help button home button Endocrine Society JCEM
HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS
Journal of Clinical Endocrinology & Metabolism, doi:10.1210/jc.2005-0014
This Article
Right arrow Abstract Freely available
Right arrow Full Text (PDF)
Right arrow All Versions of this Article:
90/9/5366    most recent
Author Manuscript (PDF)
Right arrow Submit a related Letter to the Editor
Right arrow Purchase Article
Right arrow View Shopping Cart
Right arrow Alert me when this article is cited
Right arrow Alert me when eLetters are posted
Right arrow Alert me if a correction is posted
Right arrow Citation Map
Services
Right arrow Email this article to a friend
Right arrow Similar articles in this journal
Right arrow Similar articles in PubMed
Right arrow Alert me to new issues of the journal
Right arrow Download to citation manager
Right arrow Request Copyright Permission
Citing Articles
Right arrow Citing Articles via HighWire
Right arrow Citing Articles via Google Scholar
Google Scholar
Right arrow Articles by Duncan, W. C.
Right arrow Articles by Fraser, H. M.
Right arrow Search for Related Content
PubMed
Right arrow PubMed Citation
Right arrow Articles by Duncan, W. C.
Right arrow Articles by Fraser, H. M.
Related Collections
Right arrow Female Endocrinology
The Journal of Clinical Endocrinology & Metabolism Vol. 90, No. 9 5366-5376
Copyright © 2005 by The Endocrine Society

Connective Tissue Growth Factor Expression in the Human Corpus Luteum: Paracrine Regulation by Human Chorionic Gonadotropin

W. Colin Duncan, Stephen G. Hillier, Eva Gay, Julie Bell and Hamish M. Fraser

Obstetrics and Gynecology, Department of Reproductive and Developmental Sciences, University of Edinburgh (W.C.D., S.G.H., E.G.) and Medical Research Council Human Reproductive Sciences Unit (J.B., H.M.F.), Royal Infirmary of Edinburgh-Little France, Edinburgh EH16 4SB, United Kingdom

Address all correspondence and requests for reprints to: Dr. W. Colin Duncan, Department of Reproductive and Developmental Sciences, University of Edinburgh, Royal Infirmary of Edinburgh, 49 Little France Crescent, Edinburgh EH16 4SB, United Kingdom. E-mail: w.c.duncan{at}ed.ac.uk.


   Abstract
Top
 Abstract
Introduction
 Patients and Methods
 Results
 Discussion
 References
 
Context: The molecular mechanisms of luteolysis and its inhibition during maternal recognition of pregnancy remain unclear.

Objective: The objective of this study was to investigate the differential regulation of connective tissue growth factor (CTGF) expression in human corpora lutea using in vivo and in vitro models.

Design: Corpora lutea from different stages of the luteal phase and after luteal rescue with human chorionic gonadotropin (hCG) were studied. Primary cultures and cocultures of luteinized granulosa cells and luteal fibroblast-like cells were performed.

Setting: This study was performed at the research center of a university teaching hospital.

Patients: Women with regular cycles having hysterectomy for nonmalignant conditions and women undergoing oocyte collection for assisted conception were studied.

Interventions: CTGF localization was determined by in situ hybridization, and expression by quantitative RT-PCR.

Outcomes: The outcome measures were the effect of hCG on the expression and localization of CTGF mRNA in human corpora lutea and the effect of hCG on CTGF expression in primary cultures of luteinized granulosa cells and luteal fibroblast-like cells.

Results: Luteal rescue reduced CTGF expression compared with that in the late luteal phase (P < 0.05). CTGF expression was localized to fibroblast-like cells and endothelial cells of larger blood vessels, not to steroidogenic cells. The expression of CTGF by fibroblast-like cells in vitro was not regulated by hCG. When cocultured with luteinized granulosa cells, fibroblast-like cell CTGF expression was inhibited by hCG (P < 0.001). This effect was independent of stimulated progesterone concentrations and was not blocked by follistatin or indomethacin. Both IL-1{alpha} (P < 0.05) and cAMP (P < 0.001) inhibited CTGF expression in fibroblast-like cells.

Conclusions: These results provide evidence for negative regulation of CTGF by hCG during luteal rescue mediated by paracrine signals.


   Introduction
Top
 Abstract
 Introduction
Patients and Methods
 Results
 Discussion
 References
 
THE HIGHLY VASCULAR corpus luteum is formed from the follicle, with its avascular granulosa cells, after ovulation during a period of intense tissue remodeling and angiogenesis (1, 2). In a nonconception cycle the human corpus luteum loses its functional and structural integrity to become a small fibrous avascular remnant in a matter of days. Luteolysis involves tissue remodeling with an increase in matrix metalloproteinases (3), cell death (4), vascular regression, and an influx of macrophages (5). The corpus luteum is rescued from luteolysis by human chorionic gonadotropin (hCG) during a conception cycle, and its structural and functional integrity is maintained (6). It is still not known, at a molecular level, what initiates luteolysis and how hCG, by binding to steroidogenic cell LH receptors, inhibits the underlying process.

We have developed a system for the collection of carefully dated human corpora lutea from normally cycling women (3, 5, 7). In addition, we are able to study simulated early pregnancy in women using exogenous treatment with logarithmically increasing concentrations of hCG before collection of the corpus luteum (8). This allows us to compare the early luteal phase, when the corpus luteum is forming, with the midluteal phase, when the corpus luteum is fully formed and functional. We can also compare the late luteal phase in the absence of hCG, where luteolysis is initiated, with the late luteal phase in the presence of hCG, where the corpus luteum is rescued (6).

Connective tissue growth factor (CTGF) is a heparin-binding growth factor that is a member of the CTGF/cysteine-rich 61/nephroblastoma-overexpressed family of genes. It has been implicated as a paracrine regulator of cell proliferation, angiogenesis, cellular taxis, and remodeling of the extracellular matrix (9, 10, 11). In addition, CTGF is expressed in the ovary (12, 13), where it is thought to have a paracrine role in follicular development (14). CTGF is expressed in granulosa cells of smaller follicles, and it is developmentally regulated by the actions of FSH during follicular development (15). In rodents and pigs, CTGF mRNA expression is down-regulated during FSH-induced granulosa cell maturation (12, 15). In addition, it has been suggested that granulosa cell CTGF expression is up-regulated at ovulation, and this up-regulation is associated with the intense angiogenesis seen during formation of the corpus luteum (12).

We hypothesized that if CTGF had a major role in luteal tissue and vascular remodeling, it would be differentially regulated over the life span of the corpus luteum and in the presence of exogenous hCG. In this study we report the spatio-temporal expression of CTGF mRNA in human corpora lutea and the effect of hCG on its expression in vivo and in vitro using primary cell culture models.


   Patients and Methods
Top
 Abstract
 Introduction
 Patients and Methods
Results
 Discussion
 References
 
Collection of human corpora lutea

Human corpora lutea were collected at the time of surgery from women undergoing hysterectomy for benign conditions (n = 28). All women had regular cycles and had not received any form of hormonal treatment in the 3 months before taking part in the study. Before surgery the women collected a daily early morning urine sample, and the corpora lutea were dated on the basis of the urinary LH surge as described previously (6, 7). In this study, five corpora lutea were classified as early luteal (LH+1 to LH+5), 10 as midluteal (LH+6 to LH+10), and seven as late luteal (LH+11 to LH+14). In addition, six women were given daily doubling doses of hCG (Profasi, Serono Laboratories, Welwyn Garden City, UK), starting at 125 IU, from LH+7 for 5–8 d until surgery. This regimen has been shown to rescue the corpus luteum and mimic the hormonal changes of early pregnancy (3, 8).

The collection of human corpora lutea was approved by the reproductive medicine subcommittee of the Lothian medical research ethics committee, and all women gave informed consent. At operation, the corpus luteum was quartered to ensure that each quarter contained all cellular elements of the corpus luteum. Two quarters were immediately frozen and stored at –70 C until RNA extraction was carried out. The remaining tissue was fixed in either 10% neutral buffered formalin or 4% paraformaldehyde for subsequent in situ hybridization and immunohistochemistry. In all cases an endometrial biopsy corroborated our urinary-based, tissue-dating system (16).

Collection of human luteinized granulosa cells

The reproductive medicine subcommittee of the Lothian medical ethics committee separately approved the collection of cells from patients undergoing assisted conception. With patient consent, follicular fluid was collected from women undergoing transvaginal oocyte retrieval for in vitro fertilization after ovarian stimulation using a standard procedure (17). Briefly, a long-protocol stimulated cycle was followed, using intranasal naferelin (Pharmacia Biotech, Milton Keynes, UK) for down-regulation and daily purified gonadotropins (Menopur, Ferring Pharmaceuticals, Langley, UK) for ovarian stimulation. When at least three follicles reached 18 mm in diameter, 10,000 IU hCG was administered. Transvaginal oocyte collection was performed under sonographic guidance 35 h later.

Granulosa cells were obtained from the follicular aspirates after the removal of oocytes. Individual follicles were not distinguished, and all follicular fluid from the same individual was pooled and centrifuged at 1500 rpm for 10 min. The cells were resuspended in culture medium (DMEM/Ham’s F-12 mixture, Invitrogen Life Technologies, Inc., Gaithersburg, MD), layered over a 45% Percoll/culture medium mixture (18), and centrifuged at 1200 rpm for 30 min to pellet the blood cells. Luteinized granulosa cells, visible in the interface, were collected by pipette and washed three times in PBS. The cells were resuspended in culture medium, and viable cells were counted using a trypan blue exclusion test. Eighty-thousand viable cells were plated onto each well of 24-well plates precoated with Matrigel (BD Biosciences, Bedford, MA) (17, 18) and cultured using 1 ml culture medium (DMEM/Ham’s F-12 mixture with supplements as described below) at 37 C in 5% CO2 in air.

Derivation of luteal fibroblast-like cells

To derive ovarian fibroblast-like cells, suspensions of luteinized granulosa cells were cultured in plastic flasks in DMEM/Ham’s F-12 mixture supplemented with 10% fetal bovine serum (Sigma-Aldrich Corp., St. Louis, MO) with penicillin (50 mg/liter), streptomycin (60 mg/liter), amphotericin B (2.5 mg/liter), and L-glutamine (2 mmol/liter). The medium was changed weekly, and after 6–12 wk the fibroblast-like cells had reached confluence, and the granulosa cells had disappeared. These fibroblast-like cells were then removed from the flasks using trypsin/EDTA solution, counted using trypan blue exclusion, and added to 24-well plates as described below for additional investigation.

Immunohistochemistry

The corpora lutea from each stage of the luteal phase that had been fixed in 4% paraformaldehyde for 24 h and embedded in paraffin wax (n = 12) underwent immunohistochemistry and in situ hybridization. Immunohistochemical reagents were obtained from Vector Laboratories Ltd. (Southgate, UK) or DakoCytomation Ltd. (Ely, UK). Endothelial cells were detected using a mouse monoclonal antibody to CD31 (DakoCytomation Ltd.). Five-micrometer sections were cut onto poly-L-lysine-coated slides, dewaxed, and rehydrated. The slides were then placed in a 0.01 M sodium citrate buffer, pH 6, microwaved at 450 watts for two rounds of 5 min each, and left to stand for 20 min. Tissues were permeabilized with 0.1% Triton in PBS and blocked for 1 h with normal rabbit serum diluted 1:5 with PBS and 50 mg/ml BSA. The primary antibody was applied at a concentration of 1:20 in block for 18 h at 4 C. Specific antibody binding was detected using the alkaline phosphatase-antialkaline phosphatase detection method. Rabbit antimouse antibodies diluted 1:60 in block were added for 30 min, followed by the tertiary mouse alkaline phosphatase-antialkaline phosphatase complex (diluted 1:100 in block) for an additional 30 min. Nitro blue tetrazolium was used to color regions of antibody binding blue.

Fibroblasts were detected using a mouse monoclonal antibody to human fibroblast antigen (Oncogene Research, Boston, MA). Antigen retrieval involved proteolytic digestion where sections were incubated for 30 min at 37 C in 0.1% trypsin with 0.1% calcium chloride buffered with 0.25 M Tris-HCl. Endogenous peroxidase was blocked using a 3% solution hydrogen peroxide in methanol before incubation with normal horse serum diluted 1:67 with PBS. The primary antibody was applied at a concentration of 1:15 in block and incubated for 18 h at 4 C. After washing, the slides were incubated with biotinylated horse antimouse antibodies diluted 1:200 in PBS for 30 min at room temperature. An avidin-biotin complex linked to horseradish peroxidase was added for 1 h, and antibody binding was visualized using diaminobenzidine to give a brown color. Slides then were dehydrated, mounted, and examined. Negative controls in each case consisted of similar concentrations of mouse IgG replacing the primary antibody.

Characterization of the cellular composition of primary cell cultures

To determine the cellular composition of primary cultures of luteinized granulosa cells and the fibroblast-like cells, cell suspensions from these cultures were analyzed by immunocytochemistry. Cells were prepared and cultured as described below until ready for experimentation. Cultured cells were removed from the culture wells using a trypsin and EDTA solution and resuspended in a solution of culture medium containing 10% fetal bovine serum. The cells were pelleted, washed in PBS, and fixed in 10% formaldehyde for 10 min. After centrifugation to pellet the cells, they were additionally washed in PBS and resuspended in 2% low melting point agarose prepared in PBS. The mixture was pipetted into molds and allowed to solidify at room temperature. These molds were then processed into paraffin blocks for subsequent immunocytochemistry.

Five-micrometer sections were cut onto poly-L-lysine-coated slides, dewaxed, rehydrated, and blocked with hydrogen peroxide (3%) in methanol and serum (1:5 in PBS) from the species of the secondary antibody as described above. The primary antibody was applied in serum block for 18 h at 4 C. Steroidogenic cells were identified by 3ß-hydroxysteroid dehydrogenase (3ß-HSD) immunostaining as described previously (19). Endothelial cells were identified by CD31 immunostaining as described above. Because cells were lost after antigen retrieval with proteolytic digestion, we were unable to use the fibroblast antibody described above. Fibroblasts were therefore identified by immunocytochemistry for {alpha}-smooth muscle actin ({alpha}-SMA) as described previously (19). Immune cells were identified using a mouse monoclonal antibody (1:50) to human leukocyte common antigen (LCA; CD 45, DakoCytomation Ltd). An avidin-biotin horseradish peroxidase detection system was used with diaminobenzidine detection. After dehydrating and mounting, the proportion of immunopositive cells was counted (four preparations). Negative controls were serial sections where the primary antibody had been replaced with appropriate concentrations of mouse IgG or rabbit serum (3ß-HSD).

All fibroblast-like cell cultures examined were uniform in immunostaining. They were 3ß-HSD negative, CD31 negative, {alpha}-SMA positive, and LCA negative. There was cell immunostaining heterogeneity in cultures of luteinized granulosa cells. After 6–8 d of culture, 10.1% (range, 5.4–12.3%) of cells were 3ß-HSD negative, 6.9% of cells (4.3–11.1%) were {alpha}-SMA positive, 5.7% of cells (3.4–8.3%) were CD-31 positive, and 5.95% of cells (2.7–9.3%) were LCA positive. This confirms the expected low level contamination of luteinized granulosa cells after separation with other cell types that is exploited in the derivation of fibroblast-like cells.

In situ hybridization

In situ hybridization was performed as described previously (20). The template CTGF cDNA clone corresponded to nucleotides 746-1134 of the full-length human CTGF sequence (GenBank accession no. NM_001901). Antisense and sense (used as a negative control) probes were prepared using an RNA transcription kit (Ambion, Inc., Austin, TX) and were labeled with [35S]UTP (NEN Life Science Products, Boston, MA). Deparaffinized sections (5 µm; n = 12) were treated with 0.1 N HCl and then digested in proteinase K (5 µg/ml; Sigma-Aldrich Corp.) for 30 min at 37 C. After prehybridization for 2 h at 55 C, hybridization was performed in a moist chamber overnight as described previously (20). High stringency posthybridization washings and ribonuclease (RNase) treatment were used to remove excess probe. Slides were then dehydrated, dried, and dipped in Ilford G5 liquid emulsion (Ilford Imaging, Cheshire, UK). After an exposure time of 3 wk, slides were developed (Kodak D19 developer, Eastman Kodak Co., Rochester, NY) and fixed (Kodak GBS). All slides were counterstained with hematoxylin and eosin, dehydrated, and mounted.

Preparation of cDNA from corpora lutea

Total RNA was extracted from 0.2–0.5 g frozen corpora lutea samples (n = 28) using Tri-Reagent (Sigma-Aldrich Corp.) according to the manufacturer’s instructions. The integrity of the RNA was confirmed from the absorbance at a 260/280 nm ratio and by ethidium bromide gel electrophoresis. The RNA concentration was calculated by absorbance (260 nm) measured on a GeneQuant-Pro spectrophotometer (Amersham Biosciences, Little Chalfont, UK). To eliminate genomic DNA contamination, 10 µg RNA was treated with deoxyribonuclease-I (RQ-1 RNase-free deoxyribonuclease, Promega Corp., Southampton, UK) by incubation at 37 C for 30 min, and the reaction was stopped by the addition of stop solution (20 mM EGTA; Promega Corp.) and heating to 70 C for 10 min.

RT was performed under carefully controlled conditions using a TaqMan RT kit with random hexamers (Applera UK, Warrington, UK). Samples were treated simultaneously from the same reagent mix containing 2 µl RT buffer, 2 µl deoxy-NTPs (10 mM), 1 µl each of RNase inhibitor (20 U/µl), random hexamers (50 µM), Multiscribe RT enzyme (50 U/µl), 2.4 µl MgCl2 (25 mM), and 0.2 µg deoxyribonuclease-treated RNA in a final sample volume of 20 µl. RT was performed at room temperature for 10 min and at 42 C for 60 min, and was stopped by incubation at 95 C for 10 min. The resulting cDNA was stored at –20 C until use. An equal amount of human placental total RNA (1 mg/ml; Cambridge Bioscience, Cambridge, UK) was concomitantly reverse transcribed to provide cDNA for the generation of standard calibration curves.

Investigation of CTGF expression in primary cell culture

To identify candidate molecules in the regulation of CTGF, a review of the literature on potential regulators of skin CTGF expression during wound healing, concentrations of candidate molecules, and known luteal products was undertaken. To assess the effects of these candidate molecules on fibroblast-like cell CTGF expression, 80,000 cells were added to each well of 24-well plates. After 24 h in serum-free conditions, fresh serum-free medium was added containing 1) progesterone (Sigma-Aldrich Corp.; 0.1, 1, and 10 µM and control), 2) activin A (R&D Systems, Inc., Abingdon, UK; 10, 50, and 100 ng/ml and control) (21), 3) IL-1{alpha} (R&D Systems, Inc.; 5, 50, and 500 pg/ml and control) (22), and 4) dibutyryl cAMP (VWR International Ltd., Lutterworth, UK; 0.5, 1, and 5 mM and control) (23). All controls contained the carrier solution equivalent to the highest concentration added. After 24 h, cells were collected for mRNA extraction. Two to four replicates were performed in up to five separate experiments.

Luteinized granulosa cells were cultured in medium supplemented with glutamine (2 mmol/liter), insulin (6.25 mg/liter), transferrin (6.25 mg/liter), selenious acid (6.25 µg/liter), amphotericin (2.5 mg/liter), penicillin (50 mg/liter), and streptomycin (60 mg/liter) as described previously (18). Medium was changed every 2–3 d over the course of the culture period. For the time-course experiment, cells were grown for 11 d in the presence or absence of hCG (10 ng/ml; Serono) (24) over three replicates, and mRNA was extracted on d 2, 5, 8, and 11. The dose of hCG was equivalent to the lowest dose giving maximal stimulation of progesterone in pilot dose-response experiments (24).

For coculture experiments, approximately 40,000 fibroblast-like cells were added to the wells containing the luteinized granulosa cells after culture for 6–7 d. This was based on pilot experiments showing this time to be optimal in terms of the progesterone response to hCG treatment and this cell mix to be optimal for the detection of changes in CTGF expression. After coculture for 24 h in serum-free medium, fresh serum-free medium was added containing either hCG (10 ng/ml) or the carrier control. After 24 h, mRNA was extracted from cocultures, and controls consisting of fibroblasts were added to wells without granulosa cells and with granulosa cells with no fibroblasts added. Each coculture and control treatment was replicated eight times in two to four separate experiments.

To assess the effects of candidate molecules in cocultures, the experiments were repeated with the following additions: 1) low-density lipoprotein (LDL; 50 mg/liter; Sigma-Aldrich Corp.), hCG (10 ng/ml) and LDL (50 mg/liter), hCG (10 ng/ml), and hCG (10 ng/ml) and aminoglutathamide (100 µM; Sigma-Aldrich) (25); 2) hCG (10 ng/ml) and control with no additions; 3) hCG (10 ng/ml) and control with follistatin (100 ng/ml; R&D Systems, Inc.) (26); and 4) hCG (10 ng/ml) and control with indomethacin (10 µM; Sigma-Aldrich Corp.) (27). After 24 h, medium was collected for progesterone measurement, and cells were collected for mRNA extraction. The dose of the inhibitor used was that shown to have maximal inhibition in cell culture experiments in other systems (25, 26, 27).

Measurement of progesterone

Progesterone concentrations in the culture medium collected were measured using a plate modification of a standard in-house progesterone RIA. This assay had an intraassay coefficient of variation of less than 4%, an interassay coefficient of variation of less than 11%, and a detection limit of 0.1 nmol/liter.

Preparation of cDNA from cultured cells

After removal and storage of culture medium, cells were rinsed in PBS and Tri-Reagent was added. The resulting solution was stored at –70 C until batch extraction of RNA was carried out. RNA was extracted following the manufacturer’s instructions. To remove contaminating genomic DNA, RNA was treated with deoxyribonuclease I at a concentration of 1 U/µg RNA for 30 min at 37 C. After stopping the reaction with stop solution, the samples were heated to 70 C for 10 min. Using random hexamers, 200 ng RNA was reverse transcribed in a solution containing 5.5 mM MgCl2, 2.5 µM random hexamers, 500 µM of each deoxy-NTP, 0.4 U/µl RNase inhibitor, and 1.25 U/µl Multiscribe reverse transcriptase (Applied Biosystems, Warrington, UK). Samples were incubated at room temperature for 10 min, followed by 42 C for 60 min, and 95 C for 10 min. Two controls were used: one omitted the Multiscribe enzyme, and the other omitted the template RNA.

Primer design

Oligonucleotide PCR primers for each gene investigated were designed using Primer3 software (www-genome.wi.mit.edu/cgibin/primer/primer3_www.cgi) from DNA sequences obtained from GenBank (www.ncbi.nlm.nih.gov). To minimize amplification of genomic DNA, the forward and reverse primers for each gene were chosen from different exons where possible. The sizes of PCR products were designed to be less than 250 bp to optimize the RT-PCR quantification. Primers were synthesized by MWG-AG Biotech (Milton Keynes, UK), and all products were sequenced using ABI PRISM Big Dye Terminator Sequencing Kits (Applied Biosystems) to confirm the nature of the fragment. The 5' to 3' primer sequences were: CTGF: forward, TCCCAAAATCTCCAAGCCTA; and reverse, GTAATGGCAGGCACAGGTCT; glucose-6-phosphate dehydrogenase (G6PDH): forward, CGGAAACGGTCGTACACTTC; and reverse, CCGACTGATGGAAGGCATC; LH receptor: forward, ATGAAGCAGCGGTTCTCG; and reverse, TTGACAGGGAGGTAGGCAAG; 3ß-HSD: forward, CCATGAAGAAGAGCCTCTGG; and reverse, GTTGTTCAGGGCCTCGTTTA; progesterone receptor (A and B): forward, TGGAAGAAATGACTGCATCG; and reverse, TAGGGCTTGGCTTTCATTTG; and vascular endothelial growth factor receptor-2 (VEGF-R2): forward, GTTATCCAAGCGGCAAATGT; and reverse, AAAGACACGCTCTCCTGCTC.

PCR

One microliter of the cDNA solution from cultures of fibroblast-like cells and luteinized granulosa cells was used as the template in subsequent optimized PCRs, which were carried out using Taq DNA polymerase (Promega Corp.). Reaction conditions were 0.25 U/µl Taq, 0.5 µM of each primer, 0.2 mM deoxy-NTPs, 0.1% Triton X-100, 50 mM KCl, 10 mM Tris-HCl, and 1.5 mM MgCl2. Each PCR was overlaid with mineral oil and incubated at 95 C for 5 min, followed by: for G6PDH and LH receptor, 35 cycles of 95 C for 30 sec, 62 C for 30 sec, and 72 C for 90 sec, followed by 72 C for 10 min; for 3ß-HSD, 25 cycles of (and for CTGF, 30 cycles of) 95 C for 30 sec, 59 C for 30 sec, and 72 C for 90 sec, followed by 72 C for 10 min; and for progesterone receptors, 35 cycles of 94 C for 30 sec, 57 C for 30 sec, and 72 C for 60 sec, followed by 72 C for 5 min. PCR products were separated by applying 100 V for 75 min to 1% agarose gels with ethidium bromide and were visualized under UV transillumination.

Quantitative analysis of gene expression by RT-PCR

Each assay was optimized using PCR amplification of either human placental cDNA or previously extracted and characterized human luteal cDNA. The assays were optimized for annealing temperature and MgCl2 concentration. PCR amplifications were performed using Thermostart Taq (AB Gene) in a DNA Engine gradient cycler (MJ Research, Inc., Watertown, MA). Eight PCRs, using the gradient facility of the cycler, were performed to demonstrate the PCR effectiveness at annealing temperatures between 52 and 68 C. Products were examined by gel electrophoresis to confirm the presence of a single band at the correct size and to confirm the optimal conditions. Magnesium concentrations were optimized on the light cycler between 3 and 5 mM. A melting curve analysis performed after the amplification step allowed optimization of the temperature used to quantify the level of gene expression by minimizing background fluorescence from nonspecifically amplified DNA. All assays exhibited a single DNA melting curve peak.

Quantitative real-time PCR was performed in duplicate 10-µl reaction volumes on the LightCycler using the Master Mix supplied with the LightCycler Fast Start DNA Master SYBR Green 1 kit (Roche, Lewes, UK) at the optimized conditions. A standard curve was generated with serial dilutions of standardized cDNA using the second derivative maximum method provided with the LightCycler software (version 3.3). PCR product concentrations were automatically calculated for each tissue sample by the software by comparing the sample threshold cycle to the standard curve. In all cases the level of gene expression within the samples lay within the boundaries of the corresponding standard curve. All samples were analyzed in duplicate, and assay variation was typically within 10%.

Data were normalized according to the expression level of G6PDH, determined in duplicate by reference to a serial dilution calibration curve generated for each sample using the standard LightCycler software. PCR products were extracted from the LightCycler capillaries and analyzed on 2% agarose gels to confirm product size and reaction specificity.

Statistical analysis

Statistical analysis was carried out using ANOVA after confirmation of homogeneity of variance. Where significant differences (P < 0.05) were detected, pairwise comparisons were carried out using the Bonferroni test. Coculture expression of CTGF, LH receptor, and VEGF-R2 was compared with predicted levels using the Mann-Whitney U test. The statistical test used is shown after each P value in Results.


   Results
Top
 Abstract
 Introduction
 Patients and Methods
 Results
Discussion
 References
 
Expression of CTGF in human corpora lutea

CTGF mRNA is expressed in human corpora lutea at all stages of the luteal phase (Fig. 1Go). Quantitative analysis of CTGF mRNA expression using real-time RT-PCR showed no significant differences across the normal luteal phase, although there was a trend to increasing CTGF expression from the early to the late luteal phase (P = 0.098, by ANOVA). Luteal rescue with exogenous hCG, however, resulted in a decrease in CTGF expression compared with the mid (P < 0.05, by ANOVA) and late (P < 0.05, by ANOVA) luteal phases (Fig. 1Go). In this model system, where hCG rescues the corpus luteum from luteolysis and mimics the hormonal changes of early pregnancy (6), hCG inhibits CTGF expression.



View larger version (33K):
[in this window]
[in a new window]
  		FIG. 1. Expression of CTGF mRNA in human corpora lutea using real-time quantitative RT-PCR. CTGF expression was determined in early-luteal (Early; LH+1 to LH+5; n = 5), midluteal (Mid; LH+6 to LH+10; n = 10), and late-luteal (Late; LH+11 to LH+14; n = 7) corpora lutea and in corpora lutea rescued with exogenous hCG (Resc; hCGx5 to hCGx8; n = 6). CTGF expression varied in these groups (P < 0.05, by ANOVA). The expression of CTGF after luteal rescue was decreased compared with that in the mid- and late-luteal phases (*, P < 0.05, by ANOVA).

 
Cellular localization of CTGF expression in human corpora lutea

The cellular localization of CTGF mRNA expression was assessed using isotopic in situ hybridization. The pattern of expression was similar in all corpora lutea examined, and no specific expression was detected using the sense riboprobe as a negative control. Specific CTGF expression was particularly localized to cells other than the granulosa-lutein cells at all stages of the luteal phase (Fig. 2Go). CTGF mRNA was highly expressed in individual cells. These cells were localized in the vicinity of blood vessels, in the radial spurs of connective tissue running through the steroidogenic cells and at the edge of the clot-filled, central cavity (Fig. 2Go, A and B).



View larger version (102K):
[in this window]
[in a new window]
  		FIG. 2. Localization of CTGF mRNA in sections of human corpora lutea. A, Light field of a late-luteal corpus luteum showing the antral cavity (A), granulosa-lutein cells (G), and the fibrous invagination containing blood vessels (bv) after isotopic mRNA in situ hybridization for CTGF. B, Dark field of same slide showing the bright grains of hybridization to be absent from the luteinized granulosa cells (G), but present in individual cells associated with blood vessels (bv) and neighboring cells and within the clot-filled antral cavity (A). C, Serial section of the same corpus luteum immunostained for CD31 to highlight endothelial cells (Endo). These can be seen within the luteinized granulosa cell layer (G), but not in the antral cavity. Although there are some areas of similarity of expression, not all endothelial cells express CTGF, and cells that are not endothelial cells express CTGF. D, Additional serial section immunostained for fibroblasts. These can be seen within the antrum (A) and around the vasculature in the granulosa-lutein cell layer (G). The antral fibroblasts show a similar pattern as the CTGF expression (Fib) and are likely to be the cells expressing CTGF in the antrum. Not all fibroblasts express CTGF; the perivascular ones in smaller capillaries (Vasc Fib) do not appear to express CTGF. E, A higher power view of a larger blood vessel in a different late-luteal corpus luteum showing expression of CTGF mRNA in light field. The grains can be clearly localized to endothelial cells (EC) of a larger vessel containing red blood cells (RBC) and to individual cells in the vicinity of the vessel (arrows). F, High power view of a midluteal corpus luteum showing CTGF expression around a larger blood vessel (BV) and not in the steroidogenic cells (G). G, Low-power light field of an early-luteal corpus luteum after in situ hybridization for CTGF. CTGF expression (dark spots) can be seen in the surrounding stroma (S), but not in the majority of the granulosa-lutein cell layer (G). There is intense expression in individual cells around the antrum (A; arrows), mirroring the antral expression seen in later corpora lutea. Scale bars: A–E, 50 µm; F, 25 µm; G, 100 µm.

 
The nature of the cells expressing CTGF was also investigated with serial sections immunostained for endothelial cells and fibroblasts (Fig. 2Go, C and D). The endothelial cells of the microvasculature within the steroidogenic cell layer tended not to express high levels of CTGF (Fig. 2CGo). However, CTGF expression was localized to the endothelial cells of larger vessels in the corpus luteum (Fig. 2Go, B, C, E, and F). CTGF expression was localized to fibroblasts in the antral cavity (Fig. 2Go, B and D) as well as some fibroblasts associated with the larger blood vessels (Fig. 2Go, B, D, and E). In addition, CTGF expression could be localized to fibroblasts in the peripheral stromal tissue surrounding the steroidogenic cell layer (Fig. 2GGo). This localization was consistent in corpora lutea examined at different stages of the luteal phase (Fig. 2GGo). Because LH/hCG receptors are localized to the steroidogenic cells of the corpus luteum (7), and the main cellular source of CTGF does not appear to be these cells, hCG inhibition of CTGF expression during luteal rescue may be through paracrine signaling.

Expression of CTGF in luteal cell populations

To establish the nature of the hypothesized paracrine control of CTGF expression, a primary cell culture model was developed. Luteinized granulosa cells were obtained from women undergoing oocyte recovery as part of an assisted conception program. These cells tend to express LH/hCG receptors and 3ß-HSD (Fig. 3Go) rather than CTGF (Fig. 3Go). Fibroblast-like cells were found to be a low-level contaminant of luteinized granulosa cells collected at oocyte retrieval. With prolonged culture the luteinized granulosa cells disappeared, and the fibroblast-like cells grew to confluence. These fibroblast-like cells tended to express CTGF and not LH/hCG receptors or 3ß-HSD (Fig. 3Go). Both luteinized granulosa cells and these fibroblast-like cells expressed genomic progesterone receptors (Fig. 3Go).



View larger version (45K):
[in this window]
[in a new window]
  		FIG. 3. Expression profile of various mRNA species using RT-PCR in luteinized granulosa cells after 7 d of culture (G) and in ovarian fibroblast-like cells after 6 wk of culture (F). Both cell types express the housekeeping gene G6PDH. Only the fibroblast-like cells express CTGF, whereas only the luteinized granulosa cells express LH receptors (LHR) and 3ß-HSD. Both cell types express genomic progesterone receptors (PR). Each PCR product gave a band of the expected size, and the specificity of the product was confirmed by sequencing.

 
Effect of hCG on CTGF expression in culture

CTGF expression by fibroblast-like cells was not affected by exposure to hCG in vitro (Fig. 4Go, A and B). However, CTGF expression in cocultures of fibroblast-like cells and luteinized granulosa cells was inhibited (P < 0.001, by ANOVA) by hCG (Fig, 4Go, A and B). This suggested that cocultures could function as a model system to investigate the paracrine regulation of CTGF expression.



View larger version (51K):
[in this window]
[in a new window]
  		FIG. 4. Expression of CTGF in cultures and cocultures of luteinized granulosa cells and ovarian fibroblast-like cells. A, Gel after RT-PCR showing CTGF band (confirmed by sequencing) in fibroblast-like cells in the absence (Fib; lanes 1 and 2) and presence of 10 ng/ml hCG (Fib+hCG; lanes 3 and 4) showing similar band intensities. Also shown is the expression in cocultures of fibroblast-like cells and luteinized granulosa cells in the absence (Co-Cult; lanes 5 and 6) and presence of 10 ng/ml hCG (Co-Cult+hCG). There is a suggestion of lower band intensity in the presence of hCG. Lane 9 is the marker ladder. B, Quantification of CTGF expression after the same treatments using quantitative RT-PCR standardized against G6PDH. There is no effect of hCG on fibroblast-like cell CTGF expression, but hCG inhibits CTGF expression in cocultures (P < 0.001, by ANOVA). CTGF expression in cocultures has been standardized to the expression in fibroblast-like cells. C, CTGF expression in luteinized granulosa cells (LGC) in the absence and presence (LGC+hCG) of 10 ng/ml hCG. These arbitrary units are the same as the fibroblast-like cell expression in B, showing expression approximately 10-fold lower. There is an inhibition of CTGF expression by hCG (P < 0.001, by ANOVA) to a magnitude similar to that in cocultures. Significant differences are indicated by asterisks.

 
However, analysis of CTGF expression using quantitative RT-PCR could detect low levels of expression in luteinized granulosa cells that were inhibited (P < 0.001, ANOVA) by hCG to a similar magnitude as in cocultures (Fig. 4CGo). This was not reliably detected by the less sensitive gel analysis (Fig. 3Go). This effect could be due to the low level of fibroblast contamination known to be in luteinized granulosa cells, or the luteinized granulosa cells themselves may themselves express low levels of CTGF in culture.

The effects of duration of culture and hCG on CTGF expression in luteinized granulosa cells were therefore examined. CTGF expression tended to increase with duration of culture (P < 0.05, by ANOVA). By d 11, the expression of CTGF was very variable and not statistically different from that in fibroblast-like cell cultures. The inhibitory effect of hCG had lost statistical significance by d 11. On d 2 and 5 of culture, CTGF expression was much lower than fibroblast-like cell expression (P < 0.0001, by ANOVA), and this was inhibited by hCG (P < 0.01, ANOVA). On d 8, there was greater CTGF expression than on d 5 (P < 0.05, by ANOVA), but it was markedly lower than in fibroblast cultures (P < 0.001, by ANOVA) and was inhibited by hCG (P < 0.05, by ANOVA). These results are consistent with the effect being through contaminating fibroblasts that increase in proportion as they divide and the luteinized granulosa cells are lost. However, it did not exclude a contribution of CTGF expression from both fibroblast-like cells and luteinized granulosa cells that could be affected by hCG.

Paracrine control of CTGF expression in coculture

Experiments were therefore designed to investigate the amount of inhibition of fibroblast-like cell CTGF expression by hCG in cocultures and whether the addition of luteinized granulosa cells could inhibit fibroblast-like cell CTGF expression in the absence of hCG. We hypothesized that the contribution of each cell type to CTGF expression in coculture experiments could be calculated from the quantified expression in each cell type separately and the cellular composition of the coculture. A predicted value was calculated for the expression in coculture based on this paradigm and tested using a molecular marker predominantly expressed in the fibroblast-like cells (VEGF-R2; Fig. 5AGo) and one predominantly expressed in luteinized granulosa cells (LH receptor; Fig. 5BGo). Both of these markers gave the predicted values when tested (Fig. 5Go, A and B), lending validity to this method of investigation.



View larger version (22K):
[in this window]
[in a new window]
  		FIG. 5. Expression of mRNA species by quantitative RT-PCR in cultures of fibroblast-like cells (F), luteinized granulosa cells (G), and cocultures of these cells (FG; in the ratio 1F:2G). P, Predictive expression in cocultures derived from the sum of the expression in individual cell types corrected for the proportion of cells in the coculture. A, VEGF-R2 is primarily expressed by fibroblast-like cells (P < 0.0001, by ANOVA), but cocultures show the expected expression (P > 0.05, by Mann-Whitney U test). B, LH receptors (LHR) are primarily expressed by luteinized granulosa cells (P < 0.0001, by ANOVA), but cocultures show the expected expression (P > 0.05, by Mann-Whitney U test). C, CTGF is primarily expressed by fibroblast-like cells (P < 0.0001, by ANOVA), but cocultures show less than half the expected expression in the absence of hCG (P < 0.05, by Mann-Whitney U test). D, CTGF expression in the presence of hCG (10 ng/ml), where the predicted expression in cocultures is 10 times less than anticipated (P < 0.01, by Mann-Whitney U test). This suggests that fibroblast-like cell CTGF expression is inhibited in the presence of luteinized granulosa cells and hCG in cocultures.

 
Using this paradigm it was shown that the addition of luteinized granulosa cells is associated with an inhibition of fibroblast-like cell CTGF expression (P < 0.05, by Mann-Whitney U test; Fig. 5CGo) and that hCG further decreased the expected expression (P < 0.01, by Mann-Whitney U test; Fig. 5DGo). It is therefore likely that luteinized granulosa cells inhibit fibroblast-like cell CTGF expression in a paracrine manner, and this inhibition is increased by hCG.

Regulation of CTGF expression in fibroblast-like cells

Because CTGF expression in ovarian fibroblast-like cells appears to be regulated, a candidate molecule approach was used to identify potential paracrine regulators of its expression. Candidate molecules and concentrations examined were determined by examination of the literature. Ovarian fibroblast-like cell CTGF was stimulated by activin A (Fig. 6AGo; P < 0.05, by ANOVA) and inhibited by IL-1{alpha} (Fig. 6BGo; P < 0.05, by ANOVA) and the cAMP pathway (Fig. 6DGo; P < 0.001, by ANOVA). There was no clear dose-response effect of progesterone on fibroblast-like cell CTGF expression (Fig. 6CGo). However, because there seemed to be a reduction in CTGF expression when 10 µM progesterone was compared with 0.1 µM progesterone (Fig. 6CGo), a paracrine effect of progesterone on the regulation of CTGF in this model system could not be excluded.



View larger version (36K):
[in this window]
[in a new window]
  		FIG. 6. The effects of candidate molecules on the expression of CTGF in fibroblast-like cells in culture, as measured by quantitative RT-PCR standardized against G6PDH expression. Activin A significantly increased CTGF expression (P < 0.05, by ANOVA; at a concentration of 50 ng/ml). IL-1{alpha} treatment resulted in a dose-responsive inhibition (P < 0.05, by ANOVA; at 50 and 500 pg/ml). There were no significant differences in CTGF expression from controls (C) after progesterone treatment (by ANOVA). However, treatment with 10 µM progesterone resulted in lower CTGF expression than 0.1 µM progesterone (P < 0.01, by ANOVA). Treatment with dibutyryl cAMP inhibited CTGF expression at all concentrations examined (P < 0.001, by ANOVA; at 0.5 mM or greater concentrations). Significant differences are indicated by asterisks.

 
Progesterone and paracrine regulation of CTGF

Because progesterone is an hCG-dependent secreted product of granulosa-lutein cells (24), and the fibroblast-like cells express genomic progesterone receptors (Fig. 3Go) (19), it remained an attractive candidate as a paracrine-signaling molecule. We therefore devised a system in which the progesterone concentrations in the presence of hCG could be manipulated (24) (Fig. 7Go). In the presence of hCG, CTGF expression was similarly inhibited in the presence of high (hCG plus LDL; P < 0.001, by ANOVA), intermediate (hCG; P < 0.001, by ANOVA), and control (hCG plus aminoglutathamide; P < 0.001, by ANOVA) concentrations of progesterone (Fig. 7Go). The hCG-stimulated product from luteinized granulosa cells that regulates CTGF expression in fibroblast-like cells does not appear to be progesterone.



View larger version (23K):
[in this window]
[in a new window]
  		FIG. 7. CTGF expression in cocultures of luteinized granulosa cells and fibroblast-like cells after 24 h, as measured by quantitative real-time RT-PCR. Treatment with hCG resulted in a significant reduction in CTGF expression compared with control cultures in all treatment groups examined (P < 0.001, by ANOVA). Treatment groups were designed to give different progesterone concentrations (P < 0.01, by ANOVA): hCG plus LDL gave high progesterone concentrations, hCG alone gave medium progesterone concentrations, and hCG plus aminoglutathamide gave control levels of progesterone. The effect was the same regardless of the progesterone concentration. Different letters represent significant differences.

 
Other molecules and paracrine regulation of CTGF expression

Coculture experiments designed to neutralize the effects of prostaglandins (PGs) and activin were carried out (Fig. 8Go). The inhibition of CTGF expression by hCG in cocultures was not altered in the presence of large concentrations of follistatin or indomethacin (Fig. 8Go).



View larger version (28K):
[in this window]
[in a new window]
  		FIG. 8. The effects of excess follistatin (100 ng/ml) and indomethacin (10 µM) on CTGF expression in cocultures of luteinized granulosa cells and fibroblast-like cells in the presence or absence of hCG (10 ng/ml). hCG inhibited CTGF expression (P < 0.001, by ANOVA), and this inhibition was not altered by the addition of follistatin or indomethacin (by ANOVA). Different letters indicate significant differences.

 

   Discussion
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
References
 
These experiments describe the expression and localization of CTGF in the human corpus luteum. We determined that CTGF mRNA is expressed in the human corpus luteum and is down-regulated by hCG during luteal rescue. This suggests that CTGF has a role in luteal development and that it may be one of the factors involved in luteolysis.

There is already evidence that CTGF may be involved in the regulation of ovarian function, because rat granulosa cell CTGF expression is rapidly and markedly suppressed by FSH (13). In rat (14), pig (12), and mouse (15) ovaries, CTGF is expressed in granulosa cells of the late preantral and antral follicles, and its expression is down-regulated in preovulatory follicles (12, 14), coinciding with FSH-induced granulosa cell maturation (15). Because CTGF is up-regulated at the time of ovulation in rat and pig ovaries (12, 14) and can be detected in human follicular fluid (28), it may have a role in the formation of the corpus luteum. Indeed, as well as regulating collagen deposition in fibroblasts, CTGF has been shown to have mitogenic and chemotactic roles during angiogenesis and tissue remodeling (9, 10, 29). Because CTGF continues to be expressed in the fully formed corpus luteum in the midluteal phase, its role is not restricted to luteal formation. Indeed, CTGF seems to have a particular role in the late luteal phase, because it tends to be maximal in this phase and is inhibited by hCG during luteal rescue in early pregnancy. Because luteolysis is associated with fibrosis and remodeling, and CTGF is involved in these processes during wound healing (9) [and can regulate the production or activity of matrix metalloproteases (29)], it is likely that CTGF is involved in its regulation.

We were able to localize the expression of CTGF mRNA in human corpora lutea. CTGF was expressed, in all stages of the luteal phase, in scattered individual cells. These cells appear to represent a proportion of luteal endothelial cells and fibroblasts. In rat and pig ovaries, CTGF expression in the developing corpus luteum was primarily detected in endothelial cells of the developing vasculature and in cells in the stroma and antral cavity (12, 14). It was believed that CTGF-expressing cells migrate into the corpus luteum shortly after ovulation (12). This localization is not unexpected, because CTGF is expressed by fibroblasts, vascular smooth muscle cells, and endothelial cells in other tissues (9, 30). Indeed, it can also be detected in cultured human endothelial cells and fibroblasts. It is hypothesized that CTGF acts on the same cells that produce it in a paracrine or autocrine manner (9).

We detected little CTGF mRNA expression in granulosa-lutein cells of the corpus luteum. In the rat, expression was also largely absent from these cells during luteal development (14). The same pattern is seen in the pig ovary during the formation of the corpus luteum (12). This is consistent with granulosa cell CTGF expression being inhibited during steroidogenic cell differentiation and maturation through the increased stimulation of intracellular cAMP. However, unlike what was reported in the rat and what we found in women, CTGF expression in the pig granulosa cells appears to be present in the midluteal phase at the time of maximal luteal function (12). We did not see this in any of the human corpora lutea examined throughout the luteal phase. However, there was a trend for CTGF to increase after prolonged culture of luteinized granulosa cells. Such expression, in the form of IGF-binding protein-related protein-2 (9, 31), has been previously noted in human luteinized granulosa cells cultured in serum (32). However, it is likely that this represents increasing proportions of contaminating fibroblasts rather than increasing steroidogenic cell expression, because we have demonstrated contaminants in luteinized granulosa cell cultures. Regardless, if there is CTGF expression in functional luteinized granulosa cells, it is at low levels consistent with the nonsteroidogenic cells being the major source of CTGF in the functional human corpus luteum. It would be interesting, however, to investigate CTGF expression in regressing steroidogenic cells of menstrual corpora lutea and corpora albicans.

One interesting observation that is consistent in all species examined is the marked periantral punctate expression of CTGF at the time of follicular rupture and luteal development (12, 13, 14). We have not been able to fully characterize the nature of these cells, because not all of these cells stain with endothelial cell or fibroblast antibodies. Whether these are specialized granulosa cells or less differentiated cells that act like endothelial or fibroblast precursors is not clear. However, it is likely that these cells are involved in the tissue and vascular remodeling seen in luteal development. We hypothesize that these CTGF-expressing cells are involved in the neovascularization of the corpus luteum.

Because the steroidogenic cells that express LH/hCG receptors are not the primary source of CTGF in the human corpus luteum, it was not clear how hCG inhibited CTGF expression. Using a novel coculture model, we found that hCG also inhibited fibroblast-like cell CTGF expression in vitro and that this effect was mediated indirectly by as yet unidentified paracrine molecules from luteinized granulosa cells.

One of the key components of the in vitro model is fibroblast-like cells. Because luteinized granulosa cells are terminally differentiated, with a limited life span, these cells can only be cultured for short periods of time before they disappear from culture. At oocyte retrieval, more than one cell type is present in follicular fluid. As well as luteinized granulosa cells, there are fibroblasts, red blood cells, white blood cells, ovarian surface epithelial cells, and endothelial cells. As we have demonstrated, not all of these cells are removed by the purification process; indeed, endothelial-like cells can be grown out of luteinized granulosa cell cultures (33). Contaminating fibroblast-like cells will divide in the presence of serum until they populate the cultures. Although the exact nature of these cells is not clear, they have the appearance of fibroblasts rather than endothelial cells (33) or surface epithelial cells in culture (34). Daily inspection of the cultures gave the impression that these cells were present in initial cultures and were selected for during prolonged culture rather than being derived from luteinized granulosa cells. Whatever their nature, these cells behave like fibroblasts in culture and have the expression profile expected of fibroblasts for the genes examined.

This model system has limitations. Although luteinized granulosa cells respond to hCG in culture, they may be different from the granulosa-lutein cells of the corpus luteum. Similarly, the relationship of fibroblast-like cells to luteal CTGF-secreting cells remains uncertain. However, there are clear similarities with luteal steroidogenic cells and fibroblasts in their expression profile. In addition, our model system reveals a clear paracrine effect on CTGF expression in vitro similar to that determined in vivo. Whether this paracrine effect is the same as that in the corpus luteum remains to be determined, but in the absence of better experimental systems in the human, such modeling may facilitate the development and testing of hypotheses about the paracrine regulation of luteal function.

Using this model, we confirmed the hypothesis that gene expression in coculture should be a sum of the expression in one cell type plus the expression in another cell type using cocultures and analyzing genes expressed highly in one or another cell type. This allowed us to determine that luteinized granulosa cells inhibited CTGF expression in fibroblast-like cells, and this inhibition was increased by hCG. This is not surprising, because these cells function, secreting molecules such as progesterone and VEGF (24), in the absence of hCG. It is likely that paracrine regulatory molecules are basally secreted, and this secretion is also regulated by hCG (24).

The nature of the hCG-induced factor from luteinized granulosa cells that inhibits fibroblast-like cell CTGF expression is not clear. We tested candidate molecules identified from the existing literature. There is evidence that steroids can affect CTGF expression. In the mouse ovary, androgens augment FSH-induced inhibition of CTGF expression, possibly mediated by the inhibition of cAMP catabolism (15). Dexamethasone was found to be a potent inducer of CTGF expression in cultured fibroblasts (35). In addition, CTGF is found in uterine fluids and shows cyclical changes (36). CTGF expression is stimulated by estradiol and inhibited by progesterone in the uterus of ovariectomized mice (37). We did not, however, find any evidence for progesterone regulation of luteal CTGF expression. Although we cannot exclude a role for progesterone, because it is still present in control cultures, it is clear that the down-regulation of CTGF mRNA by hCG in primary cell cocultures is independent of progesterone concentrations.

There are other candidate molecules that may be paracrine regulators of CTGF expression. It is possible that hCG may inhibit the production of a luteinized granulosa cell product that stimulates CTGF expression. CTGF is strongly induced by TGF-ß, and activin can up-regulate CTGF expression (15, 30). We have shown that activin up-regulates ovarian fibroblast-like cell CTGF expression but have failed to demonstrate any effect on follistatin in cocultures. PGE2 increases CTGF expression in rat smooth muscle cultures expressing the EP4 receptor (38). PGs, such as PGE2 through the EP2 receptor (38, 39), and Iloprost, an IP prostanoid agonist, can inhibit CTGF (40). Indomethacin, however, did not alter CTGF expression in cocultures. In addition, keratinocytes can inhibit skin fibroblast CTGF expression by paracrine molecules, and this is not affected by the presence of indomethacin (27). The marked effect of cAMP in culture, however, does suggest that anything stimulating cAMP formation in fibroblasts will down-regulate CTGF expression (29).

The factors that are involved in the inhibition of skin fibroblast CTGF expression by keratinocytes in culture are soluble and stable (27). An analysis of the cytokines secreted showed no involvement of TNF-{alpha}, but suggested that IL-1{alpha} was present. Subsequent treatment with IL-1{alpha} neutralizing antibodies negated the inhibition of CTGF by the keratinocyte-conditioned medium (27). We have confirmed that IL-1{alpha} can inhibit ovarian fibroblast-like cell CTGF expression. This and the fact that IL-4 attenuates TGF-ß-stimulated induction of CTGF mRNA (41) mean that cytokines are clear targets for additional investigation and manipulation. TNF-{alpha} can inhibit CTGF expression (42, 43). Although IL-1 is secreted by luteinized granulosa cells (44), model systems have suggested that its production may be inhibited by gonadotropins (45). Additional investigation of the paracrine role of cytokines is therefore indicated.

In summary, we have shown that CTGF is present in the human corpus luteum and have suggested roles for CTGF in the development of the luteal vasculature and the tissue remodeling associated with luteolysis. It is inhibited by hCG during maternal recognition of pregnancy, and this inhibition appears to be mediated by paracrine factors. We have developed a coculture system to investigate the effect of hCG on CTGF expression and confirmed that the change is mediated by paracrine factors. Although there are attractive candidate molecules for these paracrine factors, testing these molecules in culture has not clearly revealed the nature of the paracrine interaction. A detailed investigation into the nature of the paracrine molecules involved and whether these effects are general or specific to the ovarian cells is indicated.


   Acknowledgments
 
We acknowledge Drs. Richard Anderson and Peter Illingworth and their nursing support staff for helping to recruit the women involved in this study. Dr. Joo Thong provided the luteinized granulosa cells. Helen Wilson and Pawlina Largue provided excellent technical support.


   Footnotes
 
This work was supported by a project grant from the Wellcome Trust (to W.C.D.) and Medical Research Council Program Grant PG0000066 (to S.G.H.).

First Published Online June 7, 2005

Abbreviations: CTGF, Connective tissue growth factor; G6PDH, glucose-6-phosphate dehydrogenase; hCG, human chorionic gonadotropin; 3ß-HSD, 3ß-hydroxysteroid dehydrogenase; LCA, leukocyte common antigen; LDL, low-density lipoprotein; PGE2, prostaglandin E2; RNase, ribonuclease; {alpha}-SMA, {alpha}-smooth muscle actin; VEGF-R2, vascular endothelial growth factor receptor-2.

Received January 6, 2005.

Accepted June 1, 2005.


   References
Top
 Abstract
 Introduction
 Patients and Methods
 Results
 Discussion
 References
 

  1. Reynolds LP, Redmer DA 1999 Growth and development of the corpus luteum. J Reprod Fertil. 54(Suppl):181–191
  2. Fraser HM, Wulff C 2001 Angiogenesis in the primate ovary. Reprod Fertil Dev 13:557–566[CrossRef][Medline]
  3. Duncan WC, McNeilly AS, Illingworth PJ 1998 The effect of luteal ‘rescue’ on the expression and localization of matrix metalloproteinases and their tissue inhibitors in the human corpus luteum. J Clin Endocrinol Metab 83:2470–2478[Abstract/Free Full Text]
  4. Fraser HM, Lunn SF, Harrison DJ, Kerr JB 1999 Luteal regression in the primate: different forms of cell death during natural and gonadotropin-releasing hormone antagonist or prostaglandin analogue-induced luteolysis. Biol Reprod 61:1468–1479[Abstract/Free Full Text]
  5. Duncan WC, Rodger FE, Illingworth PJ 1998 The human corpus luteum: reduction in macrophages during maternal recognition of pregnancy. Hum Reprod 13:2435–2442[Abstract/Free Full Text]
  6. Duncan WC 2000 The human corpus luteum: remodelling during luteolysis and maternal recognition of pregnancy. Rev Reprod 5:12–17[Abstract]
  7. Duncan WC, McNeilly AS, Fraser HM, Illingworth PJ 1996 Luteinizing hormone receptor in the human corpus luteum: lack of down regulation during maternal recognition of pregnancy. Hum Reprod 11:2291–2297[Abstract/Free Full Text]
  8. Illingworth PJ, Reddi K, Smith K, Baird DT 1990 Pharmacologic "rescue" of the corpus luteum results in increased inhibin production. Clin Endocrinol (Oxf) 33:323–332[Medline]
  9. Brigstock DR 1999 The connective tissue growth factor/cysteine-rich 61/nephroblastoma overexpressed (CCN) family. Endocr Rev 20:189–206[Abstract/Free Full Text]
  10. Leask A, Abraham DJ 2003 The role of connective tissue growth factor, a multifunctional matricellular protein, in fibroblast biology. Biochem Cell Biol 81:355–363[CrossRef][Medline]
  11. Shimo T, Nakanishi T, Nishida T, Asano M, Kanyama M, Kuboki T, Tamatani T, Tezuka K, Takemura M, Matsumura T, Takigawa M 1999 Connective tissue growth factor induces the proliferation, migration, and tube formation of vascular endothelial cells in vitro, and angiogenesis in vivo. Mol Cell Biol 19:137–145
  12. Wandji S-A, Gadsby JE, Barber JA, Hammond JM 2000 Messenger ribonucleic acids for MAC25 and connective tissue growth factor (CTGF) are inversely regulated during folliculogenesis and early luteogenesis. Endocrinology 141:2648–2657[Abstract/Free Full Text]
  13. Slee RB, Hillier SG, Largue P, Harlow CR, Miele G, Clinton M 2001 Differentiation-dependent expression of connective tissue growth factor and lysyl oxidase messenger ribonucleic acids in rat granulosa cells. Endocrinology 142:1082–1089[Abstract/Free Full Text]
  14. Harlow CR, Hillier SG 2002 Connective tissue growth factor in the ovarian paracrine system. Mol Cell Endocrinol 187:23–27[CrossRef][Medline]
  15. Harlow CR, Davidson L, Burns KH, Yan C, Matzuk MM, Hillier SG 2002 FSH and TGF-ß superfamily members regulate granulosa cell connective tissue growth factor gene expression in vitro and in vivo. Endocrinology 143:3316–3325[Abstract/Free Full Text]
  16. Li TC, Rogers AW, Dockery P, Lenton EA, Cooke ID 1988 A new method of histologic dating of human endometrium in the luteal phase. Fertil Steril 50:52–60[Medline]
  17. Stamouli A, O’Sullivan MJB, Frankel S, Thomas EJ, Richardson MC 1996 Suppression of matrix metalloproteinase production by hCG in cultures of human luteinized granulosa cells as a model for gonadotrophin-induced luteal rescue. J Reprod Fertil 105:235–239
  18. Aston KE, Stamouli A, Thomas EJ, Vyas S, Iredale JP, Arthur MJP, Richardson MC 1996 Effect of gonadotrophin on cell and matrix retention and expression of metalloproteinases and their inhibitor in cultured human granulosa cells modelling corpus luteum function. Mol Hum Reprod 2:26–30[Abstract/Free Full Text]
  19. Maybin JA, Duncan WC 2004 The human corpus luteum: which cell types have progesterone receptors. Reproduction 128:423–431[Abstract/Free Full Text]
  20. Wulff C, Wilson H, Largue P, Duncan WC, Armstrong D, Fraser HM 2000 Angiogenesis in the human corpus luteum: Localisation and changes in angiopoietins, Tie-2 and vascular endothelial growth factor messenger ribonucleic acid. J Clin Endocrinol Metab 85:4302–4309[Abstract/Free Full Text]
  21. Schuldiner M, Yanuka O, Itskovitz-Eldor J, Melton DA, Benvenisty N 2000 Effects of eight different growth factors on the differentiation of cells derived from human embryonic stem cells. Proc Natl Acad Sci USA 97:11307–11312[Abstract/Free Full Text]
  22. Yong PY, Harlow C, thong KJ, Hillier SG 2002 Regulation of 11ß-hydroxysteroid dehydrogenase type I gene expression in human surface epithelial cells by interleukin-1. Hum Reprod 17:2300–2306[Abstract/Free Full Text]
  23. Matsui N, Rawano Y, Nakamura S, Miyakawa I 2004 Changes in vascular endothelial growth factor production associated with decidualization by human endometrial stromal cells in vitro. Acta Obstet Gynecol Scand 83:138–143[CrossRef][Medline]
  24. Fraser HM, Bell J, Wilson H, Taylor PD, Morgan K, Anderson RA, Duncan WC 2005 Localization and quantification of cyclic changes in the expression of endocrine gland vascular endothelial growth factor in the human corpus luteum. J Clin Endocrinol Metab 90:427–434[Abstract/Free Full Text]
  25. Fowkes RC, Chandras C, Chin EC, Okolo S, Abayasekara DR, Michael AE 2001 Relationship between the production of prostaglandins and progesterone by luteinizing human granulosa cells. J Endocrinol 171:455–462[Abstract]
  26. Jones RL, Salamonsen LA, Findlay JK 2002 Activin A promotes human endometrial stromal cell decidualization in vitro. J Clin Endocrinol Metab 87:4001–4004[Abstract/Free Full Text]
  27. Nowinski D, Hoijer P, Engstrand T, Rubin K, Gerdin B, Ivarsson M 2002 Keratinocytes inhibit expression of connective tissue growth factor in fibroblasts in vitro by an interleukin-1{alpha}-dependent mechanism. J Invest Dermatol 119:449–455[CrossRef][Medline]
  28. Yang DH, Kim HS, Wilson EM, Rosenfeld RG, Oh Y 1998 Identification of glycosylated 38-kDa connective tissue growth factor (IGFBP-related protein 2) and proteolytic fragments in human biological fluids and up-regulation of IGFBP-rP-2 expression by TGF-ß in Hs578T human breast cancer cells. J Clin Endocrinol Metab 83:2593–2596[Abstract/Free Full Text]
  29. Brigstock DR 2002 Regulation of angiogenesis and endothelial cell function by connective tissue growth factor (CTGF) and cysteine-rich 61 (CYR 61). Angiogenesis 5:153–165[CrossRef][Medline]
  30. Duncan MR, Frazier KS, Abramson S, Williams S, Klapper H, Huang X, Grotendorst GR 1999 Connective tissue growth factor mediates transforming growth factor ß-induced collagen synthesis: down-regulation by cAMP. FASEB J 13:1774–1786[Abstract/Free Full Text]
  31. Hwa V, Oh Y, Rusenfeld RG 1999 The insulin-like growth factor-binding protein (IGFBP) superfamily. Endocr Rev 20:761–787[Abstract/Free Full Text]
  32. Liu J, Kusma V-M, Vänttinen T, Hydén-Granskog C, Voutilainen R 2002 Gonadotrophins inhibit the expression of insulin-like growth factor binding protein-related protein-2 mRNA in cultured human granulosa-luteal cells. Mol Hum Reprod 8:136–141[Abstract/Free Full Text]
  33. Ratcliffe KE, Anthony FW, Richardson MC, Stones RW 1999 Morphology and functional characteristics of ovarian microvascular endothelium. Hum Reprod 14:1549–1554[Abstract/Free Full Text]
  34. Rae MT, Niven D, Critchley HO, Harlow CR, Hillier SG 2004 Antiinflammatory steroid action in human surface epithelial cells. J Clin Endocinol Metab 89:4538–4544[Abstract/Free Full Text]
  35. Dammeier J, Beer HD, Brauchle M, Werner S 1998 Dexamethasone is a novel potent inducer of connective tissue growth factor expression. Implications for glucocorticoid therapy. J Biol Chem 273:18185–18190[Abstract/Free Full Text]
  36. Ball DK, Surveyor GA, Diehl JR, Steffen CL, Uzumcu M, Mirando MA, Brigstock DR 1998 Characterization of 16- to 20-kilodalton (kDa) connective tissue growth factors (CTGFs) and demonstration of proteolytic activity for 38-kDa CTGF in pig uterine luminal flushings. Biol Reprod 59:828–835[Abstract/Free Full Text]
  37. Rageh MA, Moussad EE, Wilson AK, Brigstock DR 2001 Steroidal regulation of connective tissue growth factor (CCN2; CTGF) synthesis in the mouse uterus. Mol Pathol 54:338–346[Abstract/Free Full Text]
  38. Yu J, Prado GN, Schreiber B, Plogar P, Polgar P, Taylor L 2002 Role of prostaglandin E2 EP receptors and cAMP in the expression of connective tissue growth factor. Arch Biochem Biophys 404:302–308[CrossRef][Medline]
  39. Ricupero DA, Rishikof DC, Kuang PP, Poliks CF, Goldstein RH 1999 Regulation of connective tissue growth factor expression by prostaglandin E2. Am J Physiol. 277:L1165–L1171
  40. Stratton R, Shiwan X, Martini G, Holmes A, Leask A, Haberberger T, Martin GR, Black CM, Abraham D 2001 Iloprost suppresses connective tissue growth factor production in the skin of scleroderma patients. J Clin Invest 108:241–250[CrossRef][Medline]
  41. Rishikof DC, Ricupero DA, Kuang PP, Liu H, Goldstein RH 2002 Interleukin-4 regulates connective tissue growth factor expression in lung fibroblasts. J Cell Biochem 85:496–504[CrossRef][Medline]
  42. Villacorta L, Graca-Souza AV, Ricciavelli R, Zingg JM, Azzi A 2003 {alpha}-Tocopherol induces expression of connective tissue growth factor and antagonizes tumour necrosis factor-{alpha}-mediated downregulation in human smooth muscle cells. Circ Res 92:104–110[Abstract/Free Full Text]
  43. Abraham DJ, Shiwen X, Black CM, Sa S, Xu Y, Leask A 2000 Tumor necrosis factor {alpha} suppresses the induction of connective tissue growth factor by transforming growth factor-ß in normal and scleroderma fibroblasts. J Biol Chem 275:15220–15225