776

BIOLOGY OF REPRODUCTION 67, 776–781 (2002)DOI 10.1095/biolreprod.101.001511

Insulin-Like Growth Factor (IGF)-Independent Effects of IGF Binding Protein-4on Human Granulosa Cell Steroidogenesis1

Rebecca J. Wright,2,3 Jeffrey M.P. Holly,4 Ray Galea,5 Mark Brincat,5 and Helen D. Mason3

Department of Obstetrics and Gynaecology and Department of Physiology,3 St. George’s Hospital Medical School,Tooting, London SW17 0RE, United KingdomDepartment of Surgery,4 Bristol Royal Infirmary, Bristol University, Bristol BS2 8HW, United KingdomDepartment of Obstetrics and Gynaecology,5 University of Malta Medical School, Msida, Malta, United Kingdom

ABSTRACT

The ovarian insulin-like growth factor (IGF)/IGF binding pro-tein (IGFBP) system operates to permit maximal stimulation ofsteroidogenesis in the dominant follicle. In atretic follicles, thepredominant IGFBPs are IGFBP-2 and IGFBP-4, which appearto be selectively cleaved in healthy follicles. We have recentlydemonstrated potent inhibition by IGFBP-4 of both theca andgranulosa cell steroid production. The degree to which the in-hibition occurred suggested that it was greater than might beexpected by sequestration of IGF alone. Our study was designedto test this idea. Granulosa cells were harvested from folliclesdissected intact from patients undergoing total abdominal hys-terectomy and bilateral salpingoophorectomy. Granulosa cellswere incubated with or without gonadotropins and IGFBP-4 inthe presence or absence of either the IGF type I receptor block-er aIR3 or excess IGFBP-3 to remove the effects of endogenousIGF action. Steroid accumulation in the medium was assessed.IGFBP-4 continued to exert potent inhibitory effects when theaction of endogenous IGF was removed from the system, dem-onstrating that its actions are independent of IGF binding. Therewas no effect on cell metabolism, and the effects on steroido-genesis were reversible after IGFBP-4 removal from the culturemedium. No similar effects were seen with IGFBP-2. These rea-sults are the first evidence of IGF-independent IGFBP-4 actionsand the first evidence of IGF-independent actions of any IGFBPsin the ovary.

granulosa cells, growth factors, insulin-like growth factor recep-tor, ovary, steroid hormones

INTRODUCTION

The ovarian insulin-like growth factor (IGF) system hascome under considerable scrutiny in recent years, and it isnow clear that there are changes in the components of thesystem with follicle development. In humans, these changesinclude induction of IGF-II production [1], the increasedresponsiveness of granulosa cells to IGFs [2], and the pro-duction of IGF binding protein (IGFBP) proteases, presum-ably to increase IGF bioavailability [3, 4]. All of thesechanges are directed toward increasing the steroidogeniccapacity of the healthy follicle. We and others have shownclear differences between healthy and atretic follicles in theprofile of IGFBPs in follicular fluid and in theca-condi-tioned medium [5–7]. A consistent finding is that IGFBP-2and -4 are proteolysed in follicular fluid from healthy fol-

1This work was supported by The Wellcome Trust.2Correspondence. FAX: 44 208 725 2993; e-mail: rwright@sghms.ac.uk

Received: 16 November 2001.First decision: 3 December 2001.Accepted: 4 April 2002.Q 2002 by the Society for the Study of Reproduction, Inc.ISSN: 0006-3363. http://www.biolreprod.org

licles but are intact in follicles judged to be atretic. Like-wise, only follicular fluid from healthy follicles was ableto proteolyze radiolabeled recombinant IGFBP-4 [7]. In-creased fragmentation of IGFBP-4 was also seen in medi-um conditioned by theca from healthy compared with atret-ic follicles. Interestingly, the fragment sizes found in me-dium from our cultures of theca tissue were different fromthose present in follicular fluid, suggesting the possibilityof compartmentalization of protease activity within the ova-ry [7].

In an earlier study, we demonstrated partial inhibition ofFSH-stimulated steroidogenesis by incubation of granulosacells with IGFBP-1 or -3, which we suggested was due tothe sequestration of IGF-II in these cultures [8]. Recently,we reported an inhibitory effect of IGFBP-4 on both gran-ulosa and theca steroidogenesis that was more potent thanmight be expected by sequestration of IGF alone [9] andwas more potent than that previously seen with IGFBP-3or -1. We hypothesized that the effects of IGFBP-4 in theseexperiments are independent of its ability to bind IGF. Theseries of experiments reported here were designed to testthis hypothesis

MATERIALS AND METHODS

Patients

Ovaries were obtained from women undergoing oophorectomy fornonovarian gynecological disease. All patients were cycling regularly ex-cept one who reported polymenorrhea and one with oligomenorrhea. Cyclestage was random, and none of the patients had received medication forstimulation or suppression of ovarian function for at least 3 mo prior tosurgery. Patient details are given in Table 1. Approval for this study wasgranted by the local ethics committes and by the Ethics Committee of theFaculty of Medicine and Surgery, Medical School, Malta. Informed con-sent was obtained from each patient prior to surgery. Morphological cat-egory was assigned according to previously published criteria [10]. Mostof the ovaries collected were ovulatory and polycystic.

Cell Cultures

Follicles were microscopically dissected intact from the ovaries, thediameter was measured, and follicular fluid was aspirated. Granulosa cellswere harvested and cultured as previously described [11]. Experimentswere carried out on pooled cells from a single patient or on cells fromindividual follicles where stated. In one experiment, cells were pooledfrom two patients. Details of the size of follicles from which cells werepooled is shown in Table 1. Approximately 5 3 104 viable cells/well wereincubated in a 200-ml volume of serum-free Medium 199 (Gibco BRL,Paisley, U.K.) with the addition of antibiotics (penicillin and streptomycin;Gibco BRL) and 200 mM L-glutamine. Incubations were carried out inthe presence of 1027 M testosterone (Sigma Chemical Co., Poole, U.K.)as an aromatase substrate. One experiment was carried out using cellscollected at the time of oocyte aspiration from patients undergoing ovarianstimulation for in vitro fertilization. Following removal of the oocyte, as-pirates were pooled and granulosa cells were separated from blood cellsand other debris using the method of Abeyasekara et al. [12].

777IGF-INDEPENDENT IGFBP-4 EFFECTS IN HUMAN OVARIES

c

FIG. 1. Granulosa cell steroidogenesis. All results are the mean and SEMof triplicate wells. The size of follicles from which cells were pooled ineach case is shown in the inset. a) Left to right: Basal progesterone (P) (C,control), inhibition by IGFBP-4 (not significant), stimulation by FSH (P ,

0.05), inhibition of FSH action by IGFBP-4 (P , 0.05), stimulation byIGF-I, no significant effect of aIR-3 alone, inhibition of IGF response tobasal levels by aIR-3, no effect of aIR-3 on FSH-stimulated P, IGFBP-4inhibition of FSH-stimulated P even in the presence of aIR-3 (a vs. b, P, 0.05). Similar results were obtained for E2 in the same wells. b) Cellsfrom small follicles from the same ovary. No effect of IGFBP-4 on basalE2, significant inhibition of FSH-stimulated E2 by IGFBP-4 (P , 0.01), noeffect of aIR-3 alone, inhibition of IGF effects, no significant effects ofaIR-3 on FSH-stimulated E2, significant inhibition of FSH-stimulted E2 byIGFBP-4 in presence of aIR-3 (a vs. b, P , 0.05). c) Granulosa cells froma second ovary. IGFBP-4 inhibited LH-stimulated E2 production to basallevels. Incubation with levels of aIR-3 sufficient to inhibit the effect ofadded IGF-I did not significantly affect the inhibitory action of IGFBP-4(a vs. b, P , 0.05).

Experimental ProtocolExperiments were performed on cells from six patients, two of whom

had normal ovaries and four of whom had ovulatory polycystic ovaries.Cells were incubated in medium alone or in medium containing 5 ng/mlhighly purified human pituitary LH or FSH (Endocrine Services, Bidford-on-Avon, U.K.) with or without IGFBP-4 (Austral Biologicals, San Ra-mon, CA) at 5 or 50 ng/ml, doses that previously caused maximum in-hibition of steroid production [9]. All cultures were performed with theaddition of 5 3 1027 M testosterone as an aromatase substrate.

Although the majority of these experiments were conducted with cellsfrom small follicles in which the levels of endogenous IGF would beexpected to be very low or absent, to negate the contribution of any en-dogenous IGF to the steroid production, experiments were carried out inthe presence of the specific IGF-I receptor monoclonal antibody aIR-3(Oncogene Science, Cambridge, U.K.) [13], which is purified on a protein-G affinity column. The concentration of antibody used (100 ng/ml) hadpreviously been determined to completely inhibit the stimulatory effectsof addition of 10 ng/ml IGF-I in granulosa cells [2]. A control to dem-onstrate this inhibition was added to each experiment.

To further demonstrate IGF-independent effects, a second series of ex-periments was performed in which IGFBP-4 was added in the presenceof 50 ng/ml IGFBP-3, a concentration calculated to bind any endogenousIGF-II, previously found to be ,10 ng/ml [4]. In addition to IGFBP-4,IGFBP-2 is also selectively cleaved in follicular fluid from healthy folli-cles, and in a further experiment the effects of IGFBP-2, -3, and -4 werecompared in the same pool of cells. To investigate the degree of fragmen-tation of the IGFBPs under these experimental conditions, medium wassubsequently subjected to immunoblot for IGFBP-3 and -4, and the per-centage of fragmentation was calculated after densitometric analysis ofbands representing intact and fragmented IGFBP.

All treatments were added simultaneously. Cultures were incubated for48 h, the medium was collected, and estradiol (E2) and progesterone ac-cumulations were measured with an in-house RIA, as previously described[14, 15]. This nonextraction method was specifically designed for use withculture medium, using antibodies supplied by Guildhay (Guildford, Surrey,U.K.).

We also performed two experiments to determine the permanency ofthe IGFBP-4 inhibitory effects and to see whether IGFBP-4 also affectedcell viability or metabolism. To investigate the reversibility of IGFBP-4effects, one set of cells was incubated with FSH with and without IGFBP-4 for 24 h, after which the medium was replaced by medium containingLH alone for a further 24 h. To test for possible effects of IGFBP-4 oncell viability, medium was removed and cells were exposed to a colori-metric assay in which the cells are incubated with 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma). The assay gives aqualitative assessment of mitochondrial dehydrogenase enzyme activity,which converts MTT to dark blue formazon. The resultant color changemeasured by optical density is proportional to the metabolic activity andnumber of cells [16].

Each test substance was added to three to six wells. Results are ex-pressed as the mean and standard error. Differences between treatmentswere assessed with a Mann-Whitney U-test, with significance assumed atP , 0.05.

RESULTS

IGF Independence of IGFBP-4 Effects

Significant inhibition of basal steroid production byIGFBP-4 was seen in three of five experiments performedin various pools of cells from three different patients (Fig.1, a–c, first two bars). The degree of inhibition of bothprogesterone and E2 was variable. In the same experiments,gonadotropin stimulated steroid production, which was in-hibited by IGFBP-4 in all experiments. Figure 1a showsprogesterone results from granulosa cells from a single fol-licle 12 mm in diameter. Similar results were obtained forE2 in the same wells. Results for E2 from granulosa cells

778 WRIGHT ET AL.

FIG. 2. Effect of addition of saturatingconcentrations of IGFBP-3 on action ofIGFBP-4. a) FSH (5 ng/ml) stimulated E2

levels above those seen with testosteronealone (C); IGFBP-4 inhibited FSH-stimulat-ed E2 production to basal levels; additionof IGFBP-3 to FSH had no effect on E2

production; addition of IGFBP-4 in thepresence of IGFBP-3 inhibited E2 levels tobaseline (a vs. b and c vs. d, not signifi-cant). b) Results for progesterone weresimilar. Results are mean (SEM) of fourwells. c) Inhibition of steroidogenesis byluteinized cells by IGFBP-4 (a vs. b, P ,0.02) or IGFBP-4 plus 50 ng/ml IGFBP-3(a vs. c, P , 0.005). Note inhibitory ef-fects of IGFBP-3 in these cells, which areknown to produce IGFs.

FIG. 3. Recovery of steroidogenic capac-ity by granulosa cells treated with IGFBP-4. Bars represent mean (SEM) of six wells.First 24 h: addition of IGFBP-4 inhibitedFSH-stimulated E2 (hatched bars, left scale)and progesterone (P; open bars, rightscale) production. Medium was then re-placed with medium containing testoster-one alone (C 1 t) or LH (5 ng/ml) for afuther 24 h. Second 24 h: addition of LHresulted in recovery of P and E2 levels tothose seen in cells not previously exposedto IGFBP-4 (a vs. c and b vs. d, not signifi-cant).

pooled from small follicles from the same ovary are shownin Figure 1b and those for cells from small follicles froma second ovary are shown in Figure 1c. The results forprogesterone in these experiments were similar to those forE2.

Addition of aIR-3 in these experiments had a slight ag-onistic or no effect on basal and gonadotropin-stimulatedE2 and progesterone production, demonstrating that endog-enous IGF was either absent or at a very low level. Like-wise, addition of aIR-3 in the presence of FSH had littleeffect on the level of steroid, indicating that neither FSHnor LH stimulated IGF-II production in these experimentsand that endogenous IGF did not contribute significantly to

the steroid response. The effectiveness of the antibody inthis system was demonstrated by the fact that addition of100 ng/ml of aIR-3 completely inhibited steroid productionstimulated by addition of IGF-I.

Incubation of cells in the presence of aIR-3 did not af-fect the inhibitory action of IGFBP-4 on FSH-stimulatedsteroidogenesis (Fig. 1, a–c, last bars). Similar results wereobtained for E2 and progesterone in cells pooled from smallfollicles from a third pair of ovaries (data not shown).

The results of addition of IGFBP-4 in the presence ofIGFBP-3 on FSH-stimulated E2 and progesterone produc-tion are shown in Figure 2, a and b. IGFBP-4 inhibited thestimulatory effects of FSH in each case. Coincubation with

779IGF-INDEPENDENT IGFBP-4 EFFECTS IN HUMAN OVARIES

TABLE 1. Clinical details of patients from whom ovarian tissue was collected. The size and range of follicles fromwhich cells were pooled is also given. In three cases, some details were not known (NK).

Age(yr) Indication for surgery*

Cyclehistory

Cycleday Morphology

No. follicles,size range

(mm)

No. pooledfollicles, sizerange (mm) Figure

44 Uterine fibroid and menorrhagia Regular 7/26 ovPCO 13, 2–12 1, 12 1a12, 2–7 1b

29 CCP and pelvic congestion Regular 25/28 ovPCO 20, 2–7 15, 3–7 1c36 CCP and pelvic congestion Regular 27/28 ovPCO 21, 2–8 15, 3–8 Not

shown41 PMS Regular 9/28 Normal 7, 4–10 4, 4–6 2a

2, 8 2bNK‡ CPP Irregular ovPCO 23, 2–6 14, 3–6 335‡ CPP and pelvic congestion Regular NK Normal 5, 3–12 3, 3–6 341 PMS Irregular NK ovPCO 6, 3–14 5, 3–7 4

* CCP, Chronic pelvic pain; PMS, premenstrual syndrome; ovPCO, ovulatory polycystic ovaries.‡ Follicle cells from these two patients were pooled.

FIG. 4. Comparison of the effects of equal doses of IGFBP-2, -3, or -4on FSH-stimulated E2 production. Data are the mean (SEM) of six wells.Note lack of effect of IGFBP-2, slight inhibition by IGFBP-3, and inhibitionto basal levels with IGFBP-4 (a vs. b and a vs. c, not significant; a vs. d,P , 0.01).

IGFBP-3 had no significant effect on FSH-stimulated ste-roid production in cells from these small follicles, againsuggesting that the levels of endogenous IGF were low.IGFBP-4 continued to cause inhibition of E2 and proges-terone even in the presence of IGFBP-3, which had previ-ously been shown to negate the effects of IGF-I [8]. Similarresults were obtained in luteinized cells collected as a by-product of oocyte aspiration in patients undergoing egg col-lection for in vitro fertilization (Fig. 2c).

Recovery of Steroidogenic Capacity FollowingIGFBP-4 Treatment

In this experiment, IGFBP-4 inhibited LH-stimulatedprogesterone production during the first 24 h of incubation(Fig. 3). After the medium was removed and the cells wererestimulated with LH, there was complete recovery of pro-gesterone production to levels observed in cells that hadnot been exposed to IGFBP-4, and there was partial recov-ery of E2 production.

Comparative Effects of IGFBP-2, -3, and -4

Comparison of the effects of IGFBP-2, -3, and -4 re-vealed that only IGFBP-4 significantly inhibited granulosacell E2 production (Fig. 4). In cells from this pool of fol-licles, IGFBP-2 had no effect and IGFBP-3 caused a slightbut not significant inhibition of steroidogenesis. Westernimmunoblotting of media containing exogenous IGFBP-2and -3 revealed that the vast majority of the binding proteinremained in the intact form (data not shown). This findingwas particularly apparent for IGFBP-2, where ,1% wasfragmented.

MTT Results

There was no significant difference between any of thetreatments in the optical density of medium in the wells inthe MTT assay.

DISCUSSION

We have demonstrated that the potent inhibitory effectsof IGFBP-4 on steroidogenesis by granulosa cells occureven in the presence of levels of aIR-3 known to block theeffects of endogenous IGF. The aIR-3 antibody is highlyspecific for and binds with high affinity to the IGF-I recep-tor. Because IGF-II is known to exert its effects via thisreceptor [2] and IGF-I is not made by these cells [17], alleffects of IGF should be negated in these experiments. The

lack of effects of aIR-3 and IGFBP-3 in our experimentshowever demonstrates that the levels of endogenous IGFwere very low in these cultures. This finding suggests thatIGFBP-4 is acting on these cells in a manner independentof its ability to sequester IGF. This is the first demonstrationof IGF-independent effects of IGFBP-4 and the first dem-onstration that IGFBPs may have IGF-independent effectsin the ovary.

There was significant inhibition of basal steroid produc-tion in approximately half of our experiments, whereasIGFBP-4 consistently inhibited gonadotropin-stimulatedsteroid production. There were no consistent differences inthe composition of follicle pools that could account for theinconsistency in basal inhibition.

IGF-independent effects of IGFBPs have been demon-strated in a variety of cells types, but to date these effectshave been shown only for those IGFBPs known to interactwith cell surface or extracellular matrix binding sites. Forexample, IGFBP-1 stimulates the migration of smooth mus-cle cells by a mechanism involving binding to the integrinreceptor [18]. Nuclear localization of IGFBP-3 and -5 hasbeen recently reported, presenting an intriguing possibilityfor the mechanism of action of IGFBPs in some cell types

780 WRIGHT ET AL.

[19, 20]. At present, we have no data to suggest a mecha-nism by which IGFBP-4 may be exerting its effects.

The majority of women undergoing total abdominal hys-terectomy and bilateral salpingoophorectmy have ovulatorypolycystic ovaries (personal observation). Although cellsfrom these ovaries are poor responders to gonadotropin [10]and are therefore not ideal for experimental purposes, thesupply of tissue is not such that all studies can be performedwith normal ovaries. We considered the use of polycystictissue for these studies valid because all of the pools ofcells responded to gonadotropin and because we were notinvestigating the differential effects of IGFBP-4 in normalversuis polycystic ovaries, only on unluteinized granulosacells in general.

The inhibitory effects of IGFBP-4 were confined to thisbinding protein alone; there was no or limited inhibition ofE2 and progesterone production by IGFBP-3 or IGFBP-2in these cells. IGF-I is not produced by human granulosacells, and according to previous work the IGF-II mRNA isnot detectable until the follicle reaches 11 mm in size, i.e.,following selection [1]. There are data, however, to suggestthat IGF-II production may be stimulated in response toFSH and other factors in these cells [21, 22], which mayexplain the limited effect of IGFBP-3 on these cells. In ourexperiments, aIR-3 had little or no effect even in the pres-ence of gonadotropin, indicating that the levels of endog-enous IGF were low throughout. Proteolysis of IGFBP-2 infollicular fluid from healthy follicles has been reported con-sistently and appears to be a prerequisite for folliculoge-nesis; however, our data indicate that this proteolysis doesnot appear to be due to the antigonadotropic effects ofIGFBP-2. We have been unable to detect IGFBP-3 in gran-ulosa cell-conditioned medium, and although IGFBP-2 waspresent it was always entirely in the intact form [7]. In theexperiments reported here, the vast majority of the bindingprotein remained in the intact form at the end of the cultureperiod, demonstrating that the differences in inhibitory ac-tivity of these IGFBPs cannot be due to differences in frag-mentation in the medium. Further work is required to elu-cidate the role of IGFBP-2 in the ovary.

The degree to which IGFBP-4 was antigonadotropic inthese cells emphasizes the physiological reason for its re-moval from healthy follicles. There was complete recoveryof progesterone secretion and partial recovery of E2 secre-tion following removal of IGFBP-4, showing that the ef-fects of the binding protein were reversible. IGFBP-4 doesnot cause cell death or toxicity, as further supported by thelack of an effect on metabolism in these cells as indicatedby the results of the MMT assay. Removal of IGFBP-4 inthe ovary is achieved both by suppression of mRNA pro-duction and by proteolytic cleavage. Although we and oth-ers have demonstrated IGFBP-4 proteolytic activity in fol-licular fluid of all sizes of healthy follicles, we have onlyseen activity in one sample of granulosa cell-conditionedmedium, which was from a 25-mm preovulatory follicle[23] Proteolysis of IGFBP-4 in granulosa cells was inducedby FSH in rat granulosa cells and by FSH and IGF-I inluteinized human granulosa cells [24, 25], but addition ofIGFs in the present study had no effect on protease pro-duction by granulosa cells from small follicles.

We are currently investigating the possible presence ofan inhibitor of this activity. Others have suggested that infibroblasts at least the specific IGFBP-4 protease is preg-nancy-associated plasma protein A (PAPP-A) [26]. Thepresence of PAPP-A has been demonstrated in follicularfluid, and levels were higher in estrogen-dominant follicles

[27]. This protein and its possible role as an IGFBP-4 pro-tease will no doubt be the focus of considerable interest inthe near future.

There have been several reports of high levels of IGFBP-4 in follicular fluid from polycystic ovaries. Falling levelsof IGFBP-4 were found in a spontaneous preovulatory fol-licle from a woman with polycystic ovaries [28]. Given thepotency of the inhibitory effects of IGFBP-4 on follicularcells, it is possible to envision a role for IGFBP-4 in theapparent inability of these cells to respond to FSH in vivo.Removal of the cells from the follicular endocrine milieuby culturing them would then allow steroidogenesis to con-tinue. Alternatively, increasing the intrafollicular levels ofFSH by treatment may induce IGFBP-4 proteolysis and al-low steroidogenesis to proceed. It is impossible to deter-mine whether the follicles have stopped growing becauseIGFBP-4 levels are higher or vice versa.

IGFBP-4 has a potent IGF-independent antigonadotropiceffect on human granulosa cells. No similar effects wereseen for IGFBP-2 or -3. These results indicate a physiolog-ical reason for the preferential proteolytic cleavage ofIGFBP-4 in healthy follicles. The mode of action ofIGFBP-4 and the role of IGFBP-2 in granulosa cells arecurrently under investigation.

REFERENCES

1. Voutillainen R, Franks S, Mason HD, Martikainen H. Expression ofinsulin-like growth factor (IGF) binding protein and IGF receptormRNAs in normal and polycystic ovaries. J Clin Endocrinol Metab1996; 81:1003–1008.

2. Willis DS, Mason HD, Watson H, Franks S. Developmentally regu-lated responses of human granulosa cells to insulin-like growth factors(IGFs): IGF-I and IGF-II action mediated via the type-I IGF receptor.J Clin Endocrinol Metab 1998; 83:1256–1260.

3. Giudice L, van-Dessel HJ, Cataldo NA, Chandresekher YA, Yap OW,Fauser BC. Circulating and ovarian IGF-binding proteins: potentialroles in normo-ovulatory cycles and in polycystic ovarian syndrome.Prog Growth Factor Res 1995; 6:397–408.

4. Mason HD, Davies SC, Franks S, Holly JMP. Insulin-like growth fac-tor I and II (IGF I and II), IGF binding proteins and IGFBP proteasesare produced by normal and polycystic human ovaries. J Clin Endo-crinol Metab 1996; 81:206–213.

5. San Roman GA, Magoffin DA. Insulin-like growth factor binding pro-teins in healthy and atretic follicles during natural menstrual cycles. JClin Endocrinol Metab 1993; 76:625–632.

6. Cataldo NC, Woodruff TK, Giudice LC. Regulation of insulin-likegrowth factor binding protein production by human luteinizing gran-ulosa cells cultured in defined medium. J Clin Endocrinol Metab 1993;76:207–215.

7. Cwyfan-Hughes S, Mason HD, Franks S, Holly JMP. Modulation ofthe insulin-like growth factor binding proteins by follicle size in thehuman ovary. J Endocrinol 1997; 154:35–43.

8. Mason HD, Willis D, Holly JMP, Cwyfan-Hughes SC, Seppala M,Franks S. Inhibitory effects of insulin-like growth factor-binding pro-teins on steroidogenesis by human granulosa cells in culture. Mol CellEndocrinol 1992; 89:R1–R4.

9. Mason HD, Cwyfan-Hughes S, Franks S, Holly JMP. Potent inhibitionof human ovarian steroidogenesis by insulin-like growth factor bind-ing protein-4. J Clin Endocrinol Metab 1998; 83:284–287.

10. Mason HD, Willis DS, Beard RW, Winston RML, Margara R, FranksS. Estradiol production by granulosa cells of normal and polycysticovaries: relationship to menstrual cycle history and concentrations ofgonadotrophins and sex steroids in follicular fluid. J Clin EndocrinolMetab 1994; 79:1355–1360.

11. Mason HD, Davies S, Franks S, Holly JMP. Insulin-like growth factor-I (IGF-I) inhibits IGFBP-1 whilst stimulating estradiol production bygranulosa cells from normal and polycystic human ovaries. J ClinEndocrinol Metab 1993; 76:1275–1279.

12. Abeyasekara DR, Jones PM, Persaud SJ, Michael AE, Flint APS.Prostaglandin F2a activates protein kinase c in human ovarian cells.Mol Cell Endocrinol 1993; 91:51–57.

13. Kull FC, Jacobs S, Su YF, Svoboda ME, Van Wyk JJ, Cuatrecasas P.

781IGF-INDEPENDENT IGFBP-4 EFFECTS IN HUMAN OVARIES

Monoclonal antibodies to receptors for insulin and somatomedin C. JClin Chem 1998; 258:6561–6566.

14. Gilling-Smith C, Willis DS, Beard RW, Franks S. Hypersecretion ofandrostenedione by isolated thecal cells from polycystic ovaries. JClin Endocrinol Metab 1994; 79:1158–1165.

15. Willis DS, Mason HD, Gilling-Smith C, Franks S. Modulation byinsulin of follicle stimulating hormone and luteinising hormone ac-tions in human granulosa cells of normal and polycystic ovaries. JClin Endocrinol Metab 1996; 81:302–309.

16. Mossman T. Rapid colorimetric assay for cellular growth and survival:application to proliferation and cytotoxicity assays. J Immunol Meth-ods 1983; 65:55–63.

17. Geisthovel F, Moretti-Rojas IM, Asch RH, Rojas FJ. Expression ofinsulin-like growth factor II (IGF-II) messenger ribonucleic acid(mRNA) but not IGF-I mRNA in human preovulatory granulosa cells.Hum Reprod 1989; 4:899–902.

18. Jones JI, Clemmons DR. Insulin-like growth factors and their bindingproteins: biological actions. Endocr Rev 1995; 16:3–34.

19. Li WL, Fawcett J, Widmer HR, Fielder PJ, Rabkin R, Keller GA.Nuclear transport of insulin-like growth factor binding protein-3 inopossum kidney cells. Endocrinology 1997; 138:1763–1766.

20. Jaques G, Noll K, Wegmann B, Witten S, Kogen E, Radulescu RT,Havemann K. Nuclear localization of insulin-like growth factor bind-ing protein-3 in a lung cancer cell line. Endocrinology 1997; 138:1767–1770.

21. Voutilainen R, Miller WL. Coordinate trophic hormone regulation ofmRNAs for insulin-like growth factor II and the cholesterol side-chain

cleavage enzyme, P450scc, in human steroidogenic tissues. Proc NatlAcad Sci U S A 1987; 84:1590–1594.

22. Ramasharma K, Li CH. Human pituitary and placental hormones con-trol human insulin-like growth factor II secretion in human granulosacells. Proc Natl Acad Sci U S A 1987; 84:2643–2647.

23. Wright R, Perks C, Holly JMP, Mason HD. Production and proteolysisof IGFBP-4 in the human ovary. In: Program of the Insulin-LikeGrowth Factor Binding Protein Workshop; 2000; Terrigal, Australia(abstract).

24. Choi D, Putowski LT, Fielder PJ, Rosenreld RG, Rohan RM, AdashiEY. Characterisation and hormonal regulation of granulosa cell-de-rived insulin-like growth factor binding protein-4. J Soc GynaecolInvest 1996; 3:145–151.

25. Iwashita M, Kudo Y, Takeda Y. Effect of follicle stimulating hormoneand insulin-like growth factor-I on proteolysis of IGFBP-4 in humangranulosa cells. Mol Hum Reprod 1998; 4:401–406.

26. Lawrence JB, Oxvig C, Overgaard MT, Sottrup-Jensen L, Gleich GJ,Hays LG, Yates III JR, Conover CA. The insulin-like growth factor(IGF)-dependent IGF binding protein-4 protease secreted by humanfibroblasts is pregnancy associated plama protein-A. Proc Natl AcadSci U S A 1999; 96:3149–3153.

27. Conover CA, Oxvig C, Overgaard MT, Christiansen M, Giudice LC.Evidence that the insulin-like growth factor binding protein-4 proteasein human ovarian follicular fluid is pregnancy associated plasma pro-tein-A. J Clin Endocrinol Metab 1999; 84:4742–4745.

28. Magoffin DA, San Roman GA, Muderspach LI. Insulin-like growthfactor binding proteins in a natural pre-ovulatory follicle from a wom-an with polycystic ovary syndrome. Hum Reprod 1995; 10:2248–2252.