Ciliary neurotrophic factor (CNTF) plays an important role in regulating neuronal growth. Recently, central anorexigenic effects of this cytokine have been characterized. However, peripheral effects on tissues that actively contribute to the regulation of energy homeostasis have not been described. Here, we report direct potent and selective effects of CNTF on growth factor and metabolic signalling intermediates in mouse brown adipocytes. CNTF stimulates STAT3, MAP kinase, Akt, and p70 S6 kinase. We find that, next to mediating Akt and p70 S6 kinase activation, both phosphatidylinositol 3-kinase and protein kinase C are separately acting, main intermediates for inducing mitogen-activated protein (MAP) kinase activation. On a functional level, CNTF enhances beta3-adrenergic induction of uncoupling protein-1. Thus, these results demonstrate direct effects of CNTF on adipose tissue signalling and metabolism and suggest a novel role for this cytokine in the peripheral regulation of energy homeostasis.

Leptin is an important adipocytokine whose main regulative effects on energy metabolism are exerted via activation of signalling pathways in the central nervous system. Another important regulator of energy homeostasis is insulin. The role of direct autocrine leptin effects on adipose tissue and crosstalk with insulin, in particular in the thermogenically active brown adipose tissue, remains unclear. In the present study, we have investigated leptin secretion and interaction with insulin in highly insulin-responsive immortalised mouse brown adipocytes. Leptin was secreted in a differentiation-dependent manner, and acute leptin treatment of mature adipocytes dose- and time-dependently stimulated phosphorylation of STAT3 and MAP kinase. Interestingly, acute pretreatment of fully differentiated brown adipocytes with leptin (100 nM) significantly diminished insulin-induced glucose uptake by approximately 25%. This inhibitory effect was time-dependent and maximal after 60 min of leptin prestimulation. Furthermore, it correlated with a 35% reduction in insulin-stimulated insulin receptor kinase activity after acute leptin pretreatment. Insulin-induced insulin receptor substrate-1 tyrosine phosphorylation and binding to the regulatory subunit p85 of phosphatidylinositol 3-kinase (PI 3-kinase) were diminished by approximately 60% and 40%, respectively. Taken together, this study has demonstrated strong differentiation-dependent leptin secretion in brown adipocytes and PI 3-kinase-mediated negative autocrine effects of this hormone on insulin action. Direct peripheral leptin-insulin crosstalk may play an important role in the regulation of energy homeostasis.

SOCS (suppressor of cytokine signaling)-3 has recently been shown to be an insulin- and tumor necrosis factor (TNF)-alpha-induced negative regulator of insulin signaling. To further clarify a potential involvement of SOCS-3 in the development of insulin resistance, we measured differentiation-dependent SOCS-3 mRNA expression in 3T3-L1 adipocytes and studied its regulation by various hormones known to impair insulin signaling using quantitative real-time RT-PCR. There was a differentiation-dependent downregulation of SOCS-3 mRNA by 50% over the 9 day adipocyte differentiation course. Interestingly, besides insulin and TNF-alpha, chronic treatment of differentiated 3T3-L1 cells with 10 microM isoproterenol for 16 h stimulated SOCS-3 gene expression by about 3.5-fold. Furthermore, isoproterenol stimulated SOCS-3 mRNA expression in a dose-dependent manner with significant activation detectable at concentrations as low as 10 nM isoproterenol. Moreover, a strong 27- and 47-fold activation of SOCS-3 mRNA expression could be seen after 1 h of isoproterenol and GH treatment respectively. The stimulatory effect of isoproterenol could be almost completely reversed by pretreatment of 3T3-L1 cells with the beta-adrenergic antagonist propranolol. Finally, isoproterenol's action could be mimicked by stimulation of G(S)-proteins with cholera toxin and of adenylyl cyclase with forskolin and dibutyryl cAMP. Taken together, our results demonstrate a differentiation-dependent downregulation of SOCS-3 in adipocytes and suggest that SOCS-3 gene expression is stimulated by beta-adrenergic agents via activation of a G(S)-protein-adenylyl cyclase-dependent pathway. As SOCS-3 is a novel inhibitor of insulin signaling, the data support a possible role of this protein as a selectively regulated mediator of catecholamine-induced insulin resistance.

The mechanism by which chronic treatment with HIV (human immunodeficiency virus)-1 protease inhibitors leads to a deterioration of glucose metabolism appears to involve insulin resistance, and may also involve impaired insulin secretion. Here we investigated the long-term effects of HIV-1 protease inhibitors on glucose-stimulated insulin secretion from beta cells and explored whether altered insulin secretion might be related to altered insulin signaling. INS-1 cells were incubated for 48 h with different concentrations of amprenavir, indinavir, nelfinavir, ritonavir or saquinavir, stimulated with 20 mM d-glucose, and insulin determined in the supernatant. To evaluate insulin signaling, cells were stimulated with 100 nM insulin for 2 min, and insulin-receptor substrate (IRS)-1, -2 and Akt phosphorylation determined. Incubation for 48 h with ritonavir, nelfinavir and saquinavir resulted in impaired glucose-induced insulin secretion at 2.5, 5 and 5 μM respectively, whereas amprenavir or indinavir had no effects even at 20 and 100 μM respectively. The impaired insulin secretion by ritonavir, nelfinavir and saquinavir was associated with decreased insulin-stimulated IRS-2 phosphorylation, and, for nelfinavir and saquinavir, with decreased insulin-stimulated IRS-1 and Thr³⁰⁸-Akt phosphorylation. No such effects on signaling were observed with amprenavir or indinavir. In conclusion, certain HIV-1 protease inhibitors, such as ritonavir, nelfinavir and saquinavir, not only induce peripheral insulin resistance, but also impair glucose-stimulated insulin secretion from beta cells. With respect to the long-term effect on beta-cell function there appear to be differences between the protease inhibitors that may be clinically relevant. Finally, these effects on insulin secretion after a 48 h incubation with protease inhibitor were associated with a reduction of the insulin-stimulated phosphorylation of insulin signaling parameters, particularly IRS-2, suggesting that protease inhibitor-induced alterations in the insulin signaling pathway may contribute to the impaired beta-cell function.

Tumor necrosis factor (TNF) alpha-induced adipose-related protein (TIARP) has recently been cloned as a TNFalpha-stimulated protein expressed in adipocytes. Its expression is differentiation-dependent and potentially involved in mediating TNFalpha-induced insulin resistance. To further characterize regulation of TIARP gene expression, 3T3-L1 adipocytes were treated with key hormones modulating insulin sensitivity and influencing adipocyte metabolism, and TIARP gene expression was determined by quantitative real-time RT-PCR. Interestingly, TIARP mRNA expression was stimulated almost 9-fold after 500 ng/ml GH were added for 16 h whereas addition of 10 microM isoproterenol, 100 nM insulin and 100 nM dexamethasone for 16 h significantly decreased TIARP gene expression to between 35 and 50% of control levels. In contrast, angiotensin 2 (10 microM) and triiodothyronine (1 microM) did not have any effect. The stimulatory effect of GH was time- and dose-dependent with stimulation occurring as early as 1 h after effector addition and at concentrations as low as 5 ng/ml GH. Moreover, pharmacological inhibition of Janus kinase 2 and p42/44 mitogen-activated protein kinase reversed the stimulatory effect of GH, suggesting that both signaling molecules are involved in activation of TIARP gene expression by GH. Furthermore, an increase of TIARP mRNA could be completely reversed to control levels by withdrawal of GH for 24 h. Taken together, these results show that TIARP is not only responsive to TNFalpha but also to important other hormones influencing glucose homeostasis and adipocyte metabolism. Thus, this factor may play an integrative role in the pathogenesis of insulin resistance and its link to obesity.

Various cytokines, including tumor necrosis factor (TNF) alpha, growth hormone (GH) and interleukin (IL)-6, induce insulin resistance. Recently, it was demonstrated that induction of suppressor of cytokine signaling (SOCS)-3 by TNFalpha and GH is an important mechanism by which these cytokines impair insulin sensitivity. The current study investigated in 3T3-L1 adipocytes whether TNFalpha and GH also upregulate SOCS-1 and SOCS-6, which have both been shown to inhibit insulin signaling potently, and whether IL-6 might alter synthesis of SOCS-1, -3 and -6. Interestingly, 10 ng/ml TNFalpha, 500 ng/ml GH and 30 ng/ml IL-6 induced SOCS-1 mRNA time-dependently with maximal stimulation detectable after 8 h of TNFalpha and 1 h of GH and IL-6 addition respectively. Furthermore, TNFalpha and GH caused sustained upregulation of SOCS-1 for up to 24 h, whereas stimulation by IL-6 was only transient, with SOCS-1 mRNA returning to basal levels 2 h after effector addition. Induction of SOCS-1 was dose-dependent, and significant stimulation was detectable at concentrations as low as 3 ng/ml TNFalpha, 50 ng/ml GH and 10 ng/ml IL-6. Furthermore, stimulation experiments and studies using pharmacologic inhibitors suggested that the positive effect of TNFalpha, GH and IL-6 on SOCS-1 mRNA is, at least in part, mediated by Janus kinase (Jak) 2. Finally, SOCS-3 expression was dose- and time-dependently induced by IL-6, at least in part via Jak2, but none of the cytokines affected SOCS-6 expression. Taken together, our results show a differential regulation of SOCS mRNA by insulin resistance-inducing hormones, and suggest that SOCS-1, as well as SOCS-3, may be an important intracellular mediator of insulin resistance in fat cells and a potential pharmacologic target for the treatment of impaired insulin sensitivity.

ABSTRACT

A sensitive and specific heterologous radioimmunoassay for FSH-suppressing protein (FSP or follistatin) was applied to ovine plasma. Following a logit–log dose transformation, parallel dose–response lines were observed between purified bovine 35 kDa FSP used as standard and serial dilutions of ewe plasma. Activin-A, inhibin-A and a range of other proteins showed low (<0·5%) cross-reactivity in the assay. Daily variations in the peripheral concentrations of FSP were measured across the ovine oestrous cycle. The peripheral concentrations of plasma FSP in adult ewes revealed a significant (P <0·01) increase (33%) during the luteal phase above follicular phase levels, peaking 10 days after the LH surge. FSP concentrations were determined in arterial and venous plasma from the ovary, head, kidney and liver. A significant (P <0·05) increase across the ovary was detected with no significant differences across the head, liver and kidney.

To investigate the relationship between gonadal FSP and the pituitary, ewes underwent ovariectomy and hypophysectomy. FSP levels rose (100–110%, P <0·01) during the period of surgery for both bilateral ovariectomy and sham ovariectomy, and then decreased significantly (37–44%) at 4–6 h after surgery. A further rise in plasma FSP (180–200% increase above pretreatment levels, P < 0·001) was observed 10–12 h after ovariectomy and sham ovariectomy. FSP levels then returned to preoperative levels during the following 26 h. Plasma FSP levels in long-term ovariectomized and hypophysectomized ewes were not significantly different from preoperative levels.

To determine whether the pattern of plasma FSP seen during the ovariectomy study was due to the effect of the induction of anaesthesia, ewes were treated with sodium thiopentone and halothane or 0·9% (w/v) NaCl by procedures of similar duration to that used during surgery. Both treatments resulted in an elevation of FSP levels (33–62%) over pretreatment values only at the time of induction of anaesthesia. To examine further whether this rise in plasma FSP observed after anaesthesia was due to a stress response and therefore under the control of the pituitary-adrenal axis, ewes were treated with ACTH, dexamethasone or saline only. A further group of sheep were exposed to a barking dog for 10 min. No change in FSP levels compared with pretreatment levels or saline-treated controls were noted following any of these treatments.

It was concluded that (1) FSP is present in the peripheral circulation; (2) ovariectomy contributes to plasma FSP, but no discernible contribution to circulating levels was evident from the head, liver or kidney; (3) plasma FSP levels rose significantly during the luteal phase of the ovine oestrous cycle; and (4) FSP secretion may be associated with a stress response, perhaps related to animal handling and intensive blood sampling through mechanisms that are as yet unclear.

Journal of Endocrinology (1993) 137, 433–443

ABSTRACT

The present study was undertaken to determine (1) whether the rabbit feto-placental unit maintains corpora lutea systematically and/or locally and (2) the interrelationships between conceptus number, luteal weight, luteal progesterone concentrations and serum progesterone levels. Thirty-three does were divided into the following treatment groups: (I) bilaterally pregnant, two ovaries; (II) unilaterally pregnant, two ovaries; (III) bilaterally pregnant, one ovary; (IV) unilaterally pregnant, one ovary, contralateral and (V) unilaterally pregnant, one ovary, ipsilateral. Blood samples were obtained from all rabbits on days 6, 9, 12, 15, 18 and 21 post coitum. Does were killed on day 21, and the percentage of viable fetuses, fetal weights, and luteal weights recorded. Blood samples and corpora lutea were analysed for progesterone.

Serum progesterone levels were similar for all groups until day 9 post coitum. Levels in groups III, IV and V declined significantly between days 9 and 12 following removal of one ovary at day 9. Fetal viability, fetal weights and luteal progesterone concentrations did not differ among any of the groups. Luteal weights did not differ among groups I, III, IV and V, but luteal weights of animals in group II were lower than those of group I (P<0·05). Ratios of viable fetuses to number of corpora lutea ranged from 1:11–10:5. No differences were observed in serum progesterone, luteal weights or luteal progesterone concentrations among animals with two conceptuses and those with seven or more, but serum progesterone levels in does with only one conceptus were lower than those in does with more (P<0·05). These results indicate that the feto-placental unit maintains corpora lutea systemically and that the high rate of pregnancy failure by day 21 in does with only one conceptus is due to the inability of a single conceptus to maintain normal serum progesterone levels even though the corpus luteum weight is not affected.

J. Endocr. (1986) 109, 107–110

Recent studies suggest that high glucose concentrations impair insulin receptor phosphorylation and kinase activation in certain cell models. To examine whether such an effect of glucose can also be demonstrated in vivo, insulin receptor kinase activation was studied in erythrocytes from 11 patients with non-insulin-dependent diabetes (NIDDM), before and after reduction of hyperglycemia (from 14.6+/-1.6 to 6.6+/-0.5 mmol/l fasting plasma glucose within 8.6+/-0.6 days). For the measurement of receptor kinase activation, cells were incubated with insulin (0-400 nmol/l), solubilized and insulin receptors immobilized to microwells coated with anti-insulin receptor antibody. Kinase activity towards insulin receptor substrate-1 and insulin binding were then measured in these wells. Kinase activities (expressed as amol phosphate transferred per min and per fmol insulin binding activity) were similar before (2.4+/-0.4 and 32.2+/-2.0 amol/min per fmol with 0 and 400 nmol/l insulin, respectively) and after improvement of metabolic control (2.4+/-0.5 and 32.0+/-2.3 amol/min per fmol with 0 and 400 nmol/l insulin, respectively). Moreover, activities were also similar in 22 hyperglycemic patients with NIDDM (2.1+/-0.3 and 35.1+/-1.4 amol/min per fmol with 0 and 400 nmol/l insulin, respectively) compared with those in 21 non-diabetic control individuals (2.1+/-0.3 and 34.2+/-1.2 amol/min per fmol with 0 and 400 nmol/l insulin, respectively). We conclude that insulin activation of erythrocyte insulin receptor kinase is not impaired in NIDDM and is not influenced by hyperglycemia.

Obesity is associated with chronic inflammation. Pro-inflammatory adipokines may promote metabolic disorders and cardiovascular morbidity. However, the key mechanisms leading to obesity-related inflammation are poorly understood. The corticosteroid metabolism in adipose tissue plays a crucial role in the pathogenesis of the metabolic syndrome. Both the glucocorticoid receptor (GR) and the mineralocorticoid receptor (MR) mediate corticosteroid action in adipose tissue. The significance of the interplay of these receptors in mediating an inflammatory adipokine response is virtually unexplored. In the present study, we investigated the differential roles of the GR and MR in controlling the key adipose tissue functions including inflammatory adipokine expression and adipogenesis using selective stimulation with receptor agonists, acute receptor knockdown via RNA interference and newly generated knockout adipose cell lines. Selective GR stimulation of white adipocytes with dexamethasone inhibited the expression of interleukin 6 (IL6), monocyte chemoattractant protein-1 (MCP1 or CCL2 as listed in the MGI Database), tumour necrosis factor-α, chemerin and leptin. By contrast, selective MR stimulation with aldosterone promoted the expression of IL6, plasminogen activator inhibitor 1, chemerin and leptin. Furthermore, in the presence of an acute GR knockdown as well as in GR knockout adipocytes, corticosterone increased the gene expression of the pro-inflammatory adipokines IL6 and MCP1. Whereas GR knockout adipocytes displayed a mildly impaired adipogenesis during early differentiation, MR knockout cells completely failed to accumulate lipids. Taken together, our data demonstrate a critical role for the balance between gluco- and mineralocorticoid action in determining adipocyte responses implicated in obesity-associated inflammation and cardiovascular complications.