| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
Institute of Immunology and Transfusion Medicine,3 University of Lübeck, 23538 Lübeck, Germany
Laboratory of Immunology and Vascular Biology,4 Department of Pathology, Stanford University School of Medicine, Stanford, California 94305
ABSTRACT
Leukocyte recruitment to the pregnant mouse uterus is associated with highly regulated patterns of expression of vascular adhesion receptors. One striking observation is the localized expression of mucosal vascular addressin cell adhesion molecule (MADCAM1) and selectin, platelet (SELP, formerly P-selectin) by maternal vessels in the vascular zone (VZ) during the first half of pregnancy. From midgestation onwards, endothelial cells lining the maternal vessels of the VZ in addition express vascular cell adhesion molecule-1 (VCAM1). The predominant cell population within these vessels is monocyte-like cells. Granulocytes and low numbers of lymphocytes are also present. Murine fetal trophoblast cells are almost devoid of adhesion molecules, including SELP. In contrast, spontaneous abortions of allogeneic pregnancies are characterized by dramatic upregulation of SELP on maternal VZ vessels and on fetal trophoblast cells. Upregulation of SELP is associated with a dramatic influx of highly activated granulocytes, which infiltrate the vessels and tissue of the VZ and the trophoblast. The majority of the activated granulocytes within the trophoblast undergo nuclear fragmentation, which can be detected by TUNEL staining. To demonstrate that SELP is involved in the recruitment of granulocytes to the pregnant uterus, we undertook long-term in vivo inhibition studies using a monoclonal antibody to inhibit the contribution of SELP to leukocyte trafficking to the decidua. In addition, the pregnant uteri of syngeneic Selp–/– x Selp–/– mice were investigated and compared to the controls. Our results clearly demonstrate the importance of SELP for granulocyte access to the pregnant mouse uterus under physiological and pathological conditions.
decidua, female reproductive tract, immunology, trophoblast, uterus
Leukocytes are migratory cells, and highly specialized and regulated mechanisms exist to control their extravasation from the blood into tissues. Specificity for particular organs and tissues is provided by tissue-selective expression of vascular adhesion ligands, the vascular addressins. Tethering and rolling are usually mediated by the selectins and their carbohydrate ligands, but also by low-affinity integrin alpha 4 (ITGA4) [1]. The selectin family consists of three members: lymphocytic selectin (SELL, formerly L-selectin), selectin, platelet (SELP, formerly P-selectin), and endothelial selectin (SELE, formerly E-selectin). SELL is expressed on all leukocytes, whereas SELP and SELE are expressed on endothelial cells (EC) and SELP is also expressed on activated platelets. SELP is the only endothelial cellular adhesion molecule that is stored intracellularly in Weibel Palade bodies. The appearance of SELP on the cell surface can be induced within minutes by inflammatory agents, such as histamine and thrombin. SELP expressed on adherent platelets as well as on endothelium has been shown to support leukocyte attachment and rolling in in vitro flow assays [2–4]. To stop rolling, leukocytes must engage secondary adhesion molecules, which all belong to the integrin family. Integrins involved in leukocyte-endothelial interactions include alpha 4 integrins (ITGA4/ITGB1, formerly 
4ß1-integrin and ITGA4/ITGB7, formerly 4ß7-integrin) and beta 2 integrins (integrin alpha L [ITGAL], formerly LFA-1, and integrin alpha M [ITGAM], formerly Mac-1 or CD11b), which interact with members of the immunoglobulin-family of adhesion molecules, mucosal vascular addressin cell adhesion molecule 1 (MADCAM1), vascular cell adhesion molecule 1 (VCAM1), and intercellular adhesion molecules (ICAM). In addition, trafficking is regulated by tissue-selective leukocyte-activating and -chemoattractive factors, as well as by local tissue insults and cytokine production [1, 5].
In many instances, the specificity of leukocyte trafficking is determined as much by the leukocytes that are excluded as by those that are allowed entry. Thus, during normal pregnancy, the maternal/fetal interface represents a site in which leukocyte trafficking needs to be exquisitely regulated. This has been demonstrated in the mouse, in which microdomains of differentially expressed cellular adhesion molecules involved in leukocyte recruitment have been identified at the maternal/fetal interface, especially during the critical period of initial placenta development. Each of these microdomains is functionally correlated with the distinct localization of uterine natural killer (uNK) cells, monocyte-like cells, and granulocytes [6, 7]. Once in the decidua, maternal leukocytes have the opportunity to migrate to and interact with the invasive trophoblast. Leukocyte interaction with the trophoblast may be either deleterious or beneficial, depending upon the circumstances. In normal pregnancy, the specialized leukocytes that are allowed to access the decidua are hypothesized to communicate with each other and with the trophoblast by delivering immunoregulatory signals, exchanging growth or regulatory factors [8, 9].
Murine abortions are characterized by damaged and killed fetal trophoblast. Although the mechanisms by which the implanted fetal tissue is destroyed are not yet fully understood, the altered composition of decidual leukocytes may be crucial to this process. In mammalian pregnancies, it has been demonstrated that the rejection of fetal tissue is associated with enhanced recruitment of activated T cells to the maternal/fetal interface, upregulation of proinflammatory Th1 cytokines (interferon gamma [IFNG] and TNF), dramatic infiltration of polymorphonuclear leukocytes at sites of necrosis and along the walls of large vessels in the decidua and complement activation, reduced expression of thromboregulatory molecules, thrombosis, and hemorrhage [10–17]. Activated NK cells have been associated with early fetal loss [18]. Haddad et al. [19] have suggested a role for nitric oxide as an effector molecule in mediating early embryo loss and shown that the in situ activation of decidual macrophages is an early event that precedes spontaneous abortion.
The extravasation of leukocytes from the blood to the tissue of the pregnant uterus needs to be a well-controlled process, to ensure the recruitment of the appropriate cells to the correct location at the right time. Consequently, it is relevant to understand how these mechanisms are modulated during an abnormal pregnancy that results in a dramatically altered composition of decidual leukocytes, which may mediate the lethal effect on the trophoblast.
BALB/c and C57BL/6J mice were obtained from the Charles River Breeding Laboratories (Wilmington, MA). Selp-null (C57BL/6;129S-Selptm1Hyn) [20] mice were obtained from the Jackson Laboratory (Bar Harbor, ME). The mice were free of common viral and bacterial pathogens according to the results of routine screening procedures performed by the supplier. The pregnant mice in individual cages and the nonpregnant mice in common cages were given food and water ad libitum and were kept in an environmentally controlled room. BALB/c females were mated by overnight cohabitation with C57BL/6J males, as were the Selp–/– and C57BL/6J females with syngeneic males. Evidence of mating was determined by vaginal plug appearance on the following day, which was considered as Day 1 of pregnancy. To confirm that the mated females were pregnant, vaginal swabs were carried out and investigated each day for bleeding from Gestational Day 7 onwards. Pregnant BALB/c mice (BALB/c x C57BL/6J mating) with blood in their vaginal smears were killed and the abortive uteri were collected immediately. Abortive uteri from Gestational Days 11, 13, and 14 were evaluated (one or two abortive implantation sites in each of three animals per day of pregnancy). The abortion sites were identified by their small size in combination with a necrotic, hemorrhagic appearance compared to normal implantation sites. Uteri from normal pregnancies were collected on Days 9, 11, 13, and 14 of gestation (three animals per day of pregnancy and two implantation sites in each animal). For long-term in vivo inhibition studies, two implantation sites in each of three animals per test group were evaluated. All procedures involving animals were approved by the German Animal Care and Use Committee (no. 29/A3/97) and were conducted in accordance with the Guide for Care and Use of Agricultural Animals in Teaching and Research.
The primary rat anti-mouse mAbs (purified, directly PE- or FITC-conjugated) used for the histological stainings included: anti-THY1 mAb 30H.12; anti-CD4 mAb GK1.5; anti-CD8 mAb Lyt-2; anti-protein tyrosine phosphatase receptor type C (PTPRC) mAb 30G.12 (purchased as anti-CD45 mAb); anti-ITGAM mAb M1/70 (purchased as anti-CD11b mAb); anti-ITGA4 mAb PS/2 (purchased as anti-4-integrin mAb); anti-VCAM1 mAb MK2.7; anti-MADCAM1 mAb MECA 367 (ATCC, Rockville, MD; BD PharMingen, Heidelberg, Germany); anti-ICAM1 mAb YNI/1.7 (Dr. Takei, Karolinska Hospital, Stockholm) [21]; anti-ICAM2 mAb 3C4; anti-Gr-1 mAb RB6-8C5; anti-LY6G mAb 1A8 (BD PharMingen); anti-ITGB7 mAbs (purchased as anti-ß7-integrin mAb) FIB 21 [22] and M293 (BD PharMingen); anti-cytokeratin mAb K8 (TROMA I) (from the National Institute of Child Health and Human Development, a gift from M.B. Omary) [23]; anti-SELE mAb 10E9.6 and anti-SELP mAb RB40.34 (a gift from D. Vestweber, Max-Planck-Institute, Münster, Germany and PharMingen, San Diego, CA) [24, 25]. Rat anti-human CD44 mAb Hermes-1 (ATCC) [26] was substituted for the primary antibody as a control for nonspecific staining. For two-color staining, rat anti-mouse cytokeratin and Gr-1 mAbs were conjugated with FITC (Sigma, St. Louis, MO). PE-conjugated, affinity-purified mouse anti-rat IgG Fab2 polyclonal (Chromaprobe, Mountain View, CA) or biotin-spacer (SP)-conjugated affinity-purified donkey anti-rat IgG (H+L) (Jackson Immunoresearch Laboratories, West Grove, PA) was used as the second step. All antibodies were titered and used at saturation for staining.
For long-term inhibition studies, rat anti-MADCAM1 mAb MECA 367, rat anti-VCAM1 mAb MK2.7, rat anti-SELP RB40.34, and rat anti-human CD44 mAb Hermes-1 were used. MECA 367 binds to the first domain of MADCAM1 and blocks MADCAM1-dependent binding in vitro and lymphocyte homing to Peyer patches HEV in vivo [27]. MK.2.7 blocks the VCAM1-mediated binding function [28]. The RB40.34 antibody reacts with mouse SELP and blocks mouse SELP binding to its ligands both in vitro and in vivo [25]. Rat anti-human CD44 mAb Hermes-1 was used as a rat IgG2a negative isotype control mAb. Hermes-1 recognizes the N-terminal binding domain of CD44 and incompletely inhibits hyaluronate binding [26].
Pregnant uteri from normal and abortive pregnancies were collected and rapidly frozen in Jung Tissue Freezing Medium (Leica, Bensheim, Germany). Cryostat sections (6 µm) were air-dried for 1 h and fixed in acetone. To assist with orientation and selection and for examining tissue structure, every tenth tissue section was stained with hematoxylin and eosin. To investigate trophoblast structure, the trophoblast was marked and identified with anti-mouse cytokeratin mAb. In contrast to the normal implantation sites, the abortion sites were characterized by a smaller size accompanied by a necrotic, hemorrhagic appearance, disorganization, and degeneration of the trophoblast, as well as dramatically increased leukocyte infiltration.
For immunofluorescence staining, tissue sections were incubated with 10% normal mouse serum in PBS for 15 min and then exposed for 30 min to either purified or directly PE-conjugated individual primary mAbs. Samples were washed with PBS for 5 min. Tissue sections stained with purified primary mAb were then incubated for 30 min with PE-conjugated secondary mouse anti-rat IgG (50 µg/ml) and rinsed in PBS. For two-color staining, tissue sections were then incubated with 10% normal rat serum in PBS for 15 min and exposed for 30 min to FITC-conjugated rat anti-mouse cytokeratin, Gr-1, LY6G, CD4, CD8 or THY1 mAb. Tissue samples from the test groups treated with blocking rat-anti-mouse mAbs were stained with directly PE- and/or FITC-conjugated primary rat anti-mouse mAbs. Rat anti-human CD44 mAb Hermes-1 was substituted for the primary antibodies as controls for nonspecific staining. The staining patterns of the antibodies were tested on at least ten different tissue samples. Sections were analyzed on the day of staining. Samples were examined with a DMLM microscope (Leica) equipped with the appropriate filters. The images were processed and overlaid (Photoshop; Adobe, Mountain View, CA), as appropriate for presentation in the figures.
For immunoperoxidase staining, acetone-fixed frozen sections were exposed for 60 min to individual monoclonal antibodies diluted in PBS. Samples were washed with PBS for 5 min and then exposed to aqueous 3% H2O2 for 10 min, to quench endogenous peroxidase activity. After washing the slides in PBS, the sections were incubated for 60 min with the biotinylated secondary donkey anti-rat IgG (50 µg/ml) and rinsed in PBS. To increase sensitivity, the sections were then incubated with ABC reagent (preformed avidin and biotinylated horseradish peroxidase macromolecular complex) (Vector Laboratories, Burlingame, CA) for 30 min, washed with PBS for 5 min and then treated with diaminobenzidine/H2O2 solution (Vector Laboratories) for 2–10 min. Sections were lightly counterstained with methylene blue.
To detect apoptotic cell death at the single-cell level in the tissue of the pregnant uterus, the In Situ Cell Death Detection Kit Fluorescein, which is based on the TUNEL method, was used in accordance with the protocol of the manufacturer (Roche Diagnostics, Mannheim, Germany). After incubation with the TUNEL reaction mixture (enzyme solution and labeling solution) for 60 min at 37°C, the tissues were counterstained with PE-conjugated anti-ITGAM mAb M1/70, anti-Gr-1 mAb RB6-8C5 or anti-cytokeratin mAb Troma I and analyzed by fluorescence microscopy (DMLM microscope; Leica).
To investigate the contributions of vascular SELP, MADCAM1, and VCAM1 to granulocyte recruitment to the pregnant uterus, 500 µg of blocking anti-SELP mAb RB40.34, anti-MADCAM1 mAb MECA 367 or anti-VCAM1 mAb MK2.7 was injected i.p. daily from Day 8 through Day 10 or Day 13 of gestation. Anti-human CD44 mAb Hermes-1 (500 µg per day, injected from Day 8 through Day 10 or Day 13 of gestation) was used as a class-matched rat IgG2a negative control. Saturating serum levels were assessed by periodic analysis of serum for the presence of free mAb. The uteri were removed on Day 9, 11, or 14 of pregnancy and sectioned. The effects of blockade on the recruitment, accumulation, and phenotype of leukocytes in the pregnant uteri were assessed immunohistologically at Gestational Days 11 and 14.
Enumeration of Leukocyte Subsets
To enumerate leukocyte subsets, an eyepiece micrometer grid of 1 mm2 (equivalent to 0.25 mm2 at 200x magnification) was used. For each day of pregnancy three animals, for each animal two independent implantation sites (in abortive uteri, one or two implantation sites), and per implantation site ten different tissue samples were assessed. The number of micrometer grids counted per tissue sample ranged from 4 to 6. To avoid duplicate cell counting, sections separated by not less than 42 µm were chosen.
The results are expressed as the median and range. After evaluating the variance of data within each test group using the Kruskal-Wallis test, differences between a test group and the corresponding controls (taking into consideration normal and abortive uteri, day of pregnancy, allogeneic or syngeneic pregnancy, different leukocyte subpopulations, genetic background, and the isotype of the injected antibody) were evaluated with the unpaired Mann-Whitney U-test. In each test, P < 0.05 was considered to be statistically significant.
In the present study, we investigated the expression of vascular addressins and the phenotypes of the major infiltrating leukocyte types at the maternal/fetal interface in allogeneic BALB/c x C57BL/6J mice with spontaneous abortions at Gestational Days 11, 13, and 14 in comparison to normal stage-matched controls. The alterations in the expression of homing-associated molecules at the maternal/fetal interface during the course of normal pregnancy have been described recently [7, 29]. These results are briefly summarized to evaluate the variations in patterns of vascular homing receptors, microenvironmental relationships, and infiltration of specialized leukocyte subsets in spontaneous abortions at Days 11, 13, and 14 of pregnancy.
A schematic overview of the differences between the normal pregnant uterus and the abortive uterus, taking into consideration the locations and phenotypes of the different leukocyte types and expression patterns of the vascular addressins, is shown in Figure 1.
|
Expression of Homing-Associated Molecules at the Maternal/Fetal Interface from Day 11 Through Day 14 of Normal Allogeneic BALB/c x C57BL/6J Pregnancies
At midgestation (Gestational Days 11 and 12), the maternal blood vessels in the vascular zone (VZ) displayed combined expression of MADCAM1, SELP, ICAM1 (low expression), and ICAM2 (high expression). In addition, the vessels stained weakly with anti-VCAM1 mAb. The major infiltrating leukocyte subsets in the vessels of the VZ were monocyte-like cells. In addition, significant numbers of granulocytes, some CD4+ and CD8+ T cells and single uNK cells, were seen in the lumen. Within the tissue of the VZ, the predominant populations were uNK cells and monocyte-like cells. Low numbers of granulocytes were also present. The vessels of the central decidua basalis expressed high levels of VCAM1 and ICAM2 and low levels of ICAM1 but showed no expression of SELE, SELP or MADCAM1. These vessels were surrounded by uNK cells and contained rare scattered T cells and granulocytes (data not shown). On Days 13 and 14 of pregnancy, SELE expression peaked in trophoblast cells, which line the maternal blood spaces of the outer spongy zone. The dilated maternal blood vessels were characterized by reduced MADCAM1 expression and significantly higher expression of SELP and VCAM1. Figure 2 illustrates the reactivities of vascular MADCAM1 and SELP with the appropriate mAbs (Fig. 2, A and B). These vessels contain increased numbers of monocyte-like cells (Fig. 2C), granulocytes (Fig. 2D, Table 1) and T cells (data not shown). The invasive epithelial trophoblast was remarkably devoid of immunohistologically detectable adhesion molecules, including ICAM1, VCAM1, MADCAM1, and SELP (Fig. 2E, Fig. 3A, and data not shown). SELE, and from midgestation onwards, ICAM2 were detected on the outer trophoblast region at the border with the decidua basalis. Almost no leukocytes were observed in multiple sections within the trophoblast (Figs. 2G and 3B). Only the trophoblast-lined maternal blood spaces in the outer trophoblast region contained granulocytes, and from midgestation onwards, monocytes and scattered T cells (Fig. 2, F and G, Fig. 3B, and data not shown).
|
|
|
Upregulation of SELP in Abortive Mouse Uteri on Days 11, 13, and 14
Murine abortions are characterized by dramatic disorganization and degeneration of the trophoblast. The trophoblast was marked and identified with anti-mouse cytokeratin mAb. Spontaneous abortions of allogeneic BALB/c x C57BL/6J pregnancies were analyzed at Gestational Days 11, 13, and 14 and compared to those of normal stage-matched allogeneic BALB/c x C57BL/6J pregnancies. Immunohistological staining was performed using mAbs against the vascular addressins MADCAM1, VCAM1, ICAM1, ICAM2, SELE, and SELP. The distribution patterns of most of the vascular addressins (MADCAM1, VCAM1, ICAM1, ICAM2, and SELE), as well as the strength of their expression in all the abortive uteri investigated, were comparable to those of normal stage-matched pregnancies (Fig. 2H, and data not shown). Negative control staining with anti-human CD44 mAb Hermes-1 (isotype-matched) confirmed that the staining was specific (Fig. 2O). Interestingly, the expression of SELP was dramatically upregulated on the EC lining the maternal vessels of the VZ during fetal abortion (shown for a Day 13 abortive uterus; Fig. 2I). In addition, these vessels were characterized by strong infiltration of platelets. Surprisingly, the trophoblast, which in normal pregnancies fails to stain with mAbs against SELP, showed extremely high expression of SELP, mainly at sites of degeneration (shown for Day 13 and Day 14 abortive uteri; Figs. 2M and 3D). The expression of SELE and ICAM2 in the outer region of the trophoblast at the border with the decidua basalis was comparable to that of normal pregnancies (data not shown). The strong upregulation of SELP by maternal vessels of the VZ and by the fetal trophoblast in all the abortive uteri investigated was associated with a dramatic recruitment of myeloid cells, which displayed the granulocyte marker Gr-1hi, as well as high expression of ITGAM (ITGAMhi) (Fig. 2, J–L, Fig. 3, E and F). These ITGAMhi Gr-1hi-expressing leukocytes had infiltrated the lumina of the vessels of the VZ (Fig. 2, J–L) and could also be seen within the trophoblast at sites of degeneration associated with strong expression of SELP (Fig. 2, M and N, Fig. 3, D–F). In addition- CD4+ and CD8+ T cells were seen within the trophoblast, albeit in low numbers (Fig. 3, G and H).
Characterization of ITGAMhi Gr-1hi-Expressing Leukocytes
At the maternal/fetal interface of normal pregnancies, the monocytes are ITGAMhi Gr-1lo, whereas the granulocytes display intense reactivity with anti-Gr-1 mAb RB6-8C5 and only low reactivity with anti-ITGAM mAb M1/70 (ITGAMlo Gr-1hi). However, ITGAM expression is rapidly upregulated on granulocytes after activation [30]. The RB6-8C5 antibody reacts with a common epitope on LY6G and LY6C, also known as the myeloid differentiation antigen Gr-1, and recognizes both granulocytes (LY6G+ LY6C–) and monocytes (LY6G– LY6C+) [31, 32]. To determine whether the ITGAMhi Gr-1hi-expressing myeloid cells found in the VZ and at sites of trophoblast degeneration in murine abortions represent activated granulocytes, and to distinguish these cells from monocyte-like cells, additional stainings were performed using the monoclonal antibody 1A8 (Fig. 2, K and L). 1A8 reacts with LY6G but not LY6C and recognizes granulocytes in the periphery [31]. The numbers of infiltrating myeloid cells, consisting of both ITGAMhi LY6G–- and ITGAMhi LY6G+-expressing leukocytes (ITGAMhi cells), the numbers of monocyte-like cells (ITGAMhi LY6G– cells), and the numbers of activated granulocytes (ITGAMhi LY6G+ cells) were determined in the vessels and tissue of the VZ of abortive uteri at Days 11, 13, and 14 of pregnancy and compared to the numbers of infiltrating myeloid cells (ITGAMhi LY6G– cells and ITGAMlo LY6G+ cells), monocyte-like cells (ITGAMhi LY6G–), and granulocytes (ITGAMlo LY6G+), respectively, from the stage-matched normal allogeneic BALB/c x C57BL/6J pregnancies (Table 1). The vessels and tissue of the VZ in all the abortive uteri investigated showed dramatically increased numbers of myeloid cells (ITGAMhi) compared to the numbers of myeloid cells (ITGAMhi LY6G– cells and ITGAMlo LY6G+ cells) in stage-matched normal pregnancies (Table 1). In contrast to normal pregnancies, the great majority of these cells appeared to be activated granulocytes (ITGAMhi LY6G+; 70–80% of the infiltrating leukocytes) (Fig. 2, K and L), whereas the numbers of monocyte-like cells (ITGAMhi LY6G–) were decreased or comparable to those of normal pregnancies (Table 1).
Involvement of SELP in the Recruitment of Granulocytes and Monocyte-Like Cells to the Pregnant Uterus
The studies described above suggest a critical role for SELP in the dramatic recruitment of activated granulocytes to the vessels and tissue of the VZ of abortive uteri. To confirm the importance of this vascular addressin for the recruitment of granulocytes, we performed in vivo homing studies in normal pregnancies using a mAb that inhibits the contribution of SELP. For this purpose, the blocking anti-SELP (RB40.34) mAb was injected i.p. into normal allogeneic BALB/c x C57BL/6J pregnancies daily from Day 8 through Day 10 or from Day 8 through Day 13 of gestation, at a concentration of 500 µg per day. At Day 11 or Day 14 of gestation, the effects of mAb blockade on the contents of the granulocytes and monocyte-like cells were compared to the implantation sites of stage-matched allogeneic BALB/c x C57BL/6J controls that were either untreated or treated with the isotype-matched anti-human CD44 mAb Hermes-1. To assess the importance of vascular MADCAM1 or VCAM1 in granulocyte trafficking to the deciduas, blocking mAbs against MADCAM1 (MECA 367) or against VCAM1 (MK2.7) were injected i.p. from Day 8 through Day 10 or from Day 8 through Day 13 of gestation. In addition, pregnant uteri of syngeneic Selp–/– x Selp–/– pregnancies were investigated. To exclude potential differences between the mouse strains investigated, syngeneic pregnancies in Selp–/– mice were compared to syngeneic untreated wild-type pregnancies on the same genetic background. The uteri were removed at Day 9, 11 or 14 of pregnancy and sectioned. The effect of blockade on the recruitment of leukocyte subsets in the pregnant uteri was assessed immunohistologically. The numbers of granulocytes (ITGAMlo LY6G+) and monocyte-like cells (ITGAMhi LY6G–) were determined in the vessels and tissue of the VZ, as described above (Table 2).
|
On Days 9 and 11 of the allogeneic and syngeneic pregnancies, the MADCAM1+ SELP+ maternal vessels and tissue of the VZ contained significant numbers of monocyte-like cells. Only 2–25% of the infiltrating cells were granulocytes. At Gestational Day 14, the frequency of both cell types increased dramatically in the allogeneic pregnancies. In the syngeneic pregnancies, the number of monocyte-like cells decreased, whereas the frequency of granulocytes was comparable to that on Day 11 of pregnancy. At Day 11 or 14 of pregnancy, no significant differences in cell numbers were seen between the untreated allogeneic controls and the mice treated with the isotype-matched anti-human CD44 mAb Hermes-1 (Table 2).
The antibodies directed against SELP (RB40.34) blocked dramatically the recruitment of granulocytes, with only rare scattered granulocytes being detected in the lumina of the vessels or in the tissue of the VZ in the allogeneic pregnancies (about one cell per 0.25 mm2 on Day 11, and five cells per 0.25 mm2 on Day 14). Immunohistological analysis of the implantation sites in RB40.34-treated mice also revealed limited access of monocyte-like cells to the maternal/fetal interface on Days 11 and 14 of pregnancy. Compared to syngeneic pregnancies in the same genetic background, the implantation sites of the Selp–/– mice showed almost completely blocked recruitment of granulocytes but also limited access of monocyte-like cells on Days 11 and 14 (Table 2). Injection of blocking mAbs against MADCAM1 (MECA 367) or against VCAM1 (MK2.7) into allogeneic pregnancies did not affect the recruitment of granulocytes at Day 11 or 14 of pregnancy in comparison to untreated or Hermes-1-treated controls. The number of monocyte-like cells within the vessels and tissue of the VZ of mice treated with MECA 367 was significantly reduced at Day 11 and Day 14 of pregnancy. Treatment with anti-VCAM1 mAb MK2.7 did not influence the recruitment of monocyte-like cells on Day 11 but led to a dramatic reduction in the number of these cells on Day 14 of pregnancy in comparison to the controls (Table 2).
Apoptosis in the Degenerative Trophoblast
We also examined the fate of cells that infiltrated the trophoblast in spontaneous abortions in mice at Days 11, 13 and 14 of pregnancy in comparison to normal stage-matched pregnancies. Apoptotic cells were detected by in situ terminal deoxytransferase (TdT)-catalyzed DNA nick end-labeling (TUNEL; fluorescein) and counterstained with PE-conjugated anti-ITGAM mAb M1/70, anti-Gr-1 mAb RB6-8C5 or anti-cytokeratin mAb Troma I (Fig. 3I, and data not shown). Our results revealed extensive apoptosis of the ITGAMhi Gr-1hi myeloid cells that infiltrated the trophoblast in spontaneous abortions but also in trophoblast cells at sites of degeneration (shown for Day 14 of pregnancy; Fig. 3I). The extent of apoptosis varied between different samples. The stroma, as well as the infiltrating leukocytes of the decidua basalis in spontaneous abortions failed to demonstrate any stained nuclei, with the exception of scattered EC (data not shown). In normal stage-matched allogeneic pregnancies, the trophoblast was almost devoid of infiltrating leukocytes (Fig 3, B and C) and only rare scattered apoptotic cells could be detected in the maternal tissue and the trophoblast (Fig. 3C).
Recent papers have revealed a major role for the innate immune system during abortion that leads to local inflammation, local hemorrhaging, and fetal loss [14, 33, 34]. In most cases, these events are associated with complement activation [13, 14], reduced expression of thromboregulatory molecules [16], expression of FGL2 prothrombinase, and dramatic infiltration of activated phagocytes, which are mainly granulocytes [15, 17]. Enhanced recruitment of activated T cells, alterations in T-cell signal transduction molecules, and upregulation of proinflammatory Th1 cytokines, such as TNF and IFNG, at the maternal/fetal interface during fetal loss have also been described [10–12, 35]. Granulocytes, which are commonly seen at abortion sites, have been reported to contribute to the lysis of TNF- and IFNG-activated EC. During the course of a normal pregnancy, the specificity of decidual leukocyte composition is controlled at the level of cell trafficking [29]. The molecular determinants responsible for coordinating the recruitment of leukocytes include the cellular adhesion molecules and members of the chemokine superfamily [1]. Modulation of these mechanisms just before or during abortion may lead to extensive and altered leukocyte infiltration, which in turn may be involved in the destruction of the fetal trophoblast. In the present study, we provide preliminary support for this notion, defining vascular mechanisms that may be responsible for the extensive infiltration of leukocytes at sites of trophoblast degeneration.
The expression of homing-associated molecules at the maternal/fetal interface during the course of a normal pregnancy has been described recently [29]. One of the most striking features is the unusual coexpression of MADCAM1 and SELP on dilated maternal vessels of the VZ during the first half of pregnancy. The predominant cell population within these vessels, as well as in the surrounding tissue of the VZ, represents cells of the monocyte/macrophage lineage that show high-level expression of ITGAM and low-level expression of Gr-1. Low numbers of granulocytes (ITGAMlo Gr-1hi) are also seen. A significant number of the monocyte-like cells additionally express ITGA4/ITGB7, the ligand for vascular MADCAM1. Long-term in vivo homing studies have revealed that both MADCAM1 and SELP are involved in the recruitment of these cells, whereas ITGA4/ITGB7– monocytes as well as ITGA4/ITGB7+ cells, which do not belong to the myeloid lineage, use SELP and/or other adhesion molecules for initial contact [29]. From midgestation onwards, the EC lining the maternal vessels of the VZ additionally express VCAM1 at low levels. At Day 14 of gestation, dilated maternal vessels show increased expression of SELP and VCAM1, whereas MADCAM1 reactivity has declined. This change in vascular adhesion receptor expression is associated with increased recruitment of granulocytes and T cells [7].
In the present study, vascular adhesion molecule expression in abortive uteri of allogeneic BALB/c x C57BL/6J pregnancies at Days 11, 13, and 14 was compared to that of normal stage-matched allogeneic BALB/c x C57BL/6J pregnancies. Our results reveal that the distribution patterns of most of the vascular addressins, including MADCAM1, VCAM1, ICAM1, ICAM2 and SELE, as well as the strength of their expression in all abortive uteri investigated were comparable to those of normal stage-matched pregnancies. However, the expression of SELP was dramatically upregulated on the EC lining the maternal vessels of the VZ during fetal abortions. In addition, these vessels were characterized by strong infiltration of platelets. Surprisingly, the trophoblast, which in normal pregnancies fails to stain with mAbs against SELP, showed an unexpected high level of expression of SELP, mainly at sites of degeneration. Increased SELP expression has also been described in human decidual tissues during abortion and in murine deciduas in response to stress [10, 33]. Interestingly, recent studies have described the upregulation of FGL2 on trophoblast cells and EC in the murine uterus during abortions [15, 36, 37]. FGL2 has been shown to convert prothrombin to thrombin [36, 37]. Thrombin in turn can trigger the appearance of SELP on the surfaces of EC and platelets within minutes [38, 39]. The strong upregulation of SELP by maternal vessels of the VZ and by the fetal trophoblast in all abortive uteri investigated was associated with dramatic recruitment of activated myeloid cells, characterized by high expression of ITGAM and Gr-1. Additional stainings using the monoclonal antibody 1A8 revealed that the majority of these ITGAMhi Gr-1hi-expressing leukocytes were activated granulocytes, which were characterized by the expression of LY6G [30, 31] and upregulation of ITGAM. The number of recruited monocyte-like cells (ITGAMhi LY6G–) was decreased or comparable to those of normal pregnancies.
These results suggest a critical role for EC- and/or platelet-associated SELP in the dramatic recruitment of activated granulocytes to the vessels and tissues of the VZ of abortive uteri. To assess directly the importance of this vascular addressin in the recruitment of granulocytes, we performed in vivo homing studies in normal pregnancies, using the mAb RB40.34 to inhibit the contribution of SELP [25]. In addition, syngeneic pregnancies of mice deficient in Selp were investigated and compared to syngeneic wild-type pregnancies on the same genetic background. Antibodies against SELP almost completely blocked granulocyte (ITGAMlo LY6G+) access to the maternal/fetal interface. Identical results were obtained for the Selp–/– pregnancies. The numbers of monocyte-like cells were also reduced in allogeneic pregnancies treated with anti-SELP mAb or in syngeneic Selp–/– pregnancies compared to untreated or Hermes-1 treated allogeneic pregnancies or untreated syngeneic pregnancies, respectively. The injection of blocking monoclonal antibodies against MADCAM1 and VCAM1 did not affect the recruitment of granulocytes. The numbers of monocyte-like cells within the vessels and tissues of the VZ of mice treated with MECA 367 were significantly reduced at Day 11 and Day 14 of pregnancy. Treatment with anti-VCAM1 mAb MK2.7 did not influence the recruitment of monocyte-like cells on Day 11 but led to a dramatic reduction in these cells on Day 14 of pregnancy in comparison to the controls. The strong reduction of ITGAMhi LY6G– monocyte-like cells in mice treated with anti-SELP mAb or in Selp–/– mice suggests that monocyte-like cells also use SELP for initial contact, but tethering and rolling may be additionally controlled by other adhesion receptor/ligand interactions, such as ITGA4/ITGB7-MADCAM1 or activated ITGA4/ITGB7-VCAM1, as discussed recently [29].
The results described above clearly demonstrate that EC- and possibly platelet-derived SELP mediate the adhesion of granulocytes to EC. We hypothesize that the dramatic upregulation of SELP on the EC of VZ vessels of abortive units may be involved in the extensive recruitment of activated granulocytes to this site. However, it is not known if SELP can activate granulocyte functional responses, which is an important aspect, since granulocyte activation (including functional upregulation of ITGAL/ITGB2) occurs when they bind to EC that are stimulated by thrombin or histamine to express transiently SELP. However, in the latter situation, SELP is often coexpressed with platelet-activating factor, which is a biologically active lipid that stimulates granulocytes, and thrombin itself is known to activate granulocytes [40, 41].
Activated granulocytes were also seen within the trophoblast at sites of degeneration associated with strong expression of SELP. To determine whether granulocytes contributed to abortions, Clark et al. [36] injected pregnant mice with an anti-granulocyte mAb and found that this antibody partially reduced the spontaneous abortion rate. Girardi et al. [17] have reported that depletion of granulocytes prevents anti-phospholipid Ab-mediated fetal injury, indicating that these cells are essential effectors of tissue damage. Furthermore, local inflammation, hemorrhaging, and fetal loss are often associated with complement activation. The resulting cleavage product, C5a, in turn attracts and activates granulocytes. Interestingly, the majority of activated granulocytes within the degenerative trophoblast (as shown in the present study) undergo nuclear fragmentation. This process may be used by the organism to remove unwanted effector cells and to limit the inflammatory response to the uterus.
The current data suggest that abortions are triggered when innate immune responses or their regulators are perturbed [34]. Modulation of cell trafficking mechanisms may precede abortion and lead to uncontrolled recruitment of activated effector cells. These cells are mainly granulocytes, which appear to be key mediators of fetal injury. These observations demonstrate that controlled immune cell access is an important component of the maternal/fetal interface and is likely to play roles in immune regulation and in protection of the fetus from attack by the maternal immune system.
ACKNOWLEDGMENTS
We thank M. Bishr Omary and Dietmar Vestweber for providing antibodies.
FOOTNOTES
1Supported by grant GRK288C5 from the Deutsche Forschungsgemeinschaft (DFG). 
Correspondence: 2Andrea Kruse, Institute of Immunology and Transfusion Medicine, University of Lübeck, Ratzeburger Allee 160, 23538 Lübeck, Germany. FAX: 49 451 5002857; e-mail: kruse{at}immu.mu-luebeck.de
Received: 7 August 2006.
First decision: 21 August 2006.
Accepted: 1 December 2006.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  J. Behrends, C. M. Karsten, S. Wilke, A. Robke, and A. Kruse Identification of ITGA4/ITGB7 and ITGAE/ITGB7 Expressing Subsets of Decidual Dendritic-Like Cells Within Distinct Microdomains of the Pregnant Mouse Uterus Biol Reprod, October 1, 2008; 79(4): 624 - 632. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
