APJ-independent Metabolic Actions of Apelin-36 1 Apelin-36 Modulates Blood Glucose and Body Weight Independently of Canonical APJ Signaling

Apelin-36 was discovered as the endogenous ligand for the previously orphan receptor APJ. Apelin-36 has been linked to two major biological activities – cardiovascular (stimulation of cardiac contractility and suppression of blood pressure) and metabolic (improving glucose homeostasis and lowering body weight). It has been assumed that both of these activities are modulated through APJ. Here, we demonstrate that the metabolic activity of apelin-36 can be separated from canonical APJ activation. We developed a series of apelin-36 variants in which evolutionarily conserved residues were mutated, and evaluated their ability to modulate glucose homeostasis and body weight in chronic mouse models. We found that apelin-36(L28A) retains full metabolic activity, but is 100-fold impaired in its ability to activate APJ. In contrast to its full metabolic activity, apelin-36(L28A) lost the ability to suppress blood pressure in spontaneously hypertensive (SHR) rats. We took advantage of these findings to develop a longer-acting variant of apelin-36 that could modulate glucose homeostasis without impacting blood pressure (or activating APJ). Apelin-36-[L28C(30kD-PEG)] is 10,000fold less potent than apelin-36 at activating the APJ receptor but retains its ability to significantly lower blood glucose and improve glucose tolerance in diet-induced obese (DIO) mice. Apelin-36-[L28C(30kD-PEG)] provides a starting point for the development of diabetes therapeutics that are devoid of the blood pressure effects associated with canonical APJ activation. The apelin gene encodes a pre-pro-protein that is processed into a number of regulatory hormones. The best characterized of these peptide hormones are apelin-13, apelin-17, and apelin-36 (1). The thirteen C-terminal amino acids of these peptides (comprising apelin-13) are shared, with apelin-17 extending an additional four amino acids from the N-terminus, and apelin-36 extending a further nineteen amino acids beyond the N-terminus of apelin-17 (see Table 1). Apelin was discovered as an endogenous agonist of the G protein-coupled receptor APJ (2). Specifically, apelin-36 was purified from bovine stomach tissue extract based on its ability to stimulate signaling through APJ. Through a combination of pharmacological and genetic approaches, apelin has been linked to two major biological activities – cardiovascular (stimulation of cardiac contractility and suppression of blood pressure) and metabolic (improving glucose homeostasis and lowering http://www.jbc.org/cgi/doi/10.1074/jbc.M116.748103 The latest version is at JBC Papers in Press. Published on December 19, 2016 as Manuscript M116.748103 Copyright 2016 by The American Society for Biochemistry and Molecular Biology, Inc. by gest on Jauary 7, 2017 hp://w w w .jb.org/ D ow nladed from by gest on Jauary 7, 2017 hp://w w w .jb.org/ D ow nladed from by gest on Jauary 7, 2017 hp://w w w .jb.org/ D ow nladed from APJ-independent Metabolic Actions of Apelin-36 2 body weight) (3, 4). Both of these activities have been assumed to be mediated solely through APJ, but there is substantial evidence to suggest that APJ may have apelin-independent activities (5, 6). Here, we provide evidence that this ligandreceptor promiscuity goes in both directions, and that apelin may have APJ-independent activities as well. Specifically, we demonstrate that the metabolic activity of apelin-36 can be dissociated from canonical APJ signaling. RESULTS To evaluate the chronic metabolic activity of apelin peptides, we used an adeno-associated virus (AAV) ‘minigene’ system to drive lasting systemic expression of apelin-13, apelin-36, or a negative control [secreted green fluorescent protein (GFP)] in a mouse DIO prevention model. AAV expressing apelin-13, apelin-36, or GFP was injected through the tail vein into BDF mice that were then immediately placed on a high fat diet (HFD), followed by a series of metabolic assessments (Fig. 1A). Body weight and fasting blood glucose levels were measured weekly for eight weeks. Fasting serum insulin levels were measured during the fourth week of the study, and to evaluate the impact of apelin-13 and apelin-36 on glucose clearance, intraperitoneal glucose tolerance tests were performed at week six. Mice expressing apelin-36, but not apelin-13, gained significantly less weight on the high fat diet than the GFP control mice (Fig. 1B), exhibited lower fasting blood glucose levels (Fig. 1C), and demonstrated improved glucose tolerance (Fig. 1D). Fasting serum insulin levels were not significantly different than controls (Fig. 1E). The metabolic benefits of apelin-36 are further exemplified by its impact on serum lipids, as eight weeks post-injection, apelin-36 significantly lowered serum total cholesterol (Fig. 1F) and LDL cholesterol (Fig. 1G). Triacylglyceride (TAG) levels were not significantly different than controls (Fig. 1H). Taken together, these data indicate that apelin-36, but not apelin-13, protects mice from the negative metabolic consequences of a high-fat diet. Apelin-36 was purified based on its ability to act as an agonist at the G protein-coupled receptor APJ (2). Numerous reports have demonstrated that apelin-13 is also a full agonist of APJ, exhibiting similar, if not superior, potency to apelin-36 (2, 7, 8). We corroborated these observations, finding that both apelin-13 and apelin-36 can potently activate APJ in an APJ stable cell line [apelin-13 EC50 = 1.40 x 10 M (1.02-1.91 x 10 M, 95% confidence interval (CI)), apelin-36 EC50 = 1.02 x 10 M (0.77-1.33 x 10 M, 95% CI), Fig. 1I]. Our observations that apelin-13 and apelin-36 exhibit similar profiles of receptor activation, but that only apelin-36 can protect mice from the symptoms of metabolic syndrome, led us to examine the relationship between APJ activation and the metabolic activity of apelin peptides. To address this relationship, we generated a series of apelin-36 variants that were expected to be impaired to various degrees in APJ activation (based on previous structure activity relationship studies of apelin-13 and apelin-17 (2, 8–13)) and we evaluated their metabolic activity (using a 6week version of the paradigm illustrated in Fig. 1A). We followed up the most interesting variants by studying their APJ activation profiles. The activity profiles of these variants are summarized in Table 1. One apelin-36 variant in particular, featuring a leucine to alanine substitution at amino acid 28 [apelin-36(L28A)], exhibited a striking relationship between metabolic activity and APJ activation. A corresponding variant of apelin-13 with this leucine to alanine substitution at the same location was reported to exhibit profoundly impaired APJ activation (8, 12), and indeed, as shown in Figure 2A, while apelin-36 potently activated APJ, apelin-36(L28A) was over 100-fold less potent [apelin-36 EC50 = 2.03 x 10 M (1.642.51 x 10 M, 95% CI), apelin-36(L28A) EC50 = 1.90 x 10 M ((1.47-2.45 x 10 M, 95% CI)]. At concentrations below 10 M, apelin-36(L28A) did not substantially activate APJ. In contrast to its impaired APJ activation, apelin-36(L28A) remarkably retained full metabolic activity. Apelin-36(L28A) was able to protect mice from the negative metabolic consequences of prolonged HFD, reducing body weight (Fig. 2B) and blood glucose (Fig. 2C), improving glucose tolerance (Fig. 2D), and correcting the lipid profile (Fig. 2F, G) as effectively as wild-type apelin-36. Apelin peptides have been linked to two major biological activities – cardiovascular and metabolic. Having examined the relationship by gest on Jauary 7, 2017 hp://w w w .jb.org/ D ow nladed from APJ-independent Metabolic Actions of Apelin-36 3 between metabolic activity and APJ activation, we set out to understand the relationship of receptor activation with cardiovascular activity (specifically blood pressure). We evaluated the impact of wild-type apelin-36 and apelin36(L28A) on blood pressure in SHR rats, and found that wild-type apelin-36 caused a sharp drop in blood pressure compared to vehicle-treated animals, but apelin-36(L28A) did not (Fig. 2I). These data are consistent with apelin-36 acting through canonical APJ signaling to modulate blood pressure, but not metabolic activity. Apelin peptides are being pursued therapeutically for the treatment of heart failure (3), and have also been proposed as candidate therapeutics for diabetes and obesity (4). However, the cardiovascular activity of apelin peptides could be a liability in obese diabetic patients, as they would be expected to increase the energy demands on the heart (14). Thus, our discovery of an apelin variant that uncouples cardiovascular from metabolic activity suggests the possibility of developing apelin peptides with the ability to treat metabolic disease without impacting cardiovascular function. In addition to lacking cardiovascular activity, such a candidate therapeutic would also need to exhibit a longer half-life than wild-type apelin-36 to allow for a once daily or less frequent dosing interval (15). Appending polyethylene glycol (PEG) polymer chains to peptides has been demonstrated to increase in vivo half-life of peptides by increasing hydrodynamic size (thus reducing renal clearance) and through protection from proteolytic degradation (16). [40kDa-PEG]-apelin-36, in which a PEG moiety was appended to the Nterminus of apelin-36, has been shown to exhibit enhanced cardiovascular activity and extended half-life compared to unmodified apelin-36 (17). However, we have found that attaching PEG molecules at either the Nor C-termini of apelin36 disrupted the metabolic activity of the peptide (discussed further below and see Fig. 3). The impaired metabolic activity of these molecules may be due to steric hindrance of the Nor Ctermini by the PEG moiety. Therefore, to prevent such steric hindrance, we inserted a PEG moiety at the L28 position of apelin-36, a position that is distant from either terminus of the peptide, and which we demon