Forkhead box A3 mediates glucocorticoid receptor function in adipose tissue

Abstract

Glucocorticoids (GCs) are widely prescribed anti-inflammatory agents, but their chronic use leads to undesirable side effects such as excessive expansion of adipose tissue. We have recently shown that the forkhead box protein A3 (Foxa3) is a calorie-hoarding factor that regulates the selective enlargement of epididymal fat depots and suppresses energy expenditure in a nutritional- and age-dependent manner. It has been demonstrated that Foxa3 levels are elevated in adipose depots in response to high-fat diet regimens and during the aging process; however no studies to date have elucidated the mechanisms that control Foxa3's expression in fat. Given the established effects of GCs in increasing visceral adiposity and in reducing thermogenesis, we assessed the existence of a possible link between GCs and Foxa3. Computational prediction analysis combined with molecular studies revealed that Foxa3 is regulated by the glucocorticoid receptor (GR) in preadipocytes, adipocytes, and adipose tissues and is required to facilitate the binding of the GR to its target gene promoters in fat depots. Analysis of the long-term effects of dexamethasone treatment in mice revealed that Foxa3 ablation protects mice specifically against fat accretion but not against other pathological side effects elicited by this synthetic GC in tissues such as liver, muscle, and spleen. In conclusion our studies provide the first demonstration, to our knowledge, that Foxa3 is a direct target of GC action in adipose tissues and point to a role of Foxa3 as a mediator of the side effects induced in fat tissues by chronic treatment with synthetic steroids.

Keywords: Foxa3; GR signaling; adipose tissue; dexamethasone; fat expansion.

Fig. 1.

GCs increase Foxa3 transcriptional levels. (A) Putative GRE on the Foxa3 promoter. (B) Luciferase activity of a WT-Foxa3 promoter reporter (Foxa3-luc) and of a mutant-Foxa3 reporter containing a deletion of the putative GRE (Foxa3ΔGRE-luc) in cells treated with DMSO (control, Con) or with 1 µM Dex for 24 h. (C) EMSA analysis assessing GR binding at the Foxa3 promoter. (D) ChIP analysis of GR binding on the putative GRE identified in the Foxa3 promoter or on the Gapdh promoter in cells treated with DMSO or 1 µM Dex for 24 h. (E and F) Dex treatment increased Foxa3 mRNA levels in differentiated 10T/2 cells (E) and BAT primary cells (F) in a dose-dependent manner. (G) Dex treatment increased Foxa3 mRNA levels in adipose tissues. Data are presented as mean ± SEM; *P < 0.05; **P < 0.01 compared with the control (Con) group.

Fig. S1.

Luciferase activity of a Foxa3 promoter reporter in 10T1/2 cells and the effects of Dex treatment on Foxa3 protein levels. (A) Luciferase activity of a Foxa3 promoter reporter in 10T1/2 cells treated with DMSO (Con) or 100 nM T3, 10 µM E2, or 1 µM ATRA for 24 h. (B and C) Dex treatment increased Foxa3 protein levels in differentiated 10T/2 cells (B) and in BAT primary cells (C) in a dose-dependent manner. (D) Dex treatment increased Foxa3 protein levels in adipose tissues. Data are presented as mean ± SEM.

Fig. 2.

Foxa3 is required for GR signaling in adipose tissue. (A) Heatmap of GC signaling genes of iWAT from WT and Foxa3-null mice treated with Dex every other day in the course of 1 wk. (n = 3). (B and C) Representative gene-expression levels of Foxa3-dependent (B) and –independent (C) genes. Data are presented as mean ± SEM; *P < 0.05; **P < 0.01 compared with the control group.

Fig. 3.

Foxa3 is required for GC-mediated adipogenic gene programs in preadipocytes and for lipid metabolism gene programs in mature adipocytes. (A) Heatmap of adipogenic genes in undifferentiated primary SVF or 10T1/2 cells treated with differentiation mixture with or without Dex. (B) Representative gene-expression levels of Foxa3-dependent and -independent genes in SVF and 10T1/2 cells (n = 3). (C) Heatmap of lipid metabolism genes in primary mature adipocytes and in fully differentiated 10T1/2 cells treated with DMSO or Dex. (D) Representative gene-expression levels of Foxa3-dependent and -independent genes in mature adipocytes and differentiated 10T1/2 cells (n = 3). Data are presented as mean ± SEM; *P < 0.05; **P < 0.01 compared with the control group.

Fig. S2.

Foxa3 is required for GR signaling in preadipocytes and adipocytes. (A and B) Foxa3 mRNA levels in undifferentiated (A) and differentiated (B) 10T1/2 cells infected with control or Ad-shFoxa3 and treated with Dex or left untreated. (C and D) Foxa3 overexpression amplified GR gene targets in preadipocytes (C) and adipocytes (D). (E) Foxa3 deficiency impaired GR-induced lipolysis in adipocytes but did not affect cAMP- or PPARα-agonist–induced lipolysis. Data are presented as mean ± SEM; *P < 0.05; **P < 0.01 compared with the control group.

Fig. 4.

Foxa3 is required for GR binding to target gene promoters in preadipocytes and adipocytes. (A) Occupancy of GR, Foxa3, and AcH3 at GRE sites present on the Pparγ2 enhancer in undifferentiated 10T1/2 cells. (B) Occupancy of GR, Foxa3, and AcH4 at the GRE site of the Lipin1 and Angptl4 promoters in differentiated 10T1/2 cells. Data are presented as means ± SEM; *P < 0.05; **P < 0.01 compared with the control group.

Fig. S3.

(A and B) Occupancy of GR, Foxa3, AcH3, and AcH4 at the Gapdh promoter in undifferentiated (A) and differentiated (B) 10T1/2 cells infected with control or Ad-shFoxa3 and treated with Dex or left untreated. (C) Foxa3 and GR plasmids were cotransfected into 10T1/2 cells and treated with DMSO or Dex for 24 h. No interaction between Foxa3 and GR was detected by coimmunoprecipitation analysis. Data are presented as mean± SEM.

Fig. 5.

Foxa3-null mice are protected from increased adiposity in response to chronic GC treatment. Shown are analyses of WT and Foxa3-null mice after 6 wk of control or Dex treatment (n = 4). (A) Fat mass. (B) Adipose tissue weight. (C) Representative H&E staining of eWAT, BAT, and iWAT. (D) Quantification of adipocyte size in eWAT and iWAT. (E) Gene expression in iWAT. Data are presented as means ± SEM; *P < 0.05; **P < 0.01 compared with control group.

Fig. S4.

(A and B) Body weight, lean mass (A) and serum cholesterol and triglyceride levels (B) and related gene programs in the eWAT and BAT of WT and Foxa3-null mice treated with Dex for 6 wk or left untreated (n = 4). (C and D) Analysis of gene programs in the eWAT (C) and BAT (D) of WT and Foxa3-null mice treated with Dex for 6 wk and of untreated controls (n = 4). Data are presented as means ± SEM; *P < 0.05; **P < 0.01 compared with the control group.

Fig. 6.

Effects of GC treatment in liver, muscle, and spleen are independent of Foxa3. Shown are analyses of WT and Foxa3-null mice treated with Dex for 6 wk and untreated controls (n = 4). (A) Glucose and lipid metabolism gene expression in liver. (B) Representative Oil red O staining of liver. (C) Liver triglyceride levels. (D) Expression of genes related to muscle atrophy in gastrocnemius muscle. (E) Serum creatine kinase levels. (F) Muscle weights. (G) Representative H&E staining and relative fiber cross-sectional area of gastrocnemius muscle. (H) Spleen weights. (I) Representative H&E staining of spleen. Data are presented as means ± SEM; *P < 0.05; **P < 0.01; ns, not significant compared with the control group.

Fig. S5.

Effects of GC treatment in liver, kidney, and pancreas are independent of Foxa3. (A) Relative Foxa3 mRNA levels in tissues of WT mice treated with Dex for 6 wk or left untreated, normalized to Foxa3 levels in liver. (B) Weights of liver, kidney, and pancreas in WT and Foxa3-null mice with or without 6 wk of Dex treatment. (C and D) Insulin sensitivity measured by a glucose tolerance test (GTT) (C) and an insulin tolerance test (ITT) (D) in WT and Foxa3-null mice with or without 6 wk of Dex treatment. (n = 4 per group). (E) Heatmap of GC signaling and expression of lipid metabolism genes in livers from WT and Foxa3-null mice with or without 6 wk of Dex treatment (n = 3). Data are presented as means ± SEM; *P < 0.05; **P < 0.01; ns, not significant compared with the control group.

Fig. S6.

Diagram showing the mechanism of Foxa3 facilitating GR signaling in adipocytes. (A) Upon GC treatment, ligand-bound GR translocates into the nucleus. (B) Dimerized GRs bind to the GRE at the Foxa3 promoter and activate Foxa3 transcription. (C) Foxa3 proteins bind to the forkhead response element (FHRE) at promoters of GR target genes and recruit histone acetyltransferases (HATs) for histone acetylation and chromatin remodeling, thus facilitating GR binding at its target gene promoters for transcriptional activation and leading to the expansion of adipose tissue and impaired lipid metabolism in adipocytes.