Organo Hydrogel Hybrids. Formation of Reservoirs for Protein Delivery

Oct 5, 2005 - Fredrik Nederberg , Björn Atthoff , Tim Bowden , Ken Welch , Maria ... Håkan Rensmo, Tim Bowden, Jöns Hilborn, and Hans Siegbahn...
0 downloads 0 Views 209KB Size
Download PDF
Biomacromolecules 2005, 6, 3088-3094

3088

Organo Hydrogel Hybrids. Formation of Reservoirs for Protein Delivery Fredrik Nederberg,† Junji Watanabe,‡ Kazuhiko Ishihara,*,‡ Jo¨ns Hilborn,† and Tim Bowden*,† Department of Materials Chemistry, Polymer Chemistry, Box 538, Uppsala University, SE-751 21 Uppsala, Sweden, and Department of Materials Engineering, School of Engineering, The University of Tokyo, 7-3-1, Hongo, Bunkyo-ku, Tokyo 113-8656, Japan Received June 17, 2005; Revised Manuscript Received August 2, 2005

A biodegradable organo hydrogel hybrid material is presented, which is formed through the water uptake of a phosphoryl choline zwitterionomer (PC ionomer). The water uptake and subsequent swelling is induced by the phosphoryl choline (PC) end group functionality. The nonfunctional poly(trimethylene carbonate) is hydrophobic and as such does not absorb any water. Disks of the PC ionomer showed significant water uptake, typically above 90 wt % when fully swollen. This high water uptake triggered us to utilize the material for drug and protein loading and subsequent release. Fluorescein and fluorescein-labeled proteins were used as simple models for the loading and release characteristics of the material which was studied by fluorescence spectroscopy. The rate of release of the loaded molecules was compared, and it was shown that the release rate was similar for FITC and insulin but slightly slower for albumin. These results suggest that the PC ionomer may be used as a biodegradable and low elastic modulus material with an additional drug and/or protein release capacity. Such materials are of particular interest for use in a variety of applications in ViVo, for example as drug eluting stents. 1. Introduction In this paper we present the water uptake properties of a biodegradable phosphoryl choline zwitterionomer (PC ionomer) in combination with its ability to load and release drugs and proteins. We have in a series of recent articles described the synthesis and surface migration properties of various biomimetic phosphoryl choline (PC) functional biodegradable polymers.1 This has led to the development of a novel biodegradable PC ionomer material that through a careful synthesis protocol provides additional bulk organization of PC groups, thus forming a macroscopic physically crosslinked network consisting of PC-enriched domains.2 As further evidence of the presence of such domains we have recently found that the PC ionomer swells in water and phosphate-buffered saline (PBS, pH 7.4) solution. This discovery prompted us to use the material for drug and/or protein loading and subsequent release. The utilization of the polar PC domains is the foundation in this work and suggests that the PC ionomer could be used as an elastic, biodegradable, and hemocompatible material with an additional drug and/or protein delivery capacity. In general, hydrogels are cross-linked polymers which do not dissolve in water but swell significantly and retain a considerable amount of water within their structure. Numerous reviews have appeared in the literature which provide an excellent overview of this area.3 Hydrogels are either * Corresponding authors. E-mail: [email protected] (T.B.); [email protected] (K.I.). † Uppsala University. ‡ The University of Tokyo.

chemically or physically cross-linked, and most are formed spontaneously as they capture water within their structure. The advantage of physical hydrogels over covalently linked hydrogels is that they form their cross-linked structure by ionic, coiled-coil, or hydrophobic interactions and do not require any cross-linking agents for their formation.4 Hydrogels are frequently used in drug delivery systems due to the ease with which they can be loaded with drugs.5 The release of the loaded substance occurs either through diffusion of the drug molecule or by the continuous degradation or erosion of the gel. The degradation is dependent on the polymer and occurs either from the material bulk or from the surface. In the latter, the molecular weight is retained throughout the degradation process, whereas the former continuously decrease in molecular weight.6 The PC ionomer described in this paper consists of a poly(trimethylene carbonate) (PTMC) polymer backbone. PTMC is hydrolytically stable under in Vitro conditions but degrades in ViVo through probable enzymatic degradation.7 The PC ionomer does not swell immediately but over a period of days. The water uptake, however, is significant, typically being over 90% after 48 h, and in this way an organo hydrogel hybrid is formed. Drugs and proteins were loaded during the water uptake of manufactured PC ionomer disks, and the subsequent release was studied by fluorescence spectroscopy following the drying of the material. No polymer loss was detected during its swelling, which is consistent with the reported hydrolytic stability in Vitro,7 and as such the degradation was not measured in this model. The release was affected by diffusion of loaded drugs or proteins

10.1021/bm0504173 CCC: $30.25 © 2005 American Chemical Society Published on Web 10/05/2005

Biomacromolecules, Vol. 6, No. 6, 2005 3089

Organo Hydrogel Hybrids

and/or by redissolving the solutes during the reswelling of the material. However, in ViVo degradation of the polymer would provide a further possible release mechanism. Initially, the synthesis and swelling characteristics of the material with water is discussed. Typically, the one-pot endcapping synthesis gives quantitative conversion going from the polymer hydroxyl to PC and subsequent high yields. The material’s swelling characteristic in PBS is then discussed showing both its swell profile and the water content of the swollen PC ionomer. Furthermore, as a simple model for drug and protein loading and subsequent information on release profiles and kinetics, fluorescein (FITC), fluoresceinlabeled insulin (FITC-insulin), and flourescein-labeled albumin (FITC-albumin) were used. The results suggest that in ViVo drug and/or protein release can be achieved from the PC ionomer material in parallel to its biodegradable, hemocompatible, and low elastic modulus properties. 2. Experimental Section 2.1. Materials. Prior to use acetonitrile (Aldrich), pyridine (Aldrich), and 1,4-butanediol (Aldrich) were washed over CaH2, distilled, and stored under argon. Chloroform (VWR) was washed over a basic aluminum oxide (Al2O3) column and distilled from CaH2. Trimethylamine(g) (Aldrich) and trimethylene carbonate (Boehringer Ingelheim) were used as received. Sn(Oct)2 (tinoctoate) (Aldrich) was distilled at reduced pressure and stored in a glovebox. 2-Chloro-1,3,2dioxaphospholane-2-oxide (ethylene chloro phosphate) (Aldrich) was distilled at reduced pressure; the second fraction collected at a pressure of 0.011 mm of Hg at 48 °C and then stored in a freezer under argon. Fluorescein (FITC) obtained from WAKO was used as received. Fluoresceinlabeled albumin (FITC-Albumin), bovine serum albumin (BSA), fluorescein-labeled insulin (FITC-Insulin), and bovine pancreas insulin (all SIGMA) were all used as received. 2.2. Instrumentation. 1H NMR and 31P NMR spectra were recorded using a JEOL-ECP 400 MHz spectrometer with the solvent proton signal as an internal standard. Size exclusion chromatography (SEC) measurements were performed on a Waters Alliance GPCV 2000 with three Styragel HR columns (7.8 × 300 mm2, HR1/4/5) to determine the molecular weights and molecular weight distributions. Tetrahydrofuran (THF) was used as an eluent (40 °C), 1.0 mL/min, and a universal calibration with polystyrene standards (Shodex, Showa Denko) was performed. Millennium 32 was used to process the data from both a refractive index detector and a viscometer. Thermal analysis was performed on a differential scanning calorimeter (DSC) 6100 (Seiko Instrument Inc. Exstar). The cooling and heating rate was 5 °C/min, and the absorbed heat (Qh) of free water was detected by measuring the exotherm at 0 °C during the heating loop. The melt enthalpy (∆H) of water, which was used at 0 °C, was 333.4 J/g (not measured). The following equation was used to calculate the amount of free water (Wfw):8 Wfw )

Qh ∆H

and WH2O ) Wfw + Wbw in which Wbw is bound water and WH2O is the total amount of water. The maximum amount of free water was analyzed both on swollen and reswollen materials with an intermediate drying step. Fluorescence spectroscopy (FS) was performed on a JASCO FP 6500 spectrofluorometer at low power. A wavelength of 490 nm was used for excitation, and the emitted light was detected at 513 nm. The release curves are represented with the release at a certain time (Mt) normalized to the maximum release concentration (Mmax). Mmax was assumed to be equal to that of the initial loading concentration. Mechanical properties were analyzed on an Orientec universal testing machine using compression mode with a cross-head speed of 0.5 mm/min at ambient temperature (T ) 20 °C). The modulus from 4 ionomer disks was measured before and after incubation in phosphate-buffered saline (t ) 1 week, T ) 20 °C, pH 7.4), and the average was reported. 2.3. Procedures. 2.3.1. General Procedure for Bulk Polymerization of PTMC (1). A stir bar was added to a 50 mL two-necked Schlenk flask, and the flask was sealed with a septum. The flask was carefully flame-dried under vacuum and purged with nitrogen. For polymerization, trimethylene carbonate, 5.0 g (49.0 mmol), 1,4-butanediol, 0.11 g (1.225 mmol) for a degree of polymerization (DP) of 40 (20/arm), and Sn(Oct)2 catalyst, 0.025 g (61 µmol, 5 mol % to initiator), were combined in a glovebox and sealed with a rubber septum. Outside the glovebox, the reaction mixture was stirred at 110 °C using an oil bath. Following completion of the reaction (t ) 4 h), the poly(trimethylene carbonate) (PTMC) was dissolved in chloroform and precipitated in 500 mL of chilled methanol. The precipitate was allowed to sediment and was washed repeatedly with methanol and then dried under vacuum at 40 °C until it reached a constant weight. Yield: 95%, PDI ) 1.58. 1H NMR (CDCl3) δ ) 1.75 (m, -CH2-), 2.05 (m, -CH2-, poly), 3.70 (t, -CH2OH, ω-end), 4.22 (t, -CH2-, poly). 2.3.2. Synthesis of Phosphoryl Choline-Terminated PTMC via an Ethylene Phosphate Intermediate (2). For phosphorylation, 1.4 g (0.34 mmol) of 1 was weighed in a predried Schlenk flask and dissolved in 10 mL of dry chloroform (CHCl3). Thereafter, 4 equiv of dry pyridine, 0.11 mL (1.37 mmol), was added under a nitrogen atmosphere. The flask was attached to a predried dropping funnel and attached to a nitrogen inlet and cooled to -5 °C. Five milliliters of dry CHCl3 and 2 equivs of ethylene chlorophosphate, 63 µL (0.68 mmol), were slowly added dropwise, stirred for approximately 2 h, and gradually allowed to reach ambient temperature, after which it was stirred for a further 4 h. When the reaction was complete (monitored by 1H NMR) the solvent fraction was removed and the intermediate dissolved in dry acetonitrile. The solution was transferred to a pressure tube with two stopcocks, purged with nitrogen, sealed, and cooled to -10 °C. Approximately two equiv, or 63 µL (0.68 mmol) of trimethylamine(g) to PTMC polymer, was carefully condensed into the pressure tube and slowly

3090

Biomacromolecules, Vol. 6, No. 6, 2005

Nederberg et al.

Scheme 1. Synthetic Protocol for the Production of the PC Ionomer

heated to 65 °C. (Caution: reaction under pressure!) The solution was stirred for 45 h and left to cool to ambient temperature. The reaction product was precipitated in chilled methanol and the precipitate collected and dried until a constant weight was reached. Yield: 85%. 1H NMR (CDCl3) δ ) 1.75 (m, -CH2-), 2.05 (m, -CH2-, poly), 3.43 (s, -N(CH3)3, ω-end), 4.22 (t, -CH2-, poly). 31P NMR (CDCl3) δ ) 0. Manufacturing of Ionomer Disks. For the formation of ionomer disks, (diameter 6 mm and thickness 1 mm) initially a 1 mm thick rectangular sheet (15 × 35 mm2) was molded by heat compression using a rectangular Teflon form and two Teflon covers. The material was placed in the form between the two covers and compressed and heated simultaneously with a pressure of 1 ton and a temperature of 75 °C for 60 s. The compressed mold set was removed and stored in a freezer (-20 °C) for 5 min. The mold was dissembled and the rectangular ionomer sheet removed from the Teflon form. A circular steel tool with a diameter of 6 mm was used to hammer out the disks from the premade rectangular sheet. All the swelling and release data was obtained from the disks manufactured by this method. Procedures for Swelling and Release Studies. All the studies on the swelling of the disks were conducted on disks manufactured as described above. In all experiments a PBS solution with a pH of 7.4 was used. Disks were immersed, and the degree to which they swelled was calculated at different time points using the following formula: Swelling (%) )

(

)

mt - mo × 100 mo

in which mo is the original weight and mt is the weight at time t.

The polymer loss was also measured by comparing freezedried sample after a certain swelling time with mo. All swelling experiments were conducted at ambient temperature, T ) 20 °C. For the release studies the fluorescent probes were dissolved to a certain load concentration in PBS and the disks incubated in the solution for 1 week until fully swollen. The disks were removed and freeze-dried for 24 h prior to immersion in a fresh PBS (pH 7.4) solution at physiological temperature (37 °C). Aliquots were taken at specific times, and the fluorescence intensity of each of the samples was checked with a fluorescence spectrometer. Following each measurement the aliquot was replaced into the solution from which it was taken. The data are presented by plotting the release concentration that corresponds to a specific intensity (Mt) normalized with the maximum load concentration (Mmax) as function of time. The maximum release was equal to that of the load concentration and was validated as judged from the results from FITC since a decrease in fluorescence during the loading corresponded to the increase of fluorescence during the release. 3. Results and Discussion 3.1. Synthesis. The synthesis of the PC ionomer is shown in Scheme 1. Similar synthetic protocols have been described previousely.1,2 Initially, the polymer backbone is produced by ring-opening polymerization (ROP) of the cyclic trimethylene carbonate monomer by using 1,4-butanediol as initiator and Sn(Oct)2 as a catalyst. This afforded bifunctional polymers with a controlled molecular weight and narrow polydispersity (PDI ∼ 1.58). The number average molecular weight was measured by 1H NMR end group analysis and was calculated to be Mn ∼ 4000 g/mol (DP ) 40, 20/arm). The phosphoryl choline functionality was obtained by

Organo Hydrogel Hybrids

Figure 1.

31P

NMR spectrum of the PC ionomer.

quantitatively converting the hydroxyl end group to phosphoryl choline in a single vessel and by changing the solvent from chloroform to acetonitrile prior to the ring-opening of the phosphor triester intermediate. The PC ionomer which was formed was purified by precipitation in chilled methanol and centrifugation of the emulsified solution such that the material was obtained in high yields, typically ∼90%. The PC content of the material which was produced was about 8 wt %. Importantly, no fractionated precipitation took place as measured by 1H NMR analysis, which showed that the molecular weight was maintained at 4000 g/mol. 1 H NMR in combination with 31P NMR provides a strong and clear indication of the synthetic end group transformation from the hydroxyl of the initial polymer to the target PC functionality of the PC ionomer which is finally formed. For 1 H NMR the two strong resonances of the polymer backbone at 2.05 and 4.22 ppm in combination with the methylene group adjacent the hydroxyl at 3.70 ppm provided valuable information allowing the starting PTMC to be characterized. For the formed PC ionomer the distinct singlet from the choline at 3.42 ppm provided one further characterization handle. Moreover, 31P NMR provided further information allowing characterization such that only one singlet signal was observed for the PC group at around 0 ppm. Shown in Figure 1 is the 31P NMR spectrum of the formed PC ionomer. 3.2. Water Swelling and Characteristics of the Swollen PC Ionomer. Interestingly, the PC ionomer which was obtained was shown to swell in water and PBS solution. The swelling is not a rapid process but requires a period of days. The amount of absorbed water is, however, significant, typically above 90 wt % in 48 h such that an organo hydrogel hybrid is formed. We have in a previous publication attributed the interesting low elastic modulus and rubbery properties of the PC ionomer to the formation of a physically cross-linked network of PC groups, and the recent finding of the ability to swell in water or PBS further supports this presence.2 Since we were primarily interested in how the material behaved in ViVo we only worked under physiological PBS concentrations. Shown in Figure 2 is the resulting swell

Biomacromolecules, Vol. 6, No. 6, 2005 3091

profile for PC ionomer disks incubated in PBS (pH 7.4) at 20 °C for 1 week. Already after 24 h the water uptake was significant, almost 80%. After about 72 h equilibrium was established as the swelling reached a maximum, which was above 90%. Importantly, the nonfunctional starting PTMC did not show any water uptake, which is consistent with previous reports that demonstrated the hydrophobic behavior of the PTMC polymer.7 Therefore, the water uptake of the material can be attributed to the introduced PC functionality. The presumed mechanism, though not shown in any experiment explicitly, is that the polar PC domains swell (Figure 3) as they incorporate water into their structure and form a network of swollen aggregates and channels. The osmotic pressure is balanced during the water uptake, and at equilibrium equal pressure inside the PC enriched domains and the outside PBS solution is established. Ongoing experiments are focusing on this phenomenon, and the corresponding structure formed by water intrusion will be presented in a future publication. A further feature of the swollen ionomer is that no polymer loss was observed. This means that the weight was maintained after a swelling/drying cycle and that no polymer erosion occurred during the absorption of water. This characteristic has been observed for swelling times up to 1 week and is consistent with previous reports in the literature.7 One cannot predict that this would be the case for the in ViVo condition as PTMC was reported to degrade from probable enzymatic degradation.7 On the contrary, a more rapid degradation or material erosion due to the high water uptake in ViVo may occur. Previous reports have shown that nondegradable hydrated PC polymers with a high free water structure may encapsulate proteins reversibly and without a significant conformational change and that the adsorption of proteins is low.10-12 Thus, the free water content of water absorbing polymers is believed to be of high importance in maintaining biological properties, such as hemocompatibility. To characterize the water structure inside the swollen PC ionomer, the amount of free and bound water was analyzed using DSC at different swelling degrees. In general, free water does not take part in any interactions with the polymer structure and instead demonstrates characteristics as that of pure water, i.e., the position of the melt transition and melt enthalpy are identical. On the contrary, bound water interacts with the polymer primarily by solvating the polar parts of the molecule. Additionally, this water fraction also solvates the hydrophobic parts of the polymer and does not participate in the melting transition at 0 °C. Shown in Figure 4 is the free water fraction as function of the total water percentage inside the swollen PC ionomer. The free water fraction increases with a higher degree of swelling, and when fully swollen (above 90% total water uptake) the free water content is almost 40%. The same level was reached also on a reswollen PC ionomer with an intermediate drying step. Earlier reports have shown similar free water contents and that these polymers could maintain the protein structure following contact of the same.10 The modulus of the PC ionomer disks was measured in compression mode. For the dry ionomer, the modulus was

3092

Biomacromolecules, Vol. 6, No. 6, 2005

Nederberg et al.

Figure 2. Swelling profile of PC ionomer disks incubated in PBS (pH 7.4) at 20 °C.

Figure 3. Schematic swelling mechanism of the PC ionomer as the polar PC domains attracts water into their structure.

Figure 4. Free water percentage of the PC ionomer as function of the total water percentage.

1.4 ((0.1) MPa, which is in accordance with earlier findings of its mechanical behavior as studied by rheological measurements.2 The modulus of the swollen PC ionomer disks was measured after incubation in PBS for 1 week at 20 °C and was found to be 130 ((50) kPa. This result shows that the absorbed water had a pronounced plasticizing effect such that the material was even softer and the modulus (bulk) decreased by about 1 order of magnitude. 3.3. Drug/Protein Loading and Release. The swelling characteristics of the PC ionomer prompted us to utilize its behavior for parallel drug and/or protein loading and subsequent release studies. To investigate this further, different water soluble drugs or proteins were dissolved in PBS and loaded into the PC ionomer during the swelling process. As simple models for drug or protein release, FITC and mixtures of insulin/FITC-insulin and albumin/FITC-

albumin were used. The fluorescent probes were used to investigate whether the release process was size dependent. A difference of approximately 1 order of magnitude between each of the compounds was found with respect to the molecular weight. The molecular weight of FITC is 332.3 g/mol, whereas the molecular weights for insulin and albumin are ∼6000 and ∼65 000 g/mol, respectively. Shown in Figure 5 is the release profile for FITC showing the release concentration that corresponds to a specific intensity (Mt) normalized with the maximum load concentration (Mmax) as function of time. From Figure 5 one can see that the release of FITC is rapid and that a near quantitative release was reached after about 6 h. The release profiles of insulin and albumin were similar to that of FITC. Shown in Figure 6 are the respective release profiles for insulin and albumin.

Organo Hydrogel Hybrids

Biomacromolecules, Vol. 6, No. 6, 2005 3093

Figure 5. Release profile of FITC.

Figure 6. Release profiles of insulin (above) and albumin (below).

For insulin the release reached a maximum of about 65% after 24 h. In the case of albumin a similar plateau was reached after 2 weeks at which the release was about 60%. After 24 h the release of albumin was almost 35%. These

findings showed that the majority of the proteins were released. Additionally, the result suggests that a small protein fraction was physically entrapped inside the swollen PC ionomer. Since the bound water fraction was about 60% for

3094

Biomacromolecules, Vol. 6, No. 6, 2005

the PC ionomer, it is likely that some proteins came into contact with the hydrophobic part of the PC ionomer. Hydrophobic interactions can then take place with parts of the proteins being retained. This effect was more pronounced as the molecular weight of the protein is increased, i.e., the probability of hydrophobic interactions increases. This can be observed in the case of albumin where a longer release time is required before equilibrium is reached. For the in ViVo condition, however, we suggest that it is possible to reach a quantitative release through the continuous degradation of the PTMC backbone. The physical structure of the PC ionomer would be eroded, and as a result, the entrapped proteins would be released continuously. In one final investigation we wanted to make a comparison of the kinetics from the release process. Initially the data was fitted with the first-order equation ln(Mt/Mmax) ) krt in which kr is the rate constant. Since the initial release rate for all of the released molecules was higher and nonlinear and the R2 value below 0.9, this plot was, however, not used. A poor fit to the first-order release kinetics is expected because of an initial burst release effect, which is due to molecules adsorbed to the outer surface of the samples that are released directly upon immersion of the disks into the PBS solution, such that it provides a nonlinear area within the overall release kinetics. Processes such as swelling and change in osmotic pressure over time will in addition affect the release kinetics. To compare the release rates on a qualitative basis we used the average half-lives derived from the release data in Figures 5 and 6. Due to the initial burst release the first half-life in each series was discarded in the comparison. The average half-lives were t1/2FITC ) 8400 ((1600) s, t1/2FITC-Insulin ) 8500 ((950) s, and t1/2FITC-Albumin ) 310 000 ((68 500) s. From the obtained half-lives it was found that the release rate was similar for FITC and FITCinsulin and more than 1 order of magnitude faster than for FITC-albumin. This finding suggests that a retardation effect for larger molecules exists; however, from these data it is not possible to define the position of this threshold. The loading concentration for FITC was 10 ng/mL, while the concentrations of insulin and albumin were 35 µg/mL (10 µg/mL FITC-insulin and 25 µg/mL insulin) and 10 mg/mL (1 mg/mL FITC-albumin and 9 mg/mL albumin), respectively. 4. Conclusion With the help of various fluorescent probes we have through the course of this work demonstrated that drug or proteins may be loaded into a water-absorbing PC ionomer material. We have in addition shown that the molecules loaded into the material can be released in a subsequent step; thus, the material functions as a temporary reservoir for drug or proteins. The water content inside the swollen PC ionomer in combination with the release data have shown that the free water content is sufficient in order to promote loading

Nederberg et al.

and release of the load molecule. The bound water fraction did affect the protein release as a small fraction was retained as a result of hydrophobic interactions. This effect increased with the size of the protein. The absorbed water had a pronounced plasticizing effect on the PC ionomer as the modulus decreased approximately 1 order of magnitude; however, the water uptake did not cause any detectable polymer loss. We hypothesize that proteins may be released from the PC ionomer with a retained activity and secondary structure. It was also found that FITC and FITC-insulin was released at similar rates, while the larger protein FITCalbumin was released at a slightly lower rate. This study has formed a foundation for further research and the overall result suggests that the PC ionomer may be used as a biodegradable and low elastic modulus material and with an additional drug and/or protein release capacity. Such materials are of particular interest for a variety of uses in ViVo, for example as drug eluting stents. Acknowledgment. Uppsala University and The Swedish Foundation for International Cooperation in Higher Education and Research (STINT) are acknowledged for financial support in this work. References and Notes (1) (a) Nederberg, F.; Bowden, T.; Hilborn, J. Macromolecules 2004, 37 (3), 954-965. (b) Nederberg, F.; Bowden, T.; Hilborn, J. Polym. AdV. Technol. 2005, 16, 108-112. (2) Nederberg, F.; Bowden, T.; Hilborn, J. J. Am. Chem. Soc. 2004, 126, 15350-15351. (3) (a) Kamath, K. R.; Park, K. AdV. Drug DeliVery ReV. 1993, 11, 5984. (b) Peppas, N. A.; Bures, P.; Leobandung, W.; Ichikawa, H. Eur. J. Pharm. Biopharm. 2000, 50, 27-46. (c) Lee, K. Y.; Mooney, D. J. Chem. ReV. 2001, 101 (7), 1869-1879. (d) Hoffman, A. S. AdV. Drug DeliVery ReV. 2002, 43, 3-12. (e) Hennink, W. E.; van Nostrum, C. F. AdV. Drug DeliVery ReV. 2002, 54, 13-36. (4) (a) Nam, K. W.; Watanabe, J.; Ishihara, K. Biomacromolecules 2002, 3, 100-105. (b) Wang, C.; Steward, R. J.; Kopecek, J. Nature 1999, 397, 417-420. (c) Qu, X.; Wirsen, A.; Albertsson, A. C. J. Appl. Polym. Sci. 1999, 74, 3186-3192. (5) For example: (a) Kimura, M.; Fukumoto, K.; Watanabe, J.; Ishihara, K. J. Biomater. Sci., Polym. Ed. 2004, 15 (5), 631-644. (b) Nam, K.; Watanabe, J.; Ishihara, K. Eur. J. Pharm. Sci. 2004, 23 (3), 261270. (c) Nam, K.; Watanabe, J.; Ishihara, K. Int. J. Pharm. 2004, 275, 259-269. (d) Nam, K.; Watanabe, J.; Ishihara, K. J. Biomater. Sci., Polym. Ed. 2002, 13 (11), 1259-1269. (6) Go¨pferich, A. Biomaterials 1996, 17, 103-114. (7) (a) Matsuda, T.; Kwon, I. K.; Kidoaki, S. Biomacromolecules 2004, 5, 295-305. (b) Pego, A. P.; Poot, A. A.; Grijpma, D. W.; Feijen, A. J. J. Mater. Sci.: Mater. Med. 2003, 14 (9), 767-773. (c) Pego, A. P.; Poot, A. A.; Grijpma, D. W.; Feijen, A. J. Macromol. Biosci. 2002, 2 (9), 411-419. (d) Albertsson, A. C.; Eklund, M. J. Appl. Polym. Sci. 1995, 57 (1), 87-103. (e) Zhu, K. J.; Hendren, R. W.; Jensen, K.; Pitt, C. G. Macromolecules 1991, 24, 1736-1740. (8) Hirata, Y.; Miura, Y.; Tanaka, S.; Nakagawa, T. J. Membr. Sci. 2000, 176, 21-30. (9) Greenfield, N.; Fasman, G. D. Biochemistry 1969, 8, 4108-4115. (10) Ishihara, K.; Nomura, H.; Mihara, T.; Kurita, K.; Iwasaki, Y.; Nakabayashi, N. J. Biomed. Mater. Res. 1998, 39, 323-330. (11) Kitano, H.; Sudo, K.; Ichikawa, K.; Ide, M.; Ishihara, K. J. Phys. Chem. B 2000, 104, 11425-11429. (12) Kitano, H.; Imai, M.; Mori, T.; Gommei-Ide, M.; Yokoyama, Y.; Ishihara, K. Langmuir 2003, 19, 10260-10266.

BM0504173