11β-Hydroxysteroid Dehydrogenase Type 1(11β-HSD1) mediates insulin resistance through JNK activation in adipocytes

Glucocorticoids are used to treat a number of human diseases but often lead to insulin resistance and metabolic syndrome. 11β-hydroxysteroid dehydrogenase type 1 (11β-HSD1) is a key enzyme that catalyzes the intracellular conversion of cortisone to physiologically active cortisol. Despite the known role of 11β-HSD1 and active glucocorticoid in causing insulin resistance, the molecular mechanisms by which insulin resistance is induced remain elusive. The aim of this study is to identify these mechanisms in high fat diet (HFD) experimental models. Mice on a HFD were treated with 11β-HSD1 inhibitor as well as a JNK inhibitor. We then treated 3T3-L1-derived adipocytes with prednisone, a synthetic glucocorticoid, and cells with 11β-HSD1 overexpression to study insulin resistance. Our results show that 11β-HSD1 and JNK inhibition mitigated insulin resistance in HFD mice. Prednisone stimulation or overexpression of 11β-HSD1 also caused JNK activation in cultured adipocytes. Inhibition of 11β-HSD1 blocked the activation of JNK in adipose tissue of HFD mice as well as in cultured adipocytes. Furthermore, prednisone significantly impaired the insulin signaling pathway, and these effects were reversed by 11β-HSD1 and JNK inhibition. Our study demonstrates that glucocorticoid-induced insulin resistance was dependent on 11β-HSD1, resulting in the critical activation of JNK signaling in adipocytes.

Therefore, 11β -HSD1 is an important therapeutic target for reducing adverse effects of prescribed glucocorticoids for treatment of a variety of diseases The mechanism by which increased levels of 11β -HSD1 results in insulin resistance is not fully known. The overexpression of 11β -HSD1 gene in adipose tissue can increase levels of leptin, resistin, tumor necrosis factor-α (TNF-α ), and interleukin-6 (IL-6) 10,11 . This suggests that high local levels of glucocorticoids, at least in adipocytes, promote an inflammatory rather than the expected anti-inflammatory activity through cortisol. This inflammatory function of glucocorticoids may be regulated through c-Jun N-terminal kinases (JNK), which is believed to be a central player in the insulin signaling in diabetes and insulin resistance. Reports show that JNK knockout mice are protected against the development of insulin resistance 12,13 . Moreover, administration of small molecule or peptide inhibitors of JNK significantly improved insulin sensitivity in insulin-resistant rodents 14,15 . Multiple factors can activate JNK, such as inflammatory cytokines and free fatty acids. Importantly, glucocorticoids can also increase JNK activity in epithelial cells 16 , hippocampal cells 17 and endothelial cells 18 . Therefore, we hypothesized that glucocorticoid-induced insulin resistance is dependent on 11β -HSD1, resulting in critical activation of JNK in adipose tissue. For study, we used the high fat diet (HFD) mouse model and cultured adipocytes to investigate this potential pathophysiological mechanism under conditions of obesity. Results indicate that glucocorticoid-induced insulin resistance was dependent on 11β -HSD1, resulting in the critical activation of JNK signaling in adipocytes.

Results
Inhibition of 11β-HSD1 and JNK improved insulin sensitivity. Specific small-molecule inhibitors of 11β -HSD1 and JNK (PF00915275 19 and C66 20 , respectively) were used to investigate the causal mechanism between glucocorticoid excess and development of insulin resistance. Our results show that mice fed with HFD have significantly increased body weights compared to the mice fed standard chow (Fig. 1a). Treatment of HFD mice with PF00915275 reduced weight gain compared to the mice treated with vehicle after 8 weeks (39.87 ± 4.60 g vs 44.77 ± 6.78 g, Fig. 1a). As expected, blood glucose levels and glucose AUC were significantly increased in HFD mice (Fig. 1b,c). Administration with either PF00915275 or C66 significantly decreased blood glucose AUC in HFD mice compared to vehicle treatment (Fig. 1b,c). The effects of HFD on expression of 11 β -HSD1 and glucocorticoid receptor (GR) proteins in mice were next investigated. Western blot analysis showed that HFD significantly increased the expression of both 11 β -HSD1 and GR in subcutaneous adipose tissues (Fig. 1d). We also assessed the adipose tissue of PF00915275-and C66-treated mice for expression of insulin receptor substrate 1 (IRS1), glucose transporter-4 (GLUT4), adiponectin and fatty acid binding protein (FABP) as indices of insulin sensitivity. The mRNA levels of IRS-1 (Fig. 1e), GLUT4 (Fig. 1f) and adiponectin (Fig. 1g) were decreased in the adipose tissue of HFD mice, whereas treatment with PF00915275 and C66 prevented their decreased expression and thereby preserved insulin sensitivity ( Fig. 1e-g). In addition, HFD upregulated the expression of FABP in subcutaneous adipose tissue which was completely abolished by PF00915275 and C66 (Fig. 1h). These data confirmed that inhibition of either 11β -HSD1 or JNK significantly attenuated the HFDinduced insulin resistance in mouse adipose tissues.
We also determined JNK activaton by measuring its phosphorylation in the mouse adipose tissues. JNK phosphorylation was significantly increased in the adipose tissue of HFD mice, which was blocked by the administration of the JNK inhibitor C66 (Fig. 1i). Importantly, treatment of the mice with the 11β -HSD1 inhibitor PF00915275 also attenuated HFD-induced JNK phosphorylation (Fig. 1a). These findings support the idea that 11β -HSD1 is an upstream regulator of JNK in modulation of insulin resistance in the HFD model. 11β-HSD1 is an upstream regulator of JNK activation. We further investigated the regulation of JNK by 11β -HSD1 using cultured adipocytes differentiated from 3T3-L1 cells. Treatment of adipocytes with prednisone (a synthetic glucocorticoid) resulted in increased expression of 11β -HSD1 (Fig. 2a) and GR (Fig. 2b). This increase was blocked with pretreatment of cells with 11β -HSD1 inhibitor PF00915275. As expected from our in vivo studies, treatment of 3T3-L1 pre-adipocytes with prednisone increased phosphorylation of JNK and this increase was abolished by PF00915275 and C66 (Fig. 2c). The findings indicated that prednisone activated a sequential pathway of increased 11β -HSD1 expression/activity and JNK activity.

Inactivation of JNK by DN-JNK prevented the 11β-HSD1-dependent insulin resistance.
To exclude possibility of any nonspecific action of C66, we generated 3T3-L1 cells expressing dominant-negative JNK (DN-JNK). As shown in Fig. 4a, transfection of DN-JNK plasmid in 11β -HSD1 overexpressing cells blocked the phosphorylation of JNK. DN-JNK also improved TNF-α -and 11β -HSD1 overexpression-induced glucose uptake (Fig. 4b,d), and insulin-induced Akt phosphorylation (Fig. 4c). Furthermore, TNF-α co-incubation with insulin inhibited insulin-induced Akt phosphorylation and this, as well, was reversed in DN-JNK-transfected cells (Fig. 4c). Lastly, we tested the effects of DN-JNK on insulin resistance induced by 11β -HSD1 overexpression. 11β -HSD1 overexpression significantly decreased insulin-induced glucose uptake (Fig. 4d) and Akt PF00915275 (5 mg/kg) and C66 (5 mg/kg) were administered orally once every other day for 8 weeks to high fat diet-fed mice. (a) Body weight of HFD mice with or without PF00915275 or C66 treatment for 8 weeks showing that PF00915275 treatment reduced weight gain in mice fed HFD. (b) OGTT were performed on mice following the completion of all treatments and just before sacrifice. (c) Blood glucose AUC was calculated based on OGTT data presented in B. (d) protein levels of 11β -HSD1 and GR in subcutaneous adipose tissues from mice fed with normal chow or HFD was determined by western blot analysis. The gels were run under the same experimental conditions. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). Representative two mouse tissues from n = 8 were shown. (e-i) mRNA quantification of IRS-1 (e) GLUT (f) adiponectin (g) and FABP (h) by qPCR. (i) Akt phosphorylation levels measured in the adipose tissues. The gels were run under the same experimental conditions. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). Data was presented as mean ± S.E.M. *P < 0.05; **P < 0.01, ***P < 0.001 compared with the vehicle-treated group. n = 8 per group.

Discussion
The main finding of our study is that glucocorticoid-induced insulin resistance was dependent on 11β -HSD1, leading to JNK activation. Administration of glucocorticoids or overexpression of 11β -HSD1 in adipocytes (e-g) Cells with 11β -HSD1 overexpression were incubated with TNF-α (1 μ g/mL) or vehicle for 1 h. Cell lysates were collected and p-JNK/JNK (e), 11β -HSD1 (f) and GR (g) were detected by western blotting. (h,i) 11β -HSD1 overexpressig cells were incubated with C66 (10 μ M) or DMSO for 2 h. The expression of 11β -HSD1 (h) and GR (i) was assayed. All the gels were run under the same experimental conditions. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). *P < 0.05, **P < 0.01, ***P < 0.001, all data from 3 independent experiments.
increased JNK phosphorylation and induced features of insulin resistance, which were blocked by inhibiting 11β -HSD1 or JNK.
Glucocorticoids are widely used in the clinic as anti-inflammatory agents for a number of human diseases including autoimmune diseases, cancer, organ rejection, and asthma. However, glucocorticoids often lead to insulin resistance and diabetes, a condition referred to as steroid-induced diabetes. A number of studies have shown  Supplementary Fig. S1). (c-f) Cells were treated as in A and mRNA levels of IRS-1 (c), GLUT4 (d), Adiponectin (e) and FABP (f) were examined by qPCR. (g) Cells were treated as in B and protein levels of p-IRS1 Ser302 and IRS1 were examined by western blot analysis. The gels were run under the same experimental conditions. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). Representative blots from 3 independent experiments were shown. (h,i): Differentiated 3T3-L1 cells with or without 11β -HSD1 overexpression were treated with C66 (10 μ M) or vehicle (DMSO) for 2 h and then incubated with insulin (100 nM) for 15 min. Glucose uptake rate (h) and Akt phosphorylation (i) were determined. The gels were run under the same experimental conditions. Shown are cropped gels/ blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). *P < 0.05, **P < 0.01, ***P < 0.001; all data from 3 independent experiments. that the local/intracellular production of glucocorticoids by 11β -HSD1 is more critical than the circulating pool for the development of insulin resistance 9 . Over-expression of 11β -HSD1 in the mouse liver causes active glucocorticoid excess and induces metabolic disorders, while 11β -HSD1 knockout mice are resistant to HFD-induced insulin resistance and metabolic syndromes [21][22][23] . Thus, 11β -HSD1 is believed to an important target for the treatment of type 2 diabetes and metabolic syndromes. However, despite the well-documented role of 11β -HSD1 and active glucocorticoids in causing insulin resistance, the underlying molecular and cellular mechanisms remain unknown. Recent studies showed cross-talk between 11β -HSD1 and JNK, for example, that JNK activation was involved in IL-3-induced 11β -HSD1 expression in airway smooth muscle cells 24 , and inhibition of 11β -HSD1 mediated TNF-α -induced JNK activation and inflammation in pre-adipocytes 25 . However, our studies have demonstrated, for the first time to our knowledge, that JNK plays a central role in the insulin resistance induced by exogenous glucocorticoids or 11β -HSD1 activation (Fig. 5).
Adipose tissue plays an essential role in regulation of whole-body insulin sensitivity. And glucocorticoids have profound effects on the adipose tissue, such as driving central obesity by promoting the redistribution of fat from periphery to abdomen and inducing pre-adipocyte differentiation 26,27 . Interestingly, 11β -HSD1 is highly expressed in adipose tissues 28 . Based on the important role of 11β -HSD1 in adipose tissue and insulin (a) 3T3-L1 cells with or without 11β -HSD1 overexpression were transfected with DN-JNK or control plasmid and expression of p-JNK/JNK was assayed. The gels were run under the same experimental conditions. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). (b) Differentiated 3T3-L1 cells were transfected with DN-JNK or control plasmid and then incubated with TNF-α (1 μ g/mL) for 6 h. Following TNF-α treatment, glucose uptake rate was detected. (c) Cells expressing DN-JNK were treated with TNF-α (1 μ g/mL) for 15 min to assess the level of p-Akt/Akt. (d) 11β -HSD1 overexpressing cells were transfected with DN-JNK and incubated with insulin (100 nM) for 15 min. Following insulin treatment, glucose uptake rate was calculated. (e) Level of p-Akt/Akt in cells treated as in D. The gels from Fig. 4C and Fig. 4E were run under the same experimental conditions, respectively. Shown are cropped gels/blots (The gels/blots with indicated cropping lines are shown in Supplementary Fig. S1). *P < 0.05, **P < 0.01, ***P < 0.001; all data from 3 independent experiments.
Scientific RepoRts | 6:37160 | DOI: 10.1038/srep37160 resistance, we explored the 11β -HSD1-JNK pathway in regulation of insulin resistance. We found that inhibition of 11β -HSD1 or JNK was protective against development of insulin resistance in the HFD mouse model. A recent study has also shown that corticosterone treatment of adipocytes enhances 11β -HSD1 expression and impaired p-Akt leading to lipolysis. Furthermore, knockdown of 11β -HSD1 by shRNA attenuated corticosterone-induced lipolysis and improved insulin signaling response in adipocytes 29 . Taken together, these findings suggest that glucocorticoid signaling in adipocytes leads to an 11β -HSD1-dependent impairment of insulin.
A potential limitation of our study is that we did not examine the liver tissues in these mice. Insulin and 11β -HSD1 also function in the liver. However, the mechanisms may be shared with adipocytes and may also involve JNK. Whole body knockout of JNK or JNK inhibitor dramatically improved insulin sensitivity 12 . Interestingly, specific suppression of JNK activation in adipose tissue alone is sufficient to prevent the development of insulin resistance in HFD mice 30 . JNK interferes with insulin signaling through several mechanisms. JNK inhibits IRS1 function through phosphorylation of IRS1 at Ser 307 31 or downregulation of IRS1 expression 32 . JNK activation in adipose tissue has also been shown to decrease the production of adiponectin, which contributes to insulin resistance induced by JNK activation 30,33 . Furthermore, FABP is essential for adipocytes to produce TNF-α . Mice with FABP deletion are resistant to HFD-induced metabolic abnormality 34 . However, an independent group found that lipopolysaccharide induces FABP upregulation through activation of JNK 35 . Thus, FABP may also be implicated in JNK-induced insulin resistance. Our studies showed that glucocorticoid treatment or overexpression of 11β -HSD1 decreased IRS1 and adiponectin expression and increased FABP activation. These insulin resistance-related changes were reversed by JNK inhibition, establishing the importance of the glucocorticoid-11β -HSD1 pathway in JNK activation.
Activation of JNK by glucocorticoids and its involvement in insulin signaling has been also reported by other groups in different systems. For example, heat shock (DNAJB3) proteins have been shown to improve insulin signaling and glucose uptake through JNK repression 36 . Pigment epithelium-derived factor (PEDF) also suppresses JNK activation to improve hepatocyte insulin signaling 37 . We found that glucocorticoids activated JNK in adipocytes, resulting in aberrations in the insulin signaling pathway. Despite our findings and those by others, the specific molecular mechanisms by which the GC-GR complex activates JNK remain elusive. A previous study has demonstrated that GR could physically bind to JNK in dexamethasone-treated HeLa cells 38 . Our co-immunoprecipitation studies found that treatment with 5 μ M prednisone for 2 h induced formation of the GR-JNK complex in 3T3-L1 adipocytes (Supplementary Figure S2), which confirmed the previous results 38 . However, the co-immunoprecipitation data could not rule out the existence of other linker proteins in the complex. In addition, multiple lines of evidence have shown that glucocorticoids can directly activate protein kinase C (PKC) isoforms 17,39-43 which are upstream activators of JNK [44][45][46][47] . PKC activation has also been demonstrated to cause insulin resistance [48][49][50] . Thus, PKCs may contribute to the activation of JNK in response to glucocorticoids in adipocytes. hese observations suggest other additional mechanisms can be involved. Overall, findings from our study strongly support a role of JNK activation in GC-induced insulin resistance. An important direction for future studies is to elucidate the mechanisms by which GC/GR activate JNK.
In conclusion, our study identified a novel mechanism by which a glucocorticoid-11β -HSD1 pathway induced insulin resistance through JNK activation in adipocytes Moreover, the results suggest that 11β -HSD1 and its downstream target molecule, JNK, are key therapeutic strategies for treating insulin resistance in response to glucocorticoid excess.
Cell culture. The mouse pre-adipocyte line 3T3-L1 was obtained from the Shanghai Institute of Biochemistry and Cell Biology (Shanghai, China). Cells were cultured in Dulbecco's modified Eagle medium (DMEM) medium (Gibco, Thermo Fisher) containing 25 mM D-glucose supplemented with 10% fetal bovine serum, 100 U/mL penicillin and 100 mg/mL streptomycin at 37 °C in a humidified 5% CO 2 atmosphere. Upon confluence, cells were trypsinized and plated onto culture dishes for the measurement of kinase phosphorylation, and protein expression and mRNA levels. Prednisone was incubated for the indicated time intervals. For some experiments, cells were pre-treated with inhibitors for 2 h before prednisone treatment.

3T3-L1 cells differentiation.
To induce the differentiation of 3T3-L1 cells into adipocytes, cells were cultured to confluence. Cells were then treated with complete differentiation medium which contained 10 mg/L insulin, 1 μ M dexamethasone and 0.5 mM IBMX (isobutyl-methylxanthine). Medium was changed after 2 days with fresh medium containing 10 μ g/L insulin, then every other day for a total of 10 days. On day 11, cells were cultured in serum-free medium for 24 h and then treated with prednisone or TNF-α for 6 h to induce insulin resistance.
Adipocyte glucose uptake. The amount of glucose taken up by the differentiated 3T3-L1 adipocytes cells was measured by determining the difference of glucose concentrations in the culture medium before and after incubation with cells. Glucose levels in the medium were measured using a glucose assay kit. The results were expressed as glucose uptake rate compared to the insulin-treated group (100%).
For dominant negative JNK (DN-JNK) transfections, cells were incubated for 6 h in 1 mL serum-free medium containing 10 μ L Lipofectamine 2000 and 2.5 μ g DN-JNK or empty vector plasmid. Forty eight hours after transfection, cells were treated with 1 μ g/mL TNF-α or vehicle control for 1 h.

Animals.
Six week old male C57BL/6 mice weighing 18-22 g were obtained from the Animal Centre of Wenzhou Medical University (Wenzhou, China). Animals were housed with a 12:12 h light-dark cycle at constant room temperature, and fed standard chow diet. The mice were acclimatized to the laboratory for at least 2 weeks before initiating studies. All animal care and experimental procedures were approved by the Wenzhou Medical University Animal Policy and Welfare Committee and performed in accordance with the approved protocols and the "The Detailed Rules and Regulations of Medical Animal Experiments Administration and Implementation"(Document No. 1998-55, Ministry of Public Health, China).
Control mice (n = 8) were fed with standard chow diet (low fat diet) which contained 10 kcal.% fat, 20 kcal.% protein and 70 kcal.% carbohydrate (MediScience Diets Co., Yangzhou, Chinca; Cat. #MD12031). The high fat diet (HFD) group was fed a diet containing 60 kcal.% fat, 20 kcal.% protein and 20 kcal.% carbohydrate (Cat. #MD12033). After 8 weeks of diet intervention, the HFD mice were randomized to 3 groups for another 8 weeks of treatment as follows: i) HFD alone group (n = 8); ii) HFD + PF00915275 (11β -HSD1 inhibitor; 5 mg/kg, orally once every other day) (n = 8); and iii) HFD + C66 (JNK inhibitor; 5 mg/kg, orally once every other day) (n = 8). Body weight measurements and oral glucose tolerance test (OGTT) were performed at the end of the treatment period. For OGTT, mice were fasted for 4 h and glucose (2 g/kg body weight) was administrated by oral gavage. Blood glucose levels were measured with a glucometer at 0, 15, 30, 60 and 120 min after glucose load. At the end of the study (week 16), the mice were anesthetized using isoflurane anesthesia and sacrificed after 12 h fasting period. The subcutaneous adipose tissue was harvested for gene and protein expression analysis.
Real-time quantitative PCR. Cells or tissues were homogenized in TRIZOL (Thermo Fisher) for extraction of RNA. Reverse transcription and quantitative PCR were carried out using a two-step M-MLV Platinum SYBR Green qPCR SuperMix-UDG kit (Thermo Fisher). Eppendorf Mastercycler ep realplex detection system (Eppendorf, Hamburg, Germany) was used for q-PCR analysis. Primers were obtained from Thermo Fisher. The sequences were: AGGAGATGTTGGAATGACAGG (forward) and CTGAACGCTGAGCGATACATA (reverse) for Adiponectin, CATCAGCGTAAATGGGGATT (forward) and TCGACTTTCCATCCCACTTC (reverse) for FABP-4, GCCAGCCTACGCCACCATAG (forward) and AGCAGAGCCACGGTCATCAAG (reverse) for GLUT4, CAACAGCAGCAGCAGTCTTCC (forward) and CCGAGCCAGTCTCTTCTCTAGG (reverse) for IRS1, and CCGTGAAAAGATGACCCAGA (forward) and TACGACCAGAGGCATACAG (reverse) for β -actin. The amount of each gene was determined and normalized to the amount of β -actin.
Western immunoblot analysis. Cells or tissues were lysed and 50 μ g of lysate were separated by 10% SDS-PAGE and transferred to PVDF membranes. The membranes were pre-incubated for 1.5 h at room temperature in Tris-buffered saline, pH 7.6, containing 0.1% Tween 20 and 5% non-fat milk. Following primary antibody incubation, immunoreactive bands were detected by horseradish peroxidase-conjugated secondary antibodies and enhanced chemiluminescence reagents (Bio-Rad, Hercules, CA). The intensity of the bands was analyzed using Image J analysis software version 1.38e (National Institutes of Health, Bethesda, MD) and normalized to their respective control. Statistical analysis. The statistical significance of differences was obtained by the student's t test or ANOVA with multiple comparisons in GraphPad Pro (GraphPad, San Diego, CA). Differences were considered to be significant at P < 0.05. Data are presented as means ± SEM.