Decreases in ovarian cytochrome p450c17α activity and serum free testosterone after reduction of insulin secretion in polycystic ovary syndrome

R E D U C I N G S E R U M I N S U L I N A N D OVA R I A N CY TO C H RO M E P 4 5 0 c 1 7a I N P O LYCYST I C OVA RY SY N D RO M E
DECREASES IN OVARIAN CYTOCHROME P450c17a ACTIVITY AND SERUM
FREE TESTOSTERONE AFTER REDUCTION OF INSULIN SECRETION
IN POLYCYSTIC OVARY SYNDROME
JOHN E. NESTLER, M.D., AND DANIELA J. JAKUBOWICZ, M.D.
ABSTRACT
Background
ovarian cytochrome P450c17a activity are both fea- genism. It affects approximately 6 percent tures of the polycystic ovary syndrome. P450c17a, of women of reproductive age.1 Insulin re- which is involved in androgen biosynthesis, has sistance accompanied by compensatory hyperinsuli- both 17a-hydroxylase and 17,20-lyase activities. nemia is a common feature of the syndrome, and Increased activity of this enzyme results in exag- both obese and nonobese women with the syndrome gerated conversion of progesterone to 17a-hydroxy- are more insulin-resistant and hyperinsulinemic than progesterone in response to stimulation by gonado- age- and weight-matched normal women.2-9 tropin. We hypothesized that hyperinsulinemia Hyperinsulinemia may play a pathogenetic part in stimulates ovarian P450c17a activity.
Methods
hyperandrogenism in women with the polycystic tions during fasting and the response of serum ovary syndrome by increasing ovarian androgen pro- 17a-hydroxyprogesterone to leuprolide, a gonado- duction and decreasing the serum sex hormone- tropin-releasing hormone agonist, and performed binding globulin concentration.10-17 Serum free tes- oral glucose-tolerance tests before and after oral ad- tosterone concentrations decline in women with the ministration of either metformin (500 mg three times polycystic ovary syndrome when their insulin secre- daily) or placebo for four to eight weeks in 24 obese tion is reduced by the administration of diazoxide15 women with the polycystic ovary syndrome.
or metformin18 or by diet.19,20 Furthermore, the ob- Results
servation that adolescent girls with hyperandro- the mean (ϮSE) area under the serum insulin genism have insulin resistance9 suggests that hyper- curve after oral glucose administration decreased insulinemia may play an early and central part in the from 9303Ϯ1603 to 4982Ϯ911 mU per milliliter per minute (56Ϯ10 to 30Ϯ6 nmol per liter per minute) pathogenesis of the polycystic ovary syndrome.
(Pϭ0.004). This decrease was associated with a Cytochrome P450c17a is a bifunctional enzyme reduction in the basal serum 17a-hydroxyproges- that has both 17a-hydroxylase and 17,20-lyase activ- terone concentration from 135Ϯ21 to 66Ϯ7 ng ities, and it is a key enzyme in the biosynthesis of per deciliter (4.1Ϯ0.6 to 2.0Ϯ0.2 nmol per liter) ovarian androgens. In ovarian theca cells, P450c17a (Pϭ0.01) and a reduction in the leuprolide-stimulat- converts progesterone to 17a-hydroxyprogesterone ed peak serum 17a-hydroxyprogesterone concentra- through its 17a-hydroxylase activity, and then con- tion from 455Ϯ54 to 281Ϯ52 ng per deciliter verts 17a-hydroxyprogesterone to androstenedi- (13.7Ϯ1.6 to 8.5Ϯ1.6 nmol per liter) (Pϭ0.01). The one through its 17,20-lyase activity. Androstenedi- serum 17a-hydroxyprogesterone values increased one is then converted to testosterone by the enzyme slightly in the placebo group. In the metformin group, the basal serum luteinizing hormone concen- tration decreased from 8.5Ϯ2.2 to 2.8Ϯ0.5 mlU per Many women with the polycystic ovary syndrome milliliter (Pϭ0.01), the serum free testosterone con- have increased ovarian cytochrome P450c17a activ- centration decreased from 0.34Ϯ0.07 to 0.19Ϯ0.05 ity,21,22 as evidenced by increased 17a-hydroxylase ng per deciliter (12Ϯ3 to 7Ϯ2 pmol per liter) and, to a lesser extent, 17,20-lyase activity, resulting (Pϭ0.009), and the serum sex hormone-binding in excessive ovarian androgen production. In these globulin concentration increased from 0.8Ϯ0.2 to women, a hallmark of increased ovarian P450c17a 2.3Ϯ0.6 mg per deciliter (29Ϯ7 to 80Ϯ21 nmol per li- activity is an exaggerated serum 17a-hydroxypro- ter) (PϽ0.001). None of these values changed signif- gesterone response to stimulation by gonadotropin- Conclusions
ovary syndrome, decreasing serum insulin concen-trations with metformin reduces ovarian cytochrome P450c17a activity and ameliorates hyperandro-genism. (N Engl J Med 1996;335:617-23.)1996, Massachusetts Medical Society.
From the Departments of Internal Medicine, Obstetrics and Gynecolo- gy, and Pharmacology and Toxicology, Division of Endocrinology and Me-tabolism, Medical College of Virginia, Virginia Commonwealth University,Richmond (J.E.N.); and the Department of Internal Medicine, Hospital deClinicas Caracas, Caracas, Venezuela (D.J.J.). Address reprint requests toDr. Nestler at the Medical College of Virginia, P.O. Box 980111, Rich-mond, VA 23298-0111.
Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. T h e New E n g l a n d Jo u r n a l of Me d i c i n e releasing hormone agonists, such as nafarelin,21-23 (body-mass index [weight in kilograms divided by the square of the buserelin,24 and leuprolide.25 Whether the increased height in meters], у27.5). All had hirsutism, and 15 had acanthosis nigricans. Two women had each delivered two children, five wom- a activity in women with the poly- en had each delivered one child, and the rest were childless. All had cystic ovary syndrome is an inherited or an acquired normal serum prolactin concentrations and normal results on thy- roid-function tests. Late-onset congenital adrenal hyperplasia was We hypothesized that hyperinsulinemia stimulates ruled out by a morning serum 17a-hydroxyprogesterone concen- ovarian cytochrome P450c17a activity in women tration of less than 200 ng per deciliter (6 nmol per liter). All thewomen had findings on ultrasonography of the ovaries that were with the polycystic ovary syndrome (Fig. 1) and that consistent with the diagnosis of the polycystic ovary syndrome.28 amelioration of insulin resistance in these women None had taken any medications for at least two months, and none would return the activity of the enzyme toward nor- had diabetes mellitus. Twelve women were randomly assigned to mal. To test this hypothesis, we measured the basal receive metformin (Glafornil, North Medicamenta, Caracas, Vene- zuela) and 13 women to receive placebo. The study was approved a-hydroxyprogesterone concentration and by the institutional review board of the Hospital de Clinicas Cara- the serum 17a-hydroxyprogesterone response to the cas, and each woman gave informed consent.
administration of leuprolide in obese women withthe polycystic ovary syndrome before and after the Study Protocol
administration of metformin, which inhibits the pro- The women were evaluated during the follicular phase of the duction of hepatic glucose and enhances the sensi- menstrual cycle, as determined by a serum progesterone concen- tivity of peripheral tissue to insulin, thereby decreas- tration of less than 2 ng per milliliter (6.4 nmol per liter). On day1 the women came to the hospital after a 12-hour overnight fast, and their weight, height, waist-to-hip ratio, and blood pressurewhile supine were measured. Blood samples were drawn at 8:30, 8:45, and 9 a.m., and equal volumes of serum were pooled forthe measurement of insulin, glucose, steroids, and sex hormone- Subjects
binding globulin. At 9 a.m., 75 g of dextrose (Glycolab, Relab We enrolled 25 women who were 18 to 35 years old, 24 of Laboratory, Caracas, Venezuela) was given orally. Blood samples whom completed the study. All the women had the polycystic were collected for determinations of serum glucose and insulin ovary syndrome, as defined by oligomenorrhea (fewer than six concentrations at 60 and 120 minutes.
menstrual periods in the previous year) and hyperandrogenemia On day 2 the women ate breakfast at 9 a.m. and then fasted (elevated serum free testosterone concentrations), and were obese until 2 p.m., when a leuprolide stimulation test was performed.
Figure 1. Possible Mechanisms of Insulin Stimulation of Ovarian Cytochrome P450c17a Activity and Androgen Production.
In theca cells, insulin may directly stimulate (plus signs) ovarian cytochrome P450c17a, resulting in increased 17a-hydroxylase and, to a lesser extent, 17,20-lyase activity. This would lead to increased production of androstenedione,which is then converted to testosterone by the enzyme 17b-reductase. Alternatively or in conjunction with this, insulinmay stimulate ovarian androgen production indirectly by enhancing the amplitude of serum luteinizing hormone (LH)pulses, and luteinizing hormone may then stimulate ovarian cytochrome P450c17a activity.
Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. R E D U C I N G S E R U M I N S U L I N A N D OVA R I A N CY TO C H RO M E P 4 5 0 c 1 7a I N P O LYCYST I C OVA RY SY N D RO M E
After this test the women took 500 mg of metformin or placebo trations of sex steroids or sex hormone-binding glob- orally three times daily. They were instructed not to alter their ulin at base line (Table 1). They also did not differ at usual eating habits, physical activity, or lifestyle during the study.
base line in serum insulin or glucose values measured The women returned for studies four to eight weeks later, after a low serum progesterone value had confirmed that they were in during fasting, insulin or glucose responses after oral the follicular phase of the menstrual cycle. Five women had serum glucose administration, or basal or leuprolide-stimu- progesterone values in the postovulatory range after taking met- lated serum 17a-hydroxyprogesterone values (Table 1 formin for four weeks. One of them became pregnant despite and Fig. 2). The base-line serum luteinizing hormone long-standing infertility; she was dropped from the study and herresults were omitted from the analysis. The remaining four women concentration was higher in the metformin group continued to take metformin and were studied four weeks later than in the placebo group (8.5Ϯ2.2 vs. 3.7Ϯ0.7 when their serum progesterone values were low. In the placebo mIU per milliliter; Pϭ0.04) (Fig. 3).
group, one woman had a serum progesterone value in the post-ovulatory range after four weeks; she was studied again two weeks Anthropometric Variables
The body-mass index did not change significant- Leuprolide Stimulation Test
ly during the study in either group. The waist-to-hip After base-line blood samples had been obtained at 2 p.m. on ratio decreased slightly in the metformin group day 2, leuprolide (10 mg per kilogram of body weight; Lupron, (Pϭ0.02) but did not change substantially in the Abbott Laboratories, Takeda, Japan) was administered subcutane- placebo group. There was no significant change in ously. Blood samples for the measurement of serum luteinizing diastolic or systolic blood pressure in either group.
hormone were collected immediately before and 0.5, 1, 16, 20,and 24 hours after leuprolide was administered. Blood samples Serum Insulin and Glucose Profiles
for the measurement of serum 17a-hydroxyprogesterone werecollected immediately before and 16, 20, and 24 hours after leu- In the metformin group, the mean serum insulin prolide was administered. The women ate an evening meal on day concentration measured during fasting decreased 2 but fasted thereafter until the completion of the test. The early from 17Ϯ3 to 9Ϯ2 mU per milliliter (102Ϯ18 to response of serum luteinizing hormone was determined frompooled equal volumes of serum taken at 0.5 and 1 hour, and the 54Ϯ12 pmol per liter) (Pϭ0.03), and the area under late serum luteinizing hormone response from pooled equal vol- the serum insulin curve decreased from 9303Ϯ1603 umes of serum taken at 16, 20, and 24 hours. The serum concen- to 4982Ϯ911 mU per milliliter per minute (56Ϯ10 tration of 17a-hydroxyprogesterone measured immediately be- to 30Ϯ6 nmol per liter per minute) (Pϭ0.004) (Ta- fore the administration of leuprolide was considered the basalvalue, and the highest serum concentration of 17a-hydroxypro- ble 1). Neither of these values changed significantly gesterone that was measured after the administration of leupro- in the placebo group. The serum glucose concen- tration in fasting women did not change significantlyin either group. The area under the serum glucose curve increased in the placebo group (Pϭ0.03) but The blood samples were centrifuged immediately, and the se- did not change substantially in the metformin group. rum was stored at Ϫ20°C until it was assayed. The serum free tes-tosterone concentration was determined by radioimmunoassay Responses of Serum Luteinizing Hormone to Leuprolide
(Diagnostic Products, Los Angeles). All other hormones and sexhormone-binding globulin (measured as protein) were assayed as The basal serum luteinizing hormone concentra- previously described.15,17,29 To avoid variation among assays, all tion decreased from 8.5Ϯ2.2 to 2.8Ϯ0.5 mIU per samples were analyzed in duplicate in a single assay for each hor- milliliter (Pϭ0.01) in the metformin group but did mone. The intraassay coefficients of variation for the insulin andluteinizing hormone assays were 5.5 and 1.6 percent, respectively, not change significantly in the placebo group (Fig.
and they were less than 10 percent for all the steroid hormone 3). The early serum luteinizing hormone responses to leuprolide were lower after the administration of met-formin than at base line (17.0Ϯ2.5 vs. 40.8Ϯ11.9 Statistical Analysis
mIU per milliliter, Pϭ0.01). The late serum lutein- The results are reported as means ϮSE. Within a group, we izing hormone responses were slightly but not sig- compared the results before treatment with those after treatmentby testing for normality with the Wilk-Shapiro test and using nificantly lower after the administration of metfor- Student's two-tailed paired t-test or the Wilcoxon signed-rank min (Pϭ0.26). In contrast, in the placebo group test. Comparisons between groups were made with Student's two- the basal serum luteinizing hormone concentrations tailed unpaired t-test or the Mann-Whitney rank-sum test.
and the early and late serum luteinizing hormone We analyzed the responses of serum glucose and insulin to the oral administration of glucose and the responses of serum lutein- responses to leuprolide were virtually identical at izing hormone and 17a-hydroxyprogesterone to the administra- base line and after the administration of placebo tion of leuprolide by calculating the areas under the response curves by the trapezoidal rule using absolute values.
Serum 17a-Hydroxyprogesterone Responses
In the metformin group, the mean basal serum Base-Line Characteristics
17a-hydroxyprogesterone concentration decreased The women in the metformin and placebo groups by 51 percent, from 135Ϯ21 to 66Ϯ7 ng per deci- did not differ significantly in age, body-mass index, liter (4.1Ϯ0.6 to 2.0Ϯ0.2 nmol per liter) (Pϭ0.01), waist-to-hip ratio, blood pressure, or serum concen- but it did not change significantly in the placebo Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. T h e New E n g l a n d Jo u r n a l of Me d i c i n e TABLE 1. CHARACTERISTICS OF WOMEN WITH THE POLYCYSTIC OVARY SYNDROME AT BASE LINE
AND AFTER THE ADMINISTRATION OF METFORMIN OR PLACEBO FOR FOUR TO EIGHT WEEKS.* CHARACTERISTIC
METFORMIN GROUP (N؍11)
PLACEBO GROUP (N؍13)
Serum dehydroepiandrosterone sulfate (mg/dl) Serum sex hormone-binding globulin (mg/dl) *The mean (ϮSE) length of administration was 42Ϯ4 days in the metformin group and 32Ϯ2 days in the placebo group. Plus-minus values are means ϮSE. To convert values for insulin to picomoles per liter, multiply by 6.0; to convertvalues for glucose to millimoles per liter, multiply by 0.056; to convert values for progesterone to nanomoles per liter,multiply by 3.2; to convert values for testosterone to picomoles per liter, multiply by 34.7; to convert values for andro-stenedione to picomoles per liter, multiply by 34.9; to convert values for 17b-estradiol to picomoles per liter, multiplyby 36.7; to convert values for dehydroepiandrosterone sulfate to micromoles per liter, multiply by 0.027; and to convertvalues for sex hormone-binding globulin to nanomoles per liter, multiply by 34.7. The normal ranges for ovulatory wom-en are as follows: insulin, 5 to 20 mU per milliliter; progesterone, Ͻ2.0 ng per milliliter during the follicular phase; tes-tosterone, 22 to 70 ng per deciliter; free testosterone, 0.06 to 0.19 ng per deciliter; androstenedione, 66 to 300 ng perdeciliter; 17b-estradiol, 1 to 20 ng per deciliter (early follicular to midfollicular phase); dehydroepiandrosterone sulfate,35 to 430 mg per deciliter; and sex hormone-binding globulin, 0.6 to 4.0 mg per deciliter.
†Pϭ0.02 for the comparison with base line.
‡Pϭ0.03 for the comparison with base line.
§Values are for the area under the curve (AUC) during an oral glucose-tolerance test.
¶Pϭ0.004 for the comparison with base line.
ʈPϭ0.009 for the comparison with base line.
**PϽ0.001 for the comparison with base line.
group (Fig. 2). Similarly, in the metformin group the [139Ϯ21 vs. 210Ϯ21 nmol per liter per hour]) peak serum 17a-hydroxyprogesterone concentra- tion after leuprolide administration decreased from455Ϯ54 to 281Ϯ52 ng per deciliter (13.7Ϯ1.6 to Serum Sex Steroids
8.5Ϯ1.6 nmol per liter) (Pϭ0.01), and the area un- The administration of metformin was associated der the serum 17a-hydroxyprogesterone curve de- with a 44 percent decrease in serum free testosterone creased from 7848Ϯ945 to 4592Ϯ766 ng per deci- concentrations, from 0.34Ϯ0.07 to 0.19Ϯ0.05 ng per liter per hour (237Ϯ29 to 139Ϯ23 nmol per liter deciliter (12Ϯ2 to 7Ϯ2 pmol per liter) (Pϭ0.009), per hour) (Pϭ0.004), whereas these values increased and a threefold increase in serum sex hormone-bind- slightly in the placebo group (Fig. 2). The change ing globulin concentrations, from 0.8Ϯ0.2 to 2.3Ϯ0.6 in the area under the serum 17a-hydroxyprogeste- mg per deciliter (29Ϯ7 to 80Ϯ21 nmol per liter) rone curve in the metformin group differed signifi- (PϽ0.001) (Table 1). These values did not change sig- cantly from that in the placebo group (Ϫ3256Ϯ180 nificantly in the placebo group. The serum concentra- vs. 912Ϯ105 ng per deciliter per hour [Ϫ98Ϯ27 tions of the other measured steroids did not change vs. 28Ϯ10 nmol per liter per hour]) (PϽ0.001), and the area under the serum 17a-hydroxyproges-terone curve was significantly less after metformin DISCUSSION
administration than after placebo administration In these women with the polycystic ovary syn- (4592Ϯ766 vs. 6949Ϯ685 ng per deciliter per hour drome, the administration of metformin reduced the Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. R E D U C I N G S E R U M I N S U L I N A N D OVA R I A N CY TO C H RO M E P 4 5 0 c 1 7a I N P O LYCYST I C OVA RY SY N D RO M E
Figure 3. Mean (ϮSE) Serum Luteinizing Hormone Concentra-
tions in Women with the Polycystic Ovary Syndrome at Base Figure 2. Mean (ϮSE) Serum 17a-Hydroxyprogesterone Con-
Line and after the Administration of Metformin or Placebo for centrations in Women with the Polycystic Ovary Syndrome at Base Line and after the Administration of Metformin or Place- Metformin was administered for a mean (ϮSE) of 42Ϯ4 days, and placebo for 32Ϯ2 days. The women were studied before Metformin was administered for a mean (ϮSE) of 42Ϯ4 days, and after administration of leuprolide (10 mg per kilogram).
and placebo for 32Ϯ2 days. The women were studied before The asterisks indicate Pϭ0.01 for the comparison with the and after the administration of leuprolide (10 mg per kilogram).
base-line value in same group, and the dagger indicates Pϭ0.04 To convert values for 17a-hydroxyprogesterone to nanomoles for the comparison with the base-line value in the metformin per liter, multiply by 0.03. AUC denotes area under the curve.
serum insulin concentration during fasting and the the serum free testosterone concentration. These insulin response to oral glucose administration. Con- findings suggest that increased ovarian cytochrome comitantly, ovarian cytochrome P450c17a activity P450c17a activity in women with the polycystic decreased, as demonstrated by a substantial reduc- ovary syndrome is due to stimulation by insulin (Fig.
tion in the response of serum 17a-hydroxyprogeste- 1) and can be reversed by reducing the secretion of rone to the administration of leuprolide (to increase insulin. We intentionally did not screen the women luteinizing hormone secretion). The reduction in for the presence of insulin resistance or increased P450c17a activity was accompanied by a decline in P450c17a activity so that our results would be ap- Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. T h e New E n g l a n d Jo u r n a l of Me d i c i n e plicable to unselected women with the polycystic a decrease in the concentration of circulating andro- gens. However, raising serum androgen concentra- We cannot exclude the possibility that the decrease tions by parenteral administration in normal women35 in ovarian P450c17a activity resulted from the re- or women with the polycystic ovary syndrome36 duction in serum free testosterone or a direct action does not stimulate the secretion of luteinizing hor- of metformin, but these possibilities seem remote.
mone. Finally, some of the women in our study who Hyperandrogenism is a consequence of increased received metformin ovulated, and ovulation itself may ovarian P450c17a activity and is therefore unlikely influence the dynamics of gonadotropin secretion.37 to be the cause of the stimulated enzyme activity.
However, in our study the results in the women who Hyperandrogenism in women with the polycystic had ovulated and those who had not were similar.
ovary syndrome is ameliorated by diazoxide15 - a The metformin-induced reduction in insulin se- drug structurally unrelated to metformin that sup- cretion was associated with substantial decreases in presses insulin release and worsens glucose tolerance serum free testosterone concentrations and increases - and by diet.19,20 The common factor among these in serum sex hormone-binding globulin concentra- diverse therapies appears to be the reduction in se- tions. In women with the polycystic ovary syndrome, rum insulin concentrations. Because diazoxide is not insulin stimulates ovarian androgen production11-15 known to alter insulin sensitivity yet lowers serum and lowers serum sex hormone-binding globulin testosterone concentrations in women with the poly- concentrations.16,17 Our findings, and those of an cystic ovary syndrome,15 hyperandrogenism in such uncontrolled trial18 of metformin in women with the women appears to be related to hyperinsulinemia it- polycystic ovary syndrome, support these observa- self and not to insulin resistance; moreover, insulin tions. In contrast, Crave et al. found that neither se- stimulates ovarian androgen production in vitro.11-14 rum testosterone nor sex hormone-binding globulin The recent report by Moghetti et al.30 that hyperin- concentrations changed in women with the polycys- sulinemia may stimulate cytochrome P450c17a ac- tic ovary syndrome who were treated with a hypo- tivity in another steroidogenic tissue of women with caloric diet and metformin for four months.38 The the polycystic ovary syndrome - namely, the adrenal reasons for the discrepancies among these studies are glands - further supports our findings.
The early and late serum luteinizing hormone re- In summary, our findings suggest that two fea- sponses to leuprolide after the administration of pla- tures of the polycystic ovary syndrome - hyperin- cebo were almost identical to those at base line.
sulinemic insulin resistance and increased ovarian cy- In contrast, the administration of metformin was tochrome P450c17a activity - are pathogenetically associated with decreased basal and leuprolide-stim- linked, and that hyperinsulinemia stimulates this en- ulated serum luteinizing hormone concentrations.
zyme either directly or indirectly by increasing go- These observations raise the possibility that insulin nadotropin secretion (Fig. 1). The ability of insulin to enhances both the endogenous (basal) and the ex- stimulate ovarian cytochrome P450c17a is probably ogenous (leuprolide-stimulated) release of luteiniz- limited to women with the polycystic ovary syndrome ing hormone mediated by gonadotropin-releasing and may be a heritable abnormality, since many other hormone and that increased ovarian cytochrome obese women who also are hyperinsulinemic have P450c17a activity in women with the polycystic neither hyperandrogenism nor hyperresponsiveness ovary syndrome may be related to an insulin-induced to gonadotropin-releasing hormone.22 The clinical abnormality in the dynamics of gonadotropin secre- implication of these results is that therapeutic meas- tion rather than (wholly or partially) to direct stim- ures directed at lowering insulin secretion in women ulation of ovarian steroidogenesis by insulin (Fig.
with the polycystic ovary syndrome should amelio- 1). Insulin receptors have been identified in human pituitary tissue,31 and insulin augments the releaseof luteinizing hormone by cultured rat pituitary Supported in part by grants (RO1AG11227 and RO1CA64500) from the National Institutes of Health (to Dr. Nestler).
The secretion of luteinizing hormone is often in- We are indebted to Ms. Terre Williams, Ms. Carmen Medina, and creased in women with the polycystic ovary syn- Ms. Gladys Coz for technical assistance. drome,33 and the diurnal changes in the serum con-centrations of luteinizing hormone and insulin in REFERENCES
these women follow a similar time course.34 Prelim- 1. Franks S. Polycystic ovary syndrome. N Engl J Med 1995;333:853-61.
inary studies suggest that insulin enhances the am- plitude of serum luteinizing hormone pulses but not 2. Chang RJ, Nakamura RM, Judd HL, Kaplan SA. Insulin resistance in
nonobese patients with polycystic ovarian disease. J Clin Endocrinol Metab
their frequency in obese women with the polycystic ovary syndrome (unpublished data). An alternative 3. Dunaif A, Segal KR, Futterweit W, Dobrjansky A. Profound peripheral
possibility is that the reduction in luteinizing hor- insulin resistance, independent of obesity, in polycystic ovary syndrome. Diabetes 1989;38:1165-74.
mone secretion in the women we studied was due to 4. Dunaif A, Green G, Futterweit W, Dobrjansky A. Suppression of hyper-
Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved. R E D U C I N G S E R U M I N S U L I N A N D OVA R I A N CY TO C H RO M E P 4 5 0 c 1 7a I N P O LYCYST I C OVA RY SY N D RO M E
androgenism does not improve peripheral or hepatic insulin resistance in tection of functional ovarian hyperandrogenism in women with androgen the polycystic ovary syndrome. J Clin Endocrinol Metab 1990;70:699- excess. N Engl J Med 1992;327:157-62.
22. Rosenfield RL, Barnes RB, Ehrmann DA. Studies of the nature of
5. Carmina E, Koyama T, Chang L, Stanczyk FZ, Lobo RA. Does ethnic-
17-hydroxyprogesterone hyperresponsiveness to gonadotropin-releasing ity influence the prevalence of adrenal hyperandrogenism and insulin resist- hormone agonist challenge in functional ovarian hyperandrogenism. J Clin ance in polycystic ovary syndrome? Am J Obstet Gynecol 1992;167:1807- 23. Luppa P, Muller B, Jacob K, et al. Variations of steroid hormone me-
6. Ciaraldi TP, el-Roeiy A, Madar Z, Reichart D, Olefsky JM, Yen SSC.
tabolites in serum and urine in polycystic ovary syndrome after nafarelin Cellular mechanisms of insulin resistance in polycystic ovarian syndrome. stimulation: evidence for an altered corticoid excretion. J Clin Endocrinol J Clin Endocrinol Metab 1992;75:577-83.
7. Dunaif A, Xia J, Book CB, Schenker E, Tang Z. Excessive insulin recep-
24. White D, Leigh A, Wilson C, Donaldson A, Franks S. Gonadotrophin
tor serine phosphorylation in cultured fibroblasts and in skeletal muscle: a and gonadal steroid response to a single dose of a long-acting agonist of potential mechanism for insulin resistance in the polycystic ovary syn- gonadotrophin-releasing hormone in ovulatory and anovulatory women drome. J Clin Invest 1995;96:801-10.
with polycystic ovary syndrome. Clin Endocrinol (Oxf) 1995;42:475-81.
8. Ehrmann DA, Sturis J, Byrne MM, Karrison T, Rosenfield RL, Polonsky
25. Ibañez L, Potau N, Zampolli M, et al. Source localization of androgen
KS. Insulin secretory defects in polycystic ovary syndrome: relationship to excess in adolescent girls. J Clin Endocrinol Metab 1994;79:1778-84.
insulin sensitivity and family history of non-insulin-dependent diabetes 26. Nagi DK, Yudkin JS. Effects of metformin on insulin resistance, risk
mellitus. J Clin Invest 1995;96:520-7.
factors for cardiovascular disease, and plasminogen activator inhibitor in 9. Apter D, Butzow T, Laughlin GA, Yen SS. Metabolic features of poly-
NIDDM subjects: a study of two ethnic groups. Diabetes Care 1993;16: cystic ovary syndrome are found in adolescent girls with hyperandro- genism. J Clin Endocrinol Metab 1995;80:2966-73.
27. DeFronzo RA, Barzilai N, Simonson DC. Mechanism of metformin
10. Nestler JE. Role of obesity and insulin in the development of anovu-
action in obese and lean noninsulin-dependent diabetic subjects. J Clin En- lation. In: Filicori M, Flamigni C, eds. Ovulation induction: basic science and clinical advances. Amsterdam: Elsevier Science B.V., 1994:103-14.
28. Yeh HC, Futterweit W, Thornton JC. Polycystic ovarian disease: US
11. Barbieri RL, Makris A, Randall RW, Daniels G, Kistner RW, Ryan KJ.
features in 104 patients. Radiology 1987;163:111-6.
Insulin stimulates androgen accumulation in incubations of ovarian stroma 29. Nestler JE, Beer NA, Jakubowicz DJ, Beer RM. Effects of a reduction
obtained from women with hyperandrogenism. J Clin Endocrinol Metab in circulating insulin by metformin on serum dehydroepiandrosterone sul- fate in nondiabetic men. J Clin Endocrinol Metab 1994;78:549-54.
12. Cara JF, Rosenfield RL. Insulin-like growth factor I and insulin poten-
30. Moghetti P, Castello R, Negri C, et al. Insulin infusion amplifies
tiate luteinizing hormone-induced androgen synthesis by rat ovarian the- 17a-hydroxycorticosteroid intermediates response to adrenocorticotropin cal-interstitial cells. Endocrinology 1988;123:733-9.
in hyperandrogenic women: apparent relative impairment of 17,20-lyase ac- 13. Bergh C, Carlsson B, Olsson JH, Selleskog U, Hillensjo T. Regulation
tivity. J Clin Endocrinol Metab 1996;81:881-6.
of androgen production in cultured human thecal cells by insulin-like 31. Unger JW, Livingston JN, Moss AM. Insulin receptors in the central
growth factor I and insulin. Fertil Steril 1993;59:323-31.
nervous system: localization, signalling mechanisms and functional aspects. 14. Nahum R, Thong KJ, Hillier SG. Metabolic regulation of androgen
production by human thecal cells in vitro. Hum Reprod 1995;10:75-81.
32. Adashi EY, Hsueh AJW, Yen SSC. Insulin enhancement of luteinizing
15. Nestler JE, Barlascini CO, Matt DW, et al. Suppression of serum insu-
hormone and follicle-stimulating hormone release by cultured pituitary lin by diazoxide reduces serum testosterone levels in obese women with cells. Endocrinology 1981;108:1441-9.
polycystic ovary syndrome. J Clin Endocrinol Metab 1989;68:1027-32.
33. Berga SL, Guzick DS, Winters SJ. Increased luteinizing hormone and
16. Plymate SR, Matej LA, Jones RE, Friedl KE. Inhibition of sex hor-
a-subunit secretion in women with hyperandrogenic anovulation. J Clin mone-binding globulin production in the human hepatoma (Hep G2) cell line by insulin and prolactin. J Clin Endocrinol Metab 1988;67:460-4.
34. Yen SS, Laughlin GA, Morales AJ. Interface between extra- and intrao-
17. Nestler JE, Powers LP, Matt DW, et al. A direct effect of hyperin-
varian factors in polycystic ovarian syndrome. Ann N Y Acad Sci 1993;687: sulinemia on serum sex hormone-binding globulin levels in obese women with the polycystic ovary syndrome. J Clin Endocrinol Metab 1991;72:83- 35. Spinder T, Spijkstra JJ, van den Tweel JG, et al. The effects of long
term testosterone administration on pulsatile luteinizing hormone secre- 18. Velazquez EM, Mendoza S, Hamer T, Sosa F, Glueck CJ. Metformin
tion and on ovarian histology in eugonadal female to male transsexual sub- therapy in polycystic ovary syndrome reduces hyperinsulinemia, insulin re- jects. J Clin Endocrinol Metab 1989;69:151-7.
sistance, hyperandrogenemia, and systolic blood pressure, while facilitating 36. Dunaif A. Do androgens directly regulate gonadotropin secretion in
normal menses and pregnancy. Metabolism 1994;43:647-54.
the polycystic ovary syndrome? J Clin Endocrinol Metab 1986;63:215-21.
19. Kiddy DS, Hamilton-Fairley D, Seppälä M, et al. Diet-induced chang-
37. Blankstein J, Rabinovici J, Goldenberg M, et al. Changing pituitary re-
es in sex hormone binding globulin and free testosterone in women with activity to follicle-stimulating hormone and luteinizing hormone-releasing normal or polycystic ovaries: correlation with serum insulin and insulin-like hormone after induced ovulatory cycles and after anovulation in patients growth factor-I. Clin Endocrinol (Oxf) 1989;31:757-63.
with polycystic ovarian disease. J Clin Endocrinol Metab 1987;65:1164-7.
20. Kiddy DS, Hamilton-Fairley D, Bush A, et al. Improvement in endo-
38. Crave JC, Fimbel S, Lejeune H, Cugnardey N, Déchaud H, Pugeat
crine and ovarian function during dietary treatment of obese women with M. Effects of diet and metformin administration on sex hormone-binding polycystic ovary syndrome. Clin Endocrinol (Oxf) 1992;36:105-11.
globulin, androgens, and insulin in hirsute and obese women. J Clin En- 21. Ehrmann DA, Rosenfield RL, Barnes RB, Brigell DF, Sheikh Z. De-
Downloaded from nejm.org on December 5, 2011. For personal use only. No other uses without permission. Copyright 1996 Massachusetts Medical Society. All rights reserved.

Source: http://sites.csn.edu/science/biology/polycystic%20ovary%20syndrome.pdf