Long-Acting Release Formulation of Exendin-4 Based on Biomimetic

Apr 24, 2017 - Exendin-4 has been clinically exploited for treating type 2 diabetes, but the short circulation half-life and multiple daily injections...
3 downloads 8 Views 8MB Size
Long-Acting Release Formulation of Exendin‑4 Based on Biomimetic Mineralization for Type 2 Diabetes Therapy Wei Chen,†,‡ Guohao Wang,§ Bryant C. Yung,‡ Gang Liu,§ Zhiyong Qian,*,† and Xiaoyuan Chen*,‡ †

State Key Laboratory of Biotherapy and Cancer Center, West China Hospital, Sichuan University and Collaborative Innovation Center for Biotherapy, Chengdu 610041, People’s Republic of China § State Key Laboratory of Molecular Vaccinology and Molecular Diagnostics & Center for Molecular Imaging and Translational Medicine, School of Public Health, Xiamen University, Xiamen 361005, People’s Republic of China ‡ Laboratory of Molecular Imaging and Nanomedicine (LOMIN), National Institute of Biomedical Imaging and Bioengineering (NIBIB), National Institutes of Health (NIH), Bethesda, Maryland 20892, United States S Supporting Information *

ABSTRACT: Exendin-4 has been clinically exploited for treating type 2 diabetes, but the short circulation half-life and multiple daily injections limit its widespread application with respect to poor patient compliance, low efficacy, and high treatment cost. In this study, a potent long-acting release system based on biomimetic mineralization was constructed for biocompatible and sustained exendin-4 delivery. Similar to natural biomineralization, exendin-4 can be mineralized to form nanosized mineral solids by means of the reaction between acidic amino acid residues and calcium ions in a supersaturated environment with negligible influence on peptide bioactivity. Mineralized exendin-4 particles may be spontaneously absorbed by a living body under physiologically supersaturated conditions, resulting in gradual dissociation and sustained drug release. In such a way, the glucose level of diabetic mice may be effectively controlled for a long period of time by mineralized exendin-4 without obvious side effects. We believe this biomimetic formulation can serve as a promising candidate for future clinical applications for type 2 diabetes therapies. KEYWORDS: biomineralization, exendin-4, long-term drug release, supersaturation-based regulation, type 2 diabetes

D

(DPP IV) to generate some N-terminally truncated metabolites, followed by quick clearance.8 Exendin-4, a peptide of 39 amino acid residues, was originally isolated from the salivary secretions of the Gila monster lizard, Heloderma suspectrum, and shares around 53% sequence homology with mammalian GLP-1.9 As a result of changes in the amino acid sequence, exendin-4 resists degradation mediated by DPP-4, but shows comparable affinity to the GLP-1 receptor (compared to native incretin, GLP-1),10 resulting in promising insulinotropic potencies.11,12 However, the blood circulation time of exendin-4 remains unsatisfactory (60−90 min), requiring twice-daily subcutaneous injections for sustained glucose control.13 This drawback decreases patient compliance, leading to limited therapeutic efficacy and high treatment cost. Therefore, various

iabetes mellitus is one of the most significant health concerns worldwide. The global prevalence of this chronic disease is predicted to reach 439 million by 2030.1 A typical clinical pathological feature of diabetes is a disorder of glucose regulation, which manifests in symptoms such as nephropathy, high blood pressure, stroke, and so on, resulting in serious morbidity.2 Importantly, 90% of diabetic individuals can be classified as having type 2 diabetes (T2D),3 which exhibits two major characteristics: reduced insulin sensitivity linked to obesity and impaired insulin secretion due to β-cell dysfunction.4 One effective strategy to treat T2D is to utilize glucagon-like peptide-1 (GLP-1), which is a peptide of 30 amino acid residues with regulatory functions in glucose homeostasis.5,6 After production and secretion by the intestinal L cells in response to nutrient ingestion, GLP-1 binds and activates the pancreatic GLP-1 receptor and thereby potentiates glucose-induced insulin secretion of β-cells.7 However, native GLP-1 is rapidly degraded in vivo by dipetidylpeptidase IV © 2017 American Chemical Society

Received: March 15, 2017 Accepted: April 24, 2017 Published: April 24, 2017 5062

DOI: 10.1021/acsnano.7b01809 ACS Nano 2017, 11, 5062−5069

Article

www.acsnano.org

Article

ACS Nano

Figure 1. Schematic of m-exendin-4 generation and its disassembly in response to physiological supersaturation. Under high supersaturation (Ca-rich environment), the acidic residues in exendin-4 can easily bind to free Ca ions to provide some “nucleation sites”. Then spontaneous nucleation occurs and small crystals are generated around the peptide, which tend to form stable sphere-like clusters. Under low supersaturation, calcium phosphate spontaneously dissolves, leading to cluster dissociation to release the payloads into the bloodstream.

Figure 2. Characterization of m-exendin-4 particles. (A) Photograph of Tyndall scattering before and after Ca introduction in the reaction system. (B) UV−vis spectrum of reaction system with increasing amounts of Ca. (C) TEM images of m-exendin-4 particles at different time points. Lower right: Negative control, mineral particles formed without exendin-4. (D) Mineralization reaction kinetics based on size measurement. (E) pH changes before and after mineralization reaction.

systems mainly concentrate on direct chemical modification of drug molecules14−17 or utilization of intricate synthetic polymer structures.18−21 However, chemical modification can easily result in loss of bioactivity. For example, PEGylation may significantly decrease the cAMP activity of GLP-1 or exendin-4.28 Synthetic polymers that vary in shape, size, surface chemistry, composition, roughness, and porosity may all induce an immunological response,29,30 resulting in nonspecific adsorption of blood and tissue fluid proteins at the site of administration

devices and long-acting release (LAR) formulations have been proposed to improve exendin-4 delivery and therapeutic performance.14−22 Nano- and microsized implanted formulations for long-term antidiabetic drug delivery have garnered increasing attention,23−26 as they can provide sustained drug release to control blood glucose levels over a long time, avoiding pain associated with repeated injections and some dose-relevant adverse reactions (e.g., dose-limiting nausea and vomiting).27 Currently, LAR 5063

DOI: 10.1021/acsnano.7b01809 ACS Nano 2017, 11, 5062−5069

Article

ACS Nano or implantation.29,30 Following immune and inflammatory cell response, the implantations become walled off by a vascular collagenous fibrous capsule, which restricts the communication and interaction of implantations with their surroundings. Notably, direct use of natural biomaterials is an alternative route to maintain drug activity and bypass foreign body response to achieve controllable and biocompatible LAR systems,22,31 as many of them confer high affinity to biomolecules and can be self-absorbed by the body. Interestingly, in the natural biomineralization process, macromolecule− biomineral complexes are always involved in the assembly and disassembly process, accompanied by the “loading” and “release” of bioactive molecules.32,33 Inspired by nature, we design a biomimetic mineralization reaction to construct macromolecule−biomineral complexes (mineralized exendin-4, m-exendin-4), which respond to physiological supersaturation to allow for effective and long-term exendin-4 delivery. Like many biomineralization-related proteins, exendin-4 (isoelectric point of 4.96) contains a high proportion of acidic amino acid residues, including five glutamic acid residues (pKa = 4.25) and one aspartic acid residue (pKa = 3.86) per molecule.20 These acidic residues have been proven to chelate calcium ions from supersaturated metastable solutions, providing “nucleation sites”.34 After concentrating cationic calcium ions around the “nucleation site” to increase local supersaturation, spontaneous calcium phosphate nucleation occurs to generate tiny crystals around the peptides upon anionic phosphate participation. These crystals tend to assemble orderly to form structured spherical aggregates (Figure 1, upper left) via a “bricks and mortar” self-assembly mechanism.35 In detail, the peptide easily folds into a highly compact structure owing to intramolecular hydrogen bonds or calcium bridges. Multivalent interactions of the peptide with mineral crystals induce the unfolding of the compact structure, allowing the peptide to interact with other mineral crystal units to propagate the assembly process.35 The generated inorganic−organic hybrid structure is relatively stable under high supersaturation but has a tendency to dissociate in low supersaturation due to calcium phosphate dissolution. Notably, physiological supersaturation is relative low (0.5 to 2.5 mM),36 which can provide enough driving force to induce gradual particle destruction to achieve long-term drug release into the bloodstream (Figure 1, lower panel). Different from others, the biomineralization-inspired strategy is facile and cost-effective, which negligibly influences peptide activity. The final formula contains only biomineralized materials, which can be completely absorbed by the body, avoiding severe foreign body response. In such a way, a safe and potent platform can be well developed for long-acting exendin-4 administration in T2D therapy.

be observed under transmission electron microscopy (TEM) at different time points. Initially, very loose and small nanoclusters (around 40−50 nm) were observed, which consisted of tiny crystals (several nanometers) (Figure 2C), indicating a rapid nucleation and assembly process (Figure 2D). Over time, the clusters became larger and denser, suggesting further crystal growth and assembly (Figure 2C).38 After around 2 h of reaction, the size of the particles stabilized and could be maintained for at least 1 week (Figure S1), demonstrating that the reaction reached an equilibrium state. It should be mentioned that when exendin-4 was absent in the medium, only irregular crystals formed (Figure 2C, lower right), indicating that the organized structures came from peptide-mediated mineral generation. Moreover, the kinetic analysis revealed that the mineralization reaction was related to substrate exchange between the system and the environment. Interestingly, the open system could induce a more complete reaction and generate larger particles compared with the semisealed and completely sealed systems (Figure 2D). This phenomenon may be attributed to CO2 volatilization from the reaction system to the environment, which gradually increases solution pH to shift the ionization equilibrium of the peptide to produce more negative species to facilitate calcium sequestration (the generated m-exendin-4 particles are relatively negatively charged, Figure S2). This hypothesis was confirmed by the observation that an open system resulted in a greater pH increase from 7.4 to 8.2 (greater CO2 volatilization) during the mineralization reaction compared to other reaction conditions (semisealed and sealed samples) (Figure 2E). In such a way, by regulating the system−environment substrate exchange, a simple and controllable “peptide encapsulation” method was developed, which mimics the natural biomineralization process, providing a highly biocompatible and biofriendly peptide solidification strategy. m-Exendin-4 Compositional Analysis and Exendin-4 Release. To further observe the composition of the m-exendin-4 nanoformulation, elemental mapping of the particles was carried out using energy-dispersive X-ray spectroscopy (EDS) analysis. Results clearly showed that the particles mainly contained elements from calcium phosphate (CaP) mineral (Ca, P, O) and elements from the peptide (N, C, O). Obviously, Ca and P elements distributed within the area containing N and C elements (Figure 3A), further confirming that calcium phosphate mineral particles nucleated and grew within the organic phase (exendin-4). This method of in situ mineralization assembly was quite effective to implement on every peptide molecule simultaneously to produce uniform nanoclusters (Figure 3B), avoiding a tedious purification process and minimizing drug loss compared to conventional protein encapsulation methods.39 Fourier transform infrared (FTIR) spectroscopy of the composites showed the characteristic peaks of inorganic phosphates (1000−1100 cm−1, P−O vibration) and exendin-4 (1500−1600 cm−1, CO vibration) (Figure 3C), verifying the formation of inorganic−organic complexes. The organic components contributed around 17 wt % of the nanoparticles, while the content of CaP was 77 wt % and the remaining 6 wt % was water (Figure S3). Intriguingly, when the particles were exposed to a physiologically supersaturated environment (0.5−2.5 mM with certain fluctuations),36 the spontaneous release of encapsulated drugs occurred (Figure 3D). This could be attributed to the shifts in solubility equilibrium, which accelerate the dissolution of calcium phosphate particles, leading to dissociation of m-exendin-4 from organized particles into clusters of debris (Figure 3E,F).

RESULTS AND DISCUSSION Synthesis and Characterization of m-Exendin-4. The mineralization of exendin-4 was a facile process, conducted in the presence of a biomimetic medium, Dulbecco’s modified Eagle’s medium (DMEM), at 37 °C with 5% CO2.37 After Ca introduction, the presence of nanoparticles in the medium was verified by obvious Tyndall scattering (Figure 2A). Ultraviolet− visible (UV−vis) absorption spectra revealed that increasing Ca concentration up to 10 mM significantly elevated the optical density of the reaction system without significant peak shift, suggesting minor derangement to solution equilibrium without the formation of large aggregates (Figure 2B). The detailed assembly process of the aforementioned nanoformulation could 5064

DOI: 10.1021/acsnano.7b01809 ACS Nano 2017, 11, 5062−5069

Article

ACS Nano

Figure 3. m-Exendin-4 particle compositional analysis and exendin-4 release. (A) Elemental mapping analysis of m-exendin-4 particles. (B) Scanning electron microscopy (SEM) images of m-exendin-4 particles. Right: Magnification of indicated area in the left image. (C) FTIR analysis of exendin-4, m-exendin-4, and blank calcium phosphate particles. (D) Release profiles of m-exendin-4 under different Ca concentrations. (E) TEM images of m-exendin-4 before and after dissociation. (F) DLS analysis of m-exendin-4 particles before and after dissociation. (G) HPLC analysis of native and released exendin-4 peptide. (H) CD spectra of native and released exendin-4 peptide.

Moreover, high-performance liquid chromatography (HPLC) (Figure 3G) and electrophoresis analysis (Figure S4) revealed that no degradation products were observed during exendin-4 release, confirming the sequence integrity of the released peptide. Additionally, circular dichroism (CD) spectra (Figure 3H) demonstrated that the released exendin-4 did not exhibit changes in structural characteristics compared to the native drug, further suggesting mineralization modification and the release process to be highly biocompatible. Collectivity, all the data demonstrated that by simply adjusting environmental supersaturation, a self-assembly and dissociation process could be well regulated, which is similar to the process of biological hard tissue construction and destruction during the natural biomineralization process. Like the natural processes, biomimetic assembly and disassembly minimize the activity loss of the payload. Moreover, the decomposition products during exendin-4 release are mainly inorganic ions (e.g., calcium and phosphate ions), which act as nutrients to be absorbed by surrounding biological tissues, leading to negligible side effects. In this way, we believe biomimetic mineralization can act as an ideal strategy to “load” and “release” exendin-4 over a long time for biofriendly therapy for T2D. Long-Term in Vivo Release of Exendin-4 from Mineralized Particles. We next assessed drug release kinetics in vivo by single-dose subcutaneous injection of free exendin-4 or m-exendin-4. To label exendin-4 peptide, we used Cys40exendin-4 for ease of reaction. A near-infrared dye, IRdye 800CW, was applied to label Cys40-exendin-4 through reaction of N-hydroxysuccinimide (NHS) ester and thiol groups with the help of a linker (Figure 4A). The conjugate was characterized by HPLC (Figure S5), mass spectrometry (MS) (Figure S6), and UV−vis absorption (Figure S7). As expected, free exendin-4 quickly disappeared from the injection site within 24 h (Figure 4B), owing to quick diffusion into systemic

circulation. In contrast, m-exendin-4, as a solidified formulation, retained the drug for a much longer period of time. Although the burst release was observed within the first 0.5 h, which may be due to the initial great supersaturation difference between the injected solution and surrounding subcutaneous microenvironment, the slow drug release was achieved during gradual particle dissociation, as a result of local Ca concentration increase induced by calcium phosphate dissolution, further indicating supersaturation-dependent drug release. Notably, even after 192 h, around 5.1% of the initial fluorescence signal could still be observed at the injection site (Figure 4B,D), indicating that a single injection could provide long-term drug treatment for more than 1 week. Encouraged by the observation, we further detected the blood drug concentration based on fluorescence intensity (Figure 4C). Exendin-4 quickly disappeared from the blood within 24 h, consistent with a previous observation.13 In contrast, m-exendin-4 provided a sustained and gradual release profile to maintain blood exendin-4 concentrations up to 192 h (Figure 4C). This could be attributed to sustained drug release based on slow dissociation and self-absorption of biomineral materials. This in vivo dissociation was directly captured under Bio-TEM observation, which indicated that m-exendin-4 particles first disassembled to form small fragments and then gradually were absorbed by surrounding tissues (Figure 4E). In addition, no obvious damage or inflammatory response was observed in the skin or the muscle around the injection site (Figure 4F) during particle dissociation and absorption. This observation is probably due to the biocompatibility of calcium phosphate (Figure S8), which is the main component of bone and teeth, acting as endogenous materials to escape immune recognition and rejection. In this regard, we believe biomimetic mineral could serve as an ideal LAR system to achieve long-term delivery and safe drug therapy. 5065

DOI: 10.1021/acsnano.7b01809 ACS Nano 2017, 11, 5062−5069

Article

ACS Nano

Figure 4. Long-term in vivo release of exendin-4 from mineralized particles. (A) Schematic of exendin-4 labeling by IRdye 800CW. (B) In vivo fluorescence imaging of mice treated by exendin-4 and m-exendin-4 at different time points. The red arrow indicates the injection site; the black arrow indicates the kidneys (exendin-4 was mainly cleared by the renal route). (C) Ex vivo fluorescence imaging of blood samples from exendin-4- and m-exendin-4-treated mice at different time points. (D) Relative fluorescence intensity (RFI) of injection site after exendin-4 or m-exendin-4 treatment at different time points. (E) Bio-TEM images of m-exendin-4 particles (blue arrows) in subcutaneous tissues before and after dissociation. (F) Hematoxylin and eosin (H&E) staining of skin and muscle sections from the injection site of exendin-4- and m-exendin-4-treated mice.

In Vivo Study of Mineralized Exendin-4 for T2D Treatment. The glucose-stabilizing capabilities of m-exendin-4 were investigated by intraperitoneal glucose tolerance test (IPGTT) in C57BL/6 db/db mice (Figure 5A). As shown in Figure 5B, blood glucose level (BGL) in the control group rapidly increased and reached a maximum value of 17.8 mM at 30 min after intraperitoneal glucose challenge. Significant suppression of the increased blood glucose levels was observed after subcutaneous preadministration of exendin-4 and m-exendin-4, similar to the self-regulation of blood glucose level in healthy mice (Figure 5B,C). This phenomenon indicated that mineralization of exendin-4 did not weaken its rapid regulatory ability for acute BGL increase. Moreover, the longterm glucose-lowering effect on diabetic mice of exendin-4 and m-exendin-4 was evaluated for up to 192 h. Obviously, minimal fluctuations of BGL (approximately from 19.8 ± 0.9 to 24.3 ± 1.2 mM) were found in the CaP-treated group (Figure 5D, black curve), indicating that unloaded mineral has a negligible influence on the glucose metabolism of mice. When free exendin-4 was administered to diabetic mice, the BGL was lowered by

around 76.2% (from about 23.1 mM to 5.5 mM) within the first 6 h (Figure 5D, green line). However, this normoglycemia state could not be maintained and returned to a hyperglycemia state within 24 h due to the quick drug clearance. In contrast, m-exendin-4 showed promising BGL control (Figure 5D, red line). Even at 192 h after treatment, the BGL in m-exendin-4treated mice was still less than the original level (Figure 5D, red line). The normoglycemia duration (