Hepatic glycogen directly regulates gluconeogenesis through an AMPK/CRTC2 axis in mice

Abstract

Glycogenolysis and gluconeogenesis ensure sufficient hepatic glucose production during energy shortages. Here, we report that hepatic glycogen levels control the phosphorylation of a transcriptional coactivator to determine the amplitude of gluconeogenesis. Decreased liver glycogen during fasting promotes gluconeogenic gene expression, while feeding-induced glycogen accumulation suppresses it. Liver-specific deletion of the glycogen scaffolding protein, protein targeting to glycogen (PTG), reduces glycogen levels, increases the expression of gluconeogenic genes, and promotes glucose production in primary hepatocytes. In contrast, liver glycogen phosphorylase (PYGL) knockdown or inhibition increases glycogen levels and represses gluconeogenic gene expression. These changes in hepatic glycogen levels are sensed by AMP-activated protein kinase (AMPK). AMPK activity is increased when glycogen levels decline, resulting in the phosphorylation and stabilization of CREB-regulated transcriptional coactivator 2 (CRTC2), which is crucial for the full activation of the cAMP-responsive transcriptional factor CREB. High glycogen allosterically inhibits AMPK, leading to CRTC2 degradation and reduced CREB transcriptional activity. Hepatocytes with low glycogen levels or high AMPK activity show higher CRTC2 protein levels, priming the cell for gluconeogenesis through transcriptional regulation. Thus, glycogen plays a regulatory role in controlling hepatic glucose metabolism through the glycogen/AMPK/CRTC2 signaling axis, safeguarding efficient glucose output during fasting and suppressing it during feeding.

Keywords: Cell biology; Gluconeogenesis; Glucose metabolism; Metabolism; Signal transduction.

Figure 1. Depletion of hepatocellular glycogen induces gluconeogenic expression and glucose production.

(A) Glycogen levels in WT and PTGLKO primary hepatocytes. (B) Glucagon receptor (Gcgr) gene expression in WT and PTGLKO hepatocytes. (C) Gene expression in WT and PTGLKO primary hepatocytes treated with 100 nM glucagon for indicated times. (D) cAMP levels in WT and PTGLKO primary hepatocytes. (E) Gene expression in WT and PTGLKO primary hepatocytes treated with different doses of cell-permeable cAMP. (F) Glucagon-stimulated glucose production in WT and PTGLKO hepatocytes. n = 3–6 per group. *P < 0.05; **P < 0.01; ***P < 0.001 by unpaired Student’s t test.

Figure 2. Hepatic glycogen levels regulate gluconeogenesis in a cell-autonomous fashion.

(A) Glycogen levels in WT hepatocytes treated with vehicle (Veh) or glycogen phosphorylase inhibitor (GPI). (B) Gluconeogenic gene expression in vehicle- and GPI-pretreated hepatocytes. Cells were treated with vehicle or GPI overnight and then treated with or without 100 nM glucagon for 4 hours. (C) Schematic model of CRISPR-mediated glycogen phosphorylase (PYGL) knockdown in mice. (D) Glycogen levels in hepatocytes isolated from mice injected with single guide RNA targeting PYGL (sgPYGL) or non-targeting (sgNT) controls. (E) Gluconeogenic gene expression in primary hepatocytes isolated from sgNT and sgPYGL mice treated with or without 1 mM cAMP for 4 hours. (F and G) Glucagon-stimulated glucose production in vehicle- and GPI-treated hepatocytes (F); and cAMP-induced glucose production in sgNT and sgPYGL hepatocytes (G). n = 3–4 per group. *P < 0.05; **P < 0.01; ***P < 0.001 by unpaired Student’s t test. Panel C was created in BioRender (https://BioRender.com/x39s385).

Figure 3. Glucose metabolism is regulated by glycogen levels in vivo.

(A) Glycogen levels in liver lysates from WT and PTGLKO mice. Mice were fasted 4 hours before sacrifice. (B) Glucose levels in WT and PTGLKO mice. (C) Pyruvate tolerance test (PTT) and quantification of area under the curve (AUC). Mice were injected with 1.5 g sodium pyruvate per kilogram body weight. (D) Gluconeogenic gene expression in fasted and fed WT and PTGLKO mice. (E) Gluconeogenic gene expression in liver lysates from WT and PTGLKO mice. Mice were fasted for 1 hour, injected with 0.5 mg/kg glucagon, and sacrificed 30 minutes after. (F) PTT assay of mice injected with AAV-GFP and AAV-PTG, and quantification of AUC. (G) PTT assay of sgNT and sgPYGL mice and quantification of AUC. n = 3–8 per group. *P < 0.05; **P < 0.01; ***P < 0.001 by unpaired Student’s t test (A–C and E–G) and 2-way ANOVA (D).

Figure 4. AMPK activation promotes gluconeogenic gene expression when glycogen levels are low.

(A) Western blots of WT and PTGLKO hepatocytes. (B) Western blots of sgNT and sgPYGL hepatocytes. (C) Gene expression in WT and PTGLKO hepatocytes pretreated with vehicle or compound C followed by glucagon treatment for indicated times. (D) Glucagon-stimulated glucose production in WT and liver-specific AMPKα1/α2 knockout (AMPKLKO) hepatocytes. (E) Gene expression in WT and AMPKLKO hepatocytes treated with glucagon for 4 hours. Data were normalized to vehicle-treated group shown in Supplemental Figure 3G. (F) Gene expression in WT and PTGLKO hepatocytes pretreated with vehicle or GPI followed by vehicle or glucagon treatment. (G) Gene expression in WT and AMPKLKO liver lysates. The mice were injected with saline or 0.5 mg/kg glucagon. n = 3–7 per group. Experiments were performed at least 3 times. *P < 0.05; **P < 0.01; ***P < 0.001 by unpaired Student’s t test. ^###P < 0.001 by unpaired Student’s t test for LKO and LKO + compound C. *** is for WT and LKO comparison.

Figure 5. Phosphorylation of CRTC2 by AMPK promotes gluconeogenic gene expression.

(A) Total and glycogen-bound (AMPD fraction) proteins in WT and PTGLKO primary hepatocytes. Blot quantifications are shown in Supplemental Figure 4A. (B) Nuclear and cytosolic proteins from WT and PTGLKO hepatocytes treated with or without glucagon. Histone H3 and tubulin were used as markers for nuclear and cytosolic fractions. (C) Conserved sequence of CRTC2 across species. Serine 340 and 349 are highlighted. (D) Gluconeogenic gene expression in AML12 cells treated with cAMP. AML12 cells were transfected with vector control, WT, or S³⁴⁹D CRTC2. Data were normalized to vehicle-treated vector group shown in Supplemental Figure 4G. (E) Crtc2 gene expression in transfected AML12 cells. (F and G) Western blots of immunoprecipitated (IP) FLAG-CRTC2 and inputs. AML12 cells were transfected with FLAG-tagged indicated plasmids and treated with vehicle (F) or cAMP (G) for 1 hour. Cells were lysed to harvest input and IP proteins. n = 3–6 per group. Experiments were performed at least 3 times. ^#*P < 0.05; ^##**P < 0.01; ^###***P < 0.001 by 1-way ANOVA. Hatch marks indicate comparison with vector group; asterisks indicate comparison with WT CRTC2.

Figure 6. AMPK stabilizes CRTC2 and increases CRTC2 protein abundance.

(A) Western blots of WT and AMPKLKO hepatocytes treated with vehicle or the AMPK activator PF-739 (PF). (B) Western blots of WT and AMPKLKO liver lysates under refeeding (RF), short-fasting (SF), and long-fasting (LF) conditions. (C) Western blots of WT and PTGLKO liver lysates under SF, LF, and RF conditions. (D) Western blots of AML12 cell lysates. Cells were transfected with WT or S³⁴⁹D CRTC2 and treated with cycloheximide (CHX) for indicated times. (E) Western blots and quantification of proteins from WT and PTGLKO primary hepatocytes. Cells were treated with CHX for indicated times. RalA was used as the control. (F) Western blots of WT primary hepatocytes treated with different doses of glucagon or cAMP. Cells were isolated from C57BL/6J mice and pretreated with vehicle or PF for 2 hours followed by glucagon or cAMP treatment for 15 minutes. (G) Western blots of WT hepatocytes treated with different doses of PF. (H) Gene expression of Nr4a3 and Pgc1a in AML12 cells treated with cAMP. AML12 cells were transfected with vector or CRTC2 and pretreated with vehicle or PF. Data were normalized to vehicle-treated vector group shown in Supplemental Figure 5C. (I) Dose-response curve for the effect of cAMP on gluconeogenic gene expression in WT and AMPKLKO primary hepatocytes. n = 3 per group. Experiments were performed at least 3 times. ^#*^$P < 0.05; ^##**^$$P < 0.01; ^###***^$$$P < 0.001 by 1-way ANOVA and unpaired Student’s t test. * indicates comparison with vector; # indicates comparison with vector + PF; $ indicates comparison with CRTC2.

Figure 7. Working model.

Glycogen levels tune gluconeogenesis in response to nutritional and hormonal cues. Under fasting or glucagon stimulation, decrease in glycogen levels activates AMPK, which phosphorylates and stabilizes CRTC2, increasing its abundance to prime hepatocytes for gluconeogenesis. cAMP inhibits SIK2 activity, permitting the translocation of CRTC2 into the nucleus. The binding of CRTC2 to CREB induces gluconeogenic gene expression. Under fed conditions or insulin stimulation, glycogen accumulation suppresses AMPK activity. Meanwhile, the activation of AKT increases SIK2 activity to sequester CRTC2 in the cytosol. Thus, glycogen levels ensure efficient glucose output during energy shortage and suppress glucose production during energy surplus. This figure was created in BioRender (https://BioRender.com/q31x887).