Role of ketone signaling in the hepatic response to fasting

Abstract

Ketosis is a metabolic adaptation to fasting, nonalcoholic fatty liver disease (NAFLD), and prolonged exercise. β-OH butyrate acts as a transcriptional regulator and at G protein-coupled receptors to modulate cellular signaling pathways in a hormone-like manner. While physiological ketosis is often adaptive, chronic hyperketonemia may contribute to the metabolic dysfunction of NAFLD. To understand how β-OH butyrate signaling affects hepatic metabolism, we compared the hepatic fasting response in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown mice that are unable to elevate β-OH butyrate production. To establish that rescue of ketone metabolic/endocrine signaling would restore the normal hepatic fasting response, we gave intraperitoneal injections of β-OH butyrate (5.7 mmol/kg) to HMGCS2 knockdown and control mice every 2 h for the final 9 h of a 16-h fast. In hypoketonemic, HMGCS2 knockdown mice, fasting more robustly increased mRNA expression of uncoupling protein 2 (UCP2), a protein critical for supporting fatty acid oxidation and ketogenesis. In turn, exogenous β-OH butyrate administration to HMGCS2 knockdown mice decreased fasting UCP2 mRNA expression to that observed in control mice. Also supporting feedback at the transcriptional level, β-OH butyrate lowered the fasting-induced expression of HMGCS2 mRNA in control mice. β-OH butyrate also regulates the glycemic response to fasting. The fast-induced fall in serum glucose was absent in HMGCS2 knockdown mice but was restored by β-OH butyrate administration. These data propose that endogenous β-OH butyrate signaling transcriptionally regulates hepatic fatty acid oxidation and ketogenesis, while modulating glucose tolerance. NEW & NOTEWORTHY Ketogenesis regulates whole body glucose metabolism and β-OH butyrate produced by the liver feeds back to inhibit hepatic β-oxidation and ketogenesis during fasting.

Keywords: fasting, gluconeogenesis; ketogenesis; β-OH butyrate; β-oxidation.

Fig. 1.

Ketogenic response to fasting in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–D: hepatic HMGCS2 mRNA expression (A), serum β-OH butyrate (µM; B), hepatic acetoacetate (nmol/g liver tissue; C), and β-OH butyrate (nmol/g liver tissue; D). Direct comparisons were made between antisense oligonucleotides (ASO)-injected groups within nutritional state. NS, nonsignificant. P > 0.05. Number inside bar denotes n per group.

Fig. 2.

Glucose homeostasis in response to fasting in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–D: hepatic glycogen (mg/g liver tissue; A), serum glucose (mg/dl; B), insulin (ng/ml; C), and glucose:insulin ratio (D). Direct comparisons were made between antisense oligonucleotides (ASO)-injected groups within nutritional state. NS, nonsignificant; P > 0.05. E: glucose tolerance test in 4-h fasted mice (control and HMGCS2 KD; n = 23). *P < 0.05, significant difference between control and KD mice in a repeated measures ANOVA analysis. F–H: glucose tolerance test area under the curve (AUC; F), glucose-stimulated serum insulin (G), and insulin tolerance test AUC (H). Bars were analyzed by a two-sided unpaired t-test. Number inside bar denotes n per group.

Fig. 3.

Effect of fasting and β-OH butyrate (βHB) on lipid homeostasis in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–D: fed and fasted saline injected control and HMGCS2 KD mice. E–H: fasted saline or βHB injected control and HMGCS2 KD mice. A and E: serum nonesterified fatty acids (NEFAs; µM). B and F: serum triacylglycerol (TAG; mg/dl). C and G: hepatic NEFAs (µmol/g liver tissue). D and H: hepatic TAG (mg/g liver tissue). A–D: direct comparisons were made between antisense oligonucleotides (ASO)-injected groups within nutritional state. E–H; direct comparisons were made between injection (PBS or βHB) within ASO-treated group. NS, nonsignificant. P > 0.05. Number inside bar denotes n per group.

Fig. 4.

Hepatic gene expression in response to fasting in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–F: hepatic peroxisome-proliferator activated receptor-α (PPAR-α) mRNA expression (A), uncoupling protein 2 (UCP2) mRNA expression (B), phosphoenolpyruvate carboxykinase (PEPCK) mRNA expression (C), PEPCK enzymatic activity (D), carnitine palmitoyltransferase 1 (CPT1) mRNA (E), and CPT1 protein expression (F). Direct comparisons were made between ASO-injected groups within nutritional state. NS, nonsignificant. P > 0.05. Number inside bar denotes n per group.

Fig. 5.

Effect of β-OH butyrate (βHB) on fasting glucose homeostasis in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–D: hepatic glycogen (mg/g liver tissue; A), serum glucose (mg/dl; B), insulin (ng/m; C), and glucose:insulin ratio (D). Direct comparisons were made between injection (PBS or βHB) within antisense oligonucleotides (ASO)-treated group. NS, nonsignificant; P > 0.05. Number inside bar denotes n per group.

Fig. 6.

Effect of β-OH butyrate (βHB) on fasting hepatic gene expression in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–F: hepatic peroxisome-proliferator activated receptor-α (PPAR-α) mRNA expression (A), uncoupling protein 2 (UCP2) mRNA expression (B), phosphoenolpyruvate carboxykinase (PEPCK) mRNA expression (C), PEPCK enzymatic activity (D), carnitine palmitoyltransferase 1 (CPT1) mRNA (E), and CPT1 protein expression (F). Direct comparisons were made between injection (PBS or βHB) within antisense oligonucleotides (ASO)-treated group. NS, nonsignificant. P > 0.05. Number inside bar denotes n per group.

Fig. 7.

Effect of β-OH butyrate (βHB) on fasting ketogenesis in control and 3-hydroxy-3-methylglutaryl-CoA synthase II (HMGCS2) knockdown (KD) mice. A–D: hepatic HMGCS2 mRNA expression (A), serum βHB (µM; B), hepatic acetoacetate (nmol/g liver tissue; C), and βHB (nmol/g liver tissue; D). Direct comparisons were made between injection (PBS or βHB) within antisense oligonucleotides (ASO)-treated group. NS, nonsignificant. P < 0.05. Number inside bar denotes n per group.