INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Keywords: lung injury; bronchopulmonary dysplasia; neutrophils; cytokines

Endotoxin administered intratracheally causes an inflammatory lung injury in adult animals. Intratracheal endotoxin results in the rapid recruitment of activated white cells to the lungs and the production of tumor necrosis factor- (TNF-) and other proinflammatory cytokines (1). In adult animals responses to intratracheal endotoxin are localized to the lungs unless the lungs are ventilated with injurious ventilation styles (2). A leak of endotoxin from the airspaces to the systemic circulation and increased plasma cytokines were demonstrated in adult rabbits ventilated with high tidal volumes and no positive end-expiratory pressure (PEEP) (3).

Little is known of the responses of newborn lungs to intratracheal endotoxin. The term or preterm newborn lung may be exposed to endotoxin or other proinflammatory products as a result of chorioamnionitis and microorganisms in amniotic fluid (4). The term lung differs from the adult lung because it is structurally immature and has very few resident macrophages (5). The preterm lung has more structural immaturity and surfactant deficiency and is more easily injured by mechanical ventilation (6). The preterm fetal lung does not have increased permeability to protein relative to the mature lung (9). Therefore, the increased permeability that occurs in ventilated preterm lungs is thought to result from the effects of nonuniform inflation on a structurally immature and surfactant-deficient lung that contains insufficient collagen and elastin (10, 11). The immature lung also has decreased host defense surfactant protein (SP)-A and SP-D and has immature immune systems with depressed second messenger responses making it potentially more susceptible to endotoxin-induced injury (12, 13). Therefore, we hypothesized that intratracheal endotoxin would injure 130-d gestational age (GA) newborn lambs more than near-term lambs. We assessed the effects of mechanical ventilation on the endotoxin-induced inflammation in the immature and near-term lung and the implications for systemic inflammation.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Study Design

We evaluated the effects of intratracheal endotoxin on the lungs and systemic responses in lambs with two protocols. We delivered preterm lambs at 130 d GA and near-term lambs at 141 d GA from Suffolk ewes bred to Dorset rams (term gestation is 146 ± 1 d) (Figure 1). The protocols were approved by the Animal Care and Use Committee of the Children's Hospital Research Foundation. Based on our previous experiences with ventilation-induced lung injury in the preterm at 130 d GA (14), we gave surfactant before initiating ventilation and used a PEEP of 4 cm H₂O and continuous tidal volume monitoring and an initial target Pco₂ of 50 mm Hg to minimize lung injury. We maintained tidal volume (VT) at about 10 ml/kg even if the PCO₂ increased during the 6 h of ventilation. We randomized the 130-d GA lambs to a high dose of 10 mg/kg endotoxin or a lower dose of 0.1 mg/kg endotoxin based on our observations that doses as high as 100 mg of endotoxin given by intraamniotic injection did not harm the sheep fetus and doses less than 1 mg/kg did not consistently induce lung maturation in the fetal sheep (17, 18).

View larger version (24K): [in this window] [in a new window]   Figure 1.   Study design for preterm lambs delivered at 130 d GA and near-term lambs delivered at 141 d GA. Number of animals per group is given in parentheses. All animals were ventilated for 6 h. Interventions are indicated at birth and in the first 30 min of life by arrows.

The near-term animals were randomized to receive a saline placebo or the same high dose of 10 mg/kg of endotracheal endotoxin at birth as given to the preterm lambs. Other term lambs received this dose together with high VT ventilation for 30 min after birth to evaluate if a brief period of increased lung stretch would alter the response to endotoxin. A separate group of near-term lambs was given intravascular endotoxin (5 µg/kg) to characterize the systemic response to the endotoxin. The near-term lambs were not surfactant treated.

Delivery, Endotoxin, and Ventilation

All animals were delivered by Cesarean section, and after exposure of the fetal head and neck, an endotracheal tube was tied into the trachea (16). The fetal lung fluid that could be easily aspirated by syringe was removed, and the lambs were delivered and weighed. The lambs then received either the endotracheal saline placebo or endotoxin before the initiation of mechanical ventilation. The 130-d GA lambs received the endotoxin mixed with 100 mg/kg surfactant (Venticute; Byk Gulden, Konstanz, Germany) containing 2% recombinant human SP-C in phospholipids (19). All animals were ventilated with time-cycled and pressure-limited infant ventilators (Sechrist Industries, Anaheim, CA) with the following initial ventilator setting: fraction of inspired oxygen (FI_O2, 1.0; rate: 40 breaths/min; inspiratory time: 0.6 s; PEEP: 4 cm H₂O; peak inspiratory pressure (PIP) sufficient to yield a target VT of 10 ml/kg but with pressure limited to 35 cm H₂O. VT was monitored continuously (CP-100; Bicore Monitoring Systems, Anaheim, CA). Ventilation pressure and FI_O2 were adjusted to achieve a target Pa_CO2 of 45-55 mm Hg and a Pa_O2 of 150-200 mm Hg. For the 141-d GA lambs that were ventilated to a VT target of 25 ml/kg (high VT group) for 30 min, supplemental CO₂ was added to the ventilatory circuit to normalize the Pa_CO2. A 5 F catheter was advanced into the aorta via an umbilical artery, and a 10 ml/kg transfusion of filtered fetal blood collected from the placenta was administered within 10 min of delivery.

Dynamic compliances were calculated from VT measured with a pneumotachometer that was normalized to body weight and divided by the ventilatory pressure (PIP  PEEP) (14, 19). The ventilatory efficiency index (VEI) was calculated as VEI = 3,800  (respiratory rate × [PIP  PEEP] × Pa_CO2), where 3,800 is a CO₂ production constant ([ml × mm Hg]  [kg × min]) (20). Complete blood cell counts and differentials were performed for cord blood and for peripheral arterial blood at 2, 4, and 6 h. The arterial catheter was used for blood gas, pH, heart rate, and blood pressure recording and was infused with 10% dextrose (100 ml/kg/d). Rectal temperature was monitored and kept at 38-39° C with heating pads and radiant heat. Supplemental ketamine (10 mg/kg intramuscularly) and acepromazine (0.1 mg/kg intramuscularly) suppressed spontaneous breathing. After 6 h each animal was deeply anesthetized with 25 mg/kg pentobarbital intravenously and ventilated briefly with 100% oxygen. The endotracheal tube was clamped for 3 min to permit oxygen absorption.

Lung Gas Volume at 40 cm H₂O and Lung Processing

The thorax was opened, the lungs were inflated with air to 40 cm H₂O pressure for 1 min, and maximal lung volume was recorded (14). Lung tissue of the right lower lobe was frozen in liquid nitrogen for RNA isolation. Bronchoalveolar lavage was performed on the left lung by filling with 0.9% NaCl at 4° C until visually distended, and the lavage was repeated five times (14, 19). The bronchoalveolar lavage fluids (BALF) were pooled and aliquots saved for determination of total protein, cell number, and cell differential. Cell pellets were used for hydrogen peroxide assay and RNA isolation. Total protein in BALF was measured (21).

Alveolar Cells

BALF was centrifuged at 500 × g for 10 min, and the cells in the pellets were counted using trypan blue. Differential cell counts were performed on stained cytospin preparations (Diff-Quick; Scientific Products, McGaw Park, IN). Activation of the cells recruited to the airways was assessed by measuring hydrogen peroxide using an assay based on the oxidation of ferrous iron (Fe²⁺) to ferric iron (Fe³⁺) by hydrogen peroxide under acidic conditions (Bioxytech H₂O₂-560 assay; OXIS International, Portland, OR) (18).

Cytokine Messenger RNA

Total RNA was isolated from the right lower lung lobe, from BALF cells, and from the spleen by guanidinium thiocyanate-phenol-chloroform extraction (22). Spleen tissue was used to evaluate if the endotoxin induced a systemic inflammation. RNase protection assays were performed using RNA transcripts of ovine interleukin (IL)-1, IL-6, IL-8, IL-10, and tumor necrosis factor- (TNF-) that was described previously (18, 22). Ovine ribosomal protein L32 was the reference RNA. Densities of the protected bands were quantified on a phosphor imager using ImageQuant software (Molecular Dynamics Inc., Sunnyvale, CA).

Plasma, Endotoxin, and Cytokines

Endotoxin was quantified in plasma after 6 h ventilation with the Limulus amebocyte lysate assay (Bio Whittaker, Walkersville, MD). Enzyme-linked immunosorbent assays (ELISA) were run for IL-6 and IL-8 with antibodies from Chemicon (Temecula, CA), and recombinant proteins were kindly provided by CSIRO (Parkville, Australia) (16).

Data Analysis

Results are given as means ± SEM. Comparisons between endotoxin groups at each gestational age were by analysis of variance with Student-Newman-Keuls tests used for post hoc analyses. Comparison between 141 d GA and 130 d GA was by two-tailed t tests for control and 10 mg/kg endotoxin groups. Significance was accepted at p < 0.05.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

130-d GA Lambs

The 130-d GA lambs averaged 3.5 ± 0.3 kg and mean weights were similar for the lambs in each group. In comparison with the control lambs, the lambs that received 0.1 or 10 mg/kg endotoxin and were ventilated with similar VT had higher PCO₂ values and decreased oxygenation after 6 h ventilation (Table 1). Compliance and lung gas volumes were not changed by endotoxin exposure. Gas exchange as measured by the VEI and blood pressure decreased in the 10 mg/kg endotoxin group (Figure 2). Heart rate increased for both endotoxin-exposed groups. Protein in the BALF was increased from 69.6 ± 8.4 mg/kg for control lungs to 86.3 ± 5.1 mg/kg for the 10 mg/kg endotoxin group (p < 0.05). Therefore, the endotoxin resulted in systemic effects without a large effect on lung mechanics.

View larger version (48K): [in this window] [in a new window]   Figure 2.   Heart rates (A) and mean blood pressures (B) after 6 h of ventilation. The control lambs at 130 d GA had lower mean blood pressures than the control lambs at 141 d GA. Endotoxin doses are indicated on the horizontal axis. Endotoxin increased heart rate and decreased blood pressure in the 130-d GA lambs. Endotoxin decreased blood pressure in the 141-d GA lambs when given by intravascular injection or when accompanied by high tidal volume (VT) ventilation for the first 30 min of life. *p < 0.05 versus control at each GA; tp < 0.05 verses corresponding group at 141 d GA.

The low and high doses of endotoxin caused inflammation in the lungs of the 130-d GA lambs. Neutrophils and hydrogen peroxide production by cells in the BALF were greatly increased (Figure 3), although there were no consistent changes in lymphocytes or macrophage numbers in the BALF (data not shown). Cytokine mRNA for IL-1, IL-6, and IL-8 increased in both cell pellets and lung tissue, with qualitatively higher mRNA expression for the higher dose of endotoxin (Figure 4).

View larger version (33K): [in this window] [in a new window]   Figure 3.   Neutrophils and production of hydrogen peroxide by cells in BALF. (A) Number of neutrophils in BALF/kg body weight. The number of neutrophils increased after endotoxin administration and high tidal volume (VT) ventilation did not further increase the neutrophils in 141-d gestation lambs. Neutrophils were higher in lambs at 130 d GA than in lambs at 141 d GA given 10 mg/kg endotoxin. (B) Hydrogen peroxide production by cells in BALF increased after intratracheal endotoxin in 130-d and 141-d GA lambs. Cells from lambs at 130 d GA produced more hydrogen peroxide than cells from lambs at 141 d GA. *p < 0.05 versus control; tp < 0.05 versus corresponding group at 141 d GA.

View larger version (27K): [in this window] [in a new window]   Figure 4.   Proinflammatory cytokine mRNA in cells from BALF and in lung tissue. (A) Cytokine mRNA in cells from BALF. Results are shown as fold increase relative to value of 1 for control groups. IL-1, IL-6, and IL-8 mRNA were higher after intratracheal endotoxin in all groups. High tidal volume (VT) ventilation did not further increase the amount of mRNA in lambs at 141 d GA. (B) Cytokine mRNA in lung tissue. The mRNA for IL-1, IL-6, and IL-8 were increased for endotoxin-exposed groups relative to their controls. *p < 0.05 versus control.  IL-1; IL-6;  IL-8.

Endotoxin could be quantified in the plasma of these ventilated preterm lambs (Figure 5). The intratracheal endotoxin at either dose also decreased peripheral neutrophils to very low numbers. Large increases in mRNA for IL-1, IL-6, and IL-8 were detected in the spleens of the lambs exposed to 10 mg/kg endotoxin (Figure 6). Plasma IL-6 and IL-8 levels also were increased in response to either dose of endotracheal endotoxin.

View larger version (25K): [in this window] [in a new window]   Figure 5.   Endotoxin in plasma and blood neutrophils. (A) Quantification of endotoxin in plasma. Endotoxin was detectable in plasma from lambs at 130 d but not in lambs at 141 d GA. (B) Neutrophils in peripheral blood at 6 h. The number of neutrophils decreased in all endotoxin-exposed groups. *p < 0.05 versus control; tp < 0.05 versus corresponding group at 141 d GA.

View larger version (23K): [in this window] [in a new window]   Figure 6.   Proinflammatory cytokine mRNAs in spleen and cytokine concentrations in plasma. (A) Proinflammatory cytokine mRNAs in spleen were increased in 130-d lambs after either dose of endotoxin. IL-1, IL-6, and IL-8 mRNA were increased in the spleens of 141-d lambs after high tidal volume (VT) ventilation plus intratracheal endotoxin or intravascular endotoxin. (B) IL-6 and IL-8 concentrations in plasma were increased in lambs at 130 d GA after either endotoxin dose and in 141-d lambs after high VT ventilation plus intratracheal endotoxin and after intravascular endotoxin. *p < 0.05 versus control; tp < 0.05 versus corresponding group at 141 d GA.  IL-1; IL-6;  IL-8.

141-d GA Lambs

There were no differences in the mean birth weight of 4.4 ± 0.2 kg between the groups of lambs. Lung mechanics and gas exchange after 6 h of mechanical ventilation were similar for control lambs and lambs given 10 mg/kg endotracheal endotoxin (Table 1). These lambs had a mean VT of 9.4 ml/kg at 30 min of age. In contrast the lambs that were ventilated with a VT of 24 ± 2 ml/kg at 30 min of age and were subsequently ventilated with a mean VT of 10 ml/kg had respiratory failure at 6 h of age. These lambs had higher ventilatory pressure requirements, decreased oxygenation, decreased gas exchange as indicated by the VEI, and decreased lung compliance and lung volumes. The magnitude of the effect on lung function was similar to the lambs that received the intravascular endotoxin. The only lambs with decreased blood pressures were those that were endotoxin exposed and ventilated with high VT and those exposed to intravascular endotoxin (Figure 2).

The inflammatory responses in the lungs were similar for the three endotoxin-exposed groups. There were similar increases in neutrophils and hydrogen peroxide in the BALF (Figure 3), and cytokine mRNA for IL-1, IL-6, and IL-8 from cells in the BALF and lung tissue was increased (Figure 4). Total protein in BALF was 26 ± 1 mg/kg for the control lambs, 58 ± 5 for the lambs that received the intravenous endotoxin, 72 ± 17 for the 10 mg/kg endotoxin group, and 88 ± 20 mg/kg for the endotoxin plus high VT group (all values greater than control, p < 0.05). Therefore, a dose of 5 µg/kg of intravascular endotoxin caused the same inflammatory response in the lungs as did 10 mg endotoxin given by endotracheal injection, and the high VT ventilation for 30 min did not alter the inflammatory response in the lungs.

The three endotoxin-exposed groups had decreased neutrophils in blood, with lower values for the animals given the high VT for the first 30 min after birth and for the animals exposed to intravascular endotoxin (Figure 5). Endotoxin was not detected in the plasma of these 141-d GA lambs. Only the groups ventilated with the high VT or given intravascular endotoxin had increases in mRNA for IL-1, IL-6, or IL-8 in the spleen and had elevated cytokine levels in plasma at 6 h (Figure 6). The plasma levels of IL-6 and interferon- were increased in these two groups of lambs by 4 h, and no increases were detected in control or the 10 mg endotoxin groups (Figure 7). Therefore, 10 mg/kg of endotracheal endotoxin induced lung inflammation but minimal systemic effects in ventilated near-term lambs.

View larger version (18K): [in this window] [in a new window]   Figure 7.   IL-6 and interferon- in plasma of 141-d GA lambs. (A) IL-6 concentration increased at 4 h and remained high at 6 h after intratracheal endotoxin plus high tidal volume (VT) ventilation and after intravenous endotoxin. (B) Interferon- increased in the same groups at 4 and 6 h. *p < 0.05 versus values in cord blood at 0 time.  Control;  10 mg/kg IT;  10 mg/kg IT + High VT;  5 µg/kg IV.

Comparison of Control and 10 mg Endotoxin Groups at 130 d and 141 d GA

In comparison with the control 141-d GA lambs, the control 130-d GA lambs required higher ventilatory pressures to achieve equivalent VT, had higher Pco₂ values, and had lower lung gas volumes (Table 1). Therefore, the 130-d GA lungs were more immature functionally despite surfactant treatment. The 10 mg/kg dose of endotoxin increased neutrophils, increased hydrogen peroxide, and tended to increase cytokine mRNA in BALF cells and lung tissue in the 130-d GA lambs more than in the 141-d GA lambs (Figures 3 and 4). At the equivalent high dose of 10 mg/kg, the preterm lungs demonstrated more acute inflammation than did the near-term lungs. The 130-d GA lambs given 10 mg/kg intratracheal endotoxin had decreased mean blood pressures, a larger decrease in peripheral neutrophils, and endotoxin detected in the plasma, and these effects did not occur in the 141-d lambs (Figures 1 and 5). There were no increases in cytokine mRNA or in plasma cytokines in the 10 mg/kg endotoxin group at 141 d GA, whereas large increases were measured for the corresponding group at 130 d GA (Figure 6). These systemic effects of the endotracheal endotoxin were similar for doses of 0.1 mg/kg at 130 d GA. The pattern of responses indicates larger lung and systemic inflammatory responses of the ventilated preterm than the near-term lambs.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Preterm and term infants are susceptible to sepsis/pneumonia and infection acquired before birth (23). The majority of infants born before 30 wk GA have been exposed to chorioamnionitis and aspiration of infected amniotic fluid or amniotic fluid containing proinflammatory mediators is common (4). Airway samples from these infants contain increased numbers of neutrophils and proinflammatory mediators (24). To simulate the clinical situation we used intratracheal endotoxin as a proinflammatory stimulus and assessed the effect of the initiation of ventilation at birth on the lung and systemic inflammation.

We demonstrated that the near-term lamb lung at 141 d GA can contain a very large dose of 10 mg/kg endotoxin for the 6-h study period without a significant systemic response. We found that the low dose of 5 µg/kg endotoxin given by intravascular injection caused lung inflammation similar to the 10 mg/kg endotracheal dose of endotoxin. The systemic response was similar for the low intravascular dose or the high endotracheal dose of endotoxin in the lambs ventilated with a high VT for the first 30 min of life. The newborn lung at 141 d lost the ability to localize the inflammatory response if the lungs were stretched with the high VT for 30 min. High tidal volumes are frequently used inadvertently with resuscitation and the initiation of mechanical ventilation in infants (11, 25). High tidal volumes injure adult animal lungs and have resulted in systemic responses to intratracheal infection or inflammatory mediators (3, 26, 27). In a recent report, ventilation of adult rats with a VT of 42 ml/kg did not increase systemic cytokines unless the lungs were pretreated with intratracheal endotoxin (28). The high VT ventilation did not increase the endotoxin-induced inflammation in the lamb lung at 141 d GA when compared with the normal VT ventilation and intratracheal endotoxin group. However, the high VT group had increased proinflammatory mRNAs in spleen and cytokines in the plasma. We did not allow the PCO₂ to decrease in the high VT group because of the new information about the effects of CO₂ on lung injury (29). The near-term ventilated newborn lung can protect the systemic circulation from endotoxin unless that lung is stretched with a high VT.

In contrast, ventilation with VT monitoring and avoidance of hyperventilation by targeting the PCO₂ to 50 mm Hg for the surfactant-treated 130-d GA preterm lung resulted in a similar systemic inflammatory response to endotracheal endotoxin doses of 10 and 0.1 mg/kg. We mixed the endotoxin with the surfactant to avoid giving endotoxin to regions of the lung that did not receive surfactant and might be more easily injured with mechanical ventilation (30). Endotoxin was detected in the plasma of the preterm lambs given either dose of endotracheal endotoxin. The lung at 130 d GA did not prevent loss of endotoxin to the circulation at a dose 100 times less than the lung at 141 d GA. The clinically relevant point is that ventilation of the infected or endotoxin/inflammatory mediator-exposed preterm lung may cause showering of the systemic circulation with inflammatory mediators and result in a systemic inflammatory response. This finding helps explain the common occurrence of a sepsis syndrome with pneumonia or chorioamnionitis in sick preterm newborns (23). A caveat to our result is that the endotoxin was given with surfactant to the preterm lambs and the endotoxin was mixed with fetal lung fluid for the term lambs that were not surfactant treated. Either technique should result in a good distribution of the endotoxin (31).

The lung is susceptible to injury during the initiation of ventilation (11, 16). The airspaces are fluid filled and the newborn must clear fluid, establish a surfactant film at the rapidly expanding surface area of the air-fluid interface, and establish the appropriate functional residual capacity. We used mechanical ventilation for this transition but limited peak inspiratory pressures to less than 35 cm H₂O and VT to about 10 ml/kg, using a strategy that we found caused minimal lung injury with the initiation of ventilation in surfactant-treated preterm lambs (15, 16). Despite the adaptations required for transition from the fetal state to air breathing, the only systemic effect of exposure of the lamb lung at 141 d to the very high dose of 10 mg/kg endotoxin was lung inflammation that did not alter lung function and a decrease in blood neutrophils. This decrease in neutrophils probably resulted from neutrophil recruitment to the lungs (32, 33).

The lamb lung begins to alveolarize after about 110 d GA and lung volumes increase approximately linearly to term (6, 34). The preterm sheep lung at 130 d GA also is very surfactant deficient, and ventilation of the surfactant-deficient preterm lung causes nonuniform inflammation characterized by overinflation adjacent to regions of atelectasis (10). We treated the lambs with surfactant to minimize regional overdistention and to permit ventilation with lower pressures. Such a strategy decreases the bidirectional movement of protein between the alveolar and vascular spaces and decreases the protein content of lung lymph (7, 14, 35). Nevertheless, total protein was higher in BALF of the preterm than the near-term lambs demonstrating more vascular to alveolar leak of protein. The surfactant treatment did not prevent the systemic effects of the low dose of endotracheal endotoxin probably because the surfactant treatment did not prevent some regional overdistention of the preterm lungs with an incompletely developed collagen and elastin matrix. The functional residual capacity was not measured in this experiment. However, a 10 ml/kg VT applied to lung with an average total gas volume of 38 ml/kg is more likely to cause injury than the same VT applied to the 141-d GA lungs with a total gas volume of 67 ml/kg that have uniformly distributed endogenous surfactant.

The design of this study also allows us to compare pulmonary function and lung inflammation responses of the 130-d GA and 141-d GA lambs to 10 mg/kg endotoxin given intratracheally. The 10 mg/kg dose of intratracheal endotoxin did not alter lung mechanics in the 130-d GA lungs but gas exchange deteriorated. Although we did not measure cardiac output, heart rate increased and blood pressure fell, indicating that the likely explanation for the poor gas exchange was the sepsis syndrome. The 141-d GA lambs exposed to 10 mg/kg endotracheal endotoxin maintained lung mechanics and gas exchange. This result is consistent with the lack of a sepsis syndrome and the modest inflammation in the lungs by 6 h of age. We have no explanation for the increased lung volumes after the endotoxin in the 141-d GA lambs other than the possibility that the animals randomized to that group were somewhat more mature. The lungs at 130 d had a larger inflammatory response than did the lungs at 141 d GA. The preterms recruited more neutrophils to the lungs, the cells in the BALFs produced more H₂O₂, and more cytokine mRNA was expressed in alveolar cells and lung tissue. We had anticipated less inflammation in the preterm because of the immaturity of the immune system (36). A possible explanation is that the preterm is very deficient in the host defense proteins SP-A and SP-D (37). These proteins bind endotoxin and are known to modulate inflammatory responses in the lung (38). The surfactant that we used to treat the preterm lamb did not contain SP-A or SP-D.

The systemic inflammatory response in the lambs at 141 d GA initially ventilated with a high VT was similar to that resulting from the low dose of 5 µg/kg endotoxin given intravenously. However, the intravenous endotoxin resulted in a lung injury response similar to that measured with high VT ventilation plus intratracheal endotoxin. This result in lambs at 141 d GA is similar to the observation of Wiener-Kronisch and coworkers that the adult sheep lung is more sensitive to the endothelial damage from systemic endotoxin than from intratracheal endotoxin (39). This finding suggests how systemic inflammation and lung injury can occur in preterm newborns exposed to proinflammatory mediators in utero. Initiation of mechanical ventilation may transmit proinflammatory mediators from the lungs into the systemic circulation, initiating a systemic inflammation that injures the lung endothelium. The lung can be both the source and the target of the systemic inflammation (40). This study demonstrates that the surfactant-treated preterm lung that is not intentionally stretched or hyperventilated can transmit proinflammatory mediators from the lung to the systemic circulation.