INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Lung maturation normally occurs late in gestation just prior to term birth in animals and humans. Fetal glucocorticoid levels increase before term birth in most species, and this increase in glucocorticoids is thought to contribute to the normal late gestational maturation of the lung (1, 2). However, mice that lack corticotropin-releasing hormone or glucocorticoid receptor function have essentially normal development of the surfactant system despite a delay in distal lung structural development (3, 4). Early clinical lung maturation is the phenomenon of relatively normal gas exchange and lung mechanics in infants born prematurely and before lung maturity normally occurs. About 20% of preterm infants at very early gestations of 24-28 wk can have essentially normal lung function (5, 6). Glucocorticoids induce early maturation in experimental models, and antenatal glucocorticoid treatments decrease the incidence of respiratory distress syndrome by 40-50% (2, 7). However, perinatal complications assumed to result in fetal stress and increased fetal cortisol levels such as fetal growth restriction, preeclampsia, and preterm prolonged rupture of membranes are not consistently associated with a decreased incidence in respiratory distress syndrome (8). Chronic infection and amnionitis that often occurs with very early preterm delivery recently were associated with a decreased incidence of respiratory distress syndrome (11). Watterberg and coworkers (12) suggested that the infection/amnionitis resulted in early lung maturation because of fetal stress and increased fetal cortisol. Bry and coworkers (13) demonstrated that intraamniotic injection of the proinflammatory cytokine interleukin (IL)-1 increased the mRNA for SP-A and SP-B in preterm rabbits and increased compliance. Increased lung gas volumes and surfactant phospholipids were measured 48 h after intraamniotic IL-1 in sheep (14). We generalized the observation by demonstrating that the proinflammatory stimulus resulting from intraamniotic endotoxin injection improved lung function and increased surfactant lipids and proteins more in sheep than did antenatal glucocorticoids (15). Our hypothesis for this study was that endotoxin would induce early lung maturation by mechanisms independent of cortisol. We also anticipated that the lung maturation effects would be dose sensitive with a high dose causing injury. We quantified dose and treatment to delivery interval effects on physiological indicators of early maturation, surfactant lipid pool sizes, and umbilical arterial blood cortisol levels.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals

These studies were performed in Western Australia using date-mated singleton Merino ewes. The studies were approved by the appropriate animal care and use committees in Australia and at Children's Hospital Medical Center, Cincinnati, OH. Ewes were weighed at 100 d of pregnancy and randomized to the different doses, different treatment to delivery intervals, and repeated dose groups by ear tag number as the only indicator. Endotoxin (E. coli 055:5; Sigma, St. Louis, MO) was solubilized in saline at a concentration of 20 mg/ml and filtered through a 0.45-µm filter. The solution was diluted as necessary and used immediately or frozen one time only before use. A dose of endotoxin diluted into saline to a final volume of 2 ml (5 ml for the 100 mg endotoxin dose) was given by intraamniotic injection to ewes gently restrained in the shearing position (16). The fluid around the fetus was imaged with a 3.5-mHz sector transducer using an Aloka ultrasound machine, and a 20-gauge spinal needle was used to withdraw fluid, which was assumed to be amniotic fluid if opalescent and containing particulate material. The endotoxin or saline control dose was given. The fluid was verified to be amniotic fluid and not allantoic fluid by measurement of Na⁺ and Cl concentrations using a blood gas and electrolyte analytical instrument (Bayer Diagnostics Rapidlab 820) (17, 18). Four animals that inadvertently received intraallantoic injections of endotoxin were also studied. The animals were then released back to open paddocks.

Delivery and Postnatal Assessments

At 125 d gestation the lambs were delivered by cesarean section after securing a 4.5-mm endotracheal tube by tracheotomy as described previously (19). At delivery a fetal arterial blood sample was drawn from the placental cord for pH and blood gas analysis and for a plasma cortisol measurement using a commercial radioimmunoassay kit (ICN, Irvine, CA). The lambs were dried and ventilated with time-cycled, pressure-limited infant ventilators set to deliver 100% oxygen at a respiratory rate of 40 breaths/min, an inspiratory time of 0.7 s, and a positive end-expiratory pressure of 3 cm H₂O. Peak inspiratory pressures were initially set at 35 cm H₂O and subsequently adjusted during the 40-min postdelivery study period to target arterial PCO₂ values of 50 mmHg. Peak inspiratory pressures were limited to 40 cm H₂O to avoid pneumothorax (19). Other ventilator settings were not altered during the study period. The personnel delivering and caring for the lambs were blinded to the treatments that the animals had received. An arterial catheter was advanced into the descending aorta through an umbilical artery, and lambs received supplemental pentobarbital (15 mg/kg) by slow arterial infusion if spontaneous respirations were noted. The body temperature of each lamb was maintained at ~ 39° C with a radiant warmer and by covering the lamb with plastic wrap.

Tidal volume was measured with a pneumotachometer and ventilatory pressure was calculated as peak inspiratory pressure measured at the endotracheal tube minus the positive end-expiratory pressure of 3 cm H₂O. Ventilation also was regulated by adjusting peak ventilatory pressure to achieve a tidal volume of less than 10 ml/kg. Respiratory system compliance was calculated by dividing the tidal volume by the ventilatory pressure, and then normalized to body weight in kilograms (19). Ventilation efficiency index (VEI), an index that integrates ventilation with respiratory support, was calculated by the formula: VEI = 3800  P · f · PCO₂ (20). In this equation, 3800 is a carbon dioxide production constant (ml/mm Hg/kg/min), P is ventilatory pressure, f is the ventilation rate, and PCO₂ is the arterial partial pressure of carbon dioxide. Blood gas values were measured every 10 min to permit ventilator adjustments. After a final arterial blood sample at 40 min of age, each lamb was given pentobarbital (30 mg/kg), and the tracheal tube was clamped for 3 min to achieve atelectasis by oxygen absorption for the deflation pressure-volume curve measurement from a peak pressure of 40 cm H₂O (19).

Surfactant Lipids

The lungs were removed from the chest, each lung was weighed, and the left lung was lavaged five times by infusing and withdrawing by syringe a sufficient volume of saline at 4° C to fully fill the lungs (16). The five lavages were pooled, the total volume was measured, and aliquots were extracted with chloroform-methanol (2:1) (21). Pieces of the left lower lobe were homogenized, and aliquots were extracted with chloroform-methanol. The lipid extracts were treated with osmium tetroxide and saturated phosphatidylcholine (Sat PC) was recovered after alumina column chromatography and quantified by phosphorus assay (22, 23).

Inflammation Score

The right upper lobe of each lung was inflation fixed with 10% formalin at 30 cm H₂O pressure. The amount of inflammation was graded in a blinded fashion by scoring three 5-µm sections from each animal as zero (no inflammatory cells in tissue or airspaces), 1 (a few cells), 2 (a moderate cell infiltration), and 3 (large numbers of inflammatory cells in airspaces and tissue) (15). Average scores for airspaces and tissue were calculated for each animal.

Statistical Analysis

All values are given as means ± SE. The control group included six animals that received intraamniotic saline injections 7 d before delivery and five animals that received the same injections 2 d before delivery. Comparisons between the control group and the endotoxin treatment groups were performed by analysis of variance and the Dunnett procedure. Two-tailed t tests were used for two group comparisons.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Dose of Endotoxin

The only information available about selecting a dose of endotoxin to induce early lung maturation was our finding that 20 mg endotoxin given by intraamniotic injection 6 d before preterm delivery was effective (15). Therefore, we evaluated doses of 1 mg, 4 mg, 20 mg, and 100 mg endotoxin for a treatment to delivery interval of 7 d. The endotoxin doses did not alter birth weight or umbilical arterial blood pH and PCO₂ values relative to control values (Table 1). There were no fetal deaths. Umbilical arterial blood cortisol levels in control animals were low, and the values were not increased by any of the doses of endotoxin (Figure 1). After 40 min ventilation the four different endotoxin doses resulted in similar and significant improvements in blood gas values and tidal volumes at lower ventilation pressures calculated as peak inspiratory pressures minus the positive end expiratory pressure of 3 cm H₂O (Table 1).

View larger version (37K): [in this window] [in a new window]   Figure 1.   Effect of dose of endotoxin for a treatment interval of 7 d. Endotoxin doses are compared with the values for control lambs, *p < 0.05, **p < 0.01 versus control. All values for lung gas volumes on the pressure-volume curves for the endotoxin-treated groups were different from values for the control group (p < 0.01).

Oxygenation was not stable in these ventilated preterm lambs because of variable shunts through the ductus arteriosis and foramen ovale and lung immaturity. To estimate the maximal effect of endotoxin exposure on oxygenation, we calculated the mean PO₂ values for each treatment group using the highest PO₂ achieved of the four blood gas measurements for each animal. Endotoxin treatment with 1 mg and 100 mg significantly increased PO₂ (Figure 1). The endotoxin-exposed animals had compliances that were about 2-fold higher than control lambs, and VEI was increased about 3-fold. Lung gas volume measured at 40 cm H₂O pressure was increased from a control mean value of 10.6 ml/kg to 40-50 ml/kg by endotoxin. The amount of Sat PC in alveolar washes increased about 20-fold from a mean control value of 0.36 ± 0.06 µmol/kg to values over 6.6 ± 1.5 µmol/kg (Figure 1). The amount of Sat PC in the total lungs also doubled in the endotoxin-treated groups. The surprising result was that a low dose of 1 mg endotoxin was equivalent to 100 mg endotoxin in terms of improving lung function and increasing Sat PC pools after preterm delivery. However, the inflammation score did distinguish the doses (Figure 1). The amount of histological inflammation was minimal in the airspaces and tissue for the 1 mg dose. The higher doses caused inflammatory cells to appear in the lungs.

Interval from Treatment to Delivery

Animals received 20 mg endotoxin at intervals from 1 d to 15 d before preterm delivery at 125 d gestation. The time of dosing had no effect on birth weight or umbilical arterial blood pH and PCO₂ values, and there were no fetal deaths or abortions (Table 2). Umbilical arterial blood cortisol values were low and not different from control animals (Figure 2). We also treated two groups of four animals each 5 h and 15 h before delivery with 20 mg endotoxin. Although these animals were not ventilated or assessed for postnatal lung function, umbilical arterial blood cortisol values were 0.9 ± 0.4 µg/dl and 0.85 ± 0.01 µg/dl for treatment to delivery intervals of 5 and 15 h, respectively, values that were not different from control values. The highest mean PO₂ value was 451 ± 65 mm Hg after the 15-d treatment-to-delivery interval, significantly higher than values for animals treated for intervals of 1, 2, and 4 d and control animals. Compliance and VEI increased for the 7-d and 15-d treatment-to-delivery intervals. Although values for compliance, VEI, and lung volumes at 40 cm H₂O at 4 d were not different from the control values when analyzed by the 6-group ANOVA, all values at 4 d were significantly different from the control values by t tests (p < 0.01). The amount of Sat PC in alveolar washes increased 3-fold from the control value by 2 d (t test, p < 0.03) and increased 7.6-fold from the control value by 4 d (t test, p < 0.001). Further increases had occurred by 7 and 15 d. A similar pattern was found for the increase in Sat PC in the total lung. Total lung Sat PC increased by 25% on Day 2 (t test, p < 0.01) and increased further by 58% on Day 4 and by 94% on Day 7. Therefore, an endotoxin effect on surfactant pool sizes was detected within 2 d of endotoxin exposure, and a physiological effect was present by 4 to 7 d. The 20 mg endotoxin dose resulted in comparable numbers of inflammatory cells in the airspaces and lungs for the 1- to 7-d treatment intervals.

View larger version (37K): [in this window] [in a new window]   Figure 2.   Effect of time from treatment with 20 mg endotoxin to delivery at 125 d gestation. The time to treatment intervals are compared with control values, *p < 0.05, **p < 0.01 versus control. Values for lung gas volumes were different from control values for treatment intervals of 7 d and 15 d (p < 0.01).

Single versus Two Doses of Endotoxin

We thought that two doses of endotoxin would either further improve postnatal lung function or cause lung injury. Two doses of 20 mg endotoxin given 15 d and 7 d before delivery at 125 d gestation were compared with control animals and animals given single doses of endotoxin 7 d or 15 d before delivery (Table 3). We also evaluated a group of nine lambs exposed to 20 mg doses of endotoxin given 3 d and 7 d before delivery. The endotoxin exposure did not alter birth weights or umbilical arterial blood pH or PCO₂ values. Umbilical arterial blood cortisol values were not higher than the values for control lambs or the lambs that received the single doses of endotoxin (Figure 3). The highest PO₂ values were similar for the single dose given at 15 d and two doses given at 15 d and 7 d. The two doses improved compliance, VEI, or lung volumes measured by pressure-volume curves similarly to the single doses. The single and two-dose treatments had similar large effects on alveolar Sat PC pool sizes and the amount of Sat PC in the total lungs. Therefore, there were no apparent effects on lung physiology or surfactant pool sizes from two doses of endotoxin relative to a single dose. However, the inflation score for the tissue was increased above the control values for the two-dose treatment groups. All groups exposed to endotoxin had indications of mild inflammation in the airspaces.

View larger version (36K): [in this window] [in a new window]   Figure 3.   Single versus two doses of 20 mg endotoxin and times to delivery at 125 d gestation. Endotoxin treatment groups that received a single dose of intraamniotic endotoxin 7 d or 15 d before delivery and treatment groups that received two doses of endotoxin 7 d and 15 d (7 + 15 d) or 3 d and 7 d (3 + 7 d) before delivery are compared with the control group, *p < 0.05, **p < 0.01. The values for the treatment groups were not different. Values for lung gas volumes from the pressure-volume curves for the treatment groups were different from control values (p < 0.01) but were not different from each other.

Intraallantoic Endotoxin

Based on the Na⁺ and Cl concentrations in the fluid aspirated before endotoxin injection, three animals received 100 mg and one animal received 20 mg endotoxin injections into the allantoic fluid. These four animals were delivered and studied 7 d after fetal treatment. There were no indications of induced lung maturation based on physiological responses (compliance, 0.16 ± 0.02 ml/cm H₂O/kg and lung volume at 40 cm H₂O, 12.6 ± 3.2 ml/kg) or by measurement of Sat PC (alveolar Sat PC, 0.46 ± 0.11 µmol/kg and total lung Sat PC, 51 ± 4 µmol/kg). Cord plasma cortisol levels were 0.5 ± 0.1 µg/dl. Therefore, the allantoic and amniotic membranes defined compartments resulting in different fetal responses.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Antenatal intraamniotic endotoxin induced early lung maturation without increasing fetal plasma cortisol to levels normally associated with lung maturation responses or labor (1, 2). The lung maturational effects first occurred 2 d after treatment as evaluated by increases in Sat PC, and physiological measurements demonstrated improved lung function after a treatment-to-delivery interval of 4 d. The striking maturational effects occurred equivalently over a dose range from 1 to 100 mg intraamniotic endotoxin, and a second dose did not further augment the response. The maturational effect was associated with an inflammatory cellular response in the lung tissue and airspaces except at the low dose of 1 mg endotoxin when evaluated at 7 d.

Although multiple hormones and growth factors can influence a variety of indicators of lung maturation in vitro, glucocorticoids have been the focus of studies of lung maturation in vivo, primarily because antenatal glucocorticoids reduce the incidence of respiratory distress syndrome in preterm infants (7). Fetal stress is assumed to increase fetal cortisol, which will induce early lung maturation. A recent experimental test of this concept was the demonstration by Gagnon and coworkers (24) that placental insufficiency caused by repeated embolization in sheep resulted in very elevated fetal cortisol levels and induction of mRNA for the surfactant proteins. Proinflammatory stimuli resulting from antenatal amnionitis or the experimental administration of the proinflammatory cytokine IL-1 also promote early lung maturation (11, 13). The decreased incidence of respiratory distress syndrome following amnionitis was associated with increased cord cortisol levels in infants (12). Although cord plasma cortisol levels were not elevated 48 h after intraamniotic IL-1 in one experiment in sheep (14), a likely explanation of the induced lung maturation was that proinflammatory stimuli induced fetal stress accompanied by increased fetal cortisol (2). The endogenous cortisol then would induce the early lung maturation. In our experiments, cord plasma cortisol values did not increase from 5 h to 15 d after intraamniotic endotoxin. Therefore, there was a disassociation between the early lung maturation resulting from intraamniotic endotoxin and endogenous glucocorticoids.

The characteristics of the physiological responses of the preterm lung support the idea that intraamniotic endotoxin stimulated lung maturation by a mechanism distinct from glucocorticoids. Maternal or fetal treatments with betamethasone rapidly improved postnatal lung function 15 to 24 h after preterm birth (25). This rapid effect of glucocorticoids was associated with a structural change in the lung characterized by parenchymal thinning and an increase in lung gas volume without measurable increases in alveolar surfactant pool sizes for about 7 d (26). In contrast, alveolar and total lung Sat PC increased by Day 2, and improved lung function was first detected 4 d after intraamniotic endotoxin. We found previously that maternal glucocorticoids caused fetal growth restriction and very modest lung maturation when given at 104 d gestation (29). In contrast, intraamniotic endotoxin did not cause fetal growth restriction even at a dose of 100 mg or with two doses of 20 mg, and lung maturation was equivalent when animals were treated at 110 or 118 d gestation.

In humans, amnionitis and elevated proinflammatory cytokines are associated with the subsequent development of bronchopulmonary dysplasia and cerebral palsy after preterm birth (30, 31). Infants also may have a systemic inflammatory response that can be detected by elevated IL-6 levels in umbilical arterial blood (32). Elevated tumor necrosis factor- levels in amniotic fluid predict subsequent lung injury in preterms (33), and proinflammatory cytokine production and release by the lungs occur in lung injury models using mature lungs (34, 35). The concept is that proinflammatory cytokines either initiate and/or augment lung injury (36). The counterintuitive result in developing lungs is that intraamniotic IL-1, and endotoxin can promote early lung maturation without apparent injury. With clinical amnionitis, the proinflammatory stimulus will be continuous and cumulative, while the fetal exposure in these experiments was given once (or twice). The signal to the lungs to mature is unlikely to be endotoxin itself because a dose of 0.05 mg endotoxin given to the fetus by intramuscular injection was lethal (15) and there was no differential response to intraamniotic endotoxin over a dose range of 1 to 100 mg. This dose-response relationship suggests that the intraamniotic endotoxin triggers a secondary response that results in lung maturation, possibly as a result of the amnionitis induced by the endotoxin. The response is specific for the amniotic cavity because no effect on postnatal lung function or surfactant was detected after intraallantoic endotoxin injection.

These studies demonstrate remarkable early lung maturation that occurred independently of cortisol. This endotoxin response indicates that there are other potent lung maturation factors that remain to be identified. A reasonable hypothesis is that endotoxin induces proinflammatory cytokines that can act as maturational agents, the prototype being IL-1 (13). However, other hormones and growth factors could be acting singly or in concert to promote early lung maturation.