Resistance to Lipopolysaccharide-Induced Preterm Delivery Mediated by Regulatory T Cell Function in Mice

Abstract

Intrauterine or intraperitoneal administration of lipopolysaccharide (LPS) into normal mice at midgestation induces preterm delivery (PTD) within 24 h through a mechanism dependent on Toll-like receptor signaling and expression of inflammatory cytokines. The exact participants in the cellular network involved in PTD are not known. Although the activities of innate immune cells are thought to be important, the extent to which this process depends on T and B cells has yet to be examined. Mice deficient in T and B cells due to genetic deficiency in the recombination activating gene 1 (Rag1^−/−) were given LPS intraperitoneally on Day 15 of gestation and found to be susceptible to LPS-induced PTD. This was found to involve many of the inflammatory mediators reported as important in normal mice. Moreover, at a low dose (3 μg), pregnant Rag1^−/− mice were found to be more susceptible to PTD than a cohort of normal mice on the same genetic background. This increased susceptibility was partially reversed by transfer, on Day 10 of gestation, of whole lymphocytes or purified CD4⁺ T cells. Transfer of purified CD4⁺ T cells to Rag1^−/− mice resulted in a uterine draining node population of FOXP3⁺ cells, suggesting that these cells may contribute to resistance to LPS-induced PTD. Overall, the data suggest that, although T and B lymphocytes are not critical positive regulators of LPS-induced PTD, CD4⁺ T cells play a protective and regulatory role, and thus could be a target for preventive or therapeutic manipulation.

Mice deficient in T cells are more susceptible to Lipopolysaccharide-induced premature delivery; this is partially reversed by administration of whole lymphocytes or CD4⁺ T cells.

INTRODUCTION

Preterm delivery (PTD) continues to present a major clinical problem, even in developed countries. Approximately 12% of all deliveries in the United States are preterm, accounting for 70% of perinatal mortality, and nearly half of long-term neurological morbidity [1]. An increased incidence of PTD has been associated with multiple infectious processes, including sexually transmitted diseases, severe viral infections, and pyelonephritis [2, 3]. Infection is thought to induce PTD through activation of inflammatory responses in both maternal and fetal tissues. This process initiates via signals through receptors that recognize conserved pathogen-associated molecular patterns expressed by a wide spectrum of infectious microorganisms, including bacteria, viruses, fungi, and protozoa [4].

Toll-like receptors (TLR) are an important subset of these receptors, and are expressed in the uterus and placenta [5]. The TLR pathway also recognizes endogenous (including mammalian) DNA, RNA, heat shock proteins, interferons, interleukin (IL) 1, CD40 ligand, and breakdown products of hyaluronic acid, which may be released by injured or infected cells [6].

Activation of the TLR pathway is thought to induce local production of chemokines and proinflammatory cytokines [7], which may result in elevated monocyte and neutrophil chemotaxis. In support of this, the ligands for TLR3 (e.g., poly I:C, the viral RNA analog) or for TLR4 (e.g., bacterial lipopolysaccharide [LPS]) in human first-trimester trophoblast cultures significantly increases culture expression of IL1, IL6, and tumor necrosis factor (TNF) [8]. Furthermore, studies in humans with PTD have demonstrated higher amniotic fluid concentrations of IL6 and TNFα than in women delivered at term [9].

The proposed mechanism for PTD in humans further supposes the activation of macrophages or dendritic or epithelial cells by TLR pathways, which triggers further cycles of inflammatory response in maternal and fetal tissues mediated through proinflammatory cytokine production. This, in turn, leads to the production of prostaglandins and matrix metalloproteinases, cervical ripening, increased myometrial contractility, rupture of fetal membranes, and, ultimately, PTD [10–15].

Negative regulation of PTD is thought to occur through the activities of molecules such as antiprostaglandins [16], progestational agents [17, 18], and antiinflammatory cytokines, such as IL10 [19, 20], which may down-regulate the process on many levels.

Although a significant amount of data in the literature implicates cytokines, hormones, and other soluble factors as elements supportive of, or protective against PTD, we still do not fully understand the cellular participants in this phenomenon, nor the kinetics of their interaction.

A mouse model of PTD utilizes either intraperitoneal or intrauterine injection of LPS during the last third of gestation, thereby leading to inflammation within the pregnant uterus [13]. The human correlate of this model includes preterm labor associated with either nonspecific inflammation or the infection with specific microorganisms.

Based on the dependence on TLR activation and the cytokines produced in the response, it is thought that several cells of the innate immune response are involved as positive regulators. Depletion of natural killer (NK) cells (for examples, see [21, 22]) abrogates the response, suggesting that these cells are in the group of critical positive regulators of PTD. In addition, evidence suggests that T and B lymphocytes may not be in this group of critical positive regulators [16]. However, the extent to which this is true has not been fully tested.

Most biological responses are controlled by negative as well as positive regulators. However, the question of whether T and B cells participate in negative regulation of LPS-induced PTD has not been answered. Recently, there has been increased interest in a subpopulation of T cells, the principal function of which is to regulate immune responses [23]. Regulatory T cells are thought to play a role in responses to, for example, tumor [24] and fetal [25] alloantigens. In addition, there is evidence of T regulatory function in the control of pathologic innate immune responses [26], including recurrent early fetal resorption [27]. B-cell subsets may also be negative regulators of inflammatory immune responses (e.g., [28]). These data and others raise the hypothesis that T and/or B cells may play roles as negative regulators of LPS-induced PTD.

To test this, we examined mice with no functional recombination activating gene 1. These mice are deficient in T and B cells, but have macrophages, dendritic cells, and NK cells. Although we found that T and B cells are not critical for LPS-induced PTD, CD4⁺ T cells play a protective role in this phenomenon.

MATERIALS AND METHODS

Animals

Normal C57BL/6J (B6/J) or C57BL/6J Rag1^−/− (Rag1^−/−) mice (8–10 weeks old) were purchased from the Jackson Laboratory (Bar Harbor, ME). C57BL/6-Tg (UBC-green fluorescent protein [GFP]) mice (B6-GFP) that express enhanced GFP under the direction of the human ubiquitin C promoter were kindly provided by Dr. Philippa Marrack (National Jewish Medical Center, Denver, CO), and were used, in some cases, as a source of donor cells. All animals were treated in accordance with the Guide for Care and Use of Laboratory Animals and with the approval of the Animal Care and Use Committee of the University of Vermont. The animals were allowed free access to food and water at all times and were maintained on a 12L:12D cycle. Rag1^−/− or normal B6/J females mated to same-strain males. Timed mating procedures included 3 days of seasoning before 24 h of mating. Day 0 of 19 was considered to be the day of plugging.

On Day 15 of pregnancy, Rag1^−/− mice were injected intraperitoneally with 0.6, 3, 10, or 30 μg LPS from Escherichia coli 055:B5 (Sigma-Aldrich, St. Louis, MO) in PBS (Mediatech, Herndon VA), or PBS alone. After 24 h, the mice were euthanized and examined for delivery of at least one pup. This was done by opening the uterus and examining it for early loss (necrotic, small, no discernable fetus or placenta), intact fetal-placental units (i.e., not delivered), and placental sites with attached placental tissue consistent with Day 15 of gestation, but without intact membranes or attached fetuses (indicative of delivery). Normal B6/J mice at Day 15 of gestation were given either 3 or 30 μg LPS or PBS. In a separate set of experiments, Day 15 pregnant Rag1^−/− mice were injected with 3 μg of LPS or PBS and euthanized 2, 6, or 24 h later. Peripheral blood and uterine tissues were collected for analysis.

Preparation and Administration of Lymphocyte Populations

Whole spleen and lymph nodes were isolated from normal adult B6/J or B6-GFP females. Single-cell suspensions were prepared in PBS or serum-free medium and 10 million cells were transferred intravenously into Rag1^−/− mice at 10 days of gestation. In other experiments, single-cell suspensions of whole spleen and lymph node cells were depleted with a cocktail of antibodies to CD8, protein tyrosine phosphatase receptor type C (PTPRC, B220), CD49, killer cell lectin-like receptor subfamily B member 1C (KLRB1C, NK1.1), CD11B, LY76 (Ter119), and major histocompatability complex (MHC) class II (see Flow Cytometry section for clones and sources) and a secondary reagent conjugated to magnetic beads (Miltenyi Biotech Auburn CA; Qiagen, Valencia, CA) to obtain CD4⁺ T cells. Two million CD4⁺ T cells were administered intravenously also on Day 10 of gestation to Rag1^−/− mice. On Day 15 of gestation, mice injected with no cells, whole lymphocytes, or CD4⁺ T cells were given 3 μg LPS intraperitoneally and observed for PTD 24 h later.

Real-Time RT-PCR

Total RNA was extracted from Rag1^−/− uteri using TRIZOL reagent (Invitrogen, Carlsbad, CA), according to the manufacturer's instructions. Single-stranded cDNA was synthesized from 2 μg of total RNA using random primers. Real-time PCR on cDNA was performed using the following primer pairs (from Operon Biotechnologies Huntsville, AL) to determine expression level: IL6 (F = AGAAAGACAA AGCCAGAG TCCTTCA G, R = GTCCTTAGCCACTCCTTCTGTGACT); Nfkappab1 (F = ATGATAGCAAAGCCCCG AATG, R = GTCCCCAAATCCTTCCCAAACT); Myd88 (F = AACAAAGGAACTGGGA GGCAT, R = ACTTGGTCTGGAAGTCACATTC); and Ptgs2 (also called Cox2) (F = ATATCAGGTCATTGGTGGAGAG, R = TGGTGGC ATACATCAT CAGA).

RT-PCR was performed for samples and normalized against the level of the housekeeping gene Ywhaz [29] (F = GCA ACG ATG TAC TGT CTC TTT TGG, R=GTCCACAATTCCTTTCTTGTCATC).

Serum Cytokine Analysis

To quantify serum cytokines, whole peripheral blood was collected and stored at room temperature for 2 h. After 20 min of centrifugation at 2500 rpm, serum was removed and stored frozen until use. We used the Mouse Inflammation Antibody Array Kit (RayBiotech, Inc., Norcross, GA) to analyze cytokine expression, per the manufacturer's instruction. Briefly, duplicate anti-cytokine antibody-coated membranes were incubated in 2 ml blocking buffer for 30 min. Cryopreserved serum samples were thawed and pooled (n = 3–4) and then diluted 1:10 with blocking buffer and added to the membranes. After 24-h incubation at 4°C with agitation, membranes were washed five times and incubated with a cocktail of biotin-conjugated cytokine-specific antibodies. After a further 24-h incubation, membranes were washed five times and incubated with HRP-conjugated streptavidin. Membranes were then incubated for 5 min in the revelation solution provided with the kit and then visualized using the Chemi Doc XRS detection system (Bio-Rad, Hercules, CA). Values obtained when the signal reached plateau were used for comparison. Raw optical density values obtained from the samples were corrected by subtraction of the optical density value obtained from an internal blank. These corrected values were, in turn, normalized to an internal positive control, thereby allowing comparisons between experiments.

Single-Cell Isolation from the Uterus

Decidual cells were isolated using a modification of a protocol previously reported by Kämmerer et al. [30]. Briefly, specimens were dissected free of placenta and fetus, and washed in PBS. The uterine tissue was minced into small fragments and digested for 20 min at 37°C under slight agitation in PBS with 200 U/ml hyaluronidase (Sigma-Aldrich), 1 mg/ml liberase blendzyme 3 (Roche, Indianapolis, IN), 0.2 mg/ml DNase I (Sigma-Aldrich), and 1 mg/ml bovine serum albumin (BSA)/fraction V (Sigma-Aldrich). The cell suspension obtained was filtered through 40-μm nylon mesh (BD Falcon, Franklin Lakes, NJ), and washed twice in PBS.

Flow Cytometry

Uterine draining lymph nodes (paraaortic and ileoinguinal) and spleen were isolated and single-cell suspensions prepared in PBS with 0.1% BSA (Sigma-Aldrich). Blocking of nonspecific binding was performed by preincubation of cell suspensions with an anti-CD16/32 (FcγIII/II receptor antibody; BD Biosciences, San Diego, CA). Samples were then incubated with one or more of the following antibodies (all anti-mouse and all from BD Biosciences, unless noted): PE-anti-TCRβ chain (clone H57–597); APC Cy7- anti-CD4 (GK1.5); PE-anti-MHC class II (M5/114.15.2); fluorescein isothiocyanate-anti-CD14 (MϕP9); APC-anti-CD11C (HL-3); PECy5.5-anti-CD45.2 (104); APC-Cy7 GR1 (BB6-8C5); or Texas Red-anti-CD11B (M1/70.15; Invitrogen).

For the identification and phenotypic analysis of T regulatory cells (CD4⁺TCR⁺ FOXP3⁺), cells were first stained extracellularly with APC-anti-TCRβ (H57–597; eBioscience, Inc. San Diego, CA) and Texas Red-anti-CD4 (RM4–5; Invitrogen). This was followed by intracellular staining with PE-anti-mouse/rat FOXP3 (FJK-16a) using the mouse/rat FOXP3 Staining Kit (eBioscience, Inc.), per the manufacturer's instructions. Viable cells were selected for flow cytometric analysis (LSR II; BD, San Jose, CA) based on forward- and side-scatter light properties. The analysis was performed with FlowJo software (Tree Star Software, Inc., Portland, OR).

Statistical Analysis

Data obtained by flow cytometry, including percentages and mean fluorescence intensity, were analyzed using GraphPad PRISM (Graphpad Software, Inc., San Diego, CA). Data from groups of mice were compared by the Mann-Whitney U-test. Chi-square analyses were performed using StatTools 1.0 for Excel (Palisade Corp., Ithaca, NY). Fischer exact analyses were performed using GraphPad PRISM. Statistical significance was set at P < 0.05.

RESULTS

LPS-Induced PTD in Mice Deficient in T and B Cells

To examine the susceptibility of Rag1^−/− mice to LPS-induced PTD, timed-pregnant mice at Day 15 of gestation were injected with 0.6, 3, 5, 10, or 30 μg of LPS in PBS intraperitoneally and observed for delivery of at least one pup during the subsequent 24-h period. As observed in Figure 1, the administration of a single dose of LPS resulted in PTD in a dose-dependent manner. Thus, LPS-induced PTD occurs in the absence of T or B cells.

FIG. 1.

LPS induces premature delivery in Rag1^−/− mice in a dose-dependent manner. Rag1^−/− females were mated to same-strain males, and, on Day 15 of gestation, they were given either PBS (dose = 0) or varying doses of LPS intraperitoneally. After 24 h, the mice were euthanized and observed for delivery of pups. Y axis, proportion of pregnant females that delivered at least one pup prematurely; x axis: dose of LPS used. n, number of mice tested.

The Systemic Response to LPS in Pregnant Mice Without T and B Cells: Inflammatory Cytokines in the Serum of Rag1^−/− Mice

There is evidence that proinflammatory cytokine secretion increases during PTD in humans and also in animal models of LPS-induced PTD [13, 31, 32]. Proinflammatory cytokines have been shown to stimulate the production of prostaglandins and matrix metalloproteinases in cultured amniotic epithelial cells, amniotic fibroblasts, chorionic trophoblasts, and decidua cells. Increased levels of prostaglandins and matrix metalloproteinases, in turn, foster myometrial contractions and collagen degradation in fetal membranes [33–36]. In order to examine cytokine release into the systemic circulation in pregnant Rag1^−/− mice, we choose the lowest LPS concentration that still resulted in a high PTD rate after LPS injection (i.e., 3 μg). Serum samples were collected 6 h after injection with 3 μg LPS or vehicle (PBS); pooled samples were then analyzed by protein array for inflammatory cytokines.

The data shown in Table 1 are from a representative array (LPS, n = 4 mice; PBS, n = 4 mice). IL6, CSF3, and CXCL1 were the factors most significantly up-regulated for expression and released into the blood of LPS-injected mice, although IL12A, IL12B, MCP1 (CCL2), and RANTES (CCL5) also were up-regulated modestly. In other experiments, eotaxin (CCL11) and tissue inhibitor of metalloproteinase 1 were also modestly elevated in tissues from Rag1^−/− mice given LPS (data not shown). This suggests that the innate immune system, present in Rag1^−/− mice, mediates the response to LPS.

TABLE 1.

Serum response in Rag^−/− mice given 3 μg of LPS.

LPS Up-Regulates the TLR4 Activation Pathway in the Uterus of Rag1^−/− Mice

LPS can bind to TLR4 and its coreceptor, CD14, leading to activation of adapter protein myeloid differentiation factor 88 (MYD88)-dependent and MYD88-independent signaling pathways. Downstream signaling leads to activation of transcription factors, such as nuclear factor (NF) kappaB1 and FOS, part of AP-1, resulting in increased expression of proinflammatory cytokines, such as IL6 and TNF, and also of the enzyme prostaglandin-endoperoxide synthase (PTGS2 also called cyclooxygenase-2 [COX2]) [37–39].

To analyze expression of Myd88, Nfkappab1, Il6, and Ptgs2 (Cox2) in uterine tissue, Rag1^−/− mice were euthanized 2 h after injection with 3 μg LPS or PBS. RNA was extracted from the uterus, and quantitative RT-PCR was performed. Ywhaz mRNA levels were used to normalize the measured values. Figure 2 shows that mice injected with LPS showed increased expression of all these genes, as well as Fos, compared with mice injected with PBS. The data suggest that, following LPS injection, TLR4 pathway responses in pregnant Rag1^−/− mice are similar to those previously documented in normal pregnant and nonpregnant mice given LPS.

FIG. 2.

LPS-induced gene up-regulation in the uterus of pregnant Rag1^−/− mice. Rag1^−/− females were mated to same-strain males, and, on Day 15 of gestation, they were given either PBS or 3 μg of LPS. At 6 h after injection, the uteri were dissected away from associated placenta and used for RNA isolation. Isolated RNA was assayed for gene expression by quantitative RT-PCR. Y axis, relative expression; x axis, specific genes. Black symbols, uterine RNA from LPS-treated mice; grey symbols, uterine RNA from PBS-treated mice. Each symbol represents tissue from an individual mouse. *Significant increase compared with PBS treatment by the Mann-Whitney U-test. Lines represent the medians for each group.

Concurrent with increased activation of the TLR4 pathway in the uterus is an increase in the activation of immune cells within the uterus. At 6 h, as depicted in the example shown in Figure 3A, we observed a modest increase in the proportion of dendritic cells (i.e., cells positive for CD11C and MHC class II [13.5% versus 6.74% in the example shown]). This was consistent with a quantitatively increased expression of MHC class II in the CD11C⁺ pool. When we examined macrophages, as defined by surface expression of CD11B and CD14 (Fig. 3B, left panels), we did not observe an increased proportion with respect to the leukocyte population. We did, however, observe an increased expression of MHC class II in the CD11B-positive pool (Fig. 3B, histogram at right). Thus, LPS-induced antigen-presenting cell activation can occur in the uterus of pregnant Rag1^−/− mice.

FIG. 3.

LPS induces antigen-presenting cell activation in the uterus in Rag1^−/− mice. Rag1^−/− females were mated to same-strain males, and, on Day 15 of gestation, they were given either PBS or 3 μg of LPS. After 6 h, the mice were euthanized and uteri from 3–4 mice from each treatment group were used to isolate lymphocytes, pooled and examined by flow cytometry. A) Increase in the presence of CD11C⁺ cells that are MHC class II high. Upper plot: uteri of mice given PBS (control). Lower plot: uteri of mice given LPS. Y axis, MHC class II; x axis, CD11C. B) Increase in the expression of MHC class II on a macrophage population. Left plots: delineation of CD11B⁺CD14⁺ cells (macrophages). Upper plot, PBS-treated mice; lower plot, LPS-treated mice. The plot on the right shows expression of MHC class II in the gated cells. Dotted line, cells from PBS-treated mice; solid line, cells from LPS-treated mice.

Adoptive Transfer of Lymphoid Cell Populations from Normal Mice Reduces the Rate of PTD in Rag1^−/− Mice Injected with LPS

Our studies suggest that Rag1^−/− mice contain the cells and functioning cellular circuits needed to undergo LPS-induced PTD, and that these circuits may involve antigen-presenting cells. In addition, NK cells are likely critically involved in this response [21]. The absence of T and B cells in Rag1^−/− mice in this context raises the issue of whether these cells participate in regulatory circuits that protect against PTD. To begin to answer this question, we sought to compare Rag1^−/− mice to normal B6/J mice.

We found that Rag1^−/− and normal mice had similar numbers of grossly normal-appearing fetal-placental units and individual litter resorption (early loss) rates (data not shown). We also found that higher doses (30 μg) of LPS produce similar proportions of mothers that deliver at least one pup in Rag1^−/− and normal mice (nearly all mice; data not shown). However, when the dose of LPS given was decreased, Rag1^−/− mice were found to be more sensitive than normal animals. As shown in Figure 4A, 77% of Rag1^−/− (left bar) mice given 3 μg of LPS deliver within 24 h, as opposed to only 7% (right bar) of normal mice (Rag1^−/−, n = 26; B6/J, n = 15; P < 0.0001).

FIG. 4.

Rag1^−/− mice are more sensitive to LPS-induced premature delivery, and this is partially reversed by injection of normal whole lymphocytes or CD4⁺ T cells. A) Pregnant Rag1^−/− mice (white bars) or normal C57BL/6 mice (black bars) were injected with 3 μg of LPS and observed for delivery of pups 24 h later. B) Rag1^−/−mice were pretreated on Day 10 of gestation with either whole spleen and lymphocytes or purified CD4⁺ T cells from normal mice. On Day 15 of gestation, they were injected with LPS and observed for delivery of pups 24 h later (Day 16 of gestation). Y axis, proportion of mothers that delivered at least one pup prematurely; x axis, pretreatment and number of mice in the groups tested. P values listed are a comparison between: untreated Rag1^−/− mice and normal B6/J mice; untreated Rag1^−/− mice and Rag1^−/− mice pretreated with whole lymphocytes; or untreated Rag1^−/− mice and Rag1^−/− mice pretreated with CD4⁺ T cells alone (chi-square analysis). Significance was also obtained for each comparison if Fischer exact testing was used (P < 0.0001, P = 0.03, and P = 0.04, respectively).

The sensitivity of Rag1^−/−mice to a low dose of LPS might be due to inherent differences in nonlymphoid tissues (stromal, epithelial, or other cells), but may also be due, in part, to differences in lymphoid cells. To test this hypothesis, we adoptively transferred 10⁷ total lymphocytes from spleen and lymph node cells of B6/J wild-type mice into pregnant Rag1^−/− on Day 10 of pregnancy, 5 days before injection with LPS. A comparison of Figure 4A (left bar) and 4B (left bar) shows that, while 77% of Rag1^−/− mice given 3 μg of LPS deliver in 24 h, only 28% of mice given normal lymphocytes prior to LPS injection delivered (P = 0.016). This suggests that lymphoid cells play a role in regulating the response to LPS.

Trivial explanations to account for the protective effect of whole-lymphocyte transfer in Rag1^−/− mice could be the rapid development of antibodies to LPS, or that the transferred cells included B cells inherently productive of neutralizing antibody, which would decrease the effective dose of LPS. This has been postulated as the factor decreasing death rates in normal as compared with Rag1^−/− mice given high doses of LPS [40]. However, the development of significant titers of anti-LPS antibodies is less likely to occur in the short time frame of these experiments. Still, this raises the question of whether T cells alone could modify the effect of LPS on PTD. Within the adaptive immune response, CD4⁺ T cells are key players in the initiation and regulation of immune responses. There is increasing evidence that subsets of CD4⁺ T cells are important negative regulators of adaptive and innate immune responses [23, 26]. These regulatory T cells are not only important in autoimmune disease, cancer, or transplantation, but they may play a role in fetomaternal tolerance and recurrent early abortion as well [23, 27, 41]. We thus decided to focus on CD4⁺ T cells and their capacity to modify the response to LPS in our PTD model.

Rag1^−/− mice were pretreated with 2 × 10⁶ sorted CD4⁺ T cells isolated from B6/J or B6-GFP mice on Day 10 of pregnancy. On Day 15 of gestation, they were injected with 3 μg of LPS. Within 24 h (Fig. 4B, right bar), only 40% of mice delivered—a significant reduction compared with untreated Rag1^−/− mice (Fig. 4A, left bar; P = 0.018). This suggests that CD4⁺ T cells down-regulate LPS-induced PTD.

Diversity in the Population of CD4⁺ Lymphocytes Found in the Uterus and Uterine Draining Lymph Nodes of Pregnant Rag1^−/− Mice Injected with CD4⁺ T Cells prior to LPS Injection

The decreased LPS-induced PTD found in Rag1^−/− mice pretreated (before LPS) with CD4⁺ T cells could be due to systemic effects of soluble mediators elaborated by the CD4⁺ T cells. However, it is more likely that the CD4⁺ T cells exerted their effects in the uterus or in the tissue-draining lymph nodes through direct interaction with cells such as NK cells [21].

We confirmed the presence of CD4⁺ T cells in pregnant Rag1^−/− mice injected with these cells prior to LPS injection by isolating uterus and uterine draining nodes, generating single-cell suspensions, and examining the existing lymphocytes by flow cytometry. Figure 5 shows an example of the uteri examined as part of these studies. In the uteri of mice pretreated with CD4⁺ T cells 5 days prior to LPS injection, we found a small population of CD4⁺ cells in the lymphocyte gate (Fig. 5, upper panels). Within the CD4⁺ population, we also observed a discreet subset that was also positive for the T-cell receptor (Fig. 5, lower panels). In addition, we found that uteri from mice that delivered despite pretreatment with CD4⁺ T cells appeared to contain fewer CD4⁺ T cells as a proportion of the lymphocytes present (Fig. 5, right versus left panels). Thus, CD4⁺ T cells in the uterus negatively regulate the response to LPS in the pregnant uterus.

FIG. 5.

Presence of CD4⁺ T cells in the uterus of Rag1^−/− mice given normal CD4⁺ cells. Pregnant female Rag1^−/− mice were injected with ∼2 million B6 CD4⁺ T cells on Day 10 of gestation and then injected with 3 μg LPS on Gestational Day 15. Shown are representative examples from mice that either did not (left, upper and lower panels) or did (right, upper and lower panels) subsequently deliver 24 h later. Upper plots: the percentage of lymphocytes that were CD4⁺. Lower plots: the percentage of CD4⁺ cells that were TCR positive.

One accepted indicator of CD4⁺ T cells that can regulate immune responses, either by releasing immune suppressive cytokines or by modifying the differentiation of T cells, is the expression of the forkhead family transcription factor, FOXP3⁺ [42]. To determine if pretreatment of Rag1^−/− mice with CD4⁺ T cells 5 days before LPS allowed for the localization of CD4⁺ T cells expressing FOXP3, uterine draining lymph nodes were examined by flow cytometry for expression of the protein. An example of the data is shown in Figure 6. Normal B6/J pregnant mice injected with 3 μg LPS (top panels) did not deliver prematurely and were used as a control. Approximately 15% (the median in the mice tested) of cells from uterine draining lymph nodes from normal B6/J mice were CD4⁺ TCR⁺, and, of these, approximately 14% (the median in the mice tested) were FOXP3⁺. By comparison, the uterine draining nodes of pregnant Rag1^−/− mice pretreated with CD4⁺ T cells 5 days prior to LPS, and that did not deliver prematurely (lower panels), also developed a significant, but smaller, population of FOXP3⁺ CD4⁺ T cells (e.g., 7.4%). This suggests that FOXP3⁺ regulatory T cell subsets may contribute to the pool of CD4⁺ T cells that influence the LPS response in this system.

FIG. 6.

CD4⁺TCR⁺ FOXP3⁺ cells in the uterine draining nodes of pregnant mice given LPS. Normal mice (upper panels) were mated and given 3 μg LPS on Day 15 of gestation. After 24 h, the mice were euthanized and found to have not undergone PTD. Nodes draining the uterus were isolated and stained extracellularly for CD4 and T cell receptor, followed by intracellular staining for the molecule FOXP3. Rag1^−/− mice (lower panels) were given CD4⁺ T cells on Day 10 of gestation, and then 3 μg of LPS on Day 15. After 24 h, their tissues were isolated for study. Left and right panels: examples of the data obtained. Left panels show percentage of CD4⁺ T cells in uterine draining node lymphocytes. Right panels show percentage of CD4⁺ T cells that were found to be positive for FOXP3.

DISCUSSION

Although there exist several proposed mechanisms to explain PTD, and several soluble and membrane-bound mediators are thought to play significant roles, it is not yet clear which cells or cellular circuits are critical to the initiation, completion, or regulation of this process.

By examining Rag1^−/− mice, we first tried to understand the critical importance of T and B cells. In our study, Rag1^−/− mice delivered 24 h after LPS injection, which implies that the adaptive immune system is not critical for LPS-induced PTD. We also observed an increase in proinflammatory cytokines in the serum after injection of LPS, which is consistent with the findings by other groups studying LPS-induced PTD in normal mice [13, 43]. The inflammatory cytokine response that we observed in Rag1^−/− mice included an increase in mediators that activate and attract neutrophils and macrophages, and these, along with NK [21] cells, would be expected to be the primary cellular positive regulators of LPS-induced PTD in Rag1^−/− mice. After injection of LPS, we also observed an increase of RNA expression of key molecules, such as MyD88 and NFkappaB, thought to be critical downstream signaling molecules in the pathway linking TLR4 and PTD [43]. We also observed an increase in PTGS2 (COX2) expression, possibly by uterine-resident immune cells, which can support local levels of prostaglandins and, in turn, uterine contractions. With the caveat that expression at the protein level may be different from that at the RNA level, pregnant Rag1^−/− mice exhibit activation of the same pathway(s) in response to LPS as do mice with working adaptive immune systems.

These data, along with those of others [21], underscore the importance of the innate immune response in LPS-induced PTD, and in other models of pregnancy loss [22, 44].

Although the increase in sensitivity to low doses of LPS that we observed in Rag1^−/− mice may relate to inherent reproductive dysregulation in the absence of T and B cells, we did not observe evidence to support this, as, for example, our Rag1^−/− mice did not experience a higher rate of early loss (data not shown) compared to normal mice. To explain this difference in sensitivity, we instead hypothesized that T and B cells would serve a regulatory role in this process. This hypothesis was supported by the results obtained after transfer of whole-lymphocyte preparations from normal into Rag1^−/− mice. In light of current interest in the role of regulatory T cells in immune responses, we examined the effect of reconstitution of Rag1^−/− mice with cells significantly enriched (>90% pure) in CD4⁺ T cells. While preinjection with this cellular preparation suggests a regulatory role for CD4⁺ T cells, it does not rule out a role for “contaminating” accessory cells or B-1 B cells [40] in regulation of this response. Moreover, these studies do not shed light on whether this regulation is on the basis of soluble mediators (e.g., cytokines), or direct cell-cell contact. Finally, these experiments do not rule out a role for CD8-expressing T cells with regulatory function.

Even if pure, CD4⁺ T cells comprise a heterogeneous population of naïve, effector, and memory cells, as well as cells with true modulatory or down-regulatory function, and this may be the reason why the LPS response in CD4⁺ T cell-reconstituted Rag1^−/− mice was not completely suppressed. However, it is also clear that “regulatory T cells” also denotes a heterogeneous family of cells, potentially with differing phenotype and functional capacity. A more prominently considered subgroup of regulatory T cells expresses FOXP3 [42], and the role of this cell type in immune regulation is under intense investigation in vivo in manipulated and intact animals. Earlier studies delineated regulatory T cells as CD4⁺ CD25⁺, and have suggested a regulatory role for cells carrying this phenotype in models of recurrent pregnancy loss [27] and in normal allogeneic pregnancy [25]. Current thinking, moreover, is that the majority of CD4⁺ CD25⁺ regulatory T cells also express FOXP3. Our finding of CD4⁺ CD25⁺ FOXP3⁺ cells in Rag1^−/−mice relatively resistant to LPS-induced PTD after pretreatment with CD4 T cells suggests that so-defined regulatory T cells may be a component of the existing protective cellular circuit. It is also formally possible that, in our model, as has been described in other systems [45], this protective circuit also comprises regulatory T cells that posses key regulatory functions, but do not fit the currently accepted definitions. Finally, in this model, the idea of regulatory T cells, not as a separate lineage, but more broadly as a subset expressing cytokines (e.g., IL10) that oppose PTD, needs to be considered.

PTD represents a complex biological process that is currently difficult to accurately diagnose until its final stages, involving active labor and imminent delivery, become evident. However, as our understanding of the critical regulators and intermediate steps involved in PTD increases, potential targets for both diagnostic tests and therapeutic intervention will become apparent. The studies in this report suggest that focus on cellular components of the innate immune response activated during LPS-induced PTD may identify new diagnostic tools. Moreover, focus on related regulatory circuit(s) involving T cells may support development of novel interventions.