Article pubs.acs.org/molecularpharmaceutics
Hydrophobically Modified Polymer/α-Cyclodextrin Thermoresponsive Hydrogels for Use in Ocular Drug Delivery Daisuke Iohara,†,‡ Masanori Okubo,†,‡ Makoto Anraku,† Shunji Uramatsu,§ Toshio Shimamoto,§ Kaneto Uekama,† and Fumitoshi Hirayama*,† †
Faculty of Pharmaceutical Sciences, Sojo University, 4-22-1 Ikeda, Nishi-ku, Kumamoto 860-0082, Japan Daido Chemical Corporation, 4-4-28 Takeshima, Nishiyodogawa-ku, Osaka 555-0011, Japan
§
S Supporting Information *
ABSTRACT: We report herein on the preparation of thermoresponsive hydrogels by taking advantage of the interaction of cyclodextrins (CDs) and a hydrophobically modified polymer. A hydrophobically modified hydroxypropyl methylcellulose (HM-HPMC) gel formed thermoresponsive hydrogels when small amounts of α-CD were added to the solution. The HM-HPMC/α-CD showed reversible sol−gel transition in the physiological temperature range that was completely opposite to the temperature dependency shown by the original HM-HPMC. The thermoresponsive gelation was attributed to the temperature dependency of the interaction between α-CD and the hydrophobic moiety of HM-HPMC. The potency of the HM-HPMC/α-CD sol−gel transition system in ophthalmic formulation was tested on the eyes of a rabbit. The use of HM-HPMC/α-CD significantly improved the ocular absorption of a drug, diclofenac sodium, by virtue of the rapid formation of a gel on the ocular surface. That is, the HM-HPMC/ α-CD was in a low viscous sol state at room temperature, which made administration easy, but it rapidly formed a viscous hydrogel on the ocular surface at physiological temperature. The thermoresponsive hydrogel based on the hydrophobically modified polymer and CD promises to have widespread applications in drug delivery. KEYWORDS: cyclodextrins, sol−gel transition, hydrophobically modified polymer, thermoresponsive hydrogel, ocular drug delivery %,13 which is costly and can cause an allergic reaction or discomfort when it is used in an ophthalmic formulation. Recently, a variety of hydrophobically modified polymers that consist of a water-soluble polymer bearing hydrophobic residues along the polymer backbone or at the ends of the chain have been developed by various researchers.14−18 These polymers are now widely used in medicine, food products, and the paint industry as thickeners because of their unique physicochemical properties that permit them to form threedimensional networks through intra- and intermolecular associations of the hydrophobic residues. The networks in such polymers are usually temperature-dependent, a factor that weakens the ability of the hydrophobic residues to associate with each other, resulting in a reduction in the viscosity of the solution. Interactions of polymer side chains with cyclodextrins (CDs) have attracted significant interest, because CDs encapsulate the hydrophobic moiety, thereby modifying the association of the polymers.19−24 Introducing CDs on the polymer side chains through covalent bonds amplifies their selectivity toward guest molecules, thus permitting increased intermolecular interactions between the polymer side chains.25
1. INTRODUCTION Ocular instillation is a conventional route for the administration of drugs used for treating ocular diseases because it is easily administered and noninvasive and it provides good compliance for patients. However, eye drops suffer from poor ocular absorption because the protective barrier of the cornea and the complex structure of the eyeball cause the drug to be rapidly eliminated.1,2 Generally, less than 5% of the administered drug can penetrate through the cornea and reach the intraocular tissue. Thus, improving the bioavailability of ocular drugs and delivering the drug to the targeted site have become important issues that need to be addressed. Various strategies including stimulus-responsive hydrogels,3−5 micro/nanoparticles,6−8 and lipid based nanocarrier9−11 have been developed and evaluated to overcome such an issue in the treatment of ocular disease. Among these strategies, thermoresponsive hydrogels are one example of clinically approved formulations. These gels are typically a low viscous sol at room temperature, but form a gel at physiological temperature, thus prolonging the retention of the drug on the eye, which improves the ocular bioavailability of the drug. The most commonly used materials that show such thermal response are block copolymers having the thermodynamic property of a lower critical solution temperature.12 However, thermal gelation generally occurs at higher polymer concentrations, for example, a poloxamer 407 solution undergoes thermal gelation at concentrations above 20 w/v © 2017 American Chemical Society
Received: Revised: Accepted: Published: 2740
April 10, 2017 June 28, 2017 June 29, 2017 June 29, 2017 DOI: 10.1021/acs.molpharmaceut.7b00291 Mol. Pharmaceutics 2017, 14, 2740−2748
Article
Molecular Pharmaceutics
samples for the dynamic rheology was controlled to 20, 25, 30, and 40 ± 0.01 °C. 2.4. NMR Measurements. 1H NMR measurements were performed with a Jeol JNM-ECA500 Delta spectrometer (Tokyo, Japan) operating at 500 MHz, using a 5 mm sample tube at 25 °C. Chemical shifts are given as parts per million (ppm) downfield from that of sodium trimethylsilylpropionate with an accuracy of 0.005 ppm. HM-HPMC or HPMC was dissolved in D2O at 0.05% w/v concentration with or without CDs (0.05% w/v ∼ 0.2% w/v) for the measurements. All spectra were acquired with 72 scan counts. 2.5. Interaction of CDs with 1-Octadecylamine Hydrochloride. Solubility measurements were carried out according to the method of Higuchi and Connors.36 Excess amounts (5 mg) of 1-octadecylamine hydrochloride (OCT) were added to a test tube containing various concentrations of CD solutions (1 mL) and shaken at 20, 30, and 40 °C for 5 days. After centrifugation (8000 rpm for 5 min), an aliquot (0.5 mL) of the supernatant was withdrawn, filtered through a 0.2 μm filter, and then diluted with water. The amount of OCT in the water was detected by HPLC (Shimazu UHPLC system Tokyo, Japan) with a refractive index detector and a YMC ODS-A column (5 μm, 4.6 mm × 15 cm). The mobile phase was composed of 80% acetonitrile containing 0.1% trifluoroacetic acid, and the flow rate was 1.0 mL/min. The apparent 1:1 stability constant of CD/OCT was calculated by the equation of Kc = slope/ [intercept (1 − slope)] using slopes and intercepts of the initial straight-line portion of the phase solubility diagrams. The Kc value was plotted as a function of 1/T, and enthalpies (ΔH) were determined from the slope of a van’t Hoff plot. Entropies (ΔS) were calculated from the experimentally determined Kc and ΔH. 2.6. In Vitro Release Profile of Diclofenac Sodium (DCFNa) from HM-HPMC/α-CD. The DCFNa solution was extensively mixed with the HM-HPMC or HM-HPMC/α-CD, and the final concentrations of each component were fixed at 0.5% w/v for HM-HPMC, 0.04% w/v for α-CD, and 0.1% w/v for DCFNa, respectively. In vitro release behavior of DCFNa from the HM-HPMC/α-CD gel was studied by using a dialysis membrane. Briefly, 5 mL of HM-HPMC/α-CD containing 0.1% w/v DCFNa was placed in a dialysis membrane (molecular weight cutoff: 10000) through which the drug (MW 296) and α-CD (MW 972) can easily pass, but the HMHPMC (with a high molecular weight MW of 400,000), cannot. The membrane was immersed in 500 mL of PBS at 32 °C, and the outer medium was stirred by a paddle at 50 rpm. The temperature of the corneal surface is 34.5 ± 0.8 °C;37 thus we conducted the drug release studies at 32 °C, a temperature that is slightly lower than that of the corneal surface, because, in an actual situation, the temperature of the instilled material would be expected to be lower compared to that of corneal surface. At predetermined intervals, 1.0 mL samples were withdrawn from the outer medium and replaced with an equal volume of fresh PBS. The concentration of DCFNa was determined by UV measurement (U-2800A Hitachi, Japan) at 275 nm. The amount (%) of drug released was calculated by determining the amount of all of the drug released outside the membrane. 2.7. Instillation of HM-HPMC/α-CD Containing DCFNa on the Rabbit’s Eye. HM-HPMC/α-CD (0.5% w/v HMHPMC and 0.04% w/v α-CD) containing 0.1% w/v DCFNa was prepared with phosphate buffered saline (PBS) for the in vivo ocular penetration study. Male Japanese white rabbits (8
By introducing functional groups at the polymer side chains it is possible to prepare CD-based supramolecular gels that show a stimulus-responsiveness to pH,26,27 light,28,29 temperature,30 or oxidizing/reducing agents.31,32 However, the issue of whether it is possible to change the thermoresponsive properties of the polymers by simply adding CDs to hydrophobically modified polymers has not been investigated, and pharmaceutical applications of such CDs/hydrophobically modified polymer systems for delivering a drug to an ocular system have not been evaluated. The process of the inclusion by CDs is generally temperature-dependent and is likely to occur at a lower temperature, whereas dissociation would occur at higher temperatures.33−35 Thus, we hypothesized that thermal response of a hydrophobically modified polymer might be altered by utilizing the interaction of CDs and a hydrophobically modified polymer. In this paper, we report, for the first time, on reversing the change in the temperature-dependent viscosity of an elastic hydrophobically modified hydroxypropylmethyl cellulose (HMHPMC) by the addition of CDs. The HM-HPMC/CDs gel showed a reversible thermoresponsive sol−gel transition at low polymer concentration (0.5% w/v for polymer and 0.01% w/v for α-CD), and the mechanism responsible for the thermal gelation of HM-HPMC/CDs was studied by focusing on the interaction of CD and hydrophobic moiety of the polymer. In addition, the potency of the thermoresponsive sol−gel transition system in ocular drug delivery was tested on the eyes of a rabbit.
2. EXPERIMENTAL SECTION 2.1. Materials. Hydrophobically modified hydroxypropylmethyl cellulose (HM-HPMC) with the commercial name Sangelose 60L and HPMC (METOLOSE60SH-10000) were supplied by Daido Chemical Co. (Osaka, Japan). HM-HPMC contains a stearyl group (C18) at the hydroxypropyl ends with a degree of substitution of 0.3−0.6 wt % and has a MW of 400,000. The degree of substitution of the hydroxypropoxy and methoxy groups were 7−11 wt % and 27−30 wt %, respectively. α-, β-, and γ-CD were supplied from Nihon Syokuhin Kako Co. (Tokyo, Japan). Diclofenac sodium (DCFNa) was purchased from Wako Pure Chemical Co. (Kyoto, Japan). DCFNa commercial product, DICLOD ophthalmic solution 0.1%, was purchased from Wakamoto Pharmaceutical Co. All other reagents were of analytical grade, and deionized double-distilled water was used throughout the study. 2.2. Preparation of HM-HPMC/CD Hydrogels. HMHPMC was dispersed in hot water or phosphate buffered saline (PBS) with agitating, and then cooled down to dissolve the HM-HPMC. Various amount of α-, β-, and γ-CD were added in the HM-HPMC hydrogels and allowed to equilibrate for at least 24 h at 4 °C before the studies. 2.3. Rheological Studies. The rheological measurements of the HM-HPMC solutions were performed with a MCR-101 rheometer (Anton Paar Japan K.K., Tokyo, Japan). A cone and plate geometry with a diameter of 25 mm and a 0.998 rad cone angle was used. Steady shear rheology measurements were performed at 25 °C increasing the shear rate from 0.1 to 100 s−1. The viscosity measurements on the dynamic temperature were performed with the heating rate of 3 °C/min at shear rate of 1 s−1. Dynamic frequency data were obtained in the linear viscoelastic regime for the samples that was determined by the dynamic strain sweep measurements. The temperature of 2741
DOI: 10.1021/acs.molpharmaceut.7b00291 Mol. Pharmaceutics 2017, 14, 2740−2748
Article
Molecular Pharmaceutics
Figure 1. Effect of CDs on the viscosity of 0.5% w/v HM-HPMC. The viscosity was plotted as a function of shear rate (a−c), and the zero shear viscosity was plotted as a function of CD concentrations (d). Each point represents the mean ± SE of 3 experiments.
± SE. Kruskal−Wallis analysis of variance (ANOVA) was performed followed by Dunn’s post hoc tests for multiple comparisons. For all analyses, values of p < 0.05 were regarded as statistically significant.
week old) were used for the experiments. The care and maintenance of animals was in accordance with the institutional guidelines of the Institutional Animal Care and Use Committee of Sojo University. Fifty microliters of HM-HPMC/α-CD formulation or commercial product, DICLOD ophthalmic solution 0.1%, was instilled onto the lower conjunctival sac of the rabbit. After 0.5 h, 1 h, 2 h, and 4 h of instillation, the rabbits were sacrificed and the surface of the eyeball was washed with 10 mL of saline. Aqueous humor was collected using a 1 mL syringe with a 26G needle, and the eyeball was then removed to detect the drug in the cornea. 2.8. Detection of DCFNa in the Aqueous Humor and Cornea. Fifty microliters of the aqueous humor was mixed with 500 μL of acetonitrile and vortexed for 5 min. The solution was centrifuged at 15000 rpm for 5 min to remove proteins, the supernatant was dried, and the residue was dissolved in the HPLC mobile phase. The amount of DCFNa in the aqueous humor was detected by a Shimadzu UHPLC system (Tokyo, Japan) consisting of a pump, autoinjector, UV−vis detector, and a YMC ODS-AM column (5 μm, 4.6 mm × 25 cm, Tokyo, Japan). The mobile phase was a mixture of 60% acetonitrile containing 0.1% trifluoroacetic acid. The flow rate was 1.0 mL/min, and DCFNa was detected at a wavelength of 275 nm. Corneas were removed from the frozen eyeball and weighed. The cornea was homogenized in 2 mL of PBS, and 2 mL of acetonitrile was then added to remove proteins. The sample was vortexed and centrifuged at 1000g for 15 min. The supernatant was dried, and the residue was analyzed by HPLC after being dissolved in the HPLC mobile phase, in the same manner as was used for the aqueous humor samples. The pharmacokinetics parameters were calculated by using the Practical Pharmacokinetic Program (MULTI), a normal leastsquares program.38 2.9. Statistical Analysis. Data are presented as the median values from n samples, and the results are reported as the mean
3. RESULTS AND DISCUSSION 3.1. Effect of CDs on the Viscosity of HM-HPMC Hydrogel. We used hydrophobically modified hydroxypropyl
Figure 2. Dynamic rheology of HM-HPMC in the presence and absence of α-CD (0.04 w/v %) at 25 °C. Each point represents the mean ± SE of 3 experiments.
methylcellulose (HM-HPMC, MW 400,000) for the base of the hydrogel, which contains stearyl groups at the hydroxypropyl ends. HM-HPMC is now widely used in topical formulations as a thickener.39,40 The HM-HPMC showed a high viscosity, even 2742
DOI: 10.1021/acs.molpharmaceut.7b00291 Mol. Pharmaceutics 2017, 14, 2740−2748
Article
Molecular Pharmaceutics
Figure 3. Temperature-dependent viscosity change of HM-HPMC (0.5%% w/v) with or without CDs, measured at a shear rate of 1 s−1. Each point represents the mean ± SE of 3 experiments.
Figure 4. Reversed viscosity change of HM-HPMC and HM-HPMC (0.5% w/v)/α-CD (0.01% w/v) hydrogel upon heating and cooling cycles measured at a shear rate of 1 s−1.
at low concentrations; a 0.5% w/v polymer solution showed ca. 40,000 mPa·s at the Newtonian plateau. The high viscosity of the HM-HPMC solution decreased with shear rate, showing quasi viscous flow properties. On the other hand, the high viscosity could be markedly decreased by the addition of small amounts of α-CD, as shown in Figure 1a. The viscosity of the HM-HPMC also decreased on the addition of other β- and γCDs in a concentration-dependent manner (Figure 1b,c). The effect of CDs on the viscosity of HM-HPMC was evaluated by comparing the zero shear viscosity of the samples (Figure 1d). The decreasing effect of α-CD was greater than the corresponding values for β- and γ-CDs. The initial viscosity of 40,000 mPa·s dropped to around 10,000 mPa·s by the addition of 0.01% w/v α-CD, and the value decreased to 4000
Figure 5. Loss tangents (tan δ) of HM-HPMC (0.5% w/v) and HMHPMC (0.5% w/v)/α-CD (0.04% w/v) hydrogel at 20, 30, and 40 °C.
2743
DOI: 10.1021/acs.molpharmaceut.7b00291 Mol. Pharmaceutics 2017, 14, 2740−2748
Article
Molecular Pharmaceutics
Figure 7. Proposed mechanism for the thermoresponsive reversible sol−gel transition of HM-HPMC/CD hydrogel. Figure 6. 1H NMR spectra of HM-HPMC (0.05% w/v) with or without CDs (0.05−0.2% w/v) at 25 °C.
mPa·s when 0.06% w/v α-CD was added. γ-CD was less effective to the viscosity of the HM-HPMC, showing a high value of 10,000 mPa·s even in the presence of 0.15% w/v γ-CD. No appreciable change in viscosity was observed when glucose was added to the solution (data not shown). Figure 2 shows storage modulus G′ and loss modulus G″ values for the preparation as a function of the angular frequency ω. In the case of HM-HPMC alone, the value for G′ exceeded that for G″ over the entire range of frequencies, and both moduli were frequency independent, consistent with the rheological signature of a gel.41,42 In sharp contrast, the solution showed sol behavior in the presence of α-CD, with both moduli being dependent on frequency, and the value for G″ exceeded that for G′ in the high frequency range. These rheological responses are indicative of the transition from a gel to a sol as a result of the addition of α-CD. The transition from a gel to a low viscous sol was confirmed by its appearance (Figure 2 inset). SEM images indicated that the porous size of HM-HPMC became wider by adding α-CD (Figure S1). 3.2. Thermoresponsive Reversible Sol−Gel Transition of the HM-HPMC/CD Hydrogel. The viscosity change of the HM-HPMC hydrogel on the dynamic temperature was studied by rheological measurements (Figure 3). The viscosity of HMHPMC alone decreased with increasing temperature from 20 to 50 °C as would be expected for a typical elastic polymer solution. On the other hand, in the presence of α-CD, the viscosity increased with temperature and asymptotically approached the curve for HM-HPMC alone, showing a viscosity change with a convex curve. The temperaturedependent change in viscosity was clear at a 0.01% w/v concentration of α-CD, and the maximum peak viscosity was around 33 °C, which is acceptable for applications regarding
Figure 8. Release profile of DCFNa from the HM-HPMC/α-CD formulation in PBS at 32 °C. All formulations contain 0.1% w/v DCFNa. All the symbols were overlapped at the point of 24 h. Each point represents the mean ± SE of 3 experiments.
biological systems, as described below. The maximum peak shifted to higher temperature when the higher amounts of CD were added (43 °C at 0.04% w/v α-CD, 46 °C at 0.1% w/v αCD, and 49 °C at 0.2% w/v α-CD). Similar viscosity curves were observed for the β-CD system, showing the maximum peak around 27, 35, 42, and 47 °C for 0.01, 0.04, 0.1, and 0.2% w/v β-CD respectively. In the case of the γ-CD system, the viscosity change on the dynamic temperature was slight and a clear maximum peak was observed only for 0.2% w/v γ-CD around 30 °C. The reason for the different temperature response of the HM-HPMC/CD hydrogels is discussed below. Interestingly, the viscosity change of the HM-HPMC/CD system was reversible upon heating and cooling cycles (Figure 4). The HM-HPMC/α-CD system showed a low viscosity of around 2000 mPa·s at 20 °C, while the viscosity was twice as high at 35 °C, equivalent to the viscosity of HM-HPMC alone. The viscosity change occurred within a few minutes after heating or cooling the sample. No significant difference in viscosity was observed at each measuring point after heating or
Table 1. Apparent 1:1 Stability Constants (Kc) and Thermodynamic Parameters of OCT/CD Complexes in Water at Different Temperaturesa Kc (M−1)
a
complex
20 °C
30 °C
40 °C
ΔH (kJ/mol)
ΔS (kJ/mol·K)
α-CD β-CD γ-CD
79400 ± 900 61400 ± 700 22700 ± 200
10100 ± 90 8800 ± 50 5200 ± 30
2000 ± 40 1700 ± 10 1000 ± 4
−141 −136 −120
−0.39 −0.37 −0.32
Each value represents the mean ± SE of 3 experiments. 2744
DOI: 10.1021/acs.molpharmaceut.7b00291 Mol. Pharmaceutics 2017, 14, 2740−2748
Article
Molecular Pharmaceutics
be expected to decrease the viscosity of the preparation. The αCD probably incorporates the alkyl chain inside its small cavity, which would lead to a significant decrease in viscosity, while the large cavity of γ-CD would hold such residues more loosely, thus resulting in a small decrease in viscosity as shown in Figure 1. The interaction between the stearyl moiety of HM-HPMC and the CDs was confirmed by the 1H NMR (Figure 6). In the absence or presence of lower CD concentrations, only one broad peak, assigned to the methyl proton of the hydroxypropyl moiety, was observed at 1.06 ppm, due to the restricted motion of the moiety in the viscous solution. On the other hand, small peaks were observed at around 1.2 ppm after the addition of 0.2% w/v α-CD, which were assigned to the alkyl group of the stearyl moiety. The restriction of the stearyl moiety was recovered by the inclusion of α-CD, thus allowing these peaks to appear in the NMR spectra. In the case of β-CD, the peak at around 1.2 ppm was weaker and no peak changes were observed for γ-CD under the experimental conditions used. 1H NMR and viscosity measurements were also conducted for a HPMC preparation without stearyl groups, in the presence and absence of CDs. As expected, these CDs had no effect on the NMR peaks or the viscosity of HPMC solution (Figure S2). These results suggest that α-CD interacts strongly with stearyl moieties of HM-HPMC, compared with β- and γ-CDs. Such an inclusion process is generally temperature-dependent and is likely to occur at a lower temperature, whereas dissociation would occur at higher temperatures.33−35 The stability constant (Kc) of the stearyl moiety with CDs was determined by the solubility method at 20, 30, and 40 °C (Figure S3). 1Octadecylamine hydrochloride (OCT) was used instead of HM-HPMC in order to obtain a clear Kc value independent of the viscosity issue. The solubility of the OCT increased linearly with CD concentrations under the experimental conditions used, showing an AL type solubility diagram, however the plot deviated from the straight line (Ap type diagram) for α- and βCD at 20 °C, indicating higher order complex formation at higher CD concentrations. The Kc of the complexes and thermodynamic parameters for complexation are listed in Table 1. It was apparent that the Kc value for the α-CD complex was higher than the corresponding values for the β- and γ-CD complexes and the Kc value was temperature-dependent. Complexation with the CDs was driven by changes in enthalpy (ΔH) and entropy (ΔS). Comparing the ΔH and ΔS, ΔH appears to be the dominant contributor for complex formation. Figure 7 shows a proposed mechanism for the thermoresponsive sol−gel transition of HM-HPMC/CD hydrogel. The elastic HM-HPMC turned into a low viscous sol as a result of the interaction with CDs at the stearyl end. Heating the HMHPMC/CDs induced the dissociation of CDs from the stearyl moieties, thus the viscosity of HM-HPMC/CD hydrogel increased with temperature showing the maximum peak, where the CDs completely dissociated from the stearyl moieties (Figure 3). Further heating weakens the intermolecular association between the hydrophobic moieties of HM-
Figure 9. Concentrations of DCFNa in rabbit aqueous humor (upper) and cornea (lower) at various time points after administering a HMHPMC/α-CD formulation or a commercial product to the eyes of a rabbit. Both formulations contain 0.1% w/v DCFNa. Each point represents the mean ± SE of 3 experiments. *, p < 0.05 versus a commercial product.
cooling over repeated cycles. HM-HPMC alone showed a high viscosity of around 8000 mPa·s at 20 °C, while the viscosity dropped by half to 4000 mPa·s by heating to 35 °C. The change in viscosity with temperature was assessed using dynamic rheology (Figure 5). HM-HPMC showed a tan δ (G″/G′) of 1 at 20 °C, while the value was
