Adropin and Insulin Resistance: Integration of Endocrine, Circadian and Stress Signals Regulating Glucose Metabolism

Abstract

Dysregulation of hepatic glucose production (HGP) and glucose disposal leads to type 2 diabetes (T2D). Hyperglycemia results from declining ability of insulin to reduce HGP and increase glucose disposal, and inadequate ß-cell compensation for insulin resistance. Hyperglucagonemia resulting from reduced suppression of glucagon secretion by insulin contributes to hyperglycemia by stimulating HGP. These actions of pancreatic hormones are normally complemented by peptides secreted by cells distributed throughout the body. This regulatory network has provided new therapeutics for obesity and T2D (e.g., GLP-1). Other peptide hormones being actively investigated show promise in preclinical studies. Recent experiments using mice and nonhuman primates indicate the small-secreted peptide hormone adropin regulates glucose metabolism. We discuss recent expression profiling data indicating hepatic adropin expression increases with oxidative stress and declines with fasting or in the presence of hepatic insulin resistance, and how adropin interacts with the pancreatic hormones, insulin, and glucagon to modulate glycemic control.

With recent advances in obesity and T2D therapeutics, some might conclude that all major hormonal pathways governing energy balance and glucose homeostasis have been identified and thoroughly investigated. However, mammalian genomes contain many highly conserved short open frame sequences (sORF) encoding a multitude of secreted peptides that may regulate metabolic homeostasis (1, 2), some of which could provide further treatments for obesity and T2D.

The biological activities of the small, secreted peptide, adropin are indicative of a potentially important regulator of glucose metabolism. The sORF encoding adropin was discovered by Genentech’s Secreted Protein Discovery Initiative (2). The first experiments connecting adropin to metabolism profiled hepatic gene expression in male C57BL/6J mice (3). Expression of the Energy Homeostasis Associated (ENHO) transcript encoding adropin was suppressed with obesity and fasting, and rapidly increased when refed after overnight fasting (3, 4). These responses suggested involvement in metabolic adaptation to food intake.

Expression profiling using mouse models now suggests a more complex relationship. A variety of metabolic conditions suppress hepatic adropin expression: catabolic states induced by overnight fasting, increased hepatic lipogenesis (e.g. consumption of high sugar diets or severe hyperphagic obesity syndromes), in hypercholesterolemia, and genetic disruption of the circadian-clock activator BMAL1 (3, 4, 5, 6). Expression of adropin in nonhuman primate (NHP) and mouse liver exhibits a circadian profile with peaks coinciding with food intake (6, 7). Control of transcription is complex and appears to involve nuclear receptor components of the clock (ROR, Rev-erb) and epigenetic modification (4, 6, 7). The liver clock per se is insufficient for full expression indicating circadian signaling from extrahepatic tissues is involved (4).

The conditions that increase adropin expression are not well characterized but are correlated with activation of stress responses. While intake of high fat/low sugar diets rapidly increases adropin expression (3), expression also increases with caloric restriction (CR) (4). In the latter study, adropin expression correlated with genes involved in cellular stress responses. Hepatic adropin expression is increased by compounds that induce oxidative stress (e.g, paraquat) (4). These responses could be adaptive, as adropin actions involving NRF2 protect the liver from oxidative stress (8). One implication is that hepatic adropin expression is induced in response to cellular stress. Presumably, normal regulation of adropin in response to oxidative stress, and any potential protective role, are compromised in situations of severe obesity and insulin resistance.

An unexpected observation is a positive correlation between adropin and phosphoenolpyruvate carboxykinase 1 (PCK1), a major control point in gluconeogenesis (4). This relationship is observed in the liver transcriptome in samples collected over 24 hours from mice fed ad libitum or subjected to CR. These data provide a window into the liver transcriptome under conditions ranging from short-term energy deficit between meals in mice subject to prolonged CR, milder energy deficits during the lights-on period when nocturnal mice rest, ‘normal’ voluntary feeding in the dark phase, or recent binge-feeding (mice subject to CR consume 80–90% of daily kcal intake within 1-h of ad libitum food presentation). Overnight fasting of mice housed at room temperature induces a severe catabolic state (9). While fasting suppresses hepatic adropin expression and increases HGP, this relationship is not observed in mice subjected to a broader range of metabolic conditions involving co-activation of HGP and adropin expression.

Positive correlations with PCK1 and acute responses of liver adropin expression to fasting and refeeding suggest functions related to adapting metabolism to shifts in energy balance. Experiments performed in adropin knockout mice (AdrKO), mice treated with synthetic adropin, and in primary cultured hepatocytes all suggest adropin inhibits HGP (4, 10, 11, 12, 13). This effect is bidirectional as liver-specific AdrKO mice exhibit increased glucose excursions following ip. injections of pyruvate, suggesting increased HGP (4). Adropin appears to ‘modulate’ glucagon receptor signal transduction. Activation of the putative adropin receptor has an inhibitory effect on the induction of cAMP-protein kinase A signaling by glucagon. This could indicate an inverse relationship between liver adropin expression and the ability of glucagon to increase HGP. Alternatively, adropin has the opposite effect on insulin action, improving hepatic insulin action in conditions of obesity and insulin resistance (12, 13).

Two candidate cell-surface receptors have been identified: an orphan GPCR (GPR19) (14) and a glycosylphosphatidylinositol (GPI)-anchored neuronal membrane protein that functions as a cell adhesion molecule (NB3/CNTN6) (15). Both are predominantly expressed in the nervous system (16). GPR19 inhibits cAMP signaling in cultured cells (17), while notch signaling has been implicated in liver glucose metabolism (18). However, how adropin signaling from the cell-surface of hepatocytes affects glucose metabolism needs to be clarified.

Experiments conducted in mouse models support a role for adropin in glucose homeostasis. AdrKO mice exhibit increased HGP assessed by hyperinsulinemic-euglycemic clamp (10). Activation of hepatic PKB/AKT signaling by insulin is however normal, suggesting a potentially novel and atypical mechanism (10). Whether changes in glucagon signaling observed in liver-specific AdrKO mice explain the phenotype requires further investigation (4). Administration of recombinant adropin to diet-induced obese mice enhances insulin signaling in skeletal muscle and liver (12, 19). This effect is observed is in perfused hearts ex vivo, suggesting a direct effect of adropin (4, 12, 20). Adropin treatment is also unique because it acutely enhances oxidative glucose disposal in DIO mice (19). Other laboratories have also observed direct effects of adropin in cultured cells and isolated perfused hearts to regulate fuel substrate selection to favor the oxidation of glucose over fat (21, 22) (20). Adropin could therefore potentially represent a unique therapy targeting T2D by improving insulin sensitivity while simultaneously enhancing oxidative glucose disposal.

If the mouse data are translatable to humans, low adropin activity should correlate with hyperglycemia. In a NHP model, low plasma adropin concentrations correlate with fasting hyperglycemia, and with plasma biomarkers suggesting dysregulation of hepatic lipoprotein metabolism and increased adiposity (7, 23). Transcriptomic data in the liver of NHPs suggest adropin expression peaks around mealtime and is co-regulated with genes involved in glucose metabolism and second messenger signal transduction pathways involving G proteins and receptor tyrosine kinases. In primates, peak hepatic adropin expression may thus anticipate and/or coincide with the activation of signal transduction pathways mediating the responses of hepatocytes to endocrine and/or metabolite signals of increased nutrient flux.

Our analysis of relationships between circulating levels of adropin and circulating indices of glucose and lipid metabolism in humans are less clear. Low plasma adropin concentrations correlated with evidence of impaired insulin action in studies of subjects using narrowly defined age, weight and health criteria for inclusion (5, 24). On the other hand, in a larger study with fewer restrictions for inclusion, we reported an inverse relationship with non-HDL cholesterol that could be explained by an inhibitory effect of cholesterol on ENHO expression (6). Fructose consumption rapidly induces insulin resistance and dyslipidemia in humans and rapidly increases plasma adropin concentrations and this effects effect is most pronounced in subjects with existing dyslipidemia at baseline prior to sugar consumption (25). As with the early studies of other secreted peptides, issues of assay specificity and validation are significant caveats in the interpretation of the data on plasma adropin concentrations.

To conclude, we have discussed several mechanisms by which adropin acts to regulate glucose homeostasis. Adropin signaling in the liver through an as-yet unidentified receptor interacts with insulin and glucagon to lower HGP. Suppression of adropin expression may contribute to increased HGP in obesity. Data showing enhancement of insulin action and glucose oxidation in cardiac and skeletal muscle also suggest a role in lower glucose (19, 20). Clearly, further assessment of the clinical relevance of the current findings on regulation of adropin and its physiological actions are warranted. The identity of the cell-surface receptors for adropin and the signal transduction pathways need to be fully explored. Finally, models more relevant to humans (e.g., NHPs) should be used to explore the relationship between low hepatic adropin expression and hyperglycemia.