OBJECTIVE: The syndrome of polycystic ovaries (PCOS) is a known risk factor for type 2 diabetes. It is not known, however, whether the increase in diabetes risk is related to endocrine abnormalities associated with PCOS such as hyperandrogenemia, or whether it is a consequence of the anthropometric or metabolic alterations frequently observed in PCOS women. DESIGN: Since markers of inflammation are supposed to predict type 2 diabetes, interleukin-6 (IL-6) and C-reactive protein (CRP) in combination with parameters of obesity, insulin resistance and hyperandrogenism were determined in 57 PCOS women and in 20 age-matched healthy controls. In addition, the C-174G IL-6 promoter polymorphism was analyzed as a determinant in influencing IL-6, obesity, and androgen levels in women. RESULTS: Neither CRP nor IL-6 were significantly elevated in lean or obese PCOS women compared with age-matched lean or obese controls. In PCOS patients, variables of body composition (body mass index (BMI), waist to hip ratio, dual-energy X-ray-absorptiometry fat mass) and of insulin resistance were correlated with IL-6 or CRP, while parameters of hyperandogenism were not. Multivariate linear regression analysis revealed that obesity is the dominant force, thus explaining 18% and 24% of the IL-6 or CRP levels, respectively, in PCOS women. No association of IL-6 or BMI to a certain genotype at C-174G could be demonstrated in 50 PCOS patients. The heterozygous GC genotype, however, was associated with lower androstendione levels. Metformin treatment of 9 obese, insulin-resistant PCOS patients over a period of 6 months caused a significant decrease in body weight, body fat mass and total testosterone, but showed no significant decline in IL-6 or CRP concentrations. CONCLUSIONS: In PCOS women, plasma levels of IL-6 and CRP were not increased when compared with age- and BMI-matched controls. BMI was, however, the parameter most strongly related to IL-6 and CRP in PCOS; thus PCOS-related endocrine abnormalities do not appear to activate inflammatory parameters thereby enhancing the risk of diabetes. In PCOS, the type 2 diabetes risk may, therefore, be confined to those with obesity and/or metabolic alterations rather than affecting all women suffering from the syndrome.
M Mohlig, J Spranger, M Osterhoff, M Ristow, AF Pfeiffer, T Schill, HW Schlosser, G Brabant and C Schofl
PA Horn, M Mohlig, M Osterhoff, S Wolter, J Hofmann, C Stocking, W Ostertag, M Wahl, H Schatz and A Pfeiffer
BACKGROUND: Estrogen has been shown to have profound effects on insulin and glucose metabolism in vivo. Indeed, estrogens were recently shown to modulate ion channel and secretory activities in endocrine cells. DESIGN AND METHODS: To investigate whether estrogenic influences are caused by direct effects on pancreatic beta-cells, we equipped INS-1 insulinoma cells with estrogen receptors and monitored insulin content and Ca(2+) fluxes as well as basal and stimulated insulin secretion upon different stimuli including glucose, the Ca(2+) ionophore ionomycin, the Ca(2+) channel agonist BayK8644, the protein kinase C activator TPA, and the adenylate cyclase activator forskolin. RESULTS AND CONCLUSION: Our data reveal that estradiol has no significant direct effect on proliferation rate, insulin content, basal and stimulated insulin output as well as Ca(2+) fluxes of insulin secreting cells in vitro, indicating that in vivo responses to estrogen on insulin and glucose metabolism result from indirect betacytotropic effects.
Matthias Möhlig, Martin O Weickert, Elham Ghadamgahi, Ayman M Arafat, Joachim Spranger, Andreas F H Pfeiffer and Christof Schöfl
Adiposity, insulin resistance (IR), and hyperandrogenism are features of polycystic ovary syndrome (PCOS). Retinol-binding protein 4 (RBP4) secreted from adipose and liver tissues has been linked to IR. The impact of RBP4 on IR in PCOS and its usability to identify women with metabolic syndrome (MS) or impaired glucose tolerance ((IGT) or diabetes) were investigated.
Plasma RBP4 was determined in 115 consecutive PCOS women. Associations with IR, body composition, and hyperandrogenemia were investigated by correlation and multiple linear regression analyses in 110 non-diabetics. Receiver operating characteristic curve analysis was used to evaluate RBP4 as a parameter for identifying MS and IGT or diabetes.
RBP4 increased over tertiles of IR (P=0.009). RBP4 correlated with HOMA %S (R=−0.286, P= 0.002), waist-to-hip ratio (WHR) (R=0.233, P=0.034), and dual energy X-ray absorptiometry (DEXA)-lean body mass (R=0.282, P=0.016) but not with body mass index (BMI), DEXA-total or -trunk fat mass, hsCRP, free testosterone, DHEAS, androstenedione, and 17β-estradiol. Adjusted for age, BMI, smoking, and IGT, the association between RBP4 and HOMA %S remained significant (P=0.032). RBP4 explained 4.6% of the variation in HOMA %S. RBP4 was higher in MS and IGT or diabetes, but its ability to identify these women was low (area under the curve, AUC=0.631, P=0.041 or AUC=0.660, P=0.016).
In PCOS, RBP4 has a small independent impact on IR. It is not correlated with hyperandrogenemia, 17β-estradiol, other adrenal steroids, or with markers of adiposity in general. Furthermore, RBP4 does not appear suitable for screening MS or impaired glucose metabolism (IGT or diabetes).
Matthias Möhlig, Martin O Weickert, Elham Ghadamgadai, Andrea Machlitt, Bettina Pfüller, Ayman M Arafat, Andreas F H Pfeiffer and Christof Schöfl
Objective: Many polycystic ovary syndrome (PCOS) women suffer from adiposity and insulin resistance (IR), which play an important role in the development and maintenance of PCOS. Adipocyte fatty acid-binding protein (A-FABP) is mainly expressed in adipocytes, and circulating A-FABP has been associated with markers of obesity and IR. Thus, as observed with other adipose tissue derived factors, secreted A-FABP might be involved in the pathogenesis of obesity-associated disorders such as PCOS.
Design: Plasma A-FABP concentrations were measured in 102 non-diabetic PCOS women, and associations with markers of obesity, IR, inflammation, and hyperandrogenism were investigated by correlation and multiple linear regression analyses. The effect of lifestyle intervention on A-FABP was studied in a second cohort of 17 obese PCOS women.
Results: A-FABP correlated with body mass index (BMI; R = 0.694, P < 0.001), dual-energy X-ray-absorptiometry (DEXA) fat mass (R = 0.729, P < 0.001), DEXA lean body mass (R = 0.399, P = 0.001), HOMA %S (R = −0.435, P < 0.001), hsCRP (R = 0.355, P = 0.001), and free testosterone (fT; R = 0.230, P = 0.02). Adjusted for age, smoking, and glucose metabolism the association of A-FABP with HOMA %S was still significant (P < 0.001), whereas the associations with fT (P = 0.09) and hsCRP (P = 0.25) were not. Inclusion of BMI into the model abolished the impact of A-FABP on HOMA %S. In BMI-matched PCOS women (n = 20 pairs), neither HOMA %S (P = 0.3) nor fT (P = 0.6) were different despite different A-FABP levels (P < 0.001), and in 17 obese PCOS women undergoing a lifestyle intervention, changes in IR were not paralleled by changes in A-FABP.
Conclusions: Circulating A-FABP was correlated with markers of obesity, but had no major impact on IR, inflammation, or hyperandrogenemia in PCOS women.
M A Arafat, B Otto, H Rochlitz, M Tschöp, V Bähr, M Möhlig, S Diederich, J Spranger and A F H Pfeiffer
Objective: It is well known that i.m. glucagon administration stimulates GH and cortisol release in humans, although the mechanisms are unclear. These effects are similar to those described for ghrelin on somatotroph and corticotroph function. The aim of the present study was to investigate the role of ghrelin in mediating the stimulatory effects of glucagon and to evaluate the effect of glucagon on ghrelin secretion.
Design and methods: We studied the endocrine and metabolic response to i.m. glucagon administration in 24 subjects (14 men, 10 women; age 19–65 years; body mass index, 25.3 ± 1 kg/m2), who were shown to have an intact anterior pituitary function as evaluated before enclosure.
Results: Serum ghrelin concentrations fell significantly at 30, 60, 120 and 180 min after glucagon administration (means ± s.e.m.; baseline, 377.9 ± 34.5 pg/ml; nadir, 294.6 ± 28.3 pg/ml (60 min); P < 0.01). Conversely, i.m. glucagon elicited an increase in GH (baseline, 1.5 ± 0.4 μg/l; peak, 14.2 ± 2.7 μg/l (180 min); P < 0.01) and cortisol concentrations (baseline, 452.6 ± 35.2 nmol/l; peak, 622.1 ± 44 nmol/l (180 min); P < 0.01). The changes in ghrelin concentration at both 120 and 180 min were still significant after correction for glucose and insulin (P < 0.05).
Conclusions: We show that i.m. glucagon decreases ghrelin significantly. Therefore, the already known stimulatory effects of i.m. glucagon on cortisol and GH are not mediated by a change in ghrelin concentrations. The mechanisms underlying the ghrelin suppression after i.m. glucagon are unlikely to include glucose or insulin variations and need to be further elucidated.
Matthias Möhlig, Ayman M Arafat, Martin A Osterhoff, Frank Isken, Martin O Weickert, Joachim Spranger, Andreas F H Pfeiffer and Christof Schöfl
Low circulating testosterone concentrations have been associated with insulin resistance (IR). Androgen action is mediated by the androgen receptor (AR) whose activity is modulated by a polymorphic CAG repeat sequence within exon 1. An interaction between testosterone and CAG repeat length (CAG length) with respect to IR has been described in women.
We investigated such a putative interaction between testosterone and the CAG length with respect to IR in men with normal glucose tolerance.
In 113 non-diabetic men calculated free testosterone, the CAG length, and a multiplicative interaction term were investigated by multiple linear regression analysis for an association with IR, as indicated by homeostasis model assessment (HOMA %S).
In a multivariate regression analysis adjusted for age and body mass index, free testosterone, CAG length, and a multiplicative interaction term were significantly associated with IR (P=0.001, P=0.001, P=0.01 respectively). The model explained 36.6% of the variation of IR and predicted that in carriers with a CAG length of 23, changes in testosterone would only minimally affect IR. For CAG lengths longer than 23, however, an increase in testosterone would improve IR, namely the longer the CAG length, the greater the effect. In contrast, in the case of CAG lengths shorter than 23, the effect of increasing testosterone would be the opposite.
In men, testosterone and the AR CAG repeat length polymorphism interacted with respect to IR. The interpretation of the association between testosterone and IR seems to require consideration of the AR CAG repeat polymorphism.
Christian von Loeffelholz, Matthias Möhlig, Ayman M Arafat, Frank Isken, Joachim Spranger, Knut Mai, Harpal S Randeva, Andreas F H Pfeiffer and Martin O Weickert
To study the association of vaspin with glucose metabolism.
Cross-sectional and intervention study.
Subjects and methods
The association of serum vaspin with metabolic and anthropometric characteristics was investigated in 108 volunteers. Euglycemic–hyperinsulinemic clamps (EHC) were performed in 83 of the participants. Changes of circulating vaspin levels were additionally studied in a crossover study using 300 min EHC with lipid versus saline infusion (n=10).
Neither glucose tolerance status nor insulin sensitivity, both as measured using EHCs and using homeostasis model assessment for insulin resistance (HOMA-IR), was significantly associated with serum vaspin in the cross-sectional study. Furthermore, there was no effect of short-term lipid-induced insulin resistance due to a 300 min intravenous lipid challenge on circulating vaspin. However, circulating vaspin levels were significantly elevated in women using oral contraceptives (OC), both compared to women without OC intake (1.17±0.26 vs 0.52±0.09 ng/ml, P=0.02) and males (1.17±0.26 vs 0.29±0.04 ng/ml, P=0.01). After exclusion of OC using females and stratification according to body mass index (BMI), a significant sexual dimorphism in subjects with a BMI <25 kg/m2 was observed (males 0.21±0.04 ng/ml versus females 0.70±0.16 ng/ml, P=0.009).
Our results support the existence of a sexual dimorphism regarding circulating vaspin. The lack of an association of serum vaspin with HOMA-IR and M value indicates, however, no major role for vaspin concerning insulin sensitivity in nondiabetic humans.
S Döcke, J F Lock, A L Birkenfeld, S Hoppe, S Lieske, A Rieger, N Raschzok, I M Sauer, S Florian, M A Osterhoff, R Heller, K Herrmann, S Lindenmüller, P Horn, M Bauer, M O Weickert, P Neuhaus, M Stockmann, M Möhlig, A F H Pfeiffer and C von Loeffelholz
Adipose tissue-derived factors link non-alcoholic fatty liver disease (NAFLD) with obesity, which has also been reported for circulating chemerin. On the other hand, hepatic chemerin and chemokine-like receptor 1 (CMKLR1) mRNA expression has not yet been studied in an extensively characterized patient collective.
This study was cross-sectional and experimental in design.
Liver tissue samples were harvested from 47 subjects and histologically examined according to the NAFLD activity score (NAS). The concentrations of chemerin and CMKLR1 were measured using semi-quantitative real-time PCR, and the concentration of serum chemerin was measured using ELISA. To evaluate potential effects of chemerin and CMKLR1, cultured primary human hepatocytes (PHHs) were exposed to selected metabolites known to play a role in NAFLD (insulin, glucagon, palmitoic acid, and interleukin-6 (IL6)).
Chemerin and CMKLR1 mRNA levels were elevated in the human liver. Their expression was correlated with the NAS (R 2=0.543; P<0.001 and R 2=0.355; P=0.014 respectively) and was significantly elevated in patients with definite non-alcoholic steatohepatitis (NASH) (P<0.05 respectively). Linear regression analysis confirmed an independent association of liver fibrosis, steatosis, inflammation, and hepatocyte ballooning with hepatic chemerin mRNA expression (P<0.05 respectively). The expression of hepatic chemerin and CMKLR1 was correlated with the measures of obesity (P<0.05). The incubation of PHHs with IL6 significantly increased the expression of CMKLR1 mRNA (P=0.027), while that of chemerin remained unaffected (P>0.05). None of the other metabolites showed an influence (P>0.05).
This is the first study to show that chemerin mRNA expression is significantly elevated in the liver of NASH patients and that CMKLR1 expression is upregulated in liver inflammation, whereby IL6 could play a causal role.