INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: chorioamnionitis; inflammation; pulmonary surfactant; respiratory distress syndrome; sheep

Prenatal inflammation, as diagnosed by chorioamnionitis or elevated cytokines in amniotic fluid, is predictive of increased risk of injury to the lungs (bronchopulmonary dysplasia) and brain (periventricular leukomalacia and cerebral palsy) (1-3). However, both chorioamnionitis and fetal colonization with Ureaplasma urealyticum have been shown to benefit the very preterm infant by increasing survival and decreasing the incidence of respiratory distress syndrome (4-6). Chronic and generally asymptomatic histologic chorioamnionitis is common in preterm birth before 30 weeks of gestation (7). Cultures of chorioamnion frequently are positive for organisms that are generally considered to be of low pathogenicity and proinflammatory cytokines and granulocytes are elevated in the amniotic fluid in such cases (8). The effect of timing of the initial inflammatory exposure or its duration in the human is not well understood. The associations between inflammation during early gestation and maturation or injury of the fetal lung have not been explored in experimental models. Information from experimental models is limited in sheep to the effects of intra-amniotic endotoxin or interleukin-1 (IL-1) on lung development after 100 days of gestation (9-11). In the rabbit, IL-1 induces lung development in early-gestation lung explants and delays development or injures lung explants from late-gestation lungs (12). These in vitro results suggest that the gestational timing of inflammatory exposure may result in different effects on the developing lung.

In the fetal sheep, intra-amniotic endotoxin given after 100 days induces chorioamnionitis and results in striking improvements in lung mechanics and increases in surfactant, without altering cord blood cortisol levels (10, 13, 14). However, this maturational effect is accompanied by an inhibition of alveolarization similar to that caused by maternal glucocorticoids (15). An understanding of the effects of early gestational fetal exposures to inflammation is essential to developing strategies to avoid the adverse outcomes of cerebral palsy and bronchopulmonary dysplasia that have been associated with chorioamnionitis in humans. We hypothesized that the effects of intrauterine inflammation on the preterm lung would differ depending on the gestational timing and duration of exposure to intra-amniotic endotoxin. We exposed fetal sheep to a single dose of endotoxin during the pseudoglandular stage of lung development (60 days of gestation), the transition between the pseudoglandular and canalicular stages (80 days of gestation), and the canalicular stage of lung development (100 days of gestation) (16). We also evaluated the effects of two exposures, at 60 plus 100 days of gestation, and a 28-day continuous exposure beginning at 80 days of gestation. We report the effects of different timing of early gestational intra-amniotic endotoxin on preterm lung function, the pulmonary surfactant system, and lung morphometry after delivery at 125 days of gestation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Relevant ethics committees approved experimental procedures. Singleton-bearing ewes were randomized to endotoxin (Escherichia coli 055:B5, solubilized in saline; Sigma, St. Louis, MO) or control (saline) groups. Endotoxin (1 mg in 1 ml of saline) or saline (1 ml) was injected into the amniotic sac under ultrasound guidance (10) at 60, 80, 100, or 60 plus 100 days or was infused over 4 weeks by an osmotic pump (2ML4; Alzet, Palo Alto, CA) implanted during a brief surgical procedure (halothane anesthesia) at 80 days. Osmotic pumps were filled with approximately 2 ml of endotoxin (10 mg/ml) or saline and infused at 2.6 µl/hour, resulting in a daily endotoxin dose of 0.6 mg.

Caesarean delivery and postnatal procedures were performed at 125 days (17, 18). Amniotic fluid and lung fluid (following intubation after tracheostomy) were collected and umbilical arterial blood samples were withdrawn for measurement of blood gases and pH (Rapidlab 865; Bayer, Tarrytown, NY), and white blood cell counts and plasma cortisol (ICN Biomedical, Costa Mesa, CA). Enzyme-linked immunosorbent assays were used to measure plasma IL-8 and interferon- (IFN-).

Lambs were ventilated for 40 minutes to allow assessment of lung function, including calculation of compliance and ventilatory efficiency index (VEI) (19, 20). Lambs were then deeply anesthetized (pentobarbitone, 30 mg/kg) and exsanguinated, and deflation pressure-volume curves were measured after air inflation of the lungs to a pressure of 40 cm H₂O (15). Lungs were weighed and the left lung was lavaged five times with 4° C normal saline. Total lavage volume was measured and aliquots were saved for later analyses. A weighed piece of the left lower lobe was saved for lipid analyses and samples of the right lower lobe were frozen in N₂. The right upper lobe was inflation fixed with 10% formalin for morphometric analysis (15). Weighed lung tissue was dried to determine the wet-to-dry weight ratio. The fetal thymus-to-body weight ratio was calculated as an indicator of fetal stress.

Cells recovered from amniotic, fetal lung, and alveolar lavage fluids after centrifugation were resuspended in phosphate-buffered saline (PBS) and counted using trypan blue. Differentiation was performed on Cytospin preparations stained with Diff-Quick (Scientific Products, McGraw Park, IN). Total protein in alveolar lavages was measured by the assay of Lowry and co-workers (21). Saturated phosphatidylcholine (Sat-PC) was isolated from alveolar lavage samples and lung tissue by neutral alumina column chromatography (22) and quantified by phosphorus assay (23). Surfactant protein mRNAs were quantified by S1 nuclease protection assays (13). The amount of each surfactant protein in alveolar lavage samples was quantified by Western blot (13) and normalized to body weight with values expressed relative to control.

There were three saline control lambs treated at 60 days, two control lambs treated at 80 days, five control lambs treated at 100 days, and four control lambs that received saline from osmotic pumps. Results were similar for these saline treatment groups and data were combined to provide a single saline-treated control group. Comparisons between control lambs and each endotoxin group were performed by t test. Pressure-volume curves of the control group and each treatment group were compared by repeated measures analysis of variance and the Tukey test.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

There were no fetal deaths or pregnancy losses after intra-amniotic injections or the placement of osmotic pumps. Fetuses appeared well at delivery as indicated by their general appearance and umbilical arterial pH and cortisol measurements (Table 1). Body weights were not changed by endotoxin exposure (Table 1). Lung-to-body weight ratios were increased for the 60-day and 100-day endotoxin groups and decreased for the 80- to 108-day endotoxin group (Table 2). Wet-to-dry lung weight ratios were not different between groups (range of group means, 8.6-9.2). Therefore differences in lung-to-body weight ratios were not the result of edema for the 60-day and 100-day groups and were not caused by a decrease in fluid content after the 80- to 108-day endotoxin exposure.

Lung Function

Endotoxin treatments resulted in improvements in postnatal lung function, but differential effects were observed between groups (Table 1 and Figure 1). Ventilatory pressures (peak inspiratory pressure [PIP]  peak end-expiratory pressure [PEEP]) were lower in the 80-day, 100-day, 60- plus 100-day, and 80- to 108-day endotoxin groups than in the control lambs (control lambs versus 60-day group, p = 0.066). Improvements in arterial pH and Pa_CO2 at 40 minutes of age were observed in the 80-day and 80- to 108-day endotoxin groups. Pa_O2 was higher for only the 80- to 108-day endotoxin group. Compliance was increased in all endotoxin-exposed groups and VEI was increased in the 80-day, 60- plus 100-day, and 80- to 108-day endotoxin groups (Figure 1). All endotoxin groups had higher lung distensibility as indicated by pressure-volume curves, but the magnitude of improvement varied between groups (Figure 1). Lung volume at 40 cm H₂O was greatest in the 80- to 108-day endotoxin group. The residual lung volumes on deflation to a pressure of 0 cm H₂O were significantly different from control lambs only for the 80- to 108-day group.

View larger version (34K): [in this window] [in a new window]   Figure 1.   Lung function after intra-amniotic endotoxin treatment. Compliance (A) was significantly increased in all endotoxin treatment groups and VEI (B) was significantly improved in the 80-day, 60- plus 100-day, and 80- to 108-day endotoxin groups. Lung gas volumes (C and D) at airway pressures of 5, 10, 15, 20, and 40 cm H2O were significantly higher in all but the 80-day endotoxin group (volumes at 15, 20, and 40 cm H2O, but not at 5 and 10 cm H2O, were significantly higher in this group). Residual gas volumes at 0 cm H2O were significantly greater in the 80- to 108-day endotoxin group. *p < 0.05 versus control.

Surfactant

The amount of Sat-PC in the lungs was increased by the early gestational fetal exposures to endotoxin (Figure 2). Lung tissue Sat-PC at 125 days increased almost two-fold after the 60-day exposure and by more than two-fold after the 100-day exposure. The increases for the 60- plus 100-day and the 80- to 108-day groups were less striking. Alveolar Sat-PC pool sizes were low in the control group and did not increase for the 60-day and 80-day groups. The amount of Sat-PC in the alveolar washes for the 100-day, 60- plus 100 day, and 80- to 108-day groups was as much as 10-fold higher than in the control group, but with high variability. The early gestational exposures at 60 and 80 days resulted in the anomalous pattern of large increases in lung tissue Sat-PC without increased alveolar Sat-PC.

View larger version (35K): [in this window] [in a new window]   Figure 2.   Saturated phosphatidylcholine content of lung tissue and alveolar lavage fluid after intra-amniotic endotoxin treatment. Lung tissue Sat-PC (A) was significantly increased in all but the 60- plus 100-day endotoxin group, which showed a large degree of variability in lung and alveolar lavage Sat-PC content. Sat-PC was not increased in alveolar lavage fluid (B) after endotoxin treatment at 60 or 80 days, despite large increases in lung tissue Sat-PC. *p < 0.05 versus control.

The steady state mRNA levels for surfactant protein A (SP-A), SP-B, and SP-C were increased at 125 days of gestation by the single endotoxin exposure at 60 days (Figure 3). The most consistent result was an increase in SP-C mRNA in all groups except for the 80- to 108-day group. There was no large or consistent effect on the mRNA for SP-D after the early gestational exposures to endotoxin. The amounts of the surfactant proteins in alveolar lavages were elevated by the endotoxin exposures but there was high variability between measurements within treatment groups. The control samples had little of any of the proteins, whereas some samples from endotoxin-exposed animals had manyfold increases. A consistent finding was an increase in the proportion of SP-B recovered by alveolar lavage that was fully processed. Given the increases in total SP-B in some samples and the increase in the percent mature SP-B, the increase in mature SP-B was large.

View larger version (32K): [in this window] [in a new window]   Figure 3.   Expression of surfactant protein mRNA and alveolar surfactant protein pool sizes at 125 days of gestation, after intra-amniotic endotoxin treatment. All values are expressed relative to the mean value for the control group. Differential effects of timing of endotoxin treatment on SP mRNA and pool size were observed. Most noticeably, early gestational endotoxin treatments increased SP-C mRNA and resulted in large increases in the proportion of SP-B that was in its mature form. *p < 0.05 versus control.

Indicators of Inflammation/Injury

White blood cell numbers in umbilical arterial blood were not changed greatly by endotoxin exposure (Table 1). Fetal circulating white blood cell counts were elevated in only the 80- to 108-day endotoxin group. Circulating neutrophil numbers were elevated in the 80- to 108-day endotoxin group and lymphocyte numbers were significantly higher in 80-day and 80- to 108-day endotoxin groups than in control lambs (Table 1). Plasma cortisol was not increased by endotoxin treatment, nor was the level of IFN- in umbilical arterial blood. IL-8 concentrations in the 80- and 100-day endotoxin groups were lower than in the control group, but levels were increased in the 60- plus 100-day endotoxin group (Table 1). The thymus-to-body weight ratio was 1.65 ± 0.11 g/kg for control lambs and was not different for any of the endotoxin-exposed groups, with mean values ranging from 1.60 to 1.85 g/kg. The magnitude of the changes in systemic indicators of inflammation was low, demonstrating no consistent or large effects of the endotoxin that persisted until the time of delivery at 125 days of gestation.

Inflammatory cells (monocytes, lymphocytes, and granulocytes) in amniotic fluid were significantly greater in the 80-day, 100-day, and 80- to 108-day endotoxin groups than in control lambs (Figure 4). Monocyte numbers were elevated in the 60- plus 100-day endotoxin group, but numbers of other inflammatory cells were not increased above control levels in this group. There were increased numbers of inflammatory cells in fetal lung fluid from the 80-day, 100-day, 60- plus 100-day, and 80- to 108-day endotoxin groups. Granulocyte numbers were increased in alveolar lavage samples from the 100-day and 80- to 108-day endotoxin groups, and monocyte numbers were increased in the 80- to 108-day endotoxin group. The animals exposed to endotoxin at 60 plus 100 days had cell counts in all compartments that tended to be lower than those of the 100-day single exposure group. The amounts of protein in alveolar washes were not different for any of the groups (data not shown), indicating an absence of epithelial injury.

View larger version (20K): [in this window] [in a new window]   Figure 4.   Inflammatory cells in amniotic fluid (A), fetal lung fluid (B), and alveolar lavage fluid (C) collected at 125 days of gestation from control lambs and fetuses exposed to intra-amniotic endotoxin at 60 days, 80 days, 100 days, 60 plus 100 days, and from 80 to 108 days of gestation. *p < 0.05 versus control.

Lung Morphometry

Photomicrographs of lung histology from representative sections of control and endotoxin-treated animals are shown in Figure 5. The total volume of the fixed upper lobe of the right lung was reduced after 80- to 108-day endotoxin treatment, reflecting changes in lung weight relative to body weight in this group (Table 2). Although the parenchymal volume fraction was not affected in any of the endotoxin treatment groups, the 80- to 108-day endotoxin group had a reduced total parenchymal volume resulting from smaller lung size (Table 2). The alveolar volume fraction was reduced in the 80- to 108-day endotoxin group, resulting in a 23% reduction in the numerical density of alveoli and a reduction in total alveolar number of approximately 40% (Figure 6). Alveolar volume fraction was increased by 20% in the 60-day endotoxin group but this did not translate into an increase in the density, size, or total number of alveoli. Alveolar wall thickness was reduced in the 80- to 108-day endotoxin group. In contrast, the alveolar walls were thicker in the 60-day and 80-day endotoxin groups (Figure 6).

View larger version (132K): [in this window] [in a new window]   Figure 5.   Representative photomicrographs of lung structure in control and endotoxin-treated (60 days, 80 days, 100 days, 60- and 100 days, and 80 to 108 days) fetal lambs at 125 days. Sections (5 µm) were stained with hematoxylin and eosin. Scale bar: 50 µm.

View larger version (33K): [in this window] [in a new window]   Figure 6.   Effects of endotoxin treatment on lung structure. Intra-amniotic infusion of endotoxin from 80 to 108 days reduced alveolar surface area (A) and number (B) and caused thinning of the alveolar wall (C). Single intra-amniotic injections of endotoxin at 60 or 80 days increased alveolar wall thickness. *p < 0.05 versus control.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

These experiments demonstrate that a single proinflammatory exposure to the fetus via the amniotic fluid, as early as 65 days before preterm delivery, alters lung function and structure and the surfactant system. Lung growth was altered depending on the timing and/or duration of endotoxin exposure. All endotoxin exposures after 80 days of gestation resulted in subtle but distinct residual indicators of inflammation. A second exposure at 100 days after a 60-day exposure left fewer traces of the inflammatory response than did a single 100-day exposure. In general, the prolonged 28-day exposure from 80 days of gestation, with a daily endotoxin dose of 0.6 mg, resulted in smaller lungs with fewer alveoli and better lung function but no greater effect on the surfactant system than did a single dose of 1 mg of endotoxin at 100 days. The residual indicators of inflammation 17 days after the end of the 28-day continuous exposure were not strikingly different after the single-dose exposures at 80 or 100 days. We previously found similar lung maturation and inflammatory responses to 1, 4, 20, or 100 mg of endotoxin given at 118 days before preterm delivery at 125 days, and little differential effect with 2 doses of 20 mg of endotoxin separated by 8 or 4 days before preterm delivery (10). We used the dose of 1 mg of endotoxin in our present study because it was the lowest dose to give consistent effects on the fetal lung (10). We recognize that the chosen endotoxin dose, relative to fetal size, will be different according to gestational age at treatment. However, our intent was not to compare equivalent endotoxin doses per kilogram of fetal weight or per amniotic fluid volume. These were exploratory studies to determine whether fetal responses occurred when endotoxin was administered at early gestational ages. The fetus does respond to low doses of endotoxin early in gestation. Our present findings and those from our previous experiments (10, 11) demonstrate that the fetus also can cope with much larger and continuous exposures within the amniotic fluid, in both mid and late gestation, without a progressive or overwhelming inflammatory response causing fetal death or preterm labor. None of the treatment groups in our present study had persistent indicators of systemic inflammation.

Our studies show that intra-amniotic endotoxin for intervals from 4 days (10) to 65 days before preterm delivery at 125 days alters lung function in a manner that, from the clinical perspective, indicates induced lung maturation. This is a surprising result from the perspective of lung structure at the time of the endotoxin exposure. The 60-day fetal sheep lung is in the midpseudoglandular stage of development with no distal airspace development and no pulmonary microvasculature (16). The epithelium is lined with uniformly immature cuboidal cells. In human and mouse lungs, the cells at the tips of the branching airways in the pseudoglandular lung express SP-C mRNA but not mature SP-C, and no other surfactant proteins are detected (24, 25). Nevertheless, the intra-amniotic endotoxin programmed a response that resulted in elevated mRNAs for SP-A, SP-B, and SP-C 65 days later. The early gestational endotoxin exposure also induced the processing of SP-B to its mature form, an effect we noted previously following endotoxin exposure after 110 days (13). However, the response was not a normal maturational effect because the large increase in lung tissue Sat-PC was not accompanied by corresponding increases in alveolar Sat-PC and the surfactant proteins. In contrast, glucocorticoid or endotoxin exposure later in gestation (10, 19, 26) causes parallel increases in tissue and alveolar surfactant. This increase in lung tissue surfactant without effective secretion is similar to the surfactant abnormalities in ventilated preterm baboons developing bronchopulmonary dysplasia (27, 28) and demonstrates that these changes may exist at the time of birth, before long periods of ventilation. Therefore, the surfactant changes observed after early gestational endotoxin might indicate an injury response of the surfactant system to inflammation.

The sheep lung begins to alveolarize after 115 days of gestation (16). Alveolarization of the saccular lung can be disrupted by glucocorticoids, oxygen, and mechanical ventilation (29, 30). In transgenic mice, overexpression of several proinflammatory cytokines also disrupts alveolar septation (31). We found previously that intra-amniotic endotoxin given at 118 days of gestation caused transient proinflammatory cytokine expression in the fetal lungs and caused a 30% decrease in alveolar number at 125 days of gestation (15). The 80- to 108-day exposure that occurred before the initiation of alveolarization caused a similar decrease in alveolar number and resulted in smaller lungs. This result is of interest because many preterm infants are exposed to chronic chorioamnionitis (7) and such infants are known to be at risk of bronchopulmonary dysplasia (BPD) (3). The single endotoxin exposures at 60 days and at 80 days resulted in no alterations in alveolarization and a subtle increase in alveolar wall thickness only. The induced maturation of the fetal lung caused by glucocorticoids primarily alters lung structure by thinning the alveolar wall and inhibiting alveolarization with inconsistent and delayed effects on the surfactant system (32, 33). The single-dose early gestational exposures to endotoxin preferentially induce surfactant with minimal effects on lung structure at the level of our evaluations.

At delivery, there were no indications that an inflammatory response to the 60-day endotoxin exposure had occurred. Inflammatory cells in the amniotic fluid and lungs were not increased. The nature of the signaling of intra-amniotic endotoxin to the fetal lung is unknown. At 118 days endotoxin causes chorioamnionitis with proinflammatory cytokine expression in the chorioamnion and lung within five hours but without cytokine expression in other fetal tissues. Activated granulocytes enter the lungs within five hours and accumulate over approximately three days. Subsequently the inflammatory cells remain for many days but are no longer activated (14). Proinflammatory cells also remain in the amniotic fluid and express proinflammatory cytokine mRNA for many days. In our present study, all endotoxin exposures at ages greater than 60 days resulted in increases in granulocytes and monocytes in amniotic, fetal lung, and alveolar lavage fluids. Therefore, we infer that the 80-day, 100-day, and 80- to 108-day exposures induced initial inflammatory responses. The fact that inflammatory cells were increased 45 days after the exposure indicates an inability of the fetus and its membranes to fully clear inflammation. However, the fetus appears able to control the inflammation because two doses or a higher dose given over 28 days did not induce much larger inflammatory responses. Our data suggest the inflammatory response to endotoxin at 100 days might be blunted by prior treatment at 60 days because fewer inflammatory cells were observed in the 60- plus 100-day group than in the 100-day group. We are not aware of previous studies of the fetal response to inflammation at such early gestational ages. However, the fetal lung can respond to other stimuli at early gestations. Bunton and Plopper exposed fetal monkeys to high-dose triamcinolone at 63 to 65 days of gestation and found an inhibition of alveolar development close to term at 150 days (34).

Contrary to our hypothesis, there were no consistent differences in responses to the single-endotoxin exposure across gestation. We interpret this to indicate that endotoxin is triggering fundamental signaling pathways that alter the sequence of lung development. The nature of those signals is obscure but may be initiated by inflammatory mediators. IL-1 and IL-1 are proinflammatory cytokines that can induce lung maturation when given by intra-amniotic injection (which causes chorioamnionitis in sheep) or by direct exposure to lung explants from preterm rabbits (12, 35, 36). Therefore, there is a precedent for signaling by inflammation, although this may be indirect. Umbilical arterial cortisol, IL-8, and IFN- concentrations were not increased in our present study (with the exception of IL-8 in the 60- plus 100-day endotoxin group), and the presence of normal thymus-to-body weight ratios suggests no ongoing fetal stress. In contrast, chorioamnionitis in humans is accompanied by thymic involution (37) and elevated cord cortisol levels have been reported with chorioamnionitis terminating in preterm labor (38).

The clinical implications of our studies have two aspects: the early fetal response to an inflammatory stimulus and the lung responses. We found that the fetus and its membranes did not extinguish the indicators of inflammation even 45 days after intra-amniotic endotoxin exposure. Therefore, residual inflammatory cells in amniotic fluid may not indicate recent or ongoing inflammation. The characteristics of the inflammatory responses of early gestation fetuses remain to be defined. A fetal lung response, interpreted as either maturation or injury, also may be a reflection of a distant event that is remote from antenatal corticosteroid treatment, preterm labor, or delivery. The combination of increased tissue pools of surfactant lipids and decreased alveolar numbers that parallel findings in preterm baboons developing BPD (28, 30) suggests that the chronic 28-day endotoxin exposure is inducing progressive lung injury. Studies are now required to investigate how a lung, with development that has been altered by an early inflammatory stimulus, will respond to the subsequent events and treatments associated with preterm birth and postnatal management.