| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Regular Article |
e Department of Biomedical Sciences, Cornell University, Ithaca, New York 14853
f Department of Obstetrics and Gynecology, Stanford University Medical Center, Stanford, California 94305
ABSTRACT
A critical transition in ovarian follicular development is the selection of a dominant follicle, capable of ovulating, from a cohort of synchronously growing antral follicles. However, little is known about mechanisms and factors that regulate the selection and growth of dominant ovarian follicles. We have investigated whether a component of the insulin-like growth factor (IGF) system, namely IGFBP-4 protease, is associated with the establishment of follicular dominance in cattle. IGFBP proteases degrade IGFBPs, freeing IGFs to interact with their receptors. In experiment 1, follicular fluid from preovulatory follicles (n = 4) degraded about 80% of the added recombinant human (rh) IGFBP-4 within 18 h of incubation. The IGFBP-4 protease exhibited optimal activity at neutral/basic pH and its sensitivity to various protease inhibitors suggested a metalloprotease. The decline in the intensity of the band corresponding to intact rhIGFBP-4 was accompanied by the appearance of immunoreactive fragments of molecular weights
18 and 14 kDa, which were not detectable by ligand blot analysis. In experiment 2, follicular fluid samples were collected from dominant and subordinate follicles on Day 2 or 3 of the first follicular wave, after ovariectomy (experiment 2a, n = 3/day) or by ultrasound-guided follicular aspiration (experiment 2b, n = 4–5/day). Estradiol concentrations in follicular fluid from dominant vs. subordinate follicles confirmed their identities and indicated that the dominant follicle had been selected by Day 2 of the follicular wave. In both experiments 2a and 2b, IGFBP-4 proteolytic activity was 2- to 3.5-fold (P < 0.05) and 5-fold (P < 0.01) higher in follicular fluid from dominant than subordinate follicles on Days 2 and 3 of the follicular wave, respectively. The finding that IGFBP-4 proteolytic activity is higher in dominant, estrogen-active follicles than in subordinate follicles of the same cohort, as early as Day 2 of the follicular wave, strongly suggests a role for IGFBP-4 protease in the establishment of ovarian follicular dominance.
follicle, follicular development, FSH, growth factors, ovary
INTRODUCTION
The culmination of ovarian follicular development is the ovulation of a species-specific number of follicles at the end of each follicular phase. The ovulatory follicle(s) is selected from a larger cohort of follicles recruited for further synchronous growth by a small rise in circulating FSH [1–3], but the factors that promote the selection of an appropriate number of follicles for dominance and further development are not understood [4]. Cattle provide an experimental model of great practical utility for studying the mechanisms of selection of the dominant follicle. During the bovine estrous cycle there are two or three sequential waves of follicular development, each producing a dominant follicle capable of ovulating if luteolysis occurs naturally or is induced experimentally [5]. In addition, follicles are large enough to be followed by ultrasound imaging and hence can be sampled or isolated at specific stages of follicular development; they also yield sufficient follicular fluid and follicular cells for analysis of multiple end points within a single follicle.
When a bovine dominant follicle is just slightly larger in size (1 mm) than the largest subordinate, it has greater capacity to produce estradiol than subordinate follicles of the same cohort [6]. It has been hypothesized that differential expression of gonadotropin receptors, steroidogenic enzymes, or both may be part of the mechanisms leading to the establishment of follicular dominance [4, 7]. Although some differences in expression of mRNA for gonadotropin receptors and key steroidogenic enzymes [8–10] were observed between dominant follicles of the first wave isolated after selection and recruited follicles obtained before selection, such differences were not apparent when the dominant and the two largest subordinate follicles of the first wave were compared directly around the time of follicular selection [6]. Therefore, we hypothesized that the intrafollicular insulin-like growth factor (IGF) system plays a critical role in the enhanced response (estradiol production) of the future dominant follicle to the small rise in FSH that initiates the follicular wave. A growing body of evidence strongly suggests the involvement of the IGF system in the control of follicular development [11]. IGFs, acting synergistically with gonadotropins, are potent mitogenic and differentiation-promoting agents for ovarian follicular cells [12–14]. The binding of IGF to its receptors is strongly modulated by a family of six high-affinity IGF-binding proteins [15, 16]. It has been hypothesized that binding to low molecular weight IGFBPs (<40 kDa, mainly IGFBP-2, -4, and -5) inhibits the binding of IGFs to its receptors and its effects on gonadotropin-induced follicular growth and differentiation [12, 14, 17].
Although healthy, estrogen-active follicles and atretic follicles have similar intrafollicular concentrations of IGFs [12–14], the low molecular weight IGF-binding proteins are markedly lower in dominant, estrogen-active follicles than in atretic follicles [11, 17]. However, it is not known for any species whether changes in IGFBPs are a component of the mechanism that leads to selection of the dominant follicle. IGFBP-4 may be particularly relevant to follicular selection because 1) IGFBP-4 is the only IGFBP that consistently inhibits IGF actions in a variety of tissues and experimental conditions [18, 19]; 2) subordinate follicles have higher levels of IGFBP-4 than dominant, estrogen-active follicles [11, 17]; and 3) IGFBP-4 inhibits steroidogenesis by ovarian cells in vitro [20, 21]. In addition, Mihm et al. [22] reported recently that the follicle in the cohort of the first follicular wave of the bovine estrous cycle with the lowest intrafollicular concentration of IGFBP-4 always became the dominant follicle. The inhibitory effects of IGFBP-4 on IGF actions can be counteracted by specific IGFBP-4 proteases, as shown in a variety of systems [11, 23–27].
To begin to explore the role of changes in the IGF system in selection of the dominant follicle, we have tested the hypothesis that ovarian follicular dominance in cattle is associated with acquisition of IGFBP-4 proteolytic activity. This was achieved by determining IGFBP-4 proteolytic activity in follicular fluid samples obtained from dominant and subordinate follicles collected around the expected time of follicular selection during the first wave of follicular development of the estrous cycle. A preliminary report of these data has been presented elsewhere [28].
MATERIALS AND METHODS
Animals and Experimental Protocols
Holstein heifers with regular estrous cycles were used in accordance with procedures approved by the Cornell University Animal Care and Use Committee (protocol 86-214-99). The objective of experiment 1 was to determine whether IGFBP-4 protease activity is present in follicular fluid of preovulatory follicles. Heifers were injected with prostaglandin F2
(PGF2, 25 mg i.m.; Lutalyse, Pharmacia & Upjohn Co., Kalamazoo, MI) on Day 7 of the estrous cycle (Day 0 = day of estrus) to induce luteolysis. Luteolysis is followed by initiation of a follicular phase and differentiation of the dominant follicle of the first follicular wave into a preovulatory follicle. Follicular fluid from the preovulatory follicle (n = 4 heifers) was obtained after ovariectomy performed 24 h after injection of PGF2. In this experimental model, estrus and the LH surge occur about 48–60 h after PGF2 [29, 30]; thus follicles were obtained about 24–36 h before the expected time of the LH surge.
Experiment 2 was aimed at determining whether IGFBP-4 protease activity is associated with follicular dominance. Luteolysis, a follicular phase, and ovulation were induced by injecting heifers with PGF2 during the mid-luteal phase. Starting the day before PGF2 injection, the ovaries of each heifer were examined daily by transrectal ultrasonography using a 7.5 MHz transducer and a real time B-mode scanner (Aloka 500; Corometrics Medical Systems Inc., Wallingford, CT) to monitor follicular dynamics as described elsewhere [6].
Follicular fluid from the dominant and subordinate follicles was collected at specific times during the first follicular wave of the next estrous cycle. The day of emergence of the first follicular wave was designated as Day 0 of the wave and was retrospectively identified as the last day on which the dominant follicle was 4 or 5 mm in diameter. Dominant follicles were identified on the day of ovariectomy or aspiration as the largest follicle of the cohort, and estradiol concentrations in follicular fluid confirmed their status. In both experiments (2a and 2b), samples were collected on Day 2 of the wave from animals in which one follicle was1 mm larger in diameter (ultrasonographic measurement) than any other follicle, or on Day 3, when a distinct dominant follicle was present (
2 mm difference). In experiment 2a, follicular fluid was collected from the dominant and the two largest subordinate follicles after ovariectomy performed on Day 2 or 3 (n = 3 per group) of the first follicular wave, as reported elsewhere [6]. In experiment 2b, follicular fluid samples were collected from the dominant and the largest subordinate follicle by ultrasound-guided transvaginal follicular aspiration, performed on Day 2 or 3 (n = 4/5 per group) of the first follicular wave. Briefly, caudal epidural anesthesia was induced by injection of 5 ml of 2% lidocaine and the perineal area was washed with surgical scrub. A 5 MHz convex-array transducer was fitted with a 65-cm-long plastic handle equipped with a needle guide, to facilitate placement of the transducer and guidance of the aspiration needle to the vaginal fornix. While the transducer was positioned in the vaginal fornix, the free hand was used transrectally to position the ovary against the vaginal wall and an 18-gauge needle (60 cm long) was advanced through the vaginal wall and into the antrum of the targeted follicle. The follicular fluid was manually aspirated with a 5-ml syringe. Samples were centrifuged and stored at -80°C for later determinations.
Analysis of IGFBP-4 Proteolytic Activity
The ability of an intrafollicular protease to degrade IGFBP-4 was assessed by incubating 5 µl of follicular fluid plus substrate for 18 h at 37°C in a solution of 20 mM Tris (pH 7.5) containing 137 mM NaCl (TBS) and 0.1% BSA (final volume, 15 µl). In experiment 1, 20 or 50 ng of recombinant human (rh) IGFBP-4 (Austral Biologicals, San Ramon, CA) was used as substrate. In experiment 2, either rhIGFBP-4 cross-linked to [125I]IGF-II (experiment 2a), or rhIGFBP-4 (experiment 2b) was used as the source of IGFBP-4 substrate. IGFBP-4 was cross-linked to [125I]IGF-II (IGF-II from Austral Biologicals) by incubating 1 µg of rhIGFBP-4 with 30 µCi of 125I-labeled IGF-II (spec. act. 340–430 µCi/µg) as described previously [24]. After incubation, protease assay samples were subjected to SDS-PAGE followed by autoradiography when IGFBP-4 cross-linked to [125I]IGF-II was used as substrate. When recombinant IGFBP-4 was used as substrate, SDS-PAGE was followed by Western ligand blotting/phosphorimaging to quantify the percent of substrate loss and Western immunoblotting to detect specific proteolytic fragments.
Partial Characterization of the IGFBP-4 Protease Activity
To characterize the putative IGFBP-4 protease detected in follicular fluid from preovulatory follicles in experiment 1, rhIGFBP-4 was used as the substrate, followed by Western ligand blot analysis/phosphorimaging and Western immunoblotting. Time (in hours) and pH dependence of the IGFBP-4 degradation were assessed following incubation of 50 ng rhIGFBP-4 with follicular fluid from preovulatory follicles at 37°C. To provide an initial mechanistic classification of the IGFBP-4 protease activity present in bovine follicular fluid, the following set of standard protease inhibitors (Sigma Chemical Co., St. Louis, MO), corresponding to the four protease classes recognized by the International Union of Biochemistry [31], was used: 1,10 phenanthroline (5 mM, metalloprotease inhibitor), trans-epoxysuccinyl-L-leucylamido-(4-guanidino)-butane (E-64; 10 µM, cysteine protease inhibitor), aprotinin (2 µg/µl, serine protease inhibitor), pepstatin (0.1 mg/ml, aspartic protease inhibitor), and the nonspecific divalent cation chelator EDTA (5 mM). Inhibitors and doses used were chosen based on reported specificity and efficacy, respectively [31–33].
Western Ligand Blot Analysis
Western ligand blot analysis was performed as previously described [32]. Briefly, samples were subjected to electrophoresis in 12% SDS-PAGE under nonreducing conditions and transferred onto nitrocellulose membranes. The blotted proteins were incubated with 1.2 x 106 cpm of [125I]IGF-II (spec. act. 340–430 µCi/µg) overnight at 4°C. The nitrocellulose membranes were washed, air-dried, and subjected to autoradiography and phosphorimaging. Molecular weights of intact IGFBPs species were estimated by running the samples in parallel with protein molecular weight standards (2850–43 000 Mr range; Gibco BRL, Grand Island, NY) on the same gel.
Western Immunoblot Analysis
Immunoblot analysis was used to detect the presence of specific proteolytic fragments of IGFBP-4. The immunoblotting procedure was performed on the same blots previously subjected to ligand blot analysis. Stripping the [125I]IGF-II was accomplished by washing the nitrocellulose membranes with TBS containing 3% NP-40 for 2 h at 4°C. The blots were then reblocked with TBS containing 10% nonfat dry milk (NFDM) and 0.5% Tween 20 for 2 h at room temperature. Thereafter, nitrocellulose membranes were incubated for 1.5 h at room temperature in TBS containing 10% NFDM and 0.5% Tween and a rabbit polyclonal antibody against human IGFBP-4 (1:1000 dilution, Upstate Biotechnology Inc., Lake Placid, NY). The nitrocellulose membranes were washed and incubated with anti-rabbit horseradish peroxidase-labeled antibody (dilution 1:5000, Santa Cruz Biotechnology, Santa Cruz, CA) for 1 h at room temperature. Following washing, the membranes were incubated with enhanced chemiluminescence Western blotting reagents (Santa Cruz Biotechnology) for 1 min and exposed to x-ray films (Kodak, Rochester, NY). Molecular weights of intact and proteolytic fragments of IGFBP-4 were estimated by running the samples in parallel with protein molecular weight standards (2850–43 000 Mr range; Gibco BRL) on the same gel.
Phosphorimaging and Autoradiography
Phosphor screen autoradiography of ligand blots (72–96 h) involved scanning, digitalization, and processing of the image with a Fuji BAS1000 (Tokyo, Japan) phosphorimager. A molecular dynamics software package (Bio-Imaging Analyzer, BAS1000 MacBas, Fuji) was used to further process the data and obtain an image of the original radioactive sample in pixel values that were proportional to the amount of radioactivity. The background was subtracted from each sample. In addition, x-ray films were exposed to nitrocellulose membranes to obtain autoradiograms documenting the presence and size of bands in ligand and immunoblots.
Relative abundance (%) of IGFBP-4 was expressed as the ratio 100 x IIGFBP-4/IIGFBP, where IIGFBP-4 = intensity (arbitrary units) of the IGFBP-4 band and IIGFBP = sum of intensity (arbitrary units) for all IGFBPs (-2,-3, -4, and -5) of ligand blots for samples incubated for 0 h of each individual follicle (experiment 2b). For each follicular fluid sample, proteolytic activity was expressed as percent of substrate loss after 18 h of incubation relative to nonincubation (experiments 1 and 2b), or in arbitrary densitometric units when IGFBP-4 proteolytic fragments were analyzed by integrated laser densitometry (experiment 2a).
Radioimmunoassays and Protein Determinations
Duplicate aliquots (0.1–5.0 µl) of follicular fluid from each dominant and subordinate follicle were assayed without extraction for estradiol as described previously [6]. Total protein concentrations in samples of follicular fluid were measured in duplicate according to the Bradford method [34] using reagents purchased from Bio-Rad (Melville, NY).
Statistical Analysis
Follicular diameters were analyzed by repeated measures ANOVA using the MIXED procedure of SAS (SAS Institute, Cary, NC). Estradiol and protein concentrations in follicular fluid, relative abundance of IGFBP-4, and IGFBP-4 degradation were analyzed by ANOVA with a nested design to evaluate the effects of day of the follicular wave, animal (nested within day), follicle type (dominant or subordinate), and the interaction of day by follicle type using the GLM procedure of SAS. Bartlett test was used to test for heterogeneity of variance. When appropriate, logarithmic or square root transformations were used to yield variance homogeneity. AfterANOVA, individual means were compared by Scheffé multiple comparison test. Values are presented as means ± SEM of untransformed variables.
RESULTS
Experiment 1: Presence of an IGFBP-4 Protease in Follicular Fluid from Bovine Preovulatory Follicles
Western ligand blot analysis showed that follicular fluid from preovulatory follicles contains mainly IGFBP-3 (molecular weight 43 kDa) with no or barely detectable levels of IGFBP-4 (Fig. 1A, lane 2). When the presence of an IGFBP-4 protease activity was assessed by incubating follicular fluid from preovulatory follicles with 20 ng of rhIGFBP-4 for 18 h at 37°C, almost complete disappearance of added IGFBP-4 was observed (Fig. 1A, lane 3 vs. lane 6). To determine whether the decrease in the intact band of IGFBP-4 was accompanied by the appearance of specific proteolytic fragments, follicular fluid from the same preovulatory follicles was incubated with 50 ng rhIGFBP-4 for 18 h at 37°C and then subjected to Western immunoblotting. The almost complete disappearance of rhIGFBP-4, as observed by Western ligand blotting (Fig. 1A), was accompanied by the generation of two proteolytic fragments of molecular weights 18 and 14 kDa visualized by Western immunoblotting (Fig. 1B, lane 2 vs. 3), but not by Western ligand blotting. Degradation of rhIGFBP-4 and the resulting appearance of proteolytic fragments was inhibited by the divalent cation chelator EDTA (Fig. 1B, lane 3 vs. 4). These results show that follicular fluid from bovine preovulatory follicles obtained from natural estrous cycles contains a strong IGFBP-4 proteolytic activity, resulting in the generation of immunoreactive proteolytic fragments with no or negligible capacity to bind IGF in vitro.
|
Experiment 1: Characterization of the IGFBP-4 Protease Present in Bovine Follicular Fluid
To estimate the pH optimum of the IGFBP-4 protease, follicular fluid from preovulatory follicles was assessed for its ability to degrade rhIGFBP-4 over a broad range of pHs (pH 3–10). Proteolysis of IGFBP-4 was inhibited below pH 5; whereas weakly acid, neutral, and weakly basic pHs promoted IGFBP-4 proteolysis (Fig. 2A). Time-course experiments revealed increasing IGFBP-4 proteolysis during the first 4 h of incubation, with approximately 80% of the rhIGFBP-4 degraded after 18 h of incubation (Fig. 2B).
|
To provide initial mechanistic classification of the IGFBP-4 protease activity present in follicular fluid, we assessed its susceptibility to a set of standard protease inhibitors [31]. Follicular fluid from preovulatory follicles was preincubated for 1 h in the absence or presence of the inhibitors at the concentrations detailed in Materials and Methods. Fig. 3A shows a representative Western ligand blot of protease assay samples incubated in the absence or the presence of different inhibitors. Quantitative (phosphorimaging) results corresponding to the intensity of the remaining intact rhIGFBP-4 band were expressed as percentage of substrate loss (Fig. 3B). Preincubation with the nonspecific divalent cation chelator EDTA (5 mM) completely inhibited proteolytic activity (P < 0.001). Although this observation suggests the possibility of a metalloprotease (cation dependence), it does not rule out the possibility of a cation-activated or stabilized nonmetalloprotease. However, the metalloprotease inhibitor 1,10 phenanthroline (5 mM) proved equally effective (P < 0.001). Given that 1,10 phenanthroline has a much higher stability constant for Zn2+ (2.5 x 10-6 M-1) than for Ca2+ (3.2 x 10-1 M-1), this observation is virtually diagnostic for a Zn2+ metalloprotease [31]. Pretreatment with pepstatin A (0.1 mg/ml), an acetylated pentapeptide established as a highly selective aspartic protease inhibitor, proved ineffective (P > 0.05). Furthermore, the presence of IGFBP-4 protease activity at neutral and a basic pH argues against an aspartic protease because most aspartic proteases are inactive above pH 6.0 [31]. Similarly, E-64 (10 µM), a peptide epoxide recognized as a highly specific cysteine protease inhibitor, had no effect (P > 0.05) on the IGFBP-4 proteolytic activity. Finally, aprotinin (2 µg/µl), a serine protease inhibitor, proved ineffective (P > 0.05). Taken together, results from experiment 1 suggest that bovine follicular fluid from preovulatory follicles contains an IGFBP-4, neutral/basic pH-favoring, Zn2+ metalloprotease.
|
Experiment 2: Association Between IGFBP-4 Proteolytic Activity in Follicular Fluid and Follicular Dominance in Cattle
Experiment 2 was aimed at determining if the establishment of follicular dominance in cattle is associated with acquisition of IGFBP-4 protease activity. In the first phase of the experiment (experiment 2a), follicular fluid samples were collected by aspirating follicles obtained by ovariectomy soon after follicular selection (Day 2 of the follicular wave) or when a distinct dominant follicle was present (Day 3), as previously reported [6]. In the subsequent phase of the experiment (experiment 2b), the observations from experiment 2a were confirmed and extended with a larger number of samples, collected at the same follicular stages by a different method (ultrasound-guided follicular aspiration), and assayed for protease activity by a different method, as described in Materials and Methods.
Experiment 2a: Diameters and Intrafollicular Estradiol Concentrations of Follicles Obtained on Days 2 and 3 of the First Follicular Wave
Follicular diameters, as assessed by ultrasonography, and intrafollicular concentrations of estradiol are shown in Figure 4; further details are provided by Evans and Fortune [6]. On Day 2 of the follicular wave, the mean diameter of the dominant follicle was greater (P < 0.05) than the diameter of either of the two largest subordinate follicles, and the difference was even greater (P < 0.01) on Day 3. Estradiol concentrations were higher in follicular fluid from dominant vs. subordinate follicles on Days 2 and 3 (P < 0.05 and 0.01, respectively), confirming their status and indicating that the dominant follicle had been selected by Day 2 of the follicular wave.
|
Experiment 2a: IGFBP-4 Protease Activity in Follicular Fluid of Follicles Obtained on Days 2 and 3 of the First Follicular Wave
When IGFBP-4 proteolytic activity in follicular fluid was assessed using rhIGFBP-4 cross-linked to [125I]IGF-II as substrate, readily detectable levels of proteolysis were observed in dominant follicles collected on both Days 2 and 3 of the follicular wave, as shown by the appearance of specific proteolytic fragments of molecular weight 18 kDa (Fig. 5, A and B, lane 3 vs. 4). In contrast, incubation of the cross-linked substrate with follicular fluid from the first or second subordinate follicles, collected on Day 2 or 3 of the follicular wave, resulted in barely detectable levels of proteolytic fragments (Fig. 5, A and B, lanes 5 and 6). Densitometric analysis of proteolytic fragments showed that follicular fluid from dominant follicles had significantly higher (P < 0.05 Day 2; P < 0.01 Day 3) levels of IGFBP-4 proteolytic activity than follicular fluid from subordinate follicles as early as Day 2 of the follicular wave (Fig. 5, C).
|
Experiment 2b: Follicular Diameters and Intrafollicular Estradiol Concentrations of Follicles Sampled by Ultrasound-Guided Aspiration on Day 2 or 3 of the First Follicular Wave
The mean diameter of dominant follicles sampled on Day 3 was greater (P < 0.05) than that of all other follicles and on Day 2, the diameter of the dominant follicle was greater (P < 0.05) than the diameter of the largest subordinate follicle (Fig. 6). Estradiol concentrations in follicular fluid obtained by ultrasound-guided aspiration were higher in dominant than in subordinate follicles (P < 0.05 Day 2; P < 0.01 Day 3), confirming their status and indicating that the dominant follicle had been selected by Day 2 of the follicular wave (Fig. 6). Protein concentration was measured in follicular fluid to verify that potential differences in proteolytic activity in samples obtained by ultrasound-guided follicular aspiration were not due to unequal protein concentrations in the samples. No differences (P > 0.05) were observed in protein concentrations in follicular fluid samples (Day 2 dominant follicles = 45 ± 5, subordinate follicles = 41 ± 4; Day 3 dominant follicles = 51 ± 3, subordinate follicles = 50 ± 7 mg/ml; mean ± SEM).
|
Experiment 2b: IGFBP-4 and IGFBP-4 Proteolytic Activity in Follicular Fluid of Follicles Sampled by Ultrasound-Guided Aspiration on Day 2 or 3 of the First Follicular Wave
Relative abundance of IGFBP-4
The relative abundance of IGFBP-4 in follicular fluid collected by ultrasound-guided aspiration for experiment 2b was assessed by ligand blot analysis. Figures 7A and 8A show representative ligand blots of samples from dominant and subordinate follicles collected on Days 2 and 3, respectively. Follicular fluid collected from dominant, estrogen-active follicles on Day 2 or 3 contained IGFBP-3 and barely detectable levels of low molecular weight IGFBPs, including IGFBP-4 (Fig. 7A and 8A, lanes 4–6). In contrast, follicular fluid from subordinate follicles collected on both days of the follicular wave contained readily detectable levels of low molecular weight IGFBPs, including IGFBP-4 (Figs. 7A and 8A, lanes 10 and 12). No intact IGFBP-4 of endogenous origin was detected in immunoblots of follicular fluid from dominant follicles (Figs. 7B and 8B, lanes 4–6). In contrast, two bands of endogenous, intact IGFBP-4 were observed in immunoblots of follicular fluid from subordinate follicles (Figs. 7B and 8B, lanes 10–12) corresponding, presumably, to glycosylated and unglycosylated forms (molecular weights 28 and 24 kDa, respectively). Quantitative results, derived from phosphorimaging of ligand blots, revealed higher (P < 0.01) abundance of IGFBP-4 in follicular fluid from subordinate vs. dominant follicles collected either on Day 2 or 3 of the follicular wave (Fig. 9, upper panel).
|
|
IGFBP-4 proteolytic activity IGFBP-4 proteolytic activity in follicular fluid samples obtained by ultrasound-guided aspiration was assessed by incubating 5 µl of follicular fluid with 50 ng rhIGFBP-4, followed by Western ligand blotting/phosphorimaging and Western immunoblotting to detect specific proteolytic fragments. Figures 7 and 8 show representative ligand (panel A) and immunoblots (panel B) of protease assays run on follicular fluid samples from dominant and subordinate follicles collected on Day 2 or 3, respectively. On both Days 2 and 3 of the follicular wave, follicular fluid from the dominant follicle contained higher levels of IGFBP-4 proteolytic activity than the companion subordinate follicle, as evidenced by the significant reduction (Day 2) or almost complete disappearance (Day 3) of the band corresponding to the intact rhIGFBP-4 (Figs. 7A and 8A, lanes 1 and 3 vs. 2). In addition, the reduction in the intensity of the IGFBP-4 band detected by ligand blot analysis was accompanied by the generation of specific proteolytic fragments as determined by Western immunoblotting (Figs. 7B and 8B, lanes 1 and 3 vs. 2). In contrast, follicular fluid from subordinate follicles failed to degrade exogenous rhIGFBP-4 or endogenous IGFBP-4 to any significant extent (Figs. 7A and 8A, lanes 7 and 9 vs. 8 and lanes 10 and 12 vs. 11, respectively). As expected, specific proteolytic fragments were barely detectable when rhIGFBP-4 was incubated with follicular fluid from subordinate follicles (Figs. 7B and 8B, lanes 7 and 9 vs. 8, respectively). Quantitative results, derived from phosphorimaging of ligand blots and expressed as percent of rhIGFBP-4 degradation, are shown in Figure 9 (bottom panel). IGFBP-4 proteolysis was 3.5-fold (P < 0.01) and 4.8-fold (P < 0.001) higher in follicular fluid from dominant vs. subordinate follicles on Days 2 and 3, respectively.
DISCUSSION
The results show that dominant ovarian follicles have an IGFBP-4 protease activity that is negligible in subordinate follicles of the same cohort. The finding that the protease activity is detected very early during differentiation of the bovine dominant follicle and persists through preovulatory development suggests a critical role for IGFBPs and their proteases in follicular fate. When dominant follicles are first detected as slightly larger than companion subordinate follicles, their capacity to secrete estradiol is much greater than that of subordinates, but they do not differ in other characteristics examined [6]. This points to the acquisition of IGFBP-4 proteolytic activity by dominant, but not subordinate follicles, shown here for the first time, as a critical determinant of dominance.
Insulin-like growth factors are important mediators of some of the actions of FSH, a major survival factor for antral follicles [35], but IGFs bound to IGFBP-4 cannot bind to IGF receptors and are thus biologically inactive [26]. Therefore, we propose that IGFBP-4 protease plays a critical role in the establishment of follicular dominance in cattle by reducing intrafollicular levels of IGFBP-4 and thus providing more bioavailable IGFs to support continued growth and development of the selected follicle in response to FSH. Using an experimental model of codominant follicles induced by low doses of recombinant bovine FSH, we recently showed that IGFBP-4 protease activity in codominant bovine follicles was similar to that of single dominant follicles of control heifers, but fourfold higher than that of subordinate follicles [36]. Taken together, these observations would suggest that an FSH-inducible IGFBP-4 protease may be a determinant of follicular fate.
When follicular fluid collected from preovulatory follicles was assayed for IGFBP-4 proteolysis, almost 80% of the added rhIGFBP-4 was degraded within 18 h of incubation. These results confirm our preliminary report of an IGFBP-4 protease in bovine preovulatory follicles [28] and are consistent with a more recent report of IGFBP-4 proteolytic activity in bovine follicular fluid collected after administration of gonadotropins to induce superovulation [37]. The fact that IGFBP-4 is subjected to proteolysis during incubation with follicular fluid from preovulatory follicles suggests that levels of IGFBP-4 in estrogen-active follicles are regulated, at least in part, by an IGFBP-4-specific protease. Other studies have provided evidence for IGFBP-4 proteolytic activity in follicular fluid from dominant/estrogen-active, but not subordinate/atretic follicles, collected well after follicular selection, from women [32, 38, 39], domestic animals [33, 37, 40, 41], and rats [20], suggesting a conserved role for this protease in ovarian follicular function. The present results, combined with those of recent studies on the expression of mRNAs encoding IGFBP-2 and -4 [42] and IGF-II and type I IGF receptor [43], confirm the existence of a complete intrafollicular IGF system in the bovine species, i.e., local production of a ligand, presence of a receptor to mediate IGF actions, local production of IGFBPs, and at least one IGFBP protease contributing to the regulation of IGF bioavailability.
Characterization studies in experiment 1 showed that the IGFBP-4 protease activity in bovine follicular fluid is a neutral/basic pH-favoring, Zn2+ metalloprotease. In agreement with our results, IGFBP-4 proteases with similar characteristics have been described in follicular fluid from several species [11] as well as in a variety of systems including human fibroblasts [18, 23], bone cells [23, 26], endometrial stromal cells [24], decidual cells [44], and serum of pregnant women [27]. However, IGFBP-4 degradation by serine proteases has also been reported in a rat neuronal cell line [45] and porcine smooth muscle cells [46]. Cleavage analysis of IGFBP-4 proteolysis has shown that IGFBP-4 proteases characterized as metalloproteases cleave IGFBP-4 at Met135, Lys136 [18, 26, 27], whereas IGFBP-4 proteases characterized as serine proteases cleave IGFBP-4 at Lys120, His121 [46, 47]. Whether or not these proteases are different, some common traits are readily apparent: 1) all are IGF-dependent; ligand binding may result in a conformational change in the IGFBP-4 molecule making the cleavage site more accessible to the protease; 2) the IGFBP-4 proteolysis results in the generation of N-terminal and C-terminal immunoreactive fragments of molecular weights 18 and 14 kDa, respectively; and 3) proteolytic fragments exhibit highly reduced or negligible IGF binding capabilities, and therefore do not inhibit IGF-induced cell proliferation, differentiation, or both. The last assertion is supported by studies showing that the N-terminal sequence (Leu72–Ser91) and the C-terminal sequence (Cys205–Val214) of IGFBP-4 are necessary to form the high-affinity IGF binding domain [19], and that mutant, protease-resistant forms of IGFBP-4, with unaffected IGF binding capacity, are more potent than wild-type IGFBP-4 in inhibiting IGF-induced actions [18, 26, 46]. In the present study, the decline in the intensity of the band corresponding to intact rhIGFBP-4 was accompanied by the appearance of specific proteolytic fragments of molecular weights 18 and 14 kDa, detectable by Western immunoblotting but not by ligand blotting (experiments 1 and 2b), or by the appearance of radioactive fragments of molecular weight 18 kDa (IGFBP-4 cross-linked to [125I]IGF-II used as substrate, experiment 2a), only when follicular fluid from estrogen-active follicles was assayed. Although the biological consequences of intrafollicular IGFBP-4 proteolysis in vivo still remain unknown, in vitro studies have shown that intact IGFBP-4 [20, 21, 38], but not IGFBP-4 proteolytic fragments [38], inhibits estradiol production by isolated granulosa cells.
Almost nothing is known about the specificity and molecular identity of the IGFBP-4 proteolytic activity present in bovine follicular fluid. An earlier study showed, interestingly, strong pregnancy-associated plasma protein-A (PAPP-A) immunoreactivity in follicular fluid from hyperstimulated human ovarian follicles [48]. More recently, PAPP-A, a protein previously purified from serum of pregnant women, was reported to be identical to the IGF-dependent IGFBP-4 protease present in human follicular fluid [39]. We have shown in preliminary studies that follicular fluid from bovine preovulatory follicles failed to degrade rhIGFBP-2 to any significant extent; in contrast, rhIGFBP-5 was significantly degraded (data not shown). Based on the present results, the possibility of a common proteolytic activity for both IGFBP-4 and -5 cannot be ruled out. However, such a scenario seems unlikely considering that 1) each IGFBP contains amino- and carboxyl-terminal regions that are homologous among the different IGFBPs and a middle region that has little shared sequence between the various IGFBPs [49]; 2) regardless of the cleavage site, the fact that all of the IGFBP-4 proteases recognize the middle nonhomologous region for proteolysis underscores their high specificity [18, 47]; 3) a structural motif located on the middle nonhomologous region of IGFBP-4, which does not contain the cleavage site, seems to be essential for proteolytic degradation of IGFBP-4 [26]; 4) specific inhibition and immunodepletion of IGFBP-4 (but not IGFBP-5) protease activity by polyclonal antibodies against PAPP-A (IGFBP-4 protease) was shown in human follicular fluid [39]; and 5) the size of the proteolytic products and the inhibitor profile of the bovine IGFBP-4 protease reported in the present study suggest strong similarities to the specific IGFBP-4 protease found in human follicular fluid [32, 39]. Additional experiments to further characterize the nature of the IGFBP-4 protease present in bovine follicular fluid are needed.
Determination of the factors and mechanisms responsible for the establishment of the ovulatory quota typical of each species has proved elusive. Cattle provide a unique animal model to study follicular selection; the regular succession of follicular waves occurring at 7- to 8-day intervals throughout the estrous cycle and the feasibility of tracking growth and regression of individual follicles 4 mm in diameter allow sampling or collection of dominant and subordinate follicles as close as possible to the time of selection of the dominant follicle. Accordingly, in this and another study [6], we have examined several characteristics of dominant and subordinate follicles collected on Days 2 and 3 of the first follicular wave of the cycle. In both cases, the selected dominant follicle was distinguishable from its subordinate cohort as early as Day 2 of the wave by a slight advantage in diameter and increased capacity to produce estradiol. Surprisingly, no differences were observed on Day 2 between the dominant and the largest subordinate follicle in levels of mRNAs for gonadotropin receptors and steroidogenic enzymes [6]. However, some differences between dominant and subordinate follicles were observed on Day 3 of the wave in our experiment [6] and between recruited (prior to selection) and dominant (after selection) follicles in other studies [50]. In light of this paucity of differences between dominant and subordinate follicles close to the time of selection, it is of considerable interest that we (present study) and others [51, 52] have consistently observed lower intrafollicular levels of low molecular weight IGFBPs in bovine dominant follicles compared to companion subordinate follicles. It is interesting that when the concentrations of various biochemical markers were determined for follicles on Day 1 of the first follicular wave (before morphological selection), the follicle in the cohort with the lowest intrafollicular concentration of IGFBP-4 always became the dominant follicle [22]. In cattle, IGFBP-4 mRNA expression is restricted to theca tissue of antral follicles, with no difference in expression levels between follicles of various size classes [42]. That finding, together with the increased IGFBP-4 proteolytic activity in dominant follicles reported in the present study, suggests that post-translational modifications, rather than changes in IGFBP-4 gene expression, are responsible for the decreased levels of IGFBP-4 associated with follicular growth and selection for dominance. It is tempting to speculate that the lower intrafollicular levels of IGFBP-4 observed in preselection follicles [22] are a consequence of increased IGFBP-4 proteolysis in the follicle destined to be dominant. Further studies are necessary to assess the role of the IGFBP-4 protease in selection of the dominant follicle and its usefulness as a predictor of follicular dominance.
In conclusion, these results provide the first evidence that IGFBP-4 proteolytic activity is higher in dominant, estrogen-active follicles than in subordinate follicles of the same cohort, as early as Day 2 of the first follicular wave and strongly suggest a role for IGFBP-4 proteolytic activity in the establishment of ovarian follicular dominance in cattle, and perhaps other mammalian species. Further evidence in support of this conclusion was derived from the use of an experimental model that produces codominant follicles, as described in the companion paper [36].
|
ACKNOWLEDGMENTS
We thank B.S. Hansen for her excellent technical assistance, Dr. E. Zambrano for assistance in ultrasound-guided follicular aspirations, D. Bianchi for care of the animals, Dr. G.D. Niswender for the estradiol antiserum, and Drs. Mark S. Roberson and Phillip Bridges for their critical reading of the manuscript.
FOOTNOTES
First decision: 13 November 2000.
1 Supported by National Institutes of Health grants HD38276 to J.E.F. and HD31579 to L.C.G., a Lalor Foundation Fellowship to Y.A.C., and a Fulbright-Universidad Nacional de Río Cuarto (Argentina) Fellowship to G.M.R. 
2 Correspondence: J.E. Fortune, Department of Biomedical Sciences, Cornell University, Ithaca, NY 14853. FAX: 607 253 3476;jf11{at}cornell.edu
3 Current address: ZymoGenetics, 1201 Eastlake Avenue East, Seattle, WA 98102.
4 Current address: Department of Animal Science and Production, University College Dublin, Dublin 4, Ireland.
Accepted: February 14, 2001.
Received: September 26, 2000.
REFERENCES
4, 5 isomerase (3ß-HSD) during recruitment and selection of bovine ovarian follicles: identification of dominant follicles by expression of 3ß-HSD mRNA within the granulosa cell layer. Biol Reprod 1997; 56:1466-1473[Abstract]This article has been cited by other articles:
|
|
 |
|
  S. M Quirk, R. G Cowan, and R. M Harman The susceptibility of granulosa cells to apoptosis is influenced by oestradiol and the cell cycle. J. Endocrinol., June 1, 2006; 189(3): 441 - 453. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Matsui, B. Sonntag, S. S. Hwang, T. Byerly, A. Hourvitz, E. Y. Adashi, S. Shimasaki, and G. F. Erickson Pregnancy-Associated Plasma Protein-A Production in Rat Granulosa Cells: Stimulation by Follicle-Stimulating Hormone and Inhibition by the Oocyte-Derived Bone Morphogenetic Protein-15 Endocrinology, August 1, 2004; 145(8): 3686 - 3695. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. J. Spicer Proteolytic Degradation of Insulin-Like Growth Factor Binding Proteins by Ovarian Follicles: A Control Mechanism for Selection of Dominant Follicles Biol Reprod, May 1, 2004; 70(5): 1223 - 1230. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C.-L. Hu, R. G. Cowan, R. M. Harman, and S. M. Quirk Cell Cycle Progression and Activation of Akt Kinase Are Required for Insulin-Like Growth Factor I-Mediated Suppression of Apoptosis in Granulosa Cells Mol. Endocrinol., February 1, 2004; 18(2): 326 - 338. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. M. Quirk, R. G. Cowan, R. M. Harman, C.-L. Hu, and D. A. Porter Ovarian follicular growth and atresia: The relationship between cell proliferation and survival J Anim Sci, January 1, 2004; 82(13_suppl): E40 - 52. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Webb, P. C. Garnsworthy, J.-G. Gong, and D. G. Armstrong Control of follicular growth: Local interactions and nutritional influences J Anim Sci, January 1, 2004; 82(13_suppl): E63 - 74. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. J. Roberts and S. E. Echternkamp Insulin-like growth factor binding proteins in granulosa and thecal cells from bovine ovarian follicles at different stages of development J Anim Sci, November 1, 2003; 81(11): 2826 - 2839. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
B. Sisco, L. J. Hagemann, A. N. Shelling, and P. L. Pfeffer Isolation of Genes Differentially Expressed in Dominant and Subordinate Bovine Follicles Endocrinology, September 1, 2003; 144(9): 3904 - 3913. [Abstract] [Full Text] [PDF]
|
|
|
|
