Chitosan Hydrogel

Oct 15, 2010 - Graduate School of the Chinese Academy of Sciences, Beijing ... side chain on the microcapsules decreased with increases in the degree ...
150 downloads 5 Views 2MB Size
Download PDF
pubs.acs.org/Langmuir © 2010 American Chemical Society

Chitosan-g-MPEG-Modified Alginate/Chitosan Hydrogel Microcapsules: A Quantitative Study of the Effect of Polymer Architecture on the Resistance to Protein Adsorption Jia N. Zheng,†,‡ Hong G. Xie,† Wei T. Yu,† Xiu D. Liu,*,§ Wei Y. Xie,† Jing Zhu,† and Xiao J. Ma*,† † Laboratory of Biomedical Material Engineering, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, Dalian 116023, China, ‡Graduate School of the Chinese Academy of Sciences, Beijing 100049, China, and §College of Environment and Chemical Engineering, Dalian University, Dalian Economic Technological Development Zone, Dalian 116622, China

Received July 29, 2010. Revised Manuscript Received September 30, 2010 The chemical modification of the aginate/chitosan/aginate (ACA) hydrogel microcapsule with methoxy poly(ethylene glycol) (MPEG) was investigated to reduce nonspecific protein adsorption and improve biocompatibility in vivo. The graft copolymer chitosan-g-MPEG (CS-g-MPEG) was synthesized, and then alginate/chitosan/alginate/CSg-MPEG (ACACPEG) multilayer hydrogel microcapsules were fabricated by the layer-by-layer (LBL) polyelectrolyte self-assembly method. A quantitative study of the modification was carried out by the gel permeation chromatography (GPC) technique, and protein adsorption on the modified microcapsules was also investigated. The results showed that the apparent graft density of the MPEG side chain on the microcapsules decreased with increases in the degree of substitution (DS) and the MPEG chain length. During the binding process, the apparent graft density of CS-g-MPEG showed rapid growth-plateau-rapid growth behavior. CS-g-MPEG was not only bound to the surface but also penetrated a certain depth into the microcapsule membranes. The copolymers that penetrated the microcapsules made a smaller contribution to protein repulsion than did the copolymers on the surfaces of the microcapsules. The protein repulsion ability decreased with the increase in DS from 7 to 29% with the same chain length of MPEG 2K. CS-g-MPEG with MPEG 2K was more effective at protein repulsion than CS-g-MPEG with MPEG 550, having a similar DS below 20%. In this study, the microcapsules modified with CS-g-MPEG2K-DS7% had the lowest IgG adsorption of 3.0 ( 0.6 μg/cm2, a reduction of 61% compared to that on the chitosan surface.

1. Introduction The microencapsulation of live cells and tissues within a protective hydrogel membrane is proposed as a promising method of precluding the problems associated with immune rejection during allogenic and xenogenic transplantation.1-4 Polymeric hydrogel semipermeable membranes have a 3D network structure5 with a high water content that is formed by crosslinking polymer chains through physical, ionic, or covalent interactions. This characteristic determines to a large extent the transport properties that allow the free exchange of low-molecularweight species for cell survival and function, such as oxygen, nutrients, metabolites, and therapeutic products through porous hydrogels, while restricting the ingress of antibodies and complement fractions. Because of their nonimmunogenic, nontoxic, and biocompatible properties, alginate/chitosan/alginate (ACA) *To whom correspondence should be addressed. (X.J.M.) E-mail: maxj@ dicp.ac.cn. (X.D.L.) E-mail: [email protected]. Tel: 86-411-84379139. Fax: 86-411-84379096. (1) Chang, T. M. Science 1964, 146, 524–525. (2) Lim, F.; Sun, A. M. Science 1980, 210, 908–910. (3) Chang, T. M. S. Artif. Organs 2004, 28, 265–270. (4) Orive, G.; Gascon, A. R.; Hernandez, R. M.; Igartua, M.; Pedraz, J. L. Trends. Pharmacol. Sci. 2003, 24, 207–210. (5) Elisseeff, J. Nature 2008, 7, 271–273. (6) Zielinski, B. A.; Aebischer, P. Biomaterials 1994, 15, 1049–1056. (7) Chandy, T.; Mooradian, D. L.; Rao, G. H. R. Artif. Organs 1999, 23, 894–903. (8) Xie, H. G.; Zheng, J. N.; Li, X. X.; Liu, X. D.; Zhu, J.; Wang, F.; Xie, W. Y.; Ma, X. J. Langmuir 2010, 26, 5587–5594. (9) Limor, B.; Marcelle, M. Biopolymers 2006, 82, 570–579. (10) Tasima, H.; Hongmei, C.; Ouyang, Wei; Christopher, M.; Bisi, L.; Aleksandra, M. U.; Satya, P. Biotechnol. Lett. 2005, 27, 317–322.

17156 DOI: 10.1021/la1030203

microcapsules have attracted much attention as an immunoprotective device for transplanting cells in disease therapy.6-10 As a transplantation device, nonspecific protein adsorption on the surface of a microcapsule is the most important biological processes whenever it is integrated into a biological environment. Protein adsorption is recognized as the first event in the response of a living body to artificially implanted devices.11 The adsorbed proteins then mediate the subsequent adverse biological events, including thrombus formation, foreign body reaction, bacterial infection, and other undesirable bioresponses, which will lead to implantation failure.11 Therefore, surface modification of the artificial biomaterials to eliminate the nonspecific adsorption of proteins is an important issue in the design of biomedical devices. To reduce foreign body responses, much attention has been paid to producing a nonspecific protein-repelling surface by surfacemodification methods. Among the uncharged and water-soluble polymers, poly(ethylene glycol) (PEG) has extraordinary protein resistance arising from a number of outstanding properties such as minimum interfacial free energies with water, extensive hydration, good conformational flexibility, considerable chain mobility, and a large excluded volume.12 The mechanism underlying the protein resistance of PEG-modified surfaces is frequently attributed to (1) a tightly bound interfacial hydration layer on the surface that prevents direct contact between the surface and protein and (2) the steric stabilization effect, including the volume restriction effect of the loss of configurational entropy and the osmotic repulsion effect between PEG molecules and proteins.12 Moreover, PEG’s low (11) Blanka, R. Adv. Drug Delivery 2000, 42, 65–80. (12) Jin, H. L.; Hai, B. L.; Joseph, D. Prog. Polym. Sci. 1995, 20, 1043–1079.

Published on Web 10/15/2010

Langmuir 2010, 26(22), 17156–17164

Zheng et al.

Article

toxicity and low immunogenicity make it a suitable material for applications in the field of biomedical devices,13 and it is on the FDA’s generally recognized as safe (GRAS) list. Recent studies in eliminating protein adsorption by PEG modification have focused on the self-assembled monolayer (SAM) on a metal or crystalline surface and have demonstrated the relevance of the chain length and surface coverage of PEG in imparting protein resistance to the surface.14-20 In these studies, PEG was grafted as side chains onto a cationic polyelectrolyte backbone such as poly (L-lysine)15-17 or chitosan,14,18 which had shown spontaneous adsorption from aqueous solutions onto negatively charged surfaces. Other types of coatings were also made from PEG chains with various terminal groups assembling onto different functional surfaces.19 The results implied that there existed a critical value of the PEG graft density on the surfaces, above which the densely packed and overlapped PEG chains would stretch out of the surfaces into the solution, generating a brushlike structure and preventing protein adsorption by the steric repulsion mechanism. However, the quantitative relationship between the PEG graft density and the amount of subsequently adsorbed proteins on the hydrogel microcapsule membrane with a complex 3D network structure is still unclear. In this study, chitosan-g-MPEG (CS-g-MPEG) graft copolymers with varying degrees of substitution (DS), methoxy poly(ethylene glycol) (MPEG) chain length, and chitosan molecular weight were synthesized. The alginate/chitosan/alginate/CS-gMPEG (ACACPEG) multilayer hydrogel microcapsules modified by MPEG were fabricated by the layer-by-layer (LBL) selfassembly method. The binding amount of CS-g-MPEG on the ACACPEG microcapsules was measured by the gel permeation chromatography (GPC) technique. The quantitative relationship between the apparent graft density of CS-g-MPEG and the extent of protein adsorption on the microcapsules, which was affected by the DS, the MPEG chain length, and the molecular weight of the chitosan backbone of the copolymers, was fully investigated.

2. Materials and Methods 2.1. Materials. MPEG with Mn = 550 and 2000 g mol-1 was purchased from Sigma Aldrich Chemical Co. Chitosan with molecular weights of 22 and 79 kDa and polydispersity index (PDI) values of 1.4 and 1.6, respectively, as measured by GPC21 was degraded from the raw material (Yuhuan Ocean Biomaterials Corporation, Zhejiang, China) in our laboratory. The degree of deacetylation (DD) was 95% as determined by 1H NMR. Sodium alginate was purchased from Qingdao Crystal Salt Bioscience and Technology Corporation (Qingdao, China), of which the viscosity at a concentration of 1 wt % in a 0.9 wt % NaCl aqueous solution at 25 °C was 230 cP. Sodium cyanoborohydride (NaCNBH3) was obtained from Fluka Chemical Co. Gamma globulin (IgG, Bovine Blood) was purchased from Sigma-Aldrich Chemical Co. and used (13) Harris, J. M. Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications; Plenum Press: New York, 1992; Chapter 1. (14) Ye, Z.; Bo, L.; Natalija, G.; Ricardas, M.; Andra, D.; Per, M. C. J. Colloid Interface Sci. 2007, 305, 62–71. (15) Stephanie, P.; Susan, M.; De, P.; Janos, V.; Nicholas, D.; Marcus, T. Langmuir 2003, 19, 9216–9225. (16) Gregory, L.; Kenausis, J. V.; Donald, L. E.; Huang, N. P.; Rolf, H.; Laurence, R. T.; Marcus, T.; Jeffrey, A. H.; Nicholas, D. S. J. Phys. Chem. 2000, B104, 3298–3309. (17) Huang, N. P.; Roger, M.; Janos, V.; Marcus, T.; Rolf, H.; Antonella, R.; Donald, L. E.; Jeffrey, A. H.; Nicholas, D. S. Langmuir 2001, 17, 489–498. (18) Gorochovceva, N.; Naderi, A.; Dedinaite, A.; Makuska, R. Eur. Polym. J. 2005, 41, 2653–2662. (19) Rundqvist, J.; Hoh, J. H.; Haviland, D. B. Langmuir 2005, 21, 2981–2987. (20) Geoffrey, O.; Joseph, I.; Evgeni, P.; Ricardas, M.; Ausvydas, V.; Claesson, P. M. Langmuir 2008, 24, 5341–5349. (21) Terbojevich, M.; Cosani, A.; Focher, B.; Marsano, E. Carbohydr. Res. 1993, 250, 301–314.

Langmuir 2010, 26(22), 17156–17164

Scheme 1. Syntheses of MPEG-CHO and CS-g-MPEG

without any pretreatment. Fluorescein isothiocyanate (FITC) was purchased from Acros Organics Co. All the other reagents or chemicals were analytical grade and used without further purification. 2.2. Synthesis of MPEG-Aldehyde. MPEG-aldehyde (MPEG-CHO) was prepared by the oxidation of the terminal hydroxy of MPEG (Mn = 550 and 2000 g mol-1) with acetic anhydride in dimethyl sulfoxide (DMSO).22,23 In brief, MPEG was dissolved in a DMSO/chloroform mixture under a nitrogen environment. Acetic anhydride was then added, and the mixture was stirred for 9 h at room temperature under nitrogen protection (Scheme 1). The reaction mixture was then precipitated in excess dry, cold ethyl ether and filtered. Then the white precipitate was reprecipitated twice from CH2Cl2 with ethyl ether. The product obtained was dried in vacuum overnight at room temperature. The white powder obtained in this step was a mixture of MPEGCHO and the unreactive MPEG. The latter could be removed by dialysis and an acetone rinse in the next step. 2.3. Synthesis of CS-g-MPEG. Graft copolymer CS-gMPEG was synthesized according to the method described by Harris.22 Chitosan was dissolved in a mixture of acetic acid aqueous solution and methanol, to which an MPEG-CHO aqueous solution was subsequently added. The mixture was stirred for 1 h at room temperature, and then its pH was adjusted to 6.0-6.5 with a NaOH aqueous solution. After being stirred for another 2 h at room temperature, a NaCNBH3 aqueous solution was added dropwise and the solution was stirred for another 18 h at room temperature under nitrogen protection (Scheme 1). The reaction mixture was dialyzed with a dialysis membrane (Solarbio Co., MWCO 7 kDa) against distilled water and then freeze dried. The remaining MPEG was removed by washing several times with excess acetone. The final product was obtained after vacuum drying overnight at room temperature. By varying the molar ratio of MPEG-CHO to chitosan, graft copolymers with different DS values were synthesized.

2.4. Characterization of MPEG-CHO and CS-g-MPEG. Fourier transform infrared (FT-IR) spectra were recorded with KBr pressed disks on a Bruker Vector 22 FTIR spectrometer using 4 cm-1 resolution and 32 scans. 1H NMR spectra were recorded on a Bruker Avance 500 MHz spectrometer. An MPEG or MPEG-CHO sample (0.015 g) was dissolved in CDCl3 (0.5 mL) with TMS as an internal reference, and the spectra were recorded at room temperature. Chitosan and CS-g-MPEG samples were dissolved in 0.5 mL of 20 wt % D2O/CD3COOD, and the spectra were recorded at 70 °C. GPC was carried out using a system equipped with TSK G3000þG4000 PWXL gel size-exclusion columns (7.8 mm  300 mm, 10 μm particle diameter, Tosoh Corporation, Tokyo, Japan), a refractive index detector (Waters, model 2414, Milford, MA), and an HPLC pump (Waters, model 515). The flow rate was 0.7 mL min-1 using 0.2 M acetic acid/ 0.1 M sodium acetate buffer (pH 4.2) as the eluent. The injection volume was 20 μL, and the temperature of both the detector and (22) Harris, J. M. J. Polym. Sci., Polym. Chem. Ed. 1984, 22, 341–352. (23) Masatoshi, S.; Minoru, M.; Hitoshi, S.; Hiroyuki, S.; Yoshihiro, S. Carbohydr. Polym. 1998, 36, 49–59.

DOI: 10.1021/la1030203

17157

Article

Zheng et al.

columns was 30 °C. All data provided by the GPC system were collected and analyzed with the Empower Workstation software package (Waters Corp., MA).

2.5. Preparation of ACACPEG Hydrogel Microcapsules. The multilayer hydrogel microcapsules with CS-g-MPEG modification were prepared by the method developed by our group.24 Sodium alginate was dissolved in a 0.9 wt % NaCl solution with a final concentration of 1.5 wt %. Subsequently, the solution was filtered successively over microfilters and stored overnight in a refrigerator (4 °C) to facilitate deaeration before being used. The calcium alginate beads were prepared by extruding the sodium alginate solution through a 0.5 mm needle dropwise into a 1.1 wt % CaCl2 solution using an electrostatic droplet generator (YD-04, Dalian Institute of Chemical Physics, Chinese Academy of Sciences, China). After being gelled for 30 min, the beads were incubated in a 0.5 wt % chitosan solution dissolved in a 0.1 M sodium acetate/ 0.2 M acetic acid buffer solution (pH 4.2) for 10 min, followed by being rinsed with a physiological solution to remove excess chitosan. Then the AC microcapsules were incubated with a 0.15 wt % alginate solution for 20 min to counteract excess positive charges on the membrane. After the ACA microcapsules were rinsed, the modification procedure was carried out by incubating the ACA microcapsules in a 1.0 wt % CS-g-MPEG solution to obtain ACACPEG multilayer microcapsules in which the CS-g-MPEG layer is abbreviated as CPEG. ACAC microcapsules as a control with unmodified chitosan on the outermost layer were prepared according to the same process for ACACPEG in which the unmodified chitosan layer was abbreviated as C. The diameter of ACAC and ACACPEG microcapsules prepared by this method was about 350 μm, which was measured with a laser particle sizer (Beckman Counter LS100Q). The coefficient of variance (CV) of the diameter of the microcapsules was less than 5%.

2.6. Measurement of the Amount of Bound CS-g-MPEG. A quantitative study of the bound CS-g-MPEG on the microcapsules was carried out by measuring the CS-g-MPEG concentration decrease during the binding process with the GPC technique established by Yu el al.25 Because the membrane thickness of CS-g-MPEG bound on the microcapsules is quite small compared to the diameter of ACA microcapsules (350 μm), the bound amount was defined as the mass of copolymer per unit area of the microcapsules (μg/cm2).26 ACA microcapsules (0.1 mL) were incubated with a CS-g-MPEG solution (1.0 wt %, 1 mL) at room temperature, and then the supernatant was withdrawn at a certain time and quantified by GPC. Because the area integral of the GPC elution curve was linearly correlated to the total concentration of the polymer solution, the amount of bound CS-g-MPEG was calculated from the respective chromatograms according to the concentration-area integral standard curves of CS-g-MPEG standard solutions with concentrations from 2 to 10 mg/mL. The concentration of the CS-g-MPEG solution at time 0 was detected as soon as the polymer solution was added to the ACA microcapsules, avoiding the error introduced by dilution. The amount of bound unmodified chitosan on the microcapsules’ outermost layer was also analyzed by the same method as used for the control. The amount of bound chitosan and CS-g-MPEG (mpol, μg/cm2) was calculated using the following equation mpol ¼

ðCo - Cn ÞVt A

where Co and Cn are the initial polymer concentration and the polymer concentration in the supernatant after the binding process, respectively, Vt is the total volume of the solution (1 mL), and A is (24) Liu, X. D.; Xue, W. M.; Liu, Q.; Yu, W. T.; Fu, Y. L.; Xiong, X.; Ma, X. J.; Yuan, Q. Carbohydr. Polym. 2004, 56, 459–464. (25) Yu, W. T.; Lin, J. Z.; Liu, X. D.; Xie, H. G.; Zhao, W.; Ma, X. J. J. Membr. Sci. 2010, 346, 296–301. (26) Gugerli, R.; Cantana, E.; Heinzen, C.; Vonstockar, U.; Marison, I. W. J. Microencapsulation 2002, 19, 571–590.

17158 DOI: 10.1021/la1030203

the total sphere surface area of the microcapsules incubated in the solution. The apparent graft density of MPEG chains (nPEG, expressed as the number of molecules per unit area, chains/nm2), which was determined from mpol and DS, could be calculated from the following equation15 nPEG ¼

mpol Mchitosan þ MPEG DS

nEG ¼ N  nPEG where Mchitosan and MPEG are the molecular weights of the chitosan monomer and the MPEG chain, respectively. nEG represents the apparent graft density of the ethylene glycol (EG) monomer. N is the number of repeat units of MPEG. The values of N are considered to be 44 and 11 for MPEG2K and MPEG550, respectively. The reported data are the mean values of triplicate samples for each kind of microcapsule.

2.7. Measurement of the Permeability of ACACPEG Microcapsules. IgG was labeled with FITC according to the

literature.27 Briefly, IgG (200 mg) was dissolved in 20 mL of a NaHCO3/Na2CO3 buffer solution (pH 9.0) and mixed with 1 mL of a solution of 5 mg of FITC in 5 mL of methanol. The reaction was carried out for 2 h in the dark at room temperature. The FITC-IgG in the solution was separated from unreacted FITC by passage through a Sephadex G-25 column in Na2HPO4/KH2PO4 buffer solutions (pH 7.4). The permeability of the multilayer microcapsules was investigated with a confocal laser scanning microscope (CLSM, Leica, TCS-SP2, Germany) equipped with a Laser Sources blue laser (Ar 488 nm/5 mW) and an inverted microscope (Leica, DMIRE2, Germany). All of the microcapsule samples were analyzed in fluorescence mode with the appropriate laser wavelength (488 nm excitation and 530 nm emission). All confocal fluorescence pictures were taken with a 10 objective (HC PLAN APO 10 x/0.40 PH1 0.17/A 2.2). For background and noise reduction, the images were generated by accumulating four scans per image, and the image size was 1024 pixels  1024 pixels. Leica confocal software (Leica, Germany) was used for CLSM imaging analysis. The microcapsules (0.1 mL) were incubated in the 0.3 mL FITC-IgG solution (Na2HPO4/KH2PO4 buffer solutions, pH 7.4), which was shaken at 37 °C for 2 h, and then underwent CLSM scanning. 2.8. Measurement of Protein Adsorption. Protein adsorption measurements were carried out on the basis of the method described previously.28 IgG was used as a model protein for evaluating the protein-resistant characteristics of ACACPEG microcapsules. IgG adsorption experiments were carried out in 10 mmol/L phosphate buffer (Na2HPO4/KH2PO4, pH 7.4) in order to maintain a constant pH during the adsorption process. The microcapsules (0.1 mL) were incubated with an IgG solution (0.3 mL) with a concentration of 2.0 mg/mL at 37 °C for 2 h. The protein concentration of the centrifuged supernatant was measured by the Bradford method.29 The adsorbed amounts of protein were determined by measuring the difference in IgG concentration before and after the adsorption process.28 The mass per area (q, μg/cm2) of the protein adsorbed on the microcapsules was calculated using the following equation q ¼

ðCi - Cf ÞV S

where Ci and Cf are the initial protein concentration and the protein concentration in the supernatant after the adsorption studies, (27) Mario, B.; Thomas, F.; Ralf, B.; Frank, B.; Volker, S. Biomaterials 2006, 27, 3505–3514. (28) Xie, H. G.; Li, X. X.; Lv, G. J.; Xie, W. Y.; Zhu, J.; Thomas, L.; Ma, R.; Ma, X. J. J. Biomed. Mater. Res. 2010, A92, 1357–1365. (29) Bradford, M. M. Anal. Biochem. 1976, 72, 248–254.

Langmuir 2010, 26(22), 17156–17164

Zheng et al.

Article

Figure 1. FT-IR spectra of (A) MPEG and MPEG-CHO and (B, a) chitosan, (b) CS-g-MPEG2K-DS7%, and (c) CS-g-MPEG2KDS21%. respectively, V is the total volume of the solution (0.3 mL), and S is the total sphere surface area of the microcapsules added to the solution. The reported data were mean values of triplicate samples for each kind of microcapsule. 2.9. Statistical Analysis. Data were expressed as the mean ( the standard error of the mean. Statistical comparisons were performed using one-way analysis of variance (ANOVA). p values smaller than 0.05 were considered to indicate statistical significance.

3. Results and Discussion 3.1. Characterization of MPEG-CHO and CS-g-MPEG. The notation CS(x)-g-MPEG(y)-DS(z) for the copolymers was used to represent the molar mass of chitosan in daltons (x Da), the molar mass of MPEG (y Da), and the degree of substitution DS (z) (the percent of the number of MPEG side chains divided by the number of chitosan monomers). For example, a copolymer labeled CS79K-g-MPEG2K-DS35% had a chitosan backbone with the molecular weight of 79 kDa grafted with MPEG side chains of molecular weight 2 kDa with a DS of 35%, and in all cases where the x value of the copolymer is not mentioned, the chitosan molecular weight (x) is 22 kDa. The chemical structures of MPEG, MPEG-CHO, chitosan, and CS-g-MPEG were confirmed by FT-IR, 1H NMR, and GPC analyses. FT-IR data of MPEG and MPEG-CHO are shown in Figure 1A. The peaks at 3448 (O-H), 2887 (C-H), 1466 (formation vibration of C-H), 1350 (wagging vibration of C-H), 1280 (twisting vibration of C-H), 1242 (twisting vibration of C-H), and 1101 cm-1 (C-O-C) are characteristic of MPEG. In addition to the peak at 1739 cm-1 due to -CHO groups of MPEG-CHO, there were no differences between the FT-IR spectra of MPEG-CHO and Langmuir 2010, 26(22), 17156–17164

Figure 2. 1H NMR spectra of (A) MPEG and MPEG-CHO and (B) chitosan, CS-g-MPEG2K-DS7%, CS-g-MPEG2K-21%, and CS-g-MPEG2K-DS29%.

MPEG. Figure 1B shows the FT-IR data of chitosan and CS-gMPEG. The absorption peaks at 3435 (O-H), 2875 (C-H), 1655 (amide I), 1601 (amide II), 1381 (amide III), and 1082 cm-1 (C-OC) were attributed to the characteristics of chitosan. By comparing the FT-IR spectra of chitosan and CS-g-MPEG, it could be seen that the peak intensity corresponding to CH stretching at 2875 cm-1 was significantly strengthened after the modification and the peak intensities of amide I at 1655 cm-1 and amide II at 1601 cm-1 significantly decreased as the DS of CS-g-MPEG increased. The C-O-C stretching peaks shifted to 1095 cm-1 in comparison with that of chitosan at 1082 cm-1, and the absorption peaks associated with MPEG also appeared at 951 (wagging vibration of ether C-H), 1252 (twisting vibration of ether C-H), 1302 (twisting vibration of ether C-H), 1352 (wagging vibration of ether C-H), and 1458 cm-1 (formation vibration of ether C-H) as a result of MPEG grafting. Furthermore, the -CHO peak at 1739 cm-1 in MPEG-CHO and the peak at 1381 cm-1 (amide III) in chitosan both disappeared in the spectrum of CS-g-MPEG, indicating that MPEG-CHO had been successfully grafted to the chitosan backbone as side chains. 1 H NMR spectra of MPEG and MPEG-CHO are shown in Figure 2A. The chemical shifts at 3.49-3.79 (-CH2CH2O-), 3.38 (-OCH3) and 2.15 ppm (-OH and H2O) are due to the backbone structure of MPEG. The chemical shifts at 9.73 (-CHO) and 4.16 ppm (-CH2CHO) are attributed to the characteristic structure of MPEG-CHO. 1H NMR spectra of chitosan and CS-g-MPEG are shown in Figure 2B. The chemical DOI: 10.1021/la1030203

17159

Article

Zheng et al.

Figure 3. GPC spectra of MPEG, chitosan, CS-g-MPEG2KDS7%, CS-g-MPEG2K-DS21%, and CS-g-MPEG2K-DS29%.

shifts at 4.97 (H-1), 3.85-4.07 (H-3, H-4, H-5, H-6, H-60 ), 3.28 (H-2), and 2.14 ppm (H-a) are due to the protons of the saccharide structure of chitosan. The peak of MPEG methylene (H-b) overlapped with those of H-3, H-4, H-5, H-6, and H-60 of the monosaccharide residue in the range of 3.63-4.02 ppm. The main signals on CS-g-MPEG were similar to those observed on chitosan. Additionally, a new observed signal at 3.46 ppm (H-c) is due to the methoxyl of MPEG, and its intensity increased with the increase in DS. From the 1H NMR spectra, the DS value was evaluated from the relative peak intensities between the -OCH3 group of MPEG (3.46 ppm) and H-1 (4.93 ppm) of the monosaccharide residue.30 As shown in the GPC results in Figure 3, all of the polymers exhibited broad unimodal molecular weight distributions. The retention times of MPEG and chitosan appeared at 27 and 23 min, respectively. The CS-g-MPEG copolymers with higher molecular weights and larger steric hindrance values compared to those of chitosan had lower retention volumes and shorter retention times. Moreover, as the DS of MPEG increased, the molecular weight and steric hindrance of the copolymer increased, resulting in the retention time of the CS-g-MPEG copolymer gradually decreasing. From these combined results, it was clear that the introduction of MPEG was successfully accomplished and that the DS increased with the increase in the ratio of MPEG-CHO to chitosan. 3.2. Quantitative Analysis of CS-g-MPEG during the Binding Process by GPC. The multilayer microcapsules modified by CS-g-MPEG were fabricated by the LBL self-assembly method. The inner layer served to provide mechanical strength and immunoprotection, and the outer layer of CS-g-MPEG served to prevent protein adsorption.31 During the binding process, the supernatant in the reaction bath was collected at 0, 5, 15, 25, 35, and 45 min and analyzed by GPC. The GPC chromatograms of CSg-MPEG samples collected at different binding times are shown by differently colored lines in Figure 4. The GPC chromatograms of CS-g-MPEG samples changed asymmetrically with binding time. CS-g-MPEG with a smaller molecular weight distribution fraction (22-27 min) took part in the binding process, no matter what the DS and MPEG chain lengths were. Taking into account the polydispersity of CS-g-MPEG, the results indicated that CS-gMPEG molecules with smaller steric hindrance values diffused more rapidly to the interfaces and bonded to ACA microcapsules. The newly formed steric barrier subsequently hindered the diffusion of larger CS-g-MPEG molecules. Moreover, the ACA microcapsule (30) Bhattaraia, N.; Ramaya, H. R.; Gunna, J.; Matsenb, F. A.; Zhang, M. Q. J. Controlled Release 2005, 103, 609–624. (31) Sawhney, A. S.; Hubbell, J. A. Biomaterials 1992, 13, 863–870. (32) Olav, G.; Olav, S.; Gudmund, S. B. Biomaterials 1998, 19, 1815–1825.

17160 DOI: 10.1021/la1030203

Figure 4. GPC chromatograms of CS-g-MPEG during the binding process on the ACA microcapsules: (A) CS-g-MPEG550-DS17%, (B) CS-g-MPEG2K-DS7%, and (C) CS-g-MPEG2K-DS21%.

membrane with a certain average pore size32 limited the diffusion of the copolymers with larger steric hindrance values. The amount of bound CS-g-MPEG with various DS values, MPEG chain lengths, and molecular weights of the chitosan backbone on the ACA microcapsules during a binding time of 45 min are summarized in Table 1. The small amounts of bound CS-g-MPEG2K-DS29% and CS79K-g-MPEG2K-DS35% on the surfaces of the microcapsules were not discernible using the GPC technique because these quantities were smaller than the limit of detection in this study. The binding kinetics curves in Figure 5 showed that the apparent graft density increased, with values ranging from 0 to 317 ( 32, 97 ( 8, and 77 ( 5 chains/nm2 for the CS-g-MPEG550DS17%, CS-g-MPEG2K-DS7%, and CS-g-MPEG2K-DS21% systems, respectively, upon increasing the binding time from 0 to 45 min. Both the bound amount and binding rate of CS-g-MPEG with shorter MPEG550 side chains were significantly higher than Langmuir 2010, 26(22), 17156–17164

Zheng et al.

Article Table 1. Summary of the Binding Data for CS-g-MPEG Copolymers with Different Architectures

microcapsules

Mn MPEG (Da)

DS (%)

Mwchitosan (kDa)

binding amount (mpol, μg/cm2)

monomer density (nEG, 1/nm2)

reduction vs control (%)c

ACACa 0 22 98 ( 9 0 0 550 17 22 80 ( 8 3487 18 ACACPEG 2K 7 22 70 ( 6 4268 61 ACACPEG 2K 21 22 36 ( 2 3388 18 ACACPEG b 2K 29 22 10 ACACPEG b 2K 35 79 28 ACACPEG a Chitosan without MPEG modification was used as a control group. b The binding amount of CS-g-MPEG was undetectable by GPC. c The reduction percentage of the amount of IgG adsorption was calculated by comparing with the chitosan surface.

Figure 5. Binding kinetics curves of CS-g-MPEG550-DS17%, CS-g-MPEG2K-DS7%, CS-g-MPEG2K-DS21%, CS-g-MPEG2KDS29%, and CS79K-g-MPEG2K-DS35% on the ACA microcapsules.

those of CS-g-MPEG with longer MPEG2K side chains because the steric hindrance of MPEG2K is larger than that of MPEG550. CS-g-MPEG2K with a larger steric hindrance diffused less effectively and more slowly into the membrane with the 3D gel network structure than did CS-g-MPEG550. The larger steric hindrance made the driving force lower, which is a short-range electrostatic interaction between the amino groups (-NH3þ) and carboxyl groups (-COO-) for adsorption. Moreover, MPEG2K with a greater excluded volume cannot pack as tightly on the surface as a smaller MPEG550 molecule because of the stronger steric interactions between neighboring side chains. For CS-g-MPEG550-DS17%, an initial rapid adsorption process was followed by equilibrium after 45 min because CS-gMPEG reacted first with the free alginate chains of the external coating, resulting in a larger steric hindrance and fewer -COOanchoring sites on the surface, which increased the resistance for additional CS-g-MPEG chains to penetrate the outer layer and bind onto the microcapsules. Furthermore, CS-g-MPEG adsorption was primarily controlled by the balance between the attractive electrostatic interaction of the alginate-chitosan backbone and the steric repulsion of the MPEG side chains.16 It was thought that MPEG chains stretching from the surface had separated the chitosan backbone from alginate on the microcapsules as the result of a steric effect and shielded part of the charge restricting further electrostatic interaction.33 Moreover, the CS-g-MPEG2K-DS7% and CS-g-MPEG2KDS21% layers showed a rapid binding-plateau-rapid binding process. It was speculated that CS-g-MPEG with a larger steric hindrance bound on the ACA microcapsules would take surface (33) Wu, J.; Wang, X. F.; Keum, J. K.; Zhou, H. W.; Mikhail, G.; CarlosAlberto, A. O.; Pan, H.; Chen, W. L.; Chiao, S. M.; Hsiao, B. S.; Chu, B. J. Biomed. Mater. Res. 2007, A80, 800–812.

Langmuir 2010, 26(22), 17156–17164

rearrangements that allowed for the binding of more copolymers.16,34-36 It should be noted that these two copolymers had similar initial binding rates but the final rates decreased with increasing DS because of the increase in steric hindrance associated with the incorporation of more MPEG chains into the developing layer. Although CS-g-MPEG with a higher DS of 21% was speculated to have a higher apparent graft density because of the greater number of MPEG chains per single molecule, the copolymer with a DS of 7% had a greater number of anchoring groups (free amino groups) and a smaller steric hindrance between neighboring MPEG side chains, both of which are advantageous for CS-g-MPEG2K-DS7% binding. The apparent graft density values in this work appeared to be significantly greater than those found by others in SAMs in which the graft density of CS-g-MPEG2K chain assembled on the carboxylated gold surface was only 0.26 chains/nm2.37 The considerably high apparent graft density achieved in this work might be attributed to the fact that CS-g-MPEG could not only bind on the surface but also penetrate a certain depth into the membranes of microcapsules. A hydrogel membrane has a 3D network structure with a high water content that is formed by cross-linking alginate and chitosan chains through electrostatic complexing interactions. This characteristic determines to a large extent the transport properties that allow the diffusion of low-molecular-weight CS-g-MPEG through porous hydrogels. Therefore, the modification process includes CS-g-MPEG molecule diffusion into an ACA membrane with a 3D network structure and simultaneous binding between protonated amino groups of CS-g-MPEG and carboxyl groups of alginate. Moreover, the ability of a copolymer to diffuse into the ACA membrane32 is also probably why the binding equilibrium took more time to establish than for a SAM.15 3.3. Immunoisolation Property Analysis of ACACPEG Microcapsules. The permeability of ACACPEG microcapsules was examined by CLSM with FITC-IgG. Figure 6 shows the CLSM image of the ACACPEG microcapsules modified with CS-g-MPEG2K-DS7%. The dark appearance of the microcapsule interior, of which the fluorescence intensity was zero, contrasted with the bright-green fluorescent outside, demonstrating that the microcapsules were impermeable to IgG during the 2 h adsorption process. The ACAC microcapsules and the microcapsules modified by CS-g-MPEG with different DS values, MPEG chain lengths, and molecular weights of the chitosan backbone had the same results, which are not shown here. Therefore, in the study of protein adsorption, the total reduction of IgG in the incubation solution was all due to the (34) Xu, R. L.; Gred, D. U.; Winnik, M. A.; Marinho, J. M. G.; Oliviera, J. M. R. Langmuir 1994, 10, 2977–2984. (35) Ngankam, A. P.; Van Tassel, P. R. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 1140–1145. (36) Yan, F.; Dejardin, P.; Galin, J. C.; Schmitt, A. Colloid Polym. Sci. 1991, 269, 1021–1025. (37) Zhou, Y.; Liedberg, B.; Zhou, Y.; Liedberg, B.; Natalija, G.; Ricardas, M.; Andra, D.; Claesson, P. M. J. Colloid Interface Sci. 2007, 305, 62–71.

DOI: 10.1021/la1030203

17161

Article

Zheng et al.

Figure 6. CLSM image of ACACPEG microcapsules modified by CS-g-MPEG2K-DS7% and incubated in 0.3 mL of FITC-IgG solution (Na2HPO4/KH2PO4 buffer solutions, pH 7.4) for 2 h at 37 °C. The fluorescence intensity profile (top right) corresponds to the white line crossing one microcapsule.

amount of IgG adsorption on the surfaces of the microcapsules without diffusion into the microcapsules. 3.4. IgG Adsorption on ACACPEG Microcapsules. The effect of the apparent graft density on the amount of IgG adsorbed on the modified surface of the microcapsules was investigated. From the binding kinetics curves, the apparent graft density of CS-g-MPEG2K-DS21% increased by 170% from 22 ( 7 to 59 ( 1 chains/nm2 with binding times from 5 to 25 min as shown in Figure 7A; however, IgG adsorption on an modified microcapsule surface (7.9 ( 0.5 μg/cm2) showed no significant difference compared to that on an unmodified chitosan surface (7.7 ( 0.5 μg/cm2). As the apparent graft density reached 77 ( 5 chains/nm2, the modified surface started to reduce the IgG adsorption to 6.3 ( 0.3 μg/cm2 as shown in Figure 7B. This was due to the fact that the apparent graft density of MPEG chains was not high enough to resist the interaction of protein with the underlying microcapsule membrane when the binding time was less than 25 min in this case. There appeared a threshold of chain density above which protein resistance increased sharply because the overlapping MPEG side chains started to extend away from the surfaces of microcapsules, forming brushlike structures, and in turn started to enhance the repulsive interactions between the MPEG chains against adsorption; then the nonspecific protein adsorptions were expected to decrease. This indicated that the protein repellency was due to increasing bound amounts, resulting in an increase in the apparent graft density. Therefore, microcapsules modified by CS-g-MPEG with a binding time of 45 min were chosen to study the factors influencing the apparent graft density and protein adsorption. 3.5. Effect of CS-g-MPEG Architecture on the Apparent Graft Density and Protein Adsorption. The architectural parameters, including DS, the MPEG side-chain length, the molecular weight of the chitosan backbone, are important factors in resisting protein adsorption. The quantitative relationship between the apparent graft density and the extent of protein adsorption, which was affected by the DS, MPEG chain length, and molecular weight of chitosan in the copolymers, was fully investigated. 17162 DOI: 10.1021/la1030203

Figure 7. The apparent graft density of MPEG with binding times of 5, 25, and 45 min (A) and effect of the apparent graft density on IgG adsorption (B) on the ACACPEG microcapsules modified by CS-gMPEG2K-DS21%.

3.5.1. Effect of DS. To explore the effect of DS on the apparent graft density and subsequent protein resistance behavior, CS-gMPEG2K-DS7%, CS-g-MPEG2K-DS21%, and CS-g-MPEG2KDS29% were synthesized. As shown in Figure 8, CS-g-MPEG2KDS7% and CS-g-MPEG2K-DS21% exhibited apparent graft densities of 97 ( 8 and 77 ( 5 chains/nm2, respectively, and the binding amount of CS-g-MPEG2K-DS29% was not discernible. The results showed that CS-g-MPEG2K-DS7% demonstrated a more remarkable resistance to IgG adsorption, which decreased the IgG adsorption by 61% to 3.0 ( 0.6 μg/cm2. Obviously, the apparent graft density decreased with increasing DS from 7 to 29%, resulting in weaker protein resistance. These results are opposite to the results for the SAM15 because for a given MPEG molecular weight the increase in DS was expected to result in both a weakened electrostatic interaction between CS-g-MPEG and alginate as a result of the decrease in the number of -NH3þ anchoring sites and an increase in the steric hindrance of MPEG chains, which were both unfavorable to the copolymer diffusion and binding process. The lower the amount of bound CS-g-MPEG, the fewer the brushlike MPEG chains anchored to the surface, which led to less extended steric repulsion against protein adsorption. 3.5.2. Effect of the MPEG Chain Length. The MPEG chain length also affected the protein adsorption on the surfaces of the modified microcapsules. Recent studies of PEG modification focused on PEG with a molecular weight of 500 to 5000, which had more pronounced steric repulsion.15 The results in Figure 9 showed that with similar DS values the apparent graft density of Langmuir 2010, 26(22), 17156–17164

Zheng et al.

Figure 8. Effect of DS on the apparent graft density of MPEG and IgG adsorption.

Article

probably because the contact time was shortened.12 This indicated that longer MPEG chains had greater degrees of water hydration, larger excluded volumes, and more pronounced steric repulsion, in effect causing these chains to be more efficient at shielding the surface from protein adsorption compared to shorter chains.39 Furthermore, the results showed that longer MPEG2K chains as well as shorter MPEG550 chains with a comparative value of nEG had the same ability to reject proteins, as shown in Table 1. They indicated that both the chain length and apparent graft density affected the protein-resistant property of MPEG-based coatings, implying that the protein resistance of surfaces modified by MPEG required a sufficiently high density of EG monomer units.15 To achieve the same extent of protein resistance, the apparent graft density needed a greater quantity of shorter MPEG chains than longer ones. 3.5.3. Effect of the Molecular Weight of the Chitosan Backbone. In a previous study in our group, it was concluded that the ACA microcapsules made from chitosan with a range of molecular weight between 20 and 100 kDa had good membrane strength.24 Therefore, chitosan molecules with molecular weights of 22 and 79 kDa were chosen and modified with MPEG to study the effect of the chitosan backbone molecular weight on the apparent graft density and IgG adsorption. With similar DS values, the CS79K-g-MPEG2K-DS35% copolymer had a chitosan backbone with a larger molecular weight than did CS-gMPEG2K-DS29%. The microcapsule surface modified by the CS79K-g-MPEG2K-DS35% copolymer with 79 kDa chitosan backbone demonstrated a more remarkable inhibition of IgG adsorption (5.55 ( 0.07 μg/cm2) than did the CS-g-MPEG2KDS29% modified surface (6.9 ( 0.1 μg/cm2). The results showed that the diffusion of CS79K-g-MPEG2K-DS35% was mostly restricted by the larger steric hindrance and would bind primarily onto the surface of microcapsules forming a thin layer.30 This indicated that the MPEG chains located in the outermost layer were more effective at preventing protein adsorption than were the chains penetrating the microcapsules.

CS-g-MPEG550-DS17% (317 ( 32 chains/nm2) was 4 times that of CS-g-MPEG2K-DS21% (77 ( 5 chains/nm2) but there was no significant difference in the amount of protein adsorbed on the modified surface. Although the microcapsules modified by the higher-molecular-weight MPEG2K molecules actually had a much lower apparent graft density, they were as effective as the modified microcapsules with shorter MPEG550 chains at reducing protein adsorption. This was because MPEG2K side chains with a radius of gyration (Rg) of 1.66 nm occupy a greater excluded volume than do MPEG550 side chains with an Rg of 0.78 nm.16 Moreover, a larger MPEG2K molecule was surrounded by a hydration shell that was negligible for a lower-molecular-weight MPEG550 molecule.38 The mobility of the hydrated MPEG chains increased with chain length, and rapidly moving hydrated MPEG chains on a surface would effectively prevent the stagnation of proteins on the surface,

4. Conclusions In the present study, the quantitative relationship between the apparent graft density of MPEG and the extent of protein adsorption on the ACACPEG multilayer hydrogel microcapsules, which was affected by the DS, MPEG chain length, and molecular weight of the chitosan backbone of the copolymers, was fully investigated. On the basis of the results of our present studies, we concluded that the CS-g-MPEG copolymer binding process was primarily controlled by the balance between the attractive electrostatic interaction and the steric repulsion of the MPEG side chains. CS-g-MPEG with a large steric hindrance showed rapid binding-plateau-rapid binding behavior because it underwent surface rearrangements during the binding process. The bound amount and the binding rate decreased with increasing DS and MPEG chain length because of the steric hindrance and shielding effect. CS-g-MPEG not only bound to the surface of the ACA microcapsules but also diffused a certain depth into the membrane of the ACA microcapsules. IgG protein adsorption was drastically reduced on microcapsules that were coated with CS-g-MPEG. The protein repulsive behavior through steric stabilization effect was affected by the apparent graft density of MPEG chains on the surfaces of microcapsules. The copolymers that penetrated the microcapsules

(38) Harris, J. M. Poly(ethylene glycol) Chemistry: Biotechnical and Biomedical Applications; Plenum Press: New York, 1992; Chapter 16.

(39) Gombotz, W. R.; Wang, G.; Horbett, T. A.; Hoffman, A. S. J. Biomed. Mater. Res. 1991, 25, 1547–1562.

Figure 9. Effect of MPEG molecular weight on (A) the apparent graft density of MPEG and (B) IgG adsorption.

Langmuir 2010, 26(22), 17156–17164

DOI: 10.1021/la1030203

17163

Article

make a smaller contribution to protein repulsion than do the copolymers on the surfaces of microcapsules with MPEG side chains stretching out of the outer layer. The protein repulsion ability decreased when DS increased from 7 to 29% with the same chain length of MPEG 2K. CS-g-MPEG with a longer MPEG2K chain length was more effective at protein repulsion than was CS-g-MPEG550 with a similar DS below 20%. Acknowledgment. This study was supported by the National Key Sci-Tech Special Project of China (grant numbers 20082x10002-019), the National Basic Research Program of China (grant 2007CB714305), the National Natural Science Foundation of China (grants 20736006, 20806080, 20876018, and 10979050), and the Hi-Tech Research and Development (863) Program of China (grant 2006AA02A140).

17164 DOI: 10.1021/la1030203

Zheng et al.

Supporting Information Available: Gel permeation chromatogram of pullulan standards with a series of molecular weights and an elution time-molecular weight calibration curve of pullulan standards. Micrograph (10 objective) and graph of the diameter and size distribution of the ACACPEG microcapsules. CLSM image of 0.1 mL of ACAC microcapsules incubated in 0.3 mL of FITC-IgG solution (Na2HPO4/KH2PO4 buffer solutions, pH 7.4) for 2 h at 37 °C. Concentrationabsorbance standard curve of the IgG standard. This material is available free of charge via the Internet at http://pubs.acs.org. Note Added after ASAP Publication: This article was published ASAP on October 15, 2010. The second equation in section 2.6 has been modified. The correct version was published on October 20, 2010.

Langmuir 2010, 26(22), 17156–17164