I. Introduction

Microencapsulation was known to be a process of enclosing micron-sized solids or liquids or gases in a shell to protect the substances from the external environment and provide controlled release of the drug substance [1-11]. The microcapsules are composed of two parts, core and shell material. In general, core material contains an active ingredient while shell material shields the core material from the external environment and allows good release characteristics. Stability in physiological conditions and ability to reach local areas at therapeutic level make them very useful carriers.

Hepatocellular carcinoma (HCC) was reported to be diagnosed in >0.5 million people worldwide, including approximately 20,000 new case in the US annually [12]. Catheter-directed therapy in the form of transarterial chemoembolization has become the standard of care for well-compensated patients with HCC not amenable to resection or ablative techniques. Although the 100-300 µm and 300- 500 µm sized particles (drug-eluting beads, DEB) were widely used when performing embolization, it was reported that the 100-300 µm doxorubicin DEBs were favored over 300-500 µm doxorubicin DEBs because of lower rates of toxicity after treatment and a trend toward more complete imaging response at initial follow-up [12,13]. Yoon et al. [14] reported that HCC was sequentially treated by transarterial chemoembolization using doxorubicin-loaded beads (100~300 µm, Surrey, UK) and liver transplantation. Bashir et al. [15] also reported that the increased concentration of polymer (sodium alginate) was attributed to retardation of nifedifine (NF) drug release, suggesting that the polymer/drug ratio can be modulated to achieve the desired release rate.

Poly(ε-caprolactone) (PCL, shell material) is widely used for drug delivery and various medical applications due to its biodegradability and excellent biocompatibility [3,6,8]. Polyvinylpyrrolidone (PVP) is also a sustained-release drug delivery system [7]. It is a biocompatible polymer that has stable properties over a prolonged period of time without being affected by pH in aqueous solution. PCL, PVP, and polyethylene glycol (PEG) are widely used as shell materials in microcapsules (100~300 µm) with proper composition ratio. NF, core material, is a medication used to manage angina, hypertension, Raynaud’s phenomenon and premature labor [5,7,11]. It has a short lifetime of 2 or 3 h and is thus a suitable candidate for oral sustained release drug delivery. The microencapsulation of NF with PCL/PVP or PCL/PEG was carried out by solvent evaporation method in oil in water (O/W) emulsion system. The oil phase was formed by mixing of PCL, PVP(PEG), NF and dichloromethane (DCM) and then dispersed in polyvinyl alcohol (PVA) solution as stabilizer. In the present study, NF were microencapsulated and then the capsule size, the sustained drug release rate, and the cytotoxicity of the microcapsules were evaluated.

II. Materials and Methods

1. Materials

PCL (Mn=80,000), PVP (Mw=1,300,000), PEG (Mw=2,000, Yakuri Pure Chemicals Co., Ltd, Japan), PVA (Mw=85,000~124,000) and NF (≥98%) were purchased from Sigma-Aldrich. NF is 3,5-pyridinedicarboxyl acid, 1,4-dihydro-2, 6-dimethyl-4-(2-nitrophenyl)-,dimethyl ester, C₁₇H₁₈N₂O₅. Dissolving solvent of DCM (≥99.8%, Sigma-Aldrich, USA) was purchased and used as received without any further purification.

2. Microcapsule synthesis

The PCL solutions (4wt%~10wt%) dissolved in DCM were prepared by stirring (250 rpm) for 24 h at room temperature and then mixed with NF (10 mg) as a homogeneous solution. PCL/PVP or PCL/PEG solutions containing various weight ratios of 10/0 to 6/4 were prepared. PVA solution dissolved in distilled water was prepared by stirring with magnetic stirrer (RCT B SA0, Ika, Japan) at 500 rpm and 80℃ to obtain a viscous homogeneous aqueous dispersion [12,13]. The as-prepared PCL/PVP or PCL/PEG solutions were injected into PVA aqueous solution (0.5 to 2.0 wt%) and stirred with magnetic stirrer for 5 h at 1500 rpm to form the microcapsules. The microcapsules were screened through a 200 mesh sieve (75 µm). The capsules were cleaned in ethanol for 0.5 h and distilled water for 0.5 h to remove the residual solvent. The microcapsules were dried overnight in a vacuum oven. They were stored in a desiccator at room temperature and protected from light until use due to photosensitivity of NF [7,11]. The as-dried microcapsules were examined by using SEM (S-3000H, Hitachi, Japan) and optical microscopy (SV-55, Sometech, Korea) to investigate the morphology and the size of the capsules.

3. Characterization

10 mg of NF and 0.1 g of dried microcapsules were dissolved in 20 mL of enthanol by stirring in a filtering flask at 100 rpm and 37℃. At the given time intervals, the analytical samples (5 mL) were taken regularly from the supernatant of the effluent solution and then filtered through a 0.45 µm Millipore filter to remove the particles. The variation of adsorption at 238 nm was examined to identify the concentration of NF using an UV-vis spectrophotometer (Jasco V-670, Japan). The amount of drug release was summed up to determine the cumulative drug release rate (%) as a function of time. The crystalline phase of the capsules was analyzed by using the Xray diffraction (XRD, KFX-987228-SE, MacScience, Japan) and Fourier-transform Infrared Spectroscopy (FT-IR, Prestage 21, Shimazu, Japan). Differential scanning calorimetry (DSC, STAS 409C/31F, Netzsch, Germany) studies were performed as the temperature rose from room temperature to 200℃ at a heating rate of 10℃/min. All experiments were performed in triplicate. Values in the text were expressed as the means ± standard deviation, and p < 0.05 was considered statistically significant.

4. Cytotoxicity

4. Cytotoxicity The extract test method was conducted on the microcapsules to evaluate the potential of cytotoxicity on the base of the International Organization for Standardization (ISO 10993-5) [16-18]. The microcapsules were extracted aseptically in single strength Minimum Essential Medium (1X MEM, Dulbecco’s Modified Eagles’s Medium (Gibco) with 10% fetal bovine serum (Gibco) and 1% penicillin-streptomysin) with serum. The ratio of the microcapsules to extraction vehicle was 0.2 g/mL. Detailed experimental procedures were described elsewhere [16-20].

III. Results and Discussion

As the PCL concentration rose from 4 wt%, 6 wt%, 8wt% to 10wt%, the PCL capsule size increased linearly from 150 ± 25 µm, 340 ± 56 µm, 530 ± 36 µm, to 750 ± 80 µm probably due to higher loading of PCL, as displayed in Fig. 1. In the present study, the composition of 4 wt% PCL was chosen due to the capsule size (~150 µm) [12,13]. The effect of PVA emulsifier concentration on the size of the 4wt% PCL microcapsules was investigated, as shown in Fig. 2. The stabilizer (emulsifier) content is known to be critical to the microcapsule size and the surface state [1-11]. In general, very irregular-shaped microcapsules were produced due to low affinity with core and shell materials as a result of higher hydrophilicity and hydrophobicity of the conventional emulsifiers [3,4]. On the contrary, in the case of microcapsules prepared by using PVA emulsifier having both a hydrophilic (-OH) and a hydrophobic (-CH₂-CH-) groups simultaneously (Fig. 3) and having a large hydrophobic group molecular structure, uniform and highly stable spherical microcapsules with a narrow size distribution are likely to be synthesized [3]. This may be due to the high affinity of the PVA for the shell materials and the core material. The hydrophobic group of lipid soluble NF and PVA, the ester group of PCL (-COO) and the hydrophilic group (-OH) of PVA showed excellent selectivity. As the concentration of the stabilizer increased between 0.5 and 1.5%, the average size of the capsules exhibiting smooth and stable spheres has been reported to decrease [3]. However, our studies revealed that the PCL capsule size increased from 123±45 µm to 154 ± 25 µm with increasing the PVA concentration from 0.5 wt% to 1.0 wt% and then decreased down to 94 ± 31 µm when the PVA concentration was raised to 1.5 wt%, as depicted in Fig. 2. As the amount of emulsifier increased to 1.5 wt%, the emulsifier formed a layer on the particle surface, which may interfere adhesion of the shell material by the mutual repulsive force [3]. The PVA concentration of 1.0 wt% was chosen in the present study for the formation of PCL capsule having an average size of 154 ± 25 µm due to nearly spherical shape with a narrow size distribution.

Fig. 1. SEM images of (a) 4wt%, (b) 6 wt%, (c) 8wt%, and (d) 10 wt% PCL microcapsules.

Fig. 2. SEM images of 4wt% PCL microcapsules prepared by (a) 0.5 wt%, (b) 1.0 wt%, (c) 1.5 wt%, and (d) 2.0 wt% of PVA.

Fig. 3. Chemical structures of NF, PCL, PVA, PVP, and PEG.

The melting points of PCL, PVP and PEG are 60℃, 150℃ and 50~52℃, respectively. DSC was examined to investigate the effect of additives (PVP and PEG) on the PCL/PVP and the PCL/PEG capsules, as shown in Fig. 4. The melting temperature (T_m) of 4wt% PCL appeared approximately at 65.0℃ (Fig. 4(a)), which is in good agreement with the reference [19]. T_m of the PCL/PVP capsules increased from 65.0℃ to 66.3℃ with increasing the PVP content from 0 wt% to 20 wt% due to the blending effect of PVP (T_m = 150℃). However, T_m decreased to 64.4℃ when PVP was added to PCL more than 30 wt%, suggesting that no appreciable PVP contribution to the PCL/ PVP capsules was observed because T_m was within the error range. For the PCL/PEG capsules, T_m increased slightly from 65.0℃ to 65.5℃ with increasing the PEG content from 0 to 20 wt%, as displayed in Fig. 4(b). However, it decreased dramatically down to ~62.4℃ when PEG was added to PCL more than 30 wt%. It is likely due to the blending effect of PEG because Tm of PEG was in the range of 50℃ to 52℃. It was conceivable that the effect of PEG addition on the T_m of PCL microcapsules was more pronounced than that of PVP addition.

Fig. 4. DSC curves of various (a) PCL/PVP and (b) PCL/PEG microcapsules.

As PCL/PVP and PCL/PEG ratios were raised from 10/0 to 6/4, the capsule size increased gradually from 154 ± 25 µm to 236 ± 32 µm and 248 ± 56 µm, respectively, as shown in Figs. 5 and 6. Unlike the hydrophobic PCL, the affinity to water increased with increasing the content of the hydrophilic polymers such as PVP and PEG, resulting in an increase in the capsule size. However, the morphology was changed from sphere to airy bread-like shape. Although the mechanisms involved are not yet completely understood, they may be due to the increased porosity as a result of increased amounts of additives (PVP and PEG).

Fig. 5. SEM images of microcapsules containing various PCL/PVP ratio of (a) 10/0, (b) 9/1, (c) 8/2, (d) 7/3, and (e) 6/4, respectively.

Fig. 6. SEM images of microcapsules containing various PCL/PEG ratio of (a) 10/0, (b) 9/1, (c) 8/2, (d) 7/3, and (e) 6/4, respectively.

No characteristic PCL peaks located at 2θ = 21.4^o , 22.0^o , and 23.7^o , corresponding to the (110), (111), and (200) planes of an orthorhombic crystal, were visible (Fig. 7). However, the PVP peaks became dominant as the PVP content rose in the PCL/PVP capsules, as shown in Fig. 7(a) [19-23]. Strong PEG reflections at 2θ of 19.23^o and 23.34^o and weak reflections at 13.61^o and 27.32^o were reported [22]. In the XRD profile of PCL/PEG capsules, no appreciable XRD peak change was observed for the PCL/PEG capsules when the PCL/PEG content was varied from 9/1 to 8/2, however, the PCL peak intensity started to decrease with further increasing the PEG content, as depicted in Fig. 7(b). The dramatic increase in XRD peak located at 25.9^o was detected probably due to the increased PEG contribution to the PCL/PEG capsules.

Fig. 7. XRD patterns of (a) PCL/PVP and (b) PCL/PEG capsules having different concentration.

NF can potentially act as either proton acceptor (through the carbonyl groups, -C=O) or proton donor (through the amine group, -NH) [11]. The hydrogen bonds (H-bonds) are reported to be formed via the amine group and one of the carbonyl groups. The Hbonded NF amine and carbonyl groups were observed to have strong stretching vibration bands at 3330 cm⁻¹ (-NH) and at 1679 cm⁻¹ (-C=O), in Fig. 8. Then, the H-bonds originally formed among NF molecules were broken and replaced by those formed between NF and PCL. A new sharp stretching vibration band at 1720 cm⁻¹ was observed on the NFloaded PCL capsule because of the NF carbonyl group formed H-bonds with PCL. A new broad band at the wavenumber region of from 3140 cm−1 to 3475 cm−1 was developed in the spectra of NF-loaded PCL capsule, indicating that the NF amine group associated with PCL via H-bonds was formed [11]. The drug release behavior of NF is shown in Fig. 9. The drug release rate of PCL/PVP and PCL/PEG capsules increased dramatically from 0 to 4 h at the beginning and then reached the plateau region from 20 h. As the concentration of PVP or PEG increased, the amount of drug release increased, suggesting that the larger capsule size was attributed to the higher drug content. The capsule size affected the drug amount, but the drug release time remained almost constant regardless of capsule size, as depicted in Fig. 9.

Fig. 8. FT-IR spectra of PCL, NF, and NF-loaded PCL microcapsule.

Fig. 9. NF release behavior of (a) PCL/PVP and (b) PCL/PEG microcapsules.

A cytotoxicity test of the PCL capsules determines whether a product or compound will have a toxic effect on living cells [16-20]. The test extract with the PCL capsules with or without NF showed no evidence of causing cell lysis or toxicity, as depicted in Fig. 10. The PCL capsules with or without NF exhibited average cell amount of 118% and 110% compared to the negative control, respectively, as measured at a wavelength of 415 nm by using the microplate absorbance spectrophotometer. The qualitative morphological grading of cytotoxicity of the PCL capsules was determined to be scale 0. Therefore, it is conceivable that the PCL microcapsules with encapsulated NF have no cytotoxicity under the condition of this study and are considered to be clinically safe and effective.

Fig. 10. Photographs of cell morphologies: (a) positive control, (b) negative control, (c) PCL/PVP capsules (c) without NF and (d) with NF, respectively.

IV. Conclusions

The NF-loaded PCL/PVP or PCL/PEG microcapsules were prepared by an O/W emulsion method through proper adjustment of the process parameters. The capsule size of 4 wt% PCL/PVP and PCL/PEG capsules rose from 154 ± 25 µm to 236 ± 32 µm and 248 ± 56 µm with increasing the PVP and PEG content from 0 to 40%, respectively, due to the hydrophilicity of PVP and PEG. The NF release rate of PCL/PVP and PCL/PEG capsules increased dramatically from 0 to 4 h at the beginning and then reached the plateau region from 20 h. Although the amount of drug release increased with increasing the PVP and PEG content, the drug release rate remained constant. The PCL capsules exhibited no evidence of causing cell lysis or toxicity regardless of NF loading, implying that the microcapsules are clinically suitable for use as drug delivery systems.