) of Baylor College of Medicine; the determination was that the thoroughly deidentified samples were exempt from IRB approval for use. All samples were obtained initially by Dr Susan Guttentag under IRB approval from The Children’s Hospital of Philadelphia and released to Baylor College of Medicine with a materials transfer agreement. The fetal tissue was from previable neonates and included only tissue from fetuses with no known anomalies or exposure to antenatal steroids.

Tissues from human and murine fetuses (n = 5) were fixed overnight in 10% formalin and embedded in paraffin. Tissue sections were stained with the use of a conventional hematoxylin-eosin protocol. Immunostaining was performed by incubating slides with primary antibody overnight at 4°C, then secondary antibody for 15 minutes, followed by horseradish peroxidase label reagent. Stable DAB Plus (Diagnostic Biosystems, Pleasanton, CA) chromogen was applied for 5 minutes, followed by a Harris hematoxylin counterstain (Fisher Scientific, Pittsburgh, PA) for 15-30 seconds. Primary antibodies used were CRH rabbit polyclonal (1:400; Abcam, Cambridge, MA), surfactant protein A rabbit polyclonal (1:12000; Millipore, Temecula, CA), surfactant protein B (SFTPB) rabbit polyclonal (1:2000; Millipore), and surfactant protein C (SFPTC) rabbit polyclonal (1:4000; Millipore). Negative controls and nonspecific antibodies were included in each immunostaining procedure. Immunostained slides of fetal lungs (n = 5) were examined by 2 blinded reviewers. For each slide, 5 random high-power fields were graded with a 0–5 scale, where 0 indicated the absence of positive staining and 5 indicated intense and diffuse positive staining. Interreviewer reliability was 100%. Image-Pro Plus (Media Cybernetics, Inc., Rockville, MD) was also used to analyze 10 random high-power fields with a grading scale from 0-5 to identify the extent of positive staining (area counts) per high-power fields. The mean grade of each slide was calculated and compared across treated and untreated groups with the use of the independent sample t -test after an equal variance test was performed.

Micro computed tomography (microCT)

Fetal tissues (n = 3) were immersed in cold 4% paraformaldehyde for 24 hours then immersed in a hydrogel stabilization solution that was prepared as described for 3 days at 4°C followed by polymerization at 37°C for 3 hours to preserve tissue structure and conformation. Before imaging, samples were immersed in 0.1N iodine overnight and embedded in 1% agarose, then imaged on the SkyScan 1272 MicroCT (Bruker, Billerica, MA) with the following parameters: 70 kV, 142 μA, 0.5-mm aluminum filter, 2.161 second exposure, 180-degree rotation, and 0.2-degree step angle, which resulted in an isotropic voxel size of 3 μm at 3k total resolution. Scans were reconstructed with SkyScan NRecon (Bruker) and processed with Bruker CT-Analyser by performing binary thresholding of scans with despeckling (<25 voxels) in 3-dimensional space and sweeping for largest object. Volumes of interest (VOI) were defined by manual delineation of lung lobes followed by adaptive boundary definition and ImageJ (National Institutes of Health, Bethesda, MD) was used for semiautomated segmentation of the VOI. The CTAn 3-D analysis plugin (Bruker, Billerica, MA) was used to determine total surface area and volume, density, object porosity (volume of all open plus closed pores as a percent of the total VOI volume), and structural thickness (with a sphere-fitting measurement). Saccular space was determined by the calculation of total lung volume minus tissue volume and conducting airways. Three-dimensional reconstructions were visualized with CTVox (Bruker).

Statistical analysis

Results are given as means±standard deviation. Statistical analysis was performed with 1-way analysis of variance and the Student t tests, and graded histologic data were analyzed with Kruskal-Wallis and Mann–Whitney U tests. Correction for multiple testing used Tukey tests or false discovery rate (FDR) within each experiment with statistical significance accepted at a probability value of less than .05 with the use of the software packages SPSS (IBM Corp., Armonk, NY) and R (R Foundation for Statistical Computing, Vienna, Austria). All statistical analyses for expanded whole transcriptome analysis were performed as described earlier.

Only gold members can continue reading. Log In or Register to continue

May 2, 2017 | Posted by in GYNECOLOGY | Comments Off on Structural and transcriptomic response to antenatal corticosteroids in an Erk3-null mouse model of respiratory distress
Premium Wordpress Themes by UFO Themes