INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: respiratory distress syndrome; bronchopulmonary dysplasia; neutrophil recruitment; surfactant; cytokines

Prenatal exposure to inflammation/infection has been associated with numerous adverse outcomes that include preterm labor and delivery, increased risks of bronchopulmonary dysplasia (BPD), cerebral palsy, and neonatal sepsis (1). However, histological chorioamnionitis and fetal colonization with organisms such as Ureaplasma urealyticum also have been associated with a decreased incidence of respiratory distress syndrome (RDS) (4, 5), and in one epidemiological study chorioamnionitis predicted increased survival of infants born at gestations < 26 wk (6). There is very little information about how the fetus responds to intraamniotic infection/inflammation.

We have used intraamniotic injections of endotoxin to characterize the chorioamniotic and fetal lung responses to an inflammatory stimulus (7). Using a dose of 20 mg endotoxin, we found that inflammatory cells were recruited to the chorioamnion and the chorioamnion produced interleukin (IL)-1, IL-6, and IL-8 mRNA within 5 h (10). The expression of the mRNAs for the proinflammatory cytokines returned to control values within 2 d but inflammatory cells were increased in the chorioamnion for 25 d after the endotoxin exposure (10). The lung tissue expressed the same cytokine mRNAs within 5 h of the intraamniotic endotoxin injection but expression was higher at later time points and inflammatory cells also persisted in the lungs (10). The mRNAs for the surfactant proteins increased within 24 h of endotoxin exposure and remained increased for 15 d, although lung mechanics did not improve until 4 to 7 d after the endotoxin exposure (9). We found a minimal systemic response to the intraamniotic endotoxin. The cord blood granulocytes initially decreased and then increased moderately (10), fetal cortisol levels did not increase (11), and there was no increased cytokine expression in liver, gut, or placental tissue (10). In this study we asked whether systemic or local inflammatory changes are required for subsequent lung maturation by evaluating various endotoxin doses ranging from 0.1 mg to 10 mg, and by evaluating both the systemic and local inflammatory changes at different times after intraamniotic endotoxin injection.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animals, Endotoxin, Delivery, and Postnatal Assessment

The animal component of the studies was performed in Western Australia, as approved by the animal care and use committees from Children's Hospital Medical Center, Cincinnati, OH, the Ethics Committee of the Western Australian Department of Agriculture, and the University of Western Australia. Date bred Merino ewes with a singleton fetus were randomized for the two protocols (Table 1). To evaluate the response to endotoxin dose, animals were randomized to saline, 0.1, 1, 4, or 10 mg Escherichia coli 055:B5 endotoxin (Sigma, St. Louis, MO), given 7 d before preterm delivery at 125 d gestation. To evaluate the inflammatory responses by time interval from endotoxin exposure other animals were randomized to receive 4 mg endotoxin at 5 h, 24 h, or 72 h before delivery. The endotoxin was solubilized in saline and filtered. The intraamniotic injection was given into the amniotic cavity with ultrasound guidance (8, 11). Preterm lambs were delivered at 125 d gestational age by cesarean section (12). For evaluating lung function 7 d after endotoxin exposure, the preterm lambs were ventilated for 40 min with time-cycled, pressure-limited infant ventilators (11). Peak inspiratory pressures were limited to 40 cm H₂O. A pressure transducer (model 8507C-2; Endevco, San Juan Capistrano, CA) and pneumotachograph (model 35-597; Hans Rudolph, Kansas City, MO) were placed between the tracheostomy tube and the ventilator to measure tracheal pressure and flow, respectively. Volume was obtained by integrating flow (12).

Measurements on Lungs, Spleen, Plasma, and Cells from Amniotic Fluid and Alveolar Lavage Fluid

Lungs were removed and processed as described (13). The deflation limb of the pressure-volume curve was measured. Saturated phosphatidylcholine (Sat PC) (14, 15), total protein (13, 16), and surfactant proteins (Western blot) were measured (9). The mRNA for the surfactant proteins and for selected cytokines were quantified using sheep-specific probes (9, 10, 17). To reduce the high viscosity, amniotic fluid was incubated at 37° C with N-acetyl-L-cysteine, neuraminidase, and hyaluronidase. Cells from amniotic fluid and alveolar lavage fluid were isolated by centrifugation.

The right upper lobe was inflation fixed at 30 cm H₂O. After hematoxylin and eosin staining the tissue was assessed for inflammatory changes (8). Hydrogen peroxide was measured in cells recovered from alveolar lavage fluid of the left lung. Flow cytometry was used to measure CD11b (M subunit of integrin CR3) and CD44 (proteoglycan link protein) and the percentage of apoptotic cells was determined by annexin V and propium iodide staining (20). Enzyme-linked immunosorbent assays (ELISA) were developed for IL-6 and IL-8 with antibodies from Chemicon (Temecula, CA) and recombinant proteins as standards from CSIRO (Parkville, Australia). Interferon-gamma (IFN-) was measured by ELISA (CSL, Parkville, Australia; recombinant protein was provided by Dr. S. L. Jones, CSL).

Data Analysis

Results are given as means ± SEM. Inflammatory scores are given as medians with interquartile, 5th and 95th percentile ranges. Comparisons between endotoxin-treated and untreated lambs were by analysis of variance with Student-Newman-Keuls tests used for post-hoc analyses. Tissue scores were compared using the Mann-Whitney nonparametric tests. Significance was accepted at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Inflammation and Lung Maturation

There were no differences between the groups in body weights or pH in cord blood (Table 1). Lung function was improved after intraamniotic injection of 4 mg and 10 mg endotoxin given 7 d prior to delivery (Figure 1). Lung gas volumes measured by deflation pressure-volume curves were increased for the 4 and 10 mg dose groups. At 40 cm H₂O pressure, volumes were increased from a mean value of 11 ml/kg in control animals to 32-38 ml/kg after endotoxin exposure of 4 mg and 10 mg (Figure 1A). The compliances also were about 2-fold higher after 4 mg or 10 mg endotoxin than in control lambs (Figure 1B). The alveolar surfactant pool size was significantly increased from 0.5 µmol/kg (control) to values over 4 µmol/kg after 4 mg and 10 mg endotoxin (Figure 1C). The mRNAs for surfactant proteins A and C remained elevated 7 d after injection of 4 mg and 10 mg endotoxin (Figure 1D). The dose of 1 mg endotoxin significantly increased SP-C mRNA but not other surfactant proteins.

View larger version (30K): [in this window] [in a new window]   Figure 1.   Lung maturation and function. (A) Pressure-volume curves. After 40 min of ventilation the deflation limbs of the pressure volume curves were improved after 4 and 10 mg endotoxin. (B) Compliance after 40 min ventilation. Compliance increased after 4 mg and 10 mg endotoxin by about 2-fold. (C ) Sat PC in alveolar lavage fluid increased after 4 mg and 10 mg endotoxin. (D) mRNAs for the surfactant proteins SP-A, B, and C. The mRNAs were increased 7 d after instillation of 10 mg endotoxin, and SP-A and SP-C mRNAs were increased after 4 mg endotoxin, SP-C mRNA was increased after 1 mg endotoxin (*p < 0.05 versus control).

Indicators of Inflammation with Endotoxin Doses

The inflammatory scores in the fetal lung were increased 7 d after the intraamniotic injection of 1, 4, or 10 mg endotoxin in a dose-dependent manner (Figure 2A). The number of inflammatory cells recovered from the alveolar lavage also increased with dose (Figure 2B). The number of granulocytes and monocytes increased significantly after the 0.1 mg and higher doses of endotoxin. Therefore, inflammatory scores and inflammatory cells in alveolar lavages increased for doses of endotoxin that did not induce lung maturation. The increases were dose dependent. After 1, 4, and 10 mg endotoxin there was an increase in inflammatory cells in chorioamniotic membranes (Figure 3A). In all endotoxin-treated animals there were increased numbers of monocytes and lymphocytes in the amniotic fluid (Figure 3B). After doses of 4 mg and 10 mg there was an influx of neutrophils. The absolute number of cells increased in a dose-dependent fashion.

View larger version (24K): [in this window] [in a new window]   Figure 2.   Inflammation in the fetal lungs. (A) Inflammatory scores in the fetal lungs (airspace + tissue). Plots show median, 25th, and 75th percentile ranges as boxes and 5th to 95th percentile ranges as error bars. The numbers of inflammatory cells were ranked from few inflammatory cells (score = 1), moderate cell infiltration (score = 2), to extensive influx of cells (score = 3). Endotoxin doses of 1 mg, 4 mg, and 10 mg increased the inflammation scores. (B) Number of cells recovered from alveolar lavage. All doses of endotoxin increased monocytes and neutrophils in alveolar lavage. The number of lymphocytes was increased after doses of 4 mg and 10 mg (*p < 0.05 versus control).

View larger version (25K): [in this window] [in a new window]   Figure 3.   Inflammatory cells in fetal membranes and amniotic fluid. (A) Plots show median, 25th, and 75th percentile ranges as boxes and 5th to 95th percentile ranges as error bars. Membranes of control animals had few inflammatory cells and animals treated with 1 mg or higher endotoxin doses had an extensive influx of cells (score = 3). Most of the inflammatory cells were neutrophils. (B) Number and type of inflammatory cells in amniotic fluid. The number of inflammatory cells increased after all doses of endotoxin (*p < 0.05 versus control).

The white blood cells increased in cord blood 7 d after the 4 mg and 10 mg endotoxin doses (Figure 4A). The neutrophils were the cell type that accounted for most of the increase in cord blood white cells, and the neutrophils were also increased in the animals treated with a dose of 1 mg after 7 d (Figure 4B). The increases in white blood cells and neutrophils were dose dependent. The number of platelets increased in all groups after the 7 day exposure interval, including in the animals that received 0.1 mg endotoxin (Figure 4C).

View larger version (24K): [in this window] [in a new window]   Figure 4.   Effect of intraamniotic endotoxin on white blood cells and platelets in cord blood. (A) Doses of 4 mg and 10 mg endotoxin caused an increase in white blood cells after 7 d. (B) The number of circulating neutrophils increased in a dose-dependent manner after 1 mg and higher doses of endotoxin. (C ) The number of platelets increased in every group of animals treated with endotoxin. (D) The number of white blood cells decreased after 24 h and increased at 7 d. (E ) The circulating neutrophils increased at 7 d. (F ) The number of platelets was elevated only 7 d after endotoxin (*p < 0.05 versus control).

Time Interval for Inflammatory Response

For the animals studied at different time intervals after 4 mg endotoxin the number of total white blood cells in cord blood decreased at 24 h because the lymphocytes decreased and subsequently increased at 7 d due to increased granulocytes (Figure 4D and 4E). The increase in platelets occurred only 7 d after the 4 mg dose of endotoxin (Figure 4F). The increases in cord blood white cells and platelets were late events after endotoxin exposure.

In contrast, inflammatory cytokine mRNAs for IL-1, IL-6, IL-8, and tumor necrosis factor- (TNF-) increased in the chorioamnion within 5 h after intraamniotic endotoxin (Figure 5A). The 3- to 7-fold increases were the highest at 5 h. The increase in expression of TNF- mRNA was not large at any time, and the expression of IL-10 mRNA in chorioamnion did not change at any time. An increase in inflammatory cells in the amniotic fluid occured by 5 h after the intraamniotic endotoxin and was highest at 72 h (Figure 5C). The cells recovered from the amniotic fluid had a different time pattern of cytokine mRNA expression than did the chorioamnion (Figure 5B). The expression of IL-1 mRNA increased at 24 h and remained elevated at 7 d, IL-8 mRNA was increased at 72 h and 7 d, and IL-6 mRNA did not increase.

View larger version (33K): [in this window] [in a new window]   Figure 5.   Cytokine mRNA in fetal membranes and cells from amniotic fluid. (A) IL-1, IL-6, IL-8, and TNF- mRNAs were increased in the chorioamnion at 5 h and 24 h. IL-10 mRNA was not increased at any time. (B) IL-1 and IL-8 mRNA increased at 24 h, and 72 h and 7 d in cells from amniotic fluid. (C ) The number of monocytes, lymphocytes, and neutrophils in amniotic fluid increased at 5 h and remained increased at all time points (*p < 0.05 versus control).

Inflammatory cells were recruited into the alveolar lavage by 5 h, and peaked at 72 h (Figure 6A). IL-1, IL-8, and TNF- mRNAs were increased in lung tissue only 24 h after endotoxin exposure (Figure 6B). The mRNAs for IL-1 and IL-8 were increased in the alveolar cells recruited to the airways at 24 h and 72 h (Figure 6C). Surfactant protein mRNAs for SP-D was increased 5 h after 4 mg endotoxin and at all later time points (Figure 6D). SP-A, B, and C mRNA were increased 24 h after endotoxin exposure. Activation of cells recruited to the airways was assessed by measuring the expression of the adhesion molecules (CD11b and CD44) (Figure 7, A and B). The percentage of positive cells increased at 5 h and was highest at 72 h for both cell surface markers. The numbers of apoptotic and necrotic cells per body weight in the alveolar lavage also were the highest at 72 h after endotoxin (Figure 7, C and D). The cells recruited to the airways at 5 h were highly active based on the amount of hydrogen peroxide per cell (Figure 7E). However, relative to the number of cells recruited to the airways, the total amount of hydrogen peroxide normalized to body weight was highest at 24 h (Figure 7F).

View larger version (38K): [in this window] [in a new window]   Figure 6.   Cells in alveolar lavage, cytokine mRNAs in cells from alveolar lavage and lung, and surfactant protein mRNAs in lung tissues. (A) The number of monocytes and neutrophils in alveolar lavage increased after 5 h and remained increased to 7 d. (B) At 24 h IL-1, IL-8, and TNF- mRNA were elevated in mRNA from lung tissue. (C ) The IL-1 and IL-8 mRNA from cells recovered from alveolar lavage increased at 24 h and 72 h. (D) Time course of induction of surfactant protein mRNAs increased after 4 mg endotoxin. The SP-D mRNA was induced after 5 h, the mRNAs of all were increased after 24 h (*p < 0.05 versus control).

View larger version (27K): [in this window] [in a new window]   Figure 7.   Markers of activation of cells from alveolar lavage. (A) Percent of CD11b (M subunit of integrin CR3) positive cells. The number of CD11b positive cells increased from 5 h to 72 h. (B) The percentage of CD44 (proteoglycan link protein) positive cells also increased from 5 h to 72 h. (C, D) The number of apoptoic cells in the alveolar lavage was determined by annexin V staining and counterstaining with propium iodide to detect necrotic cells. The numbers of apoptotic and necrotic cells were highest at 72 h. (E ) The production of hydrogen peroxide per 106 cells was maximal after 5 h. (F ) The production of hydrogen peroxide per kilogram body weight was highest at 24 h when the total cell number in the alveolar lavage was also most elevated (*p < 0.05 versus control).

Septemic Inflammation

Several variables were measured to evaluate the systemic inflammation. In cord blood plasma IFN- was increased 12-fold at 24 h (Figure 8A). IL-6 was increased significantly at 5 h (Figure 8A). IL-8 was increased from 5 h to 7 d by about 3-fold, and IL-8 was the only cytokine that was elevated at all time points measured. The expression of TNF- mRNA increased 2-fold at 24 h in spleen tissue (Figure 8B). However, the overall cytokine expression pattern indicates minimal cytokine production by the spleen.

View larger version (33K): [in this window] [in a new window]   Figure 8.   Systemic inflammation. (A) Interferon gamma (IFN-) was increased in cord plasma at 24 h. Interleukin-6 was increased at 5 h and interleukin-8 was increased from 5 h to 7 d. (B) The amount of IL-1, IL-6, IL-8, and IL-10 mRNA in the spleen did not increase. The amount of TNF- increased by 2-fold at 24 h only (*p < 0.05 versus control).

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

We have shown that there was no lung maturationas indicated by increased lung gas volume, improved compliance, and increased surfactant Sat PC and increased SP mRNAswithout preceding inflammation in chorioamnion and fetal lung using various doses of endotoxin. On the other hand, mild inflammation was not followed by lung maturation. The lower doses of 0.1 mg and 1 mg endotoxin induced an influx of inflammatory cells into the chorioamnion, the amniotic fluid, lung tissue, and alveolar lavages that persisted 7 d after endotoxin exposure but lung function was not improved. The increases in white blood cells and inflammatory cells in fetal lung and alveolar lavage were dose dependent. Even the lowest dose of endotoxin induced an increase in cells in amnoitic fluid and in platelets indicating a subtle systemic response to the inflammatory stimulus given into the amniotic fluid.

The inflammatory response in the fetal lungs and cord blood increased with endotoxin dose, suggesting that a threshold amount of inflammation was needed to induce the improved lung function. A distinction between local and systemic inflammatory response could not be made after 7 d. We previously found endotoxin doses of 1, 4, 20, or 100 mg caused equivalent maturational exposure without increases in fetal cortisol (11). The variability of response to 1 mg endotoxin may be attributed to the variation in endotoxin preparations. In the fetal sheep model, a dose of 1 or 4 mg given by intraamniotic injection is the minimal stimulus needed to enhance lung function after preterm delivery. There was a clear dose- response relationship between inflammation and the amount of endotoxin. However there were no clearly graded or partial lung maturation responses.

In a previous study we found that 20 mg endotoxin given by intraamniotic injection increased IL-1, IL-6, and IL-8 mRNA in the chorioamnion at 5 h with a parallel response in the lungs beginning at 5 h that increased for about 2 d (10). For this study we used the minimal effective dose of 4 mg endotoxin for inducing lung maturation and addressed the question whether systemic versus local inflammation in the fetus was induced by intraamniotic endotoxin and which compartment was producing proinflammatory cytokines at later time points. The mRNAs for IL-1, IL-6, IL-8, and TNF- were most increased at 5 h in the chorioamnion. Inflammatory cells in the amniotic fluid were increased at 5 h, but the 40-fold increase in granulocytes was at 24-72 h, and the up to 35-fold increases in IL-1 and IL-8 mRNA occurred over the 24 h to 7 day interval, with no increase in IL-6 mRNA. The distinct differences between chorioamnion and cells in amniotic fluid suggest that the inflammation in amniotic fluid was regulated differently from the response of the chorioamnion. The cytokine response continued in amniotic fluid after it was extinguished in the chorioamnion and fetal lungs.

The mRNAs for the proinflammatory cytokines were increased in lung and alveolar lavage by 24 h, but not at 5 h as noted previously with the higher endotoxin dose (10). The first detectable effect of inflammation in the lungs was the large increase in H₂O₂ production by the few cells recruited by 5 h. At the same time the mRNA for SP-D was increased. The large number of cells in alveolar lavages at 72 h expressed the surface adhesion molecules CD11b and CD44, which mediate cell recruitment to sites of inflammation. Although the cell surface antigens indicate that many cells evaluated at 72 h were quickly recruited to the airways, they were no longer producing H₂O₂. We used the measurements of annexin V and counterstaining with propium iodide to evaluate the percentage of apoptotic and necrotic cells in the cell population. Because the total cell numbers increased, we calculated the number of necrotic and apoptotic cells per kilogram body weight, which should best reflect the timing of resolution of inflammation in the lungs. The high numbers of apoptotic and necrotic cells at 72 h indicated resolution of inflammation at 72 h (21). However, complete resolution had not occurred by 7 d, and we found residual inflammatory cells in alveolar lavage as long as 25 d after intraamniotic endotoxin (10).

None of the indicators of inflammation that were detected at 7 days correlated very well with the improved lung function after preterm delivery. Early indicators such as H₂O₂ and cytokine production in the chorioamnion and the lung cannot be evaluated because those indicators have been extinguished before the physiological maturation response occurs. In mature lungs, endotoxin and proinflammatory cytokines initially suppress the mRNAs for surfactant proteins and the surfactant proteins can rebound to higher levels than normal with healing (17, 22). Although detectable, the mRNAs for SP-A and SP-B are very low at 125 d gestation in the fetal sheep lung, and no decreases are seen after intraamniotic endotoxin. SP-C mRNA should be the most sensitive indicator because it is relatively high at 125 d gestation, and it is unchanged 5 h and increased 24 h after intraamniotic endotoxin. Inflammatory cells producing H₂O₂ and cytokines were accumulating in the lungs over 3 d, providing a source for direct lung injury. The wet-to-dry weights did not change from control values nor did the total protein in the alveolar lavages (data not shown). However, the antioxidant enzymes catalyse, glutathione peroxidase, and superoxide dismutase are induced after endotoxin exposure, indicating activation of oxidant-mediated pathways in the fetal lung (23). SP-A and SP-D may function as antioxidants (24). SP-D mRNA was increased within 5 h. A hypothesis for the events that result in lung maturation after intraamniotic endotoxin is that neutrophils producing oxidants cause a primary injury and that the fetal lung responds with lung maturation as a protective response.

We previously found that 20 mg intraamniotic endotoxin did not increase proinflammatory cytokine mRNA in the placenta, fetal liver, or fetal gut. We used the lower dose of 4 mg endotoxin to search for other indicators of a systemic inflammatory response. The TNF- mRNA increased by about 2-fold at 24 h only in the spleen, with no increase in IL-1, IL-6, or IL-8, indicating a very minimal response from the fetal spleen. In preterm lambs that were treated with intratracheal endotoxin the spleen produced high amounts of proinflammatory cytokines within 6 h demonstrating that the preterm spleen can respond to a proinflammatory stimulus (25). Nevertheless, IL-6 and IL-8 were increased in cord plasma over the extended interval from 24 h to 7 d. Interferon- also was increased in plasma 12-fold at 24 h. Using IL-8 as a representative proinflammatory cytokine, we found no increases in mRNA at 72 h or 7 d in chorioamnion, lung, or spleen. However, the mRNA for IL-8 was elevated in cells in amniotic fluid at 72 h and 7 d, suggesting that the amniotic fluid is the source of the increased cytokine in the cord plasma. As with the other components of the inflammatory response, we cannot specifically link the increased plasma cytokines at 5 h and 24 h to the lung maturation response.

The interpretation of this in vivo model relative to clinical settings is complex. The preterm human fetus at risk of delivering before 30 wk gestational age may be continuously exposed to live organisms, their products, and inflammatory mediators and the exposure may be prolonged (1). The kinetics and dynamics of chorioamniotis caused by live organisms are probably quite different from a single dose of endotoxin from gram-negative bacteria. The most common microorganism associated with chorioamnionitis is Ureaplasma urealyticum, which does not have a cell wall but induces IL-1, IL-6, and TNF- in amniotic fluid (26). In the clinical literature cord plasma elevations of IL-6 have been used to diagnose systemic inflammation in the fetus (27, 28). We found that there were subtle indications of systemic inflammation after intraamniotic injection of endotoxin. The increase of IL-6 in cord plasma within 5 h after intraamniotic endotoxin showed that the systemic inflammation was present at the same time as the local inflammation in chorioamnion and fetal lung. However, the animals were not injured in any way that was evident by the endotoxin exposure (8, 9, 11). The antiinflamatory cytokine protein IL-10 was not detectable but sustained increases in IL-8 were in airway samples from infants that developed BPD (29). We found no significant induction of IL-10 mRNA in membranes, in lung, or in spleen. IL-8 was increased in plasma at all time points after intraamniotic endotoxin in the fetal sheep. Increased cytokine proteins in cord blood were reported in clinical studies to be associated with subsequent BPD or cerebral palsy (3, 28) and interferons can cause cerebral palsy when given to infants (30). The pattern of cytokine responses in the fetal sheepelevated IL-8 and IFN- with no induction of the antiinflammatory cytokine IL-10parallels the clinical observations that have been associated with poor lung and neurodevelopmental outcomes (2, 31). In clinical studies oxidative injury has been implicated in the development of BPD after preterm birth (3). In our animal model the response of the fetal lung to intraamniotic endotoxin may be different because the oxidant stress is of short duration.

In summary, we found that intraamniotic doses of 0.1 mg and 1 mg endotoxin induced inflammatory cell recruitment to the chorioamnion and the lungs. However, this inflammation was not accompanied by physiological improvements in lung function. Higher doses of endotoxin sufficient to induce lung maturation were accompanied by systemic inflammation in the fetus within 5 h. Activated cells in the airways producing hydrogen peroxide within 5 h might cause an initial lung injury. Fetal exposure to sufficient inflammation can therefore induce lung maturation, but the precise link between the components and the intensity of the inflamatory response and lung maturation remains elusive.