All experiments were approved under the Animals - UK Scientific Procedures Act 1986, project licence number PPL 40/2293 and personal licence number PIL 40/8001. All efforts were made to minimize suffering and all surgery was performed under general anaesthesia. Time-mated Sprague-Dawley rats were used to create CDH. Day 0 of gestation was taken as the day of discovery of a vaginal plug (term = day 22). To induce left CDH (LCDH), 100 mg of nitrofen was dissolved in 2 mL of olive oil and gavage fed to dams on day 9.5 of gestation ^[10]. Untreated animals served as controls ^[22], ^[23]. Fetal lungs were harvested on days 17.5, 18.5, 20.5 and 21.5 of gestation (Figure 1). Control and nitrofen exposed dams were given a lethal dose of intraperitoneal sodium pentobarbital (100 mg/Kg) and fetuses delivered by caesarean section. Since the vasculopathy is more severe in ipsilateral CDH lung, left lungs were harvested from control and LCDH fetuses ^[24]. Postnatal lung tissue was also harvested from rat pups after a short period of air-breathing. These pups were delivered by caesarean section with the dam under general anaesthesia. Pups were rubbed dry to stimulate them and encourage the onset of spontaneous breathing, placed in a warm incubator at 30°C in ambient air and observed for respiratory effort, movement and colour. Pups from both nitrofen treated and control dams were sacrificed after fixed time periods (between 5 min and 1 h) by cervical dislocation. Control lung tissue from older animals was harvested after cervical dislocation (12 h old pups) or a lethal dose of intraperitoneal sodium pentobarbital (1 week old and adult rats). Lung tissue for Western blotting and immunoprecipitation was snap-frozen in liquid nitrogen and stored at −80°C. Lung tissue for immunohistochemistry was fixed in 4% (w/v) paraformaldehyde in phosphate-buffered saline pH 7.4 (PBS) for 4 h, washed in PBS and then cryoprotected by immersion in 20% (w/v) sucrose overnight. The tissue was then either embedded in gelatine (7.5% (w/v) gelatine in 15% (w/v) sucrose in PBS) and coated with embedding medium or directly coated with embedding medium before freezing by immersion in pre-chilled isopentone at −40°C. Frozen fixed tissue was stored at −35°C until use.

Western blotting of samples from e17.5 and e18.5 lungs comprised 2–3 pooled left lungs. A single left lung was used as an independent sample in older age groups with the exception of 1 week old and adult rats where a left lung fragment of equivalent volume to an e21.5 left lung was used. Tissue was manually homogenized in 0.5–1 mL of ice-cold modified radio-immunoprecipitation assay (RIPA) buffer (50 mM Tris-HCl pH 7.4, 1% (v/v) NP-40, 0.25% (w/v) Na-deoxycholate, 150 mM NaCl, 1 mM EDTA, 1 mM PMSF, 1 mM activated Na3VO4, 1 mM NaF and 1× complete mini protease tablet/10 mL buffer). The homogenate was left on ice for 10 min then centrifuged at 14,000 rpm for 10 min at 4°C. The insoluble material in the pellet was discarded. Total protein content of the supernatant was determined using a bicinchoninic acid kit based on the method described by Smith et al. (1985) and total protein concentration of the supernatant was adjusted to between 0.75 and 1.25 µg/mL with RIPA buffer. Electrophoresis sample buffer (5×, 250 mM Tris-HCl pH 6.8, 10% (w/v) SDS, 5% (v/v) beta mercaptoethanol, 10% (v/v) glycerol, 0.01% (w/v) bromophenol blue) was added to give a final 1× concentration of buffer and the sample was boiled for 5 min.

Samples containing approximately 24 µg of total lung protein were separated in 8, 10 and 12% (w/v) polyacrylamide gels by SDS-PAGE. A C6 BMP4-stimulated cell lysate was used as a positive control for phosphorylated SMAD1/5/8 (p-SMAD1/5/8) and phosphorylated ERK1/2 (p-ERK1/2). Separated polypeptides were electro-transferred onto nitrocellulose membrane. Membranes were blocked by incubating in 5% (w/v) non-fat milk solution/TRIS- or phosphate-buffered saline (TBS/PBS)/0.1% (v/v) Tween-20 for 1 h at room temperature. Membranes were then incubated with primary antibody at 4°C. Antibodies against p-SMAD1/5/8 and p-ERK1/2 were used at 1∶1000 dilution in TBS/0.1% (v/v) Tween-20 with 2.5% (w/v) bovine serum albumin (p-SMAD1/5/8) or 5% (w/v) non-fat milk (p-ERK1/2). Antibodies against ANG-1 and TIE-2 were used at 1∶1000 dilution in PBS/0.1% (v/v) Tween-20 with 5% (w/v) non-fat milk.

Following overnight incubation with primary antibody, membranes were washed five times in TBS/0.1% (v/v) Tween-20 (TBS/T) over 25 min, then incubated with horse-radish peroxidase (HRP) labelled anti-goat (ANG-1, TIE-2), anti-rabbit (p-SMAD1/5/8) or anti mouse (p-ERK1/2) antibodies for 1 h at room temperature. After a further wash cycle, membranes were incubated with chemiluminescence reagent and immunoreactive bands were imaged with the ChemiDoc XRS System using Quantity-One software. After imaging, membranes probed for p-ERK1/2 were incubated with stripping buffer (40 mM Tris-HCl pH 6.8, 0.67% (w/v) SDS, 0.3% beta mercaptoethanol) for 30 min at room temperature, washed in TBS/T and re-incubated overnight with antibodies against ERK1/2 or actin. Membranes were re-washed and incubated with appropriate secondary antibodies. All other membranes were simply washed, re-incubated with anti-actin followed by the appropriate secondary antibody. After a final wash cycle membranes were incubated with chemiluminescence reagent and imaged, as described above.

Lung protein extracts from e20.5 and e21.5 (800–900 µg protein) were made up to 400 µL with RIPA buffer. The protein extract was incubated overnight with 40 µL of anti-phosphorylated tyrosine cross-linked to protein-A agarose beads (clone 4G10) on a rotating wheel at 4°C. The samples were then centrifuged (14,000 rpm, 10 min, 4°C), the supernatant discarded and the agarose bead slurry washed with ice-cold RIPA buffer, then ice-cold PBS to reduce detergent levels. The beads were re-suspended in electrophoresis sample buffer to give a final 1× concentration of buffer and the mixture boiled for 5 minutes. Equal volumes of supernatant were subjected to SDS-PAGE and Western blotting with TIE-2 antibody.

Tissue sections 7 µM thick were cut using a cryostat and thaw-mounted onto slides, dried at room temperature and re-frozen until use. Slides for incubation with p-SMAD1/5/8 antibody were incubated in antigen retrieval buffer (10 mM sodium citrate, pH 6.0) at just below 100°C for 10 min, cooled in the buffer for 30 min then washed. Sections were incubated with blocking buffer (PBS, 20% (v/v) goat serum) for 30 min then incubated overnight at 4°C with p-SMAD1/5/8 antibody in buffer (PBS, 0.1% (v/v) Tween-20, 5% (v/v) goat serum). Slides for incubation with TIE-2 antibody were incubated directly with blocking buffer (PBS, 20% (v/v) rabbit serum) for 30 min then incubated overnight at 4°C with TIE-2 antibodies in buffer (PBS, 0.1% (v/v) Tween-20, 5% (v/v) rabbit serum). Slides were then washed in PBS and sections incubated with fluoroscein isothiocyanate conjugated (FITC) anti-goat antibodies (TIE-2) or anti-rabbit antibodies (p-SMAD1/5/8) in PBS for 2 h. Following further washes the slides were mounted in fluorescent mounting medium and visualized using fluorescence and the appropriate filter. Control sections were treated as above except the primary antibody was omitted.

For diaminobenzidine (DAB) staining, after incubation in blocking buffer (TBS, 20% (v/v) serum) slides were incubated in 0.3% (v/v) hydrogen peroxide in TBS for 15 minutes to block endogenous peroxide activity, then rinsed in TBS. Sections were incubated overnight at 4°C with TIE-2 or p-SMAD1/5/8 antibody in buffer (TBS, 0.1% (v/v) Tween-20, 5% (v/v) serum). The following morning slides were washed in TBS then sections were incubated with biotin conjugated anti-goat or anti-rabbit antibodies in TBS for 45 min, washed and then incubated with a streptavidin-biotinylated peroxidase mixture for a further 45 min. Slides were washed again before incubating with DAB (Fast 3,3′-diaminobenzidine tablet set) for 3–5 min and then thoroughly rinsed with running water. Slides were lightly stained with Mayer's haemalum and 1% aqueous eosin before mounting with DPX mounting medium and visualisation with light microscopy. Control sections were treated as above except the primary antibody was omitted.

Rats were purchased from Charles River Ltd., UK. Nitrofen was purchased from Zheijang Chemicals, China. Embedding medium (Cryo-M-bed) was purchased from Bright Instrument Company, Huntingdon, UK. Complete mini protease tablets were purchased from Roche, Switzerland. The bicinchoninic acid Kit, Tween-20, anti-actin antibody, FITC-conjugated anti-goat and anti-rabbit antibodies and the Fast 3,3′-diaminobenzidine tablet set were purchased from Sigma-Aldrich, UK. Antibodies against p-SMAD1/5/8, p-ERK1/2 and ERK1/2 were purchased from Cell Signaling Technology, USA. The C6 BMP4-stimulated cell lysate was donated by Cell Signaling Technology. Antibodies against ANG-1 and TIE-2 were purchased from R & D Systems, UK. The nitrocellulose membrane (Hi-bond) was purchased from Amersham Biosciences, GE Healthcare UK Ltd. SuperSignal West Dura Substrate chemiluminescence reagent was purchased from Pierce, UK. The anti-phosphorylated tyrosine antibody cross-linked to protein-A agarose beads (clone 4G10) was purchased from Upstate, Millipore Corporation, USA. Biotin conjugated anti-goat and anti-rabbit antibodies, and the streptavidin-biotinylated peroxidase mixture was purchased from Dako, Denmark.

Photon output originating from individual bands of immunoreactivity was recorded as intensity per mm2 (INT/mm2) and only measurements that were not saturated, as determined by the ChemiDoc imager, were used. Background intensity for each membrane was measured and subtracted from the intensity measurements recorded from immunoreactive bands to give normalised density. Protein level per immunoreactive band was calculated relative to actin level to give normalised relative protein levels. Each experiment was repeated between 4 and 7 times and normalised relative band densities reported as medians and inter-quartile ranges (IQR). Statistical differences between median band density for the groups (control and LCDH) at and between each gestational time-point were determined by Mann-Whitney U test using SPSS version 13. Analysis of the trends seen according to gestational age was done by Oneway ANOVA with polynomial contrasts (SPSS version 13). A p-value of <0.05 was considered significant.

To test if increased ANG-1/TIE-2 activity suppresses BMPR signalling to generate PH as postulated in human adult studies, we first tested if ANG-1/TIE-2 activity was abnormally increased in the perinatal CDH lung with its established pulmonary vascular changes typical of PH. Immunoblotting with the antibody to ANG-1 revealed a single band of immunoreactivity corresponding to 75 kDa (Figure 2a). There was a significant change in ANG-1 protein level with the developmental stage of the lung (Figure 2b & c). ANG-1 protein levels in control lung fall significantly from e17.5 and e18.5 to a low level at e21.5. From the low level at e21.5 there is a marked rise by 12 h after birth. The ANG-1 level detected in adult lung was significantly lower than in e17.5, e18.5 and e20.5 fetal lung, as well as being significantly lower than the level recorded in 12 h old lung (Figure 2c). Statistical analysis of the fall in ANG-1 between e17.5 and postnatal by Oneway ANOVA found an unweighted linear trend F value of 6.87 with a p value of 0.015, suggesting a significant linear trend. There was no significant difference between the level of ANG-1 protein between control and LCDH lung (Figure 2b). The decline of ANG-1 protein levels in control lung between the early (e17.5) and late fetal period (e21.5) was also mirrored in LCDH lung with a significant linear trend (F value 12.38, p = 0.002). Later postnatal time-points could not be examined in LCDH lung, since survival for nitrofen-exposed pups is usually less than 30 min.

Immunoblotting with an antibody to TIE-2 revealed a dominant band of immunoreactivity corresponding to a size of 145 kDa (Figure 3a). There was a significant change in the level of TIE-2 and the stage of development of the lung (Figure 3b & c). TIE-2 levels in control lung rise significantly from low levels at e17.5 to a high level in PN lung. From the high level in PN lung there is a fall by 12 h of age and then TIE-2 levels rise to a maximum in adult lung (Figure 3c). Statistical analysis of the rise in TIE-2 between e17.5 and PN by Oneway ANOVA found an unweighted linear trend F value of 21.16 with a p value of <0.0001 suggesting a significant linear trend. There was no significant difference between the level of TIE-2 protein between control and LCDH lung (Figure 3b). Again, the same trend of a rise in TIE-2 protein level was observed in LCDH lung between the early fetal period (e17.5) and postnatal. Analysis of this trend in LCDH lung also showed a significant linear trend (F value 13.17, p = 0.001).

Furthermore, TIE-2 immunoreactivity and activation appeared comparable between normal and CDH lung: TIE-2 localised to the endothelium in fetal (e20.5) and adult pulmonary arteries (Figure 4) with no gross differences in immunohistochemical staining intensity; phosphorylated (activated) TIE-2 protein levels determined by Western blotting were not significantly different between control and LCDH lung at e20.5 and e21.5 (Figures 5a & b).

In the absence of pathological changes in ANG-1/TIE-2 signalling, we next tested if BMPR signalling was being dysregulated by another pathway by comparing normal and perinatal CDH lung for the pulmonary SMAD and ERK1/2 elements of downstream BMPR signalling. Immunoblotting with antibody to p-SMAD1/5/8 revealed a dual band of immunoreactivity corresponding to a size of 55–60 kDa in control and LCDH lung extracts, and the positive control cell lysate (Figure 6a). There was no significant difference between the level of p-SMAD1/5/8 protein between control and LCDH lung at any gestation time studied (Figure 6b). p-SMAD1/5/8 immunoreactivity was localised to the vascular endothelium and the airway epithelium in fetal rat lung (Figure 7). High power images acquired using diaminobenzidine staining (Figures 7 A & B) revealed staining of the cytoplasm and nuclei of the endothelial cells of the pulmonary vasculature. Low power images acquired using FITC staining (Figures 7 C & D) revealed airway epithelial p-SMAD1/5/8 staining, but vascular endothelial staining was not detected by this method. There were no gross differences in localisation or intensity of p-SMAD1/5/8 staining between control and LCDH lung at e19.5 and e21.5.

Immunoblotting with antibody to p-ERK1/2 revealed two bands of immunoreactivity corresponding to sizes of 42 and 44 kDa in control and LCDH lung extract (Figure 8a). Using actin to calculate relative protein expression, there was no significant difference between the level of p-ERK1/2 protein between control and LCDH lung at the gestation times studied (Figure 8b). Using a total ERK1/2 antibody to calculate the protein activity relative to the total amount of ERK1/2 (Figure 8c), it was confirmed that there was no significant difference between the level of p-ERK1/2 protein between control and LCDH lung.