In vitro and in vivo characterization of a novel long-acting GLP-1 receptor agonist, exendin-4–Fc fusion protein

State Key Laboratory of Biotherapy/Collabo West China Hospital, West China Medic Section 3, South Renmin Road, Chengdu, tracy73@gmail.com; Tel: +86-18628182400 Lanzhou Institute of Biological Products Co General Medicine Department, West China 37 Guo Xue Xiang, Chengdu, Sichuan, 610 com; Tel: +86-18980601359 Pharmacy and Pharmacology Department, † Electronic supplementary informa 10.1039/c7ra10822b ‡ Two authors contributed equally to this Cite this: RSC Adv., 2017, 7, 54178


Introduction
Obesity and the associated diseases have become a growing global health problem. According to World Health Organization's 2014 report, more than 1.9 billion adults worldwide are overweight and over 600 million are obese. 1 Obesity exacerbates dyslipidemia, insulin resistance, hyperinsulinemia, and hyperglycemia. 2 There are many anti-obesity drugs and current agents available; however, most of these drugs have treatment-limiting side effects. 3 Low toxicity and long-acting drugs are required to be researched and developed to improve life quality. Glucagonlike peptide-1 (GLP-1), a polypeptide consisting of 30 amino acids, is an incretin hormone secreted by L-cell in the distal portion of the intestine in response to food ingestion. GLP-1 also stimulates insulin secretion from the pancreatic islet bcell. 4 As a potential therapeutic agent in the treatment of type II diabetes mellitus (T2DM), 5 GLP-1 enables weight control, inhibits gastric emptying, and enhances the glucose disposal; but the endogenous GLP-1 is rapidly degraded by dipeptidyl peptidase IV (DPP-IV) with a half-life of 1-2 min, therefore limiting the native GLP-1 clinical therapeutic potential. 6,7 Thus, GLP-1 analogues and their long-acting formulations have been developed.
Exendin-4 (Ex-4) is a 39 amino acid peptide isolated from the salivary glands of the Gila monster lizard (Heloderma suspectum), which shares 53% sequence homology with GLP-1. Unlike GLP-1, Ex-4 is resistant to degradation by DPP-IV, 8 so it is a stable peptide. Ex-4 treatment has been demonstrated to promote a signicant reduction in high-fat-diet (HFD) induced obesity, food intake, and gastric emptying. 9 Although Ex-4 exhibits an extended half-life compared to the native GLP-1, it still requires once-or twice-daily injections necessitating the development of many long-acting analogues. 10 For example, bydureon and lixisenatide, two analogues of GLP-1R agonists, were approved by the US Food and Drug Administration (FDA) for the treatment of T2DM in 2012 (ref. 11 and 12) and 2016, 13,14 respectively. Bydureon, a long-acting formulation of exenatide, requires once-weekly injection, whereas lixisenatide requires once-daily injection.
The aim of this study is to develop a safe, efficacious and longer acting potential anti-diabetic agent. Currently, several peptide-Fc fusion proteins have been developed 15,16 and the protein fusion technologies have contributed to the development of long-acting GLP-1 analogues. [17][18][19] In the current study, we designed a long-acting GLP-1R agonist, exendin-4-IgG4 Fc fusion protein ((Ex-4) 2 -Fc), in which Ex-4 was fused to a mutated immunoglobulin G4 (IgG4) Fc fragment. Although peptide-Fc fusion technologies have been widely used, no reports have demonstrated that two tandem copies of Ex-4 molecules were fused to a mutant human IgG4 constant heavy-chain. A single subcutaneous (s.c.) injection of (Ex-4) 2 -Fc signicantly prolonged its half-life in plasma ($122 h) in male Sprague Dawley (SD) rats, potentially allowing once-weekly dosing. In addition, diet-induced obesity (DIO) mice as well as ob/ob and db/db mice with the treatment of (Ex-4) 2 -Fc, once every six days, effectively decreased body weight and food intake and improved glucose intolerance. Our results suggest that this Fc-fusion protein can prolong the half-life, improve glucose metabolism effectively, and provide a new therapy for diabetes and obesity.

Purication of fusion proteins
The fusion proteins were produced in a 5 L scale bioreactor (Sartorius, German) using the serum-free medium for 10-12 days. The culture supernatant was concentrated by hollow-ber ultraltration (10 kDa) (GE healthcare). The fusion proteins were obtained by two-step purication with MabSelect affinity chromatography and Q anion-exchange chromatography (GE Healthcare). Then, the fusion proteins were analyzed by SDS-PAGE and western blot, sterile-ltered and stored at À20 C.

Internalization of GLP1R-EGFP stimulated with (Ex-4) 2 -Fc
Recombinant U 2 OS cells derived from stably expressing human GLP-1 receptor was fused to the N-terminus of an enhanced green uorescent protein (EGFP). The cells were plated in 96well plates at a density of 6 Â 10 3 cells per well. 100 nM Ex-4-Fc or (Ex-4) 2 -Fc was added to the cell plate and then it was incubated for 1 h at 37 C in a 5% and 95% humidity incubator, respectively. Aer 1 h, the cells were washed 4 times with 200 mL PBS per well per wash and then 100 mL 1 mM Hoechst staining solution was added. The cells were incubated at room temperature for 30 min before imaging.

Molecular docking
The model of the interaction between the extracellular domain of GLP-1 receptor and the fusion proteins of Ex-4-Fc and (Ex-4) 2 -Fc was performed using the ZDOCK module in Accelrys Discovery studio 3.5 soware.

Analysis of puried (Ex-4) 2 -Fc
The single chain samples without glycosylation were denatured and reduced using 6 M guanidine hydrochloride in 10 mM dithiothreitol (DTT) at 56 C for 30 min and then treated with PNGase F. A 5600+ QTOF mass spectrometer (AB SCIEX, USA) and an ACQUITY UPLC Protein BEH C4 Column (300Å, 1.2 mm 2.1 mm Â 100 mm (Agilent, USA)) were used to measure the molecular weights of (Ex-4) 2 -Fc fusion protein and its subunits. Approximately 10 mg of each sample was injected into the column at 80% mobile phase A (0.1% FA in water) and 20% mobile phase B (0.1% FA in acetonitrile). Aer 7 min, the percentage of mobile phase B was increased to 60% for 16 min. The column was thoroughly washed and then equilibrated with 20% mobile phase B before the next injection. Throughout the analysis, the ow-rate was set at 200 mL min À1 and the column temperature was set at 80 C. The mass spectrometer was run in the positive mode with the following settings: a scan range of 600-5000 m/z, gas temperature of 350 C, ion source gas of 30 L min À1 , and collision energy at 10 eV.

Animals and diets
All animal care and experimental procedures in this study were in compliance with the guidelines set by the Guide for the Care and Use of Laboratory Animals and approved by the Institutional Animal Care and Use Committee of Sichuan University (Chengdu, China). Six week-old male C57BL/6 mice and male Sprague-Dawley (SD) rats were purchased from Beijing Hua Fu Kang Biological Technology Co., Ltd. (Beijing, China). Male ob/ob and db/db mice were purchased from Model Animal Research Center (MARC) of Nanjing University (Nanjing, China). They were maintained at ambient temperature of 25 AE 1 C with a constant humidity and a 12 h light/dark cycle and fed ad libitum. C57BL/6 mice were fed either a high-fat diet (HFD) containing 60% of calories as fat (D12492, Research Diets, New Brunswick, NJ, USA) or a standard laboratory diet (STD, Rat & Mouse Maintenance Diet) for 12 weeks and 6 weekold male ob/ob and db/db mice were fed the standard laboratory diet for 2 weeks before treatment.

In vivo evaluation of (Ex-4) 2 -Fc
The HFD-fed obese mice were treated with (Ex-4) 2 -Fc (1.8 mg kg À1 ) or PBS as the vehicle control once every 6 days for 14 days by s.c. injection. The 3 rd group of the age-matched control mice on the standard diet received PBS for 14 days. Body weight and food intake were recorded daily. At the end of the experiment, mice were sacriced by cervical dislocation. Next, inguinal white adipose tissue (iWAT) and epididymal white adipose tissue (eWAT) were immediately dissected, weighed, xed in 4% paraformaldehyde or snap-frozen in liquid nitrogen, and then stored at À80 C. Blood was collected and stored at À80 C. Similar to the DIO mice, ob/ob and db/db mice were treated with (Ex-4) 2 -Fc or the vehicle control for 14 days and the metabolic parameters were determined.

Glucose and insulin tolerance tests
To evaluate glucose tolerance, the overnight-fasted mice were administered a single intraperitoneal (i.p.) injection of glucose (2 g kg À1 ), followed by a tail vein blood sampling at 0, 30, 60, and 120 min for the DIO mice and at 0, 15, 30, 45, 60, 90, and 120 min for ob/ob and db/db mice. The blood glucose levels were monitored using the AccuChek Active glucometer (Roche Diagnostic, Mannheim, Germany). To determine insulin tolerance, the 6 h fasted DIO mice were i.p. injected with human insulin (0.5 U kg À1 , Novo Nordisk, Bagsvaerd, Denmark), followed by the tail vein blood collection. The blood glucose levels were determined at the same intervals as described above.

Assays for plasma insulin level
Blood was collected from the tail vein, following 12 h fasting of mice before treatment and aer 14 days treatment. The plasma level of insulin was determined using Mouse Insulin ELISA (Millipore, Billerica, MA).

Histological analysis
eWAT tissues xed in 4% paraformaldehyde were processed for paraffin embedding, sectioned at 4 mm, and stained with hematoxylin and eosin (H&E). The mean adipocyte size was determined using NIH Image J soware.

Statistical analysis
All data are presented as mean AE SEM. The differences between 2 groups were analyzed by unpaired two-tailed Student's t-test and multiple independent groups were compared by 2-way ANOVA with an appropriate post hoc comparison using GraphPad Prism 6 soware (GraphPad Soware, San Diego, CA). A value of p < 0.05 was considered as statistically signicant.

Design, expression, and purication of Ex-4-Fc and (Ex-4) 2 -Fc fusion proteins
The fusion proteins, (Ex-4) 2 -Fc and Ex-4-Fc, consisting of two copies and one copy of the Ex-4 molecules, respectively, were linked via a (GGGGS) 3 linker to a mutant human IgG4 Fc segment (Fig. 1A). The puried (Ex-4) 2 -Fc and Ex-4-Fc proteins were analyzed using SDS-PAGE and Western Blotting (WB). The results showed that under the reducing electrophoretic conditions, the Ex-4-Fc and (Ex-4) 2 -Fc proteins existed as monomers of $31 kDa and 37 kDa, respectively ( Fig. 1B and C).

Bioassay of fusion proteins in vitro
The biological activity of (Ex-4) 2 -Fc and Ex-4-Fc was assessed in the U 2 OS cells derived from stably expressing GLP-1R, using an in vitro cyclic adenosine monophosphate (cAMP) bioassay. The results showed that (Ex-4) 2 -Fc and Ex-4-Fc displayed the activity with ED 50 of approximately 0.64 mg L À1 and 7.11 mg L À1 , respectively ( Fig. 2A), indicating that the bioactivity of (Ex-4) 2 -Fc was over 10-fold stronger than that of Ex-4-Fc. Then, we studied the difference between two forms of the Ex-4 fusion proteins using the structural analysis. The molecular docking between each fusion protein and the extracellular domain of GLP-1 receptor was performed using the ZDOCK module in the Accelrys Discovery Studio 3.5 soware. The structures of (Ex-4) 2 -Fc and Ex-4-Fc were rst conrmed by homology modeling method using the solution conformation of Ex-4 and the crystallography conformation of IgG4 as templates. According to the docking results, Ex-4 can form an unabridged alpha-helix when binding to the GLP-1 receptor, but the Ex-4 fragment of Ex-4-Fc can only maintain a shorter alpha-helix to conform to the binding site (Fig. 2B). Moreover, two Ex-4 arms in (Ex-4) 2 -Fc play different roles in the ligand-receptor complex: one arm was directly bound to the GLP-1 receptor and the other arm played an auxiliary role, such that it increases the stability of the (Ex-4) 2 -Fc/GLP-1 receptor complexes.
To assess the effects of the GLP-1R activation, Ex-4-Fc and (Ex-4) 2 -Fc were investigated using the recombinant U 2 OS cells that stably express human GLP-1 receptor fused to the Nterminus enhanced green uorescent protein (EGFP). The  majority of GLP-1R-EGFP of the unstimulated cell is localized in the plasma membrane and the majority of GLP-1R-EGFP of the stimulated cell is internalized in endosomes. The results showed that more GLP-1R-EGFP of cells was internalized in endosomes aer treated with (Ex-4) 2 -Fc than treated with Ex-4-Fc (Fig. 2C), suggesting that (Ex-4) 2 -Fc was more effective in activation of GLP-1R than Ex-4-Fc. These results conrmed that (Ex-4) 2 -Fc had more potent bioactivity than that of Ex-4-Fc.

Analysis of (Ex-4) 2 -Fc fusion protein
The collected fractions were rst analyzed for intact molecular weight by LC-MS to detect posttranslational modications that are present at signicant levels with substantial molecular weight differences compared to the theoretical molecular weight calculated from the known amino acids. The deconvoluted mass spectra are shown in Fig. 3A. The molecular weight of the predominant species in the main peak fraction is 74 397.6 Da, which is in agreement with the theoretical molecular weight (74 394.4 Da) of (Ex-4) 2 -Fc with G0F on both of the Fc chains. Three additional peaks correspond to (Ex-4) 2 -Fc with other glycoforms. The peak with the molecular weight of 74 559.7 Da corresponds to (Ex-4) 2 -Fc with G0F on one chain and G1F on the other chain. The peak with the molecular weight of 74 721.7 Da corresponds to (Ex-4) 2 -Fc with either G1F on both heavy chains or G0F on one chain and G2F on the other chain.
The proteins were reduced and then digested using PNGase F to generate a single chain and then analyzed by LC-MS to further conrm the observed molecular weight. The deconvoluted mass spectra are shown in Fig. 3B. The molecular weight of the peak is 35 760.5 Da, which is in agreement with the theoretical molecular weight of 35 758.7 Da. The observation of the molecular weight indicates the correct expression of the protein with no unexpected modications.

The pharmacokinetics of (Ex-4) 2 -Fc in SD rats
The pharmacokinetics of a single dose of (Ex-4) 2 -Fc was studied in SD rats in comparison to Ex-4. Aer a single dose of 0.3 mg kg À1 Ex-4 by s.c. injection, the plasma concentration of Ex-4 rapidly reached the peak value (C max of 382.27 ng mL À1 ) and declined with a half-life (t 1/2 ) of 0.56 h (Fig. 4A). Moreover, a single s.c. dose of (Ex-4) 2 -Fc at 0.45 and 1.35 mg kg À1 had a slower absorption phage and slowly reached C max of 51.20 and 128.33 ng mL À1 , respectively. The half-life of (Ex-4) 2 -Fc was 122.16 and 122.40 h, respectively, much longer than that of Ex-4 (Fig. 4B). The pharmacokinetic parameters of (Ex-4) 2 -Fc in SD rats are summarized in (Table 1). These results suggest that (Ex-4) 2 -Fc could effectively prolong the half-life compared with Ex-4.
In addition, the (Ex-4) 2 -Fc group was administered with (Ex-4) 2 -Fc once every 6 days via s.c. injection at a dose of 1.8 mg kg À1 for 14 days in the DIO mice and the vehicle control group was administered with PBS once every 6 days via s.c. injection. We found that such (Ex-4) 2 -Fc treatment signicantly (p < 0.001) reduced the HFD-induced weight gain compared to the vehicle control (Fig. 5A) and also signicantly reduced food intake (Fig. 5B). The adipocyte size in the (Ex-4) 2 -Fc treated mice fed with the HFD and the vehicle control mice fed with the STD showed signicant reduction than that observed in the vehicle control mice fed with the HFD (Fig. 5C and D). Furthermore, epididymal and inguinal fat pad weight in the (Ex-4) 2 -Fc treated mice fed with the HFD and the vehicle control mice fed with the STD were signicantly decreased compared to those in the vehicle control mice fed with the HFD (Fig. 5E). The effect of the (Ex-4) 2 -Fc group on biochemical indices is presented in Table S1. † Cholesterol (CHOl) and low-density lipoprotein (LDL) were signicantly decreased (p < 0.01) and high-density lipoprotein (HDL) was signicantly higher (p < 0.01) compared to those in the vehicle control mice fed with the HFD. There was no difference between the (Ex-4) 2 -Fc group and the vehicle control mice fed with the STD in the analysis of biochemical criterion (Table S1 †).

(Ex-4) 2 -Fc improves glucose intolerance, increases insulin sensitivity in DIO mice
Since obesity is frequently associated with insulin resistance and glucose homeostasis, 21 we next evaluated whether (Ex-4) 2 -Fc alters the systemic insulin resistance and glucose intolerance in the DIO mice by performing insulin tolerance (ITT) and glucose tolerance (GTT) tests. Interestingly, when the blood glucose levels were measured at 0, 30, 60, and 120 min aer a glucose challenge, the (Ex-4) 2 -Fc administration signicantly improved glycemic excursion and AUC compared with the vehicle control in the DIO mice (Fig. 6A), suggesting that (Ex-4) 2 -Fc improved glucose intolerance. Moreover, the (Ex-4) 2 -Fc-treated DIO mice, compared to the untreated ones, exhibited signicantly reduced blood glucose levels at 60 and 120 min aer insulin administration (Fig. 6B), indicating that (Ex-4) 2 -Fc attenuated insulin resistance. In addition, plasma insulin and glucose concentrations were normalized aer the (Ex-4) 2 -Fc administration at the end of 14 day treatment ( Fig. 6C and D). These data suggest that (Ex-4) 2 -Fc improves glucose metabolism and insulin resistance in the DIO mice.
3.7 Weight-reducing effect and glucose tolerance of (Ex-4) 2 -Fc in ob/ob and db/db mice To assess whether (Ex-4) 2 -Fc has efficacy in other mouse models of obesity, leptin-decient ob/ob or leptin receptor-defective db/ db mice were treated with (Ex-4) 2 -Fc (1.8 mg kg À1 ) by s.c. injection once every 6 days over a 14 day period. We found that the (Ex-4) 2 -Fc treatment resulted in a sustained and signicant decrease in body weight of ob/ob mice by 10% (p < 0.001) (Fig. 7A) and db/db mice by 8% (p < 0.01) (Fig. 7C), compared to the vehicle treatment. It is worth mentioning that the treatment with (Ex-4) 2 -Fc caused a marked decrease in food intake in the ob/ob and db/db mice on the rst day of treatment, but food intake was then recovered rapidly ( Fig. 7B and D). Aer 14 days of treatment, we evaluated the glucose-lowering effects of (Ex-4) 2 -Fc in ob/ob and db/db mice during an intraperitoneal glucose tolerance test (IPGTT) (2 g of glucose per kg of body weight). The blood glucose levels were monitored at 0, 15, 30, 45, 60, 90, and 120 min aer the administration of the glucose    load. The administration of (Ex-4) 2 -Fc signicantly improved glucose tolerance by 20.2% and 20.4% (assessed as total AUC during IPGTT) in ob/ob and db/db mice, respectively ( Fig. 7E and F). Moreover, aer 14 days of treatment, ob/ob and db/db mice were fasted for 6 h and then intraperitoneally injected with insulin (0.5 units perkg of body weight). Indeed, in ob/ob and db/db mice, the glucose-lowering effect aer i.p. insulin injection in the (Ex-4) 2 -Fc group was signicantly greater than that in the vehicle control group (Fig. S2 †). Collectively, the decrease in body weight, food intake, and the improvement of glucose tolerance and insulin tolerance by (Ex-4) 2 -Fc was also evident in ob/ob and db/db mice.  aim to improve the bioactivity through dimerization and to extend half-life, mediated by the neonatal Fc receptor (FcRn), for example, BR3-Fc fusion protein (briobacept), rFIXFc, and rFVIII-Fc. 25 Immunoglobulins of the IgG1, IgG2 and IgG4 subclass can exhibit a long half-life of 3-4 weeks by a FcRnmediated recycling process although their half-life is also affected by antigen-dependent elimination, such as through receptor-mediated endocytosis and intracellular degradation. 26,27 A previous study has shown that Ex-4/Fc, a fusion protein consisting of exendin-4 and mouse IgG1 heavy chain constant region (IgG1 Fc), improves energy and lipid metabolism. 17 IgG1 Fc has generally been used as a conjugate to prolong the half-life of a bioactive molecule, but has its shortcomings, such as complement-dependent and antibodydependent cell-mediated cytotoxicity. 28 When compared together, there is a 10-fold difference in the ability of IgG1 and IgG4 to bind to Fcg RI (CD64, high-affinity Fc receptor). 29 Therefore, we replaced IgG1 with a variant IgG4, containing 3 amino acid substitutions relative to native IgG4 Fc (S228P, F234A, and L235A) and deleted the C-terminal lysine of the IgG4 Fc to lower the immunogenicity and stabilize the IgG4 peptide. 30 . (Ex-4) 2 -Fc was prepared using mammalian cell lines with CHO cell lines. In addition to the extended half-life, the Fc domain could also provide other benets for the fusion protein, such as enhancing expression and secretion using Protein-A chromatography to provide a facile purication strategy as well as improving solubility and stability. The plasma pharmacokinetic proles of (Ex-4) 2 -Fc have shown a dramatically longer half-life and larger AUC 0-21 d compared with those of the Ex-4 administered as a single injection in the SD rats. The halflife of (Ex-4) 2 -Fc was prolonged by ve days, almost 250-fold longer than that of Ex-4 in rats, suggesting that the injection frequency could be reduced from daily to weekly.

Discussion
To eliminate the potential ability of body weight reduction and glucose improvement in rodents, the treatment of the DIO mice with (Ex-4) 2 -Fc for 14 days resulted in a signicant attenuation of weight gain as compared with the treatment with the vehicle control. It also signicantly reduced food intake, decreased fasting blood, insulin and glucose levels, and improved glucose tolerance and insulin sensitivity. We also used genetically modied obese ob/ob mice and db/db mice to conrm the effects of (Ex-4) 2 -Fc on reducing body weight, food intake, and glucose level. By cross-comparison of the potency of (Ex-4) 2 -Fc in three mouse models of diabetes, we noticed that compared with the vehicle control treatment, the (Ex-4) 2 -Fc treatment resulted in a greater than $20% reduction in body weight in the DIO mice and $10% reduction in body weight in ob/ob or db/db mice. Furthermore, the administration of (Ex-4) 2 -Fc modestly improved glucose intolerance and reduced food intake in ob/ob and db/db mice, which is in agreement with the ndings of the previous studies. 31 These data indicated that the (Ex-4) 2 -Fc fusion protein occurred without compromising its biological activity and the in vitro potency and in vivo efficacy were sustained.
In summary, the low-administration frequency of (Ex-4) 2 -Fc was shown to have the long-lasting effects on body weight reduction and glucose tolerance improvement in a variety of rodents. The pharmacokinetics and pharmacodynamics proles of (Ex-4) 2 -Fc strongly indicate that (Ex-4) 2 -Fc could be potentially used as a clinical treatment for obesity and diabetes.