INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Vascular endothelial growth factor (VEGF) is a relatively specific endothelial cell mitogen that regulates endothelial cell differentiation and angiogenesis in vivo (1). As an important permeability factor, VEGF plays a central role in vascular repair (3, 4). In highly vascularized tissues such as the lung, VEGF is relatively abundant and is expressed mainly in alveolar epithelial cells (5, 6).

Respiratory distress syndrome (RDS) and the development of chronic lung injury in preterm infants are characterized by early endothelial cell damage (7). Destruction of the pulmonary microvasculature is characteristic of experimental lung injury, and during hyperoxic exposure, endothelial regeneration is impaired (8). Integral to recovery from experimental lung injury is repair of microvascular endothelium, which correlates with increased expression of VEGF in alveolar epithelial cells (9). In newborn rabbits, hyperoxic lung injury reduces pulmonary expression of VEGF mRNA and protein, a reduction suggested to contribute to impaired microvascular repair of the injury (10).

Because VEGF has recently been found in lung effluents from preterm infants (11), we hypothesized that in these patients VEGF may play a role in recovery from acute respiratory distress. The aim of this study was to characterize the concentrations of VEGF in tracheal aspirates and plasma, and to evaluate the influence of perinatal factors on pulmonary VEGF in preterm infants with respiratory distress during the early postnatal period.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Patients

With the approval of the Ethics Committee of the Hospital for Children and Adolescents, University Central Hospital, Helsinki, 44 preterm infants (gestational age, 27.3 ± 2.0 wk; birth weight, 962 ± 319 g; mean ± SD) were enrolled in this study.

All infants enrolled had clinically diagnosed RDS, and were intubated at birth because of failure to establish spontaneous ventilation. They underwent mechanical ventilation during the study period. Patients with major anomalies were excluded.

Eight infants were born to mothers with established proteinuric preeclampsia. Premature rupture of the membranes (more than 24 hr ante partum) was present in 11 cases, and chorionamnionitis was present in nine; in four pregnancies, both of these complications occurred. Chorionamnionitis was diagnosed on the basis of clinical signs, leukocytosis, and increased concentration of C-reactive protein. The mothers of 36 infants had received treatment with antenatal glucocorticoids (12 mg of betamethasone twice with a 12-h interval) for 7 ± 8 d (range, 1 to 32 d) before delivery. The mean number of courses of betamethasone was 1.4 ± 0.9 (range, 1 to 3).

Of the 44, 25 infants were born by cesarean section and 19 by vaginal delivery. Mean Apgar score at 1 min was 5 ± 2 (range, 1 to 9), and at 5 min it was 7 ± 2 (range, 2 to 10). Blood gas analysis from the umbilical artery was available for 26 patients. Mean base excess was 3.9 ± 3.9 (range, 13.7 to 3.0), and mean pH was 7.26 ± 0.11 (range, 6.90 to 7.42); in two of the infants pH was below 7.1.

Surfactant was administered to 35 patients, starting at a mean age of 7 ± 7 h (range, 2 to 34 h); of these infants, 13 required two doses, and 11 required three or more. Of the 35 patients, eight received Exosurf (Glaxo Wellcome, Greenford, UK) 4 ml/kg, and 25 received Curosurf (Chiesi Farmaceutici SPA, Parma, Italy) 100 mg/kg; two received both surfactants. In 34 patients patent ductus arteriosus required treatment with indomethacin, given as four doses of 0.1 mg/kg at 12-h intervals, starting at a mean age of 23 ± 10 h (range, 7 to 54 h); for six of these patients the treatment had to be repeated. All patients received ampicillin 200 mg/kg/d and nethilmicin 6 mg/kg/d from the first day of life; for 18 of them ampicillin was changed to vancomycin 15 mg/kg/d during the study period because of clinical signs of septic infection; for three of them, diagnosis was verified by positive blood culture. Two infants had radiologically diagnosed pneumonia. To facilitate weaning from mechanical ventilation, 12 patients received treatment with dexamethasone at a dose of 0.5 mg/kg/d for 3 d, followed by 0.25 mg/kg/d for 5 to 7 d and 0.125 mg/kg/d for 7 d, starting at a mean age of 8 ± 4 d (range, 3 to 15 d). For three of the patients, dexamethasone was started during the study period (on Days 3, 3, and 5).

Thirteen patients developed bronchopulmonary dysplasia (BPD), which was defined as the need for supplemental oxygen at the age of 36 gestational weeks, in association with chest radiographic findings typical for BPD (12). Eight patients suffered from intraventricular hemorrhage, 11 developed retinopathy of prematurity, and three had necrotizing enterocolitis, which was fatal to one on Day 11. The two other deaths were from severe BPD on Days 56 and 260.

Sample Collection and Assays

Samples of tracheal apiration fluid (TAF) were collected once daily by standardized routine tracheal lavage as previously described (13). Briefly, 1 ml of sterile isotonic saline was instilled into the endotracheal tube, the patient was manually ventilated for 3 breaths, and the trachea was suctioned twice for 5 s. Recovery of lavage fluid was consistent (2 to 3 ml) in all subjects each time. For analysis of tracheal aspirates, secretions were collected into a trap and transferred into tubes containing 500 IU of aprotinin (Trasylol; Bayer, Leverkusen, Germany) and 5 mg of deferoxiamine (Desferal; Ciba, Basel, Switzerland). The tubes were stored at 20° C until analysis. Blood samples were drawn through radial artery catheters into EDTA tubes. The plasma was rapidly separated and stored at 70° C until analyzed.

A tracheal aspirate sample was taken within 3 h postnatally before treatment with surfactant for determination of lecithin/sphingomyelin ratio (LS-ratio) and presence of ceramide lactoside (14).

VEGF was analyzed by the Quantikine Human VEGF Immunoassay (R&D Systems, Oxon, UK). In order to estimate the in situ pulmonary concentration of VEGF, a correction for dilution of the tracheal aspirate sample was calculated by use of the ratio of urea-N in the tracheal aspirate sample and in the serum sample taken within 12 h.

Statistical Analysis

Patient data are given as mean ± SD and range, and the results are given as mean ± SEM. Comparisons between groups were performed with the Kruskal-Wallis one-way ANOVA, and the Mann-Whitney U test served for the post-hoc comparisons. Comparisons between unpaired items were performed with the Mann-Whitney U test. Repeated measures ANOVA was used. The chi-square served for categorical variables, and simple regression analysis for continuous variables. Logarithmic transformation of the data was performed when appropriate; p values less than 0.05 were considered statistically significant. All calculations were done with StatView 4.1.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

VEGF in TAF

A total of 189 TAF samples were collected during the first postnatal week. The mean concentration of VEGF increased from 25 ± 12 pg/ml on the first day to 526 ± 120 pg/ml on Day 7 (Figure 1). The mean concentration of VEGF was significantly lower on Days 1 to 3 than on Days 4 to 7 (106 ± 25 pg/ml, n = 72, and 342 ± 36 pg/ml, n = 117, respectively, p < 0.0001).

View larger version (12K): [in this window] [in a new window]   Figure 1.   Daily concentrations of VEGF (pg/ml) in tracheal aspirate fluid (open circles) and in plasma (closed circles) during the first postnatal week.

In patients born to mothers with premature rupture of the membranes or with chorionamnionitis, or with both, the increase in VEGF was greater, and in patients born to mothers with preeclampsia smaller than in patients without these antenatal complications (Figure 2). The differences in mean concentrations reached significance on Days 4 to 7 (Table 1).

View larger version (16K): [in this window] [in a new window]   Figure 2.   Daily concentrations of VEGF (pg/ml) in tracheal aspirate fluid in infants born to mothers with premature rupture of the membranes, or with chorionamnionitis, or with both (open circles) (n = 16), preeclampsia (closed circles) (n = 8), or with none of these complications (open squares) (n = 20). *p < 0.05.

No correlations existed between VEGF and birthweight or gestational age. VEGF associated with birth asphyxia, defined as low 1- and 5-min Apgar scores and low base-excess in blood- gas analysis from the umbilical cord artery obtained from 26 patients (Table 1).

In 32 patients a tracheal aspirate sample was obtained within 3 h after birth for determination of surfactant maturity. In these samples the LS-ratio showed a correlation with VEGF. In seven of these patients ceramide lactoside was present, and they had higher VEGF (Table 1). The presence of ceramide lactoside was associated with premature rupture of the membranes or with chorionamnionitis, or with both (p = 0.0026).

Although no correlation existed between initial arteriolar-alveolar ratio and VEGF, patients who required surfactant therapy had lower VEGF. A negative correlation was also found between VEGF and the number of surfactant doses required. On Days 4 to 7 mean inspiratory oxygen and VEGF showed a negative correlation (Table 1).

On Days 4 to 7, those 13 patients who developed BPD had a lower mean VEGF than did those surviving without BPD (Table 1). Patients developing BPD were of lower gestational age (26.6 ± 0.5 wk versus 28 ± 0.4 wk; p = 0.033), and birth weight (752 ± 34 g versus 1,097 ± 61 g; p = 0.001) and 1-min Apgar score (4 ± 0.5 versus 6 ± 0.4; p = 0.04). These patients were intubated for a longer time (25 ± 5 d versus 10 ± 1 d; p = 0.0008), and received higher mean inspiratory oxygen during the first week (0.41 ± 0.02 versus 0.32 ± 0.01; p < 0.0001).

No differences were found in VEGF concentrations between patients receiving glucocorticoid treatment (antenatal or postnatal) or not (data not shown).

A correction for dilution of the tracheal aspirates was performed in 69 of the 189 samples. In these samples the actual mean concentrations of VEGF were estimated to be 4.1 ± 0.9 ng/ml on Days 1 to 3, and 16.3 ± 2.3 ng/ml on Days 4 to 7.

VEGF in Plasma

In order to estimate the levels of VEGF in plasma, 24 blood samples were taken from nine of the patients during the first postnatal week. The concentration of VEGF in plasma remained constant, mean concentration during the first week being 48 ± 6 pg/ml (Figure 1).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We report here increasing concentrations of VEGF in the lungs of preterm infants with respiratory distress during the early postnatal period. The concentration of pulmonary VEGF increased steadily severalfold during the first postnatal week in comparison with plasma, in which it remained constant. The pulmonary concentrations of VEGF found here are in the same range as those recently shown to induce proliferation and differentation of human fetal airway epithelial cells in vitro (15).

The concentration of VEGF in TAF was significantly higher than in plasma, suggesting that VEGF is synthesized in the lungs. This is in accordance with animal studies showing VEGF to be abundantly expressed in lung tissue, including in pneumocytes and activated alveolar macrophages (5, 6). In experimental animals recovering from hyperoxic lung injury, the expression of VEGF increases specifically in alveolar type II cells, in contrast to cell types of mesenchymal origin such as alveolar macrophages and fibroblasts (9, 10). It is possible that in the preterm lung, the type II cell is a major source of pulmonary VEGF.

In RDS, the acute lung injury usually begins to resolve within the first 3 postnatal days, after which, if not complicated, a recovery phase begins (16). In premature infants with RDS, pulmonary events associated with the development of chronic lung injury occur during the recovery phase at the end of the first week (17). These events include enhanced pulmonary inflammation characterized by recruitment and activation of neutrophils, increased levels of proinflammatory cytokines, and increased fibronectin (17). The high pulmonary concentration of VEGF coincides with these events at the end of the first week. Patients developing BPD had lower VEGF during this period; this was not due to their lower gestational age because VEGF in TAF and gestational age were in no way related. Infants who eventually developed BPD required higher inspiratory oxygen concentrations. Hyperoxia has been reported to decrease VEGF expression by alveolar epithelial cells in rabbits (10). Our data are in agreement with these previous observations since lower VEGF concentrations were found in infants who later developed BPD. The increase in concentrations in VEFG seen in all preterm infants postnatally could be due to a developmental change in VEGF expression after birth.

Although no correlation was found between gestational age or birth weight and pulmonary VEGF, a correlation did exist between the functional maturity of the alveolar type II cells, defined as LS-ratio, and pulmonary VEGF, suggesting that VEGF may reflect functional maturity of the lung. Preeclampsia of the mother was associated with lower pulmonary VEGF, a finding in line with low concentrations of surfactant protein A in the preterm lung (20). Because VEGF has been shown to induce surfactant protein expression in human fetal alveolar epithelial cells in vitro, it may participate in autocrine regulation of type II cells also in the preterm lung (15).

Higher pulmonary VEGF was observed in infants born to mothers with premature rupture of the membranes and chorionamnionitis. Moreover, the presence of ceramide lactoside in tracheal aspirates, a marker for antenatal infection (21), was associated with high pulmonary VEGF. In experimental animals, pulmonary VEGF expression is induced by proinflammatory mediators, many of which have been shown to be higher in the lung of preterm infants born to mothers with premature rupture of the membranes or chorionamnionitis (17, 22). Chorionamnionitis has been associated with less severe RDS, and this effect has been ascribed to prenatal inflammation's accelerating lung maturition (22). Therefore, prenatal inflammation seems to increase pulmonary VEGF. This may result either from its specific effect on VEGF expression or by the general effect of inflammation on lung maturition.

Asphyxia was associated with high pulmonary VEGF; because hypoxia is an important stimulator of VEGF expression, it is possible that prenatal hypoxia may induce an increase in pulmonary VEGF (23, 27). A negative correlation existed between mean inspiratory oxygen and pulmonary VEGF. In experimental animals the expression of VEGF in the lung is decreased by hyperoxia, but it increases during recovery to normoxia (9, 10). On the basis of our present data it thus appears possible that in the lung of the preterm infant, both hypoxia and hyperoxia may affect the expression of VEGF.

Preterm infants receiving dexamethasone have been reported to have higher VEGF in TAF (11). On the other hand, dexamethasone has been shown to reduce induction of VEGF expression caused by hypoxia in human fetal lung in vitro (30). Our patients' values showed no correlations between antenatal or postnatal glucocorticoid treatment and pulmonary VEGF. As the use of antenatal steroids was not randomized, and only three patients received postnatal dexamethasone during the study period, the effect of glucocorticoids on VEGF cannot be elucidated on the basis of the present data.

In conclusion, the increasing pulmonary concentrations of VEGF found in these preterm infants suggest a physiologic role for this growth factor in the preterm lung. Toward the end of the first postnatal week, during recovery from acute respiratory distress, these concentrations were of the same magnitude as those that induce proliferation and differentation of human fetal airway epithelial cells in vitro (15). In the preterm lung, VEGF may thus be indicative of maturity of the type II cells. Moreover, as patients developing BPD had lower pulmonary VEGF, and VEGF participates in repair of experimental lung injury (9, 10), it may also play a role in recovery from acute lung injury in the preterm infant.