IRAK1 Is a Critical Mediator of Inflammation-Induced Preterm Birth

Intrauterine infection or intrauterine inflammation (IUI) or chorioamnionitis is a prevailing cause of preterm birth and fetal injury. In 2015, 9.7% of all births in the United States were preterm births (1), resulting in 75% of perinatal mortality and 50% of the long-term morbidity (2). Globally, approximately 1 million deaths per year are attributable to prematurity (3, 4). Notably, more than 40% of the prematurity is triggered by IUI (2, 5–7). In addition to maternal morbidity, IUI enhances the risk for fetal and newborn brain injury (8–10), necrotizing enterocolitis (11), and chronic lung disease (12).

The inflammatory mediators during IUI are strongly implicated as causative agents of prematurity (13). Currently available therapies for IUI and/or preterm labor rely on antibiotics and progesterone (14, 15). However, these treatment strategies are generally not effective, and there continues to be many adverse, long-term side effects (16). Antibiotics, the first line of treatment, often fail to prevent IUI-associated morbidities (17, 18) because of the inflammatory mediators of the infection that account for fetal and maternal injury (19). Thus, there is a pressing need for new effective therapies to treat IUI and preterm labor. A limitation in developing more effective therapies is the dearth of understanding of the molecular mechanisms of pathogenesis in IUI and preterm birth.

Recently, there have been many reports in humans, mice, and rhesus implicating IL-1β (IL-1) in the pathogenesis of IUI-associated preterm birth (20). The activation of IL-1R–associated kinase 1 (IRAK1), a downstream mediator of IL-1 and TLR signaling pathways, is involved in the generation of active IL-1, the master cytokine in inflammation (21). IL-1 and TLR signaling pathways can be activated by a variety of stimuli, including recognition of microbial pathogens/products (e.g., LPS), the presence of reactive oxygen species, recognition of DNA damage, and abnormalities in the tissue matrix caused by chronic inflammation (22, 23). IRAK1 is central in engaging immune signaling pathways and is a key signaling mediator in the TLR/IL1 pathway (24). This signaling induces proinflammatory cytokines, chemokines, and inflammasome components, which results in rupture of fetal membranes as well as induction of PGs, which enhances uterine contraction and expulsion of fetus (25). IRAK1 has also been involved in many neoplastic disorders and metabolic, cardiovascular, and inflammatory diseases (26). Thus, IRAK1 plays a critical role in innate immunity via both IL-1β production and subsequent signaling (24). Inhibition of IRAK1 kinase activity could, therefore, alter IL-1β– and TLR-driven signaling in IUI and preterm birth.

However, the role of IRAK1 in IUI and preterm birth has not been studied. We thus hypothesize that IRAK1 is activated in fetal membranes of women with IUI delivering preterm, and inhibition of IRAK1 will lead to decrease in preterm birth. As mouse pregnancy physiology differs from human pregnancy (27), we evaluated IRAK1 activation first in human fetal membranes and then in our established rhesus model of IUI and finally in a mouse model of IUI. In both humans as well as induced models of preterm birth, we found strong association of preterm birth with activation of IRAK1. Inhibition of IRAK1’s function or its absence protected from inflammation-induced preterm birth, thus presenting IRAK1 as an attractive target for preventing preterm birth in humans.

Pregnant women at pregnancies from 26 to 36^6/7 weeks were recruited. Cohorts of chorioamnionitis/IUI and no chorioamnionitis were established based on a detailed histopathologic diagnosis of chorioamnionitis based on Redline’s (28) criteria and cytokine profile. Patient characteristics are shown in Table I. Membranes were separated within few hours after delivery. Decidua was scrapped and amnion and chorion separated and frozen at −80°C.

Separated amnion and decidua from term C-section patient not in labor were cultured in DMEM plus 10% FCS plus penicillin/streptomycin for 4–5 h and then treated with LPS with mentioned doses and duration.

H&E staining was performed for human fetal membrane sections, and staining was photographed. H&E-stained sections of human fetal membranes were scored in a blinded manner (by S.G.K.) for chorioamnionitis using criteria outlined by Redline et al. (28) based on numbers and depth of neutrophil infiltration of the tissue.

Normally cycling, adult female rhesus macaques (Macaca mulatta) were time mated. At ∼130 d of gestation (∼80% of term gestation), the pregnant rhesus received either 1 ml saline solution or 1 mg LPS (Escherichia coli O55:B5; MilliporeSigma) in 1 ml saline solution by ultrasound-guided intra-amniotic (IA) injection (Table II). Dams were surgically delivered 16 or 48 h later. After delivery, fetuses were euthanized with pentobarbital, and fetal tissues were collected. There were no spontaneous deaths or preterm labor in the animals.

Male and female C57BL/6 mice were bred overnight, and copulation plug was identified in the morning and defined as 0.5 d postcoitum. On embryonic day 15.5, dams were given ketamine and xylazine anesthesia, and a minilaparotomy was performed. Ultrapure LPS (E. coli O111:B4; InvivoGen), 250 μg [as originally described by Lyttle et al. (29)], or 10 μg [as described by Edey et al. (30)] in a 100 μl volume or sterile saline solution was infused between two gestational sacs in the lower right uterine horn. Uterus was then returned to abdomen; the fascia and the skin were sutured. The entire procedure took ∼5 min/mouse. The mice were observed under infrared camera for preterm birth. IRAK1 inhibitor (407601; Sigma-Aldrich) was given 4 h before the surgery. Mice genetically deficient for IRAK1 were created on a C57BL/6 background as previously described (26). For mechanistic studies, mice were sacrificed 12 h after LPS injection.

Both rhesus and human fetal membranes were dissected away from the placenta. Amnion were peeled off from the choriodecidua layer. Tissues were washed after harvesting with ice-cold PBS and then lysed in protein extraction buffer containing 1% Nonidet P-40 (v/v), 20 mM Tris-HCl (pH 7.5), 2 mM EDTA, 150 mM NaCl, 1 mM DTT, protease inhibitor mixture (P8340; Sigma-Aldrich), phosphatase inhibitor mixture 2 (P5726; Sigma-Aldrich), and phosphatase inhibitor mixture 3 (P0044; Sigma-Aldrich). The lysates were centrifuged at 13,200 rpm for 10 min, and insoluble debris was discarded. Tissue extracts were resolved through SDS-PAGE using 4–12% separating gel (Invitrogen). Proteins were transferred to Hybond ECL nitrocellulose membrane (Amersham Pharmacia Biotech) using a semidry transfer system (Bio-Rad Laboratories) and blocked with 5% dried milk or 5% BSA and 0.1% Tween 20 (MilliporeSigma). Blots were probed with anti–p-IRAK1 Ab (bs-319R; BIOSS); cyclooxygenase-2 (COX-2) Ab (Cayman Chemical aa 584–598); anti-IRAK1 (4504), anti-TAK1 (5206 and 4505), anti–p-SAPK/JNK (4668 and 9251), anti-SAPK/JNK (9253 and 9252), anti–p-p38 MAPK (4511 and 9211), and anti–p38 MAPK (9211, 9212) (Cell Signaling Technology, Beverly, MA); and anti–p-TAK1 (06-1425) and β-actin Ab (A5060; MilliporeSigma, Temecula, CA) overnight at 4°C. Binding of HRP-labeled goat anti–rabbit Ab (sc-2004; Santa Cruz Biotechnology, Santa Cruz, CA) was determined using SuperSignal West Chemiluminescent Substrate (Thermo Fisher Scientific, Rockford, IL). Blots were stripped with Restore Western Blot Stripping Buffer (Thermo Fisher Scientific), as required. Because IRAK1 undergoes degradation following phosphorylation, we examined β-actin expression for loading control instead of total IRAK1.

Dissected amnion from human and rhesus and the whole placenta from mice were homogenized and the tissue extract incubated with anti-IRAK1 (4504; Cell Signaling Technology) and magnetic Sepharose beads (Thermo Fisher Scientific) in immunoprecipitation buffer containing 0.1% SDS (per the manufacturer’s instructions). The precipitates were resolved by SDS-PAGE and subjected to immunoblot analyses with Abs against anti-K63 (Apu3, 05-1308; MilliporeSigma) or anti-K48 linkage–specific polyubiquitin Ab (Apu2, 05-1307; MilliporeSigma).

Total RNA was extracted from uterus after homogenizing in TRIzol (Invitrogen). RNA concentration and quality were measured by NanoDrop spectrophotometer (Thermo Fisher Scientific). Reverse transcription of the RNA was performed using qScript One-Step qRT-PCR Kit (Quantabio), following the manufacturer’s protocol. Quantitative PCR was run in a QuantStudio 3 Real-Time PCR System (Thermo Fisher Scientific). Quantitative PCR assays were performed with rhesus TaqMan Gene Expression primers (PTGS2, Rh02787802_m1; Invitrogen). Eukaryotic 18S rRNA (Invitrogen) was used as endogenous control for normalization. The values were expressed relative to the average value of the control group.

All animal procedures were approved by the Institutional Animal Care and Use Committee at the Cincinnati Children’s Hospital and Medical Center. All animal procedures done on rhesus monkeys were approved by the Institutional Animal Care and use Committee at the University of California, Davis. Pregnant women provided a written informed consent from 2014 to 2017 under a protocol approved by the Institutional Review Boards of Cincinnati Children’s Hospital and University of Cincinnati (no. 2013-2243).

SPSS v21.0 (IBM, Chicago, IL) and GraphPad Prism v8 software (GraphPad, San Diego, CA) were used to analyze data. Values were expressed as means ± SEM. Data were tested for normality (Kolmogorov–Smirnov test). Results that passed the normality assumption were analyzed by an unpaired two-tailed Student t test. Data that failed the normality assumption were analyzed by the nonparametric Mann–Whitney U test with post hoc analysis and Fisher exact test for categorical variables was used to determine differences between groups. Results were considered significant for p ≤ 0.05.

The human fetal membrane consists of inner amnion, middle chorion (fetal components), and outer decidua parietalis (maternal component). Phosphorylation of IRAK1 is a critical event that determines downstream signal transduction following activation of TLRs and receptors of the IL-1 family. We thus compared phosphorylation status of IRAK1 (p-IRAK1) in human fetal membranes harvested after preterm delivery in those subjects who had IUI or chorioamnionitis and those without (n = 4–6 per group) (Table I). We found minimal activation of IRAK1 in amnion (Fig. 1A) and decidua parietalis (Fig. 1C) in subjects without chorioamnionitis, which was substantially increased in those with chorioamnionitis. In addition, there was baseline activation of IRAK1 seen in the chorion layer in preterm subjects without chorioamnionitis (Fig. 1B), but significantly higher activation was seen in those with chorioamnionitis. Similarly, there was significant activation seen in the whole thickness placenta of those with chorioamnionitis (Fig. 1D). However, the abundance of IRAK1 activation was much less in the decidua parietalis layer of the fetal membranes compared with that seen in amnion or chorion layers in both with and without chorioamnionitis (Fig. 1).

Because we found IRAK1 to be phosphorylated in fetal membranes obtained from preterm births, we examined acute IRAK1 activation following exposure to canonical TLR ligands. We harvested the fetal membranes from women delivering at term via C-section and those not in labor. We then examined IRAK1 activation in LPS-treated cultured explants in amnion and decidua given the significant activation seen in preterm human subjects with chorioamnionitis (n = 3). In the early phase of LPS treatment (10 and 30 min), there was significant IRAK1 phosphorylation and degradation compared with controls in the amnion (Fig. 2A) and less in decidua (Fig. 2B). These results were similar to those seen in human preterm subjects with chorioamnionitis (Fig. 1), suggesting a potential upstream activation of TLRs. These results prompted us to directly examine the role of IRAK1 in preterm birth induced by intrauterine LPS administration.

In our previously established model of preterm rhesus IUI that is comparable to human IUI (31), IA LPS was given, and the fetus was delivered 16 or 48 h later (Table II). Similar to humans, we observed a significant increase of p-IRAK1 following LPS treatment at 16 h, which persisted at 48 h in the amnion, chorion, and the full thickness placenta (Fig. 3A). We did not see significant phosphorylation of IRAK1 in the decidua layer, thus replicating the observations in human decidua (Fig. 1). Given the significant increase in the activation status of IRAK1 seen in human and rhesus amnion, we then further analyzed the downstream mediators of IRAK1 in the rhesus amnion and found enhanced phosphorylation of TAK1 and JNK at 16 h, following LPS administration (Fig. 3B). This phosphorylation persisted at 48 h, whereas the downstream mediator p38 MAPK was elevated at 16 h after LPS injection and decreased at 48 h (Fig. 3B).

We also analyzed the placenta of mice after intrauterine LPS injection for IRAK1 activation, and compared with saline control, we saw significant increase in p-IRAK1 with LPS at 12 h, which persisted at 24 h (Fig. 3C).

We infused Ultrapure LPS (250 μg/mouse) into the uterus of embryonic day 15.5 mice via minilaparotomy, and the dosage of LPS was selected based on a previous study (29). Preterm birth was measured as delivery within 24 h of LPS injection. As the dosage of the IRAK1 inhibitor treatment increased, we saw a significant reduction in frequency of preterm birth and increase in live birth. We also noticed increased maternal death due to dystocia, which was significantly reduced with increased dosage of the IRAK1 inhibitor treatment (Fig. 4A, 4B). We saw similar results in the IRAK1 knockout mice when compared with wild-type (WT) mice that were administered intrauterine LPS, with a significant reduction in preterm birth and maternal death as well as an increase in mice giving live birth (Fig. 4B). Given the increased rate of maternal death due to dystocia, we repeated the experiment using a lower 10 μg of LPS as used in other studies (n = 5–6 per group) (30). Interestingly, in IRAK1 knockout mice, no preterm delivery was seen, and all these animals delivered live pups at the end of gestation (Fig. 4C). These data clearly establish that a lower dosage of LPS causes an IRAK1-dependent inflammatory response that results in preterm birth without causing maternal dystocia and IRAK1 knockout mice are completely protected from such LPS-induced preterm delivery.

Once IRAK1 is phosphorylated, it undergoes modification/degradation via ubiquitination. A delicate balance of IRAK1 activation is mediated by E3 ubiquitin ligase with phosphorylation of lysine at 63 (K63) indicating activation, whereas K48 phosphorylation targets IRAK1 for proteasomal degradation (32). Because of predominant IRAK1 activation in the amnion of women delivering preterm with chorioamnionitis, we examined K63-linked IRAK1 ubiquitination in the amniotic tissues by ubiquitination assays. We found significant increases in K63 linkage–specific IRAK1 ubiquitination in women delivering preterm with chorioamnionitis as compared with no-chorioamnionitis subjects. In contrast, there was no significant difference between chorioamnionitis and no-chorioamnionitis groups for K48 linkage–specific IRAK1 ubiquitination that accounts for IRAK1 degradation seen (Fig. 5A). In the rhesus model of IUI, a significant increase in K63 linkage–specific IRAK1 ubiquitination in the amnion at 16 h after IA LPS was observed (Fig. 5B). Similarly, in the mouse model of IUI, we observed an increase in K63 linkage–specific IRAK1 ubiquitination in LPS-infused experimental group (Fig. 5B). Overall, these data establish that IRAK1 is functionally activated by upstream signals, and activation of IRAK1 is tightly linked to the preterm births observed in humans and experimental models.

PGs are potent dilators of cervix and uterotonic agents leading to preterm birth (33). To identify the mechanism(s) by which IRAK1 activation leads to preterm birth, we examined for COX-2 expression, which is a rate-limiting enzyme for PG synthesis. In our rhesus model, we found a significant increase in COX-2 mRNA expression in the uterus 16 h after LPS injection (Fig. 6A), consistent with increases in PGs PGE2 and PGF2α seen in the amniotic fluid after IA LPS (31). We then evaluated COX-2 protein expression in our mouse IUI model. We saw a significant increase in COX-2 levels in the uterus of WT mice, 12 h after intrauterine LPS injection compared with PBS control group. However, COX-2 levels were significantly decreased in the IRAK1 knockout mice after intrauterine LPS administration when compared with WT mice (Fig. 6B). These data provide a direct mechanistic link between activation of IRAK1 and preterm birth. Furthermore, we can conclude that upstream TLR/IL-1R signaling leads to premature production of COX-2, which are known to induce PG synthesis that act on uterus and cause preterm birth.

Birth reflects transition from a propregnancy state of immunological tolerance toward the fetus allograft to a prolabor, proinflammatory state. It has been suggested that although term labor is a physiological activation of this pathway, preterm labor is the result of a pathological activation of the same pathway. Many causes of preterm birth have been identified. However, latest evidence suggests that inflammation plays a significant role in all types of labors, regardless of the presence of infection, the timing of the delivery, or other causes (25).

There have been several studies that have correlated the increase in proinflammatory cytokines, principally IL-1β, with preterm birth in humans as well as several animal species (20) as well as with spontaneous delivery at term in humans (34). It is thus thought that IL-1 overproduction heralds labor, regardless of the presence of infection (25). Moreover, elevated IL-1β blood concentration in human neonates has been associated with preterm birth (35). IL-1β concentration and bioactivity increases in amniotic fluid of women with preterm labor and infection, and elevated maternal plasma levels of IL-1β are associated with preterm labor (36). As such, IL-1β is now considered a key mediator of inflammation in preterm labor, and IL-1 is a potential therapeutic target for chorioamnionitis and associated morbidities (31).

IRAK1 is central in engaging immune signaling pathways and is a key signaling mediator in the TLR/IL-1 pathway. Although there are four different IRAK proteins that participate in downstream signaling, IRAK1 has been recently shown to be the most prominent molecule in humans (24). Furthermore, we have shown that IRAK1 is critical for linking TLRs to inflammasome activation that results in IL-1β production (26). Our focus on IRAK1 was prompted by these studies, and we began with a very novel finding of increased activation of IRAK1 in both human preterm fetal membrane with chorioamnionitis as well as in rhesus and mouse models of IUI. We did not observe IRAK1 activation in the amnion and decidua of preterm subjects without chorioamnionitis, but we saw increased abundance of p-IRAK1 in the chorion and the whole thickness placental layer of patients that had given preterm birth without chorioamnionitis. Studies on TNF/IL-1β–induced fetal membrane weakening have demonstrated that the initial tissue and cellular targets of these agents are in the choriodecidua component of the fetal membrane rather than the amnion (37). Thus, it is likely that there is some degree of IRAK1-mediated inflammation even during nonchorioamnionitis preterm birth at the chorion layer, signifying the importance of IRAK1 signaling in all preterm births. However, during active inflammation, ligands act on the choriodecidua layer, which overwhelmingly releases proinflammatory cytokines via the IRAK1 pathway during chorioamnionitis, acting on all the fetal membranes layers, which then leads to its weakening and subsequent preterm birth (31).

We examined the activation of IRAK1 in rhesus macaques due to the similarities in reproductive biology and placenta and fetal immunology between human and rhesus (38, 39). In our previously established model of IUI, the histology and inflammatory mediators after IA LPS infusion in the rhesus is comparable to women delivering preterm infants with histologic chorioamnionitis. Also, the choriodecidual plane was the histological location of neutrophil invasion in both human and rhesus IUI (31). Similar to humans, we saw activation of IRAK1 with LPS, which, in amnion, decreased at 48 h but persisted in chorion and the placenta. This supports that the activation of IRAK1 at the chorion layer propagates the inflammatory cascade and enhances the detrimental effects of the downstream inflammatory mediators.

The intrauterine injection of LPS to cause inflammation and induce preterm labor in mice is thought to mimic inflammation due to intrauterine infection (40). IL-1 blockade (41) and TLR-4 blockade (42, 43) decreased inflammation-induced preterm birth in studies using rodent models. However, in some studies using IL-1R knockout mice, no reduction in inflammation-induced preterm birth was seen, likely because of compensatory mechanisms (44). Thus, our focus in the current study was to identify specific mechanisms of IUI that may lead to preterm birth, as direct targeting of effector molecules of inflammation such as by IL-1R antagonist might incur detrimental side effects due to simultaneous inhibition of homeostatic actions. IRAK1 inhibitor and knockout mice have been shown to reduce LPS-induced kidney and renal damage (23) as well as reduce mortality from sepsis (45). We found that using IRAK1 inhibitor treatment and in IRAK1 knockout mice, preterm birth and maternal death because of dystocia was significantly reduced, and increase in live birth was observed. The IRAK1 inhibitor also attenuates the activity of IRAK4, but the lack of preterm birth in IRAK1 knockout mice further implicate IRAK1 as a critical mediator of inflammation-induced preterm birth. Increased mortality due to dystocia was a surprise finding, as previous model using 250 μg intrauterine LPS did not show any mortality (29). This difference could be due to the type of LPS or the strain of mice used. In addition, the higher dose of LPS could have contributed to preterm birth in the knockout model, as previous studies have shown increased frequency of preterm birth with higher doses intrauterine injection of heat-killed E. coli in TLR4 knockout mouse models of preterm birth (46). Reassuringly, we found no incidence of preterm birth or maternal mortality with the lower dosage of LPS in the IRAK1 knockout mice IUI model.

Targeting of E3 ligases and subsequent inhibition of linkage-specific ubiquitination of IRAK1 would be an alternative therapeutic approach to prevent inflammation. Following engagement of immune receptor with relevant ligand, IRAK1 undergoes phosphorylation that triggers Pellino 2, an E3 ligase in promoting IRAK1 activation via K63-linked ubiquitination. This is followed by β-TrCP (E3 ligase)–dependent, K48-linked ubiquitination and degradation/inactivation of IRAK1 (47). Further, Pellino 3b (E3 ligase)–mediated, K63-linked ubiquitination competes with β-TrCP–dependent, K48-linked IRAK1 ubiquitination at the same ubiquitination site (Lys134) of IRAK1 and thereby negatively regulates IL-1–induced NF-κB activation (48). Together, ubiquitination by various E3 ligases regulates the stability and actions of IRAK1. We have previously shown that suppression of IRAK1 ubiquitination is as a novel mechanism for restraining acute inflammatory reactions (32). In our study, we found increased upregulation of K63 linkage–specific IRAK1 ubiquitination in the fetal membrane of preterm women with chorioamnionitis and rhesus and mouse models of IUI. Thus, IRAK1 is regulated via E3 ligase action in IUI, and preterm birth and targeting of such E3 ligases could be a potential therapeutic option to suppress the inflammatory mediator actions in preterm birth.

COX-2–mediated PGs are important mediators for cervical dilation, uterine contraction, and preterm birth, and COX-2 inhibition prevents inflammation-induced preterm labor (33, 49). We saw COX-2 levels were significantly increased after LPS in both the rhesus and mice IUI models but decreased in IRAK1 knockout mice. These data suggest a novel association of IRAK1 and COX-2. Thus, IRAK1 likely induces preterm birth by premature and increased expression of COX-2 enzyme.

In summary, our results suggest activation of IRAK1 in the amnion and chorion of women delivering preterm with IUI, in chorion of those delivering preterm without IUI as well as in human ex vivo membranes, and rhesus and mouse models of IUI. Furthermore, K63-linked ubiquitination that signifies activation of IRAK1 was substantially increased in fetal membranes of humans, rhesus monkeys, and mice subjected to IUI. Inhibition of IRAK1 in mice by an inhibitor or by gene knockout significantly reduced preterm birth and increased live birth. More strikingly, we found that IRAK1 knockout mice had reduced expression of COX-2 following IUI, thus establishing a mechanistic and direct link between IRAK1 signaling and preterm birth. These data overwhelmingly suggest that IRAK1 is a critical mediator in inflammation-induced preterm birth and is a potential therapeutic target for IUI or chorioamnionitis and preterm birth.

We thank Dr. Alan Jobe for thoughtful insight and mentorship for this project. We would like to acknowledge all the mothers who participated in the study.

The authors have no financial conflicts of interest.