Estrogen Deficiency Promotes Hepatic Steatosis via a Glucocorticoid Receptor-Dependent Mechanism in Mice

Abstract

Glucocorticoids (GCs) are master regulators of systemic metabolism. Intriguingly, Cushing's syndrome, a disorder of excessive GCs, phenocopies several menopause-induced metabolic pathologies. Here, we show that the glucocorticoid receptor (GR) drives steatosis in hypogonadal female mice because hepatocyte-specific GR knockout mice are refractory to developing ovariectomy-induced steatosis. Intriguingly, transcriptional profiling revealed that ovariectomy elicits hepatic GC hypersensitivity globally. Hypogonadism-induced GC hypersensitivity results from a loss of systemic but not hepatic estrogen (E2) signaling, given that hepatocyte-specific E2 receptor deletion does not confer GC hypersensitivity. Mechanistically, enhanced chromatin recruitment and ligand-dependent hyperphosphorylation of GR underlie ovariectomy-induced glucocorticoid hypersensitivity. The dysregulated glucocorticoid-mediated signaling present in hypogonadal females is a product of increased follicle-stimulating hormone (FSH) production because FSH treatment in ovary-intact mice recapitulates glucocorticoid hypersensitivity similar to hypogonadal female mice. Our findings uncover a regulatory axis between estradiol, FSH, and hepatic glucocorticoid receptor signaling that, when disrupted, as in menopause, promotes hepatic steatosis.

Keywords: estrogen receptor; follicle-stimulating hormone; glucocorticoid receptor; lipid metabolism; liver; menopause; metabolic syndrome; steatosis.

Figure 1. GCs Drive Metabolic Dysfunction and Steatosis in Ovariectomized Mice

(A) Serum corticosterone levels in sham-operated, ovariectomized, and OVX+ADX mice for 3 months. n = 4 per group. (B) Body weight in sham, ovariectomized, and OVX+ADX mice. n = 5 per group. (C) Body composition measured by dual-energy X-ray absorptiometry (DEXA)-scan of sham, ovariectomized, and OVX+ADX mice. n = 3 per group. (D) Liver weight of sham, ovariectomized, and OVX+ADX mice. n = 4 per group. (E) Glucose levels in sham, ovariectomized, and OVX+ADX mice fasted overnight. n = 3 per group. (F) TG levels measured from sham, ovariectomized, and OVX+ADX mice. n = 3 per group. (G) Representative H&E staining of livers from sham, ovariectomized, and OVX+ADX mice. Scale bars represent 50 μm. (H) Representative images of oil red O-stained livers from sham, ovariectomized, and OVX+ADX mice. Scale bars represent 50 μm. that this pathway may be an alternative therapeutic target for the treatment of menopause-induced metabolic dysfunction. (I) Immunoblot for GR in GR^fl/fl and H-GRKO mice. (J) Body weight of sham and ovariectomized GR^flox/flox and H-GRKO mice. n = 3–8 per group. (K) Representative oil red O staining of livers from ovariectomized GR^flox/flox and H-GRKO mice 3 months after OVX. Scale bars represent 100 μm. (L) Schematic of the experimental design for ADX rescue of ovariectomized mice. (M) Final body weight at the end of the rescue experiment. n = 6–8 mice per group. (N) Percent weight gain at the end of the rescue experiment. n = 6–8 mice per group. (O) Representative oil red O staining of livers from ovariectomized mice receiving either sham or ADX surgery 3 months after OVX. Scale bars represent 100 μm. Data are expressed as mean ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001.

Figure 2. The GR-Governed Hepatic Lipid Metabolism Network Is Reprogrammed in Hypogonadal Female Mice

(A) Correlation heatmap of adrenalectomized ovary-intact and ADX+OVX mice treated with vehicle or dexamethasone for 6 hr. (B) Scatterplot of dexamethasone-regulated genes commonly regulated between ovary-intact and ovariectomized mice. Each dot represents a gene. The x axis represents the dexamethasone (dex)-induced fold change in ADX alone, and the y axis represents the dexamthasone-induced fold change in ADX+OVX mice. (C) GSEA of the GC-controlled lipid metabolism reactome in adrenalectomized and ADX+OVX mice treated with dexamethasone. (D) Heatmap of fragments per kilobase per million (FPKM) values from adrenalectomized and ADX+OVX mice treated with vehicle or dexamethasone, showing the lipogenic (top) (GO_0008610) and lipolytic pathways (bottom) (GO_0016042). Hyper-induced GC targets are outlined in a yellow box. (E–G) FPKM fold change elicited by GCs in ADX and ADX+OVX mice of genes involved in lipid synthesis (E), lipid transport (F), and β-oxidation (G).

Figure 3. Ovariectomy Promotes Enhanced Chromatin Recruitment of GR

(A) qPCR of PLIN5, LCN2, and LPIN1 mRNA in adrenalectomized ovary-intact and ADX+OVX mice treated with either vehicle or dexamethasone for 6 hr. PLIN5, LCN2, and LPIN1 were normalized to PPIB mRNA; n = 3 per group. Data are expressed as relative mRNA to PPIB ± SEM. (B) Chromatin immunoprecipitation of GR to putative GREs in the PLIN5, LCN2, and LPIN1 loci in response to hormone treatment for 1 hr in adrenalectomized ovary-intact and ADX+OVX mouse livers. n = 4 per group. Data are expressed as fold recruitment of GR over adrenalectomized ovary-intact vehicle-treated mice ± SEM. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001.

Figure 4. Systemic but Not Hepatic Estrogen Signaling Deficiency Promotes GC Hypersensitivity

(A) Chromatin immunoprecipitation of GR to the GRE in the PLIN5 loci in hypogonadal female mice treated with dexamethasone for 1 hr, primed with and without estradiol for 72 hr. n = 3–4 animals per group. Data are expressed as fold recruitment of GR over vehicle-treated ovariectomized mice ± SEM. (B) PLIN5 mRNA in ovariectomized mice with and without estradiol priming for 72 hr following 6 hr dexamethasone treatment. Data are expressed as relative PLIN5 mRNA normalized to PPIB mRNA ± SEM. n = 4 animals per group. (C) PLIN5 mRNA in vehicle- and dexamethasone-treated ER-floxed and H-ERKO mice with and without estradiol priming. Data are expressed as relative PLIN5 mRNA normalized to PPIB mRNA ± SEM. n = 3 independent animals per group. *p < 0.05, **p < 0.01, ***p < 0.001, ****p < 0.0001. (D) Representative oil red O staining of livers from ovariectomized ER^flox/flox and H-ERKO mice 2 months after OVX. Scale bars represent 100 μm. n = 3–4 mice per group.

Figure 5. Systemic Estrogen Deficiency Promotes Ligand-Dependent Phosphorylation of GR at Serine211 in the Liver

(A) Western blot analysis of phospho-serine 211 of GR in vehicle- and dexamethasone-treated (1 hr) adrenalectomized ovary-intact and ADX+OVX mice. Data are expressed as percent of vehicle-treated adrenalectomized ovary-intact mice of p211 normalized to total GR ± SEM. n = 4 individual animals per group. (B) Immunoblot for phosphoserine 211 of GR in vehicle- and dexamethasone-treated (1 hr) ovari-ectomized mice with and without estradiol priming for 48/72 hr. Data are expressed as percent of vehicle-treated adrenalectomized ovary-intact mice of p211 normalized to total GR ± SEM. n = 5–6 individual animals per group. **p < 0.01, ****p < 0.0001.

Figure 6. FSH Enhances GR Phosphorylation and Transcriptional Induction of Target Genes

(A) Immunoblot of FSHR and LHR in female mouse livers. β-Actin was used as a loading control, and mouse ovary was used as a positive control. (B) Circulating FSH levels in sham-operated mice and ovariectomized mice treated with either vehicle or estradiol for 48 and 72 hr. Data are expressed as milli-international units per milliliter of FSH ± SEM. n = 5–9 individual animals per group. (C) Immunoblot for p211 GR and total GR in livers of vehicle-, dexamethasone-, and FSH+dex-amethasone-treated adrenalectomized ovary-intact mice. FSH was administered 5 min prior to dexamethasone injection, and livers were harvested 1 hr after dexamethasone injection. Data are expressed as percent of vehicle-treated adrenalectomized ovary-intact mice of p211 normalized to total GR ± SEM. n = 3–4 animals per group. (D) PLIN5, LCN2, and LPIN1 mRNA in vehicle-, dexamethasone-, and FSH+dexamethasone-treated adrenalectomized ovary-intact mice. FSH was administered 5 min prior to dexamethasone injection, and livers were harvested 1 hr after dexamethasone injection. Data are expressed as relative mRNA to PPIB mRNA ± SEM. n = 3–8 individual mice per group. *p < 0.05, **p < 0.01, ***p < 0.001.