Leptin Activation of Corticosterone Production in Hepatocytes May Contribute to the Reversal of Obesity and Hyperglycemia in Leptin-Deficient ob/ob Mice

Female C57BL/6J obese (ob/ob) mice and lean littermates were purchased at 4 weeks of age from Jackson Laboratory and housed in a room illuminated daily from 0700 to 1900 (12:12 h light/dark cycle). All animal protocols were approved by the Charles R. Drew University Humane Care and Use of Laboratory Animal Committee. Temperature was maintained at 21 ± 1°C, and humidity was maintained at 55 ± 5%. Mice were caged individually and allowed free access to tap water and standard laboratory food. Body weight and food intake were recorded daily.

Two weeks before the experiment, obese mice and lean littermates were divided randomly into three groups: 1) ob/ob mice treated with saline (n = 9), 2) ob/ob mice treated with leptin (n = 8), and 3) lean mice treated with saline (n = 8). Recombinant murine leptin (Sigma, St. Louis, MO) (1 mg/kg) or the saline vehicle was injected intraperitoneally twice each day (at 0700 and 1900) for 2 weeks. In some mice, at the end of the second week after leptin or saline treatment, an intraperitoneal insulin tolerance test was performed by administration of 1.0 units insulin/kg.

ob/ob mice were treated with a single intraperitoneal injection of STZ (180 mg/kg body wt, freshly dissolved in 10 mmol/l citrate buffer, pH 4.5) (25,26). An equal volume of citrate buffer (pH 4.5) was given to control ob/ob mice that were matched for body weight and food consumption. Glucose concentrations in blood samples collected via the tail vein were measured by the glucose oxidase method (Boehringer Mannheim, Mannheim, Germany). Three weeks later, STZ-treated ob/ob mice showed hyperglycemia and relatively low insulin levels because of pancreatic β-cell damage. STZ-treated ob/ob mice were then subdivided into two groups: STZ-ob/ob mice treated with saline (n = 9) and STZ-ob/ob mice treated with leptin (1 mg/kg i.p.) (n = 7), both twice daily for an additional 2 weeks. Non-STZ-treated ob/ob mice (n = 8) also received vehicle for an additional 2 weeks.

Liver was removed surgically under ether anesthesia and frozen immediately in liquid nitrogen. All tissue samples were stored at −80°C until use. Blood samples were collected from each mouse, kept in an ice bath during processing, and then stored at −80°C until measurement of insulin, corticosterone, and glucose concentrations as described previously (27). Liver corticosterone levels were measured with a corticosterone radioimmunoassay kit for mice (ICN Biomedicals) as reported previously (28).

Hepatocytes were isolated from 10-week-old female ob/ob or db/db mice by a two-step collagenase perfusion method (0.5 mg/ml in Hanks’ balanced salt solution) as described previously (29). Primary hepatocytes were plated at 1 × 10⁶ cells/dish in 3 ml Dulbecco’s modified Eagle’s medium (DMEM)/F-12 with 10% fetal bovine serum and incubated at 37°C for 4 h. Cells were then washed with PBS, and the medium was changed to DMEM/F-12-A without fetal bovine serum. After 24 h, hepatocytes were treated with increasing concentrations (10–1,000 ng/ml) of recombinant mouse leptin for 48 h. A leptin concentration occurring in the serum of obese humans (100 ng/ml) (30) was used to treat hepatocytes for different times. In some experiments, insulin (10 nmol/l) was added to ob/ob hepatocytes in the absence or presence of leptin (10 ng/ml). The medium was changed daily, and fresh hormones were added. Specific assays for 11β-HSD1 mRNA expression and enzyme activity were then performed.

The 11β-HSD1 isozyme expresses stable dehydrogenase activity (converting corticosterone [B] to 11-dehydrocorticosterone [A]), which reflects 11β-HSD1 protein levels and reductase activity in vivo (18,20). Thus, we measured 11β-dehydrogenase activity in liver as previously reported (31,32,33). Briefly, liver was homogenized in Krebs-Ringer buffer solution at 4°C in a Dounce tissue grinder. Protein concentrations of the supernatant were measured with a Bradford assay (Bio-Rad protein assay kit), and the supernatants were diluted to yield uniform protein concentrations. Liver homogenates were incubated with 2 μmol/l [³H]B (specific activity, 90 Ci/mmol; New England Nuclear), 3.4 mmol/l NADP⁺, and Krebs-Ringer buffer solution (containing 0.2% glucose and BSA, pH 7.4) at 37°C for 10 min in a shaking bath. Steroids were separated by thin-layer chromatography in a chloroform-ethanol (9:1) system. The conversion of [³H]B to [³H]A was calculated from the radioactivity in each fraction.

11β-HSD activity was measured in intact isolated hepatocytes in six-well tissue culture dishes as described above by measuring the rate of conversion of [³H]A to [³H]B. [³H]A was synthesized in vitro by incubation of rat kidney NRK-52E cells with 50 μCi [³H]B for 24 h in DMEM at 37°C as previously reported (34). [³H]A was separated from [³H]B by high-performance liquid chromatography with a 24–35% methanol gradient over 55 min at a flow rate of 1.5 ml/min. The medium was then removed and replaced with fresh medium containing 18 nmol/l unlabeled A with 2 nmol/l [³H]A as tracer. An aliquot of medium was removed at intervals between 30 min and 24 h after the addition of steroids. Steroids were separated by thin-layer chromatography, and levels were estimated by scintillation counting. Enzyme activity was expressed as the percent conversion rate of [³H]A to [³H]B. All experiments were carried out on at least three separate occasions in triplicate.

Total RNA was isolated with a single-step extraction method (RNAzol B) (Invitrogen). The cDNA probe for mouse 11β-HSD1 was provided by Dr. K.E. Chapman (35), and that for PEPCK was provided by Dr. Yoo Warren (36). Probes were labeled with [³²P]dCTP by nick translation (Nick Translation System; Life Technologies). Autoradiography was performed by exposure of membranes to X-ray film with an intensifying screen at −70°C for up to 7 days. For densitometric measurements, autoradiographic signals were standardized to signals measured from 18S rRNA (37).

11β-HSD1 mRNA levels in primary hepatocytes were measured using quantitative fluorescent real-time PCR. Total RNA (200 ng) was reverse-transcribed with Superscript II RT (Life Technologies). Reactions in which RNA was omitted served as negative controls. Amplification of each target cDNA was then performed with TaqMan PCR reagent kits in the ABI Prism 7700 Sequence Detection System according to the protocols recommended by the manufacturer. The sequences of 11β-HSD1 primers/probe were as follows (38): forward primer 5′-AAGCAGAGCAATGGCAGCAT3′, reverse primer 5′-GAGCAATCATAGGCT GGGTCAT-3′, and probe 5′-CGTCATCTC CTCCTT GGC TGG GAA-3′. The 18S primer and probe were supplied by PE Biosystems. All reactions were carried out using the following cycling parameters: 55°C for 2 min and 95°C for 10 min, following by 30 cycles of 95°C for 15 s and 60°C for 1 min. Threshold cycle (Ct) readings for each of the unknown samples were then used to calculate the amount of 11β-HSD1 or 18S rRNA relative to the standard.

All data are expressed as the mean ± SE for all of the determinations. Comparisons between the different experimental groups were carried out by an unpaired Student’s t test and ANOVA. Values at P < 0.05 were considered statistically significant.

As shown in Table 1, ob/ob mice had very high plasma corticosterone, insulin, and blood glucose levels. The body weight of ob/ob mice was significantly higher than that of lean controls (P < 0.001). However, the 11β-HSD1 activity in the liver of ob/ob mice was markedly decreased compared with that of lean littermates (15.34 ± 1.1 vs. 31.7 ± 2.1%, P < 0.001) (Fig. 1A). When examining both lean and ob/ob mice together, there was an inverse correlation between hepatic 11β-HSD1 activity and body weight (R² = 0.958). The decrease in enzyme activity was paralleled by 11β-HSD1 mRNA expression, which was decreased to 32% that of lean animals (Fig. 1A and B). In contrast, the level of PEPCK mRNA in liver of ob/ob mice was significantly higher than that in lean mice (P < 0.01) (Fig. 2).

Treatment of ob/ob mice with leptin restored the blood glucose concentration to that of lean mice and reduced serum insulin and corticosterone levels, although the levels were still higher than those in lean mice (Table 1). Body weight loss in ob/ob mice after leptin treatment was only 27.6% that of untreated ob/ob mice (P < 0.01). In contrast, leptin treatment caused a significant increase in 11β-HSD1 activity in liver of ob/ob mice (15.34 ± 1.1 to 24.7 ± 1.6%, P < 0.01), but the levels of this enzyme were lower than those in lean controls (P < 0.05) (Fig. 1A). Similarly, leptin treatment induced a twofold increase in 11β-HSD1 mRNA levels in liver of ob/ob mice compared with that of untreated ob/ob mice (P < 0.01), although it did not restore enzyme levels to those of lean controls (Fig. 1B) (P < 0.05). Northern blot analysis also showed that leptin significantly reduced the hepatic PEPCK mRNA level in ob/ob mice compared with that in untreated ob/ob mice (P < 0.01) (Fig. 2).

Levels of liver corticosterone in lean control, ob/ob, and leptin-treated ob/ob mice are shown in Fig. 3A. Treatment of ob/ob mice with leptin significantly reduced the corticosterone level in liver of ob/ob mice, but did not restore it to normal (P < 0.05).

Results of insulin tolerance tests are shown in Fig. 3B. ob/ob mice showed significant inhibition of the glucose-lowering effect of insulin compared with that of lean controls at all time points after insulin challenge (P < 0.01). Administration of leptin to ob/ob mice restored the glucose-lowering effect of insulin (Fig. 3B).

Blood glucose levels in STZ-treated ob/ob mice were significantly higher than those of untreated ob/ob mice (P < 0.001, Table 1). As expected, serum insulin concentrations in STZ-treated ob/ob mice were reduced by 41% that of ob/ob controls (P < 0.01). The body weight of STZ-treated ob/ob mice was significantly lower than that of untreated ob/ob mice (P < 0.01). In contrast, STZ-treated ob/ob mice had plasma corticosterone concentrations that were higher than those of controls (P < 0.01). Moreover, 11β-HSD1 activity in liver of STZ-treated ob/ob mice increased significantly by 37% that of vehicle-treated ob/ob mice (P < 0.01) (Fig. 4A). Enzyme activity correlated with the increased levels of enzyme mRNA shown in Fig. 4A and B. Leptin treatment of STZ-treated ob/ob mice reversed the elevated blood glucose and plasma corticosterone concentrations to those of controls (Table 1). However, leptin treatment did not significantly reduce serum insulin levels, but it did induce a loss of body weight to a level 26% lower than that in STZ-treated ob/ob mice (P < 0.01, Table 1). In addition, with leptin, hepatic levels of 11β-HSD1 activity and mRNA were markedly higher than levels observed in STZ-treated ob/ob animals (Fig. 4).

In primary culture, hepatocytes from ob/ob mice showed only minimal dehydrogenase activity (<5% conversion of [³H]B to [³H]A in intact cells, even after 24-h incubations). However, 11β-reductase activity was clearly detected with 40–60% conversion of [³H]A to [³H]B. Thus, 11β-HSD function is primarily expressed as a reductase in intact ob/ob mouse hepatocytes. As shown in Fig. 5A, treatment of primary cultures of ob/ob hepatocytes with increasing doses of leptin led to a dose-dependent induction of 11β-HSD1 activity from 20.2 ± 1.1% (control) to 28.4 ± 2.1% at 10 ng/ml leptin, 37.8 ± 2.9% at 100 ng/ml leptin, and 49.3 ± 3.5% at 1,000 ng/ml leptin. At a concentration of leptin that occurs in obese patients in vivo (100 ng/ml), 11β-HSD1 activity in primary cultures of ob/ob hepatocytes increased significantly after treatment with leptin for periods ranging from 18 to 72 h (Fig. 6A). However, in primary cultures of db/db hepatocytes, leptin had no significant effect on 11β-HSD1 activity (Figs. 5C and 6C).

The results of real-time quantitative PCR for 11β-HSD1 mRNA expression in primary hepatocytes treated with 10–1,000 ng/ml leptin are shown in Figs. 5B and 6B. After leptin treatment of ob/ob hepatocytes, the 11β-HSD1 mRNA levels increased two- to sixfold above control levels (Fig. 5B). Leptin (100 ng/ml) treatment for 18–72 h resulted in a one- to fourfold increase in 11β-HSD1 gene expression in ob/ob hepatocytes compared with control levels (Fig. 6B). In contrast, leptin treatment did not have a significant effect on 11β-HSD1 mRNA expression in db/db primary hepatocytes (Figs. 5D and 6D).

In addition, real-time quantitative PCR analysis also demonstrated that insulin (10 nmol/l) reduced 11β-HSD1 mRNA levels in hepatocytes (P < 0.05, Fig. 7). Co-treatment with insulin and leptin attenuated the leptin-mediated changes in 11β-HSD1 mRNA levels in primary hepatocytes (P < 0.01, Fig. 7).

Accumulating evidence has shown that glucocorticoids play a fundamental role in the development of obesity and insulin resistance. In ob/ob mice, removal of adrenal steroids reverses the obese phenotype, and corticosterone replacement results in reappearance of the obesity syndrome (39,40), highlighting the crucial role of glucocorticoids in the pathogenesis of obesity and insulin resistance. Importantly, the role of glucocorticoids in obesity depends not only on circulating glucocorticoid levels, but also on intracellular pre-receptor metabolism, which is regulated by 11β-HSD1. Enhanced expression of 11β-HSD1 in adipose tissue promotes central obesity in obese humans and other obese animals (17,18,19,20,21). However, obese patients have increased metabolic clearance of cortisol and enhanced ACTH-cortisol secretion (9,10), which is believed to be due to impaired 11β-HSD1 activity in liver. Similarly, obese animals also have impaired hepatic 11β-HSD1 activity that may stimulate glucocorticoid-induced obesity and insulin resistance (20,21). These data demonstrate that obesity might be correlated with a selective inhibition of 11β-HSD1 in the liver. Consistent with these findings, we observed that obesity in ob/ob mice was associated with an impairment of hepatic 11β-HSD1 expression. 11β-HSD1 activity in hepatocytes from ob/ob mice was significantly lower than that in hepatocytes from lean mice. This reduction in 11β-HSD1 activity corresponded to elevated serum corticosterone concentrations and body weight, suggesting that impairment of hepatic 11β-HSD1 expression contributes to the development of obesity in ob/ob mice.

In ob/ob mice, genetic leptin deficiency results in glucocorticoid-dependent obesity and diabetes. Treating ob/ob mice with recombinant leptin normalizes circulating corticosterone levels and reverses the obese phenotype (23,24). The present study showed that physiological concentrations of leptin increase 11β-HSD1 activity and mRNA expression at the transcriptional levels in isolated ob/ob mouse hepatocytes through a direct action. However, db/db mice are insensitive to leptin because of a mutation in the leptin receptor (41). Leptin did not influence 11β-HSD1 expression in primary hepatocytes from db/db mice, demonstrating that the effects of leptin on hepatic 11β-HSD1 are mediated by the leptin receptor. Moreover, we also found that treatment of ob/ob mice with leptin restores levels of hepatic 11β-HSD1 expression to normal levels and reverses the increase in plasma corticosterone levels and body weight. Therefore, we conclude that decreased hepatic 11β-HSD1 activity and mRNA expression play crucial roles in the obesity observed in ob/ob mice. The reversal or attenuation of the obese phenotype of ob/ob mice by leptin may be mediated, at least in part, by activation of hepatic 11β-HSD1 expression. Hepatic induction of 11β-HSD1 corresponded to reduced levels of circulating glucocorticoids and body weight, indicating that hepatic 11β-HSD1 may be involved in the response of the glucocorticoid system to exogenous leptin in modulating the obese phenotype of ob/ob mice. The reduced synthesis of intracellular corticosterone via impaired hepatic 11β-HSD1 expression in ob/ob mice may contribute to the leptin deficiency-induced activation of the HPA axis, which is associated with elevated circulating glucocorticoid levels and body weight. This is consistent with several studies that showed that hepatic 11β-HSD1 mediates the negative feedback regulation of the HPA axis by endogenous glucocorticoids (11,21). Lack of hepatic 11β-HSD1 expression in null mice enhances the response of the adrenal glands to ACTH and elevates circulating corticosterone levels (11,33). Similarly, a recent study of obese Zucker rats also indicated that inhibition of hepatic 11β-HSD1 might stimulate the HPA axis to increase circulating glucocorticoids and body weight (20,21). Moreover, activation of the HPA axis in obese patients is thought to be associated with impaired hepatic 11β-HSD1 activity (16,17,18,19). These important findings may clarify how the impairment of hepatic 11β-HSD1 expression contributes to glucocorticoid secretion and ultimately the obese phenotype of ob/ob mice. The present study provides experimental evidence that alteration of hepatic 11β-HSD1 activity and gene expression may be an important component in the development of obesity and diabetes. In addition, enhanced 11β-HSD1 expression in adipose tissue is observed in ob/ob mice (data not shown). This is consistent with studies in obese rats where elevated circulating glucocorticoid levels induce adipose 11β-HSD1 activity (21). Thus, the possibility that facilitation of local glucocorticoid action in adipose tissue is involved in the development of obesity in ob/ob mice must also be considered.

Leptin treatment of ob/ob mice also reversed the elevated serum insulin levels and induction of hepatic 11β-HSD1 expression. Indeed, early studies observed that insulin reduced 11β-HSD1 activity in skin fibroblasts and liver cells (13,42). We recently reported that low circulating levels of insulin induce renal 11β-HSD1 activity and mRNA expression in STZ-treated diabetic rats (37). In the present study, we observed that insulin reduces 11β-HSD1 activity and attenuates the leptin-induced expression of 11β-HSD1 in primary cultures of hepatocytes. Our observation that leptin has a direct effect on 11β-HSD1 indicates not only that leptin has insulin-independent actions, but also that the elevated serum insulin levels in ob/ob mice may be involved in the impairment of hepatic 11β-HSD1. In contrast, reduced serum insulin levels after leptin treatment might contribute to the activation of this enzyme in the liver of ob/ob animals. Our conclusion of both the insulin-independent and insulin-dependent effect of leptin on hepatic 11β-HSD1 is consistent with several studies suggesting that insulin antagonizes the actions of leptin and that some effects of leptin in peripheral tissues can be partly obtained in the setting of reduced circulating insulin levels (43,44).

It has become clear that leptin can improve glucose metabolism in both normal and obese animals and that leptin-induced reduction of blood glucose levels in these animals does not require increased secretion of insulin (43,44,45). Such a concept is supported by data from studies in which leptin acted independently of insulin to restore euglycemia and normalize glucose turnover in insulin-deficient STZ-induced diabetic rats (46,47). In the present study, we observed that leptin treatment of ob/ob and STZ-treated ob/ob mice causes marked elevation of hepatic 11β-HSD1 expression, reduction of body weight and circulating corticosterone levels, and reversal of hyperglycemia within 2 weeks without the use of insulin. Moreover, we also observed that leptin treatment of ob/ob mice increased insulin sensitivity and reduced the expression of PEPCK, a gluconeogenic enzyme that is correlated with the acceleration of gluconeogenic pathways in the ob/ob mouse liver (48). These results provide new evidence that leptin plays insulin-independent roles in the control of glucose metabolism and that modulation of hepatic 11β-HSD1 expression may be involved in the glucose-lowering effects of leptin in obese mice. This is consistent with recent reports that insulin resistance is associated with decreased intracellular cortisol production via impaired hepatic 11β-HSD1 activity and that expression of 11β-HSD1 in liver facilitates the maintenance of glucose homeostasis in obese humans (17,18). Leptin correction of impaired hepatic 11β-HSD1 expression may restore the physiological levels of intracellular corticosterone and thereby promote the restoration of circulating glucocorticoids, which is correlated with the reduction of hepatic corticosterone levels and improvement of insulin resistance in leptin-treated ob/ob mice. Although overexpressing 11β-HSD1 may increase local corticosterone levels, maintaining the physiological activity of intracellular 11β-HSD1 may not elevate tissue glucocorticoid levels, because it reflects the metabolic balance between active corticosterone and inactive 11-dehydrocorticosterone (11,28). This result is supported by recent findings that lean rats have high expression of hepatic 11β-HSD1 compared with obese rats, yet have normal metabolic features, including glucocorticoid homeostasis, insulin sensitivity, and hepatic glucose production (21).

In summary, we demonstrated that impairment of 11β-HSD1 expression by hepatocytes is associated with development of the obese phenotype of ob/ob mice. We also found that physiological concentrations of leptin increase hepatic 11β-HSD1 activity and mRNA expression and that these actions are mediated by the leptin receptor. The liver-specific association of leptin with 11β-HSD1 expression may be involved in the modulation of obesity and insulin resistance in ob/ob mice. This suggests that selective enzyme induction within the liver is a new pathway for understanding the importance of tissue-specific dysregulation of 11β-HSD1 in the development of obesity and diabetes.

T.C.F. was supported by National Institutes of Health Grant DA-00276 and Center of Clinical Research Excellence Grant U54 RR14616-01 (to the Charles R. Drew University of Medicine & Sciences). This work was also supported by a grant from the Research Committee on Disorders of Adrenal Hormones under the auspices of the Ministry of Heath and Welfare of Japan.