Impaired Expression of the Inducible cAMP Early Repressor Accounts for Sustained Adipose CREB Activity in Obesity Open Access

All procedures on mice were performed according to the Swiss legislation for animal experimentation. Six-week-old C57Bl6-Rj male mice were obtained from the Janvier Breeding Centre (Le Genest-Saint-Isle, France). Diets and hyperinsulemic-euglycemic clamps were realized as previously described (15). Epididymal visceral adipose tissue (VAT) was extracted for total RNA and protein preparation in killed animals.

Approximately 5 cm³ of VAT was obtained at the level of the omentum from five obese white women (BMI >35 kg/m²) who were referred for weight reduction surgery and five nonobese white women (24< BMI <28 kg/m²) (16). All patients provided informed consent, and the study was approved by the institutional review board (16). The criteria for exclusion and phenotyping are those previously described (16). Total RNA was isolated from adipose tissues and different cell fractions with the TriPure isolation reagent (Roche) as previously described (17). The procedure for preparation of adipocytes and stroma vascular fraction (SVF) was performed as previously described (18).

The antibodies against CREM that recognize ICER were provided by E. Lalli (Institut de Pharmacologie Moléculaire et Cellulaire, Nice University, Valbonne, France). The antibodies against hairy and enhancer of Split-1 (HES-1), CREB1, ATF3, and histone deacetylase 1 (HDAC1) were obtained from Santa Cruz Biotechnology (Santa Cruz, CA).

Culture and differentiation of 3T3-L1 cells were conducted as previously described (19). The 19-nucleotide small interfering RNA (siRNA) duplex (siICER) was synthesized by Mycrosynth AG (Balgach, Switzerland). For overexpression and silencing of Icer, 3T3-L1 adipocytes on day 7 postdifferentiation were electroporated with either the plasmid coding for Icer Iγ (ICER) or siRNAs (12) using the GenePulser XCell (Bio-Rad).

Nuclear protein extracts, Western blotting, EMSA, and real-time PCR were conducted as previously described (12). PCR assays were carried out on a Bio-Rad MyiQ single-color, real-time PCR detection system using the Bio-Rad iQ SYBR Green Supermix, with 100 nmol/L primers and 1 μl of template per 20 μl of PCR and an annealing temperature of 60°C. The primer sequences are available in the Supplementary Data.

2-Deoxyglucose (2-DOG) uptake assays was conducted on fully differentiated 3T3-L1 adipocytes (days 7 and 8) as previously described (20). Adipocytes were treated without (basal) or with insulin (100 pmol/L) for 10 min. 2-[³H]DOG (0.1 μCi; final concentration, 0.01 mmol/L) and 5 mmol/L cold 2-DOG were then added for an additional 10 min at 37°C.

The experiments including two groups were analyzed by Student t test or with the nonparametric equivalent Wilcoxon.

Obese mice fed with an HFD display similar metabolic disorders to those observed in humans (21,22). The diet indeed led to increased mouse body weight, hyper-insulinemia, hyperglycemia, and insulin resistance, as indicated by a marked decrease in glucose infusion rate compared with littermates fed with chow diet (Supplementary Table 1). In this model, elevation in the transcriptional activity of Atf3 downregulates the expression of Glut4 and Adipoq in white adipose tissue (WAT), thereby contributing to systemic insulin resistance (9). In agreement with these data, we found an increase in the abundance of Atf3 and decreased expression of Glut4 and Adipoq in WAT of obese mice when compared with control animals (Fig. 1A and C). The rise of Atf3 was associated with a striking drop in Icer content in WAT of obese mice when compared with control animals both after fasting (Fig. 1A and C) or fed conditions (data not shown), indicating that the reduction in Icer in HFD-induced obese mice is independent from the nutritional status. To confirm the data obtained in mice, we analyzed adipose tissues from human nonobese and obese subjects. In agreement with the data found in animals, we detected a smaller amount of ICER mRNA in the VAT of obese subjects (Fig. 1B). In contrast, the expression of ICER was significantly increased in subcutaneous adipose tissue of obese individuals (Supplementary Fig. 1). These results indicate that VAT and subcutaneous adipose tissue dysfunctions are likely to rely on distinct mechanisms. As it is the excess of visceral rather than subcutaneous fat that is more associated with an increased risk of developing cardiovascular disease and metabolic syndrome (23), we chose to concentrate our study on VAT. Reduction of ICER mRNA in VAT of obese individuals was associated with an increase in the expression of the CREB target gene ATF3 and decreased levels of ADIPOQ and GLUT4 (Fig. 1B). EMSAs were next done to confirm changes in binding activities of Icer and Creb. Nuclear proteins from WAT of obese mice were incubated with a labeled oligonucleotide corresponding to the cAMP response element (CRE) (12). A fast migrating band was observed in WAT of control mice, which was markedly reduced in obese mice (Fig. 1D). On the other hand, a slower migrating pattern appeared in WAT of obese mice that is almost undetectable in control animals (Fig. 1D). The lower band was identified as Icer because it comigrates with exogenously expressed ICER-Iγ, and the formation of this complex was selectively abrogated by the addition of an anti-CREM antibody (Fig. 1E). In the presence of an anti-Creb1 antibody, the migration of the upper band observed in WAT of obese mice was slower (Fig. 1F). This result indicates that nuclear extracts of WAT from obese mice display an increase in Creb binding activity that coincides with a reduction in Icer expression. In support of this result, although the binding pattern corresponding to ICER in VAT was different from WAT, the activity of this factor was reduced in obese individuals (Fig. 1G and H). The diminution in ICER activity was associated with a slower migrating complex in VAT from obese individuals, which was identified as the CREB factor (Fig. 1I and J). Various stimuli, including insulin, can elicit Creb activity (24,25). In line with this, we found that insulin infusion elicited a 16-fold increase in the expression of Icer in control mice (Supplementary Fig. 2). As anticipated, insulin-induced Icer expression was abolished in insulin-resistant WAT of obese mice, indicating that induction of Icer is also defective in obese animals.

The drop of Icer in VAT and WAT of obese human and mice, respectively, was restricted to the adipocyte fraction (Fig. 2A and B). Indeed, ICER expression was higher in SVF of obese subjects (Fig. 2B) whereas it was unchanged in SVF of obese mice (Fig. 2A). The differences in SVF Icer levels in humans and mice may possibly indicate a differential regulation between species associated with obesity. As anticipated, the decrease in Icer/ICER level in adipocytes paralleled the rise in the mRNA level of Atf3/ATF3 (Fig. 2C and D) and the diminution in Adipoq/ADIPOQ and Glut4/GLUT4 levels (Fig. 2C and D).

To confirm the regulation of Atf3 by Icer, a plasmid encoding for Icer-Iγ was introduced by electroporation into 3T3-L1 adipocytes. Western blotting and EMSA experiments revealed a high expression of Icer (Supplementary Fig. 3). Overexpression of Icer-Iγ led to reduction in Atf3 mRNA (Fig. 3A) and, consequently, an increase of Glut4 and Adipoq levels (Fig. 3A). On the other hand, silencing of Icer by RNA interference elevated the contents of Atf3 and diminished the abundance of Glut4 and Adipoq both in 3T3-L1 adipocytes (Fig. 3B and C) and fibroblasts (Supplementary Fig. 4). As expected, gene dysregulation evoked by Icer silencing resulted in impairment of insulin-induced 2-DOG uptake in 3T3-L1 adipocytes (Fig. 3D).

Augmentation of Creb activity in adipocytes is responsible for the elevation in Atf3 expression in HFD-induced obese mice (9). In turn, the rise of Atf3 reduces Adipoq and Glut4 contents, thus promoting systemic insulin resistance (9). In this report, we provide evidence that defective expression of ICER, a natural antagonist of the CREB transcriptional activity (10) is responsible for increased expression of ATF3 and adipocyte dysfunction caused by CREB in obesity (Fig. 4). We found that diminution of ICER/Icer correlates with an increase in Creb/CREB activity and expression of Atf3/ATF3 in adipose tissues from HFD-induced obese mice and human obese individuals. The adipose reduction of Icer/ICER was specific to adipocytes, indicating that the rise in Atf3 in this fraction is caused by the loss of the repressor. In vitro manipulation of the level of the repressor in 3T3-L1 adipocytes validated Atf3 as a target gene of Icer and the subsequent modification in the expression of Adipoq and Glut4. Finally, silencing the expression of Icer hampered DOG uptake, thus confirming the requirement of appropriate Icer levels for adipocyte function. In contrast to adipocytes, the expression of Icer increased in SVF of human obese individuals when compared with control individuals. This observation suggests that the regulation of ICER differs among cell types, possibly leading to different modulation of CREB activity in other cells located within the adipose tissues. Additional studies will be needed to determine which stromal cell displays increased ICER levels and to investigate the potential functional impact of this elevation.

Creb activity, and consequently the expression of Icer, can be induced by various stimuli, including insulin (24,25). With this respect, we found that insulin elicited a rise in the expression of Icer in control animals whereas insulin-induced expression of the repressor was abrogated in insulin-resistant WAT of obese mice. The fact that the expression of Icer in WAT of obese animals is reduced both in fed and fasting conditions, supports an impaired induction of Icer and, therefore, an abnormal negative control of Creb activity under stimulatory conditions. This dysregulation could result in persistently increased CREB activity, in turn leading to constitutively elevated Atf3 expression and decreased Glut4 and Adipoq levels. Further studies are now needed to identify the mechanism leading to impaired expression of Icer in adipocytes in obesity. We believe that such investigations may pave the way for the design of novel therapeutic strategies or for improvement of existing approaches for recovering normal adipose function and insulin sensitivity.

This work was supported by grants from the Swiss National Foundation (310000-121999, 31003A-127254, and 310030-122586 to A.A., R.R., and G.W., respectively). The authors were also supported by the Placide Nicod and Octav Botnar Foundations. A.A. is supported by the Chair of Excellence from the French National Agency for Research (ANR-10-CEXC-005-01).

No potential conflicts of interest relevant to this article were reported.

D.Fav., A.A., and E.L.G. researched data, contributed to discussion, wrote the manuscript, and reviewed and edited the manuscript. D.Fah., C.V., and V.G. researched data. R.R. researched data, contributed to discussion, and reviewed and edited the manuscript. G.C.-G., B.S., R.C., F.P., K.-A.L., and L.T. researched data and reviewed and edited the manuscript. P.V. and G.W. contributed to discussion and reviewed and edited the manuscript.

The authors are grateful to Eric Bernardi (University of Lausanne) for editorial assistance.