Exenatide loaded PLGA microspheres for long-acting antidiabetic therapy: preparation, characterization, pharmacokinetics and pharmacodynamics

Abstract

Clinical application of exenatide, a peptide drug widely used for the treatment of type 2 diabetes mellitus, is greatly limited due to its short plasma half-life of 2.4 hours. To prolong the half-life of exenatide, a water in oil in oil (W/O/O) method was employed to prepare exenatide-loaded PLGA microspheres with a satisfactory particle size of 75.22 μm and a span value of 1.13. Optimized exenatide-loaded microspheres were endowed with high entrapment efficiency (83.8 ± 1.3%) and low initial burst release (1.31 ± 0.13%). Degradation pattern of the peptide-loaded microspheres was investigated by monitoring the changes in molecular weight, appearance, and release profile. Histological study proved the safety of the microspheres. In the pharmacokinetics and pharmacodynamics study, the long acting formulation had relative bioavailability of 70.31% and it achieved a hypoglycemic activity for up to 3 weeks in diabetic mice. A single subcutaneous injection of exenatide-loaded microspheres exhibited a comparable effect on controlling blood glucose to exenatide solution, which was injected twice per day at the same dose of exenatide. In conclusion, the exenatide-loaded microspheres might be a promising long acting formulation for glycemic control with low initial burst release and reduced risk of gastrointestinal intolerance and hypoglycemia.

---

Introduction

Biomacromolecules, such as gene, protein or peptide drugs, have long been developed as therapeutics, however, their poor stability, low bioavailability and short half-life present challenges to clinical applications.^1–3 Proteins and peptides could be used as medicines if an efficient drug delivery system can in place allow the bioactive molecules to reach their target sites at the right time.⁴

As a severe global health problem, diabetes currently affect millions of people worldwide with type 2 diabetes accounting for around 90% of cases.⁵ The pathogenesis of type 2 diabetes is characterized both by insulin resistance in muscle, fat, and liver and a relative failure of the pancreatic β cell.⁶ As the first reported incretin mimetics, exenatide is approved, by FDA and European Medicines Agency to treat type 2 diabetes. Consisting of 39 amino acids, exenatide shares about 53% homology with mammalian gut hormone (GLP-1).^7,8 Exenatide's glucose-dependent enhancement of insulin secretion may be mediated by exenatide binding to the pancreatic GLP-1 receptor.⁹ In diabetic animal models and insulin-secreting cell lines, exenatide and GLP-1 are reported to improve β-cell function by increasing the expression of key genes involved in insulin secretion, insulin biosynthesis, and β-cell mass through multiple mechanisms.¹⁰ Pre-clinical studies also indicate that exenatide and GLP-1 reduce food intake, cause weight loss, and sensitize insulin responsiveness.¹¹ Due to its relatively longer half-life, i.e., 2.4 h for exenatide versus 1–2 min for GLP-1,^12,13 the former can be subcutaneously administered twice daily. Currently, commercial exenatide formulations includes a twice-daily and a weekly subcutaneous injection preparation (Byetta™ and Bydureon™).^14,15 Many researchers strive to further prolong the duration of exenatide by developing novel sustained-release drug delivery systems, modifying its structure, and other strategies.¹⁶

Poly(D,L-lactic-co-glycolic acid) (PLGA), approved by the Food and Drug Administration (FDA) as a polymer material with desirable safety,¹⁷ has gained tremendous attention due to its great biodegradability and biocompatibility. Moreover, there are many cases showing the successful application of PLGA in sustained release of proteins/polypeptides.^18–20 Several peptides, i.e., leuprolide acetate, goserelin acetate and octreotide acetate, have been incorporated into PLGA and commercialized as sustained-release products, i.e. Lupron Depot®, Zoladex®, and Sandostatin LAR®.²¹

For preparation of microspheres, solvent evaporation, solvent extraction and co-solvent methods are among the most popular approaches.²² However, injectable microspheres still release 10–80% drug rapidly at the very beginning of administration. This so-called “initial burst” phenomenon poses a severe toxicity threat and is a major hurdle for the development of microsphere products. The initial burst is widely believed to be the result of rapid release of drug from the microsphere surface.²³ As PLGA is lipophilic and exenatide hydrophilic, exenatide-loaded PLGA microspheres always have limitations, such as low drug loading content and high initial burst release. The initial burst release, therefore, turns out to be the crucial drawback of PLGA microspheres while the long-acting preparation usually carries much more drug than conventional one. Uncontrollable initial release of sustained release preparation may lead to an overdose to patients.^24,25 Nevertheless, a poor initial release control, i.e., about 10 percent of the drug released on the first day, could be found in most literatures.^1,26,27

In this study, exenatide-loaded PLGA microspheres with low initial burst release and long acting activity were produced by a W/O/O method. The degradation pattern of exenatide release from drug-loaded microspheres was appraised and the pharmacokinetics and pharmacodynamics profiles of the microspheres were investigated.

Materials and methods

Materials

PLGA (5050 4A, inherent viscosity 0.35–0.45 dL g⁻¹) was received from Lakeshore Biomaterials (Birmingham, AL, USA). Exenatide (purity 98.5%) was purchased from Sunlida Biological Technology Co. (Nanjing, China). Streptozotocin (STZ) was purchased from Sigma-Aldrich (St. Louis, Missouri). All other reagents used were of at least analytical grade.

Male Sprague-Dawley rats weighing 190–210 g and Kun Ming (KM) mice with body weight of 18–20 g were purchased from Qinglongshan Animal Center (Nanjing, China). All animal experiments were conducted in accordance with the terms regulated by the National Institute of Health Guide for the Care and Use of Laboratory Animals and approved by the Animal Ethics Committee of China Pharmaceutical University.

Preparation of exenatide-loaded PLGA microspheres

A W/O/O method was used to fabricate PLGA microspheres (Fig. 1). Briefly, exenatide was first dissolved in sucrose solution (2%) to obtain exenatide solution (0.225 mL, 57 mg mL⁻¹), then it was further emulsified by addition of 3 mL of PLGA solution in dichloromethane (DCM) (80 mg mL⁻¹); the resultant solution was sonicated for 60 s in ice bath to get primary emulsion (W/O). Afterwards, 4 mL of silicon oil (0 °C) was added into primary emulsion with continuous stirring for 3 min in ice bath to get the unformed microspheres. The resultant emulsion was immediately transferred into a special container, in which 100 mL of mixture of heptane and ethanol (10 : 1, w/w) was prior kept at 4 °C, with continuous stirring for 15 min. After this preparation was left under a standing condition for 5 min, the supernatant liquid was discarded and the remnant were further washed with fresh heptane to extract remaining DCM. This washing procedure was repeated for three times. After then the obtained crude microspheres were withheld in a 600-mesh sieve (0.025 mm) and vacuum dried at 4 °C for 12 hours then at 37 °C for 24 hours. Finally, the dried microspheres were passed through a 100-mesh sieve (0.150 mm).

Fig. 1 Schematic illustration of the W/O/O method used to prepare the exenatide microspheres.

Morphology and particle size distribution of microspheres

The shape and morphology of PLGA microspheres were observed using an ultra-high resolution scanning electron microscope (SU8010, Hitachi high-tech, Japan). Particle size distribution was measured by laser diffraction using Mastersizer 2000 (Malvern, UK).²⁸ The microspheres were dispersed in 0.2% tween-80 solution and particle size was analyzed with sonication.

The volume mean diameter (d_4,3) was evaluated according to the following equation, where n_i is the number of particles of diameter d_i.²⁹

d_4.3 = ∑n_id_i⁴/∑n_id_i³.

Span value was calculated as follows:

D_v,90%, D_v,50% and D_v,10% are volume size diameters at 90%, 50% and 10% of the cumulative volume, respectively. The smaller span value indicates the narrower particle size distribution.

Drug loading content (DLC) and entrapment efficiency (EE)

The DLC and EE of exenatide in microsphere was determined by dissolving 10 mg of exenatide microsphere in 5 mL of acetonitrile followed by same quantity of acetate buffer solution (pH 4.5). This mixture was centrifuged at 2500 rpm for 5 min and the supernatant was collected as the sample solution.

The amount of exenatide was quantified using high performance chromatography (HPLC) (Shimadzu, Japan); chromatographic study was performed at 30 °C on a reversed-phase column (Xtimate® SEC-300, Welch Materials, Inc, USA). Mobile phase was a mixture of 0.1% trifluoroacetic acid (TFA) in 2% sodium sulfate –0.1% TFA in acetonitrile (75 : 25, v/v) pumped at a flow-rate of 0.8 mL min⁻¹ (LC-20AT pump, Shimadzu, Japan). Detection performed at a wavelength of 210 nm (Ultra Violet detector, Shimadzu, Japan). Specificity, linearity, recovery, precision, accuracy and solution stability were performed for validation.³⁰

DLC and EE of the microspheres were calculated by following equations:³¹

Circular dichroism (CD) spectroscopic study

CD spectroscopy (Jasco J-810 spectrometer, Jasco, Japan) was exerted to investigate the secondary structure of exenatide.^32–34 The peptide was extracted from microspheres by dissolving about 5 mg of exenatide microspheres in 1 mL acetonitrile followed by addition of 3 mL of distilled water. The presence of acetonitrile from the solution was removed by rotary evaporation. Supernatant was collected as sample solution through centrifugation. Exenatide was collected by same method for the purpose of comparison and contrast as control substance. Its secondary structural integrity was compared with unprocessed one by CD analysis. The detection was performed using a 0.1 cm path cuvette at a scanning speed of 50 nm min⁻¹.

Polymer molecular assay

The average molecular weight of PLGA was determined using gel permeation chromatography (GPC).^35,36 Drug-loaded microspheres were dissolved in tetrahydrofuran (THF) followed by filtering through a nylon 66 micro-filtration membrane to remove insoluble peptides, then the separation was performed at 35 °C on a Shodex KF-803 column. THF was used as mobile phase. 20 μL sample was injected and the mobile phase was pumped at a flow-rate of 1 mL min⁻¹ (LC-20AT, Shimadzu, Japan). Detection was performed using a RID-10A refractive index detector (Shimadzu, Japan). The obtained chromatogram was interpreted using the GPC software.

Initial burst and in vitro release

Burst release was determined by incubating 10 mg of microspheres in 1 mL of buffer solutions with different pH, i.e., 4.5, 7.4 and 9.1, at 37 °C and 100 rpm. After 24 h incubation, all samples were centrifuged at 10 000 rpm for 5 min and the supernatant was collected for HPLC analysis of exenatide concentration.

For in vitro release of exenatide, 10 mg of microspheres was placed in 1 mL of buffer solutions with different pH, i.e., 4.5, 7.4 and 9.1. All samples were incubated at 37 °C and 100 rpm. At designated time points of 7 d, 14 d, 21 d, 28 d, 35 d, 42 d, 49 d and 56 d, three individual samples at each pH were taken out from the incubator. Supernatant was discarded after centrifugation at 10 000 rpm for 5 min. The remaining sediments of microspheres in tubes were washed with distilled water for three times, followed by collection for lyophilization. White powders were yielded as microsphere residuals. Then HPLC analysis were conducted to determine the remaining amount of exenatide in microsphere residuals following the method mentioned above. All tests were performed in triplicate.

Evaluation of tissue reaction

For tissue reaction study, male KM mice were taken and divided into three groups. First group was treated with injectable solvent (3% CMC-Na, 0.9% NaCl, 0.1% tween 20), second group was treated with exenatide solution, and third group was treated with exenatide loaded microsphere solution prepared into above injectable solvent. Each group was subcutaneous injected with the same concentration of drug, i.e., 0.3 mg mL⁻¹, and the dose was maintained at 1.6 mg kg⁻¹. Mice were sacrificed at designated time points, i.e. 1 d, 2 d, 4 d, 7 d, 14 d and 28 d, and tissues of injected site were excised and fixed in a 10% formaldehyde solution. The fixed tissues were then embedded in paraffin and sectioned using a 4 μm-thick microtome. The sectioned tissues mounted on glass slides were stained with haematoxylin and eosin (H&E) to evaluate the cellular reaction at injection site.^37–39

In vitro cytotoxicity study

MTT assay was performed for cytotoxicity analysis.^40,41 In brief, Mef (mouse embryonic fibroblast) cells were seeded in microplates at a concentration of 5 × 10⁴ cells per well in 100 μL medium, and incubated in incubator under an atmosphere of 5% CO₂ and 37 °C for 24 hours. Then 100 μL of DMEM media containing microspheres with different concentrations, i.e., 0, 5, 25, 50, 100 and 200 μg mL⁻¹, were added. After incubation for 24 hours and 48 hours at 37 °C, the culture medium was discarded, and 20 μL of MTT at a concentration of 0.5 mg mL⁻¹ was added for incubation for another 4 h at 37 °C. Subsequently, the culture medium was discarded, 100 μL of dimethylsulfoxide (DMSO) was added to wells. After dissolving the dark blue formazan crystals, the plates were read on a Multiscan spectrophotometer at a wavelength of 570 nm.

In vivo release pharmacokinetics

In vivo release study was performed in male SD rats weighing approximately 200 g. The animals were housed (n = 3 per cage) and kept under controlled conditions of 12 : 12 h light: dark cycle, temp. 22 ± 2 °C and RH 50 ± 15%. Food and water were provided ad libitum.

The rats were divided into two groups treated with exenatide solution (100 μg kg⁻¹) and exenatide loaded microspheres (1.4 mg kg⁻¹, equivalent to the dose of twice-daily injection of exenatide solution for 1 week²²), respectively. Blood samples from each group were collected in heparin sodium containing tubes at different time points. All samples were centrifuged at 2000 rpm for 20 min and the supernatant was stored at −20 °C. Quantification of exenatide was executed with the exendin-4 EIA kit (EK-070-94, Phoenix Pharmaceuticals, USA) as per the manufacturer's instructions.

Pharmacokinetic study and relative bioavailability

Pharmacokinetic study was conducted in normal SD rats as described in the in vivo release study. The relevant pharmacokinetic parameters, C_max, t_max, AUC were calculated based on a software, PK Slover.⁴² The relative bioavailability was calculated using following equation:⁴³
where, A represents the exenatide microspheres; B represents the free exenatide.

Pharmacodynamic study in diabetic mice

For drug efficacy study, the type 2 diabetes were induced in 8 weeks-old male KM mice by treatment with high-fat diet followed by STZ injection.⁴⁴ Briefly, the mice were randomly divided into normal diet control group and type 2 diabetes model group. All mice were housed under standard conditions with a 12 h dark/light cycle and an ambient temperature of 22–25 °C, and in diabetes model group, mice get free access to a high-fat diet (20% sucrose, 10% lard, 2.5% cholesterol, 1% cholate, and 66.5% normal chow diet; Qinglongshan Animal Center, Nanjing, China) different from normal diet control group and water was allowed throughout the study period. After 3 weeks, the mice were injected intraperitoneally with 100 mg kg⁻¹ STZ in 1% (w/v) citrate buffer (pH 4.3) after an overnight fast (free access to water). These mice had free access to food and water after the STZ injection. One week later, mice were fasted for 4 h and One Touch Basic Glucose Meter (OneTouch Ultra 2, Lifescan, USA) was used to measure the blood glucose. Mice with blood glucose level 11.1 mmol L⁻¹ were accepted for study.

18 diabetic mice were randomly divided into three groups, saline solution, exenatide solution (108 μg kg⁻¹) and exenatide microspheres (1.52 mg kg⁻¹, equivalent to the dose of twice-daily injection of exenatide solution for 1 week²²) were subcutaneously injected at flank. For saline and drug solution groups, mice are subcutaneously injected twice daily for 7 days, while exenatide microspheres were given only once on the first day. Thereafter, on designated time points, blood glucose levels were monitored for all mice under a non-fasting condition for 30 days. The relative change in glucose was calculated by the following equation.⁴⁵

Initial blood glucose means the blood glucose concentrations measured before exenatide injection. Two-way ANOVA was performed for statistical analysis and significance was labelled as *** for P < 0.001.

Results and discussion

Preparation of exenatide-loaded microspheres

In the process of microsphere preparation, water phase (sucrose solution) containing exenatide and oil phase (DCM) containing PLGA could be homogenated by Ultra-sonication to obtain a small-size W/O primary emulsion. While PLGA could not be dissolved in silicon oil, addition of this oil would drive the concentration of PLGA on the surface of dispersive aqueous drops due to miscibility of silicon oil and DCM. Furthermore, because of the surface tensions of different phases, the size of dispersive aqueous drops increased by incorporation to yield so-called unformed microspheres where exenatide was entrapped. And the particle size of microspheres could be easily regulated by controlling the time of dropping silicon oil. Thereafter, the above emulsion (no more than 8 mL) was transferred rapidly into the mixture of heptane and ethanol (100 mL, 10 : 1, w/w) to discontinue the incorporation of unformed microspheres and make them hardened and precipitated by immediate and complete substitution of DCM and silicon oil with heptane. Meanwhile, ethanol acted as the plasticizer of PLGA to swell the polymer matrix followed by a better remove of DCM. In the whole preparation process, exenatide was always entrapped in the unformed microspheres or in the deep polymer matrix, which is supposed to give a reasonable explanation to the low burst release of our product. Finally, residual heptane could be removed in the following drying stage.

Characterization of microspheres

As one of the most important characters of microspheres, particle size affects degradation rate, drug loading and initial burst release of microspheres. For injectable microsphere formulations, large particles caused severe incompliance as large needles should be used. However, small particles generally suffered from initial burst release due to its large surface area. In this study, the volume mean diameter of exenatide microspheres were 75.22 μm with a span value of 1.13, and the D_v,90%, D_v,50% and D_v,10% were 118.72 μm, 70.13 μm and 39.80 μm, respectively. In addition, the microspheres prepared had a unimodal distribution (Fig. 2A). Based on these results, a 22-gauge needle, commonly used in clinical transfusion, was large enough for subcutaneous administration of exenatide-loaded microspheres.

Fig. 2 Characterization of microspheres. (A) Size distributions of exenatide microspheres prepared by PLGA5050 4A. (B) Scanning electron micrographs of the surface morphology.

Higher encapsulation efficiency and lower initial release are another two critical parameters in the development of sustained release microspheres containing water-soluble drugs.⁴⁶ The effect of sonication power on drug loading, drug loading content, initial release, and particle size is presented in Table 1. In general, the size of microspheres declined with the increase of sonication power during preparation of primary emulsion. It was noted that, at the power of 30%, maximum drug loading content (83.8 ± 1.3%) and minimum initial drug release (1.31 ± 0.13%) were obtained. Further increasing the power to 50% led to smaller microspheres in size (57.47 μm); besides, the initial drug release was increased to 2.71 ± 0.85%.

Table 1 DLC, EE, initial burst release and volume mean diameter of exenatide microspheres prepared by different ultra phonic power

Ultrasonic power (650 W)	Drug loading content (DLC) (%)	Entrapment efficiency (EE) (%)	Burst initial release (%)	Volume mean diameter (μm)
10%	3.73 ± 0.08	74.6 ± 1.6	1.70 ± 0.30	122.94
20%	3.85 ± 0.10	77.0 ± 2.0	2.95 ± 1.20	103.63
30%	4.19 ± 0.07	83.8 ± 1.3	1.31 ± 0.13	75.22
40%	4.07 ± 0.08	81.5 ± 1.5	2.20 ± 0.51	61.57
50%	3.85 ± 0.11	77.0 ± 2.2	2.71 ± 0.85	57.47

---

It is well known that polymers with higher molecular weight produce colloids with higher viscosity. To investigate the influence of molecular weights of PLGA polymer on properties of prepared microspheres, different polymers, i.e., PLGA 5050 2A (low viscosity), 4A (medium viscosity) and 6A (high viscosity), were used in this study. As expected, different drug loading contents and different initial releases were observed as shown in Table 2. Maximal entrapment efficiency along with a minimal initial release were obtained when using PLGA polymer with a moderate viscosity (41.82 ± 1.63 dL g⁻¹). On the contrary, PLGA 2A microspheres resulted in a substantial initial release (6.33 ± 1.11%), presumably due to its small size.

Table 2 DLC, EE and burst initial release of exenatide microspheres prepared by PLGA with different inherent viscosity

Polymers	Inherent viscosity (dL g^−1)	Drug loading content (DLC) (%)	Entrapment efficiency (EE) (%)	Burst initial release (%)	Volume mean diameter (μm)
PLGA 5050 2A	0.22 ± 0.003	4.11 ± 0.24	82.1 ± 4.8	6.33 ± 1.11	54.64
PLGA 5050 4A	0.42 ± 0.016	4.19 ± 0.07	83.8 ± 1.3	1.31 ± 0.13	75.22
PLGA 5050 6A	0.64 ± 0.012	3.88 ± 0.31	77.6 ± 6.1	1.54 ± 0.22	114.59

---

Under the same condition of sonication power, there is no difference in initial release when compared PLGA 4A microspheres and PLGA 6A microspheres. However, with the increase of the microspheres size, it becomes difficult, if not possible, to pass through the syringe to inject the drug into the body. Besides, because of the increase in size and lipophilicity as well as hydration effect,⁴⁷ the increase of viscosity will reduce the degradation rate of the microspheres in its release. Experiment showed that about 50% of microspheres residue remains (data not shown) after 2 month release. Hence, PLGA5050 4A with medium viscosity is the best choice.

The SEM images showed that exenatide microspheres were spheres with a rough surface and small pores spreading across the surface (Fig. 2B).

Bio-stability of exenatide within microspheres

Exenatide extracted from microspheres was analyzed by CD spectroscopy to monitor the changes in secondary structure. As shown in Fig. 3 there was no difference between the extracted exenatide and the unprocessed one. This result is not surprising, as unlike a protein a peptide should exist in a relatively dynamic conformation. In detail, minima near 208 and 222 nm indicated presence of α-helix.^32,33 It was implied that the preparation process did not affect the chemical integrity of exenatide. The bio-stability of exenatide was highly preserved during the preparation process.

Fig. 3 CD spectra of exenatide extracted from PLGA microspheres compared with unprocessed one (control).

In vitro release profiles

As shown in Fig. 4A, peptide release was investigated at different pH, i.e., 4.5, 7.4 and 9.1. Under all conditions, no burst peptide release occurred in the first two weeks, and microspheres presented a similar lag time for up to 21 days with no more than 20% of peptide released. Thereafter, acidic condition (pH 4.5) led to a burst drug release starting from the 21^st day, while sustained peptide release could be observed at the other two pH, especially pH 7.4, revealing possible risk control of hypoglycemia under physiological conditions. In addition, more than 80% of peptides were released in the first 42 days, and the percentage achieved nearly 100% at the last assigned time point.

Fig. 4 (A) In vitro release profiles of exenatide microspheres under different pH, i.e., 4.5, 7.4 and 9.1 (n = 3). (B) Polymer molecular weight change and cumulative release change under pH 9.1. (C) SEM pictures of exenatide-loaded PLGA microspheres at different times.

Degradation pattern of in vitro release

As described in Fig. 4B, microspheres displayed distinctive appearances through the degradation process. The morphological change of PLGA microsphere was also monitored via SEM. Small pores came out on the surface of microspheres in the very beginning, and during the first week, the pores grew up and the microspheres maintained the spherical appearance. After one week, the pores on the surface became larger and the numbers of inner pores were increased, thereby inducing particle disruption and loss of spherical shape. After three weeks, the microspheres completely lost their original spherical shape and turned into a random shape.

The degradation pattern of exenatide release from microspheres was appraised along with polymer molecular weight and particle morphology (Fig. 4C). The in vitro release dynamics demonstrated that drug release is primarily controlled by erosion, following a typical three-phase release mechanism.⁴⁸

Phase 1: initial release stage (0–24 h). When the microspheres were exposed to aqueous medium, the surface of the microspheres became hydrated, and exenatide on and under the surface with weak interactions was dissolved and released into the medium. Thus, the primary release stage is appraised as very small burst release into the medium. Exenatide release rate at this stage can be optimized by adjusting preparation parameters and techniques.

Phase 2: hydrate stage (24 h to 3 weeks). During phase 1, a small amount of exenatide released into the medium. At this stage, the microspheres continued to be hydrated, and, due to hydrolysis, the molecular weight of polymers started to decrease. Under the scan electronic microscope (SEM), holes growing inside microspheres as well as erosion of polymer base can be obviously observed. In this phase, the rate of drug release was relatively slow at neutral and basic pH, most of the drugs still stay in the microspheres.

Phase 3: collapse stage (after 3 weeks). Starting from the 21^st day, the polymer molecular weight reached the lower threshold value level (about 10000 Da). Multi-hole erosion caused the collapse of microspheres, followed by complete release of exenatide.

In vitro cytotoxicity study

To mimic the physiological condition of microspheres' application, mouse embryonic fibroblast (Mef) cells were used as fibroblast is the main component of the loose connective tissue which is widely distributed in dermis layer of animal skin. The cytotoxicity of blank and exenatide-loaded microspheres was determined by a typical MTT manner.^49,50 As shown in Fig. 5, cell viability was at round 100% across the concentration ranging from 5 to 200 μg mL⁻¹ for both groups, indicating negligible cytotoxicity and desirable safety of blank and exenatide-loaded microspheres.

Fig. 5 Cell viability of blank microspheres and exenatide-loaded microspheres against Mef cells assayed by MTT.

Histological study

Long acting in situ preparations are prone to cause greater or lesser damage to peripheral tissues. At appointed time points after drug administration to KM mice, the injection sites were collected periodically to perform the histological study. As shown in Fig. 6, compared with exenatide solution, a mild inflammatory response was observed in both tissue-injected exenatide microspheres and solvent groups, evidenced by a focal minimal foreign body reaction with a small number of neutrophils in subcutaneous tissue. The inflammation disappeared 7 days after injection. The short-term inflammation was mainly caused by the overdose of injective solvent. After the injective solvent was absorbed by tissues, local physiological milieu was recovered. The high molecular weight of the polymer resulted in slow degradation of PLGA, which generate lower levels of acidic oligomers. Although an acidic environment in the interior of microspheres may develop due to degradation products degraded from PLGA polymers, it is expected that a small dose would cause negligible stimulation to tissues. In conclusion, the exenatide microspheres were biocompatible and safe for subcutaneous injection.

Fig. 6 Histology of injection sites after subcutaneous administration with exenatide solution, injective solvent and exenatide microspheres in different times (A) exenatide solution (2d)-control, (B) injective solvent (2d)-control, (C) exenatide microspheres (1 day), (D) exenatide microspheres (2 days), (E) exenatide microspheres (4 days), (F) exenatide microspheres (1 week), (G) exenatide microspheres (2 weeks), (H) exenatide microspheres (4 weeks).

In vivo pharmacokinetics

The profile of the plasma concentration of exenatide versus time obtained from the in vivo study clearly showed that the microspheres were able to maintain a constant drug level over a 30 day duration. A comparison of the drug plasma concentration between exenatide solution and exenatide microspheres was shown in Fig. 7A. As for the peptide solution group, rapid absorption and elimination of exenatide led to a short t_max of 0.75 h but an extreme high C_max of 102.2 ng mL⁻¹, probably inducing hypoglycemia, while the exenatide microspheres group showed a sustained release for up to 30 days. In consistence with data from in vitro drug release study, the highest plasma concentration of exenatide released from microspheres appeared at the end of the second hour due to an initial burst release in the very beginning. However, the C_max, no more than 7.53 ng mL⁻¹, would not induce any risk when compared with that of the peptide solution group. Thereafter, hydration and erosion induced peptide release started and reached a plateau in the following 20 days after the first week.

Fig. 7 In vivo pharmacokinetic profiles of exenatide-loaded microspheres. (A) The profile of the plasma exenatide concentration vs. time after subcutaneous drug delivery in SD rats (n = 5). (B) Relative change in glucose of exenatide solution, exenatide microsphere and saline groups (n = 5). Exenatide solution-treated groups received twice-daily subcutaneous injections of exenatide in PBS at an exenatide dose of 108 μg kg⁻¹ for 7 days. The exenatide microspheres treated groups received a single subcutaneous injection of an aqueous suspension of exenatide microspheres at an exenatide dose of 1.52 mg kg⁻¹.

As shown in Fig. 7A, the stable exenatide concentration in plasma indicated that the drug can maintain a constant therapeutic effect during a period of 4 weeks in vivo. Because of the hydration stage as described in section of release degradation pattern, the release rate was relatively slow. The drug concentration was low in the first 2 days after initial release. Compared with in vitro study, PLGA microspheres degraded faster in vivo due to non-specific enzymatic catalysis.⁵¹ By comparing the AUC obtained from two groups, a relative bioavailability of 70.31% was obtained, suggesting that most of the incorporated exenatide in microspheres was released within 32 days following injection. In spite of high bioavailability, a small amount of peptide was degraded and inactivated due to the acid microenvironment in PLGA microspheres during such a long period.^52,53

Pharmacodynamic study in diabetic mice

In order to assess the potential of the exenatide microspheres for clinical use, STZ induced diabetic mice were used as a type 2 diabetes model to investigate the pharmacodynamics. After injection, both groups receiving exenatide solution and exenatide microspheres exhibited strong therapeutic effects when compared with control (saline group). As shown in Fig. 7B, during the first 7 days, the exenatide solution group showed an obvious effect on blood glucose control, whereas discontinuing peptide solution treatment on the 8^th day made the blood glucose escalate to normal level.

For the microsphere group, blood glucose went down quickly after administration due to initial burst release. After going back to normal level at the end of the first day, blood glucose was well controlled for up to 20 days, which is greatly consistent with the results in pharmacokinetic study (Fig. 7A). After the lowest blood glucose level appeared on the 10^th day, the antidiabetic effect became weaker and weaker along with the disintegration of microspheres till the end of the 3^rd week. In contrast, the saline group still presented significantly high blood glucose throughout the entire experimental period.

For exenatide microspheres, both pharmacokinetic and pharmacodynamic studies presented an extremely similar profile. In spite of initial burst release, it is mild enough to induce side effects, such as hypoglycemia and shortened acting time. Oppositely, a moderate antidiabetic effect can be observed for up to 20 days since the 8^th day. Despite high blood glucose occasionally in the first weeks, this issue can be well resolved by giving a combination of exenatide solution and microspheres before stable peptide release and having a rational schedule of microspheres administration to cover the slow peptide release period. In all, exenatide microspheres may be considered a promising long-acting formulation for antidiabetic therapy with great compatibility and compliance compare to any commercial product.

Conclusions

In this study, a W/O/O method for preparation of exenatide microspheres with long acting as well as high safety was established. The initial burst release can be reduced to 1.31 ± 0.13% after preparation optimization. In vitro study showed an up to 7 week peptide release, and no serious inflammatory reactions were found in histological study. For in vivo study of exenatide microspheres, a similar profile can be plotted from the data of both pharmacokinetic and pharmacodynamic studies. What's more, compared to its solution product, exenatide microspheres owned great advantages in risk control of hypoglycemia as well as frequency of administration, thereby exhibiting desirable safety and compliance for potential clinical use.

Conflicts of interest

The authors report no conflicts of interest in this work.

Acknowledgements

This work was funded by the National Natural Science Foundation of China (no. 81501579), the Natural Science Foundation of Jiangsu Province (no. BK20150702), and the Priority Academic Program Development of Jiangsu Higher Education Institutions.

References

1. C. Zhu, Y. Huang, X. Zhang, L. Mei, X. Pan, G. Li and C. Wu, Colloids Surf., B, 2015, 132, 103–110.
2. Y. Wen, H. R. Kolonich, K. M. Kruszewski, N. Giannoukakis, E. S. Gawalt and W. S. Meng, Mol. Pharmaceutics, 2013, 10, 1035–1044.
3. Y. Ding, Y. Wang, J. Zhou, X. Gu, W. Wang, C. Liu, X. Bao, C. Wang, Y. Li and Q. Zhang, Biomaterials, 2014, 35, 7214–7227.
4. V. R. Sinha and A. Trehan, J. Controlled Release, 2003, 90, 261–280.
5. L. Chen, D. J. Magliano and P. Z. Zimmet, Nat. Rev. Endocrinol., 2012, 8, 228–236.
6. R. N. Kulkarni, J. C. Brüning, J. N. Winnay, C. Postic, M. A. Magnuson and C. R. Kahn, Cell, 1999, 96, 329–339.
7. R. A. DeFronzo, R. E. Ratner, J. Han, D. D. Kim, M. S. Fineman and A. D. Baron, Diabetes Care, 2005, 28, 1092–1100.
8. M. A. Nauck, Am. J. Med., 2009, 122, S3–S10.
9. R. Göke, H.-C. Fehmann, T. Linn, H. Schmidt, M. Krause, J. Eng and B. Göke, J. Biol. Chem., 1993, 268, 19650–19655.
10. L. L. Nielsen, A. A. Young and D. G. Parkes, Regul. Pept., 2004, 117, 77–88.
11. B. R. Gedulin, S. E. Nikoulina, P. A. Smith, G. Gedulin, L. L. Nielsen, A. D. Baron, D. G. Parkes and A. A. Young, Endocrinology, 2005, 146, 2069–2076.
12. E. Jimenez-Solem, M. H. Rasmussen, M. Christensen and F. K. Knop, Curr. Opin. Mol. Ther., 2010, 12, 790–797.
13. G. M. Keating, Drugs, 2005, 65, 1681–1692 ; discussion 1693–1685.
14. J. J. Cunningham, M. J. Kirchmeier and S. Mittal, Pharmaceutical Dosage Forms-Parenteral Medications: Volume 1: Formulation and Packaging, 2010, vol. 1, p. 158.
15. D. L. Graham, K. Erreger, A. Galli and G. D. Stanwood, Mol. Psychiatry, 2013, 18, 961.
16. Y. Cai, L. Wei, L. Ma, X. Huang, A. Tao, Z. Liu and W. Yuan, Drug Des., Dev. Ther., 2013, 7, 963–970.
17. R. A. Jain, Biomaterials, 2000, 21, 2475–2490.
18. J. Emami, H. Hamishehkar, A. R. Najafabadi, K. Gilani, M. Minaiyan, H. Mahdavi and A. Nokhodchi, J. Pharm. Sci., 2009, 98, 1712–1731.
19. Z. Ghalanbor, M. Körber and R. Bodmeier, Eur. J. Pharm. Biopharm., 2013, 85, 624–630.
20. C. Engman, Y. Wen, W. S. Meng, R. Bottino, M. Trucco and N. Giannoukakis, Clin. Immunol., 2015, 160, 103–123.
21. J. H. Ahn, E. J. Park, H. S. Lee, K. C. Lee and D. H. Na, AAPS PharmSciTech, 2011, 12, 1220–1226.
22. F. Qi, J. Wu, T. Yang, G. Ma and Z. Su, Acta Biomater., 2014, 10, 4247–4256.
23. J. Wang, B. M. Wang and S. P. Schwendeman, J. Controlled Release, 2002, 82, 289–307.
24. H. Hezaveh and I. I. Muhamad, Chem. Eng. Res. Des., 2013, 91, 508–519.
25. M. Rahman, A. Mohamed, K. Haliza and R. H. Ansary, in Malaysia International Biological Symposium 2014 (i-SIMBIOMAS 2014), 2014, Putrajaya, p. 270–272.
26. F. Qi, J. Wu, Q. Fan, F. He, G. Tian, T. Yang, G. Ma and Z. Su, Colloids Surf., B, 2013, 112, 492–498.
27. H. J. Yang, I. S. Park and K. Na, Colloids Surf., A, 2009, 340, 115–120.
28. Z. Yang, B. Yu, J. Zhu, X. Huang, J. Xie, S. Xu, X. Yang, X. Wang, B. C. Yung, L. J. Lee, R. J. Lee and L. Teng, Nanoscale, 2014, 6, 9742–9751.
29. Y.-y. Fang, L.-j. Wang, D. Li, B.-z. Li, B. Bhandari, X. D. Chen and Z.-h. Mao, Carbohydr. Polym., 2008, 74, 379–384.
30. C. Sun, Y. Shen, D. Sun, T. Hang and J. Tu, J. Chromatogr. Sci., 2012, 50, 343–348.
31. N. Samadi, A. Abbadessa, A. Di Stefano, C. F. van Nostrum, T. Vermonden, S. Rahimian, E. A. Teunissen, M. J. van Steenbergen, M. Amidi and W. E. Hennink, J. Controlled Release, 2013, 172, 436–443.
32. Y. Wen, S. L. Roudebush, G. A. Buckholtz, T. R. Goehring, N. Giannoukakis, E. S. Gawalt and W. S. Meng, Biomaterials, 2014, 35, 5196–5205.
33. Y. Zheng, Y. Wen, A. M. George, A. M. Steinbach, B. E. Phillips, N. Giannoukakis, E. S. Gawalt and W. S. Meng, Biomaterials, 2011, 32, 249–257.
34. C. Sun, W.-C. Shen, J. Tu and J. L. Zaro, Mol. Pharmaceutics, 2014, 11, 1583–1590.
35. Y.-P. Li, Y.-Y. Pei, X.-Y. Zhang, Z.-H. Gu, Z.-H. Zhou, W.-F. Yuan, J.-J. Zhou, J.-H. Zhu and X.-J. Gao, J. Controlled Release, 2001, 71, 203–211.
36. J. Song, J. Xie, C. Li, J.-h. Lu, Q.-f. Meng, Z. Yang, R. J. Lee, D. Wang and L.-s. Teng, Int. J. Pharm., 2014, 472, 296–303.
37. L. K. Rogers, M. Robbins, D. Dakhlallah, Z. Yang, L. J. Lee, M. Mikhail, G. Nuovo, G. S. Pryhuber, G. McGwin, C. B. Marsh and T. E. Tipple, Ann. Am. Thorac. Soc., 2015, 12, 1506–1513.
38. X. Ling, Z. Huang, J. Wang, J. Xie, M. Feng, Y. Chen, F. Abbas, J. Tu, J. Wu and C. Sun, J. Mater. Chem. B, 2016, 4, 1787–1796.
39. C. Li, S. Li, T. Tu, X. Qi, Y. Xiong, S. Du, Y. Shen, J. Tu and C. Sun, Polym. Chem., 2015, 6, 2740–2751.
40. X. Ling, C. Zhao, L. Huang, Q. Wang, J. Tu, Y. Shen and C. Sun, RSC Adv., 2015, 5, 81668–81681.
41. Y. Ding, W. Wang, M. Feng, Y. Wang, J. Zhou, X. Ding, X. Zhou, C. Liu, R. Wang and Q. Zhang, Biomaterials, 2012, 33, 8893–8905.
42. Y. Zhang, M. Huo, J. Zhou and S. Xie, Comput. Methods Programs Biomed., 2010, 99, 306–314.
43. J. Xuan, Y. Lin, J. Huang, F. Yuan, X. Li, Y. Lu, H. Zhang, J. Liu, Z. Sun and H. Zou, Peptides, 2013, 46, 172–179.
44. S.-y. Wu, G.-f. Wang, Z.-q. Liu, J.-j. Rao, L. Lü, W. Xu, S.-g. Wu and J.-j. Zhang, Acta Pharmacol. Sin., 2009, 30, 202–208.
45. F. Qi, J. Wu, D. Hao, T. Yang, Y. Ren, G. Ma and Z. Su, Pharm. Res., 2014, 31, 1566–1574.
46. F. Cui, D. Cun, A. Tao, M. Yang, K. Shi, M. Zhao and Y. Guan, J. Controlled Release, 2005, 107, 310–319.
47. G. Mittal, D. Sahana, V. Bhardwaj and M. R. Kumar, J. Controlled Release, 2007, 119, 77–85.
48. B. Liu, Q. Dong, M. Wang, L. Shi, Y. Wu, X. Yu, Y. Shi, Y. Shan, C. Jiang, X. Zhang, T. Gu, Y. Chen and W. Kong, Chem. Pharm. Bull., 2010, 58, 1474–1479.
49. Y. Ding, Y. Wang, Y. Opoku-Damoah, C. Wang, L. Shen, L. Yin and J. Zhou, Biomaterials, 2015, 72, 90–103.
50. Y. Shen, J. Wang, Y. Li, Y. Tian, H. Sun, O. Ammar, J. Tu, B. Wang and C. Sun, RSC Adv., 2015, 5, 46464–46479.
51. Y. Geng, W. Yuan, F. Wu, J. Chen, M. He and T. Jin, J. Controlled Release, 2008, 130, 259–265.
52. A. G. Ding and S. P. Schwendeman, Pharm. Res., 2008, 25, 2041–2052.
53. K. Fu, D. W. Pack, A. M. Klibanov and R. Langer, Pharm. Res., 2000, 17, 100–106.

---