Interleukin (IL)-1β-deficient (IL-1β−/−) mice were assessed for cytokine production during pregnancy. A significant reduction in nuclear factor (NF)-κB p65 protein content was observed in the uteri and spleens of pregnant IL-1β−/− mice, as demonstrated by immunohistochemistry and Western immunoblot analysis. In addition, electromobility gel shift assay revealed less DNA binding activity of NF-κB p65-containing complex in pregnant IL-1β−/− mice. To investigate differences in cytokine production regulated by NF-κB, the levels of tumor necrosis factor-α, macrophage inflammatory protein-1α, and interferon-γ were measured in the uterine wall, spleen homogenates, and spleen cell cultures obtained from pregnant mice. Endocervical administration of lipopolysaccharide (LPS) increased cytokine levels in both wild-type (IL-1β+/+) and IL-1β−/− animals, but in IL-1β−/− mice this response was 50–75% lower. Splenocytes from nonpregnant mice exhibited decreased LPS-induced cytokine production when primed in vitro with progesterone. This suppression was 25% greater in IL-1β−/− than in IL-1β+/+ mice. These data suggest that constitutive NF-κB p65 protein synthesis is regulated by IL-1β, particularly during pregnancy.

  • p65 nuclear factor-κB
  • lipopolysaccharide
  • inflammation

interleukin (IL)-1 is a multifunctional cytokine, which induces several genes associated with inflammation, infection, and tissue injury (reviewed in Ref. 14). IL-1 stimulates the synthesis of other proinflammatory cytokines, including tumor necrosis factor (TNF)-α, IL-6, and chemokines (23, 30, 34, 38, 43). Surprisingly, in response to systemic administration of lipopolysaccharide (LPS), IL-1β-deficient (IL-1β−/−) mice do not exhibit altered circulating levels of TNF-α or IL-6 (18). In addition, no differences between wild-type (IL-1β+/+) and IL-1β−/− mice in LPS-induced toxicity have been observed (16). The only difference reported to date is the failure of IL-1β−/− mice to increase leptin mRNA in fat tissue and to elevate leptin levels in the circulation after systemic LPS administration (15).

We have recently reported that late in pregnancy IL-1β−/− mice exhibit reduced levels of uterine TNF-α, macrophage inflammatory protein-1α (MIP-1α), and IL-6 in response to endocervical injection of LPS compared with IL-1β+/+mice (37). However, the mechanism for this reduction remains unknown. IL-1β is a potent activator of nuclear factor (NF)-κB nuclear translocation, and NF-κB and Ap-1 binding motifs are functional in the IL-1β promoter (1, 9, 19, 20, 27). Therefore IL-1β and NF-κB are linked by a positive feedback loop serving to amplify inflammatory signals (21). Furthermore, both IL-1α and IL-1β have been reported to increase steady-state levels of NF-κB2 (p52) transcript and protein (25). In pregnancy, the uterine levels of IL-1 increase as parturition approaches (12). However, despite this increase, responses mediated by NF-κB are relatively suppressed due to the negative interaction between the NF-κB p65 subunit and progesterone (24, 44,47). NF-κB, and particularly the p65 subunit, is required for the transcription and production of TNF-α, IL-6, and MIP-1α (42, 44,48, 49). Therefore the altered activation of NF-κB may be responsible for reduction in cytokine production (37).

We hypothesized that IL-1β deficiency associated with a pregnancy-related endocrine environment may lead to an alteration of NF-κB-mediated cytokine production in response to LPS. In the present study, differences in NF-κB p65 subunit protein content and nuclear translocation in pregnant IL-1β−/− and pregnant IL-1β+/+ mice challenged with LPS were examined.



IL-1β−/− and IL-1β+/+ mice of mixed C57BL/6 and 129Sv(ev) background were obtained from Dr. H. Zheng (Merck, Rahway, NJ) (50) and bred in the University of Colorado Health Sciences Center animal facility. Mating (using one male per three females) was verified by the presence of a vaginal plug. All mice were allowed free access to food and water and exposed to 12:12-h light-dark cycles before and after experimentation was initiated.

Chemicals, reagents, and instruments.

Antibodies to NF-κB were raised in rabbits against the carboxy terminus of the human p65 or p50 subunit of NF-κB (Santa Cruz Biotechnology, Santa Cruz, CA). Both antibodies cross-react with the mouse NF-κB subunits. The enhanced chemiluminescence kit was obtained from Amersham (Arlington Heights, IL). RPMI and penicillin-streptomycin sterile solutions were purchased from Cellgro (Waukesha, WI); fetal bovine serum was from Life Technologies (Pascagoula, MS). Lyophilized LPS (a phenol-extracted preparation from Escherichia coli 055:B5) as well as other chemicals including water-soluble progesterone was purchased from Sigma Chemical (St. Louis, MO). The endoscope (arthroscope; angle 0°; diameter 1.7 mm; length 90 mm) was obtained from Comeg Endoscopy (Aurora, CO).

Experimental protocol.

Animal experimental studies were approved by the Animal Use and Care Committee of the University of Colorado Health Sciences Center. Onday 14 or15 of pregnancy (approximately 70% gestation), both IL-1β+/+ and IL-1β−/− mice were subjected to inhalation anesthesia with methoxyflurane (Metofane; Mallinckrodt Veterinary, Mundelein, IL) using an inhalation chamber and subsequent nasal anesthetic cone. Mice were placed in a dorsal supine position and restrained with paper tape. After the perineal area was washed with 70% isopropanol, the endoscope was inserted approximately 6–8 mm into the vagina until the cervix was visualized. A needle, attached to the syringe and bent to an angle of 30°, was guided through the vagina and visually advanced 3 mm into the cervix and 100 μl of LPS in saline (5 mg/kg) or corresponding amounts of sterile saline were injected intracervically. This dose of LPS is not lethal but required for cytokine response in C57BL mice injected intraperitoneally (18). Both IL-1β+/+ and IL-1β−/− pregnant and nonpregnant mice were injected intracervically. Nonpregnant and nonchallenged mice were included in this study for comparison. Animals were killed by cervical dislocation 24 h after injection. Gestational tissue and spleen specimens were collected. Uterine tissue (myometrium and decidua) was separated from placenta. The spleen was dissected from surrounding fat and fascia. Specimens were weighed and snap frozen in liquid nitrogen or directly embedded in tissue freezing medium (Triangle Biomedical Sciences, Durham, NC), frozen using dry ice-cold 2-methylbutane, and then stored at −70°C before use as will be described.

NF-κB staining of uterine tissue.

Transverse 5-mm cryosections were prepared with a cryostat (IEC Minotome Plus, Needham Heights, MA) and collected on poly-l-lysine-coated slides (Becton Dickinson, Franklin Lakes, NJ). Sections were fixed with a 70% methanol-30% acetone solution for 10 min at −20°C. Slides were washed three times with PBS for 5 min each, blocked in 10% normal goat serum for 25 min at room temperature, and then incubated for 1 h with rabbit polyclonal anti-NF-κB p65 antibody [Santa Cruz Biotechnology; 1:40 dilution with PBS containing 1% bovine serum albumin (BSA)]. After three washes with PBS, sections were incubated for 45 min with Cy3-labeled goat anti-rabbit IgG (Jackson Immuno Research Laboratories, West Grove, PA; 1:250 dilution with PBS-1% BSA), then washed three times with PBS, and counterstained with 2.5 mg/ml bis-benzimide (Sigma) for nuclear staining and 5 μg/ml fluorescein-labeled wheat-germ agglutinin (Molecular Probes, Eugene, OR) for cell surface staining. Sections were then mounted with aqueous antiquenching medium. To assess the specificity of immunostaining, adjacent sections were incubated with nonimmune rabbit IgG (5 μg/ml in PBS containing 1% BSA) as the replacement of the primary antibody and then processed identically. All sections were stained at the same time using the same antibody preparation. Images were observed at the same confocal conditions and photographed using a Leica DMRXA confocal microscope.

Macrophage staining of uterine tissue.

Rat anti-mouse macrophage F4/80 antigen monoclonal antibody (Caltag Laboratories, Burlingame, CA) at 10 μg/ml was used as the primary antibody for macrophage labeling (7), and Cy3-labeled goat anti-rat IgG (Jackson Immuno Research Laboratories; 1:125 dilution with PBS-1% BSA) was used as secondary antibody. The samples were stained as previously described.

Isolation and culture of spleen cells.

After aseptic removal of spleens, cell suspensions in RPMI with 10% fetal bovine serum were prepared as previously described (11). Spleen cells were cultured in 1.0 ml at 5 × 106 cells/ml in 24-well, flat-bottom culture plates (Becton Dickinson) in the presence or absence of LPS (10 μg/ml). Cultures were incubated at 37°C in a humidified 5% CO2 atmosphere. In some experiments, cells were preincubated with progesterone (50–200 ng/ml) for 4 h before LPS treatment. Cultures incubated for 24 h were subjected to three freeze-thaw cycles at −70°C. Samples were then centrifuged for 10 min at 10,000g, and cytokines were measured in the supernatants.

Tissue cytokine extraction.

Tissue extracts were obtained using a modified method of Nishiyama et al. (36). Briefly, tissue samples were homogenized using Tissue Tearor 985–370 (Biospec Products) at 30,000 in 10 volumes of 0.1% Tween 20 in 0.01 M PBS (pH 7.4) for 1 min on ice. After centrifugation at 13,000 g for 15 min at 4°C, supernatants were assayed for cytokine levels.

Cytokine assays.

Concentrations of TNF-α and MIP-1α in tissue extracts or cell culture medium were measured by liquid-phase electrochemiluminescence method (17) with a lower limit of detection of 10 pg/ml for MIP-1α and 60 pg/ml for TNF-α. Interferon-γ was measured using ELISA kits specific for murine cytokines (Endogen, Woburn, MA) with lower limit of detection of 10 pg/ml. Cytokine levels were normalized by weight of fresh tissue before homogenization.

Western immunoblotting.

Tissue was homogenized, as previously explained, in five volumes of buffer containing (in mM) 25 Tris ⋅ HCl, 2 EDTA, and 1 phenylmethanesulfonyl fluoride, pH 7.4. After centrifugation at 4°C at 13,000 g for 15 min, the supernatants were collected. Protein concentration was measured by Coomassie Plus Protein Assay (Pierce, Rockford, IL) using BSA as the standard. Samples, containing 20 μg of protein, were mixed 1:1 with sample-prep buffer (Bio-Rad, Richmond, CA) and boiled for 5 min. Electrophoresis was performed on 4–20% linear gradient SDS-polyacrylamide gels (Bio-Rad, Hercules, CA). After electrophoretic transfer to nitrocellulose membrane (Bio-Rad), membranes were stained with Ponceau S (Sigma) to confirm the equal amount of protein between samples, then washed and blocked for 1 h with PBS containing 0.1% Tween 20 and 5% nonfat dried milk (antibody buffer), and then incubated for 1.5 h at room temperature with rabbit polyclonal anti-NF-κB p65 antibody (1:200 dilution with antibody buffer; Santa Cruz Biotechnology). After sequential washing in 0.1% Tween 20 in PBS, membranes were incubated at room temperature for 1 h with horseradish peroxidase-linked goat anti-rabbit secondary antibody (Santa Cruz Biotechnology; 1:5,000 dilution with antibody buffer) and detected using the enhanced chemiluminescence system. Quantification of the immunoblot was performed by computer-assisted densitometry (software available from National Institutes of Health Application 1.599b4).

Detection of NF-κB by electrophoretic mobility shift assay.

Snap-frozen tissues (approximately 100 mg for uterus or 10 mg for spleen) were thawed at 4°C and washed with ice-cold PBS before homogenization in 500 μl lysis buffer containing 10 mM HEPES (pH 7.9), 10 mM KCl, 1 mM EGTA, 1 mM DTT, and one tablet of Complete Protease Inhibitor (Boehringer Mannheim, Indianapolis, IN). Nuclear proteins were then extracted as previously described (41). Briefly, homogenized tissue was incubated for 15 min on ice, and NP-40 (Sigma) was added to a final concentration of 0.5% followed by vortexing for 10 s. Samples were centrifuged at 8,000g for 15 min at 4°C and the nuclear pellet was resuspended in 50 μl of nuclear extraction buffer [20 mM HEPES (pH 7.9), 0.4 M NaCl, 1 mM EGTA, 1 mM DTT, and one tablet of Complete Protease Inhibitor] and incubated on ice for 30 min with gentle vortexing every 10 min. The nuclear extract was then clarified by centrifugation at 12,000g for 5 min at 4°C. The supernatants containing the nuclear proteins were quantified with the Coomassie Plus protein assay (Pierce).

NF-κB consensus oligonucleotide (AGTTGAG̅G̅G̅G̅A̅C̅T̅T̅T̅- CCCAGGC, binding site underlined, Promega, Madison, WI) was 5′ end-labeled with [γ-32P]ATP (NEN, Boston, MA) using T4 polynucleotide kinase. Unincorporated nucleotide was separated using a NucTrap Probe purification column (Stratagene, La Jolla, CA). Ten micrograms of nuclear protein were incubated with labeled oligonucleotide (150,000 counts/min) in binding buffer [10 mM Tris ⋅ HCl (pH 7.5), 50 mM NaCl, 0.5 mM EDTA, 1 mM MgCl2, 0.5 mg poly(dI-dC), 1% NP-40, and 4% glycerol] for 25 min at room temperature in a final volume of 25 μl. Subsequently, the products were separated by electrophoresis on a 4% polyacrylamide-0.5 times Tris-Borate-EDTA gel. The gel was then dried onto Whatman paper and exposed to an X-ray film overnight at −70°C with an intensifying screen.

For supershift studies, 1 μg of antibody to either p50 or p65 (Santa Cruz Biotechnology) was added and incubated for 10 min at room temperature before the addition of labeled oligonucleotide. Binding of the antibody to NF-κB was indicated by a supershift in the electrophoretic mobility shift assay. To further demonstrate specificity, excess unlabeled oligonucleotide was used as a specific competitor.

Statistical analysis.

Unpaired Student's t-test was used to analyze differences between experimental groups. Statistical significance was accepted within 95% confidence limits.


Diminished uterine NF-κB p65 in pregnant IL-1β−/− mice.

Although immunostaining with anti-NF-κB p65 antibody revealed immunoreactivity in uterine sections from both IL-1β+/+ and IL-1β−/− mice, a marked reduction of intracellular NF-κB protein was observed in endometrial tissue from IL-1β−/− mice compared with IL-1β+/+. As depicted in Fig.1 A, intense intracellular NF-κB staining (red) was detected in endometrial tissue from IL-1β+/+mice following LPS-induced pregnancy loss. Because the tissue was taken 24 h after the intracervical LPS challenge, most of this immunoreactivity was observed in cytoplasm. In contrast, IL-1β−/− mice demonstrated significantly less intracellular NF-κB immunoreactivity in endometrial tissue (Fig. 1 B). A similar difference in NF-κB immunoreactivity was observed in myometrium of IL-1β+/+ and IL-1β−/− mice; however, this difference was not as dramatic as that observed in endometrial tissue (data not shown).

Fig. 1.

Immunohistochemical localization of endometrial nuclear factor (NF)-κB. Endometrium from pregnant interleukin (IL)-1β wild-type (IL-1β+/+;A) and IL-1β-deficient (IL-1β−/−;B) mice were collected 24 h after endocervical injection of lipopolysaccharide (LPS; 5 mg/kg) and stained with anti-NF-κB p65 antibody. Anti-NF-κB p65 antibody stained red (Cy3) and bis-benzamide for nuclei stained blue.

Specific staining for macrophages confirmed that these differences between IL-1β+/+ and IL-1β−/− are not due to unequal numbers of macrophages in endometrial tissue (data not shown).

Diminished NF-κB p65 protein in the uterus and spleen of pregnant IL-1β−/− mice.

With Western immunoblotting, NF-κB p65 protein was detected in the uterus of both pregnant and nonpregnant IL-1β+/+ and IL-1β−/− mice. As depicted in Fig.2 A, endocervical LPS injection was associated with an increase in uterine NF-κB levels in both IL-1β+/+and IL-1β−/−pregnant mice. Nevertheless, compared with IL-1β+/+ pregnant mice, pregnant IL-1β−/− animals, both saline and LPS-injected, exhibited lower uterine content of NF-κB p65 protein. Densitometric analysis revealed that 24 h after LPS administration, NF-κB uterine levels in IL-1β−/− were 50–60% of those in IL-1β+/+ mice.

Fig. 2.

NF-κB protein content in uterine and spleen tissue. Western blotting of proteins extracted from uteri (A) and spleen (B) of IL-1β−/− and IL-1β+/+ pregnant mice 24 h after endocervical injection of either saline or LPS (5 mg/kg). Anti-NF-κB p65 protein was used as the primary antibody. InA, the vertical axis indicates reading of samples in densitometric units. Data depict 2 of 4 saline-injected mice and 4 of 6 LPS-challenged pregnant IL-1β−/− and IL-1β+/+ mice.

Similar differences in NF-κB protein content were observed in spleen homogenates of LPS-challenged IL-1β+/+ and IL-1β−/− mice (Fig.2 B). Twenty-four hours after LPS injection, NF-κB levels in spleen of IL-1β+/+ mice were approximately twofold higher than those of IL-1β−/− mice.

Electrophoretic mobility shift assay demonstrates the different activity of NF-κB p65-containing complex in pregnant uterus.

The binding of NF-κB p50 and p65 protein-containing complex was detected in the uterus of both IL-1β+/+ and IL-1β−/− pregnant mice. As depicted in Fig. 3, compared with IL-1β+/+ mice, NF-κB p50-containing complexes in IL-1β−/− animals were reduced by 10%, whereas the level of p65-containing complexes appeared to be reduced by 60%.

Fig. 3.

Electrophoretic mobility shift assay detection of active NF-κB. Polyacrylamide gel electrophoresis of nuclear extracts after binding to labeled NF-κB probe. Lanes 1–3and 4–6 represent binding of uterine NF-κB obtained from LPS-treated IL-1β+/+ and IL-1β−/− mice, respectively. Lanes 2 and5 represent binding of NF-κB p50-containing complex (top band) and NF-κB complexes not recognized by anti-p50 antibody (bottom band).Lanes 3 and 6 represent binding of NF-κB p65-containing complex (top band) and NF-κB complexes not recognized by anti-p65 antibody (bottom band). Data in table below gel depict amount of super-shifted NF-κB (top bands only) estimated by densitometry. Densitometry readings are normalized by individual background subtraction. Data represent 1 of 3 similar experiments.

IL-1β deficiency is associated with diminished uterine cytokine production in response to LPS administration in pregnant mice.

Compared with saline, administration of LPS to IL-1β+/+ and IL-1β−/− pregnant mice induced an increase in uterine content of MIP-1α and TNF-α (Fig. 4, Aand B); however, in pregnant IL-1β−/− mice, the LPS-induced elevation of uterine TNF-α and MIP-1α was two- to threefold less (P < 0.05) than that in IL-1β+/+ mice. A similar finding was observed for interferon-γ; however, a 1.6-fold difference did not reach statistical significance.

Fig. 4.

Cytokine concentrations in uterine tissue. Uterine tissues were obtained 24 h after either intracervical LPS (5 mg/kg) or saline injection. After homogenization, clarified extracts were assayed for macrophage inflammatory protein (MIP)-1α (A) and tumor necrosis factor (TNF)-α (B). Cytokines from pregnant IL-1β+/+(n = 5) or IL-1β−/−(n = 3) mice 24 h after saline injection were set at 1.0. Data shown represent change ± SE in mice 24 h after LPS. Ten animals for LPS-injected IL-1β+/+ mice (open bar) and 15 animals for LPS-injected IL-1β−/− mice (solid bar) are shown. Data obtained from 3 saline-injected IL-1β+/+ and IL-1β−/− mice, 3 LPS-injected IL-1β+/+ mice, and 5 LPS-injected IL-1β−/− mice were previously reported (37).

Spleens of pregnant IL-1β−/− mice exhibit diminished cytokine production in response to LPS administration in vivo.

Twenty-four hours after in vivo LPS administration, the content of MIP-1α in spleens of pregnant IL-1β+/+ mice was 40% lower than the corresponding level in nonpregnant IL-1β+/+ mice (Fig.5 A). However, in IL-1β−/−mice MIP-1α levels were 80% less (P< 0.01) than that in pregnant IL-1β+/+ mice (Fig.5 A). Spleen TNF-α also was 50% lower in pregnant IL-1β−/− mice compared with pregnant IL-1β+/+mice; however, this difference was not statistically significant (P = 0.1).

Fig. 5.

LPS-induced MIP-1α production by splenocytes in vivo (A) and in vitro (B).A: spleens were obtained from nonpregnant IL-1β+/+ and pregnant IL-1β+/+ or IL-1β−/− mice 24 h after intracervical LPS (5 mg/kg) injection. After homogenization, clarified extracts were assayed for MIP-1α. Data represent change ± SE from nonpregnant IL-1β+/+ control mice injected with LPS and set at 1.0. Four mice per group were used.B: splenocytes were cultured from nonchallenged, nonpregnant IL-1β+/+ and pregnant IL-1β+/+ or IL-1β−/− mice. Cultures were subjected to 3 freeze-thaw cycles and centrifuged at 10,000 g, and MIP-1α was measured 24 h after in vitro incubation with LPS (10 μg/ml). Data represent change ± SE from control splenocytes (set at 1.0) obtained from nonpregnant IL-1β+/+ mice and stimulated with LPS in vitro for 24 h. Four mice per group are shown.

In addition, concentrations of MIP-1α and TNF-α were measured in culture medium 24 h after in vitro LPS stimulation of splenocytes obtained from untreated pregnant mice. Although the differences did not reach statistical significance, a trend toward lower MIP-1α (Fig.5 B) and TNF-α levels was observed in IL-1β−/− spleen cells compared with IL-1β+/+splenocytes. No differences in cytokine production were observed when spleen cells from nonpregnant IL-1β+/+ and IL-1β−/− mice were stimulated in vitro with LPS.

Progesterone treatment of splenocyte cultures from nonpregnant IL-1β−/− mice reveals diminished cytokine production in response to LPS administration in vitro.

Concentrations of MIP-1α and TNF-α were measured in culture medium from splenocytes obtained from untreated, nonpregnant IL-1β+/+ and IL-1β−/− mice. Splenocytes were initially incubated for 4 h in the presence of progesterone; LPS was then added. Preincubation with progesterone suppressed production of MIP-1α and TNF-α, measured 24 h after the in vitro addition of LPS, in a dose-dependent manner (Fig.6). Progesterone pretreatment at 200 ng/ml reduced LPS-induced MIP-1α and TNF-α production in splenocytes from IL-1β+/+ by 10 and 30%, respectively, whereas in IL-1β−/− mice this cytokine production was additionally reduced by 25% compared with IL-1β+/+ mice.

Fig. 6.

Effect of progesterone on LPS-induced cytokines splenocytes. Splenocytes were cultured from nonchallenged, nonpregnant IL-1β+/+ and IL-1β−/− mice. Cells were preincubated with progesterone (200 or 50 ng/ml) for 4 h and then LPS (10 mg/ml) was added. Cultures were subjected to 3 freeze-thaw cycles 24 h after incubation at 37°C, centrifuged at 10,000g and MIP-1α (A) and TNF-α (B) were measured. Data represent change ± SE from control set at 1.0 for in vitro LPS-induced splenocytes not pretreated with progesterone (* P < 0.05).


We observed that pregnancy in IL-1β−/− mice is associated with decreased LPS-induced uterine and splenic cytokines compared with similarly challenged pregnant wild-type mice. With the sole exception of leptin (15), there is no difference in the response to LPS between IL-1β−/− and IL-1β+/+ nonpregnant mice (18). Thus the present findings were unexpected. The most likely explanation for the decrease in LPS-inducible cytokines in pregnant IL-1β−/− mice is a reduction in the amount of the p65 component of NF-κB. In pregnant IL-1β−/− mice, we observed a marked decrease of p65 immunostaining in the endometrium. A 40–50% reduction in uterine p65 protein levels was also detected by Western blotting in IL-1β−/− compared with IL-1β+/+ pregnant mice. These findings were confirmed by gel super-shift electromobility assay of nuclear extracts from pregnant uteri. These changes corresponded to decreased LPS-induced NF-κB-mediated uterine cytokine production in IL-1β−/− pregnant mice. Thus by four separate methods, we conclude that IL-1β participates in the regulation of steady-state levels of NF-κB p65 in murine endometrium and splenocytes.

Specific staining of endometrial tissue for macrophages using the anti-F4/80 antibody demonstrated that endometrial macrophages are not the major source of NF-κB in uterus. In normal pregnancy, uterine levels of IL-1 increase as parturition approaches (12). In the absence of pregnancy and thus in the absence of high levels of progesterone and estrogens, differences in NF-κB protein levels between IL-1β+/+ and IL-1β−/− mice were not detected. During pregnancy, progesterone, known to suppress the function of NF-κB (24, 44, 45, 47), is elevated. As pregnancy progresses, there is increased production of IL-1β and likely a concomitant increase in NF-κB synthesis. This IL-1β-orchestrated increase in p65 overcomes the suppression of NF-κB activity by progesterone. However, in the absence of IL-1β, this apparently does not take place and thus a relative lower level of p65 was observed. This finding implicates a role for IL-1β in the steady-state level of p65 that was not observed in nonpregnant IL-1β−/− mice.

The transcription factor NF-κB regulates gene expression of a variety of proteins induced during immune and inflammatory responses, including several cytokines. Typically, NF-κB is characterized as a heterodimer comprised of a 50-kDa (p50) and a 65-kDa (p65) subunit (10). In addition to the heterodimer p50-p65, homodimers also recognize the common DNA sequence binding motif. Although p50-p50 or p65-p65 homodimers both have been previously proposed to be involved in gene expression by selective activation of genes, p65-containing complexes are most frequently reported as initiating transcription factors (4,35). In contrast to p65 homo- or heterodimers, the p50 homodimer is considered to function as an inhibitory complex displacing the p50-p65 or p50-c-rel heterodimers from the NF-κB binding sites (8). The reduction in LPS-induced cytokines and slightly reduced p50 binding in the gel-shift assay suggest that a partial reduction in p50 may also take place in IL-1β−/− mice. This is consistent with the report that the p65 subunit increases activation of p50 subunit gene expression (46). NF-κB is activated by many stimuli, including oxidants, bacterial products, viruses, inflammatory cytokines, and immune stimuli (2). This activation involves the multistep process based on the interaction of dimerized DNA-binding subunits with the inhibitor of NF-κB (I-κB) proteins in the cytosol with subsequent phosphorylation of I-κB and followed by nuclear translocation of NF-κB (29, 40). However, in the present study, we show that reduced LPS-induced cytokines are associated with relatively lower amounts of p65.

The gene encoding for NF-κB is constitutively autoregulated by NF-κB (10). Also, in most situations, NF-κB activity is regulated posttranscriptionally (2). Nevertheless, in different cells NF-κB-mediated response to various stimuli may be either dependent or independent of novel protein synthesis. For example, in HL-60 cells, phorbol ester-induced NF-κB nuclear translocation was found to be dependent on de novo protein synthesis, whereas in murine 70Z/3 cells, NF-κB activation in response to phorbol esters was protein synthesis independent (22). Recently, several messengers, including Ets proteins, were reported to bind the NF-κB p50 promoter. It was suggested that these proteins were responsible for a fine tuning of NF-κB gene expression in different cell types (26).

The present study demonstrates a possible involvement of IL-1β in the regulation of steady-state levels of NF-κB/p65 in murine endometrium. IL-1β is a potent activator of NF-κB translocation, inducing its activity rapidly via a protein synthesis-independent mechanism (6). IL-1α and IL-1β also have been reported to increase steady-state levels of NF-κB2 (p52) transcript and protein (25). The difference in endometrial NF-κB protein levels and in NF-κB-mediated cytokine production revealed during pregnancy may be due to rising levels of progesterone during gestation. This hypothesis was supported by the greater reduction of in vitro LPS-induced MIP-1α and TNF-α production in progesterone-treated IL-1β−/−splenocytes. There were no differences in LPS-induced cytokine responses in splenocytes cultured from nonpregnant IL-1β+/+ or IL-1β−/− mice, whereas a 4-h exposure to progesterone at a concentration similar to levels circulating in pregnancy (3) decreased LPS-induced cytokine production in IL-1β−/− mice 25% greater than in IL-1β+/+ mice.

Progesterone has been reported to modulate macrophage-dependent immune responses (32). In murine macrophages, progesterone inhibits inducible nitric oxide synthase gene expression and nitric oxide production (31). It also has been reported that progesterone modulation of macrophage responses, and particularly of LPS-induced nitrite release by murine peritoneal macrophages, is a receptor-mediated process (39). Progesterone suppresses TNF-α mRNA and protein synthesis in LPS-activated murine macrophages and human peripheral blood mononuclear cells (28, 33). These effects may be explained by1) the negative interaction between the NF-κB p65 subunit and the progesterone receptor (24, 47) and/or2) the progesterone-induced increase of I-κBα protein translocated to the nucleus (33). We presume that the relative reduction of NF-κB protein associated with IL-1β deficiency determines the greater sensitivity of LPS-challenged IL-1β−/− mice to the anti-inflammatory effects of progesterone. In splenocytes from IL-1β+/+ mice, blocking IL-1 with IL-1Ra reduced LPS-induced TNF-α and MIP-1α production by 10–15%; however, in the presence of progesterone, this reduction is greatly enhanced (unpublished observation). Associated with this reduction, we observed a reduction in p65 by Western blotting after 24-h incubation of IL-1Ra plus progesterone (unpublished observation).

Although NF-κB has been well described as an important regulator of proinflammatory cytokine production (5, 42, 44, 48, 49), the observed changes in NF-κB protein associated with IL-1β deficiency might be considered as only a partial contribution to the complex mechanism of cytokine regulation in vivo. However, along with the potential effects of other transcription factors or recently described involvement of IL-1β in the regulation of I-κB kinase activity (13), the influence of IL-1β on NF-κB protein levels may lend an additional insight into the regulation of the cytokine network in vivo.


Constitutive NF-κB gene expression and protein levels are present in nearly all types of cells. In the present understanding, the regulation of NF-κB-mediated events, including proinflammatory cytokine production, takes place posttranscriptionally when NF-κB complexes are released from I-κB and translocate to the nucleus. Changes in steady-state NF-κB protein levels could be considered another regulatory mechanism that remained unknown. Thus the role of IL-1β in the modulation of NF-κB protein levels, demonstrated by reduced level of NF-κB p65 in IL-1β−/− mice and decreased proinflammatory cytokine response to LPS in pregnancy, offers an alternative regulatory mechanism. The importance of some proinflammatory cytokines may not be apparent until “pressured” under specific circumstances. In the case of pregnancy, a thoroughly physiological circumstance, “pressuring” constitutive NF-κB expression takes place but is revealed by IL-1β deficiency. At issue is that “nature” would not leave NF-κB synthesis, such an important regulator of the inflammatory response during pregnancy, to just one gene product, either progesterone or IL-1β. Hence understanding the regulation of NF-κB is one of uncovering the components of this regulation. That during pregnancy inflammatory diseases are suppressed is well established. Now we are seeing part of this mechanism at the molecular level.


We thank Dr. H. Zhang for providing IL-1β-deficient mice and Dr. F. Gamboni-Robertson for assistance.


  • Address for reprint requests and other correspondence: C. A. Dinarello, Univ. of Colorado Health Sciences Center, Div. of Infectious Diseases, Campus Box B-168, 4200 East Ninth Ave., BRB 401, Denver, CO 80262 (E-mail: leonid.reznikov{at}uchsc.edu).

  • This work was supported by the National Institutes of Health Grants AI-15614 to C. A. Dinarello and AI-2532359 to L. L. Reznikov and Colorado Cancer Center Grant CA-46943.

  • The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. §1734 solely to indicate this fact.


View Abstract