Human Fetal Membranes Generate Distinct Cytokine Profiles in Response to Bacterial Toll-Like Receptor and Nod-Like Receptor Agonists

ABSTRACT

Bacterial infection-associated inflammation is thought to be a major cause of preterm premature rupture of membranes. Proinflammatory cytokines, such as interleukin 1B (IL1B), can weaken fetal membranes (FM) by upregulating matrix metalloproteinases and inducing apoptosis. The mechanism by which infection leads to inflammation at the maternal–fetal interface and subsequent preterm birth is thought to involve innate immune pattern recognition receptors (PRR), such as the Toll-like receptors (TLR) and Nod-like receptors (NLR), which recognize pathogen-associated molecular patterns (PAMPs). The objective of this study was to determine the cytokine profile generated by FMs in response to the bacterial TLR and NLR agonists peptidoglycan (PDG; TLR2), lipopolysaccharide (LPS; TLR4), flagellin (TLR5), CpG ODN (TLR9), iE-DAP (Nod1), and MDP (Nod2). PDG, LPS, flagellin, iE-DAP, and MDP triggered FMs to generate an inflammatory response, but the cytokine profiles were distinct for each TLR and NLR agonist, and only IL1B and RANTES were commonly upregulated in response to all five PAMPs. CpG ODN, in contrast, had a mild stimulatory effect only on MCP-1 and primarily downregulated basal FM cytokine production. IL1B secretion induced by PDG, LPS, flagellin, iE-DAP, and MDP was associated with its processing. Furthermore, FM IL1B secretion in response to TLR2, TLR4, and TLR5 activation was caspase 1-dependent, whereas Nod1 and Nod2 induced IL1B secretion independent of caspase 1. These findings demonstrate that FMs respond to different bacterial TLR and NLR PAMPs by generating distinct inflammatory cytokine profiles through distinct mechanisms that are specific to the innate immune PRR activated.

Human fetal membranes respond to different bacterial Toll-like receptor and Nod-like receptor agonists by generating specific and distinct inflammatory cytokine profiles through distinct mechanisms.

INTRODUCTION

Preterm premature rupture of membranes (PPROM), defined as rupture of membranes prior to 37 wk, complicates 2%–4% of singleton and 7%–20% of twin pregnancies [1]. PPROM is the leading identifiable cause of preterm birth and accounts for 18%–20% of all perinatal deaths in the United States [2]. Preterm birth affects 11.7% of live pregnancies in the United States [3], and PPROM occurs in 25%–30% of these births [4–6]. Although risk factors for PPROM have been described, little is known about the mechanisms responsible for fetal membrane (FM) weakening and rupture. Given that FMs typically rupture over the cervix [7, 8] and that PPROM is more common in the setting of intrauterine infection [1, 2, 9], it is likely that inflammation serves as the final common pathway leading to PPROM and its sequelae, preterm birth [10, 11]. An intrauterine bacterial infection can gain access to FMs primarily by ascending into the uterus from the lower tract but also by descending into the uterus from the peritoneal cavity or via maternal circulation [12]. The way in which infection can lead to prematurity is thought to involve innate immune responses toward the pathogen, mediated by pattern recognition receptors (PRRs), leading to inflammation [6]. Furthermore, proinflammatory cytokines such as tumor necrosis factor alpha (TNFA) and interleukin 1beta (IL1B) are thought to damage FMs by inducing mediators of membrane weakening such as matrix metalloproteinases, prostaglandins, and proapoptotic proteins [4, 13–22].

Two main families of PRRs have been described: the Toll-like receptors (TLRs) and the Nod-like receptors (NLRs). TLRs are transmembrane proteins that can recognize pathogen-associated molecular patterns (PAMPs) either at the extracellular level or within endosomes [23, 24]. The extracellular domain of each TLR corresponds to a specific PAMP, such that the collective TLR family can recognize a range of pathogens. In terms of bacterial PAMPs, TLR2, in cooperation with TLR1, TLR6, or TLR10, senses Gram-positive peptidoglycan (PDG) as well as other bacterial PAMPs [25]. TLR4 is the receptor for Gram-negative bacterial lipopolysaccharide (LPS) [26, 27], TLR5 is the receptor for bacterial flagellin, and TLR9 recognizes unmethylated bacterial DNA CpG motifs [28]. The primary adapter protein involved in TLR signaling in response to bacteria is myeloid differentiation factor 88 (MyD88), which triggers downstream NFκB and mitogen-activated protein kinase (MAPK) activity [29]. However, because of the use of coreceptors and divergence in intracellular signaling, distinct responses can be generated depending upon the stimuli used, the TLR activated, and the cell type involved [30].

NLRs, which are cytoplasm-based, include sensors of bacterial PAMPs and infections such as the Nod proteins Nod1 and Nod2, which recognize peptides derived from degradation of bacterial peptidoglycan [31, 32], and the inflammasome components Nalp1 and Nalp3, which are involved in IL1B production [33, 34]. Activation of Nod1 by γ-d-glutamyl-meso-diaminopimelic (iE-DAP) acid, or of Nod2 by muramyl dipeptide (MDP), triggers a signaling cascade through the common adapter protein receptor-interacting serine/threonine-protein kinase 2 (RICK), resulting in NFκB and MAPK and an inflammatory cytokine/chemokine response [35–38]. Activation of the Nalp protein Nalp1 (NLRP1) or Nalp3 (NLRP3) leads to formation of the inflammasome, a protein platform also containing apoptosis-associated speck-like protein containing a CARD (ASC), and caspase 1, which mediates the processing of intracellular pro-IL1B into its active, secreted form [33, 34]. However, unlike the TLRs and Nod proteins, specific bacterial agonists have not been identified for Nalp1 or Nalp3.

Normal human term FMs stimulated in vitro with various bacteria or bacterial components are known to produce elevated levels of proinflammatory cytokines (IL1B, IL6, IL8, TNFA) [39–45], a profile also associated clinically with PPROM and preterm birth [4, 6, 21]. These studies suggest that, like other gestational tissues, the chorioamniotic membranes can sense and respond to microbes through PRRs [30]. Indeed, TLRs and the Nod proteins Nod1 and Nod2 are known to be expressed by FMs [42, 46–49]. Moreover, animal models have demonstrated roles for TLRs and NLRs in infection-induced preterm birth [50–56]. However, little is still known about the precise function of these proteins in the chorioamnion. Although some investigators have assessed the response of TLR and Nod protein activation in FMs, those studies have measured only a limited panel of cytokines, and most measurements were performed on either separated FMs or on isolated chorionic or amnionic cells [42, 48, 57–60]. Because, similar to other gestational tissues [30, 61], FMs may generate different cytokine profiles in response to different PRR agonists, the objective of the current study was to perform a comprehensive characterization of the full cytokine profile generated by intact FM explants after specific TLR and NLR activation by bacterial components.

MATERIALS AND METHODS

Fetal Membrane Collection, Preparation, and Stimulation

Fetal membranes (n = 10) were collected from uncomplicated, normal, term pregnancies (39–41 wk) delivered by elective cesarean section, without signs of labor, infection, or premature rupture of membranes. No patients received prostaglandins or any other induction agent prior to cesarean section. Sample collection was approved by Yale University's Human Investigation Committee. After the FMs were washed with sterile PBS supplemented with penicillin (100 U/ml) and streptomycin (100 μg/ml; Gibco, Grand Island, NY), adherent blood clots were removed, and sections where both the chorion and amnion were intact were cut using a 6-mm biopsy punch. The explants were placed in 0.4-μm cell culture inserts, chorion facing up (BD Falcon, Franklin Lakes, NJ) with 500 μl of Dulbecco modified Eagle medium (DMEM; Gibco) supplemented with 10% fetal bovine serum (FBS; Hyclone, Logan, UT), and these were place in a 24-well plate containing 500 μl of DMEM for 24 h, as previously described [19]. The next day, the medium was removed and replaced with serum-free OptiMeM medium (Gibco). FM explants were then treated with no treatment (NT) or the following agonists: the TLR2 agonist Gram-positive bacterial PDG from Staphylococcus aureus at 10 μg/ml (Invivogen, San Diego, CA); the TLR4 agonist Gram-negative LPS from Escherichia coli O111:B4 at 100 ng/ml (Sigma-Aldrich, St. Louis, MO) L2630); the TLR5 agonist Gram-negative bacterial flagellin from Salmonella typhimurium at 1 μg/ml (Invivogen); the TLR9 agonist bacterial CpG ODN at 5 μM (code 2216; Invivogen); the Nod1 agonist Gram-positive/negative bacterial iE-DAP at 100 μg/ml (Invivogen); or the Nod2 agonist Gram-positive/negative bacterial MDP at 10 μg/ml (Invivogen). In some experiments, the FM explants were pretreated for 1 h with the specific caspase 1 inhibitor Z-WEHD-FMK at 1 μM (R&D Systems, Minneapolis, MN). The optimal TLR/NLR agonist concentrations were determined in preliminary experiments (data not shown). After 24 h of treatment with TLR/NLR agonists, cell-free culture supernatants were collected, and explants were snap frozen. Supernatants and tissues were then stored at −80°C until further analysis was performed. Because of the varying amounts of usable tissue and, consequently, the number of punch biopsies obtained from each specimen, not all treatments were performed for some FM preparations.

Cytokine Analysis

FM supernatants were analyzed by ELISA for IL1B (R&D Systems) and IL8 (Assay Designs, Farmingdale, NY). Supernatants were further measured for the following cytokines/chemokines by using multiplex analysis (Bio-Rad, Hercules, CA): IL2, IL4, IL6, IL10, IL12, IL17, G-CSF (CSF3), GM-CSF (CSF2), IFNG, MCP-1 (CCL2), MIP1A, MIP1B, RANTES (CCL5), TNFA, VEGF, and GROα as previously described [19].

Western Blot Analysis

FM explant biopsies were homogenized and proteins extracted as previously described [19]. Protein concentrations were measured using BCA protein assay (Pierce, Rockford, IL). Samples were diluted with gel loading buffer and boiled for 5 min, after which they were resolved under reducing conditions on SDS-polyacrylamide gels and then transferred onto polyvinylidene membranes (PerkinElmer, Boston, MA). Membranes were blocked with either 5% fat-free powdered milk or 5% goat serum in PBS-0.05% Tween-20 (PBS-T). After being washed with PBS-T, membranes were incubated overnight at 4°C with primary antibody in either PBS-T-1% fat-free powdered milk or PBS-T-1% goat serum. Membranes were probed with the following anti-human primary antibodies: active IL1B rabbit polyclonal (code 2022; Cell Signaling, Boston, MA); ASC rabbit polyclonal (code ST1121; Calbiochem, Gibbstown, NJ); caspase 1 rabbit polyclonal (code 06–503; Millipore, Billerica, MA); Nod1 rabbit polyclonal (code AL184; Alexis Biochemicals, San Diego, CA); Nod2 mouse monoclonal antibody (mAb; 2D9; eBioscience, San Diego, CA); Nalp1 rabbit polyclonal (code 4990; Cell Signaling); Nalp3 rabbit mAb (code 13158; Cell Signaling); MyD88 rabbit polyclonal (code 14–6223; eBioscience); TIR domain-containing adapter protein-induced IFNB (TRIF) rabbit polyclonal (code 4596; Cell Signaling); or RICK rabbit polyclonal (code 14–6275; eBioscience) as previously described [52, 62–66]. Following incubation with primary antibody, membranes were washed as described above and then incubated with either the horse anti-mouse IgG or goat anti-rabbit IgG secondary antibody conjugated to peroxidase (Vector Labs, Burlingame, CA). Following washes with PBS-T and then with distilled water, the peroxidase-conjugated antibody was detected by enhanced chemiluminescence (PerkinElmer). β-Actin was used as a loading control (Sigma). Images were recorded using Gel Logic 100 (online) and Kodak MI software (Eastman Kodak, Rochester, NY).

Quantitative Real Time RT-PCR

FM explant biopsies were homogenized, and total RNA was extracted as previously described [49, 67]. Quantitative real-time PCR was performed using KAPA SYBR Fast qPCR kit (Kapa Biosystems, Woburn, MA), and PCR amplification performed with the Bio-Rad CFX Connect real time system (Bio-Rad). Detection of human TLR1–10 was performed using primer sequences previously described [67]. For relative abundance of TLR mRNA levels, expression was normalized to an internal positive control and GAPDH (Invivogen) [49, 67].

Statistical Analysis

Experiments were performed at least three times and data are means ± SEM. Prism software (Graphpad Software, Inc., La Jolla, CA) was used to calculate significance (P < 0.05). Statistical analysis was performed using either the paired t-test or, for multiple comparisons, one-way ANOVA.

RESULTS

Fetal Membranes Express TLRs, NLRs, and Associated Adapter Signaling Proteins

Normal, term, unstimulated human FMs expressed all 10 TLRs at the mRNA level (Fig. 1A). At the protein level, FMs expressed the NLRs Nod1, Nod2, Nalp1, and Nalp3 as well as the TLR adapter proteins MyD88 and TRIF, the Nod1/Nod2 adapter protein RICK, and the Nalp1/Nalp3 inflammasome components ASC and caspase 1 (Fig. 1B).

FIG. 1.

Human fetal membranes express TLRs, NLRs, and their associated signaling proteins. A) FM explants (n = 3) were evaluated for expression of TLR1–10 mRNA by using qRT-PCR. Bar chart shows relative abundance after normalization to the housekeeping gene and a positive control. B) FM explants (n = 4) were evaluated for Nod1, Nod2, Nalp1, Nalp3, MyD88, TRIF, RICK, ASC, caspase 1, and β-actin protein expression by Western blotting.

Fetal Membranes Generate Distinct Cytokine Profiles in Response to TLR and NLR Bacterial Agonists

The bacterial components PDG, LPS, flagellin, iE-DAP, and MDP are agonists for TLR2, TLR4, TLR5, Nod1, and Nod2, respectively. FMs treated with PDG secreted significantly higher levels of IL1B, IL6, IL8, IL10, G-CSF, MIP1A, and RANTES than the NT control (Fig. 2 and Table 1). All other cytokines tested were not significantly changed after PDG treatment (Table 1). FMs treated with LPS secreted significantly higher levels of IL1B, IL2, IL8, IL10, RANTES, TNFA, and GROA than the NT control (Fig. 3 and Table 1). Exposure of FM explants to flagellin significantly upregulated the tissue's secretion of IL1B, IL6, IL8, G-CSF, MIP1A, MIP1B, RANTES, TNFA, and GROA (Fig. 4 and Table 1). Although activation of TLR2, TLR4, and TLR5 induced cytokine production by FM explants, treatment with the TLR9 agonist CpG ODN had a differential response. As shown in Figure 5 and Table 1, treatment of FM explants with CpG ODN significantly upregulated the secretion of MCP-1 and, concurrently, significantly decreased the basal production of G-CSF, IFNG, MIP1A, MIP1B, RANTES, and VEGF.

FIG. 2.

Cytokine profile of fetal membranes is shown after activation of TLR2 by Gram-positive bacterial PDG. FM explants treated with PDG (10 μg/ml) for 24 h secreted significantly higher levels of IL1B, IL6, IL8, IL10, G-CSF, MIP1A, and RANTES than the NT control (*P < 0.05; n = 5).

TABLE 1.

Fetal membrane cytokine profiles in response to TLR2, TLR4, TLR5, TLR9, Nod1, and Nod2 activation by PDG, LPS, Flagellin, CpG ODN, iE-DAP, and MDP, respectively.

^*Up arrow indicates fold-change increase relative to the untreated control (P < 0.05), down arrow indicates a decrease, and a dash indicates no significant change or a level below the assay's detection limit.

FIG. 3.

Cytokine profile of fetal membranes is shown after activation of TLR4 by Gram-negative bacterial LPS. FM explants treated with LPS (100 ng/ml) for 24 h secreted significantly more IL1B, IL2, IL8, IL10, RANTES, TNFA, and GROA than the NT control (*P < 0.05, **P < 0.001; n = 10).

FIG. 4.

Cytokine profile of fetal membranes is shown after activation of TLR5 by bacterial flagellin. FM explants treated with flagellin (1 μg/ml) for 24 h secreted significantly more IL1B, IL6, IL8, G-CSF, MIP1A, MIP1B, RANTES, TNFA, and GROA than the NT control (*P < 0.05; n = 5).

FIG. 5.

Cytokine profile of fetal membranes is shown after activation of TLR9 by CpG ODN. FM explants treated with CpG ODN (5 μM) for 24 h secreted significantly more MCP-1 and significantly less G-CSF, IFNG, MIP1A, MIP1B, RANTES, and VEGF than the NT control (*P < 0.05; n = 5).

The profile of cytokines produced by FM explants in response to Nod1 and Nod2 activation was also distinct. iE-DAP treatment of FMs resulted in a significant increase in the secretion of IL1B, IL2, IL6, IL8, IFNG, MCP-1, G-CSF, MIP1A, MIP1B, and RANTES (Fig. 6 and Table 1), whereas MDP treatment induced elevated secretion of IL1B, IL2, IL6, IFNG, G-CSF, MIP1A, MIP1B, RANTES, TNFA, and GROA (Fig. 7 and Table 1). When all the cytokine profiles generated by the TLR/NLR agonists were compared, only 2 cytokines were common to all agonists except CpG ODN: IL1B and RANTES. As shown in Table 1, LPS and PDG induced higher levels of IL1B and RANTES than the levels induced by flagellin, iE-DAP, and MDP.

FIG. 6.

Cytokine profile of fetal membranes is shown after activation of Nod1 by bacterial iE-DAP. FM explants treated with iE-DAP (100 μg/ml) for 24 h secreted significantly more IL1B, IL2, IL6, IL8, IFNG, MCP-1, G-CSF, MIP1A, MIP1B, and RANTES than the NT control (*P < 0.05; n = 7).

FIG. 7.

Cytokine profile of fetal membranes is shown after activation of Nod2 by bacterial MDP. FM explants treated with MDP (10 μg/ml) for 24 h secreted significantly more IL1B, IL2, IL6, IFNG, G-CSF, MIP1A, MIP1B, RANTES, TNFA, and GROA than the NT control (*P < 0.05; n = 7).

TLRs and NLRs Activate FM IL1B Processing

IL1B, unlike most other cytokines, is first produced as a proprotein and is then proteolytically cleaved into its active mature form. In this way, secretion of this highly proinflammatory cytokine is tightly regulated [68]. Western blot analysis was used to examine whether IL1B processing occurred in the FM explants after exposure to the TLR and Nod protein agonists. As shown in Figure 8A, FM tissue that had not been treated (NT) expressed little if any active IL1B (17 kDa). However, after stimulation with PDG, LPS, flagellin, iE-DAP, or MDP, there was greater expression of active IL1B (Fig. 8A).

FIG. 8.

TLRs and NLRs induce IL1B processing and secretion in a caspase-1-dependent and -independent manner. A) FM explants were treated with PDG, LPS, flagellin, iE-DAP, or MDP for 24 h after which lysates were analyzed for active IL1B (17 kDa) expression by Western blotting (n = 3). As shown in the representative blot, the NT control had no detectable levels of active IL1B, but active IL1B expression was induced after treatment with PDG, LPS, flagellin, iE-DAP, and MDP. B) FM explants (n = 3) were treated for 24 h with PDG, LPS, flagellin, iE-DAP, or MDP in the presence of media or a caspase 1 inhibitor (1 μM). IL1B secretion was significantly induced after exposure to all bacterial products compared to that of the NT control (*P < 0.05, **P < 0.001). The presence of the caspase 1 inhibitor significantly reduced PDG-, LPS-, and flagellin-induced IL1B secretion (*P < 0.05) but not iE-DAP- or MDP-induced IL1B (ns, not significant).

FM IL1B Secretion in Response to TLR Activation is Caspase 1-Dependent and Caspase 1-Independent in Response to Nod Protein Activation

The most common mediator of IL1B processing is the nonapoptotic caspase 1 [69]. Therefore, its role in FM IL1B production in response to the TLR and NLR agonists was tested. In the presence of a caspase 1 inhibitor, IL1B secretion induced by PDG, LPS, or flagellin was significantly reduced by 72.7% ± 15.5%, 56.5% ± 22.6%, and 16.0% ± 8.9%, respectively (Fig. 8B). In contrast, the presence of the caspase 1 inhibitor had no significant effect on the levels of IL1B secreted in response to either iE-DAP or MDP (Fig. 8B).

DISCUSSION

PPROM and subsequent preterm birth have been strongly associated with bacterial infection and inflammation [1, 2, 9–11]. Although FMs express PRRs such as TLRs and Nod proteins [42, 46–49] and the ability of FMs to generate an inflammatory response toward bacteria or bacterial components is known [39–45], the precise role of innate immune PRRs is not fully appreciated. Furthermore, previous studies in which the function of PRRs has been investigated in the FMs have been limited by their focus on a small panel of factors and/or by studying isolated cells rather than intact tissue [42, 48, 57–60]. Because the chorion and amnion are normally in direct contact with each other and act as a physiologic unit, such communication may be an important factor governing the specific type of response generated. In the current study, we have demonstrated the basal expression of TLR1–10 and NLRs and their associated signaling proteins by normal human term FMs. Moreover, using an in vitro FM explant system, we established a broad cytokine profile secreted by the FMs in response to bacterial PAMPs that specifically activate TLR2, TLR4, TLR5, TLR9, Nod1, and Nod2. All six PRRs are functional in human FMs and trigger distinct cytokine profiles. Furthermore, we have demonstrated a role for the inflammasome component caspase 1 in mediating FM IL1B production in response to TLR but not Nod protein activation.

TLR2 senses Gram-positive bacterial PDG in cooperation with its coreceptors TLR1, TLR6, and TLR10 [25, 50]. In this study, we demonstrated that TLR2 and its coreceptors are expressed under basal conditions by normal human term FMs. In cases of chorioamnionitis, FM TLR2 expression is elevated [47], and, in women with microbial invasion of the amniotic cavity, soluble TLR2 levels in the amniotic fluid are increased [70]. Treatment of FM with PDG resulted in a significant increase in the secretion of the proinflammatory cytokines IL1B, IL6, and G-CSF; the inflammatory chemokines IL8, MIP1A, and RANTES; and the anti-inflammatory cytokine IL10. The production of IL10 suggests, in addition to an inflammatory profile, the initiation of a regulatory anti-inflammatory response. Whereas most of the factors identified as being induced by PDG have not previously been assessed, our findings that PDG induces IL6 and IL8 secretion are in keeping with those of studies by Fortunato et al. [44] and Gillaux et al. [59], the latter of which reported that the TLR2/TLR6 agonist MALP2 induces IL6 and IL8 production by isolated amniotic epithelial cells.

Similar to TLR2, the expression of the sensor for Gram-negative bacterial LPS, TLR4, is upregulated in the FMs in patients with chorioamnionitis [47] and becomes located to the amniotic epithelial basal membrane [46]. Also similar to TLR2, activation of TLR4 by LPS induced the secretion of the inflammatory cytokines IL1B and IL6, the inflammatory chemokine RANTES, and the anti-inflammatory cytokine IL10. However, unlike the response to PDG, there was no significant elevation in IL6, G-CSF, or MIP1A secretion. Instead, we observed the upregulation of IL2, GROA, and TNFA after FM treatment with LPS. Thus, TLR2 and TLR4 activation by their respective bacterial agonists trigger distinct FM cytokine profiles, yet both responses are a combination of pro- and anti-inflammatory cytokines and chemokines. In support of our findings, previous studies using separated human chorion and amnion explants and isolated chorionic cells have reported that LPS induces TNFA [48, 58] and that intact FM explants treated with LPS secreted IL1B, TNFA, IL8, and IL10 [43]. However, in contrast to our findings, a study using chorionic cells showed that LPS induced MCP1 production [58], and a study using intact FM explants showed that LPS induced elevated levels of IFNG and IL6 [43]. These differences in the intact FM explant cytokine profiles may be a reflection of different LPS strains used. In our current study, we used E. coli LPS strain O111:B4, whereas 2 other studies showing LPS-induced IL6 production by FMs used E. coli LPS strain O55:B5 and LPS from Salmonella typhimurium [71, 72]. Moreover, a study in non-human primates reported differential cytokine responses from chorioamnion exposed to different LPS structural variants [60].

To our knowledge, only 1 study has evaluated the function of TLR5 and TLR9 in FMs by using isolated amniotic epithelial cells rather than intact tissue [59]. Gillaux et al. [59] reported that, whereas flagellin induced the secretion of IL6 and IL8, TLR9 had no such effect, and it was therefore suggested to be nonfunctional. In our studies using intact FM explants, we also found flagellin to increase the secretion of IL-6 and IL-8, as well as IL1B, TNFA, G-CSF, MIP1A, MIP1B, RANTES, and GROA, indicating an overall proinflammatory response. However, we found that TLR9 did generate an FM response, albeit atypical. Treatment with unmethylated CpG induced a mild stimulatory effect on MCP1 secretion and, in parallel, downregulated the basal secretion of G-CSF, IFNG, MIP1A, MIP1B, RANTES, and VEGF. This suggests that FM TLR9 activation may generate more of a regulatory response associated with tissue repair, rather than an inflammatory one [73, 74]. While it is possible that a longer treatment time could yield a more robust TLR9 response in FM [75], CpG motifs can promote protective innate immune responses and suppress inflammatory cytokine production [75, 76].

Together our studies show that in FMs, TLRs in response to different bacterial PAMPs can trigger varied cytokine/chemokine responses. Indeed, while the 4 TLRs examined in this study can all signal through the common adapter protein MyD88, the reality is that these TLRs have distinct signaling capacities. As already discussed, TLR2 functions by dimerizing either with itself or its coreceptors TLR1, TLR6, or TLR10 [25, 50]. TLR2 and TLR4 signal through MyD88 in cooperation with TIR-associated protein (TIRAP/Mal), and TLR4 can also signal through the other bridging adapter proteins TRIF and TRIF-related adapter protein (TRAM) [29, 77]. While TLR5 and TLR9 both signal through MyD88 alone to activate IRAK and subsequently NFκB [78, 79], TLR9 also triggers activation of IRF7 [80]. Thus, these divergences in signaling may explain the differences in cytokine profiles and in the magnitudes of responses generated by the FM explants after treatment with the specific TLR2, TLR4, TLR5, and TLR9 agonists.

Although Nod1 and Nod2 activation by iE-DAP and MDP, respectively, also induced cytokine production by the FMs, the number of factors upregulated by these NLRs was greater than after TLR2, TLR4, TLR5, or TLR9 activation. Furthermore, and similar to TLR stimulation, there were differences in the cytokine profiles generated by iE-DAP and MDP. Both Nod1 and Nod2 activation upregulated the secretion of IL1B, IL2, IL6, G-CSF, IFNG, MIP1A, MIP1B, and RANTES. However, Nod1, but not Nod2, activation also upregulated IL8 and MCP1 secretion, whereas Nod2, but not Nod1 activation, upregulated TNFA and GROA. These findings further underscore the fact that even though a common adapter protein might be involved in the signaling cascade, in this case RICK [38], the downstream responses may still differ. Although our findings support a recent study by Lappas [42] showing iE-DAP induces IL6 and IL8 secretion by FM explants, unlike this report, in our study we observed elevated IL6, but not IL8 production, in response to MDP.

Although PDG, LPS, flagellin, iE-DAP, and MDP all triggered FMs to generate a strong inflammatory cytokine response, only 2 factors were commonly upregulated, IL1B and RANTES. IL1B is a potent proinflammatory cytokine associated with mediators of membrane weakening [4, 14, 15, 17, 18, 22], PPROM, and preterm birth [81–85]. It is also a cytokine that is under tight regulation. Prior to being secreted, IL1B must first be processed from its pro form into its cleaved, active form, and this process is mediated primarily by the inflammasome component caspase 1 [69]. Indeed, elevated caspase 1 is detected in the amniotic fluid from patients with preterm labor in the presence and absence of infection, compared with those delivering at term [86]. In the current study, we demonstrated that FM tissue expresses little, if any, active IL1B under basal conditions, but when treated with the bacterial components PDG, LPS, flagellin, iE-DAP, or MDP, active IL1B expression is induced, suggesting that TLR and Nod protein activation results in IL1B processing. However, whereas TLR-induced IL1B production was reduced by the presence of a caspase 1 inhibitor, Nod1- and Nod2-induced IL1B production appeared to be independent of caspase 1. How Nod1 and Nod2 trigger IL1B processing might involve other nonapoptotic caspases such as caspase-4 or caspase-5 [34, 87]. This observation highlights why clinically the field of PPROM and preterm birth is so challenging and that different bacterial components can induce a common proinflammatory cytokine, IL1B, to trigger PPROM and preterm birth, yet the mechanisms involved are distinct.

In summary, we have demonstrated that human FMs express TLRs, NLRs, and their associated adapter signaling proteins. In response to specific bacterial PAMPs, human FMs generate specific and distinct inflammatory and regulatory cytokine profiles through distinct mechanisms that are dependent upon the type of innate immune PRR involved. The true in vivo scenario, however, represents a polymicrobial state where multiple bacterial components are present [88], and subsequently, multiple pattern recognition receptors may be simultaneously engaged. Moreover, within a single bacterium there is more than 1 TLR or NLR agonist expressed. Thus, activation of more than 1 TLR/NLR in the fetal membranes by a combination of bacterial components of varying levels may give rise to further differential cytokine profiles. Together, these findings suggest that infection-associated PPROM and subsequent preterm birth may arise through a number of different signaling pathways.