| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
research-article |
Department of Biochemistry and Molecular & Cellular Biology,3 Vincent T. Lombardi Comprehensive Cancer Center,4 Georgetown University Medical Center, Washington, DC 20057
ABSTRACT
Platelet-derived growth factors (PDGFs) and their receptors (PDGFRs) play a vital role in regulating cell growth and angiogenesis. In this study, the expression of the family of PDGFs and PDGFRs in the ovarian corpus luteum were identified and characterized, and an effect of their activity on development of the corpus luteum revealed. Gonadotropin-stimulated immature rats were utilized as a model of induced ovulation, luteogenesis, and pseudopregnancy. Levels of ovarian mRNA for Pdgfb and Pdgfd, and their receptor, Pdgfrb, increased significantly as early as 4 h after human chorionic gonadotropin (hCG) injection in immature rats primed with equine chorionic gonadotropin (eCG). Gonadotropin regulation of Pdgfb expression was confirmed by in vitro promoter-reporter assays, which showed a 2- to 3-fold increase in Pdgfb promoter activity in response to luteinizing hormone (LH). Inhibition studies implicated protein kinase A, phosphatidylinositol 3-kinase and mitogen activated protein kinase signaling pathways in the LH-induced upregulation. In the corpus luteum, PDGFA, PDGFB, PDGFC, and PDGFRA were localized to a population of luteal parenchymal/steroidogenic cells. PDGFRB was expressed primarily in what appeared to be cells of the luteal microvasculature. Intraovarian injection of an inhibitor of PDGF receptor activity, the tyrphostin AG1295, prior to injection of hCG in eCG-primed immature rats resulted in a significant 21.86% ± 11.15% decrease in corpora lutea per treated ovary in comparison to the contralateral vehicle-injected control ovary. In addition, the treated ovary of 3 of 16 rats showed widespread hemorrhage throughout the entire ovary, indicating a possible role for PDGF receptor activity in maintenance of the ovarian vasculature.
corpus luteum, growth factors, luteinizing hormone, ovary
The ovulatory surge of luteinizing hormone (LH) triggers extensive structural and biochemical changes in the preovulatory, or Graafian, follicle leading to rupture of the follicle and expulsion of the mature oocyte at ovulation. In the rat, ovulation occurs about 12–16 h after the LH surge [1–3], after which the follicle collapses inwards, generating inward folds of the granulosa cell layer around the follicular antrum and carrying the theca layer, including blood vessels lying within. By this time, the basement membrane is degraded, allowing the invasion of thecal cells and blood vessels into the granulosa cell layer [4–6]. A period of intense angiogenesis follows [7–10] whereby new blood vessels develop from the theca vasculature, extending initially towards the antrum and followed by extensive lateral branching of capillaries through rapid cell proliferation [8, 11, 12]. The newly formed vascular network is so extensive that in the mature rat corpus luteum, capillary lumina comprise 22% of the total volume of the corpus luteum and the majority of luteal parenchymal cells are either directly adjacent to capillaries (59%) or adjacent to interstitial space in close proximity to capillaries (37%) [13, 14]. This process is completed within a few days following ovulation, and by the mid-luteal phase the ovary is one of the most highly vascularized tissues in the body, receiving one of the greatest rates of blood flow per weight of any tissue [15–17]. The extensive microvascular network functions to supply the steroidogenic cells with cholesterol-rich lipoproteins as the substrate for steroid synthesis and collects the massive amounts of progesterone secreted by these cells. After ovulation the remaining cells of the ruptured follicle form a transient endocrine structure, the corpus luteum (yellow body). Whereas the follicle was compartmentalized, the corpus luteum is populated by a heterogeneous cell population that intermingles extensively in most mammals (except humans) and consists of parenchymal (steroidogenic) cells, endothelial cells, pericytes, fibroblasts, and cells of the immune system. Two different populations of luteal parenchymal cells with different morphology and functions can be identified in the corpus luteum of many species [18], and although the two subpopulations have been reported in dissociated rat corpora lutea [19, 20], they cannot be distinguished morphologically in tissue sections [21]. These two populations of luteal steroidogenic cells originate from the theca and ganulosa cells of the preovulatory follicle. Although the rapid transformation of the previously compartmentalized estrogen-producing pre-ovulatory follicle into a progesterone-producing corpus luteum is initiated by the LH surge, numerous intraovarian factors are essential for various stages of this process, resulting in tight regulation of the cycles of intense angiogenesis, luteinization, luteolysis, and regression of vasculation.
The family of platelet-derived growth factors (PDGFA, PDGFB, PDGFC, and PDGFD) are plieotropic factors that are expressed in a variety of tissues. They exert their effects through binding and subsequent activation of two structurally related tyrosine kinase receptors, PDGF receptor alpha (PDGFRA) and PDGF receptor beta (PDGFRB) [22, 23]. PDGFA, PDGFB, and PDGFC bind to PDGFRA, while PDGFB and PDGFD bind to PDGFRB [24, 25]. PDGFs are dimeric molecules that consist of either homodimers (AA, BB, CC, DD) or a heterodimer (AB) that bind two receptors simultaneously, thereby inducing a wide variety of cell responses, including proliferation, survival, and chemotaxis [23]. PDGFs and their receptors are often expressed by neighboring cells, suggesting a paracrine interaction. Expression of all four isoforms of PDGFs and both PDGFRs have previously been characterized in the rat ovary [26]. PDGF has been demonstrated to stimulate in vitro proliferation of theca cells from antral follicles from both rat [27] and pig [28, 29] while inhibiting thecal LH-induced steroid hormone synthesis [29]. In addition, the classic PDGF isoforms, PDGFA and PDGFB, were shown to contribute towards growth of preantral follicles [26].
The expression of platelet-derived growth factors and receptors in the mature follicle gives rise to the question of the possible involvement of this system in the post-LH life of the follicle. It has also been well established that PDGFB and PDGFRB are components of the vasculature and play an important role in angiogenesis [30–32]. The properties of platelet-derived growth factors and receptors in inducing cell proliferation and migration suggest that signaling through these receptors may be important in the development of the corpus luteum, either by affecting migration and/or proliferation of thecal cells and/or cells of the vasculature in response to the ovulatory surge of LH. In this study, the presence of platelet-derived growth factors and receptors in the corpus luteum was characterized and a role for PDGF receptor activity in luteogenesis was identified in vivo by the effects on formation of corpora lutea subsequent to a preovulatory blockade of PDGF receptor activity.
The Institutional Animal Care and Use Committee of Georgetown University approved all procedures. Immature Sprague-Dawley rats (Day 26) were injected with 5 I.U. equine chorionic gonadotropin (eCG; CalBiochem, La Jolla, CA), a nonpituitary hormone with FSH and some LH properties, followed 54 h later with an injection of 10 I.U. human chorionic gonadotropin (hCG; CalBiochem, La Jolla, CA), a nonpituitary hormone resembling LH in its action, to induce ovulation and formation of corpora lutea. Treatment of immature rats in this way has been shown to induce corpora lutea, which are then maintained for 2 wk [1]. Three rats were killed at specified time points after eCG and/or hCG injection. One ovary from each rat was placed in RNAlater (QIAGEN Inc. Valencia CA), surrounding fat and bursa dissected away, frozen in dry ice, and stored at –80°C until processed. The other ovary from each rat was placed in 10% neutral buffered formalin for 24 h and processed for immunohistochemistry.
Preparation of RNA and reverse transcription. Ovarian RNA was prepared using the QIAGEN RNeasy Mini Kit (QIAGEN, Inc., Valencia, CA) and included a DNase digestion step using the QIAGEN RNase-free DNase set (QIAGEN, GmbH, Hilden, Germany). RNA was eluted in 30 µl RNase-free water and aliquots stored at –80°C. Ovarian RNA (1 µg) was reverse transcribed to cDNA using random hexamers and the Taqman Reverse Transcription reagents (Applied Biosystems, Foster City, CA) in a total volume of 50 µl per reaction. The reverse transciption reaction was performed in a BioRad iCycler (BioRad, Hercules, CA) at 25°C 10 min, 37°C 30 min, and 95°C 5 min. Reaction products were stored at –20°C.
Real-time PCR. For real-time PCR, 4 µl of equal dilutions of each reverse transcription reaction product was added to 5 µl 2X SYBR Green PCR Master Mix (Applied Biosystems, Foster City, CA) and made up to a final volume of 10 µl with primers and water. The concentration of each set of primers was determined during prior validation experiments using a series of dilutions of each primer set. Forward and reverse 18S rRNA and Pdgfb primers were used at 200 nM, forward and reverse Pdgfa, Pdgfc, Pdgfd, and Pdgfra primers were used at 300 nM, and forward and reverse Pdgfrb primers were used at 600 nM final concentrations. Primers used were: Pdgfa forward 5'- CCC CGG GAG TTG ATC GA -3', reverse 5'- GGT TTG TCT CCA AGG CAT CCT -3'; Pdgfb forward 5'- CTC CAT CCG CTC CTT TGA TG -3', reverse 5'- CAG AAT GTG CTC GGG TCA TGT -3'; Pdgfc forward 5'- CTC GGG CTG AGT CCA ACC T -3', reverse 5'- TGC ACT CCG TTC TGC TCC TT- 3' Pdgfd forward 5'- CAA CAG CTA CCC GAG GAA CCT -3', reverse 5'- TGG TCA AAG GCC AGC TGT ATC -3'; Pdgfra forward 5'- TTC CCC TGC CAG ACA TTG AC -3', reverse 5'- ATG GCG CTC TCT TCC GAA GT -3'; Pdgfrb forward 5'- GTG AGC GGA AGC GCA TCT A -3', reverse 5'- TCA CTC GGC ATG GAA TCG T -3'; 18S rRNA forward 5'- AGG ATC CAT TGG AGG GCA AG -3', reverse 5'- AAC TGC AGC AAC TTT AAT ATA CGC TAT T -3'. Reaction conditions were 50°C for 2 min, 95°C for 10 min, and 40 cycles of 95°C for 15 sec and 60°C for 1 min using an ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA).
CT values (the PCR cycle number at which the PCR fluorescence product is detectable above an arbitrary threshold) for each reaction were determined using ABI PRISM SDS 2.1 analysis software. Quantitation of results was determined using the delta delta CT method [33]. Briefly, Pdgf mRNA levels were normalized to 18S rRNA levels for each quadruplicate sample to compensate for errors in total RNA amounts (Delta Ct). Fold differences of mRNA per µg total ovarian RNA were determined by comparison of each Delta Ct to an arbitrary control (eCG time-0) (Delta Delta Ct), and the relative value (RQ) determined by raising 2 to the power of the negative Delta Delta Ct value [33]. Single and specific PCR products, as well as lack of contamination or primer dimer formation, were determined by a heat dissociation curve assay at the end of the 40 cycles. Using a 386-well plate, each real-time PCR reaction included quadruplicate samples of mRNA from one ovary from each of three rats at each time point; thus, one real-time PCR reaction yielded complete data for levels of mRNA for each of the four Pdgfs, two Pdgfrs, and 18S rRNA. To determine equal efficiency of reactions for Pdgf or Pdgfr and 18S rRNA, each reaction also included a standard curve of log of transcript concentrations versus CT, based upon a number of dilutions of ovarian cDNA prepared for this purpose; each of the seven real-time PCR reactions yielded efficiencies of 95%–100%.
Statistics. Comparisons of the fold change in Pdgf and Pdgfr message per µg total ovarian RNA at each time point compared with time-0 were carried out using specific contrasts within a generalized estimating equations (GEE) analysis [34]. This analysis calculated the appropriate variance taking into account the correlation between replicates from the same ovary. A Bonferroni adjustment [35] for significance was utilized such that any P values between 0.05 and 0.0029 were considered interesting, and P < 0.0029 was considered significant. Additionally, a two-sided binomial test of whether the proportion of time points with fold-changes greater than time-0 was performed for each isoform and receptor. If P < 0.05, then the isoform or receptor was considered to significantly change following treatment.
Pdgfb Promoter–Reporter Assays
Cell transfection and reporter assays. The Pdgfb promoter – luciferase reporter construct (pRALuc) [36] spanning from –396 to the mRNA cap site of the promoter (+1), cloned upstream of the firefly luciferase reporter gene was a gift from Dr. Lee Ratner (Washington University). The Thymidine Kinase promoter–luciferase reporter construct (pRL-TK vector; Promega Corporation, Madison, WI) was used as an internal control reporter, and consists of cDNA encoding the herpes simplex virus thymidine kinase promoter and the Renilla luciferase reporter gene (Rluc).
The concentration of internal control reporter construct (pRL-TK) utilized in these experiments was determined based on a series of validation experiments with varying concentrations in the presence of a constant amount (1 µg) of experimental reporter construct (pRALuc). In the absence of an LH-responsive thecal cell line, the promoter-reporter assays were performed using the LH-responsive MA10 Leydig cell line obtained from Mario Ascoli (University of Iowa) [37]. MA10 Leydig cells were cultured in DMEM/Hams F12 medium (Invitrogen Corporation, Grand Lakes, NJ) containing 5% donor horse serum (DHS) (Invitrogen Corporation, Grand Lakes, NJ) and 2.5% fetal calf serum (FCS) (Invitrogen Corporation, Grand Lakes, NJ), at 5% CO2 in a humidified chamber.
For transfection, cells were plated in 6-well plates at 5 x 104 cells in 1 ml medium per well. The next day cells were co-transfected with pRALuc and pRL-TK using Lipofectamine 2000 reagent (Invitrogen Corporation, Carlsbad, CA) according to the manufacturer's protocol. For each well of a 6-well plate, pRL-TK (1 ng) and pRALuc (1 µg) were mixed with 50 µl Opti-MEM I (Invitrogen Corporation, Grand Lakes, NJ). Twenty-four hours following transfection, medium was replaced with 1 ml DMEM/Hams F12 medium plus 1X ITS, 2 mg/ml Albumax and with or without an inhibitor of the following signaling pathways: Src family tyrosine kinase inhibitors, PP2 (10 µM; CalBiochem, La Jolla, CA) or SU6656 (1 µM; CalBiochem, La Jolla, CA); phosphotidylinositol 3-kinase (PI 3-kinase) inhibitor, LY294002 (50 µM; CalBiochem, La Jolla, CA); mitogen activated protein kinase (MAPK) (42/44) inhibitor, PD98059 (30 µM); protein kinase A inhibitors, H-89 (10 µM; CalBiochem, La Jolla, CA) and PKA-I (100 µM; CalBiochem, La Jolla, CA); and a protein kinase C inhibitor, Calphostin C (1 µM; CalBiochem, La Jolla, CA). Ovine LH (20 ng/ml; NIADDK-NIH, Bethesda, MD) was added to each well 30 min later.
After 24 h, cells were lysed and luciferase activity determined using the Dual Luciferase Reporter Assay System (Promega Corporation, Madison, WI). PDGFB promoter activity, as determined by firefly luciferase activity, was normalized for transfection efficiency to Renilla luciferase activity. In one experiment, firefly luciferase activity was normalized to protein concentration to confirm results.
Statistics. Statistical significance of differences in Pdgfb promoter activity, as determined by firefly luciferase activity, in the presence of LH compared with promoter activity in the presence of LH and inhibitors was determined by ANOVA followed by t-test analyses. Data was extracted from three experiments.
Paraffin-embedded ovaries were sectioned (3 µm) and deparaffinized in xylene for 2 x 5 min, followed by rehydration in decreasing concentrations of ethanol (100%, 90%, 70%, and 50%) and finally in water for 5 min. Antigen retrieval was performed in Retrievit (BioGenex, San Ramon, CA) at pH 5.5 (PDGFRA, PDGFA, PDGFC), pH 6 (PDGFRB), or pH 7 (PDGFB) in a microwave at 20% maximum for 10 min. After cooling for 30 min at room temperature, sections were washed in PBS pH 7.6 and endogenous peroxidase activity was quenched in 0.3% H2O2 in PBS for 5 min at room temperature. Sections were washed in PBS and blocked for 45 min at room temperature with 10% normal goat or rabbit serum (Vector Laboratories, Burlingame, CA) in PBS, washed and incubated for 24–48 h at 4°C in a 1:100 dilution of 1% BSA in PBS of antibody specific for PDGFRA (sc338; Santa Cruz Biotechnology, Santa Cruz, CA), PDGFRB (sc432; Santa Cruz Biotechnology, Santa Cruz, CA), PDGF-AA (BioDesign International, Saco, Maine), PDGFB (sc7878; Santa Cruz Biotechnology, Santa Cruz, CA), or PDGFC (sc18228; Santa Cruz Biotechnology, Santa Cruz, CA).
After washing in PBS, sections were exposed to a 1:200 dilution of biotinylated secondary antibody (anti-goat or anti-rabbit; Vector Laboratories, Burlingame, CA) in 1% BSA in PBS for 1 h at room temperature. Sections were washed with PBS and immunoreactivity visualized by 20 min incubation in peroxidase-conjugated streptavidin (Biogenex, San Ramon, CA) at room temperature followed by peroxidase reduction of AEC substrate (Zymed, San Francisco, CA). Sections were counterstained with hemotoxylin.
In Vivo PDGF Receptor Blockade
For in vivo PDGF receptor blockade, immature (26-day-old) female Sprague-Dawley rats received 5 I.U. eCG by intraperitoneal injection. Fifty hours later, they were anesthetized and their ovaries exposed via bilateral 5 mm incisions made on the dorsum 5 mm lateral to each side of the spine and 5 mm posterior to the rib cage. A PDGF receptor–selective inhibitor, AG1295 (CalBiochem, La Jolla, CA) was injected into the right ovarian bursa (20 µl of a 5 µg/µl solution in DMSO). The contralateral ovary served as a control, receiving DMSO (Sigma Chemical Co., St Louis, MO) only. Following recovery from surgery, rats received 10 I.U. hCG by intraperitoneal injection. Rats were killed 72 h later, ovaries collected, fixed in 10% neutral buffered formalin, and processed for paraffin embedding. Ovaries were sectioned (5 µm) and every 20th section stained with hemotoxylin and eosin and analyzed for formation and numbers of corpora lutea present. Individual corpora lutea were followed through every 20th section analyzed, and the total number of corpora lutea per ovary recorded. A total of 16 ovary pairs were analyzed.
Statistics. For each animal, the number of corpora lutea in the treated ovary was represented as a percentage difference to the number of corpora lutea in the contralateral control ovary. A one-sample t-test was performed on the total percentage differences, and the result represented as a percentage difference to control ovaries (100%) ± SEM.
Effect of hCG on mRNA Levels of PDGFs and Receptors
We have previously described the expression of the various PDGFs and receptors in the rat ovary from birth until post-natal Day 28 [26]. To determine whether levels of PDGFs and PDGF receptors are regulated in any way by gonadotropins during the estrous cycle, a rat model of pseudopregnancy was employed. Following hCG injection, significant changes were observed in levels of mRNA per µg total ovarian mRNA for several Pdgfs and Pdgfrs. Increases in Pdgfra mRNA were observed following hCG injection, with a 2- to 2.5-fold increase at both 4 h and 18 h (Fig. 1A), although at these numbers of samples (n = 3), this increase was considered interesting but not significant. Pdgfrb showed a significant overall increase in mRNA after hCG injection (P = 0.001) with 2.5- to 4.5-fold increases from 18 h to 2 wk following treatment (Fig. 1B). Levels of Pdgfa mRNA, which appeared to have decreased prior to hCG injection, increased to control levels and then did not vary significantly (Fig. 1C).
|
Pdgfc, the protein product of which also binds to PDGFRA, showed a significant overall decrease in mRNA present in the ovary following treatment (P = 0.001), diminishing to 50% of eCG time-0 levels as early as 2 h after the gonadotropin surge. This level was maintained during the period of corpus luteum formation and decreased still further thereafter (Fig. 1E). In contrast, transcripts of ligands for PDGFRB, namely Pdgfb and Pdgfd, both exhibited significant overall increases in mRNA following hCG injection (P = 0.001 and 0.01 respectively) (Fig. 1, D and F). Pdgfb message increased approximately 3-fold as early as 4 h post hCG, and levels peaked at 4- to 5-fold by 2 days following hCG injection. Pdgfb mRNA dropped slightly thereafter, but was maintained at 3- to 4-fold over pre-hCG levels for 2 wk, corresponding to the duration of the corpus luteum (Fig. 1D). An approximate 3-fold increase in Pdgfd message was observed 18 h post hCG injection and maintained at 2- to 3-fold during the early luteal phase (Fig. 1F).
In Vitro Pdgfb Promoter Activity in Response to LH and Signal Transduction Inhibitors
The rapid in vivo increase in Pdgfb message in response to hCG treatment suggests that LH may directly induce Pdgfb gene transcription. In order to determine if LH directly activates the Pdgfb promoter, MA-10 cells were co-transfected with a plasmid containing the Pdgfb gene promoter – luciferase reporter construct (pRALuc) [36] and a thymidine kinase – renilla luciferase reporter construct as an internal control reporter. After transfection, cells were cultured in the presence or absence of LH, and Pdgfb promoter activity was determined by firefly luciferase activity and normalized for transfection efficiency by renilla luciferase activity. Treatment of MA10 cells with LH for 24 h resulted in a 2- to 2.5-fold increase in Pdgfb promoter activity compared with control treated cells (Fig. 2).
|
To further explore gonadotropin regulation of Pdgfb expression, MA10 cells were cultured in the presence of LH plus inhibitors of various signaling pathways, including Src family tyrosine kinases (PP2 and SU6656), PI 3-kinase (LY294002), MAP p42/44 kinase (PD98059), protein kinase A (H-89 and PKA-I) and protein kinase C (Calphostin C). Cells transfected with the Pdgfb promoter and exposed to LH showed inhibition of LH-stimulated Pdgfb promoter activity when preincubated in the presence of inhibitors of PI 3-kinase (LY294002), MAP kinase (PD98059), and protein kinase A (PKA-I, HA-89) (Fig. 2). In contrast, inhibition of either Src family tyrosine kinases (PP2 and SU6656) or protein kinase C (Calphostin C) had no significant effect on LH-induced Pdgfb promoter activity (Fig. 2).
Localization of PDGFs and Receptors in the Corpus Luteum
A previous study demonstrated the expression of PDGFs and PDGF receptors in both granulosa and thecal compartments of the developing and preovulatory rat ovary [26]. To determine which cells within the ovary express PDGFs and PDGFRs during the ovulatory process and luteogenesis, immunohistochemistry using antibodies to PDGFs and PDGFRs was performed on ovarian sections from rats at various time points after treatment with eCG and hCG.
At 8 h post-hCG, a subpopulation of granulosa cells in some secondary and antral follicles were immunoreactive for PDGFRA (data not shown). At 18 h post-hCG, theca and granulosa cell layers were no longer separated by a basement membrane and luteogenesis was underway as indicated by the hypertrophy of some of the cells within the developing corpus luteum. PDGFRA was expressed by a population of cells within the developing corpus luteum (Fig. 3, A and B). Three days after hCG injection, fully formed corpora lutea were observed and PDGRA immunoreactivity was apparent in a subset of luteal parenchymal cells and possibly also in cells of the vasculature (Fig. 3, C and D). A similar pattern of localization was observed at 8 days post-hCG (Fig. 3, E and F). PDGFRB was expressed in what appear to be cells of the vasculature in the developing and mature corpus luteum at 48 h after hCG injection (Fig. 3, H and I). At 3 days post-hCG, PDGFA (Fig. 4, A and B), PDGFB (Fig. 4, E and F), and PDGFC (Fig. 4, H and I) were also localized to a population of luteal parenchymal cells, and similarly at 8 days after injection with hCG (PDGFA, Fig. 4, C and D).
|
|
In Vivo PDGF Receptor Blockade
In an effort to determine the in vivo effect of PDGF signaling inhibition on ovarian function during the periovulatory period, a rat model of gonadotropin-induced ovulation was utilized. Rats (n = 16) were first injected with eCG to induce follicle growth, and 50 h later with hCG to emulate the LH surge. Two hours prior to injection with hCG, the PDGF receptor-selective inhibitor, AG1295, which inhibits both PDGFRA and PDGFRB activity, was injected into the ovarian bursa of the right ovary of each rat, and the contralateral ovary was injected with an equal volume of vehicle (DMSO).
Of 16 rats, four did not respond to hormone treatments, as shown by an absence of development of corpora lutea. In 10 of the 12 rats that responded to hormone injections, as evidenced by formation of corpora lutea, a significant reduction (21.86% ± 11.5% SEM) was observed in the average percentage of corpora lutea in AG1295-treated ovaries relative to the percentage of corpora lutea in control ovaries (100%) (Fig. 5A). Photomicrographs of a section from a control and a treated ovary from one rat are shown in Figure 5, B and C, respectively. Control ovaries from the remaining two rats exhibited six and seven corpora lutea in the early stages of formation with cells still in the early stages of luteinization, while the contralateral treated ovaries from these two rats revealed an absence of corpora lutea and widespread hemorrhage throughout each entire ovary. In all, widespread hemorrhage occurred in the treated ovary of three rats (Fig. 6, A–C). This was observed as a mass extravasation of red blood cells into surrounding regions throughout the entire ovary (Fig. 6, D–G). No corpora lutea were observed in any of the hemorrhagic AG1295-treated ovaries. The three contralateral control ovaries varied in their responses, with one control ovary also exhibiting a total absence of corpora lutea and the other two control ovaries having six and seven developing corpora lutea still in the early stages of luteogenesis, as previously mentioned.
|
|
We have previously demonstrated the presence of PDGFs and receptors in developing pre-antral and antral follicles of the rat ovary [26]. Furthermore, in an in vitro cell culture system, PDGFB stimulated thecal cell growth [27–29], and in an in vitro follicle culture system, PDGFA and PDGFB stimulated follicular growth [26]. The current study demonstrates the expression of PDGFs and PDGFRs in the rat corpus luteum. In the developing and mature corpus luteum, PDGFA, PDGFB, and PDGFC were all localized to luteal parenchymal cells, PDGFRB was expressed in what appeared to be cells of the microvasculature, and PDGFRA was observed in both luteal parenchymal and vascular cells, suggesting autocrine and paracrine functions of the PDGF ligands.
The LH surge in the estrous cycle induces a pleiotropic array of effects within the ovary. These include regulation (both positive and negative) of the expression of numerous gene products [3], resulting in structural changes in the preovulatory follicle to allow expulsion of the ovum and subsequent luteinization of the follicle to form the transient endocrine gland, the corpus luteum. The present study demonstrated a rapid increase in Pdgfb and Pdgfrb message following hCG injection to eCG primed immature rats, as soon as 4 h following hCG injection, a time frame observed for numerous gene products in response to the LH surge [3], and suggesting a possible direct regulation of Pdgfb and Pdgfrb in the ovary by the LH surge. This observation was further strengthened by the demonstration that LH increased Pdgfb promoter-driven luciferase activity in the LH-responsive MA10 cell line.
The LH-induced increase in Pdgfb message appears to be mediated by the cAMP protein kinase A (PKA) pathway, as incubation of LH-responsive stereoidogenic cells in the presence of both LH and each of two PKA inhibitors resulted in a reduction in the LH-induced increase in Pdgfb promoter activity. Conversely, cholera toxin, an activator of adenylate cyclase and increased intracellular cAMP, has been observed to induce activation of the Pdgfb promoter (Taylor, unpublished data). The PI3-kinase and MEK – ERK signaling pathways are also implicated in regulation of LH-stimulated Pdgfb promoter activity, as inhibition of either of these kinases at least partially inhibited the LH-stimulated increase in promoter activity. Recent studies have identified the activated LH receptor to signal through a number of signaling pathways other than the classic cAMP-PKA pathway, including the MEK – ERK [38–40] and the PI3-kinase – Akt [38] pathways. LH activation of the MEK – ERK pathway in ovarian steroidogenic cells has been implicated in downregulation of the steroidogenic machinery [39, 40]. While previous reports indicate that protein kinase C induces Pdgfb promoter activity in several cell types [41–43], inhibition of PKC did not have any effect on promoter activity in the cell type utilized in the current studies. Thus data from the present study indicate that LH may utilize multiple pathways to maximally activate the Pdgfb promoter and demonstrate crosstalk between G protein and growth factor signaling pathways. Pdgfd mRNA also increased significantly in response to hCG, as demonstrated by real-time PCR, but as the pattern of expression differed to that of Pdgfb, it is possible that different intracellular pathways regulate LH-induced Pdgfd expression, although this was not explored in this study.
PDGFRA, PDGFA, PDGFB, and PDGFC protein expression were localized by immunohistochemistry to subpopulations of parenchymal/steroidogenic cells within the corpus luteum, but it could not be determined from these studies whether these proteins were expressed in the same or in different subpopulations. Although two different populations of luteal parenchymal cells with different morphology and functions can be identified in the corpus luteum of many species, we and others have not been able to distinguish morphologically the two populations in rat tissue sections [21]. However, as all these proteins are expressed on theca cells of the preovulatory follicle [26], it is possible that the group of parenchymal cells expressing these proteins originate from the thecal layer. The expression of PDGFRA and PDGFA in the granulosa cells of preovulatory follicles [26] also implicates these cells as the possible origin of the luteal parenchymal cells expressing these two proteins. Following the LH surge, an increase in cell proliferation occurs, involving both follicle cells and cells of the vasculature (i.e., endothelial cells and pericytes). Thus, the observed increases in levels of Pdgfb, Pdgfd, and Pdgfr mRNA following hCG injection may reflect an increase in expression of these ligands and receptors on theca and granulosa cells leading to cell proliferation in response to signaling between ligands and receptors. In view of the evidence indicating proliferation of luteal parenchymal cells [21, 44–48], it is also possible that during luteogenesis and luteal maintenance, PDGF signaling may be important for luteal parenchymal cell proliferation and/or survival. In line with these observations, it is interesting to note that PDGFA and PDGFB have been identified in human ovarian follicular fluids from females undergoing ovarian stimulation, and PDGF receptors have been identified by Western blotting in extracts from granulosa-luteal cells isolated from follicular fluids from these patients [49].
PDGFRB immunoreactivity was observed in cells of the corpus luteum that exhibited a similar morphology and distribution pattern to that of cells of the microvasculature [31]. It has been well established that PDGFRB is expressed by both pericyte progenitors and by pericytes of the microvasculature [50, 51] and that these cells play an important role in vessel formation and stabilization during angiogenesis [52–56]. Capillary pericytes have been proposed to be the primary vascular cells that initiate the angiogenic process, and in both the bovine and ovine corpus luteum, pericytes invade the granulosa folds of the ruptured follicle either before or coincident with the thecal endothelial cells [57–59]. A high rate of pericyte proliferation has been observed in the early ovine corpus luteum, representing 22% of proliferating cells 2 days after ovulation [59]. It is likely that the increase in Pdgfrb mRNA at 18 h after hCG injection reflects the increase in Pdgfrb expressing pericytes during luteogenesis, in addition to the possible aforementioned increased expression in theca cells.
Thus, the differential localization of the platelet-derived growth factors in luteal parenchymal cells and PDGFRB in the vascular elements suggest a paracrine activity whereby PDGFs promote the recruitment and development of vascular elements in the developing corpus luteum. PDGFB is an angiogenic factor involved in the recruitment of pericytes to newly forming vessels during angiogenesis as well as inducing pericyte proliferation [30–32]. Although not yet fully investigated, it is possible that PDGFD may be involved in the angiogenic process; studies based on embryonic expression have led to the hypothesis that PDGFB and PDGFD may provide different types of signals to PDGFRB-expressing pericytes [60]. The increase in Pdgfb and Pdgfd transcripts following hCG injection observed in the current studies may reflect a role in angiogenesis. Thus, during the periovulatory period and in the newly forming corpus luteum, expression of PDGFB (and possibly also PDGFD) by theca cells and then luteal parenchymal cells, likely the invading theca cells of the developing corpus luteum, may result in the proliferation and migration of pericytes to surround the developing capillaries [58]. It is also likely that PDGFC expression by theca cells and luteal parenchymal cells contributes towards this angiogenic process [24, 61].
To define the importance of the PDGF ligand-receptor system in the corpus luteum, an inhibitor of PDGF receptor signaling was injected into the ovary prior to injection of hCG in a rat model of gonadotropin-induced ovulation. Results varied, and this could be attributed to the difficulty of injecting a small volume under the ovarian bursa and into the ovary. Although it has been reported that the tyrphostin AG1295 is rapidly taken up by cells [62], the concentration of AG1295 most effective for inhibition of receptor activity throughout the ovary has not been determined, just as it is not clear how extensively the inhibitor circulates throughout the ovary, or how long the inhibitor remains active following injection. It is possible that differences in response to the treatment could depend on where the inhibitor entered the ovary and consequently how effectively it was circulated. However, it is interesting to note that AG1295-treated ovaries showed a significant average 21.86% ± 11.15% reduction in corpora lutea when compared with the contralateral vehicle-injected control ovaries (100%). This average percentage reduction is still greater when taking into account the two control ovaries that revealed six and seven corpora lutea still in the early stages of luteogenesis with both contralateral treated ovaries exhibiting widespread hemorrhage and no corpora lutea. While variations in numbers of corpora lutea between normal contralateral ovaries are regularly observed, it is noteworthy in these studies that the treated ovaries show an overall significant reduction in numbers of corpora lutea when compared with the non-treated ovaries.
It is not clear from these studies why inhibition of PDGF receptor signaling affects formation of corpora lutea. It is possible that inhibition of signaling could result in lack of pericyte recruitment and proliferation and/or function and therefore lead to inhibition of development of the microvascular system of the Graafian follicle. Classic studies of ovarian morphology recognized that capillary growth may be important in the selection and growth of ovulatory follicles, as it was observed that the capillary network of preovulatory follicles is more extensive than that of other follicles [14, 58]. More recently, morphometric measurements have shown increased vascularity of dominant follicles [63, 64] and a subsequent greater selective in vivo uptake of gonadotropin by dominant follicles. It has been suggested that increased vascularity may be a primary determinant in the selection and maturation of dominant follicles [64, 65], thereby allowing greater access to nutrients, hormones, and factors essential for final development and ovulation. In contrast, insufficient vascular supply could limit further growth and lead to follicle degeneration or atresia [66]. Thus, inhibition of PDGF receptors by AG1295 just prior to the LH surge could impair the growth and/or stability of the capillary network in the thecal layer of the Graafian follicle, reducing subsequent gonadotropin uptake and compromising final development and subsequent ovulation of the Graafian follicle. Similarly, interference with another angiogenic factor, VEGF, has also been demonstrated to affect subsequent ovulation and luteal function in primates [66–68].
In the current study, it is interesting to note that massive hemorrhage was observed in a proportion of treated ovaries, again suggesting a role of PDGF signaling in ovarian angiogenesis and possibly vascular stabilization. It is possible the hemorrhage was the result of mechanical damage during injection, however, such widespread hemorrhage was never observed in sham-injected control ovaries. Indeed, the observation of hemorrhage in these treated ovaries is similar to the observed widespread microvascular leakage and hemorrhage observed in PDGFB and PDGFRB knockout mice [69, 70]. This phenotype has been attributed to a severe deficit in pericytes resulting from a failure of the expansion and spreading of the existing pericyte population due to impaired PDGFB/PDGFRB signaling [50, 51]. It is worth noting that only one third of homozygous PDGFRB mutant mice at E16.5 and E18.5 exhibit an overt phenotype of purpura, an accumulation of blood under the surface of the dermis, while the remaining two thirds of mutant embryos at both time points lack this phenotype, despite sharing the microvasculature defects as previously described [70]. Similarly, in the current study only three of 16 AG1295-treated ovaries exhibited an overt hemorrhagic phenotype. The current study also raises the possibility that the inhibition or impairment of normal PDGFB/PDGFRB signaling by AG1295 treatment interferes with the normal function of existing pericytes and leads to destabilization of existing blood vessels in the ovary, the extent of the effect depending on the extent of penetration of the inhibitor within the ovary. Recently two groups have shown that inhibition of PDGF receptor activity by administration of a PDGF receptor inhibitor or an anti-PDGFRB monoclonal antibody leads to pericyte loss through apoptosis, with subsequent effects on endothelial cell viability in rat retina and neonatal mouse glomeruli [71, 72].
Although the current study demonstrates that pharmacologic blockade of PDGF signaling appears to have significant effects on ovulation and luteal development, it must be noted that AG1295 also inhibits activity of c-kit, albeit at a higher effective dose than for the PDGF receptors [62]. As yet, the role of c-kit and its ligand, stem cell factor, in the corpus luteum has not been investigated. In mice, the cyclic secretion of LH was followed by an immediate and dramatic elevation of stem cell factor in mural granulosa cells and decreased levels of c-kit transcripts in theca and interstitial cells [73]. The presence of stem cell factor protein has been observed in the large and small luteal cells of the ovine corpus luteum on Days 3 and 10 of the luteal phase [74], with levels of mRNA constant throughout the luteal phase. C-kit protein is present in small luteal cells and endothelial cells of the ovine corpus luteum, and mRNA levels are low in the early luteal phase, increasing to maximum levels on Day 13 (midluteal) and decreasing by Day 16 [75]. Thus, although c-kit is present in the corpus luteum, the low levels of mRNA following the LH surge and during the early stages of luteogenesis suggest that it might not play an important role in the development of the corpus luteum.
In summary, these studies have elucidated the presence of components of the PDGF ligand-receptor system in the corpus luteum of the rat. The LH-induced increase in levels of mRNA for several platelet-derived growth factors and receptors suggests a role for this growth factor system during luteogenesis. This is confirmed by the observation that inhibition of PDGF receptor signaling prior to ovulation and luteogenesis resulted in a significant decrease in the average percentage of corpora lutea relative to the contralateral control ovaries, and by the widespread hemorrhage observed in a proportion of treated ovaries. Further investigation is required to identify the luteal steroidogenic cell population(s) that synthesize this group of growth factors and receptors, as well as the roles of these growth factors and receptors in luteal parenchymal and luteal microvascular cell proliferation and migration.
ACKNOWLEDGMENTS
The authors would like to acknowledge Rebecca Slack, M.S., for statistical analyses, and Robert Koos, Ph.D., and Jodi Flaws, Ph.D., for helpful discussions.
FOOTNOTES
1Supported by NIH grants HD36013 and HD39523. 
Correspondence: 2Leanne Sleer, Georgetown University Medical Center, 3970 Reservoir Rd., Washington, D.C. 20057. FAX: 202 687 8434; e-mail: sleerl{at}georgetown.edu
Received: 17 May 2006.
First decision: 23 June 2006.
Accepted: 3 November 2006.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||
