Regulation of hepatic energy metabolism and gluconeogenesis by BAD

Abstract

The homeostatic balance of hepatic glucose utilization, storage, and production is exquisitely controlled by hormonal signals and hepatic carbon metabolism during fed and fasted states. How the liver senses extracellular glucose to cue glucose utilization versus production is not fully understood. We show that the physiologic balance of hepatic glycolysis and gluconeogenesis is regulated by Bcl-2-associated agonist of cell death (BAD), a protein with roles in apoptosis and metabolism. BAD deficiency reprograms hepatic substrate and energy metabolism toward diminished glycolysis, excess fatty acid oxidation, and exaggerated glucose production that escapes suppression by insulin. Genetic and biochemical evidence suggests that BAD's suppression of gluconeogenesis is actuated by phosphorylation of its BCL-2 homology (BH)-3 domain and subsequent activation of glucokinase. The physiologic relevance of these findings is evident from the ability of a BAD phosphomimic variant to counteract unrestrained gluconeogenesis and improve glycemia in leptin-resistant and high-fat diet models of diabetes and insulin resistance.

Figure 1. Hepatic glucose metabolism in the absence of BAD

(A) Glucose-stimulated lactate production by primary Bad +/+ and −/− hepatocytes (n=5–8). (B) Glucose release by Bad +/+ and −/− hepatocytes treated with lactate/pyruvate (n=6). (C) PTT in Bad +/+ and −/− mice (n=14–20). (D) Lactate production in Bad knockdown hepatocytes 8 hr after glucose stimulation (n=9). (E) Glucose production in Bad knockdown hepatocytes treated with lactate/pyruvate (n=3–5). (F) PTT in C57BL/6J mice after hepatic knockdown of Bad (n=12–17). (G–H) Relative hepatic mRNA levels of Pck1 (G) and L-Pk (H) in fed and overnight fasted C57BL/6J mice after hepatic knockdown of Bad (n=10). (I) PTT in C57BL/6J mice after hepatic knockdown of Gk (n=8–12). (J–K) Relative hepatic mRNA levels of Pck1 (J) and L-Pk (K) in fed and overnight fasted C57BL/6J mice after hepatic knockdown of Gk (n=5). Error bars, ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001. See also Figure S1.

Figure 2. The effect of BAD phosphorylation on gluconeogenesis

(A) Immunoblot analysis and quantification of relative BAD phosphorylation on S155, S136 and S112 in liver samples derived from C57BL/6J mice fasted overnight and re-fed for 6 hr after overnight fasting. (B) PTT in Bad +/+ and S155A knock-in mice (n=10–11). (C) PTT in Bad +/+ and −/− mice following hepatic reconstitution with the indicated adenoviruses (n=16–24). Asterisks in (C) compare Bad −/−: BAD S155D vs. Bad −/−: GFP and the symbols # compare Bad −/−: BAD WT vs. Bad −/−: GFP. (D) Glucose production in primary Bad −/− hepatocytes reconstituted with the indicated adenoviruses and treated with lactate/pyruvate (n=7–12). Error bars, ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001; n.s., non-significant. See also Figure S2.

Figure 3. BAD-dependent modulation of hepatic energy metabolism

(A) Mitochondrial OCR in primary Bad +/+ and −/− hepatocytes transduced with the indicated adenoviruses and treated with lactate/pyruvate (n=4). (B) Mitochondrial OCR in primary Bad +/+ and −/− hepatocytes treated with palmitate (n=10). (C) Etomoxir (Etx) inhibition of OCR in primary hepatocytes treated with lactate/pyruvate (n=5). (D) Respiratory quotient in Bad +/+ and −/− mice (n=8). (E–F) Comparison of mitochondrial mass as examined by the protein levels of Voltage Dependent Anion Channel (VDAC) using western blot analysis (E) and by the fluorescence signal intensity of MitoTracker Green in Bad +/+ and −/− primary hepatocytes (F) (n=3). Scale bar in F, 10 μm. Error bars, ± SEM. **p < 0.01, ***p < 0.001; n.s., non-significant. See also Figure S3 and Table S1.

Figure 4. Hepatic function of BAD is GK-dependent

(A) Mitochondrial OCR in GK-deficient hepatocytes treated with lactate/pyruvate or palmitate (n=8–10). (B–C) Mitochondrial OCR (B) and glucose production (C) in GK-deficient hepatocytes transduced with the indicated adenoviruses (n=9). Error bars, ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001; n.s., non-significant. See also Table S1.

Figure 5. Hepatic insulin resistance in the absence of BAD

A) Modulation of hepatic BAD phosphorylation by insulin in vivo. Immunoblot analysis and quantification of relative BAD phosphorylation on S155, S136 and S112 in liver samples derived from C57BL/6J mice fasted overnight and injected with saline or insulin. (B–G) Euglycemic-hyperinsulinemic clamp analysis in Bad −/− and Bad +/+ mice (n=7–9), showing plasma glucose (B), glucose infusion rate (GIR) (C), hepatic glucose production (HGP) (D), relative mRNA abundance of hepatic gluconeogenic genes (E), hepatic glycogen synthesis rates (F), and peripheral insulin sensitivity (G). Error bars, ± SEM. *p < 0.05; **p < 0.01, Bad−/− vs. Bad+/+ mice. See also Figure S4.

Figure 6. Altered insulin signaling upon hepatic manipulation of BAD and GK

(A) IRS-1-associated PI3K activity in liver of Bad +/+ and −/− mice injected with saline or insulin after an overnight fast (n=10 per genotype). (B) Phosphorylation of the insulin receptor (IR) and AKT in liver of Bad +/+ and −/− mice treated as in (A). Bar graphs quantitate relative insulin induction of IR and AKT phosphorylation (n=4). (C–D) Insulin signaling in liver samples isolated from C57BL/6J mice following hepatic knockdown of Bad (C) and Gk (D). Bar graphs quantitate relative insulin induction of IR, AKT, and GSK-3 phosphorylation (n=4). Error bars, ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001. See also Figure S5.

Figure 7. Metabolic effects of BAD in ob/ob and HFD-treated mice

(A–D) Fasting glucose levels (A), liver Pck1 mRNA abundance (B), pyruvate (C) and glucose (D) tolerance tests following hepatic reconstitution of ob/ob mice with the indicated adenoviruses (n=10–14). Asterisks in (C–D) compare ob/ob mice treated with BAD S155D vs. GFP adenoviruses. (E) GTT in C57BL/6J mice subjected to high-fat diet for 10 weeks prior to hepatic reconstitution with the indicated adenoviruses (n=9). Asterisks in (E) compare HFD mice treated with BAD S155D vs. GFP adenoviruses. Error bars, ± SEM. *p < 0.05; **p < 0.01, ***p < 0.001. See also Figure S6.