Main

Chorioamnionitis, an intrauterine inflammatory response of the chorioamniotic membranes to microorganisms, is a common antenatal exposure for very preterm infants that affects up to 70% of preterm deliveries before 30-wk gestation (1). In chorioamnionitis, the fetus is exposed to inflammation through direct contact with amniotic fluid or via the placental–fetal circulation (2,3). Often, chorioamnionitis is identified only by histological examination of the placental and fetal membranes after birth. Consequently, the duration of fetal exposure to an inflammatory stimulus is seldom known.

Chorioamnionitis affects neonatal outcomes in an organ-dependent way (3), but it is primarily associated with alterations in lung development (4). Chorioamnionitis is associated with a decrease in the incidence of respiratory distress syndrome (5), probably due to induced structural and functional maturity of the lung (6). Conversely, chorioamnionitis is associated with an increase in bronchopulmonary dysplasia (4,7,8) due to inflammatory disruption of lung alveolar and vascular development (9,10). Chorioamnionitis may also negatively affect neonatal outcomes by modulating the fetal immune system (3,11,12). The fetal lung, gut, and skin are directly exposed to the inflammation associated with chorioamnionitis, although other organs and the immune system, which are not in direct contact with the amniotic fluid, can also be affected adversely (3).

Neutrophils, monocytes, and lymphocytes are increased in the alveolar lavages of preterm lambs after a single dose of lipopolysaccharide (LPS) (13). Additionally, intra-amniotic LPS exposure induces (i) maturation of monocytes to alveolar macrophages with increased inflammatory response potentials by inducing hematopoietic transcription factor PU.1 (14) and (ii) cross-tolerance to different Toll-like receptor (TLR) agonists (15). Thus, the naive fetal immune system is capable of complex modulations that can alter subsequent immune function (13). However, our knowledge of the characteristics of these immune responses is limited to short-term exposure to proinflammatory agonists. The persistence of fetal immune modulations following a proinflammatory stimulus and how the duration of exposure to antenatal inflammation may affect the fetal response as gestation advances till near term is poorly understood (13,16). Whether such changes in the immune system persist after a proinflammatory stimulus may enhance understanding of whether preterm/term infants are more susceptible to infections, sepsis, and/or pneumonia later in life (17).

We used an ovine model of chorioamnionitis induced by intra-amniotic LPS to evaluate how exposure interval changed the lung as the fetus matured. Our hypothesis was that intra-amniotic LPS exposures would cause persistent fetal pulmonary responses as the lungs develop in utero.

Results

Characterization of Fetuses and Lungs

The body weights and lung weights of LPS-exposed lambs were similar to those of the control group at each gestational age (GA; Table 1 ), indicating similar growth. The cord blood pH and pCO2 were not different between groups.

Table 1 Description of animals
Full size table

As anticipated, the LPS-exposed lambs delivered at 125 d GA had higher lung gas volumes compared with the control group ( Table 1 ). Increased lung volume after LPS exposure persisted with delivery at 133 d GA. At 140 d GA, the control and the treatment groups had comparable pressure–volume deflation curves. The number of LPS injections did not influence lung gas volumes at any GA.

Pulmonary Inflammation

The total number of leukocytes increased after exposure to LPS at 125 and 133 d GA. Neutrophils, in particular, increased significantly in the bronchoalveolar lavage fluid (BALF) after two injections of LPS ( Figure 1a ). Monocytes also increased at both 125 and 133 d GA ( Figure 1b ). Lymphocytes increased only at 133 d GA after a single LPS exposure ( Figure 1c ). No difference in the total number of neutrophils, monocytes, and lymphocytes after LPS exposure was detected at 140 d GA ( Figure 1a – c ).

Figure 1
figure 1

Pulmonary inflammation. Leukocytes were increased in BALF after LPS exposure at gestational ages of 125 and 133 d, with a significant increase of (a) neutrophils and (b) monocytes at 125 and 133 d GA. (c) Lymphocytes were increased after a single dose of LPS at 133 d GA. (b and c) The y-axis was adjusted for the low number of monocytes and lymphocytes. Black bars = sal/sal; grey bars = sal/LPS; white bars = LPS/LPS. *P < 0.05 vs. control group of the same gestational age. BALF, bronchoalveolar lavage fluid; GA, gestational age; LPS, lipopolysaccharide; sal, saline.

PowerPoint slide

Full size image

PU.1-Expressing Cells in Lung Tissue

Cells expressing PU.1 increased in the lung tissue after two LPS injections at each gestation ( Figure 2a ). Compared with PU.1 expression in control lambs ( Figure 2b ), the expression of PU.1 was high in both alveolar and interstitial cells ( Figure 2c ). In contrast, a single LPS exposure resulted in only a transiently increased PU.1 expression: increased PU.1 level was evident at 125 d GA, but PU.1 expression was lost at 133 and 140 d GA ( Figure 2a ).

Figure 2
figure 2

PU.1 expression in lung tissue. (a) The PU.1-positive cells in lung tissue exposed to LPS differed significantly from controls at each gestation. Representative staining (original magnification ×200, bar = 100 µm) is shown (b) for 125 d GA control animals and (c) after two doses of LPS. The high-magnification (original magnification ×400, bar = 25 µm) insets illustrate the strong positivity of cells for PU.1 after exposure to LPS compared with that of controls. Red arrows indicate cells expressing PU.1. Black bars = sal/sal; grey bars = sal/LPS; white bars = LPS/LPS. *P < 0.05 vs. control group of the same gestational age. GA, gestational age; LPS, lipopolysaccharide; sal, saline.

PowerPoint slide

Full size image

TLR mRNA in Lung Tissue

TLR1 mRNA levels increased after one and two doses of LPS at 125 d GA ( Figure 3a ). mRNA levels of TLR4 and TLR6 decreased after the two exposures to LPS and subsequent delivery at 125 d GA ( Figure 3c – d ). However, increased mRNA levels of TLR2 and TLR6 were detected at 133 d GA ( Figure 3b – d ). No significant differences in TLR mRNA levels were detected at 140 d GA.

Figure 3
figure 3

mRNA levels of TLR 1, 2, 4, and 6. (a) TLR1 mRNA level was increased after a single dose and after double doses of LPS at 125 d GA. At 125 d GA, mRNA levels of (c) TLR 4 and (d) TLR 6 were decreased after a double dose of LPS. At 133 d GA, increased mRNA levels of (b) TLR 2 and (d) TLR 6 were detected. No differences in mRNA levels were detected at 140 d GA. Black bars = sal/sal; grey bars = sal/LPS; white bars = LPS/LPS. *P < 0.05 vs. control group of the same gestational age. GA, gestational age; LPS, lipopolysaccharide; sal, saline; TLR, Toll-like receptor.

PowerPoint slide

Full size image

Mast Cells and Clara Cells in Lung Tissue

Mast cells were present in the bronchial wall but not in other regions of lung tissue ( Figure 4b , c ). At 125 d GA, there was a trend for a decreased number of mast cells with LPS ( Figure 4a ). The decrease in mast cells was significant after two doses of LPS at 133 d GA. The number of mast cells increased as GA advanced ( Figure 4a ). A representative Clara cell staining ( Figure 4d , e ) was performed to demonstrate the specificity of the mast cell staining.

Figure 4
figure 4

Representative staining and number of mast cells counted in the bronchial wall. There were no changes in the number of mast cells (a) as a result of exposure to LPS with one or two doses compared with the control group at 125 and 140 d GA. At 133 d GA, the number of mast cells decreased after repetitive LPS administration. Representative staining (original magnification ×200, bar = 100 µm) is shown in panel (b) for 133-d-GA control animals and in panel (c), after two doses of LPS. The high-magnification insets (original magnification ×400, bar = 25 µm) illustrate the loss of mast cells after exposure to LPS compared with the number in controls. Panels (d) and (e) represent the staining for Clara cells in the bronchial wall. Red arrows indicate mast cells, and green arrows indicate Clara cells. Black bars = sal/sal; grey bars = sal/LPS; white bars = LPS/LPS. *P < 0.05 vs. control group of the same gestational age. GA, gestational age; LPS, lipopolysaccharide; sal, saline.

PowerPoint slide

Full size image

Elastin Expression in the Lung

The elastin foci on alveolar septa were counted on sections from the right upper lobe ( Figure 5a ). Elastin foci were reduced in number after two LPS injections at 125 d GA. The elastin was localized more diffusely in LPS-exposed lungs ( Figure 5c ), compared with control lungs, at 125 d GA ( Figure 5b ). Elastin foci were not different in number or localization in lungs of LPS-exposed and control groups at 133 and 140 d GA.

Figure 5
figure 5

Expression of elastin foci on alveolar septa in lung tissue. Decreased numbers of (a) elastin foci were recorded at 125 d GA after a double dose of LPS. At 133 and 140 d GA, the lungs showed no differences after LPS exposure compared with control animal lungs. The high-magnification insets (original magnification ×400, bar = 25 µm) show that the focal depositions of elastin on the tips of alveolar septa, which are strongly present in (b) the lungs of controls at 125 d GA, are diminished and diffusely expressed in (c) lungs after a double dose of LPS at 125 d GA. (b and c) High-magnification images (original magnification ×200, bar = 100 µm). Red arrows indicate location of focal expression of elastin, and green arrows indicate diffusely expressed elastin. Black bars = sal/sal; grey bars = sal/LPS; white bars = LPS/LPS. *P < 0.05 vs. control group of the same gestational age. GA, gestational age; LPS, lipopolysaccharide; sal, saline.

PowerPoint slide

Full size image

Discussion

To better characterize the effects of LPS on the fetal lung, we measured a number of variables that could contribute to longer-term adverse effects of fetal exposure to LPS. We used the single exposure at 118 d and the two exposures at 118 and 123 d to parallel our previous reports that demonstrated fetal immune tolerance to LPS (13,16,18). The deliveries at 2, 10, and 17 d after the second exposure allowed evaluation of the persistence of responses during late fetal development.

LPS in the amniotic fluid has a half-life of 1.7 d, with inflammation markers remaining measurable for at least 15 d after administration (19). In this study, intra-amniotic LPS exposure resulted in changes in the fetal lung inflammatory responses. This is consistent with our previous findings at 125 d GA regarding treatment after one or two doses of LPS (4,20,21). Our findings show that the increase in neutrophils at 133 d was evidence of a persistent response to the initial inflammatory stimulus. Furthermore, two doses of LPS were required to induce significant differences in BALF neutrophil content at 125 d GA. This finding suggests that the pulmonary immune response may be dependent on the intensity and/or persistence of the LPS exposure. The local immune system of the lung seems to remain active until 133 d GA, with increased number of monocytes and lymphocytes. Interestingly, a macrophage response was involved in the pulmonary inflammatory response to LPS in the course of gestation, shown by an increased PU.1 expression. Normally, the fetal lung contains few macrophages, and monocytes mature to macrophages with PU.1 expression after term birth (14,22,23). The lungs of fetal sheep exposed to intra-amniotic LPS expressed PU.1 within 1–2 d, and vacuolated alveolar macrophages could be recovered by lavage several days later (14). We reported that PU.1 expression increased 7 d after a single exposure to LPS but that expression was lost 15 and 22 d later. In contrast, the second exposure 5 d following the first exposure resulted in high PU.1 expression, which persisted until 140 d GA. These observations demonstrate that monocyte to macrophage maturation in the fetal lung can be transient or persistent depending on the characteristics of the inflammatory exposure. In vitro, fetal alveolar macrophages responded to TLR agonists similarly to alveolar macrophages from the adult lung (15). The persistence of macrophages could promote inflammation in the fetal lung or in the lung following delivery after TLR stimulation and may lead to an increased susceptibility for neonatal sepsis and pneumonia (17).

TLR mRNA levels were low in the fetal sheep lung relative to the adult lung, and changes in expression are minimal during late gestation (24). Our results demonstrated that changes in TLR mRNA levels with LPS exposure differed by TLR, exposure, and interval from exposure. TLR4 mRNA levels decreased 2 d, but not 7 d, following intra-amniotic injection of LPS. The delayed increases in TLR 2 and 6 mRNA levels 15 d after the initial exposure to the TLR4 agonist LPS might have clinical implications for the chronic infection and often polymicrobial organisms associated with chorioamnionitis (17). Interestingly, increased TLR2 and 6 mRNA levels seem to be related to increased monocyte numbers in BALF 15 d after a single injection, indicating an early development of the monocytes (25). It is not clear why there was a resolution of TLR2 and TLR6 to term gestation.

The homeostasis of the airway is also maintained by mast cells. Mast cells in the lower airways contribute to the development of asthma and allergic diseases (26). Mast cells are of interest because of a possible relation between chorioamnionitis, preterm birth, and the development of asthma later in life (27,28). Mast cell activation results in degranulation, leading to the rapid release of inflammatory mediators, such as histamine, proteoglycans, and cytokines, which stimulate the recruitment of other inflammatory cells (29). Mast cells are mostly present near epithelial surfaces such as mucosa of the lungs and the digestive tract (29). In this study, mast cells were found in the bronchial walls of preterm fetal lamb lungs. Although exposure to LPS resulted in a decrease of mast cells in the more immature lambs, mast cell numbers had returned to normal for lambs delivered near term. Our finding of reduced mast cell populations soon after LPS exposure contrasts with our hypothesis that LPS exposure would increase mast cells. However, repeated fetal exposures to LPS decreased airway responses to methacholine challenge in 7-wk-old lambs (12). The relationship between prenatal inflammation and the mast cell response in the fetal airways has not been studied. Increased levels of mast cells were associated with the development of BPD in one autopsy study (30). Mast cell regulation is complex, depending on the timing of the fetal and postnatal stimuli. Our results demonstrated that fetal exposure to LPS could change mast cell numbers in the fetal lungs. However, the long-term effect of intrauterine exposure to inflammation on mast cell presence and function in postnatal life needs to be determined.

To examine lung structure, elastin foci on alveolar septa were counted on lung sections. Although a double LPS exposure resulted in more diffusely localized elastin foci in the more immature fetal lambs, a more normal distribution with localization of elastin foci at the tips of alveolar septa was reestablished near term, indicating recovery with advancing gestation. Whereas a single LPS exposure can induce structural maturation (4,31), a second LPS exposure may increase the abnormal elastin distribution and alveolar development. In addition, it was demonstrated that prenatal LPS exposure upregulated tropoelastin (soluble precursor of elastin) mRNA levels in lung tissue at the time of secondary septa development, resulting in a more diffuse elastin expression along the alveolar cell wall (32). This might contribute to the abnormal lung structure. The mechanism by which the fetal lung repairs the abnormal elastin response is unknown. Some of the variables in the “recovery” response, detected in part in this study, may be GA at assessment, GA at time of initial exposure (33), and magnitude of the inflammatory responses and/or antenatal steroids (34).

A limitation of our study was that no additional exposure to LPS was given immediately before delivery at the different GAs to challenge the fetal immune system. In human pregnancies affected by chorioamnionitis, fetuses are exposed to bacteria until delivery, which challenges the immune system continuously. In our model, the fetus had a single or a second dose of LPS within a confined GA range, and the proinflammatory stimuli might be less than those occurring in human infants. In this study, we focused on the lungs. From literature, it is known that chorioamnionitis is a multiorgan disease that affects the central nervous system, skin, and the gut in clinics and experimental models (3,35,36). The interaction of the different organs is not very well understood, but the role of the immune system of the fetus, as well as the role of endotoxins in human chorioamnionitis, warrants further research.

In summary, the fetal lung responded to intra-amniotic LPS with changes that might have long-term effects on lung and immune function. Although the abnormalities in elastin distribution did not persist, monocyte to macrophage maturation, selective TLR expression, and mast cell numbers could change at later times after the LPS exposure. How preterm delivery might interact with these changes remains to be determined.

Methods

Antenatal Treatment

All animal procedures were approved by the Animal Ethics Committee of the University of Western Australia. Pregnant Merino ewes received an intra-amniotic injection with saline (sal, 2 ml; controls) or 10 mg LPS–(Escherichia coli 055:B5; Sigma–Aldrich, St. Louis, MO) at 118 and 123 d GA ( Figure 6 ). The fetuses were delivered operatively at 125 d (n = 21), 133 d (n = 19), or 140 d (n = 20) gestation (term = 150 d). Each lamb was killed at delivery with an intravenous injection of pentobarbital (100 mg/kg, Valabarb; Jurox, Rutherford, NSW, Australia).

Figure 6
figure 6

Animal model of intra-amniotic LPS exposure. All animals were exposed to saline (sal) or LPS at 118 and 123 d GA, resulting in three treatment groups: sal/sal (control), LPS/sal, and LPS/LPS. Animals were delivered at three different GAs: 125, 133, and 140 d. GA, gestational age; IA, intra-amniotic; LPS, lipopolysaccharide.

PowerPoint slide

Full size image

Processing of the Fetal Lung

Lung compliance was assessed by measuring lung gas volumes from pressure–volume curves as a measure of lung maturation. The thorax was opened with a midline incision, and a tracheal tube was inserted and connected to a manometer (37). The maximal volume measured at a pressure of 40 cm H2O was recorded as a measure of lung compliance. To collect BALF, the left lung was lavaged three times with 0.9 % NaCl (sal). The BALF samples were pooled and centrifuged at 500 rpm for 5 min. Differential cell counts were obtained on cytospin preparations after a Pappenheim staining (May–Grünwald, Giemsa (20)). Tissue from the right lower lobe was snap-frozen for mRNA analyses. The bronchus to the right upper lobe was cannulated for airway fixation with 10% formalin for 24 h at 30 cm H2O pressure, followed by transfer into phosphate buffered saline.

RNA Extraction and Reverse Transcription

Snap-frozen lung tissue from the right lower lobe was cut in 30-mg pieces. The tissue was transferred into lysis buffer (RLT buffer Qiagen, Hilden, Germany) for RNA extraction. After vortexing, the suspension was transferred to an RNeasy mini column (Qiagen) and processed according to the manufacturer’s protocol. The purity and yield of the RNA were photometrically determined using a Nanodrop spectrophotometer (Thermo Scientific, Wilmington, DE), and 1 μg total RNA was reverse transcribed by addition of 2 μl oligo(dT) primers (10 μmol/l), 1 μl RNase inhibitor (10 units/μl), 2 μl deoxynucleotide Mix (5 mmol/l), and Omniscript transcriptase (0.2 U/μl; Qiagen) and incubation for 1 h at 37 °C. The cDNA was denatured at 93 °C for 5 min.

Real-Time PCR

The design of real-time PCR primers was based on published cDNA sequences (24,38). All PCRs were performed using 1 μg/μl cDNA per reaction in duplicates. Real-time PCR reactions were performed with the LightCycler 480 SYBR Green I Master mix (4707516001, Roche-Applied, Mannheim, Germany) on a LightCycler 480 Instrument according to the manufacturer’s instructions. Real-time PCR results of TLR1, TLR2, TLR4, and TLR6 were normalized to the housekeeping gene ovine ribosomal protein S15 (ovRPS15), and mean fold changes in mRNA expression were calculated by the ΔΔCt method (39).

Immunohistochemistry (PU.1 and Clara Cell Secretory Protein)

Paraffin-embedded lung sections were deparaffinized in an ethanol series, and endogenous peroxidase activity was blocked by incubation with 3% H2O2 in milli-Q water. Sections were incubated overnight at 4 °C with the diluted primary antibody. Staining for hematopoietic transcription factor PU.1 (Santa Cruz Biotechnology, Santa Cruz, CA) was performed as described (14,16). After incubation with the appropriate secondary antibody, immunostaining was enhanced with the Vectastain ABC peroxidase Elite kit (Vector Laboratories, Burlingame, CA) and stained with nickel-diaminobenzidine. After counterstaining with Mayer’s hematoxylin, the sections were washed and dehydrated. Sections were scored for PU.1 staining with a semiquantitative scoring system: 1, little staining; 2, some staining; and 3, heavy staining. The analysis was performed in a blinded fashion.

A lung-specific Clara cell secretory protein (40) was detected in lung sections that were incubated overnight at 4 °C with the diluted primary antibody (rabbit anti-uteroglobin, Abcam, Cambridge, UK). After incubation with the appropriate secondary antibody, immunostaining was enhanced as described above, stained with diaminobenzidine, and counterstained with Mayer’s hematoxylin.

Mast Cell Staining

Mast cell staining was performed on paraffin-embedded lung sections (4 μm, transverse) from the right upper lobe (41). Briefly, the sections were stained with Weigert’s iron hematoxylin working solution for 10 min. After rinsing with water, the sections were stained with 1% toluidine blue in 35% ethanol for 1 min. Subsequently, the sections were quickly dehydrated in ethanol and xylol. Microscopic images were taken at ×100 and ×200 magnifications using a Leica microscope and Leica Qwin Pro, version 3.4.0, software (Leica Microsystems, Mannheim, Germany). The circumference of all bronchi was measured at ×100 magnification using ImageJ software, version 1.45 (National Institutes of Health, Bethesda, MD). Mast cells were counted at ×200 magnification in six representative sections per animal. The analysis was performed in a blinded fashion.

Elastin Staining

Sections of lung were incubated for 20 min in Weigert’s resorcin–fuchsin (Chroma GmbH, Münster, Germany) at 60–70 °C (41,42). After rinsing with water, the sections were incubated for 3 min in a tartrazine solution at room temperature. Subsequently, the sections were washed and dehydrated in ethanol and xylol. Microscopic images were taken at ×200 magnification, and elastin foci were counted in six representative sections per animal. The analysis was performed in a blinded fashion.

Data Analysis

Results are given as mean ± SE. Comparisons between groups were performed by two-way analysis of variance with Bonferroni post hoc analysis (GraphPad Prism, version 5, San Jose, CA). Significance was accepted at P < 0.05.

Statement of Financial Support

This work was supported by National Institute of Child Health and Human Development Grant HD-57869 (S.G.K.).

Disclosure

The authors declare no conflict of interest.