Apolipoprotein A-IV is induced by high-fat diets and mediates positive effects on glucose and lipid metabolism

Abstract

Objective: Low-carbohydrate, high-fat diets under eucaloric conditions are associated with several health-beneficial metabolic effects in humans, particularly in the liver. We recently observed that apolipoprotein A-IV (apoA-IV), a highly abundant apolipoprotein, was among the most upregulated proteins in circulation after six weeks of consuming a high-fat diet in humans. However, the impact of dietary changes in regulating apoA-IV, and the potential effects of apoA-IV on regulation of glucose- and lipid metabolism remain to be fully established.

Methods: We investigated the regulation of circulating fasting concentrations of apoA-IV in humans in response to diets enriched in either fat or carbohydrates. Moreover, to study the whole-body and tissue-specific glucose and lipid metabolic effects of apoA-IV, we administrered apoA-IV recombinant protein to mice and isolated pancreatic islets.

Results: We demonstrate that in healthy human individuals high-fat intake increased fasting plasma apoA-IV concentrations by up to 54%, while high-carbohydrate intake suppressed plasma apoA-IV concentrations. In mice, administration of apoA-IV acutely lowered blood glucose levels both in lean and obese mice. Interestingly, this was related to a dual mechanism, involving both inhibition of hepatic glucose production and increased glucose uptake into white and brown adipose tissues. In addition to an effect on hepatic glucose production, the apoA-IV-induced liver proteome revealed increased capacity for lipoprotein clearance. The effects of apoA-IV in the liver and adipose tissues were concomitant with increased whole-body fatty acid oxidation. Upon glucose stimulation, an improvement in glucose tolerance by apoA-IV administration was related to potentiation of glucose-induced insulin secretion, while apoA-IV inhibited glucagon secretion ex vivo in islets.

Conclusions: We find that apoA-IV is potently increased by intake of fat in humans, and that several beneficial metabolic effects, previously associated with high fat intake in humans, are mimicked by administration of apoA-IV protein to mice.

Keywords: Adipose tissue; Chylomicron; Diet; Fatty acid oxidation; Hepatic glucose production; Incretin hormone; Insulin secretion; Liver.

Figure 1

Dietary regulation of fasting plasma apoA-IV concentration in human individuals. Fasting plasma concentration of apoA-IV was measured in the overnight fasted state following either A) five days eucaloric intake of high-fat low-carbohydrate (LC) diet (70E% fat, 14E% carbohydrate) or low-fat high-carbohydrate (HC) diet (14E% fat, 70E% carbohydrate), and B) three days intake of a high-fat unsaturated diet (HFUNSAT) (78E% fat, 10 E% carbohydrate, 12 E% protein), a high-fat saturated diet (HFSAT) (83E% fat, 9E% carbohydrate, 12E% protein), or a high-carbohydrate diet (CHO) (80 E% carbohydrate, 11 E% protein, 9 E% fat) during 75% caloric excess, compared with a eucaloric control diet (24E% fat, 62E% carbohydrate, 14E% protein). Data are means +/− SEM. Two-way RM ANOVA was applied in A, with Sidak’s multiple comparisons post-hoc test when interaction was detected by ANOVA, paired t-test was applied in B. ∗p < 0.05, ∗∗p < 0.01 difference between the respective diet intervention and control diet/pre-intervention. #p < 0.05 difference between post-LC and post-HC. n = 11 in A, n = 8 in B, except for CHO group, where n = 5.

Figure 2

Acute apoA-IV administration lowers fasting glucose levels in an insulin-, sympathetic activity- and GFRAL independent manner and relates to increased adipose tissue glucose disposal in mice. Plasma concentration of human recombinant apoA-IV protein (A) and fasting blood glucose level (B) at indicated time points in mice injected at 0 h with vehicle (veh) or 1 mg/kg apolipoprotein A-IV (apoA-IV) (injection timing indicated by black arrow) (n = 8 in all groups). Fasting plasma insulin concentration (C) at 1 h–4 h following injection. D. Basal fasting glucose clearance into indicated tissues (BAT: brown adipose tissue, SKM: skeletal muscle, eWAT: epididymal white adipose tissue) assessed 150 min after vehicle or apoA-IV injection. E. Blood glucose before and 3 h after vehicle or apoA-IV-injection, with or without sympathetic inhibition by propranolol (pro). F. Representative blots. G. Phosphorylation level and protein content of proteins in eWAT (G–H) and BAT (I–J). K. Plasma growth differentiation factor 15 (GDF15) concentration 4 h after vehicle or apoA-IV-injection. L. Blood glucose before and 2, 3, and 4 h after vehicle or apoA-IV-injection in wildtype GDNF family receptor alpha like (GFRAL) knock-out mice. Data from western blot analyses are presented relative to the vehicle group. In all experiments, female mice were fed 60% HFD for eight weeks, and fasted prior to the experiments. Data are means ± SEM. Two-way RM ANOVA was applied in B–C, E and L, with Sidak’s post hoc test whenever interactions were detected. Unpaired t-test was applied in D, G-K. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 difference between apoA-IV and vehicle (or vehicle + vehicle in E and within genotype in L). ˆˆˆ p < 0.001 difference between apoA-IV and vehicle + pro. (For interpretation of the references to color in this figure legend, the reader is referred to the Web version of this article.)

Figure 3

Acute apoA-IV administration leads to suppression of basal hepatic glucose production and ex vivo pancreatic glucagon secretion, while changing liver protein abundance in direction of plasma lipoprotein clearance. A. Fasting hepatic glucose production, measured over 75 min by continuous 3-³H-glucose infusion in anesthetized mice, i.e., at 150 min–225 min following vehicle (vehicle, n = 7) or 1 mg/kg apolipoprotein A-IV (apoA-IV, n = 9) administration. B. Glucagon secretion of isolated islets incubated with vehicle or apoA-IV (n = 6). Plasma glucagon (C) and blood glucose (D) concentration after injection with vehicle or apoA-IV (n = 10). E + F. Phosphorylation level and protein content of liver proteins obtained in the fasted state 4 h following vehicle or apoA-IV treatment (n = 9). G. Representative western blot of liver samples. H. Volcano plot illustrating changes in the liver proteome with apoA-IV compared with vehicle administration, with liver tissue obtained 6 h post-injection. The significance level was set at p < 0.05. Mice were in all experiments fed 60% HFD for 8 weeks and fasted prior to the experiments. Data are means +/− SEM. Two-way RM ANOVA was applied in B-D, with Sidak’s post hoc test whenever interactions were detected by ANOVA. Unpaired t-test was applied in A, E-F. ∗p < 0.05, ∗∗p < 0.01 ∗∗∗p < 0.001 difference between apoA-IV and vehicle (or main effect of apoA-IV). ##p < 0.01 effect of glucose concentration.

Figure 4

Acute administration of apoA-IV increases glucose tolerance in obese mice, associated with increased glucose-stimulated but not basal fasting insulin secretion. Blood glucose (A), blood glucose incremental area under curve (iAUC) (B) and plasma insulin concentration (C) in mice subjected to an intraperitoneal glucose tolerance test (GTT) (2 g glucose/kg BW) 4 h after injection with vehicle (veh) or apolipoprotein A-IV (apoA-IV; 1.0 mg/kg apoA-IV). Blood glucose (D), plasma insulin (E) and plasma C-peptide (F) concentrations measured for 4 h at fasting conditions and 20 min after intraperitoneal administration of 2 g/kg glucose. The effect of 0.5 mg/kg BW apoA-IV, 5 mg/kg BW apoA-I, and co-administration of 0.5 mg/kg apoA-IV + 5 mg/kg apoA-I were compared in G-I, showing blood glucose (G), blood glucose incremental area under curve (iAUC) (H) and plasma insulin concentration (I) before and during a 2 g glucose/kg intraperitoneal GTT 4 h after injection. Mice were in all experiments fed 60% HFD for 16 weeks, and fasted prior to the experiments. n = 8 in A-F, n = 6 in G-I. Data are means +/− SEM. Two-way RM ANOVA was applied in A, C, D-I, with Sidak’s multiple comparisons post-hoc test when interactions were detected by ANOVA. Unpaired t-test was applied in B. ∗∗p < 0.01, ∗∗∗p < 0.001 are differences between apoA-IV and vehicle. ˆ p < 0.05, ˆˆˆp < 0.001 apoA-I versus vehicle. ##p < 0.01, ###p < 0.001 apoA-IV + apoA-I versus vehicle. $p < 0.05 apoA-IV versus apoA-IV + apoA-I.

Figure 5

Acute administration of apoA-IV increases whole-body fatty acid oxidation. Metabolic chamber oxygen uptake (VO₂) (A + D), respiratory exchange ratio (RER) (B + E) and activity level (C + F) in lean chow-fed mice (A–C) and in obese high-fat diet fed mice (D–F) measured for 20 h post-injection in the light and dark cycle after injection with vehicle (veh) or 1 mg/kg apolipoprotein A-IV (apoA-IV) at 11 a.m. (n = 8 in all groups). Mice had ad libitum access to their respective diets. Data are means +/− SEM. Two-way RM ANOVA was applied in A-F, with Sidak’s post hoc test when interactions were detected by ANOVA. ∗p < 0.05, ∗∗p < 0.01 difference between vehicle and apoA-IV treated mice.

Figure 6

Acute administration of apoA-IV lowers food intake, and prolonged daily administration leads to weight loss and increased glucose tolerance in obese mice. A. Female mice fed 60% HFD were injected with vehicle (veh) or 1 mg/kg apolipoprotein A-IV (apoA-IV) and cumulative food intake was assessed after 24 h (B). Three groups of mice were either ad libitum fed the HFD and vehicle-treated (vehicle), ad libitum fed HFD and apoA-IV treated (apoA-IV) and vehicle-treated while pair fed the HFD to the energy intake of the apoA-IV-treated mice (vehicle pair fed) for eight days, with daily injection at 12 a.m. (C). Body weight is shown as % of initial body weight (D). The total %weight lost is shown in E. At day nine, mice were fasted in the morning and subjected to an intraperitoneal glucose tolerance test (2 g glucose/kg BW) with blood glucose levels shown in F. n = 8 in all groups. Data are means +/− SEM. Unpaired t-test was applied in B. Two-way RM ANOVA was applied in D + F, with Sidak’s post hoc test when interactions were detected by ANOVA. One-way ANOVA was applied in E. ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001 difference between vehicle and apoA-IV. #p < 0.05 difference between vehicle and vehicle pair fed mice.