Abstract
AMP-activated protein kinase (AMPK) is a regulator of energy balance at both the cellular and the whole-body levels. Direct activation of AMPK has been highlighted as a potential novel, and possibly safer, alternative to treat type II diabetes and obesity. In this study, we aimed to design and characterize novel peptides that mimic the αG region of the α2 AMPK catalytic domain to modulate its activity by inhibiting interactions between AMPK domains or other interacting proteins. The derived peptides were tested in vivo and in tissue culture. The computationally predicted structure of the free peptide with the addition of the myristoyl (Myr) or acetyl (Ac) moiety closely resembled the protein structure that it was designed to mimic. Myr-peptide and Ac-peptide activated AMPK in muscle cells and led to reduced adipose tissue weight, body weight, blood glucose levels, insulin levels, and insulin resistance index, as expected from AMPK activation. In addition, triglyceride, cholesterol, leptin, and adiponectin levels were also lower, suggesting increased adipose tissue breakdown, a result of AMPK activation. On the other hand, liver weight and liver lipid content increased due to fat retention. We could not find an elevated pAMPK:AMPK ratio in the liver in vivo or in hepatocytes ex vivo, suggesting that the peptide does not lead to AMPK activation in hepatocytes. The finding that an AMPK-derived peptide leads to the activation of AMPK in muscle cells and in adipose tissue and leads to reduced glucose levels in obese mice, but to fat accumulation in the liver, demonstrates the differential effect of AMPK modulation in various tissues.
Introduction
Type II diabetes is among the main threats to human health in the twenty-first century, and in most countries, diabetes is now one of the leading causes of mortality through its effects on the cardiovascular system. Type II diabetes is characterized by insulin resistance and/or abnormal insulin secretion, resulting in elevated blood glucose levels (Zimmet et al. 2001). Current trends in drug prescribing for the treatment of insulin resistance indicate a move away from agents that stimulate insulin secretion, such as sulfonylureas, toward agents that increase insulin sensitivity, such as the biguanide metformin and thiazolidinediones (Hardie 2007). Due to concerns regarding cardiovascular side effects of thiazolidinediones (Nissen & Wolski 2007), the US Food and Drug Administration (FDA) issued the toughest safety warning on two diabetes drugs, Avandia and Actos. Thus, nowadays metformin is the first-line treatment for type II diabetes, in particular, in overweight and obese people and those with normal kidney function.
Recent findings indicate that metformin and thiazolidinediones indirectly activate the AMP-activated protein kinase (AMPK; Batandier et al. 2006). Direct activation of AMPK has been highlighted as a potential novel, and possibly safer, alternative for type II diabetes and obesity drug development (Hardie 2007). AMPK is a regulator of energy balance at both the cellular and the whole-body levels (Carling et al. 2011, Hardie 2011). Typically quiescent under normal conditions, AMPK is activated in response to hormonal signals or stress, such as hypoglycemia, strenuous exercise, and ischemia. The activation of AMPK is dependent on the ratio of AMP to ATP: while AMP activates AMPK, ATP has an inhibitory affect. When intracellular ATP levels are low, active AMPK inhibits numerous ATP-consuming pathways while activating the mobilization of intracellular energy stores to produce ATP. Conversely, when ATP levels are high, AMPK activity is inhibited, leading to the enhancement of energy storage pathways and increased energy usage for biosynthetic pathways (Kahn et al. 2005). Hypothalamic AMPK regulates energy intake by mediating the opposing effects of orexigenic and anorexigenic signals (Minokoshi et al. 2004). In peripheral tissues, AMPK is phosphorylated and activated under elevated AMP levels by liver kinase B1 (LKB1; Sanders et al. 2007) following hormonal signals, such as adiponectin (Yoon et al. 2006). Consequently, activated AMPK phosphorylates and, thus, inactivates acetyl-CoA carboxylase (ACC), the key enzyme in fatty acid synthesis. Reduced fatty acid synthesis results in mitochondrial fatty acid oxidation and activation of catabolic pathways (Yoon et al. 2006). In vitro, AMPK can be activated by 5-aminoimidazole-4-carboxamide ribonucleotide (AICAR), an AMP analog, and inhibited by Compound C, an ATP-competitive inhibitor (Viollet et al. 2006). In the present study, we aimed to design novel AMPK modulators.
Most traditional drug discovery campaigns aim to develop small-molecular weight, orally available compounds, targeting catalytic or ligand-binding sites (Arora & Ansari 2009). Due to the constant decrease in productivity of drug design efforts (Booth & Zemmel 2004, Scannell et al. 2012), novel approaches, including targeting of protein–protein interactions (PPIs), are emerging in the drug discovery field. PPIs may be targeted by small molecules (Wells & McClendon 2007) or by peptidic and peptidomimetic compounds (Arora & Ansari 2009, Rubinstein & Niv 2009). We have recently designed novel and highly specific kinase-inhibiting peptides (Niv et al. 2004, Wexler et al. 2005, Mack et al. 2008, Rubinstein & Niv 2009) for the prevention of insulin resistance and type II diabetes. In the present study, we designed peptides that mimic the αG region of the AMPK catalytic domain, a region typically involved in PPIs of kinases (Niv et al. 2004), to modulate its activity by inhibiting interactions with other AMPK domains or with interacting proteins, and examined the metabolic effects of such modulation.
Materials and methods
Animals, treatments, and tissues
Eight-week-old C57BL/6 or leptin-deficient ob/ob male mice were housed in a temperature- and humidity-controlled facility (23–24 °C and 60% humidity). Mice were entrained to a light–dark cycle of 12 h light:12 h darkness cycle for 2 weeks with food available ad libitum. After 2 weeks, mice were fed a high-fat (HF) diet for 10 weeks. The HF diet was based on soybean oil and palm stearin (fatty acid composition: C:12 0.3%, C:14 1.3%, C:16 55%, C:18 5.1%, C:18-1 29.5%, C:18-2 7.4%, and C:18-3 0.7%) and contained 22% w/w fat (42% kcal from fat). After 10 weeks of HF feeding, at an average body weight of ∼31 g, C57BL/6 mice were injected subcutaneously with 0.4 mg/g body weight per day of myristoylated (Myr) or acetylated (Ac)-peptide or 0.4 mg/g body weight of Myr-peptide or Ac-peptide every other day for 12 days (n=10 in each group). The ob/ob mice were injected with 0.06 mg/g body weight of Myr-peptide for 12 days. Depending on the experiment, during the injection period, mice were given either a HF diet or regular chow. Daily food intake and body weight were monitored throughout the experiment. Mice were fasted for 10 h before tissue removal. Mice were anesthetized with isoflurane and blood, liver, gastrocnemius muscle, and white adipose tissue (WAT) were removed at the end of the experiment. After tissue removal, mice were humanely killed. Fasting blood glucose levels were determined using a glucometer (Optium Xceed; Abbott Laboratories). Tissues were immediately frozen in liquid nitrogen and stored at −80 °C until further analysis. The joint ethics committee (IACUC) of the Hebrew University and Hadassah Medical Center approved this study.
Tissue culture experiments
Mouse-derived C2C12 myoblasts were maintained in DMEM supplemented with 20% (v/v) bovine calf serum. Differentiation of cells to myotubes was achieved by allowing the cells to reach confluence in 0.1% gelatin-coated flasks in DMEM supplemented with 10% (v/v) fetal calf serum (FCS). When the cells reached confluence, the medium was replaced with DMEM supplemented with 2% horse serum. Differentiation was achieved after 72–96 h. AML-12 hepatocytes were grown in DMEM supplemented with 10% (v/v) FCS and 200 mM glutamine for 3 days. Peptides were added to the medium at 100 μM for 1 h, after which the cells were harvested and total and phosphorylated AMPK and ACC were measured by western blotting.
Serum separation and ELISA
Blood samples were kept at room temperature for 30 min for clotting and consequently centrifuged at 2000 g for 15 min. Serum was collected and stored at −80 °C for further analysis. Serum hormone levels were determined for insulin (Mercodia, Uppsala, Sweden), corticosterone (Assaypro, St Charles, MO, USA), and leptin and adiponectin (R&D Systems, Inc., Minneapolis, MN, USA) using ELISA kits. Assays were performed according to the manufacturers' instructions.
Enzymatic colorimetric tests
Serum triglyceride, total cholesterol, HDL-cholesterol, alanine aminotransferase (ALT), and aspartate transaminase (AST) levels were determined using Cobas kits (Roche Diagnostics) and analyzed with a Roche/Hitachi analyzer (Roche Diagnostics). Assays were performed according to the manufacturers' instructions.
Homeostasis model assessment of insulin resistance
The insulin resistance index from fasting serum insulin and plasma glucose levels was determined using the homeostasis model assessment (HOMA) parameter: HOMA=fasting serum insulin (μU/ml)×fasting plasma glucose (mg/dl)/405.
Circular dichroism spectroscopy
Circular dichroism (CD) spectra of the peptide were recorded on a JASCO J-810 spectrophotometer (JASCO, Tokyo, Japan) using the supplied Spectra Manager software. The temperature was kept constant using a temperature-controlled water bath. Samples (400 μl) containing 100 μM peptide and 10 mM phosphate buffer were placed in 0.2 cm quartz cells (Starna, Atascadero, CA, USA), and spectra were recorded in the wavelength range λ=195–260 nm, with 5–10 accumulations for each measurement and a data pitch of 0.1 nm. Background CD spectra of the buffer were recorded and subtracted from each spectrum.
Peptide design
The peptide was derived from specificity-determining patches in AMPK catalytic domain, the αG helix in the human catalytic domain of AMPK α2 (accession P54646). The corresponding 215HVPTLFKKIRG225 sequence was amidated and conjugated either with Ac-Gly (Ac-peptide Ac-GHVPTLFKKIRG-NH2) or with Myr-Gly (Myr-peptide Myr-GHVPTLFKKIRG-NH2) as has been reported previously (Niv et al. 2004, Mack et al. 2008). The peptides were synthesized by GL Biochem at purity >95%.
Peptide structure prediction
Starting from an extended conformation, peptides were energy-minimized using conjugate gradient until the maximal force was below 1 kJ/mol or when no change in energy occurred. Subsequently, peptides were subjected to annealing molecular dynamics simulation of 300 ns with repeated cycles of the following heating and cooling sequence protocol: annealing time (ps): 0, 11, 70, 246, 257, 268, 327, 503, 522, 533, 592, 768, and 799; annealing temperature (oC): 275, 290, 525, 290, 275, 290, 431, 290, 275, 290, 375, 290, and 275. The simulations were carried out with the GROMACS package version 4.5 (Pronk et al. 2013), in implicit solvent conditions (Onufriev et al. 2004), the AMBER96 force field, and virtual hydrogen representations that allow a large integration time step of 5 fs. No cutoffs were used; snapshots were saved every 1 ps. Bonds were constrained with the Lincs algorithm using Lincs order 2, with six iterations, to cancel out the possibility of the ‘flying ice cube’ phenomenon (Harvey et al. 1998). Snapshots along the trajectory at temperatures of 290–300 oC were clustered using the Gromos method, with backbone root mean square deviation (RMSD) cutoff of 1 Å. The cluster-representative structures were sorted by the average energy of the best 15 poses of the cluster.
Lipid quantification
Lipid quantification in the liver was performed using the Folch reagent. Liver tissue samples weighing 100 mg were homogenized in a mixture of chloroform:methanol (2:1) and left shaking for 30 min. The samples were then centrifuged at 2500 g for 30 min, and water was added. The samples were left open for evaporation and the test tubes were weighed. Lipid quantification in AML-12 cells was performed using Oil Red O staining. The cells were fixed in 10% formaldehyde in aqueous phosphate buffer overnight, washed with 60% isopropanol, and stained with Oil Red O solution (in 60% isopropanol) for 10 min. The cells were then repeatedly washed with water and destained in 100% isopropanol for 15 min. The optical density of the isopropanol solution was measured at 500 nm.
Western blot analysis
C2C12 or AML-12 cells were harvested, and liver and WAT samples (∼200 mg) were homogenized in lysis buffer (20 mM Tris, 145 mM NaCl, 10% glycerol, 5 mM EDTA, 1% Triton-X, 0.5% Nonidet P-40, 100 μM phenylmethylsulfonyl fluoride, 50 μM NaF, and 1 mM sodium orthovanadate). The samples were run on a 10% SDS–polyacrylamide gel for AMPK, pAMPK (pThr172), ACC, and pACC (pSer79). After electrophoresis, proteins were semi-dry-transferred onto nitrocellulose membranes. Blots were incubated with anti-mouse AMPK/pAMPK and ACC/pACC, polyclonal antibody (Cell Signaling Technology, Beverly, MA, USA). Anti-mouse actin antibody (MP Biomedicals, Solon, OH, USA) was used to detect actin, the loading control. The membranes were washed and reacted with horseradish peroxidase (HRP)-conjugated anti-goat antibodies (Santa Cruz Biotechnologies). The immune reaction was detected by enhanced chemiluminescence (Santa Cruz Biotechnologies). Finally, bands were quantified by scanning and densitometry and expressed as arbitrary units.
Statistical analyses
All results are expressed as means±s.e.m. Levels were assessed by one-way ANOVA, and a least-significant difference Tukey–Kramer post hoc analysis was used for comparison among the groups. For all analyses, the significance level was set at P<0.05. Statistical analyses were carried out with the JMP 7.0.2 software (SAS Institute, Inc., Cary, NC, USA).
Results
Design of AMPK-derived peptides
With the goal of specifically modifying PPIs of AMPK, we designed a peptide mimicking the αG region of the catalytic kinase domain (Fig. 1A and B). This region is typically involved in PPIs of kinases (Niv et al. 2004, Scholz et al. 2009). To evaluate the most stable conformations of the peptide in solution, we carried out molecular dynamics simulations of the Myr-peptide and Ac-peptide. We found that the most stable conformations of both the Myr-peptide and Ac-peptide were helical. The simulated structures perfectly fit the helical secondary structure of the native AMPK protein (Fig. 1C). The only deviation observed was the orientation of the first histidine residue, which is emphasized in stick representation in Fig. 1C. Excluding the histidine, the predicted lowest-energy structure of the largest conformational cluster was very similar to the native structure (the backbone and heavy atom RMSD of as low as 1 and 2.6 Å respectively). Independent simulations with different possible protonation states of the histidine (Ben-Shimon et al. 2013) yielded similar results (not shown). CD analysis of the peptide confirmed the helical conformation (Fig. 1D). We conclude that the derived peptide closely resembles the protein structure that it was designed to mimic, and the addition of the Myr or Ac moiety seems to influence only the orientation of the N-terminal (histidine) residue in the peptide.
Design and structure of the Myr-peptide. (A) The peptide was derived from the αG helix of the AMPK kinase catalytic domain α2. AMPK (pdb 2H6D) is shown in gray and the region from which the peptide was derived in orange. (B) Peptide sequence of Myr-peptide. (C) Peptide structure prediction using simulated annealing molecular dynamics protocol (SAMDp). Three most energetically stable conformations obtained from molecular dynamics simulation of the peptide in implicit solvent are superimposed on the corresponding native AMPK region (orange) and colored blue, pale blue, and cyan. RMSD of the peptide calculated with respect to the native AMPK region vs potential energy of the peptide as sampled during the SAMDp procedure is shown. The myristoyl group bonded to the first glycine residue, which is not part of the native AMPK sequence, was included in the simulations, but is not shown. (D) CD analysis of Myr-peptide. CD results obtained at 25, 37, and 50 °C confirmed the predicted helical content of this peptide.
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
Activity characterization of Myr-peptide and Ac-peptide in C2C12 myotubes
We next studied the effect of Myr-peptide and Ac-peptide on AMPK activity in a muscle cell line (C2C12 myotubes). Peptides were added to the medium at 100 μM for 1 h, after which the cells were harvested and measured by western blotting for the levels of total AMPK and ACC and their phosphorylated forms. Both Myr-peptide and Ac-peptide elevated AMPK phosphorylation, which is reflective of its activation, with Myr-peptide having the highest activity. In parallel, the phosphorylation of the AMPK substrate, ACC, the rate-limiting enzyme in fatty acid synthesis, also increased (P<0.05, Tukey's HSD) (Fig. 2). The effect of Myr-peptide was comparable to the activating effect of the in vitro reagent AICAR, namely a ∼2.4-fold increase in AMPK phosphorylation (Fig. 2A) and a ∼3.4-fold increase in ACC phosphorylation (Fig. 2B). Taken together, these results demonstrate that Myr-peptide and Ac-peptide activate AMPK in muscle cells.
Modulation of AMPK and ACC in C2C12 myotubes treated with AMPK-derived peptides or AICAR. (A) pAMPK:AMPK ratio. (B) pACC:ACC ratio. Peptides were added to the medium of C2C12 cells at 100 μM for 1 h or 1 μM AICAR, after which the cells were harvested and measured by western blotting for the levels of total AMPK and ACC and their phosphorylated forms. Different letters denote statistical significance (P<0.05).
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
Effect of Myr-peptide and Ac-peptide on body weight and food intake in mice
As AMPK activation leads to overall satiety and lipid oxidation, we next studied the effect of Myr-peptide and Ac-peptide in vivo using diet-induced obese C57BL mice. Our results showed that obese mice treated with s.c. peptide injection of 0.4 mg/g body weight per day for 12 days gained significantly less weight (P<0.05, Tukey's HSD) (Fig. 3A) with no reduction in average food intake (Fig. 3C). This reduction in body weight correlated with decreased adipose tissue weight (Fig. 3E). Injection of 0.4 mg/g body weight every other day led to an almost significant change in body weight (P=0.06, Tukey's HSD) (Fig. 3B) with no change in food intake (Fig. 3D) and no reduction in adipose tissue weight (Fig. 3F). The activity of Myr-peptide was also confirmed in genetically obese leptin-deficient ob/ob mice. These mice treated with s.c. peptide injection of 0.06 mg/g body weight per day (a typical dosage used for metformin) gained less weight than saline-injected controls (Fig. 3G) and consumed less food (Fig. 3H). Taken together, these results demonstrate that Myr-peptide and Ac-peptide lead to reduced adipose tissue weight and body weight, suggesting the activation of AMPK, at least in adipose tissue.
Effect of Myr-peptide or Ac-peptide injection on the body weight, food intake, and adipose tissue of mice fed a HF diet. (A) Mean body weight of mice injected every day for 12 days. The final body weight is shown in the inset. (B) Mean body weight of mice injected every other day for 12 days. The final body weight is shown in the inset. (C) Food intake of mice injected every day for 12 days. (D) Food intake of mice injected every other day for 12 days. (E) Mean adipose tissue weight of mice injected every day for 12 days. (F) Mean adipose tissue weight of mice injected every other day for 12 days. (G) Mean body weight of ob/ob mice injected every other day for 12 days. (H) Food intake of ob/ob mice injected every other day for 12 days. After 10 weeks of HF feeding, at an average body weight of ∼31 g, C57BL mice were injected subcutaneously with 0.4 mg/g body weight per day of Myr-peptide (Myr), Ac-peptide (Ac), or saline for 12 days or 0.4 mg/g body weight every other day for 12 days. The ob/ob mice were injected every other day with 0.06 mg/g body weight. During the injection period, mice were given a HF diet. n=10 in each group. *P<0.05.
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
Effect of Myr-peptide on metabolic parameters
Our goal was to inject the lowest concentration possible and our preliminary results revealed that glucose levels change after every-other-day injections. In addition, as Myr-peptide and Ac-peptide were comparable in their effects, we carried out subsequent experiments only with Myr-peptide injected every other day. Injection of Myr-peptide every other day led to reduced levels of blood glucose (Fig. 4A), insulin (Fig. 4B), HOMA of insulin resistance (HOMA-IR) (Fig. 4C), HDL-cholesterol (Fig. 4D), triglycerides (Fig. 4E), and total cholesterol (Fig. 4F) (P<0.05, Tukey's HSD). Corticosterone (Fig. 4G) levels did not differ between the saline-injected group and the Myr-peptide-injected group (P>0.05, Tukey's HSD). Leptin (Fig. 4H) and adiponectin (Fig. 4I) levels were lower in the Myr-peptide-injected group (P<0.05, Tukey's HSD), correlating with the almost significant reduction in body weight and adipose tissue weight. By contrast, liver weight increased (Fig. 4J), although no change in liver functions, determined by serum AST:ALT ratio, was detected (Fig. 4K). However, liver lipid content increased significantly (P<0.05, Tukey's HSD) (Fig. 4L). Serum parameters demonstrated a beneficial metabolic effect typical of AMPK activation. By contrast, liver fat accumulation suggested no AMPK activation in the liver.
Serum parameters, liver weight, and fat content in Myr-peptide-injected mice fed a HF diet. (A) Glucose. (B) Insulin. (C) HOMA-IR. (D) HDL-cholesterol. (E) Triglycerides. (F) Total cholesterol. (G) Corticosterone. (H) Leptin. (I) Adiponectin. (J) Liver weight. (K) AST:ALT ratio. (L) Liver lipid content. After 10 weeks of HF feeding, at an average body weight of ∼31 g, mice were injected subcutaneously with 0.4 mg/g body weight of Myr-peptide (Myr) or saline every other day for 12 days. During the injection period, mice were given a HF diet. All measurements were made on the last day of the experiment. n=10 in each group. *P<0.05.
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
To determine whether liver fat accumulation was a result of HF feeding during the peptide injection period, we fed mice a HF diet, and once peptide injections were commenced, mice were switched to a low-fat (LF) diet. At the end of the injection period (12 days), there was no difference in body weight between the saline-injected group and the Myr-peptide-injected group (P>0.05, Tukey's HSD) (Fig. 5A). Serum glucose levels were not significantly different from day 0 in the saline-injected group. However, glucose levels were significantly higher in the Myr-peptide-injected group (P<0.05, Tukey's HSD) (Fig. 5B) (see Discussion). Triglyceride levels were significantly lower in the Myr-peptide-injected group than in the saline-injected group (P<0.05, Tukey's HSD) (Fig. 5C). As in the HF-fed mice, liver weight increased (Fig. 5A), although no change in liver functions, determined by serum AST:ALT ratio, was detected (Fig. 5D). In addition, liver lipid content increased significantly (P<0.05, Tukey's HSD) (Fig. 5E). These results suggest that HF feeding during the injection period was not the reason for the accumulation of fat in the liver.
Effect of Myr-peptide injected every other day for 12 days on the body, liver, and adipose tissue weight and serum parameters of mice fed a LF diet. (A) Mean body, liver, and adipose tissue weight. (B) Mean glucose levels on day 0 and day 12. (C) Triglycerides. (D) AST:ALT ratio. (E) Liver lipid content. After 10 weeks of HF feeding, at an average body weight of ∼31 g, mice were injected subcutaneously with 0.4 mg/g body weight of Myr-peptide (Myr) or saline every other day for 12 days. During the injection period, mice were given a LF diet. All measurements were made on the last day of the experiment. n=10 in each group. *P<0.05 and **P<0.05 between groups.
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
Activity of Myr-peptide in mouse liver and AML-12 hepatocytes
As liver fat accumulation does not typify AMPK activation, we next examined whether liver AMPK was activated as a result of Myr-peptide injections. Unlike in C2C12 cells (Fig. 2), liver pAMPK:AMPK ratio of the Myr-peptide-injected group was not different from that of the saline-injected group (P>0.05, Tukey's HSD) (Fig. 6A). Moreover, pACC:ACC ratio of the Myr-peptide-injected group was even lower than that of the saline-injected group (P<0.05, Tukey's HSD) (Fig. 6B), suggesting increased fat synthesis. To determine whether this holds for hepatocytes in general, we studied the effect of Myr-peptide on AMPK activation in a liver cell line (AML-12 hepatocytes). Myr-peptide was added to the medium at 100 μM for 1 h, after which the cells were harvested and measured by western blotting for total AMPK and ACC and their phosphorylated forms. Myr-peptide did not elevate AMPK catalytic activity or ACC phosphorylation (P>0.05, Tukey's HSD) (Fig. 6C and D). These findings suggest that Myr-peptide does not activate AMPK in hepatocytes and together with the decreased ratio of pACC:ACC can explain liver fat accumulation.
Modulation of AMPK and ACC in liver and AML-12 hepatocytes treated with Myr-peptide. (A) Liver pAMPK:AMPK ratio. (B) Liver pACC:ACC ratio. (C) AML-12 cell pAMPK:AMPK ratio. (D) AML-12 cell pACC:ACC ratio. (E) Lipid content in AML-12 hepatocytes treated with metformin, Myr-peptide (Myr), lipid emulsion (LE), AICAR, Compound C (CC), and their combinations. (F) Muscle pAMPK:AMPK ratio. (G) WAT pAMPK:AMPK ratio. After 10 weeks of HF feeding, at an average body weight of ∼31 g, mice were injected subcutaneously with 0.4 mg/g body weight of Myr-peptide (Myr) or saline every other day for 12 days. During the injection period, mice were given a HF diet. Liver, muscle, and WAT were removed and measured by western blotting for the levels of total AMPK or ACC and their phosphorylated forms. Peptides were added to the medium of AML-12 cells at 100 μM for 1 h, after which the cells were harvested and measured by western blotting for the levels of total AMPK and ACC and their phosphorylated forms. AML-12 cells were treated with 1 mM metformin, 100 μM Myr-peptide, 0.1% LE, 1 μM AICAR, 40 mM Compound C, and their combinations for 6 h. Lipid quantification in AML-12 cells was performed using Oil Red O staining. Different letters denote statistical significance (P<0.05). *P<0.05.
Citation: Journal of Endocrinology 221, 1; 10.1530/JOE-13-0625
Lipid accumulation in AML-12 hepatocytes
To enable the estimation of liver lipid accumulation, AML-12 cells were stained with Oil Red O after application of Myr-peptide and compared for the effect of metformin, lipid emulsion (LE), AICAR, Compound C, and their combinations (Fig. 6E). As expected, LE, but not metformin, caused lipid accumulation in AML-12 cells (P<0.05, Tukey's HSD). Myr-peptide caused fat accumulation in AML-12 cells, but to a lower degree than LE. Similarly, AICAR, a direct activator of AMPK, also caused fat accumulation in AML-12 cells (P<0.05, Tukey's HSD). Metformin reduced the level of fat accumulation when combined with LE. Compound C, an AMPK direct inhibitor, reduced the levels of fat accumulation when combined with Myr-peptide or AICAR (P<0.05, Tukey's HSD) (Fig. 6E). These results indicate that Myr-peptide causes fat accumulation in hepatocytes.
Activity of Myr-peptide in mouse muscle and adipose tissue
In light of the results obtained in the liver cell line, we studied AMPK activation in muscle and adipose tissue. Although AMPK was activated in C2C12 myotubes treated with Myr-peptide (Fig. 2), we could not detect such activation in muscle tissue in vivo (P>0.05, Tukey's HSD) (Fig. 6F). By contrast, in WAT, increased AMPK phosphorylation was achieved in mice injected with Myr-peptide (P<0.05, Tukey's HSD) (Fig. 6G). These results correlated with the reduced adipose tissue weight (Fig. 3).
Discussion
In this study, we designed a peptide based on the αG region of the α2 isoform of the kinase domain of AMPK and examined its effect on protein activity both in vivo and in cell culture. The αG region is typically involved in PPIs of kinases and has been used previously to design modulatory peptides (Niv et al. 2004, Wexler et al. 2005). We have also previously shown that αG region, together with the preceding H–J loop, undergoes rigidification upon substrate binding (Cheng & Niv 2010). In AMPK, in particular, homo-oligomerization and activation have been shown to be mediated by the αG region (Scholz et al. 2009). To evaluate the most stable conformations of the AMPK-mimicking peptide in solution, we carried out molecular dynamics simulations of Myr-peptide and Ac-peptide. Computational predictions of the most energetically stable conformations of peptides have evolved immensely with the advance of computational power and algorithms, tackling problems of increasing size and complexity (Dror et al. 2012, Ben-Shimon et al. 2013, Rizzuti & Daggett 2013). In the present study, we used a structure prediction procedure based on simulated annealing molecular dynamics in implicit solvent conditions and large integration time steps that efficiently reproduced the NMR structures of peptides (Ben-Shimon and Niv, unpublished). Our computational results, supported by CD experiments, indicate that the derived peptide closely resembles the protein structure that it was designed to mimic.
The effect of the AMPK-mimicking peptides was next explored ex vivo. A muscle cell line treated with the peptide exhibited the activation of AMPK alongside phosphorylation of its substrate, ACC. This activation was similar to that of AICAR, a known activator of AMPK. Obese mice injected every day with either Myr-peptide or Ac-peptide exhibited reduced body weight at 0.4 mg/g body weight. Reduction of weight in mice injected every other day was achieved as well, albeit with a higher P value (P=0.06), suggesting that the effective concentration to yield significant reduction in body weight is ∼0.3 mg/g body weight per day. Interestingly, leptin-deficient obese mice exhibited significant reduction in body weight with fivefold less concentrated peptide. This result suggests that leptin may antagonize the effect of the peptide so that higher concentrations of peptide are needed to achieve reduction in body weight. Reduction in adipose tissue mass was observed in obese mice injected with the peptide while being fed a HF diet. Moreover, obese mice that were fed a LF diet while being treated with the peptide exhibited a greater reduction in fat depots. In light of the lack of change in food intake, these results suggest that the peptide indeed leads to the breakdown of fat depots. Indeed, AMPK is activated in adipose tissue, leading, in turn, to increased fatty acid oxidation and reduced fatty acid synthesis. However, when fat is supplied in the diet, it counteracts this breakdown.
Injection of Myr-peptide every other day led to reduced levels of blood glucose, insulin, and HOMA-IR, suggesting increased insulin sensitivity and reduced hepatic glucose output. Interestingly, when fed a HF diet during the injection period, blood glucose levels increased in both peptide-injected and saline-injected mice. However, the increase in blood glucose levels was lower in the peptide-injected group. When fed a LF diet during the injection period, blood glucose levels increased in the peptide-injected group, but not in the saline-injected mice compared with day 0. These results suggest that the peptide leads to a smaller increase in hepatic glucose output under HF feeding, and this hepatic glucose output is not influenced by the diet. It is plausible that the retention of lipids in the liver led to a constant hepatic glucose output regardless of the type of diet.
Importantly, our results suggest a dual effect in vivo. On the one hand, blood glucose levels, insulin levels, and insulin resistance index were reduced, as expected from AMPK activation. In addition, triglyceride, cholesterol, leptin, and adiponectin levels were lower, suggesting increased adipose tissue breakdown, also expected from AMPK activation. On the other hand, liver weight and liver fat content increased, indicating fat retention in the liver. The activation of AMPK in the liver leads to decreased gluconeogenesis and glucose output that could explain the reduction in blood glucose levels. However, we did not find an elevated pAMPK:AMPK ratio in the liver in vivo or in hepatocytes ex vivo. This finding suggests that the peptide does not lead to AMPK activation in hepatocytes. Lack of AMPK activation would lead to lack of ACC phosphorylation and under HF feeding would lead to fat synthesis and accumulation in the liver. Indeed, the pACC:ACC ratio was reduced in the liver, suggesting increased fatty acid synthesis.
The effect of fat accumulation in the liver as a result of the peptide administration could also be observed in the hepatocyte cell line. Similarly, AICAR also led to fat accumulation in this cell line. Indeed, it has recently been shown that AICAR activates atypical protein kinase C together with AMPK in human hepatocytes, leading to an increase in lipogenic enzyme levels (Sajan et al. 2013). Nevertheless, other reports have shown significant reduction in hepatic triglyceride accumulation in both chow- and HF diet-fed animals treated with AICAR (Tomita et al. 2005, Henriksen et al. 2013).
Interestingly, muscle tissue exhibited no activation of AMPK after Myr-peptide treatment. These results contradict those obtained in C2C12 myotubes. The differences could stem from the different isoforms found in C2C12 myotubes vs gastrocnemius muscle. Indeed, it has been shown that α2 is the main isoform expressed in muscle tissue (Stapleton et al. 1996), whereas the isoform induced through C2C12 differentiation is α1 (Niesler et al. 2007). Similarly to muscle tissue, no AMPK activation was found in the liver. This finding could probably stem from the fact that both tissues express both α1 and α2 isoforms (Stapleton et al. 1996, Merrill et al. 2012). However, in adipose tissue, the predominant AMPK isoform expressed is α1 (Daval et al. 2005). The sequence corresponding to the HVPTLFKKIRG stretch, the α2 αG helix mimicked in the synthetic peptide in the present study, is HVPTLFKKICD in the α1 isoform. The difference in the last two amino acids of the peptide, cysteine aspartate vs arginine glycine, may lead to differential effects of the peptide on the two isoforms. The peptide may induce the activation of AMPK by interacting with a phosphatase that inactivates AMPK and, thus, disrupting inactivation. Alternatively, it may stem from the prevention of the domain–domain interactions within the AMPK, thus relieving self-inhibition and leading to increased activity. Alternatively, the effect may be due to the disruption of binding at the αG-helix that would promote the formation of homodimers or monomers instead of less-active higher oligomers (Scholz et al. 2009). Such interactions and disruptions thereof may very well be isoform specific. Discerning the mechanism of action of these novel peptides and their effects on α1 and α2 isoforms is the subject of a subsequent study. The difference in the metabolic response to agents that activate AMPK in other cell lines and tissues is of extreme importance. These findings reiterate the differential effect that AMPK activation has on different tissues and the tissue-dependent and nutrition-dependent effects that may be elicited by AMPK modulators.
Conclusion
An AMPK-derived peptide leads to the activation of AMPK in adipose tissue and myotubes and to reduced adipose tissue mass and reduced glucose levels in obese mice. On the other hand, this peptide leads to fat accumulation in the liver. These results demonstrate the differential effect of AMPK modulation in various tissues.
Declaration of interest
The authors declare that there is no conflict of interest that could be perceived as prejudicing the impartiality of the research reported.
Funding
O F and M Y N received funding from Yissumiot and A B-S received funding from the Valazzi-Pikovsky fund.
Author contribution statement
N C, Y G, and A B-S conducted the experiments and analyzed the data. O F, N C, M Y N planned the experiments and wrote the paper.
Acknowledgements
The authors thank Prof. Daniel Harries and Shahar Sukenik for their assistance in the production of the CD results.
References
Arora PS & Ansari AZ 2009 Chemical biology: a Notch above other inhibitors. Nature 462 171–173. (doi:10.1038/462171a)
Batandier C, Guigas B, Detaille D, El-Mir MY, Fontaine E, Rigoulet M & Leverve XM 2006 The ROS production induced by a reverse-electron flux at respiratory-chain complex 1 is hampered by metformin. Journal of Bioenergetics and Biomembranes 38 33–42. (doi:10.1007/s10863-006-9003-8)
Ben-Shimon A, Shalev DE & Niv MY 2013 Protonation States in molecular dynamics simulations of peptide folding and binding. Current Pharmaceutical Design 19 4173–4181. (doi:10.2174/1381612811319230003)
Booth B & Zemmel R 2004 Prospects for productivity. Nature Reviews. Drug Discovery 3 451–456. (doi:10.1038/nrd1384)
Carling D, Mayer FV, Sanders MJ & Gamblin SJ 2011 AMP-activated protein kinase: nature's energy sensor. Nature Chemical Biology 7 512–518. (doi:10.1038/nchembio.610)
Cheng S & Niv MY 2010 Molecular dynamics simulations and elastic network analysis of protein kinase B (Akt/PKB) inactivation. Journal of Chemical Information and Modeling 50 1602–1610. (doi:10.1021/ci100076j)
Daval M, Diot-Dupuy F, Bazin R, Hainault I, Viollet B, Vaulont S, Hajduch E, Ferre P & Foufelle F 2005 Anti-lipolytic action of AMP-activated protein kinase in rodent adipocytes. Journal of Biological Chemistry 280 25250–25257. (doi:10.1074/jbc.M414222200)
Dror RO, Dirks RM, Grossman JP, Xu H & Shaw DE 2012 Biomolecular simulation: a computational microscope for molecular biology. Annual Review of Biophysics 41 429–452. (doi:10.1146/annurev-biophys-042910-155245)
Hardie DG 2007 AMP-activated protein kinase as a drug target. Annual Review of Pharmacology and Toxicology 47 185–210. (doi:10.1146/annurev.pharmtox.47.120505.105304)
Hardie DG 2011 Signal transduction: how cells sense energy. Nature 472 176–177. (doi:10.1038/472176a)
Harvey S, Tan RKZ & Cheatham TE 1998 The flying ice cube: velocity rescaling in molecular dynamics leads to violation of energy equipartition. Journal of Computational Chemistry 19 726–740. (doi:10.1002/(SICI)1096-987X(199805)19:7%3C;726::AID-JCC4%3E;3.0.CO;2-S)
Henriksen BS, Curtis ME, Fillmore N, Cardon BR, Thomson DM & Hancock CR 2013 The effects of chronic AMPK activation on hepatic triglyceride accumulation and glycerol 3-phosphate acyltransferase activity with high fat feeding. Diabetology & Metabolic Syndrome 5 29. (doi:10.1186/1758-5996-5-29)
Kahn BB, Alquier T, Carling D & Hardie DG 2005 AMP-activated protein kinase: ancient energy gauge provides clues to modern understanding of metabolism. Cell Metabolism 1 15–25. (doi:10.1016/j.cmet.2004.12.003)
Mack E, Ziv E, Reuveni H, Kalma R, Niv MY, Jorns A, Lenzen S & Shafrir E 2008 Prevention of insulin resistance and β-cell loss by abrogating PKC epsilon-induced serine phosphorylation of muscle IRS-1 in Psammomys obesus. Diabetes/Metabolism Research and Reviews 24 577–584. (doi:10.1002/dmrr.881)
Merrill JF, Thomson DM, Hardman SE, Hepworth SD, Willie S & Hancock CR 2012 Iron deficiency causes a shift in AMP-activated protein kinase (AMPK) subunit composition in rat skeletal muscle. Nutrition and Metabolism 9 104. (doi:10.1186/1743-7075-9-104)
Minokoshi Y, Alquier T, Furukawa N, Kim YB, Lee A, Xue B, Mu J, Foufelle F, Ferre P & Birnbaum MJ et al. 2004 AMP-kinase regulates food intake by responding to hormonal and nutrient signals in the hypothalamus. Nature 428 569–574. (doi:10.1038/nature02440)
Niesler CU, Myburgh KH & Moore F 2007 The changing AMPK expression profile in differentiating mouse skeletal muscle myoblast cells helps confer increasing resistance to apoptosis. Experimental Physiology 92 207–217. (doi:10.1113/expphysiol.2006.034736)
Nissen SE & Wolski K 2007 Effect of rosiglitazone on the risk of myocardial infarction and death from cardiovascular causes. New England Journal of Medicine 356 2457–2471. (doi:10.1056/NEJMoa072761)
Niv MY, Rubin H, Cohen J, Tsirulnikov L, Licht T, Peretzman-Shemer A, Cna'an E, Tartakovsky A, Stein I & Albeck S et al. 2004 Sequence-based design of kinase inhibitors applicable for therapeutics and target identification. Journal of Biological Chemistry 279 1242–1255. (doi:10.1074/jbc.M306723200)
Onufriev A, Bashford D & Case DA 2004 Exploring protein native states and large-scale conformational changes with a modified generalized born model. Proteins 55 383–394. (doi:10.1002/prot.20033)
Pronk S, Pall S, Schulz R, Larsson P, Bjelkmar P, Apostolov R, Shirts MR, Smith JC, Kasson PM & van der Spoel D et al. 2013 GROMACS 4.5: a high-throughput and highly parallel open source molecular simulation toolkit. Bioinformatics 29 845–854. (doi:10.1093/bioinformatics/btt055)
Rizzuti B & Daggett V 2013 Using simulations to provide the framework for experimental protein folding studies. Archives of Biochemistry and Biophysics 531 128–135. (doi:10.1016/j.abb.2012.12.015)
Rubinstein M & Niv MY 2009 Peptidic modulators of protein–protein interactions: progress and challenges in computational design. Biopolymers 91 505–513. (doi:10.1002/bip.21164)
Sajan MP, Ivey RA III & Farese RV 2013 Metformin action in human hepatocytes: coactivation of atypical protein kinase C alters 5′-AMP-activated protein kinase effects on lipogenic and gluconeogenic enzyme expression. Diabetologia 56 2507–2516. (doi:10.1007/s00125-013-3010-1)
Sanders MJ, Grondin PO, Hegarty BD, Snowden MA & Carling D 2007 Investigating the mechanism for AMP activation of the AMP-activated protein kinase cascade. Biochemical Journal 403 139–148. (doi:10.1042/BJ20061520)
Scannell JW, Blanckley A, Boldon H & Warrington B 2012 Diagnosing the decline in pharmaceutical R&D efficiency. Nature Reviews. Drug Discovery 11 191–200. (doi:10.1038/nrd3681)
Scholz R, Suter M, Weimann T, Polge C, Konarev PV, Thali RF, Tuerk RD, Viollet B, Wallimann T & Schlattner U et al. 2009 Homo-oligomerization and activation of AMP-activated protein kinase are mediated by the kinase domain αG-helix. Journal of Biological Chemistry 284 27425–27437. (doi:10.1074/jbc.M109.047670)
Stapleton D, Mitchelhill KI, Gao G, Widmer J, Michell BJ, Teh T, House CM, Fernandez CS, Cox T & Witters LA et al. 1996 Mammalian AMP-activated protein kinase subfamily. Journal of Biological Chemistry 271 611–614. (doi:10.1074/jbc.271.2.611)
Tomita K, Tamiya G, Ando S, Kitamura N, Koizumi H, Kato S, Horie Y, Kaneko T, Azuma T & Nagata H et al. 2005 AICAR, an AMPK activator, has protective effects on alcohol-induced fatty liver in rats. Alcoholism, Clinical and Experimental Research 29 240S–245S. (doi:10.1097/01.alc.0000191126.11479.69)
Viollet B, Foretz M, Guigas B, Horman S, Dentin R, Bertrand L, Hue L & Andreelli F 2006 Activation of AMP-activated protein kinase in the liver: a new strategy for the management of metabolic hepatic disorders. Journal of Physiology 574 41–53. (doi:10.1113/jphysiol.2006.108506)
Wells JA & McClendon CL 2007 Reaching for high-hanging fruit in drug discovery at protein–protein interfaces. Nature 450 1001–1009. (doi:10.1038/nature06526)
Wexler ID, Niv MY & Reuveni H 2005 Sequence-based protein kinase inhibition: applications for drug development. Biotechniques 39 575–576. (doi:10.2144/000112045)
Yoon MJ, Lee GY, Chung J-J, Ahn YH, Hong SH & Kim JB 2006 Adiponectin increases fatty acid oxidation in skeletal muscle cells by sequential activation of AMP-activated protein kinase, p38 mitogen-activated protein kinase, and peroxisome proliferator-activated receptor (α). Diabetes 55 2562–2570. (doi:10.2337/db05-1322)
Zimmet P, Alberti KG & Shaw J 2001 Global and societal implications of the diabetes epidemic. Nature 414 782–787. (doi:10.1038/414782a)
