Specific binding sites for insulin-like growth factor I in the ovarian stroma of women with polycystic ovarian disease and stromal hyperthecosis Manubai Nagamani, MD, and Charles A. Stuart, MD Galveston, Texas Women with polycystic ovarian disease and hyperthecosis have hyperinsulinemia and insulin resistance. It is possible that insulin in supraphysiologic concentration exerts its steroidogenic action on ovarian stromal cells through insulin-like growth factor I receptors. We undertook this study to investigate whether the ovarian stroma of women with hyperthecosis has specific binding sites for insulin or insulin-like growth factor I. Ovarian stromal tissue was obtained from seven women with normal ovulatory cycles and from five women with hyperthecosis of the ovaries. Binding studies with insulin tagged with iodine 125 and insulin-like growth factor I tagged with iodine 125 revealed specific binding sites both for insulin and insulin-like growth factor I in the ovarian stroma. The binding of insulin tagged with iodine 125 in the ovarian stroma of women with hyperthecosis (3.4% ± 1.1 % (± SE) per 100 fLg protein) was significantly (p < 0.04) lower than that observed in normal premenopausal women (8.3% ± 1.6% per 100 fLg protein). By contrast, the specific binding of insulin-like growth factor I tagged with iodine 125 in the ovarian stroma of women with hyperthecosis (7.1 % ± 1.7% per 100 fLg protein) was higher than that observed in the ovarian stroma from normal women (4.5% ± 1.7% per 100 fJ..g protein), although the difference was not statistically significant. The affinity constants for these high-affinity receptors were similar (1.2 to 3.6 x 1D· Llmol) in the two groups of women. These results indicate that (1) the ovarian stroma has specific binding sites both for insulin and for insulin-like growth factor I; (2) in women with hyperthecosis, the ovarian stroma has decreased binding sites for insulin but has normal concentrations of insulin-like growth factor I receptors; and (3) in women with hyperthecosis, stimulation of ovarian androgen synthesis by insulin may be mediated through the insulin-like growth factor I receptors. (AM J OaSTET GVNECOL 1990;163:1992-7.)
Key words: Insulin-like growth factor I receptors, insulin receptors, ovarian stroma, hyperthecosis, polycystic ovarian disease Insulin resistance and hyperinsulinemia are often seen in association with ovarian hyperandrogenism. I -3 A positive correlation between peripheral insulin levels and ovarian vein androgen levels has been observed in women with hyperandrogenism as a result of hyperthecosis of the ovaries! In vitro studies indicate that insulin stimulates ovarian stromal androgen synthesis. 5 However, patients with hyperinsulinemia have extreme insulin resistance that is a result of either defective insulin receptors or a postreceptor defect.5. 7 It is hard to explain how insulin exerts its steroidogenic action on the ovarian stroma in these women because its action in other target tissues is defective. It has been shown From the Department of Obstetrics and Gynecology and Department of Medicine, University of Texas Medical Branch. Supported by National Institutes ofHealth Research Grant Nos. RO lea 45181 and R01-DK 33749 and a grant from the National Institutes of Health general clinical research centers (M 0 1-RR0073). Presented at the Thirty-seventh Annual Meeting of the Society for Gynecologic Investigation, St. Louis, Missouri. March 21-24. 1990. Reprint requests: Manubai Nagamani. MD, Department ofObstetries and Gynecology, University of Texas Medical Branch, Galveston. TX
77550. 6/6/24720
1992
that insulin in high concentration is capable of binding to insulin-like growth factor I receptors." It is possible that insulin, which is present in supraphysiologic concentrations in women with polycystic ovarian disease and hyperthecosis, exerts its steroidogenic action on the ovarian stromal cells through insulin-like growth factor I receptors. We undertook this study to investigate whether normal ovarian stroma has specific binding sites for insulinlike growth factor I and insulin and whether the number of these receptors is increased or decreased in women with hyperthecosis.
Material and methods Patients. We obtained ovarian stromal tissue from seven normal premenopausal women undergoing hysterectomy for nonendocrine problems (myoma uterus, endometriosis, and cervical dysplasia). All of them had normal ovulatory cycles and had no hirsutism. The mean age was 39 years (range, 30 to 46 years) and the mean body weight was 160 pounds (range, 126 to 198 pounds). The fasting glucose and insulin levels were normal. Ovarian stromal tissue was also obtained from
Insulin-like growth factor I receptors in ovarian stroma
Volume 163 Number 6, Part I
1993
Table I. Clinical characteristics of women with hyperthecosis Patient
Age (yr)
% Ideal body weight
A B C D E
38 32 30 30 38
160 164 200 168 163
Hirsutism
Severe Severe Severe Severe Severe
Virilization
Clitorimegaly Clitorimegaly; temporal balding Temporal balding Clitorimegaly; temporal balding Clitorimegaly
Table II. Hormonal studies in women with hyperthecosis Patient
A B
C
D E
Normal range
Testosterone (ng/dl)
DHEAS (nglml)
185 612 138 249 400 20-90
720 300 700 627 550 500-3000
LHIFSH (mIUlml)
17113
23/5 13/8
4110 7112 5-15
Fasting insulin (p.mml)
Insulin response*
38 212 52 42 51 3-22
317 2634 685 2290 650 33-167
LH, Luteinizing hormone; FSH, follicle-stimulating hormone.
*Cumulative sum of serum insulin concentrations 1,2, and 3 hours after oral administration of 75 gm glucose. five virilized women with a long history of hirsutism who were undergoing laparotomy. Clinical characteristics and endocrine studies of these five women are summarized in Tables I and II. All these women were obese, had anovulatory cycles, and had been hirsute for many years. All of them were nulliparous and had failed to ovulate with clomiphene citrate (Clomid). None had acanthosis nigricans. The testosterone levels were high. The dehydroepiandrosterone (DHEA) sulfate levels were in the low-normal range. After a highcarbohydrate (300 gm) diet was observed for 3 days, a standard oral glucose tolerance test was performed in all subjects around 8 AM. Blood samples for glucose and insulin levels were obtained before and hourly for 3 hours after oral glucose (75 gm) administration. Fasting insulin levels and insulin response to oral glucose were elevated in all patients whereas the glucose levels were normal. All except patient D underwent hysterectomy with bilateral salpingo-oophorectomy for severe progressive hirsutism and virilization that did not respond to ovarian suppression with birth-control pills. A wedge biopsy of the ovaries was performed in patient D to rule out virilizing ovarian tumor because the testosterone levels were in the tumor range. In all patients, the ovaries were found to be enlarged to two to three times the normal size. Cut section of the ovaries revealed abundant ovarian stroma with few small follicular cysts in the cortex of the ovary. Histologic examination revealed extensive luteinization of the ovarian stroma in patients B, D, and E. Nests of luteinized cells were present in the ovarian stroma of patients A and C. Binding studies with ovarian stroma. To prepare ovarian microsomal membranes, we used with mi-
nor modification a technique developed for cartilage plasma membranes." Ovarian stromal tissue (400 to 500 mg) was homogenized in 15 ml 0.01 mollL imidazole, 0.25 mollL sucrose, and 0.005 mollL ethylenediaminetetraacetic acid (EDTA), pH 7.5 (ISE buffer) at 00 C for 60 seconds at half-speed with a polytron (Brinkman Instruments, Westbury, N.Y.). The homogenate was centrifuged at 1000 g for 15 minutes to remove nuclei, most mitochondria, and tissue fragments. The resulting supernatant was centrifuged at 35,000 g for 60 minutes to provide a final pellet that was resuspended in 0.001 mollL imidazol, pH 7.5. Insulin and insulin-like growth factor I binding studies were performed in 0.1 mollL HEPES acid, 0.120 mollL sodium chloride, 0.0012 mollL magnesium sulfate, 0.001 mollL EDTA, 0.01 mollL glucose, 0.015 mollL sodium acetate, and 1% bovine serum albumin, pH 8.0, as previously described. \0 Studies designed for Scatchard analysis of binding parameters consisted of eight different concentrations of insulin performed in triplicate. The concentration of insulin tagged with iodine 125 was approximately 0.05 x 10- 9 mollL. Unlabeled porcine insulin was added to achieve concentrations of 0, 0.2, 0.5, 1, 2, 10, 20, and 1000 x 10- 9 moll L and the membrane protein concentration was ::::; 1 mg / ml. For insulin-like growth factor I binding studies, unlabeled insulin-like growth factor I and ['25I]insulin-like growth factor I were used in similar concentrations. The mixtures were incubated at 4 C for 15 to 18 hours and the membranes were sedimented by centrifugation at 18,000 g for 15 minutes and washed once in binding buffer. The results were expressed as the percentage of total radioal':tivity bound per 100 tJ-g membrane protein. Protein was quantitated 0
1994 Nagamani and Stuart
-- 12
~
0 0
....
10
'Q)
Q.
~ -...;>
8
0
z
:::> 0
6
OJ
Z
::J :::> en
4
z 2 T
December 1990 Am J Obstet Gynecol
0 - 0 normal
e-e hyperthecosis
l~!
I1 I ~
"! ~"6 -......,'6
i 0.20
0.40
0 0.80
0.80
1.00
BOUND (pmoI/mg)
It)
....
N
0
o
0.1
1
10
UNLABELED INSULIN (nM) Fig. 1. Mean (± SE) specific insulin binding to ovarian stromal membranes from seven premenopausal women and five women with ovarian hyperthecosis. Inset shows Scatchard plot of same data.
by the method of Lowry et al. II Nonspecific binding was determined in the presence of at least 200-fold molar excess of unlabeled insulin and insulin-like growth factor I, respectively. Analysis of binding data was performed with a computer program designed to provide the best fit of a two-receptor model to the Scatchard plot. The program performs an iterative curvestripping procedure in which a curvilinear Scatchard plot is recalculated stepwise, subtracting an increasing number of counts due to low-affinity specific binding until the optimal high-affinity straight line is achieved. '2 The independent Student t test was used to compare differences in specific binding between control women and women with hyperthecosis. A p value < 0.05 was considered significant. Specificity studies. To evaluate the specificity of insulin-like growth factor I binding, ovarian stromal membranes were incubated in the presence of a constant amount of (1 x 105 counts per minute) of [125I]insulin-like growth factor I, with or without increasing concentrations of unlabeled (I to 100 ngltube) insulin-like growth factor I or insulin (1 to 1000 ng/ml).
Results Insulin binding studies. Our results indicate the presence of specific binding sites for insulin in normal ovarian stromal tissue. The percentage of total insulin binding in normal stroma was 8.3% ± 1.6% (± SE) per 100 fLg protein and ranged from 4.3% to 15.2% per 100 fLg protein. In women with hyperthecosis, the percent binding of [125I]insulin (3.4% ± 1.1 % ± SE) per
100 fLg protein; range, 0% to 6.7% per 100 fLg protein) was significantly lower (p < 0.04) than that observed in normal premenopausal women. The ovarian stroma of patient D did not have any specific binding sites for insulin. Fig. 1 shows the competition curves for binding of [12 5 I]insulin to ovarian stromal membranes of normal premenopausal women and women with hyperthecosis. The inset of Fig. 1 shows the mean data for Scatchard plots. The Scatchard plots were curvilinear, indicating either two binding sites of differing affinity or "negative cooperativity." The Scatchard plots of the high-affinity segments of the data (first five points) were parallel, indicating that the difference in insulin binding was a result of a difference in the concentration of the receptors rather than differences in affinity. The affinity constants for the specific high-affinity insulin receptors from these preparations were similar (3.5 ± 1.0 x 10" L/mol; normal ovarian stroma; and 3.5 ± 0.8 x 10" L/moll hyperthecosis). The mean receptor concentration in the ovarian stroma of women with hyperthecosis (0.16 ± 0.07 pmollmg membrane protein) was lower than that in the normal ovarian stroma (0.48 ± 0.18 pmollmg membrane protein). Insulin-like growth factor I binding studies. Specific binding sites for insulin-like growth factor I were present in the ovarian stroma from normal premenopausal women. The percentage of total binding of insulin-like growth factor I to normal ovarian stroma was 4.5% ± l.7% per 100 fLg protein (range, 2.0% to 7.4% 1100 fLg protein). The percentage of total binding insulin-like growth factor I in the ovarian stroma of women with hyperthecosis (7.1 % ± l. 7% per 100 fLg
Insulin-like growth factor I receptors in ovarian stroma
Volume 163 Number 6. Pan I
,...... [ o o ..-
12
10
8 o
z o
:::> CD
I
6
T
ION
..-
I
~
. - . hyperthecosis
'"-----l
4
0.110
T
~.,.
0.00
~_L-l---'--'--_'--_'-----'
?---Ol-""l~,:..:.·~·m ,~ 1
v
2
0
1.00
j::
I
........
I.&..
(!)
0 - 0 normal
1995
1.00
~ ,,~(,moV"'" T
? o
0.1
1
10
UNLABELED IGF-I (nM) Fig. 2. Mean (± SE) specific insulin-like growth factor I (IGF-I) binding to ovarian stromal membranes from premenopausal women and women with hyperthecosis. Inset shows Scatchard plot of same data.
protein) was higher than that observed in normal ovarian stroma, but the difference was not statistically significant (Fig. 2). Specific binding sites for insulin-like growth factor I were present in all the ovaries with hyperthecosis (range, 4.1 % to 12.7% per 100 f,Lg protein). In patient D, the ovarian stroma did have binding sites for insulin-like growth factor I, even though there was no demonstrable insulin binding. The inset of Fig. 2 shows the mean data from Scatchard plots. The affinity constants for these insulin-like growth factor I receptors were 3.6 ± 1.5 x 10" Llmol in hyperthecosis and 1.2 ± 0.3 x 10 9 LI mol in the normal ovarian stroma. The difference was not statistically significant. Specificity studies. Fig. 3 shows the competition curves for binding of [12;I]insulin-like growth factor I to ovarian stromal membranes prepared from normal ovarian stroma and women with hyperthescosis in the absence and presence of varying concentrations of unlabeled insulin-like growth factor I, insulin-like growth factor II, and insulin. Increasing concentrations (1 to 100 nmol!L) of unlabled insulin-like growth factor I resulted in dose-dependent saturable displacement of specific insulin-like growth factor I binding. Insulin was nonreactive at the final concentration of 1000 nmol! L and insulin-like growth factor II was only weakly effective. Results of our specificity studies indicate that human ovarian stromal cells possess separate receptors for insulin and insulin-like growth factor I. Comment
Specific binding sites for insulin have been shown in the ovarian stroma of normal and polycystic ova-
ries.I:\· 14 However, we are not aware of any previous reports that compare the receptor concentrations in the normal and polycystic ovaries. Our present data indicate that the specific binding sites for insulin are significantly decreased in the ovarian stroma of women with hyperthecosis. Lack of a difference in receptor affinity, as shown by the Scatchard plots, indicates that the reduced [12 5 I]insulin binding is primarily due to a reduction in the number of receptors. This decrease in insulin receptor concentrations could be a result of either a primary receptor defect or a down regulation in response to elevated insulin levels. Poretskey et a\. I', observed a 40% decrease in insulin binding in ovarian tissue when the concentration of insulin in the preincubation medium was 10 ng/ml and complete obliteration of [12C'I]insulin binding when the concentration was >50 ng/m\. Maximal plasma insulin concentration of 300 to 1000 f,LU Iml that we observed in our patients after oral glucose ingestion is similar to the dosages used in this study (50 ng/ml- 1250 f,LU/ml). In the present study, all our patients with hyperthecosis had hyperinsulinemia with normal glucose levels, a finding that indicates insulin resistance and defective insulin action on carbohydrate metabolism. The ability of insulin to alter ovarian function in the face of insulin resistance could be due to point mutations in the insulin receptor that lead to an inability of the insulin receptor to modulate glucose metabolism but do not affect the ability of insulin to bind to the receptor and modulate other functions such as steroidogenesis. The action of hyperinsulinemia on ovarian androgen production may be a result of direct effects of insulin on its own
1996
Nagamani and Stuart
December 1990 Am J Obstet Gynecol
,-...
X
c
E 0
110 INSULIN
100
+"
C
Q)
0
L.
90
IGF-II
Q)
a.
'-'"
Cl
z
80
:J
0
CD
70
I
lL. C,!)
I
Ii) N
......
60 50
IGF-I
0
0.1
1
10
100
1000
UNLABELED LIGAND CONCENTRATION (nM) Fig. 3. Specificity of insulin-like growth factor I (IGF-I) binding
receptor or cross-over stimulation via the insulin-like growth factor I receptor. Our specificity data shown in Fig. 3 do not support crossover action in this subject because insulin concentrations as high as 1000 nmollL did not displace [! 25 I]insulin-like growth factor I. However, in many tissues the relative potency of insulin in the prevention of binding of insulin-like growth factor I is 1/50 to 1/200, suggesting that hyperinsulinemia could result in activation of the insulin-like growth factor I receptor." This supposition is further supported by the fact that patients with type B severe insulin resistance, who have antibodies to insulin receptors, have ovarian hyperandrogenism. i!; Insulin-like growth factor I receptors have previously been shown to be present in the rat interstitial and granulosa cells and swine granulosa cells.I7·!" Specific binding sites for insulin-like growth factor I have also been shown to be present on human granulosa cells. 20 We showed for the first time that there are insulin-like growth factor I receptors in the normal human ovarian stroma. Unlike insulin receptors, the mean receptor concentration of insulin-like growth factor I in ovarian stroma of women with hyperthecosis was higher than that in normal ovarian stroma, even though the difference did not achieve statistical significance. It has been shown in the rat that experimetnal hyperinsulinemia upregulates ovarian insulin-like growth factor I receptors. 2 ! Small differences in the receptor concentration between the stroma from normal ovary and that from the polycystic ovary could also be a result of the differences in the cell type. Studies by Erikson et al. 22 indicate that the ovarian stromal tissue is heteroge-
to
ovarian stromal membranes.
neous and contains primary interstitial cells that have little or no capacity to synthesize steroids and secondary interstitial cells that secrete androgens. In the ovarian stroma of polycystic ovaries, there is a marked increase in the number of steroidogenically active secondary interstitial cells. It is possible that these secondary interstitial cells have an increased number of binding sites for insulin-like growth factor I and the stimulation of androgen synthesis by insulin is mediated through insulin-like growth factor I receptors. The only previous attempt to identify insulin-like growth factor I receptors in the human ovarian stroma is that of Poretsky and colleagues.!4 In that study, minced ovarian tissue rather than stromal membrane preparation was incubated with [! 25 I]insulin-like growth factor I, and the failure to identify insulin-like growth factor I receptors could be because of the difference in methods. After modification of the assay conditions, these authors were able to show specific, high-affinity insulin-like growth factor I binding sites on human granulosa cells. 20 Inasmuch as the human ovary has both insulin-like growth factor I and insulin receptors, insulin could mediate its action on ovarian steroidogenesis either through its own receptor or through insulin-like growth factor I receptors. In the normal ovary under physiologic conditions, insulin may exert its effect mainly through its own receptors, whereas in polycystic ovarian disease and hyperthecosis, when the insulin receptors may be defective and insulin is present in pharmacologic concentrations, insulin might mediate its action through insulin-like growth factor I receptors. This is further supported by the fact that the ovarian
Volume 163 Number 6, Part 1
stroma of patient D, who had severe hyperinsulinemia and ovarian hyperandrogenism, had no measurable insulin receptors. In conclusion, our present data indicate that (1) the ovarian stroma has specific binding sites both for insulin and for insulin-like growth factor I; (2) in women with hyperthecosis, the ovarian stroma has decreased binding sites for insulin, but has normal or increased concentration of insulin-like growth factor I receptors; and (3) in women with hyperthecosis and hyperinsulinerr ia, stimulation of ovarian androgen synthesis by insuhL may be mediated through the insulin-like growth factor I receptors. REFERENCES 1. Chang Rj, Nakamura RM,judd HL, Kaplan SA. Insulin resistance in nonobese patients with polycystic ovarian disease. j Clin Endocrinol Metab 1983;57:356-9. 2. Shoupe D, Kumar DD, Lobo RA. Insulin resistance in polycystic ovary syndrome. AMj OBSTET GYNECOL 1983; 147:588-92. 3. Burgen GA, Given jR, Kitabchi AE. Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. j Clin Endocrinol Metab 1980;50: 113-6. 4. Nagamani M, Dinh TV, Kelver ME. Hyperinsulinemia in hyperthecosis of the ovaries. AM j OBSTET GYNECOL 1986; 154:384-9. 5. Barbieri RL, Makris A, Randall RW, Daniels G, Kistner RW, Ryan KJ. Insulin stimulates androgen accumulation in incubation of ovarian stroma obtained from women with hyperandrogenism. j Clin Endocrinol Metab 1986; 62:904-10. 6. Podskalny jM, Kahn CR. Cell culture studies on patients with extreme insulin resistance I receptor defects on cultured fibroblasts. j Clin Endocrinol Metab 1982;54: 261-8. 7. Bar RS, Muggeo M, Roth j, Kahn CR, Kavrankova j, Imperato-McGinley J. Insulin resistance, acanthosis Nigricans, and normal insulin receptors in a young woman: evidence for a post receptor defect. j Clin Endocrinol Metab 1978;47:620-5. 8. Rechler MM. The nature and regulation of the receptors for insulin-like growth factors. Annu Rev Physiol 1985; 47:425-42. 9. Stuart CA, Furlanetto RW, Lebovitz HE. The insulin re-
Insulin-like growth factor I receptors in ovarian stroma
10.
11. 12. 13. 14.
15. 16.
17.
18.
19.
20.
21.
22.
1997
ceptor of embryonic chick cartilage. Endocrinology 1979; 105: 1293-302. Nagamani M, Hannigan EV, Dinh TV, Stuart CA. Hyperinsulinemia and stromal luteinization of the ovaries in postmenopausal women with endometrial cancer. j Clin Endocrinol Metab 1988;67:144-8. Lowry OH, Rosebrough Nj, Farr AL, Randall RJ. Protein measurement with the folin phenol reagent. j Bioi Chern 1951; 193:267. Stuart CA. Phylogenetic distance form mean correlates with immunologic cross-reactivity among liver insulin receptors. Comp Biochem Physiol 1986;84B:167-72. Poretsky L, Smith D, Seibel M, Pazianos M, Moses AC, Flier jS. Specific insulin binding sites in human ovary. j Clin Endocrinol Metab 1984;59(4):809-11. Poretsky L, Grigorescu F, Seibel M, Moses AC, Flier jS. Distribution and characterization of insulin and insulinlike growth factor I receptors in normal human ovary. j Clin Endocrinol Metab 1985;61 :728-34. Poretsky L, Bhargava G, Kalin MF, Wolf SA. Regulation of insulin receptors in the human ovary: in vitro studies. j Clin Endocrinol Metab 1988;67:774-8. Taylor SI, Dons RF, Hernandez E, Roth j, Gordon P. Insulin resistance associated with androgen excess in women with auto antibodies to the insulin receptor. Ann Intern Med 1982;97:851-5. Hernandez RE, Resnick CE, Svolboda EM, VanWyk jJ. Somatomedin-C/insulin-like growth factor I as an enhancer of androgen biosynthesis by cultured rat ovarian cells. Endocrinology 1988;122:1603-12. Adashi EY, Resnick EC, Hernandez ER, Svoboda ME, VanWyk jJ. 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. Veldhuis jD, Furlanetto RW. Tropic action of human somatomedin C/insulin-like growth factor I on ovarian cells: in vitro studies with swine granulosa cells. Endocrinology 1985; 116: 1235-42. Gates GS, Bayer S, Seibel M, Poretsky L, Flier Sj, Moses AC. Characterization of insulin-like growth factor binding to human granulosa cells obtained during in vitro fertilization. j Recept Res 1987;7:885-902. Poretsky L, Glover B, Laumas V, Kalin M, Dunaif A. The effects of experimental hyperinsulinemia on steroid secretion, ovarian [125 Ilinsulin binding, and ovarian 125 [ Ilinsulin-like growth factor I binding in the rat. Endocrinology 1988;122:581-5. Erickson GF, Magoffin DA, Dyer CA, Hofeditz C. The ovarian androgen producing cells: a review of structure function relationship. Endocrinol Rev 1985;6:371-99.
Key words: Insulin-like growth factor I receptors, insulin receptors, ovarian stroma, hyperthecosis, polycystic ovarian disease Insulin resistance and hyperinsulinemia are often seen in association with ovarian hyperandrogenism. I -3 A positive correlation between peripheral insulin levels and ovarian vein androgen levels has been observed in women with hyperandrogenism as a result of hyperthecosis of the ovaries! In vitro studies indicate that insulin stimulates ovarian stromal androgen synthesis. 5 However, patients with hyperinsulinemia have extreme insulin resistance that is a result of either defective insulin receptors or a postreceptor defect.5. 7 It is hard to explain how insulin exerts its steroidogenic action on the ovarian stroma in these women because its action in other target tissues is defective. It has been shown From the Department of Obstetrics and Gynecology and Department of Medicine, University of Texas Medical Branch. Supported by National Institutes ofHealth Research Grant Nos. RO lea 45181 and R01-DK 33749 and a grant from the National Institutes of Health general clinical research centers (M 0 1-RR0073). Presented at the Thirty-seventh Annual Meeting of the Society for Gynecologic Investigation, St. Louis, Missouri. March 21-24. 1990. Reprint requests: Manubai Nagamani. MD, Department ofObstetries and Gynecology, University of Texas Medical Branch, Galveston. TX
77550. 6/6/24720
1992
that insulin in high concentration is capable of binding to insulin-like growth factor I receptors." It is possible that insulin, which is present in supraphysiologic concentrations in women with polycystic ovarian disease and hyperthecosis, exerts its steroidogenic action on the ovarian stromal cells through insulin-like growth factor I receptors. We undertook this study to investigate whether normal ovarian stroma has specific binding sites for insulinlike growth factor I and insulin and whether the number of these receptors is increased or decreased in women with hyperthecosis.
Material and methods Patients. We obtained ovarian stromal tissue from seven normal premenopausal women undergoing hysterectomy for nonendocrine problems (myoma uterus, endometriosis, and cervical dysplasia). All of them had normal ovulatory cycles and had no hirsutism. The mean age was 39 years (range, 30 to 46 years) and the mean body weight was 160 pounds (range, 126 to 198 pounds). The fasting glucose and insulin levels were normal. Ovarian stromal tissue was also obtained from
Insulin-like growth factor I receptors in ovarian stroma
Volume 163 Number 6, Part I
1993
Table I. Clinical characteristics of women with hyperthecosis Patient
Age (yr)
% Ideal body weight
A B C D E
38 32 30 30 38
160 164 200 168 163
Hirsutism
Severe Severe Severe Severe Severe
Virilization
Clitorimegaly Clitorimegaly; temporal balding Temporal balding Clitorimegaly; temporal balding Clitorimegaly
Table II. Hormonal studies in women with hyperthecosis Patient
A B
C
D E
Normal range
Testosterone (ng/dl)
DHEAS (nglml)
185 612 138 249 400 20-90
720 300 700 627 550 500-3000
LHIFSH (mIUlml)
17113
23/5 13/8
4110 7112 5-15
Fasting insulin (p.mml)
Insulin response*
38 212 52 42 51 3-22
317 2634 685 2290 650 33-167
LH, Luteinizing hormone; FSH, follicle-stimulating hormone.
*Cumulative sum of serum insulin concentrations 1,2, and 3 hours after oral administration of 75 gm glucose. five virilized women with a long history of hirsutism who were undergoing laparotomy. Clinical characteristics and endocrine studies of these five women are summarized in Tables I and II. All these women were obese, had anovulatory cycles, and had been hirsute for many years. All of them were nulliparous and had failed to ovulate with clomiphene citrate (Clomid). None had acanthosis nigricans. The testosterone levels were high. The dehydroepiandrosterone (DHEA) sulfate levels were in the low-normal range. After a highcarbohydrate (300 gm) diet was observed for 3 days, a standard oral glucose tolerance test was performed in all subjects around 8 AM. Blood samples for glucose and insulin levels were obtained before and hourly for 3 hours after oral glucose (75 gm) administration. Fasting insulin levels and insulin response to oral glucose were elevated in all patients whereas the glucose levels were normal. All except patient D underwent hysterectomy with bilateral salpingo-oophorectomy for severe progressive hirsutism and virilization that did not respond to ovarian suppression with birth-control pills. A wedge biopsy of the ovaries was performed in patient D to rule out virilizing ovarian tumor because the testosterone levels were in the tumor range. In all patients, the ovaries were found to be enlarged to two to three times the normal size. Cut section of the ovaries revealed abundant ovarian stroma with few small follicular cysts in the cortex of the ovary. Histologic examination revealed extensive luteinization of the ovarian stroma in patients B, D, and E. Nests of luteinized cells were present in the ovarian stroma of patients A and C. Binding studies with ovarian stroma. To prepare ovarian microsomal membranes, we used with mi-
nor modification a technique developed for cartilage plasma membranes." Ovarian stromal tissue (400 to 500 mg) was homogenized in 15 ml 0.01 mollL imidazole, 0.25 mollL sucrose, and 0.005 mollL ethylenediaminetetraacetic acid (EDTA), pH 7.5 (ISE buffer) at 00 C for 60 seconds at half-speed with a polytron (Brinkman Instruments, Westbury, N.Y.). The homogenate was centrifuged at 1000 g for 15 minutes to remove nuclei, most mitochondria, and tissue fragments. The resulting supernatant was centrifuged at 35,000 g for 60 minutes to provide a final pellet that was resuspended in 0.001 mollL imidazol, pH 7.5. Insulin and insulin-like growth factor I binding studies were performed in 0.1 mollL HEPES acid, 0.120 mollL sodium chloride, 0.0012 mollL magnesium sulfate, 0.001 mollL EDTA, 0.01 mollL glucose, 0.015 mollL sodium acetate, and 1% bovine serum albumin, pH 8.0, as previously described. \0 Studies designed for Scatchard analysis of binding parameters consisted of eight different concentrations of insulin performed in triplicate. The concentration of insulin tagged with iodine 125 was approximately 0.05 x 10- 9 mollL. Unlabeled porcine insulin was added to achieve concentrations of 0, 0.2, 0.5, 1, 2, 10, 20, and 1000 x 10- 9 moll L and the membrane protein concentration was ::::; 1 mg / ml. For insulin-like growth factor I binding studies, unlabeled insulin-like growth factor I and ['25I]insulin-like growth factor I were used in similar concentrations. The mixtures were incubated at 4 C for 15 to 18 hours and the membranes were sedimented by centrifugation at 18,000 g for 15 minutes and washed once in binding buffer. The results were expressed as the percentage of total radioal':tivity bound per 100 tJ-g membrane protein. Protein was quantitated 0
1994 Nagamani and Stuart
-- 12
~
0 0
....
10
'Q)
Q.
~ -...;>
8
0
z
:::> 0
6
OJ
Z
::J :::> en
4
z 2 T
December 1990 Am J Obstet Gynecol
0 - 0 normal
e-e hyperthecosis
l~!
I1 I ~
"! ~"6 -......,'6
i 0.20
0.40
0 0.80
0.80
1.00
BOUND (pmoI/mg)
It)
....
N
0
o
0.1
1
10
UNLABELED INSULIN (nM) Fig. 1. Mean (± SE) specific insulin binding to ovarian stromal membranes from seven premenopausal women and five women with ovarian hyperthecosis. Inset shows Scatchard plot of same data.
by the method of Lowry et al. II Nonspecific binding was determined in the presence of at least 200-fold molar excess of unlabeled insulin and insulin-like growth factor I, respectively. Analysis of binding data was performed with a computer program designed to provide the best fit of a two-receptor model to the Scatchard plot. The program performs an iterative curvestripping procedure in which a curvilinear Scatchard plot is recalculated stepwise, subtracting an increasing number of counts due to low-affinity specific binding until the optimal high-affinity straight line is achieved. '2 The independent Student t test was used to compare differences in specific binding between control women and women with hyperthecosis. A p value < 0.05 was considered significant. Specificity studies. To evaluate the specificity of insulin-like growth factor I binding, ovarian stromal membranes were incubated in the presence of a constant amount of (1 x 105 counts per minute) of [125I]insulin-like growth factor I, with or without increasing concentrations of unlabeled (I to 100 ngltube) insulin-like growth factor I or insulin (1 to 1000 ng/ml).
Results Insulin binding studies. Our results indicate the presence of specific binding sites for insulin in normal ovarian stromal tissue. The percentage of total insulin binding in normal stroma was 8.3% ± 1.6% (± SE) per 100 fLg protein and ranged from 4.3% to 15.2% per 100 fLg protein. In women with hyperthecosis, the percent binding of [125I]insulin (3.4% ± 1.1 % ± SE) per
100 fLg protein; range, 0% to 6.7% per 100 fLg protein) was significantly lower (p < 0.04) than that observed in normal premenopausal women. The ovarian stroma of patient D did not have any specific binding sites for insulin. Fig. 1 shows the competition curves for binding of [12 5 I]insulin to ovarian stromal membranes of normal premenopausal women and women with hyperthecosis. The inset of Fig. 1 shows the mean data for Scatchard plots. The Scatchard plots were curvilinear, indicating either two binding sites of differing affinity or "negative cooperativity." The Scatchard plots of the high-affinity segments of the data (first five points) were parallel, indicating that the difference in insulin binding was a result of a difference in the concentration of the receptors rather than differences in affinity. The affinity constants for the specific high-affinity insulin receptors from these preparations were similar (3.5 ± 1.0 x 10" L/mol; normal ovarian stroma; and 3.5 ± 0.8 x 10" L/moll hyperthecosis). The mean receptor concentration in the ovarian stroma of women with hyperthecosis (0.16 ± 0.07 pmollmg membrane protein) was lower than that in the normal ovarian stroma (0.48 ± 0.18 pmollmg membrane protein). Insulin-like growth factor I binding studies. Specific binding sites for insulin-like growth factor I were present in the ovarian stroma from normal premenopausal women. The percentage of total binding of insulin-like growth factor I to normal ovarian stroma was 4.5% ± l.7% per 100 fLg protein (range, 2.0% to 7.4% 1100 fLg protein). The percentage of total binding insulin-like growth factor I in the ovarian stroma of women with hyperthecosis (7.1 % ± l. 7% per 100 fLg
Insulin-like growth factor I receptors in ovarian stroma
Volume 163 Number 6. Pan I
,...... [ o o ..-
12
10
8 o
z o
:::> CD
I
6
T
ION
..-
I
~
. - . hyperthecosis
'"-----l
4
0.110
T
~.,.
0.00
~_L-l---'--'--_'--_'-----'
?---Ol-""l~,:..:.·~·m ,~ 1
v
2
0
1.00
j::
I
........
I.&..
(!)
0 - 0 normal
1995
1.00
~ ,,~(,moV"'" T
? o
0.1
1
10
UNLABELED IGF-I (nM) Fig. 2. Mean (± SE) specific insulin-like growth factor I (IGF-I) binding to ovarian stromal membranes from premenopausal women and women with hyperthecosis. Inset shows Scatchard plot of same data.
protein) was higher than that observed in normal ovarian stroma, but the difference was not statistically significant (Fig. 2). Specific binding sites for insulin-like growth factor I were present in all the ovaries with hyperthecosis (range, 4.1 % to 12.7% per 100 f,Lg protein). In patient D, the ovarian stroma did have binding sites for insulin-like growth factor I, even though there was no demonstrable insulin binding. The inset of Fig. 2 shows the mean data from Scatchard plots. The affinity constants for these insulin-like growth factor I receptors were 3.6 ± 1.5 x 10" Llmol in hyperthecosis and 1.2 ± 0.3 x 10 9 LI mol in the normal ovarian stroma. The difference was not statistically significant. Specificity studies. Fig. 3 shows the competition curves for binding of [12;I]insulin-like growth factor I to ovarian stromal membranes prepared from normal ovarian stroma and women with hyperthescosis in the absence and presence of varying concentrations of unlabeled insulin-like growth factor I, insulin-like growth factor II, and insulin. Increasing concentrations (1 to 100 nmol!L) of unlabled insulin-like growth factor I resulted in dose-dependent saturable displacement of specific insulin-like growth factor I binding. Insulin was nonreactive at the final concentration of 1000 nmol! L and insulin-like growth factor II was only weakly effective. Results of our specificity studies indicate that human ovarian stromal cells possess separate receptors for insulin and insulin-like growth factor I. Comment
Specific binding sites for insulin have been shown in the ovarian stroma of normal and polycystic ova-
ries.I:\· 14 However, we are not aware of any previous reports that compare the receptor concentrations in the normal and polycystic ovaries. Our present data indicate that the specific binding sites for insulin are significantly decreased in the ovarian stroma of women with hyperthecosis. Lack of a difference in receptor affinity, as shown by the Scatchard plots, indicates that the reduced [12 5 I]insulin binding is primarily due to a reduction in the number of receptors. This decrease in insulin receptor concentrations could be a result of either a primary receptor defect or a down regulation in response to elevated insulin levels. Poretskey et a\. I', observed a 40% decrease in insulin binding in ovarian tissue when the concentration of insulin in the preincubation medium was 10 ng/ml and complete obliteration of [12C'I]insulin binding when the concentration was >50 ng/m\. Maximal plasma insulin concentration of 300 to 1000 f,LU Iml that we observed in our patients after oral glucose ingestion is similar to the dosages used in this study (50 ng/ml- 1250 f,LU/ml). In the present study, all our patients with hyperthecosis had hyperinsulinemia with normal glucose levels, a finding that indicates insulin resistance and defective insulin action on carbohydrate metabolism. The ability of insulin to alter ovarian function in the face of insulin resistance could be due to point mutations in the insulin receptor that lead to an inability of the insulin receptor to modulate glucose metabolism but do not affect the ability of insulin to bind to the receptor and modulate other functions such as steroidogenesis. The action of hyperinsulinemia on ovarian androgen production may be a result of direct effects of insulin on its own
1996
Nagamani and Stuart
December 1990 Am J Obstet Gynecol
,-...
X
c
E 0
110 INSULIN
100
+"
C
Q)
0
L.
90
IGF-II
Q)
a.
'-'"
Cl
z
80
:J
0
CD
70
I
lL. C,!)
I
Ii) N
......
60 50
IGF-I
0
0.1
1
10
100
1000
UNLABELED LIGAND CONCENTRATION (nM) Fig. 3. Specificity of insulin-like growth factor I (IGF-I) binding
receptor or cross-over stimulation via the insulin-like growth factor I receptor. Our specificity data shown in Fig. 3 do not support crossover action in this subject because insulin concentrations as high as 1000 nmollL did not displace [! 25 I]insulin-like growth factor I. However, in many tissues the relative potency of insulin in the prevention of binding of insulin-like growth factor I is 1/50 to 1/200, suggesting that hyperinsulinemia could result in activation of the insulin-like growth factor I receptor." This supposition is further supported by the fact that patients with type B severe insulin resistance, who have antibodies to insulin receptors, have ovarian hyperandrogenism. i!; Insulin-like growth factor I receptors have previously been shown to be present in the rat interstitial and granulosa cells and swine granulosa cells.I7·!" Specific binding sites for insulin-like growth factor I have also been shown to be present on human granulosa cells. 20 We showed for the first time that there are insulin-like growth factor I receptors in the normal human ovarian stroma. Unlike insulin receptors, the mean receptor concentration of insulin-like growth factor I in ovarian stroma of women with hyperthecosis was higher than that in normal ovarian stroma, even though the difference did not achieve statistical significance. It has been shown in the rat that experimetnal hyperinsulinemia upregulates ovarian insulin-like growth factor I receptors. 2 ! Small differences in the receptor concentration between the stroma from normal ovary and that from the polycystic ovary could also be a result of the differences in the cell type. Studies by Erikson et al. 22 indicate that the ovarian stromal tissue is heteroge-
to
ovarian stromal membranes.
neous and contains primary interstitial cells that have little or no capacity to synthesize steroids and secondary interstitial cells that secrete androgens. In the ovarian stroma of polycystic ovaries, there is a marked increase in the number of steroidogenically active secondary interstitial cells. It is possible that these secondary interstitial cells have an increased number of binding sites for insulin-like growth factor I and the stimulation of androgen synthesis by insulin is mediated through insulin-like growth factor I receptors. The only previous attempt to identify insulin-like growth factor I receptors in the human ovarian stroma is that of Poretsky and colleagues.!4 In that study, minced ovarian tissue rather than stromal membrane preparation was incubated with [! 25 I]insulin-like growth factor I, and the failure to identify insulin-like growth factor I receptors could be because of the difference in methods. After modification of the assay conditions, these authors were able to show specific, high-affinity insulin-like growth factor I binding sites on human granulosa cells. 20 Inasmuch as the human ovary has both insulin-like growth factor I and insulin receptors, insulin could mediate its action on ovarian steroidogenesis either through its own receptor or through insulin-like growth factor I receptors. In the normal ovary under physiologic conditions, insulin may exert its effect mainly through its own receptors, whereas in polycystic ovarian disease and hyperthecosis, when the insulin receptors may be defective and insulin is present in pharmacologic concentrations, insulin might mediate its action through insulin-like growth factor I receptors. This is further supported by the fact that the ovarian
Volume 163 Number 6, Part 1
stroma of patient D, who had severe hyperinsulinemia and ovarian hyperandrogenism, had no measurable insulin receptors. In conclusion, our present data indicate that (1) the ovarian stroma has specific binding sites both for insulin and for insulin-like growth factor I; (2) in women with hyperthecosis, the ovarian stroma has decreased binding sites for insulin, but has normal or increased concentration of insulin-like growth factor I receptors; and (3) in women with hyperthecosis and hyperinsulinerr ia, stimulation of ovarian androgen synthesis by insuhL may be mediated through the insulin-like growth factor I receptors. REFERENCES 1. Chang Rj, Nakamura RM,judd HL, Kaplan SA. Insulin resistance in nonobese patients with polycystic ovarian disease. j Clin Endocrinol Metab 1983;57:356-9. 2. Shoupe D, Kumar DD, Lobo RA. Insulin resistance in polycystic ovary syndrome. AMj OBSTET GYNECOL 1983; 147:588-92. 3. Burgen GA, Given jR, Kitabchi AE. Correlation of hyperandrogenism with hyperinsulinism in polycystic ovarian disease. j Clin Endocrinol Metab 1980;50: 113-6. 4. Nagamani M, Dinh TV, Kelver ME. Hyperinsulinemia in hyperthecosis of the ovaries. AM j OBSTET GYNECOL 1986; 154:384-9. 5. Barbieri RL, Makris A, Randall RW, Daniels G, Kistner RW, Ryan KJ. Insulin stimulates androgen accumulation in incubation of ovarian stroma obtained from women with hyperandrogenism. j Clin Endocrinol Metab 1986; 62:904-10. 6. Podskalny jM, Kahn CR. Cell culture studies on patients with extreme insulin resistance I receptor defects on cultured fibroblasts. j Clin Endocrinol Metab 1982;54: 261-8. 7. Bar RS, Muggeo M, Roth j, Kahn CR, Kavrankova j, Imperato-McGinley J. Insulin resistance, acanthosis Nigricans, and normal insulin receptors in a young woman: evidence for a post receptor defect. j Clin Endocrinol Metab 1978;47:620-5. 8. Rechler MM. The nature and regulation of the receptors for insulin-like growth factors. Annu Rev Physiol 1985; 47:425-42. 9. Stuart CA, Furlanetto RW, Lebovitz HE. The insulin re-
Insulin-like growth factor I receptors in ovarian stroma
10.
11. 12. 13. 14.
15. 16.
17.
18.
19.
20.
21.
22.
1997
ceptor of embryonic chick cartilage. Endocrinology 1979; 105: 1293-302. Nagamani M, Hannigan EV, Dinh TV, Stuart CA. Hyperinsulinemia and stromal luteinization of the ovaries in postmenopausal women with endometrial cancer. j Clin Endocrinol Metab 1988;67:144-8. Lowry OH, Rosebrough Nj, Farr AL, Randall RJ. Protein measurement with the folin phenol reagent. j Bioi Chern 1951; 193:267. Stuart CA. Phylogenetic distance form mean correlates with immunologic cross-reactivity among liver insulin receptors. Comp Biochem Physiol 1986;84B:167-72. Poretsky L, Smith D, Seibel M, Pazianos M, Moses AC, Flier jS. Specific insulin binding sites in human ovary. j Clin Endocrinol Metab 1984;59(4):809-11. Poretsky L, Grigorescu F, Seibel M, Moses AC, Flier jS. Distribution and characterization of insulin and insulinlike growth factor I receptors in normal human ovary. j Clin Endocrinol Metab 1985;61 :728-34. Poretsky L, Bhargava G, Kalin MF, Wolf SA. Regulation of insulin receptors in the human ovary: in vitro studies. j Clin Endocrinol Metab 1988;67:774-8. Taylor SI, Dons RF, Hernandez E, Roth j, Gordon P. Insulin resistance associated with androgen excess in women with auto antibodies to the insulin receptor. Ann Intern Med 1982;97:851-5. Hernandez RE, Resnick CE, Svolboda EM, VanWyk jJ. Somatomedin-C/insulin-like growth factor I as an enhancer of androgen biosynthesis by cultured rat ovarian cells. Endocrinology 1988;122:1603-12. Adashi EY, Resnick EC, Hernandez ER, Svoboda ME, VanWyk jJ. 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. Veldhuis jD, Furlanetto RW. Tropic action of human somatomedin C/insulin-like growth factor I on ovarian cells: in vitro studies with swine granulosa cells. Endocrinology 1985; 116: 1235-42. Gates GS, Bayer S, Seibel M, Poretsky L, Flier Sj, Moses AC. Characterization of insulin-like growth factor binding to human granulosa cells obtained during in vitro fertilization. j Recept Res 1987;7:885-902. Poretsky L, Glover B, Laumas V, Kalin M, Dunaif A. The effects of experimental hyperinsulinemia on steroid secretion, ovarian [125 Ilinsulin binding, and ovarian 125 [ Ilinsulin-like growth factor I binding in the rat. Endocrinology 1988;122:581-5. Erickson GF, Magoffin DA, Dyer CA, Hofeditz C. The ovarian androgen producing cells: a review of structure function relationship. Endocrinol Rev 1985;6:371-99.
