Conceived and designed the experiments: TGAMW WB LJIZ BWK. Performed the experiments: TGAMW BZ RMJM JD GT BWK. Analyzed the data: TGAMW WB BZ RMJM JD GT EV MG YG LJIZ BWK. Contributed reagents/materials/analysis tools: EV MG YG. Wrote the paper: TGAMW WB EV LJIZ BWK.
The authors have declared that no competing interests exist.
Chorioamnionitis is the most significant source of prenatal inflammation and preterm delivery. Prematurity and prenatal inflammation are associated with compromised postnatal developmental outcomes, of the intestinal immune defence, gut barrier function and the vascular system. We developed a sheep model to study how the antenatal development of the gut was affected by gestation and/or by endotoxin induced chorioamnionitis.
Chorioamnionitis was induced at different gestational ages (GA). Animals were sacrificed at low GA after 2d or 14d exposure to chorioamnionitis. Long term effects of 30d exposure to chorioamnionitis were studied in near term animals after induction of chorioamnionitis. The cellular distribution of tight junction protein ZO-1 was shown to be underdeveloped at low GA whereas endotoxin induced chorioamnionitis prevented the maturation of tight junctions during later gestation. Endotoxin induced chorioamnionitis did not induce an early (2d) inflammatory response in the gut in preterm animals. However, 14d after endotoxin administration preterm animals had increased numbers of T-lymphocytes, myeloperoxidase-positive cells and gammadelta T-cells which lasted till 30d after induction of chorioamnionitis in then near term animals. At early GA, low intestinal TLR-4 and MD-2 mRNA levels were detected which were further down regulated during endotoxin-induced chorioamnionitis. Predisposition to organ injury by ischemia was assessed by the vascular function of third-generation mesenteric arteries. Endotoxin-exposed animals of low GA had increased contractile response to the thromboxane A2 mimetic U46619 and reduced endothelium-dependent relaxation in responses to acetylcholine. The administration of a nitric oxide (NO) donor completely restored endothelial dysfunction suggesting reduced NO bioavailability which was not due to low expression of endothelial nitric oxide synthase.
Our results indicate that the distribution of the tight junctional protein ZO-1, the immune defence and vascular function are immature at low GA and are further compromised by endotoxin-induced chorioamnionitis. This study suggests that both prematurity and inflammation in utero disturb fetal gut development, potentially predisposing to postnatal intestinal pathology.
Chorioamnionitis is a bacterial infection of the amniotic fluid, placenta and membranes that is very frequently diagnosed after preterm birth
Swallowing of the bacterial contaminated amniotic fluid is likely to result in antenatal exposure of the premature intestine to bacteria and their toxins in utero. Whether this exposure results in inflammatory and or developmentary changes of the gut of preterm babies is subject of the current study. We hypothesized that prematurity and inflammation, either alone or combined, disturb maturation of the gut barrier, the innate immune defense and vascular function thereby potentially increasing the risk to postnatal intestinal pathologic conditions. To test our hypothesis, chorioamnionitis was induced by intraamniotic endotoxin injection at low gestational age (GA) and animals were delivered either preterm or near term. In this model, the distribution of the tight junctional protein zonula occludens protein 1 (ZO-1) was determined since no data exists on its expression during gestation. Besides the TLR4 and MD-2 mRNA expression, also the distribution of either myeloperoxidase (MPO) expressing cells (polymorphonuclear leukocytes (PMN) as marker for early inflammation) or T-lymphocytes and gammadelta T-cells (cells known to be crucial in monitoring and maintaining integrity of epithelial tissues; late inflammation) were assessed. Finally, the contraction or relaxation of third-generation mesenteric arteries was studied. Relaxation was studied in the absence and presence of a NO donor and concomitantly the expression of endothelial nitric oxide synthase (eNOS) was assessed.
In this study, we show that a low gestational age and endotoxin induced chorioamnionitis both prevent maturation of the intestinal barrier function, the innate immune defense and vascular function.
This study was performed according to the guidelines of the Animal Care Committee of the University of Maastricht, which approved the protocol. The study was conducted after approval and in compliance with the guidelines of the ethical committee. Time-mated Texel ewes with singleton fetuses were randomly assigned to groups of four or five animals, to receive a single dose of 10 mg endotoxin (
Gestational development and effects of chorioamnionitis were studied at two different gestational ages. At 125d GA, fetuses were comparable to 27 weeks of human gestation. Lambs were almost near term at 140d GA since term gestation is 147d. Chorioamnionitis was induced by a single injection of endotoxin under ultrasound guidance at 111d GA and at 123d GA. Animals were delivered at 125d GA and animals of the control group underwent the same procedure with an injection of saline. Animals of 140d GA had a separate control group to assess gestational changes.
All animals were delivered by caesarean section. Lambs received a lethal i.v. injection of phenobarbiturate. The stomach content was collected for analysis. Samples of the small intestine were collected for snap freezing and fixation. Third generation mesenteric arteries of the small intestine were prepared for analysis of vascular reactivity.
Intestinal tissue was analysed with the following antibodies: monoclonal antibody against endothelial nitric oxide synthase (eNOS) (BD Transduction laboratories, San Jose, CA); rabbit antibodies against human MPO and CD3 Dakocytomation (Glostrup, Denmark) or Zonula Occludens protein 1 (ZO-1) (Invitrogen, San Francisco, CA); monoclonal antibody IL-A29 for the detection of gammadelta T-cells (VMRD, Pullman, WA). Secondary antibodies, biotin conjugated rabbit anti-mouse and Texas red conjugated goat anti-rabbit were purchased from Dakocytomation and peroxidase conjugated goat anti-rabbit from Jackson (West Grove, PA).
Small intestinal biopsies, fixed overnight in 4% formaldehyde, were embedded in paraffin and sectioned at 3 µm. For staining of CD-3 and eNOS expressing cells, slides were heated in 10 mM Na-citrate (pH 6.0) for 20 min. Endogenous peroxidase activity was blocked with 0.3% H2O2 in methanol and subsequently, slides were blocked with either normal goat serum (eNOS and MPO) or bovine serum albumin (CD-3) for 30 min at room temperature. Slides were incubated with the primary antibody of interest for 1 h at RT (CD-3 and MPO) or overnight at 4°C (eNOS). After washing, sections were incubated with the appropriate secondary conjugated antibody. Binding of the eNOS antibody was demonstrated by the streptavidin-biotin system (Dakocytomation) and binding of antibodies against CD-3 and MPO was detected using a peroxidase conjugated secondary antibody. Positive staining was visualized by applying 3-amino-9-ethylcarbazole (AEC, Sigma); nuclei were counterstained with haematoxylin. The number of cells exhibiting immunostaining for MPO and CD-3 were counted per high power field (200x). Figures were expressed as number of MPO or CD-3 expressing cells per high power field.
Frozen tissue sections were first fixed in filtered, ice-cold acetone and next in 4% paraformaldehyde. Endogenous peroxidase was blocked with 0.3% H2O2 in methanol. After blocking with 10% normal goat serum, sections were incubated for 2 h with a mab against gamma delta T-cells. After washing, an appropriate biotin conjugated secondary antibody was used. Binding of the primary antibody was demonstrated by the streptavidin-biotin system (Dakocytomation) and visualized by AEC. Nuclei were counterstained with haematoxylin. All slides were photographed by a Nikon eclipse E800 microscope with a Nikon digital camera DXM1200F and were analysed by the computer program, Lucia G color image analysis.
Tight junction distribution was examined by immunofluorescent staining of frozen sections (3 µm) with an antibody to Zonula Occludens protein 1 (ZO-1). Ileum sections were fixed with 4% paraformaldehyde. After blockage of non-specific binding sites with 10% goat serum, slides were incubated with anti-ZO-1 for 1 h. Thereafter, the sections were incubated 45 min with Texas red conjugated goat anti-rabbit antibody, followed by 2 min incubation with 4′,6-diamino-2-phenyl indole (DAPI), dehydrated in ascending ethanol series and mounted in Fluorescence Mounting Solution (Dakocytomation). The distribution of ZO-1 was recorded at a magnification of 200x using the Metasystems Image Pro System (black and white charge-couple device camera; Metasystems, Sandhausen, Germany) mounted on a Leica DM-RE fluorescence microscope (Leica, Wetzler, Germany). All images were taken at equal time-exposures after being normalized to negative control sections without primary antibody, to exclude non-specific binding of the secondary antibody or autofluorescence. At least 25 microscopic fields for each tissue section were studied.
For RT-PCR total RNA was extracted from ileal tissues using the SV Total RNA isolation system (Promega, Madison, WI). Next, isolated total RNA was treated with RQ1 RNase-Free DNase (Promega) and reverse transcribed using oligo (dT) primer and Moloney murine leukemia virus reverse transcriptase (Life technologies, Paisley, United Kingdom) according to the supplier's recommendations. cDNA samples were standardized based on the content of GAPDH cDNA as housekeeping gene. GAPDH cDNA was evaluated by performance of a GAPDH PCR on multiple dilutions of each cDNA sample. The amount of amplified product was estimated by densitometry of ethidiumbromide stained 1.2% agarose gels using a CCD camera and Imagemaster VDS software (Amersham, Uppsala, Sweden). Primer sequences for GAPDH were:
Since ischemia as a consequence of vascular compromise is one of the frequently proposed risk factors for impaired intestinal homeostasis, responsiveness to contractile and endothelium-dependent and -independent relaxant agonists was assessed in isolated third-generation mesenteric arteries. Artery rings were mounted in a wire myograph, bathed in Krebs solution (gassed with 95% air/5% CO2, pH 7.4), and normalized to a resting pretension corresponding to an intraluminal pressure of 40 mm Hg (125d GA) or 50 mm Hg (140d GA). The contractions induced by KCl (62.5 mM), the thromboxane A2 mimetic U46619 (10−10–10−7M) and the adrenergic receptor agonist norepinephrine (NE; 10−10–10−4M) was assessed. In addition, the endothelium-dependent relaxation (evoked by acetylcholine, Ach;10−9–10−4M)
The number of cells exhibiting immunostaining for MPO and CD-3 were counted per high power field. A Mann-Whitney
Tight junctions have an important role in maintaining the barrier integrity by connecting enterocytes
At 125d GA, a fragmented staining pattern for ZO-1 was seen in control animals (A) or lambs exposed to endotoxin for 2 (B) or 14d (C). At 140d GA, a normal ZO-1 distribution was detected in control animals (D) that became disrupted upon endotoxin exposure for 30d (E). Magnification 200x. For inset, 1000x magnification was used.
In preterm control animals of 125d GA, hardly any MPO expressing cells were detected (data not shown). At this GA, endotoxin exposure for 2d did not result in a significant increase of infiltrating intestinal MPO positive cells (data not shown). Interestingly, at 14 days post endotoxin treatment, large numbers of MPO positive cells were present in the lamina propria of the immature intestine (
At 140d GA, some MPO positive cells were detected in control sections (C). For each experimental group mean cell counts of MPO positive cells are depicted per high-power field (D).
The recruitment of CD-3 positive T-cells into the fetal gut paralleled that of PMN: in preterm lambs exposed to endotoxin for 2d, no significant increase of CD-3 expressing T-cells in the lamina propria was observed when compared with time matched control tissues of 125d GA (data not shown). In contrast, increased numbers of CD-3 expressing cells were seen at 14d and 30d post endotoxin exposure in both preterm and near term animals of 125 and 140d GA respectively, with the highest density in the 14d group (
Significant increase of CD-3 immunoreactivity was seen at 14 and 30d after endotoxin exposure (A+B). For each experimental group mean cell counts of CD-3 positive cells are depicted per high-power field (C).
Next, the distribution of gammadelta T-cells was evaluated, cells known to be crucial in monitoring and maintaining integrity of epithelial tissues
After 2d or 14d of endotoxin exposure, hardly any gammadelta T-cells were detected in the upper mucosa of the preterm intestine (A+B). At 140d GA, gammadelta T-cells were not detected in the lamina propria whereas lymphoid follicles in the basal mucosa became populated with gamma delta T-cells (C+D). 30d after endotoxin injection, increased numbers of gammadelta T-cells were found within areas of mucosal damage (E).
The molecular complex responsible for the recognition and signalling of endotoxin is formed by TLR4 and its indispensable cofactor MD-2
Representative DNA fragments are shown for each group (n = 4). cDNAs were standardized for GAPDH content. Quantitative data were obtained by densitometric evaluation of RT-PCR products which were compared to a standard curve obtained by amplification of a serial dilution of highly concentrated cDNA. At 125d GA, TLR4 and MD-2 mRNA levels were significantly (*) lower compared to animals at 140d GA. TLR4 and MD-2 mRNA expression was significantly reduced following intraamniotic endotoxin injection for 2, 14 and 30d.
Contractions of third-generation mesenteric arteries, evoked by the thromboxane receptor agonist U46619 increased with gestational age (
(A). Concentration-dependent contractile effects of norepinephrine tend to increase (p>0.05) during gestation but remained unaltered during chorioamnionitis (B). Concentration-dependent relaxation of mesenteric arteries induced by acetylcholine was unaltered during gestation and was significantly (*) inhibited at 14d but not after 30d of endotoxin exposure (C). The relaxation response curve for sodium nitroprusside remained identical during gestation or chorioamnionitis (D).
Either ACh (
Interestingly, the presence of SNP completely reversed the impaired relaxation of premature animals exposed to endotoxin (
Since the findings described above point at endothelial dysfunction due to impaired synthesis of NO, the distribution of endothelial nitric oxide synthase (eNOS) was studied in mesenteric arteries. Constitutive eNOS expression was absent in preterm and near term control tissues of 125 and 140d GA respectively (data not shown). However, 2 days after intraamniotic endotoxin injection in preterm animals, eNOS was weakly expressed by submucosal vascular endothelia (
Weak eNOS staining was seen at 2d post endotoxin treatment (A) while expression increased in animals exposed to endotoxin for 14 or days 30d (B-C).
This study shows for the first time the effects of gestation either alone or combined with antenatal exposure to bacterial toxins on the development of the fetal gut. A sheep model was used that closely resembles the human situation of preterm babies that are in utero exposed to chorioamnionitis by bacterial toxins resulting in adverse outcomes
To our knowledge, this is the first report showing that the cellular distribution of tight junction protein ZO-1 was underdeveloped at 125d GA in sheep. Although ZO-1 is not the only tight junction protein responsible for paracellular barrier sealing, it plays a central role in maintenance of paracellular intestinal barrier integrity by connection of the intercellular tight junction proteins (claudins and occludin) with the cell cytoskeleton.
The presence of fragmented tight junctions in the gut of preterms (125d GA) is inherent to immaturity and correlates with diminished gut barrier function in preterm infants, as assessed by sugar absorption tests
This study shows that bacterial toxins impair maturation of tight junctions during gestation, an effect which lasted as long as 30d post endotoxin administration. These data are in line with earlier in vitro and in vivo findings showing that the permeability of the intestinal epithelium is impaired after exposure to LPS
Interestingly, at 14 and even 30d after intraamniotic endotoxin injection, a strong intestinal inflammatory response was present with a massive influx of MPO positive cells and T-lymphocytes. Paradoxically, at 14 and 30d post endotoxin exposure, TLR4 mRNA was still undetectable while MD-2 expression levels remained reduced. Whether this sustained inflammatory process is the result of a broken or absent tolerance to endotoxin remains to be elucidated
At 30d post endotoxin exposure, inflammatory parameters were decreased compared to the 14d group, and this was paralleled by migration of gammadelta T-cells from the lower mucosa towards compromised epithelial surfaces. These findings are supported by a report from Chen et al. who showed that gammadelta T-cells preserve the integrity of stressed or injured intestinal mucosa
This study shows that intrauterine exposure to endotoxin induced a transient increase in the mesenteric vascular responsiveness to the thromboxane receptor agonist U46619. In addition a profound suppression of ACh-induced endothelium-dependent relaxation in mesenterial arteries was found during chorioamnionitis at low GA a process that was completely reversed by administration of the NO donor SNP. Our results strongly suggest the presence of an endotoxin-evoked transient endothelial damage and it could be speculated that a similar mechanism of endothelial dysfunction and diminished NO production might be partially responsible for the vascular vulnerability of the intestine in preterm infants
The decreased NO availability could have several explanations. First, clinical reports showed that the intestinal synthesis of the important NO precursor arginine is strongly reduced in premature infants
In summary, in this study evidence is provided that both prematurity and chorioamnionitis are associated with impaired development of the intestinal innate immune defence, the tight junctional distribution and vascular function in utero. Although at present, we can only speculate about the postnatal consequences of these developmental disorders, immaturity of the parameters investigated in this study are all associated with neonatal intestinal pathology and NEC in particular
We thank Leon Janssen, Isabelle Meijssen, Nona Jongmans and Verena Lambermont for the excellent technical help.