Modified UCN2 Peptide Acts as an Insulin Sensitizer in Skeletal Muscle of Obese Mice

Exercise and diet are potent lifestyle interventions to combat metabolic dysfunction by improving weight management and glucose homeostasis. In particular, exercise enhances skeletal muscle insulin sensitivity and mitochondrial function (1). Nevertheless, such lifestyle interventions have poor adherence, requiring pharmacological advances to alleviate obesity and prevent metabolic disease. Consequently, efforts are under way to develop insulin sensitizers and weight-reducing pharmacological agents for the treatment of diabetes (2,3).

Skeletal muscle is an important tissue involved in maintaining glucose homeostasis under insulin-stimulated conditions and is a major site of insulin resistance in type 2 diabetes (4,5). Although precise mechanisms of skeletal muscle insulin resistance are not fully elucidated, impaired insulin signaling and reduced glucose uptake are major aspects (4,5). Insulin resistance is present at all pathogenic stages of type 2 diabetes progression. Consequently, efforts to maintain skeletal muscle insulin sensitivity to prevent/delay type 2 diabetes are warranted. In addition to lifestyle modifications, including diet and exercise, new therapeutic routes to directly enhance skeletal muscle insulin sensitivity, either as monotherapy or in combination with other drugs, are of interest to treat type 2 diabetes.

The corticotropin-releasing factor (CRF) urocortin (UCN) family of neuropeptides is a direct modulator of the hypothalamic-pituitary-adrenal axis both centrally and peripherally (6). Within this family are four peptides (CRF and UCN 1, 2, and 3) that are structurally related but encoded by separate genes (7). UCN peptides signal through two different G-protein–coupled receptors: corticotropin-releasing hormone receptors (CRHRs) 1 and 2 (8). These peptides and receptors are differentially expressed in central and peripheral tissues (7,8). UCN1 binds to both receptors, while UCN2 and UCN3 are selective for CRHR2. Skeletal muscle has high expression levels of both UCN2 and its receptor CRHR2 (9). While emerging evidence suggests that CRF peptides regulate cardiovascular and renal function and inflammatory processes (10), their role in metabolic diseases is unclear.

UCNs and CRHR2 play a role in glucose homeostasis. Ucn2 and Crhr2 knockout mice have enhanced glucose tolerance and increased insulin sensitivity and are protected from high-fat diet (HFD)–induced obesity (11,12). HFD and/or elevated stress states upregulate skeletal muscle CRHR2 (13), while CRHR2 activation inhibits insulin signaling (14). Thus, increased CRHR2 activity impairs glucose homeostasis. In contrast, whole-body Ucn3 transgenic mice are protected from HFD-induced obesity (15), and transient overexpression of Ucn3 in skeletal muscle enhances glucose metabolism through increased insulin signaling (16). In addition, Ucn2 overexpression through systemic virus delivery improves whole-body insulin sensitivity in HFD-fed rodents (17). Accordingly, activating CRHR2 during obesity can also enhance glucose homeostasis. While the UCN-CRHR axis appears to regulate skeletal muscle metabolism, the predominant effects remain unclear.

Observations of aerobic training–like phenotypes in transgenic mice (18–20) has ignited interest in developing pharmacological therapies to combat insulin resistance in patients with type 2 diabetes (21). Given the role of the UCN-CRHR axis in skeletal muscle metabolism, we hypothesized that UCN peptides act as insulin sensitizers in skeletal muscle. Thus, we investigated the effects of a modified UCN2 peptide acting on the CHRH2 in HFD-induced obese mice, with a specific focus on skeletal muscle.

Compound A (a PEGylated peptide analog of human UCN2) was synthesized using established solid-phase peptide synthesis protocols. After final cleavage of the peptide from the resin, the peptide was purified using reversed-phase chromatography and lyophilized to obtain peptide powder as trifluoroacetate salt. The peptide was conjugated to a 20-kDa functionalized polyethylene glycol (PEG) polymer through an acetamide-based linker. Formulated aliquots of the peptide conjugate in PBS were stored at −20°C. Working solutions were freshly prepared from thawed stock aliquots diluted with 0.5% pan-albumin/0.9% NaCl.

Pharmacokinetics were determined in mice after a single subcutaneous administration of compound A. Plasma concentrations of compound A were determined through liquid chromatography with tandem mass spectrometry. Pharmacokinetic parameters were calculated by noncompartmental analysis using Phoenix WinNonlin 6.3 software.

HEK293 cells transfected with mouse CRHR1 or CRHR2β plasmid were plated in 96-well plates at 2,000 cells per well and allowed to attach overnight. Serial dilutions of human UCN2 or compound A were placed onto the cells for 15 min. cAMP levels were measured using a cAMP cell-based assay kit (Cisbio).

L6 rat myoblasts expressing human GLUT4 with an exofacial Myc-epitope tag were cultured in a 96-well plate and incubated in the absence or presence of 100 nmol/L insulin, 100 nmol/L compound A, or 100 nmol/L clenbuterol for 30 min. Cell surface density of GLUT4-Myc was measured as previously described (22). Fluorescence intensity was obtained using a LI-COR Odyssey eXL (LI-COR Biosciences, Lincoln, NE).

Experiments were approved by the Stockholm North animal ethics committee or the Eli Lilly institutional animal care and use committee. Male mice (C57BL/6J) were purchased from Charles River Laboratories (Sulzfeld, Germany) or Envigo (Somerset, NJ) at 5 weeks of age. Mice were maintained under a 12-h light/dark cycle and had free access to water and standard rodent chow (4% kcal from fat, R34; Lantmännen, Kimstad, Sweden). At 6 weeks of age, mice were placed on either a standard rodent chow or an HFD (60% kcal from fat, TD.06414; Harlan Laboratories) ad libitum for 20 weeks and were single housed after 19 weeks. After 20 weeks on an HFD, mice received daily subcutaneous injections of vehicle before the onset of the dark period (0.5% pan-albumin/0.9% NaCl) or compound A (0.3 mg/kg body weight) for 14 days. Injections were performed in the intrascapular region or hind leg on alternating days to minimize discomfort.

HFD-fed mice were randomized into sedentary or wheel running groups. Wheel running mice were acclimatized to the running wheels for 7 days, and all groups were weight and running matched before injections. Body weight and food intake were recorded daily. Activity of the mice on the running wheels (35-cm diameter) was monitored by a magnetic switch affixed to each wheel, which recorded the number of revolutions. Data were captured by an automated computer monitoring system (VitalView application software; Mini-Mitter Company). Physical activity was recorded continuously as wheel revolutions per 5-min interval.

Extensor digitorum longus (EDL) muscles were dissected from 4-h–fasted mice anesthetized with an intraperitoneal injection of 16 μL/g body weight 2.5% 2,2,2-tribromoethanol and tertiary amyl alcohol. Muscles were incubated with Krebs-Henseleit buffer under continuous gassing (95% O₂/5% CO₂) at 30°C in the absence (basal) or presence of 0.36 nmol/L insulin (Actrapid; Novo Nordisk), and 2-deoxy-d-glucose uptake was determined as previously described (23). Results are expressed as μmol/L glucose × mg protein⁻¹ × 20 min⁻¹.

Soleus muscles from 8-week-old, chow-fed mice were incubated in the absence or presence of compound A (63.3 nmol/L) with or without a submaximal insulin dose (0.18 nmol/L) to assess insulin sensitivity. Glucose uptake was determined as described above.

Mice fed an HFD for 20 weeks received daily subcutaneous injections of compound A (0.3 mg/kg) or vehicle for 6 days. Fasted mice (4 h) were anesthetized with isoflurane. Mice received 10 μCi [³H]2-deoxy-d-glucose (PerkinElmer) ± 0.5 units/kg insulin (Humilin R; Eli Lilly) by retro-orbital injection. Blood samples were taken at 2, 5, 10, 15, 20, and 30 min after injection, treated with Ba(OH)₂, and precipitated with ZnSO₄ for determination of blood-specific activity. After centrifugation, the supernatant was collected, and radioactivity was determined using liquid scintillation counting (Beckman LSC). Animals were euthanized, and tissues were frozen. Tissue homogenates were mixed with either water to determine total 2-deoxy-d-glucose or Ba(OH)₂/ZnSO₄ to determine unphosphorylated 2-deoxy-d-glucose as previously described (24).

Glucose tolerance and body composition were determined on day 11 of the treatment. Glucose (2 g/kg body weight) was administered by intraperitoneal injection in 4-h–fasted mice. Blood was sampled through the tail vein to assess glucose (OneTouch Ultra 2 glucose meter; LifeScan) and insulin (Insulin ELISA Kit; Crystal Chem). Total lean and fat mass was assessed in conscious mice using the EchoMRI-100 system (Echo Medical Systems).

Chow-fed male mice (7–9 weeks of age) were anesthetized with isoflurane and a solution of 100 μL of hyaluronidase (Sigma H-3506) (2 μg/μL in Tyrode’s buffer) was injected into the triceps surae and tibialis anterior (TA) (two separate injections) in each leg 1 h before the DNA injection. Mice were then injected intramuscularly in the tibialis and triceps surae with 100 μg human UCN2 plasmid construct (catalog number RC201333, RefSeq NM_033199.3; Origene) (two separate injections; 1 μg/μL in Tris-EDTA [TE] buffer) and an equal amount of TE buffer in the contralateral leg as a control. Thereafter, the leg was subjected to electroporation (mode LV, 99 ms/500 V, voltage 150 V, four 20-ms pulses one per second, 150 V/cm) using a BTX 830M electroporation unit (BTX, Holliston, MA) fitted with gene caliper electrodes (BTX). Four days after electroporation, mice were subjected to in vivo glucose uptake.

Glycogen and triacylglyceride (TAG) content in liver and TA muscle were measured in 4-h–fasted mice using a Glycogen Assay Kit (ab65620; Abcam) or Triglyceride Quantification Assay Kit (ab65336; Abcam) according to the manufacturer’s protocol. Plasma free fatty acid (ab65341; Abcam) and plasma leptin (MOB00; R&D Systems) were analyzed using assay kits according to the manufacturers’ protocol.

Western blot analysis was performed as previously described (25). Primary antibodies used are listed in Supplementary Table 1. Bands were quantified using Quantity One 1-D analysis software (Bio-Rad) and normalized to total protein staining with Ponceau S (Sigma-Aldrich).

Significant differences were determined by one-way, two-way, or two-way repeated-measures ANOVA with Sidak multiple comparison post hoc test as indicated in the figure legends. Chow-fed mice were excluded from statistical analysis because they served as a control for the HFD. Comparisons were considered significant at P < 0.05. Analyses were performed using GraphPad 7 statistical software (GraphPad Software).

The modified human UCN2 peptide (compound A) is based on the previously reported compound 8 (26). In contrast to compound 8, compound A includes a cysteine residue at position 29, where a PEG 20,000 is attached through an acetamide-based linker. The potency of compound A was assessed by cAMP production in HEK293 cells transfected with mouse CRHR1 or CRHR2β plasmid (the predominant skeletal muscle isoform) (Table 1). CRHR2β-transfected cells treated with serial dilutions of compound A for 15 min had a half-maximal effective concentration value of 0.31 nmol/L compared with 0.08 nmol/L for human UCN2, while there was no cAMP production in cells transfected with CRHR1. Time to maximum plasma concentration after mice were treated with a single subcutaneous injection of compound A was ∼4 h, while the clearance was estimated at 5.94 mL/h/kg. Compound A has a half-life of 22.3 h compared with 15 min for the native UCN2 peptide (27).

To investigate the metabolic effects of modified UCN2 peptide, male mice were fed an HFD for 20 weeks before daily subcutaneous injections of vehicle or compound A (0.3 mg/kg) for 14 days. Body weight was reduced in compound A–treated mice after day 1, with a cumulative weight loss of 7.5 g over the 14-day treatment period (Fig. 1A and B and Table 2). Weight loss in compound A–treated mice was accompanied by an 84% initial reduction in food intake compared with controls, which was followed by a 28% decrease at the end of the treatment period (Fig. 1C and Table 2). Thus, we noted a large decrease in food intake in the first several days and a smaller, but still significant decrease later in the treatment. The reduction in food intake after the first injection with compound A is likely due to an initial reduction in gastric emptying whereby a single compound A injection reduced gastric emptying by 52% compared with vehicle (Supplementary Fig. 1). Feed efficiency (the ratio of body weight change to food intake) was initially reduced 34% with compound A treatment, which was followed by a significant reduction at the end of the treatment period (Fig. 1D). Weight loss in compound A–treated mice was attributed to a decrease in fat mass without alteration in lean mass (Fig. 1E and Table 2). Compound A treatment reduced TAG content in TA muscle compared with vehicle-treated mice (Fig. 1F). Compound A treatment reduced liver weight without altering hepatic TAG or glycogen content (Table 1).

To investigate whether compound A treatment has a synergistic effect with physical activity, mice were given free access to running wheels over the 14-day treatment period. In vehicle-treated mice, wheel running reduced body fat mass (Fig. 1E) without altering lean mass (Table 2), and this was associated with reduced final body weight compared with sedentary mice (Table 2). Wheel running reduced the absolute and percent weight loss (Fig. 1A and B) and final body weight in compound A–treated mice compared with vehicle-treated wheel running mice (Table 2), despite less distance ran (Supplementary Fig. 2). The weight loss was attributed to decreased fat mass (Fig. 1E) without altering lean mass, liver weight (Table 2), or TA TAGs (Fig. 1F) compared with vehicle-treated wheel running mice. There were no synergistic effects with wheel running in the phenotypic improvements seen with compound A treatment alone. This highlights the potent nature of compound A on weight loss and fat mass reduction.

Next, we investigated the whole-body glucose homeostasis of compound A–treated mice. Compound A–treated mice had reduced fasting blood glucose (Fig. 2A) and fasting plasma insulin (Fig. 2B) compared with HFD-fed vehicle-treated mice. Despite reduced plasma insulin (Fig. 2C), compound A–treated mice cleared the same amount of blood glucose as vehicle-treated control mice during a glucose tolerance test (GTT) (Fig. 2D and E). Moreover, in vivo insulin-stimulated glucose uptake into red and white quadriceps and EDL from compound A–treated mice was enhanced compared with insulin-stimulated vehicle-treated control muscles, while there was no alteration in insulin-stimulated glucose uptake in brown adipose tissue (BAT) (Fig. 2F).

Although wheel running in vehicle-treated mice did not affect fasting blood glucose levels compared with sedentary vehicle-treated mice (Fig. 2A), it reduced fasting plasma insulin levels (Fig. 2B). Wheel running vehicle-treated mice cleared the same amount of glucose during a GTT as sedentary vehicle-treated mice (Fig. 2D and E), even with reduced plasma insulin (Fig. 2C). These data suggest that wheel running improves insulin sensitivity in HFD-fed mice.

Wheel running in compound A–treated mice reduced fasting blood glucose (Fig. 2A) and fasting plasma insulin (Fig. 2B) compared with vehicle-treated wheel running mice. Blood glucose clearance was increased in compound A–treated wheel running mice compared with wheel running controls (Fig. 2D and E), although similar plasma insulin levels were observed during the GTT (Fig. 2C). We found no differences between compound A–treated sedentary and compound A–treated wheel running mice in the aforementioned parameters of in vivo glucose homeostasis. Thus, while wheel running in vehicle-treated mice improved glucose homeostasis, compound A treatment alone was more potent, supporting the notion that UCN2 peptide treatment in HFD-fed mice improves skeletal muscle and whole-body insulin sensitivity with increased glucose uptake.

To further investigate potential mechanisms of action for the positive effect of compound A on whole-body glucose homeostasis, we assessed insulin sensitivity of EDL muscle ex vivo after compound treatment. Insulin-stimulated Akt phosphorylation at Ser⁴⁷³ and Thr³⁰⁸ was increased in EDL from compound A–treated mice compared with vehicle-treated mice (Fig. 3A and B). TBC1D4 phosphorylation at Ser³¹⁸ was increased above basal levels in compound A–treated EDL muscle (Supplementary Fig. 3A), whereas insulin-stimulated phosphorylation at Thr⁶⁴² was decreased compared with controls (Supplementary Fig. 3B). A trend for increased insulin-stimulated GSK3α phosphorylation at Ser²¹ was observed in EDL muscle from compound A–treated mice compared with vehicle-treated EDL (P = 0.087) (Supplementary Fig. 3C). Total Akt protein was increased in EDL muscle from compound A–treated mice (Fig. 3D), whereas protein abundance of downstream targets, such as GLUT 4, GSK3α, GSK3β, or glycogen synthase, were unaltered (Supplementary Fig. 3D–F). Insulin-stimulated glucose transport into the EDL muscle was increased in compound A–treated mice compared with vehicle-treated mice (Fig. 3C).

To assess the potential direct effect of UCN2 on skeletal muscle metabolism, without the confounding factors of reduced adiposity and food intake with in vivo compound A treatment, TA and triceps surae muscles from lean chow-fed mice were electroporated with vectors containing human UCN2 or TE buffer in the contralateral leg. Expression of human UCN2 was detected in TA and soleus muscle after transfection with human UCN2 vectors (Supplementary Fig. 4A). UCN2 overexpression in mouse muscle was associated with a modest reduction in endogenous Ucn2 expression (Supplementary Fig. 4A). UCN2 overexpression increased glucose transport into soleus muscle compared with the control contralateral leg (Supplementary Fig. 4B), confirming a positive regulation of skeletal muscle glucose transport by UCN2. In L6-GLUT4-Myc myoblasts, compound A (100 nmol/L) enhanced GLUT4 translocation to the membrane to levels comparable to insulin stimulation (100 nmol/L) (Fig. 4A and B). Compound A stimulation also increased glucose uptake ex vivo in skeletal muscle. Isolated soleus muscle from lean chow-fed mice was incubated with compound A (63.3 nmol/L), with or without submaximal insulin (0.18 nmol/L) for 1 h. Compound A increased insulin-stimulated glucose uptake into soleus muscle compared with insulin-stimulated vehicle treatment (Fig. 4C). Akt phosphorylation at both Ser⁴⁷³ and Thr³⁰⁸ was increased in response to insulin and compound A stimulation compared with insulin stimulation alone (Fig. 4C–E). Mammalian target of rapamycin (mTOR) phosphorylation at Ser²⁴⁴⁸ and Ser²⁴⁸¹ was increased in response to compound A in an insulin-independent manner (Fig. 4D, G, and H).

In the context of changing demographic patterns and the aging population, current pharmacological treatments to combat the majority of lifestyle-related conditions are inadequate. In particular, pharmacological treatments for type 2 diabetes that specifically target skeletal muscle to increase insulin sensitivity and preserve skeletal muscle function are lacking. In this regard, CRHR2 agonists may improve skeletal muscle substrate metabolism and mitigate aging-associated disorders. Here, we determined the effects of a modified UCN2 peptide in HFD, obese mice. We show that compound A treatment of HFD-fed mice results in an initial reduction in food intake and rapid weight loss, which was accompanied by improved whole-body glucose tolerance and insulin-stimulated glucose uptake into skeletal muscle. Mechanistically, this could be due to an effect on skeletal muscle because ex vivo stimulation of soleus muscle from lean chow-fed mice with compound A increased glucose uptake and insulin signaling. Thus, UCN2 peptides may be efficacious in the treatment of type 2 diabetes by acting as insulin sensitizers.

Genetic manipulation of CRF family members alters body weight in mouse models. While body weight in Ucn2 knockout mice is unaltered after 16 weeks on an HFD, fat mass is reduced, and lean mass is increased (11). Conversely, overexpression of Ucn3, which also signals through CRHR2, increases body weight, with increased lean mass in chow-fed transgenic mice, whereas the HFD-fed transgenic mice are obesity resistant (15). These genetic models are at the whole-body level, and therefore, the contribution of a centrally mediated effect on metabolism cannot be excluded. Here, we provide evidence that pharmacological activation of CRHR2 with a modified UCN2 peptide reduced body weight in HFD-fed mice. These results are primarily localized to peripheral tissues because the PEGylated compound cannot cross the blood-brain barrier. Thus, strategies to activate CRHR2 in peripheral tissues appear to have positive effects on energy homeostasis. In accordance with our results, Ucn2 adeno-associated virus gene transfer attenuates weight gain in HFD-fed mice (17). Our results represent a pharmacological approach to activating the CRHR2 peripherally with a PEGylated UCN2 compound, whereas genetic models/approaches represent a supraphysiological event that may not portray the normal activity of the pathway. Our results also highlight potential discrepancies between activating CRHR2 with genetic models from birth versus transient activation of CRHR2 with pharmacological treatments.

Exercise and diet are considered a first-line treatment of insulin resistance and type 2 diabetes. For many patients, pharmacological intervention is required to manage this disease, yet effective insulin sensitizers are lacking from the current diabetes pharmacopeia. Here, we provide evidence that UCN2 peptide treatment reduced fasting hyperglycemia and hyperinsulinemia in obese mice. Despite lower insulin levels during a GTT, compound A treatment enhanced glucose tolerance during a GTT compared with vehicle-treated obese mice, indicating that UCN2 treatment improves insulin sensitivity. This is consistent with an earlier study reporting that Ucn2 gene transfer improves glycemia and insulin sensitivity in HFD-fed and db/db mice (17). Collectively, these results implicate peripheral action of UCN2 therapies for the treatment of obesity and insulin resistance. Nevertheless, we cannot exclude the possibility that compound A acts on the pancreas. UCN3, but not UCN2, is expressed in the β-cells of the pancreas and acts in an autocrine manner on CRHR2 to regulate glucose-stimulated insulin production and secretion, particularly in conditions of nutrient excess (28). Theoretically, the UCN2 peptide used here may activate pancreatic β-cell CRHR2 and stimulate insulin secretion; however, this remains to be determined.

The mechanistic basis of UCN2 treatment may involve enhanced insulin signaling. We have reported that insulin signaling and glucose transport are impaired in skeletal muscle from patients with type 2 diabetes (4,5). Thus, targeting components of the canonical insulin signaling cascade or GLUT4 transport machinery in skeletal muscle may improve glucose homeostasis (29,30). Indeed, we found that UCN2 peptide treatment in HFD-fed mice increased Akt phosphorylation and protein abundance in skeletal muscle concomitant with enhanced insulin-stimulated glucose uptake. Consistent with this, transient Ucn3 overexpression in skeletal muscle increases protein abundance of IRS1, Akt, TBC1D4, and GSK3α/β (16). Insulin-stimulated TBC1D4 phosphorylation at Ser³¹⁸ was increased in response to compound A treatment, while phosphorylation at Thr⁶⁴² was reduced compared with controls. However, the exact mechanisms by which CRHR2 affects insulin signaling are unknown. Nonetheless, insulin-stimulated glucose uptake is increased in response to either acute or chronic compound A treatment. Moreover, we found that compound A acutely promoted GLUT4 translocation, which may account for the increased glucose clearance during the GTT as well as enhanced glucose uptake in isolated skeletal muscle. Improved glucose homeostasis by Ucn2 gene transfer in HFD-fed mice was attributed to increased GLUT4 translocation (17). Along with increased skeletal muscle glucose uptake, glycogen content was unaltered (data not shown), suggesting that compound A improved glucose metabolism by increasing glucose oxidation. Thus, the enhanced glucose uptake and metabolism in compound A–treated mice is not only due to a weight loss effect but also due to direct action on skeletal muscle.

CRHR2 is differentially expressed in peripheral tissues, including cardiac and skeletal muscle (31), adipose tissue (32), skin (9), and the gastrointestinal tract (33), where it serves diverse functions. Activation of CRHR2 in the gastrointestinal tract is involved in gastric motility (34) and intestinal inflammation (35), while activation in cardiac tissue is involved in blood pressure regulation (36). In the current study, subcutaneous administration of compound A targets CRHR2, which is present throughout the periphery and could therefore have numerous effects in multiple organs controlling whole-body glucose and energy homeostasis. Selective agonists for UCN2 and UCN3 reduce gastric emptying (34,37). As such, we observed an initial decrease in food intake and a corresponding reduction in wheel running after the first day of treatment, which could be attributed to a decrease in gastric emptying and the accompanying malaise. By the end of the treatment period this effect was attenuated; however, without the inclusion of a pair-fed control group, the proportion of the metabolic effect during the in vivo treatment that was related to the reduced food intake is uncertain. Given the effect of compound A to reduce body weight, altered energy expenditure or thermogenesis could play a role. The hypothalamus is unlikely to be a direct target of compound A because of the PEGylation, which results in poor blood-brain barrier drug penetration. Thus, any potential neuroendocrine effect of compound A on energy homeostasis at the level of the hypothalamus is likely to be secondary. However, without the inclusion of a PEG-vehicle control, we cannot fully exclude the possibility of a central component of compound A on the regulation of energy homeostasis. We also do not believe that BAT is a major target of compound A because basal- or insulin-stimulated glucose uptake was unaltered. Nevertheless, we cannot exclude the possibility that non-insulin-mediated metabolic processes in BAT are affected. Skeletal muscle appears to be a direct target of compound A. Compound A directly increases GLUT4 translocation in L6 cells and increases insulin-stimulated glucose uptake and insulin signaling in isolated soleus muscle from chow-fed mice. Additionally, electroporation of skeletal muscle with a UCN2 plasmid increases glucose uptake. Thus, compound A has a direct and immediate effect on skeletal muscle metabolism independent of changes in adiposity. Our main findings related to the physiological effects of this approach to control skeletal muscle insulin sensitivity and body weight is schematically highlighted (Fig. 5).

CRHR2 activation in skeletal muscle enhances AMPK signaling, which increases glucose disposal (16,38) while also activating AMPK in cardiac tissues (39). However, in the current study, AMPK and downstream signaling, such as pACC (a surrogate marker for AMPK activation), was not altered (data not shown). In contrast, compound A treatment increased mTOR phosphorylation, implicating a role in anabolic processes. A role for IGF-I signaling in UCN3-mediated hypertrophy of soleus, tibialis cranialis, and gastrocnemius muscle and glucose disposal has also been proposed (15,16). However, plasma IGF-I levels after UCN2 treatment were unaltered (data not shown). Differences between these studies may be accounted for by the models studied (Ucn3 transgenic mice and overexpression in rats vs. subcutaneous injection in HFD mice) or the specific ligand used to activate CRHR2 (UCN3 vs. UCN2), resulting in different signaling/downstream effects. In support of this, signaling through either CRHR or G-protein–coupled receptors confers distinct conformational changes, which elicit different coupling of the G-proteins and activation of signaling cascades (40–42). Specifically, UCN1 binding to CRHR1 or CRHR2 leads to CREB and mitogen-activated protein kinase phosphorylation, whereas CRF binding does not (43,44). Furthermore, Ucn2 gene transfer increases glucose disposal in mice, while Ucn3 gene transfer has no effect (45). Thus, the use of specific ligands may fine-tune specific effects on metabolic or gene regulatory pathways to influence glucose or energy homeostasis.

Exercise training increases insulin sensitivity and glucose uptake in skeletal muscle of obese patients and prevents type 2 diabetes progression (46,47). We determined whether UCN2 treatment and voluntary wheel running have a synergistic effect on skeletal muscle insulin sensitivity. In vehicle-treated obese mice, wheel running reduced hyperinsulinemia and increased insulin sensitivity during a GTT, while addition of compound A treatment produced negligible effects over the treatment alone. Thus, increased physical activity does not further enhance the insulin sensitizing effects of compound A possibly because of the potent nature of compound A treatment alone.

G-protein–coupled receptors are the target of many modern pharmaceutical drugs. There are currently no pharmacological agents that target skeletal muscle for the treatment of type 2 diabetes. An agent that not only increases skeletal muscle insulin sensitivity but also reduces body weight would be highly desired to treat the growing metabolically perturbed population. Indeed, acute UCN2 peptide infusion is currently being tested clinically as an adjunct treatment in patients with heart failure (48–50), although a treatment for type 2 diabetes requires a more long-term regimen. In conclusion, our results fill a therapeutic void by providing new evidence for a treatment for type 2 diabetes that acts on skeletal muscle to enhance insulin sensitivity and glucose transport.

Funding. Vetenskapsrådet (Swedish Research Council) (2011-3550, 2015-00165); Swedish Diabetes Foundation (DIA2015-032); Stiftelsen för Strategisk Forskning (Swedish Foundation for Strategic Research) (SRL10-0027); and Novo Nordisk Foundation, Strategic Research Programme in Diabetes, at Karolinska Institutet (Swedish Research Council grant number 2009-1068) supported this research. M.L.B. is supported by the Swedish Society for Medical Research.

Duality of Interest. L.G., M.W., J.A.-F., R.M., A.R., S.B., T.C., E.O., E.M.N., and J.T.B. are employees of Eli Lilly. J.R.Z. received compound A as a gift from Eli Lilly. No other potential conflicts of interest relevant to this article were reported.

Author Contributions. M.L.B., J.M., M.S., T.D.C.B., L.G., M.W., J.A.-F., R.M., A.R., S.B., T.C., E.O., E.M.N., A.V.C., H.K.K., and J.T.B. researched data. M.L.B., J.M., H.K.K., J.T.B., and J.R.Z. analyzed and interpreted the data. M.L.B., H.K.K., J.T.B., and J.R.Z. designed the study. M.W., A.K., and J.T.B. contributed to the discussion and reviewed and edited the manuscript. M.L.B. and J.R.Z. wrote the manuscript. All authors approved the manuscript. J.R.Z. is the guarantor of this work and, as such, had full access to all the data in the study and takes responsibility for the integrity of the data and the accuracy of the data analysis.