Paracrine Regulation of Insulin-Like Growth Factor I ... - Oxford Journals

66 downloads 0 Views 253KB Size Report
Synthesis by Human Corpus Luteum in Vitro: A Possible. Balance of Luteotropic and Luteolytic Effects. ROSANNA APA, FIORELLA MICELI, EMILIA PIERRO, ...
0021-972X/99/$03.00/0 The Journal of Clinical Endocrinology & Metabolism Copyright © 1999 by The Endocrine Society

Vol. 84, No. 7 Printed in U.S.A.

Paracrine Regulation of Insulin-Like Growth Factor I (IGF-I) and IGF-II on Prostaglandins F2a and E2 Synthesis by Human Corpus Luteum in Vitro: A Possible Balance of Luteotropic and Luteolytic Effects ROSANNA APA, FIORELLA MICELI, EMILIA PIERRO, FRANCESCA MINICI, PIERLUIGI NAVARRA, ALESSANDRO CARUSO, MADDALENA NAPOLITANO, SALVATORE MANCUSO, AND ANTONIO LANZONE Departments of Obstetrics and Gynecology (A.P., M.F., P.E., M.F., C.A., M.S.) and Pharmacology (N.P.), Universite` Cattolica del Sacro Cuore, 00168 Rome; the Department of Experimental Medicine and Pathology, Universite` La Sapienza (N.M.), Rome; and OASI Institute for Research (L.A.), Troina, Italy ABSTRACT The existence of a complete intraovarian insulin-like growth factor (IGF) system replete with ligands, receptors, and binding proteins has been demonstrated as well as the ability of IGF-I to positively affect steroidogenesis in human granulosa cells. Furthermore, we recently showed that IGF-I and IGF-II stimulate progesterone secretion by human luteal cells. As the PGs, PGE2 and PGF2a, are classically known to have luteotropic and luteolytic effects, we wanted to determine whether the IGFs could affect the human luteal phase by influencing the PG system. For this reason, human luteal cells were cultured for different times (12, 24, and 48 h) with IGF-I, IGF-II

P

ROGESTERONE (P) secreted by the human corpus luteum (CL) of the menstrual cycle is essential for an adequate length of the luteal phase, implantation, and maintenance of early embryo development (1). Its different secretion during the three different phases of the luteal period is the result of complex interactions between stimulatory (luteotropic) and inhibitory (luteolytic) factors still not completely known. In recent years evidence has been accumulated to favor a role for PGs in several functional processes in the ovary, mainly in the CL physiology. PGF2a seems to be essential for luteolysis in certain animal species (2), and many recent findings suggest a key role for it also in human CL regression (3–5). In fact, a negative correlation between the concentrations of P and PGF2a throughout the menstrual cycle has been demonstrated (6), and in vitro the intraluteal injection of PGF2a caused both an immediate fall in serum P and a shortening of the luteal phase (7). Conversely, there are several reports suggesting a luteotropic action for PGE2. In animals this PG stimulates luteal P secretion (8); in sheep, experiments using chronic intrauterine injections of PGE2 prevented mechanically induced luteolysis created by insertion of an intrauterine device (9). In humans, a positive corReceived December 14, 1998. Revision received March 19, 1999. Accepted April 8, 1999. Address all correspondence and requests for reprints to: Apa Rosanna, M.D., Department of Obstetrics and Gynecology, Universite` Cattolica del Sacro Cuore, Largo A. Gemelli 8, 00168 Rome, Italy.

(10 –100 ng/mL), and GH (100 ng/mL), and both PGs were assayed in the medium culture. We found that both IGF-I and IGF-II were able to stimulate PGE2 synthesis in a time- and dose-dependent way, whereas they both inhibited PGF2a production. GH, too, significantly reduced PGF2a synthesis; this effect was IGF-I mediated because it was reverted by increasing dilutions of an anti-IGF-I antibody. On the contrary, no GH effect was observed on PGE2 production. In conclusion, based on these data and on our previous results, we speculate that IGFs could influence luteal steroidogenesis through PG system. (J Clin Endocrinol Metab 84: 2507–2512, 1999)

relation between P and PGE2 has been demonstrated throughout the entire menstrual cycle (10); in vitro, PGE2 was able to stimulate cAMP (4, 11) and P (11, 12) synthesis in isolated human CL. Among the several factors regulating the complex luteal physiology, the pivotal roles of LH and gonadal steroids have been well documented (13). However, new substances have recently provoked interest in this regard. Of these, the insulin-like growth factor (IGF) system has received considerable attention. An increasing body of information is compatible with the existence of a complete intraovarian IGF system replete with ligands, receptors, and binding proteins (14), and a direct steroidogenic action of IGF-I has been shown, as it is able to positively affect steroidogenesis in human granulosa (15) and rabbit and rat luteal cells. Furthermore, in addition to these results, we recently demonstrated that IGF-I as well as IGF-II stimulate P secretion when cultured with human luteal cells (16). Based on this information we wanted to continue our investigation about the influence of IGFs on luteal physiology and determine whether these factors could also affect PGs luteal synthesis. Materials and Methods Luteal cell culture preparation CL were obtained at the time of hysterectomy performed for nonendocrine gynecological diseases (leiomyomatosis) in the midluteal phase of the menstrual cycle (days 5– 6 from ovulation). Patients were

2507

2508

JCE & M • 1999 Vol 84 • No 7

APA ET AL.

between 30 – 43 yr of age. All of them had a history of regular menstrual cycles. A total of eight experiments were performed. Informed consent was obtained from each patient, and the study obtained approval from the internal review board. The age of the CL was determined as follows. All patients were monitored until ovulation by daily measurement of basal body temperature and ultrasound examination of follicular growth. When the maximal follicular diameter had reached 18 mm, daily determination of plasma P values was made. The time of ovulation (day 0) was detected by the biphasic pattern of basal body temperature, the typical ultrasound disappearance of the dominant follicle or the echographic detection of CL, and the rise of plasma P concentrations. At the time of surgery, plasma samples were collected immediately before anesthesia to determinate plasma P concentrations. The removed luteal tissue was immediately freed from blood vessels and ovarian stroma under a dissecting microscope, dissected, and minced. Human CL cultures were established as previously described (16). The luteal tissue was placed in Ham’s F-12 medium containing 10 mmol/L HEPES buffer and 5 mg/mL collagenase type IV, then incubated at 37 C in a shaking water bath for 2 h. The cell suspension was filtered through 40-mm pore size nylon mesh, centrifuged, and resuspended in fresh medium several times to obtain highly purified luteal cells. Cells were counted, and viability was determined by the trypan blue test. The cells were diluted to a final concentration of 80,000 –100,000 live cells/mL medium supplemented with 2 mmol/L l-glutamine, 100 IU penicillin, and 100 mg/mL streptomycin with and without 10% FCS and were cultured in multiwell plates for 24 h in 5% CO2 and 95% air at 37 C. After this time the cells attached to the wells; the medium was then removed, and fresh serum-free medium supplemented with 5 ng/mL insulin, 5 mg/mL transferrin, and 40 ng/mL hydrocortisone was added to the cultures. At this point the cells were incubated with IGF-I or IGF-II (1–100 ng/mL) or with GH (4 mg/mL) alone or in association with an anti-IGF-I antibody at different dilutions (from 1:1000 to 1:8000). At the end of culture, the cells were stained for lipids with oil red O14 and counted. More than 90% of the luteal cells stained positive for lipids. The remainder of the cells did not stain for lipids, and there were occasional vascular cells, including erythrocytes and leukocytes. The medium was harvested after 12, 24, and 48 h of culture and was stored at 220 C until assay for PGs.

data were then analyzed by one-way ANOVA with Bonferroni correction to perform pairwise comparisons between group means.

Results

First we investigated the effects of different doses of IGF-I and IGF-II (1–100 ng/mL) on PGF2a and PGE2 production by human luteal cells cultured for 24 h. As shown in Fig. 1a, both IGFs were able to progressively reduce PGF2a synthesis com-

PG assays The RIAs for PGE2 and PGF2a were first characterized for measurement of the prostanoids in human urine (17) and later were used successfully to measure PGs produced and released by several cell types in vitro, including cells from human ovaries (18). Briefly, for each assay, incubation mixtures of 1.5 mL were prepared in disposable plastic tubes in which 100 or 20 mL (for PGE2 or PGF2a, respectively) incubation medium were diluted to 250 mL with 0.025 mol/L phosphate buffer (pH 7.5). Tritiated PGE2 or PGF2a (2500 –3500 cpm) and appropriately diluted antisera were added together to a final volume of 1.25 mL. The antisera, provided by Prof. G. Ciabattoni, were employed at a final dilution of 1:100,000 or 1:150,000 (for PGE2 or PGF2a, respectively). The standard curves ranged from 2–100 pg/tube. A duplicate standard curve was run for each assay. All tubes were incubated for 12–18 h at 4 C. Separation of antibody-bound prostanoids was obtained with 3 mg charcoal (NoritA), which absorbs 95–98% of free PGs; charcoal suspension (3 mg/100 mL) in 0.025 mol/L phosphate buffer, pH 7.5, was added to each tube after the addition of 100 mL 5% BSA. The tubes were briefly shaken and then centrifuged for 10 min at 4 C. Supernatants were decanted into 10 mL scintillation liquid. Radioactivity was measured by liquid scintillation counting. The detection limit of the assay was 2 pg/tube in all cases. The inter- and intraassay variability coefficients were 2.7 and 2.9 for PGE2 and 3.2 and 2.8 for PGF2a, respectively. Pure human IGF-I and IGF-II were obtained from Boehringer Mannheim (Mannheim, Germany). Human GH was obtained from Serono (Rome, Italy), and the IGF-I antibody (monoclonal antibody against human somatomedin C/IGF-I) was obtained from the National Hormone and Pituitary Program, NIH. [3H]PGE2 and [3H]PGF2a were purchased from DuPont-New England Nuclear (Milan, Italy).

Data analyses Data were first analyzed by the Kolmorogov-Smirnov test to assess differences in the general shapes of distribution. Normally distributed

FIG. 1. Dose-dependent effects of IGF-I and IGF-II on PGF2a (a) and PGE2 (b) production by human luteal cells. Luteal cells were cultured for 24 h with medium alone (CTR) or with increasing concentrations of IGF-I and IGF-II (1–100 ng/mL). Data represent the mean 6 SEM of eight experiments. Significance vs. control: *, P , 0.05; **, P , 0.01; ***, P , 0.001.

IGFs AFFECT PGs SYNTHESIS IN HUMAN CORPUS LUTEUM

pared to the control value; this inhibitory effect was dose dependent and statistically significant at a dose of 100 ng/mL (IGF-I vs. control, P , 0.001; IGF-II vs. control, P , 0.05). Conversely, both factors significantly stimulated PGE2 synthesis by the same cells (Fig. 1b). Also in this case, the effect was dose dependent; the amount of PGE2 induced by IGF-I was statistically different from the control value at a dose of 10 ng/mL (P , 0.01), whereas for IGF-II, the effective significant dose was 100 ng/mL (P , 0.01). We next cultured luteal cells with a single dose of IGFs (100 ng/mL) for different times. Interestingly, the negative effect of IGF-I on PGF2a production was statistically present at 12 h (P , 0.05) and was still present after 48 h of culture (P , 0.05; Fig. 2a, left panel). On the contrary, the inhibitory action of IGF-II was present at 24 h (P , 0.05), whereas no effect was seen at either

2509

12 or 48 h (Fig. 2a, right panel). A positive effect was observed on PGE2 production for IGF-I and IGF-II, as both of them were able to significantly stimulate, in a time-dependent way, PGE2 synthesis at 24 and 48 h (IGF-I, P , 0.001 at 24 h and P , 0.05 at 48 h; Fig. 2b, left panel; IGF-II, P , 0.01 at 24 h and P , 0.05 at 48 h; Fig. 2b, right panel). It is well known that GH is able to induce intraovarian IGF-I production (19) and that many of its effects are mediated by IGF. To evaluate whether GH was able to affect PG production, human luteal cells were cultured for 24 h with 4 mg/mL GH. As shown in Fig. 3a, GH was able to significantly reduce PGF2a production (P , 0.01) in a manner similar to that of the two IGFs, whereas it did not show any effect on PGE2 synthesis (Fig. 3b). At this point, to determine whether the action of GH on PGF2a was direct or mediated

FIG. 2. Time-dependent effects of IGF-I (left panel) and IGF-II (right panel) on PGF2a (a) and PGE2 (b) synthesis by human luteal cells. Luteal cells were cultured for 12, 24, or 48 h with medium alone (CTR) or with 100 ng/mL IGF-I or IGF-II. Data represent the mean 6 SEM of eight experiments for the 24-h period and of six experiments for the 12- and 48-h periods. Significance vs. control: *, P , 0.05; **, P , 0.01; ***, P , 0.001.

2510

JCE & M • 1999 Vol 84 • No 7

APA ET AL.

FIG. 4. Dose-dependent reversal of the GH inhibition of PGF2a synthesis by an anti-IGF-I antibody. Human luteal cells were incubated for 24 h with medium alone (CTR) or GH (4 mg/mL) alone or combined with increasing dilutions of the antibody (from 1:1000 to 1:8000). The antibody was also used alone at the highest concentration (1:1000). Data represent the mean 6 SEM of six experiments. Significance vs. control: ***, P , 0.001. Significance vs. GH: G, P , 0.05; GG, P , 0.01.

progressive reduction of the IGF-I stimulatory action on PGE2 production (Fig. 5b); both effects disappeared at the highest dilution. Discussion

FIG. 3. Effects of GH on PGF2a (a) and PGE2 (b) synthesis by human luteal cells. Luteal cells were cultured for 24 h with medium alone (CTR), IGF-I (100 ng/mL), IGF-II (100 ng/mL), or GH (4 mg/mL). Data represent the mean 6 SEM of eight experiments. Significance vs. control: *, P , 0.05; **, P , 0.01; ***, P , 0.001.

by IGF-I, luteal cells were incubated with GH (4 mg/mL) alone or combined with increasing dilutions of an anti-IGF-I antibody (from 1:1000 to 1:8000). As shown in Fig. 4, the inhibitory GH effect was progressively blocked by coincubation with the antibody, which significantly reduced the GH effect at a dilution of 1:2000 (P , 0.05). When used at a dilution of 1:1000 the IGF-I antibody brought the GHinduced PGF2a level to values similar to those observed in untreated luteal cells, whereas when incubated alone, it had no effect on basal PGF2a production. Figure 5 shows the specificity of the IGF-I antibody. Used in association with IGF-I (100 ng/mL), it induced a progressive reduction of the IGF-I inhibitory effect on PGF2a synthesis (Fig. 5a) and a

Cyclic ovarian follicular development is dependent on gonadotropins and growth factors that act synergistically with or as mediators of gonadotropin action (20). In fact, intraovarian regulatory systems have been postulated to be involved in gonadal regulation, as classic endocrine theories for gonadotropins cannot fully explain the development of the dominant follicle, the simultaneous demise of the remaining cohort, and the CL activity. The IGF- system, comprised of IGF-I and IGF-II peptides, IGF-binding proteins (IGFBP), IGFBP proteases, and IGF receptors (21), is one of the several growth factor systems that probably serves this adjunctive role in ovarian activities (22). The presence and the influence of IGF-I and IGF-II in the human follicular phase has been abundantly demonstrated, whereas little was known about their role in the luteal phase. IGF-II gene expression was the first to be detected in human luteal cells as well as type I and II IGF receptor messenger ribonucleic acids (14). Successive studies also demonstrated the presence of IGF-I messenger ribonucleic acid in the human CL (23). Furthermore, in vivo, there were marked differences in IGF-I concentration between human ovarian veins; IGF-I levels were significantly higher in the ovarian vein bearing the CL (24). Therefore, the human CL can be considered as a site of secretion, action, and reception of IGFs. We were the first to demonstrate that both IGFs were able to positively affect human CL steroidogenesis either by acting directly on luteal cells or by amplifying the classic stimulatory effect of hCG

IGFs AFFECT PGs SYNTHESIS IN HUMAN CORPUS LUTEUM

FIG. 5. Dose-dependent reversal of an anti-IGF-I antibody on the IGF-I effects on PGF2a (a) and PGE2 (b) synthesis by human luteal cells. Luteal cells were cultured with medium alone (CTR) or with IGF-I (100 ng/mL) alone or combined with increasing dilutions of the antibody (from 1:1000 to 1:8000). The antibody was also used alone at the highest concentration (1:1000). Values represent the mean 6 SEM of six experiments. Significance vs. control: ***, P , 0.001. Significance vs. IGF: G, P , 0.05; GG, P , 0.01; GGG, P , 0.001.

(16). A similar positive role was demonstrated for GH (25), which, however, required the intermediacy of IGF-I to accomplish its effect (16). The aim of the present study was to continue our research about IGFs and CL to clarify another possible mechanism(s) through which they can affect luteal physiology. In the complex regulation of CL physiology, PGs seem to play an important role. It is well known that human CL produces both PGE2 and PGF2a (26, 27), and in a large number of species, including the human, PGF2a seems to be of importance for luteolysis (2–5), whereas there is accumulat-

2511

ing evidence that the effect of PGE2 on luteal function is luteotropic. In the present study we demonstrate the ability of IGF-I and IGF-II to stimulate PGE2 and to inhibit PGF2a production by human luteal cells in a dose- and timedependent manner. The stimulatory effect of both IGFs on PGE2 production was stronger than their inhibition of PGF2a synthesis, and it was reached at lower doses. It is unlikely that this effect can be explained by a possible mitogenic action of IGF-I and IGF-II, because luteal cells are highly differentiated cells, and mitosis was not observed during the culture period (28). Based on these results it is tempting to speculate that IGFs could influence luteal steroidogenesis through PGs. We know that the PGE2 to PGF2a ratio in the human CL varies during the menstrual cycle, being higher in the early and midluteal phases than in the late luteal phase (10, 26). Among the factors influencing this balance, which, in turn, influences the different luteal phases, IGFs can play an interesting role. It has been demonstrated that dominant follicles contain significantly higher concentrations of IGF-I than their cohorts, and that after the LH surge a further significant increase in dominant follicular fluid IGF-I occurs (29). Similarly, follicular fluid IGF-II levels are higher in the dominant follicles and they are positively correlated with the amount of estradiol (30). Based on these and our data, a possible mechanism(s) of IGF action could be the following. Around and immediately after ovulation, high intraovarian IGF-I and IGF-II levels could stimulate PGE2 production while inhibiting PGF2a synthesis, resulting in high P levels. Conversely, a progressive reduction in IGF levels occurring in the late luteal phase could be at least in part responsible for the change in the PGE2/PGF2a ratio, resulting in a fall in P levels. In addition, the fact that intrafollicular IGFBP-1 production increases predominantly in the largest follicles around luteinization, with a possible consequent modulation of IGFs and therefore PG levels (30) indirectly confirms this hypothesis. Furthermore, the recently demonstrated ability of PGF2a to stimulate the release of IGFBP-3 from human granulosa-luteal cells (31) with a corresponding reduction in IGF levels suggests another mechanism relating the two systems by which they can modulate each other. Much information has been accumulated indicating the ovary as a site of GH action and reception. The potential interest in the GH effect in gonadal function arises from clinical and experimental studies. In human in vivo studies, GH was found to vary in fertile women, with an increase in the late follicular phase (32), whereas anovulatory women appear to be mildly GH insufficient compared to ovulatory women (33). Furthermore, GH significantly increased the ovarian response to gonadotropin stimulation (34). In animal in vitro studies, GH augmented FSH-stimulated LH receptor formation and P biosynthesis (35) and stimulated plasminogen activator synthesis (36) and oocyte maturation (37, 38). In humans, GH stimulates estradiol production by granulosa cells (39) and P synthesis by luteal cells (25); this latter effect was obtained through the intermediacy of IGF-I (16). In fact, it seems that GH can accomplish its effects either directly or by stimulating ovarian production of IGF-I, which, in turn, mediates GH effects. In our study we found that GH is able to inhibit PGF2a synthesis. This effect could be one of the mechanisms by which GH influences luteal steroidogenesis;

2512

APA ET AL.

furthermore, the disappearance of the GH effect we observed in the presence of antibodies raised against IGF-I clearly indicates that at least in this case GH inhibition was obtained through the IGF-I intermediacy. Recently, in the human CL, a GH receptor bearing no similarity to PRL or placental lactogen receptors has been identified (40). Therefore, given the presence of pure GH receptors, it is reasonable to believe that lactogenic receptors are not involved in GH action. Finally, no GH effect on PGE2 synthesis was observed. At the moment we do not have any explanation for this dualistic GH action on PGs. It can simply be that GH is not necessarily involved in all of the effects of IGFs; however, a better knowledge of the intraovarian PG and IGF systems through molecular biology will allow us in the future to answer questions such as this. In conclusion, this work has probably added another small piece to the complex mosaic regarding CL physiology. To our knowledge, this is the first study correlating IGFs and PGs, even though it is clear that much work remains to further elucidate the mechanism(s) of action of both IGFs and PGs and their relationships. In this regard, experiments are currently in progress in our laboratory. References 1. Csapo AI, Pulkkinen MO, Wiest WG. 1973 Effects of lutectomy and progesterone replacement therapy in early pregnant patients. Am J Obstet Gynecol. 115:759 –765. 2. Horton EW, Poyser NL. 1976 Uterine luteolytic hormone: a physiological role for prostaglandin F2a. Physiol Rev. 56:595– 601. 3. Korda AR, Shutt DA, Smith ID, Shearman RP, Lyneham RC. 1975 Assessment of possible luteolytic effect of intraovarian injection of prostaglandin F in the human. Prostaglandin. 9:443– 447. 4. Dennefors B, Sjogren A, Hamberger L. 1982 Progesterone and 39,59-monophosphate formation by isolated human corpora lutea of different ages. Influence of human chorionic gonadotropin and prostaglandin. J Clin Endocrinol Metab. 55:102–108. 5. Pathwardhan VV, Lanthier A. 1984 Effect of prostaglandin F2a on the hCGstimulated progesterone production by human corpora lutea. Prostaglandins. 27:465– 470. 6. Vijayakumar R, Walters WA. 1987 Ovarian stromal and luteal tissue prostaglandins, 17b-estradiol and progesterone in relation to the phases of the menstrual cycle in woman. Am J Obstet Gynecol. 156:947–951. 7. Bennegard B, Hahlin M, Wennberg E, Nore´n H. 1991 Local luteolytic effect of prostaglandin F2a in the human corpus luteum. Fertil Steril. 56:1070 –1076. 8. Weems CW, Reynolds LP, Huie JM, Hoyer GL, Behrman HR. 1985 Effect of prostaglandin E1 or E2 on luteal function and binding of luteinizing hormone in non pregnant ewes. Prostaglandins. 29:161–164. 9. Weems C, Huie M, Magress R, Hoyer G, Whysong G. 1985 Inhibition of prostaglandin E2 in sheep. Prostaglandins. 30:573–576. 10. Vijayakumar R, Walters WA. 1983 Human luteal tissue prostaglandins, 17bestradiol and progesterone in relation to the growth and senescence of the corpus luteum. Fertil Steril. 39:298 –303. 11. Marsh JM, Lemaire WJ. 1974 Cyclic AMP accumulation and steroidogenesis in the human corpus luteum: effect of gonadotropins and prostaglandins. J Clin Endocrinol Metab. 38:99 –105. 12. Hahlin M, Dennefors B, Johanson C, Hamberger L. 1988 Luteotropic effects of prostaglandin E2 on the human corpus luteum of the menstrual cycle and early pregnancy. J Clin Endocrinol Metab. 66:909 –914. 13. Gospodarowicz D, Gospodarowicz F. 1975 The morphological transformation and inhibition of growth of bovine luteal cells in tissue culture induced by luteinizing hormone and dibutyryl cyclic AMP. Endocrinology. 96:458 – 467. 14. Hernandez ER, Hurwitz A, Vera A, Pellicer A, Adashi EY, LeRoith D. 1992 Expression of the genes encoding the insulin-like gowth factors and their receptors in the human ovary. J Clin Endocrinol Metab. 74:419 – 425.

JCE & M • 1999 Vol 84 • No 7

15. Erickson G, Garzo G, Magoffin D. 1991 Progesterone production by human granulosa cells cultured in serum free medium: effects of gonadotropins and insulin-like growth factor. Hum Reprod. 6:1074 –1081. 16. Apa R, Di Simone N, Ronsisvalle E, et al. 1996 Insulin-like growth factor (IGF)-I and IGF-II stimulate progesterone production by human luteal cells: role of IGF-I as mediator of growth hormone action. Fertil Steril. 66:235–239. 17. Ciabattoni G, Pugliese F, Spaldi M, Cinotti GA, Patrono C. 1979 Radioimmunoassay measurement of prostaglandin E2 and F2 in human urine. J Endocrinol Invest. 2:173–179. 18. Navarra P, Andreani CL, Lazzarin N, et al. 1996 Increased production and release of prostaglandin E2 by human granulosa cells from polycystic ovaries. Prostaglandins. 52:187–197. 19. Hsu CJ, Hammond JM. 1987 Concomitant effect of growth hormone on secretion of insulin-like growth factor I and progesterone by cultured porcine granulosa cells. Endocrinology. 121:1343–1348. 20. Hillier SG. 1991 Ovarian endocrinology. Oxford: Blackwell; 1–356. 21. Jones JI, Clemmons DR. 1995 Insulin-like growth factors and their binding proteins: biological actions. Endocr Rev. 16:3–34. 22. Giudice LC. 1992 Insulin-like growth factors and ovarian follicular development. Endocr Rev. 13:641– 669. 23. Johnson MC, Devoto L, Retamales I, Kohen P, Troncoso JL, Aguilera G. 1996 Localization of insulin-like growth factor (IGF-I) and IGF-I receptor expression in human corpora lutea: role on estradiol secretion. Fertil Steril. 65:489 – 494. 24. Devoto L, Kohen P, Castro O, Vega M, Troncoso JL, Charreau E. 1995 Multihormonal regulation of progesterone synthesis in cultured human midluteal cells. J Clin Endocrinol Metab. 80:1566 –1570. 25. Lanzone A, Di Simone N, Castellani R, Fulghesu AM, Caruso A, Mancuso S. 1992 Human growth hormone enhances progesterone production by human luteal cells in vitro: evidence of a synergistic effect with human chorionic gonadotropin. Fertil Steril. 57:92–96. 26. Pathwardhan VV, Lanthier A. 1985 Luteal phase variations in endogenous concentrations of prostaglandins PGE and PGF and in the capacity for their in vitro formation in the human corpus luteum. Prostaglandins. 30:91–98. 27. Challis JRG, Calder AA, Dilley S, et al. 1976 Production of prostaglandin E and F by corpora lutea, corpora albicantes and stroma from the human ovary. J Endocrinol. 68:401– 407. 28. Retamales I, Carrasco I, Troncoso JL. 1994 Morphofunctional study of human luteal cells subpopulation. Hum Reprod. 9:591–596. 29. Eden JA, Jones J, Carter GD, Alaghband-Zadeh. 1988 A comparison of follicular fluid levels of insulin-like growth factor-I in normal dominant and cohort follicles, polycystic and multicystic ovaries. Clin Endocrinol (Oxf). 29:327–336. 30. Van Dessel T, Chandrasekher Y, Yap S, et al. 1996 Serum and follicular fluid levels of Insulin-like growth factor I (IGF-I), IGF-II and IGF-binding protein-1 and -3 during the normal menstrual cycle. J Clin Endorinol Metab. 81:1224 –1231. 31. Sarvas K, Angervo M, Koistinen R, Tiitinen A, Seppala M. 1994 Prostaglandin F2a stimulates release of insulin-like growth factor binding protein-3 from cultured human granulosa-luteal cells. Hum Reprod. 9:1643–1646. 32. Faria ACS, Bekenstein LW, Booth RA, Vaccaro VA, Asplin CM, Veldis JD. 1991 Clin Endocrinol (Oxf). 36:591–596. 33. Ovesen P, Moller J, Moller N, Christiansen JS, Jorgenson JOL, Orskov H. 1992 Fertil Steril. 57:97–101. 34. Hibrahim ZHZ, Matson PL, Buck P, Lieberman BA. 1991 Fertil Steril. 55:202–204. 35. Jia XC, Kalmijn J, Hsuesh AJW. 1986 Growth hormone enhances folliclestimulating hormone-induced differentiation of cultured rat granulosa cells. Endocrinology. 118:1401–1409. 36. Apa R, Lanzone A, Miceli F, Caruso A, Mancuso S, Canipari R. 1994 Growth hormone induction of rat granulosa cell tissue-plasminogen activator expression and progesterone synthesis. Mol Cell Endocrinol. 99:153–159. 37. Yoshimura Y, Nakamura Y, Koyama N, Iwashita M, Adachi T, Takeda Y. 1993 Effects of growth hormone on follicle growth, oocyte maturation and ovarian steroidogenesis. Fertil Steril. 59:917–923. 38. Apa R, Lanzone A, Miceli F, et al. 1994 Growth hormone induces in vitro maturation of follicle- and cumulus-enclosed rat oocytes. Mol Cell Endocrinol. 106:207–212. 39. Mason HD, Mertikainen H, Beard RW, Anyaoku V, Franks S. 1990 Direct gonadotrophic effect of growth hormone on oestradiol production by human granulosa cells in vitro. J Endocrinol. 126:R1–R4. 40. Sharara FI, Nieman LK. 1994 Identification and cellular localization of growth hormone receptor gene expression in the human ovary. J Clin Endocrinol Metab. 79:670 – 672.