Esculentin-2CHa-Related Peptides Modulate Islet Cell Function and Improve Glucose Tolerance in Mice with Diet-Induced Obesity and Insulin Resistance

The frog skin host-defense peptide esculentin-2CHa (GFSSIFRGVA10KFASKGLGK D20LAKLGVDLVA30CKISKQC) displays antimicrobial, antitumor, and immunomodulatory properties. This study investigated the antidiabetic actions of the peptide and selected analogues. Esculentin-2CHa stimulated insulin secretion from rat BRIN-BD11 clonal pancreatic β-cells at concentrations greater than 0.3 nM without cytotoxicity by a mechanism involving membrane depolarization and increase of intracellular Ca2+. Insulinotropic activity was attenuated by activation of KATP channels, inhibition of voltage-dependent Ca2+ channels and chelation of extracellular Ca2+. The [L21K], [L24K], [D20K, D27K] and [C31S,C37S] analogues were more potent but less effective than esculentin-2CHa whereas the [L28K] and [C31K] analogues were both more potent and produced a significantly (P < 0.001) greater maximum response. Acute administration of [L28K]esculentin-2CHa (75 nmol/kg body weight) to high fat fed mice with obesity and insulin resistance enhanced glucose tolerance and insulin secretion. Twice-daily administration of this dose of [L28K]esculentin-2CHa for 28 days had no significant effect on body weight, food intake, indirect calorimetry or body composition. However, mice exhibited decreased non-fasting plasma glucose (P < 0.05), increased non-fasting plasma insulin (P < 0.05) as well as improved glucose tolerance and insulin secretion (P < 0.01) following both oral and intraperitoneal glucose loads. Impaired responses of isolated islets from high fat fed mice to established insulin secretagogues were restored by [L28K]esculentin-2CHa treatment. Peptide treatment was accompanied by significantly lower plasma and pancreatic glucagon levels and normalization of α-cell mass. Circulating triglyceride concentrations were decreased but plasma cholesterol and LDL concentrations were not significantly affected. The data encourage further investigation of the potential of esculentin-2CHa related peptides for treatment of patients with type 2 diabetes.

membrane depolarization and increase of intracellular Ca 2+ . Insulinotropic activity was attenuated by activation of K ATP channels, inhibition of voltage-dependent Ca 2+ channels and chelation of extracellular Ca 2+ . The [L21K], [L24K], [D20K, D27K] and [C31S,C37S] analogues were more potent but less effective than esculentin-2CHa whereas the [L28K] and [C31K] analogues were both more potent and produced a significantly (P < 0.001) greater maximum response. Acute administration of [L28K]esculentin-2CHa (75 nmol/kg body weight) to high fat fed mice with obesity and insulin resistance enhanced glucose tolerance and insulin secretion. Twice-daily administration of this dose of [L28K]esculentin-2CHa for 28 days had no significant effect on body weight, food intake, indirect calorimetry or body composition. However, mice exhibited decreased non-fasting plasma glucose (P < 0.05), increased non-fasting plasma insulin (P < 0.05) as well as improved glucose tolerance and insulin secretion (P < 0.01) following both oral and intraperitoneal glucose loads. Impaired responses of isolated islets from high fat fed mice to established insulin secretagogues were restored by [L28K]esculentin-2CHa treatment. Peptide treatment was accompanied by significantly lower plasma and pancreatic glucagon levels and normalization of α-cell mass. Circulating triglyceride concentrations were decreased but plasma cholesterol and LDL concentrations were not significantly affected. The data encourage further investigation of the potential of esculentin-2CHa related peptides for treatment of patients with type 2 diabetes.

Introduction
Skin secretions of many species of Anura (frogs and toads) represent a valuable source of peptides with therapeutic potential. These peptides, whose primary function is believed to be hostdefence, are best known for their antimicrobial properties and there are many examples of components with potent activity against multidrug-resistant pathogenic Gram-positive and Gram-negative bacteria and fungi (reviewed in [1][2][3][4]). However, frog skin host-defense peptides are multifunctional and may also display anti-tumor, antiviral, immunomodulatory, and chemoattractive activities (reviewed in [5]). In addition, myotropic peptides produced in the skin, such as those related to mammalian tachykinins, bradykinin and CCK-8, may play a role in deterring ingestion by predators [6]. There are no conserved structural domains in amphibian host-defense peptides that are responsible for biological activity but with few exceptions they are cationic and have the propensity to adopt an amphipathic, α-helical conformation in a membrane-mimetic solvent or in the environment of a phospholipid vesicle [2].
One such host-defense peptide with therapeutic potential is esculentin-2CHa (GFSSIFRGV AKFASKGLGKDLAKLGVDLVACKISKQC), first isolated from norepinephrine-stimulated skin secretions of the Chiricahua leopard frog Lithobates chiricahuensis (Ranidae) [7]. This compound shows potent broad-spectrum antimicrobial properties including activity against clinical isolates of multidrug-resistant strains of Staphylococcus aureus, Acinetobacter baumannii, and Stenotrophomonas maltophilia. The peptide also stimulates the release of the antiinflammatory cytokine IL-10 by mouse lymphoid cells and displays high cytotoxic potency against human non-small cell lung adenocarcinoma A549 cells but relatively low hemolytic activity against human erythrocytes [8]. Interestingly, such actions on cytokine production by mouse lymphoid cells were also extended to TNF-alpha which may impact on beta cells [8]. Structure-activity studies indicate that removal of the either hydrophobic N-terminal hexapeptide (GFSSIF) or the cyclic C-terminal domain (CKISKQC) and replacement of the Cys 31 and Cys 37 residues by serine results in appreciable decreases in cytotoxicity against microorganisms and mammalian cells. In contrast, increasing cationicity by substitution of the Asp 20 and Asp 27 residues by L-Lysine resulted in a modest increase in potency against all microorganisms tested (up to 4-fold) [8].
The current pandemic of type 2 diabetes has necessitated a search for new types of therapeutic agents and naturally occurring incretin peptides that stimulate insulin release in response to high circulating glucose concentration are receiving increasing attention. Several long-acting analogues of the potent incretin GLP-1 are currently in clinical use [9]. A number of frog skin peptides that were first identified on the basis of their ability to inhibit the growth of microorganisms have subsequently been shown to possess the ability to release insulin from the BRIN-BD11 clonal β-cells and isolated mouse islets at low concentrations that are not cytotoxic to the cells and to improve glucose tolerance in mice following acute administration (reviewed in [5,10]). The high fat fed mouse exhibits obesity, glucose intolerance and insulin resistance and so is a useful model for a preliminary investigation of the therapeutic potential of peptides in treatment of patients with type 2 diabetes [11]. More recently, it has been shown that twice daily treatment of high-fat fed mice for up to 28 days with tigerinin-1R [12], magainin-AM1 [13], and CPF-SE1 [14] results in an improvement in glucose tolerance, insulin sensitivity, and islet β-cell secretory responsiveness.
In the present study, the antidiabetic potential of esculentin-2CHa and selected analogues with increased cationicity (Table 1) was assessed in vitro using BRIN-BD11 cells and isolated mouse islets and in vivo in studies using the high fat fed mouse.

Peptide synthesis and purification
Synthetic esculentin-2CHa and its analogues (Table 1) were purchased in crude form GL Biochem Ltd (Shanghai, China) and purified to near homogeneity (> 98% pure) by reversedphase HLPC on a (2.2 cm x 25 cm) Vydac 218TP1022 (C18) column equilibrated with acetonitrile/water/triflouroacetic acid (21.0/78.9/0.1 v/v) mobile phase at a flow rate of 6 ml/min. The concentration of acetonitrile in the eluting buffer was raised to 56% (v/v) over 60 min. The molecular masses of the peptides were confirmed using MALDI-TOF mass spectrometry.

In vitro insulin-releasing studies
In vitro insulin-releasing effects of esculentin-2CHa and its analogues were assessed using BRIN-BD11 rat clonal β-cells and mouse islets. In the first set of experiments, BRIN-BD11 cells were incubated with the peptides in the concentration range (1 pM-3 μM) in Krebs-Ringer bicarbonate buffer containing 5.6 mM glucose for 20 min at 37°C as previously described [15][16][17]. Insulin-releasing effects were also assessed using buffer supplemented with 1.4 and 16.7mm glucose concentrations and with established modulators of insulin release as previously described [15][16][17]. In a second set of experiments, islets from NIH Swiss mice, isolated by collagenase digestion [18,19], were incubated with a range of concentrations of esculentin-2CHa or its analogues (0.1 nM-1 μM) for 1 h in Krebs-Ringer bicarbonate (KRB) buffer supplemented with 1.4, 5.6 or 16.7 mM glucose. Following test incubations, aliquots of buffer were retrieved and stored at -20°C for measurement of insulin by radioimmunoassay [20]. In the experiments with BRIN-BD11 cells, the release of lactate dehydrogenase (LDH) was measured as an indicator of the integrity of the plasma membrane using commercially available CytoTox 96 non-radioactive cytotoxicity assay kit (Promega, Madison, WI, USA) according to the manufacturer's recommended protocol.

Membrane potential studies and intracellular calcium ([Ca 2+ ] i )
Effects of esculentin-2CHa (1 μM) and its analogues on membrane potential and [Ca 2+ ] i in BRIN-BD11 cells over a period of 5 min were assessed using commercially available FLIPR membrane or calcium assay kit (Molecular Devices, USA) as previously described [15]. Data were captured using a Flexstation 3 microplate reader equipped with automatic fluid transfer unit (Molecular Devices, USA).

Laboratory animals
Male National Institutes of Health (NIH) Swiss mice (Harlan Ltd, UK) were housed separately in an air-conditioned room (22 ± 2°C) with relative humidity of 51 ± 5% and a 12-hour light:

Acute in vivo studies
For acute in vivo studies, overnight fasted, high fat fed mice (n = 8) received an intraperitoneal injection of glucose alone (18 mmol/kg body weight) or in combination with esculentin-2CHa or its analogues (75 nmol/kg body weight). This peptide dose was selected on the basis of a pilot study that examined acute effects of different doses of esculentin-2CHa on glucose tolerance in lean mice. Blood samples were collected as described previously before injection and at times indicated in the Figures for the measurement of plasma glucose and insulin concentrations.

Longer-term in vivo studies
High-fat fed mice with clearly manifested features of obesity and hyperglycaemia received twice daily injections of either saline vehicle (0.9% (w/v) (high fat fed controls) or [L28K]esculentin-2CHa (75 nmol/kg body weight) for 28 days. This peptide was chosen based on its potent in vitro and acute in vivo actions. Mice (n = 8) fed standard rodent diet and injected with saline were used as lean controls. Energy intake, bodyweight, non-fasting blood glucose and plasma insulin concentrations were monitored every 72 h throughout the duration of the study. At the end of the 28 day treatment period, glucose tolerance (18 mmol/kg body weight, intraperitoneal or oral, overnight fasted) and insulin sensitivity (25 U/kg body weight) were assessed as previously described [12][13][14]. Indirect calorimetry and energy expenditure in treated and control mice were measured using the Comprehensive Laboratory Animal Monitoring System (CLAMS) metabolic chambers (Columbus Instruments, Columbus, OH, USA). Total body lean and fat mass, bone mineral density and bone mineral content were also measured using DXA scanning (Piximus Densitometer, USA) [21]. Improvement in beta cell function in [L28K]esculentin-2CHa-treated and control mice were evaluated from the insulin secretory responses of islets isolated from these animals to established insulin secretagogues and incretin hormones (16.7 mM glucose, 1 μM GLP-1, 1 μM GIP, 10 mM alanine, 10 mM arginine and 30 mM KCl). Changes in islet morphology were assessed using pancreatic tissues excised from mice treated with [L28K]esculentin-2CHa or saline for 28 days as previously described [12][13][14].

Biochemical measurements
Pancreatic tissues were homogenized in 20 mM Tris-HCl, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA and 0.5% Triton X 100; pH 7.5 as previously described [14]. Blood samples (approximately 150 μl), collected from the cut tip of the tail vein of unanesthetized mice at intervals indicated in the Figures were used for blood glucose measurements and determination of plasma insulin as previously described [13,14]. Blood glucose was measured using a hand-held Ascencia Contour meter (Bayer Healthcare, UK). Plasma and pancreatic insulin were determined by radioimmunoassay [20]. Plasma and pancreatic glucagon contents were determined by ELISA using commercially available kit (Millipore, MA, USA). Plasma creatinine, alanine transaminase (ALT), aspartic acid transaminase (AST) and alkaline phosphatase (ALP) were measured using commercially available kits (Randox Laboratories, UK) as indicators of renal and liver function. Plasma triglyceride and cholesterol concentrations were measured using an automated clinical chemistry analyser (I-lab 650, and reagents purchased from Instrumentation Laboratory (Warrington, UK). The assay kit for cholesterol was obtained from Randox Laboratories (Antrim, UK). Estimation of LDL cholesterol concentrations was achieved using the Friedewald equation as described previously [22].

Statistical analysis
Results are expressed as mean ± S.E.M. Values were compared using one-way ANOVA followed by Student-Newman-Keuls post hoc test. Area under the curve (AUC) analysis was performed using the trapezoidal rule with baseline correction. P < 0.05 was considered statistically significant.

Effects of 28 day administration of [L28K]esculentin-2CHa on body weight, food intake and non-fasting plasma glucose and insulin concentrations
Based on its superior in vitro and acute in vivo effects, [L28K]esculentin-2CHa was selected for longer-term studies using high fat fed mice. These animals exhibited increased body weight, food intake, plasma glucose and insulin concentrations compared with control animals fed a standard diet (Fig 4). Treatment with [L28K]esculentin-2CHa for 28 days did not affect body weight or food intake but resulted in 25% (P < 0.05, Fig 4C) reduction in non-fasting plasma glucose and a 42% (P < 0.05, Fig 4D) increase in non-fasting plasma insulin compared with saline-treated controls.

Effects of 28 day administration of [L28K]esculentin-2CHa on indirect calorimetry, energy expenditure and body composition in high fat fed mice
Oxygen (O 2 ) consumption, CO 2 production and energy expenditure in high fat fed mice increased by 30% (P < 0.001), 22% (P < 0.001) and 27% (P < 0.001) respectively but these parameters together with respiratory exchange ratio were not affected by treatment with

Effects of 28 day administration of [L28K]esculentin-2CHa on pancreatic and plasma glucagon content, lipid profile and both renal and liver function
The elevated pancreatic and plasma glucagon levels generated by high fat feeding were reduced by 20% (P < 0.05) and 35% (P < 0.05) in [L28K]esculentin-2CHa-treated mice (Fig 7A and 7B). Plasma total cholesterol were similar in all groups while high fat fed mice exhibited reduced plasma HDL cholesterol (23%, P < 0.05) and elevated plasma triglycerides (1.2-fold, P < 0.05) (Fig 7C). Treatment with [L28K]esculentin-2CHa for 28 days had no effect on HDL cholesterol but significantly (P < 0.05) decreased plasma triglycerides. Plasma levels of AST and ALT were similar in all groups of mice, plasma ALP was decreased (43%, P < 0.01) and creatinine increased (140%, P < 0.001) in high fat fed mice compared with lean controls (Fig 7D and 7E). Administration of [L28K]esculentin-2CHa for 28 days resulted in 43% (P < 0.05, Fig 7D) increase in plasma ALP and 40% (P < 0.05, Fig 7E) decrease in plasma creatinine levels.

Effects of 28 day administration of [L28K]esculentin-2CHa on islet morphology, beta cell and alpha cell area
Islets isolated from saline-treated high fat fed mice were significantly larger than those isolated from lean mice (Fig 8A). The number of islets per mm 2 of pancreas as well as beta and alpha cells area were also significantly (P 0.05) increased in high fat fed mice (Fig 8B-8E). Treatment with [L28K]esculentin-2CHa did not affect the number of islets per mm 2 of pancreas but significantly reduced total islet area (44%, P<0.001) as well as beta (43%, P < 0.001) and alpha (48%, P < 0.001) cell areas.

Discussion
The increasing incidence of type 2 diabetes together with ever mounting challenges posed by associated complications have motivated efforts to discover novel antidiabetic agents with better therapeutic outcomes [23]. The present study provides further evidence for the potentially beneficial actions of the esculentin-2CHa family of peptides and highlights their possible utility in the management of type 2 diabetes. Consistent with observations using several other amphibian host defence peptides [10], esculentin-2CHa stimulates insulin release from BRIN-BD11 cells in a dose-dependent and glucose-responsive manner at concentrations that are non-toxic to the cells. However, magnitude of the effects of the peptide, particularly at the highest concentration tested (3 μM), were appreciably greater than those previously reported for other amphibian host-defence peptides [9]. Indeed, the stimulatory effect of esculetin-2CHa on isolated mouse islets is comparable in magnitude to that of GLP-1 under the same experimental conditions [24,25]. Membrane depolarization and enhanced intracellular Ca 2+ concentrations together with the reduced insulin release in the presence of diazoxide, verapamil or Ca 2+ chelation suggest that the insulinotropic action of esculentin-2CHa, in common with other but not all previously identified insulinotropic amphibian skin peptides [5,10], may involve the activation of the K ATP -dependent pathway of insulin secretion [26]. This is fully supported by initial patch clam experiments evaluating effects of [Arg4]tigerinin-1R in clonal BRIN-BD11 beta cells (R. C. Moffett et al., unpublished observations). This analogue derived from tigerinin-1R isolated from the skin secretions of Hoplobatrachus rugulosus has similar action prolife to esculentin-2CHa. However, effects on other elements of beta cell signal transduction pathways such as glucose metabolism (glycolysis or mitochondrial metabolism), intracellular cyclic AMP generation or activation of phospholipase C should not be ruled out. Indeed, insulin secretion was largely preserved in the presence of diazoxide or verapamil, as well as under Ca 2+ free conditions, indicating that almost half of the insulin secretory response induced by esculentin-2CHa is dependent neither on K ATP channel closure nor L-type Ca 2+ channel activation. We have demonstrated previously in studies using the amphibian peptides pseudin-2 [15], B2RP [27], alyteserin-2a [28] and tigerinin-1R [29] that site-specific amino acid substitutions that result in increasing cationicity may lead to significant enhancement of their in vitro insulin releasing actions. In this study, such amino acid substitutions resulted in increased net charge and isoelectric point (Table 1)   Values are mean ± SEM with n = 8 mice. *P < 0.05, *** P < 0.001 compared to high fat fed control. All parameters were significantly lower in lean mice than high fat fed control mice (P < 0.05 -P < 0.001). Consistent with previous studies [11], mice fed a high fat diet exhibited glucose intolerance and insulin resistance. Acute injection with esculentin-2CHa significantly improved both oral and intraperitoneal glucose tolerance and increased insulin secretion. These in vivo actions were greater with the [D28K] and [C31S] analogues while the in vitro insulinotropic actions of the other peptide analogues were not replicated in vivo. The in vivo effects of [L28K] and [C31K]esculentin-2CHa in these animals were more pronounced than those previously reported for other amphibian skin peptides [12][13][14]30]. Twice-daily administration of [L28K] esculentin-2CHa for 28 days did not affect body weight and food intake but reduced hyperglycaemia and elevated plasma insulin concentrations in the non-fasting state. This was associated with improved glucose tolerance following oral and intraperitoneal glucose administration accompanied by enhanced insulin secretion. These observations are broadly consistent with effects of tigerinin-1R [12], [I10W]-tigerinin-1R [30] and magainin-AM2 [13], suggesting a possible similar spectrum of actions in this animal model of obesity and diabetes. We have previously shown that the antidiabetic effects of magainin-AM2 [13] and [I10W]-tigerinin-1R [30] were accompanied by improvement in insulin sensitivity. However, in the present study, the insulin resistance induced by high fat feeding was not alleviated in mice treated with [L28K]esculentin-2CHa. It is probable, therefore, that the improvement of glucose homeostasis observed in this study is largely due the effects of the peptide on islet cell function. Consistent with this view, the reduced insulin-secretory responses to incretin hormones and other insulin secretagogues observed with islets isolated from saline treated high fat fed mice were reversed by treatment with [L28K]esculentin-2CHa. . Insulin sensitivity tests were performed using 25 U/kg body weight of insulin injected intraperitoneally. All tests were conducted following twice-daily treatment of mice with saline or [L28K]esculentin-2CHa (75 nmol/kg body weight) for 28 days. Values are mean ± SEM with n = 8 mice. *P < 0.05, **P < 0.01 and ***P < 0.001 compared with saline-treated lean mice. Δ P < 0.05, ΔΔ P < 0.01 compared with high-fat fed control.  *P < 0.05,**P < 0.01, ***P < 0.001 compared with the response of islets isolated from the same group of mice at 16.7 mM glucose; Δ P < 0.05, ΔΔ P < 0.01, ΔΔΔ P < 0.001 compared with the respective response of islets isolated from lean control. For Fig 6D, stimulation index refers to fold-increase in insulin secretion from 1.4 mM glucose to 16.7 mM glucose or from 16.7 mM glucose to that observed in the presence of each secretagogue. *P < 0.05, **P < 0.01 compared with the stimulation index of islets isolated from each group of mice at 16.7mM glucose. Δ P < 0.05, ΔΔ P < 0.01 compared the stimulation index of lean control. Unrestrained hepatic glucose output and elevated glucagon secretion make an important contribution to the development of hyperglycaemia in type 2 diabetes [31]. Interesting, the elevation of pancreatic and plasma glucagon concentrations as well as α-cell area in saline-treated high fat fed mice were normalized by treatment with [L28K]esculentin-2CHa. Although the exact molecular mechanism through which the peptide inhibits α-cell function is not yet known, our results suggest that such effects contribute to the improvement of hyperglycaemia and glucose tolerance. Thus, further studies are required to assess whether [L28K]esculentin-2CHa acts directly on β cells or by indirect effects mediated via enteroendocrine secretions such as GLP-1. Increased islet size accompanying enhanced insulin output and defective β-cell function caused by high fat feeding were also reversed in [L28K]esculentin-2CHa-treated mice. Our preliminary work with analogues of the parent peptide indicate that this may be due to correction of the ratio of beta cell proliferation to apoptosis.
Energy metabolism was not affected by treatment with [L28K]esculentin-2CHa. Similarly, the peptide did not affect adipose tissue deposition, bone mineral content or bone mineral density in high fat fed mice. Bone area and lean body mass were increased in [L28K]esculentin- 2CHa-treated mice, but this did not translate to changes in overall body weight. Further studies are required to evaluate long-term impact of [L28K]esculentin-2Cha on circulating cytokines and insulin signalling pathways. Importantly, no adverse effects were observed in mice treated with the peptide, including unchanged circulating concentrations of AST, ALT and ALP together with improved creatinine clearance and lower circulating triglycerides. Although longer term toxicological studies are obviously needed, these observations suggest both benefits and lack of toxicity of [L28K]esculentin-2CHa.

Conclusions
In conclusion, this study has demonstrated the beneficial effects of esculentin-2CHa family of peptides in mice with diet-induced obesity-diabetes.
[L28K]esculentin-2CHa exhibited better in vitro insulinotropic potential compared with the native peptide and significantly improved glucose tolerance and enhanced insulin secretion by positive effects on beta cell function in high fat fed mice. These observations together with the previous demonstrated antimicrobial actions [8] encourage further evaluation of the potential of [L28K]esculentin-2CHa as a novel agent for managing patients with type 2 diabetes and associated microbial infections. Values are mean ± SEM for 8 observations (~120 islets per group). *P < 0.05, **P < 0.01, ***P < 0.001 compared to lean control. ΔΔΔ P < 0.001 compared to high fat control. doi:10.1371/journal.pone.0141549.g008 Supporting Information S1 Fig. Effects of [L28K]esculentin-2CHa on O 2 consumption (A, B), CO 2 production (C, D), respiratory exchange ratio (E, F) and energy expenditure (G, H) in lean and high-fat fed treated with saline or [L28K]esculentin-2CHa (75nmol/kg bw) for 28 days. Mice were placed in CLAMS metabolic chambers, and O 2 consumption or CO 2 production were measured for 30s at 15min intervals. RER was calculated by dividing VCO 2 by VO 2 . Energy expenditure was computed using the formula (3.815 + 1.232 x RER) x VO 2 . Values are means ± SEM for 6 mice. ÃÃ P<0.01, ÃÃÃ P<0.001 compared with saline-treated lean mice. Shaded bar indicates dark phase.