Article pubs.acs.org/Biomac

Dithiol-PEG-PDLLA Micelles: Preparation and Evaluation as Potential Topical Ocular Delivery Vehicle Jian Yang, Jing Yan, Zhihan Zhou, and Brian G. Amsden* Department of Chemical Engineering, Queen’s University, Kingston, Ontario, Canada ABSTRACT: Thiol-modified nanoparticles have potential applications in mucoadhesive drug delivery and have been examined in this regard for topical ocular delivery. In this paper we provide a simple method for the synthesis of a dithiol terminated amphiphilic diblock copolymer. Bidentate dithiolpoly(ethylene glycol)-poly(D,L-lactide) (SH2-PEG-PDLLA) was synthesized and micelles with dithiol-containing coronas were prepared from this block copolymer via the emulsion method. In vitro release studies indicated that the presence of the thiol groups at the surface did not affect the rate of release of dexamethasone, used as a representative ocular drug. The micelles also showed low cytotoxicity to human corneal epithelial cells (HCEC) and murine fibroblast cells (3T3 cells). A hydrophobic red fluorophore, Nile red, was loaded into the core of micelles and confocal microscopy was used to study HCEC uptake and retention of the micelles. The micelles were rapidly endocytosed by the HCEC, with intracellular micelle levels remaining unchanged with incubation times from 5 to 120 min. Interestingly, Nile red was eliminated significantly more slowly from HCECs treated with the thiolated micelles. These results suggest that these dithiolated micelles may be effective for topical ocular drug delivery.

■

INTRODUCTION Topical drug therapy is the main method to treat ocular diseases. However, the conventional means of achieving topical drug delivery through the use of eye drops is limited by short residence times and the anterior segment barriers that limit drug transport; less than 5% of a drug can be absorbed by eye drop administration.1,2 Moreover, poor water solubility makes a great quantity of hydrophobic drugs unsuitable for aqueous eye drop formulations. Many formulation strategies, therefore, are being developed to overcome these challenges, including gels,3,4 ointments,5 and microemulsions.6 In particular, micelles as nanoparticulate drug delivery systems have been examined as a means of improving topical delivery of hydrophobic drugs to the eye.7−10 To improve residence time on the ocular surface, mucoadhesivity can be imparted to the micelles through incorporation of a reactive group capable of binding to chemical groups present in the mucin layer.11 An interesting property of the thiol group is that it can combine with cysteine, which is abundant on the surface of mucin protein11,12 and so materials with thiol groups adhere to the mucin protein and increase the retention time. One approach then is to prepare micelles using thiol-terminated hydrophilic polymers. For example, Shen et al. recently showed improved retention of cyclosporine-loaded cysteine-poly(ethylene glycol)-stearate micelles on the ocular surface of the cul-de-sac of rabbit eyes over that of nonthiolated micelles.10,13 Thiolated micelles were retained on the ocular surface for up to 6 h. Such an improvement in retention is notable; nevertheless, it would be desirable to achieve retention times of 24 h or longer. © 2014 American Chemical Society

We hypothesized that increased retention of drug-loaded micelles might be possible if a bidentate thiol group was incorporated onto the terminus of the poly(ethylene glycol) (PEG) capable of increased interactions with thiol-reactive groups within mucin. Toward this objective, we utilized the conjugation strategy of Uyeda et al.14 to generate α-lipoic acidω-hydroxyl-PEG (LA-PEG-OH), then employed this polymer as a macroinitiator in the ring-opening polymerization of D,Llactide (Scheme 1). Lipoic acid is abundant in nature and is synthesized in the human body where it is used as an antioxidant.15 Following reduction of the lipoic acid to 6,8dimercaptooctanoate, these diblock copolymers were then used to form micelles. We disclose herein the preparation and characterization of this diblock copolymer, the preparation and characterization of micelles using this copolymer, demonstrate that these micelles are noncytotoxic at concentrations relevant for in vivo use, and that the dimercapto groups on the surface do not impede the release of a model drug, dexamethasone. Dexamethasone was chosen for preliminary assessment of the approach as dexamethasone eye drops and ointment are used to treat a number of eye conditions associated with inflammation of the eye (e.g., allergic conjunctivitis). Received: December 20, 2013 Revised: March 7, 2014 Published: March 10, 2014 1346

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

Scheme 1. Synthesis Route for the Preparation of 6,8-Dimercaptooctanoate-PEG-PDLLA (SH2-PEG-PDLLA) Block Copolymer

■

pipet, 10 mL of cold water was added to the oil, and the mixture was heated again. The procedure was repeated three times to remove unreacted dilipoic acid modified PEG. After lyophilization, 0.84 g of block polymer was obtained as a white powder. Reduction of LA-PEG-PDLLA. A 600 mg aliquot of LA-PEGPDLLA (0.076 mmol) was dissolved in 20 mL of a mixture of ethanol and water (1:4, v/v), then 66 mg of TCEP was added, and the mixture was stirred for 2 h under the protection of argon. After reaction, 50 mL of dichloromethane was added and the solution was washed by water (50 mL × 3) then dried by magnesium sulfate. The hydrolyzed 6,8dimercaptooctanoate-PEG-PDLLDA diblock copolymer (SH2−PEGPDLLA) was obtained as a white powder. The reduced polymer was stored at −20 °C under argon until used to form micelles. Synthesis of Methoxy-PEG-PDLLA. PEG monomethyl ether (5000 g/mol) was used to synthesize methoxy capped PEG-PDLLA diblock copolymer (MPEG-PDLLA). A total of 1.0 g of PEG monomethyl ether (0.2 mmol), 1.0 g of D,L-lactide (6.9 mmol), and 2.8 mg of stannous 2-ethylhexanoate (0.0069 mmol) were loaded into a 10 mL glass ampule. The ampule was sealed under vacuum, and placed in an oven at 130 °C for 15 h. The product was purified by dissolving in dichloromethane and precipitating in cold diethyl ether. The white precipitate was filtered and dried in vacuum to obtain 1.92 g methoxy-PEG-PDLLA block polymer. Polymer Characterization. Structure, composition, and number average molecular weight of the polymers were determined by 1H NMR using a Bruker spectrometer (400 MHz) and CDCl3 as the solvent. The resulting peaks were compared to the solvent peak relative to a tetramethylsilane (TMS) reference. Molecular weight and molar-mass dispersity were also analyzed by gel permeation chromatography (GPC) using a Waters chromatography system mounted with Styragel HR 1−4 columns (Waters, U.S.A.) and connected to a Waters 410 differential refractometer. The flow rate was set at 1 mL/min, and tetrahydrofuran at 30 °C was the continuous

MATERIALS AND METHODS

PEG diol (4000 Da), PEG mono methyl ether (5000 Da), 4dimethylaminopyridine (DMAP), Ellman’s reagent (5,5′-dithiobis-(2nitrobenzoic acid) or DTNB) and dexamethasone were purchased from Aldrich (Mississauga, ON, Canada). Lipoic acid and 1-ethyl-3-(3dimethylaminopropyl)carbodiimide (EDC) were purchased from Fisher (Mississauga, ON, Canada), D,L-lactide, tris-(2-carboxyethyl)phosphine hydrochloride (TCEP) was from Soltec Ventures (Beverly, MA, U.S.A.), and stannous 2-ethylhexanoate was from Alfa Aesar (Ward Hill, MA, U.S.A.). Synthesis of Lipoic Acid PEG (LA-PEG-OH). A total of 4.0 g of PEG diol 4000 (1.0 mmol) was dissolved in 200 mL of dichloromethane, then 279 mg EDC (1.8 mmol), 19 mg DMAP (0.16 mmol), and 310 mg lipoic acid (1.5 mmol) were added into the solution and stirred under the protection of argon. After 24 h, 93 mg EDC (0.6 mmol), 6.5 mg DMAP (0.05 mmol), and 103 mg lipoic acid (0.5 mmol) were added into the mixture and stirred for 24 h. The product was purified by dialysis against deionized water (membrane MWCO: 3500 Da) after removing the solvent by vacuum evaporation. The mixture of monolipoic acid terminated PEG (LA-PEG-OH) and dilipoic acid modified PEG (LA-PEG-LA) was collected from the water by lyophilization. Synthesis of Lipoic Acid PEG-b-poly(D,L-lactide) (LA-PEGPDLLA). A total of 1.1 g of the mixed LA-PEG-OH and LA-PEG-LA product, 0.45 g of D,L-lactide (3.125 mmol) and 1.31 mg of stannous 2ethylhexanoate (0.0032 mmol) were loaded into a 10 mL glass ampule. The ampule was sealed under vacuum and placed in an oven at 130 °C for 15 h. After cooling, the product was dissolved in 5 mL of dichloromethane and precipitated in 250 mL of cold diethyl ether. After removing the solvent, the white solid was dissolved in 15 mL of deionized water in a 50 mL glass beaker. The water solution was heated to 80 °C with stirring until a colorless oil precipitated to the bottom (approximately 3 min). The supernatant was removed by glass 1347

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

phase. Polystyrene samples of known molecular weights were used as standards for calibration. Micelle Preparation and Characterization. Drug-free micelles were prepared by dissolving 30 mg of MPEG-PDLLA and 10 mg SH2PEG-PDLLA in 400 μL acetonitrile (thiolated micelles). The polymer solution was dripped into 2 mL of deionized water slowly with stirring. After adding the polymer, the mixture was stirred for 3 h to evaporate the acetonitrile. Drug-loaded micelles were prepared in a similar fashion, but with 2 mg/mL dexamethasone in the acetonitrile. To remove dexamethasone crystals and large particles, the micelle suspension was filtered through a 0.2 μm syringe filter. The drug loaded, nonthiolated micelles were made with MPEG-PDLLA diblock copolymer using the same method. The micelles were obtained by lyophilization and stored at −20 °C under the protection of argon until used. The dexamethasone content in the micelles was determined by HPLC. Freeze-dried, drug-loaded micelles were dissolved in acetonitrile/water (50:50 v:v) to destabilize the micelles and liberate the drug. Samples were prepared by diluting 100 μL of micelle solution with 900 μL of mobile phase (a mixture of methanol/water (80:20) with 0.05% trifluoroacetic acid). A Waters 600 controller, equipped with a Waters 2487 Detector (U.S.A.) and a C18 column was used. The flow rate was 1.2 mL/min, and the injection volume was 200 μL. Dexamethasone was detected at 254 nm with a retention time of 6 min. A calibration curve (coefficient of determination = 0.9994) was obtained with dexamethasone standard solutions with dexamethasone concentration from 5 to 20 μg/mL. The drug loading efficiency was calculated by the following equation:

loading efficiency% =

1% (w/v) penicillin/streptomycin (100 U/mL penicillin and 0.1 mg/ mL streptomycin). The cells were grown in 5% CO2 at 37 °C. The HCEC and 3T3 cells were seeded separately in 96-well plates (5000 cells per well) with culture medium in 5% CO2 at 37 °C overnight to adhere. Cells were treated with thiolated micelles or nonthiolated micelles diluted with culture medium at serial concentrations (5 − 20 mg/mL) for 24 h at 37 °C. Subsequently, the medium was removed and gently washed with PBS (pH = 7.4). A total of 120 μL of MTT (1 mg/mL in serum free DMEM medium) was supplied to each well followed by incubation at 37 °C for another 4 h, and finally, the formazan crystals were dissolved in 150 μL. The optical density (OD) was measured immediately at 570 nm in a microplate reader (Bio-Tek Instruments, Inc., Winooski, VT, U.S.A.). Nontreated cells assumed to possess 100% viability were used as controls and all measurements were made in quintuplicate. Cell viability was calculated as follows: cell viability(%) = (ODsample − OD blank )/(ODcontrol − OD blank ) × 100 Cellular Uptake Study. HCEC were incubated in a 96-well plate (10000 cells per well) with culture medium in 5% CO2 at 37 °C overnight to adhere. To investigate micelle cellular uptake, HCEC cells were stained with 10 μM calcein AM for 30 min and gently washed three times with sterile PBS buffer to remove the excess dye solution. Nile red loaded micelles were prepared as described above but the dexamethasone/acetonitrile solution was replaced with Nile Red/ acetonitrile solution (50 μg/mL). The stained HCEC were then exposed to the Nile Red loaded micelles (5 mg/mL in culture medium) in 5% CO2 at 37 °C for 5 and 120 min. The cells were thoroughly washed three times with sterile PBS buffer to remove residual micelles. The cells were then imaged immediately using an Olympus FV 1000 laser scanning confocal microscope. To investigate the intracellular retention of the micelles, HCEC were incubated in a 96-well plate (10000 cells per well) with culture medium in 5% CO2 at 37 °C overnight then exposed to Nile Red loaded micelles (5 mg/mL in culture medium) in 5% CO2 at 37 °C for 2 h. After washing three times with fresh sterile PBS buffer, groups of HCEC were incubated in fresh culture medium for either another 2 or 24 h. The HCEC were then stained with 10 μM calcein AM for 30 min and subsequently washed three times with sterile PBS buffer. The cells were imaged immediately under the confocal microscope. The images obtained were also used to quantify the degree of intracellular retention of the micelles. During the acquisition of all images, the microscope settings were kept constant in order to allow a comparison of the micelle retention at different times. The area of cells (Acells) and retained micelles (Amicelles) were obtained by analyzing the green and red channel signals on the micrographs using Image J 2.0 software. The area percentage of micelles within the cells was calculated as Amicelles/Acells × 100%. The standard deviation was calculated from three independent measurements. Statistical analyses were performed by one-way ANOVA with a Tukey’s posthoc comparison of the means using OriginPro8.0 Software. Differences were considered statistically significant at p < 0.05.

mass of drug loaded in the micelles × 100 total mass of micelles and drug

The average hydrodynamic diameter and the polydispersity (PDI) of the micelles were measured by Dynamic Laser Scattering (DLS) using a Zetasizer HS 3000 (Malvern Instruments, U.K.) at 25 °C. The micelle samples were diluted to 0.2 mg/mL with deionized water before the measurements. The mean values of three measurements of three runs were calculated. The morphology of the polymeric micelles was assessed from images obtained using transmission electron microscopy (TEM; FEI TecnaiTM G2 Sphera, U.S.A.). Samples were deposited on copper grids, stained with 1% uranyl acetate, and analyzed after drying. Assessment of the Stability of the Micelles and the Thiol Groups. Freshly prepared water suspensions of thiolated and nonthiolated micelles (10 mg/mL) were placed in a 4 °C refrigerator or a 37 °C incubator. At particular time intervals the size of the micelles was measured by DLS, as described above. To determine the stability of the thiol group, freshly made water suspensions of thiolated micelles (10 mg/mL) were placed at room temperature in ambient air. At specific time intervals, the presence of thiol groups was measured colorimetrically using Ellman’s reagent.16 In Vitro Dexamethasone Release. A total of 10 mg of dexamethasone-loaded thiolated micelles or nonthiolated micelles were suspended in 1 mL of PBS buffer (pH = 7.2). The solution was loaded into dialysis tubing (MWCO: 3500). The tubing was then immersed into 50 mL of PBS buffer. The device was placed on a rotating plate in a 37 °C incubator. At frequent time intervals 10 mL of the external PBS buffer was removed and 10 mL fresh buffer was added. The drug content in the PBS buffer was measured by HPLC as described above. For control samples, the dexamethasone-loaded micelles were replaced by free dexamethasone with a concentration of 100 μg/mL. Every measurement was performed in triplicate. In Vitro Cytotoxicity Study. The hybrid adenovirus 12-SV40 immortalized human epithelial corneal cell line (HCEC), a gift of Dr. Heather Sheardown of McMaster University, Canada, was grown in keratinocyte serum-free medium (KSFM, Gibco #17005−042) supplemented with bovine pituitary extract, epidermal growth factor, 100 U/mL penicillin/streptomycin (Gibco #15140−122), and 10 U/ mL gentamycin (Gibco #15710−064). 3T3 (ATCC CCL 92) cells were grown in Dulbecco’s Modified Eagle’s Medium (DMEM, Gibco #15140−122) supplemented with 10% (v/v) fetal bovine serum and

■

RESULTS AND DISCUSSION Polymer Synthesis. Scheme 1 shows the synthesis route used to prepare the lipoic acid terminated PEG-PDLLA (LAPEG-PDDLA) block copolymer via the EDC coupling method. Compared to amidation reactions, EDC-mediated coupling gives low yields for esterification. To avoid the byproduct of a triblock polymer, it was important to eliminate PEG diol in the esterification reaction products before the ring-opening polymerization step. Uyeda et al.14 reported the synthesis of monolipoic acid modified low molecular weight PEG (less than 1000 Da) by a dicyclohexylcarbodiimide (DCC)-mediated esterification reaction and used excess PEG, with the unreacted PEG removed by column chromatography. For high molecular 1348

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

weight PEG such as the PEG 4000 used in this work, column chromatography was unable to separate the PEG diol and monolipoic acid modified PEG because the polarities of the two macromolecules were very close. Therefore, excess lipoic acid was used in an EDC mediated coupling reaction to obtain a mixture of monolipoic acid (LA-PEG-OH) and dilipoic acid terminated PEG (LA-PEG-LA). The product was then purified by precipitation in cold diethyl ether and dialysis against water. The mass ratio of LA-PEG-OH and LA-PEG-LA was consistently approximately 45:55, as measured from GPC. The two-peak trace in Figure 1 (dashed line) is the mixture of LA-PEG-OH and LA-PEG-LA obtained following esterification. This product mixture was used directly for the subsequent polymerization step.

Figure 2. 1H NMR spectra of LA-PEG-PDLLA (top) and SH2-PEGPDLLA (bottom) block copolymers with peak assignments.

spectrum shows that the block copolymers obtained were free of residual monomer and catalyst, and confirmed that the lipoic acid was present at the PEG terminus. The number average molecular weight (Mn) was calculated from the integration of the methine protons on the PDLLA block (δ = 5.1 ppm) and the ethylene protons on the PEG block (δ = 3.8−4.0 ppm), and using the Mn of PEG as 4000 Da. The Mn calculated in this way was 7650 Da, which compared favorably with the GPC measurement value, and the Mn of the PDLLA block in the LAPEG-PDLLA was 3650 Da. Uyeda et al. reduced the disulfide in the dithiolane ring on the lipoic acid using sodium borohydride.14 When sodium borohydride was used to reduce the disulfide bond on the LAPEG-PDLLA, however, it also reduced the ester bonds of the PDLLA residues resulting in a reduction in molecular weight. A similar result was reported by Timbart et al. when sodium borohydride was used to reduce a ketone modified poly(εcaprolactone).17 We therefore used tris-(2-carboxyethyl)phosphine hydrochloride (TCEP) to reduce the LA-PEGPDLLA at room temperature with ethanol/water as medium. TCEP was chosen because alkylphosphines are selective for disulfides and unreactive toward many other functional groups.18 TCEP effectively reduced the disulfide to yield the 6,8-dimercaptooctanoate-terminated PEG-PDLLA (SH2-PEGPDLLA) without cleavage of the ester bonds, either between the PEG and PDLLA and within the PDLLA block. This was confirmed from an analysis of the 1H NMR spectra of the reduced diblock copolymers. The lower spectrum in Figure 2 shows a representative 1H NMR spectrum of the SH2-PEGPDLLA diblock copolymer. Compared to the spectrum obtained for the LA-PEG-PDLLA, the proton peaks of the opened five-unit ring all shift (i′, j′, k′) and peaks appear for the thiol groups at 1.25 and 1.3 ppm, indicating that the reduction was accomplished completely. Micelle Fabrication and Drug Loading. Monomethoxy PEG 5000 was used as initiator to synthesize MPEG-PDLLA diblock polymers, which were then used to control the surface

Figure 1. GPC traces of the reaction products following esterification (mixture of OH-PEG-LA and LA-PEG-LA, dashed line) and ringopening polymerization following purification using water washes (solid line). The bimodal trace obtained before washing shows the presence of LA-PEG-LA (elution time 6.7−7.5 min), which is completely removed following water washing to yield LA-PEGPDLLA.

Ring-opening polymerization was employed to synthesize a monolipoic acid-terminated PEG-PDLLA diblock copolymer using stannous 2-ethylhexanoate as catalyst and LA-PEG-OH as the macroinitiator. LA-PEG-LA was inert in the polymerization reaction. After removing residual monomer and catalyst by precipitation in cold diethyl ether, LA-PEG-PDLLA and LAPEG-LA were separated by hot water washing. LA-PEGPDLLA is a thermosensitive block copolymer, and at room temperature the polymer chains self-assemble into micelles in water but at higher temperature the micelles are unstable and the polymer precipitates out from the water as an oil. Figure 1 also shows a characteristic GPC trace of the purified polymerization product (solid line), which shows that the LA-PEG-LA has been completely removed. As measured by GPC, the LA-PEG-PDLLA was typically obtained with a number average molecular weight of 7850 Da and a molar mass dispersity of less than 1.3. Figure 2 shows a representative 1H NMR spectrum of LAPEG-PDLLA (upper spectrum) with peak assignments. The 1349

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

micelles.20 These results indicate that the presence of thiol groups on the surface of the micelles had no detectable influence on the physical properties and the morphology of the micelles. Quantification and Stability of the Thiol Groups and Stability of Micelles. Ellman’s test was employed to quantify the thiol groups on the surface of the micelles. Table 2 shows

thiol concentration of the micelles. The number average molecular weight and molar-mass dispersity of MPEG-PDLLA block polymer were 9600 Da and 1.14, respectively. Mixtures of the MPEG-PDLLA and SH2-PEG-PDLLA diblock copolymers of different weight ratios were used to fabricate micelles via the emulsion method,19 with and without dexamethasone (DEX) present. The characteristics of the micelles obtained, as well as the DEX loading efficiency, are provided in Table 1. Changing

Table 2. Stability of Thiol Groups on the Micelle Surfacea

Table 1. Properties of Micelles With or Without Loaded DEX without DEX mass ratio of SH2-PEGPDLLA and MPEG-PDLLA

size (nm)

1:0 6:1 3:1 1:1 1:3 1:6 0:1

37.5 36.3 39.5 41.2 40.7 39.0 42.5

0 hb

with DEX

PDI

size (nm)

PDI

DEX content (%)

0.31 0.31 0.26 0.22 0.23 0.16 0.12

43.6 45.4 43.2 44.7 47.3 45.1 49.5

0.28 0.29 0.23 0.19 0.20 0.17 0.10

2.1 1.7 2.0 1.9 2.1 1.9 2.0

3h

6h

size (nm) PDI thiol content size (nm) PDI thiol content size (nm) PDI thiol content

thiolated micelles

nonthiolated micelles

42 0.24 1 44 0.21 0.93 44 0.26 0.60

39 0.30 42 0.27 41 0.25

a

The quantification of thiol group was done using the Ellman’s test with the amount of thiol at each time point normalized to that at time 0. bTime zero was taken as the end of stirring of the micelle suspension to remove the organic solvent.

the ratio of MPEG-PDLLA to SH2-PEG-PDLLA did not have an influence on the average size of the micelles, which ranged from 36 to 43 nm. Table 1 also shows that the micelles containing a greater weight fraction of MPEG-PDLLA had a lower size dispersity, which can be explained by the narrower molar-mass dispersity of MPEG-PDLLA compared to SH2PEG-PDLLA. Figure 3 shows the TEM images of nonthiolated micelles (a) and thiolated micelles (b; MPEG-PDLLA/SH2-PEG-PDLLA = 3:1). Both groups of micelles were spherical in shape and of a similar size. Dynamic light scattering measurements (Table 1) also showed that there was no significant difference between the diameters of nonthiolated micelles and thiolated micelles. Loading the hydrophobic drug enlarged the diameter of the hydrophobic core and increased the diameters of the micelles by about 10 nm (Table 1). Moreover, with or without thiol groups the micelles encapsulated a similar amount of dexamethasone into their cores, which was approximately 2%. This loading content is consistent with that recently reported for dexamethasone loading into PEG-poly(ε-caprolactone)

that after 3 h of stirring to remove the organic solvent, the quantity of the thiol groups was close to the theoretical quantity of the thiol groups on the DLA-PEG-PLA block polymer used for the micelles fabrication. Kalarickal et al. reported the aggregation of micelles prepared with monothiol-PEG-polyesters due to the formation of intermicelle disulfide bonds under oxidative conditions.21 We therefore studied the stability of the nonthiolated and thiolated micelles. After 3 and 6 h stirring in air the quantity of thiol groups remaining on the thiolated micelles decreased by 30 and 60%, respectively (Table 2). Nevertheless, DLS measurements indicated that there was no obvious aggregation of the micelles. This result indicates that under oxidative conditions, some of the 6,8-dimercaptooctanoate termini were oxidized and the dithiolane ring reformed, but that there was no intermicelle disulfide bonds formed. Disulfide bond reformation was eliminated when appropriate storage of the micelles was used. Storing the micelles at −20 °C under argon protected the thiol groups from oxidation for 30 days.

Figure 3. Representative TEM images of thiolated micelles and nonthiolated micelles. The scale bar is 100 nm. The thiolated micelles were prepared with a 3:1 ratio of MPEG-PDLLA/SH2-PEG-PDLLA. 1350

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

released in this phase is likely from the surface of the micelles; some DEX encapsulated into the micelles possibly interacts with the PEG block by secondary forces such as Van de Waals’s forces or hydrogen bonding between PEG chain and DEX molecules. After 72 h the DEX release from the micelles was more sustained, with approximately 1.5−2 μg DEX released per day for more than 10 days. The similarity of the drug release from nonthiolated and thiolated micelles indicated that the thiol group on the surface of micelles did not influence the partitioning of the hydrophobic drug from the core to the aqueous environment. While increasing the release duration of DEX from these micelles was not an objective of the present study, the kinetics of DEX release from these micelles is also consistent with that reported for release from PEG-poly(εcaprolactone) micelles by Lu et al.,20 suggesting that little improvement in release duration would be possible by altering the nature of the polyester block of the diblock copolymer. Possible strategies for increasing DEX release duration for future work include incorporating degradable covalent linkages between DEX and the hydrophobic polymer block22 or utilizing a silica shell around the hydrophobic core,20 both of which have been shown to increase the release duration of DEX from micelles. In Vitro Cytotoxicity. The assessment of the interaction between corneal epithelium cells and the thiolated micelles is important for assessing the potential for topical ophthalmic drug delivery, while the 3T3 murine fibroblast cell line is a standard cell type for material cytotoxicity assessment. Therefore, the cellular toxicity of micelles to human corneal epithelial cells (HCEC) in comparison to the 3T3 fibroblast cell line was evaluated by MTT assay (Figure 6). At or below 10 mg/mL, cell viability was approximately 85% following 24 h incubation with either thiolated or nonthiolated micelles for both cell lines examined, and there was no significant difference between cell response to either thiolated or nonthiolated micelles (p < 0.01). As the micelle concentration was increased to 20 mg/mL, the thiolated micelles induced significantly greater cell death than did the nonthiolated micelles. Again, the cellular response was consistent and not significantly different (p < 0.05) for both the 3T3s and the HCECs. These micelles can therefore be considered to possess relatively low cytotoxicity, with an upper limit in application to the ocular surface of approximately 10 mg/mL. Cell Uptake and Retention. The interaction between micelles and cells is important for the micelles as a drug carrier. Corneal epithelial cells form the outer eye barrier; therefore the rate of endocytosis and duration of intracellular micelle retention within the HCECs was examined to probe potential in vivo responses to these micelles. To investigate the endocytosis rate of the micelles, confocal laser scanning microscopy was used to monitor the location of micelles within the HCECs at different times. In brief, the micelles were tracked by a fluorescent hydrophobic probe, Nile red, while the live HCECs were tracked by the polyanionic dye calcein AM with green fluorescence. Nile Red fluoresces only in a hydrophobic environment23 and so the images only show intact micelles, as disassembled micelles and free Nile Red would not be visible. In these experiments a micelle concentration of 5 mg/mL was utilized, based on the cytotoxicity findings. Representative confocal images of HCECs incubated with nonthiolated and thiolated PEG-PDLLA micelles for 5 min and 2 h are shown in Figure 7A. After 5 min of incubation with both

Micelle stability in the aqueous environment is important for the proposed drug delivery application. Figure 4 shows the

Figure 4. Time variation of micelle average diameter following suspension in water at 4 and 37 °C.

change in average diameter of thiolated and nonthiolated micelles suspended in deionized water at both 4 and 37 °C. The diameter of the micelles increased slightly over the first 72 h then remained stable over the next 7 days. This result also supports the conclusion that intermicelle disulfide bond formation does not occur. In Vitro Dexamethasone Release. The release profile of dexamethasone (DEX) from both nonthiolated and thiolated micelles is given in Figure 5. Also shown in Figure 5 is the

Figure 5. Cumulative dexamethasone release from both thiolated and nonthiolated micelles.

dissolution kinetics of free DEX crystals. Within the first hour 78% of free DEX crystals had dissolved and diffused through the dialysis tubing, and by 72 h the free DEX had been completely transported from within the dialysis tubing. DEX released from nonthiolated and thiolated micelles had similar release curves. In the first hour about 57% of drug was released and after 8 h 90% of the DEX had been released. The DEX 1351

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

Figure 6. Variation of cellular viability of the HCEC (a) and 3T3 cell lines (b) after 24 h incubation with micelles with different concentrations. The viability of cells incubated in free micelle-free medium as control was taken as 100%. Values are mean ± standard deviation, n = 5.

Figure 7. (A) Confocal images of HCECs incubated with 5 mg/mL thiolated and nonthiolated micelles for different times. (B) Relative % area of micelles within the cytoplasm of the HCECs.

thiolated and nonthiolated micelles, there is no evidence of extracellular red fluorescence, indicating that the washing step

effectively removed nonendocytosed micelles. Moreover, the merged red and green channel images clearly showed yellow in 1352

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

Article

Figure 8. (A) Confocal images of HCECs incubated with 5 mg/mL thiolated and nonthiolated micelles for 120 min, followed by washing and incubation in culture medium for 2 and 24 h. (B) Relative % area of micelles within the cytoplasm of the HCECs as a function of time. “#” represents a significant difference between sample populations for p < 0.05.

the cytoplasm, indicating that both nonthiolated and thiolated micelles were rapidly endocytosed by the HCECs. Additional incubation of up to 2 h with micelles did not increase the red fluorescence noticeably. Cellular uptake of nanoparticles depends on the nanoparticle size, shape, surface charge, and surface chemistry, often in an interrelated fashion,24 making comparison of the cellular uptake rate in this study to that found in other studies difficult. Nevertheless, this rate of uptake of micelles is similar to that previously reported by Allen et al., who observed that 25 nm diameter, spherical micelles composed of poly(ε-caprolactone)-b-poly(ethylene glycol) were internalized within PC12 cells within 5 min.25 There was no significant difference in cellular uptake of the thiolated versus nonthiolated micelles. These observations were confirmed by comparing the relative area of intracellular red fluorescence to overall green fluorescence area (Figure 7B).

Thus, there likely is not a different mechanism for micelle uptake into HCECs for thiolated micelles, such as interactions with protein receptors on the surface, than there is for nonthiolated micelles. The rate of cell uptake of micelles has received a considerable amount of attention, but is only one factor governing effective drug response. Intracellular retention of the drug is also important,26 yet there have been fewer studies investigating this phenomenon. To investigate the degree of intracellular retention of a hydrophobic drug from the thiolated versus nonthiolated micelles, HCECs were incubated with Nile red loaded micelles for 2 h, then washed and incubated with fresh medium for either an additional 2 or 24 h before being imaged with the confocal microscope. Representative images are shown in Figure 8A. The images clearly show that the Nile red was retained within the cell cytoplasm for at least 24 h, and there 1353

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354

Biomacromolecules

■

ACKNOWLEDGMENTS Funding for this project was provided by the Natural Sciences and Engineering Research Council of Canada through the 2020 Ophthalmic Biomaterials Network.

appeared to be a greater retention of the Nile red in the cells initially incubated with the thiolated micelles. To confirm this observation, the relative area of red fluorescence to green fluorescence was used to quantify the amount of intracellular Nile red at each time point following incubation (Figure 8B). This analysis demonstrated that elimination of the Nile red from HCECs treated with the thiolated micelles was slower than that from HCECs treated with nonthiolated micelles. In particular, at 2 h the Nile red concentration within the HCECs treated with thiolated micelles was more than twice that for the nonthiolated micelle treated group. By 24 h, both thiolated and nonthiolated micelle treated groups had statistically equivalent amounts of intracellular Nile red, which were significantly lower than the levels following treatment with the micelles and washing. The reasons for the greater retention of the Nile red within the HCECs following treatment with the thiolated micelles is presently unknown. Micelles are endocytosed through either calveolae or clathrin mediated processes,27 and it may be that the thiol groups at the surface of the micelle coronas form temporary disulfide bonds with calveolae or clathrin proteins on the cell surface. These temporary bonds may lead to enhanced micelle cytoplasmic retention and thus greater Nile red intracellular retention. It has also recently been demonstrated that micelles whose structure has been stabilized through cross-linking are more rapidly eliminated from cells through exocytosis.26 Thus, it may be that the thiolated micelles are less structurally stable than the nonthiolated micelles. Further study is necessary to clarify the reasons behind this result. Enhanced retention within cells has potential therapeutic advantages in anterior ocular therapy. For example, immunomodulators/anti-inflammatories such as glucocorticoids, cyclosporine A, and tacrolimus are useful in the treatment of topical conditions such as uveitis, necrotizing scleritis, vernal keratoconjunctivitis, and thyroid ophthalmopathy.28−30 These compounds work through inhibiting intracellular protein signaling pathways within inflammatory cells.31,32 However, these compounds are poorly water-soluble,33 and so effective intracellular access through conventional topical applications, which involve release of the drug into the aqueous medium surrounding the cells, is suboptimal. A delivery approach that provides direct intracellular access could improve drug efficacy and potentially reduce dosing requirements.

■

REFERENCES

(1) Gaudana, R.; Jwala, J.; Boddu, S. H. S.; Mitra, A. K. Pharm. Res. 2009, 26, 1197−1216. (2) Urtti, A. Adv. Drug Delivery Rev. 2006, 58, 1131−1135. (3) Agrawal, A. K.; Das, M.; Jain, S. Exp. Opin. Drug Delivery 2012, 9, 383−402. (4) Cao, Y.; Zhang, C.; Shen, W.; Cheng, Z.; Yu, L. L.; Ping, Q. J. Controlled Release 2007, 120, 186−194. (5) Lang, J. C. Adv. Drug Delivery Rev. 1995, 16, 39−43. (6) Chan, J.; Maghraby, G. M. M. E.; Craig, J. P.; Alany, R. G. Int. J. Pharm. 2007, 328, 65−71. (7) Nagarwal, R. C.; Kant, S.; Singh, P. N.; Maiti, P.; Pandit, J. K. J. Controlled Release 2009, 136, 2−13. (8) Civiale, C.; Licciardi, M.; Cavallaro, G.; Giammona, G.; Mazzone, M. G. Int. J. Pharm. 2009, 378, 177−186. (9) Di Tommaso, C.; Bourges, J.-L.; Valamanesh, F.; Trubitsyn, G.; Torriglia, A.; Jeanny, J.-C.; Behar-Cohen, F.; Gurny, R.; Möller, M. Eur. J. Pharm. Biopharm. 2012, 81, 257−264. (10) Shen, J.; Wang, Y.; Ping, Q.; Xiao, Y.; Huang, X. J. Controlled Release 2009, 137, 217−223. (11) Ludwig, A. Adv. Drug Delivery Rev. 2005, 57, 1595−1639. (12) Khutoryanskiy, V. V. Macromol. Biosci. 2010, 11, 748−764. (13) Shen, J.; Deng, Y.; Jin, X.; Ping, Q.; Su, Z.; Li, L. Int. J. Pharm. 2010, 402, 248−253. (14) Uyeda, H. T.; Medintz, I. L.; Jaiswal, J. K.; Simon, S. M.; Mattoussi, H. J. Am. Chem. Soc. 2005, 127, 3870−3878. (15) Moini, H.; Packer, L.; Saris, N.-E. L. Toxicol. Appl. Pharmacol. 2002, 182, 84−90. (16) Riddles, P. W.; Blakeley, R. L.; Zerner, B. Methods Enzymol. 1983, 91, 49−60. (17) Timbart, L.; Amsden, B. G. J. Polym. Sci., Part A-1: Polym. Chem. 2008, DOI: 10.1002/pola.23117. (18) Burns, J. A.; Butler, J. C.; Moran, J.; Whitesides, G. M. J. Org. Chem. 1991, 56, 2648−2650. (19) Kataoka, K.; Harada, A.; Nagasaki, Y. Adv. Drug Delivery Rev. 2012, 64, 37−48. (20) Lu, C.; Yoganathan, R. B.; Kociolek, M.; Allen, C. J. Pharm. Sci. 2012, 102, 627−637. (21) Kalarickal, N. C.; Rimmer, S.; Sarker, P.; Leroux, J.-C. Macromolecules 2007, 40, 1874−1880. (22) Howard, M. D.; Ponta, A.; Eckman, A.; Jay, M.; Bae, Y. Pharm. Res. 2011, 28, 2435−2446. (23) Greenspan, P.; Fowler, S. D. J. Lipid Res. 1985, 26, 781−789. (24) Treuel, L.; Jiang, X.; Nienhaus, G. U. J. R. Soc. Interface 2013, 10, 20120939. (25) Allen, C.; Yu, Y.; Eisenberg, A.; Maysinger, D. Biochim. Biophys. Acta 1999, 1421, 32−38. (26) Kim, Y.; Pourgholami, M. H.; Morris, D. L.; Lu, H.; Stenzel, M. H. Biomater. Sci. 2013, 1, 265−275. (27) Verma, A.; Stellacci, F. Small 2010, 6, 12−21. (28) Nussenblatt, R. B.; Palestine, A. G. Surv. Ophthalmol. 1986, 31, 159−169. (29) Hogan, A. C.; McAvoy, C. E.; Dick, A. D.; Lee, R. W. J. Ophthalmology 2007, 114, 1000−1006. (30) Cholkar, K.; Patel, S. P.; Vadlapudi, A. D.; Mitra, A. K. J. Ocul. Pharmacol. Ther. 2013, 29, 106−123. (31) Vichyanond, P.; Kosrirukvongs, P. Curr. Allergy Asthma Rep. 2013, 13, 308−314. (32) Ayroldi, E.; Cannarile, L.; Migliorati, G.; Nocentini, G.; Delfino, D. V.; Riccardi, C. FASEB J. 2012, 26, 4805−4820. (33) Budavari, S.; O’Neil, M.; Smith, A.; Heckelman, P.; Kinneary, J. The Merck Index, 12th ed.; Merck & Co., Inc.: Whitehouse Station, NJ, 1996.

■

CONCLUSIONS Micelles containing a bidentate dithiol-terminated PEG corona were effectively prepared using newly synthesized 6,8dimercaptooctanoate-poly(ethylene glycol)-b-poly(D,L-lactide). The release of dexamethasone from these micelles was unaffected by the presence of the dithiol group and below a concentration of 10 mg/mL, the micelles possessed low cytotoxicity. The micelles were rapidly endocytosed by human corneal epithelial cells at a rate that was independent of the presence of dithiol groups on the PEG corona. However, micelles possessing the bidentate dithiol groups were retained to a greater degree within the cytoplasm of these cells. These results suggest that this formulation approach may have potential for topical ocular drug delivery.

■

Article

AUTHOR INFORMATION

Notes

The authors declare no competing financial interest. 1354

dx.doi.org/10.1021/bm4018879 | Biomacromolecules 2014, 15, 1346−1354