Altered hepatic glucose homeostasis in AnxA6-KO mice fed a high-fat diet

Abstract

Annexin A6 (AnxA6) controls cholesterol and membrane transport in endo- and exocytosis, and modulates triglyceride accumulation and storage. In addition, AnxA6 acts as a scaffolding protein for negative regulators of growth factor receptors and their effector pathways in many different cell types. Here we investigated the role of AnxA6 in the regulation of whole body lipid metabolism and insulin-regulated glucose homeostasis. Therefore, wildtype (WT) and AnxA6-knockout (KO) mice were fed a high-fat diet (HFD) for 17 weeks. During the course of HFD feeding, AnxA6-KO mice gained less weight compared to controls, which correlated with reduced adiposity. Systemic triglyceride and cholesterol levels of HFD-fed control and AnxA6-KO mice were comparable, with slightly elevated high density lipoprotein (HDL) and reduced triglyceride-rich lipoprotein (TRL) levels in AnxA6-KO mice. AnxA6-KO mice displayed a trend towards improved insulin sensitivity in oral glucose and insulin tolerance tests (OGTT, ITT), which correlated with increased insulin-inducible phosphorylation of protein kinase B (Akt) and ribosomal protein S6 kinase (S6) in liver extracts. However, HFD-fed AnxA6-KO mice failed to downregulate hepatic gluconeogenesis, despite similar insulin levels and insulin signaling activity, as well as expression profiles of insulin-sensitive transcription factors to controls. In addition, increased glycogen storage in livers of HFD- and chow-fed AnxA6-KO animals was observed. Together with an inability to reduce glucose production upon insulin exposure in AnxA6-depleted HuH7 hepatocytes, this implicates AnxA6 contributing to the fine-tuning of hepatic glucose metabolism with potential consequences for the systemic control of glucose in health and disease.

Fig 1. Liver function, lipid and insulin-regulated glucose metabolism are normal in chow-fed AnxA6-KO mice.

(A) Spider diagram presentation of total protein, albumin, aspartate aminotransferase (AST), alanine aminotransferase (ALAT), lactate dehydrogenase (LDH), as well as total cholesterol, triglycerides (TAG), low density lipoprotein (LDL), high density lipoprotein (HDL), and free fatty acids (free FA) in the plasma of 8–12 weeks old wildtype (WT) and AnxA6-KO animals (AnxA6^-/-) on a chow diet (n = 4). The data are presented relative to WT (100%). For absolute values, see text and [28] for further details. (B) H&E staining of liver sections from WT and AnxA6-KO mice on a chow diet. (C) Liver homogenates from 8–12 weeks old WT and AnxA6-KO mice were analyzed for triglyceride, cholesterol and phospholipid content as described in Material and Methods (mean ± SEM, n = 8). SM, sphingomyelin; PC, phosphatidylcholine; PE, phosphatidylethanolamine; PS, phosphatidylserine. (D) WT and AnxA6-KO mice on a chow diet were starved for 4 h prior to receive glucose via oral gavage (1 g/kg body weight). Blood glucose levels were determined 0–120 min after glucose administration (mean ± SEM, n = 6). (E) WT and AnxA6-KO mice received insulin (1.5 U/kg bodyweight) or saline via intraperitoneal injection. Blood glucose was measured at t = 0–120 min post-injection (mean ± SEM, n = 6). (D-E) Data points for t = 0 and 30 min have been described previously [28].

Fig 2. Reduced weight gain and less adiposity of AnxA6-KO mice after HFD feeding.

8 weeks old male WT and AnxA6-KO (AnxA6^-/-, A6^-/-) C57BL/6J mice were fed a high fat diet (HFD) for 17 weeks. (A) Weekly weight gain over the 17 week HFD feeding period. (B) Cumulative weight gain after 17 week HFD feeding (mean ± SEM, n = 25–27). (C) Organ weight comparison after 17 week HFD feeding. AnxA6-KO mice had significantly less white adipose tissue (WAT; epididymal + subcutaneous WAT) than WT mice (mean ± SEM, n = 7–8 for heart, spleen, kidney, BAT; n = 14 for WAT; n = 22 for liver). Brown adipose tissue (BAT). (D) Body temperature over 6 weeks HFD feeding (mean ± SEM, n = 14). (E) Food intake over the HFD feeding period. AnxA6-KO mice consumed slightly more in week 6, 9, 12, and 17, but the area under the curve (AUC) of total food intake over the 17 week HFD feeding period was comparable. The metabolic efficiency (% food stored as body weight) was reduced in AnxA6-KO mice (mean ± SEM, n = 14). (F) Serum from HFD-fed WT and AnxA6-KO mice was analyzed for leptin and adiponectin levels using ELISA as described (mean ± SEM, n = 14). ** P < 0.01, * P <0.05 (Student’s T-test).

Fig 3. Serum triglyceride and cholesterol profiling in HFD-fed WT and AnxA6-KO mice.

(A, C) Total cholesterol and triglyceride levels showed no significant differences between genotypes. (B, D) Pooled serum samples from 3 mice of each genotype were separated by fast-performance liquid chromatography (FPLC) and each fraction was analyzed for cholesterol and triglyceride content. (B) Cholesterol profiling with HDL (fractions #18–20) and neighboring cholesterol-rich fractions (see arrow; #13–17), likely representing HDL- or LDL subspecies. (D) Triglyceride-rich lipoproteins (TRL, fraction #5) and free glycerol (fractions #31–34).

Fig 4. Glucose and insulin tolerance tests (OGTT, ITT) and insulin signaling in HFD-fed WT and AnxA6-KO mice.

(A, B) HFD-fed WT and AnxA6-KO (AnxA6^-/-) mice were starved for 4 h prior to receive glucose via oral gavage (1 g/kg body weight). (A) Blood glucose levels were determined 0–120 min after glucose administration (mean ± SEM, n = 25–27). (B) Plasma insulin levels before and 15 min after oral glucose gavage (mean ± SEM, n = 14). (C-D) HFD-fed WT and AnxA6-KO animals (2 per group) were fasted for 4 h to receive insulin (1.5 U/kg bodyweight) or saline via intraperitoneal injection. (C) Blood glucose was measured at t = 0 min and t = 15 min post-injection. (D) Livers were removed and cytosolic fractions were analyzed by western blotting for phosphorylated and total ribosomal S6 kinase (P-/total S6), Akt kinase (P-/total Akt), insulin receptor substrate 1 (P-/total IRS1) and β-actin. (E-G) Relative levels of activated S6, Akt and IRS1 were quantified and normalized to total S6, Akt and IRS1, respectively. The mean values (± SEM) are shown. * P < 0.05 (Student’s T-test).

Fig 5. Pyruvate tolerance test (PTT) in HFD-fed WT and AnxA6-KO mice.

(A-B) HFD-fed and chow fed (C) WT and AnxA6-KO (AnxA6^-/-, A6^-/-) mice were starved for 4 h prior to intraperitoneal pyruvate administration (2 g/kg body weight). Blood glucose levels were monitored for 0–180 min (mean ± SEM; n = 19 at 0 min, n = 15 at 15, 30, 60 min, n = 10 at 90 min, n = 5 at 120, 180 min); (B-C) Area under the curve (AUC, mean ± SEM) of glucose levels during the PTT of (B) HFD-fed and (C) chow-fed WT and AnxA6-KO mice. (D) Plasma insulin levels of HFD-fed WT and AnxA6-KO mice during the PTT at t = 0, 60 and 120 min following pyruvate administration (mean ± SEM; n = 4–5 for each time point). * P < 0.05, ** P < 0.01; * P <0.05 (Student’s T-test).

Fig 6. Insulin signaling during the PTT of HFD-fed WT and AnxA6-KO mice.

(A) Lysates from liver samples from WT and AnxA6-KO (AnxA6^-/-) mice before (0 min; WT lane 1–4, AnxA6-KO lane 5–8) and 120 min after pyruvate administration (120 min; WT lane 9–12, AnxA6-KO lane 13–16; n = 4 per group) were analyzed by western blotting for phosphorylated/total mTOR (Ser2448), Akt (Ser473), AMPKα (Thr172), and ERK1/2 (Thr202/204). β-actin served as loading control. Molecular weight markers are shown. Black arrowheads point at P-/Total mTOR, respectively. (B) Relative levels of activated mTOR, Akt, AMPKα, and ERK1/2 were quantified and normalized to total mTOR, Akt, AMPKα, and ERK1/2, respectively. The mean values (± SEM) relative to WT at 0 min are shown (* P < 0.05, *** P < 0.001 for Student’s T-Test).

Fig 7. Expression of nuclear transcription factors during the PTT of HFD-fed WT and AnxA6-KO mice.

(A) Nuclear fractions from liver samples from WT and AnxA6-KO (AnxA6^-/-) mice before (0 min; WT lane 1–4, AnxA6-KO lane 5–8) and 120 min after pyruvate administration (120 min; WT lane 9–12, AnxA6-KO lane 13–16; n = 4 per group) were prepared and analyzed by western blotting for the transcription factors FoxO1, PPARα, SREBP1 and LXR. β-actin served as loading control. Molecular weight markers are shown. Arrowhead points at LXR. (B) Relative levels of FoxO1, PPARα, SREBP1 and LXR were quantified and normalized to β-actin expression. The mean values (± SEM) relative to WT at t = 0 min are shown. (C) RNA from HFD-fed WT and AnxA6-KO livers before (0 min) and 120 min after pyruvate administration was isolated (n = 4 per group). cDNA was generated and RT-PCR for AnxA6, fibroblast growth factor 21 (Fgf21), glucose-6 phosphatase (G6P) and insulin induced gene 1 (Insig1) was performed as described in Material and Methods. Relative mRNA expression was normalised to the housekeeper Tbp levels using the ΔΔCT method. The expression relative to the WT at t = 0 min is shown. * P < 0.05, ** P < 0.01, *** P < 0.001 (Student’s T-Test).

Fig 8. Lipid and glycogen accumulation in HFD-fed AnxA6-KO mice.

(A-B, D) Liver tissue sections isolated from (A-B) HFD-fed and (D) chow-fed WT and AnxA6-KO (AnxA6^-/-) mice were analyzed by electron microscopy. Representative images from 4 animals per group are shown. Bar is 0.5–2 μm as indicated. (A) HFD-fed AnxA6-KO mice show fewer, but larger lipid droplets. (B, D) Glycogen accumulation (red arrowheads) and glycogen autophagy (red cross) in HFD- and chow-fed AnxA6-KO mice. Bile canaliculi (BC), lipid droplet (LD), mitochondria (mit), endoplasmic reticulum (ER). (C) Glycogen content in livers from chow-fed WT and AnxA6-KO mice. (E) Primary hepatocytes from chow-fed WT and AnxA6 KO-mice were isolated and lysed to measure glycogen content. The media was collected to determine release of ketone bodies (2-hydoxybutyrate). (F-G) Primary hepatocytes from chow-fed WT and AnxA6 KO-mice were isolated, fasted for 6 h without glucose and then incubated with 2 mM pyruvate or 20 mM glycerol for 24 h as indicated. The media was collected and glucose secretion was determined. Cells were lysed and glycogen was extracted (see Methods for details; mean ± SEM; n = 3). * P <0.05 (Student’s T-test).