Prevention of lipopolysaccharide-induced preterm labor by the lack of CX3CL1-CX3CR1 interaction in mice

Preterm labor (PTL) is the most common cause of neonatal death and long-term adverse outcome. The pharmacological agents for PTL prevention are palliative and frequently fail to prevent PTL and improve neonatal outcome. It is essential to fully understand the molecular mechanisms of PTL in order to develop novel therapeutic methods against PTL. Several lines of evidence indicate some chemokines are expressed in gestational tissues during labor or PTL. To reveal the pathophysiological roles of the CX3CL1-CX3CR1 axis in PTL, we performed present study using LPS-induced PTL mice model in CX3CR1-deficient (Cx3cr1-/-) mice. We indicated that PTL was suppressed in Cx3cr1-/- mice and immunoneutralization of CX3CL1 in WT mice. From immunohistochemical and the gene expression analyses, the CX3CL1-CX3CR1 axis has detrimental roles in PTL through intrauterine recruitment of macrophages and the enhancement of macrophage-derived inflammatory mediators. Thus, the CX3CL1-CX3CR1 axis may be a good molecular target for preventing PTL.


Definition of clinical situations and sampling of blood and placentas from patients
Human sera and placentas were obtained at the Department of Obstetrics and Gynecology, Wakayama Medical University between 2011 and 2016. The clinical data of patients were extracted from medical records. As described previously [26], the study groups were defined as follows; deliver at term in which sera were obtained at the gestation of 28 to 32 weeks (preterm control), preterm with labor at the gestation of 28 to 32 weeks (PTL), and term with (TIL) or without (TNL) spontaneous labor. PTL group included threatened premature labor and preterm premature rupture of membrane (pPROM). Patients in PTL group were treated with antibiotics and tocolysis (ritodorine hydrochloride). Only when ritodorine hydrochloride was ineffective, magnesium sulfate (MgSO 4 ) was additionally administered. Labor was defined by the presence of regular uterine contraction at a frequency of less than 10 min interval with cervical changes or rupture of membrane resulting in delivery [27]. In TIL and PTL groups, sera were obtained at the hospitalization. In TNL group, sera were collected before caesarian section. Patients who delivered neonates with congenital or chromosomal abnormalities, twin, and those with preeclampsia or gestational diabetes were excluded based on the clinical records. The demographic and clinical characteristics of each group are shown in Table 1. Placentas were obtained at the delivery, and several tissue sections consisting of the chorioamniotic membranes, umbilical cord, and placenta were evaluated for diagnosis of chorioamnionitis and

Mice
Specific pathogen-free 8-12-week-old female C57BL/6 mice were purchased from SLC Japan Inc (Shizuoka, Japan). and designated as WT mice in this study. Cx3cr1 -/mice were a generous gift from Drs. P.M. Murphy and J.L. Gao (National Institute of Allergy and Infectious Diseases, National Institutes of Health, Bethesda, MD) [29]. Age-and sex-matched Cx3cr1 -/mice, which were backcrossed to C57BL/6 mice at least 8 generations, were used in the following experiments. All mice were housed individually in cages under specific pathogen-free conditions during the whole course of the study.

Animal model
LPS-induced PTL was conducted similarly as described previously [16]. In brief, two female mice were co-housed with a male one of the same genotype and checked for vaginal plugs as the evidence of mating at the next morning. The day of vaginal plug detection was designated as gd 0.5 of pregnancy, and the pregnant mice were removed from the male. All pregnant mice delivered their pups on gd 19 or 20. At gd 15.5, the pregnant mice were intraperitoneally administered with LPS (25 μg in 200 μl of PBS) or PBS 200 μl, followed by the observation until delivery. The presence of intact or partial fetal tissue in the cage was noted as the evidence for delivery, and the delivery before gd 18.5 was judged as pre-term labor, and pups born before gd 18.5 could not survive >24 h after birth [30]. In our dose, 25 μg was the highest dose of LPS that reproducibly induced PTL without significant morbidity or mortality in gravid C57BL/6 females, based on the established method [31,32]. In some experiments, pregnant WT mice were i.p. injected with anti-CX3CL1 Ab (200 μg/mouse) or control IgG at 3 h after LPS injection, and were observed until delivery. The percentage of preterm labor was calculated as follows: percentage (%) = 100 x (number of LPS-treated mice that delivered until gd 18.5/number of LPS-treated pregnant mice). All animal experiments were approved by the Committee on Animal Care and Use at Wakayama Medical University.

Blood and tissue sampling
At gd 5.5, 10.5, 15.5, 18.5, or 6 h after LPS or PBS injection, mice were anesthetized by the intraperitoneal injection of Avertin (tribromoethanol, 20 μg/g), and whole blood samples were taken, followed by centrifugation (3,000 x g for 15 min) to collect serum. In each pregnant mouse, uterus was removed, and viable fetuses were dissected from the amniotic sac and umbilical cord, to weigh the obtained fetuses and placentas. Gestational tissues including uterus (from between implantation sites), placenta and amnion were stored at -80˚C until processing for the subsequent analyses.

Histopathological and immunohistochemical analyses
Human and mouse gestational tissues were fixed in 4% formaldehyde buffered with PBS (pH 7.2) and paraffin-embedded sections (4-μm thick) were made, followed by conventional H&E staining. Immunohistochemical analyses were performed as described previously [33]. Briefly, deparaffinized sections were immersed in 3% H 2 O 2 in PBS for 5 min to eliminate endogenous peroxidase activity. The sections were further treated with PBS containing 1% normal serum corresponding to the secondary Abs and 1% BSA to reduce nonspecific reactions, and incubated with the primary Abs at the optimal conditions. After incubation with biotinylated secondary Abs, immune-complexes were visualized by Catalyzed Signal Amplification system (CSA) or Labelled Streptavidin biotin (LSAB) system (Dako, Kyoto, Japan), followed by counterstaining with hematoxylin. The positive cells were enumerated on 5 randomly-chosen visual fields at x400 magnifications, the total numbers in the five fields were combined. All measurements were carried out by two independent researchers without a prior knowledge of the experimental protocols.

Double-color immunofluorescence analyses
A double-color immunofluorescence analysis was performed to identify the types of cells expressing CX3CR1-, IL-1β-, IL-6-, TNF-α-, or COX-2 in mouse tissues as described previously [34]. Deparaffinized sections were performed by enzymatic digestion with trypsin for 10 min at 37˚C for antigen retrieval. The slides were incubated with PBS containing 1% normal donkey serum and 1% BSA to reduce nonspecific reactions. Thereafter, the slides were incubated overnight at 4˚C with pairs of primary Abs. After washing, the slides were incubated with fluorochrome-conjugated secondary Abs at room temperature for 1 hr. The slides were then observed under a fluorescence microscopy.

Cell culture
WT mice were i.p. injected with 2 ml of 4% thioglycolate (Sigma-Aldrich, St. Louis, MO), and intraperitoneal macrophages were harvested at 3 days later as described previously [29]. The resultant cell population consisted of more than 95% macrophages based on the morphological criteria. The cells were suspended in DMEM medium (containing 10% FBS and penicillin) and incubated at 37˚C in six-well cell culture plates. Two h later, non-adherent cells were removed, and the medium was replaced. Thereafter, the cells were stimulated with CX3CL1 (1, 10 and 100 ng/ml) for 3 h and were subjected to subsequent analyses.

Extraction of total RNAs and real-time RT-PCR
Total RNA was extracted from gestational tissue or peritoneal macrophages using ISOGEN II (Nippon Gene, Tokyo, Japan), according to the manufacturer's instructions. One μg of total RNA was reverse transcribed into cDNA at 37˚C for 15 min using PrimeScript RT reagent Kit with gDNA Eraser (Takara Bio, Shiga, Japan). Thereafter, generated cDNA was subjected to real-time PCR analysis using SYBR Premix Ex Taq II kit (Takara Bio, Shiga, Japan) with the specific primer sets ( Table 2) obtained from Takara Bio. Amplification and detection of mRNA was performed using Thermal Cycler Dice Real Time System (TP800; Takara Bio, Shiga, Japan) according to the manufacturer's instructions. Relative quantity of target gene expression to β-actin gene was measured by comparative Ct method as described previously [35].

Statistics
The means and SEMs were calculated and presented for all parameters determined in this study. Statistical significance was evaluated using, Mann-Whitney's U test, Kruskal-Wallis test, Steel-Dwass test, or Chi-square test. P<0.05 was accepted as statistically significant.

Elevated serum CX3CL1 levels in PTL patients
Serum CX3CL1 levels were not increased in term labor with spontaneous labor (TIL) and term labor without spontaneous labor (TNL), compared with preterm control group, indicating that labor itself cannot increase serum CX3CL1 levels ( Fig 1A). On the contrary, the PTL group exhibited significantly elevated serum CX3CL1 level compared with those in the other groups ( Fig 1A). CX3CL1 proteins were observed by immunohistochemistry in amnion epithelial cells and trophoblasts obtained from the placenta of patients who suffered from preterm labor arising from chorioamnionitis (Fig 1B-1D). On the contrary, chemokines, serum CCL2 and CCL5 levels were relatively but not significantly elevated in PTL group, compared to preterm control group (S1 Fig). Moreover, serum CCL3 levels could not be detected in all cases. These observations implied the involvement of CX3CL1 in the pathophysiology of PTL, particularly one that is associated with intrauterine infection.

CX3CL1 and CX3CR1 expression in murine gestational tissues after LPS injection
LPS treatment enhanced Cx3cl1 gene expression in the gestational tissues of pregnant WT mice at gestational day (gd) 15.5 (Fig 2A). Consistently, immunohistochemical analysis detected CX3CL1 proteins in amnion epithelial cells and cytotrophoblasts within the labyrinth zone in the placenta of pregnant WT mice at gd 15.5 ( Fig 2B). In contrast, serum CX3CL1 levels fluctuated but did not change significantly until gd 18.5 compared with that in non-pregnant mice (Fig 2C). LPS injection at gd 15.5, however, markedly increased serum CX3CL1 levels, at 6 h after injection ( Fig 2C) and induced preterm labor in most WT animals (7 cases of PTL in 8 LPS-treated WT mice). Moreover, a large number of CX3CR1-expressing cells appeared in uterus at 6 h after LPS injection compared with PBS treatment (Fig 2D and 2E). Furthermore, a double-color immunofluorescence analysis identified most CX3CR1-expressing cells in the uterus as F4/80-positive macrophages ( Fig 2F). These observations implied that LPS-induced PTL was accompanied by the infiltration of CX3CR1-expressing macrophages into the uterus. https://doi.org/10.1371/journal.pone.0207085.t002 The roles of CX3CL1-CX3CR1 in preterm labor of mice

Crucial roles of the CX3CL1-CX3CR1 axis in LPS-induced PTL
In order to evaluate the roles of the CX3CL1-CX3CR1 axis in PTL, we injected WT and Cx3cr1 -/mice with LPS since CX3CR1 is a single specific receptor for CX3CL1. WT and Cx3cr1 -/mice went through labored at term either without treatment (S2 Fig) or with PBS treatment (Fig 3A).
On the contrary, LPS injection markedly reduced pregnancy duration and caused PTL in most WT (6 out of 7, 85.7%) but a minority of Cx3cr1 -/-(1 out of 8 mice, 12.5%) mice (Fig 3A and 3B). WT and Cx3cr1 -/mice delivered pups with the same body weights when they delivered spontaneously at full-term ( Fig 3C). Moreover, LPS administration at gd 15.5 had few impacts on the fetal body and placenta weights in both WT and Cx3cr1 -/mice until 6 h after the injection (Fig 3D and  3E). In order to exclude the possibility that genetic deficiency of CX3CR1 may cause secondary effects on labor processes, we next examined the effects of anti-CX3CL1 on LPS-induced PTL, since CX3CL1 is the sole ligand for CX3CR1. Anti-CX3CL1 injection significantly reduced LPSinduced PTL in WT mice compared with that in control IgG treatment (anti-CX3CL1, 3 out of 8 mice, 37.5%; control IgG, 7 out of 8 mice, 87.5%) (Fig 3F). These observations implicated the crucial involvement of the CX3CL1-CX3CR1 axis in LPS-induced PTL. WT mice at 6 h after the injection, compared with PBS treatment (Fig 4A-4H). Consistent with CX3CR1 expression on F4/80 + macrophages (Fig 2E), macrophage infiltration was not enhanced in Cx3cr1 -/mice after LPS treatment (Fig 4C and 4D). On the contrary, neutrophil and T cell infiltration was not significantly suppressed in Cx3cr1 -/mice (Fig 4A, 4B, 4E and 4F). These observations implicated that CX3CR1 deficiency may prevent LPS-induced PTL by suppressing intrauterine macrophage recruitment.

Intrauterine gene expression of inflammatory mediators
Accumulating evidence has indicated the vital roles of pro-inflammatory cytokines including IL-1β, IL-6, and TNF-α, and prostaglandin E 2 in LPS-induced PTL [37]. Hence, we examined the intrauterine gene expression of Il1b, Il6, Tnfa, and Ptgs2. Both WT and Cx3cr1 -/mice exhibited similar levels of intrauterine mRNA expression of these molecules at 6 h after PBS treatment at gd 15.5 (Fig 5). LPS treatment markedly enhanced intrauterine gene expression of these molecules in both WT and Cx3cr1 -/mice, but the enhancement was significantly attenuated in Cx3cr1 -/mice compared with that in WT ones (Fig 5). Moreover, a double-color immunofluorescence analyses revealed that IL-1β, IL-6, TNF-α, and COX-2 proteins were expressed mostly in F4/80 + macrophages, which infiltrated into the uterus at 6 h after LPS injection (Fig 6). Furthermore, CX3CL1 treatment augmented the gene expression of these molecules in WT mice-derived peritoneal macrophages in a dose-dependent manner (Fig 7). These observations indicated that CX3CL1 induced the infiltration of CX3CR1-expressing macrophages in the uterus and the simultaneous expression of these pro-inflammatory molecules, thereby promoting PTL development.

Discussion
PTL, a major obstetrical complication, is associated commonly with systemic and/or local bacterial infection during pregnancy. Various types of cells including inflamed endothelial cells, macrophages, neurons, and glial cells express CX3CL1 [17], which exerts its activity by binding to its specific receptor, CX3CR1 [19]. We previously provided definitive evidence to indicate the crucial involvement of the CX3CL1-CX3CR1 in several bacterial infection models [29,38]. Moreover, in line with previous reports [39,40,41], we could find that CX3CL1 was The roles of CX3CL1-CX3CR1 in preterm labor of mice immunohistochemically expressed in amnion epithelial cells and trophoblasts in both human and murine samples. Hence, we examined the pathogenic roles of CX3CL1 in PTL. Indeed, in the present clinical and experimental studies, elevated serum CX3CL1 levels were related to PTL. CX3CR1-expressing cells markedly infiltrated the gestation tissues in mouse PTL model and genetic disruption of CX3CR1 gene or anti-CX3CL1 antibody administration prevented LPS-induced PTL in mice. Collectively, the CX3CL1-CX3CR1 axis can have important pathophysiological roles in infection-induced PTL. Prior to delivery arising from either term labor or PTL, gestational tissues exhibit inflammatory responses [36], which induces uterine contraction, an indispensable step for labor [8,42]. Inflammatory responses are associated with infiltration of various type of leukocytes including neutrophils, NK cells and macrophages. However, neutrophil recruitment had no influence on LPS-induced PTL [43]. The involvement of NK cells was proposed as evidenced  The roles of CX3CL1-CX3CR1 in preterm labor of mice by NK cell dysregulation in preterm labor in human [44] and NK cell depletion mediated delayed term in mice [45]. This assumption, however, could not be applied to the present model as evidenced by the absence of differences in intrauterine NK cells between WT and Cx3cr1 -/mice. Macrophages consistently reside in mouse decidual tissues but are increased as gestational age advances, reaching the maximal level at gd 18, immediately before term labor. Indeed, the systemic depletion of macrophages could significantly prevent LPS-induced PTL in mice [37]. In the present study, macrophage recruitment was significantly suppressed The roles of CX3CL1-CX3CR1 in preterm labor of mice together with PTL prevention in Cx3cr1 -/mice compared with WT mice. Thus, macrophage infiltration may be associated with LPS-induced PTL.
Like other types of leukocytes, macrophage recruitment is also regulated under the guidance of chemokines systems [46]. Indeed, F4/80 + macrophages express various chemokine receptors including CX3CR1, and genetic disruption or immunoneutralization of CX3CR1 impaired macrophage recruitment in atherosclerosis, wound healing, and renal diseases, thereby preventing or alleviating pathological changes [47][48][49][50][51]. On the contrary, the absence of CX3CR1 gene had few effects on macrophage infiltration in other disease models such as thioglycolate-induced or cecal ligation and puncture-induced peritonitis [29,52,53]. Thus, CX3CR1-mediated signals may have distinct roles in macrophage infiltration in a contextdependent manner. In the present study, Cx3cr1 -/mice showed impaired macrophage infiltration after LPS treatment, indicating the essential involvement of the CX3CL1-CX3CR1 interaction in macrophage recruitment in this PTL model.
Before delivery, the expression of several pro-inflammatory mediators is increased in the gestational tissues, together with concomitant intrauterine leukocyte influx [6,54,55]. Indeed, the expression of several pro-inflammatory cytokines is increased in pregnancies complications caused by infection or PTL in humans and other species [56][57][58]. These observations suggest the essential involvement of these mediators in PTL. In support of this notion, IL-1 and TNF-α administration induced both preterm labor and intermediate steps in the labor The roles of CX3CL1-CX3CR1 in preterm labor of mice cascade, and genetic disruption of IL-1 or TNF-α receptors significantly prevented LPSinduced PTL [59][60][61]. In addition, the inhibition of IL-6 signaling or IL-6 production suppressed LPS-induced PTL [30,62]. On the contrary, the absence of IL-10, a major anti-inflammatory cytokine, exaggerated LPS-induced PTL with elevated expression of inflammatory cytokines such as IL-1, IL-6 and TNF-α [63]. Thus, proinflammatory cytokines and IL-10 can promote and alleviate PTL, respectively. In the present study, PTL was alleviated in pregnant Cx3cr1 -/mice, while intrauterine gene expression levels of Il1b, Il6 and Tnfa were attenuated. Moreover, IL-1β, IL-6, and TNF-α were expressed in CX3CR1-expressing F4/80 + macrophages, and CX3CL1 enhanced their expression. Thus, CX3CR1-mediated signals could promote PTL, partly by inducing macrophage infiltration and their pro-inflammatory cytokine expression.
In both term labor and PTL, PGs play essential roles in the processes of uterine contractility, membrane rupture, and cervical ripening [64][65][66]. PGs are generated by the enzymatic action of PTGS, known as COX. PTGS1 (COX1) is constitutively expressed in most cell types, whereas PTGS2 (COX2) is induced in response to inflammatory stimuli [67]. The expression of PTGS2 but not PTGS1 is increased at term labor in human gestational tissues such as amnion, choriodecidua and myometrium [68,69]. The administration of PTGS2 (COX-2) inhibitor can effectively suppress LPS-induced PTL [70,71], implying the crucial roles of COX-2-mediated PGs. We proved that the lack of CX3CR1 decreased the intrauterine gene expression of Ptgs2 after LPS treatment. In vitro, CX3CL1 treatment augmented the gene expression of Ptgs2, in addition to that of Il1b, Il6, and Tnfa, in WT-derived peritoneal macrophages in a dose-dependent manner. Because of the cytotoxicity, the highest dose of CX3CL1 at 100 ng/ml had no effects on the expression of these molecules in vitro. Moreover, a doublecolor immunofluorescence analysis detected COX-2 in F4/80 + macrophages, and CX3CL1 induced COX-2 expression in macrophages. Thus, the CX3CL1-CX3CR1 interaction prematurely induced COX-2-expressing macrophage infiltration and further enhanced COX-2 expression in macrophages, thereby inducing PTL.
Currently, uterine relaxant or tocolytic drugs are mainly employed to prevent PTL [2]. However, after systemic and/or local infection advances to inflammatory cascades, these drugs cannot effectively prevent PTL. Thus, it is necessary to develop more specific and effective therapeutic agents against infection-induced PTL. The blockade of CX3CL1-CX3CR1 axis significantly inhibited inflammation-induced PTL but did not result in any adverse effects on uninfected and uncomplicated labor. Thus, CX3CL1-CX3CR1 axis may be a novel molecular target for the development of therapies that prevent infection-related PTL.