| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
a Nuffield Department of Obstetrics and Gynaecology, University of Oxford, The Women's Centre, John Radcliffe Hospital, Headington, Oxford, OX3 9DU, United Kingdom
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
The expression of several cytokines at the materno-fetal interface and their possible functions in reproduction has been described [1]. Uterine epithelial cells express colony stimulating factor 1 (CSF-1) [2, 3], granulocyte macrophage-colony stimulating factor (GM-CSF), and interleukin (IL) 6 at high levels before implantation [4, 5]. The CSF-1 receptor has been identified by in situ hybridization in murine trophoblast cells [6], and receptors for granulocyte-colony stimulating factor (G-CSF) and GM-CSF (G-CSFR and GM-CSFR, respectively) have been detected on first-trimester human trophoblast cells [7, 8]. There is also increasing evidence that cytokines may be involved in trophoblast development since DNA synthesis in cultured murine trophoblast cells is increased in response to CSF-1 and GM-CSF [9]. Furthermore, GM-CSF and IL-3 given to abortion-prone (CBA/J x BALB/c) mice causes an increase in placental mass and fetal weight and reduces fetal resorption rates [10]. Leukemia inhibitory factor (LIF) is essential for implantation of mouse blastocysts [11], and null mutation of the LIF receptor disrupts normal placental development, leading to poor intrauterine nutrition and eventual death of the offspring [12].
The placenta is one of the few nonhemopoietic tissues to express G-CSF. In the mouse, G-CSF has been detected in placental spongiotrophoblast and labyrinthine trophoblast [13], and both first- and third-trimester human placental explants secrete biologically active G-CSF [14, 15]. Although its function in the placenta is unknown, in the hemopoietic system G-CSF acts to regulate the growth, differentiation, and survival of neutrophilic granulocytes [16].
The biological function of G-CSF is mediated via its interaction with membrane-bound G-CSFR, which bind G-CSF exclusively [17]. These receptors belong to the cytokine receptor superfamily, members of which share amino acid homology in their extracellular sequences [18]. At least four G-CSF receptors exist; three are membrane-bound, and the fourth is soluble [19, 20]. The three G-CSFR isotypes possess identical extracellular domains containing the ligand-binding region, but lack amino acid homology within the transmembrane and cytoplasmic domains [19, 20]. In contrast to human granulocytes, which express only the class I membrane-bound G-CSFR [20], three distinct cDNAs corresponding to class I, class III, and D-7 membrane-bound G-CSFR molecules have been isolated from human placental libraries [19, 20]. Differential expression of G-CSFR isotypes may therefore underlie different functional properties of G-CSF.
We have recently demonstrated the expression of G-CSF and G-CSFR in the human placenta [7]. In this report, we describe further the temporal and spatial regulation of G-CSFR mRNA and protein expression in villus and in extravillous trophoblast cells during the development of the human placenta.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Placental tissues were obtained from 11 first- (6–12.5 wk gestation) and 10 second-trimester (13–20 wk gestation) terminations of normal, uncomplicated pregnancies, 7 third-trimester vaginal deliveries (29–40 wk gestation), and one third-trimester elective cesarean section. Three patients undergoing first- or second-trimester terminations received 1–4 mg of prostaglandin (PG) E1 analogue (either gemeprost or Cervagem; Farillon, Romford, UK) with or without a titrated i.v. infusion of oxytocin (Syntocinon; Sandoz, Surrey, UK). Three patients undergoing first-trimester medical terminations at 6–8 wk received 200 mg of the progesterone receptor antagonist mifepristone (Mifegyne; Roussel Laboratories Ltd., Uxbridge, UK), followed by 800 µg of misoprostol (Cytotec; Searle, High Wycombe, UK). Four patients who delivered in the third trimester received infusions of oxytocin to augment labor. All tissues were obtained in accordance with the requirements of the Central Oxford Research Ethics Committee.
Samples of chorionic villus attached to the decidual plate were taken for immunohistochemistry. Amniotic and chorionic membranes from third-trimester placentas were selected, and membrane rolls were prepared for immunohistochemistry as previously described [21]. Samples of chorionic villus from immediately below the decidual plate were obtained for RNA and protein analysis. All tissues were washed in endotoxin-free PBS, snap-frozen in liquid nitrogen, and stored at -80°C for subsequent immunohistochemistry, ribonuclease protection assays (RPA), or Western blotting as described below.
Immunohistochemistry
Immunohistochemical localization of G-CSFR, cytokeratin 18, or mouse IgG was achieved by the use of monoclonal antibodies (mAbs) LM832 (G-CSFR) or JMB2 (cytokeratin 18), or mouse IgG, and the alkaline phosphatase-anti-alkaline phosphatase (APAAP) method as described previously [7].
RPA
A 496-basepair (bp) BamHI/EaeI fragment of G-CSFR2145–2722 cDNA, which lacks 81 bp between nucleotide positions 2209 and 2290, was obtained from pQB3s (kindly donated by Dr. S. Nagata Osaka Bioscience Institute, Japan). The fragment, corresponding to the membrane-bound receptor lacking a 27-amino acid insertion in the cytoplasmic domain of the mature polypeptide, was cloned into the BamHI/EagI site of pBluescript 2 KS+ (Stratagene, La Jolla, CA) by the use of T4 deoxyribonuclease ligase (MBI, Sunderland, UK), generating the clone pBSG-CSFR. The sequence of the insert was confirmed using Sequenase 2 (Amersham Life Science Ltd., Little Chalfont, UK).
The G-CSFR (595 bases) and human glyceraldehyde-3-dehydrogenase (GAPDH, 226 bases) cRNA probes for use in RPA were generated by the use of T3 polymerase and the Promega Transcription Kit (Promega Corporation, Madison, WI) according to manufacturer's instructions and were purified by centrifugation through Qiaquick PCR purification columns (Qiagen, Valencia, CA).
Total RNA was isolated from 11 first-trimester, 10 second-trimester, and 8 third-trimester placental tissues according to a standard procedure [22]. Placental RNA (15 µg) was subjected to RPA using the G-CSFR and GAPDH probes and the HybSpeed RPA Kit (Ambion, Witney, UK) according to the manufacturer's instructions. The protected RNA hybrids were subjected to 6% denaturing PAGE. Densitometric analysis was performed using BioImage Intelligent Quantifier software (ImproVison, Coventry, UK). Statistical analyses were performed using the Mann Whitney U-test for nonparametric variables, and p < 0.05 was considered statistically significant.
Western Blotting
Cell membrane preparations from 11 first-, 10 second-, and 8 third-trimester frozen placentas were processed from tissue homogenates and subjected to Western blotting using LM832 and anti-actin antibody (Sigma Chemical Company, St. Louis, MO) as previously described [7]. Actin and G-CSFR were detected by the enhanced chemiluminescence (ECL) Western Blotting System (Amersham Life Science Ltd.) according to the manufacturer's instructions. Densitometric analyses of G-CSFR and actin bands was performed as above.
|
RESULTS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Immunohistochemical localization of G-CSFR in first-, second-, and third-trimester placental tissue is shown in Figure 1. Interstitial cytotrophoblast embedded in dense maternal connective tissue and decidual tissue distal and proximal to the materno-fetal interface in first- and second-trimester placental samples was positive for G-CSFR. The intensity of staining for G-CSFR on interstitial cytotrophoblast was weaker in the first trimester than in the second trimester (compare Fig. 1, a and c). Third-trimester interstitial trophoblast showed weak or absent staining for G-CSFR (Fig. 1e). Trophoblast in adjacent sections was identified by staining with antibody JMB2 (Fig. 1, b, d, and f). Adjacent sections treated with mouse IgG were negative (data not shown). Staining for G-CSFR was detected on the syncytial layer of first- and third-trimester placental tissues (Fig. 1, a and e) but not in second-trimester tissues (Fig. 1c), thus confirming our previous data [7]. The staining for G-CSFR was particularly intense on the apical surface of these cells (Fig. 1, a and e), with weak or absent staining in the underlying cytotrophoblast of first-trimester chorionic villi (Fig. 1, a and e). In contrast, G-CSFR staining on first- (Fig. 1a) and second- (Fig. 1, c and g) trimester, cytokeratin-positive interstitial trophoblast was diffuse. High-power magnification of sections of second-trimester decidual tissue showing interstitial trophoblast positive for G-CSFR and cytokeratin is shown in Figure 1, g and h, respectively. This staining pattern was similar in first-trimester interstitial trophoblast. Adjacent sections treated with mouse IgG were negative (data not shown).
|
G-CSFR Was Expressed on Undifferentiated Cytotrophoblast Cells of Term Fetal Membranes
The staining pattern obtained for G-CSFR and cytokeratin in fetal membranes is shown in Figure 2. The trophoblast cells in fetal membranes exhibited diffuse staining for G-CSFR, which, in this respect, was similar to the staining pattern observed in first- and second-trimester interstitial cytotrophoblast (Fig. 1g). The intensity of staining was similar to that observed on second-trimester interstitial trophoblast. Adjacent sections treated with mouse IgG were negative (Fig. 2c)
Expression of Placental G-CSFR mRNA and Protein Was Gestationally Regulated
Quantitation of levels of G-CSFR mRNA in placental tissue throughout gestation was achieved by RPA (Fig. 3). The expected protected RNA fragments obtained with hybridization of RNA with the 32P-labeled G-CSFR cRNA probe are shown in Figure 3c. A G-CSFR-protected fragment corresponding to the class I G-CSFR [19], which lacks the 81-bp insertion, could be detected in all placental tissues examined (Fig. 3a). Fragments corresponding to other G-CSFR transcripts were not detected. Two GAPDH fragments of 120 bp and ~125 bp were detected, and variations of band intensity indicated differences in loading. The intensity of G-CSFR and GAPDH bands in placental RNA samples was estimated by densitometry, and levels of G-CSFR mRNA were expressed as a ratio of G-CSFR band intensity to the sum of the two GAPDH band intensities as shown in Figure 3b. Levels of G-CSFR mRNA were highest in third-trimester placental samples (Fig. 3d), with a 4-fold increase in intensity with respect to levels in the first trimester (p = 0.0002). Levels of G-CSFR in second-trimester tissues were 4-fold lower than those of first-trimester tissues (p = 0.0045).
|
Western blotting of cell membrane protein preparations from first-, second-, and third-trimester placentas was performed to detect G-CSFR and actin (Fig. 4). Two distinct bands of 150 kDa and 120 kDa, which correspond to G-CSFR [23], could be detected with mAb LM832 in all first- and third-trimester placental proteins examined (Fig. 4a). These bands could also be detected in early second-trimester placental protein (13–13.5 wk); however, between 14 and 20 wk gestation, G-CSFR levels were low or undetectable (even after 30-min exposure of the filters to autoradiographic paper). There was a remarkably sharp drop in levels of G-CSFR protein between 13 and 14 wk gestation, which was particularly prominent in the samples analyzed in Figure 4a. One band of 43 kDa, which corresponds to actin, could be detected in all placental protein. The intensity of G-CSFR and actin bands was estimated by densitometry. The levels of G-CSFR protein were expressed as the ratio of the sum of the two G-CSFR band intensities to the actin band intensity as shown in Figure 4b. Levels of G-CSFR protein were 5-fold higher in third-trimester placental samples relative to first- and early second-trimester placentas (Fig. 4c; p = 0.0007).
|
The levels of the two G-CSFR immunoreactive bands were estimated by densitometry and expressed as the ratio of the 150-kDa band to the 120-kDa band (Fig. 4d). The values obtained when the ratios of the two bands were compared were consistently higher in the third trimester, with a 5-fold increase in intensity relative to first-trimester levels (p = 0.0007).
Expression of Placental G-CSFR Was Similar in Tissues from Patients Treated with Oxytocin, Prostaglandin E1 Analogues, or Mifepristone
Detection of G-CSFR by immunohistochemistry and Western blotting, using mAb LM832, in the chorionic villi of first-trimester placentas from patients treated with 200 mg of mifepristone is shown in Figure 5, a and b, respectively. Staining of G-CSFR was diffuse throughout the syncytial layer of the villi, with intense apical membrane staining (Fig. 5a), similar to the pattern obtained for first-trimester placentas obtained by surgical treatment (Fig. 1a, and [7]). Two immunoreactive bands for G-CSFR were detected by LM832 in Western blotting of placental protein from mifepristone-treated patients, similar to term placenta (Fig. 5b). The pattern of expression of placental G-CSFR, as detected by immunohistochemistry, in other tissues was similar within each gestational group, irrespective of drug regimes the patients had received (data not shown). In these respects, expression of G-CSFR was similar to that observed in first-trimester placental samples obtained from patients who had not received drugs.
|
Levels of G-CSFR protein and mRNA from placental tissue from patients who received similar drug regimes before first- and second-trimester terminations and third-trimester deliveries are shown in Figure 5c. The major factor associated with modulation of levels of both G-CSFR protein and mRNA throughout placentation was gestational age. For example, expression of both G-CSFR protein and mRNA was lower in second-trimester placental tissue from patients who had received 1 mg of the PGE1 analogue than in first-trimester placental tissue from patients receiving a similar dose. In third-trimester placental tissue, both G-CSFR protein and mRNA levels were similar in placental tissue from 1 spontaneous and 4 oxytocin-induced deliveries, and were greater than in the first trimester. Similar levels for G-CSFR protein and mRNA were detected in 5 other third-trimester spontaneously delivered placentas (data not shown).
|
DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
We detected by RPA one form of G-CSFR throughout pregnancy, corresponding to the Class I receptor [19]. The G-CSFR mRNA levels were relatively high in the first trimester of pregnancy, decreased early in the second, and increased in the third. Although it has previously been shown by reverse transcription-polymerase chain reaction that both class I and class III G-CSFR exist in placental tissue [19, 20], the mRNA corresponding to the class III receptor [19], a membrane-bound G-CSFR containing a 27-amino acid insertion in the cytoplasmic domain, was not detected by RPA. The class III receptor may be expressed at low levels in placental tissue and therefore be undetectable in the assays we used in this study. Both G-CSFR mRNAs are derived from the same gene by differential splicing [24]. This raises the possibility that the regulation of alternative splicing may be tissue type-specific, with the class III receptor specifically expressed in nonhemopoietic tissues such as the placenta.
Western blotting analyses of G-CSFR in placental membrane indicated that placental G-CSFR protein expression is modulated according to gestational age and coincides with changes in G-CSFR mRNA levels. These data suggest that expression of G-CSFR at the materno-fetal interface may be subject to transcriptional regulation. The ratio of the two distinct 120-kDa and 150-kDa G-CSFR proteins detected by Western blotting analysis changed with gestation. Both G-CSFR molecules have equal affinity for G-CSF [23]. Nine possible N-glycosylation sites exist in G-CSFR, most on the extracellular domain [20], and the 120-kDa molecule may thus be a lesser glycosylated form of the 150-kDa molecule. Alternatively, it may be a proteolytic fragment of the 150-kDa molecule as described by others [23]. Affinity cross-linking experiments have shown that a 150-kDa G-CSFR is expressed on human neutrophilic granulocytes [25], and three G-CSFR molecules of 145, 135, and 115 kDa have been detected by Western blotting of a human, G-CSF-responsive monocytic leukemic cell line [26]. Thus the 120-kDa G-CSFR may be specific to placental tissue. Although it is not known whether the structural differences in G-CSFR molecules affect receptor function, the fact that the larger form of G-CSFR is highest in third-trimester placental tissues may reflect a functional difference between first- and third-trimester placental G-CSFR.
The mechanisms for the control of G-CSFR expression in placental chorionic villi throughout gestation are unknown. In particular, there is a remarkably sharp drop in levels of G-CSFR protein between 13 and 14 wk gestation, suggesting that G-CSFR expression is strictly regulated. The levels of other cytokines and cytokine receptors in the murine and human uterus are influenced by circulating and exogenous hormones [2, 27]. Our RPA and Western blotting analyses of placental tissues from patients subjected to different drug regimes revealed that there was no correlation between the type of drugs received and levels of G-CSFR mRNA and protein. Furthermore, although second-trimester chorionic villi exhibited strict down regulation of G-CSFR expression, independent of drug treatments, interstitial cytotrophoblast in the decidual plate from the same patients continued to express G-CSFR. These data suggest that trophoblast cells are subject to cell type-specific regulation of G-CSFR expression throughout pregnancy and that this is not modulated directly by circulating levels of oxytocin, PGE1 analogues, or mifepristone.
Undifferentiated cytotrophoblast cells of term fetal membranes also express G-CSFR yet do not have invasive properties. On the contrary, trophoblast cells in fetal membranes act as a physiological barrier between the fetus and the mother [28], and until parturition there is continual exchange of hormones across this barrier, which is necessary for fetal development. In the hemopoietic system, G-CSF is necessary for neutrophilic granulocyte colony formation from bone marrow precursor cells in culture. Subsequent growth in G-CSF-depleted media leads to cell death [29], while growth in the presence of G-CSF increases cell life span considerably. Placental G-CSFR may mediate similar functions of cell maintenance in trophoblast.
In conclusion, we have demonstrated strict cell type- and developmental stage-specific expression of G-CSFR mRNA and protein during the development of the human placenta. The pattern of expression of placental G-CSFR suggests that G-CSF may have diverse functions in trophoblast invasion and maintenance during placentation, and that the regulated expression of specific G-CSFR proteins may be an important mechanism for facilitating these functions.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
2 Correspondence. FAX: 44 1865 769141; hmardon{at}molbiol.ox.ac.uk
Accepted: December 3, 1998.
Received: February 24, 1998.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  C. Ferretti, L. Bruni, V. Dangles-Marie, A.P. Pecking, and D. Bellet Molecular circuits shared by placental and cancer cells, and their implications in the proliferative, invasive and migratory capacities of trophoblasts Hum. Reprod. Update, March 1, 2007; 13(2): 121 - 141. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. Salmassi, A.G. Schmutzler, S. Schaefer, K. Koch, J. Hedderich, W. Jonat, and L. Mettler Is granulocyte colony-stimulating factor level predictive for human IVF outcome? Hum. Reprod., September 1, 2005; 20(9): 2434 - 2440. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 X Yang, F Liu, Z Xu, C Chen, X Wu, G Li, and J Li Expression of granulocyte colony stimulating factor receptor in human colorectal cancer Postgrad. Med. J., May 1, 2005; 81(955): 333 - 337. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
