Mechanical ventilation with 40% oxygen reduces pulmonary expression of genes that regulate lung development and impairs alveolar septation in newborn mice

Richard D. Bland, Lucia M. Mokres, Robert Ertsey, Berit E. Jacobson, Shu Jiang, Marlene Rabinovitch, Liwen Xu, Eric S. Shinwell, Feijie Zhang, Matthew A. Beasley

Abstract

Mechanical ventilation with 40% oxygen reduces pulmonary expression of genes that regulate lung development and impairs alveolar septation in newborn mice. Am J Physiol Lung Cell Mol Physiol 293: , 2007. First published August 17, 2007; — Mechanical ventilation (MV) with O2-rich gas offers life-saving treatment for extremely premature infants with respiratory failure but often leads to neonatal chronic lung disease (CLD), characterized by defective formation of alveoli and blood vessels in the developing lung. We discovered that MV of 2- to 4-day-old mice with 40% O2 for 8 h, compared with unventilated control pups, reduced lung expression of genes that regulate lung septation and angiogenesis (VEGF-A and its receptor, VEGF-R2; PDGF-A; and tenascin-C). MV with air for 8 h yielded similar results for PDGF-A and tenascin-C but did not alter lung mRNA expression of VEGF or VEGF-R2. MV of 4- to 6-day-old mice with 40% O2 for 24 h reduced lung protein abundance of VEGF-A, VEGF-R2, PDGF-A, and tenascin-C and resulted in lung structural abnormalities consistent with evolving CLD. After MV with 40% O2 for 24 h, lung volume was similar to unventilated controls, whereas distal air space size, assessed morphometrically, was greater in lungs of ventilated pups, indicative of impaired septation. Immunostaining for vimentin, which is expressed in myofibroblasts, was reduced in distal lung after 24 h of MV with 40% O2. These molecular, cellular, and structural changes occurred without detectable lung inflammation as evaluated by histology and assays for proinflammatory cytokines, myeloperoxidase activity, and water content in lung. Thus lengthy MV of newborn mice with O2-rich gas reduces lung expression of genes and proteins that are critical for normal lung growth and development. These changes yielded lung structural defects similar to those observed in evolving CLD.

  • bronchopulmonary dysplasia
  • neonatal chronic lung disease
  • lung growth and development
  • vascular endothelial growth factor
  • VEGF receptor 2 (fetal liver kinase-1)
  • platelet-derived growth factor
  • tenascin-C
  • pulmonary inflammation

infants who are born very prematurely often exhibit respiratory failure because of their incompletely formed lungs, primitive respiratory drive, and susceptibility to infection. Life-saving treatment of such infants using assisted ventilation with O2-rich gas frequently leads to a chronic form of lung disease that was first described 40 yr ago as bronchopulmonary dysplasia (BPD; Ref. 36). The clinical and pathological features of this condition have changed in recent years as a result of major advances in perinatal care, including widespread use of antenatal glucocorticoid therapy, postnatal surfactant replacement, and improved respiratory and nutritional support. These changes in newborn intensive care have increased survival of the tiniest premature infants whose very immature lungs are susceptible to injury from lengthy mechanical ventilation (MV) with O2-rich gas. Thus the incidence of BPD and its associated morbidity remain high, reflecting improved survival of infants whose incompletely developed lungs are at greatest risk for injury from prolonged cyclic stretch and exposure to increased O2. Sometimes called the “new BPD,” this form of neonatal chronic lung disease (CLD) is the leading cause of lengthy hospitalization and recurrent respiratory illness in infants who have been born at <28 wk of gestation (6, 23).

Authentic animal models of this disease, featuring prolonged MV of surfactant-treated, premature newborn baboons and lambs, have provided important insights regarding the pathophysiology and treatment of CLD (4). Lung histopathology after 2–4 wk of positive-pressure ventilation with O2-rich gas includes failed formation of alveoli and lung capillaries, excess disordered accumulation of elastin, overgrowth of vascular and airway smooth muscle, chronic inflammation, and interstitial edema (1, 5, 13, 37). These abnormal histological findings are similar to those that have been described in very premature infants with CLD (11, 18, 19, 32, 44). The challenge now is to improve understanding of the molecular mechanisms that regulate normal lung growth and development and to clarify how prolonged cyclic stretch with O2-rich gas disrupts lung structure and function during development and in response to injury.

To this end, we investigated the effects of MV with either air or 40% O2 on pulmonary expression of genes that regulate formation of alveoli and blood vessels in lungs of newborn mice, in which alveolarization and angiogenesis occur mainly after birth at term gestation (2). In the initial series of studies, positive-pressure MV of 2- to 4-day-old mice for 8 h with 40% O2 reduced lung mRNA expression of VEGF-A and one of its receptors, VEGF-R2, in addition to PDGF-A and tenascin-C (TN-C), all of which are genes known to affect formation of alveoli and blood vessels in the developing lung. MV for 8 h with air reduced lung mRNA expression of PDGF-A and TN-C but did not cause a significant decrease in lung mRNA for VEGF-A or VEGF-R2. In a subsequent series of studies, MV of 4- to 6-day-old mice for 24 h with 40% O2, at either 60 or 180 breaths/min (bpm), reduced lung content of VEGF-A, VEGF-R2, PDGF-A, and TN-C proteins and yielded lung structural changes indicative of diminished septation compared with the lungs of unventilated control pups that spontaneously breathed 40% O2 for 24 h. These molecular and structural changes occurred without apparent lung inflammation as assessed by histology and measurements of proinflammatory cytokines, myeloperoxidase (MPO) activity, and lung water content.

METHODS

Animal Experiments

This investigation included two sets of studies, both of which used healthy BALB/c mice that were 2–6 days old and weighed ∼2–5 g after spontaneous vaginal birth at term gestation. The first series of studies focused on lung expression of genes known to affect alveolar formation in 2- to 4-day-old mouse pups that received positive-pressure MV for 8 h at either 60 or 180 bpm with either 40% O2 or air compared with unventilated control pups that spontaneously breathed either 40% O2 or air, respectively, for 8 h. These studies were designed to test the hypothesis that MV during a critical stage of lung development would adversely affect expression of genes known to be important in regulating formation of alveoli and lung capillaries. We expected that the effects would be most apparent when the lungs were exposed to cyclic stretch with O2-rich gas, more so than with air. The second series of studies focused on lung abundance of relevant proteins and changes in lung structure after MV at either 60 or 180 bpm with 40% O2 for 24 h compared with unventilated controls that spontaneously breathed 40% O2 for 24 h. These studies were designed to test the hypothesis that prolonged cyclic stretch of the developing lung with O2-rich gas would alter lung content of specific proteins that affect lung growth and development (VEGF-A, VEGF-R2, PDGF-A, and TN-C) and induce structural changes in lung septation as seen in CLD. All surgical and animal care procedures and experimental protocols were reviewed and approved by the Stanford University Institutional Animal Care and Use Committee.

Eight-hour studies.

We studied six groups of mice that were 2–4 days old and weighed 1.5–3.5 g (average 2.3 ± 0.4 g). Four groups of pups (n = 8–10/group) underwent a sterile tracheostomy after intramuscular anesthesia with ketamine (∼75 μg/g body wt) and xylazine (∼15 μg/g body wt) in preparation for MV. Two groups of control pups (n = 8–10/group), which came from the same litters as those on MV, also received ketamine-xylazine anesthesia before undergoing a small, superficial neck incision in preparation for breathing either air or 40% O2 without MV. Pups that were randomly selected to receive MV had a midline neck incision done under a dissecting microscope followed by a tracheal slit and insertion of a custom-made polyvinyl catheter [intratracheal segment: inner diameter (ID), 0.28 mm; outer diameter (OD), 0.61 mm; length, 1–2 mm; Tygon, Akron, OH] that was securely sutured in the trachea. Pups received positive-pressure MV for 8 h with either air or 40% O2 delivered at a rate of either 60 bpm using a time-cycled, pressure-limited, small animal respirator (model 2000; Mallard Medical, Redding, CA) or at a rate of 180 bpm using a custom-designed, piston-type, small animal respirator (MicroVent Model 848 Ventilator; Hugo Sachs Elektronik/Harvard Apparatus, Holliston, MA). We used an inspired O2 concentration (FIO2) of 0.4 based on our (5) earlier observation that preterm lambs with CLD required, on average, ∼40% O2 to maintain a normal partial pressure of O2 in arterial blood. The rationale for using one rate that was slower and one that was faster than the normal breathing rate of newborn mice (∼110 bpm; Ref. 34) was to compare effects on the lung of different respiratory patterns featuring different tidal volumes and inflation pressures. The ventilation circuit was configured as a loop of tubing with a cut-off 20-gauge Luer stub glued into the loop and connected to the tracheostomy, yielding a dead space volume of <5 μl.

FIO2 was verified as either 0.21 or 0.40 at the start of each experiment with a portable, calibrated O2 analyzer (MAXO2 OM-25AE; Maxtec, Salt Lake City, UT). Peak inspiratory pressure (PIP) and mean airway pressure (MAP) were measured continuously from an in-line pressure monitor (Bunnell, Salt Lake City, UT) that was connected to the ventilation circuit of the pressure-limited respirator. When using the MicroVent volume respirator, PIP and MAP were obtained from pressure measurements made at the start of each experiment with the ventilation circuit and in-line tracheostomy tube connected to the pressure monitor at the specific volume settings used in the study. Because of fluctuations in delivered gas volume when the MicroVent circuit was connected to the pressure monitor, we did not obtain continuous measurements of airway pressures but relied instead on specific tidal volume settings and observations of chest wall and abdominal excursions associated with lung inflations.

For all ventilation studies, the pups rested supine with their limbs gently taped to an underlying gauze pad, which rested on an electric warming blanket, with an overhead lamp that kept ambient temperature at ∼30–33°C. The mice were covered with a sheet of clear plastic wrap to reduce heat and water losses through the skin during MV. They were observed continuously, with close monitoring of chest movement, skin color, temperature, and response to tactile stimulation assessed throughout the period of MV. In many studies, heart rate (HR) and systolic blood pressure (BPs) were monitored at frequent intervals using a noninvasive tail-cuff microsensor device as described below. Pups received additional intramuscular doses of a solution containing ketamine, 10 μg/g body wt, and xylazine, 2 μg/g body wt, as needed for agitation, which was detected by body movement or an exaggerated response to tactile stimulation. Control pups were unrestrained within a warm plastic chamber and breathed either air or 40% O2 that flowed into the chamber. Control and ventilated pups appeared to have milk in their stomach from the start and therefore were not fed during the 8-h studies. At the end of each study, we injected pentobarbital, 100 μg/g body wt, before opening the chest to harvest the lungs or before intratracheal instillation of fixative.

In preparation for these studies, we obtained estimates of delivered tidal volume from each respirator by connecting the ventilation circuit and in-line tracheostomy tube to a low-resistance water-filled catheter that was attached to the barrel of a 100-μl Hamilton syringe. This enabled us to determine the approximate delivered tidal volume by measurement of water displacement resulting from gas entry into the ventilation circuit from each respirator within the range of peak-pressure settings and tidal volume settings that were used in these studies. For the MicroVent volume ventilators that we used at a rate of 180 bpm, delivered tidal volume averaged ∼40% of the tidal volume that the unit was set to deliver.

We performed several studies to determine pH and PCO2 values in blood aspirated percutaneously from the heart after 8 h of MV. After giving pentobarbital anesthesia, we withdrew blood (0.15–0.25 ml) from the heart through a 25-gauge needle attached to a heparinized 1-ml plastic syringe. The blood was immediately transferred to a portable analyzer (i-STAT; Heska, Fort Collins, CO) for measurement of pH and PCO2. This helped to verify that ventilator settings (PIP and tidal volume) were appropriate for maintaining respiratory gas exchange within the physiological range. Some measurements were made on heart blood of pups as small as 4 g and as young as 4 days old; some were made on heart blood from larger pups that were 7–14 days old during MV, with airway pressures and tidal volume/kg body wt similar to those used in our 2- to 4-day-old pups.

In several studies of both mechanically ventilated and unventilated, spontaneously breathing pups, we measured BPs and HR at frequent intervals using a custom-designed tail-cuff microsensor connected to a rodent blood pressure monitor (MK-2000A; Muromachi Kikai, Tokyo, Japan). This device has been used to measure BPs and HR in adult mice with bleomycin-induced lung fibrosis (35) and in young, anesthetized mice on MV (43), but there are no reports of its previous application to measure BPs and HR in newborn mice.

Twenty-four-hour studies.

We studied three groups of mice that were 4 days old and weighed 2.5–4.5 g (average 3.3 ± 0.5 g) and two groups of mice that were 6 days old and weighed 3.2–5.6 g (average 4.5 ± 0.6 g). Two groups of 4-day-old pups (n = 8–10/group) and two groups of 6-day-old pups (n = 6–7/group) had a sterile tracheostomy after intramuscular anesthesia with ketamine (∼75 μg/g body wt) and xylazine (∼15 μg/g body wt) in preparation for MV with 40% O2 at either 60 or 180 bpm for 24 h. The pups also had a polyvinyl catheter (0.28 mm ID, 0.61 mm OD, 5 cm long; Tygon) inserted gently through the mouth into the stomach for subsequent feedings. Additional intramuscular doses of ketamine (10–20 μg/g body wt) and xylazine (2–4 μg/g body wt) were given, as needed, to prevent discomfort and associated agitation. During MV, pups received feedings of a rodent milk replacer (KMR; PetAg, Hampshire, IL), ∼30–40 μl every 2–3 h, designed to provide a daily fluid intake of ∼100–120 μl/g body wt with a caloric intake of ∼70–80 kcal/g body wt. Care of the mice during MV and ventilator settings used in these 24-h studies were similar to those used during the 8-h studies. Control pups (n = 12) received intramuscular ketamine-xylazine in preparation for a sham neck incision, after which they spontaneously breathed 40% O2 for 24 h in a warm plastic chamber except for a 4-h interval during the second 12 h in which they were kept with the dam in a larger plastic chamber with continuous flow of 40% O2, thereby enabling them to feed without a change in FIO2. All pups received antibiotics, 200 μg/g body wt of intramuscular ampicillin and 4 μg/g body wt of intramuscular gentamicin, to reduce the risk of infection.

Postmortem Studies

RNA extraction and quantitative RT-PCR.

At the end of several 8-h studies (n = 6/group), lungs were excised, rapidly frozen in liquid N2, and stored at −80°C for subsequent two-step mRNA extraction using TRIzol reagent (Invitrogen, Carlsbad, CA) and purification with RNeasy Mini Kit columns (Qiagen, Valencia, CA) as previously described (7). Lung mRNA expression of genes linked to lung development (VEGF-A, VEGF-R2, PDGF-A, TN-C) and inflammation (TNF-α, IL-1β, and IL-6) was assessed by quantitative RT-PCR (qRT-PCR) using commercially available primers and probes (TaqMan Gene Expression Assays; Applied Biosystems, Foster City, CA). Ribosomal RNA (18S) was used as an internal control. Additional details of the methods used for qRT-PCR are described in the information supplement online at the AJP-Lung Cellular and Molecular Physiology web site.

Protein extraction and Western immunoblots.

At the end of several 24-h studies (n = 4–8/group), lungs were excised, rapidly frozen in liquid N2, and stored at −80°C for later measurement of VEGF-A, VEGF-R2, PDGF-A, and TN-C protein content by Western blot analysis as previously described (7). Specific details of the immunoblot methods used for measuring lung content of the above proteins are included in the online supplement.

Microarray-based cytokine multiplex analysis for proinflammatory mediators.

Protein extracts from samples of frozen lung were obtained as described above, and supernatants were used to measure total protein by the bicinchoninic acid assay, and proinflammatory mediators (TNF-α, IL-1β, and IL-6) by a microarray-based cytokine multiplex assay (Mouse Cytokine Microarray Kit 2.2 Fluorescent, cat. no. PA4104; Allied Biotech, Ijamsville, MD). Details of this method are described in the online supplement.

Immunohistochemical localization of specific proteins in lung.

Standard immunohistochemical (IHC) techniques were applied for localization of PDGF-A and TN-C proteins in 5-μm sections of lung from 4-day-old mice that received MV with 40% O2 for 24 h compared with controls. We also used IHC to define the distribution of vimentin, which is expressed in myofibroblasts (MFBs), in sections of lung obtained from these mice. These methods were reported previously (7) and are briefly summarized in the online supplement.

Lung water content.

At the end of several 8-h studies (n = 4–8/group), lungs were excised, blotted to remove surface fluid, weighed, dried for 72 h in a vacuum oven at 65°C, and reweighed to determine lung water content as (wet lung weight minus dry lung weight)/dry lung weight.

MPO activity.

At the end of several 8-h studies (n = 3/group), lung tissue was frozen in liquid N2 and stored at −80°C for later assay of MPO activity by a modification of a previously described method (41). Details of this method are included in the online supplement.

Processing of lungs for quantitative histology.

To obtain lungs for histopathology at the end of the 24-h studies, 4-day-old pups were euthanized with pentobarbital, and the diaphragm was punctured via the abdomen to allow the lungs to expand as they were filled with fixative via the tracheostomy tube. For unventilated control pups, a tracheostomy and diaphragm puncture were performed after death so that the lungs could be filled with fixative. Carnoy's solution was instilled to fix the lungs for morphometric measurements, and PBS-buffered 4% paraformaldehyde (PFA) solution was used to fix the lungs for IHC. After the lungs were filled with fixative, the tracheostomy tube was connected through a plastic catheter to a column of freshly prepared fixative solution at a pressure of 20 cmH2O, which was maintained for 30 min at room temperature for fixation with Carnoy's solution or overnight at 4°C for fixation with 4% PFA. The trachea was ligated to prevent leakage of fixative, and the chest was opened to excise the lungs and measure their volume by fluid displacement (40). The lungs then were transferred to a vial containing 70% ethanol for brief storage before tissue embedding and sectioning.

Fixed lungs were embedded in paraffin for morphometry and IHC with right and left lungs embedded separately. Isotropic uniform random (IUR) orientation was achieved by cutting the paraffin blocks at random angles in two perpendicular planes. Random angles were selected using a random number chart and applied using an orientator (20). The IUR paraffin blocks were re-embedded onto mounting blocks before sectioning. A random number generator was used to determine a number from 1 to 100, defining how many micrometers from the leading edge of lung tissue serial sectioning would begin. A series of six 5-μm sections was generated every 200 μm, and the first consecutive pair of each series of sections was stained with hematoxylin and eosin for histology and morphometry.

Lung morphometry to measure dimensions of distal air spaces.

Three stained tissue sections from both the right and left lung series taken from each mouse were randomly selected using a random number table. Using a Nikon Eclipse E800 microscope and Bioquant Osteo II image analysis software (cat. no. BQ-OST II, Bioquant, Nashville, TN), the diameter, perimeter, and area of all complete distal air spaces (nonbranching, terminal air spaces with complete circumferential walls) were measured using a 10× calibrated objective. Diameter was determined by measuring the shortest distance between the walls of the air space, thereby avoiding measurement of the longitudinal axis of terminal airways. We did not include regions of atelectasis, where air space walls could not be delineated. Measured air spaces were marked so that they would not be measured again in overlapping fields of view. Data for air space diameter, perimeter, and area from right and left lungs of each mouse were combined, and their respective mean and SD values were calculated for comparison between experimental groups. Analyses were done without knowledge of the group from which the lung section was derived.

Statistical Analysis of Data

Data in the text, table, and figures are expressed as mean ± SD unless denoted otherwise. For comparison of 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 control and mechanically ventilated groups of mice. For data sets that had a skewed, non-Gaussian distribution, we used the nonparametric Mann-Whitney test to assess for 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 unventilated control pups that breathed either air or 40% O2 and mechanically ventilated pups that breathed either air or 40% O2 at either 60 or 180 bpm (46). Statistical analysis was done using the Prism 4 software package (GraphPad, San Diego, CA). Differences were considered significant if the P was <0.05.

RESULTS

Physiological Data for 8-h Studies

Table 1 shows summary data for the groups of newborn mice that received MV for 8 h with air or 40% O2 at either 60 or 180 bpm. As expected, PIP and tidal volume were greater in pups that received MV at 60 bpm than in pups that had MV at 180 bpm. End-expiratory pressure was maintained at a level close to atmospheric pressure after pilot studies showed rapid onset of cyanosis, hepatomegaly, and death from circulatory failure when the end-expiratory pressure was raised above 1 cmH2O, which we attributed to impaired venous return coupled with a miniscule stroke volume at a HR of >500 beats/min. MAP did not differ between the groups of ventilated pups as the inspiratory-to-expiratory (I:E) ratio was greater with the volume ventilator (I:E = 1:1.2) used to deliver 180 bpm than it was for the time-cycled ventilator (I:E = 1:2.3) used to deliver 60 bpm. Airway pressures and tidal volumes were similar in mice that received MV with air compared with 40% O2 (data not shown).

View this table:
Table 1.

Respiratory data for newborn mice that received MV for 8 h at either 60 or 180 bpm

BPs and HR were measured during MV for 8 h in 12 newborn mice that were 3–4 days old and weighed 3.0 ± 1.0 g. BPs averaged 41 ± 13 mmHg, and HR averaged 523 ± 47 beats/min, with no significant differences between animals that received air compared with 40% O2 or between those that received MV at 60 bpm compared with 180 bpm. In 7 unventilated control pups that were 3–4 days old and weighed 2.8 ± 0.9 g, BPs averaged 47 ± 5 mmHg, and HR averaged 521 ± 42 beats/min. These averages derive from 100 measurements made on 12 mice that were on MV and 39 measurements made on 6 unventilated controls. Adverse hemodynamic effects of MV and anesthesia likely accounted for the small reduction in BPs observed in pups on MV.

Terminal measurements of pH and PCO2 on heart blood were not possible in these 2- to 4-day-old mice owing to their very small heart size. We, however, obtained heart blood at the end of 24 8-h studies of MV done with 35 mice that were 5–14 days old and weighed 6.4 ± 1.9 g. Ten of these pups weighed <5 g. Fourteen pups had MV at 60 bpm, and 21 at 180 bpm; 14 breathed air, and 21 breathed 40% O2 during MV. PIP and estimated tidal volume/g body wt were similar to settings that were applied in the studies of smaller newborn pups. Blood pH averaged 7.30 ± 0.12, and PCO2 averaged 37 ± 11 mmHg, indicative of a mild metabolic acidosis, presumably from positive-pressure breathing and anesthesia, compounded by the terminal barbiturate injection. There were no significant differences in pH or PCO2 between pups that received air, compared with 40% O2, nor were there differences related to the applied respirator rate.

mRNA Expression of Genes that Regulate Lung Septation

Lung mRNA expression of VEGF-A and VEGF-R2 decreased significantly, compared with controls, after MV for 8 h with 40% O2 at either 60 or 180 bpm but not after MV with air at either respirator rate (Fig. 1, A and B). Lung expression of PDGF-A mRNA decreased, compared with controls, after MV for 8 h with either 40% O2 or air at 60 or 180 bpm (Fig. 1C). MV at a rate of 180 bpm with either 40% O2 or air was associated with a significant decrease in TN-C mRNA (Fig. 2D). There was a similar trend in TN-C mRNA expression after 8 h of MV with either air or 40% O2 at 60 bpm, but these differences were not statistically significant. There were no significant differences in lung mRNA expression of VEGF-A, VEGF-R2, PDGF-A, or TN-C between unventilated control groups that breathed either 40% O2 or air for 8 h (Fig. 1).

Fig. 1.

AD: lung mRNA abundance, expressed relative to 18S rRNA, for genes that are known to regulate alveolar septation: VEGF-A (A), VEGF-R2 (B), PDGF-A (C), and tenascin-C (TN-C; D). All data are means ± SD for n = 6 pups/group at the end of 8 h; 40% O2-breathing groups on left, air-breathing groups on right. *Significant difference compared with relevant control group, P < 0.05. MV, mechanical ventilation (in breaths/min).

Fig. 2.

Lung histology in newborn mice after breathing 40% O2 (top) or air (bottom) either without MV (control) or with MV at 60 or 180 breaths/min for 8 h is shown. The lungs were fixed intratracheally with Carnoy's solution at 20 cmH2O. Magnification was ×100. There was no evidence of inflammation or edema after MV with air or 40% O2 at either respiratory rate.

Lung Histology and Water Content After MV for 8 h

We examined sections of lung obtained from 3–4 newborn mice in each of the six groups that were studied for 8 h. There was no evidence of lung inflammation or edema after MV at either 60 or 180 bpm with either 40% O2 or air (Fig. 2). Measurements of lung water content showed no significant difference between control mice that spontaneously breathed either air (4.42 ± 0.20, n = 6) or 40% O2 (4.21 ± 0.19, n = 7) compared with mice that received MV with either air or 40% O2 (4.40 ± 0.33 g/g dry lung, n = 20). These results, confirming the absence of pulmonary edema after 8 h of MV at either 60 or 180 bpm or after 8 h of mild hyperoxia, indicate that the modest inflation pressures and FIO2 used in these studies did not cause detectable lung injury.

Effects of MV with 40% O2 for 24 h

We extended the duration of studies to 24 h to determine the effects of prolonged cyclic stretch with O2-rich gas on lung content of the proteins for which we measured mRNA levels at 8 h. Lung content of VEGF-A, VEGF-R2, PDGF-A, and TN-C proteins, expressed relative to β-actin, decreased after 24 h of MV with 40% O2 at both 60 and 180 bpm compared with unventilated control pups that breathed 40% O2 for 24 h (Fig. 3, AD). Immunostaining of lung sections for PDGF-A and TN-C confirmed the immunoblot results, showing reduced expression of both these proteins in the walls of distal air spaces after MV for 24 h with 40% O2 compared with controls that breathed 40% O2 for 24 h (Fig. 4, A and B).

Fig. 3.

AD: protein abundance of VEGF-A (A), VEGF-R2 (B), PDGF-A (C), and TN-C (D), measured by immunoblot analysis and expressed relative to β-actin protein, in lungs of 4- to 6-day-old mice that breathed 40% O2 for 24 h either without MV (control) or with MV at 60 or 180 breaths/min. All data are means ± SD; n = 4–7 pups/group. *Significant difference compared with the relevant control group, P < 0.05; †significant difference compared with the relevant control group, P = 0.05.

Fig. 4.

A and B: immunostaining for PDGF-A (A) and TN-C (B) in lungs of newborn mice that breathed 40% O2 for 24 h either without MV (control; left) or with MV at 180 breaths/min (right). Arrows point to brown immunostaining, showing reduced PDGF-A protein in distal lung parenchyma after lengthy MV with O2-rich gas, with similar staining of vascular and airway smooth muscle in control and ventilated lungs (A), and reduced TN-C protein in distal lung cells after prolonged MV with 40% O2 (B). Magnification was ×400.

The second aim of this series of studies was to examine the effects of lengthy MV with O2-rich gas on structural features of the lung during development. Figure 5 shows representative sections of lung taken from 5-day-old mouse pups that had breathed 40% O2 for 24 h either without MV (controls) or with MV at 60 or 180 bpm. Although total lung volume (LV) was similar in the three groups of animals (no MV, LV = 163 ± 22 μl; MV at 60 bpm, LV = 168 ± 14 μl; MV at 180 bpm, LV = 162 ± 26 μl), there appeared to be fewer septa and larger air spaces after MV at 60 or 180 bpm, compared with unventilated controls, with little or no evidence of inflammation or injury associated with MV. Quantitative image analysis of air space size showed a >2-fold increase in the average cross-sectional area of distal respiratory units in lungs of ventilated compared with unventilated control pups that breathed 40% O2 for 24 h (Fig. 6). This finding, coupled with similar LV for ventilated and control mice, suggests that prolonged MV with O2-rich gas resulted in less lung septation, thereby yielding fewer air spaces than in the lungs of control pups that breathed 40% O2 without MV. As the 4-day-old mice on which we assessed lung structure were just emerging from the saccular stage of lung development, they exhibited scant alveolar formation, even among control pups. Hence, alveolar counts were not done.

Fig. 5.

Lung histology in newborn mice after breathing 40% O2 for 24 h either without MV (control) or with MV at 60 or 180 breaths/min (bpm) is shown. The lungs were fixed intratracheally with Carnoy's solution at 20 cmH2O. Magnification was ×100 for top, ×400 for bottom. Arrows point to alveolar septa. Lung volume was similar for the 3 groups: control, 163 ± 22 μl; MV @ 60 bpm, 168 ± 14 μl; MV @ 180 bpm, 162 ± 26 μl.

Fig. 6.

Increased air space perimeter and area after MV with 40% O2 for 24 h at both 60 and 180 breaths/min are shown. Air space size was not significantly different between the 2 groups of pups that received MV at the different respirator rates. All data are means ± SD; n = 4–5 pups/group. *Significant difference compared with control group, P < 0.05.

The apparent reduction in lung septation after prolonged MV, coupled with reduced lung expression of PDGF-A mRNA and protein, caused us to consider the possibility that decreased migration of MFBs to distal lung could have contributed to a decrease in alveolar formation associated with MV. Previous studies showed that mutant mice deficient in PDGF-A exhibited failed alveolarization that was attributed to impaired MFB migration to septal tips as a result of defective PDGF-A signaling (29). We therefore used IHC and quantitative image analysis to compare the amount and distribution of vimentin, expressed in MFBs, that was present in distal lung of newborn mice after 24 h of MV with 40% O2 compared with unventilated control pups. In preliminary studies, cultured lung MFBs harvested from newborn mice stained positive for both vimentin and α-smooth muscle actin (α-SMA; see online supplement). Because α-SMA also stained smooth muscle cells, whereas vimentin did not, we used vimentin to detect MFBs in lung tissue sections. Quantitative image analysis showed ∼20% less vimentin staining of distal lung tissue after MV compared with unventilated controls (Fig. 7, A and B). Results were similar for α-SMA (see online supplement).

Fig. 7.

A: immunostaining for vimentin in lungs of newborn mice that breathed 40% O2 for 24 h either without MV (control) or with MV at 180 breaths/min. Arrows point to brown staining in the walls of distal air spaces. Magnification was ×400. B: quantitative image analysis showed reduced abundance of vimentin stain (area), expressed as percentage of total tissue area, after 24 h of MV with 40% O2 at 180 breaths/min compared with unventilated control group. All data are means ± SD; n = 5 pups/group. *Significant difference compared with control group, P < 0.05.

The finding that MV with 40% O2 yielded little or no apparent influx of neutrophils or macrophages into the lungs prompted us to conduct a more rigorous assessment of pulmonary inflammation. Measurements of MPO activity (Fig. 8A) and lung expression of proinflammatory cytokines (Fig. 8, BD) confirmed a lack of inflammation after 8–24 h of MV.

Fig. 8.

A: myeloperoxidase activity (measured as OD450, optical density at 450 nm) was not significantly different in lungs of control and ventilated groups of newborn mice (n = 3/group). B: lung mRNA content of TNF-α, expressed relative to 18S rRNA, was significantly less after 8 h of MV with 40% O2 than it was in unventilated control pups, but there was no significant difference between groups in lung protein abundance of TNF-α after 24 h of MV. C: lung content of IL-1β mRNA and protein were not significantly different between control and mechanically ventilated groups of mice that breathed 40% O2 (mRNA measured after 8 h, protein measured after 24 h). D: lung content of IL-6 mRNA and protein were similar in control and mechanically ventilated groups of mice that breathed 40% O2 (mRNA measured after 8 h, protein measured after 24 h). All data are means ± SD; n = 4–6 pups/group. *Significant difference compared with control group, P < 0.05.

DISCUSSION

A Unique Murine Model to Study Lung Development

This report is the first to describe long-term MV of newborn mice and thereby establishes a new experimental model that can be used to define mechanisms of disordered lung septation and angiogenesis as observed in neonatal CLD. The model enabled us to examine the specific effects of prolonged cyclic stretch, with or without associated adjustments of inspired O2, on lung growth, which is germane to the development of CLD. In this study we show that MV of newborn mice with O2-rich gas for 24 h reduces lung expression of genes and proteins that regulate formation of alveoli and pulmonary capillaries and induces structural changes indicative of impaired lung septation without causing significant inflammation. The study provides a basis for applying this model to mutant neonatal mice as a way of defining the genomics of normal and defective lung development and disease. This model also can be applied to help clarify pathogenesis and evaluate strategies for treating a number of life-threatening respiratory disorders including neonatal CLD.

Earlier reports described MV of neonatal mice using a negative-pressure, body-enclosing device designed to provide brief respiratory support and prolong survival of mutant newborn mice that had respiratory depression from lack of functional N-methyl-D-aspartate (NMDA) receptors in the brain (25, 38). In these studies, NMDA-null neonatal mice, which normally die from respiratory failure soon after birth, were kept alive briefly (∼1 h) using negative-pressure assisted ventilation. Technical difficulties with this ventilation strategy and the device that was created to apply it, however, prevented its use for long-term respiratory support of newborn mice. Another group recently reported using brief (1-h) positive-pressure MV of 10-day-old mice to assess metabolic and neuronal effects of isoflurane anesthesia in hypoxic-ischemic brain injury (30).

The experimental approach described here takes advantage of the fact that alveoli and lung capillaries in mice form mainly after birth at term gestation, thus enabling us to examine the impact of lengthy MV on genes and proteins that regulate lung growth and development and to assess structural changes in the lung that occur in response to prolonged cyclic stretch with O2-rich gas. To help make our findings relevant to the situation that prevails during development of BPD, we used a ventilation strategy similar to the current clinical approach that has been adopted in many newborn intensive care units to treat tiny infants with respiratory failure. This approach uses small tidal volumes, low inflation pressures, and modest concentrations of inspired O2 to maintain adequate respiratory gas exchange while minimizing the risk of lung injury.

There are, however, important differences between newborn mice that receive MV after birth at term gestation compared with premature infants and experimental models of BPD that have been created in premature baboons and lambs. These include a more fully developed respiratory drive, a more competent surfactant system, and the need for a tracheostomy in lieu of an endotracheal tube to deliver MV in newborn mice. Despite these differences, prolonged MV with O2-rich gas yielded similar changes in expression of genes that regulate formation of alveoli and pulmonary capillaries in newborn mice compared with preterm animal models of CLD (7, 31).

Gene Regulation of Alveolar Formation

Previous studies showed defective lung septation and pulmonary emphysema in mice that were rendered deficient in VEGF or where its receptor, VEGF-R2, was blocked (15, 22, 28, 33), indicating that VEGF and VEGF-R2 have important roles in forming and maintaining normal alveolar structure. Another report used gene microarray analysis to link downregulation of VEGF-R2 with failed formation of alveoli in lungs of newborn mice that had been treated with dexamethasone, which inhibits lung septation (12).

Our finding of decreased expression of VEGF-A and VEGF-R2 (also called fetal liver kinase-1, Flk-1, or kinase domain receptor, KDR) in lungs of newborn mice that received prolonged MV with O2-rich gas is consistent with earlier reports of decreased pulmonary expression of VEGF and one of its receptors, fms-like tyrosine kinase (also called Flt-1 or VEGF-R1), in both premature baboons and human infants with CLD (3, 31) and with reduced lung expression of VEGF-A and VEGF-R2 observed in preterm lambs with CLD after 3 wk of MV with O2-rich gas (7). 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 (27, 42).

The observation that PDGF-A mRNA was reduced in lungs of newborn mice after MV with either 40% O2 or air suggests that cyclic stretch, rather than hyperoxia, acting on the developing lung is the main modifier of PDGF-A expression in this model. It is noteworthy that the decreased PDGF-A mRNA and protein in the lungs of these mice treated with MV is similar to the PDGF-A changes noted in preterm lambs with CLD after 3 wk of MV (7). These findings, however, differ from results of previous studies of hyperoxic lung injury in newborn rats in which lung expression of PDGF-A mRNA was unaffected by 14 days of exposure to either 60% or 85% O2, which also inhibits alveolar septation and induces a form of lung injury that, in some respects, resembles the pathology observed in CLD (10, 16, 17). This apparent difference in lung expression of PDGF-A between newborn mice exposed to MV with 40% O2 compared with newborn rats exposed to 60–85% O2 suggests that reduced pulmonary expression of PDGF-A may be related more to mechanical distention of the lung during MV rather than hyperoxia.

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 (9, 29). Reduced lung abundance of PDGF-A, coupled with air space enlargement and diminished expression of vimentin in distal lung of newborn mice after 24 h of MV, is consistent with the view that PDGF-A signaling may be important in directing MFBs to septal tips where they provide a matrix for alveolar formation during lung development (8).

Our finding of reduced TN-C mRNA and protein in lungs of newborn mice after prolonged MV with either 40% O2 or air suggests that mechanical stress rather than increased O2 alone is the primary determinant of TN-C expression in this model. TN-C is an extracellular matrix glycoprotein that has been shown to play an important role in embryonic lung development (24, 45, 47). Its localization at the growing tips of future airways and at the interface of epithelial and mesenchymal cells has caused investigators to conclude that TN-C may be involved in matrix organization, epithelial-mesenchymal cell interactions, and branching morphogenesis in the developing lung. Lungs of mutant fetal mice lacking TN-C did not branch in organ culture, and assessment of lung morphology at postnatal day 2 showed enlarged air spaces in TN-C-null mice compared with lungs of wild-type pups, suggesting that TN-C helps to regulate alveolar formation as well as airway branching (39). Our observation that prolonged and repetitive mechanical stretch of the newborn mouse lung reduces lung expression of TN-C and leads to enlarged air spaces is consistent with the view that TN-C has a role in regulating alveolar formation during the early postnatal period. Other studies have demonstrated that TN-C, regulated by the paired-related homeobox gene Prx1, also contributes to the formation of the pulmonary vasculature (21). Thus reduced lung expression of TN-C might contribute to the impaired pulmonary microvascular development observed in infants dying with BPD (3) and the reduced lung capillary surface density described in authentic animal models of CLD (5, 31).

Dysregulation of Lung Septation without Apparent Inflammation

The discovery of reduced pulmonary expression of several genes that regulate lung septation led us to consider the possibility that lengthy MV with O2-rich gas may have induced lung inflammation, which is considered to be a major determinant of failed alveolar septation in neonatal CLD. Lack of an inflammatory response after 24 h of MV with 40% O2 is consistent with a previous study which showed reduced susceptibility to ventilator-induced lung injury in newborn rats compared with adult rats (26). In this study, MV with air, using a PIP of 30 cmH2O for 90 min, resulted in diminished lung compliance, pulmonary edema, and increased lung lavage concentration of TNF-α in adult rats but not in newborn rats. In a related study, MV of newborn and adult rats for up to 3 h with air using a high tidal volume (25 ml/kg) yielded a significant reduction in respiratory system compliance, increased lung water content, and histological evidence of lung injury in adults but not in newborns (14). It is therefore not surprising that MV with more modest tidal volumes and inflation pressures, albeit with a greater FIO2 and for a longer duration, would not cause lung inflammation. In the absence of obvious inflammation, however, there was reduced lung expression of several genes that regulate alveolar septation. These findings suggest that exposure of developing lungs to cyclic stretch and mild hyperoxia can lead to defective alveolar formation without apparent inflammation or injury.

We expected that MV at a relatively slow respirator rate (60 bpm) with relatively large tidal volumes (∼10 μl/g body wt) would have a greater adverse effect on expression of genes that regulate lung growth than the more rapid respirator rate (180 bpm) with smaller tidal volumes (∼5 μl/g). Lack of a detectable difference in molecular or structural endpoints between these two ventilation strategies may reflect the fact that MAP was similar for the two groups and that neither strategy was sufficient to produce significant lung injury. These results are consistent with previous studies that showed little or no difference in the severity of CLD in preterm lambs that received MV for 3 wk at a respirator rate of 20 bpm and tidal volume of ∼15 ml/kg compared with a rate of 60 bpm and tidal volume of ∼5 ml/kg (1, 5). Thus exposure to positive-pressure MV with O2-rich gas, irrespective of the inflation pattern by which it is delivered, is likely to be the critical determinant of impaired lung septation.

Application of this experimental model to neonatal mutant mice bearing gene modifications that impact lung septation and angiogenesis is likely to yield important insights to help define mechanisms by which MV can cause defective lung development and disease. Such studies also may provide rationale for devising novel strategies to treat various respiratory disorders including neonatal CLD.

GRANTS

This study was supported by National Heart, Lung, and Blood Institute Grants HL-62512 (R. D. Bland) and HL-56401, SCOR Project V (R. D. Bland), and the Vera Moulton Wall Cardiopulmonary Research Center, Stanford University.

Footnotes

  • 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.

REFERENCES

View Abstract