Langmuir 2006, 22, 3945-3949
3945
Controlled Release of Model Drugs through a Molecular Recognition Ion Gating Membrane in Response to a Specific Ion Signal Taichi Ito† and Takeo Yamaguchi* Department of Chemical System Engineering, The UniVersity of Tokyo, 7-3-1 Hongo, Bunkyo-ku Tokyo, 113-8656 Japan ReceiVed NoVember 28, 2005. In Final Form: January 21, 2006 A controlled-release device that responds to a specific molecular signal is an ideal goal in drug delivery and tissue engineering. A molecular recognition ion gating membrane, in which a copolymer of N-isopropylacrylamide and benzo[18]-crown-6-acrylamide was grafted onto the surface of the porous polyethylene film, was used to control the permeability of vitamin B12 and lysozyme in response to a specific ion signal. The observed response depended on the amount of grafted copolymer. When the grafting ratio was below 15%, the membrane pores opened by Ca2+ and closed by Ba2+. The permeability of model drugs became higher by opening of the pores. On the other hand, when the grafting ratio was above 15%, the properties of the membrane changed. The permeability of model drugs became lower by Ca2+ due to dehydration of the grafted copolymer. The opposite responses were observed at different grafting ratios.
Introduction The technology to deliver drugs, hormones, and growth factors in response to a specific molecular signal is extremely important. In tissue engineering,1 the induction of cell differentiation requires the proper delivery of growth factors in response to the cell differentiation process. In drug delivery engineering,2 the controlled-release kinetics in response to a molecular signal from given cells can contribute to enabling an optimally minimal dosage of a drug to be delivered. Hydrogels are known to exhibit a volume phase transition in response to an external stimulus, such as temperature,3 ion,4 glucose,5 antigen,6,7 pH,8 and light.9 Hydrogels are a promising candidate for delivery devices. Poly-N-isopropylacrylamide (NIPAM) changes its volume reversibly at 32 °C.3 Hoffman et al.,10 Okano et al.,11 Feil et al.,12 and Palasis et al.13 have reported on the thermosensitive release of drugs from NIPAM hydrogels. The release kinetics of these drugs depended on the hydrophobicity and hydrophilicity of the drug13 and also on the shape, size, and water content of the hydrogel.12,13 Both the solubility of a drug in a hydrogel and the diffusivity of a drug through a hydrogel are changed at lower * To whom correspondence should be addressed. Tel: +81-3-5841-7345. Fax: +81-3-5841-7227. E-mail:
[email protected]. † Present address: Department of Chemical Engineering, Massachusetts Institute of Technology, 77 Massachusetts Avenue, Massachusetts 02142. (1) Langer, R.; Vacanti, J. P. Science 1993, 260 (5110), 920-926. (2) Anderson, D. G.; Burdick, J. A.; Langer, R. Science 2004, 305 (5692), 1923-1924. (3) Hirokawa, Y.; Tanaka, T. J. Chem. Phys. 1984, 81, 6379-6380. (4) Irie, M.; Misumi, Y.; Tanaka, T. Polymer 1993, 34, 4531-4535. (5) Kataoka, K.; Miyazaki, H.; Okano, T.; Sakurai, Y. Macromolecules 1994, 27, 1061-1062. (6) Miyata, T.; Asami, N.; Uragami, T. Nature 1999, 399, 766-769. (7) Lu, Z. R.; Kopeckova, P.; Kopecek, J. Macromol. Biosci. 2003, 3, 296300. (8) Annaka, M.; Tanaka, T. Nature 1992, 355, 430-432. (9) Sumaru, K.; Kameda, M.; Kanamori, T.; Shinbo, T. Macromolecules 2004, 37, 4949-4955. (10) Hoffman, A. S.; Afrassiabi, A.; Dong, L. C. J. Controlled Release 1986, 4, 213-222. (11) Okano, T.; Bae, Y. H.; Jacobs, H.; Kim, S. W. J. Controlled Release 1990, 11, 255-265. (12) Feil, H.; Bae, Y. H.; Feijen, J.; Kim, S. W. J. Membrane Sci. 1991, 64, 283-294. (13) Palasis, M.; Gehrke, S. H. J. Controlled Release 1992, 18, 1-11.
critical solution temperatures (LCST). Peng et al.14 and Chu et al.15,16 grafted poly-NIPAM onto the surface of a porous film and a microcapsule and showed the rapid response of drug diffusivity to changes in temperature. The drug release kinetics with changes in temperature depended on the degree of NIPAMgrafting. A copolymer of NIPAM and benzo[18]-crown-6-acrylamide (BCAm) has been reported to show a phase separation in response to specific ions, such as the potassium ion.4 We grafted this copolymer onto the surface of a porous polyethylene film and developed a molecular recognition ion gating membrane.17,18,19 This membrane controlled pressure-driven flow,17 solute rejection,18 and osmotic pressure19 in response to a specific ion signal, such as K+ or Ba2+. In this paper, we report on the controlled release function of this molecular recognition ion gating membrane. The membrane controlled the permeability of a drug in response to a specific ion signal, as shown in Figure 1. In addition, our gating membrane was a pore-filling type membrane,20 which is different from a cross-linked hydrogel, and thus, the swelling of the grafted NIPAM and BCAm copolymer was suppressed so that the water content of the grafted polymer was kept at a low value. Therefore, its response properties were unique compared with cross-linked hydrogel, and were dependent on the amount of grafted polymer. Experimental Section Membrane Preparation. The membrane was prepared using the peroxide plasma graft polymerization method, as described in an earlier publication.17,18 A porous cross-linked polyethylene (CLPE) film (Nitto-Denko, Japan, pore size ) 0.2 µm, thickness ) 20 µm, porosity ) 40%) was used as a substrate, and the plasma treatment power and time were 30 W and 1 min, respectively. Various graft (14) Peng, T.; Cheng, Y. L. J. Appl. Polym. Sci. 1998, 70, 2133-2142. (15) Chu, L. Y.; Park, S. H.; Yamaguchi, T.; Nakao, S. J. Membrane Sci. 2001, 192, 27-39. (16) Chu, L. Y.; Yamaguchi, T.; Nakao, S. AdV. Mater. 2002, 14, 386-389. (17) Yamaguchi, T.; Ito, T.; Sato, T.; Shinbo, T.; Nakao, S. J. Am. Chem. Soc. 1999, 121, 4078-4079. (18) Ito, T.; Hioki, T.; Yamaguchi, T.; Shinbo, T.; Nakao, S.; Kimura, S. J. Am. Chem. Soc. 2002, 124, 7840-7846. (19) Ito, T.; Yamaguchi, T. J. Am. Chem. Soc. 2004, 126, 6202-6203. (20) Yamaguchi, T.; Nakao, S.; Kimura, S. Macromolecules 1991, 24, 55225527.
10.1021/la053206s CCC: $33.50 © 2006 American Chemical Society Published on Web 02/28/2006
3946 Langmuir, Vol. 22, No. 8, 2006
Ito and Yamaguchi
Figure 3. FT-IR spectra of molecular recognition ion gating membranes, with various grafting ratios: (a) substrate (0%), (b) 13.8%, (c) 68.8%, and (d) 177.7%.
Figure 1. Schematic representation of diffusion control in response to an ion signal across a molecular recognition ion gating membrane, for (a) a low grafting ratio and (b) a high grafting ratio.
was set in an incubator at 41.5 °C. A volume of 170 mL of a mixed aqueous solution of the selected drug and signal ion was poured into the feed chamber, and 170 mL of the aqueous signal ion solution was poured into the permeation chamber. Both chambers were stirred at 400 rpm to prevent any concentration polarization effect. Vitamin B12 (Mw ) 1355.38) and lysozyme (Mw ) 14 600) were used as the model drugs. The concentration of the model drug was measured using UV-vis spectroscopy (Hitachi, Japan, model U3310). The permeability of the model drug was calculated using the following equation:21 ln ∆C )
V1 + V2 PA t, ∆C ) Cf - Cp V 1V 2 L
(1)
where P is permeability, V1 and V2 are the volume of the aqueous solution in the feed chamber and permeation chamber, respectively, Cf and Cp are the concentration of the model drug in the feed chamber and in the permeation chamber, respectively, and L, A, and t are the membrane thickness, membrane area, and time, respectively. The diffusivity was determined using the time-lag method as follows: D)
L2 6θ
(2)
where D is the diffusivity, and θ is the lag time. The partition coefficient was determined based on the solutiondiffusion model as follows: P ) KD Figure 2. Schematic diagram of: (a) filtration experiments, and (b) dialysis experiments. Key: 1 ) pump; 2 ) pressure gauge; 3 ) thermometer; 4 ) test cell; 5 ) permeation solution; 6 ) flow meter; 7 ) pressure valve; 8 ) feed tank; 9 ) thermostat; 10 ) stirrer; 11 ) stirrer chip; 12 ) membrane; and 13 ) rubber plug. polymerizations onto the pore surface were attempted, and the grafting ratio was defined as the volume of the grafted polymer divided by the pore volume of the substrate. Morphological Analysis. The polymerization was confirmed using Fourier transform infrared analysis (FT-IR, Nicolet, USA, Model MAGNA550). The surface of the prepared membrane was observed using a scanning electron microscope (SEM, Hitachi, Japan, model S-900). The membrane area and thickness were measured both in the dry and the wet state, where the membranes were immersed in pure water at 43.5 and 21.5 °C. The volume flux of pure water from the pressure-driven flow, Lp, was calculated from filtration experiments at a filtration pressure of 1 kgf/cm2 at 40.5 and 22.8 °C using the apparatus shown in Figure 2a. Permeability Measurements. A side-by-side chamber cell was used to measure the permeability, as shown in Figure 2b. The cell
(3)
where K is the partition coefficient.
Results and Discussion Membrane Morphology. The membrane morphology changed with grafting ratio, and the membrane morphology was closely related to the transport properties of the membrane. Thus, in our discussion, we will focus on the relationship between the membrane morphology and the grafting ratio. Figure 3 shows the FT-IR absorbance of the membranes prepared using plasma graft polymerization. The grafting ratio of each membrane was 0% (substrate), 13.8%, 68.8%, and 177.7%. The ratio of the peak height of the NIPAM absorbance at 1647 cm-1 to the PE absorbance at 1460 cm-1 increased with increasing grafting ratio. It seems strange that a value of the grafting ratio over 100% was obtained. However, in plasma graft polymerization, the grafted polymer can grow outside the pore and can extend to the membrane thickness, as discussed below. (21) Yasuda, H.; Lamaze, C. E. J. Macromol. Sci. Phys. 1971, B 5, 111.
Controlled Release of Model Drugs
Langmuir, Vol. 22, No. 8, 2006 3947
Figure 4. FE-SEM photographs of molecular recognition ion gating membranes, with various grafting ratios.
Figure 5. Relationship between the permeability of pure water measured using pressure-driven filtration at 40.5 and 22.8 °C. The grafting ratio varied from 0% to 14.1%.
Figure 4 shows SEM pictures of membranes with grafting ratios of 0% (substrate), 15.4%, and 49.7%. The pore size in the dry state decreased with increasing grafting ratio. At a grafting ratio of about 50%, no large pores were observed. Figure 5 shows the relationship between volume flux of the pressure-driven flow and the grafting ratio. The grafted copolymer is poly-NIPAN-co-BCAm, and this copolymer shrinks at 40 °C and swells at 20 °C. Therefore, below a grafting ratio of 15%, the volume flux of pure water changed in response to changes in temperature. However, its change in flux was smaller than observed previously.17,18 This was because the distribution in pore size of our present CLPE substrate was larger than that of the porous substrate used in our previous work. On the other hand, the flux was too small to be measured at both low and high temperatures at grafting ratios above 15%. Hence, the pores of the membrane were fully closed at a grafting ratio of about 15% in the wet state due to the swelling of the grafted copolymer, although pores were observed in the dry state, as shown by the SEM photograph in Figure 4. Figure 6 shows the relationship between the area and thickness of a membrane and its grafting ratio. Poly-NIPAM hydrogel is known to change its volume and water content dramatically. However, the mechanical strength of the CLPE substrates was strong enough to suppress any swelling of the grafted polymer below grafting ratios of 70%. Accordingly, the membrane area and thickness were constant with different grafting ratios. On the other hand, the membrane thickness increased above grafting ratios of 70%. On the basis of our previous study using another substrate,20 we think that the grafted polymer grew on the surface of the membrane after the pores had filled with grafted polymer. However, there is no difference of the membrane thickness between at dry state and wet state. Thus, further study will be needed to make clear the structure of the membrane above grafting ratios of 70%. As mentioned later, the maximum grafting ratio of the membrane which used for permeability measurements was 68.8%. Therefore, almost all of the entire grafted polymer existed inside the pores at least in the case of the membrane used for permeability measurements.
Figure 6. Thickness and area of a molecular recognition ion gating membrane, which was measured under dry and wet conditions at 21.5 and 43.5 °C: (a) membrane thickness and (b) membrane area.
Figure 7. Lysozyme permeability through a molecular recognition ion gating membrane, with various grafting ratios at 41.0 °C. Aqueous 0.1 M BaCl2 or CaCl2 were added to aqueous lysozyme solutions as signal ions.
Permeability Controlled by Pore Opening and Closing at Low Grafting Ratios. Figure 7 shows the permeability of lysozyme through a gating membrane in the presence of 0.1 M Ba2+ or Ca2+ ions. In the presence of Ca2+ ions, lysozyme permeated through the membrane as fast as it did with pure water. On the other hand, the permeability of lysozyme in the presence of Ba2+ ions was lower than that in the presence of Ca2+ ions. That is to say, when the grafted copolymer shrank in a solution containing Ca2+ ions, lysozyme passed through the opened pores. However, the grafted copolymer swelled in a solution containing Ba2+ ions, and the pores closed, and so the permeability of lysozyme decreased. At a grafting ratio of 15%, the permeability of lysozyme in solutions containing Ba2+ and Ca2+ ions was so small that it could not be measured. With
3948 Langmuir, Vol. 22, No. 8, 2006
Figure 8. Permeability of vitamin B12 through a molecular recognition ion gating membrane, with various grafting ratios at 41.0 °C. Aqueous 0.1 M BaCl2 or CaCl2 were added to aqueous vitamin B12 solutions as signal ions.
Figure 9. Release kinetics of vitamin B12 from a molecular recognition ion gating membrane, with a grafting ratio of 68.8% at 41.0 °C. Aqueous 0.1 M BaCl2 or CaCl2 were added to aqueous vitamin B12 solutions as signal ions.
higher grafting ratios, the pores became filled with grafted copolymer both in the swollen and shrunken states. However, the smaller molecules tested did not show clear difference between the two ions such as lysozyme. Figure 8 shows the permeability of vitamin B12 through the gating membrane in the presence of 0.1 M Ba2+ or Ca2+ ions. Below a grafting ratio of 15%, the permeability of vitamin B12 in a solution containing Ca2+ ions was almost same with that in a solution containing Ba2+ ions. This was because the distribution in pore size of the substrate CLPE porous film was wide, and a more dynamic difference can be achieved using a homogeneous pore size substrate film. Permeability Controlled by Changes in Solubility at High Grafting Ratios. The permeability response through the membrane to the ion signal changed at high grafting ratios. Figure 9 shows the example of time courses of Cp, and in this case the membrane whose grafting ratio was 68.8%. In opposite behavior to that with the low grafting ratio membrane, the permeation through the high grafting ratio membrane in a solution containing Ba2+ ions was faster than that of a solution containing Ca2+ ions. In addition, the permeability of vitamin B12 of this membrane was calculated as 1.03 × 10-15 m2/s in the solution containing Ba2+ ions and 9.88 × 10-17 m2/s in the solution containing Ca2+ ions from the release kinetics data of Figure 9 using eq 1. The same measurements and calculations were performed using gating membranes, with grafting ratios of 4.7, 9.1, 13.8, 14.1, 29.8, and 40.4%, and summarized in Figure 8. Figure 8 shows that the permeability of vitamin B12 in a solution containing Ba2+ ions was higher than that of a solution containing Ca2+ ions. From the pressure-driven pure water volume flux as shown in Figure
Ito and Yamaguchi
Figure 10. Diffusivity of vitamin B12 through a molecular recognition ion gating membrane, with various grafting ratios at 41.0 °C. The diffusivity was measured using the time-lag method.
Figure 11. Partition coefficient of vitamin B12 through a molecular recognition ion gating membrane, with various grafting ratios at 41.0 °C. The partition coefficient was determined based on the solution-diffusion model.
5, there apparently were no pores that were filled with the grafted copolymer in either the swollen state in a solution containing Ca2+ ions or the shrunken state in a solution containing Ba2+ ions. In addition, as can be observed from the SEM data, there were no pores in the dry state at a grafting ratio of 49.7%. Hence, pore opening and closing functions did not occur and could not control the permeability. Instead of being controlled by changes in pore size, the permeability response was controlled by either the solubility of the drug compared to the grafted copolymer, or the diffusivity of the drug through the insides of the grafted copolymer pores. Figures 10 and 11 show the diffusivity and partition coefficient, respectively, of vitamin B12 through the membrane. The diffusivity in a solution containing Ca2+ ions was higher than that in a solution containing Ba2+ ions. On the other hand, the partition coefficient of a solution containing Ba2+ ions was higher than that of a solution containing Ca2+ ions. Being a function of the product of these two effects, the permeability of a drug through the membrane was mainly controlled by the change in partition coefficient, as shown in Figure 11. Difference in Response Mechanisms of a Hydrogel and a Gating Membrane. Palasis et al.13 reported that the partition coefficient of hydrophilic model drugs, such as vitamin B12, of poly-NIPAM hydrogels changed at the LCST; when the hydrogel was in the swollen state below the LCST, the partition coefficient was larger than when the hydrogel was in the shrunken state above the LCST. This tendency is in agreement with our results. When the grafted polymer was in a swollen state in a solution containing Ba2+ ions, the partition coefficient was larger than when the grafted polymer was in the shrunken state in a solution containing Ca2+ ions.
Controlled Release of Model Drugs
On the other hand, there exists a large structural difference between the cross-linked hydrogel membrane and our fillingtype gating membrane. In the case of our work, the swelling of the grafted copolymer inside the pores was suppressed by the CLPE substrate,20 and thus the water content of the polymer was extremely low. The membrane area and thickness remained almost constant, as shown in Figure 6. In the case of a cross-linked hydrogel,13 the water content changes with swelling and shrinking of the hydrogel. However, regardless of a constant water content, the degree of hydration of the grafted copolymer changes on capture of a signal ion by the crown ether receptors contained in the grafted copolymer.21 That is to say, because of the suppression of the grafted copolymer by the substrate, the grafted copolymer did not shrink or swell in reaction to the presence of the ion signal, but the degree of hydration from water molecules did change. As a result, the partition coefficient and the diffusivity of vitamin B12 changed in response to the ion signal.
Conclusions We have fabricated a gating membrane, capable of molecular recognition, that controlled permeability spontaneously in
Langmuir, Vol. 22, No. 8, 2006 3949
response to a specific ion signal. Its response mechanism depended on the grafting ratio. At low grafting ratios, the response was characterized by the opening and closing of the pores. At high grafting ratios, the response was determined by the change in the degree of hydration from water molecules. The control of the permeability of a drug through such a membrane is very important for a drug delivery system and for tissue engineering. Many signal transduction molecules exist and, in particular, the control of the permeability of low molecular weight molecules is very difficult using existing technologies. This type of membrane is expected to contribute to an advance in this area. Note Added after ASAP Publication. This article was published ASAP on February 28, 2006. A change has been made to equation 1. The correct version was posted on March 29, 2006. Acknowledgment. We thank the Nitto Denko Co. Ltd., Japan, for supplying the CLPE porous substrate, and the Kozin Co., Japan, for supplying the NIPAM. LA053206S
