| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Articles |
a Reproductive Medicine Unit, Department of Obstetrics and Gynaecology, The University of Adelaide, The Queen Elizabeth Hospital, Woodville, South Australia 5011, Australia
b Department of Obstetrics and Gynaecology, The University of Adelaide, Adelaide, South Australia 5005, Australia
c Department of Obstetrics and Gynaecology, Göteborg University, S-41345, Göteborg, Sweden
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
) (reviewed in [1, 2]). Cytokines are implicated in regulating the abundant populations of leukocytes that infiltrate ovarian tissues. By acting directly and indirectly through these leukocytes, they are thought to influence steroidogenesis, extracellular matrix (ECM) remodeling, apoptosis, angiogenesis, vascular permeability, and the induction of immune responses [3]. However, little is known about the precise molecular mechanisms governing the spatial and temporal dynamics of leukocyte recruitment and activation in ovarian tissue.
Recently, granulocyte-macrophage colony-stimulating factor (GM-CSF) has been detected in the ovary. Both protein and mRNA encoding components of the GM-CSF signaling system are present within the human [4, 5], rat [6, 7], and mouse ovary [8]. GM-CSF targets leukocytes of the myeloid lineages that express both the GM-CSF-specific, low-affinity -chain of the GM-CSF receptor and the affinity-converting ß-chain [9]. The ß-chain has no affinity for GM-CSF in the absence of the -chain and is shared with the IL-3 and IL-5 receptor complexes [10]. Although GM-CSF was originally identified as a regulator of myeloid hemopoiesis, it is now clear that the survival, proliferation, differentiation, and functional activation of mature mononuclear phagocytes, granulocytes, and dendritic cells [11] are determined by this cytokine. GM-CSF may also promote the infiltration of myeloid leukocytes into tissues by acting as a specific chemoattractant [12] and through its ability to promote integrin expression in endothelial cells [13]. In addition, GM-CSF may influence the functional activities of these target cells that span a spectrum of cellular and biochemical profiles including phagocytosis, antigen processing and presentation, ECM remodeling, cytotoxicity, and angiogenesis in endothelial cells (reviewed in [14]).
During the processes of folliculogenesis, follicle atresia, ovulation, and corpus luteum development and demise, leukocytes of the macrophage and granulocyte lineages display a close physical association with follicles and corpora lutea (reviewed in [3, 15]). We have proposed that GM-CSF contributes to regulation of the tissue remodeling events and immunological changes within the ovary through its actions on resident and infiltrating myeloid cells.
This study was undertaken to determine whether GM-CSF gene ablation alters parameters of ovarian function including estrous cycle length, ovulation rate, steroidogenesis, corpus luteum development, and ovarian leukocyte recruitment and activation. In vivo ovulation rates were assessed in immature mice after priming with exogenous gonadotropins and in adult naturally cycling mice. Ovulation was also assessed in an in vitro perfusion model, which enables measurement of ovarian-derived secretions released during the periovulatory period. Circulating progesterone was assessed following ovulation in vivo and in vitro and on Day 4 of pregnancy. The duration of pseudopregnancy and the weight of the ovary on Day 4 of pregnancy was determined as a measure of corpus luteum development. We also examined the spatial and temporal dynamics of leukocyte populations in the ovary during the periovulatory period and pregnancy in an immunohistochemical analysis with leukocyte lineage-specific monoclonal antibodies.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Mice were housed in a specific pathogen-free facility at the University of Adelaide with controlled temperature and lighting (12L:12D). Food and water were supplied ad libitum. Mice deficient in GM-CSF (GM-/-) were generated using gene targeting techniques in 129 (mouse type, not number of) embryonic stem cells [16], and breeding pairs were kindly provided by Ashley Dunn (Ludwig Institute for Cancer Research, Melbourne, Australia). Conditioned media generated from spleen and uterine cells of GM-/- mice contained no bioactive or immunoactive GM-CSF [16, 17]. Control mice (GM+/+) were derived from F2 offspring of GM-/- females crossed with wild-type 129 males, and they were identified as homozygotes for the wild-type gene by diagnostic polymerase chain reaction [16].
Immature, 3- to 4-wk-old mice were primed to ovulate by the s.c. administration of 5 IU eCG (Intervet, Boxmeer, The Netherlands) to stimulate follicle growth, followed by i.p. administration of 5 IU hCG (Intervet) 48 h later to initiate the ovulatory process. All adult females were > 8 wk of age and were virgins at the onset of the experiment.
Estrous Cycle Tracking
Estrous cycle length was assessed in naturally cycling adult GM+/+ (n = 14) and GM-/- (n = 16) mice. The mice were housed 4 per cage and in close proximity to stud males. Daily vaginal smears were examined as wet, unstained preparations to determine the stage of the cycle as described by Bronson [18]. Mean estrous cycle length was determined as an average over 5 completed cycles.
In Vivo Ovulation Rate
In vivo ovulation rate was assessed in immature 129 (n = 8), GM+/+ (n = 6), and GM-/- (n = 7) mice, primed to ovulate by the administration of 5 IU eCG followed by 5 IU hCG 48 h later. Mice were killed on the morning of ovulation between 0900 and 1200 h, and oocytes in the ampulla region of each oviduct were counted. In vivo ovulation rate was assessed in naturally cycling adult mice after estrous cycle tracking. At metestrus-1, following a minimum of 5 completed estrous cycles, mice were killed, and the oviduct was dissected free from the ovary and uterus. Oocytes were dissected from the ampulla or flushed from the oviduct with Ca2+- and Mg2+-free Dulbecco's PBS, pH 7.4 (hereafter referred to as PBS), using a blunt 30-gauge needle attached to a 2-ml syringe.
In Vitro Perfusion of Ovaries
Adult female GM+/+ and GM-/- mice were synchronized to ovulate by the i.p. administration of 20 µg of an LHRH agonist (des-Gly10,[D-Ala6]-LH-RH ethylamide; Sigma Chemical Co., St. Louis, MO), followed 48 h later by the s.c. administration of 5 IU eCG. On the morning prior to ovulation (48 h post-eCG treatment), the right ovary with its vasculature was surgically isolated and perfused as described by Brännström and Flaherty [19]. After 1 h of perfusion, 0.1 µg/ml ovine LH (LH-26, lot #AFP-5551B, specific activity 2.3 U/mg; National Hormone and Pituitary Program, National Institute of Diabetes and Digestive and Kidney Diseases, Bethesda, MD), either alone or in combination with 4 ng/ml recombinant murine GM-CSF (rmGM-CSF; specific activity 107 U/mg; Pharmingen, San Diego, CA), was added to the perfusion media at Time 0 h. Samples of the recirculating perfusion medium were collected at 0, 1, 2, 3, 4, 8, and 22 h, stored at -20°C, and subsequently assayed for progesterone, androstenedione, and estradiol. The number of oocytes ovulated was assessed at 22 h by counting the number released into the bursa. Eleven GM+/+ mice were set aside and allowed to ovulate in vivo under the same hormone priming regime as mice whose ovaries were perfused. These mice were killed 16–20 h after i.p. administration of 5 IU hCG, and the number of oocytes in the ampulla region of each oviduct was counted.
Corpora Lutea Development
Ovarian weight and the number of corpora lutea present in the ovary were assessed on Day 4 of pregnancy in adult GM+/+ (n = 11) and GM-/- (n = 11) mice mated with studs of the same genotype. On Day 4 of pregnancy, mice were killed; their ovaries were dissected free from mesentery and fixed in 10% formalin in PBS. The wet weight of ovaries was determined, and then the ovaries were embedded in paraffin. Serial sections were cut at 50-µm intervals and counterstained with hematoxylin and eosin. The total number of corpora lutea was counted in both ovaries.
Duration of Pseudopregnancy
The duration of pseudopregnancy was assessed in naturally cycling adult GM+/+ (n = 9) and GM-/- (n = 10) mice. One GM+/+ and 1 GM-/- adult female mouse were housed per cage with a vasectomized adult male until the completion of three consecutive pseudopregnancies. The presence of a vaginal plug was regarded as Day 1 of pseudopregnancy. The duration of pseudopregnancy was calculated as the number of days lapsed until a second vaginal plug was detected. During the second pseudopregnancy, luteolysis was induced by the i.p. administration of 3 µg of prostaglandin F2 (PGF2) (Jurox, Silverwater, Australia) in 0.9% NaCl between 0900 and 1200 h on Days 4 and 5 after sighting of the second vaginal plug according to a protocol adapted from Bartke and coworkers [20]. The number of days elapsed until the third, and finally the fourth, vaginal plug was also determined.
Immunohistochemical Localization of Ovarian Leukocytes
Ovaries were collected from GM+/+ (n = 5) and GM-/- mice (n = 5) and were embedded in Tissue-Tek OCT Compound (Miles, Elkhart, IN), frozen in liquid nitrogen-cooled isopentane, and stored at -20°C. Immature ovaries were collected from immature unprimed mice, preovulatory ovaries from mice 48 h after 5 IU eCG, and ovulatory ovaries at 9 h after 5 IU eCG/hCG. Luteinized ovaries were collected from immature primed mice postovulation on Day 1 of pregnancy (Day 1), Day 3 of pregnancy (Day 3), and 24–48 h postpartum (PP). Fresh frozen tissue sections (6 µm) were fixed in 96% ethanol for 10 min at 4°C, and leukocytes were detected using monoclonal antibodies (mAbs; American Type Culture Collection, Rockville, MD), including CD45 (LCA) antigen (all leukocytes), CD11b (Mac-1) (monocyte/macrophages, neutrophils, and some B and T lymphocytes and natural killer cells), F4/80 (macrophages), and class II major histocompatibility complex (MHC) (Ia; activated macrophages, dendritic cells, and some B lymphocytes). CD4 (L3T4), CD5, and CD8 (Lyt 2) antigens expressed by subpopulations of T lymphocytes and some B lymphocytes, as well as the Thy 1.2 antigen, were also localized. Sections were incubated for 3 h at 4°C in undiluted hybridoma supernatant plus 10% normal mouse serum (NMS). This was followed by a 2-h incubation with biotinylated rabbit -rat IgG (Dako, Glostrup, Denmark) or for antibody CD8, biotinylated goat -rat IgM (Pierce Chemical Company, Rockford, IL), both diluted 1/300 in PBS plus 1% PBS-BSA and 10% NMS. Biotinylated antibodies were subsequently labeled during a 40-min incubation in avidin-horseradish peroxidase (Dako) diluted 1/400 in PBS-BSA plus 10% NMS. Bound antibody was detected after a 10-min incubation at room temperature with 3,3'-diaminobenzidine peroxidase substrate (DAB; Sigma). Slides were counterstained in hematoxylin, cleared, and mounted. Negative control sections were incubated as described but with omission of the primary antibody, while spleen tissue and/or Day 1 pregnant uterus were used as positive controls. The specificity of binding of all antibodies was validated in spleen, lymph node, and reproductive tract tissues against isotype-specific control antibodies. Staining was quantified using a video image analysis system and Video Pro software (Leading Edge, Adelaide, Australia). The intra- and interassay coefficients of variation were 8.9% and 10.3%, respectively. Mean percentage positive stain was defined as percentage brown stain (DAB)/total stain (hematoxylin+DAB) and was determined by analyzing multiple fields within the ovarian stroma (n = 4–8 fields per ovary), theca (n = 4 fields per follicle, diameter > 350 µm), and corpora lutea (n = 4 fields per corpus luteum). Between 1 and 5 follicles or corpora lutea were measured per mouse.
Blood Sampling
Blood was collected from adult GM+/+ and GM-/- mice following natural ovulation (n = 12 mice per group) and on Day 4 of pregnancy (n = 10 mice per group) via cardiac puncture in mice anesthetized with 2% Avertin (15 µl/g BW; Aldrich Chemical Company, Milwaukee, WI). Blood was collected into EDTA-coated tubes (Terumo, Piscataway, NJ), and the cellular component was pelleted by centrifugation at 13 000 x g for 15 min. Plasma was stored at -20°C and subsequently assayed for progesterone and estradiol.
Assays
Plasma progesterone concentrations were analyzed by Amerlex-M RIA (Amersham International, Amersham, UK) from a 1.3–127 nM standard curve. The intra- and interassay coefficients of variation were 
5.1% and 9.4%, respectively, and the sensitivity of the assay was 0.25 nM. Progesterone concentrations in perfusion media were analyzed by ACS chemiluminescent immunoassay (Ciba Corning Diagnostics, Medfield, MA) from a 1.9–127.2 nM standard curve. The intra- and interassay coefficients of variation were calculated to be 10% and 12%, respectively [21], and the assay sensitivity was 0.35 nM. Androstenedione concentrations in perfusion media were analyzed by RIA (Diagnostic Systems Laboratories, Webster, TX) from a 0.34–34.5 nM standard curve. The intra- and interassay coefficients of variation were calculated to be 5.9 and 7%, respectively, and the assay sensitivity was 0.069 nM. Plasma estradiol concentrations were analyzed by Clinical Assays Estradiol-2 RIA (Sorin Biomedica Diagnostics, Vercelli, Italy) from a 37–7340 pM standard curve. The intra- and interassay coefficients of variation were 8% and 10%, respectively, and the assay sensitivity was 18 pM. Estradiol concentrations in perfusion media were analyzed by ACS chemiluminescent immunoassay (Ciba Corning Diagnostics) from a 108–5000 pM standard curve; the intra- and interassay coefficients of variation were 7% and 10%, respectively [22], and the assay sensitivity was 36.7 pM. Nitrate/nitrite was detected as a measure of the accumulation of nitric oxide (NO) in perfusion media by conversion of nitrate to nitrite and analysis of total nitrite in a colorimetric assay (Cayman Chemical Company, Ann Arbor, MI). A standard curve ranging from 5 to 35 µM was calibrated by inclusion of nitrate, and the intraassay coefficient of variation was 4%.
Statistics
Data are presented as mean ± SE, and data expressed as percentages were arcsin transformed for statistical analysis. Graphpad Instat version 2.04a (Graphpad Software, San Diego, CA) was used to analyze all data except perfusion data that were ranked prior to assessment by ANOVA and Tukey Honestly Significant Difference post hoc test using SPSS (SPSS, Chicago, IL). Statistical significance was accepted when P < 0.05.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
To investigate the effect of GM-CSF deficiency on the estrous cycle, the cellular characteristics of daily vaginal smears were examined in GM+/+ and GM-/- adult female mice, and the length of the estrous cycle was determined. Vaginal smears reflected changes in the structure of the vaginal epithelium and followed a regular and predictable sequence across the course of the cycle. The mean length of the estrous cycle in GM+/+ mice was 4.9 ± 0.3 days (n = 14). In contrast, GM-/- mice exhibited an estrous cycle of 6.5 ± 0.5 days (n = 16), an extension of approximately 1.5 days (P = 0.02, Mann-Whitney test) (Fig. 1A). The duration of the metestrus-2 and/or diestrus stage of the cycle was found to be significantly longer in GM-/- than in GM+/+ females (Fig. 1B).
|
Effect of GM-CSF Deficiency on Ovulation Rate
To investigate ovulation rates, oocytes were collected and counted from immature mice primed with gonadotropins and naturally cycling adult mice. Comparable numbers of oocytes were ovulated from GM+/+ and GM-/- mice regardless of genotype or hormone priming. The numbers of oocytes ovulated from immature primed GM+/+ and 129 mice were similar (Table 1) (P = 0.94, ANOVA). Comparable numbers of oocytes were also ovulated from naturally cycling adult GM+/+ and GM-/- mice (Table 1) (P = 0.98, unpaired t-test). In naturally cycling adult mice, oocytes were retrieved in equal proportions from the ampulla and from the flushed isthmus of the oviduct (P = 0.41, Fisher's exact test), suggesting that the rate of transport through the oviduct was similar in GM+/+ and GM-/- mice. Together, these results suggest that the timing of ovulation and the number of follicles ovulated are unaltered in GM-CSF-deficient mice.
|
The effect of GM-CSF deficiency on ovulation rate was also examined in an in vitro perfusion system. Of the GM+/+ ovaries perfused for 22 h with LH, 70% ovulated one or more oocytes compared to 93% of GM-/- ovaries (Table 2). The mean number of oocytes ovulated per GM-/- ovary was 2.8-fold higher than that for the GM+/+ ovary (P = 0.08, ANOVA). The addition of rmGM-CSF did not alter the ovulation rate in GM-/- ovaries. However, ovulation in GM+/+ ovaries was very inefficient in the in vitro perfusion system, with 8.6-fold fewer oocytes per ovary released in vitro than in vivo (17.2 ± 2.7 oocytes per ovary), despite use of the same hormone priming regime. These results suggest that the mechanisms inhibiting ovulation in vitro may be ameliorated to some extent by GM-CSF deficiency.
|
NO liberation is important for successful ovulation in the perfusion system [23]; so to investigate the impact of GM-CSF deficiency on NO production in the ovary at ovulation, nitrate/nitrite concentrations were measured in the perfusion media at 22 h after the addition of the treatments (Table 2). Addition of LH did not alter nitrate/nitrite secretion by GM+/+ ovaries as compared with GM+/+ control ovaries perfused with media alone (P = 0.99, ANOVA). Nitrate/nitrite secretion by GM-/- ovaries perfused with LH was significantly greater than that in control and LH-perfused GM+/+ ovaries. Addition of rmGM-CSF to GM-/- ovaries did not increase nitrate/nitrite secretion (P = 0.57, ANOVA).
Effect of GM-CSF Deficiency on Ovarian Steroid Hormone Synthesis
To investigate the effect of GM-CSF deficiency on ovarian steroid hormone synthesis, plasma was collected from adult mice following natural ovulation. Comparable plasma progesterone levels were recorded from mice irrespective of genotype (9.1 ± 1.3 vs. 10.2 ± 1.2 nM plasma in GM+/+ and GM-/- mice, respectively) (P = 0.54, unpaired t-test). Only 50% (n = 6 of 12) of GM+/+ and 42% (n = 5 of 12) of GM-/- mice had detectable plasma estradiol following natural ovulation; but where this was measurable, there appeared to be no difference (50 ± 2.6 vs. 46 ± 2.4 pM plasma for GM+/+ and GM-/- mice, respectively) (P = 0.30, unpaired t-test).
To further investigate the impact of GM-CSF deficiency on steroid production by the ovary, plasma was collected from adult mice on Day 4 of natural pregnancy. Plasma progesterone was significantly reduced in GM-/- (116.5 ± 6 nM) compared to GM+/+ mice (141.6 ± 10.3 nM) (P = 0.05, t-test). Estradiol was not detected in the plasma of mice on Day 4 of pregnancy.
Release of steroid hormones was also measured from the ovaries perfused in vitro. The concentrations of cumulative progesterone, androstenedione, and estradiol in perfusion media from GM+/+ and GM-/- ovaries were determined over the 22-h period of perfusion. The addition of LH resulted in an approximate 20-fold increase in progesterone secretion from the GM+/+ ovaries, with maximum concentrations in the perfusion media 2–3 h after LH addition and levels subsequently declining due to rapid adsorption of steroid into the perfusion equipment (Fig. 2A). This profile of secretion was observed for all perfused ovaries except those perfused with media alone. Progesterone concentrations in the media of ovaries from GM+/+ and GM-/- mice perfused with LH were comparable regardless of genotype and the addition of rmGM-CSF, but they were significantly greater than those of the control group at all time points examined.
|
The addition of LH resulted in an approximate 4.5-fold increase in androstenedione secretion from the GM+/+ ovary, with maximum concentrations detected in the perfusion media 2 h after the addition of LH (Fig. 2B). The addition of LH to ovaries of both genotypes stimulated androstenedione secretion above that of the control at the majority of sampling times. Androstenedione concentrations in perfusion media of GM+/+ and GM-/- ovaries perfused with LH were comparable. The addition of rmGM-CSF to GM-/- ovaries resulted in a further increase in mean androstenedione production 3.7-fold greater than that of the GM-/- ovaries perfused with LH alone, although because of the large variability between individual ovaries, this increase was not statistically significant (P = 0.13, ANOVA).
The addition of LH resulted in a 2.3-fold increase in mean estradiol secretion from the GM+/+ ovary, with concentrations in the perfusion media reaching a maximum 3 h later (Fig. 2C). Peak secretion of estradiol by GM-/- ovaries perfused with LH+rmGM-CSF occurred 1 h later at 4 h. The addition of rmGM-CSF to the ovaries from GM-/- mice stimulated a 2-fold increase in estradiol production compared with that for GM-/- ovaries, although this difference was not statistically significant (P = 0.39, ANOVA).
Therefore, GM-CSF deficiency was not associated with major alterations in steroid secretion at ovulation, but it was associated with a reduction in progesterone production on Day 4 of pregnancy. In addition, exogenous GM-CSF may promote production of androstenedione.
Effect of GM-CSF Deficiency on Corpus Luteum Development and Demise
To investigate the impact of GM-CSF deficiency on development of the corpus luteum, ovarian weight and the number of corpora lutea present in the ovary were assessed on Day 4 of natural pregnancy in adult mice. Ovarian weight was significantly reduced in GM-/- mice compared with GM+/+ mice (Table 3). However, the number of corpora lutea on Day 4 of pregnancy was similar in ovaries from GM+/+ and GM-/- mice.
|
To further investigate whether GM-CSF deficiency may influence the kinetics of the development and/or demise of the corpus luteum, the durations of natural pseudopregnancy and pseudopregnancy terminated by PGF2 administration were examined. The number of days prior to the sighting of the first vaginal plug in GM+/+ mice (on Day 2.3 ± 0.4) and GM-/- mice (on Day 1.6 ± 0.2) was comparable (P = 0.23, Mann-Whitney test). The length of natural pseudopregnancy in GM+/+ and GM-/- mice was also comparable (Fig. 3). The administration of PGF2 induced luteal regression and premature termination of pseudopregnancy as seen by the onset of estrus generally within 2 days of administration in both GM-/- and GM+/+ mice (P < 0.05, Kruskal-Wallis ANOVA). The duration of the third cycle of pseudopregnancy after PGF2 treatment was also similar for the two genotypes and was comparable to that seen before PGF2 treatment, illustrating no delayed effect of PGF2 administration in either group. These results suggest that recruitment of follicles for ovulation is not disrupted in genetically GM-CSF-deficient mice, and that the time required for full completion of the processes of luteinization and subsequent luteolysis is similarly not affected by a lack of this cytokine.
|
Effect of GM-CSF Deficiency on the Recruitment and Activation of Ovarian Leukocytes During the Periovulatory Period
Hematoxylin-stained sections of periovulatory ovaries collected from GM+/+ and GM-/- mice comprised follicles from primordial to antral stages. Ovaries collected from pregnant and postpartum mice also comprised follicles but predominantly corpora lutea. There were no overt abnormalities in the histology of the ovaries collected from GM-/- mice at any of the stages examined.
To assess the effect of GM-CSF deficiency on the distribution and activation of leukocytes in the ovary, immunohistochemistry using an indirect immunoperoxidase technique was performed on sections cut from immature and periovulatory GM+/+ and GM-/- ovaries. Leukocyte common antigen (CD45)-positive cells were localized predominantly within stromal tissue and surrounding blood vessels but were also found within the theca cell layer of follicles, adjacent to the basement membrane, in both GM+/+ and GM-/- ovaries. Seldom were leukocytes found within the granulosa cell layer of follicles, and in such cases these follicles exhibited atretic characteristics [24]. There was no significant effect of GM-CSF deficiency on the distribution or intensity of CD45-positive staining in the stroma or theca during the periovulatory period (Fig. 4A).
|
Evaluation of sections labeled with lineage-specific monoclonal antibodies showed that the leukocytes present in the ovary were predominantly macrophages, neutrophils, and/or dendritic cells. Macrophages were localized using the macrophage-specific monoclonal antibody F4/80, which shows reactivity with comparatively large cells with membranous processes. Cells of similar macrophage-like morphology were also CD11b reactive, while smaller CD11b-reactive cells with granular nuclei were presumed to be neutrophils. The proportion of positive staining in the stroma and theca during the periovulatory period was relatively constant, and there were no significant differences in CD11b- or F4/80-positive staining between GM+/+ and GM-/- mice (Figs. 4B and C). A large population of MHC class II (Ia) antigen-expressing cells with dendritic morphology was identified in both the stroma and theca, and potentially included both activated macrophages and dendritic cells. The extent of Ia staining was significantly diminished in the stromal region of ovulating ovaries in GM-/- mice compared with GM+/+ mice; however, no significant differences were observed in the theca (Figs. 4D and 5). Eosinophils were identified in only a few of the negative control sections owing to their characteristic endogenous peroxidase activity. CD4- and CD5-positive T lymphocytes were also very sparsely distributed in the stromal regions and were not quantified. Interestingly, anti-Thy 1.2 was found to react with a large number of cells within the stroma and theca of periovulatory ovaries. These cells did not have the characteristic shape of T lymphocytes or natural killer cells but were elongated with a fibroblast-like morphology. The degree of reactivity varied in relation to the developmental stage of follicles, with the most strongly labeled cells surrounding preantral follicles. A progressive decrease in the intensity of labeling was evident in follicles of increasing developmental stage. There was no effect of GM-CSF genotype on the numbers or location of eosinophils, T lymphocytes, or Thy 1.2-positive cells in ovaries.
Effect of GM-CSF Deficiency on the Recruitment and Activation of Ovarian Leukocytes in the Pregnant Ovary
To assess the influence of GM-CSF deficiency on the distribution and activation of leukocytes in the ovary during pregnancy, immunohistochemistry was performed on sections cut from ovaries collected from pregnant and postpartum GM+/+ and GM-/- mice. Leukocytes (CD45) were localized predominantly within stromal tissue and in corpora lutea of the GM+/+ and GM-/- ovaries. Large populations of CD11b-, F4/80-, and Ia-positive cells were evident in the stroma and corpora lutea throughout pregnancy (Fig. 6, A–C). While the numbers of these cells were comparable in Day 1 and Day 3 pregnant ovaries, they increased dramatically in ovaries collected from postpartum mice. CD11b-positive staining was significantly reduced in the stroma of GM-/- ovaries. In contrast, no significant differences were observed in Ia-positive staining between GM+/+ and GM-/- mice during pregnancy. Eosinophils were identified on a few of the negative control sections, although generally only one cell was observed per 1–2 sections. CD5- and CD8-positive T lymphocytes were also very sparsely distributed. Anti-Thy 1.2 was found to react with a large number of fibroblast-like cells in the stroma surrounding corpora lutea and theca of immature follicles.
|
Therefore, GM-CSF deficiency was not associated with a major alteration in ovarian leukocyte populations; but GM-CSF may contribute to the regulation of MHC class II expression in cells at the time of ovulation, and to the recruitment of macrophages and neutrophils during the postpartum period.
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
There are a variety of potential mechanisms whereby altered leukocyte behavior may lead to disruption of the rhythmicity of the estrous cycle. However, it is most reasonable to speculate that the underlying cause might be inadequate or delayed tissue remodeling due to insufficient or inappropriate macrophage activation. Indeed, ovarian dysregulation of a similar but more severe phenotype is observed in CSF-1-deficient cfmsop/cfmsop mice, which are almost totally devoid of ovarian macrophages [25].
Our previous studies show that the corpus luteum is the major site of GM-CSF synthesis and suggests that GM-CSF expression is maximal during and after ovulation. Together with the current finding of diminished ovarian size and progesterone secretion on Day 4 of pregnancy in GM-/- mice, this indicates that GM-CSF is likely to have a principal role in the formation of the corpus luteum, particularly the proliferation of luteal cells and the regulation of their functional capacity. Our observation that GM-/- mice become stalled in the metestrus-2 and/or diestrus stages of the cycle also supports a role for this cytokine in the events associated with remodeling of the postovulatory follicle during nonpregnant cycles. The location of expression of the GM-CSF receptor subunits correlates with this proposed role [4, 5, 8]. GM-CSF deficiency does not appear to alter the time frame of corpus luteum demise, however, as the length of pseudopregnancy was unaltered in the case of GM-CSF deficiency.
A role for GM-CSF in the proliferation or differentiation of luteal cells would also be consistent with the observed timing of the effects on progesterone secretion. In both the in vivo- and in vitro-perfused ovulation models, progesterone levels were unaltered by GM-CSF deficiency. It was not until Day 4 of pregnancy, when corpus luteum secretion of progesterone should be maximal, that progesterone was significantly reduced in the GM-/- mouse. A tendency for reduced progesterone secretion was evident in the in vitro perfusion system when results were expressed per ovulated oocyte. These data suggest that while progesterone secretion during an estrous cycle may occur in the absence of GM-CSF, the increased demand in the case of pregnancy may not be able to be met at a time when the corpus luteum should be fully functional. In addition to effects on progesterone, our experiments in in vitro-perfused ovaries demonstrated that GM-CSF may stimulate androstenedione synthesis. The mechanism for this effect is unknown, but the differential effects of GM-CSF on progesterone and androstenedione suggest an enzyme P45017-dependent pathway. Estradiol synthesis was moderately increased, probably as a consequence of increased precursor.
A principal role for GM-CSF in the corpus luteum as opposed to a role in follicle development is consistent with the observation that ovulation is unperturbed in GM-/- mice. Our findings of similar numbers of oocytes recovered from both immature primed and naturally cycling adults are in agreement with reports of a lack of effect of GM-CSF on in vivo ovulation rate in the rat [26] and of comparable numbers of implantation sites in GM+/+ and GM-/- mice on Day 8 of pregnancy [17]. However, the developmental competence of oocytes ovulated from GM-/- mice was not addressed in this study, and the possibility of altered maturation kinetics remains.
The precise identity of the cells targeted by GM-CSF in the developing corpus luteum remains unclear. Although GM-CSF may act directly on luteal cells, or perhaps through promoting vascularization via endothelial cell migration and proliferation [27], it is more likely that indirect actions through ovarian leukocytes influence progesterone secretion. For example, it is known that GM-CSF can promote the secretion of macrophage products such as IL-1ß and TNF, and inhibit the secretion of IL-1 inhibitors and NO, all of which can influence steroidogenesis. Our finding of increased NO secretion from GM-/- ovaries is consistent with this proposal. Interestingly, synthesis of IL-1, NO, superoxide radicals, and hydrogen peroxide is also known to be disregulated in macrophages from GM-/- mice [28–30], suggesting that macrophage responses to stimulation can be inadequate in the absence of GM-CSF. An indirect effect of GM-CSF mediated through macrophages would agree with findings that GM-CSF had no effect on LH-stimulated progesterone synthesis by luteinized granulosa cells in vitro [7, 31]. Macrophages are not the only ovarian leukocytes likely to be targeted by GM-CSF in the preovulatory rat ovary; for example, it has been found that GM-CSF enhances IL-1ß-induced release of histamine, presumably from mast cells [32].
GM-CSF deficiency appears to have little effect on leukocyte recruitment into the ovarian tissue, since leukocytes accumulate in the ovary in relatively normal numbers in the absence of GM-CSF. This contrasts with the situation for CSF-1, which appears to be essential for macrophage recruitment and survival, since macrophage populations are vastly reduced in the ovaries of cfmsop/cfmsop mice [25, 33]. Rather, GM-CSF is implicated in driving the phenotypic maturation of macrophages that occurs subsequent to their recruitment. In particular, changes in MHC class II (Ia) expression and secretory profile of macrophages [28–30] appear to result from the absence of this cytokine. MHC class II (Ia), an indicator of macrophage activation or antigen-presenting activities in dendritic cells, was found in the current study to be clearly diminished in ovarian leukocytes. This suggests that altered phenotype in macrophages and dendritic cells is a feature common to reproductive tract [34] as well as lymphoid and inflammatory tissues in GM-CSF-deficient mice. The function of the MHC class II-expressing cells in the ovary is unclear, but it may be that antigen processing and presentation are required to promote lymphocyte circulation through the tissue and perhaps even to facilitate immunoregulated killing of superfluous cells. Whatever their precise role, these macrophages and/or dendritic cells activated by GM-CSF appear to be less critical than the CSF-1-dependent macrophage population, since the effect of GM-CSF deficiency on ovarian function is relatively moderate compared with the major disruption seen in csfmop/csfmop mice [25]. The clear differences between the moderate infertility of GM-CSF knockout mice [17, 35] and greatly compromised fertility of csfmop/csfmop mice [36] may be partially attributable to the heterogeneous functions performed in the ovary by macrophages influenced by GM-CSF as opposed to CSF-1 [37].
In summary, our results indicate that GM-CSF is not essential for ovulation. Instead, this cytokine is likely to be involved in the formation of the corpus luteum from follicle remnants and in the regulation of luteal phase steroidogenesis, via the activation of macrophages at the time of ovulation. Despite the subtlety of the ovarian lesion in GM-CSF-deficient mice, there may be significant physiological consequences of this disruption in corpus luteum function. The reduced fetal size and increased postnatal mortality seen in GM-/- mice [17] may originate partly as a consequence of inadequate luteal support. Although implantation rates are normal on Day 8 of pregnancy in GM-CSF-deficient mice, it is not unreasonable to propose that moderate reductions in circulating progesterone at the time of implantation could contribute to the less-than-optimal placental development that eventually causes fetal demise in these mice. Indeed, this view is consistent with the growing recognition that subtle alterations in the uterine environment at the time of implantation can manifest as diminished health and viability of the fetus later in gestation.
|
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 This study was supported by NHMRC (Australia) and The University of Adelaide. M.J.J. was supported by a Reproductive Medicine Scholarship and Queen Elizabeth Hospital Scholarship. Presented in part at the annual meetings of the Australian Society for Reproductive Biology (1995), Federation of Immunological Societies of Asia-Oceania (1996), and Fertility Society of Australia (1997). 
2 Correspondence. FAX: 61 8 8222 7521; robert.norman{at}adelaide.edu.au
Accepted: October 20, 1999.
Received: May 28, 1999.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and ß receptor messenger ribonucleic acid and protein in human ovarian tissue. Biol Reprod 1995; 53:923–930.[Abstract]This article has been cited by other articles:
|
|
 |
|
  W. V. Ingman and R. L. Jones Cytokine knockouts in reproduction: the use of gene ablation to dissect roles of cytokines in reproductive biology Hum. Reprod. Update, March 1, 2008; 14(2): 179 - 192. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Tonello and G. Poli Tubal ectopic pregnancy: macrophages under the microscope Hum. Reprod., October 1, 2007; 22(10): 2577 - 2584. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Wu, S. Fujii, N.K. Ryan, K.H. Van der Hoek, M.J. Jasper, I. Sini, S.A. Robertson, R.L. Robker, and R.J. Norman Ovarian leukocyte distribution and cytokine/chemokine mRNA expression in follicular fluid cells in women with polycystic ovary syndrome Hum. Reprod., February 1, 2007; 22(2): 527 - 535. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
K.R. Barnett, C. Schilling, C.R. Greenfeld, D. Tomic, and J.A. Flaws Ovarian follicle development and transgenic mouse models Hum. Reprod. Update, September 1, 2006; 12(5): 537 - 555. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Wang, Y. Wen, M. L. Polan, R. Boostanfar, M. Feinman, and B. Behr Exogenous granulocyte-macrophage colony-stimulating factor promotes follicular development in the newborn rat in vivo Hum. Reprod., October 1, 2005; 20(10): 2749 - 2756. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. Gangnuss, M. L. Sutton-McDowall, S. A. Robertson, and D. T. Armstrong Seminal Plasma Regulates Corpora Lutea Macrophage Populations During Early Pregnancy in Mice Biol Reprod, October 1, 2004; 71(4): 1135 - 1141. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Wu, K. H. Van der Hoek, N. K. Ryan, R. J. Norman, and R. L. Robker Macrophage contributions to ovarian function Hum. Reprod. Update, March 1, 2004; 10(2): 119 - 133. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. N. Sferruzzi-Perri, S. A. Robertson, and L. A. Dent Interleukin-5 Transgene Expression and Eosinophilia Are Associated with Retarded Mammary Gland Development in Mice Biol Reprod, July 1, 2003; 69(1): 224 - 233. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. F. Seymour and J. J. Presneill Pulmonary Alveolar Proteinosis: Progress in the First 44 Years Am. J. Respir. Crit. Care Med., July 15, 2002; 166(2): 215 - 235. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A. Robertson, C. Sjöblom, M. J. Jasper, R. J. Norman, and R. F. Seamark Granulocyte-Macrophage Colony-Stimulating Factor Promotes Glucose Transport and Blastomere Viability in Murine Preimplantation Embryos Biol Reprod, April 1, 2001; 64(4): 1206 - 1215. [Abstract] [Full Text]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
