Failed alveolar formation and excess, disordered elastin are key features of neonatal chronic lung disease (CLD). We previously found fewer alveoli and more elastin in lungs of preterm compared with term lambs that had mechanical ventilation (MV) with O2-rich gas for 3 wk (MV-3 wk). We hypothesized that, in preterm more than in term lambs, MV-3 wk would reduce lung expression of growth factors that regulate alveolarization (VEGF, PDGF-A) and increase lung expression of growth factors [transforming growth factor (TGF)-α, TGF-β1] and matrix molecules (tropoelastin, fibrillin-1, fibulin-5, lysyl oxidases) that regulate elastin synthesis and assembly. We measured lung expression of these genes in preterm and term lambs after MV for 1 day, 3 days, or 3 wk, and in fetal controls. Lung mRNA for VEGF, PDGF-A, and their receptors (VEGF-R2, PDGF-Rα) decreased in preterm and term lambs after MV-3 wk, with reduced lung content of the relevant proteins in preterm lambs with CLD. TGF-α and TGF-β1 expression increased only in lungs of preterm lambs. Tropoelastin mRNA increased more with MV of preterm than term lambs, and expression levels remained high in lambs with CLD. In contrast, fibrillin-1 and lysyl oxidase-like-1 mRNA increased transiently, and lung abundance of other elastin-assembly genes/proteins was unchanged (fibulin-5) or reduced (lysyl oxidase) in preterm lambs with CLD. Thus MV-3 wk reduces lung expression of growth factors that regulate alveolarization and differentially alters expression of growth factors and matrix proteins that regulate elastin assembly. These changes, coupled with increased lung elastase activity measured in preterm lambs after MV for 1–3 days, likely contribute to CLD.
- bronchopulmonary dysplasia
- lung development
- mechanical ventilation
- alveolar, lung capillary, and elastic fiber formation
neonatal chronic lung disease (CLD), a modified form of what Northway et al. (53) described as bronchopulmonary dysplasia (BPD), usually develops from lung injury after premature birth. The clinical and pathological features of this condition have changed in recent years owing to major advances in perinatal care, including widespread use of antenatal glucocorticoid and postnatal surfactant treatments and improved respiratory and nutritional support. Animal models of this disorder, featuring long-term mechanical ventilation (MV) of surfactant-treated premature baboons and lambs, have provided important insights on the pathophysiology and treatment of this disease (1, 2, 6–8, 17, 42, 44, 46, 55, 63, 72, 73).
In the ovine model of CLD, lambs that were delivered by cesarean section at ∼80% of term gestation and mechanically ventilated with O2-rich gas for 3 wk had increased lung vascular and respiratory tract resistances compared with control lambs born at term. These physiological abnormalities were associated with increased smooth muscle and elastin in pulmonary arteries and airways, fewer alveoli and lung microvessels, with diminished capillary surface area, and excess, disordered lung elastin compared with term controls (2, 6, 55). Abnormal abundance and distribution of elastin was especially notable in blunted secondary crests, where focal deposits of distally situated elastin normally define loci of future alveoli during lung development. These functional and structural abnormalities of the lung mimic those that have been described in infants with BPD (15, 16, 28, 31, 45, 59, 65).
Abnormalities of lung elastin have been well documented in babies with BPD. Urinary excretion of desmosine, a biomarker of elastin degradation, was shown to increase during the first week of MV in infants with evolving BPD (13). Breakdown of lung elastin in this disease has been attributed to inflammation associated with infection and hyperoxia (12). Lung pathology of infants who have died with severe BPD showed increased accumulation of thickened, tortuous, and irregularly distributed elastic fibers in the distal lung parenchyma, which was associated with reduced septation and fewer alveoli compared with control lungs (45, 62). In a study of very premature infants who were highly susceptible to lung injury and subsequent BPD, treatment with an α1-proteinase inhibitor during the first 2 wk of life showed a significant decrease in the incidence of pulmonary hemorrhage, with a nearly significant reduction (P = 0.06) in the incidence of BPD, defined as O2 dependence at 36 wk postconception (58).
There is considerable evidence that elastin plays a critical role in normal mammalian lung development, contributing to the structural integrity and distensibility of airways, alveoli, blood vessels, and extracellular matrix. Deletion of the elastin gene in mice leads to neonatal death from cardiorespiratory failure associated with reduced terminal airway branching and defective vasculogenesis in the lungs (38, 69). The specific way that abnormal elastin synthesis contributes to failed alveolar and lung capillary formation in BPD is unclear. Several studies have shown that cyclic stretch may induce tropoelastin gene expression in the developing lung, which, in turn, may yield increased elastin deposition (24, 32, 51). There is also evidence that lung inflammation, which typically occurs with acute and chronic neonatal respiratory failure, is associated with increased elastase activity (12, 27, 49, 54, 57, 68), which, in turn, can disrupt normal elastin deposition and alveolar formation.
Our laboratory previously reported that there were fewer alveoli and microvessels and more elastic fibers in lungs of preterm lambs compared with term lambs that had received MV for 3 wk (7). As preterm and term lambs are at different phases of lung development with respect to alveolar and vascular formation and elastic fiber assembly, we surmised that an extended period of cyclic stretch, coupled with O2 exposure, might yield different expression patterns of genes considered important in lung development. We, therefore, tested the hypothesis that prolonged MV with O2-rich gas would reduce pulmonary expression of genes that regulate alveolar and lung vascular formation, specifically vascular endothelial growth factor (VEGF)-A and one of its receptors, VEGF-R2 (also known as fetal liver kinase-1), and platelet-derived growth factor (PDGF)-A and one of its receptors, PDGF-Rα, while increasing lung expression of genes that regulate synthesis and assembly of elastin, namely transforming growth factor (TGF)-α, TGF-β1, tropoelastin, fibrillin-1, fibulin-5, lysyl oxidase-like 1, and lysyl oxidase. We postulated that these effects would be most apparent in preterm lambs, whose incompletely developed lungs, despite surfactant treatment at birth, required greater inflation pressures to achieve normal respiratory gas exchange than did the more mature lungs of lambs born at term.
Animals and Preparation for Experiments
We used mixed-breed newborn lambs, some of which were delivered prematurely by cesarean section at ∼125 days of gestation (term = 147 days), and some that were born vaginally after spontaneous labor at term gestation, followed by MV for either 1 day, 3 days, or 3 wk after birth. Table 1 lists descriptive data for the two groups of lambs. Weight gain was greater in term compared with preterm lambs that received MV for 3 wk, presumably the result of feeding intolerance and reliance mainly on intravenous (iv) nutrition in preterm lambs. For the short-term studies of MV (1–3 days), there were 16 preterm and 16 term lambs, 8 per group at 1 day and 3 days respectively, that were managed in a similar way, as described below. Control animals included groups of four to six preterm fetuses whose unventilated lungs were harvested at ∼125 days of gestation (same gestational age as the ventilated preterm lambs were at birth), and groups of term fetuses whose unventilated lungs were harvested at ∼145 days of gestation (same gestational age as the ventilated term lambs were at birth). In addition, we used groups of four to six unventilated, spontaneously breathing term lambs that were killed either 1 day or 3 wk after birth to serve as respective controls for the preterm and term lambs that had MV for up to 3 wk. Lung physiological and structural data for four of the term lambs that received MV for 3 wk was included in an earlier report (7).
Preparation of animals for experiments.
For operative delivery of preterm lambs, time-dated pregnant ewes received spinal anesthesia for a midline hysterotomy, through which catheters were placed in a carotid artery, jugular vein, and hindlimb artery and vein of the fetus under 1% lidocaine local anesthesia. Residual lung liquid was withdrawn from the fetal trachea, followed by injection of 10 ml of calf lung surfactant (35 mg/ml Infasurf, gift of Ony, Amherst, NY) into the lung lumen through a 4.5- or 5-mm endotracheal tube just before delivery.
Lambs received MV for up to 3 wk, initially with a high-frequency jet ventilator (Bunnell, Salt Lake City, UT) set at a frequency of 420 cycles/min, mean airway pressure 7–8 cmH2O, sufficient inflation pressure to keep arterial Pco2 (PaCO2) at ∼35–45 Torr, and sufficient inspired O2 to keep arterial Po2 (PaO2) at 50–90 Torr. We used high-frequency MV immediately after birth to minimize tidal volumes and peak inflation pressures, thereby reducing risk of extrapulmonary air leaks and early death while attaining adequate respiratory gas exchange.
After an initial period of stabilization, usually ∼24 h after birth, the lambs were switched to conventional MV with a time-cycled, pressure-limited infant respirator (model IV-100B, Sechrist, Anaheim, CA) set at a rate of 20 breaths/min, inspiratory time 0.75 s, and end-expiratory pressure ∼5 cmH2O. Peak inflation pressure was adjusted to keep PaCO2 at ∼35–45 Torr, and the fraction of inspired oxygen was adjusted to keep PaO2 at ∼50–90 Torr. The intent of the switch from early high-frequency MV to sustained conventional MV, using a slow rate, long inspiratory time, and relatively low target level of PaCO2, was to induce cyclic lung stretch with O2-rich gas over a prolonged period and thereby create CLD.
Term lambs were born vaginally after spontaneous labor. They did not receive surfactant, as they were regarded to be mature newborns with a normal complement of endogenous surfactant in their lungs. With the use of 1% lidocaine local anesthesia, we placed polyvinyl catheters in their carotid artery, jugular vein, femoral artery, and saphenous vein soon after birth, followed by MV for the predetermined duration of study.
All lambs were managed on a padded platform bed beneath a radiant warmer. Arterial blood pressure was monitored continuously with a calibrated pressure transducer connected to an electronic recorder, and during the stabilization period the lambs received an iv solution of glucose and saline to maintain normal blood glucose and sodium concentrations. The lambs also received buprenorphine (Buprenex, Recritt and Coleman Pharmaceuticals, Richmond, VA), 0.01 mg/kg iv, soon after birth and as often as needed thereafter to prevent agitation.
Lambs that were selected before birth to receive MV for 3 wk had surgery within 2 days of birth to allow subsequent measurements of pulmonary vascular resistance (PVR). They received general anesthesia with iv fentanyl (Abbott Laboratories, North Chicago, IL), an initial dose of 15–20 μg/kg followed by doses of 10 μg/kg as often as needed to prevent intraoperative tachycardia or hypertension. As previously described (6–8), we did a left thoracotomy to ligate the ductus arteriosus and place catheters in the pulmonary artery and left atrium. A 10- or 12-mm ultrasonic flow probe (Transonics Systems, Ithaca, NY) was implanted around the main pulmonary artery for subsequent measurement of cardiac output (CO) via a detachable cable that was connected to a calibrated flow meter (model T106; Transonics Systems). We placed an 8-Fr catheter in the left pleural space for postoperative drainage of air and liquid, and a silicone rubber balloon catheter that was used for later measurement of pleural pressure during assessment of lung mechanics, as described below.
Postoperatively, the lambs received buprenorphine, 0.01 mg/kg iv, every 4–6 h for analgesia. Lambs that had MV for only 1 or 3 days did not undergo a thoracotomy and therefore did not have measurements of pulmonary vascular pressures or CO from which to calculate PVR. All lambs received iv buprenorphine and either pentobarbital (Vet Lab, Lenexa, KS), 3–5 mg/kg, or phenobarbital (Wyeth Laboratories, Philadelphia, PA), 10 mg/kg, as needed for sedation during MV. We sampled arterial blood hourly to measure pH, PaO2, and PaCO2, and adjusted ventilator settings accordingly. Hematocrit was measured twice daily, and filtered maternal blood was transfused if the hematocrit was <35%. Maternal plasma or isotonic saline was infused if the mean arterial blood pressure was <40 mmHg.
Lambs that had MV for 3 wk received iv penicillin, 100 mg/kg every 12 h, and gentamicin, 2.5 mg/kg every 24 h, for at least a week after birth. If signs of sepsis developed thereafter, alternative broad-spectrum antibiotics were given. Lambs that had MV for either 1 or 3 days were treated with penicillin and gentamicin for the duration of their studies. Nutrition was provided with iv solutions containing glucose, protein (trophamine), electrolytes, trace metals, and vitamins, and with ewe's milk that was given through an orogastic tube. Because term lambs tolerated enteral feedings much better than preterm lambs did, they required less iv nutrition than did the preterm lambs, although they received ample amounts of iv glucose solution to supplement their feedings. All lambs were weighed daily to monitor fluid balance and nutritional status. Serum electrolytes were measured at least once each day with ion-selective electrodes (Na/K/Cl Stat Analyzer, model 644, Ciba Corning Diagnostics, Medfield, MA) to guide fluid and electrolyte management, and blood glucose concentrations were monitored with a rapid detection device (Exactech Medisense, Waltham, MA). Urine output was determined from diaper weights before and after each voiding. All surgical and animal care procedures and experimental protocols were approved by the Institutional Animal Care and Use Committees at the University of Utah and State University of New York at Buffalo, where the animal studies were performed.
Physiological studies, which were conducted weekly on the lambs that received MV for 3 wk, included measurements for 4–8 h of steady-state mean pulmonary arterial (Ppa) and left atrial (Pla) pressures and CO, which was monitored continuously via the pulmonary artery flow probe to calculate PVR [(Ppa − Pla)/CO]. We measured vascular pressures with calibrated pressure transducers (BT3DC, Statham Instruments) connected to an eight-channel amplifier recorder (model 7D, Grass Instruments, Quincy, MA). Respiratory variables were assessed weekly from simultaneous measurements of proximal airway and pleural pressures and gas flow that was measured using a calibrated pneumotachograph connected to a PEDS Pulmonary Evaluation and Diagnosis System (Medical Associated Services, Hatfield, PA) (5). Measured variables included tidal volume, minute ventilation, dynamic lung compliance, and expiratory resistance. The endotracheal tube cuff was inflated for these studies. Arterial blood pH, PaO2, and PaCO2 were measured before and after assessment of PVR and lung mechanics.
Postmortem Studies of Lung Tissue
At the end of each study, the lamb was anesthetized with iv pentobarbital, 35 mg/kg, followed by a midline sternotomy to open the chest and resect the lungs at peak inflation pressure so that lung histopathology could be related to the physiological studies that were done just before death. Lung lobes were double-clamped and excised with the clamps still attached to retain air volume and vascular and air space contents. The clamped right middle lobe was immersed in Carnoy's fixative for quantitative histology, and the left lingula was immersed in 10% buffered formalin for subsequent immunohistochemistry. Portions of lung were frozen in liquid nitrogen and stored at −80°C for later measurements of enzyme activity and tissue levels of various genes and proteins that affect alveolar and elastic fiber formation, as described below.
We applied quantitative image analysis to compare specific structural features of lungs obtained from lambs that had MV for 3 wk and term control lambs that breathed without assistance for 1 day or 3 wk. To measure radial alveolar counts as an index of alveolar number, and to measure elastic fiber density as an index of parenchymal elastin content, we used a design-based method to systematically and randomly sample the lungs for quantitative histology, as previously described (2, 9, 20). Methodological details are described in the online supplement. (The online version of this article contains supplemental data.)
Serine elastase and myeloperoxidase activity assays.
Lung tissue was snap-frozen in liquid N2 and stored at −80°C for measurement of serine elastase activity by a modification of a previously described method (64, 74) that uses DQ-elastin substrate (Molecular Probes, Eugene, OR). We applied a previously described method (56) to measure myeloperoxidase (MPO) activity in lungs harvested from preterm and term lambs that had MV for up to 3 wk after birth. Details of these assays are described in the online supplement.
RNA extraction and quantitative real-time PCR.
Total RNA was extracted from frozen lung tissue with TRIzol (Invitrogen Life Technologies, Carlsbad, CA) and purified with RNeasy Mini Kit columns (Qiagen, Valencia, CA). cDNA was synthesized from 2 μg of purified RNA using reverse transcriptase (SuperScript III Kit, Invitrogen), according to the manufacturer's instructions. Amplification was performed in triplicate at 50°C for 2 min, 95°C for 10 min, followed by 40 cycles at 95°C for 15 s and 60°C for 1 min. cDNA products were used as templates for quantitative real-time polymerase chain reaction (qRT-PCR) to measure gene expression levels using either a SYBRgreen System (Applied Biosystems, Foster City, CA) or the preverified Assays-on-Demand TaqMan (Applied Biosystems) primer/probe set (used to measure fibulin-5 mRNA and PDGF-Rα mRNA). Reactions without template and/or enzyme served as negative controls. Standard curves were plotted for each target gene and internal control (GAPDH mRNA or 18S rRNA). RNA quantity for each gene was expressed relative to the internal control. In separate experiments, we found that GAPDH mRNA and 18S rRNA did not change significantly in lungs harvested from the various groups of fetal and newborn lambs that were studied (online supplement, Figs. 1A and 2A). Primer sequences used for qRT-PCR with the SYBRgreen system are listed in the online supplement.
Protein extraction and Western immunoblots.
Frozen lung tissue was homogenized at 4°C in a solution containing 50 mM Tris·HCl, pH 7.5, 20 mM 3-[(3-cholamidopropyl)dimethylammonio]-1-propanesulfonate, 3 mM DTT, and Halt protease inhibitor cocktail (Pierce Biotechnology, Rockford, IL). Samples were mixed at 4°C for 1 h on a rocker platform, centrifuged at 14,000 g for 15 min, and supernatants were stored at −80°C. Protein concentration was measured in aliquots of tissue extracts using the DC Protein Assay Kit (Bio-Rad Laboratories, Hercules, CA). Proteins were separated on SDS-PAGE gels (Invitrogen) and blotted electrophoretically onto 0.45-μm nitrocellulose membranes (Invitrogen). Membranes were blotted with a solution of 5% nonfat milk and 0.05% Tween 20 in PBS buffer at room temperature for 1 h. Membranes were incubated with specific antibodies, diluted in a solution of 5% nonfat milk and 0.05% Tween 20 in PBS buffer under conditions described below.
After incubation with the diluted primary antibody solution, the membranes were washed three times in a solution containing 0.05% Tween 20 and PBS buffer. Membranes were incubated with a dilute solution of a secondary antibody (details below) at room temperature for 2 h and washed three times in a solution containing 0.05% Tween 20 and PBS buffer. Images were detected by chemiluminescence (ECL or ECL Plus Detection Kit, Amersham Life Science, Piscataway, NJ) and quantified by densitometry (Bio-Rad).
Immunoblot analysis was applied to measure lung proteins as follows: VEGF, using a 1:100 dilution of VEGF mouse monoclonal antibody (sc-7269, Santa Cruz Biotechnology, Santa Cruz, CA); PDGF-A, using a 1:200 dilution of mouse monoclonal PDGF-A antibody (sc-9974, Santa Cruz); PDGF-Rα, using a 1:400 dilution of rabbit polyclonal PDGF-Rα antibody (sc-338, Santa Cruz); fibulin-5, using a 1:500 dilution of rabbit polyclonal fibulin-5 antibody (the generous gift of R Mecham, Washington University, St. Louis, MO); and lysyl oxidase, using a 1:300 dilution of rabbit polyclonal lysyl oxidase antibody (39). All antibodies were incubated overnight at 4°C. Visualization of bound antibodies was performed using chicken anti-mouse IgG conjugated to horseradish peroxidase (sc-2954, Santa Cruz) or donkey anti-rabbit IgG conjugated to horseradish peroxidase (NA934, Amersham). Membranes were stripped and reprobed with a 1:10,000 dilution of rabbit polyclonal anti-β-actin antibody (no. ab8227, Abcam, Cambridge, MA) and incubated at room temperature for 1 h to provide an internal loading control. For some of the genes that we studied, there were no available antibodies that cross-reacted with sheep tissue for immunoblots, and we therefore report results for mRNA without corresponding quantitative protein measurements (e.g., VEGF-R2, TGF-α).
Tissue localization of specific proteins (TGF-α, TGF-β1, and PDGF-Rα) was done by immunostaining 5-μm-thick sections of formalin-fixed, paraffin-embedded lung. Briefly, tissue sections were deparaffinized and prepared for immunoperoxidase staining using the Elite VectaStain ABC kits for mouse and rabbit primary antibodies (Vector Laboratories, Burlingame, CA). Antigen detection was enhanced by treatment with pronase (Roche, Basel, Switzerland), 1 mg/ml, in PBS at room temperature for 10 min (TGF-α and PDGF-Rα), or with sodium citrate, 10 mM, pH 6, at 95°C for 10 min, followed by cooling for 20 min (TGF-β1), before applying the primary antibody. Endogenous peroxidase activity was removed by incubation with 3% H2O2 for 30 min. Sections were treated for 1 h at room temperature with blocking serum (10% goat serum/0.1% Tween 20 in PBS for rabbit polyclonal antibodies, 10% rabbit serum/0.1% Tween 20 in PBS for mouse monoclonal antibodies) before the primary antibody was applied. The following primary antibodies were used: TGF-α, 10 μg/ml, mouse monoclonal anti-human antibody (no. GF10, Calbiochem, San Diego, CA); TGF-β1, 1:300 dilution, rabbit polyclonal antibody (sc-146, Santa Cruz); and PDGF-Rα, 1:400 dilution, rabbit polyclonal antibody (sc-338, Santa Cruz). Primary antibodies were applied for 16 h at 4°C. Immune complexes were stained by incubation in a solution containing 0.5 mg/ml of 3,3′-diaminobenzidine in 50 mM Tris buffer (pH 7.4) and 0.015% H2O2 (DAB Peroxidase Substrate Kit, Vector Laboratories, catalog no. SK4100). Sections were counterstained with hematoxylin. Negative control slides were stained using the same procedure, substituting buffer solution for primary antibody.
Data in the text, tables, and figures are expressed as means ± SD, unless denoted otherwise. When comparing data sets that displayed a normal Gaussian distribution, we used Student's unpaired t-test to assess for significant differences in physiological and histological data between the two groups of preterm and term lambs that received MV for 3 wk (75). For data sets that had a skewed non-Gaussian distribution, we applied the nonparametric Mann-Whitney test to assess for significant differences. We used one-way analysis of variance and Student-Newman-Keuls multiple-comparison tests to identify differences in gene expression and protein abundance in lungs of preterm and term lambs over time. Statistical analysis was done using the Prism 4 software package (GraphPad, San Diego, CA). Differences were considered significant if the P value was <0.05.
Table 2 shows results for respiratory and pulmonary vascular variables that were measured at the end of weeks 1, 2, and 3 in the groups of preterm and term lambs that received MV for 3 wk. While there were no significant differences in fraction of inspired oxygen, arterial pH, or PaCO2 between groups at the end of week 3, peak inflation and mean airway pressures were significantly greater in the preterm lambs, as were tidal volume and lung resistance. Dynamic lung compliance was similar in preterm and term lambs after 3 wk of MV. There were no significant differences between groups at the end of week 3 with respect to Ppa or Pla, whereas pulmonary blood flow was greater in term lambs than it was in preterm lambs, thereby yielding a lower PVR in the more mature animals. If, however, pulmonary blood flow at the end of week 3 is expressed relative to lung weight, PVR was not significantly different in preterm and term lambs.
Lung Elastin in Preterm and Term Lambs
As noted in Fig. 1, after 3 wk of MV, elastin accumulation in lung parenchyma was five times greater in preterm compared with term lambs (elastic fiber density, as percent lung parenchyma, 19 ± 3 vs. 4 ± 1%, significant difference, P < 0.05), and the distribution of elastin was very different between the two groups. Elastin was expressed mainly at the tips of alveolar septa and around blood vessels in the term lung, whereas elastic fibers were prominent throughout the walls of distal respiratory units in the preterm lung, in which there was a striking paucity of alveolar septa. Alveolar number, as assessed by radial alveolar counts, was nearly three times greater in the lungs of term compared with preterm lambs (10.2 ± 2.3 vs. 3.6 ± 1.5 per terminal respiratory unit, significant difference, P < 0.05).
VEGF-A, VEGF-R2, PDGF-A, and PDGF-Rα Gene Expression in Lungs of Preterm and Term Lambs
Because of our laboratory's previous observations that radial alveolar counts and capillary surface density were significantly less in chronically ventilated preterm lambs compared with term lambs (7), we used qRT-PCR to measure expression of VEGF-A mRNA and VEGF-R2 mRNA, genes that are known to affect alveolar septation and angiogenesis, in lung tissue obtained from preterm and term lambs that had received MV for 1 day, 3 days, or 3 wk, compared with unventilated fetal controls. As shown in Fig. 2, both VEGF-A mRNA and VEGF-R2 mRNA were significantly reduced, compared with fetal controls, in both preterm and term lambs that had MV for 3 wk. Lung mRNA expression of VEGF-A and VEGF-R2 in lambs killed after 3 wk of MV was significantly less than the corresponding lung mRNA levels measured in unventilated control lambs that were killed 1 day or 3 wk after birth at term (respective postconceptional age controls for preterm and term lambs with CLD, data not shown).
VEGF-A protein abundance also was less in the lungs of the chronically ventilated lambs compared with controls, with similar expression patterns seen for both preterm (Fig. 3A) and term (Fig. 3B) lambs. The only difference in VEGF-A protein abundance between preterm and term lambs was noted after 1 day of MV, when VEGF-A protein was significantly less in preterm lambs than it was in term lambs (data not shown).
We measured lung expression of PDGF-A mRNA in our chronically ventilated preterm and term lambs, based on the reported observation that mutant mice lacking PDGF-A exhibit failed alveolar septation and abnormal lung elastin, with subsequent development of pulmonary emphysema (11, 40). The pattern of expression of PDGF-A mRNA was similar in lungs of preterm and term lambs (Fig. 4A), showing a significant reduction after 3 wk of MV. Lung abundance of PDGF-A protein also was reduced in the preterm lambs with CLD (Fig. 4B).
Because of a previous report showing that PDGF-A/PDGF-Rα signaling is required for mammalian lung growth and alveolar formation (10), we measured lung expression of PDGF-Rα in preterm lambs with CLD compared with preterm and term fetal controls (Fig. 5A). Abundance of both PDGF-Rα mRNA and protein was significantly reduced after 3 wk of MV, consistent with the reduced lung expression of PDGF-A mRNA and protein in lambs with CLD. Lung sections that were immunostained for PDGF-Rα showed abundant staining of airways and walls of distal air spaces in preterm and term fetuses, with little or no immunostaining detected in distal lung of preterm lambs with CLD (Fig. 5B).
Lung Expression of Elastin-Related Genes in Preterm and Term Lambs
The discovery of marked differences in abundance and distribution of elastin in lungs of preterm compared with term lambs that had MV for 3 wk prompted us to measure lung expression of several genes known to be important in elastin synthesis and assembly, including tropoelastin, fibrillin-1, fibulin-5, lysyl oxidase-like 1, and lysyl oxidase. We also measured lung mRNA expression of specific growth factors that are known to affect lung elastin accumulation, namely TGF-α and TGF-β1.
Tropoelastin gene expression increased abruptly on the first day after birth and then decreased between 3 days and 3 wk after birth in preterm lambs (Fig. 6A). In lambs born at term, the increase in tropoelastin mRNA was less in both magnitude and duration than it was in the preterm lambs. Thus tropoelastin gene expression in lung was significantly greater at all three postnatal time points in preterm compared with term lambs. Lung expression of fibrillin-1 mRNA, which has an important role in establishing and maintaining the microfibril scaffold on which tropoelastin is assembled, also increased in preterm lambs after 1 day of MV, but then decreased after 3 days and 3 wk of MV (Fig. 6B). Fibrillin-1 mRNA did not change at any of the postnatal time points in lambs born at term, such that lung expression of this gene was significantly greater in preterm than in term lambs at 1 and 3 days after birth. Lung expression of fibulin-5 mRNA did not change significantly in either preterm or term lambs at any of the three postnatal time points at which measurements were made (Fig. 6C). Fibulin-5 mRNA was significantly greater in preterm fetuses and 1-day-old newborns than in term fetuses and newborns, respectively. There was no significant difference in fibulin-5 protein in lungs of preterm lambs with CLD compared with fetal controls (Fig. 7A). Lung mRNA expression of lysyl oxidase-like 1 increased in both preterm and term lambs after 1 and 3 days of MV, and then fell to near fetal levels after 3 wk (Fig. 6D). Lung expression of lysyl oxidase mRNA transiently increased in preterm lambs after 1 day of MV and then decreased to less than fetal levels after 3 wk (Fig. 6E). In lambs born at term, lysyl oxidase mRNA also increased transiently and then decreased, but lung expression of lysyl oxidase mRNA was consistently greater in preterm than in term lambs. Lysyl oxidase protein was less in lungs of 3-wk-old preterm lambs with CLD than it was in preterm control fetuses (Fig. 7B).
Lung mRNA expression for both TGF-α and TGF-β1 increased significantly after 1 and 3 days of MV in preterm lambs, but did not change significantly during MV of term lambs (Fig. 8, A and B). Sections of preterm fetal lung tissue showed scant immunostaining for TGF-α protein in the walls of distal air spaces, whereas sections of lung obtained from preterm lambs that had received MV for 1 day showed abundant TGF-α protein that was distributed throughout the walls of distal air spaces and in airway epithelium (Fig. 8C). Likewise, TGF-β1 immunostaining was minimal in distal lung parenchyma of preterm fetal lambs, whereas immunostaining was prominent in the walls of distal air spaces, in septa, and in airway epithelium of preterm lambs that had received MV for 1 day (Fig. 8D).
Serine Elastase Activity in Lungs of Preterm and Term Lambs During MV
The finding that elastin deposition was very different in the chronically ventilated preterm lung compared with the term lung, coupled with differences in the expression patterns of elastin-related genes and growth factors, led us to investigate whether there might be differences in lung elastase activity between premature and term lambs, which could contribute to the observed differences of lung elastin abundance and distribution in preterm compared with term lambs. We hypothesized that excessive proteolytic activity in a premature lung might inhibit the normal assembly of elastin needed for alveolar and lung capillary formation. Lung elastase activity increased soon after birth and remained up at 3 days, but then decreased to fetal values after 3 wk of MV (Fig. 9). In similar studies carried out with term lambs, lung elastase activity also increased after 3 days of MV and increased further by 3 wk, such that lung elastase activity was more than 10-fold greater in the chronically ventilated term lambs compared with the chronically ventilated preterm lambs.
MPO Activity in Lungs of Preterm and Term Lambs After 3 Wk of MV
The discovery that serine elastase activity was so much greater in lungs of term compared with preterm lambs after 3 wk of MV prompted us to measure lung MPO activity as an index of pulmonary inflammation in the two groups of chronically ventilated lambs. There were no significant differences in MPO activity between preterm and term lambs after 1 and 3 days of MV (see online data, Fig. 3A), but MPO activity was greater in lungs of term compared with preterm lambs after MV for 3 wk (term = 0.080 ± 0.087 vs. preterm = 0.023 ± 0.004 change in optical density at 450 nm/min, significant difference by nonparametric Mann-Whitney test, P < 0.05).
Structural Differences in Lungs of Chronically Ventilated Preterm vs. Term Lambs
This study confirms previous reports of failed alveolar formation and excess, disordered lung elastin in preterm lambs with CLD. In lambs that were delivered at term, however, prolonged MV with O2-rich gas had little or no apparent effect on lung septation or elastin accumulation, as alveolar number and lung elastin content were similar to measurements made in unventilated lambs that were born at term (2, 6, 55).
There are several possible explanations for these differences in lung structure between chronically ventilated preterm and term newborn sheep. Formation of alveoli, lung capillaries, and the extracellular matrix progresses rapidly during the last month of gestation in sheep (2, 19), and therefore developmental differences in lung structure at the time of birth likely accounted for the greater number of alveoli and microvessels seen in the lungs of lambs that were allowed to remain in utero until term. These structural differences, coupled with surfactant sufficiency and less lung liquid present at term birth, likely contributed to better lung function and adequate respiratory gas exchange using considerably smaller tidal volumes and lower airway pressures during MV of the term compared with the preterm newborns.
Gene Regulation of Alveolar Formation
Previous studies demonstrated defective lung septation and pulmonary emphysema in mice that were rendered deficient in VEGF or PDGF-A, or where receptors for these growth factors were deleted or blocked (10, 11, 23, 30, 37, 40, 48), indicating that both VEGF and PDGF-A have important roles in the formation and maintenance of normal alveolar structure. While lung expression of VEGF, VEGF-R2, and PDGF-A were reduced in both preterm and term lambs after 3 wk of MV, the greater number of alveoli and lung capillaries present at birth in the term lambs would have minimized the structural impact of these postnatal changes in lung expression of growth factors in the animals born at term.
Our finding of decreased expression of VEGF and VEGF-R2 (also called fetal liver kinase-1) in lungs of preterm sheep with CLD is consistent with earlier reports of decreased pulmonary expression of VEGF and one of its receptors, fms-like tyrosine kinase receptor (also called Flt-1, or VEGF-R1), in both premature baboons and human infants with CLD (4, 44). Recent studies showing that treatment with VEGF can overcome the adverse effects of prolonged hyperoxia on alveolar and lung capillary formation in newborn rats underscores the importance of VEGF in regulating lung septation and angiogenesis during normal development and during repair of neonatal lung injury (35, 61).
Our finding of reduced PDGF-A expression in lungs of preterm lambs with CLD contrasts with previous reports that lung expression of PDGF-A mRNA was unaffected by exposure of newborn rats to 14 days of either 60% or 85% O2, which induces a form of lung injury that, in some respects, resembles the pathology seen in CLD (14, 25, 26). It is noteworthy, however, that lung mRNA expression of PDGF-A in our preterm lambs remained unchanged during the first 3 days of MV, with a subsequent decrease below fetal control levels after 3 wk of MV. Moreover, it is possible that reduced lung expression of PDGF-A mRNA and protein in CLD is related more to mechanical distention associated with prolonged cyclic stretch of the lung, rather than hyperoxia, as most of our lambs breathed considerably less than 60% O2 during the course of these studies. The fact that PDGF-A-deficient mice that survive the newborn period acquire pulmonary emphysema from failed alveolar formation clearly attests to the importance of PDGF-A in postnatal lung development.
Because measurements of lung gene and protein expression were made in whole lung, rather than in specific cell types, it is not possible to define either the magnitude or the loci of expression of the reported changes at the sites of septation, where presumably the changes may have been greater than those reported. Moreover, as lung abundance of growth factors was assessed at only three postnatal time points, we cannot define the precise timing of changes in lung expression of these proteins during the 3-wk course of these studies. It is fair to assume, however, that the measured changes in lung abundance of VEGF-A and PDGF-A, and the relevant receptors, VEGF-R2 and PDGF-Rα, occurred at a critical stage of alveolar septation, based on the fivefold increase in alveolar number observed in sheep during the last 3 wk of gestation (2).
Gene Regulation of Elastin Synthesis and Assembly in Preterm Lambs With CLD
Elastin is known to play a critical role in alveolarization and angiogenesis during lung development. Elastin-null mice die soon after birth from cardiorespiratory failure associated with smooth muscle overgrowth in pulmonary arteries, defective airway branching, and lack of lung septation to form alveoli (38, 69). Mutations in several genes that are known to affect elastin synthesis and assembly, such as fibulin-5, fibrillin-1, lysyl oxidase, and lysyl oxidase-like 1, result in offspring that exhibit abnormal lung development, in some cases leading to pulmonary pathology that resembles emphysema in those mice that survive beyond the perinatal period (29, 41, 43, 52, 70).
We, therefore, tested the hypothesis that MV with O2-rich gas, when applied to the lungs of preterm lambs with evolving CLD, would induce changes in lung expression of genes that affect elastin synthesis and assembly, which might account for the abnormal accumulation of lung elastin and perhaps contribute to the impaired lung septation seen in CLD. Because term lambs that received MV for 3 wk did not exhibit these structural abnormalities, we expected to see different patterns of expression of elastin-related genes in their lungs, which were exposed to lower inflation pressures and tidal volumes than the lungs of the chronically ventilated preterm lambs. It was not surprising that lung expression of tropoelastin mRNA was significantly greater in preterm compared with term lambs during these studies, with maximal expression observed after 1 day of MV. The discordant expression patterns of genes that impact elastin assembly in lungs of preterm lambs, however, was unexpected. The observation that lung mRNA and protein abundance of fibulin-5 and lysyl oxidase did not increase in preterm lambs after 3 wk of MV, as lung tropoelastin expression and elastin fiber density increased, offers a plausible explanation for the abnormal abundance and distribution of elastin in the lungs of these animals. Deficiencies of either fibulin-5 or lysyl oxidase result in abnormal elastin assembly, which is associated with defective lung septation and emphysema in mutant mice (29, 41, 43, 52, 70).
Early postnatal increases of TGF-α and TGF-β1 gene expression in lungs of preterm, but not term, lambs also is consistent with the differences in lung structure and elastin accumulation seen after 3 wk of MV in the two groups of animals. Previous studies showed that mutant mice that overexpressed TGF-α acquired a lung phenotype that included aberrant alveolarization and disorganized elastic fibers, analogous to the pathology seen in pulmonary emphysema, and similar to the lung remodeling that occurs in BPD (36). Persistently increased lung expression of TGF-α also was reported in association with O2-induced lung inflammation and subsequent pulmonary fibrosis in newborn rabbits (67). TGF-β1 has been shown to increase tropoelastin mRNA and soluble elastin protein content in cultured neonatal lung fibroblasts (47). Previous reports that overexpression of TGF-α and TGF-β1 in newborn rodents yields lung pathology similar to the pathology observed in BPD (22, 36, 66) support the notion that increased lung expression of these growth factors likely contributed to the abnormal alveolar formation and excess, disordered elastin seen in our preterm lambs with CLD. The observation that after 1 day of MV both TGF-α and TGF-β1 protein were highly expressed throughout the walls of distal air spaces, where elastin accumulation was prominent in the preterm lambs with CLD, supports a role for these mitogens in modifying elastin assembly during the course of MV of the lung during development.
Reduced lung expression of PDGF-A and its receptor, PDGF-Rα, may help to explain the abnormal distribution of elastin observed in the lungs of our preterm lambs with CLD. As noted previously, PDGF-A-deficient mice exhibit defective alveolar formation, a finding that was attributed to lack of distal migration of lung myofibroblasts, cells that express PDGF-Rα and which make and secrete elastin (10, 11, 40). The paucity of PDGF-A and PDGF-Rα in the lungs of preterm lambs with CLD is consistent with impaired migration of pulmonary interstitial myofibroblasts and abnormal distribution of lung elastin.
Serine Elastase Activity in Lungs of Chronically Ventilated Preterm and Term Lambs
Several studies have shown evidence of inflammation and increased proteolytic activity in the lungs of infants with evolving BPD (21, 49, 54, 57, 68), as well as in authentic animal models of this disease (3, 17, 60, 71). We were not surprised to discover that serine elastase activity in lung increased abruptly after the onset of MV in both preterm and term lambs. It was unexpected, however, that elastase activity would continue to increase in term lambs, but decrease in preterm lambs, such that elastase activity was several-fold greater in term than in preterm lambs after 3 wk of MV. This finding likely reflects greater inflammation in the lungs of the chronically ventilated term lambs compared with the preterm lambs, as lung MPO activity was significantly greater in the more mature animals at the end of the 3-wk studies. The reason for this apparent difference in lung inflammation is unclear, as the term lambs did not have evidence of systemic or pulmonary infection, as assessed by hourly measurements of temperature, pulse and respiratory rates, daily measurements of circulating white blood cell counts, and terminal studies of pulmonary mechanics and cardiovascular function. The associated increase in lung elastase activity after 3 wk of MV in the term lambs could have contributed to the apparent absence of newly synthesized elastin in the more mature animals. These observed differences present the intriguing possibility that early postnatal stimulation of lung proteolytic activity may reflect an inflammatory process, perhaps associated with release of elastase from lung parenchymal cells, leading to breakdown of matrix proteins, whereas later induction of lung proteolytic activity may facilitate tissue remodeling to help sustain important structural components of the lung that formed during fetal life.
Working Model of Dysregulated Synthesis and Assembly of Lung Elastin in Neonatal CLD
Figure 10 depicts our current notion of how elastic fiber synthesis and assembly is thought to occur at the cell surface of lung myofibroblasts, and the putative impact that our reported lung changes (defined in red) of gene and protein expression and elastase activity might have on elastic fiber formation in preterm lambs with CLD. Tropoelastin is secreted from the cell and deposited onto microfibril scaffolds that attach to the cell surface through the interaction of fibrillins, microfibril-associated glycoproteins, and integrin receptors located on the cell membrane. Fibulin-5 serves as a bridging molecule for tropoelastin, helping to tether it to the cell surface by interacting with matrix proteins and integrins located on the plasma membrane. Lysyl oxidase-like 1, which is secreted by fibroblasts and smooth muscle cells, also helps to bind tropoelastin and matrix proteins to the cell surface. In addition, it has an important role in elastic fiber homeostasis, converting tropoelastin into a lysyl-deaminated form that allows for covalent cross-linking of tropoelastin to yield mature elastic fibers (18, 41). Lysyl oxidase, an extracellular copper-dependent enzyme that is secreted by fibrogenic cells, oxidizes lysine residues in both elastin and collagen, and thereby initiates the formation of covalent cross-linkages that stabilize these fibrous proteins, which provide resilience and extensibility to the lung's respiratory units and blood vessels (33, 43).
In our model, increased pulmonary expression of TGF-α and TGF-β1 induce tropoelastin secretion in lung myofibroblasts, which are located within septal crests. Pulmonary expression of PDGF-A, which is thought to guide migration of PDGF-Rα-expressing myofibroblasts to the septal crests and thereby facilitate alveolar formation, is diminished after prolonged MV, which could contribute to failed lung septation in CLD. Increased lung expression of tropoelastin persists throughout the period of MV, without a corresponding increase of fibulin-5 or lysyl oxidase. As molecular interactions of tropoelastin and several extracellular matrix proteins, including fibulin-5, lysyl oxidase, and lysyl oxidase-like 1, are crucial for normal elastic fiber formation in the lung (34, 41), discordant pulmonary expression of these closely linked proteins can lead to defective elastin assembly. These findings, coupled with increased proteolytic activity associated with lung inflammation during the initial days of MV, may lead to increased production of poorly assembled elastic fibers that contribute to the abnormal lung structure and function observed in neonatal CLD. Validation of this working model awaits further inquiry to determine how mechanical forces applied to the lung, with or without moderate hyperoxia, uncouple the finely orchestrated programs of myofibroblast migration and of elastin synthesis and assembly that are necessary for normal alveolar septation and lung development.
This work was supported by National Heart, Lung, and Blood Institute Grants HL-62512 (to R. D. Bland) and HL-56401, Specialized Center of Research Project V (to R. D. Bland), and by funding from the Vera Moulton Wall Cardiopulmonary Research Center at Stanford University.
This work would not have been possible without the help of numerous people who are not listed as authors. We are especially grateful for the technical assistance of Melinda Williams and MarJanna Dahl, and for the steadfast efforts of the many medical students, part-time technicians, and respiratory therapists who assisted in the daily management of the lambs at the University of Utah and State University of New York at Buffalo. We also thank Dr. Edmund Egan (Ony) for generously providing bovine surfactant (Infasurf), and Dr. Robert Mecham at Washington University in St Louis for giving us the fibulin-5 antibody.
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2007 the American Physiological Society