Angiopoietin-like protein 4 is an exercise-induced hepatokine in humans, regulated by glucagon and cAMP

Abstract

Objective: Angiopoietin-like protein-4 (ANGPTL4) is a circulating protein that is highly expressed in liver and implicated in regulation of plasma triglyceride levels. Systemic ANGPTL4 increases during prolonged fasting and is suggested to be secreted from skeletal muscle following exercise.

Methods: We investigated the origin of exercise-induced ANGPTL4 in humans by measuring the arterial-to-venous difference over the leg and the hepato-splanchnic bed during an acute bout of exercise. Furthermore, the impact of the glucagon-to-insulin ratio on plasma ANGPTL4 was studied in healthy individuals. The regulation of ANGPTL4 was investigated in both hepatic and muscle cells.

Results: The hepato-splanchnic bed, but not the leg, contributed to exercise-induced plasma ANGPTL4. Further studies using hormone infusions revealed that the glucagon-to-insulin ratio is an important regulator of plasma ANGPTL4 as elevated glucagon in the absence of elevated insulin increased plasma ANGPTL4 in resting subjects, whereas infusion of somatostatin during exercise blunted the increase of both glucagon and ANGPTL4. Moreover, activation of the cAMP/PKA signaling cascade let to an increase in ANGPTL4 mRNA levels in hepatic cells, which was prevented by inhibition of PKA. In humans, muscle ANGPTL4 mRNA increased during fasting, with only a marginal further induction by exercise. In human muscle cells, no inhibitory effect of AMPK activation could be demonstrated on ANGPTL4 expression.

Conclusions: The data suggest that exercise-induced ANGPTL4 is secreted from the liver and driven by a glucagon-cAMP-PKA pathway in humans. These findings link the liver, insulin/glucagon, and lipid metabolism together, which could implicate a role of ANGPTL4 in metabolic diseases.

Keywords: Diabetes; Insulin; Liver; Muscle; Myokine.

Figure 1

Nine healthy male subjects performed 2 h (0–2 h) of one-legged knee extensor exercise. A: Arterial (●) and femoral venous concentrations of angiopoietin-like protein 4 (ANGPTL4) from both the resting (Δ) and the exercising (▲) leg (the curves are superimposed) (MANOVA, Time: P < 0.0001, Group: P = 0.9929, Time × Group: P = 1.000). B: Arterial to venous differences over the resting (Δ) and the exercising (▲) leg of ANGPTL4 (Two-way ANOVA, Time: P = 0.0073, Group: P = 0.6064, Time × Group: P = 0.9954). C: Production (μg/min) or clearance of ANGPTL4 from the resting (Δ) and the exercising (▲) leg (Two-way ANOVA, Time: P = 0.1734, Group: P = 0.4460, Time × Group: P = 0.5284). D: ANGPTL4 mRNA content in the skeletal muscle biopsies from both the resting (○) and the exercising (●) leg (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.9761, Time × Group: P = 0.7075). #: Significant changes from time point 0 h by a 2-way ANOVA followed by a Dunnett's post hoc test. P < 0.05 was considered significant.

Figure 2

Hepato-splanchnic release of angiopoietin-like protein 4 (ANGPTL4) in healthy human subjects during exercise. A: ANGPTL4 concentration in the artery (●) and the hepatic vein (○) before, during (0–2 h) and into recovery after exercise (2–6 h) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.3727, Time × Group: P = 0.9883). B: Arterial-to-venous (μg/l) difference over the hepato-splanchnic bed. A negative value indicates a release whereas a positive value indicates an uptake (One-way ANOVA, Time: P = 0.0017). C: Hepato-splanchnic production of ANGPTL4 is calculated as arterial-to-venous difference multiplied by hepatic plasma flow (μg/min) (One-way ANOVA, Time: P = 0.0014). A positive value indicates a release into the circulation and a negative an uptake. #: Significant changes from time point 0 h after an 2-way ANOVA followed by a Dunnett's post hoc test. (*) designates a borderline significance of the post hoc Dunnett's test at P = 0.08 and P = 0.06 respectively at time point 3 h. The area under the curve for both arterial-to-venous difference (B) and production (C) were significantly different from zero (P < 0.05) by a t-test.

Figure 3

Inhibition of the increase in glucagon-to-insulin ratio by somatostatin infusion in healthy male subjects (n = 5) blunts the exercise-induced increase in plasma angiopoietin-like protein 4 (ANGPLT4) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.0002, Time × Group: P < 0.0001). The subjects performed 2 h of bicycling exercise (0–2 h) on two separate days. On the control day saline was infused (○) and on the pancreatic clamp day somatostatin (●) was infused, see Ref.  for details. $: Significant difference between groups by 2-way ANOVA; *: Significant change from the 0 h time point by one-way ANOVA. P < 0.05 was considered significant.

Figure 4

Effect of increasing the glucagon-to-insulin ratio on plasma angiopoietin-like protein 4 (ANGPTL4) and free fatty acids (FFA) in resting healthy subjects (n = 10). For details, including the glucagon-to-insulin ratio of the study, see Ref. . A–C: Changes in plasma ANGPTL4 concentration by infusion of glucagon and somatostatin (●) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.5002, Time × Group: P < 0.0001) (A), by infusion glucagon (●) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.6127, Time × Group: P < 0.0001) (B), by infusion of somatostatin (●) (Two-way ANOVA, Time: P = 0.0004, Group: P = 0.4160, Time × Group: P < 0.0001) (C) and by saline infusion as control (○) (A–C). Plasma concentrations of glucagon and insulin are presented in Ref. . D–F: Changes in FFA concentration by infusion of glucagon and somatostatin (●) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.0152, Time × Group: P = 0.0515) (D), by infusion of glucagon (●) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.2870, Time × Group: P < 0.0001) (E), by somatostatin (●) (Two-way ANOVA, Time: P < 0.0001, Group: P = 0.0179, Time × Group: P < 0.0001) (F), and by saline (○) (D–F). $: Significant difference between groups by 2-way ANOVA. *: Significant change from the 0 h time point by one-way ANOVA. P < 0.05 was considered significant.

Figure 5

Regulation of angiopoietin-like protein 4 (ANGPTL4), phosphoenolpyruvate carboxykinase (PCK1), and glucose-6-phosphatase (G6PC) mRNA content in HepG2 cells by the cAMP-protein kinase A (PKA) pathway. A–D: Cells were treated with 100 μM 8-(4-chlorophenylthio) (CPT)-cAMP (cAMP), 20 μM forskolin (F) 30 μM H89, 100 μM Rp8-8-bromoadenosine-3′,5′-cyclic monophosporothioate (Rp-8), or 100 nM insulin (Ins) as indicated for 2 h. mRNA abundance related to β-actin is shown as mean ± SEM. *: Significant difference evaluated by t-test. P < 0.05 was considered significant.

Figure 6

The effect of AMPK activation on angiopoietin-like protein 4 (ANGPTL4) mRNA abundance in differentiated human skeletal muscle cells. A: Cells were treated with 250 μM palmitate and 100 μM A769662 for 24 h. B: Cells were treated with 100 μM A769662 or 1 mM 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR) for 3 or 24 h. C: Representative immunoblots of protein lysates (in duplicate) using the indicated antibodies. Relative band intensities normalized to total protein are shown as fold change compared with vehicle-treated cells (mean ± SEM). D–F: Cells were transfected with siRNA oligonucleotides against AMPK catalytic subunit α1 (PRKAA1) and α2 (PRKAA2) and treated with 250 μM palmitate and oleate for 24 h. ANGPTL4 mRNA abundance related to TATA box binding protein (TBP) measured by qPCR is shown as mean ± SEM. *: Significant t-test. P < 0.05 was considered significant.