Preterm labor is associated with inflammation and infection. The mechanisms underlying the role of omega-3 fatty acid in inflammasome activation and prevention of preterm labor remain unknown. We hypothesized that omega-3 fatty acid can reduce the rate of preterm birth induced by infection and trophoblast inflammation. In the present study, we found that inflammasome-related molecules and IL-1β in trophoblasts were activated by TNF-α derived from lipopolysaccharide (LPS)-stimulated THP-1 cell-conditioned medium (CM) and recombinant TNF-α protein. The results demonstrated that stimulation with TNF-α caused lysosomal rupture in trophoblasts, which accelerated cathepsin S (CTSS) diffusion from lysosomes to the cytosol and activated NLRP1 (nacht domain-leucine-rich repeat, and pyd-containing protein 1) and absent in melanoma 2 (AIM2) inflammasomes, thereby increasing IL-1β secretion. Moreover, in response to LPS challenge, TNF-α increased trophoblast cell death and decreased cell viability through inflammasome and CTSS activation. Stearidonic acid (SDA; 18:4n−3) and docosahexaenoic acid (DHA; 22:6n−3) inhibited inflammasome-related molecule synthesis and CTSS and caspase-1 activation, which further reduced the preterm delivery rate of pregnant mice induced by LPS (92.9 compared with 69.7% (DHA); 92.9 compared with 53.5% (SDA)). Higher expression of TNF-α, IL-1β, prostaglandin E2, and CTSS, but lower resolvin D1 expression, was observed in preterm pregnant mice than in controls. Similarly, resolvin D1 was highly expressed in women with term delivery compared with women with preterm delivery. Thus, SDA and DHA may attenuate macrophage-derived TNF-α inducing CTSS and inflammasome activation, IL-1β secretion, and placental trophoblast cell death. These functions are implicated in the preventive effects of SDA and DHA on preterm labor.
Intrauterine infection is a common cause of preterm birth, and Escherichia coli is the main pathogen identified in placental infection leading to preterm birth . Innate immune responses to a pathogen are often triggered after the host recognition of microbial pathogen-associated molecular pattern molecules, such as lipopolysaccharide (LPS). Exposure to LPS increases levels of proinflammatory cytokines in animal models of preterm delivery, implicating the role of LPS in preterm birth .
In mammals, 18-carbon and longer omega-3 (18:n−3) polyunsaturated fatty acids (PUFAs) are produced from the dietary essential fatty acid α-linolenic acid (ALA; 18:3n−3) through a series of desaturation and elongation reactions catalyzed by various enzymes [3,4]. Docosahexaenoic acid (DHA; 22:6n−3) is a long-chain n−3 PUFA. Stearidonic acid (SDA; 18:4n−3) is a rate-limiting metabolic intermediate in the conversion of ALA into DHA, catalyzed by elongase and desaturase. Plant-derived SDA can be a potential alternative to marine-based n−3 PUFAs .
Epidemiological studies have revealed that n−3 PUFA intake during pregnancy exerts protective effects on the gestation length . DHA supplementation increases the gestation length and results in greater birth weight, and fewer infants were born before 34 weeks of gestation . The mean gestational age at delivery significantly increased by 4.5 days in the n−3 PUFA supplementation group compared with the control group . Dietary supplementation with n−3 PUFA during pregnancy increased fetal and placental weights at delivery in rats . However, the mechanisms underlying the lengthening of pregnancy by the n−3 PUFA (i.e. DHA) remain unknown. Furthermore, the effect of SDA in pregnancy has not been reported.
The host recognition of a pathogen results in the secretion of proinflammatory cytokines, such as TNF-α, IL-6, IL-8, and IL-1β. Amongst these proinflammatory cytokines, IL-1β is a major cytokine that activates immune cells to phagocytose the invading pathogen and induces adaptive cellular responses. IL-1β is secreted as an inactive precursor, and the processing of pro-IL-1β depends on cleavage by proteases, such as caspase-1, which in turn is activated by inflammasomes. Inflammasomes are inducible multiprotein cytosolic complexes that assemble on recognition of inflammatory signals; these complexes include proteins of the nucleotide-binding oligomerization domain (NOD)-like receptor (NLR) family  or pathogen-associated compounds . NLRs, such as NLRP1 (nacht domain-leucine-rich repeat, and pyd-containing protein 1) and NLRP3 or absent in melanoma 2 (AIM2) can bind to caspase-1 through the apoptosis-associated speck-like protein containing a caspase recruitment domain (ASC) adaptor protein . The resulting macromolecular complex recruits and activates pro-caspase-1 into caspase-1, which in turn promotes the processing and secretion of the proinflammatory cytokine IL-1β. Preterm labor resulting from infection is associated with a high concentration of caspase-1 in the amniotic fluid, suggesting that infection induces caspase-1 production and inflammasome activation. Sequential activation of inflammasomes and caspase-1 with IL-1β processing and production can be a pathway leading to the activation of parturition .
Inflammasome assembly and activation have been implicated in lysosomal damage, mitochondrial dysfunction, reactive oxygen species production, and calcium signaling. Moreover, excessive phagocytosis triggers lysosomal destabilization, resulting in lysosomal rupture, which in turn stimulates numerous cellular processes, such as cell death and inflammasome activation .
The molecular link between inflammation in preterm labor and inflammasome-mediated IL-1β release has not been well established thus far. The role of PUFAs in preventing preterm labor remains unclear. We hypothesized that n−3 PUFA inhibits preterm birth induced by infection and trophoblast inflammation. Thus, the present study investigated whether SDA and DHA exert anti-inflammatory effects in human pregnancy and parturition. We observed that exogenous recombinant TNF-α or LPS-stimulated THP-1 conditioned medium (CM) caused inflammasome activation, which was blocked by treatment of trophoblasts with SDA and DHA. Furthermore, lysosomal damage and cathepsin S (CTSS) activation triggered inflammasome signaling, in turn increasing trophoblast cell death.
Materials and methods
Cell culture and treatment
We used the human THP-1 cell line (established from a patient with acute monocytic leukemia; TIB-202, American Type Culture Collection (ATCC), Manassas, VA); the HEK-293 cell line (CRL-1573, ATCC); and trophoblast cell lines, namely JEG-3 cells (HTB-36, ATCC), and 3A-sub-E cells (CRL-1584, ATCC). THP-1 cells, HEK-293 cells, and JEG-3 cells were grown in Dulbecco`s Modified Eagle Medium (DMEM; Gibco, Life Technologies) and 3A-sub-E cells in Minimum Essential Medium (MEM; Gibco, Life Technologies), were all supplemented with 10% FBS (Biological Industries, Kibbutz Beit–Haemek, Israel), 100 U/ml penicillin, and 100 mg/ml streptomycin. The cells were incubated at 37°C and 5% CO2. THP-1 cells were differentiated with 100 nM/ml PMA (Millipore, Billerica, MA) for 24 h. These cells were challenged with various stimulators either alone or in combination as indicated. The stimulators used in the present study were LPS (100 ng/ml, serotype 0111:B4; Sigma–Aldrich, St. Louis, MO), TNF-α (700 ng/ml, PeproTech, Rocky Hill, NJ), caspase-1 inhibitor (CASP1i, 50 μM, Millipore), PUFA (SDA, 18:4n−3, 100 μM; DHA, 22:6n−3, 25 μM; dihomo-γ-linolenic acid (DGLA), 20:3n−6, 100 μM; and oleic acid (OA), 18:1n−9, 100 μM; Cayman Chemical, Ann Arbor, MI), ATP diphosphatase (apyrase, 0.5–3 U, Sigma–Aldrich), cathepsin inhibitor (CATi, 30 μM, Millipore), specific CTSS inhibitor (Z-FL-COCHO, 5–10 μM, AdooQ BioScience, Irvine, CA), specific inhibitor of vacuolar type H+-ATPase (bafilomycin A1 (Baf A1), 0.1–0.5 nM, Millipore), lysosomal acidification inhibitor (dexamethasone (DEX), 10–30 μM, Sigma–Aldrich). Before their use for treatment, all fatty acids were dissolved in ethanol. For control treatments, ethanol was used as the vehicle control.
Isolation and culture of primary trophoblasts
Isolation and purification of villous cytotrophoblasts from term placentas of women with normal pregnancies were performed, as described previously . Purified cytotrophoblasts were maintained at 37°C in a 1:1 mixture of DMEM and F12 medium (Gibco, Life Technologies) supplemented with 4% serum replacement (Sigma–Aldrich), and these cytotrophoblasts were used in the experiments described subsequently.
CM is the medium collected from cultured cells after culturing them for a specific time. THP-1, JEG-3, and 3A-sub-E cells (3 × 105) and primary trophoblasts (5 × 106) were grown in 6-cm cell culture dishes. The cells were washed with PBS and incubated in regular medium containing FBS or in serum-free medium with or without LPS stimulation. At the end of LPS treatment for 24–72 h, the CM of LPS-stimulated cells were collected and centrifuged at 1500×g to eliminate intact cells, followed by concentration by using spin columns with a molecular mass cut-off of 3 kDa (Amicon Ultra, Millipore), and the concentrated medium was stored at −80°C until use .
According to the manufacturer protocol, ELISA kits were used to assess IL-1β, IL-6, monocyte chemoattractant protein (MCP)-1, IL-8, TNF-α, prostaglandin E2 (R&D Systems, Minneapolis, MN), CTSS (MyBioSource, San Diego, CA), and resolvin D1 (Cayman Chemical) concentrations in the THP-1, 3A-sub-E, JEG-3, and primary trophoblast CM; pregnant mouse plasma; and plasma samples of pregnant women with preterm births and controls with term delivery.
Plasma of women with preterm births and controls with term delivery was collected at delivery, which was approved by the Institutional Review Board of MacKay Memorial Hospital, Taipei, Taiwan.
Immunofluorescence staining was performed as previously described . Serial sections of placental tissue were stained with anti-human NLRP1 (1:50, Santa Cruz Biotechnology, Dallas, Texas), AIM2 (1:100, Atlas Antibodies, Bromma, Sweden), or ASC (1:50, Santa Cruz Biotechnology) antibody to mark trophoblast inflammasome-related molecules, followed by staining with FITC– or rhodamine–conjugated donkey secondary antibody (1:50; Millipore). For double immunofluorescence, the sections were stained with specific antibodies against CD163 (1:250, Atlas Antibodies; a placental macrophage (Hofbauer cell) marker) and cytokeratin 7 (1:500, Dako, Santa Clara, CA), followed by staining with FITC–conjugated donkey anti-rabbit secondary antibody (1:500; Millipore) and rhodamine–conjugated donkey anti-mouse antibody (1:500; Millipore). Finally, the sections were stained with DAPI (Sigma–Aldrich) diluted to 1: 5 × 105 in PBS.
Placental tissues of pregnant women with preterm births and controls with term delivery were collected at delivery, as approved by the Institutional Review Board of Mackay Memorial Hospital, Taipei, Taiwan. Preterm birth was defined as delivery between 20 and 37 weeks of gestation. Inclusion criteria were singleton pregnancy delivered between 20 and 41 weeks of gestation without clinical evidence of infection or chorioamnionitis. Exclusion criteria included still birth, fetal anomaly, multiple pregnancies, preeclampsia, chronic hypertension, placenta previa, placental abruption, chorioamnionitis, other maternal medical complications, and fetal distress indicated preterm delivery.
THP-1 and 3A-sub-E cells were treated with 100 ng/ml LPS (Sigma–Aldrich) for 24 h or 700 ng/ml TNF-α (PeproTech) for 12–48 h. Total RNA of LPS- or TNF-α-treated THP-1 and 3A-sub-E cells was extracted using TRIzol reagent (Life Technologies) according to the manufacturer’s instructions, and cDNA was synthesized using oligodeoxythymidine (Promega Corporation, Madison, WI) and Superscript II Reverse Transcriptase (Life Technologies). Real-time PCR was performed using a 15-μl reaction mixture, containing 750 nM forward and reverse primers, varying amounts of template, and 1× SYBR Green reaction mix (Applied Biosystems, Foster City, CA). SYBR Green fluorescence was determined using the ABI PRISM 7500 detection system (Applied Biosystems). Primers were designed using Primer Express Software (Applied Biosystems). The primers used are listed in Supplementary Table S1.
|Treatment||Number of mice delivered|
|Preterm||Term||Preterm delivery rate (%)||P-value||Delivery day (mean ± S.D.)|
|n||17 dpc1||18 dpc||19 dpc||20 dpc||21 dpc|
|Negative control||6||0||0||0||6||0||0||20.0 ± 0.0|
|PBS+LPS||28||24||2||0||2||0||92.9||17.3 ± 0.8|
|SDA+LPS||15||7||1||2||5||0||53.3||<0.012||18.3 ± 1.4|
|DHA+LPS||33||17||6||8||1||1||69.7||0.0233||17.9 ± 1.1|
|DGLA+LPS||10||10||0||0||0||0||100||>0.9994||17.0 ± 0.0|
|OA+LPS||10||9||0||0||1||0||90||>0.9995||17.3 ± 0.9|
|Treatment||Number of mice delivered|
|Preterm||Term||Preterm delivery rate (%)||P-value||Delivery day (mean ± S.D.)|
|n||17 dpc1||18 dpc||19 dpc||20 dpc||21 dpc|
|Negative control||6||0||0||0||6||0||0||20.0 ± 0.0|
|PBS+LPS||28||24||2||0||2||0||92.9||17.3 ± 0.8|
|SDA+LPS||15||7||1||2||5||0||53.3||<0.012||18.3 ± 1.4|
|DHA+LPS||33||17||6||8||1||1||69.7||0.0233||17.9 ± 1.1|
|DGLA+LPS||10||10||0||0||0||0||100||>0.9994||17.0 ± 0.0|
|OA+LPS||10||9||0||0||1||0||90||>0.9995||17.3 ± 0.9|
The mice received SDA (500 nmole/kg), DGLA (500 nmole/kg), OA (500 nmole/kg), or DHA (125 nmole/kg) once per day at dpc 14–16, followed by stimulation with LPS (25 μg) at dpc 16 by intraperitoneal injection.
dpc, days post coitum.
2The P-value of SDA+LPS was compared with PBS+LPS (SDA+LPS compared with PBS+LPS).
3DHA+LPS compared with PBS+LPS.
4DGLA+LPS compared with PBS+LPS.
5OA+LPS compared with PBS+LPS.
Total cell extracts (20 μg of protein) were fractionated on 10% SDS/polyacrylamide gel, transferred onto an Immobilon polyvinylidene difluoride membrane (Amersham Biosciences, Piscataway, NJ), and probed with one of the specific primary antibodies against NLRP1 (Cell Signaling Technology, Danvers, MA), AIM2 (Cell Signaling Technology), ASC (Santa Cruz Biotechnology), caspase-1 (GeneTex, Irving, CA), IL-1β (Cell Signaling Technology), CTSS (GeneTex), p-IκBα (Cell Signaling Technology), IκBα (Cell Signaling Technology), p-p65 (Cell Signaling Technology), p65 (Cell Signaling Technology), and tubulin (Millipore) overnight at 4°C. Subsequently, the membrane was hybridized with appropriate horseradish peroxidase–conjugated secondary antibodies for 1 h at room temperature. Finally, immune complexes were visualized using the chemiluminescence method with an ECL detection kit (Millipore) on Fujifilm LAS-3000 (Fujifilm, Tokyo, Japan).
Depletion of TNF-α in THP-1 cell CM
To remove TNF-α from the CM of LPS-treated THP-1 cells (Sigma–Aldrich), the CM (2 ml) was incubated with 0.1 μg/ml anti-TNF-α (Dako) antibody and rotated overnight at 4°C. Subsequently, the reactive complex was incubated with 100 μl of protein G-agarose beads (50% slurry; Thermo Scientific, Rockford, IL) and rotated for 1 h at 4°C. The CM containing agarose beads was centrifuged at 800×g for 3 min, and the supernatant was collected and used immediately.
3A-sub-E trophoblasts were treated with 700 ng/ml TNF-α (PeproTech) for 15 min to 7 h. At the end of this treatment, the cells were lysed with lysis buffer (25 mM Tris phosphate (pH 7.8), 2 mM DTT, 2 mM CDTA, 10% glycerol, and 1% Triton X-100), and total cell extracts were incubated with 1× reaction buffer (100 mM DTT, 10 mM d-luciferin, and 5 mg/ml firefly luciferase). ATP levels were measured using the firefly luciferase-based method with an ATP determination kit (Invitrogen, Grand Island, NY), according to the manufacturer’s instructions. ATP levels (pM) were determined at 560 nm on an EnSpire Multimode Plate Reader (PerkinElmer, Waltham, MA).
Immunofluorescence staining and confocal microscopy
3A-sub-E trophoblasts were seeded on coverslips and treated with 700 ng/ml TNF-α (PeproTech) for 24 h. At the end of this treatment, the cells were incubated with LysoTracker Red DND-99 (500 nM, Life Technologies) for 4 h at 37°C to mark lysosomes and were washed three times in PBS containing 0.1% Tween 20. Subsequently, the cells were fixed in 3.7% methanol for 15 min at room temperature, blocked with 5% nonfat powdered milk for 1 h at room temperature, and immunostained with anti-CTSS antibody (GeneTex) overnight at 4°C, followed by incubation with FITC–conjugated donkey anti-rabbit secondary antibody (AP182F, Millipore) for 1 h at room temperature. Subsequently, the cells were washed three times in PBS and stained with Hoechst 33342 (Sigma-Aldrich) diluted to 1:1 × 104 in PBS for 10 min at room temperature. After three washes in PBS, coverslips were mounted using Fluoromount-G medium (Southern Biotechnology Associates, Birmingham, AL). Immunofluorescence images were acquired using a TCS SP5 laser scanning confocal microscope (Leica Microsystems, Mannheim, Germany). Green (FITC) fluorochromes were excited using a laser beam at 488 nm, and emissions were sequentially acquired using a photomultiplier tube with emission filters of 505–515 nm. Red fluorochromes were excited by a laser beam at 577–651 nm, and emissions were acquired sequentially using a photomultiplier tube with emission filters of 615–667 nm.
Cloning of pro-IL-1β, caspase-1, ASC, AIM2, and NLRP1
Total RNA was extracted using TRIzol reagent (Life Technologies), according to the manufacturer’s instructions and was reverse transcribed using Superscript II Reverse Transcriptase (Invitrogen) and oligodeoxythymidine to synthesize template cDNA. The coding region of pro-IL-1β, caspase-1, ASC, AIM2, and NLRP1 was amplified through 30 PCR cycles. These ORFs were ligated into a pcDNA 3.0 expression vector (Promega), and the resulting construct was sequenced to confirm the presence of the gene. The primers used are shown in Supplementary Table S2.
Inflammasome reconstitution assay
HEK-293, 3A-sub-E, and JEG-3 cells (2 × 105) were seeded in 6-cm cell culture dishes and were transfected transiently with pro-IL-1β, caspase-1, ASC, AIM2, or NLRP1-containing plasmids alone or in combination, as indicated, by using jetPRIME transfection reagent (Polyplus-transfection, New York, NY), according to the manufacturer’s instructions. After 24 h of transfection, the cell extract and cultured medium were collected, and the secreted ASC, AIM2, NLRP1, and IL-1β levels were determined through Western blotting and ELISA (R&D Systems). The cells were fully viable after transfections, with high transfection efficacy (>90%; not shown).
Cell viability assay
Cell proliferation rates were examined using the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTS) assay (Promega). 3A-sub-E and JEG-3 cells (2 × 103) were seeded in 96-well plates and were treated with 700 ng/ml TNF-α for 48–72 h in the presence of specific CTSS or caspase-1 inhibitors (Z-FL-COCHO and CASP1i). At the end of the treatment, 20 μl of MTS solution (10× dilution of 5 mg/ml MTS in MEM and DMEM without serum) was added to each well and incubated for 3 h at 37°C. Subsequently, 150 μl of lysis buffer (6 M HCl and 10% SDS) was added to solubilize formazan crystals, and the mixture was incubated overnight at 37°C. Finally, absorbance at 490 nm was measured on a SpectraMax microplate reader (Molecular Devices, Sunnyvale, CA).
Cell death assay
Cell death was examined using propidium iodide (PI; Sigma-Aldrich, St. Louis, MO) staining to identify cells with loss of plasma membrane integrity. 3A-sub-E and JEG-3 cells (3 × 104) were seeded in 6-cm cell culture dishes and were treated with TNF-α (700 ng/ml) at the indicated times in the presence of specific CTSS or caspase-1 inhibitors (Z-FL-COCHO and CASP1i). At the end of this treatment, the cells were collected and centrifuged at 3000×g for 3 min, and the pellets were stained with PI (Sigma-Aldrich) in 250 μl of PBS (0.5% Triton X-100, 0.125 mg of DNase-free RNase A, and 6.25 μg of PI) containing DNase-free RNase A (Roche, South San Francisco, CA) and PI (Sigma-Aldrich) for 15 min at 4°C in the dark. Nonviable cells were measured using flow cytometry. Data were collected and analyzed using FACScan (Becton Dickinson, San Jose, CA) run on CellQuest software (Becton Dickinson).
Timed-pregnant FVB mice purchased from a national animal center were used in the present study. Gestational age (±12 h) was determined by the presence of the vaginal plug, and the day of vaginal plug detection was designated as day 0 of pregnancy. All experiments were approved by the Animal Research Committee of MacKay Memorial Hospital, Taiwan. On day post coitum (dpc) 16, the mice were intraperitoneally (IP) injected with PBS or purified E. coli LPS (serotype 0111:B4, Sigma–Aldrich) dissolved in PBS. Based on preliminary studies, the LPS dose of 25 μg/mouse (1 mg/kg) was chosen for experiments.
On dpc 14, pregnant mice were assigned to Groups 1–6: Group 1 comprised control mice, receiving 0.2 ml of PBS through IP injection. Group 2 comprised LPS mice, receiving 25 μg of LPS and PBS IP on dpc 16. Groups 3–6 were injected with different fatty acids: 500 nmol/kg SDA (Group 3), 125 nmol/kg DHA (Group 4), 500 nmol/kg DGLA (Group 5), or 500 nmol/kg OA (Group 6) once per day on dpc 14–16; followed by stimulation with LPS (25 μg; IP) on dpc 16. After LPS or PBS administration, the mice were closely observed (every 30 min) for any signs of vaginal bleeding or preterm delivery (presence of pups in the cage). Spontaneous labor occurred on the morning of dpc 20 ± 1. Preterm delivery was defined as the delivery of at least 1 pup within 48 h of LPS exposure or delivery before day 19. The beginning of preterm delivery was defined by the delivery of the first pup. To evaluate the clinical relevance of SDA, DHA, DGLA, or OA treatment, the preterm delivery rate of the different groups was measured. The dams were followed up every 4 h during the first 12 h after treatment and then continuously until delivery. The pregnant mice without preterm delivery were killed on dpc 19–21 for sample collection. The timing for maternal mouse blood sample collection was at delivery or at maternal mouse scarification.
All values are reported as means ± S.D. Differences were evaluated using the Student’s t, chi-square, or Fisher’s exact tests as well as two-way ANOVA (main effect – treatment (LPS compared with various PUFAs), delivery (preterm compared with term), and treatment × delivery), if interaction term (treatment × delivery) <0.05, and followed by simple main effect between delivery (Student’s t test, Bonferroni’s correction, α/3 = 0.0167). The P-values of <0.05 were considered statistically significant.
First, we analyzed the levels of various cytokines expressed by macrophages after stimulation with LPS, a major outer surface membrane component in Gram-negative bacteria and thus a strong stimulator of innate immunity . ELISA revealed that LPS elevated IL-1β, IL-6, MCP-1, IL-8, and TNF-α levels in THP-1 cells at 24 h (Figure 1A,B). Inflammasomes are critical signaling platforms that recognize pathogenic microorganisms and sterile stressors, and they activate proinflammatory cytokines . Inflammasomes are cytosolic multiprotein complexes consisting of numerous proteins, including the NLR family of cytosolic pattern recognition receptors (such as NLRP1, NLRP3, or NLRC4) and the pyrin and HIN200 (hematopoietic interferon-inducible nuclear antigens with 200 amino-acid repeats) domain-containing protein (PYHIN) family of proteins such as AIM2. NLRP1 or AIM2 recognition receptors can interact with ASC and caspase-1 to form a complex to produce IL-1β and IL-18 cytokines . Hence, in the present study, we analyzed whether the expression of inflammasome-related genes is involved in LPS-induced inflammation in macrophages. Real-time PCR revealed that several inflammasome-related genes such as IL-1β, caspase-1, ASC, NLRP1, and AIM2 were up-regulated in THP-1 cells by LPS at 24 h (Figure 1C). The immune system plays a crucial role in regulating both term and preterm delivery, which is characterized by significantly altered proinflammatory cytokine levels in gestational tissues [19,20]. Macrophages play a vital role in the innate defense system. Activated macrophages secrete cytokines or chemokines to influence the immune response and recruit immune cells to the damaged tissue or infection site . Thus, placental trophoblasts may be influenced by macrophage-secreted molecules.
CM of THP-1 cells stimulated with LPS and TNF-α trigger inflammasome expression in trophoblasts
In the present study, according to ELISA, the secreted IL-1β level was approximately 200 pg/ml in THP-1 CM containing FBS after stimulation with 100 ng/ml LPS for 24 h (Figure 1A). IL-1β expression significantly increased in 3A-sub-E trophoblasts cultured in THP-1 CM containing FBS with LPS stimulation (THP-1-LPS CM) compared with trophoblasts cultured in THP-1 CM containing FBS without LPS treatment (THP-1 CM; Figure 1D). Western blotting revealed that inflammasome-related molecules, such as NLRP1, AIM2, ASC, and pro-IL-1β in the lysate and caspase-1 in the CM were activated in THP-1-LPS CM-treated 3A-sub-E cells (Figure 1E). Because we found that TNF-α was significantly up-regulated by LPS in THP-1 CM (Figure 1B), we next examined whether TNF-α is a crucial factor triggering trophoblast inflammasome activation in THP-1-LPS CM. Expectedly, the TNF-α level expressed by THP-1 cells was increased by LPS, but the level decreased after TNF-α depletion through blocking with anti-TNF-α antibody (Figure 1F). The secreted IL-1β level was elevated in THP-1-LPS CM, but the effect was abolished after TNF-α depletion (corresponding CM was referred to as THP-1-LPS CM-dTNF-α) in 3A-sub-E trophoblasts and the primary culture of trophoblasts from term placenta (Figure 1G,H). IL-1β expression in 3A-sub-E trophoblasts was time-dependently up-regulated by recombinant TNF-α (Figure 1I). A similar effect was observed in JEG-3 trophoblasts (Supplementary Figure S1). The CM of THP-1 cells after LPS stimulation contained elevated levels of TNF-α, which induced trophoblast inflammasome activation and IL-1β expression. The effect was inhibited after TNF-α depletion from LPS-stimulated THP-1 CM. Thus, TNF-α may activate trophoblast inflammasomes in LPS-stimulated THP-1 CM.
We investigated whether inflammasome-related molecules in trophoblasts are regulated by TNF-α. In 3A-sub-E and JEG-3 cells, inflammasome-related molecules, including NLRP1, AIM2, ASC, caspase-1, and pro-IL-1β, were activated by TNF-α (Figure 2A,B). Furthermore, we determined whether the regulated inflammasome-related molecules both in the cell lysate and CM were activated by the LPS-treated THP-1 CM with TNF-α depletion (THP-1-LPS CM-dTNF-α) at 24 and 48 h in 3A-sub-E and JEG-3 trophoblasts and the primary culture of trophoblasts from term placenta (Figure 2C–E). Based on the findings, we suggest that TNF-α activated inflammasome-related gene expression and caspase-1 activation in placental trophoblasts. We also observed that after treatment with a caspase-1 inhibitor, TNF-α-induced IL-1β secretion was abolished in 3A-sub-E and JEG-3 cells and the primary culture of trophoblasts from term placenta (Figure 2F–H).
PUFAs attenuate TNF-α-induced inflammasome expression in trophoblasts
We determined whether PUFA modulates TNF-α-triggered inflammasome-related gene expression in trophoblasts. The secreted IL-1β induced by THP-1-LPS CM for 24 h was attenuated by the n−3 fatty acids SDA and DHA in 3A-sub-E cells, as revealed by ELISA (data not shown). We found the n−3 fatty acids SDA and DHA inhibited TNF-α-induced IL-1β secretion in 3A-sub-E and JEG-3 trophoblasts and the primary culture of trophoblasts. However, the effect was not observed in n−6 DGLA- and n−9 OA-treated trophoblasts (Figure 2I–K). Inflammasome-related molecules including NLRP1, AIM2, ASC, and pro-IL-1β in the cell lysate and caspase-1 in CM were activated by TNF-α treatment for 24–48 h, and SDA and DHA treatment inhibited the activation of these molecules in 3A-sub-E cells, JEG-3 cells, and primary trophoblast culture. The effect was not observed in n–6 DGLA- and n–9 OA-treated trophoblasts (Figure 2L–N).
ATP has been identified as a key factor influencing IL-1β secretion . Therefore, we determined whether ATP plays a role in TNF-α-triggered inflammasome-related gene expression in trophoblasts. The results showed that contrary to our expectations, the ATP level in 3A-sub-E cells was not altered after treatment with 700 ng/ml TNF-α for 15 min to 7 h (Figure 3A), and the increased intracellular ATP content decayed rapidly over time . Moreover, we treated 3A-sub-E cells with different doses of apyrase, which is ATP diphosphohydrolase. The TNF-α-induced IL-1β level was not influenced by apyrase (Figure 3B). Furthermore, cathepsins have been demonstrated to be involved in inflammasome complex formation . Thus, in this study, we examined whether the cathepsin family plays a role in TNF-α-induced inflammasome-related gene expression. Amongst the members of the cathepsin family, only CTSS expression in 3A-sub-E cells was increased by TNF-α (700 ng/ml) at 24 h, as demonstrated by real-time PCR (Figure 3C). TNF-α time-dependently induced CTSS expression in 3A-sub-E and JEG-3 cells at both RNA and protein levels (Figure 3D–F). Cathepsins are found in lysosomes and are translocated from lysosomes to the cytosol through lysosomal rupture . Therefore, in this study, we examined whether lysosomes ruptured after TNF-α stimulation. To detect CTSS and lysosomal rupture, the cells were incubated with green fluorescent stain and red LysoTracker, respectively. The lysosome was observed as red circular form under normal condition, CTSS (green) would diffuse into cytosol with lysosome rupture after TNF-α stimulation for 24 h in 3A-sub-E cells, as observed under a confocal microscope (Figure 3G). This result indicated that TNF-α caused lysosomal rupture, which in turn accelerated the diffusion of CTSS from lysosomes to the cytosol in placental trophoblasts. We then determined the role of PUFA in TNF-α-stimulated CTSS expression. The CTSS level decreased after SDA and DHA pretreatment for 1 h, followed by TNF-α stimulation, in 3A-sub-E and JEG-3 cells compared with the level after TNF-α treatment alone. However, the effect was not observed after DGLA and OA pretreatment (Figure 3H,I).
TNF-α induces CTSS activation in trophoblasts
We subsequently analyzed whether CTSS plays an essential role in TNF-α-mediated inflammasome-related gene expression. We treated 3A-sub-E cells with 30 μM CATi, a cathepsin B, L, S and papain inhibitor, the THP-1-LPS CM- and TNF-α-induced IL-1β secretion were attenuated in the 3A-sub-E cells. A similar effect was observed in JEG-3 trophoblasts and the primary culture of trophoblasts (data not shown). Furthermore, we stimulated with a specific CTSS inhibitor (Z-FL-COCHO, 5–10 μM). The results showed that TNF-α-induced IL-1β secretion was dose dependently and significantly inhibited in 3A-sub-E and JEG-3 cells and the primary culture of trophoblasts (Figure 4A–C). Human cysteine cathepsins are lysosomal proteases and play roles in inflammasome activation [25,26]. Lysosomal damage has been predicted to be involved in inflammasome signaling . Hence, in this study, trophoblasts were pretreated with a specific inhibitor of vacuolar type H+-ATPase (Baf A1, 0.1–0.5 nM) to inhibit lysosome acidification and DEX (10–30 μM) to stabilize lysosome integrity and lysosomal enzymatic release for 1 h, followed by TNF-α stimulation for 24–72 h. TNF-α-upregulated IL-1β secretion was significantly inhibited in 3A-sub-E cells, JEG-3 cells, and the primary culture of trophoblasts. TNF-α-regulated CTSS was also dose dependently and significantly inhibited by Baf A1 and DEX, as revealed by Western blotting (Figure 4D–I). Based on these findings, we suggest the TNF-α induces inflammasome activation through CTSS activation and lysosomal damage.
TNF-α induces inflammasome expression via CTSS activation in trophoblasts
To examine the functional significance of ASC, AIM2, and NLRP1 in IL-1β production, we performed an inflammasome reconstitution assay in HEK-293 cells, which are deficient in endogenous inflammasomes but can produce active IL-1β after reconstitution of inflammasomes [27,28]. In the present study, we reconstituted inflammasomes by transiently transfecting the cells with pro-IL-1β, pro-caspase-1, ASC, AIM2, and NLRP1, and the amount of secreted IL-1β in cell CM was analyzed. Inflammasomes containing ASC, and AIM2 or NLRP1 caused the processing and secretion of IL-1β. However, in the absence of ASC, AIM2, or NLRP1, the IL-1β level decreased in HEK-293 cells (Figure 5A). Total cell lysates were used to determine the expression of transfected inflammasome-related molecules through Western blotting (Figure 5A). Similar results were observed in 3S-sub-E and JEG-3 trophoblasts (Figure 5B,C). Collectively, the results indicate that ASC, AIM2, and NLRP1 inflammasomes are involved in IL-1β processing and secretion in trophoblasts. Additionally, caspase activation can regulate the proteolytic maturation of IL-1β cytokines, and it causes inflammasome-associated cell death called pyroptosis, which limits endogenous and exogenous microorganism invasions . Hence, in the present study, we investigated whether the TNF-α-triggered inflammasome expression influences trophoblast cell viability and death. Cell viability decreased after TNF-α stimulation, but the effect was abrogated in 3A-sub-E and JEG-3 cells after treatment with CTSS and caspase-1 inhibitors (Z-FL-COCHO and CASP1i) (Figure 5D,E). TNF-α accelerated cell death, but the effect was inhibited in 3A-sub-E and JEG-3 cells after treatment with Z-FL-COCHO and CASP1i, as demonstrated by PI staining (Figure 5F,G). Thus, we suggested that in response to exogenous LPS challenge, TNF-α increased trophoblast cell death through inflammasome and CTSS activation.
The CTSS is essential in TNF-α-influenced cell viability and death in trophoblasts
To further delineate inflammasome-related gene expression induced by TNF-α, we treated 3A-sub-E and JEG-3 cells with TNF-α for 12–48 h; the results revealed that the levels of p-IκBα and p-p65 were up-regulated in these cells after TNF-α treatment (Figure 6A,B). TNF-α-induced NF-κB signaling and p-IκBα and p-p65 up-regulation were blocked by SDA and DHA, but not DGLA and OA, in 3A-sub-E and JEG-3 cells (Figure 6C,D). Therefore, we suggested that SDA and DHA inhibited TNF-α-activated inflammasome-related gene expression through the NF-κB signaling pathway. We then treated trophoblasts with 7.5 μM BAY11-7082, a NF-κB signaling inhibitor. TNF-α-stimulated inflammasome-related molecules such as NLRP1, AIM2, ASC, CTSS, and pro-IL-1β in the lysate and CTSS and caspase-1 in CM were significantly decreased in 3A-sub-E and JEG-3 cells (Figure 6E,F, lane 2 compared with 4). Therefore, NF-κB signaling may be a crucial pathway in TNF-α-induced inflammasome-related gene expression in trophoblasts. Finally, we determined whether CTSS plays a role in the synthesis of TNF-α-activated inflammasome-related molecules. Western blotting showed that the expression of inflammasome-related molecules such as NLRP1, AIM2, ASC, CTSS, and pro-IL-1β in the lysate and caspase-1 in CM decreased after 30 μM CATi pretreatment for 1 h, followed by TNF-α stimulation, in 3A-sub-E and JEG-3 cells compared the expression after TNF-α treatment alone (data not shown). According to Western blotting, a similar effect was observed after treatment with Z-FL-COCHO, a specific CTSS inhibitor (Figure 6G,H: lane 2 compared with 4). In 3A-sub-E and JEG-3 cells, the TNF-α-activated NF-κB signaling pathway involving p-IκBα and p-p65 was not influenced by treatment with 30 μM CATi (Figure 6I,J: lane 2 compared with 4). These results indicate CTSS serves as the downstream signal of the NF-κB pathway to modulate inflammasome-related molecules.
The NF-κB signaling and CTSS involved in TNF-α-induced inflammasome activation in trophoblasts
Placenta inflammation is associated with preterm birth. In the present study, we used a pregnant mouse model to investigate whether n−3 PUFA prevents infection-induced preterm delivery. The detailed information of the pregnant mouse model is illustrated in Table 1. The preterm delivery rate of LPS-injected mice was 92.9% (26/28). The preterm birth rate reduced to 53.3% (8/15) and 69.7% (23/33) after SDA and DHA pretreatment, but DGLA and OA did not cause any reduction in the preterm birth rate (Table 1). Resolvin D1 was identified as a DHA metabolite and anti-inflammatory lipid mediator . Before spontaneous labor, increased prostaglandin concentrations in the maternal circulation were observed, and the administration of exogenous prostaglandins induced uterine contractions and cervical dilation . Prostaglandin E2 is highly expressed in preterm cases and can be a risk factor for preterm labor . In the LPS-induced preterm mouse model with SDA, DHA, DGLA, or OA pretreatment, compared with mice with term delivery, inflammatory-related cytokines including TNF-α, IL-1β, prostaglandin E2, and CTSS were highly expressed in mice with preterm delivery. By contrast, the expression of the DHA metabolite resolvin D1 was higher in term mice with SDA and DHA pretreatment (Figure 7). Similarly, TNF-α and IL-1β plasma levels increased in preterm birth women, but resolvin D1 was highly expressed in women with term pregnancies (Figure 8). Immunofluorescence staining showed that the expression of NLRP1, AIM2, and ASC and the number of CD163-positive (Hofbauer) cells (indicated by arrows) increased in preterm birth placenta compared with controls (Figure 8D). Based on these findings, we suggested that NLRP1 and AIM2 inflammasomes were induced, and that the number of placental macrophages (CD163-positive cells) increased in preterm birth placenta. Thus, the activation of inflammasome in trophoblasts and macrophages may play a role in preterm birth. The characteristics of the pregnant women investigated in the present study are shown in Supplementary Table S3.
The inflammation-related cytokines, CTSS, resolvin D1, and prostaglandin E2 (PGE2) were examined in pregnant mouse plasma at delivery
The inflammation-related cytokines and DHA metabolite, resolvin D1, were examined in plasma samples of pregnant women with preterm births (n=31) and term controls (n=54)
The role of n−3 PUFA in infection of maternal–fetal interface and preterm labor remains unclear. Our findings showed that SDA and DHA attenuated the macrophage-accelerated inflammation of trophoblasts through the regulation of NLRP1 and AIM2 inflammasomes; these inflammasomes were regulated by inhibiting the synthesis of NLRP1, AIM2, ASC, caspase-1, and IL-1β and CTSS and caspase-1 activation. SDA and DHA significantly reduced the preterm birth rate in pregnant mice with LPS stimulation. TNF-α was significantly induced by LPS in THP-1 CM. The inflammasome mechanism was triggered by LPS-treated THP-1 CM in 3A-sub-E and JEG-3 cells and the primary culture trophoblasts. The effect was abolished after TNF-α depletion. However, various immune cells at the maternal–fetal interface, such as macrophages, natural killer cells, T cells, and dendritic cells , may produce TNF-α. Other mechanisms with different immune cells not assessed by our experiments may underlie trophoblast inflammasome activation. However, we demonstrated at least one pathway through which macrophages induce trophoblast inflammasome activation.
The study results indicate that macrophages are involved in the inflammation of the maternal–fetal interface through TNF-α expression and thus cause trophoblast phenotypes such as decreased cell viability and inflammasome assembly. In a previous study, IL-33 was detected in the culture supernatants of human placental macrophages and was identified as a critical factor elevating trophoblast proliferation through PI3K and MEK1/2 pathways . In addition, macrophage-secreted TNF-α attenuated the invasion of trophoblasts by inhibiting the urokinase system of plasminogen activators, a network of proteases that modulate cellular invasiveness . Our data support previous reports that macrophages can interact with trophoblasts to influence the trophoblast phenotype and pregnancy outcome. Cathepsins are found in lysosomes and are translocated from lysosomes to the cytosol through lysosomal rupture ; cathepsins can influence inflammasome complex formation . In the present study, amongst cathepsin family members, only CTSS expression in trophoblasts was up-regulated by TNF-α. TNF-α induced CTSS release into the trophoblast cytoplasm. Inhibition of CTSS activation blocked TNF-α-triggered IL-1β secretion and inflammasome activation. Thus, CTSS activation and lysosomal damage or rupture may be involved in the TNF-α-induced activation of inflammasomes in placental trophoblasts.
IL-1β is a major cytokine involved in the activation of immune cells and proinflammatory signaling pathways in gestational tissue. Mulla et al.  showed that inflammasome components including NLRP1, NLRP3, ASC, caspase-1, and IL-1β are expressed in trophoblasts, and IL-1β secretion is increased by uric acid. Furthermore, they observed antiphospholipid antibody-induced IL-1β secretion is dependent on ASC and NLRP3 inflammasomes in first-trimester human trophoblasts . IL-1β processing and secretion induced by Chlamydia trachomatis infection is dependent on the Nod1, but not ASC, inflammasome in human trophoblasts . These findings support that inflammasomes are involved in IL-1β secretion by trophoblasts and in placental inflammation. In the present study, we further demonstrated that inflammasome activation was triggered by macrophage-secreted TNF-α through lysosomal damage and CTSS activation. In response to LPS stimulation, TNF-α increased trophoblast cell death through inflammasome and CTSS activation. This finding is similar to that in a previous mouse model, in which trophoblastic apoptosis was observed in preterm birth placentas after LPS stimulation. IL-1β and TNF-α levels were also significantly increased in mouse serum . In response to LPS stimulation, the initiation of innate immunity involves essential co-operation between TLRs and NLRPs in macrophages and gestational tissues. LPS activates TLR4 to increase the production of pro-IL-1β and the inflammasome component NLRP through IκB kinase/NF-κB signaling . Our data revealed that the synthesis of inflammasome-related molecules such as NLRP1, AIM2, ASC, caspase-1, IL-1β, and CTSS was regulated through NF-κB signaling. Studies have shown that DHA and eicosapentaenoic acid interact with G-protein-coupled receptor 120 and G-protein-coupled receptor 40 involved in downstream signaling to prevent inflammation and metabolic disorder and that they modulate insulin sensitivity and antidiabetic effects by suppressing macrophage-influencing tissue inflammation [41,42]. In this study, we observed that the n−3 fatty acids SDA and DHA blocked TNF-α-triggered inflammasome expression by affecting NF-κB signaling and associated molecule synthesis.
n−3 PUFAs inhibit inflammation and the synthesis of prostaglandins, but inconsistent results have been obtained regarding the role of n−3 PUFA supplementation in the prevention of preterm birth [43–46]. For example, randomized controlled trials conducted in Australia and the United States have found a significant reduction in the preterm birth rate before 34 weeks of gestation . The Cochrane Review showed that women who received fish oil supplementation had a mean gestation age of 2.6 days, which was longer than the mean gestation age of controls . n−3 PUFA supplementation was associated with a lower risk of spontaneous preterm delivery in smokers with a history of at least singleton spontaneous preterm delivery . In a fat-1 mouse model, n−3 PUFA reduced the incidence of preterm birth induced by LPS. The gene expression of IL-6 and IL-1β in uteruses and the number of cervical infiltrating macrophages were reduced . By contrast, in a meta-analysis of randomized controlled trials of 3854 asymptomatic women with singleton gestations, no significant differences were observed in the preterm birth rate before 37 weeks of gestation in the n−3 PUFA treatment group, except for the lower perinatal death rate (0.3 compared with 1.2%) in women who received n−3 PUFA before 21 weeks of gestation . n−3 PUFA supplementation was provided to women with asymptomatic singleton gestations with a previous preterm birth to prevent recurrent preterm birth; the n−3 PUFA groups had significantly longer latency (mean difference: 2.10 days) and higher birth weight (mean difference: 102.52 g), but n−3 PUFA supplementation did not reduce the preterm birth rate before 37 weeks of gestation or before 34 weeks of gestation . However, the study populations from different countries were heterogeneous. The type and dietary intake of n−3 PUFA were not controlled across the trials. The use of different obstetric interventions to delay preterm birth could not be controlled in the different study populations, which may be a main confounding factor in these studies [50,51]. Further research evaluating the effect of n−3 PUFA supplementation in pregnancy is needed. In the present study, n−3 PUFA pretreatment decreased the preterm birth rate of LPS-treated mice and reduced TNF-α, IL-1β, prostaglandin E2, and CTSS expression. These results were also supported by the finding of higher resolvin D1 levels in the plasma samples of women with term delivery.
Collectively, the present data revealed that n−3 PUFAs, such as SDA and DHA, may provide protective effects during the early phases of bacterial challenge. Our findings support the potential role of SDA and DHA in attenuating the macrophage-accelerated inflammation of trophoblasts through the regulation of the NLRP1 and AIM2 inflammasomes; this regulation is caused by inhibiting NLRP1, AIM2, ASC, caspase-1, and IL-1β synthesis; caspase-1 and CTSS activation; and lysosomal damage. n−3 PUFA supplementation during pregnancy is feasible and can be used as a prophylactic intervention for preterm birth.
The molecular link between inflammation in preterm labor and inflammasome-mediated IL-1β release has not been well established. Macrophages secrete TNF-α to influence trophoblast inflammasome assembly and decrease cell viability at the maternal–fetal interface.
SDA and DHA reduce the macrophage-induced inflammation of trophoblasts by inhibiting inflammasome-related molecule expression and CTSS activation. SDA and DHA may provide protection effects during the early phases of bacterial challenge.
Supplementation with SDA and DHA during pregnancy can be used as a prophylactic intervention for preterm birth; such supplementation reduces preterm birth by inhibiting inflammasome activation in trophoblasts.
This work was supported by the Ministry of Science and Technology [grant number MOST 103-2314-B-195-009-MY3 (to C.-P.C.)]; and the MacKay Memorial Hospital [grant number MMH-E 105001 (to C.-P.C.)].
C.-P.C. conceived and designed the experiments. C.-Y.C. and C.-P.C. wrote the paper. C.-Y.C., C.-Y.C., and C.-C.L. performed the experiments. C.-Y.C., C.-C.L., C.-Y.C., and C.-P.C. analyzed the data. C.-P.C. contributed reagents/materials/analysis tools.
The authors declare that there are no competing interests associated with the manuscript.
absent in melanoma 2
apoptosis-associated speck-like protein containing a caspase recruitment domain
American Type Culture Collection
- Baf A1
Dulbecco`s Modified Eagle Medium
day post coitum
haematopoietic interferon-inducible nuclear antigens with 200 amnio-acid repeats
monocyte chemoattractant protein-1
Minimum Essential Medium
nucleotide-binding oligomerization domain-like receptor
nacht domain-leucine-rich repeat, and pyd-containing protein 1
nucleotide-binding oligomerization domain
polyunsaturated fatty acid