Main

Late gestational maturation prepares the fetal lung for postnatal survival with the development of saccules/alveoli and a lung microvasculature, increases in surfactant and increases in lung antioxidant enzymes (AOE) (1,2). A reduction in respiratory distress syndrome in preterm infants, reflecting early lung maturation, has been associated with antenatal treatment with corticosteroids and with intrauterine infection and/or inflammation (3,4). This inflammation-induced fetal lung maturation has been explored in fetal sheep using intra-amniotic injections of the pro-inflammatory agents E.coli endotoxin and IL-1 (5,6).

We previously showed that intra-amniotic endotoxin caused chorioamnionitis and induced expression of IL-1 in the chorioamnion and fetal lung (7,8). Using surgical procedures to separate the fetal lung from the amniotic fluid, we further demonstrated that direct lung exposure to endotoxin could induce both inflammation and lung maturation without chorioamnionitis (9). However, endotoxin-induced production of IL-1 or other mediators by the chorioamnion or fetal lung may be important for initiating or amplifying the inflammation and maturation signals. Intra-amniotic IL-1 also causes chorioamnionitis and induces lung maturation (6). We hypothesized that IL-1 effects on the fetal lung are mediated by direct signaling without the requirement of chorioamnionitis.

Although the endotoxin TLR-4 receptor and the IL-1 receptor are similar and share components (10), the agonists for these receptors do not induce equivalent responses in the fetal lung. For example, endotoxin induces inflammatory cell recruitment to the fetal lung by a CD-18 dependent mechanism while inflammatory cell recruitment by IL-1 is CD-18 independent (11). Endotoxin effects on the fetal lung include potentially adverse responses such as inflammation, decreased alveolar septation and microvascular injury and potentially beneficial effects such as induction of AOEs, modulation of innate immunity, and induced lung maturation (7,12–14). The pattern of responses of the fetal lung to the direct administration of IL-1 has not been evaluated. The objectives of this study were to evaluate inflammatory, morphologic, vascular, AOE and surfactant changes to direct IL-1 infusion in the fetal lung.

METHODS

Animals, surgery, and delivery.

The Animal Experimentation Ethics Committees at the Cincinnati Children's Hospital Medical Center and the Western Australia Department of Agriculture approved these studies with fetal sheep. Date-bred singleton Merino ewes in Western Australia were pre-anesthetized with ketamine (20 mg/kg, im) and xylazine (0.2 mg/kg, im) and anesthesia was maintained with 1.5–2% halothane in O2. Aseptic surgery was performed to isolate the fetal respiratory system from the amniotic sac, silicone tubing was secured in the fetal trachea to allow collection of the fetal lung fluid into a 2-L sialastic bag (9). In the intratracheal IL-1 (IT IL-1) group, an osmotic infusion pump (Alzet 2001D, Alza Corp, Palo Alto, CA) was used to infuse 10 μg of recombinant ovine IL-1α (6) or saline over 24 h via a short length of vinyl tubing in the distal trachea as described previously (9). In a second experiment that was previously reported, 1 mg E. coli endotoxin (IT endo) or saline over 24 h was infused into the trachea (9). We report here previously unpublished measurements of AOE activity for this second group of animals. The ewes were sedated at gestational d 124 ± 1 for delivery 1d or 7d after the surgery with ketamine (1g IM) and xylazine (25 mg IM) followed by spinal anesthesia (2% lidocaine, 3 mL). Each fetus was delivered by cesarean section, given a lethal dose of pentobarbital and exsanguinated. Deflation pressure-volume curves then were performed after inflating the lungs with air to 40 cm H20 pressure (5).

Processing of lungs.

The lungs were removed from the chest and weighed. The left lung was lavaged three times with a volume of saline at 4°C that fully distended the lungs (5). Total cell counts in bronchoalveolar lavage fluid (BALF) were performed using Trypan blue exclusion to identify dead cells, followed by differential counts on cytospin preparations stained with Diff-Quick (Dade Behring, Dudington, Switzerland) (15). BALF was used for surfactant lipid analysis. Tissue from the right lower lobe was frozen in liquid nitrogen for later protein and mRNA analysis. The right upper lobe was inflation fixed by bronchial instillation with 10% formalin at 30 cm H2O pressure for morphometric analysis.

Surfactant lipid measurements.

The lavage fluid was pooled, lipids extracted with chloroform:methanol, and saturated phosphatidylcholine (Sat PC) isolated using osmium tetroxide (16) and quantified by phosphorus assay (17).

Antioxidant enzyme measurements.

Minimal residual blood remained in the lung tissue because the lambs had been exsanguinated (14). Thawed lung tissue was weighed, homogenized in cold PBS (1:10-1:20 wt/vol; Brinkmann Polytron, Westbury, NY), and centrifuged at 15,000 rpm for 45 min. The supernatants were assayed for AOEs by standard spectrophotometric techniques. Total superoxide dismutase (SOD) was measured using the xanthine/xanthine oxidase method, whereby SOD activity is assayed by inhibition of the reduction of cytochrome c in the xanthine oxidase reaction (18). Catalase (CAT) was measured as the rate of reduction of hydrogen peroxide at 240 nm (19). Glutathione peroxidase (GP) was quantified as the rate of oxidation of NADPH at 340 μm using cumene hydroperoxide as substrate (20). Lung homogenates were also assayed for total DNA content (21), and AOE activities were expressed as units/mg lung tissue DNA. Assays were performed in parallel for tissue collected from the lambs exposed to intratracheal endotoxin (9).

mRNA analyses.

Total RNA isolated from lung tissue was used for multiprobe RNAse protection analysis (7). Sheep specific riboprobes for IL-1β, Il-6, IL-8 (7), VEGF, PECAM-1, Tie-1 (12) Tie-2 and VEGFR2 were described (15). For all the RNAse protection assays, ovine ribosomal protein L32 was used as an internal standard. Aliquots of 10 μg of total RNA were hybridized with excess [32P] UTP (Life Sciences Products, Boston, MA) radiolabeled probes at 55°C for 18h. Single-stranded RNA was digested with RNAse A/RNAse T1 (Pharmingen, San Diego CA). RNAse was then inactivated and protected RNA was precipitated using RPA III™ inactivation/precipitation buffer (Ambion, Austin TX). The protected fragments were electrophoresed on a 6% polyacrylamide-urea sequencing gel and visualized by autoradiography. The protected bands were quantified on a PhosphorImager using ImageQuant software (Molecular Dynamics, Sunnyvale, CA).

Evaluations of vascular proteins.

Frozen lung samples were homogenized in ice-cold buffer containing protease inhibitors (12). Briefly, the samples were sonicated, centrifuged and the protein content in the supernatant was determined by the BCA method. Fifty μg of protein sample per lane was resolved by 3% to 8% Tris-Acetate gel electrophoresis (Novex precast gels, Invitrogen, Carlsbad CA) and proteins from the gels were transferred to PVDF membrane by electroblotting (Invitrogen, Carlsbad CA). The blots were incubated with the primary antibody overnight at 4°C using one of the following antibodies diluted in blocking buffer: mouse monoclonal anti-eNOS antibody (Transduction Labs, Lexington KY), goat polyclonal anti-PECAM-1 antibody (sc-1505), rabbit polyclonal anti-Tie-2 antibody (sc-324) or rabbit polyclonal anti-VEGFR2 antibody (sc-504) (all antibodies from Santa Cruz Biotechnology, CA). Blots were incubated for 1 h at room temperature with the appropriate conjugated secondary antibody. After washing, bands were visualized by chemiluminescence (ECL kit; Amersham Pharmacia Biotech, Buckinghamshire, UK) and radiographed. The autoradiographs were scanned at high resolution and images acquired using Adobe Photoshop software. Densitometric quantitation was performed using ImageQuant v1.2 software (Molecular Dynamics, Sunnyvale, CA).

Lung morphometry.

As lung maturation is known to vary between regions of the lung all morphometric analysis was performed on the right upper lobe (RUL). Morphometric analysis was performed on lungs receiving either intratracheal IL-1 or saline. Fixed lobe volume (FLV) was measured by the volume displacement method of Scherle (22). Each RUL was cut into 5-mm serial transverse sections and 3 randomly selected and uniformly spaced sections were then paraffin-embedded. The paraffin-embedded sections were then cut (5-μm thickness) and stained with hematoxylin and eosin (H&E) for morphometric measurements.

Stained sections of lung were enlarged and printed onto photographic paper. A cycloid point counting grid (72 lines/144 points) was superimposed onto each photographic image to calculate the volume fraction of lung parenchyma (alveoli and alveolar ducts), non-parenchyma (blood vessels and conducting airways), interlobular septa (interstitial tissue forming distinct lobulation of the lung) and pleura. The volume fraction of these variables was calculated using the formula Pi/Pt where Pi is the number of points hitting the structure of interest and Pt is equal to the total number of points that where superimposed over the entire lung. The parenchymal, non-parenchymal, interlobular septa and pleura volumes per lobe were calculated by multiplying the FLV by the appropriate volume fraction variable.

From each 5-μm H & E-stained section, 10 non-overlapping digital images were captured at a final magnification of ×850 using a Spot RT digital camera interfaced with an Olympus BX51 microscope and computer. The number of points that fell on alveolar septal tissue and airspace, and the number of air-tissue intercepts where counted by superimposing a linear point counting grid (40 lines/80 points) using the method described by Willet et al. (23,24). Surface fraction (Sv) which is the surface area of alveoli and alveolar ducts per unit volume of alveolar tissue, total alveolar surface area of the RUL, alveolar wall thickness, numerical density, total number of alveoli in the RUL and average alveolar volume were determined using the methods and equations described previously (23,24).

Statistics.

Values are expressed as means ± SE. Normally distributed data were compared by ANOVA, using the Student-Newman-Keuls method for multiple comparisons. Data not normally distributed were compared using Kruskal-Wallis one-way ANOVA of ranks, with pair wise multiple comparisons using Dunn's procedure. A p-value < 0.05 was considered statistically significant.

RESULTS

Physiologic variables at birth.

The control lambs given saline and lambs exposed to IL-1α had similar body weights, lung to body weight ratios and cord blood gases (Table 1). There were no fetal deaths.

Table 1 Physiological variables of preterm lambs at birth
Full size table

Inflammation.

Fetal lung inflammation was evaluated by assessing recruitment of leukocytes to the BALF and expression of pro-inflammatory cytokines in fetal lung tissue. Relative to controls, neutrophils and monocytes in BALF were increased on 1 d and 7 d after intratracheal IL-1α (Fig. 1A). The mRNAs for pro-inflammatory cytokines IL-1β, IL-8 and IL-6 were induced 4-11 fold over control 1d after intratracheal IL-1α (Fig. 1B).

Figure 1
figure 1

Intratracheal IL-1 induces lung inflammation. (A) Inflammatory cells in bronchoalveolar lavage (BAL) fluid of preterm lambs were counted 1d and 7d after exposure to intratracheal IL-1. Cell counts are expressed per kilogram body weight. Neutrophils (□) and monocytes (▪) were <104/kg in control BAL fluid. (4–9 animals per group, *p < 0.05 versus control). (B) IL-1β, IL-6 and IL-8 mRNA in lung tissue from preterm lambs 1d after intratracheal IL-1 exposure were quantified by RNAse protection assays. The mRNA values were normalized to L32 (ribosomal protein mRNA). The mean mRNA signal in control animals was given a value of 1 (shown as a dashed line) and levels for each mRNA were expressed as values relative to controls (4–6 animals per group, *p < 0.05 versus control).

Full size image

Expression of vascular proteins.

Expression of proteins known to be essential for vascular development were evaluated both at the mRNA and protein level. VEGFR2, Tie-1, Tie-2, PECAM-1 and VEGF 165 isoform mRNA expression were not different between groups (Table 2). There was a 40% decrease in VEGF 188 isoform mRNA expression at 1d but not at 7d after IL-1α exposure. Of the proteins evaluated, only Tie-2 decreased 30–40% at both 1d and 7d after IL-1α exposure, but eNOS, VEGFR2 and PECAM-1 protein expression were unchanged. This pattern of responses suggests minimal and transient effects of the intratracheal IL-1 on the developing lung vasculature.

Table 2 Effect of IT IL-1 exposure on expression of vascular markers in fetal lung
Full size table

AOE activity.

AOE activity was measured for the intratracheal IL-1 exposed lambs and the previously reported parallel experiment using intratracheal endotoxin (9). We measured AOE enzyme activity at 7d because AOE enzymes do not increase at 1d after intra-amniotic endotoxin (14). Intratracheal IL-1 exposure did not increase SOD, CAT or GP activities 7d after exposure (Fig. 2). In contrast, intratracheal exposure to endotoxin increased CAT and GP activities at 7d.

Figure 2
figure 2

Intratracheal IL-1 did not increase AOE activity. Anti-oxidant enzyme activity in lung tissue from fetal sheep 7d after intratracheal infusions of IL-1α or endotoxin. The antioxidant enzyme activities of superoxide dismutase (SOD), catalase (CAT) and glutathione peroxidase (GP) were not increased by IL-1α (▪) relative to controls (□). The endotoxin exposure (□) increased CAT and GP relative to control and IL-1α. *p < 0.05

Full size image

Lung morphometry.

The right upper lobe volumes (mL/kg) were 4.0 ± 0.6 for the saline group versus 4.5 ± 0.4 for the IT IL-1α exposed group (p = 0.5 NS). Lung surface area, alveolar numbers, alveolar numeric density and alveolar wall thickness were not different between IL-1α and control groups 7d after the surgery and intratracheal exposures (Fig. 3).

Figure 3
figure 3

Intratracheal IL-1 did not cause morphometric changes in the fetal lung. Lung morphometry was performed for 5 controls (□) and for 5 lungs 7d after exposure to intratracheal IL-1 (IT IL-1 - ▪). There were no differences between the groups for alveolar surface area (A), alveolar number (B), alveolar wall thickness (C), or alveolar numeric density (D).

Full size image

Maturation.

Compared with controls, the Sat PC pool size increased by 4-fold at 1d and by 19-fold at 7d after IL-1α (Fig. 4). Compared with controls, IL-1α increased lung gas volumes at 7d but not at 1d.

Figure 4
figure 4

Intratracheal IL-1 induced lung maturation. (A) Amount of saturated phosphatidyl choline (Sat PC) in bronchoalveolar lavage fluid (BALF) was expressed as μmol/kg body weight for controls (□) and intratracheal IL-1α at 1d and 7d (▪). (B) Lung volumes as (ml/kg body weight) versus pressure in cm H2O. Intratracheal IL-1 increased BAL Sat PC and lung volumes at 7d (▿) relative to controls (O) or a 1d intratracheal IL-1α exposure (□)). (4–9 animals per group, * p < 0.05 versus control).

Full size image

DISCUSSION

Chorioamnionitis, caused by ureaplasmas, Gram-negative or Gram-positive bacteria, is common in cases of premature delivery (25). Analysis of amniotic fluid of women with chorioamnionitis demonstrates increases in multiple cytokines and pro-inflammatory mediators including IL-1 (26). The fetal effects associated with chorioamnionitis include lung injury, lung maturation and brain injury in humans (27,28). In this study, we analyzed lung injury and lung maturation responses to direct lung exposure to a single cytokine IL-1 α without the confounding effects of the cascade of inflammatory mediators resulting from chorioamnion inflammation. In fetal sheep, IL-1β mRNA increases soon after intra-amniotic endotoxin or IL-1α (29). While either IL-1α or IL-1β when given by intra-amniotic injection can induce lung maturation, IL-1α is more potent and was chosen as the agonist for these experiments (6).

Intratracheal IL-1 induced acute lung inflammation at 1d (increased cytokines and inflammatory cells) and at 7d monocytes were increased in BALF. This inflammatory response is similar quantitatively and qualitatively to that induced by intratracheal as well as intra-amniotic endotoxin as reported previously by our group (9,30). Interestingly, the fetal lung responds to IL-1 with both inflammation and maturation, but it has no inflammatory or maturational responses to TNFα (29). While endotoxin inhibited microvascular development and alveolar septation (12,23), the IL-1 induced inflammation did not have much of an effect on markers of microvascular injury and did not interfere with alveolar septation. However, the degree of lung maturation, as indicated by the induction of Sat PC and increases in lung volumes, was similar to intratracheal or intra-amniotic endotoxin-exposed animals (9). However, intratracheal IL-1α did not induce AOE activity. This lack of response differs from the increase in AOEs in the fetal lung in response to intra-amniotic endotoxin (14). This comparison of the IL-1 and endotoxin responses demonstrates that inflammation can induce lung maturation in the absence of microvascular and septation abnormalities. We speculate that the increased injury and perhaps oxidant stress caused by endotoxin exposure is needed to induce AOE. Antenatal endotoxin exposure results in changes in postnatal inflammatory responses to ventilation (31). There is no information about how antenatal IL-1 might impact postnatal injury responses.

Although endotoxin receptor TLR-4 signaling and IL-1 receptor signaling share common mediators, there are important differences between the signal transduction mechanisms. Activation of TLR-4 recruits myeloid differentiation factor 88 (MyD88) and other adaptor molecules to the receptor complex (32). The IL-1 receptor associated kinases and tumor necrosis factor receptor-associated factor 6 then are activated leading to activation and nuclear translocation of nuclear factor-κB (NF-κB) resulting in transcription of cytokines and other genes. MyD88 gene knock-out studies in mice demonstrate that IL-1 receptor signaling is completely dependent on MyD88 while LPS signaling can occur by a MyD88-independent pathway (32,33). While the TLR-4 signaling via the MyD88-dependent pathway causes an early NF-κB activation, TLR-4 signaling via the MyD88-independent pathway leads to delayed NF-κB activation and transcription of interferon and interferon responsive genes. The signal transduction pathways initiated by TLR-4 or IL-1 receptor activation have not been studied in fetal tissues.

Previous studies from our laboratory demonstrate that responses of the fetal lung to IL-1 and endotoxin differ in a number of ways. The recruitment of inflammatory cells to the fetal lung is CD-18 (β2 integrin) dependent for endotoxin, but not for IL-1 (11). Intra-amniotic endotoxin induces the interferon-responsive genes IP-10 and MIG in the fetal lung, while intra-amniotic IL-1 does not (34). We also previously reported that intra-amniotic endotoxin strikingly induces the AOEs (14) and the results from this study show that intratracheal endotoxin causes modest increases in CAT and GP. In contrast, intratracheal IL-1α had no effect on AOE. Other signaling pathways also can induce both inflammation and lung maturation. For example, live ureaplasmas, which presumably signal via TLR2 and TLR6 receptors, and endotoxin from periodontal organisms also can induce lung maturation in the fetal sheep (35,36). These results demonstrate that inflammatory responses, downstream injury, and maturational responses in the fetal lung can differ with different pro-inflammatory agents. Further, these new results demonstrate that lung maturation can result from the direct signaling of the fetal lung by a single cytokine and the antecedent inflammatory cascade from chorioamnionitis is not required.

In most experimental models, an increase in AOEs was linked to other lung maturational responses (2). Maternal or fetal exposures to corticosteroids induce the AOEs in sheep and all other species evaluated to date (37). We report that AOEs are induced by intra-amniotic and intratracheal endotoxin, but not intratracheal IL-1. Of note, the induction of AOE by intratracheal endotoxin was less striking than the response to intra-amniotic endotoxin (14). The inflammatory response as assessed by inflammatory cells in BALF did not differ much for intratracheal and intra-amniotic endotoxin compared with intratracheal or intra-amniotic IL-1 (9). A difference in the characteristics of the inflammatory responses or in oxidant products of those responses may explain why intratracheal IL-1 did not induce the AOE. The important point is that AOEs may not increase despite lung inflammation and lung maturation. The clinical corollary is that inflammation induced lung maturation may not protect the infant from oxygen mediated injury. Fetal lung AOEs increase in late gestation in parallel with increases in plasma cortisol and surfactant lipids (2). In contrast to corticosteroids, thyroid hormones induce the surfactant system but suppress the simultaneous increases in lung AOEs (38,39). Thus, maturational agents need not induce parallel increases in surfactant and AOE.