| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Clinical Research Center Studies |
Departments of Medicine (D.A.E., M.K.C., J.I., R.L.R., K.S.P.) and Pediatrics (R.L.R.), University of Chicago, Chicago, Illinois 60637; and the Department of Medicine (D.J.S., B.E.S.), University of Vermont, Burlington, Vermont 00000
Address all correspondence and requests for reprints to: David A. Ehrmann, M.D., Department of Medicine, Section of Endocrinology, University of Chicago Pritzker School of Medicine, 5841 South Maryland Avenue, MC 1027, Chicago, Illinois 60637. E-mail: dehrmann{at}medicine.bsd.uchicago.edu
Abstract
Women with polycystic ovary syndrome (PCOS) are characterized by defects in insulin action, insulin secretion, ovarian steroidogenesis, and fibrinolysis. We administered the insulin-sensitizing agent troglitazone to 13 obese women with PCOS and impaired glucose tolerance to determine whether attenuation of hyperinsulinemia ameliorates these defects. All subjects had oligomenorrhea, hirsutism, polycystic ovaries, and hyperandrogenemia. Before and after treatment with troglitazone (400 mg daily for 12 weeks), all had 1) a GnRH agonist (leuprolide) test, 2) a 75-g oral glucose tolerance test, 3) a frequently sampled iv glucose tolerance test to determine the insulin sensitivity index and the acute insulin response to glucose, 4) an oscillatory glucose infusion to assess the ability of the ß-cell to entrain to glucose as quantitated by the normalized spectral power for the insulin secretion rate, and 5) measures of fibrinolytic capacity [plasminogen activator inhibitor type 1 (PAI-1) and tissue plasminogen activator].
There was no change in body mass index (39.9 ± 1.4 vs. 40.2 ± 1.4 kg/m2) or body fat distribution after treatment. Both the fasting (91 ± 3 vs. 103 ± 3 mg/dL; P < 0.001) and 2 h (146 ± 8 vs. 171 ± 6 mg/dL; P < 0.02) plasma glucose concentrations during the oral glucose tolerance test declined significantly. There was a concordant reduction in glycosylated hemoglobin to 5.7 ± 0.1 from a pretreatment level of 6.1 ± 0.1% (P < 0.03). Insulin sensitivity increased from 0.58 ± 0.14 to 0.95 ± 0.26 10-5 min-1/pmol·L (P < 0.01) after treatment as did the disposition index (745 ± 135 vs. 381 ± 96; P < 0.05). The ability of the ß-cell to appropriately detect and respond to an oscillatory glucose infusion improved significantly after troglitazone treatment; the normalized spectral power for the insulin secretion rate increased to 5.9 ± 1.1 from 4.3 ± 0.8 (P < 0.05). Basal levels of total testosterone (109.3 ± 15.2 vs. 79.4 ± 9.8 ng/dL; P < 0.05) and free testosterone (33.3 ± 4.0 vs. 21.2 ± 2.6 pg/mL; P < 0.01) declined significantly after troglitazone treatment. Leuprolide-stimulated levels of 17-hydroxyprogesterone, androstenedione, and total testosterone were significantly lower posttreatment compared to pretreatment. The reduction in androgen levels occurred independently of any changes in gonadotropin levels. A decreased functional activity of PAI-1 in blood (from 12.7 ± 2.8 to 6.3 ± 1.4 AU/mL P < 0.05) was associated with a decreased concentration of PAI-1 protein (from 64.9 ± 9.1 to 44.8 ± 6.1 ng/mL; P < 0.05). No change in the functional activity of tissue plasminogen activator (from 5.3 ± 0.4 to 5.1 ± 0.5 IU/mL) was observed despite a decrease in its concentration (from 9.6 ± 0.9 to 8.2 ± 0.7 ng/mL; P < 0.05). The marked reduction in PAI-1 could be expected to improve the fibrinolytic response to thrombosis in these subjects.
We conclude that administration of troglitazone to women with PCOS and impaired glucose tolerance ameliorates the metabolic and hormonal derangements characteristic of the syndrome. Troglitazone holds potential as a useful primary or adjunctive treatment for women with PCOS.
INSULIN resistance and its concomitant, hyperinsulinemia, are characteristic of women with polycystic ovary syndrome (PCOS) (1, 2, 3, 4, 5, 6, 7). The insulin resistance is typically profound (2, 5, 7) and acts conjointly with alterations in ß-cell function (7, 8) to impart a high risk for the development of early-onset noninsulin-dependent diabetes mellitus (NIDDM) (5, 6, 7, 9, 10). Hyperinsulinemia also contributes to the ovarian androgen overproduction in PCOS (11, 12, 13) and may accelerate the development of coronary and other vascular disease (14, 15, 16, 17, 18) in these women. Although the mechanism for accelerated atherosclerosis has not been elucidated in this population, hyperinsulinemia may play an important role, as it has been shown to enhance the production of plasminogen activator inhibitor type 1 (PAI-1) (19, 20, 21, 22). It follows from these observations that attenuation of hyperinsulinemia may have a salutary effect on the metabolic and hormonal sequelae of PCOS.
Troglitazone is a novel insulin-sensitizing agent that improves oral glucose tolerance and insulin resistance in individuals with impaired glucose tolerance (IGT) (23, 24). More recently, troglitazone was shown to improve defects in ß-cell function in subjects with IGT (23) and to lower androgen levels in women with PCOS (25). This reduction in hyperandrogenemia may result from a reduction in insulin-mediated ovarian androgen production, although this remains speculative. The impact of troglitazone on fibrinolytic capacity has not previously been examined. The present study was thus undertaken to determine whether attenuation of insulin resistance with subsequent reduction of hyperinsulinemia ameliorates abnormalities in insulin secretion, ovarian steroidogenesis, and fibrinolysis in women with PCOS who have IGT.
Materials and Methods
Selection and definition of study subjects
Fifteen women with hyperandrogenemia [plasma free testosterone,
>10 pg/mL (34.6 pmol/L)] (26) and IGT (2 h plasma glucose
concentration, 140–199 mg/dL during OGTT) were recruited. All subjects
were anovulatory (cycle length, >35 days) and had a history of
infertility, hirsutism, or acne. The ovarian source of androgen excess
was established in all subjects by a dexamethasone androgen suppression
test, as previously described (26). To avoid sample bias, the results
of GnRH agonist testing were not used to identify or select subjects.
Dexamethasone suppression of androgen levels was considered abnormal if
the plasma free testosterone concentration remained elevated [
8
pg/mL (27.7 pmol/L)] after the administration of dexamethasone (2 mg
daily for 4 days) (26). Although ultrasound evidence of polycystic
ovaries was not used as a diagnostic or inclusion criterion, all
subjects had polycystic ovaries by transvaginal ultrasonography using a
5-MHz/900 phased array sector vaginal probe (GE RT 3000,
General Electric Corp., Milwaukee, WI) (27).
Exclusion criteria included the use of medications known to alter insulin secretion or action and endocrinopathies, including nonclassic 21-hydroxylase deficiency congenital adrenal hyperplasia, Cushing’s syndrome, hyperprolactinemia, or thyroid dysfunction. Women with a prior history of gestational diabetes or NIDDM (28) were excluded, as were those who had normal glucose tolerance or NIDDM at baseline oral glucose tolerance testing. A negative screening pregnancy test was required.
The racial/ethnic composition of the study subjects was representative of that of women attending our clinics: eight (62%) were Caucasian, four (31%) were African-American, and one (7%) was Asian. All protocols were approved by the institutional review board of the University of Chicago. Written informed consent was obtained from each subject.
Because of the possible effect of variations in insulin secretion (29, 30) and ovarian function at different phases of the menstrual cycle, women were studied in the follicular phase (days 1–8) of the cycle or after 2 or more months of amenorrhea. After baseline studies (see below) were completed, 400 mg troglitazone were administered orally once daily for 12 weeks, at which time the pretreatment studies were repeated. Compliance with the medication was confirmed by pill count.
Because of the impact of body fat distribution on androgen levels, glucose tolerance, and possibly insulin secretion (31), waist to hip ratios (WHR) were measured, and body composition and fat distribution were assessed by dual energy x-ray absorptiometry scan (32) before and after treatment.
Experimental protocols
Studies were performed after an overnight fast. Intravenous catheters were placed in antecubital veins. Where appropriate, one catheter was used for administration of iv glucose or other secretagogues, whereas the catheter in the contralateral forearm was used for blood sampling. The blood-sampling arm was heated to obtain arterialized venous samples.
GnRH agonist test. Blood samples were obtained before (i.e. at baseline, 20-min intervals, four times) and 0.5, 1, 16, 20, and 24 h after the administration of 10 µg/kg leuprolide acetate (TAP Pharmaceuticals, Deerfield, IL). Serum LH and FSH were measured at baseline, 0.5, and 1 h, and steroid hormones were measured at 0, 16, 20, and 24 h. The plasma steroids assayed included total testosterone, estradiol, 17-hydroxyprogesterone, dehydroepiandrosterone (DHEA), androstenedione, and progesterone.
Oral glucose tolerance test (OGTT). Blood samples were obtained at baseline and at 30-min intervals for 3 h for measurement of glucose and insulin after the ingestion of a 75-g glucose load. Glucose tolerance was evaluated using the criteria of the WHO (28).
Frequently sampled iv glucose tolerance test (IVGTT). A frequently sampled IVGTT was performed as previously described by Bergman et al. (33). Blood was collected during the IVGTT at -20, -15, -10, -5, and 0 min, at which time 0.3 g/kg glucose was administered as an iv bolus. Blood was then obtained at 2, 3, 4, 5, 6, 8, 10, 12, 14, 16, and 19 min. At 20 min, iv tolbutamide (125 mg/m2; Orinase, Upjohn, Kalamazoo, MI) was administered. Thereafter, blood was sampled at 22, 24, 25, and 27 min, with subsequent sampling every 10 min from 30–100 min and then every 20 min from 100–240 min.
Insulin secretory responses to oscillatory glucose administration
To determine whether the ability of the ß-cell to be entrained (i.e. respond to repetitive increases and decreases in glucose with parallel changes in insulin secretion) is affected by troglitazone, a 20% glucose solution was administered in an oscillatory pattern with a period of 144 min, as previously described (34, 35). This period most effectively brings out defective entrainment of pulsatile insulin secretion by glucose (34). Samples were drawn every 15 min for the remaining 12 h for the measurement of glucose, insulin, and C peptide.
Measurements of fibrinolytic parameters
Fibrinolytic capacity was characterized by measurement of both the concentration and the functional activity of tissue plasminogen activator (t-PA) and plasminogen activator inhibitor type 1 (PAI-1) in the blood of the subjects at baseline, 6 weeks, and 12 weeks. The mean of the values at 6 and 12 weeks for both t-PA and PAI-1 represents the value during treatment.
Lipid measurements
Serum was obtained for measurement of total cholesterol, high density lipoprotein (HDL) cholesterol, low density lipoprotein (LDL) cholesterol, and triglycerides before and after treatment with troglitazone.
Assay methods
Plasma glucose was measured immediately using a glucose analyzer (YSI model 2300 STAT, Yellow Springs Instruments Co., Yellow Springs, OH). The coefficient of variation of this method is less than 2%. Glycosylated hemoglobin was measured by boronate affinity chromatography with an intraassay coefficient of variation of 4% (Bio-Rad Laboratories, Hercules, CA). Serum insulin was assayed by a double antibody technique (36), with a lower limit of sensitivity of 20 pmol/L and an average intraassay coefficient of variation of 6%. The cross-reactivity of proinsulin in the RIA for insulin is approximately 40%. Plasma C peptide was measured as previously described (37). The lower limit of sensitivity of the assay is 0.02 pmol/mL, and the intraassay coefficient of variation averaged 6%.
Serum LH and FSH concentrations were measured by fluoroimmunometric assay (Delfia, Wallac, Gaithersburg, MD). The minimal detectable concentration of both LH and FSH was 0.05 U/L. The interassay coefficient of variation was 8% for LH and 5% for FSH. Plasma concentrations of testosterone, estradiol, and DHEA sulfate were measured using kits obtained from Diagnostic Products Corp. (Los Angeles, CA), Pantex (Santa Monica, CA), and Diagnostic Systems (Webster, TX), respectively. The free fraction of plasma testosterone and the concentration of sex hormone-binding globulin (SHBG) were measured by a competitive protein binding assay as previously described (38). The intra- and interassay coefficients of variation were 3.8% and 8.7%, respectively. Plasma levels of 17-progesterone, androstenedione, and DHEA were determined by RIA after these steroids had been chromatographically purified as previously reported (39). The precision of these assays averaged 7% (intraassay coefficient of variation) and 12% (interassay coefficient).
The concentrations of t-PA and PAI-1 were determined by enzyme-linked immunosorbent assay (Tintelize, Biopool, Ventura, CA) with the use of antibodies that detected t-PA and PAI-1 in both the free and the complexed state. For the determination of PAI-1, the intraassay coefficient of variation was 2.5%, and the interassay coefficient of variation was 3.6%. For determination of t-PA, the intraassay coefficient of variation was 4.3%, and the interassay coefficient of variation was 9.5%. The functional fibrinolytic capacity within an aliquot of plasma was determined with a chromogenic substrate of plasmin as previously described (40, 41). The functional activity of PAI-1 was determined based on the extent to which an aliquot of plasma inhibited t-PA-induced activation of plasminogen. The plasmin generated was exposed to a chromogenic substrate [S2251, Chromogenix (Pharmacia Hepar), Franklin, OH]. A standard curve was generated with a specific amount of t-PA. One arbitrary unit (AU) of PAI-1 is that amount of PAI-1 able to inhibit 1 IU (based on the WHO standard) of t-PA. In a similar fashion, the concentration of t-PA in plasma was determined with the use of a functional assay based on the ability of an aliquot of sample to induce activation of plasminogen. A colorimetric assay was performed with the chromogenic substrate S2251, and a standard curve was generated with a specific amount of t-PA.
Measurements of total cholesterol, HDL cholesterol, and triglycerides were performed using a Boehringer Mannheim Hitachi 747–100 automated analyzer (Indianapolis, IN).
Data analysis and statistical methods
Summary measures derived from IVGTT.
Summary measures
derived from the IVGTT included 1) the first phase insulin secretion
(AIRg) in response to glucose, calculated as the mean increment above
basal of insulin values measured at 2, 3, 4, 5, 6, 8, and 10 min; 2)
the insulin sensitivity index (Si), calculated using the MINMOD program
provided by Dr. R. N. Bergman (33, 42) (the Si represents the increase
in net fractional glucose clearance rate per unit change in plasma
insulin concentration after the iv glucose load); and 3) the
relationship between the AIRg relative to the degree of insulin
resistance (Si), calculated from the equation Z
= {ln[(Si x
10-5) x AIRg)] 3.802}/0.5613, which defines the
relationship for these parameters in normal subjects (43). The
percentile ranking based on the value for Z is obtained from the
table of the standard normal distribution. The product of Si and AIRg
(disposition index) was also derived and provides a measure of ß-cell
secretory function adjusted for insulin sensitivity, previously
described as a constant in any individual (43).
Determination of insulin secretion rates (ISRs). Standard kinetic parameters for C peptide clearance were used to derive ISRs. These standard parameters were calculated from 200 decay curves of biosynthetic C peptide obtained in normal, obese, and NIDDM subjects (44). Age, sex, and body surface area were accounted for. These parameters were derived by application of a two-compartment model of C peptide distribution proposed by Eaton et al. (45). These parameters were used to derive, in each 15-min interval between blood sampling, the ISR from the plasma C peptide concentrations by deconvolution as previously described (46). The C peptide profiles were smoothed with a two-point moving average before calculation of the ISRs.
Spectral analysis. Each individual ISR and glucose profile from the oscillatory glucose infusion protocol was submitted to spectral analysis to determine whether the insulin oscillations were entrainable, as previously reported (47, 48). Each spectrum was normalized assuming the total variance of each series to be 100% and is expressed as the normalized spectral power. Slow trends were removed by the first difference filter before each spectrum was calculated with the window-closing procedure, using a Tukey window with a width of 24 data points, as described by Jenkins and Watts (48).
Statistical analysis. The significance of differences between pretreatment and posttreatment measures was determined with the use of paired Student’s t tests. Logarithmic transformation of data was performed for measures that were not normally distributed. For all analyses, a two-tailed P < 0.05 was considered to indicate statistical significance. Unless otherwise noted, all results are expressed as mean \ SEM. Data analysis was performed using StatView 4.5 for the Macintosh (Abacus Concepts, Berkeley, CA).
Results
Clinical characteristics of study subjects
Thirteen subjects completed the protocol. Two subjects were
removed from the study before its completion because of noncompliance
with testing procedures. All study subjects were obese, and most were
markedly so [mean body mass index (BMI), 39.9 \ 1.4
kg/m2; range, 32.0–47.3 kg/m2; Table 1
]. Because alterations in body fat distribution can
affect insulin sensitivity independently of body weight (49), precise
assessment of regional fat distribution was made by dual energy x-ray
absortiometry scan, in addition to the standard waist/hip ratio
measure. With the exception of one subject who had a WHR below 0.80,
all subjects had the characteristic central obesity of PCOS. Body
weight, BMI, and body fat distribution remained stable over the course
of the study (Table 1).
|
Glucose tolerance.
All subjects had glycosylated hemoglobin
levels below 7.2%, the upper limit of normal (mean, 6.1 \ 0.1%;
range, 5.4–7.0%). After treatment with troglitazone, there was a
significant reduction in the level of glycosylated hemoglobin (5.7
\ 0.1 vs. 6.1 \ 0.1%; P < 0.03) as
well as in both the plasma glucose and serum insulin responses to the
oral glucose challenge (Fig. 1). Fasting plasma glucose
concentrations declined by an average of 12% (91 \ 3
vs. 103 \ 3 mg/dL; P < 0.001),
whereas the 2 h glucose levels were 15% lower, on the average,
after troglitazone treatment (146 \ 8 vs. 171 \ 6 mg/dL;
P < 0.02). The area under the glucose
response curve declined similarly by 15% (P < 0.002).
There was a remarkably similar 56% decline in both fasting serum
insulin levels (176 \ 22 vs. 397 \ 63 pmol/L;
P < 0.001) and the area under the insulin response
curve (P < 0.001). Eight of 13 (62%) subjects
exhibited normal glucose tolerance after treatment, whereas 5 (38%)
had persistent, albeit lesser, IGT.
|
). By comparison, women
with normal glucose tolerance and similar BMI values (32–47
kg/m2) have Si values between approximately 2.0–5.5
x 10-5 min-1/pmol·L (43). The improvement
in Si in our subjects occurred in the absence of a reduction in weight.
The first phase AIRg increased modestly after treatment (876 \ 212
vs. 1161 \ 244 pmol/L), but did not reach
statistical significance. However, when the insulin secretory response
was analyzed after accounting for the degree of insulin resistance
(i.e. disposition index), an improvement in ß-cell
function could be appreciated. That is, the product of Si and AIRg was
greater after troglitazone treatment compared to the baseline (745
\ 135 vs. 381 \ 96; P < 0.05).
|
).
|
Leuprolide-stimulated levels of all steroids were significantly lower
posttreatment than pretreatment (Table 1
). There was a significant
correlation between the change in the leuprolide-stimulated androgen
(androstenedione plus testosterone) responses and the change in the
area under the insulin response curve during OGTT (r = 0.71;
P < 0.02) and in basal ISRs before and after treatment
(r = 0.65; P < 0.02). The reduction in androgen
levels occurred independently of any changes in gonadotropin levels.
The LH and FSH levels as well as the ratio of LH to FSH were within the
normal range at baseline and did not change after treatment. Likewise,
the early LH response to leuprolide was not altered after troglitazone
treatment.
Fibrinolytic capacity
Treatment with troglitazone resulted in a 31% reduction in the
concentration of PAI-1 protein in blood (44.8 \ 6.1
vs. 64.9 \ 9.1 ng/mL; P < 0.05) and a
50% reduction in the functional activity of PAI-1 (6.3 \ 1.4
vs. 12.7 \ 2.8 AU/mL; P < 0.05; Table 1 and Fig. 4). There was a significant correlation
between the difference in insulin levels during the OGTT before and
after treatment and the difference in PAI-1 activity levels pre- and
posttreatment (r = 0.61; P < 0.03).
|
). Despite this decrease, the fibrinolytic activity attributable
to t-PA in blood did not change after treatment with troglitazone. Lipid measurements
There was no significant change in levels of total cholesterol, HDL cholesterol, or LDL cholesterol after troglitazone treatment. Although triglyceride levels declined from 164 \ 33 to 115 \ 15 mg/dL, this did not achieve statistical significance.
Discussion
We have found that administration of troglitazone to women with
PCOS and IGT results in a marked attenuation of hyperinsulinemia that
is associated with improvements in insulin secretion, ovarian androgen
biosynthesis, and enhanced fibrinolytic system capacity. Although the
precise mechanism of action of troglitazone remains incompletely
defined, it appears to enhance insulin action without directly
stimulating insulin secretion (50) and acts as a selective ligand for
the peroxisome proliferator-activated receptor-
(51). Peroxisome
proliferator-activated receptor- is a nuclear hormone receptor
expressed predominantly in adipose tissue, where it plays a central
role in the control of adipocyte gene expression and differentiation
(52, 53, 54, 55).
The results of previous studies have shown that troglitazone attenuates hyperinsulinemia in a number of diverse populations of obese individuals (24), including women with PCOS (25). We recently found that troglitazone acts not only to enhance insulin action, but also to ameliorate defects in insulin secretion in subjects with IGT. These combined effects resulted in improved oral glucose tolerance (23). In the present study, we found a similar improvement in glucose tolerance, as reflected in both a significantly lower concentration of glycosylated hemoglobin after treatment (5.7 \ 0.1 vs. 6.1 \ 0.1%; P < 0.05) and markedly reduced glucose concentrations throughout an oral glucose tolerance test. As before (23), this improvement resulted from the effects of troglitazone on both insulin action and insulin secretion. Improvement in insulin action was evidenced by a 30% decline in the basal ISR (P < 0.02) and a 64% increase (P < 0.05) in the insulin sensitivity index derived from a frequently sampled IVGTT. It is important to note, however, that although insulin sensitivity improved in response to treatment, it remained profoundly depressed in these subjects and was disproportionate to the degree of obesity, as described previously (2). Improved ß-cell function was evidenced by an enhanced ability of the ß-cell to appropriately secrete insulin in response to an oscillatory infusion of glucose (34). The latter is expressed quantitatively by the improvement in normalized spectral power for insulin secretion (5.9 \ 1.1 posttreatment vs. 4.3 \ 0.8 pretreatment; P < 0.05) (34). Enhanced ß-cell function was also reflected in results from the IVGTT, which provides an index of ß-cell function (AIRg) that can be adjusted for the degree of ambient insulin resistance (Si). Because the product of these measures is presumed to be constant (33, 42, 43), a compensatory decrease in AIRg would be expected to accompany the increase in Si observed after troglitazone treatment. In fact, there was a significant increase in the product of these parameters. We interpret this finding to represent an improvement in ß-cell function that can be quantified as both the absolute product of these measures as well as the z-score/percentile rank.
We previously postulated (23) that the improvement in ß-cell function resulting from troglitazone treatment may result from one of a number of mechanisms. These include reduction in hyperglycemia (i.e. elimination of so-called ß-cell glucotoxicity) (56, 57, 58), improvement in insulin resistance, or improvement in lipid parameters. It is also possible that troglitazone may improve defects in insulin secretion through direct actions on the pancreatic ß-cell (59). Our results do not provide further insights into the mechanisms responsible for the improved ß-cell function. In the absence of a significant decline in triglyceride levels, however, we can conclude that hypertriglyceridemia is an unlikely explanation. In addition, glucotoxicity seems an implausible explanation, given that fasting plasma glucose and glycosylated hemoglobin concentrations were within the normal range. Although mild hyperglycemia was seen after glucose challenge, such modest alterations in glucose metabolism have not been linked to ß-cell glucotoxicity.
It is important to note that our current and previous (23) finding of improved ß-cell function in IGT contrasts with the results of a recent trial of similar design in which women with IGT and a prior history of gestational diabetes were treated with troglitazone (60). In that study, troglitazone attenuated insulin resistance, but did not improve glucose tolerance or ß-cell function (60). The basis for the difference in these results is not obvious, but may relate to the fact that ß-cell dysfunction was likely to have been present for a protracted period of time in women with a history of gestational diabetes. Alternatively, there may be significant differences in the pathogenetic mechanisms that underlie the insulin secretory dysfunction in these two populations.
In the only previous study in which troglitazone was used to treat women with PCOS, there was no significant improvement in oral glucose tolerance, despite significant improvements in insulin resistance (25). It is important to note, however, that nearly two thirds of the subjects enrolled in that trial had normal glucose tolerance at baseline and thus might not be expected to demonstrate significant improvement in glucose tolerance.
It has been shown previously that troglitazone administration lowers plasma androgen levels in women with PCOS in correlatation to a reduction in hyperinsulinemia (25). However, it was not possible to determine whether ovarian androgen production per se was reduced or, alternatively, whether free androgen concentrations declined because of the insulin-related increase in serum levels of SHBG. We, too, found significant reductions in basal levels of both total and free testosterone that were associated with a significant increase in the concentration of SHBG. In addition, the leuprolide-stimulated levels of total testosterone, androstenedione, 17-hydroxyprogesterone, and DHEA declined significantly after treatment with troglitazone, indicative of decreases in ovarian steroid secretion. Our results are consistent with the hypothesis that insulin modulates the 17-hydroxylase and 17,20-lyase activities of the ovarian steroid-forming P450c17 (6, 61). This enzyme is characteristically abnormally regulated in women with the ovarian androgen excess of PCOS (6, 61), as evidenced by a characteristic pattern of steroid secretion in response to the endogenous gonadotropin surge induced by a single injection of GnRH agonist (26, 39, 62, 63, 64, 65). Two lines of evidence suggest that the reduction in leuprolide-stimulated steroid levels results primarily from the reduction in insulin levels. First, there were no changes in either basal or stimulated concentrations of LH or FSH, or in the ratio of LH to FSH after troglitazone treatment, which could account for the attenuated steroidogenic response. Second, there was a significant correlation between the decline in insulin levels and the decline in leuprolide-stimulated androgen (androstenedione and testosterone) secretion (r = 0.71; P < 0.02). Such a correlation was not demonstrable for individual steroids, possibly because the conversion of androstenedione to testosterone is enhanced in PCOS (66, 67).
Two previous studies have examined the relationship between the reduction of hyperinsulinemia and ovarian steroidogenesis in response to a GnRH agonist (5, 68). In both cases, metformin was administered to attenuate hyperinsulinemia. Where this was achieved (68), there was a demonstrable reduction in glucose-stimulated insulin secretion, which correlated with a reduction in leuprolide-stimulated serum levels of 17-hydroxyprogesterone. In the instance in which insulin levels did not decline after metformin treatment (5), there was no decline in ovarian androgen biosynthesis. Although this discordant effect of metformin on hyperinsulinemia cannot readily be explained, taken together, the results from the previous and present studies are consistent with the hypothesis that attenuation of hyperinsulinemia can reduce ovarian steroid production.
Treatment with troglitazone led to decreased concentrations of PAI-1 in blood. A concordant decrease in the activity of PAI-1 in blood was observed. The decrease in the concentration of t-PA in response to troglitazone was not associated with a change in the functional activity of t-PA in the blood of these subjects with PCOS. Although no change in the activity of t-PA was observed under basal conditions, an improved fibrinolytic response to thrombosis could be anticipated because of the substantial decrease in the concentration and activity of PAI-1 after treatment with troglitazone. Accordingly, treatment with troglitazone improves the fibrinolytic capacity in the blood of these subjects with insulin resistance.
Increased concentrations of PAI-1 have been observed in subjects with insulin-resistant states, including diabetes (69), obesity (70), hypertension (71), and PCOS (15, 72). Weight loss (70, 73) and treatment with metformin (74) have both been shown to decrease circulating concentrations of insulin as well as PAI-1. Insulin has been shown to directly augment expression of PAI-1 in vivo and in vitro (19, 22, 75). We have recently reported a synergistic interaction between insulin and triglyceride associated with very low density lipoprotein (20). The positive correlation between the reduced secretion of insulin in response to glucose and the decreased plasma concentration of PAI-1 after treatment with troglitazone (r = 0.61; P < 0.03) is consistent with the proposed direct role of insulin in modulating expression of PAI-1. In the subjects with PCOS reported here, the combination of a reduction in the secretion of insulin and the decreased concentrations of triglyceride in plasma after treatment with troglitazone is likely to account for the reduction in blood concentrations of PAI-1 observed.
In summary, we have found that administration of troglitazone to women with PCOS and IGT ameliorates each of the metabolic and hormonal derangements that characterize these women. Given the high prevalence of the disorder (6) and the significant risk it imparts for the development of glucose intolerance (5, 6, 7, 9) and cardiovascular morbidity and mortality (14, 15, 16, 17, 18), troglitazone may prove to be a useful adjunct in the treatment of women with PCOS.
Acknowledgments
The authors thank Jody Dushay, Eunice Yim, and Anupam Basu for their assistance in coordinating this study. Dr. Zubie Sheikh kindly performed and evaluated the transvaginal ultrasonography. The nursing staff of the Clinical Research Center provided expert care of the subjects who participated in the study. The authors appreciate the critical review of the manuscript by Roy Weiss, M.D., Ph.D.
Footnotes
1 This work was supported in part by Grants DK-02315, DK-31842,
DK-20595, HD-06308, and DK-07011–17; General Clinical Research Center
Grant MO1-RR-00055, and in part by Parke-Davis Pharmaceuticals. 
Received February 4, 1997.
Revised March 28, 1997.
Accepted April 8, 1997.
References
as the cause of polycystic ovary
syndrome. Fertil Steril. 53:785–791.[Medline]
activity and serum free testosterone after
reduction of insulin secretion in polycystic ovary syndrome. N
Engl J Med. 335:617–623.This article has been cited by other articles:
 |
|
 |
|
  H. TFAYLI and S. ARSLANIAN Menstrual Health and the Metabolic Syndrome in Adolescents Ann. N.Y. Acad. Sci., June 1, 2008; 1135(1): 85 - 94. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E. Minge, B. D. Bennett, R. J. Norman, and R. L. Robker Peroxisome Proliferator-Activated Receptor-{gamma} Agonist Rosiglitazone Reverses the Adverse Effects of Diet-Induced Obesity on Oocyte Quality Endocrinology, May 1, 2008; 149(5): 2646 - 2656. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Rasouli, P. A Kern, E. A. Reece, and S. C. Elbein Effects of pioglitazone and metformin on beta-cell function in nondiabetic subjects at high risk for type 2 diabetes Am J Physiol Endocrinol Metab, January 1, 2007; 292(1): E359 - E365. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 S. E. Kahn, S. M. Haffner, M. A. Heise, W. H. Herman, R. R. Holman, N. P. Jones, B. G. Kravitz, J. M. Lachin, M. C. O'Neill, B. Zinman, et al. Glycemic Durability of Rosiglitazone, Metformin, or Glyburide Monotherapy N. Engl. J. Med., December 7, 2006; 355(23): 2427 - 2443. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. DIAMANTI-KANDARAKIS and F. ECONOMOU Stress in Women: Metabolic Syndrome and Polycystic Ovary Syndrome Ann. N.Y. Acad. Sci., November 1, 2006; 1083(1): 54 - 62. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Z. T. Bloomgarden Third Annual World Congress on the Insulin Resistance Syndrome: Associated conditions. Diabetes Care, September 1, 2006; 29(9): 2165 - 2174. [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]
|
|
|
|
 |
|
 R. Pasquali and A. Gambineri Insulin-sensitizing agents in polycystic ovary syndrome. Eur. J. Endocrinol., June 1, 2006; 154(6): 763 - 775. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 K. Rautio, J.S. Tapanainen, A. Ruokonen, and L.C. Morin-Papunen Endocrine and metabolic effects of rosiglitazone in overweight women with PCOS: a randomized placebo-controlled study Hum. Reprod., June 1, 2006; 21(6): 1400 - 1407. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. M. Kanaya, C. Wassel Fyr, E. Vittinghoff, T. B. Harris, S. W. Park, B. H. Goodpaster, F. Tylavsky, and S. R. Cummings Adipocytokines and incident diabetes mellitus in older adults: the independent effect of plasminogen activator inhibitor 1. Arch Intern Med, February 13, 2006; 166(3): 350 - 356. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J.-P. Baillargeon and J. E. Nestler Polycystic Ovary Syndrome: A Syndrome of Ovarian Hypersensitivity to Insulin? J. Clin. Endocrinol. Metab., January 1, 2006; 91(1): 22 - 24. [Full Text] [PDF]
|
|
|
|
|
|
N. A. Cataldo, F. Abbasi, T. L. McLaughlin, M. Basina, P. Y. Fechner, L. C. Giudice, and G. M. Reaven Metabolic and ovarian effects of rosiglitazone treatment for 12 weeks in insulin-resistant women with polycystic ovary syndrome Hum. Reprod., January 1, 2006; 21(1): 109 - 120. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Lemay, S. Dodin, L. Turcot, F. Dechene, and J-C. Forest Rosiglitazone and ethinyl estradiol/cyproterone acetate as single and combined treatment of overweight women with polycystic ovary syndrome and insulin resistance Hum. Reprod., January 1, 2006; 21(1): 121 - 128. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. E Minge, N. K. Ryan, K. H. V. D. Hoek, R. L. Robker, and R. J. Norman Troglitazone Regulates Peroxisome Proliferator-Activated Receptors and Inducible Nitric Oxide Synthase in Murine Ovarian Macrophages Biol Reprod, January 1, 2006; 74(1): 153 - 160. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Glintborg, R. K. Stoving, C. Hagen, A. P. Hermann, J. Frystyk, J. D. Veldhuis, A. Flyvbjerg, and M. Andersen Pioglitazone Treatment Increases Spontaneous Growth Hormone (GH) Secretion and Stimulated GH Levels in Polycystic Ovary Syndrome J. Clin. Endocrinol. Metab., October 1, 2005; 90(10): 5605 - 5612. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
S. A.R. Doi, M. Al-Zaid, P. A. Towers, C. J. Scott, and K. A.S. Al-Shoumer Irregular cycles and steroid hormones in polycystic ovary syndrome Hum. Reprod., September 1, 2005; 20(9): 2402 - 2408. [Abstract] |
