Chorioamnionitis induces hepatic inflammation and time-dependent changes of the enterohepatic circulation in the ovine fetus

Chorioamnionitis, inflammation of fetal membranes, is an important cause of preterm birth and a risk factor for the development of adverse neonatal outcomes including sepsis and intestinal pathologies. Intestinal bile acids (BAs) accumulation and hepatic cytokine production are involved in adverse intestinal outcomes. These findings triggered us to study the liver and enterohepatic circulation (EHC) following intra-amniotic (IA) lipopolysaccharide (LPS) exposure. An ovine chorioamnionitis model was used in which circulatory cytokines and outcomes of the liver and EHC of preterm lambs were longitudinally assessed following IA administration of 10 mg LPS at 5, 12 or 24h or 2, 4, 8 or 15d before preterm birth. Hepatic inflammation was observed, characterized by increased hepatic cytokine mRNA levels (5h – 2d post IA LPS exposure) and increased erythropoietic clusters (at 8 and 15 days post IA LPS exposure). Besides, 12h after IA LPS exposure, plasma BA levels were increased, whereas gene expression levels of several hepatic BA transporters were decreased. Initial EHC alterations normalized over time. Concluding, IA LPS exposure induces significant time-dependent changes in the fetal liver and EHC. These chorioamnionitis induced changes have potential postnatal consequences and the duration of IA LPS exposure might be essential herein.

Ribosomal protein S15 tBAs Total bile acids TNF-α Tumor necrosis factor alpha Premature birth, before 37 weeks of gestation, is a major cause of morbidity and mortality among infants worldwide 1 . Chorioamnionitis, inflammation of the chorion and amnion during pregnancy, resulting from an ascending bacterial infection through the birth canal, is associated with premature birth [2][3][4] . During chorioamnionitis, the fetus can develop the fetal inflammatory response syndrome (FIRS), characterized by increased fetal systemic IL-6 and IL-8 levels 5 . Chorioamnionitis, prematurity and FIRS can contribute to significant rates of neonatal morbidity and mortality and are associated with the development of adverse neonatal outcomes including early-onset sepsis and intestinal pathologies [6][7][8][9] . Sepsis-associated cholestasis is a common complication in infants and is commonly caused by lipopolysaccharide (LPS), derived from Gram negative bacteria 10 . This inflammation-induced cholestasis results from a reduced mRNA expression of various hepatic bile acid (BA) transporters, such as the Na + -taurocholate cotransporting polypeptide (NTCP) and the bile salt export pump (BSEP) 10 . In specific, LPS activates cytokine production of Kupffer cells such as tumor necrosis factor alpha (TNF-α) and interleukin 1 beta (IL-1β). These cytokines bind to their receptors on hepatocytes that activates an intracellular signaling cascade resulting in altered nuclear transcription factors and reduced mRNA expression 10 . Cholestasis can lead to hepatocyte injury as BAs have been shown to be hepatotoxic during cholestasis 11 . Several mechanisms may account for this hepatotoxity; BAs could disrupt cell membranes through their detergent action on lipid components and also promote the generation of reactive oxygen species that oxidatively modify lipids, proteins and nucleic acids, which eventually cause hepatocyte necrosis and apoptosis 11 .
The liver and intestine are closely connected via the enterohepatic circulation (EHC) of BAs. In line, BAs are recently found to be critical regulators of intestinal epithelial function 12 . Several studies have found that elevated ileal BAs and an altered expression of several BAs transporters may result in ileal damage, contributing to the development of necrotizing enterocolitis (NEC). More precisely, full necrosis in the ileum was found in rats as the result of intraluminal accumulation of conjugated BAs, which was similar to the histopathological findings in an experimental rat NEC model 13,14 . Increased BA synthesis might be the cause of these elevated intraluminal BAs concentrations 15 . In addition, preterm infants with NEC and rodents with induced NEC were found to have an increased expression of the apical sodium-dependent bile acid transporter (ASBT), a protein involved in intestinal BAs uptake, suggesting increased BAs uptake by enterocytes 14,16 . In addition, an insufficient transport from the apical to basolateral side of the enterocyte was suggested since rats with induced NEC had a decreased expression of ileal bile acid-binding protein (IBABP) in their ileum 14 . Altogether, this results in BAs accumulation within enterocytes, with concomitant enterocyte damage 17 , resulting from similar mechanisms as in cholestasis induced hepatocyte injury 11 . The production of hepatic cytokines was increased in neonatal rats with induced NEC and this correlated with the progression of intestinal damage during NEC development 18 , which further underlines the important role of the gut-liver axis in NEC pathogenesis.
Fecal BA levels were found to be higher in preterm infants in the week preceding NEC manifestation compared with gestation matched controls 17 . Moreover, it was shown that fetuses exposed to endotoxin-induced chorioamnionitis develop hepatic inflammation and a disturbed lipid metabolism in utero 19 , which may persist into adolescence 20 . These combined findings suggest that the mentioned earlier liver and EHC alterations might already have their origin in utero.
Recent research showed that the fetal liver is an active immune organ with the ability of inducing an early and robust innate immune response activation, and immune activation is already initiated within 1 h to 5 h after an in utero inflammatory challenge 21 . The changes during intra-uterine infection, including liver inflammation and EHC alterations, are likely to be time-dependent as inflammation is a dynamic process and the vulnerability of the fetus to injurious hits during the complex intra-uterine development varies. The aim of this study is, given the involvement of the liver in sepsis and adverse gastro-intestinal outcomes, to evaluate the time-dependent effects of intra-uterine administration of one bolus of LPS from 15 days to 5 hours before premature delivery, on the liver and EHC of premature sheep.  5,[22][23][24] . In short, date-mated merino ewes pregnant with singleton fetuses were randomly assigned to eight different groups. After dropout or exclusion (fetuses in the control group with extremely high systemic IL-6 and/or IL-8 levels), the different groups consisted of 6 to 7 animals per group. The pregnant ewes received a single intra-amniotic (IA) injection under ultrasound guidance of 10 mg Escherichia coli-derived lipopolysaccharide (LPS) (O55:B5; Sigma-Aldrich, St. Louis, MO, USA) dissolved in saline. These injections were given at 5, 12 or 24 hours or 2, 4, 8 or 15 days before preterm delivery at 125 days of gestation (corresponds with 30-32 weeks of human gestation; term gestation in sheep is around 150 days). The control group received IA injections of saline at variable gestational ages comparable to the time points of the LPS injections before preterm delivery (Fig. 1).
The preterm lambs were delivered by cesarean section at 125 days of gestation and euthanized with intravenous pentobarbital (100 mg/kg). During necropsy, blood, liver and terminal ileum samples were sampled. Liver and ileum samples were fixed in paraformaldehyde and subsequently embedded in paraffin or liver and ileum samples were snap frozen in nitrogen.

Qualitative analysis of liver histology. A Hematoxylin and Eosin (H&E) staining was performed which
an independent pathologist blinded to the experimental set-up analyzed as previously described 25 . In short, the slides were qualitatively scored pathologically on a 0 to 4 scale for hepatic sinusoidal dilatation, shape and size of central veins and number and location of extramedullary hematopoietic clusters. Animals that had been assigned with scoring 0 had no sinusoidal dilatation, no divergent shape or size of central veins and no altered extramedullary hematopoietic clusters. Animals that had been assigned with scoring 4 had pronounced sinusoidal dilatation throughout the parenchyma, large central veins with venous stowing throughout the parenchyma and a severely increased pathologic score of extramedullary hematopoietic clusters. An increased pathologic score of extramedullary hematopoiesis, reflecting clustered and conflated hepatic erythropoiesis in the parenchyma, is a hallmark of increased erythropoiesis as a response to inflammation 26 . Total bile acid assay. As previously described 25 , total bile acids (tBAs) were measured in plasma, liver homogenate and terminal ileum homogenate by an enzymatic cycling method, according to the manufacture protocol (Total Bile Acids Assay kit, Diazyme Laboratories, Poway, CA, USA). tBAs in liver and ileum homogenate were corrected for protein content, which were measured with a BCA protein assay kit (Thermo Fisher Scientific, Waltham, MA, USA).

RNA extraction and real-time PCR.
RNA extraction and quantitative real-time polymerase chain reaction (qPCR) was performed as previously described 25 . In short, RNA was extracted from snap frozen liver and terminal ileal tissue using TRI reagent (Thermo Fisher Scientific)/chloroform extraction. Hereafter, with the use of a sensifast cDNA Synthesis kit (Bioline, London, UK), RNA was reverse transcribed into cDNA. With the specific primer in Sensimix SYBR & Fluorescein Kit (Bioline), a qPCR was performed using a 384-wells qPCR plate. qPCR reactions were executed in a LightCycler 480 Instrument (Roche Applied Science, Basel, Switzerland) for 45 cycles. mRNA expression levels of tumor necrosis factor alpha (TNF-α), interleukin 1 beta (IL-1β), interleukin-8 (IL-8) and interleukin-18 (IL-18) were measured to assess liver inflammation. IL-18 mRNA expression was also measured in ileal samples to asses ileal inflammation. Gene expression levels of cholesterol 7 alpha-hydroxylase (CYP7A1), cytochrome P450 family 27 subfamily A member 1 (CYP27A1), Na + -taurocholate cotransporting polypeptide (NTCP), bile salt export pump (BSEP), apical sodium-dependent bile acid transporter (ASBT), fibroblast growth factor 19 (FGF19), ileal bile acid-binding protein (IBABP) and organic solute transporter alpha-beta (OSTα-β) were measured to assess alterations in the EHC of BAs. To calculate the mRNA expression levels, LinRegPCR software (version 2016.0, Heart Failure Research Center, Academic Medical Center, Amsterdam, the Netherlands) was used. The geometric mean of the expression levels of three reference genes (ribosomal protein S15 (RPS15), glyceraldehyde 3-phosphate dehydrogenase (GAPDH) and peptidylprolyl isomerase A (PPIA)) was calculated and used as a normalization factor. Data are shown as fold increase over the control value; arbitrary unit (AU). Primer sequences are displayed in Table 1. Data analysis. Data are presented as median with interquartile range (IQR). Statistical analyses were carried out with GraphPad Prism (version 6.01, GraphPad Software Inc., La Jolla, CA, USA). To analyze significant differences between the groups, a nonparametric Kruskal-Wallis test followed by Dunn's post hoc test was performed. At P ≤ 0.05, differences were considered statistically significant. Given the small study groups and the potential biological relevance, differences with a P < 0.10 are also taken into account, and described as tendencies as previously described 27 . This assumption will increase the chance of a type I error, but will decrease the chance of a type II error.

Results
Severe liver inflammation at day 8 and day 15 after IA LPS administration. Histologically, no sinusoidal dilatation or divergent shape or size of central veins was observed in any of the groups. An increased qualitative pathological score of extramedullary hematopoiesis was observed in the liver of animals 8 days and 15 days after LPS administration, compared to control (both P < 0.05; Fig. 2A-D) reflecting clustered and conflated hepatic erythropoiesis in the parenchyma (Fig. 2B,C). Hepatic IL-8 mRNA levels tended to be increased 5 hours after IA LPS administration, compared to control (P = 0.09; Fig. 3A). Furthermore, TNF-α mRNA levels tended to be increased 12 hours after IA LPS, compared to control (P = 0.07; Fig. 3B). Moreover, IL-18 mRNA levels were increased 24 hours and 2 days after IA LPS administration (both P ≤ 0.05; Fig. 3C). Hepatic IL-1β mRNA levels did not differ between the groups (Supplementary Figure S1). Enterohepatic circulation alterations due to IA LPS exposure. Gene expression of NTCP, the transporter responsible for BAs uptake from the portal circulation into the hepatocyte, was decreased in the liver of animals at 12 hours after IA LPS administration, compared to control (P < 0.05; Fig. 4a). In addition, gene expression of BSEP, the pump responsible for BAs excretion from the hepatocyte into the bile canaliculi for export into the gastrointestinal tract, was also decreased in the liver of animals at 12 hours after IA LPS administration (P < 0.05; Fig. 4b).
Ethical approval. All procedures performed in studies involving animals were in accordance with the ethical standards of the institution at which the studies were conducted and ethical approval was obtained from the Animal Ethics Committees at The University of Western Australia (Perth, Australia; permit number: RA/3/100/928). All institutional and national guidelines for the care and use of laboratory animals were followed.

Discussion
In this study, we evaluated the time-dependent effects of intra-uterine administration of LPS from 15 days to 5 hours before premature delivery, on the liver and EHC of premature sheep. Recent research showed that the fetal liver is an active immune organ with the ability of inducing an early and robust innate immune response activation, and immune activation is already initiated within 1 hour to 5 hours after an in utero inflammatory challenge 21 . In our study, hepatic cytokine mRNA expression was increased from 5 hours until 2 days after IA LPS administration and these time points overlap with the ongoing fetal systemic  www.nature.com/scientificreports/ immune response, characterized by increased circulatory IL-6 levels 5 , suggesting that FIRS and local cytokine production in the liver are associated with each other. In accordance with these findings, chorioamnionitisinduced hepatic inflammation was associated with FIRS in a similar ovine model in which fetuses were assessed at day 2 and day 14 after IA LPS administration 19 . Accordingly, in utero, the liver might thus be one of the first organs to induce an innate immune response by early cytokine production as a result of chorioamnionitis. An increased qualitative pathologic score of extramedullary hematopoiesis was observed in the liver parenchyma at 8 and 15 days after IA LPS administration. This increased score, reflecting clustered and conflated hepatic erythropoiesis in the parenchyma, is a hallmark of increased erythropoiesis in response to inflammation 26 . Interestingly, in human stillborns, a similar histological pattern was strongly associated with chorioamnionitis 29 , suggesting that antenatal inflammation alters fetal extramedullary hematopoiesis. Although hepatic TNF-α, IL-8 and IL-18 levels have all been normalized at day 4 after IA LPS administration and no alterations in extramedullary hematopoiesis were observed until day 8 after IA LPS administration, it is possible that the early hepatic cytokine response contributed to the altered fetal erythropoiesis at day 8 and day 15 after IA LPS administration. In addition, fetal ileal inflammation was observed in the current and previous studies 24 with the most evident signs of inflammation 2 days and 4 days after IA LPS exposure. Transport of inflammatory mediators from the gut to the liver via the portal vein may therefore also contribute to hepatic inflammation, illustrated by extramedullary hematopoiesis. As seen in a previous study, it has been shown that upon infection and resultant immune responses, various hematopoietic factors including TLR ligands and cytokines promote extramedullary hematopoiesis in the liver 30 . Therefore, the ongoing fetal systemic inflammatory response 5 or the direct exposure to inflammatory mediators through the transport from the gut via the portal vein, or a combination of both are likely the cause of the observed altered fetal extramedullary hematopoiesis in our study. Interestingly, increased production of hepatic cytokines was also found in neonatal rats with NEC and correlated with the progression of intestinal damage during disease development 18 . Since increased hepatic TNF-α, IL-8 and IL-18 levels preceded and showed overlap with intestinal inflammation, it is likely that the early hepatic cytokine response contributed to intestinal inflammation. Whether hepatic cytokines contribute to intestinal inflammation in utero via the periphery or the bile remains to be elucidated 18 .
The important role of the liver and the gut-liver axis in NEC pathogenesis is further emphasized by the fact that elevated ileal BAs and an altered expression of several BAs transporters can contribute to the development of NEC [13][14][15][16][17]31 . In our study, however, ileal BA levels were not changed. Only at 12 hours after IA LPS administration, plasma BA levels were increased, probably as a result of decreased NTCP expression. This indicates overflow of BAs returning via the portal circulation into the systemic circulation, with a rise in systemic BA levels as a result. Interestingly, cytokines are found to be key mediators in regulating hepatic expression of BA transporters during inflammation 32 . Specifically, IL-6 can suppress NTCP and BSEP expression levels 10,32 . The decreased expression of NTCP and BSEP in our study was thus most likely the result of the increased systemic IL-6 levels 5 . Moreover, it was shown that LPS activates cytokine production of Kupffer cells, such as TNF-α, which in turn bind to their receptors on hepatocytes resulting in reduced mRNA expression of several BAs transporters 10 . Therefore, the increased hepatic TNF-α expression, in our study, might also have contributed to the decreased mRNA expression of NTCP and BSEP 10,15,33 . This suggests that both FIRS and liver inflammation are causally related to the changes in the hepatic BA transporters.
Reduced mRNA expression of various hepatic excretory BA transporters, e.g. BSEP, was observed in many animal models upon LPS treatment and also in patients with inflammation-induced cholestasis 10 . The decreased www.nature.com/scientificreports/ NTCP expression might thus be a compensatory mechanism to protect the liver against damage during sepsis as BAs have been shown to be hepatotoxic during cholestasis 11 . Our results probably show an earlier disease state in which LPS induced sepsis not yet has progressed to cholestasis since hepatic BA levels were unchanged, indicating that these animals were not cholestatic. Another compensatory mechanism might be the unchanged BAs synthesis, which in the normal situation would be expected to increase to remain a constant BA pool.
Intestinal ASBT mRNA expression was increased in animals 24 hours after IA LPS administration, which might be a response to the decreased transport of BAs from the liver to the gut 12 hours after IA LPS administration. No changes in BAs concentrations in the ileum were observed at any of the time points, suggesting that the increased ASBT expression may be a compensatory mechanism to the postulated lower intraluminal BAs supply to keep a constant BA pool in the EHC, as ASBT is known to play a major role in BA homeostasis 34 . Moreover, it might also be related to the increased cytokines such as ileal TNF-α and IL-18 levels, which in previous studies  www.nature.com/scientificreports/ have been shown to upregulate ASBT expression 35,36 . Of note, intestinal BAs did not contribute to intestinal inflammation 24 , as they were not increased. The elevated BA levels in serum following IA LPS exposure were also measured in neonates with NEC in combination with depressed biliary BA levels, suggesting a failure of BA transport from the hepatocytes into the bile canaliculi 37 . In addition, increased ASBT expression has been found in preterm infants and rodents with NEC 14,16 . However, in our study, initial alterations to the EHC normalized after IA LPS administration, suggesting that the duration of IA inflammation is important for the hepatic outcome. There is a defined window of vulnerability in which additional inflammatory hits, a premature born child is likely to encounter, might induce (additional) injury. This study highlights that the liver must be considered in neonatal care of preterm infants that suffered from inflammatory stress. Importantly, additional inflammation may have further impact on the liver and EHC and as a consequence the host in general. Based on our current findings, immune modulatory interventions such as nutrition and cytokine inhibitors might have the potential to improve neonatal wellbeing. In this context, timing of treatment initiation on the liver and EHC should be further studied.
A limitation of this study is the set-up in which the fixed moments of premature birth cannot exclude a potential influence of gestational age at start of IA infection. Furthermore, an unavoidable shortcoming of large animal studies is the relatively low number of animals per group.
In summary, in utero, the liver might be one of the first organs to induce an innate immune response by early cytokine production as a result of chorioamnionitis. An altered fetal erythropoiesis as a reaction to inflammation was detected at day 8 and day 15 after IA LPS administration that might be the result of hepatic cytokine production, ileal inflammation and ongoing FIRS. This ongoing fetal systemic inflammation and liver inflammation most likely also caused the changes in the decreased expression of several hepatic BAs transporters and resultant increased plasma BA levels that were observed 12 hours after IA LPS administration. Initial alterations to the EHC normalized over time, suggesting that the duration of IA inflammation is important for the hepatic outcome. Since a premature born child is likely to encounter additional postnatal inflammatory hits that might induce (additional) injury, this study highlights that the liver must be considered in neonatal care of preterm infants that suffered from inflammatory stress.

Data availability
The datasets generated during and/or analysed during the current study are available from the corresponding author on reasonable request.