Edinburgh Research Explorer Regulation of Adipocyte 11-Hydroxysteroid Dehydrogenase Type 1 (11-HSD1) by CCAAT/Enhancer-Binding Protein (C/EBP) Isoforms, LIP and LAP Regulation of Adipocyte 11 b -Hydroxysteroid Dehydrogenase Type 1 (11 b -HSD1) by CCAAT/Enhancer-Binding Protein (C/EBP) b Isoforms, LIP and LAP

11 b -hydroxysteroid dehydrogenase type 1 (11 b -HSD1) catalyses intracellular regeneration of active glucocorticoids, notably in liver and adipose tissue. 11 b -HSD1 is increased selectively in adipose tissue in human obesity, a change implicated in the pathogenesisofmetabolicsyndrome.Withhighfat(HF)-feeding,adiposetissue11 b -HSD1isdown-regulatedinmice,plausibly to counteract metabolic disease. Transcription of 11 b -HSD1 is directly regulated by members of the CCAAT/enhancer binding protein (C/EBP) family. Here we show that while total C/EBP b in adipose tissue is unaltered by HF diet, the ratio of the C/EBP b isoforms liver-enriched inhibitor protein (LIP) and liver-enriched activator protein (LAP) (C/EBP b -LIP:LAP) is increased in subcutaneous adipose. This may cause changes in 11 b -HSD1 expression since genetically modified C/EBP b ( + /L) mice, with increased C/EBP b -LIP:LAP ratio, have decreased subcutaneous adipose 11 b -HSD1 mRNA levels, whereas C/EBP b D uORF mice, with decreased C/EBP b -LIP:LAP ratio, show increased subcutaneous adipose 11 b -HSD1. C/EBP b -LIP:LAP ratio is regulated by endoplasmic reticulum (ER) stress and mTOR signalling, both of which are altered in obesity. In 3T3-L1 adipocytes, 11 b -HSD1 mRNA levels were down-regulated following induction of ER stress by tunicamycin but were up-regulated following inhibition of mTOR by rapamycin. These data point to a central role for C/EBP b and its processing to LIP and LAP in transcriptional regulation of 11 b -HSD1 in adipose tissue. Down-regulation of 11 b -HSD1 by increased C/EBP b -LIP:LAP in adipocytes may be part of a nutrient-sensing mechanism counteracting nutritional stress generated by HF diet.

Introduction 11b-hydroxysteroid dehydrogenase type 1 (11b-HSD1) is highly expressed in liver and adipose tissue where it catalyses the regeneration of active glucocorticoids (corticosterone, cortisol) from inert 11keto-forms (11-dehydrocorticosterone, cortisone) thus increasing intracellular glucocorticoid action [1]. 11b-HSD1 expression is elevated selectively in adipose tissue of obese humans and in monogenic rodent genetic obesity, whereas levels in liver are unaffected or even decreased [2,3,4]. Transgenic overexpression of 11b-HSD1 in adipose tissue recapitulates the metabolic syndrome in mice, with visceral obesity, dyslipidemia, insulin resistance/diabetes and hypertension [2,5]. In contrast, 11b-HSD1-deficiency or inhibition causes insulin-sensitization (including in humans), lowers fasting plasma glucose and lipid levels, reduces visceral adipose tissue mass and attenuates atherosclerosis [6,7,8]. Unexpectedly, high fat (HF) diet downregulated 11b-HSD1 selectively in adipose tissue in mice and rats [9,10,11]. This down-regulation is greatest in obesity-resistant strains [9] suggesting it may be a mechanism to minimise metabolic disease with adiposity. Understanding the mechanisms of adipose-specific control of 11b-HSD1 is crucial to dissecting the pathogenesis of sensitivity/resistance to obesity.
In vivo, C/EBPa is the major known inducer of 11b-HSD1 transcription in liver, where C/EBPb acts as a relative repressor [12]. In contrast, in adipose tissue, C/EBPb is an activator [29]. Moreover, C/EBPb is required for glucocorticoid-induction of 11b-HSD1 in A549 cells [16] as well as regulation by IL-1b, cAMP and ceramide in fibroblasts and 3T3-L1 preadipocytes [13,14,30]. However, the crucial transcriptional regulation of 11b-HSD1 in adipose tissue and mature adipocytes remains unexplored. Here, we have tested the hypothesis that C/EBPs are modulated by diet in mice and mediate the regulation of 11b-HSD1 in adipose.

Materials and Methods
Animals C57BL/6J mice (Harlan UK Ltd., Oxon, UK) were housed in standard conditions on a 12 h light, 12 h dark cycle (lights on at 0700 h) at 2161uC. Adult male mice (n = 16/group) were fed control diet (11% calories as fat; diet D12328, Research Diets, Inc., New Brunswick, NJ) or HF diet (58% calories as fat; diet D123331, Research Diets) for 6 weeks. Adult male C/EBPb mutant mice, heterozygous C/EBPb (+/L) and homozygous C/EBPb DuORF and respective wild-type (WT) littermate control mice (n = 6-7/group) were generated as previously described [31,32] and fed standard chow diet. All animal experiments were conducted in strict accord with accepted standards of humane animal care under the auspices of the Animal (Scientific Procedures) Act UK 1986 and following prior approval by the Home Office in UKor following prior approval by the Institutional Animal Care and Use Committee in Berlin, Germany.

3T3-L1 Adipocyte Differentiation and Transfection with siRNA
3T3-L1 murine preadipocytes (ATCC, American Type Culture Collection) were maintained and differentiated into mature adipocytes as previously described [33]. Briefly, 2d after reaching confluence, 3T3-L1 cells were induced to differentiate by the addition of 0.5 mM dexamethasone, 500 mM 3-isobutyl-1-methylxanthine and 5 mg/ml insulin for 2d. Thereafter, 3T3-L1 cell differentiation continued in medium supplemented with 5 mg/ml insulin alone. Experiments were performed 8-12d following induction of differentiation, unless otherwise stated. 3T3-L1 adipocytes were transfected with siRNA using DeliverX Plus (Panomics Inc., Fremont, CA) according to the manufacturer's protocol. Fully differentiated adipocytes were gently detached by a brief wash with trypsin followed by collagenase type I (0.5 mg/ml; Invitrogen, Paisley, UK), and reseeded at 1.5610 5 cells per well in a 12-well plate then transfected the next day with siRNA (32 pmol) in serum-free medium for 3 h followed by 21 h incubation in growth medium. All siRNAs were purchased from Applied Biosystems (Warrington, UK) and were scrambled (AM4611, negative control), C/EBPa (ID 101889), C/ EBPb (ID 288793) and CHOP (ID 288792). In the experiments designed to induce ER stress, 3T3-L1 adipocytes were treated with tunicamyin (4 mg/ml). Rapamycin (100 nM and 500 nM, 24 h) was used to inhibit the mammalian target of rapamycin (mTOR) in 3T3-L1 adipocytes.

Statistical Analysis
All data were analyzed by Student's t test or ANOVA followed by post hoc Tukey, Fisher LSD or Dunnet tests using SigmaStat 2.03 statistical software. Significance was set at p#0.05.

Effect of HF Diet on 11b-HSD1 and C/EBP Expression in Mouse Adipose Tissue
Mice fed HF diet for 6 weeks were heavier (HF, 30.860.52 vs control diet, 26.360.34 g; p,0.01) with increased subcutaneous adipose tissue weight (HF, 0.02260.001 vs control, 0.01260.001 (w/w) corrected for body weight; p,0.01), while liver weight was unchanged (HF, 0.04960.001 vs control, 0.04660.003 (w/ w) corrected for body weight). 11b-HSD1 mRNA was downregulated in adipose tissue by HF diet both in subcutaneous ( Fig. 1A) and visceral (mesenteric) depots (data not shown), but was unchanged in liver (data not shown), consistent with previous data showing down-regulation of 11b-HSD1 mRNA and enzyme activity in adipose tissue of mice fed HF diet [9,11]. To test whether altered C/EBP expression may underlie the dietary regulation of 11b-HSD1, we examined C/EBPa, b, d and CHOP expression in adipose tissue. HF diet did not change C/EBPa, b or d mRNA levels, while CHOP mRNA levels were increased in subcutaneous adipose tissue (Fig. 1A) and mesenteric [11] adipose tissue (data not shown). Because 11b-HSD1 and C/EBP mRNA levels showed the same pattern of changes in both adipose depots of HF-fed (vs control) animals, subcutaneous adipose, which is more abundant, was used for subsequent analyses. Consistent with mRNA levels, western blot analysis showed an increase in CHOP protein levels but no alteration in total C/EBPa (p42+ p30 isoforms), total C/EBPb (LAP* + LAP + LIP) or C/EBPd (Fig. 1B-E) protein levels with HF diet. However, HF diet reduced adipose tissue levels of the C/EBPb-LAP*+LAP isoforms, concomitantly increasing levels of C/EBPb-LIP (Fig. 1C), resulting in a significant increase in the C/EBPb-LIP:LAP ratio (Fig. 1C, inset).
C/EBPa and C/EBPb, but not CHOP, Regulate Expression of 11b-HSD1 in 3T3-L1 Adipocytes Increased CHOP levels or increased C/EBPb-LIP:LAP ratio could plausibly reduce adipose 11b-HSD1 expression. The effect of CHOP on 11b-HSD1 expression has not been reported. Moreover, although C/EBPb regulates 11b-HSD1 expression during preadipocyte differentiation [13,29], any role in mature adipocytes has not been tested. To investigate the requirement for C/EBPa, C/EBPb and CHOP in 11b-HSD1 expression in adipocytes, 3T3-L1 preadipocytes were fully differentiated into mature adipocytes, which show high endogenous expression of 11b-HSD1 [33], and then transfected with siRNAs to decrease levels of C/EBPa, C/EBPb or CHOP. Measurement of mRNA and protein levels confirmed the reduction in expression of appropriate C/EBPs with siRNA, notably reducing all isoforms of C/EBPa and C/EBPb ( Fig. 2A, B) but increasing C/EBPb- Figure 3. Binding of C/EBPa and C/EBPb to the promoter of 11b-HSD1 throughout 3T3-L1 adipocyte differentiation. (A) Relevant C/ EBP binding sites in the mouse 11b-HSD1 promoter; FP1 to 4 and 9. The start of transcription (+1) [12] is indicated with a bent arrow. The indicated C/ EBP binding sites are highly conserved between the rat and mouse 11b-HSD1 promoters [12,13] and are named according to the sites in the rat promoter [12]. PCR products amplified in the ChIP experiment are indicated (a to c). (B) and (C) ChIP assays carried out to examine C/EBP binding to the 11b-HSD1 promoter in (B) differentiating 3T3-L1 adipocytes (8 h after induction of differentiation) and (C) fully differentiated adipocytes (9d after induction of differentiation). Left panels show representative gels of ChIP assays following immunoprecipitation with C/EBPa or C/EBPb antibody. Control reactions contained rabbit IgG. PCR reactions (35 cycles) were carried out on input DNA (input) or on immunoprecipitated DNA with primers spanning FP1/2 of the 11b-HSD1 promoter (lanes marked a), FP3/4 (lanes marked b), FP9 (lanes marked c), an intronic sequence from the 11b-HSD1 gene not predicted to bind C/EBP (lanes marked d) and a C/EBPb binding site in the C/EBPa promoter [44] (lanes marked e). In (B), all PCR products were visualized in the same gel but not in adjacent lanes, whereas in (C), all PCR reactions were performed at the same time but were visualized in different gels. Right panels show real-time PCR quantification of C/EBPa and C/EBPb binding to the FP1/2 and FP3/4 regions of the 11b-HSD1 promoter. Reactions were performed with the same primers used for standard PCR as detailed in Methods. ND, not detected. Values are mean6SEM from 2 to 3 independent experiments and each sample was analysed in triplicate. doi:10.1371/journal.pone.0037953.g003 LIP:LAP ratio (scrambled RNA, 0.560.03 vs C/EBPb siRNA 0.760.02). The results also showed that reduced expression of one of the C/EBPs in mature 3T3-L1 adipocytes did not affect mRNA levels of the other C/EBPs tested ( Fig. 2A). Both C/ EBPa and C/EBPb siRNA reduced 11b-HSD1 mRNA levels in 3T3-L1 adipocytes (Fig. 2C), demonstrating that both these members of the C/EBP family play a key role in maintaining adipocyte expression of 11b-HSD1. In contrast, although levels of CHOP were reduced in 3T3-L1 adipocytes by siRNA ( Fig. 2A, B), this did not alter 11b-HSD1 mRNA levels (Fig. 2C).

C/EBPb Locates to the Promoter of 11b-HSD1 throughout Adipocyte Differentiation Whilst C/EBPa Binds Only in Differentiated Adipocytes
Previous reports showed C/EBPb binding to the 11b-HSD1 promoter in undifferentiated 3T3-L1 preadipocytes [13,14], but this has not been described in adipocytes. To determine if C/EBPs interact with the 11b-HSD1 promoter in immature and fully mature adipocytes, binding of C/EBPa and b was assayed by chromatin immunoprecipitation (ChIP). The proximal promoter of the rat 11b-HSD1 gene contains 4 C/EBP binding sites between -196 and +44; footprints (FP) 1 to 4, which are conserved in the mouse promoter [12,13] (Fig. 3A). C/EBPb was bound to the proximal 11b-HSD1 promoter early during adipocyte differentiation (when 11b-HSD1 expression is low) (Fig. 3B) as well as in fully differentiated 3T3-L1 adipocytes (Fig. 3C). In contrast, C/EBPa was only bound to the 11b-HSD1 proximal promoter in fully differentiated 3T3-L1 adipocytes, which highly express 11b-HSD1 [33]. Neither C/EBPa nor C/EBPb were detected at FP9 in the promoter of 11b-HSD1 (also conserved between rat and mouse) (Fig. 3B, C). Thus C/EBPb interacts directly with the 11b-HSD1 promoter in mature adipocytes.

11b-HSD1 mRNA Levels are Down-regulated by ER Stress in 3T3-L1 Adipocytes
Elevated C/EBPb-LIP:LAP ratio and CHOP levels, observed in adipose tissue of HF-fed mice, may be a result of ER stress [34,35]. Treatment of differentiated 3T3-L1 adipocytes with tunicamycin, an inducer of ER stress [35], significantly reduced 11b-HSD1 mRNA levels at 6 h, with a more pronounced effect at 16 h (Fig. 5A). Similar results were obtained with an alternative ER stress-inducer, thapsigargin (data not shown). Down-regulation of 11b-HSD1 mRNA was accompanied by up-regulation of genes increased by ER stress; CHOP and the 78 kDa glucose-regulated protein (GRP78) (Fig. 5C, E), though calreticulin mRNA levels were unaffected at this stage (Fig. 5D). C/EBPb mRNA levels were not altered (Fig. 5B) whilst, as expected, C/EBPb-LIP:LAP ratio was increased, and both C/EBPb-LAP and -LIP were upregulated (Fig. 5F). These results show that induction of ER stress, which increases the C/EBPb-LIP:LAP ratio, also down-regulates 11b-HSD1 in 3T3-L1 adipocytes.
mTOR Regulates the Expression of 11b-HSD1 in 3T3-L1 Adipocytes The mTOR pathway, which is responsive to nutrient availability and growth factors, controls many aspects of cellular and -LIP (20 kDa) isoforms in the adipose tissue of C/ EBPb mutants C/EBPb (+/L) (+/L), C/EBPb DuORF (DuORF) and control (Con) mice. Blots were stripped and reprobed with b-tubulin antibody, as loading control. All samples were analysed in the same gel but not all in adjacent lanes. (B) Real-time PCR measurement levels of mRNA encoding 11b-HSD1 in adipose tissue of wild type control mice (Con, white bar), C/EBPb (+/L) (+/L, black bar) and C/EBPb DuORF (DuORF, grey bar). C/EBPb (+/L) mice are heterozygous for an allele of C/EBPb in which the normal gene has been replaced by C/EBPb-LIP (a ''knock-in'') [31] and C/EBPb DuORF is homozygous for the deletion of upstream ORF codon [32]. Adipose 11b-HSD1 mRNA levels, normalized to TBP, are expressed relative to levels in control mice (arbitrarily set to 100%) and are mean6SEM; n = 6-9/group. *, p#0.05; **, p,0.001. doi:10.1371/journal.pone.0037953.g004 metabolism and regulates the production of C/EBPb-LIP and -LAP isoforms [24]. To test the importance of this pathway for the transcriptional regulation of 11b-HSD1, rapamycin was used to inhibit mTOR in differentiated 3T3-L1 adipocytes. Rapamycin did not alter C/EBPb mRNA levels (Fig. 6A) or C/EBPb-LAP, but decreased C/EBPb-LIP, resulting in a lower C/EBPb-LIP:LAP ratio (Fig. 6C). This was associated with increased 11b-HSD1 mRNA levels (Fig. 6B). These results show that the mTOR signalling pathway regulates 11b-HSD1 transcription in adipocytes, plausibly via modulation of the C/EBPb-LIP:LAP ratio.

Discussion
This study emphasises the key role played by members of the C/EBP family in the regulation of 11b-HSD1 transcription in adipocytes. However, the main finding here is that the striking down-regulation of 11b-HSD1 in adipose tissue of mice in response to HF feeding is likely due to changes in posttranscriptional processing of C/EBPb generating an increased C/EBPb-LIP:LAP ratio. Indeed, knock-down of C/EBPb in fully differentiated mature 3T3-L1 adipocytes showed that C/EBPb is required for normal 11b-HSD1 expression in differentiated adipocytes, consistent with a previous report of reduced 11b-HSD1 expression in adipose tissue of C/EBPb-deficient mice [29]. In differentiating 3T3-L1 adipocytes, C/EBPb binds to the C/ EBPa promoter and induces its expression [20]. Importantly, in contrast, knock-down of C/EBPb in fully differentiated adipocytes did not affect levels of C/EBPa, suggesting a direct role for C/ EBPb in adipocyte 11b-HSD1 expression, independent of its effect on C/EBPa expression. Indeed, C/EBPa mRNA and protein levels were unchanged in adipose tissue of HF diet-fed mice. However, it cannot be excluded that C/EBPb may also exert its effect by regulating C/EBPa transactivation potential. Nevertheless, supporting a direct role, C/EBPb was bound to the 11b-HSD1 promoter in mature adipocytes as well as in 3T3-L1 preadipocytes during the differentiation process. C/EBPa, in contrast, was only bound to the 11b-HSD1 promoter in Figure 6. Inhibition of mTOR in 3T3-L1 adipocytes reduces C/EBPb-LIP:LAP ratio and increases 11b-HSD1 expression. Real-time PCR measurement of levels of mRNA encoding (A) C/EBPb and (B) 11b-HSD1 in 3T3-L1 adipocytes (white bars) or following treatment with 100 nM (grey bars) or 500 nM (black bars) rapamycin for 24 h. Data are normalized to TBP and expressed relative to levels of control cells (arbitrarily set to 100%). (C) Representative western blot (40 mg protein/lane) and quantification showing levels of C/EBPb-LAP (38 kDa LAP* +35 kDa LAP) and -LIP (20 kDa) isoforms and inset showing C/EBPb-LIP:LAP ratio in untreated 3T3-L1 adipocytes (C, white bars) or following treatment with 100 nM (100, grey bars) or 500 nM (500, black bars) rapamycin for 24 h. Blots were stripped and reprobed with b-tubulin antibody, as loading control. All samples were analysed in the same gel but not in adjacent lanes. C/EBPb levels were quantified relative to b-tubulin and are expressed in arbitrary units (AU). Values are mean6SEM of up to 3 independent adipocyte differentiations, each treatment tested in triplicate. *, Significantly different from control. *, p#0.05; **, p,0.001. doi:10.1371/journal.pone.0037953.g006 differentiated adipocytes, as expected from its later appearance during differentiation [20]. Thus, both C/EBPa and C/EBPb play a direct and positive role in regulation of 11b-HSD1 transcription in differentiated adipocytes, but C/EBPb alone is crucial in differentiating cells. This is likely to impact on the 11b-HSD1 mRNA levels in these cells, where C/EBPa is absent or in low levels. In adipose tissue, C/EBPb and CHOP were the only C/EBPs changed by HF diet. Although CHOP generally represses transcription [36], it can also be a co-activator [37]. However, siRNA-mediated knock-down of CHOP showed that it plays no role in 11b-HSD1 expression in fully differentiated 3T3-L1 adipocytes, although we cannot rule out that increased CHOP levels may contribute to the down-regulation of 11b-HSD1 in adipose tissue with HF diet.
An altered C/EBPb-LIP:LAP ratio, like the HF-induced increase in CHOP, might be a response to ER stress as a consequence of a high demand for ER activity with nutrient excess [34]. The C/EBPb-LIP:LAP ratio is controlled by the double stranded RNA-dependent protein kinase (PKR, a target of ER stress [38]) and mTOR signalling pathways through the eukaryotic translation factors eIF-2a and eIF-4E, respectively [24]. Indeed, diet-induced obesity in mice results in the activation of ER stress pathways in metabolically relevant tissues, such as adipose and liver, with increased phosphorylation of the PKR-like kinase and its substrate eIF-2a, which are key indicators of ER stress [34]. mTOR may contribute to the ER stress response as well [39]. Our results show that both induction of ER stress and inhibition of mTOR signalling alter C/EBPb-LIP:LAP ratio and modulate 11b-HSD1 mRNA levels in 3T3-L1 adipocytes. ER stress and mTOR signalling pathways are sensitive to cellular nutrients and energy homeostasis, suggesting that the increased C/EBPb-LIP:LAP ratio in obese adipose tissue may result from activation of these nutrient/stress sensing mechanisms leading to the downregulation of 11b-HSD1.
Alterations in C/EBPb-LIP:LAP ratio are likely to impact upon other metabolically-important genes, many of which, like 11b-HSD1, are regulated by C/EBPb [40,41]. Indeed, the gene encoding phosphoenolpyruvate carboxykinase, a C/EBPb-regulated enzyme with key roles in hepatic gluconeogenesis and adipose glycerogenesis, is down-regulated by an increased LIP:LAP ratio [26]. C/EBPb-null mice have decreased gluconeogenesis and lipolysis during fasting or diabetes [40,41], resist dietinduced obesity [42] and exhibit attenuated ER stress and inflammatory responses [43]. Importantly, mice deficient in C/ EBPb lack both LIP and LAP isoforms and so are not informative regarding the relevance of the C/EBPb-LIP:LAP ratio. In contrast, gene targeted C/EBPb (+/L) and C/EBPb DuORF mice, have altered levels of the C/EBPb isoforms. C/EBPb DuORF mice show activation of acute-phase response genes after partial hepatectomy, suggesting that C/EBPb-LIP may be a natural restraint of inflammation [32]. However, the impact of the C/ EBPb-LIP:LAP ratio on 11b-HSD1 expression has not been previously reported. Here, we demonstrate that manipulation of the C/EBPb-LIP:LAP ratio regulates the expression of 11b-HSD1, in agreement with the modulation of 11b-HSD1 mRNA levels observed in mouse adipose tissue by diet.