Adropin – a circulating factor in metabolic control or a drop in the ocean?

KEYWORDS:

• Biologic
• peptides
• diabetes
• insulin resistance
• glucose tolerance
• obesity

Over 115 million Americans either have type 2 diabetes or exhibit impaired fasting glucose and evidence of insulin resistance (‘prediabetes’) and are thus at high risk of developing the disease within 10 years. These statistics clearly indicate the need to develop new strategies to improve insulin action. We identified adropin (‘sets-fire to fat’) as a secreted peptide linked to metabolic control in 2008. Two recent papers from our laboratory suggest that a 43 amino acid peptide fragment (adropin^34–76) may be a promising lead in the development of insulin-sensitizing agents. Treatment of diet-induced obese (DIO) mice with adropin^34–76 improved glucose tolerance, oxidative glucose metabolism and insulin receptor signaling in skeletal muscle. This editorial discusses current knowledge, obstacles and controversies in the field of adropin physiology.

Secreted peptides regulated by nutrient-sensing signals have important and diverse roles in metabolic control. These peptides regulate glucose-stimulated insulin secretion, glucose metabolism (production and disposal), appetite, lipoprotein metabolism and the ‘browning’ of adipocytes [1–7]. Identifying and characterizing new peptides has added to knowledge of how homeostasis is maintained, and provided potential therapeutic targets for treating the metabolic diseases of obesity.

Discovery of Adropin –The Energy Homeostasis Associated (ENHO) transcript encoding adropin was identified using microarrays investigating fatty liver and insulin resistance in genetically induced obese C57BL/6J (B6) mice [8]. Liver Enho expression was suppressed in mutant and diet-induce obese (DIO) mice, with suppression being secondary to obesity. In silico analysis indicated a highly conserved open reading frame (ORF) encoding a 76-residue protein. A putative signal sequence targeting the secretory pathway resided in residues 1–33; secretion was initially confirmed in cultured cells expressing adropin fused to a c-terminal FLAG epitope tag [8]. FLAG-immunoreactivity in sera from mice infected with an adenovirus expressing the adropin-FLAG construct suggested secretion into the circulatory system. Expression profiling by northern blot in adult mouse and human tissues indicated high levels of expression in the liver and central nervous system relative to other tissues. Collectively, these observations suggested a secreted peptide with autocrine/paracrine, neuropeptide and classical endocrine functions.

Liver adropin expression responds rapidly to nutritional conditioning. Fasting suppresses while feeding rapidly stimulates liver Enho expression [8]. In the fed state, the level of expression is dependent on the macronutrient composition of the diet, with high fat/low carbohydrate diets (HFD) having the most potent effect. Plasma adropin concentrations also correlate with intake of dietary fat (positive) and carbohydrate (negative) relative to other macronutrients in humans, suggesting clinical relevance [9].

Regulation by nutrient sensors suggests roles related to metabolic control. Over expression of adropin or treatment using synthetic adropin³⁴⁻⁷⁶ alters the metabolic profile of mice. Adropin transgenic mice and DIO mice treated with synthetic adropin³⁴⁻⁷⁶ exhibit lean phenotypes, reduced fasting insulin and improved glucose tolerance [8,10]. Furthermore, two recent papers by independent laboratories suggest adropin regulates glucose and lipid metabolism in rats [11] and lipoprotein metabolism in fish [12].

To determine whether suppression in obesity contributes to metabolic diseases, an adropin knockout mouse was engineered using the Cre/Lox system [13]. Insertion of LoxP sites to flank the adropin ORF in the Enho gene generated “flox’d” mice (Enho^lox/lox). In chow-fed conditions, homozygous carriers of the null Enho gene (Enho –/–) exhibit modest but significant increases in fat mass (FM) determined by NMR. Enho –/– mice were also insulin resistant, determined by hyperinsulinemic-euglycemic clamp at the Mouse Metabolic Phenotyping Center at Vanderbilt University. While exhibiting normal DIO, propensity for fasting hyperinsulinemia, impaired fasting glucose, glucose intolerance and dyslipidemia were all increased in Enho –/– mice. Importantly, heterozygous (Enho±) mice exhibited intermediate phenotypes, indicating partial suppression of adropin function can impact metabolic control.

Developing a theoretical model of adropin function

While significant in suggesting the identification of a new factor in metabolic control, these initial observations provided no indication of how adropin regulates metabolic activity. Also absent was a theoretical framework explaining a physiological role of adropin. As adropin expression is regulated by dietary fat, Gao et al. tested whether adropin regulates fatty acid oxidation (FAO). Surprisingly, initial studies indicated a marked suppression of FAO in skeletal muscle lysates from adropin transgenics, and in DIO mice treated with adropin³⁴⁻⁷⁶. At the time the first experiments were completed (2006), coupling improved metabolic control with inhibition of FAO was startling, and inconsistent with popular theories linking insulin resistance to low fat oxidation and accumulation of lipid intermediates [14]. However, the observation that adropin expression is stimulated by feeding and suppressed by fasting led to the hypothesis that adropin might mediate changes in fuel selection between these conditions [15].

It has long been known that FAO is limited and glucose oxidation enhanced in the fed condition, while FAO is enhanced and glucose utilization limited in muscle and liver limited to spare glucose for the brain during fasting [16]. Regulation of fuel selection involves local signals of metabolic flux within mitochondria, and also responds to signals of systemic metabolic condition provided by circulating factors (primarily insulin and glucagon) [16]. Starvation and diabetes increase expression of pyruvate dehydrogenase kinase 4 (PDK4) in skeletal muscle [17]. Phosphorylation of pyruvate dehydrogenase (PDH), a component of the pyruvate dehydrogenase complex linking glycolysis to the citric acid cycle by converting pyruvate to acetyl-CoA, by PDK4 is inhibitory. Based on data from adropin transgenic and knockout mice [15], Gao et al. proposed adropin suppresses PDK4 expression in the fed condition, increasing PDH activity and enhancing pyruvate oxidation. In contrast, a decline in adropin with fasting inhibits pyruvate oxidation due to the loss of a stimulatory input. At the same time, adropin inhibits fat oxidation by suppressing the expression and activity of carnitine palmitoyl transferase-1b (CPT1B), a mitochondrial enzyme that prepares fatty acids for translocation into mitochondria for oxidation. PGC-1α is a common element involved in regulating CPT1B and PDK4. Inhibition of PGC-1α, a transcriptional co-activator regulating expression of genes involved in carbohydrate and lipid metabolism, would hence reduce CPT1B (thereby restricting the flow of fatty acid from the cytosol into mitochondria) and PDK4 (thereby enhancing pyruvate oxidation through increased PDH activity). In addition, by reducing CD36 expression adropin could reduce uptake of fatty acids into the cytosol, thereby countering the effect of CPT1B suppression of cytosolic accumulation of fatty acid intermediates [10].

Controversies and contradictions in the field

Whether adropin is actively secreted is controversial, as a recent publication failed to replicate experiments showing secretion of epitope-tagged adropin [18]. However, publications too numerous to be cited here have reported measurements of adropin in the circulation of humans, nonhuman primates, cows, rats and mice. Results from a collaboration between several laboratories using the Bachem enzyme immunoassay suggest circulating adropin levels correlate with liver Enho expression in mice [13,19]; This group also measured plasma adropin concentrations in humans [9,19,20]. Collectively, these early results suggested a decline in adropin concentrations with aging and obesity, a negative association between plasma adropin concentrations and fasting triglycerides, and a transient increase in levels following bariatric surgery [19]. Plasma adropin concentrations also correlates positively with saturated fat intake in human female participants of a sleep-restriction study [9]. Fish oil supplements that attenuated hypertriglyceridemia and insulin resistance associated with fructose consumption for 6 months in nonhuman primates also attenuated a decrease in plasma adropin concentrations [21]. On the other hand, humans consuming fructose as 25% of daily energy requirements for 2–10 weeks exhibited increases in serum triglyceride and plasma adropin concentrations [20]. The exact physiological significance of these results is unclear at this time. It is possible that exclusion criteria in the original studies from which these samples were obtained resulted in bias. Clearly, further experiments are needed to understand how environmental and metabolic factors affect plasma adropin concentration in humans.

It is also worth mentioning that results from studies measuring adropin in serum or plasma should be viewed with caution. Little is known about specificity and function of the various commercially available assays. In addition, the stability of adropin in serum or plasma is likewise not known, and whether protease inhibitors are required is also unclear [22]. For the tests using the Bachem assay, samples collected with a cocktail of protease inhibitors were initially used. EDTA-plasma was ultimately used as this provides inhibition from a broad-spectrum of metalloproteases. A ‘spike-recovery’ test using human EDTA plasma and lack of signal in plasma from Enho –/– mice indicated both specificity and linearity over a broad range [19]. However, the assay does have a high intra-assay coefficient of variation. As is the case for other peptide hormones, with irisin being a recent example [23], extensive validation and further assay development would significantly benefit the field.

Another laboratory produced Enho –/– mice [18], and while a similar reduction of physical activity was observed there was no evidence of insulin resistance. Failure to replicate metabolic phenotypes in a mouse models between laboratories is not unprecedented, as genetic background and housing conditions can significantly affect phenotype [24]. In this case, differences in approach by the two laboratories are worth noting. Wong et al.’s strategy involved inserting a neomycin selection cassette into the 5ʹUTR of the Enho gene [18]. In contrast, Kumar et al. flanked the ORF in exon 2 with LoxP sites and removed the neomycin selection cassette to minimize potential effects on expression of adjacent genes [13]. Wong et al. reported that a transcript was still expressed based on RT-PCR using primers targeting the 3ʹUTR in exon 2, while expression in Kumar et al.’s mice was undetectable by RT-PCR. Wong et al. used 129  Sv/J embryonic stem (ES) cells, backcrossing founders for six generations onto a B6 background, whereas Kumar et al. used B6 ES cells. Kumar et al. used the hyperinsulinemic-euglycemic clamp to assess glucose homeostasis, whereas Wong et al. relied on less sensitive glucose and insulin tolerance tests. Finally, Kumar et al. exposed mice to HFD for 6 weeks whereas they exposed their mice for 24 weeks, possibly masking subtle differences in glucose control and insulin sensitivity.

Obstacles and future directions

Is the Diabetes article reporting a role for adropin in regulating fuel selection the final word in how adropin regulates glucose and fatty acid metabolism? Unlikely, given that the pathways involved in regulating oxidative metabolism are complex. Indeed, while observing that treatment of DIO mice with adropin³⁴⁻⁷⁶ also suppresses PDK4 expression and enhances PDH activity, this response does not involve the same signaling mechanism. Many questions remain about adropin physiology that requires further research. For example, it remains to be established whether circulating adropin is indeed of hepatic origin, and whether adropin produced by liver regulates fuel selection in skeletal muscle. The role of adropin in regulating vascular function is also of interest [25], in terms of maintaining vascular health and in potentially contributing to metabolic control. Moreover, the relationship between plasma adropin concentrations and metabolic risk factors in humans may be more complicated then first thought [9,19,20].

Financial & competing interests disclosure

The National Institute of Diabetes and Digestive and Kidney Diseases (DK073189), American Diabetes Association (1-04-JF09; 7-08-RA16), Biomeasure (now named Ipsen Bioscience, Inc.), and Novo Nordisk have provided financial support for research in the author’s laboratory. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed.

Acknowledgements

The authors acknowledge the work of many individuals who have made significant contributions of time and material to the adropin program. While it is not possible to provide the names of everyone, the author acknowledges the intellectual and material inputs of Professor Peter Havel (UC Davis, Davis, CA), Dr Kimber L. Stanhope (UC Davis), Dr Marie-Pierre St-Onge (Columbia University, New York, NY), and Professor Eric Ravussin (Pennington Biomedical Research Center, LSU System, Baton Rouge, LA).