Role of Notch signaling during lipopolysaccharide-induced preterm labor

Notch signaling is activated during LPS-induced PTL, as angiogenesis factors are down-regulated.

Abstract

Notch signaling pathways exert effects throughout pregnancy and are activated in response to TLR ligands. To investigate the role of Notch signaling in preterm labor, Notch receptors (Notch1–4), its ligand Delta-like protein-1, transcriptional repressor hairy and enhancer of split-1, and Notch deregulator Numb were assessed. Preterm labor was initiated on gestation d 14.5 by 1 of 2 methods: 1) inflammation-induced preterm labor: intrauterine injection of LPS (a TLR4 agonist) and 2) hormonally induced preterm labor: subcutaneous injection of mifepristone. Delta-like protein-1, Notch1, and hairy and enhancer of split-1 were elevated significantly, and Numb was decreased in the uterus and placenta of inflammation-induced preterm labor mice but remained unchanged in hormonally induced preterm labor compared with their respective controls. F4/80⁺ macrophage polarization was skewed in the uterus of inflammation-induced preterm labor toward M1-positive (CD11c⁺) and double-positive [CD11c⁺ (M1) and CD206⁺ (M2)] cells. This process is dependent on activation of Notch signaling, as shown by suppression of M1 and M2 macrophage-associated cytokines in decidual macrophages in response to γ-secretase inhibitor (an inhibitor of Notch receptor processing) treatment ex vivo. γ-Secretase inhibitor treatment also diminished the LPS-induced secretion of proinflammatory cytokines and chemokines in decidual and placental cells cultured ex vivo. Furthermore, treatment with recombinant Delta-like protein-1 ligand enhanced the LPS-induced proinflammatory response. Notch ligands (Jagged 1 and 2 and Delta-like protein-4) and vascular endothelial growth factor and its receptor involved in angiogenesis were reduced significantly in the uterus and placenta during inflammation-induced preterm labor. These results suggest that up-regulation of Notch-related inflammation and down-regulation of angiogenesis factors may be associated with inflammation-induced preterm labor but not with hormonally induced preterm labor.

Introduction

Infection and the associated inflammatory processes required for host defense comprise a major challenge to successful pregnancy outcome. In women, PTL affects ∼12% of all deliveries. Approximately 40% of cases of PTL are associated with infection at the feto-maternal interface [1, 2]. In the event of IU infection, inflammatory processes lead to premature delivery of the baby [3]. Animal models of PTL, resulting from inflammation or premature progesterone withdrawal, have relevance to human gestation [4, 5]. Well-studied PTL models include those involving an inflammatory stimulus (e.g., administration of TLR ligands, such as LPS, a TLR4 ligand [6, 7]) and withdrawal of progesterone’s effects on pregnancy maintenance (e.g., MIF, a progesterone receptor antagonist [8]).

Our previous studies showed that PTL, induced by the combination of PGN + poly(I:C), is associated with decidual macrophage polarization toward a double-positive state [i.e., CD11c⁺ (M1) and CD206⁺ (M2)], accompanied by a robust inflammatory response and apoptosis. These processes are linked with reduction of the a2V [9]. The inhibition of a2V is associated with defective endolysosomal functions and incomplete turnover of autophagosomes in the macrophage and decidual and placental cells [10]. Our study also showed that this reduction or inhibition of a2V leads to alteration in autophagic flux and hyperinflammation during inflammation-induced PTL [10]. V-ATPase is known to be involved in the regulation of various cellular processes, such asendocytosis, vesicular trafficking, and protein degradation, thereby influencing various signaling pathways [11]. V-ATPase-dependent acidification is also critical for the Notch signaling pathway [12]. V-ATPase knockdown causes defects in endosomal acidification and affects the levels of Notch signaling-associated molecules (e.g., Notch1) [12–15].

Four distinct Notch receptors (Notch1–4) and 5 different Notch ligands (DLL-1/3/4 and Jagged 1/2) have been identified in mammals [16, 17]. Ligation of these ligands to Notch receptors leads to proteolytic cleavage of the receptor and liberates the NICD from the membrane. NICD moves to the nucleus, where it activates the recombining binding protein suppressor of hairless transcription factor, thus allowing the recruitment of coactivators, leading to transcription of Notch target genes. The best described Notch target genes in mammals are Hes and hairy/enhancer-of-split related with YRPW motif protein [18, 19], and other targets include NF-κB [19]. Numb (a deregulator of Notch signaling)–Notch interaction regulates cell fate in various tissues [20]. Notch signaling has been proven to decide the fate of immune cells and is involved in T and B cell activation, in Th cell differentiation [19], and in polarization of macrophages to the M1 subtype (inflammatory macrophages) [21]. LPS treatment activates the Notch signaling through a JNK-dependent pathway [22], which enhances NF-κB phosphorylation [23] and secretion of proinflammatory cytokines (IFN-γ and TNF-α) [24].

Decreased Notch signaling is reported to be associated with endometriosis and impaired decidualization in the human [25], and inhibition of Notch signaling is associated with pre-eclampsia in the mouse [18]. Notch ligands (DLL-4, Jagged 1 and 2), possibly through regulation of angiogenesis via VEGF, have essential roles in embryogenesis and placental development [26, 27]. However, their role in PTL remains undetermined.

We hypothesized that inflammation-induced PTL and hormonally induced PTL (resulting from progesterone withdrawal) are associated with defects in Notch signaling and a2V function. We found that during inflammation-induced PTL in the mouse model production of the Notch ligand DLL-1 and of the Notch1 receptor, nuclear translocation of Hes1 was greatly increased. This amplifies the inflammatory immune response via increased M1 macrophage polarization and production of associated cytokines. In contrast, during hormonally induced PTL, Notch signaling remained unaffected. Paralleling these in vivo observations, ex vivo LPS induces DLL-1 ligand, Notch1 receptor, and Hes1 in decidual macrophages, an effect that was amplified with blockade of a2V. Furthermore, GSI (an inhibitor of Notch receptor processing) significantly diminished the levels of LPS-induced M1 macrophage-associated cytokines in cultured decidual macrophages. The similar results were observed in decidual and placental cells. LPS treatment also suppressed Notch-dependent angiogenesis (production of Jagged 1 and 2, DLL-4, and VEGF). We have demonstrated that Notch-related inflammation is up-regulated, and angiogenic factors are down-regulated during inflammation-induced PTL but not during hormonally induced PTL. The activation of Notch signaling is associated with decidual macrophage polarization to M1 and enhancement of inflammatory responses, and this activation might be a reflective event of decreased a2V activity. The suppression of Notch signaling has the potentially beneficial effect of suppressing inflammation.

MATERIALS AND METHODS

Mice

All procedures involving animals were approved by the NorthShore University HealthSystem Animal Care and Use Committee and conform to the Guide for Care and Use of Laboratory Animals (National Academy of Sciences, 1996). Mice used in the present studies were a CD-1 strain (Harlan Laboratories, Madison, WI, USA). Female mice in estrus were selected by the gross appearance of the vaginal epithelium [28] and were impregnated naturally. Mating was confirmed by the presence of a vaginal plug, and the day of plug formation was counted as d 0.5 of pregnancy.

Inflammation-induced PTL model

Labor was induced in mice on gestation d 14.5, as described previously [6, 29, 30]. In brief, animals were anesthetized with 0.015 ml/g body weight of Avertin (2.5% tribromoethyl alcohol and 2.5% tert-amyl alcohol in PBS). A 1.5 cm midline incision was made in the lower abdomen. In the mouse, the uterus is a bicornuate structure in which the fetuses are arranged in a "beads-on-a-string" pattern. IU injections were performed in the midsection of the right uterine horn at a site between 2 adjacent fetuses, taking care not to inject individual fetal sacs. Mice underwent injection of the following: 1) LPS (TLR4 agonist, extracted from Salmonella enterica, L2262; Sigma, St. Louis, MO, USA; 0.025 mg/mouse) or 2) PBS. Surgical procedures lasted ∼10 min. The abdomen was closed in 2 layers, with 4-0 polyglactin sutures at the peritoneum and wound clips at the skin. Animals recovered in individual, clean cages in the animal facility. The doses of LPS used cause delivery within 18–24 h after injection and preterm delivery rate were 90%.

Hormonally induced PTL

MIF (RU486; Enzo Life Sciences, Farmingdale, NY, USA), a progesterone receptor antagonist, was used to induce premature progesterone withdrawal [8, 31]. MIF was solubilized in DMSO and stored in −20°C. This stock was diluted in 1 M acetic acid and diluted further in water at the time of use. A dose of 150 µg/animal was injected s.c. on gestation d 14.5, predictably causing delivery within 18–24 h after injection, and preterm delivery rate was 100%. The same volume of DMSO in 1 M acetic acid in water was used as a vehicle control.

Tissue harvest

Animals were euthanized 10 h after surgery or injection. The inoculated/right horn was incised longitudinally along the anti-mesenteric border. Gestational tissues uteri (decidual caps underlying placental attachment sites) and placentas were harvested, washed in ice-cold PBS, flash frozen in liquid nitrogen, and stored at −85°C for mRNA and protein extraction or fixed in 10% neutral-buffered formalin for immunohistochemistry.

Decidual and placental cell preparation

Uteri were dissected on d 14.5 of pregnancy, and decidual caps and placentas were collected. Decidua and placenta were prepared as single-cell suspensions, as described previously [32]. In brief, tissues were minced in HBSS (Life Technologies, Thermo Fisher Scientific, Grand Island, NY, USA), mechanically dispersed through a 100 µm nylon filter, and centrifuged at 1500 rpm. The remaining pellet was dispersed in RPMI medium at 10⁷ cells/ml in 48-well plates. Before plating, placental suspensions underwent red cell lysis by incubation with RBC lysis buffer (BioLegend, San Diego, CA, USA). The above specimens were incubated at 37°C in 5% CO₂/95% air for 1 h. Viability of ex vivo-cultured cells was >95%, as assessed by use of the Trypan blue dye exclusion test.

Extraction of decidual macrophages

Decidual macrophages were isolated as described previous by Co et al. [33], with slight modifications. Decidual caps collected on d 14.5 of pregnancy (i.e., in unmanipulated mice) were minced gently and incubated in 50 ml PBS containing 30 U collagenase type II (Gibco, Grand Island, NY, USA) in a shaking water bath at 37°C for 20 min. The collagenase reaction was stopped by washing with PBS, supplemented with 10% FBS. Cells were strained through a 70 μm nylon strainer to remove debris, washed with PBS, layered over 15 ml Ficoll-Paque (GE Healthcare Life Sciences, Pittsburgh, PA, USA), and centrifuged at 1200 rpm for 20 min at 4°C. A crude decidual leukocyte fraction was collected from the supernatant-Ficoll interface and washed twice by centrifugation in HBSS at 250 rpm for 5 min. For purification of decidual macrophages, F4/80⁺ macrophages were flow sorted from the decidual leukocyte fraction by use of anti-F4/80-allophycocyanin antibody (BioLegend) on FACSAria with FACSDiva software (BD Biosciences, San Jose, CA, USA). Isolated decidual macrophages (4 × 10⁵ cells/well) were cultured in DMEM high glucose (Gibco), supplemented with 10% FBS, 1% streptomycin, and 1% penicillin in 48-well plates at 37°C in 5% CO₂/95% air for 1 h.

Ex vivo treatment

Decidual and placental cells and decidual macrophages were incubated for 2 h in the presence of PBS or LPS (10 ng/ml), followed by 1 of the following experiments: 1) 10 h treatment with GSI (an inhibitor of Notch receptor processing; 20 μM; EMD Millipore, Billerica, MA, USA) or control (solvent for GSI; DMSO diluted in PBS at ∼1:1300); 2) 10 h treatment with or without mouse rDLL-1 (1 μg/ml; R&D Systems, Minneapolis, MN, USA), as described by Tsao et al. [22]; or 3) 2 h with anti-a2V (5 µg/ml; Covance, Denver, PA, USA) or IgG (same concentration as anti-a2V). All experiments were conducted in triplicate and repeated twice (i.e., 3 triplicate experiments).

Real-time PCR

Total RNA was extracted after homogenization in TRIzol reagent (Life Technologies, Thermo Fisher Scientific), according to the manufacturer’s protocol. For ex vivo experiments, cells were lysed in culture dishes, or cell pellets were homogenized in TRIzol. cDNA was prepared by use of qScript cDNA SuperMix (Quanta BioSciences, Gaithersburg, MD, USA). Duplex RT-PCR was performed with 1 primer pair amplifying the gene of interest and the other, an internal reference (GAPDH) in the same tube by use of the StepOne Real-Time PCR Systems (Applied Biosystems, Thermo Fisher Scientific). Semiquantitative analysis of gene expression was done by use of the ∆∆ comparative threshold method, normalizing expression of the gene of interest to Gapdh. Prevalidated TaqMan gene-expression assays were used for Dll-1 (Mm01279269_m1), Notch1 (Mm00435249_m1), Notch2 (Mm00803077_m1), Notch3 (Mm01345646_m1), Notch4 (Mm00440525_m1), Hes1 (Mm01342805_m1), a2V (Atp6v0a2, Mm00441848_m1), Jagged 1 (Mm00496902_m1), Jagged 2 (Mm01325629_m1), Dll-4 (Mm00444619_m1), VEGF (Mm01281449_m1), and control Gapdh (4352339E; Applied Biosystems, Thermo Fisher Scientific). Real-time PCR was performed by use of the Universal PCR Master Mix reagent (Applied Biosystems, Thermo Fisher Scientific).

Immunohistochemistry and immunofluorescence

Tissues fixed in 10% neutral-buffered formalin were processed as described previously [9, 10]. Sections (5 µm) from tissues were mounted onto silane-coated glass slides (Dako, Carpinteria, CA, USA) and stored at −80°C until used. The sections were submerged in sodium citrate buffer (pH 6) and heated for 10 min in a microwave oven for antigen retrieval.

For immunohistochemistry, sections were incubated with rabbit anti-Numb antibody (Abcam, Cambridge, MA, USA), overnight at 4°C. After washing, Dako EnVision + dual-link System, HRP (DAB+; Dako) was used to stain the frozen sections according to the manufacturer’s instructions and as described previously [9, 10]. The sections were counterstained with Mayer’s hematoxylin and mounted in Faramount Mounting Medium, Aqueous (Dako). The level of immune staining was evaluated by light photomicroscopy (Carl Zeiss, Weesp, The Netherlands) by use of a high-resolution camera (Canon G10; Canon, Tokyo, Japan).

Tissue immunostaining results were scored negative if no immunopositive tissue was present. The total score was based on the percentage of stained tissue and immunostaining intensity. The percentage of stained tissue and immunostaining intensity was calculated according to the method described in Teixeira Gomes et al. [34] and as described previously [9, 10]. The ISIS was generated by use of the following equation, SAS multiplied by the IIS: (ISIS = SAS × IIS).

For immunofluorescence, staining sections were incubated with goat anti-DLL-1, rabbit anti-Notch1, rabbit anti-Hes1, rat anti-F4/80 antibody, rabbit anti-Jagged 1, and rabbit anti-VEGF receptor (Abcam) or isotype-matched controls, overnight at 4°C. Primary antibodies were followed by secondary antibody rabbit anti-rat DyLight 488, goat anti-rabbit AF594, or FITC and donkey anti-goat AF594 (Abcam). Cells were fixed in ProLong Gold Antifade Reagent with DAPI (Invitrogen, Thermo Fisher Scientific) to visualize the nuclei. Antigen distribution was examined by use of a Nikon Eclipse TE2000-S florescence microscope (Nikon Instruments, Melville, NY, USA).

For semiquantitative analysis of immunofluorescence images, 5 squares with a fixed area (250 × 250 μm²) covering different regions of uterine decidua or placenta were positioned. The mean IDV within a defined threshold was measured by use of image analysis software (ImageJ; NIH Image; National Institutes of Health, Bethesda, MD, USA). Immunofluorescent measurements were obtained from a minimum of 6 sections and averaged. The average mean IDV were plotted [35].

Protein extraction and Western blot analysis

For protein extraction, cells were lysed, or tissues were homogenized in ice-cold 1× radioimmunoprecipitation assay buffer (Santa Cruz Biotechnology, Dallas, TX, USA), containing protease and phosphatase inhibitor (Roche Applied Science, Indianapolis, IN, USA). Lysates were incubated on ice for 30 min and centrifuged at 10,000 g for 10 min at 4°C. Supernatant fluid was collected and used as total cell lysates for protein assays. Samples were stored at −80°C until further use. Protein concentration was measured by bicinchoninic acid protein assay (Thermo Fisher Scientific).

Western blot analysis was performed as described previously [9, 10]. Equal amounts of protein (5–20 μg) from total cell lysates were separated by 4–12% SDS-PAGE and blotted onto polyvinylidene difluoride transfer membranes, which were blocked at room temperature for 1 h in 5% nonfat dry milk in TBST. Blots were incubated with primary antibody to rabbit anti-Notch1 (Santa Cruz Biotechnology) and/or rabbit anti-GAPDH (Cell Signaling Technology, Danvers, MA, USA), overnight at 4°C, and followed by donkey anti-rabbit IRDye 800CW secondary antibody (LI-COR Biosciences, Lincoln, NE, USA) for 1 h at room temperature. Quantification of bands was performed by use of ImageQuant software (Molecular Dynamics, Sunnyvale, CA, USA). Background intensity was subtracted from each sample, and then fold change was determined. Fluorescent blots were imaged on the Odyssey Infrared Imaging System (LI-COR Biosciences). To verify equal loading, membranes were reprobed for loading control GAPDH.

Flow cytometry

To evaluate macrophage polarization in decidual cells, suspensions were stained with F4/80-Pacific Blue (general macrophage marker), CD11c-FITC (M1 macrophage), and CD206-PE (M2 macrophage) antibodies with matched isotype-control (BioLegend) and analyzed on a BD LSR II flow cytometer (BD Biosciences) with FlowJo software (Tree Star, Ashland, OR, USA). For assessment of proportions of M1 and M2 macrophages out of total F4/80⁺ cells, gating was set for F4/80⁺CD11c⁺CD206⁺. Propidium iodide staining was performed to determine cell death, and there was no significant difference between control and treated groups. Three animals were used in each group.

Cytokine/chemokine assay

A panel of mouse proinflammatory cytokines (IL-1β, TNF-α, IL-6, IL-15, and IFN-γ), chemokines (MIP-1β, CCL5, and CCL2), anti-inflammatory cytokine (IL-10), and VEGF was assayed in medium or in total cell lysate by Milliplex Map kit (EMD Millipore) and assayed on a MAGPIX instrument (EMD Millipore), as per the instructions provided by manufacturer. Equal amounts of protein (50 μg) or medium were used for the assay, which was repeated 3 times with duplicates.

Statistical analysis

Continuous variables (e.g., relative mRNA levels) were assessed with 2-tailed Student’s t test or ANOVA. When data were not normally distributed, 2 groups were compared with Mann-Whitney U test.

RESULTS

DLL-1 ligand and Notch1 increase in inflammation-induced PTL but not in hormonally induced PTL

We tested the expression of Notch ligand (DLL-1), Notch receptors (Notch1–4), and transcriptional repressor (Hes1) during inflammation-induced PTL and hormonally induced PTL. A significant increase of DLL-1 mRNA expression was observed in the uterus from inflammation-induced PTL compared with control (Fig. 1). Expression remained unchanged in the uterus and placenta from the hormonally induced PTL group compared with their respective control (Fig. 1 and Supplemental Fig. 1). Immunofluorescence shows that the DLL-1 ligand is increased in uterine decidua and placenta from inflammation-induced PTL (Fig. 1 and Supplemental Fig. 1).

Figure 1. DLL-1 ligand increases during inflammation-induced PTL.

(A) The mRNA expression of DLL-1 in the uterus recovered from PTL groups [inflammation (LPS)- and hormonally (MIF) induced PTL] and controls (PBS for inflammation-induced PTL and mifepristone-control (MIF-C)/diluted DMSO for hormonally induced PTL); n = 7–11/group. (B) Immunofluorescence staining with corresponding mean IDV of DLL-1 (red) in the uterus recovered from inflammation-induced PTL and hormonally induced PTL. Nuclei stained with DAPI (blue) in merged images; n = 4–5/group. Six sections/animal were analyzed. Original magnification, 200×; original scale bars, 10 μm. PBS and LPS, IU injections on d 14.5; MIF-C, s.c. DMSO control; MIF, s.c. MIF in DMSO on d 14.5. Error bars ± sem. **P ≤ 0.01, significant difference vs. respective control.

Immunofluorescence shows the up-regulation of Notch1 protein in uterine decidua (Fig. 2A, B, and D) and placenta (Supplemental Fig. 2) in inflammation-induced PTL but no change in hormonally induced PTL. Western blot analysis correlated with immunofluorescence analysis and also showed that expression of Notch1 was increased in uterus (Fig. 2C and E) and placenta (Supplemental Fig. 2) in inflammation-induced PTL but no change in hormonally induced PTL. Notch2 and Notch4 mRNA levels remained unchanged (except lower Notch4 mRNA expression observed in the uterus) with undetectable Notch3 mRNA in the uterus and placenta from both of the groups studied (Supplemental Fig. 3).

Figure 2. Notch receptor increases during inflammation-induced PTL.

Immunofluorescence staining (A) with corresponding mean IDV (B and D) of Notch1 (green) in the uterus and nuclei stained with DAPI (blue) in merged images. Six sections/animal were analyzed. Original magnification, 200×; original scale bars, 10 μm. Sample Western blots with corresponding densitometric analysis (C and E) of Notch1 and GAPDH in the uterus. PBS group set to 1; n = 4–5/group. PBS and LPS, IU injections on d 14.5; MIF-C, s.c. DMSO control; MIF, s.c. MIF in DMSO on d 14.5. *P ≤ 0.05; **P ≤ 0.01, significant difference vs. respective control.

The nuclear localization of Hes1 was increased in uterine decidua of inflammation-induced PTL but remained unchanged in hormonally induced PTL compared with their respective control, as shown by immunofluorescence (Fig. 3 and Supplemental Fig. 3C). Isotype controls are also shown (Supplemental Fig. 4).

Figure 3. Hes1 expression increases with nuclear localization during inflammation-induced PTL.

Immunofluorescence staining of Hes1 (green) in the uterus recovered from inflammation-induced PTL. Nuclei stained with DAPI (blue) in merged images; n = 4–5/group. Six sections/animal were analyzed. Original magnification, 200× and 400×; original scale bars, 10 μm. PBS and LPS, IU injections on d 14.5.

The sum of the above results shows that Notch signaling is up-regulated in decidua and placenta from inflammation-induced PTL but not in hormonally induced PTL.

Numb decreases in inflammation-induced PTL, not in hormonally induced PTL

Numb is a protein complex with multiple cellular functions and a deregulator of Notch signaling. The Numb–Notch interaction regulates cell fate in various tissues, and loss of Numb results in activation of Notch signaling [20]. Therefore, we checked the expression of Numb in the uterus and placenta during inflammation-induced PTL and hormonally induced PTL. The expression of Numb was down-regulated in uterus and placenta from inflammation-induced PTL with no change in hormonally induced PTL (Fig. 4) compared with their respective control. Isotype control is also shown (Supplemental Fig. 4).

Figure 4. Numb expression decreases during inflammation-induced PTL.

Immunohistochemistry of Numb (brown) in the uterus and placenta recovered from inflammation-induced PTL and hormonally induced PTL. Corresponding ISIS are also shown (ISIS = SAS × IIS); n = 4–5/group. Six sections/animal were analyzed. Original magnification, 200×; original scale bars, 10 μm. PBS and LPS, IU injections on d 14.5; MIF-C, s.c. DMSO control; MIF, s.c. MIF in DMSO on d 14.5. **P ≤ 0.01, significant difference vs. respective control.

Inhibition of Notch signaling suppresses the inflammatory response in decidual and placental cells

Inhibition of Notch signaling by GSI (an inhibitor of Notch receptor processing) reduces inflammatory responses via suppression of activation of NF-κB [22, 36, 37]. To study the role of Notch signaling in PTL in the context of inflammation, decidual and placental single-cell suspensions were prepared from mouse tissues obtained on d 14.5 of pregnancy. Cells were cultured ex vivo and incubated with LPS or PBS for 2 h, followed by treatment with control or GSI for 10 h. GSI treatment suppresses Hes1 mRNA expression, showing the effect of GSI on Notch signaling suppression in decidual cells (Fig. 5A). GSI significantly decreased LPS-induced secretion of proinflammatory cytokines (TNF-α, IL-6, IL-15, and IFN-γ), the chemokine MIP-1β, and the anti-inflammatory cytokine IL-10 in decidual cells (Fig. 5B). Overall, similar findings were obtained in placental cells (Fig. 6).

Figure 5. Inhibition of Notch signaling prevents inflammatory responses in decidual cells.

(A) The mRNA expression of Hes1 after GSI (an inhibitor of Notch receptor processing) treatment showing the effect of GSI on Notch signaling suppression in decidual cells. Proinflammatory cytokines, chemokines, and anti-inflammatory cytokine were measured by Luminex assay in protein extracted from decidual cells recovered from mouse on d 14.5 of pregnancy, cultured ex vivo, and treated with PBS and LPS for 2 h, followed by treatment with control, (B) GSI for 10 h, or (C) rDLL-1 for 10 h; n = 3/group. Error bars ± sem. *P ≤ 0.05; **P ≤ 0.01, significant difference between LPS treated with control/GSI or control/DLL-1.

Figure 6. Inhibition of Notch signaling suppresses inflammatory responses in placental cells.

Proinflammatory cytokines and chemokines were measured by Luminex assay in protein extracted from placental cells recovered from mouse on d 14.5 of pregnancy, cultured ex vivo, and treated with PBS and LPS for 2 h, followed by treatment with control or GSI (an inhibitor of Notch receptor processing) for 10 h; n = 3/group. Error bars ± sem. *P ≤ 0.05; **P ≤ 0.01, significant difference between LPS treated with control/GSI.

DLL-1 ligand amplifies the inflammatory response in LPS-treated decidual cells

We explored further the role of Notch signaling in the enhancement of the inflammatory response. We measured the levels of different inflammatory cytokines and chemokines in the LPS-treated decidual cells (recovered on d 14.5 of pregnancy) in the presence or absence of rDLL-1 ligand. Cells were cultured ex vivo and incubated with LPS or PBS for 2 h, followed by treatment with or without rDLL-1 for 10 h. DLL-1 ligand significantly increased LPS-induced secretion of proinflammatory cytokines (TNF-α, IL-6, and IFN-γ) and chemokines (CCL5 and MIP-1β) in decidual cells. However, the level of anti-inflammatory cytokine IL-10 remained unchanged (Fig. 5C). Similar findings were observed in placental cells (data not shown) and in the mouse macrophage cell line [22]. These data further provide the evidence that Notch signaling ligand DLL-1 plays a critical role in the enhancement of inflammatory responses during inflammation-induced PTL.

Notch signaling molecules are associated with LPS-induced polarization of decidual macrophages to an M1 profile

We have reported a large infiltration of macrophages in mice in a PTL model induced by the combination of PGN + poly(I:C) [9]. Among several lines of evidence suggesting that macrophages are critical cells for the process of infection- and inflammation-induced PTL is a report by Gonzalez et al. [38], showing that macrophage depletion by anti-F4/80 antibody prevented preterm delivery in LPS-treated mice. Therefore, we studied Notch signaling in decidual macrophages during inflammation-induced PTL. Double-immunofluorescence staining of F4/80 and DLL-1 ligand showed that decidual macrophages have higher levels of DLL-1 ligand during inflammation-induced PTL (Fig. 7A). No infiltration of macrophages was observed in hormonally induced PTL (data not shown); this finding is also supported by a report from a different group [39]. Isotype controls are also shown (Supplemental Fig. 4).

Figure 7. DLL-1 ligand increases in decidual macrophages during inflammation-induced PTL and LPS-induces polarization of decidual macrophages to M1 profile.

(A) Colocalization of macrophages stained with F4/80 (green) with DLL-1 (red), and merged images of F4/80 and DLL-1 in the uterus from PBS- and LPS-treated groups; n = 4–5/group. Six sections/animal were analyzed. Original magnification, 400×; original scale bars, 10 μm. PBS and LPS, IU injections on d 14.5. (B) Representative flow cytometry analysis of decidual cells, freshly recovered from control and inflammation-induced PTL mouse on d 14.5 of pregnancy. F4/80⁺ macrophages were gated for M1 (CD11c-FITC), M2 (CD206-PE), and M1 and M2 double-positive macrophages. Q1, M1 macrophages; Q2, M1⁺M2⁺ macrophages; Q4, M2 macrophages. PBS and LPS, IU injections on d 14.5; n = 3/group.

We have reported that during PGN + poly(I:C)-induced PTL, decidual macrophages polarize to a M1⁺M2⁺ (i.e., both CD11c⁺ and CD206⁺) double-positive phenotype [9]. Activation of Notch signaling promotes the polarization of macrophages to the IFN-γ-producing M1 macrophage subtype [21]. Therefore, we studied the polarization of decidual macrophages during inflammation-induced PTL. Flow cytometry analysis showed that F4/80⁺ decidual macrophages from inflammation-induced PTL are polarized to M1 (CD11c⁺, 35.7%, 261 cells) with an increase of M1⁺M2⁺ (CD11c⁺ and CD206⁺, 32.8%, 248 cells) double-positive phenotype (Fig. 7B) compared with control (M1, 16.4%, 113 cells; M1⁺M2⁺, 16.2%, 112 cells). Isotype controls are also shown. Therefore, we further explored the role of Notch signaling in decidual macrophage polarization.

The mRNA expression of DLL-1 ligand, Notch1, and Hes1 was significantly higher in decidual macrophages extracted from mice on d 14.5 of pregnancy and treated with LPS ex vivo (Fig. 8A). Notch signaling inhibitor GSI significantly decreases LPS-induced cytokines, confirming to the M1 profile (IL-1β, TNF-α, IL-6, and IFN-γ) and the M2 profile (IL-10) and also chemokine (CCL2). However, the secretion of IL-15 (M1 profile) remained unchanged (Fig. 8B). These data provide evidence that Notch signaling might be involved in the secretion of M1- and M2-associated cytokines and suggest the importance of Notch signaling in decidual macrophage polarization during inflammation-induced PTL. We have reported earlier [10] and confirmed here that during hormonally induced PTL, the secretion of proinflammatory and anti-inflammatory cytokines remained at base level. This observation is also supported by a report from a different group [39].

Figure 8. Inhibition of Notch signaling prevents the inflammatory responses in decidual macrophages.

(A) The mRNA expression of DLL-1, Notch1, and Hes1 in decidual macrophages recovered from mouse on d 14.5 of pregnancy, cultured ex vivo, and treated with PBS and LPS for 2 h. **P ≤ 0.01, significant difference between LPS and PBS. (B) M1 and M2 macrophage profile-associated cytokines and chemokines were measured by Luminex assay in medium from decidual macrophages treated with PBS and LPS for 2 h, followed by treatment with control or GSI for 10 h; n = 3/group. *P ≤ 0.05; **P ≤ 0.01, significant difference between LPS treated with control/GSI. Error bars ± sem.

a2V blockade induces Notch signaling in decidual macrophages

We have reported previously that a2V is decreased significantly in the uterus and placenta of inflammation-induced PTL mice. Moreover, the blocking of a2V by anti-a2V antibody enhances the expression and secretion of proinflammatory cytokines at baseline and after treatment with TLR ligands in RAW 264.7 macrophages [9, 10] and in decidual and placental cells [10]. Here, we correlated this effect with Notch signaling. Decidual macrophages were treated with PBS or LPS for 2 h, followed by 2 h incubation with IgG control or anti-a2V antibody. LPS treatment suppresses the mRNA expression of a2V (Fig. 9A). LPS induces mRNA expression of DLL-1 ligand (Fig. 9B) and mRNA and protein expression of Notch1 receptor (Fig, 9C), an effect that was amplified with blockade of a2V (Fig. 9B and C). This suggests that reduction of a2V activity might be associated with activation of Notch signaling.

Figure 9. a2V blockade induces Notch signaling in decidual macrophages.

(A) The mRNA expression of a2V in decidual macrophages recovered from mouse on d 14.5 of pregnancy, cultured ex vivo, and treated with PBS and LPS for 2 h. **P ≤ 0.01, significant difference between LPS and PBS. The mRNA expression of (B) DLL-1 and (C) Notch1 and sample Western blots with corresponding densitometric analysis of Notch1 and GAPDH (PBS group set to 1) in decidual macrophages treated with PBS and LPS for 2 h, followed by treatment with PBS, IgG control, or anti-a2V antibody for 2 h; n = 3/group. **P ≤ 0.01, significant difference. Error bars ± sem.

Angiogenic factors decrease during inflammation-induced PTL, not in hormonally induced PTL

Regulation of angiogenesis is another important function of the Notch signaling pathway, apart from the regulation of inflammation [40]. The Notch ligands Jagged 1 and 2 and DLL-4 are indispensable for vascular development during fetal growth [40–42]. VEGFs are other specific angiogenic factors critical for angiogenesis [43, 44] and are regulated by Notch ligands Jagged 1 and 2 and DLL-4 [45, 46]. Decreased VEGF causes reduction in placental vascularization during late pregnancy and leads to restricted fetal growth [47] and preterm birth [48]. Therefore, we determined the expression of these factors in the uterus and placenta during inflammation-induced PTL and hormonally induced PTL.

The mRNA expression of Jagged 1 and 2 and DLL-4 was decreased significantly in the uterus and placenta during inflammation-induced PTL (except undetectable Jagged 2 in the uterus and no change in DLL-4 in the placenta), with no change in hormonally induced PTL (Figs. 10A and 11). Immunofluorescence staining confirms decreased protein expression of Jagged 1 in the placenta during inflammation-induced PTL with no change in hormonally induced PTL (Fig. 10B).

Figure 10. Jagged 1 decreases during inflammation-induced PTL.

(A) The mRNA expression of Jagged 1 in the placenta recovered from inflammation-induced PTL and hormonally induced PTL; n = 7–11/group. (B) Immunofluorescence staining with corresponding mean IDV of Jagged 1 (green) in the placenta. Nuclei stained with DAPI (blue) in merged images; n = 4–5/group. Six sections/animal were analyzed. Original magnification, 200×; original scale bars, 10 μm. Error bars ± sem. *P ≤ 0.05; **P ≤ 0.01, significant difference versus respective control. PBS and LPS, IU injections on d 14.5; MIF-C, s.c. DMSO control, MIF, s.c. MIF in DMSO on d 14.5.

Figure 11. Expression of angiogenesis-associated Notch receptors during inflammation-induced PTL and hormonally induced PTL.

The mRNA expression of Jagged 1 and DLL-4 in the uterus (A) and Jagged 2 and DLL-4 in the placenta (B) recovered from PTL groups; n = 7–11/group. PBS and LPS, IU injections on d 14.5; MIF-C, s.c. DMSO control, MIF, s.c. MIF in DMSO on d 14.5. Error bars ± sem. *P ≤ 0.05; **P ≤ 0.01, significant difference vs. respective control.

The mRNA expression of VEGF was decreased in the placenta during inflammation-induced PTL, with no change in hormonally induced PTL (Fig. 12A). Moreover, LPS treatment in ex vivo-cultured placental cells significantly down-regulates VEGF secretion (Fig. 12B).

Figure 12. VEGF decreases during inflammation-induced PTL.

(A) The mRNA expression of VEGF in the placenta recovered from inflammation-induced PTL and hormonally induced PTL; n = 7–11/group. (B) Concentration of VEGF in protein extracted from placental cells recovered from mouse on d 14.5 of pregnancy, cultured ex vivo, and treated with PBS and LPS for 2 h; n = 3/group. (C) Immunofluorescence staining with (D) corresponding mean IDV of VEGF-R (green) in the placenta. Nuclei stained with DAPI (blue) in merged images; n = 4–5/group. Six sections/animal were analyzed. Original magnification, 200×; original scale bars, 10 μm. PBS and LPS, IU injections on d 14.5. Error bars ± sem. **P ≤ 0.01, significant difference vs. respective control or PBS.

VEGF-R was also measured in the placenta during inflammation-induced PTL. Immunofluorescence staining shows that protein expression of VEGF-R was decreased in the placenta during inflammation-induced PTL (Fig. 12C and D).

DISCUSSION

Here, we have reported that during inflammation-induced PTL, the regulation of Notch signaling plays a critical role in the regulation of inflammation during PTL. Several lines of evidence identified in these studies support the proposal of an essential role of Notch signaling in inflammation-induced PTL (Fig. 13). 1) In inflammation-induced PTL, the DLL-1 ligand, Notch 1 receptor, and Notch transcription repressor Hes1 are increased, and Notch deregulator Numb is decreased in the uterus and placenta but remained unchanged in hormonally induced PTL. 2) Activation of Notch signaling is involved in polarization of decidual macrophages (a key mediator of cells in inflammation-induced PTL) to M1, and inhibition of Notch signaling suppresses the secretion of M1-associated (proinflammatory) cytokines. 3) The increase of Notch signaling is associated with reduced activity of a2V. 4) Notch inhibitor GSI suppresses the inflammation in vitro. 5) During inflammation-induced PTL, angiogenesis-associated Notch ligands (Jagged 1 and 2 and DLL-4) and VEGF are reduced in the placenta and are required for the proper placental development and pregnancy maintenance. These results suggest that the activation of the Notch signaling-mediated inflammatory response and suppression of angiogenesis-associated Notch ligands simultaneously play a key role in the induction of inflammation-induced PTL.

Figure 13. Schematic summary of the main findings from this study.

Increased expression of DLL-1 ligand on macrophages plays a critical role in the induction of IFN-γ [49]. Here, we have shown that DLL-1 ligand and Notch1 were up-regulated during inflammation-induced PTL, which activates Notch signaling at the feto-maternal interface, as shown by increased levels of Hes1. Our observation is that blockade of Notch signaling by GSI suppresses the LPS-induced inflammatory responses in decidual and placental cells. Furthermore, the addition of exogenous ligand DLL-1 synergistically stimulates the LPS-induced inflammatory responses in decidual and placental cells. Together, these data suggest that up-regulation of Notch signaling enhances the inflammatory responses during inflammation-induced PTL. The role of Notch signaling activation in the enhancement of inflammatory response is in agreement with other studies in different models [22, 36, 50].

a2V is specifically expressed on the plasma membrane [51, 52], early endosomes, and lysosomes [53], suggesting a possible role in the receptor endocytosis pathway. Previously, we have reported that a2V expression was decreased significantly at the feto-maternal interface during inflammation-induced PTL, and its suppression alters the autophagic flux [10] and enhances the inflammatory response [9]. One of the key functions of V-ATPases is to maintain normal autophagic flux via acidification of lysosomes [10, 51]. Autophagy is the process by which cellular organelles and proteins undergo turnover and recycling. Inhibitors of V-ATPase, such as bafilomycin A1, hinder autophagosome turnover by inhibiting fusion with lysosomes [10, 54]. It has been reported previously that disturbed endocytosis pathways induce the retention of Notch signaling molecules in endosomal vesicles, which enhances NICD cleavage and enhances Notch signaling [12–15, 55]. These distinct outcomes could be associated with the fact that endocytosis is required not only for the activation of internalized receptors but also for degradation of the receptor in the lysosomes [56, 57]. The up-regulation of Notch signaling molecules in inflammation-induced PTL might be associated with blocking of the endolysosomal system via a reduced level of a2V. In the present study, we show that a2V inhibition also increases the expression of the DLL-1 ligand and Notch1 receptor in decidual macrophages. Previously, we have shown that the inhibition of a2V leads to hyperinflammation during inflammation-induced PTL via altering the autophagic flux [10]. Collectively, these data suggest the possible role of a2V in the enhancement of inflammatory responses, in part, via activation of Notch signaling.

Activation of Notch signaling promotes the polarization of macrophages toward the M1 subtype [21]. Here, we confirm that Notch signaling is involved in the polarization of decidual macrophages. Our macrophage phenotype analysis showed that LPS treatment induces the macrophages polarization toward M1 macrophages and M1⁺M2⁺ double-positive macrophages, a development suppressible by the Notch inhibitor GSI, as shown by the suppression of M1 and M2 cytokines secretion. These double-positive macrophages and their cytokine output (both M1 and M2 cytokines) may contribute to recruitment and activation of innate-immune defenses at the inflamed sites of uterus and thus, might be helpful for rapid promotion and resolution of inflammation. Similar double-positive macrophages have been observed during PGN + poly(I:C)-induced PTL [9].

We explored further another crucial Notch-regulated process required for placental development, i.e., angiogenesis [26, 27]. Notch ligands (DLL-4, Jagged 1 and 2), angiogenic factor (VEGF), and VEGF-R (previously shown to be essential for normal trophoblast function and placental angiogenesis) were reduced significantly during inflammation-induced PTL and upon LPS treatment ex vivo. Therefore, it seems reasonable to suggest that inappropriate placental expression of angiogenic factors (Notch ligands DLL-4, Jagged 1 and 2, and VEGF) may contribute to placental vascular defects and placental dysfunction and thereby, be an important cause of creating hypoxia and hence, PTL.

In summary, we have demonstrated that the simultaneous activation of a Notch signaling-mediated inflammatory response and suppression of angiogenic Notch ligands may be key mechanisms involved in the initiation or progression of inflammation-induced PTL (but not during hormonally induced PTL). The activation of Notch signaling is associated with macrophage polarization and decreased a2V activity. Notch inhibition suppresses the inflammation, and its use in the treatment of PTL remains to be explored.

AUTHORSHIP

V.A. and M.K.J. designed and performed research, analyzed data, and wrote the paper. S.P., G.K.K., and A.K. performed research and analyzed data. A.G.-S., E.H., and K.D.B. designed research, analyzed data, and helped in writing the paper.

Glossary

a2V

a2 isoform of vacuolar ATPase

DLL

Delta-like protein

GSI

γ-secretase inhibitor

Hes1

hairy and enhancer of split-1

IDV

integrated density value(s)

IIS

immunostaining intensity score

ISIS

immunostaining index score

IU

intrauterine

MIF

mifepristone

NICD

Notch intracellular domain

PGN

peptidoglycan

poly(I:C)

polyinosinic:polycytidylic acid

PTL

preterm labor

SAS

stained area score

s.c.

subcutaneous

V-ATPase

vacuolar-type H+-ATPase

VEGF

vascular endothelial growth factor

DISCLOSURES

The authors declare no conflicts of interest.