Cathelicidins and the Onset of Labour

Preterm birth, defined as delivery before 37 weeks of gestation, is the leading cause of neonatal mortality and morbidity. Infection and inflammation are frequent antecedents of spontaneous preterm birth. Cathelicidin, an antimicrobial host defence peptide, is induced by infection and inflammation and although expressed in the reproductive tract and fetal tissues, its role in the pathogenesis of spontaneous preterm birth is unknown. Here we demonstrate that cathelicidin expression is increased at RNA and protein level in the mouse uterus in a model of inflammation-induced labour, where ultrasound guided intrauterine injection of lipopolysaccharide (LPS) at E17 stimulates preterm delivery within 24 hours. Cathelicidin-deficient (Camp−/−) mice are less susceptible to preterm delivery than wild type mice following intrauterine injection of 1 μg of LPS, and this is accompanied by a decrease in circulating IL-6, an inflammatory mediator implicated in the onset of labour. We also show that the proportion of cathelicidin expressing cells in the myometrium is higher in samples obtained from women in labour at term than pre-labour. Together, these data suggest that cathelicidin has roles in mediating pro-inflammatory responses in a murine model of inflammation-induced labour, and in human term labour.

parturition, and activation of inflammation in the uterus is thought to have a dominant role in promoting the change from uterine quiescence to a contractile phenotype. In most other mammals including mice, progesterone levels fall immediately prior to parturition. However, an intrauterine injection of lipopolysaccharide (LPS) is sufficient to activate inflammation and overcome the repressive effects of high circulating progesterone concentrations, inducing PTB 19,22 . This has become an established model for parturition research 22,23 .
In this study we used Camp −/− mice, which lack the murine cathelicidin peptide, mCRAMP, to determine whether cathelicidin plays a role in the onset of PTB using an established LPS-induced PTB mouse model 22 . To determine if human cathelicidin levels change in response to labour we investigated mRNA levels of CAMP in myometrial samples from women obtained at caesarean section either before labour onset, or during labour; at preterm or at term gestations.

Results
Intrauterine LPS injection induces mCRAMP expression in the mouse uterus at mRNA and protein level. We have previously shown that intrauterine administration of 20 μg LPS at gestation day 17 induced PTB in wild type C57Bl/6 mice, with an increase in pro-inflammatory cytokines and chemokines at the maternal-fetal interface 22 . To determine the minimal dosage required to induce PTB, an LPS dose response was performed in wild type mice ranging from 0.3-20 μg/dam ( Supplementary Fig. S1). 1 μg LPS was the lowest dose found to induce PTB showing a similar time to delivery as 20 μg LPS. However, 1 μg LPS showed a significantly reduced expression of inflammatory genes Cxcl1 and IL-6 compared to 20 μg LPS ( Supplementary Fig. S2).
We found that Camp mRNA expression and mCRAMP peptide levels were increased in the mouse uterus 6 hours after an intrauterine injection of both a 1 μg and 20 μg dose of LPS (Fig. 1). Immunofluorescence showed that mCRAMP is found in the uterine epithelium in PBS-treated control mice (Fig. 2a). Following a 1 μg LPS injection, mCRAMP was found in the uterine epithelium, stromal compartment and neutrophils as confirmed with dual staining with Ly6G (Fig. 2b).

Camp −/− mice have significantly decreased circulating IL-6 following intrauterine LPS injection.
Having shown that Camp −/− mice are less susceptible to LPS-induced PTB, we explored differences in maternal cytokine levels and uterine inflammatory gene expression following 1 μg LPS injection in wild type and Camp −/− mice. Maternal blood of Camp −/− mice had significantly lower circulating IL-6 compared to wild type following LPS injection (694 ± 1045 pg/ml vs. 2608 ± 1678 pg/ml; p = 0.045) (Fig. 4). No differences were seen in circulating CXCL10 or TNF levels (p = 0.996 and p > 0.999, respectively).
The gene expression of prostaglandin-endoperoxide synthase 2 (PTGS2), chemokine (C-X-C motif) ligand 1 (CXCL1) and tumor necrosis factor (TNF) were unaltered in the uteri of Camp −/− mice compared to wild type mice, irrespective of whether 1 μg of LPS, PBS or no treatment was given (Fig. 5).

CAMP gene expression and quantification of LL-37 positive cells in human myometrial samples.
Myometrial samples were classified into term pre-labour caesarean section (Term No Labour [TNL]), term labour (TL), preterm pre-labour caesarean section (Preterm No Labour [PTNL]) preterm labour (PTL) ( Table 1). Labour was defined as presence of regular uterine contractions leading to progressive cervical dilatation.
The myometrial samples were analysed for CAMP gene expression as well as the percentage of LL-37 positive cells (by immunofluorescence) in relation to total number of nuclei counted. Although no differences were found in CAMP gene expression between the groups (Fig. 6a), the percentage of LL-37 positive cells present in the human myometrium was higher in TL samples in comparison to TNL (35.5% vs 2.02% p < 0.0001), PTNL (35.5% vs 2.44% p < 0.0001) and PTL (35.5% vs 2.66% p < 0.0001) myometrial samples (Fig. 6b). Representative images of LL-37 immunofluorescent staining used to determine the percentage of LL-37 positive cells are shown in Fig. 7.

Discussion
In this study, we demonstrate that mouse cathelicidin, mCRAMP, may play a role in the onset of inflammation associated PTB stimulated by LPS. We show that mCRAMP and IL-6 are upregulated in the uterus 6 hours following a labour-inducing intrauterine LPS injection stimulus. Camp −/− mice, which lack the mCRAMP protein, are less likely to go into labour following intrauterine LPS and show reduced levels of LPS-induced IL-6. We also show that human cathelicidin, LL-37, is highly expressed in myometrium from women who have had spontaneous labour at term, but we do not show this increase in myometrium from women who experienced preterm labour.
Cathelicidin is abundantly found in the reproductive tract 10 and has been characterised as having both anti-inflammatory and pro-inflammatory roles 3 . Therefore, the role of this pleiotropic anti-microbial peptide in the context of PTB (well characterized as an inflammatory event) 16,24 , could be beneficial or detrimental. The direct microbicidal and anti-endotoxic properties of cathelicidins might be expected to be protective against PTB and the capacity of cathelicidin to bind to LPS and block activation of TLR4 might have been predicted to inhibit induction of PTB 3,25,26 . However, the level and timing of LPS and cathelicidin exposure is crucial to the response; cathelicidin is also capable of inducing inflammasome activation and IL-1β and IL-18 production in LPS-primed www.nature.com/scientificreports www.nature.com/scientificreports/ upon inflammation. Interestingly, it has been shown that endogenous mouse cathelicidin does not protect against LPS-induced shock, with similar or improved survival rates of Camp −/− compared to wild type mice 30 , suggesting that anti-endotoxic properties of the endogenous peptide may not be the dominant effect in an in vivo inflammatory setting 31 .  . Time to delivery in Camp−/− and wild type mice after intra-uterine PBS or LPS injection. Time to delivery of first pup after intra-uterine injection with PBS, 1 µg or 20 µg of LPS in cathelicidin-deficient mice (Camp −/− ) and wild type (C57BL/6J) mice. Significance representing LPS treated mice compared to PBS controls from same genotype (*p < 0.05, **p < 0.01, ***p = 0.001), # p < 0.05 difference between genotype in the same treatment group, Two-way ANOVA). No symbol signifies no significant difference between the groups (p > 0.05). Preterm Birth (PTB) rates, defined as delivery before 24 hours rates in percentage are shown below. Significance is calculated by using Fisher's Exact test.  www.nature.com/scientificreports www.nature.com/scientificreports/ The pro-inflammatory effect of cathelicidin can be beneficial in certain circumstances. Indeed, we have shown that cathelicidin can promote pulmonary clearance of Pseudomonas aeruginosa in an in vivo murine model by enhancing the neutrophil response 13 . In our murine LPS-induced PTB model, intrauterine LPS injected Camp −/− mice, compared to wild type mice, showed significantly less circulating maternal IL-6; a pro-inflammatory cytokine shown to increase in myometrial cells after LPS stimulation 32 and known to play an important role in the development of preterm and term labour [33][34][35] . This may suggest that mCRAMP influences the inflammatory profile to some extent and may have the potential to trigger the inflammatory cascade that ultimately leads to labour (Fig. 8). However, it may contribute to the onset of PTB through alternative pathways as CXCL10 and TNF, also known pro-inflammatory cytokines, were found to be similar in both wild type and Camp −/− mice following an intrauterine LPS injection. Transcription of PTGS2 (encoding COX-2 enzyme), CXCL1 (encodes the chemokine CXCL1 which regulates the recruitment of neutrophils and basophils during inflammation 36,37 ) were also investigated in the mouse myometrium, however, these were found to be unaltered following 1 μg LPS injection at 6 hours.
Similar delivery rates were seen in wild type and Camp −/− mice when a high concentration of LPS (20 μg) was administered, suggesting that a sufficiently strong inflammatory stimulus can override the protective effect of mCRAMP deficiency.
It has been previously shown that cathelicidin is increased in a variety of tissues in various inflammatory conditions, and upon LPS exposure 3,[38][39][40] . This is consistent with our observations in the mouse uterus with mCRAMP significantly increased 6 hours after LPS injection. Although mouse pregnancy is considerably different from human pregnancy, studies have indicated that the mouse model can potentially be useful for preclinical studies of human labour 19,22,41 . In our study we used human myometrium as an initial insight into the expression of LL-37 in context of labour as it forms the majority of the uterine wall and is central to labour as it is responsible for uterine contractions. We observed an increase in LL-37 protein levels during human term labour, compared to non-labour. These findings are in line with those of Lim et al. who found an increase in LL-37 protein levels in human myometrium from term women in labour compared to term women undergoing a pre-labour caesarean section 42 . They found LL-37 localised to myometrial muscle fibres in both labouring and pre-labouring samples,  www.nature.com/scientificreports www.nature.com/scientificreports/ and in labouring samples LL-37 was also found in leukocytes 42 . LL-37 is known to be predominantly produced by neutrophils and epithelial cells 3 and as there is an influx of neutrophils in term and pre-term labouring myometrium 17,43 , we hypothesise that the LL-37 increase is at least in part due to this infiltration 20,21 . However, we were unable to confirm whether neutrophils were a source of the LL-37 increase in our human myometrial samples, due to technical failure of dual immunofluorescent staining on paraffin embedded tissues. To overcome this limitation in future experiments, collected samples could be frozen for staining on frozen tissue, which may work better in this type of analyses. Furthermore, flow cytometry could be utilised to identify the LL-37 positive infiltrate cells. Importantly, future investigations should also include the study of other reproductive tissues, such as the endometrium and the amnion/chorion, as inflammatory mechanisms vary in different tissues.
Despite a greater proportion of LL-37 positive cells present in labouring human myometrium, we did not observe an equivalent increase in myometrial CAMP gene expression, suggesting that the LL-37 we observe in these human samples may primarily be stored and released from granules, as opposed to being synthesized de novo. Indeed, LL-37 in neutrophils is known to be predominantly pre-synthesised and stored as pre-peptides in granules 9,44 . This may potentially explain the discrepancy between the levels of mRNA and of protein seen in the human labouring myometrium. This contrasts our findings in the mouse, where LPS stimulated the production of mCRAMP at both mRNA and protein level. We did not detect a difference in LL-37 levels in the human myometrium when we compared samples from women in preterm labour and those having pre-labour preterm www.nature.com/scientificreports www.nature.com/scientificreports/ delivery. As all preterm deliveries are abnormal, finding an appropriate control group for the study of gene and protein expression changes during preterm labour is notoriously difficult. Many of the conditions that indicated iatrogenic PTB in our control, pre-labour caesarean section preterm group, are themselves associated with inflammation (e.g. pre-eclampsia, placental insufficiency and abruption). Thus, the lack of a detectable difference may reflect increased inflammation in both groups. In the present study, we cannot exclude that the regulation of LL-37 in the myometrium is different in preterm labour compared to preterm pre-labour, future research is needed for this, using appropriate control groups which are difficult to obtain. However, there was less LL-37 protein expression in preterm labour samples compared to term labour samples which may indicate differences in the pathophysiology of preterm and term labour, or may reflect other differences between the groups (e.g. duration of labour at time of sampling). The lack of an LL-37 upregulation in myometrium of the PTNL women might be considered inconsistent with our findings in the mouse model where we see an upregulation of mCRAMP 6 hours post-LPS injection. However, a previous study has shown that changes in the transcriptome in mouse myometrium induced by intrauterine LPS injection mirror those in the human myometrium seen during term parturition more closely than those seen during preterm labour 19 . Our data supports this finding as we also see a parallel between human TL myometrium and mouse uteri exposed to LPS, in that both show an upregulation of cathelicidin. In the mouse term labour is initiated through progesterone withdrawal and is associated with much less inflammation 45 . This corresponds with our observation in preliminary experiments examining mCRAMP expression throughout mouse gestation, where we found no evidence of upregulation in term labour.
In conclusion, we show that LPS increases the expression of cathelicidin in wild type mouse uterus and an absence of cathelicidin in Camp −/− mice reduces the incidence of PTB triggered by a low dose of LPS. In humans, cathelicidin protein levels are increased in myometrium undergoing term labour. These data suggest that cathelicidin plays a role in mediating some of the effects which lead to inflammatory PTB in mice, and may also be relevant in the onset of human labour.

In vivo experiments.
In vivo studies were conducted under a UK Home Office license to J.E.N. (60/4241) in accordance with the Animals Scientific Procedures Act (1986). C57Bl/6 virgin female mice were purchased from Charles River Laboratories (Margate, UK). Camp tm1Rig mice 46 were backcrossed to congenicity onto a C57Bl/6J OlaHsd strain, and bred from homozygous matings in house in specific pathogen free facilities, in individually ventilated cages at the University of Edinburgh. Following matings, mice were housed in grouped cages of up to 6 mice per cage under specific pathogen free conditions at the same facility. Mice were permitted food and water as required and maintained in a 12-hour cycle of light and dark. All mice were acclimatized for a minimum of 1 week before use. Female mice were mated with stud males from the same genotype. The presence of a vaginal copulatory plug indicated day 1 of gestation. Pregnant mice with no appearance of a plug were not used in experiments, as gestational age was uncertain.
In vivo preterm birth model. The ultrasound guided intrauterine injection mouse model of PTB previously optimised in our group was used 22 . A dose response curve using wild type mice was performed to determine the minimal dosage needed to induce PTB (1 μg/dam), which was used parallel with the highest dose previously used (20 μg/dam) to determine the effects on PTB rates of the different dosages, (Fig. S1). Differences in the inflammatory response were also investigated (Fig. S2). Space between two intra-amniotic sacs was located with ultrasound where the injection was administered. 25 μl of sterile DPBS or LPS (20 μg/dam or 1 μg/dam) was injected at gestational age day 17. After recovery, mice were monitored using closed circuit television cameras and a video recorder. Time of delivery was the time between injection and delivery of first pup. PTB was defined as delivery within 24 hours of intrauterine injection.

In vivo timed collection experiments.
A separate cohort of mice was used to assess the inflammatory effect of the stimuli used on maternal tissues. Mice were sacrificed by lethal exposure to carbon dioxide 6 hours after the intrauterine injection. A group of "no treatment" mice were also included. These mice were 17 days of gestation and received no treatment, and were culled in the same manner described above. Maternal serum was acquired by collecting whole blood through puncture of the aorta using 21G needles in Brand Tubes Microtainers tubes and spun for 5 minutes at 5.9 × g for ELISA analysis. Uterine samples were dissected and snap frozen in RNA-free tubes for protein and RNA analysis and separate samples were fixed in 4% neutral buffered formalin and embedded in paraffin blocks for immunofluorescence analysis.
Enzyme-linked Immunosorbent Assay (ELISA). EBiosceince duosets for IL-6 and TNF-α and an R&D Duoset for CXCL-10 was used for cytokine analysis. Duosets were used according to the manufacturer's protocols. Upon collection, samples were placed into RNAlater solution (R0901 Sigma-Aldrich) for 24 hours at 4 °C, then taken out and stored at −80 °C prior to use. Inclusion criteria was singleton pregnancies that were term (>37 weeks of gestation) or preterm (<37 weeks of gestation) whilst exclusion criteria were age under 16 and any blood borne infections. Women either underwent an elective pre-labour caesarean section (no labour group) or an emergency CS in labour due to maternal and/or fetal indications (e.g. delay in labour, pre-eclampsia, fetal distress). Labour was defined as regular uterine contractions with cervical dilation.