| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Department of Reproductive Medicine, University of California-San Diego, La Jolla, California 92093
Address all correspondence and requests for reprints to: Dr. R. Jeffrey Chang, Department of Reproductive Medicine, University of California-San Diego, School of Medicine, 9500 Gilman Drive, La Jolla, California 92093-0633. E-mail: rjchang{at}ucsd.edu.
 |
Abstract |
|---|
 Top Abstract IntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
|
Introduction |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Recently, we observed that increased LH secretion in women with PCOS as well as in normal women was unaltered by prolonged insulin infusion (19). Pulsatile LH release and gonadotropin responses to multidose GnRH were similar before and during a 12-h hyperinsulinemic, euglycemic clamp. In contrast to previous in vitro studies, these findings suggested that insulin did not exert an influence on pituitary gonadotropin secretion and, accordingly, may not contribute to elevated serum LH levels in PCOS women. Alternatively, the lack of an insulin effect may have been the result of insulin resistance, which is a common feature of PCOS (20, 21). To investigate this possibility, we extended our earlier studies and examined LH secretion and gonadotropin responses to GnRH in PCOS women after administration of an insulin-lowering drug, pioglitazone, for 12 wk. Subsequently, treatment was continued for an additional 4–8 wk, after which identical studies were repeated during insulin infusion by the hyperinsulinemic, euglycemic clamp method.
|
Subjects and Methods |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Nine women with PCOS were recruited for this study. Each subject exhibited clinical or biochemical evidence of hyperandrogenism, was oligomenorrheic or amenorrheic, and had ultrasound evidence of bilaterally enlarged polycystic ovaries. Late-onset congenital adrenal hyperplasia was excluded by a serum 17-hydroxyprogesterone (17-OHP4) level of less than 3 ng/ml (<9.1 nmol/liter). TSH and prolactin levels were also normal in all subjects. The mean body mass index (BMI) in this study population was 33.6 ± 2.54, and the average age was 31 ± 1.16 yr. None of the subjects had received any hormonal medication for at least 3 months before the start of the study. The study was approved by the institutional review board at University of California-San Diego, and written informed consent was obtained from each subject before the study.
Procedure
This work is an extension of an earlier study designed to systematically examine the role of insulin on gonadotropin secretion in women with PCOS (19). The PCOS subjects in the current investigation represent a subset of those women who underwent additional similar studies after being treated with pioglitazone. For all subjects, the entire protocol is provided in the following description. Each PCOS subject was admitted to the General Clinical Research Center (GCRC) at University of California-San Diego for 2 d of study on four separate occasions. The subjects were tested at random. On the first day of the study, an iv cannula was inserted, and after 30 min, baseline blood samples were drawn. At 0800 h, blood samples were obtained at 10-min intervals for 12 h. On the second day of the study, after at least an overnight rest, beginning at 0800 h, three successive doses of GnRH (2, 10, and 20 µg) were administered iv at 4-h intervals over a continuous 12-h period. The sequence of GnRH dosing was intentional and was not randomized to minimize potential residual increases in serum LH before administration of the next dose of GnRH. GnRH (Factrel) was provided by Wyeth-Ayerst Pharmaceuticals (St. Davids, PA). None of the subjects had experienced recent ovulation, as evidenced by changes in their menstrual patterns or scant bleeding episodes unaccompanied by premenstrual molimina. In addition, at the time of the study each woman had serum progesterone (P4) levels less than 1 ng/ml (<3.0 nmol/liter) at the baseline sample. Frequent blood samples were obtained before and for up to 120 min after each dose of GnRH. After an interval of at least 1 month, each subject was readmitted to the GCRC, and the 2-d study protocol was repeated during a 12-h euglycemic hyperinsulinemic clamp. Blood samples were obtained before initiation of the insulin infusion for baseline hormone measurements. Subsequently, each subject was administered pioglitazone (45 mg/d) for 20 wk. At the end of wk 12 of treatment, each subject was admitted to the GCRC, and the 2 d of study were repeated as described above. Between 16–20 wk of pioglitazone treatment, the 2 d of study were again repeated during a 12-h hyperinsulinemic, euglycemic clamp. One subject withdrew from the study after 16 wk of pioglitazone therapy and was not included in the data analysis that compared the effect of hyperinsulinemic, euglycemic clamps on LH responses to GnRH before and after treatment.
Hyperinsulinemic-euglycemic clamp
Studies were performed in the morning after a 12-h overnight fast. At 2100 h, an 18-gauge iv catheter was inserted into an antecubital vein, and an infusion of normal saline was started. At 0700 h, another iv catheter was inserted in a retrograde fashion in a hand vein, with the hand placed in a hand warmer for sampling of arterialized blood. An iv infusion of insulin (Humulin; Eli Lilly & Co., Indianapolis, IN) diluted in 0.15 mol/liter saline containing 1% (wt/vol) human albumin was then begun at a rate of 80 mU/m2·min, which was started 2 h before the first GnRH dose and continued for 12 h. Potassium and phosphate were given iv to compensate for the intracellular movement of these ions and to maintain normal blood levels. A variable infusion of 20% glucose was delivered to maintain a plasma glucose concentration of 4.72 mol/liter (85 ng/dl). Blood samples were obtained every 5 min for measurement of plasma glucose with a glucose analyzer (YSI 2700 analyzer; YSI, Inc., Yellow Springs, OH). During the last 30 min of insulin infusion, blood samples were obtained at 10-min intervals for determination of plasma glucose concentrations.
The glucose infusion rate in each patient was calculated as the amount of glucose (milligrams) infused per kilogram of body weight during the last 30 min of the clamp study. The mean steady-state insulin level achieved at the end of the clamp is known to suppress hepatic glucose output; therefore, the glucose infusion rate was equivalent to the glucose disposal rate.
Assays
Serum LH and FSH concentrations were measured by RIA, with intra- and interassay coefficients of variation (CVs), respectively, of 5.4% and 8.0% for LH and 3.0% and 4.6% for FSH (Diagnostic Products Corp., Los Angeles, CA). Serum concentrations of estrone (E1), estradiol (E2), androstenedione (A), and testosterone (T) were measured by well established RIAs with intraassay CVs less than 7%. Serum P4, 17-OHP4, and dehydroepiandrosterone sulfate (DHEA-S) were measured by RIA, with intraassay CVs less than 7% (Diagnostic Systems Laboratories, Webster, TX). Serum insulin levels were measured by a double antibody RIA with an assay sensitivity of 2 µU/ml and intra- and interassay CVs of 7% and 9%, respectively. Plasma glucose levels were determined by the glucose oxidase method (YSI, Inc.) with an intraassay CV less than 2% and an intraassay CV of 3%.
Pulse analysis
LH pulse activity was analyzed using the Cluster pulse detection algorithm (22). A cluster configuration of 2 x 2 and t statistics of 2.45 x 2.45 were chosen to minimize false-positive and false-negative errors. Dose-dependent intrasample variance was assessed by employing a second-degree polynomial regression of SD as a function of the hormone concentration. Pulse number per 12 h and mean pulse amplitude (difference in serum concentration between the preceding nadir and the pulse peak) were determined for each subject.
Statistics
Depending on the analysis, LH responses were measured as the difference between the maximal and baseline levels (maximal increment) and the maximal percent change from baseline. A log transformation was applied when appropriate, and a square root transformation was used for the percent change in LH response. To determine the interaction between conditions and dose, two-way repeated measures ANOVA was used. Additionally, to test individual interactions between conditions, paired t tests with the application of a Bonferroni correction were used. Comparisons of mean baseline values for steroids as well as mean LH pulses before and after pioglitazone were performed using a paired t test (SPSS 10.1 software; SPSS, Inc., Chicago, IL).
|
Results |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
Baseline hormone levels in PCOS women demonstrated significant (P < 0.05) elevations in mean (±SE) circulating LH, T, A, E1, E2, and 17-OHP4 and fasting insulin compared with those in normal women as previously reported (data not shown) (19). Serum FSH, DHEA-S, P4, and glucose levels were similar between groups.
Hyperinsulinemic-euglycemic clamp
Mean steady-state plasma insulin levels resulting from the hyperinsulinemic clamps before and during pioglitazone administration were 235 ± 25.5 and 269 ± 33.1 µU/ml, respectively, with equivalent infusion rates and similar serum glucose concentrations. Steady-state serum glucose levels were maintained between 85 and 90 mg/dl during the clamp studies. The mean glucose disposal rate in PCOS subjects after pioglitazone administration (6.5 ± 1.13 mg/kg·min) was significantly (P < 0.05) greater than that observed before treatment (4.9 ± 1.00 mg/kg·min), which indicated increased insulin sensitivity.
Effect of pioglitazone administration on LH secretion and gonadotropin responses to GnRH
Composite 12-h mean LH levels, LH pulse frequency, and LH pulse amplitude as determined over 12 h of rapid sampling in PCOS women before and after administration of pioglitazone for 3 months were not significantly different (Table 1
). During GnRH stimulation in untreated women, baseline LH levels before the initial 2-µg injection of GnRH were correspondingly higher than those observed after treatment. In addition, the absolute serum LH concentrations before each subsequent GnRH dose exhibited progressive increases that were greater than those of treated subjects. However, despite higher pretreatment baseline levels of LH in untreated subjects compared with those after treatment, the difference in these values did not reach statistical significance. Absolute peak serum LH responses to multidose GnRH before pioglitazone treatment were of significantly (P < 0.05) greater magnitude than those observed after treatment, as shown in Fig. 1 and Table 2. These increments can be attributed to differing baseline levels, because the amplitude of maximal LH responsiveness to GnRH has been shown to be dependent on the antecedent baseline (23). As a result, the percent increments in LH responses to multidose GnRH before and after pioglitazone treatment were similar and not significantly different. In addition, treatment effects of pioglitazone on FSH responses to GnRH were not observed (data not shown).
|
|
|
As shown in Table 1
, administration of insulin to PCOS subjects was not associated with significant alterations in composite mean serum LH or pulsatile LH secretion compared with that without insulin infusion, as previously reported (19). After pioglitazone treatment, there were no detectable changes in composite mean serum LH levels or LH pulse frequency and amplitude with or without insulin infusion as a result of treatment. In addition, gonadotropin responses to multidose GnRH were not altered by elevated insulin levels, because maximally stimulated LH and FSH levels during the hyperinsulinemic clamp were similar before and after pioglitazone administration (Fig. 2).
|
Comparison of mean baseline values of circulating steroid hormone concentrations before and after 12 wk of pioglitazone treatment are displayed in Table 3. Pioglitazone therapy was associated with a significant (P < 0.05) reduction in the mean serum T level compared with the pretreatment value. Effects of pioglitazone on serum A, E1, E2, 17-OHP4, and DHEA-S were not detected.
|
. Insulin administration was associated with a statistically significant (P < 0.05) decline in mean serum E1 in women after pioglitazone treatment, whereas circulating E1 levels were unaltered by insulin in the untreated state. A trend toward decreased serum A4 levels as a result of insulin infusion was noted both before and during pioglitazone therapy, although statistical significance was not achieved. The remaining steroid hormones were unchanged by insulin administration regardless of treatment. Of note, comparison of basal T levels before insulin infusion before and after pioglitazone treatment did not reveal a significant decrement in serum T, which was attributed to a spontaneous reduction of the preinfusion value in untreated individuals.
|
During pioglitazone therapy, three PCOS subjects started experiencing cyclic bleeding. Serum P4 determinations were not determined to confirm ovulation. One subject withdrew from the study after 16 wk of treatment. The other two subjects were subsequently studied in the early follicular phase of their menstrual cycles to avoid possible feedback effects of P4 on LH secretion. Improvement in hirsutism was not observed in any subject. Weight gain was observed in each individual; the mean pretreatment BMI (33.57 ± 2.54 kg/m2) increased to 35.30 ± 2.79 kg/m2, which was statistically significant (P < 0.01).
|
Discussion |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
These findings have confirmed and extended our previous observations to indicate that increased LH secretion in PCOS women is not influenced by insulin administration. Our results demonstrated that even after improvement of insulin sensitivity with pioglitazone treatment, there was no difference in baseline LH values, LH pulsatility, or maximally stimulated percent LH increment after GnRH with or without insulin infusion in women with PCOS. In untreated women, absolute increases in stimulated LH were significantly higher than those observed after pioglitazone treatment. These increases were attributed to subtle elevations of preinjection levels of LH, which did not achieve statistical significance compared with baseline values observed during treatment. It has been previously well documented that the magnitude of the LH response to GnRH is directly correlated to the prevailing baseline LH concentration (1, 23, 24). As a result, the treatment effect on LH responsiveness to GnRH was assessed according to the percent change from baseline, which was not different between untreated and treated individuals.
Other studies in women with PCOS also have shown that hyperinsulinemia induced by insulin infusion is not associated with an increase in basal or stimulated LH release, although subsequent testing after reduction of insulin resistance was not performed (8, 9). Our results are consistent with in vivo experiments conducted in normal rats, in which insulin administration on a daily basis failed to increase LH responsiveness to GnRH despite a 3-fold increase in plasma insulin (25). Interestingly, in streptozotocin-treated (insulin deficient) diabetic rats, LH responses to GnRH have been shown to be 2- to 3-fold greater than those in control normal animals (26). In the same study, insulin-deficient diabetic rats treated with insulin exhibited a 2-fold decrease in the mean LH response to GnRH that was comparable to responses observed in the normal control animals, which suggested inhibitory regulation of gonadotropin production by insulin. In contrast, previous in vitro studies have clearly shown a facilitatory effect of insulin on pituitary LH production and release. In rat anterior pituitary cells cultured in serum-free medium, it has been demonstrated that LH responsiveness to GnRH stimulation was enhanced by insulin in a dose-dependent manner (5, 6, 7). While interpreting these in vitro data, consideration must be given to the possibility that anterior pituitary cell cultures incubated for 36–48 h may have been influenced by trophic effects of insulin, thereby leading to the observation of increased hormone production.
Clinically, indirect evidence for an insulin effect on the pituitary gonadotrope has been suggested from the decrement in serum LH after administration of insulin-lowering drugs or dietary caloric restriction (10, 11, 12, 13, 14, 27, 28, 29, 30). However, in some instances, decreased mean levels of serum LH have not been found (31). Interpretation of these findings is potentially confounded by several factors. First, most of the patients studied were obese, and it has been shown recently that obesity is inversely correlated to LH secretion in PCOS (15). Second, hyperinsulinemia is positively correlated with BMI in women with this syndrome (32). Third, reduced LH secretion may have been the result of steroid negative feedback after restoration of ovarian follicular activity and ovulation in these studies.
Alternatively, the disparity between the in vitro and in vivo effects of insulin on pituitary gonadotropin release may reflect the intact reproductive-metabolic environment to which the pituitary gonadotrope is exposed in PCOS women. Indeed, in the presence of serum-supplemented medium, incubated rat pituitary cells failed to demonstrate increased LH release after GnRH stimulation (5). Thus, the possibility exists that an endogenous factor may be responsible for blocking an effect of hyperinsulinemia on LH secretion. Notably, it has been previously reported that composite 24-h serum LH concentrations and LH responses to GnRH are inversely correlated to hyperinsulinemia as well as obesity in PCOS and normal women (15). Because the vast majority of PCOS subjects in the current study exhibited substantial obesity and gained weight during treatment, an effect of insulin infusion on LH secretion may have been obscured by excessive weight or weight gain. Resolution of the impact of hyperinsulinemia on pituitary LH may require similar studies to be performed in nonobese PCOS women. Current investigation is underway to explore this issue.
Another possible explanation for our results is that the dose of pioglitazone was inadequate to abrogate the effect of insulin resistance in obese PCOS women. In vitro studies have clearly demonstrated that increases in pituitary LH responsiveness to GnRH were insulin dose dependent (5). As a result, there may have been residual insulin resistance that was sufficient to nullify an effect of insulin administration after pioglitazone treatment. Despite this consideration, the glucose disposal rate was significantly increased during pioglitazone treatment compared with that observed before treatment.
An effect of pioglitazone on serum T was apparent, because the mean concentration after 12 wk of treatment was significantly decreased compared with the pretreatment value. However, a reduction in T was not observed when comparing basal levels before insulin infusion before and after pioglitazone. Other researchers have reported lowered circulating levels of T after administration of other thiazolidinediones, such as troglitazone and rosiglitazone (10, 31). Similar results have been obtained after weight reduction and improved insulin sensitivity (11, 16). The effect of pioglitazone on serum T levels may be attributed to reduced hyperinsulinemia and improved insulin sensitivity, as indicated by increased glucose disposal rates after treatment. Alternatively, it is also possible that clearance may have been affected, which could have accounted for the reduced T levels. Although our study was not designed to evaluate clearance, this is unlikely, because any effect of pioglitazone would be expected to increase SHBG (10, 33) and lower T clearance. Therefore, a decrease in secretion rate would seem to be a more plausible explanation for decreased T levels.
Because thiazolidinediones have been shown to have a direct effect on ovarian steroidogenesis, reduced serum T may have been mediated by inhibition of the steroidogenic enzymes P450c17 and 3ß-hydroxysteroid dehydrogenase (34, 35, 36). However, pharmacological concentrations well beyond therapeutic doses would be required to achieve this effect. In light of the failure of pioglitazone to alter gonadotropin secretion in our study and the aforementioned data, we conclude that lowered T levels resulted primarily from decreased hyperinsulinemia.
With pioglitazone therapy, a decrease in E1 levels was seen during insulin infusion, whereas changes in E2 were not detected during insulin infusion before treatment. The decline in E1 may have reflected reduced conversion of androgen to estrogen, because the related thiazolidinedione troglitazone has been shown to inhibit aromatase activity in luteinized granulosa cells obtained from women undergoing in vitro fertilization (37). In contrast, an effect of troglitazone on aromatase activity has not been demonstrated in unstimulated porcine granulosa cells (35). Whether pioglitazone has an effect on aromatase activity in our study is unclear, because serum E2 levels were unaltered by insulin both before and after pioglitazone treatment. Moreover, circulating levels of A, an aromatase substrate, trended downward during insulin infusion both pre- and posttreatment with pioglitazone.
In summary, our results have failed to demonstrate an effect of insulin infusion on LH secretion or gonadotropin responses to GnRH in women with PCOS before and after improved insulin sensitivity during pioglitazone treatment. These findings strongly suggest that in this disorder increased LH secretion is not due to insulin resistance and compensatory hyperinsulinemia.
|
Acknowledgments |
|---|
|
Footnotes |
|---|
First Published Online January 11, 2005
Abbreviations: A, Androstenedione; BMI, body mass index; CV, coefficient of variation; DHEA-S, dehydroepiandrosterone sulfate; E1, estrone; E2, estradiol; 17-OHP4, 17-hydroxyprogesterone; P4, progesterone; PCOS, polycystic ovary syndrome; T, testosterone.
Received June 2, 2004.
Accepted January 5, 2005.
|
References |
|---|
|
TopAbstractIntroductionSubjects and MethodsResultsDiscussionReferences |
|---|
activity and serum free testosterone after reduction of insulin secretion in polycystic ovary syndrome. N Engl J Med 335:617–623This article has been cited by other articles:
 |
|
 |
|
  V. R. Aroda, T. P. Ciaraldi, P. Burke, S. Mudaliar, P. Clopton, S. Phillips, R. J. Chang, and R. R. Henry Metabolic and Hormonal Changes Induced by Pioglitazone in Polycystic Ovary Syndrome: A Randomized, Placebo-Controlled Clinical Trial J. Clin. Endocrinol. Metab., February 1, 2009; 94(2): 469 - 476. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
M. A. Lawson, S. Jain, S. Sun, K. Patel, P. J. Malcolm, and R. J. Chang Evidence for Insulin Suppression of Baseline Luteinizing Hormone in Women with Polycystic Ovarian Syndrome and Normal Women J. Clin. Endocrinol. Metab., June 1, 2008; 93(6): 2089 - 2096. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. C. Kent, C. L. Gnatuk, A. R. Kunselman, L. M. Demers, P. A. Lee, and R. S. Legro Hyperandrogenism and Hyperinsulinism in Children of Women with Polycystic Ovary Syndrome: A Controlled Study J. Clin. Endocrinol. Metab., May 1, 2008; 93(5): 1662 - 1669. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S.K. Blank, C.R. McCartney, and J.C. Marshall The origins and sequelae of abnormal neuroendocrine function in polycystic ovary syndrome Hum. Reprod. Update, July 1, 2006; 12(4): 351 - 361. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
J. C. Marshall Obesity in Adolescent Girls: Is Excess Androgen the Real Bad Actor? J. Clin. Endocrinol. Metab., February 1, 2006; 91(2): 393 - 395. [Full Text] [PDF]
|
|
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Endocrinology | Endocrine Reviews | J. Clin. End. & Metab. |
| Molecular Endocrinology | Recent Prog. Horm. Res. | All Endocrine Journals |
