- Email: [email protected]
Circulating leptin concentrations in women with hirsutism
FERTILITY
AND
Vol. 68, No. 5, November
STERILITY@
Copyright
o 1997 American
Society for Reproductive
Published
by Elsevier Science Inc.
1997
Printed on acid-free paper in U. S. A.
Medicine
Circulating leptin concentrations in women with hirsutism
Hector F. Escobar-Morreale, M.D., Ph.D. Joaquin Serrano-Gotarredona, M.D.* Cesar Varela, M.D., Ph.D. Department
of Endocrinology,
Rafael Garcia-Robles, M.D., Ph.D. Jose M. Sancho, M.D., Ph.D.
Hospital Ramcin y Cajal, Madrid,
Spain
Objective: To evaluate serum leptin concentrations in hirsute women. Design: Controlled clinical study. Setting: Tertiary institutional hospital. Patient(s): Thirty-three hirsute women and 11 healthy female controls. Intervention(s): Serum samples were obtained at baseline and on day 1 (gonadal stimulation) and day 21 (gonadal suppression) after the IM injection of a single 3.75mg dose of triptorelin. Main Outcome Measure(s): Leptin, T, sex hormone-binding globulin (SHBG), insulin, and glucose levels and free androgen index. Result(s): Leptin levels were increased in hirsute women in comparison with control subjects at baseline and on day 1. Leptin levels increased on day 1 compared with baseline and then decreased to baseline by day 21. Leptin levels correlated with body mass index (r = 0.76), SHBG levels (r = -0.52), free androgen index (r = 0.381, insulin levels (r = 0.461, and the glucose/ insulin ratio (r = -0.38). When the effect of obesity on these results was removed by analysis of covariance and partial correlation analysis, leptin levels remained elevated only on day 1 and the only correlations that remained significant were those of leptin with insulin (r = 0.24) and the glucose/insulin ratio (r = -0.24). Conclusion(s): The increased leptin levels found in hirsute women are related mainly to obesity and also to insulin resistance. Leptin levels increased during gonadal stimulation and returned to baseline during gonadal suppression, suggesting that leptin also is influenced by 0 1997 by American Society for Repro(Fertil Sterila 1997;68:898-906. the gonadal axis. ductive Medicine.) Key Words: Leptin, hirsutism, hyperandrogenism, polycystic ovary syndrome, insulin resistance, sex hormone-binding globulin, free androgen index, testosterone, GnRH analogues
Leptin, the product of the ob gene, is a protein hormone secreted by adipocytes that has been shown to increase energy expenditure and decrease appetite (1). Genetically obese oblob mice, in which functional leptin is deficient, progressively develop obesity, insulin resistance, diabetes, and infertility (2). Infertility in obese oblob mice is characterized by low gonadotropin concentrations, which increase with
Presented at the 79th Annual Meeting of the Endocrine Society, Minneapolis, Minnesota, June 11-14, 1997. Reprint requests: Hector F. Escobar-Morreale, M.D., Ph.D., Department of Endocrinology, Hospital Ramon y Cajal, Carretera de Colmenar Km. 9,100, 28034 Madrid, Spain (FAX: 34-1-3369016; e-mail: [email protected]). * Present address: Section of Endocrinology, Hospital General Universitario de Alicante, Alicante, Spain. 898
the administration of exogenous leptin, promoting follicular development and restoring fertility (3). The recent discovery of leptin receptor messenger RNA in the brain and the ovary suggests that these effects might be related to leptin actions on the hypothalamic-pituitary axis as well as to direct actions on the ovary (4, 5). As opposed to oblob mice, obese humans show elevated serum leptin concentrations, suggesting that most obese persons are insensitive to endogenous leptin (6). Obesity and body fat seem to be the major factors regulating circulating leptin, which also is influenced by gender and age (7). Leptin is related closely to insulin and glucose metabolism. Insulin appears to upregulate leptin secretion chronically but indirectly by its trophic action on aclipocytes (81. Moreover, pancreatic receptors for leptin have been described, sug0015-0282/97/$17.00 PI1 SOO15-0282(97)00336-l
gesting that leptin also may regulate insulin release as part of an adipoinsular feedback (9). Although acute hyperinsulinemia does not influence leptin secretion either in vivo or in vitro (8, lo), insulin resistance, and the resultant chronic hyperinsulinemia, are associated with an increased secretion of leptin that is independent of body fat mass (11) and that may be reversed by insulin-sensitizing drugs, at least in vitro (12). The influence of nutritional status on gonadal function has been recognized for more than 30 years (13). Leptin recently has been proposed as the metabolic signal produced by the adipose tissue to influence the reproductive axis (14). Female mice treated with human recombinant leptin show an earlier onset of puberty than placebo-treated mice (14) and, as stated previously, treatment with exogenous leptin in leptin-deficient mice restores fertility (3). Polycystic ovary syndrome (PCOS), a syndrome characterized by hyperandrogenism, oligomenorrhea or amenorrhea, and infertility, frequently is associated with obesity and insulin resistance. Recently, Brzechffa et al. (15) reported that serum leptin concentrations are increased in some women with PCOS to a greater extent than expected for their body mass index (BMI) and insulin sensitivity, suggesting that abnormalities in leptin signaling to the reproductive system may be involved in the pathogenesis of this syndrome. However, it is still not clear whether the elevated serum leptin levels in patients with PCOS are a consequence of the higher prevalence of obesity in these women (16-19), and the relation between leptin, insulin resistance, and sex steroids in patients with PCOS remain controversial (17- 19). Moreover, to our knowledge, a possible influence of gonadal function on the regulation of serum leptin concentrations in hirsute women has not been reported. To provide new insights into the possible pathophysiologic role of leptin in hirsutism, we evaluated serum immunoreactive leptin concentrations in a group of women with this disorder. The possible influence of the changes in gonadal function induced by the administration of the long-acting GnRH analogue triptorelin on circulating leptin and the relation of serum leptin to obesity, insulin resistance, and hyperandrogenism also were studied. MATERIALS AND METHODS Subjects
Thirty-three women (mean age ?SE, 23.6 ? 1.3 years) who were referred to our department for the evaluation of hirsutism were included in the study. Hirsutism was defined by the presence of excessive Vol. 68, No. 5, November
1997
body hair distributed in an androgen-dependent pattern, with a Ferriman-Gallwey score of >7 (mean age +- SE, 15.6 ? 0.9 range, 8-34). Menstrual cycle intervals were evaluated on recall for every patient. Oligomenorrhea was defined by the presence of three or more cycles of ~35 days in the previous 6 months, and amenorrhea by the lack of vaginal bleeding for 3 months. None of the patients was hypertensive or had evidence of Cushing’s disease or drug-induced hirsutism. Hyperprolactinemia, thyroid disease, acromegaly, and nonclassic congenital adrenal hyperplasia were ruled out by appropriate testing. The reference values for the analytic procedures were obtained from a control group of 11 healthy menstruating women (mean age &SE, 27.1 + 1.8 years) without signs and symptoms of hyperandrogenism or a family history of endocrine diseases. None of the patients or control subjects had been receiving hormonal medications, including contraceptive pills, for the last 6 months. The study was approved by the hospital ethics committee, and informed consent was obtained from each patient and control subject. Experimental Design
Studies were performed during the follicular phase, between days 5 and 10 of the menstrual cycle, or during amenorrhea after excluding gestation by appropriate testing. After undergoing an abdominal and pelvic ultrasonogram (US), the patients reported to the endocrine-metabolic testing room between 8:00 A.M. and 9:00 A.M., having undertaken a 12-hour overnight fast. An indwelling IV line was placed in a forearm vein and, after 15-30 minutes, basal blood samples were obtained for the measurement of leptin, cortisol (F), total T, LH, FSH, Es, sex hormone-binding globulin (SHBG), glucose, and insulin levels. Immediately after the basal samples were taken, a 250~pg lV bolus of l-24 ACTH (Synacthen; CibaGeigy, Basel, Switzerland) was injected and blood samples were obtained at 0 and 60 minutes. After the 60-minute sample of the ACTH test was obtained, a 100~pg IV bolus of GnRH (Luforan; Serono, Madrid, Spain) was injected and blood samples for LH and FSH were obtained at 0 and 30 minutes. A single dose (3.75 mg IM) of triptorelin (D-Trp’GnRH, Decapeptyl; LASA-Ipsen, Barcelona, Spain) then was administered. Basal sampling and the ACTH test were repeated on the next day (day 1) and, together with a GnRH test, 21 days after triptorelin administration (day 2 l), when the patients were in gonadal axis stimulation and suppression states, respectively (20). To avoid changes in body weight, the patients were inEscobar-Morreale
et al.
Circulating
leptin
in hirsutism
899
strutted to maintain their usual caloric intake during the 21 days after triptorelin administration. Basal E2 and basal and GnRH-stimulated LH and FSH levels were used to confirm the stimulation and suppression of the gonadal axis after triptorelin administration. In nine randomly selected hirsute women, insulin sensitivity was determined in vivo 1 week before triptorelin administration, using a modified endogenous insulin secretion-suppression test (21). In brief, this test consisted of a constant infusion for 150 minutes of glucose, insulin, and somatostatin. Insulin sensitivity was estimated by the steady-state plasma glucose concentrations during the last 60 minutes of the test. The insulin sensitivity index (in dL/kg per minute) was calculated by the following formula: (glucose infusion rate/ steady-state plasma glucose) x 103, as previously described (21). In control women, triptorelin was not administered because they considered ovarian suppression unacceptable. Some of these women were included in the control groups of previous publications (22, 23). The basal and ACTH-stimulated adrenal steroid profiles at baseline and during ovarian suppression of some of the patients have been published previously (22, 23). Thus, the results of the ACTH test will not be described further here. All blood samples were centrifuged immediately and serum was separated and frozen at -20°C for this and future studies until assayed. Assays Serum leptin concentrations were measured by direct RIA (Linco Research Inc., St. Charles, MO), with intraassay and interassay coefficients of variation (CVs) of 6.5% and 4.5%, respectively. Insulin was measured by direct RIA with standards calibrated against the World Health Organization 66/304 international standard (SORIN Biomedica S.p.A., Saluggia, Italy), with intraassay and interassay CVs of 2.5% and 10.9%, respectively. Plasma glucose was assayed by the glucose oxidase method (Beckman Glucose Analyzer 2; Beckman Instruments, Fullerton, CA). Cortisol, T, and SHBG were assayed as reported previously (22, 23) and the free androgen index was calculated using the following formula: [T (nmol&) x lOOl/SHBG (nmol/L). Luteinizing hormone, FSH, and E, were assayed by a commercial chemiluminescent enzyme immunoassay (Immulite; Diagnostic Products Corporation, Los Angeles, CA), with intraassay and interassay assay 999
Escobar-Morreale et al.
Circuhtin~ httn in h.ir~tkm
CVs of 5.7% and 9.6% for LH, 6.4% and 7.5% for FSH, and 9.3% and 10.6% for Ez, respectively. The molar glucose/insulin ratio was calculated by the following formula: [glucose (mg/dL) x 0.055611 [insulin (@/mL)
X
7.1751,
and was expressed in a 10’ order of magnitude (mmol/pmol). All hormone determinations for each individual were measured in duplicate within a single assay. Ultrasonography Transparietal abdominal and pelvic US was performed in every patient using a Toshiba 3.75-MHz transducer (Tosbee SSA-240; Toshiba Medical Systems, Tokyo, Japan), as previously described (22,23). Statistical Analysis Results are expressed as means ? SE. Repeated measures multivariate analysis of variance (ANOVA) was used to evaluate the evolution of each patient’s hormonal variables during sequential gonadal stimulation and suppression. Paired t-tests were used to identify these differences only when the result of the repeated measures multivariate ANOVA was statistically significant. Repeated measures multivariate two-way ANOVA also was used to compare obese and lean patients and was followed by one-way ANOVA when significant changes were found. One-way ANOVA followed by Dunnet’s post hoc procedure was used to compare patients with control subjects. Analysis of covariance was used to evaluate the influence of obesity, measured by the BMI, on the differences in serum leptin concentrations between patients and control subjects. The relation between the different variables were evaluated by correlation analysis, and Pearson’s coefficient was calculated. Partial correlation analysis, introducing the BMI as the control variable, was used to remove the influence of obesity on the correlations observed. P co.05 was considered to be statistically significant. RESULTS Evolution of Sex Steroids and Gonadotropins During Gonadal Stimulation and Suppression
The administration of triptorelin resulted in an increase in basal LH, FSH, and E2 levels by day 1, which was followed by a marked suppression of these Fertility
and Sterility@
hormones by day 21 (Fig. 1). Compared with baseline, GnRH-stimulated LH and FSH levels were reduced markedly by day 21 (LH, 27.1 2 5.4 versus 3.3 f 0.3 mIU/mL; FSH, 8.1 + 0.6 versus 2.2 2 0.3 mIU/mL; P
1
90
124
ID
I
425 405 1
I’
T
_II 1 Wb' I
323
I
1
AmO.81,P-3.42,P.zO.O~O I
torelin in suppressing gonadotropin secretion. The repeated measures multivariate ANOVA showed significant changes in T and SHBG levels and in the free androgen index after triptorelin administration (Fig. 1, Table 11,which consisted of the following: 1. During gonadal stimulation (day l), the free androgen index increased and the SHBG level decreased, with respect to baseline, whereas the T level did not change (Fig. 1, Table 1). To rule out a possible influence of the acute ACTH stimulation test performed at baseline on the hormonal changes observed by day 1, basal F concentrations were measured and showed no increase with respect to baseline (baseline, 15.9 t 1.2pg/dL versus day 1, 14.6 + 1.1 &dL; paired t-test = 1.18, nonsigni6cant). 2. During gonadal suppression (day 211, the T level and the free androgen index decreased with respect to both baseline and day 1 and the SHBG level remained similar to baseline (Fig. 1, Table 1). The BMI did not change throughout the study (baseline: 25.5 + 0.9 versus day 21, 25.4 2 0.9, in kg/m’). The repeated measures multivariateANOVA showed no significant differences in insulin and in the glucose/ insulin ratio between baseline, day 1, and day 21 (Table 1). Finally, when comparing the group of hirsute patients with the control subjects,the T level was elevated at baseline and day 1 and the free androgen index was elevated and the SHBG level was decreased at baseline, day 1, and day 21, whereas the insulin level was elevated only at baseline (Table 1). Serum Leptin Concentrations
I
I
I
50
1
*
Bl
21 Day of measurement
Figure 1 Changes in basal serum leptin T, SHBG, Ez, LH, and FSH concentrations and in the free androgen index at baseline (B), day 1 (I), and day 21(21) after IM administration of a single 3.75mg dose of triptorelin. When the multivariate repeated measures ANOVA (included in the panels) resulted in statistically significant differences between the different days of measurement, repeated paired t-tests were applied to identify these differences. *At least P < 0.05 with respect to baseline. TAt least P < 0.05 with respect to day 1. .$At least P < 0.05 with respect to baseline and day 1. Vol. 68, No. 5, November
1997
Serum leptin concentrations were elevated in hirsute women compared with control subjects at baseline and on day 1, but not on day 21 (Table 1). Serum leptin concentrations also were different in hirsute patients depending on the day that the concentrations were measured (Fig. 1, Table 1); serum leptin levels increased by day 1 with respect to baseline and returned to baseline levels by day 21 (Fig. 1, Table 1). The serum leptin concentrations of the patients did not vary according to the presence of polycystic ovaries on US, which were found in 16 patients or to the presence of oligomenorrhea or amenorrhea, which occurred in 15 and 1 patients, respectively (data not shown). Moreover, the 19 hirsute patients with functional ovarian or adrenal hyperandrogenism (22) had leptin concentrations similar to those of the 14 patients in whom serum androgen levels were within the normal range (data not shown). Correlation of Circulating Levels of Leptin With Other Variables
Serum leptin levels showed a strong correlation with BMI when both patients and control subjects Escobar-Morreale
et al.
Circulating
leptin
in hirsutism
901
Table 1 Comparison of T, Insulin, SHBG, and Serum Leptin Concentrations, Free Androgen Index, and Glucose/Insulin Ratio of Hirsute Patients With the Reference Values Derived From the Control Group Controls (n = 111 Leptin (ng/mL) T (ng/mL) Free androgen index SHBG ( pg/mLl Insulin ( @/mLl Glucose/insulin ratio
(X
10’)
9.6 38.0 2.7 542 6.9 0.089
5 5 2 2 2 2
1.4 4.3 0.7 63 0.7 0.009
Patients (n = 331 Day 1
Baseline 17.8 70.3 8.2 377 12.4 0.062
+ 2 5 2 c ?
20.3 75.5 9.7 351 11.2 0.066
2.0* 5.1* 1.33* 33” 1.3* 0.007
Day 21
2 2.3* 2 5.5* -c 1.4* 2 34* k 1.0 5 0.008
17.5 52.2 6.2 366 10.9 0.065
2 z t k lr 2
2.3 4.8 l.O* 33* 1.0 0.005
* At least P < 0.05 compared with the control group by one-way ANOVA and Dunnet’s test for comparison with a control group.
were considered (r = 0.76, P < 0.001). The correlation of serum leptin concentrations with hormonal and metabolic variables was evaluated, including the values of the patients at baseline and on days 1 and 21, together with the baseline values obtained from the control subjects. Serum leptin concentrations showed a significant correlation with SHBG levels (r = -0.52, P < O.OOl),the free androgen index (r = 0.38, P < O.OOl), insulin levels (r = 0.46, P < O.OOl),and the glucose/insulin ratio (r = -0.38, P < O.OOl>,but not with T or E2 levels (Fig. 2). The increase in serum leptin levels that occurred in response to triptorelin administration, expressed as the percent increase with respect to baseline values, correlated with the percent increases in the free androgen index and T levels (r = 0.35, P < 0.05 and r = 0.39, P < 0.03, respectively) but not with the percent increases in EP, LH, or FSH levels, or with the percent decrease in SHBG levels. When the control subjects were studied separately, the only correlation that persisted was the one between serum leptin levels and the free androgen index (r = 0.65, P < 0.05). Although the insulin secretion-suppression test was performed in a small number of patients (n = 9; insulin sensitivity index 54.1 t 8.9 dL/kg per minute, range, 20.6-107.6 dL/kg per minute), serum leptin levels showed a strong inverse correlation with the insulin sensitivity index (r = -0.70, P < 0.040). The insulin sensitivity index also presented near-significant correlations with SHBG levels (r = 0.62, P = 0.077) and with the free androgen index (r = -0.61, P = 0.079). No correlations were found between the insulin sensitivity index and baseline insulin levels or the glucose/insulin ratio (data not shown). Influence of Obesity on Serum Leptin and Sex Steroid Concentrations and on the Correlations Between Serum Leptin Concentrations and Other Variables
Although there were no significant differences in BMI between patients and control subjects (25.5 5 902
Escobar-Morreale et al.
CirculatingZeptin in hirsutism
0.9 [range, 17.0-37.21 versus 22.8 ? 0.6 [range, 20.2-27.3, in kg/m’]; F = 2.89, not significant), when correcting for the influence of BMI by analysis of covariance only a near-significant tendency (P = 0.065) to higher leptin levels in hirsute women compared with control subjects persisted at baseline, whereas the increase in serum leptin by day 1 remained significant (P < 0.02). When comparing obese and lean hirsute women, using a cutoff BMI value of 25 (kg/m2), obese patients showed higher
200
,
I
I
, 600
I
I
lb
I
0
20
40
60
:6
6
2;
4b
6b
&
Figure 2 Correlations of circulating leptin levels with serum SHBG, insulin, T and Es concentrations, and with the free androgen index and the glucose/insulin ratio, including the values from patients at baseline, day 1 and 21, and the values from control subjects. The coefficients of correlation and their statistical significance are described in the text. Fertility and Sterility@
leptin concentrations at baseline, day 1, and day 21 than did lean patients (Table 2). Insulin levels also were higher in obese patients at baseline, day 1, and day 21 compared with lean patients, whereas no differences in the glucose/insulin ratio were observed (Table 2). Despite the fact that T levels were the same in obese and lean hirsute women, SHBG levels were reduced markedly at baseline, day 1, and day 21 and the free androgen index was increased at baseline and day 1 in obese patients compared with lean hir-
Table 2
Comparison
sute women (Table 2). Serum LH, FSH, and E2 concentrations were not different between obese and lean hirsute women (Table 2). Moreover, the stimulator-y effect of triptorelin on the gonadal axis and on serum leptin levels, observed by day 1, did not appear to be influenced by obesity because no differences were found between obese and lean patients in the percent increases in serum leptin, T, LH, FSH, and Ez levels and in the free androgen index, or in the percent decrease in SHBG levels, compared with baseline values (Table 2).
of Obese and Lean Hirsute Patients
Value
Obese patients (n = 14)
BMI (kg/m’) Ferriman-Gallwey score Leptin (ng/mL) Baseline Day 1 (% increase)? Day 21 T (ng/mL) Baseline Day 1 (o/o increase) Day 21 Free androgen index Baseline Day 1 (% increase) Day 21 SHBG ( /&dL) Baseline Day 1 (% increase) Day 21 Insulin ( pU/mL)fi Baseline Day 1 Day 21 Glucose/insulin ratio (X log)111 Baseline Day 1 Day 21 LH (mIU/mL)II Baseline Day 1 (% increase) Day 21 FSH (mIU/mL)/l Baseline Day 1 (% increase) Day 21 Ez (pg/mL)Il Baseline Day 1 (% increase) Day 21
Lean patients (n = 19)
30.2 2 1.2 16.6 + 1.7
22.1 2 0.5* 14.8 t 1.0
27.0 -c 3.0 31.2 2 3.68 (17.4 2 6.2) 26.9 5 4.15
11.0 2 1.2* 12.2 -c 1.2** (13.4 t 4.5) 10.5 2 1.1*
71.2 2 10.6 78.3 2 10.2 (24.1 + 18.9) 51.6 lr 8.2
69.6 t 4.4 73.4 t 6.2 52.6 2 5.8
11.3 2 2.8 13.0 t 2.6 8.4 t l.S$
(39.5 + 20.8)
266 2 36 252 t 39 271 5 44
(-6.0
(11.7 ? 6.7)
5.9 2 0.8* 7.3 t 1.4* (18.7 2 8.6) 4.6 -c 0.9$§ 459 2 43* 423 -e 44*$ 436 2 42*
? 3.9)
14.5 ? 1.6 14.8 + 1.5 13.8 2 2.0
10.8 2 1.9* 8.5 2 0.8* 8.8 2 0.8*
0.050 * 0.009 0.048 -c 0.007 0.054 -+ 0.007
0.071 * 0.009 0.079 2 0.011 0.073 t 0.008
(-8.2
2 3.3)
4.0 2 0.7 27.1 2 4.5 1.3 2 0.2
(777 2 172)
6.0 t 1.0 44.5 c 8.9 1.8 t 0.2
(674 2 75)
4.5 2 0.4 11.8 2 1.2 1.6 2 0.2
(164 + 21)
4.7 +- 0.4 14.1 + 1.4 1.5 * 0.3
(392 2 191)
37 ? 5 136 + 17 15 -+ 2
(305 2 52)
45 ? 6 185 ? 26 15 * 2
(381 ? 48)
Note: A cutoff BMI value of 25 (kg/m’) was used to define obese and lean patients. Data were analyzed by multivariate repeated measures two-way ANOVA, including the day of measurement as the within-subjects factor and being obese or lean as the betweensubjects factor. When statistically significant differences were present depending on the day of measurement, these differences were identified by paired t-test. The differences between obese and lean patients were identified further by one-way ANOVA. * At least P < 0.05 with respect to the obese group. t Values in parentheses are the percent increase or decrease at day 1 with respect to baseline, in response to triptorelin administration. No differences between obese and lean patients were observed in these percent changes for any of the variables studied. t At least P < 0.05 with resnect to baseline. g At least P < 0.05 with respect to day 1. /IBecause no significant differences between obese and lean patients were observed, the changes depending on the day of measurement were evaluated including the whole group of patients, and the significance of the changes in LH, FSH, and Ez levels is included in Figure 1. fi No statistically significant differences were observed depending on the day of measurement, and thus the percent changes with respect to baseline were not calculated. Vol. 68, No. 5, November
1997
Escobar-Morreale
et al.
Circulating
leptin in hirsutism
903
Finally, after controlling for the influence of obesity, as expressed by the BMI, the partial correlation analysis showed that the only correlations that remained significant were those of serum leptin levels with insulin (r = 0.24, P < 0.02) and with the glucose/insulin ratio (r = -0.24, P < 0.02). In contrast, the correlations of serum leptin with SHBG levels, the free androgen index, and the insulin sensitivity index were no longer significant, suggesting that these correlations actually resulted from the much stronger correlation between serum leptin levels and the BMI. DISCUSSION
There is increasing evidence that the metabolic effects of leptin are not restricted to the control of body weight and fat composition. As stated previously, leptin stimulates reproductive function, affording an explanation for the influence of nutritional status on the gonadal axis (3, 14). In addition to its physiologic effects, leptin also might play a role in pathologic abnormalities of the reproductive system and might be the link between obesity and hyperandrogenism. If the latter hypothesis is correct, leptin should be increased in hirsute women, and this increase should be related not only to obesity but also to the degree of hyperandrogenism. In our series, serum leptin concentrations were increased in hirsute patients compared with control subjects, and detailed study of the patients disclosed that this increase was present mostly in obese patients. The increase in leptin levels appeared to be influenced by insulin sensitivity and also to be related to the degree of hyperandrogenism for the following reasons. First, serum leptin levels still showed a direct correlation with fasting insulinemia, indicating a direct correlation with the degree of insulin resistance and an inverse correlation with the glucose/insulin ratio, which is an index of insulin sensitivity. Second, serum leptin levels in both patients and control subjects correlated with the free androgen index but not with T and E2 levels. This might suggest that the relation between leptin and the free androgen index is mediated by SHBG, which in turn depends on insulin action and/ or insulin sensitivity, closing the loop between obesity, insulin resistance, and hirsutism. However, the correlation between the free androgen index and leptin also was present when the control group was analyzed separately, whereas leptin did not correlate with SHBG in these subjects. Thus, a direct relation between leptin and the bioavailability of T to the pilosebaceous unit could not be excluded completely. It is surprising that serum leptin levels increased 904
Escobar-Morreale
et al.
Circulating Zeptin in hirsutisn
further by day 1, and this increase appeared to be related to the acute stimulatory effect of triptorelin on the gonadal axis. Because glucocorticoids increase leptin messenger RNA expression and leptin secretion (24), it also would appear that the increase in serum leptin levels on day 1 might have been related to an increase in glucocorticoid secretion resulting from the ACTH test performed the day before. However, the lack of an increase in F levels in our patients suggests that the stimulatory effect on the adrenal of the ACTH test performed at baseline did not persist until day 1. Moreover, two recent studies showed that serum leptin levels do not increase during acute adrenal stimulation. In fact, circulating leptin showed a minimum decrease after IV ACTH administration in patients with normal adrenal function, and this decrease was independent of BMI or sex (25). In the second study, serum leptin did not increase after the administration of IV ovine corticotropin-releasing hormone, despite marked increases in ACTH and F levels, either in patients with obesity and chronic hypercortisolism due to Cushing’s disease or in lean, normal control subjects (26). To our knowledge, the possibility that the reproductive system modulates leptin concentrations has not been explored previously. Our present results suggest that, in addition to the influences of leptin on the gonadal axis, the gonadal axis itself might upregulate leptin secretion acutely, in what could be considered an adipogonadal axis. This is supported by the finding of a statistically significant correlation between the acute percent increase in circulating leptin levels after triptorelin administration and the corresponding percent increases in the free androgen index and in T levels. This correlation further suggests that leptin is involved in the development of hyperandrogenism in hirsute women, or at least that it is regulated by the same factors responsible for their hyperandrogenemia. The percent increase in leptin showed no correlation with the corresponding increase in Ez , as would have been expected from the previously reported finding of higher leptin levels in premenopausal women compared with postmenopausal women and with men (27). This lack of a relation between leptin and E, also might explain why serum leptin levels only returned to baseline 21 days after triptorelin administration, instead of showing a greater decrease. In contrast to Ea, the decrease in circulating androgens was much less pronounced during gonadal suppression, further suggesting that, at least in hirsute women, leptin levels correlate better with androgens than with estrogens. As expected, obesity explained the increase in circulating leptin levels observed in hirsute women Fertility and Sterility@
and, with the exception of the correlation of leptin with insulin and the glucose/insulin ratio, obesity justified all the significant correlations observed between leptin and other hormones. Moreover, in addition to higher leptin concentrations, obese hirsute women had higher insulin levels, a higher free androgen index, and lower SHBG levels compared with lean patients. Thus, our experimental design does not permit us to conclude whether the increase in serum leptin is involved directly in the development of hirsutism or merely reflects the presence of obesity in these women, without any pathophysiologic significance. However, the changes in serum leptin levels that occurred in response to triptorelin administration, paralleling the changes in serum androgen levels, were not different between obese and lean patients, providing indirect evidence of a certain relation between the gonadal axis and leptin metabolism, with possible biologic significance. To date, studies regarding the influence of obesity on serum leptin concentrations in women with PCOS have produced conflicting results. Brzechffa et al. (15) found an increase in serum leptin concentrations in a selected population of hirsute women with PCOS, which was greater than expected for their obesity and insulin sensitivity. In contrast, several recent studies (16-19) have shown that the increased circulating leptin levels found in women with PCOS, as well as the correlations of serum leptin levels with serum androgen and SHBG levels (19), are related primarily to obesity, in agreement with the data presented here. In contrast to previous studies, our series also includes hirsute women without polycystic ovaries on US or menstrual disturbances, in addition to patients with PCOS. This result suggests that the relation between obesity, leptin, and hyperandrogenism is not limited to patients with complete PCOS and also might be involved in milder forms of hirsutism. In conclusion, serum leptin levels are increased in hirsute women with or without PCOS. This increase is dependent mostly on obesity, and also is related independently to the insulin resistance present in hirsute women. However, the correlation of the increase in serum leptin levels with the increase in the free androgen index and in T levels in response to triptorelin administration also suggests that leptin might play a role in the development of hyperandrogenism. Moreover, the increase in circulating leptin levels during gonadal stimulation and the tendency to normalization of serum leptin levels during gonadal suppression suggest that serum leptin also is influenced by the gonadal axis. Further studies are needed to determine whether an adipogonadal axis Vol. 68, No. 5, November
1997
exists, which might explain the frequent association of changes in body weight and reproductive abnormalities. Acknowledgments. The authors thank Mrs. Genoveva Gonzalez and Mrs. M” Dolores Carreto for their technical assistance. REFERENCES 1. Weigle DS, Bukowski T, Foster D, Holderman S, Kramer JM, Lasser G, et al. Recombinant ob protein reduces feeding and body weight in the oblob mouse. J Clin Invest 1995;96: 2065-70. 2. Coleman DL. Obesity and diabetes: two mutant genes causing diabetes-obesity syndromes in mice. Diabetologia 1978; 14:141-8. 3. Chehab FF, Lim ME, Lu R. Correction of the sterility defect in homozygous obese female mice by treatment with the human recombinant leptin. Nat Genet 1996; 12:318-20. 4. Lee GH, Proenca R, Montez JM, Carroll KM, Darvishzadeh JG, Lee JI, et al. Abnormal splicing of the leptin receptor in diabetic mice. Nature 1996;379:632-5. 5. Tartaglia LA, Dembski M, Weng X, Deng N, Culpepper J, Devos R, et al. Identification and expression cloning of a leptin receptor, OB-R. Cell 1995;83:1263-71. 6. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292-5. 7. Ostlund RE, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age, and metabolic covariates. J Clin Endocrinol Metab 1996; 81:3909-13. 8. Kolaczynski JW, Nyce MR, Considine RV, Boden G, Nolan JJ, Henry R, et al. Acute and chronic effects of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes 1996;45:699-701. 9. Kieffer TJ, Heller RS, Habener JF. Leptin receptors expressed on pancreatic beta-cells. Biochem Biophys Res Commun 1996;224:522-7. 10. Dagogo JS, Fanelli C, Paramore D, Brothers J, Landt M. Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes 1996;45:695-8. 11. Segal KR, Landt M, Klein S. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 1996;45:988-91. 12. Nolan JJ, Olefsky JM, Nyce MR, Considine RV, Caro JF. Effect of troglitazone on leptin production. Studies in vitro and in human subjects. Diabetes 1996;45:1276-8. 13. Kennedy CG, Mitra J. Body weight and food intake as initiating factors for puberty in the rat. J Physiol 1963;166:40818. 14. Chehab FF, Mounzih K, Lu R, Lim ME. Early onset of reproductive function in normal female mice treated with leptin. Science 1997;275:88-90. 15. Brzechffa PR, Jakimiuc AJ, Agarwal SK, Weitsman SR, Buyalos P, Magoffin DA. Serum immunoreactive leptin concentrations in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1996;81:4166-9. 16. Chapman IA, Wittert GA, Norman F&J.Circulating leptin concentrations in polycystic ovary syndrome: relation to anthropometric and metabolic parameters. Clin Endocrinol1997;46: 175-81. 17. Mantzoros CS, Dunaif A, Flier JF. Leptin concentrations in the polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:1687-91. Escobar-Morreale
et al.
Circulating
leptin in hirsutism
905
18. Laughlin GA, Morales AJ, Yen SSC. Serum leptin levels in women with polycystic ovary syndrome: the role of insulin resistance/hyperinsulinemia. J Clin Endocrinol Metab 1997; 82:1692-6. 19. Rouru J, Anttila L, Koskinen P, Penttila T-A, Irjala K, Huupponen R, et al. Serum leptin concentrations in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:1697-700. 20. Castelo Branco C, Martinez de Osaba MJ, Martinez S, Fortuny A. Effects of a long-acting gonadotropin-releasing hormone analog on the pituitary-ovarian-adrenal axis in women with severe hirsutism. Metabolism 1996;45:24-7. 21. Piedrola G, Novo E, Serrano-Gotarredona J, EscobarMorreale HF, Villa E, Luna JD, et al. Insulin resistance in patients with recent diagnosis of coronary artery disease. J Hypertens 1996; 14:1477-82. 22. Escobar-Morreale HF, Serrano-Gotarredona J, Garcia-Robles R, Sancho J, Varela C. Lack of an ovarian function influence on the increased adrenal androgen secretion present in women with functional ovarian hyperandrogenism. Fertil Steril 1997;67:654-62.
906
Escobar-Morreale
et al.
Circulating
leptin
in hirsutism
23. Escobar-Morreale HF, Serrano Gotarredona J, Garcia Robles R, Sancho J, Varela C. Mild adrenal and ovarian steroidogenie abnormalities in hirsute women without hyperandrogenemia: does idiopathic hirsutism exist? Metabolism 1997; 46:902-7. 24. Papaspyrou-Rao S, Schneider SH, Petersen RN, Fried SK. Dexamethasone increases leptin expression in humans in vivo. J Clin Endocrinol Metab 1997;82:1635-7. 25. Lahlou N, Panneciere MT, Bouvatier C, Chourrout P, Le Fourn J, Bougneres PF, et al. Decrease in leptin secretion during acute adrenal stimulation: A direct effect of ACTH? Proceedings of the 79th Annual Meeting of the Endocrine Society, Minneapolis, MN, 1997, p 197. 26. Cizza G, Lotsikas A, Chrousos GP. Plasma leptin and cortisol levels in Cushing patients before and after correction of hypercortisolism and in lean controls. Proceedings of the 79th Annual Meeting of the Endocrine Society, Minneapolis, MN, 1997, p 119. 27. Rosenbaum M, Nicolson M, Hirsch J, Heymsfield SB, Gallagher D, Chu F, et al. Effects of gender, body composition, and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996;81:3424-7.
Fertility
and Sterility@
AND
Vol. 68, No. 5, November
STERILITY@
Copyright
o 1997 American
Society for Reproductive
Published
by Elsevier Science Inc.
1997
Printed on acid-free paper in U. S. A.
Medicine
Circulating leptin concentrations in women with hirsutism
Hector F. Escobar-Morreale, M.D., Ph.D. Joaquin Serrano-Gotarredona, M.D.* Cesar Varela, M.D., Ph.D. Department
of Endocrinology,
Rafael Garcia-Robles, M.D., Ph.D. Jose M. Sancho, M.D., Ph.D.
Hospital Ramcin y Cajal, Madrid,
Spain
Objective: To evaluate serum leptin concentrations in hirsute women. Design: Controlled clinical study. Setting: Tertiary institutional hospital. Patient(s): Thirty-three hirsute women and 11 healthy female controls. Intervention(s): Serum samples were obtained at baseline and on day 1 (gonadal stimulation) and day 21 (gonadal suppression) after the IM injection of a single 3.75mg dose of triptorelin. Main Outcome Measure(s): Leptin, T, sex hormone-binding globulin (SHBG), insulin, and glucose levels and free androgen index. Result(s): Leptin levels were increased in hirsute women in comparison with control subjects at baseline and on day 1. Leptin levels increased on day 1 compared with baseline and then decreased to baseline by day 21. Leptin levels correlated with body mass index (r = 0.76), SHBG levels (r = -0.52), free androgen index (r = 0.381, insulin levels (r = 0.461, and the glucose/ insulin ratio (r = -0.38). When the effect of obesity on these results was removed by analysis of covariance and partial correlation analysis, leptin levels remained elevated only on day 1 and the only correlations that remained significant were those of leptin with insulin (r = 0.24) and the glucose/insulin ratio (r = -0.24). Conclusion(s): The increased leptin levels found in hirsute women are related mainly to obesity and also to insulin resistance. Leptin levels increased during gonadal stimulation and returned to baseline during gonadal suppression, suggesting that leptin also is influenced by 0 1997 by American Society for Repro(Fertil Sterila 1997;68:898-906. the gonadal axis. ductive Medicine.) Key Words: Leptin, hirsutism, hyperandrogenism, polycystic ovary syndrome, insulin resistance, sex hormone-binding globulin, free androgen index, testosterone, GnRH analogues
Leptin, the product of the ob gene, is a protein hormone secreted by adipocytes that has been shown to increase energy expenditure and decrease appetite (1). Genetically obese oblob mice, in which functional leptin is deficient, progressively develop obesity, insulin resistance, diabetes, and infertility (2). Infertility in obese oblob mice is characterized by low gonadotropin concentrations, which increase with
Presented at the 79th Annual Meeting of the Endocrine Society, Minneapolis, Minnesota, June 11-14, 1997. Reprint requests: Hector F. Escobar-Morreale, M.D., Ph.D., Department of Endocrinology, Hospital Ramon y Cajal, Carretera de Colmenar Km. 9,100, 28034 Madrid, Spain (FAX: 34-1-3369016; e-mail: [email protected]). * Present address: Section of Endocrinology, Hospital General Universitario de Alicante, Alicante, Spain. 898
the administration of exogenous leptin, promoting follicular development and restoring fertility (3). The recent discovery of leptin receptor messenger RNA in the brain and the ovary suggests that these effects might be related to leptin actions on the hypothalamic-pituitary axis as well as to direct actions on the ovary (4, 5). As opposed to oblob mice, obese humans show elevated serum leptin concentrations, suggesting that most obese persons are insensitive to endogenous leptin (6). Obesity and body fat seem to be the major factors regulating circulating leptin, which also is influenced by gender and age (7). Leptin is related closely to insulin and glucose metabolism. Insulin appears to upregulate leptin secretion chronically but indirectly by its trophic action on aclipocytes (81. Moreover, pancreatic receptors for leptin have been described, sug0015-0282/97/$17.00 PI1 SOO15-0282(97)00336-l
gesting that leptin also may regulate insulin release as part of an adipoinsular feedback (9). Although acute hyperinsulinemia does not influence leptin secretion either in vivo or in vitro (8, lo), insulin resistance, and the resultant chronic hyperinsulinemia, are associated with an increased secretion of leptin that is independent of body fat mass (11) and that may be reversed by insulin-sensitizing drugs, at least in vitro (12). The influence of nutritional status on gonadal function has been recognized for more than 30 years (13). Leptin recently has been proposed as the metabolic signal produced by the adipose tissue to influence the reproductive axis (14). Female mice treated with human recombinant leptin show an earlier onset of puberty than placebo-treated mice (14) and, as stated previously, treatment with exogenous leptin in leptin-deficient mice restores fertility (3). Polycystic ovary syndrome (PCOS), a syndrome characterized by hyperandrogenism, oligomenorrhea or amenorrhea, and infertility, frequently is associated with obesity and insulin resistance. Recently, Brzechffa et al. (15) reported that serum leptin concentrations are increased in some women with PCOS to a greater extent than expected for their body mass index (BMI) and insulin sensitivity, suggesting that abnormalities in leptin signaling to the reproductive system may be involved in the pathogenesis of this syndrome. However, it is still not clear whether the elevated serum leptin levels in patients with PCOS are a consequence of the higher prevalence of obesity in these women (16-19), and the relation between leptin, insulin resistance, and sex steroids in patients with PCOS remain controversial (17- 19). Moreover, to our knowledge, a possible influence of gonadal function on the regulation of serum leptin concentrations in hirsute women has not been reported. To provide new insights into the possible pathophysiologic role of leptin in hirsutism, we evaluated serum immunoreactive leptin concentrations in a group of women with this disorder. The possible influence of the changes in gonadal function induced by the administration of the long-acting GnRH analogue triptorelin on circulating leptin and the relation of serum leptin to obesity, insulin resistance, and hyperandrogenism also were studied. MATERIALS AND METHODS Subjects
Thirty-three women (mean age ?SE, 23.6 ? 1.3 years) who were referred to our department for the evaluation of hirsutism were included in the study. Hirsutism was defined by the presence of excessive Vol. 68, No. 5, November
1997
body hair distributed in an androgen-dependent pattern, with a Ferriman-Gallwey score of >7 (mean age +- SE, 15.6 ? 0.9 range, 8-34). Menstrual cycle intervals were evaluated on recall for every patient. Oligomenorrhea was defined by the presence of three or more cycles of ~35 days in the previous 6 months, and amenorrhea by the lack of vaginal bleeding for 3 months. None of the patients was hypertensive or had evidence of Cushing’s disease or drug-induced hirsutism. Hyperprolactinemia, thyroid disease, acromegaly, and nonclassic congenital adrenal hyperplasia were ruled out by appropriate testing. The reference values for the analytic procedures were obtained from a control group of 11 healthy menstruating women (mean age &SE, 27.1 + 1.8 years) without signs and symptoms of hyperandrogenism or a family history of endocrine diseases. None of the patients or control subjects had been receiving hormonal medications, including contraceptive pills, for the last 6 months. The study was approved by the hospital ethics committee, and informed consent was obtained from each patient and control subject. Experimental Design
Studies were performed during the follicular phase, between days 5 and 10 of the menstrual cycle, or during amenorrhea after excluding gestation by appropriate testing. After undergoing an abdominal and pelvic ultrasonogram (US), the patients reported to the endocrine-metabolic testing room between 8:00 A.M. and 9:00 A.M., having undertaken a 12-hour overnight fast. An indwelling IV line was placed in a forearm vein and, after 15-30 minutes, basal blood samples were obtained for the measurement of leptin, cortisol (F), total T, LH, FSH, Es, sex hormone-binding globulin (SHBG), glucose, and insulin levels. Immediately after the basal samples were taken, a 250~pg lV bolus of l-24 ACTH (Synacthen; CibaGeigy, Basel, Switzerland) was injected and blood samples were obtained at 0 and 60 minutes. After the 60-minute sample of the ACTH test was obtained, a 100~pg IV bolus of GnRH (Luforan; Serono, Madrid, Spain) was injected and blood samples for LH and FSH were obtained at 0 and 30 minutes. A single dose (3.75 mg IM) of triptorelin (D-Trp’GnRH, Decapeptyl; LASA-Ipsen, Barcelona, Spain) then was administered. Basal sampling and the ACTH test were repeated on the next day (day 1) and, together with a GnRH test, 21 days after triptorelin administration (day 2 l), when the patients were in gonadal axis stimulation and suppression states, respectively (20). To avoid changes in body weight, the patients were inEscobar-Morreale
et al.
Circulating
leptin
in hirsutism
899
strutted to maintain their usual caloric intake during the 21 days after triptorelin administration. Basal E2 and basal and GnRH-stimulated LH and FSH levels were used to confirm the stimulation and suppression of the gonadal axis after triptorelin administration. In nine randomly selected hirsute women, insulin sensitivity was determined in vivo 1 week before triptorelin administration, using a modified endogenous insulin secretion-suppression test (21). In brief, this test consisted of a constant infusion for 150 minutes of glucose, insulin, and somatostatin. Insulin sensitivity was estimated by the steady-state plasma glucose concentrations during the last 60 minutes of the test. The insulin sensitivity index (in dL/kg per minute) was calculated by the following formula: (glucose infusion rate/ steady-state plasma glucose) x 103, as previously described (21). In control women, triptorelin was not administered because they considered ovarian suppression unacceptable. Some of these women were included in the control groups of previous publications (22, 23). The basal and ACTH-stimulated adrenal steroid profiles at baseline and during ovarian suppression of some of the patients have been published previously (22, 23). Thus, the results of the ACTH test will not be described further here. All blood samples were centrifuged immediately and serum was separated and frozen at -20°C for this and future studies until assayed. Assays Serum leptin concentrations were measured by direct RIA (Linco Research Inc., St. Charles, MO), with intraassay and interassay coefficients of variation (CVs) of 6.5% and 4.5%, respectively. Insulin was measured by direct RIA with standards calibrated against the World Health Organization 66/304 international standard (SORIN Biomedica S.p.A., Saluggia, Italy), with intraassay and interassay CVs of 2.5% and 10.9%, respectively. Plasma glucose was assayed by the glucose oxidase method (Beckman Glucose Analyzer 2; Beckman Instruments, Fullerton, CA). Cortisol, T, and SHBG were assayed as reported previously (22, 23) and the free androgen index was calculated using the following formula: [T (nmol&) x lOOl/SHBG (nmol/L). Luteinizing hormone, FSH, and E, were assayed by a commercial chemiluminescent enzyme immunoassay (Immulite; Diagnostic Products Corporation, Los Angeles, CA), with intraassay and interassay assay 999
Escobar-Morreale et al.
Circuhtin~ httn in h.ir~tkm
CVs of 5.7% and 9.6% for LH, 6.4% and 7.5% for FSH, and 9.3% and 10.6% for Ez, respectively. The molar glucose/insulin ratio was calculated by the following formula: [glucose (mg/dL) x 0.055611 [insulin (@/mL)
X
7.1751,
and was expressed in a 10’ order of magnitude (mmol/pmol). All hormone determinations for each individual were measured in duplicate within a single assay. Ultrasonography Transparietal abdominal and pelvic US was performed in every patient using a Toshiba 3.75-MHz transducer (Tosbee SSA-240; Toshiba Medical Systems, Tokyo, Japan), as previously described (22,23). Statistical Analysis Results are expressed as means ? SE. Repeated measures multivariate analysis of variance (ANOVA) was used to evaluate the evolution of each patient’s hormonal variables during sequential gonadal stimulation and suppression. Paired t-tests were used to identify these differences only when the result of the repeated measures multivariate ANOVA was statistically significant. Repeated measures multivariate two-way ANOVA also was used to compare obese and lean patients and was followed by one-way ANOVA when significant changes were found. One-way ANOVA followed by Dunnet’s post hoc procedure was used to compare patients with control subjects. Analysis of covariance was used to evaluate the influence of obesity, measured by the BMI, on the differences in serum leptin concentrations between patients and control subjects. The relation between the different variables were evaluated by correlation analysis, and Pearson’s coefficient was calculated. Partial correlation analysis, introducing the BMI as the control variable, was used to remove the influence of obesity on the correlations observed. P co.05 was considered to be statistically significant. RESULTS Evolution of Sex Steroids and Gonadotropins During Gonadal Stimulation and Suppression
The administration of triptorelin resulted in an increase in basal LH, FSH, and E2 levels by day 1, which was followed by a marked suppression of these Fertility
and Sterility@
hormones by day 21 (Fig. 1). Compared with baseline, GnRH-stimulated LH and FSH levels were reduced markedly by day 21 (LH, 27.1 2 5.4 versus 3.3 f 0.3 mIU/mL; FSH, 8.1 + 0.6 versus 2.2 2 0.3 mIU/mL; P
1
90
124
ID
I
425 405 1
I’
T
_II 1 Wb' I
323
I
1
AmO.81,P-3.42,P.zO.O~O I
torelin in suppressing gonadotropin secretion. The repeated measures multivariate ANOVA showed significant changes in T and SHBG levels and in the free androgen index after triptorelin administration (Fig. 1, Table 11,which consisted of the following: 1. During gonadal stimulation (day l), the free androgen index increased and the SHBG level decreased, with respect to baseline, whereas the T level did not change (Fig. 1, Table 1). To rule out a possible influence of the acute ACTH stimulation test performed at baseline on the hormonal changes observed by day 1, basal F concentrations were measured and showed no increase with respect to baseline (baseline, 15.9 t 1.2pg/dL versus day 1, 14.6 + 1.1 &dL; paired t-test = 1.18, nonsigni6cant). 2. During gonadal suppression (day 211, the T level and the free androgen index decreased with respect to both baseline and day 1 and the SHBG level remained similar to baseline (Fig. 1, Table 1). The BMI did not change throughout the study (baseline: 25.5 + 0.9 versus day 21, 25.4 2 0.9, in kg/m’). The repeated measures multivariateANOVA showed no significant differences in insulin and in the glucose/ insulin ratio between baseline, day 1, and day 21 (Table 1). Finally, when comparing the group of hirsute patients with the control subjects,the T level was elevated at baseline and day 1 and the free androgen index was elevated and the SHBG level was decreased at baseline, day 1, and day 21, whereas the insulin level was elevated only at baseline (Table 1). Serum Leptin Concentrations
I
I
I
50
1
*
Bl
21 Day of measurement
Figure 1 Changes in basal serum leptin T, SHBG, Ez, LH, and FSH concentrations and in the free androgen index at baseline (B), day 1 (I), and day 21(21) after IM administration of a single 3.75mg dose of triptorelin. When the multivariate repeated measures ANOVA (included in the panels) resulted in statistically significant differences between the different days of measurement, repeated paired t-tests were applied to identify these differences. *At least P < 0.05 with respect to baseline. TAt least P < 0.05 with respect to day 1. .$At least P < 0.05 with respect to baseline and day 1. Vol. 68, No. 5, November
1997
Serum leptin concentrations were elevated in hirsute women compared with control subjects at baseline and on day 1, but not on day 21 (Table 1). Serum leptin concentrations also were different in hirsute patients depending on the day that the concentrations were measured (Fig. 1, Table 1); serum leptin levels increased by day 1 with respect to baseline and returned to baseline levels by day 21 (Fig. 1, Table 1). The serum leptin concentrations of the patients did not vary according to the presence of polycystic ovaries on US, which were found in 16 patients or to the presence of oligomenorrhea or amenorrhea, which occurred in 15 and 1 patients, respectively (data not shown). Moreover, the 19 hirsute patients with functional ovarian or adrenal hyperandrogenism (22) had leptin concentrations similar to those of the 14 patients in whom serum androgen levels were within the normal range (data not shown). Correlation of Circulating Levels of Leptin With Other Variables
Serum leptin levels showed a strong correlation with BMI when both patients and control subjects Escobar-Morreale
et al.
Circulating
leptin
in hirsutism
901
Table 1 Comparison of T, Insulin, SHBG, and Serum Leptin Concentrations, Free Androgen Index, and Glucose/Insulin Ratio of Hirsute Patients With the Reference Values Derived From the Control Group Controls (n = 111 Leptin (ng/mL) T (ng/mL) Free androgen index SHBG ( pg/mLl Insulin ( @/mLl Glucose/insulin ratio
(X
10’)
9.6 38.0 2.7 542 6.9 0.089
5 5 2 2 2 2
1.4 4.3 0.7 63 0.7 0.009
Patients (n = 331 Day 1
Baseline 17.8 70.3 8.2 377 12.4 0.062
+ 2 5 2 c ?
20.3 75.5 9.7 351 11.2 0.066
2.0* 5.1* 1.33* 33” 1.3* 0.007
Day 21
2 2.3* 2 5.5* -c 1.4* 2 34* k 1.0 5 0.008
17.5 52.2 6.2 366 10.9 0.065
2 z t k lr 2
2.3 4.8 l.O* 33* 1.0 0.005
* At least P < 0.05 compared with the control group by one-way ANOVA and Dunnet’s test for comparison with a control group.
were considered (r = 0.76, P < 0.001). The correlation of serum leptin concentrations with hormonal and metabolic variables was evaluated, including the values of the patients at baseline and on days 1 and 21, together with the baseline values obtained from the control subjects. Serum leptin concentrations showed a significant correlation with SHBG levels (r = -0.52, P < O.OOl),the free androgen index (r = 0.38, P < O.OOl), insulin levels (r = 0.46, P < O.OOl),and the glucose/insulin ratio (r = -0.38, P < O.OOl>,but not with T or E2 levels (Fig. 2). The increase in serum leptin levels that occurred in response to triptorelin administration, expressed as the percent increase with respect to baseline values, correlated with the percent increases in the free androgen index and T levels (r = 0.35, P < 0.05 and r = 0.39, P < 0.03, respectively) but not with the percent increases in EP, LH, or FSH levels, or with the percent decrease in SHBG levels. When the control subjects were studied separately, the only correlation that persisted was the one between serum leptin levels and the free androgen index (r = 0.65, P < 0.05). Although the insulin secretion-suppression test was performed in a small number of patients (n = 9; insulin sensitivity index 54.1 t 8.9 dL/kg per minute, range, 20.6-107.6 dL/kg per minute), serum leptin levels showed a strong inverse correlation with the insulin sensitivity index (r = -0.70, P < 0.040). The insulin sensitivity index also presented near-significant correlations with SHBG levels (r = 0.62, P = 0.077) and with the free androgen index (r = -0.61, P = 0.079). No correlations were found between the insulin sensitivity index and baseline insulin levels or the glucose/insulin ratio (data not shown). Influence of Obesity on Serum Leptin and Sex Steroid Concentrations and on the Correlations Between Serum Leptin Concentrations and Other Variables
Although there were no significant differences in BMI between patients and control subjects (25.5 5 902
Escobar-Morreale et al.
CirculatingZeptin in hirsutism
0.9 [range, 17.0-37.21 versus 22.8 ? 0.6 [range, 20.2-27.3, in kg/m’]; F = 2.89, not significant), when correcting for the influence of BMI by analysis of covariance only a near-significant tendency (P = 0.065) to higher leptin levels in hirsute women compared with control subjects persisted at baseline, whereas the increase in serum leptin by day 1 remained significant (P < 0.02). When comparing obese and lean hirsute women, using a cutoff BMI value of 25 (kg/m2), obese patients showed higher
200
,
I
I
, 600
I
I
lb
I
0
20
40
60
:6
6
2;
4b
6b
&
Figure 2 Correlations of circulating leptin levels with serum SHBG, insulin, T and Es concentrations, and with the free androgen index and the glucose/insulin ratio, including the values from patients at baseline, day 1 and 21, and the values from control subjects. The coefficients of correlation and their statistical significance are described in the text. Fertility and Sterility@
leptin concentrations at baseline, day 1, and day 21 than did lean patients (Table 2). Insulin levels also were higher in obese patients at baseline, day 1, and day 21 compared with lean patients, whereas no differences in the glucose/insulin ratio were observed (Table 2). Despite the fact that T levels were the same in obese and lean hirsute women, SHBG levels were reduced markedly at baseline, day 1, and day 21 and the free androgen index was increased at baseline and day 1 in obese patients compared with lean hir-
Table 2
Comparison
sute women (Table 2). Serum LH, FSH, and E2 concentrations were not different between obese and lean hirsute women (Table 2). Moreover, the stimulator-y effect of triptorelin on the gonadal axis and on serum leptin levels, observed by day 1, did not appear to be influenced by obesity because no differences were found between obese and lean patients in the percent increases in serum leptin, T, LH, FSH, and Ez levels and in the free androgen index, or in the percent decrease in SHBG levels, compared with baseline values (Table 2).
of Obese and Lean Hirsute Patients
Value
Obese patients (n = 14)
BMI (kg/m’) Ferriman-Gallwey score Leptin (ng/mL) Baseline Day 1 (% increase)? Day 21 T (ng/mL) Baseline Day 1 (o/o increase) Day 21 Free androgen index Baseline Day 1 (% increase) Day 21 SHBG ( /&dL) Baseline Day 1 (% increase) Day 21 Insulin ( pU/mL)fi Baseline Day 1 Day 21 Glucose/insulin ratio (X log)111 Baseline Day 1 Day 21 LH (mIU/mL)II Baseline Day 1 (% increase) Day 21 FSH (mIU/mL)/l Baseline Day 1 (% increase) Day 21 Ez (pg/mL)Il Baseline Day 1 (% increase) Day 21
Lean patients (n = 19)
30.2 2 1.2 16.6 + 1.7
22.1 2 0.5* 14.8 t 1.0
27.0 -c 3.0 31.2 2 3.68 (17.4 2 6.2) 26.9 5 4.15
11.0 2 1.2* 12.2 -c 1.2** (13.4 t 4.5) 10.5 2 1.1*
71.2 2 10.6 78.3 2 10.2 (24.1 + 18.9) 51.6 lr 8.2
69.6 t 4.4 73.4 t 6.2 52.6 2 5.8
11.3 2 2.8 13.0 t 2.6 8.4 t l.S$
(39.5 + 20.8)
266 2 36 252 t 39 271 5 44
(-6.0
(11.7 ? 6.7)
5.9 2 0.8* 7.3 t 1.4* (18.7 2 8.6) 4.6 -c 0.9$§ 459 2 43* 423 -e 44*$ 436 2 42*
? 3.9)
14.5 ? 1.6 14.8 + 1.5 13.8 2 2.0
10.8 2 1.9* 8.5 2 0.8* 8.8 2 0.8*
0.050 * 0.009 0.048 -c 0.007 0.054 -+ 0.007
0.071 * 0.009 0.079 2 0.011 0.073 t 0.008
(-8.2
2 3.3)
4.0 2 0.7 27.1 2 4.5 1.3 2 0.2
(777 2 172)
6.0 t 1.0 44.5 c 8.9 1.8 t 0.2
(674 2 75)
4.5 2 0.4 11.8 2 1.2 1.6 2 0.2
(164 + 21)
4.7 +- 0.4 14.1 + 1.4 1.5 * 0.3
(392 2 191)
37 ? 5 136 + 17 15 -+ 2
(305 2 52)
45 ? 6 185 ? 26 15 * 2
(381 ? 48)
Note: A cutoff BMI value of 25 (kg/m’) was used to define obese and lean patients. Data were analyzed by multivariate repeated measures two-way ANOVA, including the day of measurement as the within-subjects factor and being obese or lean as the betweensubjects factor. When statistically significant differences were present depending on the day of measurement, these differences were identified by paired t-test. The differences between obese and lean patients were identified further by one-way ANOVA. * At least P < 0.05 with respect to the obese group. t Values in parentheses are the percent increase or decrease at day 1 with respect to baseline, in response to triptorelin administration. No differences between obese and lean patients were observed in these percent changes for any of the variables studied. t At least P < 0.05 with resnect to baseline. g At least P < 0.05 with respect to day 1. /IBecause no significant differences between obese and lean patients were observed, the changes depending on the day of measurement were evaluated including the whole group of patients, and the significance of the changes in LH, FSH, and Ez levels is included in Figure 1. fi No statistically significant differences were observed depending on the day of measurement, and thus the percent changes with respect to baseline were not calculated. Vol. 68, No. 5, November
1997
Escobar-Morreale
et al.
Circulating
leptin in hirsutism
903
Finally, after controlling for the influence of obesity, as expressed by the BMI, the partial correlation analysis showed that the only correlations that remained significant were those of serum leptin levels with insulin (r = 0.24, P < 0.02) and with the glucose/insulin ratio (r = -0.24, P < 0.02). In contrast, the correlations of serum leptin with SHBG levels, the free androgen index, and the insulin sensitivity index were no longer significant, suggesting that these correlations actually resulted from the much stronger correlation between serum leptin levels and the BMI. DISCUSSION
There is increasing evidence that the metabolic effects of leptin are not restricted to the control of body weight and fat composition. As stated previously, leptin stimulates reproductive function, affording an explanation for the influence of nutritional status on the gonadal axis (3, 14). In addition to its physiologic effects, leptin also might play a role in pathologic abnormalities of the reproductive system and might be the link between obesity and hyperandrogenism. If the latter hypothesis is correct, leptin should be increased in hirsute women, and this increase should be related not only to obesity but also to the degree of hyperandrogenism. In our series, serum leptin concentrations were increased in hirsute patients compared with control subjects, and detailed study of the patients disclosed that this increase was present mostly in obese patients. The increase in leptin levels appeared to be influenced by insulin sensitivity and also to be related to the degree of hyperandrogenism for the following reasons. First, serum leptin levels still showed a direct correlation with fasting insulinemia, indicating a direct correlation with the degree of insulin resistance and an inverse correlation with the glucose/insulin ratio, which is an index of insulin sensitivity. Second, serum leptin levels in both patients and control subjects correlated with the free androgen index but not with T and E2 levels. This might suggest that the relation between leptin and the free androgen index is mediated by SHBG, which in turn depends on insulin action and/ or insulin sensitivity, closing the loop between obesity, insulin resistance, and hirsutism. However, the correlation between the free androgen index and leptin also was present when the control group was analyzed separately, whereas leptin did not correlate with SHBG in these subjects. Thus, a direct relation between leptin and the bioavailability of T to the pilosebaceous unit could not be excluded completely. It is surprising that serum leptin levels increased 904
Escobar-Morreale
et al.
Circulating Zeptin in hirsutisn
further by day 1, and this increase appeared to be related to the acute stimulatory effect of triptorelin on the gonadal axis. Because glucocorticoids increase leptin messenger RNA expression and leptin secretion (24), it also would appear that the increase in serum leptin levels on day 1 might have been related to an increase in glucocorticoid secretion resulting from the ACTH test performed the day before. However, the lack of an increase in F levels in our patients suggests that the stimulatory effect on the adrenal of the ACTH test performed at baseline did not persist until day 1. Moreover, two recent studies showed that serum leptin levels do not increase during acute adrenal stimulation. In fact, circulating leptin showed a minimum decrease after IV ACTH administration in patients with normal adrenal function, and this decrease was independent of BMI or sex (25). In the second study, serum leptin did not increase after the administration of IV ovine corticotropin-releasing hormone, despite marked increases in ACTH and F levels, either in patients with obesity and chronic hypercortisolism due to Cushing’s disease or in lean, normal control subjects (26). To our knowledge, the possibility that the reproductive system modulates leptin concentrations has not been explored previously. Our present results suggest that, in addition to the influences of leptin on the gonadal axis, the gonadal axis itself might upregulate leptin secretion acutely, in what could be considered an adipogonadal axis. This is supported by the finding of a statistically significant correlation between the acute percent increase in circulating leptin levels after triptorelin administration and the corresponding percent increases in the free androgen index and in T levels. This correlation further suggests that leptin is involved in the development of hyperandrogenism in hirsute women, or at least that it is regulated by the same factors responsible for their hyperandrogenemia. The percent increase in leptin showed no correlation with the corresponding increase in Ez , as would have been expected from the previously reported finding of higher leptin levels in premenopausal women compared with postmenopausal women and with men (27). This lack of a relation between leptin and E, also might explain why serum leptin levels only returned to baseline 21 days after triptorelin administration, instead of showing a greater decrease. In contrast to Ea, the decrease in circulating androgens was much less pronounced during gonadal suppression, further suggesting that, at least in hirsute women, leptin levels correlate better with androgens than with estrogens. As expected, obesity explained the increase in circulating leptin levels observed in hirsute women Fertility and Sterility@
and, with the exception of the correlation of leptin with insulin and the glucose/insulin ratio, obesity justified all the significant correlations observed between leptin and other hormones. Moreover, in addition to higher leptin concentrations, obese hirsute women had higher insulin levels, a higher free androgen index, and lower SHBG levels compared with lean patients. Thus, our experimental design does not permit us to conclude whether the increase in serum leptin is involved directly in the development of hirsutism or merely reflects the presence of obesity in these women, without any pathophysiologic significance. However, the changes in serum leptin levels that occurred in response to triptorelin administration, paralleling the changes in serum androgen levels, were not different between obese and lean patients, providing indirect evidence of a certain relation between the gonadal axis and leptin metabolism, with possible biologic significance. To date, studies regarding the influence of obesity on serum leptin concentrations in women with PCOS have produced conflicting results. Brzechffa et al. (15) found an increase in serum leptin concentrations in a selected population of hirsute women with PCOS, which was greater than expected for their obesity and insulin sensitivity. In contrast, several recent studies (16-19) have shown that the increased circulating leptin levels found in women with PCOS, as well as the correlations of serum leptin levels with serum androgen and SHBG levels (19), are related primarily to obesity, in agreement with the data presented here. In contrast to previous studies, our series also includes hirsute women without polycystic ovaries on US or menstrual disturbances, in addition to patients with PCOS. This result suggests that the relation between obesity, leptin, and hyperandrogenism is not limited to patients with complete PCOS and also might be involved in milder forms of hirsutism. In conclusion, serum leptin levels are increased in hirsute women with or without PCOS. This increase is dependent mostly on obesity, and also is related independently to the insulin resistance present in hirsute women. However, the correlation of the increase in serum leptin levels with the increase in the free androgen index and in T levels in response to triptorelin administration also suggests that leptin might play a role in the development of hyperandrogenism. Moreover, the increase in circulating leptin levels during gonadal stimulation and the tendency to normalization of serum leptin levels during gonadal suppression suggest that serum leptin also is influenced by the gonadal axis. Further studies are needed to determine whether an adipogonadal axis Vol. 68, No. 5, November
1997
exists, which might explain the frequent association of changes in body weight and reproductive abnormalities. Acknowledgments. The authors thank Mrs. Genoveva Gonzalez and Mrs. M” Dolores Carreto for their technical assistance. REFERENCES 1. Weigle DS, Bukowski T, Foster D, Holderman S, Kramer JM, Lasser G, et al. Recombinant ob protein reduces feeding and body weight in the oblob mouse. J Clin Invest 1995;96: 2065-70. 2. Coleman DL. Obesity and diabetes: two mutant genes causing diabetes-obesity syndromes in mice. Diabetologia 1978; 14:141-8. 3. Chehab FF, Lim ME, Lu R. Correction of the sterility defect in homozygous obese female mice by treatment with the human recombinant leptin. Nat Genet 1996; 12:318-20. 4. Lee GH, Proenca R, Montez JM, Carroll KM, Darvishzadeh JG, Lee JI, et al. Abnormal splicing of the leptin receptor in diabetic mice. Nature 1996;379:632-5. 5. Tartaglia LA, Dembski M, Weng X, Deng N, Culpepper J, Devos R, et al. Identification and expression cloning of a leptin receptor, OB-R. Cell 1995;83:1263-71. 6. Considine RV, Sinha MK, Heiman ML, Kriauciunas A, Stephens TW, Nyce MR, et al. Serum immunoreactive-leptin concentrations in normal-weight and obese humans. N Engl J Med 1996;334:292-5. 7. Ostlund RE, Yang JW, Klein S, Gingerich R. Relation between plasma leptin concentration and body fat, gender, diet, age, and metabolic covariates. J Clin Endocrinol Metab 1996; 81:3909-13. 8. Kolaczynski JW, Nyce MR, Considine RV, Boden G, Nolan JJ, Henry R, et al. Acute and chronic effects of insulin on leptin production in humans: studies in vivo and in vitro. Diabetes 1996;45:699-701. 9. Kieffer TJ, Heller RS, Habener JF. Leptin receptors expressed on pancreatic beta-cells. Biochem Biophys Res Commun 1996;224:522-7. 10. Dagogo JS, Fanelli C, Paramore D, Brothers J, Landt M. Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes 1996;45:695-8. 11. Segal KR, Landt M, Klein S. Relationship between insulin sensitivity and plasma leptin concentration in lean and obese men. Diabetes 1996;45:988-91. 12. Nolan JJ, Olefsky JM, Nyce MR, Considine RV, Caro JF. Effect of troglitazone on leptin production. Studies in vitro and in human subjects. Diabetes 1996;45:1276-8. 13. Kennedy CG, Mitra J. Body weight and food intake as initiating factors for puberty in the rat. J Physiol 1963;166:40818. 14. Chehab FF, Mounzih K, Lu R, Lim ME. Early onset of reproductive function in normal female mice treated with leptin. Science 1997;275:88-90. 15. Brzechffa PR, Jakimiuc AJ, Agarwal SK, Weitsman SR, Buyalos P, Magoffin DA. Serum immunoreactive leptin concentrations in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1996;81:4166-9. 16. Chapman IA, Wittert GA, Norman F&J.Circulating leptin concentrations in polycystic ovary syndrome: relation to anthropometric and metabolic parameters. Clin Endocrinol1997;46: 175-81. 17. Mantzoros CS, Dunaif A, Flier JF. Leptin concentrations in the polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:1687-91. Escobar-Morreale
et al.
Circulating
leptin in hirsutism
905
18. Laughlin GA, Morales AJ, Yen SSC. Serum leptin levels in women with polycystic ovary syndrome: the role of insulin resistance/hyperinsulinemia. J Clin Endocrinol Metab 1997; 82:1692-6. 19. Rouru J, Anttila L, Koskinen P, Penttila T-A, Irjala K, Huupponen R, et al. Serum leptin concentrations in women with polycystic ovary syndrome. J Clin Endocrinol Metab 1997;82:1697-700. 20. Castelo Branco C, Martinez de Osaba MJ, Martinez S, Fortuny A. Effects of a long-acting gonadotropin-releasing hormone analog on the pituitary-ovarian-adrenal axis in women with severe hirsutism. Metabolism 1996;45:24-7. 21. Piedrola G, Novo E, Serrano-Gotarredona J, EscobarMorreale HF, Villa E, Luna JD, et al. Insulin resistance in patients with recent diagnosis of coronary artery disease. J Hypertens 1996; 14:1477-82. 22. Escobar-Morreale HF, Serrano-Gotarredona J, Garcia-Robles R, Sancho J, Varela C. Lack of an ovarian function influence on the increased adrenal androgen secretion present in women with functional ovarian hyperandrogenism. Fertil Steril 1997;67:654-62.
906
Escobar-Morreale
et al.
Circulating
leptin
in hirsutism
23. Escobar-Morreale HF, Serrano Gotarredona J, Garcia Robles R, Sancho J, Varela C. Mild adrenal and ovarian steroidogenie abnormalities in hirsute women without hyperandrogenemia: does idiopathic hirsutism exist? Metabolism 1997; 46:902-7. 24. Papaspyrou-Rao S, Schneider SH, Petersen RN, Fried SK. Dexamethasone increases leptin expression in humans in vivo. J Clin Endocrinol Metab 1997;82:1635-7. 25. Lahlou N, Panneciere MT, Bouvatier C, Chourrout P, Le Fourn J, Bougneres PF, et al. Decrease in leptin secretion during acute adrenal stimulation: A direct effect of ACTH? Proceedings of the 79th Annual Meeting of the Endocrine Society, Minneapolis, MN, 1997, p 197. 26. Cizza G, Lotsikas A, Chrousos GP. Plasma leptin and cortisol levels in Cushing patients before and after correction of hypercortisolism and in lean controls. Proceedings of the 79th Annual Meeting of the Endocrine Society, Minneapolis, MN, 1997, p 119. 27. Rosenbaum M, Nicolson M, Hirsch J, Heymsfield SB, Gallagher D, Chu F, et al. Effects of gender, body composition, and menopause on plasma concentrations of leptin. J Clin Endocrinol Metab 1996;81:3424-7.
Fertility
and Sterility@