| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Article |
a Center for Perinatal Biology,
b Departments of Physiology,
c Microbiology and Molecular Biology, and
d Division of Anatomy, Loma Linda University School of Medicine, Loma Linda, California 92350
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
By far, the most common and widely distributed resident immune cell in the uterus is the mononuclear phagocyte. Defined as a macrophage, though some possess monocytic characteristics [5, 9], these cells are abundant in human decidua during the first and third trimesters of pregnancy [10, 11]. In the rodent, uterine macrophages are distributed throughout the pregnant endometrium, as well as in stroma and connective tissue around muscle bundles in the myometrium [6, 12]. On the basis of estimates from cell suspensions of pooled murine uteri at various stages of pregnancy, macrophages may account for nearly 10% of cells in the virgin uterus and comprise upwards of 22% of cells from the pregnant uterus [13, 14]. As confirmed by cell count estimates in tissue sections, uterine macrophage numbers appear to increase as pregnancy progresses [9]. However, a recent study of cell counts during the peripartum period has indicated that the uterine macrophage population is reduced on the day preceding birth by more than 50% from that 4 days before term [12]. This decline in macrophage numbers correlates with reduced production of inducible nitric oxide synthase (iNOS) in the uterus near term [15]. Macrophages are the major source of iNOS, a factor that promotes vasodilatation and uterine smooth muscle relaxation [16, 17].
Although loss of resident macrophages may eliminate a major restraint on uterine contractile activity, myometrial quiescence may also be lost by other means. Macrophages may traffic between endometrial and myometrial compartments, a possibility not investigated in previous studies of total numbers of resident cells [12, 14]. Alternatively, local or select activation of macrophages before the initiation of labor may modulate myometrial contractility. For example, macrophages produce specific inflammatory cytokines and prostaglandins [18–20], concentrations of which are elevated in maternal circulation and amniotic fluid in association with induced preterm labor [1, 21, 22]. These findings raise the possibility that activation of resident macrophages in the peripartum myometrium may contribute to a local immune process that initiates parturition. Therefore, the goal of the present study was to test the hypothesis that enhanced numbers or activation of uterine macrophages within the myometrium precedes parturition. To address this hypothesis, macrophage distribution in the uterus and expression of CD54, an intercellular adhesion molecule (ICAM)-1 marker of macrophage activation [23], were assessed in the pregnant and postpartum uterus during the peripartum period.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Time-dated pregnant C3N/HeN mice were obtained on Day 12 of pregnancy from Harlan Sprague Dawley, Inc. (Indianapolis, IN). Mice were killed by cervical dislocation on the morning of gestational Day 15 or 18, on the day pups were born (postpartum: Day 19 postmating), and 1 day postpartum. Strips of complete uterine tissue (approximately 2 x 5 mm) from the mesometrium between placental discs were obtained as previously described [24]. Tissue was formalin-fixed and embedded in paraffin (n = 4–9/group in duplicate). Uteri from virgin mice (nonpregnant) in the estrous phase of the reproductive cycle served as controls. In nonpregnant mice, the estrous phase of the reproductive cycle was confirmed by the presence of endometrial thickening, glandular development, and columnar epithelium following histological analysis of hematoxylin-stained uterine tissue. The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committees.
Tissue Processing and Macrophage Counts
Paraffin-embedded uterine strips were sectioned at 6 µm, then immunostained by an automated "sandwich" technique with a monoclonal pan-macrophage marker, BM8 (affinity-purified rat IgG2a antibody; Bachem Biosciences, King of Prussia, PA) [25]. As has been previously described [12, 24], this antibody produces dark, high-contrast resolution of uterine macrophages, the distribution and number of which are comparable to those in earlier studies with the F4/80 antibody [6, 14]. Tissue sections processed in the absence of primary or secondary antibody, or with a blocking antibody (Dako, Carpenteria, CA), displayed no appreciable nonspecific stain. Briefly, macrophages were enumerated by systematically counting cells in adjacent visual fields of a tissue section (every one half 10x field) using a 40x lens and a 10 x 10 grid in the ocular reticle; one section from each of the two uterine horns was analyzed for each mouse. These stereological methods preclude duplication of counts, and their implementation has been previously described [12]. Cell counts from tissue sections avoid cell loss and changes in cellular characteristics that may accompany enzyme-digested tissue suspensions; as such, macrophage numbers cannot be directly compared to those in previous reports [12, 14]. A macrophage was defined as a distinct BM8-positive soma (brown) with a hematoxylin-counterstained nucleus (blue); reaction product in pseudopodia and cellular fragments not associated with a nucleus were excluded from cell counts. Tissue boundaries, i.e., endometrium versus myometrium, were determined by histological examination. Endometrial and myometrial macrophage numbers were then calculated from BM8+ cell counts. To compensate for hypertrophy and hyperplasia associated with pregnancy, BM8+ cell counts were normalized to nuclear area as measured in digitized images (Image Pro Plus; Media Cybernetics, Silver Spring, MD). Frozen sections (20 µm) were also processed by immunohistochemistry to visualize stained macrophages for photomicrographs. Staining of splenic macrophages was used as a positive control for BM8 specificity (not shown).
Assessment of Macrophage Activation Using Laser-Scanning Cytometry (LSC)
Paraffin-embedded tissue sections immediately adjacent to those in which BM8-stained cells were enumerated, i.e., subsequent serial sections, were incubated with fluorescein isothiocyanate (FITC)-conjugated anti-CD54 monoclonal antibody (Southern Biotechnology Assoc., Inc., Birmingham, AL). As an activation marker, CD54 (ICAM-1) is expressed by macrophages, epithelium, and vascular endothelium [6, 23, 26, 27]. CD54 was also chosen as an indicator of macrophage activation because it represents a mechanism of cell-to-cell communication, as well as an indicator of cellular activity that is linked to immune surveillance and enhanced antigen presentation [23]. Numbers of positively stained cells, as well as density of expression per cell, are considered markers of activation [28, 29]. Other cell types express CD54 [30–32] but are not typically present in tissue compartments within the uterus. Sections were counterstained with propidium iodine (PI) at a concentration of 50 µg/ml in PBS with 100 µg/ml ribonuclease type A (Sigma Chemical, St. Louis, MO). PI is a DNA-intercalating fluorescent stain that was used to identify nuclei from CD54-positive cells for macrophage quantification and cell cycle determination by LSC [33]. Slides were wet-mount coverslipped with Permafluor mounting solution (Lipshaw Immunon, Pittsburgh, PA). Each microscopy slide was scanned with argon-ion and helium-neon lasers, and images were digitized at 0.5-µm spatial intervals along the raster scan path. (CompuCyte, Cambridge, MA). On the basis of fluorescent light intensity, contours were automatically drawn to define individual PI-labeled objects. A minimum threshold (inclusion/exclusion gate) was defined by a contour parameter of pixel area that established the size of a standard nucleus. Settings and threshold adjustments were confirmed by visual inspection using a digital camera that was attached to the microscope. Once limits were defined, subsequent sets nearly always fell within the same range; parameters were stored in a protocol file and retrieved for all additional runs. A second integrating contour was drawn to assess intensity of PI fluorescence from objects that expressed CD54 fluorescence. Finally, a third contour was drawn to measure nonspecific background fluorescence. Background was automatically subtracted from the integrated fluorescence of CD54+ cells, a calculation based on gated channels of green fluorescence (FITC). The PI and CD54+ cell population was gated out from other cells and set up in an isolated window for cell cycle program analysis, which automatically proportioned G1, S, and G2 cell cycle components. Intensity of CD54 expression per cell was defined as green channel florescence as a proportion of total fluorescence; tissues from pregnant and nonpregnant mice were analyzed, but postpartum groups were not available for this analysis. Whether fluorescent-labeled cells were in endometrium or myometrium could not be distinguished because of analytic limitations. Cell cycle was determined by measuring intensity of PI fluorescence, an indication of chromatin condensation and nuclear size.
Statistical Analysis
Cell count data were evaluated by one-way ANOVA (SPSS, Chicago, IL). When statistically significant, individual comparisons were made using Bonferroni's test (P < 0.05 was considered significant). If Levene's test for homogeneity of variance was significant, then the Kruskal-Wallis test was used and multiple comparisons were performed.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The pregnant endometrium (mesometrium) was characterized by a loosely organized matrix of cells, blood vessels, and glands, juxtaposed to a tightly packed columnar epithelial cell layer at the interface of the uterine lumen (Fig. 1A). By comparison, the myometrium was a highly organized region composed of an inner layer of circular smooth muscle fascicles and an outer layer of longitudinal smooth muscle bundles. In the nonpregnant uterus, endometrial and myometrial regions were densely packed with cells (Fig. 1B).
|
In the pregnant endometrium, BM8-labeled cells were present, often adjacent to glandular or vascular tissue. Stained cells were commonly found in the endometrial stroma either associated with the subepithelial region or proximal to blood vessels (Fig. 1C; Day 15 of gestation). Though the distribution of macrophages was similar in pregnant and postpartum endometrium, as described below, fewer macrophages were observed in mice on Day 18 of pregnancy, the day before parturition.
In myometrium, labeled cells were exclusively present in connective tissue or stromal areas between the muscle fascicles or bundles (Fig. 1, D–F). On Day 15, clusters of macrophages were present between circular and longitudinal muscle layers (Fig. 1D). Specific stromal staining in cells and their processes around longitudinal muscle fascicles or circular muscle bundles was enhanced in peripartum groups (Fig. 1, E and F; postpartum Day 1 and Day 18, respectively). This network of stained cells and pseudopodia was most apparent in thick frozen tissue sections.
Peripartum Macrophage Numbers
In endometrium, macrophage numbers increased nearly 5-fold on Day 15 of pregnancy compared to those in nonpregnant controls (Fig. 2). Macrophages declined 70% by Day 18, then rebounded postpartum to levels that were not significantly different from those on Day 15 (2784 cells/mm2 of nucleus). In myometrium, macrophage numbers were also increased on Day 15 of pregnancy. Cell numbers remained elevated on the day before birth and postpartum compared to those in the nonpregnant group.
|
Activated Macrophages in the Peripartum Uterus
CD54 was expressed by a significant number of cells in specific compartments within the uterus and in the luminal epithelium. In the nonpregnant mice, FITC-conjugated CD54+ cells were sparsely distributed in subepithelial endometrium and myometrium (Fig. 3, A and B, respectively). In the pregnant uterus, CD54+ cells were dispersed throughout the endometrium and were frequently aligned with, or adjacent to, the luminal epithelium and endothelial lining of the vasculature (Fig. 3, C and D). In pregnant and postpartum myometrial sections, CD54+ cells were most evident in the stromal regions between circular and longitudinal muscle bundles (Fig. 3, E and F, respectively). Expression of CD54 by vascular epithelium was minimal in these regions of the uterus.
|
In addition, increased numbers of resident cells expressed the activation marker CD54 during pregnancy compared to those in nonpregnant mice (Fig. 4, left). At Day 15 of pregnancy, CD54-labeled cells were significantly increased compared to those in all other groups, nearly 5-fold higher than those in uteri from nonpregnant mice. Enhanced numbers of CD54+ cells persisted through the postpartum period compared to those of nonpregnant controls. Technical limitations in the LSC analyses precluded anatomic resolution within uterine compartments of individual fluorescent cells. Although not shown, the average integrated intensity of fluorescence by individual CD54+ cells and by the luminal epithelium was increased by 50% in pregnant uteri (both gestational Day 15 and Day 18 groups) compared to those in uteri from nonpregnant mice (P < 0.05, ANOVA; tissue from postpartum groups was not available for analysis).
|
Although the number of activated uterine macrophages was highest on Day 15 of pregnancy, whether a change occurred during the peripartum period in the percentage of cells that expressed CD54 relative to total number of macrophages remained to be determined. BM8+ cells that were labeled with CD54+ could not be directly assessed since separate serial tissue sections were analyzed by different methods, i.e., brightfield microscopy for BM8 and LSC for CD54. To address the question of percentage of activated macrophages present in the peripartum uterus, CD54+ cell number/mm2 of nucleus for each tissue section was estimated and then normalized to density of nuclei to correct for variations in uterine hypertrophy across groups. These CD54 cell number estimates were divided by BM8+ cell count/mm2 of nucleus. In nonpregnant uteri, CD54+ cells represented approximately 45% of the total number of BM8-stained cells. By contrast, 25% of the total number of BM8-stained cells expressed CD54 in uteri from mice on Day 15 of pregnancy and in postpartum groups. However, on Day 18, the day before birth, CD54 cells represented 60% of BM8-labeled cell numbers. These estimates provide an indication that activation in the macrophage population peaks on the day before birth.
The proportion of cells double-labeled with FITC and PI in various stages of the cell cycle did not vary with respect to reproductive status. Cell cycle analysis across groups indicated that approximately half of the CD54+ cells were in S phase, i.e., actively synthesizing DNA (Fig. 4, right). Of the remaining double-labeled cells, 40% were in G0/G1 phase, recently divided cells or non-DNA synthesizing, while 10% were in mitosis, in G2/M phase. Thus, activated uterine macrophages in nonpregnant, pregnant, and postpartum groups were in a comparable state of proliferation and were not terminally differentiated.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Rather, numbers of activated macrophages in the endometrium were diminished on the day before birth. The loss of endometrial macrophages may signify a divestiture of cells that contribute to immune suppression and uterine quiescence during gestation. This loss could be due to emigration or death of a select population of macrophages. Remaining macrophages, those that reside in the myometrium, may shift the balance of activity to produce inflammatory factors that promote uterine contractile activity. Thus, the data raise the possibility that loss of a critical number or subset of macrophages, accompanied by activation of macrophages that remain in the uterus, is part of the process of parturition.
The results of this study extend previous reports that indicate macrophage numbers are increased during the quiescent phase that characterizes the majority of pregnancy [9, 12, 14]. Peak numbers occur after the midpoint of gestation, around Day 12 in the mouse [14]. On Day 15 of gestation, labeled macrophages are still present in high numbers in both the endometrium and myometrium compared to those in the nonpregnant uterus, particularly in association with uterine epithelium, glandular tissue, and stroma between smooth muscle bundles. Numbers of activated macrophages—cells expressing the activation marker CD54—demonstrate changes parallel to total macrophage numbers. Thus at this stage of pregnancy, enhanced numbers of activated macrophages correspond temporally to elevated iNOS production in the uterus in other rodents [15, 39, 40]. Since iNOS promotes uterine quiescence, the present study provides support for the hypotheses by Buhimschi et al. [16] and others [15, 17] that enhanced production of iNOS may restrain uterine contractile activity at this stage of pregnancy [24]. Furthermore, these data focus attention on the possibility that this inhibitory mechanism involves a macrophage population that principally resides in the endometrium.
As pregnancy nears conclusion, endometrial macrophage numbers decline, but a stable population is maintained in the myometrium. These data raise the possibility that macrophage depletion from the endometrium may account for decreased iNOS production at term [15]. Thus, emigration or death of uterine macrophages may remove a mechanism that promotes uterine quiescence. Eliminating an impediment to uterine activity does not itself explain why contractile activity is enhanced at term in the mouse [24].
Since macrophages in the myometrium are sustained at high numbers relative to those in nonpregnant mice, it is conceivable that resident myometrial macrophages may shift activities towards production of inflammatory mediators that enhance local contractile capabilities. Macrophage processes appear to surround or envelop individual muscle bundles to a greater extent before and after the day of birth than on Day 15 of pregnancy [12]. Throughout this period, a high percentage of activated macrophages are in S phase of the cell cycle; i.e., they are proliferative and not terminally differentiated. This finding is congruent with other studies that described monocyte-like cells or resident macrophages as proliferative in various tissues [41, 42] and capable of cell division in the uterus [5, 9]. Whether a subset of activated macrophages is proliferating on Day 18 and contributes to enhanced uterine contractility just before the onset of labor cannot be excluded from consideration.
Use of two independent methods to count cells precludes direct analysis of activation marker expression by an individual macrophage. However, a preliminary indication based upon the number with CD54 fluorescence per tissue area suggests that a greater percentage of macrophages were activated on Day 18, the day before birth. These activated immunophenotypes are presumably in the myometrium since fewer cells reside in the endometrium on day of pregnancy. As an activation marker, enhanced CD54 expression is thought to facilitate extravasation of leukocytes into areas of inflammation by the binding of CD54 to the integrin, leukocyte function-associated antigen-1 [43, 44]. Whether an interaction between activated macrophages and other constituent cells in endometrium or myometrium may regulate cytokine production at the appropriate time during pregnancy is not known, but synthesis of interferon 
and tumor necrosis factor 
, potent inflammatory mediators, is increased in the murine uterus at term [4, 9, 45]. Since the macrophage is known to produce such inflammatory cytokines and prostaglandins [18, 45, 46], the findings raise the possibility that resident macrophages may influence uterine contractile activity at term.
After birth, macrophage numbers increase in the postpartum uterus. Immigration of macrophages into the evacuated uterus is likely to restore elevated macrophage numbers in the postpartum endometrium. Macrophages are found in the vicinity of the detachment wound in the uterus, known as the postpartum nodule, for up to 3 mo [47]. Influx of these cells is suggested to participate in wound healing, tissue repair, and the restoration of the uterus to the nonpregnant state in preparation for the next cycle of postpartum estrus. Infiltration of leukocytes into the human myometrium follows the onset of labor [48]. However, continued proliferation (50% of CD54 cells were in S phase) also suggests that cell division may temporally coincide with increased macrophage numbers in the aftermath of parturition.
In summary, at the conclusion of pregnancy, the macrophage is proposed to be important for the shift from immune suppression, associated with the majority of pregnancy, to inflammatory immune activity at the onset of labor [4, 5, 9]. This study provides evidence of a modified distribution of macrophages in the uterus before and after birth. A decline in the number of activated macrophages in the endometrium, but not myometrium, on the day before birth suggests that withdrawal of these immune cells may participate in an escape from contractile inhibition at term. One destination for immigration of endometrial macrophages may be the cervix, to facilitate relaxation there, because the population of this immunophenotype increases in the cervix on the day before birth [12]. At the same time, an established, and perhaps transformed, number of activated macrophages in the myometrium raises the possibility that this subpopulation of cells may contribute to amplification of contractile activity that culminates in the onset of labor. Thus, the process of parturition in the mouse may be characterized as a transition from quiescence to inflammation associated, in part, with a shift in the uterine macrophage population at term.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Correspondence. FAX: 909 824 4029; syellon{at}som.llu.edu 
Accepted: December 13, 1999.
Received: June 22, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
gene expression in cycling and pregnant mouse uterus: temporal aspects and cellular localization. J Leukocyte Biol 1998; 64:393–400.[Abstract] on prostaglandin synthesis of oestrogenized rat uterus is mediated by nitric oxide. Prostaglandins Leukot Essent Fatty Acids 1998; 58:413–416.[CrossRef][Medline]This article has been cited by other articles:
|
|
 |
|
  A. S.-Y. Leong, J. E. Norman, and R. Smith Vascular and Myometrial Changes in the Human Uterus at Term Reproductive Sciences, January 1, 2008; 15(1): 59 - 65. [Abstract] [PDF]
|
|
|
|
|
|
L. S. Kirby, M. A. Kirby, J. W. Warren, L. T Tran, and S. M. Yellon Increased Innervation and Ripening of the Prepartum Murine Cervix Reproductive Sciences, December 1, 2005; 12(8): 578 - 585. [Abstract] [PDF]
|
|
|
|
 |
|
 A. M. Mackler, T. C. Ducsay, C. A. Ducsay, and S. M. Yellon Effects of Endotoxin and Macrophage-Related Cytokines on the Contractile Activity of the Gravid Murine Uterus Biol Reprod, October 1, 2003; 69(4): 1165 - 1169. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. M. Yellon, A. M. Mackler, and M. A. Kirby The Role of Leukocyte Traffic and Activation in Parturition Reproductive Sciences, September 1, 2003; 10(6): 323 - 338. [Abstract] [PDF]
|
|
|
|
|
|
K. Tsuboi, A. Iwane, S. Nakazawa, Y. Sugimoto, and A. Ichikawa Role of Prostaglandin H2 Synthase 2 in Murine Parturition: Study on Ovariectomy-Induced Parturition in Prostaglandin F Receptor-Deficient Mice Biol Reprod, July 1, 2003; 69(1): 195 - 201. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
