Langmuir 2005, 21, 2495-2504
2495
Oligo(ethylene glycol) Containing Polymer Brushes as Bioselective Surfaces Luisa Andruzzi,† Wageesha Senaratne,†,‡ Alexander Hexemer,§ Erin D. Sheets,‡,|,⊥ B. Ilic,# Edward J. Kramer,§ Barbara Baird,‡,| and Christopher K. Ober*,†,| Materials Science and Engineering, Cornell University, Ithaca, New York 14853, Department of Materials, University of California at Santa Barbara, Santa Barbara, California 93106, Chemistry and Chemical Biology, Cornell University, Ithaca, New York 14853, Cornell Nanofabrication Facility, Cornell University, Ithaca, New York 14853, and Cornell Nanobiotechnology Center, Cornell University, Ithaca, New York 14853 Received September 30, 2004. In Final Form: December 6, 2004 The nitroxide-mediated polymerization of styrenic monomers containing oligo(ethylene glycol) (OEGn) moieties was chosen for the preparation of biocompatible polymer brushes tethered to silicon oxide surfaces due to the broad range of monomer structures available and the use of a nonmetallic initiator. These surfaces were characterized by near-edge X-ray absorption fine structure and water contact angle measurements. The biocompatibility of these grown polymer brushes was studied and compared with deposited assemblies of surface-bound OEGn-terminated silanes with selected chain lengths. Grown polymer brushes with short OEGn side chains suppressed protein adsorption significantly more than the deposited assemblies of short OEGn chains, and this was attributed to higher surface coverage by the brushes. Cell adhesion studies confirmed that OEGn-containing polymer brushes are particularly effective in preventing nonspecific adhesion. Studies of protein adsorption and cell localization carried out with specific ligands on surfaces patterned demonstrated the potential of these surface-tethered polymer brushes for the formation of micro- and nanoscale devices.
Introduction The tendency of proteins or cells to physically adsorb onto a substrate without specific receptor recognition is known as nonspecific adsorption. This type of contamination typically reduces the efficacy of enzyme-linked immunoassays, western blot tests, protein microarrays, biosensors, single molecule detection, and single cell analysis. Undesirable features are high background noise and “false positives”. Poly(ethylene glycol) (PEG) and PEGbased polymeric materials have been used for many biological applications because of PEG’s capacity to resist protein and cell adhesion, as well as its nontoxicity and its nonimmunogenicity.1,2 Nevertheless, the fabrication of PEG surfaces with optimal selectivity, biocompatibility and minimal nonspecific interactions remains a key goal in the development of many diagnostic and biosensing applications. Most commonly studied surfaces in current nanobiotechnology applications are oligo(ethylene glycol) (OEGn, n ) number of EG units) alkanethiol assemblies on gold due to the simplicity of the surface chemistry and the * To whom correspondence may be addressed. cober@ ccmr.cornell.edu (C.K.O.). † Materials Science and Engineering, Cornell University. ‡ Chemistry and Chemical Biology, Cornell University. § Department of Materials, University of California at Santa Barbara. | Cornell Nanobiotechnology Center. ⊥ Current address: Department of Chemistry, Penn State University, University Park, PA 16802. # Cornell Nanofabrication Facility, Cornell University. (1) Harris, J. M. Poly(ethyleneglycol) chemistry: Biotechnical and Biomedical Applications; Plenum Press: New York, 1992 and references therein. (2) Lee, J. H.; Lee, H. B.; Andrade, J. D. Prog. Polym. Sci. 1995, 20, 1043-1079.
formation of ordered assemblies. Prime and Whitesides3 prepared OEGn alkanethiol self-assemblies on gold of different chain lengths and characterized these films for protein adsorption. They showed that these self-assembled monolayers had high surface coverage and were able to suppress protein adhesion regardless of the EG chain length. Mrksich et al.4 verified the use of OEGn assemblies by tailoring the interfacial characteristics of the monolayers to control protein and cell adhesion. The authors used soft lithographic techniques to lay out simple tri(ethylene glycol)-terminated alkanethiol monolayers that could control protein and cell adhesion remarkably well on gold substrates. Zhu et al.5 used long and short chain OEGn alkanethiol assemblies to investigate the differences in cell response. The results showed that cell adhesion was sensitive to the concentration of proteins in the culture medium and to the length of the OEGn chains. In particular, short assemblies on gold were found to be not very effective. Only a few examples of OEGn-alkylsilanes on SiOX have been reported in the literature mostly due to the difficulty of controlling surface chemistry with active silane end groups, formation of less ordered films, and lack of chemical versatility.6,7 Nevertheless, alkylsilane assemblies on SiOX show higher chemical and thermal stability and easier substrate preparation, and they are optically transparent if prepared on glass.8 (3) Prime, K. L.; Whitesides, G. M. J. Am. Chem. Soc. 1993, 115, 10714-10721. (4) Mrksich, M.; Chen, C. S.; Xia, Y.; Dike, L. E.; Ingber, D. E.; Whitesides, G. M. Proc. Natl. Acad. Sci. 1996, 93, 10775-10778. (5) Zhu, B.; Eurell, T.; Gunawan, R.; Leckband, D. J. Biomed. Mater. Res. 2001, 56, 406-421. (6) Mrksich, M.; Whitesides, G. M. TIBTECH 1995, 13, 228-235. (7) Fadeev, A. Y.; McCarthy, T. J. Langmuir 1999, 15, 3759-3766.
10.1021/la047574s CCC: $30.25 © 2005 American Chemical Society Published on Web 02/05/2005
2496
Langmuir, Vol. 21, No. 6, 2005
Lee and Laibinis9 prepared OEGn-terminated alkylsilanes and tested them with proteins of different size. They showed that these OEGn assemblies were resistant toward smaller proteins such as albumin, hexokinase, and insulin. But for larger proteins such as fibrinogen, they demonstrated that short chain OEGn silanes had lower protein resistance than their thiol counterparts. They believed that this was due to the difficulties associated with the silane-based coatings. Sofia et al.10 studied the protein resistance of PEG by using it in linear and star (n > 3400) forms grafted to SiOX and evaluated the dependence of protein adsorption to PEG grafting density. On linear PEG surfaces, it was found that a nonoverlapping monolayer was not sufficient to significantly prevent protein adsorption due to low surface coverage. They also reported that protein adsorption depended not only on the grafting density but also on the grafting density for a specific molecular weight PEG. In contrast, grafted star PEG chains allowed for a high protein resistance regardless of the grafting density as long as open spaces between grafted molecules were smaller than the effective size of proteins. Nevertheless, the tedious synthesis of star polymers does not make it a practical choice. The present work was directed toward creating OEGn surfaces with high resistance to protein and cell adsorption for the fabrication of biological and chemical micro- and nanoscale devices. This effort utilizes the surface-initiated nitroxide-mediated polymerization of styrenic monomers containing short OEGn segments. Surface-initiated polymerization (SIP) has attracted much interest in recent years due to its versatility and lack of polymer desorption. This method holds advantages over the “grafting to” technique where the functional end of a polymer chain is covalently attached to the reactive sites of a surface. In fact the SIP process does not involve entropic factors at the surface due to crowding of initial grafting chains that prevent further insertion of polymer onto the surface. Thus SIP allows for assembly of well-defined nanoscopic structures which have found application in a variety of areas. Several synthetic methods have been developed for this purpose including cationic11 and anionic polymerization,12 ring opening and ring opening metathesis polymerization,13,14 free radical,15,16 nitroxidemediated,17,18 and atom transfer radical polymerizations.19-21 (8) Ulman, A. An introduction to ultrathin organic thin films: from Langmuir Blodgett to Self-Assembly; Academic Press: San Diego, CA, 1991. (9) Lee, S. W.; Laibinis, P. E. Biomaterials 1998, 19, 1669-1675. (10) Sofia, S. J.; Premnath, V.; Merrill, E. W. Macromolecules 1998, 31, 5059-5070. (11) Zhao, B.; Brittain, W. J. J. Am. Chem. Soc. 1999, 121, 35573558. (12) Jordan, R.; Ulman, A.; Kang, J. F.; Rafailovich, M. H.; Sokolov, J. J. Am. Chem. Soc. 1999, 121, 1016-1022. (13) Jordan, R.; Ulman, A. J. Am. Chem. Soc. 1998, 120, 243-247. (14) Kim, N. Y.; Jeon, N. L.; Choi, I. S.; Takami, Y.; Harada, Y.; Finnie, K. R.; Girolami, G. S.; Nuzzo, R. G.; Whitesides, G. M.; Laibinis, P. E. Macromolecules 2000, 33, 2793. (15) Prucker, O.; Ru¨he, J. Macromolecules 1998, 31, 592-601. (16) De Boer, B.; Simon, H. K.; Werts, M. P. L.; van Der Vegte, E. W.; Hadziioannou, G. Macromolecules 2000, 33, 349-356. (17) Blomberg, S.; Ostberg, S.; Harth, E.; Bosman, A. W.; Van Horn, B.; Hawker, C. J. J. Polym. Sci., Part A: Polym. Chem. 2002, 40, 13091320. (18) Ignatova, M.; Voccia, S.; Gilbert, B.; Markova, N.; Mercuri, P. S.; Galleni, M.; Sciannamea, V.; Lenoir, S.; Cossement, D.; Gouttebaron, R.; Jerome, R.; Jerome, C. Langmuir 2004, 20, 10718-10726. (19) Shah, R. R.; Merreceyes, D.; Husemann, M.; Rees, I.; Abbott, N. L.; Hawker, C. J.; Hedrick, J. L. Macromolecules 2000, 33, 597-605. (20) Feng, W.; Brash, J.; Zhu, S. J. Polym. Sci., Part A: Polym. Chem. 2004, 42, 2931-2942. (21) Ma, H.; Hyun, J.; Stiller, P.; Chilkoti, A. Adv. Mater. 2004, 16, 338-341.
Andruzzi et al.
Our synthetic approach leads to SiOX surface tethered homopolymer and block-copolymer brushes with properties that are superior to those of the OEGn or PEG assemblies studied to date, and it allows the formation of uniform surfaces with controlled chemical architecture, high surface coverage and thickness, and good chemical and thermal stability.19,22 Chemical stability of surfaceinitiated polymer brushes is especially important in patterning applications, and this concept is highlighted in the work by Shah et al., where the authors carry out a pattern transfer onto gold substrates using polymer brushes as barriers to various wet chemical etchants of gold.19 The preference for nitroxide-mediated polymerization (NMP) over atom transfer radical polymerization (ATRP) was dictated by the fact that this process does not utilize metal-based catalysts. We believe that metal impurities involved in ATRP are not easy to completely eliminate from the polymerization product and therefore may be undesirable in many biological applications.23 Polymer brushes were characterized by near-edge X-ray absorption fine structure (NEXAFS) and water contact angle measurements. Protein adsorption and cell adhesion studies were carried out on these surfaces and on OEGn silane assemblies for comparison. To demonstrate the use of the surface-grown OEGn-containing polymer brushes for selective protein adsorption and cell localization on patterned media, we have fabricated and tested structures with a recently developed polymer lift-off 24 patterning method and also with a standard lithography process. Experimental Section Materials. All reagents and stable free radical 2,2,6,6tetramethylpiperidinyloxy (TEMPO) were purchased from Aldrich and used without further purification unless noted. Styrene was washed several times with a 30% NaOH aqueous solution and then distilled over CaH2 (bp 45 °C/30 mmHg). Diglyme was distilled over sodium (bp 62 °C/17 mmHg). Tetrahydrofuran (THF) was distilled over sodium at 65-67 °C under nitrogen atmosphere. AlexaFluor 488 labeled IgG antibodies, bovine serum albumin (BSA), and collagen were obtained from Molecular Probes. Fibronectin (Sigma Chemicals) was labeled with AlexaFluor488 dye (Molecular Probes) following manufacturer protocol. Anti-dinitrophenyl (DNP) IgE antibodies produced at the Hybridoma Center at Oklahoma State University were purified as previously described25 and labeled with AlexaFluor488 according to manufacturer’s protocol. Silicon wafers were obtained from Montco Silicon Technologies, Inc. TEMPO-based initiators, 1 and 2, were synthesized by modifying a procedure described in the literature26 (see Supporting Information). Experimental details are given below. Preparation of Monomers. The OEGn-containing styrenic monomers, 3a and 3b, were synthesized by reaction of pchloromethylstyrene with commercially available monomethoxy oligo(ethylene glycol) (99.6% diol free) derivatives in anhydrous THF and in the presence of NaH according to a procedure described elsewhere27 (see Supporting Information). Preparation of |-[STY-PEGn]x Brushes. The preparation of |-[STY-PEG3]34 is described as a representative example. A (22) Andruzzi, L.; Hexemer, A.; Li, X.; Ober, C. K.; Kramer, E.; Galli, G.; Chiellini, E.; Fisher, D. Langmuir 2004, 20, 10498-10506. (23) Youngblood, J. P.; Andruzzi, L.; Ober, C. K.; Hexemer, A.; Kramer, E. J.; Callow, J. A.; Finlay, J. A.; Callow, M. E. Biofouling 2003, 19, 91-98. (24) Ilic, B.; Craighead, H. G. Biomedical Microdevices 2000, 2, 317322. (25) Posner, R. G.; Lee, B.; Conrad, D. H.; Holowka, D.; Baird, B.; Goldstein, B. Biochemistry 1992, 31, 5350-5356. (26) Husseman, M.; Malmstro¨m, E. E.; McNamara, M.; Mate, M.; Mecerreyes, D.; Benoit, D. G.; Hedrick, J. L.; Mansky, P.; Huang, E.; Russell, T. P.; Hawker, C. J. Macromolecules 1999, 32, 1424-1431. (27) Andruzzi, L.; Senaratne, W.; Hexemer, A.; Ober, C. K.; Kramer, E. J. PMSE. Prepr. (ACS, Division of PMSE) 2003, 88, 604-605.
Biocompatible Polymer Brushes round-bottom flask was charged under nitrogen with a small magnetic bar, 1 × 1 cm2 silicon wafer modified with a TEMPObased initiator 1, 8 mg (0.03 mmol) of free TEMPO-based initiator 2, 0.5 g (1.04 mmol) of styrenic monomer 3a, and 0.5 mL of freshly distilled diglyme. The free initiator was used to allow parallel formation of polymer in solution and thereby estimation of the molecular weight of the grafted chain. Then, the mixture was subjected to three freeze-thaw cycles and subsequently reacted at 125 °C for 48 h under nitrogen. After the polymerization was quenched in ice water, the silicon wafer was separated from the liquid feed, washed repeatedly with chloroform (CHCl3), and finally ultrasonicated in CHCl3 for 2 min. The dried wafer was then characterized by water contact angle measurements. In parallel, the soluble polymer was precipitated twice into hexane from CHCl3, vacuum-dried, and analyzed by GPC in THF. Preparation of |-[PS]y Brushes. Above experimental process was used with a 1 × 1 cm2 silicon wafer modified with a TEMPObased initiator 1, 12 mg (0.046 mmol) of free TEMPO-based initiator 2, and 0.5 g (4.8 mmol) of freshly distilled styrene. Preparation of |-[PS]y-[STY-PEGn]x Brushes. Above experimental process was used with a TEMPO end-capped polystyrene brush modified silicon chip, 12 mg (0.046 mmol) of free TEMPO-based initiator 2, 0.5 g (1.04 mmol) of styrenic monomer 3a, and 0.5 mL of freshly distilled diglyme. Synthesis of 4n and Preparation of |-OEGn Assemblies. Syntheses of 4n and the surface-attached OEGn assemblies (n ) 3, 5, 12, 17, 44) were previously reported (see Supporting Information).28 Protein Adsorption Studies. Fluorescently labeled collagen, fibronectin, BSA, or IgG antibodies were diluted with phosphatebuffered saline (PBS: 10 mM sodium phosphate pH 7.4, 150 mM sodium chloride, 1 mM sodium EDTA) (all with the dye/protein ratios between 4 and 5). The experimental procedure is described in Supporting Information. Mast Cell Adhesion Studies. RBL cells were maintained and harvested as previously described.29 Prior to harvesting, the cells were sensitized with fluorescently labeled anti-DNP IgE (1 µg mL-1) for 4 h and washed to remove unbound IgE. IgE binds irreversibly to its high affinity receptor during the time frame of this experiment.25 The cells were then plated onto the 1 × 1 cm2 silicon chips and prepared as described below at a density of 1 × 106 cells mL-1 in minimum essential medium (MEM, fetal bovine serum, 20% FBS, L-glutamine, gentamicin). Prior to plating the cells, the silicon chips were placed in tissue culture dishes immersed in 70% ethanol for 20 min in a sterile environment. Then the silicon chips were washed with deionized (DI) water followed by growth medium. Each silicon chip was incubated with a 2 mL aliquot of the sensitized cells in growth medium and incubated at 37 °C, for 3 h. For comparison, cells were seeded onto polystyrene spin-coated wafers, glass, and untreated SiOX. Unattached cells were removed by aspiration, and the substrates were washed once with medium by adding 2 mL to the outside of the silicon chip. Then the samples were carefully inverted and gently swirled in the medium. Afterward these were turned upright and washed once again with 2 mL of medium and finally with 2 mL of buffered saline solution (BSS: 20 mM HEPES, pH 7.4, 135 mM sodium chloride, 5 mM potassium chloride, 1.8 mM calcium chloride, 1 mM magnesium chloride, and 5.6 mM glucose). The cells were fixed using 3.7% formaldehyde in PBS for 10 min and followed by additional washing with PBS. The samples were mounted on glass slides and characterized for cell adhesion with fluorescence microscopy as described for the protein adhesion studies, except that the average density of adherent cells was determined for a particular substrate. CHO Cell Adhesion Studies. To further verify resistance toward cell adhesion to these polymer brushes, experiments were carried out using a fibroblast cell line, Chinese Hamster Ovary (CHO) cells that have been transfected with an IgE receptor.30 (28) Senaratne, W.; Andruzzi, L.; Sheets, E. D.; Holowka, D.; Ilic, B.; Hexemer, A.; Baird, B.; Kramer, E. J.; Ober, C. K. PMSE Prepr. (ACS, Division of PMSE) 2003, 88, 337-338. (29) Sheets, E. D.; Holowka, D.; Baird, B. J. Cell Biol. 1999, 145, 877-887. (30) Vonakis, B. M.; Chen, H.; Haleem-Smith, H.; Metzger, H. J. Biol. Chem. 1997, 272, 24072-24080.
Langmuir, Vol. 21, No. 6, 2005 2497 The same protocol described above was used for labeling them with anti-DNP IgE (1 µg mL-1) and used at a density of 1 × 106 cells mL-1 in F-12 nutrient mixture (Ham’s F-12 containing 0.5 g L-1 geneticin G418) at 37 °C for 3 h. The experiment was carried out on polymer brushes |-[STY-PEGn]x and |-[PS]y-[STY-PEGn]x (n ) 3, 7) and polystyrene spin-coated wafers, glass, and untreated SiOX as controls as described above. Cell Localization Studies of Patterned Surfaces. The patterned polymer brushes were incubated with a dinitrophenyl derivatized BSA (DNP/BSA)31 protein solution, 20 µg mL-1 in a dark, humidified environment at room temperature for 1 h. The samples were washed thrice with PBS and incubated with IgEsensitized cell suspensions (1 × 106 cells mL-1) in BSS containing 1 mg mL-1 BSA for 1 h. Then the samples were fixed and washed with BSS/BSA after the aspiration of unattached cells. The patterned polymer brushes were directly incubated with IgEsensitized CHO cell suspension (0.5 × 106 cells mL-1) in medium for 3 h at 37 °C without the prior addition of the DNP-BSA. Then the samples were fixed and washed with medium (3×) and then BSS after the aspiration of unattached cells. Characterization. Gel permeation chromatography (GPC) was carried out using four Waters Styragel HT columns operating at 32 °C. Contact angles were measured using a NRL contact angle goniometer model 100-00 (Rame´-Hart Inc.) at 20 °C. The NEXAFS experiments were carried out on the NIST/Dow materials characterization end station on the U7A beamline at the National Synchrotron Light Source at Brookhaven National Laboratory (for further details see Supporting Information).
Results and Discussion Major requirements for a biological device such as a sensor are the stability and specificity of the attachment of biomolecules to selected regions of a patterned surface with negligible nonspecific interactions in adjacent regions. PEG is commonly used to prevent nonspecific interaction with proteins or cells. This work centers around the fabrication of surfaces modified with OEGn-containing polymer brushes and the investigation of their usefulness in biotechnology applications. A variety of living polymerization methods including nitroxide-mediated polymerization along with ATRP32,33 and reversible additionfragmentation chain transfer (RAFT),34 free radical,15 anionic,12 and cationic polymerizations11,35 have all been widely used for production of surface-tethered poly(styrene), poly(methyl methacrylate), or poly(methylacrylate) brushes with controlled architecture and defined surface morphology. In this work a TEMPO-based silicon oxideanchored initiator 1 in conjunction with a “free” TEMPObased initiator 2 was synthesized (according to general procedures reported in the literature26,22) and used for the surface-initiated polymerization of OEGn-containing styrene monomers (Scheme 1). Both homopolymer and diblock copolymer OEGn-containing brushes were prepared. 4-(Oligoethyleneoxy)oxymethylstyrene monomers were prepared by reaction of 4-chloromethylstyrene and monomethoxy oligo(ethylene glycol) derivatives of varying lengths (n ) 3, 7) in the presence of sodium hydride. Surface-tethered homopolymer brushes, |-[STY-PEGn]x, were produced by reaction of surface-bound initiator 1 in the presence of an OEGn-containing styrene monomer and a controlled amount of “free” initiator 2 at 125 °C for 48 h, as shown in Scheme 1. The “free” initiator was used (31) Xu, K.; Holowka, D.; Baird, B. J. Immunol. 1998, 160, 32253235. (32) Ejaz, M., S.; Yamamoto, S.; Ohno, K.; Tsujii, Y.; Fukuda, T. Macromolecules 1998, 31, 5934-5936. (33) Matyjaszewski, K.; Miller, P. J.; Shukla, N.; Immaraporn, B.; Gelman, A.; Luokala, B. B.; Siclovan, T. M.; Lickelbick, G.; Vallant, T.; Hoffmann, H.; Pakula, T. Macromolecules 1999, 32, 8716-8724. (34) Tsujii, Y.; Ejaz, M.; Sato, K.; Goto, A.; Fukuda, T. Macromolecules 2001, 34, 8872-8878. (35) Zhao, B.; Brittain, W. J. Macromolecules 2000, 33, 8813-8820.
2498
Langmuir, Vol. 21, No. 6, 2005
Andruzzi et al.
Scheme 1. Synthesis of Tethered OEGn-Containing Polymer Brushes and |-OEGn Assemblies on SiOX
to control the polymerization process both at the surface and in the polymerization feed and also to allow parallel formation of homopolymer in solution for characterization purposes and comparison with the parent surface-grafted homopolymer.22 A small amount of diglyme was used as a solvent to prevent formation of a viscous gel during reaction. After the reaction was quenched in ice water, the brush-modified surface |-[STY-PEGn]x was separated from the polymerization feed, repeatedly washed with CHCl3, and finally sonicated for 2 min in CHCl3. This procedure provided clean smooth surfaces as characterized by atomic force microscopy (data not shown). Surfacetethered block copolymer brushes, |-[PS]y-[STY-PEGn]x, were similarly produced by reaction of a surface-bound TEMPO end-capped polystyrene macroinitiator |-[PS]x in the presence of an OEGn-containing styrenic monomer and a controlled amount of “free” initiator at 125 °C for 48 h, as shown in Scheme 1. Also in this case the “free” initiator was used to control the polymerization process both at the wafer surface and in the polymerization feed and to allow formation of polymer in solution for characterization purposes.22 GPC analysis was carried out on all polymers formed in solution and showed a low polydispersity Mw/Mn ) 1.111.19 (Table 1). The brush-modified silicon wafers were characterized by water contact angle measurements (Table 1). The tethered homopolymer brushes (both |-[STYPEG3]34 and |-[STY-PEG7]25) and the polystyrene brush
|-[PS]128 showed advancing contact angle values of about 48° and 94°, typical of PEG and polystyrene surfaces, respectively. The tethered block copolymer brushes |-[PS]128-[STY-PEG3]40 and |-[PS]128-[STY-PEG7]32 also possessed a contact angle of about 48° typical of PEG surfaces. For all PEG polymer brushes, the hysteresis between the advancing and receding contact angles was about 10° suggesting that the brushes were subjected to surface reorganization upon exposure to water. Captive air bubble measurements resulted in contact angles in the range of 35-40°, indicating that sample exposure to water allows the PEG side chains to rearrange at the film surface. Ellipsometry measurements yielded polymer brush thickness values z* in the range of 10-15 nm and the ratio z*/Rg was estimated for all brushes based on the radius of gyration, Rg, of the same homopolymer or block copolymer chains that were grafted. The molecular weights were determined for the polymers formed from free initiator in the same solution and were polystyrene equivalent molecular weights. Rg was roughly estimated as ax(N/6), where N is the degree of polymerization of the PS equivalent chain and a is the statistical segment length of PS (about 0.7 nm). The values of z*/Rg were all above 1 as shown in Table 1, suggesting that the brushes are stretched. The large ratio of brush thickness to radius of gyration, as well as the dramatic change in the contact angle in passing from a polystyrene brush to the corre-
Biocompatible Polymer Brushes
Langmuir, Vol. 21, No. 6, 2005 2499
Table 1. Characterization Data of Surface-Tethered Polymer Brushes and OEGn Assemblies sample |-[A]n-[B]pa |-[ STY-PEG3]34 |-[ STY-PEG7]25 |-[ PS]128 |-[PS]128-[STY-PEG3]40 |-[PS]128-[STY-PEG7]32 |-OEG3 |-OEG7 |-OEG12 |-OEG17 |-OEG44
Mnb (kg/mol)
Mw/Mnb
13 12 13 15 16 n. d.g n. d. n. d. n. d. n. d.
1.19 1.19 1.11 1.19 1.19 n. d. n. d. n. d. n. d. n. d.
z*c (nm)
θAd (deg)
θRd (deg)
θaire (deg)
10 11 8 16 15 2 2 3 4 6
48 47 94 49 48 65 54 50 46 42
35 36 80 37 36 50 40 39 37 30
40 38 87 38 35 50 45 42 39 39
z*/Rg(calc)f 6.02 7.74 3.23 4.32 4.15 n. d. n. d. n. d. n. d. n. d.
a Degrees of polymerization (n, p) are estimated from GPC polymerization degrees of homopolymers separated from polymerization feed or, in case of insoluble homopolymers, from the fraction of unreacted monomer. b Molecular weight and polydispersity index of the last block to be grafted determined by GPC in THF of homopolymers separated from polymerization feed. c Thickness of the brush determined by ellipsometry. d Contact angles determined by goniometry measurements with water as wetting liquid. e Contact angles determined by goniometry measurements on a captive air bubble deposited on the film surface under water. f z* is the thickness of the brush and Rg is the radius of gyration of the polymer as formed in solution and is estimated as ax(N/6) with a ) 0.7 being the statistical segment length (a) of polystyrene and N the degree of polymerization. In the case of block copolymers, Rg is calculated as 〈Rg2〉 ) 〈Rg2〉A + 〈Rg2〉B. g n.d. ) not determined.
sponding OEGn-containing block copolymer pointed to high brush density and high surface coverage. OEGn assemblies on silicon oxide were prepared according to a synthetic procedure that is also outlined in Scheme 1. Characterization data are reported in Table 1. The dry layer thickness of the assemblies was about 2-6 nm. The thicknesses for OEGn (n ) 7-44) are smaller than the estimated lengths of the fully extended all-trans configuration of the molecule but are similar to the range of thicknesses reported by others for different OEGn-based assemblies on silicon oxide.9,36 The measured water contact angle values ranged between 30° and 65°. This difference in contact angle probably reflects interaction between water and ethylene glycol that moderates the effect of the terminal group. Methoxy-terminated OEG3 has shown larger contact angles as reported by others as well.9,37 The hysteresis between the advancing and receding water contact angles was about 15° and those of captive air bubble contact angles were in the range 40-50°, suggesting that the OEGn assemblies were also subject to surface reorganization upon exposure to water. It should be noted that the captive air bubble contact angles of polymer brushes were about 10° smaller than corresponding OEGn assemblies. This indicates a faster surface reorganization of polymer brushes upon exposure to water, likely due to a more dense and homogeneous OEGn surface coverage. NEXAFS Characterization of Surface-Tethered Brushes. The surface structure of the surface-tethered polymer brushes was studied by near-edge X-ray absorption fine structure (NEXAFS) analysis. This technique examines the upper 3 nm of a surface and provides information on bond type and molecular orientation of the chemical groups populating this surface region.38 Figure 1 shows representative partial electron yield (PEY) versus photon energy data at an angle θ ) 55° between the polarization vector of the X-ray beam and the surface normal of polymer brushes |-[STY-PEG7]25 and |-[PS]128[STY-PEG7]32 as compared to the parent |-OEG7 assembly and to a spun-cast [STY-PEG7]25 homopolymer film of 100 nm thickness as a reference. The angle used for the analysis, 55°, represents the magic angle where the (36) Yang, Z.; Galloway, J. A.; Yu, H. Langmuir 1999, 15, 84058411. (37) Benesch, J.; Svedhem, S.; Svensson, S. T.; Valiokas, R.; Liedberg, B.; Tengvall, P. J. Biomater. Sci., Polym. Ed. 2001, 12, 581-597. (38) Xiang, M.; Li, X.; Ober, C. K.; Char, K.; Genzer, J.; Sivaniah, E.; Kramer, E. J.; Fisher, D. A. Macromolecules 2000, 33, 6106-6119.
intensity of the PEY peaks is independent of molecular orientation, and it was chosen because no molecular orientation of the OEGn side chains was expected to occur at the surface. The polymer brushes and the spun-cast reference film showed the same type of spectrum, indicating the presence of a uniform OEGn surface. In particular, these spectra showed a peak at 284.5 eV due to the 1s f π* transition of the phenyl ring and peaks at 289.5 and 293 eV, due to the 1s f σ* transitions of the C-O and C-C bonds, respectively. In the case of block copolymer brushes, these results support the idea that the brushes were dense enough that the outermost OEGn-containing block completely covers the underlying polystyrene block despite its much higher surface energy. In contrast, the |-OEG7 assembly showed a broad signal between 287 and 295 eV, as well as a low intensity peak at 284.5 eV likely due to silicon oxide background absorption. The lack of definite signals for the C-O and C-C bonds in the 287295 eV region is likely due to a less dense and less homogeneous OEGn surface coverage in the |-OEGn assemblies with respect to the polymer brushes. Protein Adsorption Studies. Studies were carried out with representative proteins of the sort used in biosensor applications. These proteins spanned a wide range of sizes and, depending on their isoelectric point (pI), are expected to have different properties at solvent accessible surfaces. Fluorescently labeled proteins were incubated with surface-bound polymer substrates. Fluorescence data for the adsorption of fibronectin (500 kDa, pI 5.5-6.3), collagen (340 kDa, pI 4-5), IgG antibodies (150 kDa, pI 6.1-8.5), and BSA (65 kDa, pI 5.2) on the polymer brushes and OEGn assemblies are shown in Figure 2. The fluorescence of the samples for each set of proteins was normalized to 100% corresponding to a positive SiOX control. Polymer brushes |-[STY-PEG3]34 and |-[PS]128-[STY-PEG3]40 resisted adsorption of all four proteins. On the other hand, the |-OEGn assemblies showed high protein adsorption with short chain length OEGn units but showed decreased protein adsorption with longer chain length OEGn units. The OEG3 showed only 35-45% reduction in adsorption to larger proteins IgG, collagen, and fibronectin but lower adsorption to the smallest protein BSA. One explanation for the higher adsorption could be the film quality of the OEG3 surface; the difficulties associated with silane chemistry could provide poor coverage as well as irreproducibility of surfaces. All other OEGn films showed >70% reduction to protein adsorption. But the surface-tethered polymer
2500
Langmuir, Vol. 21, No. 6, 2005
Andruzzi et al.
Figure 1. NEXAFS partial electron yield (PEY) intensity vs X-ray photon energy, at an angle θ ) 55° between the electric field vector of the polarized X-rays and the sample normal at EGB ) -150 V, of (a) |-[STY-PEG7]25, (b) |-[PS]128-[STY-PEG7]32, (c) |-[OEG]7, and (d) spun-cast [STY-PEG7]25.
brushes with different chain length ethylene glycol units performed better in inhibiting protein adsorption (3 h) to promote adhesion. Here we observed that a 3 h period was sufficient to see strong adhesion to the SiOX regions on the patterned PEG brush surfaces. Fluorescence micrographs of patterns with different feature sizes for CHO cells are shown in Figure 4d-f. As for the RBL cells, some CHO cells were observed
Biocompatible Polymer Brushes
Langmuir, Vol. 21, No. 6, 2005 2503
Figure 4. Schematic of the parylene patterning carried out on the polymer brush modified surface. Surface-initiated polymerization is first carried out on the SiOX surface and then parylene is vapor deposited onto the substrate. Standard photolithography process yields a fabricated surface. For subsequest biological material testing, the parylene film was mechanically peeled off the substrate yielding a polymer brush patterned surface. (top) Optical micrograph of the parylene deposited surface after pattern fabrication but prior to mechanical peeling off of the film. (bottom) Optical micrograph shows the fabricated pattern after peeling of the parylene. (a-f) Fluorescence micrographs of mast cell and CHO cell attachment to patterned |-[STY-PEG3]34 (a-c) and |-[PS]128[STY-PEG3]40 (d-f) polymer brush surfaces having patterned lines of SiOX. (a-c) These patterned substrates were exposed to DNP-BSA, an antigenic protein, and then exposed to fluorescently labeled anti-DNP IgE sensitized RBL cells. The patterned regions consist of (a) 10 µm lines/40 µm pitch, (b) 50 µm lines/100 µm pitch, and (c) 90 µm lines/140 µm pitch. Epifluorence micrographs were imaged at room temperature with 60 × 1.4 NA oil immersion objective. (d-f) Patterned substrates were exposed to fluorescently labeled anti-DNP IgE sensitized CHO cells (without DNP-BSA). The patterned regions consist of (d) 10 µm lines/40 µm pitch, (e) 50 µm lines/100 µm pitch, and (f) 90 µm lines/140 µm pitch. Epifluorescence micrographs were imaged at room temperature with a 20 × 0.7 NA objective.
2504
Langmuir, Vol. 21, No. 6, 2005
in the PEG spaces between the 10 µm SiOX lines. For the wider features, the patterned surfaces showed a high degree of control by localizing cells in the SiOX regions. Conclusions OEGn-containing styrene-based homopolymer and block copolymer brushes were produced with controlled chemical architecture and high surface coverage using surfaceinitiated nitroxide-mediated polymerization on a SiOX surface modified with a chlorosilyl-functionalized TEMPO derivative. Biocompatibility of these surfaces was studied by comparison between the polymer brushes and OEGn assemblies and controls. Adsorption of proteins and RBL cells was studied on these surfaces using fluorescence microscopy. Polymer brushes possessed a superior ability to inhibit protein (less than 6%) and cell (less than 3%) adhesion when compared to surface assemblies with the same OEGn length. This was attributed to the greater thickness and surface coverage of polymer brushes with respect to monolayers. Longer monolayer assemblies show a >70% reduction in protein adsorption but short chain OEGn behave somewhat poorly. We believe that the very nature of silane chemistry may lead to inhomogeneous coverage and therefore nonreproducible OEGn assembly surface properties. In contrast, the surface-initiated polymerization of STY-PEGn monomers represents an easy and straightforward method of creating a homogeneous PEG surface. The “grafting from” nature of the SIP method together with its “living” character also gives us the ability to create thicker and more robust films. Cell localization studies were carried out on polymer brush surfaces patterned by a modified photolithographic method. Patterned surfaces with feature sizes varying between 10 and 90 µm initially exposed to DNP-BSA and then to anti-DNP IgE sensitized RBL cells, showed specific cell localization on the DNP-BSA regions. Similarly, CHO cells preferred to adhere to the SiOX micropatterns rather than the PEG brush regions. These results provide strong
Andruzzi et al.
evidence that the OEGn-containing polymer brushes described in this paper have potential in the fabrication of devices for the use as biological and chemical microand nanodevices. Acknowledgment. This work was supported by the National Science Foundation, Division of Materials Research, Polymers Program (Grant Nos. DMR-0307233 and DMR-0208825), Chrysalis Inc., and in part by the Nanobiotechnology Center (NBTC), an STC Program of the National Science Foundation under Agreement No. ECS9876771. We greatly appreciate the advice and discussions with Dr. David Holowka and for the invaluable assistance in mast cell and CHO cell studies. We also greatly appreciate the use of the fluorescence microscope at the Microscopy and Imaging Facility (MIF) in the BioResource Center as well as helpful discussions with Carol Bayles (MIF, Cornell University). The use of the facilities of the Cornell Nanofabrication Facility (CNF) and the NIST/ Dow materials end-station at the National Synchrotron Light Source at Brookhaven National Laboratory is acknowledged. Both the Cornell Center for Materials Research and the Materials Research Laboratory of UCSB (both funded by the NSF-DMR-MRSEC Program), are acknowledged for partial support of this work. Dr. Xuefa Li (Cornell University and presently at Argonne National Laboratories) is also thanked for his help with NEXAFS measurements and Ms. Norah Smith for the assistance with the CHO cell labeling. Supporting Information Available: Preparation of surface-bound TEMPO-based initiatiors, preparation of 4-triethyleneoxymethylstyrene, synthesis of methyl(undec-10-en-1yl) oligo(ethylene glycol), synthesis of methyl[(1-trichlorosilyl)undec-11-yl] oligo(ethylene glycol), preparation of OEGn assemblies, protein adsorption studies, parylene depositions and pattern fabrication, and procedure details. This material is available free of charge via the Internet at http://pubs.acs.org. LA047574S