Human Reproduction vol.13 no.3 pp.547 553, 1998 C-peptide and insulin, but not C 19 -steroids, support the predictive value of body mass index on leptin in serum of premenopausal women* Franz Geisthövel 1,3, Astrid Meysing 1 and Georg Brabant 2 1 Institute for Gynaecological Endocrinology and Reproductive Medicine, Kaiser-Joseph-Strasse 168, D-79098 Freiburg and 2 Department of Internal Medicine and Dermatology, Medizinische Hochschule, Hannover, Germany 3 To whom correspondence should be addressed Hyperleptinaemia is known to be positively associated with obesity in females. Therefore, circulating leptin concentrations are predicted by body mass index (BMI). Additional effects of endogeneous C 19 -steroids, sex hormone binding globulin (SHBG), luteinizing hormone (LH), follicle stimulating hormone (FSH), C-peptide and insulin on the predictive value of BMI on serum leptin were investigated in 56 hyperandrogenaemic and/or hyperinsulinaemic and/or obese premenopausal women. Serum concentrations (after an overnight 12 h fast) of leptin, total, free, SHBG, dehydroepiandrosterone sulphate (DHEAS), LH, FSH, and oestradiol as well as serum concentrations of C-peptide and insulin prior to, and 1 h after, an oral 100 mg glucose load (1 h values) were determined by immunoassays. Subjects with regular menstrual cycles were studied in the mid-follicular phase while the remainder were studied at random. Nineteen normotestosteronaemic, normoinsulinaemic, lean and ovulatory volunteers served as controls; in order to determine the effect of different stages of the menstrual cylce, serum concentrations of leptin (and of oestradiol in 12 out of the 19 individuals) were determined at the preovulatory, the mid-luteal and the following mid-follicular phase. Significant differences between the patients versus control were not found possibly because of the heterogeneity in the patient group. Multiple regression indicated a hyperbolic correlation between BMI and leptin concentrations. As expected, BMI was the major determinant responsible for >50% (R 2 0.51) of the elevation of leptin concentrations. The combination of BMI with fasting C-peptide or fasting insulin enhanced the R 2 up to 0.59. The multiple regression with two explaining parameters showed a significant regression coefficient for BMI at the 0.001 level, and for fasting C-peptide and fasting insulin at the 0.01 level, which was as statistically significant as the combination of BMI with the 1 h values of C-peptide and of insulin. In contrast, *Some of the data in this study were presented at the 32nd Jahrestagung der Deutschen Diabetes-Gesellschaft, Lübeck, Germany, May 8 10, 1997, and at the 13th Annual Meeting of the ESHRE, Edinburgh, UK, June 22 25, 1997. total, free, SHBG, free /SHBG ratio, DHEAS and LH/FSH ratio had no effect. Similarly, models with more than two variables did not measurably improve the explained variation. In the control group, leptin concentrations were significantly higher in preovulatory and mid-luteal phases than the two midfollicular phases (P 0.05) and must be considered when determining sampling time. In conclusion, hyperandrogenaemia does not have a predictive value on leptin concentrations in premenopausal subjects but hyperinsulinaemia exerts an effect independent of obesity that is the strongest predictor for elevation of leptin concentrations. Hyperinsulinaemia might contribute to the hyperbolic correlation of circulating leptin in obese patients. Key words: insulin/leptin/menstrual cycle/obesity/pcod Introduction The obesity (ob) gene product, leptin, is an adipocyte-derived 16 kda protein (Zhang et al., 1994; Frederich et al., 1995; Masuzaki et al., 1995). Important target organs of leptin action in the mouse are the choroid plexus and the hypothalamus. These express in high abundance ob-receptor (R) mrna, while an ob-r homologue has been identified in human infant brain (Tartaglia et al., 1995). Leptin is thought to be a classical endocrine factor informing the hypothalamus about the state of stored adipocyte fat (see reviews by Caro et al., 1996; Campfield et al., 1996; Sørensen et al., 1996). Activation of the hypothalamus leptin receptor seems to interfere with the neuropeptide Y receptor in the rat arcuate nucleus (Schwartz et al., 1996), a feedback loop system thought to orchestrate body weight homeostasis (e.g. appetite, food intake, thermogenesis, energy expenditure, fat mass). Two obese mouse mutations in the ob gene (the ob/ob mouse strains) present either a premature stop codon or complete absence of the ob mrna (Zhang et al., 1994), which lead to biological or complete aleptinaemia. In contrast, circulating leptin concentrations in humans are highly correlated with the mass of body fat (Maffei et al., 1995; Dagogo- Jack et al., 1996; Geldszus et al., 1996; Haffner et al., 1996). This paradox obesity in the presence of hyperleptinaemia is observed in another obese mouse mutant (the db/db mouse strain) which is totally leptin resistant due to a central leptin receptor mutation (Chen et al., 1996). In between these models is the diet-induced obesity (DIO) mouse model which appears to be more applicable to human obesity (Collins and Surwit, 1996). It has been proposed (see review by Caro et al., 1996) that leptin resistance in human obesity might be due to European Society for Human Reproduction and Embryology 547
F.Geisthövel, A.Meysing and G.Brabant changes in leptin-induced regulation of food intake and energy consumption generated by an altered transport of the leptin signal to the hypothalamus nuclei. Spliced variants of the transmembrane leptin receptor have been identified (Lee et al., 1996) which are widely distributed in the central nervous system including leptominges, choroid plexus, vascular endothelial cells and hypothalamic nuclei (Mercer et al., 1996; Friedman, personal communication, 1997). Mutations of the receptor (Lee et al., 1996) or an acquired dysfunction might cause central insensitivity to leptin action resulting in a compensatory hyperleptinaemia and in an energy dysbalance in favour of energy uptake versus energy expenditure in obese subjects. Ovarian hyperandrogenaemia, the most common endocrine symptom of the so-called polycystic ovarian disease (PCOD), is frequently associated with obesity and hyperinsulinaemia (Geisthövel et al., 1994, 1996). Recently, it has been proposed that adipocyte leptin secretion is abnormally regulated in some PCOD patients (Brzechffa et al., 1996). The aim of the present study was to investigate whether C 19 -steroids, sex hormone binding globulin (SHBG), luteinizing hormone (LH), follicle stimulating hormone (FSH) as well as C-peptide and insulin have additional effects on the predictive value of body mass index (BMI) on leptin concentration in serum of hyperandrogenaemic and/or hyperinsulinaemic and/or obese premenopausal patients. These compounds were selected as they have been established to be the most important endocrine factors in the evaluation of patients with these conditions. Furthermore, determinations of serum leptin were performed at various phases of the menstrual cycle in the control group; since diurnal changes with a nocturnal rise of leptin have been found in humans (Sinha et al., 1996), cyclic variations in circulating leptin might also exist. It is important to investigate this in order to determinate the appropiate time for blood sampling in the patient group. Materials and methods Subjects The patient group consisted of 56 premenopausal women. Anovulatory cycles (monophasic body temperature curve) were found in nine cases and oligoamenorrhoeic cycle disturbances in 47 cases. The patients had the following criteria: (i) either hyperandrogenaemia and/or hyperinsulinaemia (see below) and/or obesity ( 27 kg/m 2 ) (Hirsch and Leibel, 1991) were present, (ii) progesterone concentrations corresponded to those found in the mid-follicular phase of the control group (see below), and (iii) serum concentrations of 17- OH-progesterone were in the normal range (see below). Body fat distribution was defined as android if the waist-to-hip-ratio (WHR: maximal abdominal and gluteal circumferences) was 0.85 (Evans et al., 1983). The subjects had not taken any drugs during the preceding 3 months. The control group consisted of 19 lean highly selected volunteers. These healthy subjects had a history of three or more regular menstrual cycles (range 28 34 days) in the absence of any drugs; mean progesterone concentrations obtained on days 10 and 6 of the expected menstrual bleeding were 18 nmol/l. Major blood sampling (see below) was performed around days 7 9 of the following second cycle. The mid-follicular phase was hormonally confirmed by serum 548 concentrations of LH 12 IU/l and of progesterone 2.0 nmol/l. The physiological nature of the entire cycle was verified by (i) sonographical visualization of the preovulatory follicle (diameter 18 mm), (ii) preovulatory LH surge 20 IU/l, and (iii) mean progesterone concentrations 18 nmol/l measured on day 4 and 8 following ovulation. A further blood sample was obtained around days 7 9 of the following third cycle. In the volunteers, concentrations of leptin (and of oestradiol in 12 out of the 19 individuals) were additionally determined at the preovulatory and at the mid-luteal phase of the second and at the mid-follicular phase of the third cycle. Procedures and assays After an overnight (12 h) fast, blood samples were obtained between 0800 and 0900 h at 20 min intervals through an indwelling catheter placed in the antecubital vein. The blood samples were centrifuged, and the sera were pooled and stored at 20 C until determinations of leptin, total, free, SHBG, dehydroepiandrosterone sulphate (DHEAS), 17-OH-progesterone, LH, FSH, oestradiol, and progesterone as well as of fasting C-peptide and fasting insulin were performed. The blood sampling was followed by administration of a standard oral glucose load (100 g Dextro O.G.-T.Saft ; Boehringer, Mannheim, Germany), and another single blood sample was obtained 1 h later and centrifuged, and sera were stored at 20 C for determination of stimulated (1 h) C-peptide and 1 h insulin. Leptin was determined by a radioimmunoassay as described in detail elsewhere (Horn et al., 1996). The sensitivity of the assay was ~6 pmol/l, and the inter- and intra-assay coefficients of variation were 8.3%. All other parameters were determined by commercially available immunoassay kits: total (RIA-mat Testosteron ; Byk-Santec Diagnostica, Dietzenbach, Germany), free (Freies Testosteron-RIA ; DPC Biermann GmbH, Bad Nauheim, Germany), SHBG (SHBG ; BioChem ImmunoSystems GmbH, Freiburg, Germany), DHEAS (DHEA-Sulfat-RIA ; DPC Biermann GmbH), 17-OH-progesterone (17α-OH-Progesteron ; DSL Deutschland GmbH, Sinsheim, Germany), LH (LH Axsym ; Abott GmbH Diagnostika, Wiesbaden-Delkenheim, Germany), FSH (FSH Axsym ; Abott GmbH Diagnostika), oestradiol (Östradiol-Immulite ; DPC Biermann GmbH), progesterone (Progesterone Axsym ; Abott GmbH Diagnostika), C-peptide (C-Peptid-Immulite ; DPC Biermann GmbH) and insulin (Insulin IMx ; Abott GmbH Diagnostika). The intra- and inter-assay variations for all assays were 6% and 10%, respectively. Based on the statistical evaluation of the highly selected control group (see above and Statistics), hyperandrogenaemia was defined if total and/or free and/or DHEAS were 2.2 nmol/l and/or 9.4 pmol/l and/or 10.4 µmol/l, respectively; 17- OH-progesterone 0.6 nmol/l was defined as normal; hyperinsulinaemia was diagnosed if fasting insulin and/or 1 h insulin were 100 pmol/l and/or 1100 pmol/l, respectively. Statistics Patients and controls were characterized by mean and 95% confidence limits (CL) for the normally distributed age, BMI, WHR and FSH. Residual skew distributed parameters were logarithmically transformed, and mean and CL were computed; these characteristics were retransformed and presented as geometric mean and 95% CL. Using these data, the multiple regression model (MRM) was performed in order to select variables (total, free, SHBG, free /SHBG ratio, DHEAS, LH/FSH ratio, fasting C-peptide, 1 h C-peptide, fasting insulin, 1 h insulin, BMI) which might predict alterations in circulating leptin, the dependent variable in the MRM. Concentrations of leptin and oestradiol during physiological cycles of the controls were evaluated statistically by performing the
Serum leptin of premenopausal women Table I. Anthropometric and endocrine characteristics in hyperandrogenaemic and/or hyperinsulinaemic and/or obese premenopausal patients and in controls Parameters Patients (n 56) Control (n 19) Mean Factor a Lower Upper Mean Factor a Lower Upper Age (years) b 28 16 40 30 22 38 BMI (kg/m 2 ) b 30 17 41 21 18 24 WHR b 0.96 0.80 1.15 0.70 0.60 0.80 Leptin (pmol/l) 296 4.3 68 1281 74 3.2 23 241 Total (nmol/l) 2.7 2.9 0.9 7.7 1.0 2.1 0.5 2.2 Free (pmol/l) 11.1 2.9 3.8 32 4.9 1.9 2.4 9.4 SHBG (nmol/l) 15.2 6.3 2.4 95 39 3.0 12.9 119 Free /SHBG 8.6 9.3 0.9 79 1.6 4.1 0.003 6.12 DHEAS (µmol/l) 5.9 3.8 1.6 22 4.5 2.3 1.9 10.4 17-OH-Progesterone (nmol/l) 0.3 0.7 0.1 0.59 0.2 0.6 0.1 0.57 LH (IU/l) 9.2 5.7 1.6 52 4.1 2.8 1.5 11.6 FSH (IU/l) b 5.5 2.0 9.0 4.8 1.7 7.9 LH/FSH 1.7 3.7 0.5 6.4 0.9 2.2 0.4 2.0 Oestradiol (pmol/l) 152 3.1 49 472 254 2.5 102 629 Progesterone (nmol/l) 1.0 2.3 0.5 2.0 0.8 2.4 0.3 2.0 C-peptide (pmol/l) 17 2.4 6.9 41 12.6 2.4 5.2 30 1 h C-peptide (pmol/l) 75 2.2 33 168 51 2.4 22 121 Insulin (pmol/l) 69 4.2 16 289 40 2.6 15 100 1 h insulin (pmol/l) 489 6.5 76 3158 356 3.3 101 1100 Values are geometric mean and lower and upper 95% confidence limits for logarithmically transformed factors. a Factor of retransformation. b Original data are used for these normally distributed values (see Statistics). BMI body mass index; WHR waist/hip ratio; SHBG sex hormone binding globulin; DHEAS dehydroepiandrosterone sulphate; LH luteinizing hormone; FSH follicle stimulating hormone. Wilcoxon signed rank test and the Friedman two-way of analysis of variance. P 0.05 was considered to be significant. Results Volunteers with physiological cycle Anthropometric and endocrine characteristics of the midfollicular phase data of 19 volunteers with proven ovulatory cycles are presented in Table I. Cyclic changes of leptin concentrations were observed with significantly elevated concentrations at the preovulatory and the mid-luteal phase compared with the two mid-follicular phases (P 0.05) (Figure 1A); cyclic variations of leptin concentrations paralleled those of circulating oestradiol (Figure 1B). Hyperandrogenaemic and/or hyperinsulinaemic and/or obese subjects Hyperandrogenaemia, hyperinsulinaemia, and obesity were found in 47, 23 and 38 patients, respectively, thereby one symptom was observed in 19 cases, and two or three symptoms were present in 23 and 14 individuals respectively. Anthropometric and endocrine characteristics of the 56 patients are summarized in Table I. Significant differences in mean and 95% CL between patients versus controls were not found, possibly because of the heterogenity in the patient group. Using MRM, a positive hyperbolic correlation between leptin concentrations and BMI was found (Figure 2). There was a wide range of leptin concentrations at any BMI and vice versa; e.g. leptin concentrations of ~375 pmol/l were found at BMI ranging between 19 and 35 kg/m 2 ; however, the leptin concentration of only two patients with BMI 27 kg/m 2 was found to fit within the range of concentrations of the control, and in addition 13 out of 38 obese patients presented relatively low leptin concentrations fitting within the range of concentrations of non-obese patients. Body mass index was the major determinant responsible for 50% (R 2 0.51) of the elevation of leptin concentrations. The combination of BMI with fasting C-peptide or fasting insulin enhanced the R 2 up to 0.59. The MRM with two variables (regression parameters) showed a significant regression coefficient for BMI at the 0.001 level, and for fasting C- peptide and fasting insulin at the 0.01 level which was not additionally increased by the combination of BMI with the 1 h values of C-peptide and insulin. In contrast, total, free, SHBG, free /SHBG ratio, DHEAS and LH/FSH ratio failed to have any effect. Similarly, models with more than two variables did not measurably improve the explained variation (Table II). Estimates of the regression parameters BMI and fasting C-peptide as well as BMI and fasting insulin, the corresponding SE and P-values are presented in Table III; by these statistical means, the prediction of individual leptin concentrations based on the individual values of BMI, fasting insulin and fasting C-peptide can be performed. Discussion This paper suggests that leptin concentrations in premenopausal women are not predicted by hyperandrogenaemia (total, free, free /SHBG ratio and DHEAS including the associated LH/FSH ratio) but by hyperinsulinaemia (fasting and 1 h values of C-peptide and insulin) supporting independently the strongly predictive value of 549
F.Geisthövel, A.Meysing and G.Brabant Table II. Multiple regression models for dependent variable: circulating leptin No. of variables Variables in model a R 2 in model (regression parameters) Figure 1. (A) Cyclic changes of leptin concentrations in 19 normoandrogenaemic, normoinsulinaemic, lean and ovulatory volunteers. Mid-follicular phase 1 and 2 (FP 1, FP 2). Preovulatory (Preov). Postovulatory day 4 and day 8 (Postov 4; Postov 8). Data are expressed as geometric mean ( 34 ) and 95% upper and lower confidence limits (s s) (see Statistics). *P 0.05 compared with mid-follicular phase. (B) Cyclic changes of oestradiol in 12 out of the 19 ovulatory volunteers; see A. 1 BMI 0.505 1 f Ins 0.352 1 f CP 0.300 1 1 h CP 0.182 1 1 h Ins 0.174 2 BMI/fCP 0.592 2 BMI/fIns 0.585 2 BMI/1 h CP 0.554 2 BMI/1 h Ins 0.539 2 BMI/SHBG 0.518 3 BMI/fIns/fCP 0.601 3 BMI/fCP/total 0.597 3 BMI/fCP/free 0.594 3 BMI/fCP/SHBG 0.593 3 BMI/fCP/1 h Ins 0.593 4 BMI/fIns/total /free 0.623 4 BMI/fCP/total /free 0.620 4 BMI/fIns/fCP/total 0.606 4 BMI/fIns/fCP/free 0.603 4 BMI/fIns/fCP/1 h Ins 0.601 5 BMI/fIns/fCP/total /free 0.631 5 BMI/fIns/SHBG/total /free 0.631 5 BMI/fIns/1 h CP/total /free 0.627 5 BMI/fIns/1 h Ins/total /free 0.624 5 BMI/fCP/SHBG/total /free 0.623 fins fasting insulin; fcp fasting C-peptide. a The five leading variables are listed. Table III. Parameter estimates Variable Parameter SE P estimate Figure 2. Correlation obtained by the multiple regression model (MRM) between body mass index (BMI) and serum leptin concentrations in 56 hyperandrogenaemic and/or hyperinsulinaemic and/or obese premenopausal women ( ). Control data (19 volunteers, see Figure 1A) (s) are not used in the MRM. Limit between normal body weight and obesity at 27 kg/m 2 ( ). Limit of maximal serum leptin concentrations in patients 27 kg/m 2 (---). Original data were used for BMI (normal distribution) and logtransformed and retransformed data are used for leptin concentrations (skew distribution). obesity (BMI) on the elevation in circulating leptin. In addition, a cyclic variation of leptin concentrations with enhanced concentrations at the preovulatory and the mid-luteal phase compared with the mid-follicular phase of ovulatory volunteers was observed. 550 Fasting C-peptide (fcp) Intercept 1.365 0.139 0.0001 BMI 0.031 0.005 0.0001 fcp 0.524 0.156 0.0015 Fasting insulin (fins) Intercept 1.289 0.142 0.0001 BMI 0.029 0.005 0.0001 fins 0.329 0.103 0.0024 BMI body mass index; fins fasting insulin; fcp fasting C-peptide Our finding that the BMI is the dominant predictor of elevated leptin concentrations is supported by results of several previous studies (Maffei et al., 1995; Dagago-Jack et al., 1996; Geldszus et al., 1996; Haffner et al., 1996; Rosenbaum et al., 1996) and by studies using percentage body fat mass (Considine et al., 1996; Rosenbaum et al., 1996) or absolute fat mass (Rosenbaum et al., 1996). Consequently, weight loss due to food restriction resulted in a decline in leptin concentrations (Maffei et al., 1995; Considine et al., 1996; Geldszus et al.,
Serum leptin of premenopausal women 1996), and changes in leptin concentrations paralleled changes in weight during the follow-up period (Wing et al., 1996). In agreement with these clinical studies, ob mrna expression was found to be elevated in human adipocytes (Lönnqvist et al., 1995; Considine et al., 1996) thereby larger adipocytes had higher ob expression than smaller adipocytes, supporting the idea that cell stretching itself may be the signal (Hamilton et al., 1995; Collins and Surwit, 1996). The hyperbolic correlation between BMI and circulating leptin found in the present study indicates that another factor in conjunction with BMI, such as insulin, might be involved in leptin elevation. In long-term cultured human fat cells, insulin provoked a rise in ob mrna (Kolaczynski et al., 1996) and in leptin protein production (Kolaczynski et al., 1996; Wabitsch et al., 1996) that was completely abolished by coincubation with the antidiabetic drug, troglitazone (Nolan et al., 1996). Long-term, not short-term exogeneous hyperinsulinaemia in vivo had an effect on circulating leptin in humans (Dagogo-Jack et al., 1996; Kolaczynski et al., 1996). A 58% increase in leptin concentrations in normal individuals after prolonged (8.5 h) exposure to insulin and a rapid decline after insulin withdrawal (Malmström et al., 1996) and during a 52 h fast (Boden et al., 1996) have been described. A positive correlation between leptin and fasting insulin was found in non-obese or diabetic women (Ryan and Elahi, 1996). These data are confirmed by our results in premenopausal women showing for the first time that chronic endogeneous hyperinsulinaemia per se does have a significant impact on the increase in leptin concentrations (hyperinsulinaemia was studied by four parameters: fasting and 1 h values of C-peptide and insulin). These findings are in contrast to those reported by Havel et al. (1996) who did not find a significant correlation in postmenopausal women between leptin and insulin independent of BMI by using a partial regression analysis; the reason for these incompatible results might be the differences in the patient groups investigated in the two studies. The observations by Brzechffa et al. (1996), who did not find a correlation between PCOS patients and insulin sensitivity, might be explained by the fact that their patients were defined by pathomorphological criteria (polycystic ovaries) rather than by endocrine means as in our study. Concerning our results, insulin might be responsible in part for the hyperbolic correlation between leptin and BMI. Consistent with these data, hypoleptinaemia in amenorrhoeic athletes was found in the presence of low insulin concentrations (Laughlin and Yen, 1997). All these data indicate that insulin acts either trophically (Kolaczynky et al., 1996) or directly on adipocyte leptin production. Several studies showed that leptin concentrations were higher in women than in men when the groups were defined by BMI (Maffei et al., 1995; Dagogo-Jack et al., 1996; Considine et al., 1996; Haffner et al., 1996; Rosenbaum et al., 1996); however, when leptin was compared with percentage body fat, there were no significant differences between the genders (Maffei et al., 1995; Considine et al., 1996) probably because of the higher body fat content of women (Pouliot et al., 1994). Nevertheless, leptin concentrations were found to be highest in premenopausal individuals, lower in post- menopausal women and lowest in men suggesting that C 18 - and C 21 -steroids might increase, and C 19 -steroids might decrease, leptin concentrations (Rosenbaum et al., 1996). These results are in agreement with our study in physiological cycles which indicated significantly elevated leptin concentrations at the preovulatory and the mid-luteal phase suggesting that leptin underlies not only diurnal (Sinha et al., 1996) but also cyclic variations in dependency on circulating oestradiol, which might up-regulate circulating leptin. Concerning the effects of C 19 - steroids, one would expect that hyperandrogenaemia is correlated with relative low leptin concentrations. However, there was no predictive value of total, free, free /SHBG ratio and DHEAS and the associated LH/FSH ratio on leptin concentrations in our patients, suggesting that moderate hyperandrogenaemia (of ovarian, adrenal or extraglandular origin) does not influence leptin concentrations. These findings are compatible with the results by Brzechffa et al. (1996) describing a lack of correlation between free and leptin in PCOS patients; our data differ from their results showing that insulin-resistant PCOS women have significantly higher leptin concentrations than insulinresistant controls of similar BMI. Keeping in mind that adipocytes are important sources of extraglandular oestrogens due to their high aromatizing activity (Edman and MacDonald, 1978), it is possible that putative inhibitory effects of moderately elevated androgens on adipocyte-dependent leptin denovo synthesis are in part counteracted in an auto- or paracrine manner by locally converted oestrogens, and that therefore male-like concentrations of C 19 -steroids may be required for leptin-suppressive effects. From these data, it may be assumed that adipocyte-dependent leptin synthesis and secretion are more sensitive to stimulatory effects of C 18 -steroids than to suppressive effects of C 19 -steroids under physiological or moderately dysregulated conditions in premenopausal women. Because of cyclic and diurnal (Sinha et al., 1996) changes of circulating leptin concentrations, strict definitions of blood sampling are required. Although a direct relationship between leptin and ovarian or adrenal androgens was not found in our study it seems likely that fat mass interacts with the hypothalamus pituitary ovarian/adrenal axis. The lack of ob gene expression in female ob/ob-mice is associated not only with increased food intake, reduced energy expenditure and obesity, but also with hypothalamic pituitary deficiency, sexual immaturaty and infertility which can be restored by intraperitoneal administration of leptin (Chebab et al., 1996). In addition, leptin treatment in normal pubertal female mice results in an early onset of reproductive function (Chebab et al., 1997). Hypoleptinaemia was a major finding in amenorrhoeic athletes, and an adverse relationship between serum cortisol and leptin was found in these individuals (Laughlin and Yen, 1997). Because of the adjacent site of leptin target organs and gonadotrophin hormone releasing hormone (GnRH)-secreting neurons in the arcuate nucleus it might be reasonable to propose that leptin is involved in the prepubertal maturation of GnRH biosynthesis and secretion and also in a pathological condition such as amenorrhea in undernutrition. Since the ob protein is found to modulate synaptic transmission in the hypothalamus (Glaum 551
F.Geisthövel, A.Meysing and G.Brabant et al., 1996) it has been postulated that leptin acts at the hypothalamic level as a coordinator within the neural pleiotypic network of behavioural, metabolic and neuroendocrine interactions (Campfield et al., 1996). Therefore, dysregulations of central leptin actions might be in part responsible for the tonic elevation of LH secretion in hyperleptinaemic, leptin-resistant patients. Hyperleptinaemia may play a role in prepubertal onset of PCOD in obese girls. Adipocyte leptin signalling might also be linked directly with ovarian function. An isoform of the leptin receptor was detected in the murine ovary (Cioffi et al., 1996) and in murine granulosa cells of the preovulatory follicle (Cioffi et al., 1996: data not shown). Furthermore, leptin is able to impair the insulin-like growth factor I-mediated augmentation of FSHstimulated C 18 -steroid synthesis by rat granulosa cells (Zachow and Magoffin, 1997). Further studies have to clarify: (i) local regulation of adipocyte leptin biosynthesis, (ii) interactions between leptin and the hypothalamic pituitary ovarian axis, (iii) interactions between leptin and liver function, (iv) whether determinations of leptin will be a clinical tool in the evaluation and control of hyperandrogenaemic and/or hyperinsulinaemic and/or obese patients, and (v) whether there are subjects who benefit in terms of metabolic, endocrine and reproductive disturbances from the administration of genetically synthesized ob protein. Acknowledgements We want to thank Prof. Dr J.Schulte-Mönting (Institute of Statistics and Biometry of the University of Freiburg, Freiburg, Germany) for his intensive statistical work and S.Bruncken (correspondence address) and R.Horn (G.Brabant s address) for their excellent assistance in carrying out this study. The investigation received generous financial support from Ferring Arzneimittel GmbH, Kiel, Germany. References Boden, G., Chen, X., Mozzoli, M. and Ryan, I. (1996) Effect of fasting on serum leptin in normal human subjects. J. Clin. Endocrinol. Metab., 81, 3419 3423. Brzechffa, P.R., Jakimiuk A.J., Agarwal, S.K. et al. (1996) Serum immunoreactive leptin concentrations in women with polycystic ovary syndrome. J. Clin. Endocrinol. Metab., 81, 4166 4169. Caro, J.F., Sinha, M.K., Kolaczynki, J.W. et al. (1996) Leptin: the tale of an obesity gene. Diabetes, 45, 1455 1462. Campfield, L.A., Smith, F.J. and Bum, P. (1996) The ob protein (leptin) pathway a link between adipose tissue mass and central neural networks. Horm. Metab. Res., 28, 619 632. Chehab, F., Lim, M.E. and Lu, R. (1996) Correction of the sterility defect in homozygous obese female mice by treatment with the human recombinant leptin. Nature Genet., 12, 318 320. Chehab, F., Mounzih, K., Lu, R. and Lim, M.E. (1997) Early onset of reproductive function in normal female mice treated with leptin. Science, 275, 88 90. Chen, H., Charlat, O., Tartaglia, L.A. et al. (1996) Evidence that the diabetes gene encodes the leptin receptor: identification of a mutation in the leptin receptor gene in db/db mice. Cell, 84, 491 495. Cioffi, J.A., Shafer, A.W., Zupancic, T.J. et al. (1996) Novel B219/ob receptor isoforms: possible role of leptin in hematopoiesis and reproduction. Nature Med., 2, 585 589. Collins, S. and Surwit, R.S. (1996) Pharmacologic manipulation of ob expression in a dietary model of obesity. J. Biol. Chem., 16, 9437 9440. Considine, R.V., Sinha, M.K., Heiman, M.L. et al. (1996) Serum 552 immunoreactive-leptin concentrations in normal weight and obese humans. N. Engl. J. Med., 334, 292 295. Dagogo-Jack, S., Fanelli, C., Paramore, D. et al. (1996) Plasma leptin and insulin relationships in obese and nonobese humans. Diabetes, 45, 695 698. Edman, C.D. and MacDonald, P.C. (1978) Effect of obesity on conversion of plasma androstendione to estrone in ovulatory and anovulatory young women. Am. J. Obst. Gynecol., 130, 456 461. Evans, D.J., Hoffmann, R.G., Kalkhoff, R.K. and Kissebah, A.H. (1983) Relationship of androgenic activity to body fat topography, fat cell morphology, and metabolic aberrations in premenopausal women. J. Clin. Endocrinol. Metab., 57, 304 310. Frederich, R.C., Löllmann, B., Hamann, A. et al. (1995) Expression of ob mrna and its encoded protein in rodents. J. Clin. Invest., 96, 1658 1663. Geisthövel, F., Olbrich, M., Frorath, B. et al. (1994) Obesity and hypertestosteronaemia are independently and synergistically associated with elevated insulin concentrations and dyslipidaemia in pre-menopausal women. Hum. Reprod., 9, 610 616. Geisthövel, F., Frorath, B. and Brabant, G. (1996) Acarbose reduces elevated serum concentrations in hyperinsulinaemic premenopausal women: a pilot study. Hum. Reprod., 11, 2377 2381. Geldszus, R., Mayr, B., Horn, R. et al. (1996) Serum leptin and weight reduction in female obesity. Eur. J. Endocrinol., 135, 659 662. Glaum, S.R., Hara, M., Bindokas, V. et al. (1996) Leptin, the obese gene product, rapidly modulates synaptic transmission in the hypothalamus. Mol. Pharmacol., 50, 230 235 Haffner, S.M., Stern, M.P., Miettinen, H. et al. (1996) Leptin concentrations in diabetic and nondiabetic Mexican-Americans. Diabetes, 45, 822 824. Hamilton, B.S., Paglia, D., Kwan, A.Y.M. and Deitel, M. (1995) Increased obese mrna expression in omental fat cells from massively obese humans. Nature Med., 1, 953 956. Havel, P.J., Kasim-Karakas, S., Mueller, W. et al. (1996) Relationship of plasma leptin to plasma insulin and adiposity in normal weight and overweight women: effects of dietary fat content and sustained weight loss. J. Clin. Endocrinol. Metab., 81, 4406 4413. Hirsch, J. and Leibel, R.L. (1991) Clinical review 28: A biological basis for human obesity. J. Clin. Endocrinol. Metab., 73, 1153 1157. Horn, R., Geldszus, R., Pötter, E., et al. (1996) Radioimmunoassay for the detection of leptin in human serum. Exp. Clin. Endocrinol. Diabetes, 104, 454 458. Kolaczynski, J.W., Nyce, M.R., Considine R.V. et al. (1996) Acute and chronic effect of insulin on leptin production in humans. Diabetes, 45, 699 701. Laughlin, G.A. and Yen, S.S.C. (1997) Hypoleptinemia in women athletes: absence of a diurnal rhythm with amenorrhea. J. Clin. Endocrinol. Metab., 82, 318 321. Lee, G.H., Proenca, R., Montez, J.M. et al. (1996) Abnormal splicing of the leptin receptor in diabetic mice. Nature, 379, 632 635. Lönnqvist, F., Arner, P., Nordfors, L. and Schalling, M. (1995) Overexpression of the obese (ob) gene in adipose tissue of human obese subjects. Nature Med., 1, 950 953. Maffei, M., Halaas, J., Ravussin, E. et al. (1995) Leptin levels in human and rodent: Measurement of plasma leptin and ob RNA in obese and weightreduced subjects. Nature Med., 1, 1155 1161. Malmström, R., Taskinen, M.-R., Karonen, S.-L. and Yki-Järvinen, H. (1996) Insulin increases plasma leptin concentrations in normal subjects and patients with NIDDM. Diabetologia, 39, 993 996. Masuzaki, H., Ogawa,Y., Isse, N. et al. (1995) Human obese gene expression. Diabetes, 44, 855 858. Mercer, J.G., Hoggard, N., Williams, L.M. et al. (1996) Localization of leptin receptor mrna and the long form splice variant (ob-rb) in mouse hypothalamus and adjacent brain regions by in situ hybridization. FEBS Lett., 387, 113 116. Nolan, J.J., Olefsky, J.M., Nyce, M.R. et al. (1996) Effect of Troglitazone on leptin production. Diabetes, 45, 1276 1278. Pouliot, M.-C., Després, J.-P., Lemieux, S. et al. (1994) Waist circumferences and abdominal sagittal diameter: Best simple anthropometric indexes of abdominal visceral adipose tissue accumulation and related cardiovascular risk in men and women. Am. J. Cardiol., 73, 460 468. Rosenbaum, M., Nicolson, M., Hirsch, J. et al. (1996) Effects of gender, body composition, and menopause on plasma concentrations of leptin. J. Clin. Endocrinol. Metab., 81, 3424 3427. Ryan, A.S. and Elahi, D. (1996) The effects of acute hyperglycaemia and hyperinsulinaemia on plasma leptin levels: its relationship with body fat, visceral adiposity, and age in women. J. Clin. Endocrinol. Metab., 81, 4433 4438.
Serum leptin of premenopausal women Schwartz, M.W., Seeley, R.J., Campfield, L.A. et al. (1996) Identification of targets of leptin action in rat hypothalamus. J. Clin. Invest., 98, 1101 1106. Sinha, M.K., Ohannesian, J.P., Heiman, M.L. et al. (1996) Nocturnal rise of leptin in lean, obese, and non-insulin-dependent diabetes mellitus subjects. J. Clin. Invest., 97, 1344 1347. Sørensen, T.I., Echwald, S.M. and Holm, J.C. (1996) Leptin in obesity. Br. Med. J., 313, 953 954. Tartaglia, L.A., Dembski, M., Weng, X. and Deng, N. (1995) Identification and expression cloning of a leptin receptor, OB-R. Cell, 83, 1263 1271. Wabitsch, M., Jensen, P.B., Blum, W.F. et al. (1996) Insulin and cortisol promote leptin production in cultured human fat cells. Diabetes, 45, 1435 1438. Wing, R.R., Sinha, M.K., Considine, R.V. et al. (1996) Relationship between weight loss maintenance and change in serum leptin levels. Horm. Metab. Res., 28, 698 703. Zachow, R.J. and Magoffin, D.A. (1997) Direct intraovarian effects of leptin: impairment of the synergistic action of insulin-like growth factor-i on follicle-stimulating hormone-dependent estradiol-17β production by rat ovarian granulosa cells. Endocrinology, 138, 847 850. Zhang, Y., Proenca, R., Maffei, M. et al. (1994) Positional cloning of the mouse obese gene and its human homologue. Nature, 372, 425 432. Received on June 18, 1997; accepted on December 11, 1997 553