BIOLOGY OF REPRODUCTION 56, 508-515 (1997)
Activin-Attenuated Expression of Transcripts Encoding Granulosa Cell-Derived Insulin-Like Growth Factor Binding Proteins 4 and 5 in the Rat: A Putative Antiatretic Effect' DooSeok Choi, 3,4 Richard M. Rohan, 4 Ron G. Rosenfeld,5 Tomoko Matsumoto,5 Sharron E. Gargosky,5 and Eli Y. Adashi 2 4,
Division of Reproductive Endocrinology, 4 Department of Obstetrics and Gynecology, University of Maryland School of Medicine, Baltimore, Maryland 21201 Department of Pediatrics,5 Oregon Health Sciences University, Portland, Oregon 97201 activin may play an antiatretic role in the dynamic process of follicular selection.
ABSTRACT Given the suggestion that intraovarian insulin-like growth factor (IGF)-binding proteins (IGFBPs) may constitute markers of follicular atresia, we investigated the possibility that activin, a putative antiatretic principle, may modulate granulosa cellderived IGFBPs. Untreated granulosa cells cultured for 72 h exhibited a progressive increase in the steady-state levels of transcripts corresponding to IGFBP-4 and IGFBP-5 (1.5-fold and 12-fold, respectively). Transcript levels corresponding to IGFBP-5 were consistently higher than their IGFBP-4 counterparts. Treatment with activin-A (50 ng/ml) for 72 h produced significant ( < 0.05) decrements in the steady-state levels of IGFBP-4 and IGFBP-5 transcripts (46% and 79%, respectively) as compared to controls. Thus, treatment with activin-A appears to be capable of blocking the spontaneous increase in IGFBP-4 and IGFBP-5 transcripts exhibited by untreated cultured granulosa cells. Consistent activin-A-induced decrements were also observed in the accumulation of the IGFBP-5 (but not the IGFBP-4) protein. Dose-response analysis revealed monophasic dose dependence (half maximal inhibitory doses of 16.2 and 7.8 ng/ml for IGFBP-5 and IGFBP-4 transcripts, respectively). The addition of increasing concentrations of the putative activinbinding protein, follistatin, produced dose-dependent reversal of the activin-A effect on IGFBP transcripts (IGFBP-5 > IGFBP-4). Activin-B was as effective as activin-A in reducing IGFBP-4 transcripts (31% decrement, p < 0.05) whereas it had little or no effect on IGFBP-5 transcripts (21% decrement, p > 0.1). No apparent effect was observed for the corresponding proteins. Activin-A action was specific in that treatment with transforming growth factor (TGF)-P,1 , inhibin-A, or Miillerian-inhibiting substance (MIS)-all related peptides-failed to produce statistically significant alterations in the steady-state levels of IGFBP-4 and IGFBP-5 transcripts. Taken together, these observations reveal that activin-A exerts a substantial, relatively rapid, follistatin-neutralizable, dose- and time-dependent inhibitory effect on granulosa cell-derived IGFBP transcripts (IGFBP-5 > IGFBP-4). Other members of the TGFIB superfamily (e.g., inhibin-A, TGFI3, and MIS) were without significant effect on the expression of IGFBP-4 and IGFBP-5. To the extent that the inhibition of IGFBP-4 and IGFBP-5 expression is associated with, and possibly causally related to, the promotion of follicular health, the present observations are in keeping with the proposition that
INTRODUCTION
Accepted September 5, 1996. Received June 6, 1996. 'Supported in part by NIH Research Grants HD-19998 and HD-30288 (E.Y.A.) and HD-28703 (R.G.R.). 2Correspondence: Division of Reproductive Sciences, Department of Obstetrics and Gynecology, University of Utah Health Sciences Center, Suite 23200, 50 N.Medical Drive, Salt Lake City, UT 84132. FAX: (801) 539-1770; e-mail:
[email protected] 3 Current address: Department of Obstetrics/Gynecology, Samsung Medical Center, II Won-dong Kangnam-Ku, Seoul, Korea. 508
Among potential intraovarian regulators, insulin-like growth factor (IGF)-I has been the subject of particularly intense investigation [1]. In this context, studies relevant to the rat disclosed the existence of an intraovarian autocrine control loop, wherein IGF-I may serve as the central signal, and the granulosa cell as its site of production [2-5], reception [6-9], sequestration [10-19], and action [20-24]. Stated differently, there exists a complete intrafollicular IGF-I system, replete with a ligand (IGF-I), a receptor (type I), and IGF-binding proteins (IGFBPs 4 and 5). According to current views, the intrafollicular IGF-I system may be a determinant of follicular fate [25]. It is hypothesized that the net (i.e., bioavailable) intrafollicular IGF-I activity (as defined by the intrafollicular IGF-I: IGFBP ratio) may effect follicular selection by way of FSH amplification [26]. Such notions are in keeping with the suggestion that intrafollicular IGFBP expression is inversely related to the health status of the follicle [16-18]. Given the potentially central role of intrafollicular IGFBPs in determining the net bioavailable intrafollicular IGF-I content [19], serious consideration must be given to their possible involvement in the process of follicular selection. To gain further insight into the workings of intrafollicular IGFBPs, we investigated the possibility that activin, an established regulator of ovarian function [27], as well as related peptides may be in a position to modulate granulosa cell-derived IGFBPs. Activin, a granulosa cell-derived principle, has been suggested as an autocrine regulator of granulosa cells, promoting differentiation during the preantral and early antral stages of folliculogenesis and preventing premature luteinization in the later stages of antral follicle development [27]. Overall, in vitro observations support the view that activin may serve as a promoter of gonadotropin-supported follicular development [27, 28]. However, data derived from in vivo experiments reveal that activin-A reduces the proportion of follicles > 350 Jxm in diameter and reduces thymidine incorporation rates into granulosa cells [29]. The latter data were taken to mean that activin-A may be an atretic agent. Although several explanations have been put forth in an effort to reconcile these diametrically opposed outlooks, no consensus has yet been reached. It is the purpose of this communication to indirectly reassess the potential relevance of activin to the process of follicular selection by examining its impact on granulosa cell-derived IGFBPs, which have been proposed to play a role in follicular selection [16-18].
ACTIVIN AND OVARIAN IGFBPs
509
Immature (21-23 days old) intact Sprague-Dawley female rats purchased from Zivic-Miller Laboratories (Zelienople, PA) were killed by CO 2 asphyxiation when they were 24-26 days old, following three days of diethylstilbestrol (DES) priming achieved by the s.c. placement of silastic DES-containing capsules by the vendor. The project was approved by the Institutional Animal Care and Use Committee.
A portion of the rat IGFBP-5 mRNA was amplified by RT-PCR of total RNA obtained from immature rat ovaries as described for rat IGFBP-6 [301. Amplification of a 432-bp product was achieved with an oligonucleotide primer (5'-CTGGATCC ACCAGCAAACCAAGATAGA-3') specific to the published sequence of the rat IGFBP-5 mRNA [31] in conjunction with an oligonucleotide primer common to all IGFBP mRNAs [30]. The PCR product was cloned in the sense and antisense orientations in pCR1000 vector (Invitrogen, San Diego, CA), and its identity was verified by DNA sequencing.
Hormones and Reagents
RNA Extraction
Recombinant human activin-A and activin-B as well as recombinant human inhibin-A and recombinant human TGFPI3 were generously contributed by Jennie P Mather, Genentech, Inc. (South San Francisco, CA). Recombinant human Millerian inhibiting substance (MIS) and plasmincleaved recombinant human MIS was the generous gift of Richard Cate, BioGen (Cambridge, MA). Human follistatin (lot #B3904) was provided by Nicholas Ling through the NHPP, NIDDK, NIH (Bethesda, MD). Basic fibroblast growth factor (FGF) was generously provided by Andreas Sommer, Synergen, Inc. (Boulder, CO). Synthetic IGF-II was generously provided by the late C.H. Li, University of California at San Francisco (San Francisco, CA). McCoy's 5a medium (modified, without serum), penicillin-streptomycin solution, and L-glutamine were obtained from GibcoBRL Life Technologies, Inc. (Gaithersburg, MD).
Total cellular RNA was extracted with RNAZOL-B as recommended by the manufacturer (Tel Test, Inc., Friendswood, TX). Precipitated RNA from individual culture tubes was suspended in 8 xl diethyl pyrocarbonate-treated distilled deionized H 20. A constant fraction (1.4 1) of each sample was electrophoresed in 1% agarose/2.2 M formaldehyde gels and stained with ethidium bromide in order to visually assess the integrity and intensity of 28S and 18S rRNAs. The entire remainder was used for Northern blot analysis.
MATERIALS AND METHODS Animals
In Vitro Studies Granulosa cells from DES-primed intact immature rats were cultured in 12 x 75-mm polypropylene tubes (Falcon Plastics, Los Angeles, CA) unless indicated otherwise at a density of 2 x 106 viable cells/tube. Obtained by follicular puncture as previously described [6], the cells were maintained for up to 72 h in 2 ml of serum-free McCoy's 5a medium supplemented with 2 mM L-glutamine, 100 U/ml penicillin, and 100 jIg/ml streptomycin sulfate. A temperature of 37°C and a water-saturated atmosphere of 95% air: 5% CO 2 were maintained throughout. All reagents were dissolved in sterile culture media and added in 50-pld aliquots. At the conclusion of the incubation period, the media were removed and stored at -20°C for subsequent Western ligand blotting where indicated. The corresponding cellular pellets in turn were subjected to RNA extraction as described below, IGFBP-4 and IGFBP-5 Probe Construction Using the published nucleotide sequence of the rat IGFBP-4 cDNA [13], two synthetic oligonucleotide primers were designed for reverse-transcriptase (RT) polymerase chain reaction (RT-PCR). The sequence of the 5' primer was 5'-ATAGGATCCTGGGCGACGAAGCCATCCA-3'. The sequence of the 3' primer was 5'-AGCGAATTCTGGCAGTCCAGCTCCCCCTT-3'. Total RNA (1 pLg) from rat liver was reverse-transcribed with the 3' primer using MMLV-RT (Gibco BRL Life Technologies, Inc.). The resulting cDNA was then amplified by PCR. The expected product (697 bp) was gel-purified, digested with EcoRI and BamHI (restriction site contained in the primers), and cloned into pGEM7Zf(+) plasmid (Promega, Madison, WI). The identity of the cloned insert was verified by Hae III digestion and partial DNA sequencing.
Preparation of 32P-Labeled Probes The IGFBP-4 and IGFBP-5 inserts were excised by digestion with the appropriate restriction endonucleases. Inserts were then purified by agarose gel electrophoresis followed by GELase (Epicentre Technologies, Madison, WI) digestion. DNA was radioactively labeled by the random priming method (Pharmacia, Piscataway, NJ) using alpha32 P-dCTP (NEN, Boston, MA). Probes were purified by gel exclusion column chromatography (Pharmacia) before hybridization. The blots were also probed with a 600-bp EcoRI-BamHI fragment of the hamster Chinese clone B (CHOB) cDNA originally described by Harpold et al. [32]. The CHOB mRNA was used to normalize the IGFBP-4 and IGFBP-5 data for possible variation in RNA loads and/or transfer. This strategy was based on the validation of CHOB as a largely hormonally independent transcript previously used to normalize Northern blots of RNAs from rat ovaries after gonadotropin treatment [33, 34]; from rat liver, heart, muscle, and fat after streptozotocin-induced diabetes [35, 36]; and from rat liver after glucocorticoid treatment [37]. Northern Blot Analysis Northern blotting was carried out as previously described [38]. Western Ligand Blots Conditioned media were electrophoresed on SDS-PAGE (12.5%) under nonreducing conditions. The size-fractionated proteins were then electroblotted onto nitrocellulose for 1 h. Thereafter, the filter-immobilized proteins were blocked, incubated overnight at 4C with 1 X 106 cpm of [125 I]IGF-I and II, washed, and visualized by autoradiography according to the method of Hossenlopp et al. [39]. Synthetic IGF-II was iodinated by a modification of the chloramine-T technique to specific activities of up to 250 ipCi/,lg. The iodinated peptide was then purified by gel filtration over a Sephadex G-50 column (1.0 X 120 cm) at 4°C and eluted with 100 mM Hepes buffer, pH 7.4, containing 0.5% BSA, 120 mM NaC1, 1.2 mM MgSO 4, 5 mM
510
CHOI ET AL.
FIG. 1. Granulosa cell IGFBP gene expression: effect of treatment with activin-A. Granulosa cells were cultured under serum-free conditions in the absence or presence of activin-A (50 ng/ml) for 72 h. After incubation, cellular pellets were subjected to RNA extraction followed by Northern blot analysis. The blots were simultaneously probed with a hamster CHOB cDNA to normalize for possible variation in RNA loading or transfer. The intensity of RNA bands was quantified as described in Materials SE of 18 experiments and Methods. The top panel depicts the mean compiled from data displayed in Figures 2 (n = 4), 3 (n = 3), 5 (n = 3), 6 (n = 3), 7 (n = 2), and 9 (n = 3), which have been assembled in the interest of enhanced statistical power. The lower panel shows an autoradiograph from a representative experiment.
KC1, 50 mM Na acetate, and 10 mM dextrose. Molecular weights were estimated using prestained protein standards. Densitometric analysis of autoradiograms was performed on an Ultrascan XL laser densitometer (LKB, Stockholm, Sweden).
Statistical Analysis Each experiment contained duplicate or triplicate samples. The amount of IGFBP was corrected for the amount of CHOB (IGFBP:CHOB ratio) and averaged for each replicate. Within each experiment, the IGFBP:CHOB ratio was further normalized to the 0-h (see Fig. 4) or 72-h (all others) control. Differences between treatment groups and the normalizing group were assessed by paired Student's t-tests. Differences between other treatment groups were assessed by nonpaired Student's t-tests. Median effective doses were calculated by regression analysis of the log dose versus the IGFBP:CHOB ratio using SAS (Statistical Analysis Systems, Cary, NC) software. RESULTS To examine the ability of activin-A to regulate the expression of granulosa cell IGFBP-4 and -5 under in vitro circumstances, granulosa cells from immature DES-primed rats (2 X 106 viable cells/tube) were cultured under serumfree conditions in the absence or presence of activin-A (50 ng/ml) for 72 h. As shown (Fig. 1; n = 18), treatment with activin-A produced significant (p < 0.05) decrements in the steady-state levels of transcripts corresponding to IGFBP-4 and -5 (46% and 79%, respectively). Thus, the overall maximal suppression of IGFBP-5 transcripts (79%; p = 0.001) in response to 50 ng/ml of activin-A (the dose used in all subsequent experiments) far exceeded the value observed for IGFBP-4 (46% inhibition; p = 0.001). Dose response analysis (Fig. 2; n = 3-4) revealed largely monophasic
FIG. 2. Activin-A-attenuated granulosa cell IGFBP expression: dose-dependence. Granulosa cells were cultured under serum-free conditions in the absence or presence of increasing concentrations (0-100 ng/ml) of activin-A. Following incubation, cellular pellets were subjected to RNA extraction and Northern blot analysis. The top panel depicts the mean ± SE of 4 experiments. Data corresponding only to the 30- and 100-ng/ml doses of activin-A reflect a total of 3 experiments. The lower panel shows an autoradiograph of a replicate from an experiment.
dose-dependent inhibition of IGFBP-5 expression, an effect characterized by a median effective inhibitory dose (EC 50) of 16.2 ng/ml. The comparable EC50 figure for IGFBP-4 (7.8 ng/ml) was judged to be less reliable in light of the narrower range of inhibition. As shown below (see Fig. 8), treatment with activin-A was capable of producing a decrease in IGFBP-5 protein levels. The levels of the IGFBP-4 protein were not consistently altered in this study even though fluctuation of mRNA levels was documented. To examine the time dependence of the activin-A effect, granulosa cells (2 x 106 viable cells/tube) were cultured under serum-free conditions in the absence or presence of activin-A (50 ng/ml) for up to 72 h. In the absence of treatment (Fig. 3), the steady-state levels of transcripts corresponding to IGFBP-4 and IGFBP-5 displayed time-dependent increments (1.5-fold and 12-fold, respectively). However, treatment with activin-A produced nearly maximal suppression (relative to untreated controls) of IGFBP-5 transcripts as early as 24 h into the experiment, with no further decrements being observed thereafter. However, suppression of IGFBP-4 transcripts was not significant until 72 h. Qualitatively comparable results were noted for the media content of the IGFBP-5 (but not IGFBP-4) as assessed by ligand blotting. Parallel assessment of the accumulation of the corresponding proteins in a single experiment revealed activin-A to attenuate the accumulation of IGFBP-5 as early as 48 h into the experiment, with limited inconsistent alterations observed for the IGFBP-4 protein (not shown). To better characterize the early action of activin-A, we set out to evaluate the effect of treatment with activin-A on IGFBP-4 and IGFBP-5 transcripts during the first 24 h of culture. Granulosa cells (2 x 106 viable cells/tube) were cultured under serum-free conditions for the duration indicated in the absence or presence of activin-A (50 ng/ml). In cultures of untreated granulosa cells (Fig. 4; n = 3),
ACTIVIN AND OVARIAN IGFBPs
FIG. 3. Activin-A-attenuated granulosa cell IGFBP expression: time-dependence. Granulosa cells were cultured under serum-free conditions in the absence or presence of activin-A (50 ng/ml) for the durations indicated for up to 72 h. At the indicated time points, cellular pellets were subjected to RNA extraction and Northern blot analysis. The top panel depicts the mean SE of 3 experiments. The lower panel shows an autoradiograph from a single representative experiment. **Control samples (-activin-A) significantly different (p < 0.05) from the 72-h control. *Activin-A-treated samples significantly different (p < 0.05) from the corresponding untreated control for that time point.
transcripts corresponding to IGFBP-4 displayed a temporary decline culminating in a 6-h nadir followed by a relatively prompt recovery to levels comparable to those observed at the outset of culture (p < 0.05 for 4, 6, and 12 h). Similar changes in IGFBP-4 transcripts were seen in the presence of activin-A. IGFBP-5 transcripts, in turn, displayed spontaneous time-dependent increments beginning with a rapid rise after 6 h (p < 0.05 for 12 and 24 h). Addition of activin-A completely abrogated this rise in IGFBP-5 mRNA. Importantly, the inhibitory effect of activin-A on IGFBP-5 mRNA became apparent as early as 12 h into the experiment (p < 0.05). To examine the interactions between activin-A and follistatin, its putative binding protein [40], granulosa cells (2 x 106 viable cells/tube) were cultured under serum-free conditions for 72 h in the absence or presence of activin-A (50 ng/ml), with or without increasing concentrations (0100 ng/ml) of follistatin. As shown (Fig. 5; n = 3), treatment with activin-A by itself (in the absence of follistatin), produced 41% and 72% decreases (vs. control) in the relative expression of IGFBP-4 and IGFBP-5 transcripts, respectively (p < 0.05). However, given the concurrent presence of increasing concentrations of follistatin, note was made of progressive dose-dependent reversal of the activin-A effect on IGFBP-5, with an ED 50 of 8.0 ng/ml. The corresponding figure for IGFBP-4, 13 ng/ml, was judged to be less reliable given the narrower inhibition range. These observations confirm the ability of follistatin to neutralize, presumably by way of sequestration, the activity of activin-A, thereby leading to complete inhibition of its effect (at least for IGFBP-5). To compare the activity of activin-A with activin-B, granulosa cells (2 106 viable cells/tube) were cultured under serum-free conditions for 72 h in the absence or presence of activin-A (50 ng/ml) or activin-B (50 ng/ml). As shown (Fig. 6; n = 3), treatment with activin-A produced
511
FIG. 4. Activin-A-attenuated granulosa cell IGFBP expression: the first 24 h of culture. Granulosa cells were cultured under serum-free conditions for up to 72 h. At the indicated time points, cellular pellets were subjected to RNA extraction and Northern blot analysis as described for Figure 1. The top panel depicts the mean SE of 3 experiments. The lower panel shows an autoradiograph from a representative experiment.
the projected decrements in IGFBP-4 (42%) and IGFBP-5 (83%) transcripts. Activin-B, in turn, produced a comparable decrement in IGFBP-4 (31%, p < 0.05) but was incapable of inhibiting IGFBP-5 expression (21% decrement, p = 0.135).
To examine the specificity of action of activin-A and to determine possible interactions with inhibin-A, granulosa
FIG. 5. Activin-A-attenuated granulosa cell IGFBP expression: effect of follistatin. Granulosa cells were cultured under serum-free conditions in the absence or presence of activin-A (50 ng/ml), with or without increasing concentrations (0-100 ng/ml) of follistatin. Following incubation, cellular pellets were subjected to RNA extraction and Northern blot analysis as described for Figure 1. The top panel depicts the mean SE of 3 experiments. The lower panel shows an autoradiograph from a representative experiment.
512
CHOI ET AL.
FIG. 8. Activin-A-attenuated granulosa cell IGFBP release. Granulosa cells were cultured under serum-free conditions in the absence or presence of activin-A (50 ng/ml), inhibin-A (50 ng/ml), or activin-B (50 ng/ml) for 72 h. Media so conditioned were subjected to Western Ligand blotting. Data reflect a representative experiment, comparable data having been obtained in 3 additional experiments.
FIG. 6. Granulosa cell IGFBP gene expression: comparison of activin-A with activin-B. Granulosa cells were cultured under serum-free conditions for 72 h in the absence or presence of activin-A (50 ng/ml) or activin-B (50 ng/ml). Following incubation, cellular pellets were subjected to RNA extraction and Northern blot analysis as described for Figure 1. The top SE of 3 experiments. The lower panel shows panel depicts the mean an autoradiograph from a representative experiment.
cells (2 x 106 viable cells/tube) were cultured under serumfree conditions for 72 h in the absence or presence of activin-A (50 ng/ml), inhibin-A (50 ng/ml), or both. As shown (Fig. 7), treatment with activin-A by itself (n = 4) produced the expected decrease (p < 0.05) in IGFBP-4 expression (37% inhibition). Interestingly, the provision of inhibin-A by itself (n = 3) produced a more modest inhibitory effect that approached significance (p = 0.096). However, the combined application of both activin-A and inhibin-A (n = 3) produced a level of inhibition (30%) of IGFBP-4 that (although statistically insignificant) did not differ substantially from the level of inhibition observed for activin-A by itself. As expected, treatment with activin-A by itself produced profound inhibition of the steady-state levels of transcripts corresponding to IGFBP-5 (80% inhibition; p < 0.01). In contrast, however, no significant inIGFBP-5
IGFBP-4 .
*4_,
M0
8z
"°
U) 8
Kgo
IO
*:l IfM. i T
60
60 40"
20 U l
0u4 0.1). Likewise, treatment of granulosa cells with MIS did not produce a statistically significant effect (36% and 28% increments for IGFBP-4 and IGFBP -5, respectively). Interestingly, prior plasmin cleavage increased the magnitude of this potential stimulatory effect (126% and 59% increments, respectively), although this result was highly variable (p > 0.1). Lastly (Fig. 9), treatment with bFGF, an unrelated growth factor, tended to increase the abundance of IGFBP-4 transcripts (120% increment, p > 0.1) but not of IGFBP-5 transcripts. DISCUSSION
"
A +
FIG. 7. Activin-A-attenuated granulosa cell IGFBP transcripts: effect of inhibin-A. Granulosa cells were cultured under serum-free conditions for 72 h in the absence or presence of activin-A (50 ng/ml; n = 4), inhibin-A (50 ng/ml; n = 3), or both (n = 3). Following incubation, cellular pellets were subjected to RNA extraction and Northern blot analysis as described for Figure 1. *Significantly different (p < 0.05) from control.
The present findings document for the first time the ability of activin-A to suppress the expression of granulosa cell-derived IGFBP transcripts. In this respect, our findings expand the list of independent actions attributed to activin such as the induction of FSH receptors [41, 42], the enhancement of follistatin expression [43], and the promotion of inhibin biosynthesis [44]. To the extent that activin constitutes an autocrine promoter of the differentiation of im-
ACTIVIN AND OVARIAN IGFBPs
FIG. 9. Activin-A-attenuated granulosa cell IGFBP expression: specificity studies. Granulosa cells were cultured under serum-free conditions for 72 h in the absence or presence of activin-A (50 ng/ml), TGFP, (50 ng/ml), bFGF (50 ng/ml), MIS (5 jig/ml), and PC-MIS (PC, plasmin-cleaved) (5 jag/ ml). Following incubation, cellular pellets were subjected to RNA extraction and Northern blot analysis as described for Figure 1. The top panel depicts the mean + SE of 3 experiments. The lower panel shows an autoradiograph from a single representative experiment.
mature granulosa cells, the ability of activin to suppress the expression of granulosa cell-derived IGFBP-4 and IGFBP-5 transcripts must be viewed as a positive facilitatory phenomenon given the projected resultant increase in bioavailable IGF-I. According to this view, the inhibition of the expression of granulosa cell-derived IGFBPs may well be yet another action of activin that appears to be in the best interest of the differentiating granulosa cell. In assessing the characteristics of the activin-A effect, note was made of dose-dependent responsiveness and EC 50 s of 16.2 and 7.8 ng/ml for IGFBP-5 and IGFBP-4 transcripts, respectively. Although these EC 50 s are similar to those reported for other activin-A end points, it is not immediately apparent why the overall suppression of IGFBP-5 transcripts (-80%) far exceeded the value observed for IGFBP-4 (< 50%). Likewise, it remains unclear why the apparent decrease in the steady-state levels of IGFBP-4 transcripts was not associated with a decrease in the accumulation of the corresponding protein. The above notwithstanding, treatment with activin-A produced nearly maximal suppression of IGFBP-5 as early as 12 h and IGFBP-4 by 72 h, presumably by blocking the increase in IGFBP-4 and IGFBP-5 mRNAs that was observed to occur after the start of culture. The precise mechanism(s) through which activin exerts an inhibitory effect on the expression of IGFBP-4 and IGFBP-5 transcripts remains uncertain. It is clear, however, that activin-A is capable of blocking the spontaneous rise of transcripts corresponding to IGFBP-4 and IGFBP-5 exhibited by untreated cultured granulosa cells. In principle, it is conceivable that this phenomenon is attributable to the ability of activin-A to attenuate the transcriptional rate of the corresponding genes. Alternatively, treatment with activin-A may be in a position to diminish the half-life of
513
the relevant transcripts. A combination of the preceding phenomena also constitutes a possibility. Given that the transduction cascade associated with activin receptors remains to be clarified, it is difficult at this time to deduce the possible DNA response elements potentially involved in the phenomena under study. As expected, the concurrent provision of follistatin, a putative activin-binding protein [38], produced dose-dependent reversal of the activin-A effect. These observations confirm the ability of follistatin to neutralize, presumably by way of sequestration, the activity of activin-A, thereby leading to partial abrogation of the effect on IGFBP-4 and complete abrogation of the effect on IGFBP-5. It follows then that the net intrafollicular content of IGFBPs under in vivo circumstances may depend not only on the relative level of expression of activin-A but also on the relative abundance of its binding protein. In contrast to the ability of follistatin to block activin action, inhibin-A had no effect (Fig. 7). This apparent inaction of inhibin-A is unusual since inhibin is dominant over activin with respect to most other end points [27]. Although the precise molecular and cellular reasons underlying inhibin's lack of effect remain uncertain, it is notable that inhibin-A proved to be singularly incapable of modulating granulosa cell-derived IGFBPs. It is hoped that a more detailed analysis of signal transduction pathways and promoter activity may shed additional light on this otherwise inexplicable phenomenon. The present observations confirm and extend the recognition that IGFBP-4 and IGFBP-5 are differentially regulated. First, it must be noted that the overall maximal suppression of IGFBP-5 transcripts by activin-A far exceeded the value observed for IGFBP-4, and that the reverse was observed for activin-B. Second, the limited ability of inhibin-A, TGF3, MIS, or bFGF to modulate the expression of granulosa cell-derived IGFBPs showed that IGFBP-4 was unique in this effect. Consequently, it appears that the regulation of granulosa cell-derived IGFBPs is characterized by a level of selectivity and at times exclusivity. Taken together, these observations suggest a level of complexity not previously envisioned, the precise physiological significance of which remains to be determined. The realization that the inhibitory action of activin-A on elaboration of the IGFBP-5 protein is shared only with FSH [11, 12, 26] suggests that activin-A may be an intermediate for FSH. However, to the extent that the intrafollicular IGFBP status may be a determinant of follicular fate, our current and previous observations support the view that the net intrafollicular content of IGFBPs (particularly IGFBP-5) reflects the combined input of several regulatory agents, some of which may be inhibitory (e.g., FSH, activin-A) while others, including IGF-I [45], may be stimulatory. Subject to limitations inherent in the extrapolation of in vitro experiments to the in vivo state, and to the extent that the inhibition of IGFBP-4 and IGFBP-5 expression is associated with and possibly causally related to the promotion of follicular health [16-18, 26], the present observations are in keeping with the proposition that activin may play an antiatretic role in the dynamic process of follicular selection. Such a conclusion is compatible with the view that the suppression of intrafollicular IGFBPs may well constitute a requirement for the successful completion of follicular maturation.
514
CHOI ET AL.
ACKNOWLEDGMENT The authors wish to thank Ms. Cornelia T. Szmajda for her invaluable assistance in the preparation of this manuscript.
REFERENCES 1. Adashi EY, Rohan RM. Intraovarian regulation: peptidergic signalling systems. Trends Endocrinol Metab 1992; 3:243-248. 2. Hansson H-A, Nilsson A, Isgaard J, Billig H, Isaksson O, Skottner A, Andersson IK, Rozell B. Immunohistochemical localization of insulin-like growth factor I in the adult rat. Histochemistry 1988; 89: 403-410. 3. Oliver JE, Aitman TJ, Powell JF, Wilson CA, Clayton RN. Insulinlike growth factor I gene expression in the rat ovary is confined to the granulosa cells of developing follicles. Endocrinology 1989; 124: 267-269. 4. Hernandez ER, Roberts CH Jr, LeRoith D, Adashi EY. Rat ovarian insulin-like growth factor I (IGF-I) gene expression is granulosa cellselective: 5'UT mRNA variant representation and hormonal regulation. Endocrinology 1989; 125:572-574. 5. Zhou J, Chin E, Bondy C. Cellular pattern of insulin-like growth factor-I (IGF-I) and IGF-I receptor gene expression in the developing and mature ovarian follicle. Endocrinology 1991; 129:3281-3288. 6. Adashi EY, Resnick CE, Svoboda ME, Van Wyk JJ. Follicle-stimulating hormone enhances somatomedin-C binding to cultured rat granulosa cells: evidence for cAMP-dependence. J Biol Chem 1986; 261: 3923-3926. 7. Davoren JB, Kasson BG, Li CH, Hsueh AJW. Specific insulin-like growth factor (IGF) I- and II-binding sites on rat granulosa cells: relation to IGF action. Endocrinology 1986; 119:2155-2162. 8. Adashi EY, Resnick CE, Hernandez ER, Svoboda ME, van Wyk JE Characterization and regulation of a specific cell membrane receptor for somatomedin-C/insulin-like growth factor-I in cultured rat granulosa cells. Endocrinology 1988; 122:194-201. 9. Adashi EY, Resnick CE, Rosenfeld RG. Insulin-like growth factor-I (IGF-I) and IGF-II hormonal action in cultured rat granulosa cells: mediation via type I but not type II IGF receptors. Endocrinology 1989; 126:216-222. 10. Ui M, Shimonaka M, Shimasaki S, Ling N. An insulin-like growth factor-binding protein in ovarian follicular fluid blocks follicle stimulating hormone-stimulated steroid production by ovarian granulosa cells. Endocrinology 1989; 125:912-916. 11. Adashi EY, Resnick CE, Hernandez ER, Hurwitz A, Rosenfeld RG. Follicle-stimulating hormone inhibits the constitutive release of insulin-like growth factor binding proteins by cultured rat ovarian granulosa cells. Endocrinology 1990; 126:1305-1307. 12. Adashi EY, Resnick CE, Hernandez ER, Hurwitz A, Rosenfeld RG. Ovarian granulosa cell-derived insulin-like growth factor (IGF) binding proteins: constitutive release of low molecular weight, high-affinity IGF-selective species. Mol Cell Endocrinol 1990; 74:175-185. 13. Shimasaki S, Shimonaka M, Ui M, Inouye S, Shibata E Ling N. Structural characterization of a follicle-stimulating hormone action inhibitor in porcine ovarian follicular fluid. J Biol Chem 1990; 165:2198-2202. 14. Shimasaki S, Uchiyama E Shimonaka M, Ling N. Molecular cloning of the cDNAs encoding a novel insulin-like growth factor binding protein from rat and human. Mol Endocrinol 1990; 4:1451-1458. 15. Bicsak TS, Shimonaka M, Malkowski M, Ling N. Insulin-like growth factor-binding protein (IGFBP) inhibition of granulosa cell function: Effect on cyclic adenosine 3',5'-monophosphate, deoxyribonucleic acid synthesis, and comparison with the effect of an IGF-I antibody. Endocrinology 1990; 126:2184-2189. 16. Nakatani A, Shimasaki S, Erickson GF, Ling N. Tissue-specific expression of four insulin-like growth factor-binding proteins (1, 2, 3, and 4) in the rat ovary. Endocrinology 1991; 129:1521-1529. 17. Erickson GE Nakatani A, Ling N, Shimasaki S. Cyclic changes in insulin-like growth factor binding protein-4 messenger ribonucleic acid in the rat ovary. Endocrinology 1992; 130:625-636. 18. Erickson GE Nakatani A, Ling N, Shimasaki S. Localization of insulin-like growth factor binding protein-5 messenger ribonucleic acid in rat ovaries during the estrous cycle. Endocrinology 1992; 130: 1867-1878. 19. Adashi EY, Resnick CE, Ricciarelli E, Hurwitz A, Hernandez ER, Rosenfeld RG, Carllson-Skwirut C, Francis GL. Granulosa cell-derived insulin-like growth factor (IGF) binding proteins are inhibitory to IGF-I hormonal action: evidence derived from the use of a truncated IGF-I analog. J Clin Invest 1992; 90:1593-1599.
20. Adashi EY, Resnick CE, Svoboda ME, van Wyk JJ. A novel role for somatomedin-C in the cytodifferentiation of the ovarian granulosa cell. Endocrinology 1984; 115:1227-1229. 21. Davoren JB, Hsueh AJW, Li CH. Somatomedin-C augments FSHinduced differentiation of cultured rat granulosa cells. Anal J Physiol 1985; 249:E26-E33. 22. Hutchinson LA, Findlay JK, Herington AC. Growth hormone and insulin-like growth factor-I accelerate PMSG-induced differentiation of granulosa cells. Mol Cell Endocrinol 1988; 55:61-69. 23. Bicsak TA, Tucker EM, Cappel S, Vaughan J, Rivier J, Vale W, Hsueh AJW. Hormonal regulation of granulosa cell inhibin biosynthesis. Endocrinology 1986; 119:2711-2719. 24. Zhien ZZ, Carson RS, Herington AC, Lee VWK, Burger HG. Folliclestimulating hormone and somatomedin-C stimulate inhibin production by rat granulosa cells in vitro. Endocrinology 1987; 120:1633-1638. 25. Adashi EY. Regulation of intrafollicular IGFBPs: possible relevance to ovarian follicular selection. In: Baxter RC, Gluckman PD, Rosenfeld RG (eds.), The Insulin-Like Growth Factors and Their Regulatory Proteins, Excerpta Medica International Congress Series 1056. Amsterdam, The Netherlands: Elsevier Science B.V.; 1995: (in press). 26. Liu X-J, Malkowski M, Guo Y, Erickson GE Shimasaki S, Ling N. Development of specific antibodies to rat insulin-like growth factorbinding proteins (IGFBP-2 to -6): analysis of IGFBP production by rat granulosa cells. Endocrinology 1993; 132:1176-1183. 27. Findlay JK. An update on the role of inhibin, activin, and follistatin as local regulators of folliculogenesis. Biol Reprod 1993; 48:15-23. 28. Li R, Phillips DM, Mather JP. Activin promotes ovarian follicle development in vitro. Endocrinology 1995; 136:849-856. 29. Woodruff TK, Lyon RJ, Hansen SE, Rice GC, Mather JP Inhibin and activin locally regulate rat ovarian folliculogenesis. Endocrinology 1990; 127:3196-3205. 30. Rohan RM, Ricciarelli E, Kiefer MC, Resnick CE, Adashi EY. Rat ovarian insulin-like growth factor binding protein-6: a hormonallyregulated theca-interstitial selective species with limited antigonadotropic activity. Endocrinology 1993; 132:2407-2512. 31. Shimasaki S, Shimonaka M, Zhang H-P, Ling N. Identification of five different insulin-like growth factor binding proteins (IGFBPs) from adult rat serum and molecular cloning of a novel IGFBP-5 in rat and human. J Biol Chem 1991; 266:10646-10653. 32. Harpold MM, Evans RM, Salditt-Georgiff M, Darnell JE. Production of mRNA in chinese hamster cells: relationship of the rate of synthesis to the cytoplasmic concentration of nine specific mRNA sequences. Cell 1979; 17:1025-1035. 33. Park O-K, Mayo KE. Transient expression of progesterone receptor messenger RNA in ovarian granulosa cells after the preovulatory luteinizing hormone surge. Mol Endocrinol 1991; 5:967-978. 34. Camp T, Rahal JO, Mayo KE. Cellular localization and hormonal regulation of follicle-stimulating hormone and luteinizing hormone receptor messenger RNAs in the rat ovary. Mol Endocrinol 1991; 5:1405-1417. 35. Garvey WT, Huecksteadt TP, Birnbaum MJ. Pretranslational suppression of an insulin-responsive glucose transporter in rats with diabetes mellitus. Science 1989; 245:60-63. 36. Griffiths SL, Knowler JT, Houslay MD. Diabetes-induced changes in guanine-nucleotide-regulatory-protein mRNA detected using synthetic oligonucleotide probes. Eur J Biochem 1990; 193:367-374. 37. Qulali M, Crabb DW. Corticosterone induces rat liver alcohol dehydrogenase mRNA but not enzyme protein or activity. Alcohol Clin Exp Res 1992; 16:427-431. 38. Putowski LT, Choi DS, Mordacq J, Scherzer WJ, Mayo ME, Kiefer ND, Adashi EY, Rohan RM. In vivo hormonal regulation of insulinlike growth factor binding protein-5 mRNA in the immature rat ovary. J Soc Gynecol Invest 1995; 2:735-742. 39. Hossenlopp P Seurin D, Segovia-Quinson B, Hardouin S, Binoux M. Analysis of serum insulin-like growth factor binding proteins using western blotting: use of the method for titration of the binding proteins and competitive binding studies. Anal Biochem 1986; 154:138-143. 40. Michel U, Famworth P, Findlay JK. Follistatins: more than follicle-stimulating hormone suppressing proteins. Mol Cell Endocrinol 1993; 91:1-11. 41. Xiao S, Robertson DM, Findlay JK. Effects of activin and folliclestimulating hormone (FSH)-suppressing protein/follistatin on FSH receptors and differentiation of cultured rat granulosa cells. Endocrinology 1992; 131:1009-1016. 42. Hasegawa Y, Miyamoto K, Abe Y, Nakamura T Sugino H, Eto E, Shibai H, Igarashi M. Induction of follicle stimulating hormone receptor by erythroid differentiation factor on rat granulosa cell. Biochem Biophys Res Commun 1988; 156:668-674. 43. Michel U, McMaster JW, Findlay JK. Regulation of steady state fol-
ACTIVIN AND OVARIAN IGFBPs listatin messenger RNA levels in rat granulosa cells in vitro. J Mol Endocrinol 1992; 9:147-156. 44. LaPolt PS, Soto D, Su JG, Campen CA, Vaughan J, Vale W, Hsueh AJW. Activin stimulation of inhibin secretion and messenger RNA levels in cultured granulosa cells. Mol Endocrinol 1989; 3:1666-1673.
515
45. Adashi EY, Resnick CE, Rosenfeld RG. IGF-I stimulates granulosa cell-derived insulin-like growth factor binding protein-5: evidence for mediation via type-I IGF receptors. Mol Cell Endocrinol 1994; 99: 279-284.
