Functionalizable and nonfouling zwitterionic carboxybetaine hydrogels with a carboxybetaine dimethacrylate crosslinker

Biomaterials 32 (2011) 961e968

Contents lists available at ScienceDirect

Biomaterials journal homepage: www.elsevier.com/locate/biomaterials

Functionalizable and nonfouling zwitterionic carboxybetaine hydrogels with a carboxybetaine dimethacrylate crosslinker Louisa R. Carr, Hong Xue, Shaoyi Jiang* Department of Chemical Engineering, University of Washington, Box 351750, Seattle, WA 98195, USA

a r t i c l e i n f o

a b s t r a c t

Article history: Received 19 July 2010 Accepted 30 September 2010 Available online 20 October 2010

We have introduced a dimethacrylate carboxybetaine-based crosslinker that has excellent compatibility with zwitterionic hydrogel systems. Poly(carboxybetaine methacrylate) (pCBMA) hydrogels prepared with the new CBMA crosslinker (CBMAX) result in considerably improved solubility, homogeneity, and mechanical properties (up to 8 MPa compressive modulus) over those prepared with the commercially available N,N’-methylenebis(acrylamide) (MBAA) crosslinker. The zwitterionic nature of the CBMAX crosslinker provides continuity of ordered hydration in the CBMA hydrogel and retains its nonfouling properties. CBMAX-crosslinked CBMA hydrogels had lower cell fouling than MBAA-crosslinked CBMA hydrogels and in fact reduced cell adhesion by about 90% relative to pHEMA hydrogels. Furthermore, unlike pHEMA, CBMA hydrogels are readily functionalizable. Cell adhesion on nonfouling CBMA hydrogels was controlled by cRGD functionalization. Ó 2010 Elsevier Ltd. All rights reserved.

Keywords: Hydrogel Mechanical properties Biocompatibility Cross-linking

1. Introduction Hydrogels have long been of interest for biological and biomaterial applications due to their high water content, which mimics the interstitial tissue environment, ensures high diffusive permeability, and provides biomimetic mechanical strengths [1e4]. Several excellent reviews provide complete lists and discussion of attributes of commonly used biomaterials for hydrogels [4e6]. Particular interest has been given to PEG and poly(2-hydroxyethyl methacrylate) (pHEMA) [1] hydrogels because, in addition to the general properties of hydrogels, they are also commonly considered be low fouling, bioinert, and versatile. pHEMA hydrogels have found use in and been studied for applications such as contact lenses, artificial cornea, drug delivery vehicles, cartilage substitutes, and tissue scaffolds, among others [1,7e9]. The hydration of pHEMA, however, is lower than that of native tissue, and its fouling, while low, is higher than other nonfouling materials. Furthermore, pHEMA functionalization via the hydroxyl group is generally difficult. PEG hydrogels are routinely used, and can only be modified for applications that require a bioinert background with specific added bioactive functionalities for controlled in vitro and in vivo uses when additional functional groups are introduced into PEG hydrogels. However, it has been found that PEG is subject to oxidation [10e15]. The susceptibility of PEG to oxidative damage reduces its utility for applications * Corresponding author. Tel.: þ1 206 616 6509; fax: þ1 206 543 3778. E-mail address: [email protected] (S. Jiang). 0142-9612/$ e see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2010.09.067

that require long-term material stability. For applications in which maximal biological stability and nonfouling are required, however, PEG-based materials are insufficient. Recently, we demonstrated [16e25] that zwitterionic compounds, including poly(carboxybetaine methacrylate) (pCBMA, Scheme 1, structure 1), are ultra-low-fouling, meaning that surfaces coated with these polymers allow less than 5 ng/cm2 protein adsorption [26e28]. It was also demonstrated that surfaces coated with zwitterionic poly (carboxybetaine methacrylate) highly resist non-specific protein adsorption, even from undiluted blood plasma and serum [18,19,22,29], and prohibit long-term bacterial colonization by Pseudomonas aeruginosa for up to 10 days at room temperature [25,30]. The ultra-low-fouling of zwitterionic materials is due to high hydration around the opposing charges and the high energetics required to remove that hydration layer [31,32]. Furthermore, CBMA is functionalizable through conventional EDC/NHS chemistry. Because of the high hydration and ultra-low fouling properties of zwitterionic materials, zwitterionic hydrogels are of interest as hydrogels with superior suitability for biomedical applications. In previous work on zwitterionic hydrogels, we have demonstrated low protein adhesion on sulfobetaine methacrylate (SBMA) and mixed charge hydrogels [16,17], and low cell adhesion on carboxybetaine methacrylate gels [16]. The zwitterionic hydrogels studied so far, however, have shown low mechanical strength [16,33], which limits their potential biological uses. Efforts are needed to improve the mechanical properties of functionalizable pCBMA hydrogels.

962

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

In this paper, we describe the synthesis of a CBMA-based crosslinker, CBMAX, and examine and compare the MBAA- and CBMAX-crosslinked pCBMA hydrogels by measuring their equilibrium water content, compressive moduli, functionalizability, and their ability to resist non-specific cell adhesion. The unlimited solubility and compatibility of the CBMAX crosslinker is expected to improve the mechanical properties of the pCBMA hydrogel by increasing the crosslinker content, and will not compromise the nonfouling of the polymer chains in the hydrogel. 2. Experimental methods 2.1. Materials

Scheme 1. Chemical structures of carboxybetaine monomer and crosslinkers. The monomer used in this study is carboxybetain methacrylate (CBMA), 1. The crosslinkers used are N,N’-methylenebis(acrylamide) (MBAA), 2, and carboxybetain dimethacrylate (CBMAX), 3.

Another fundamentally limiting feature of these zwitterionic hydrogels is the dearth of hydrophilic crosslinkers. The most commonly used of the commercially available “hydrophilic” crosslinker is N,N’-methylenebis(acrylamide) (MBAA, Scheme 1, structure 2). Water-soluble at very low concentrations, this crosslinker is only moderately soluble at crosslinker concentrations around 10%, especially in the salt solutions that are ideal for zwitterionic hydrogel formation. Additionally problematic for polymerization with pSBMA and pCBMA is the inherent incompatibility of the polymerizable moieties: the greatly different chemical structure of the crosslinker may result in poor incorporation into the growing methacrylate polymer chains. Perhaps the most unacceptable feature of MBAA as a crosslinker in zwitterionic hydrogels is that it does not structure water the way the zwitterionic monomers do. Structured water around the opposing charges in a zwitterionic material provides the nonfouling mechanism [31,32]; MBAA will disrupt the ordered water and present locations where proteins, bacteria, and even cells, may bind and foul the hydrogel. Furthermore, the MBAA crosslinker is not functionalizable. In order to improve the mechanical properties of nonfouling pCBMA hydrogels without using a fouling crosslinker, we designed and synthesized a CBMA-based dimethacrylate crosslinker, CBMAX. The structure of this crosslinker is shown in Scheme 1, structure 3. This chemical structure is identical to the CBMA monomer except that there is a one carbon spacer between the cationic quaternary amine and the anionic carboxyl group instead of two, and that one of the two methyl groups on the quaternary amine is replaced by a second methacrylate group. It is hypothesized that this crosslinker will be incorporated more efficiently into the growing pCBMA chains because of its chemical similarity to CBMA, relative to MBAA. It also exhibits excellent solubility in 1 M salt solutions, and its zwitterionic group ensures that it, unlike MBAA, will provide continuity of structured water across the crosslinks in a pCBMA hydrogel. Goda et al. [34,35], studied the water structure of zwitterionic phosphorylcholine (PC)-based hydrogels made with a custom-made phosphorylcholine-based crosslinker. They reported a qualitative difference in the water in PC-crosslinked zwitterionic hydrogels compared to MBAA-crosslinked zwitterionic hydrogels; hydrogels made from zwitterionic monomers and zwitterionic crosslinkers had more ice-like bound water than hydrogels made from zwitterionic monomers and MBAA. They attributed improved mechanical properties of the zwitterionically crosslinked zwitterionic hydrogels to the increased order of the water.

N,N’-Methylenebis(acrylamide), ammonium persulfate, sodium metabisulfite, 2-(N-morpholino)ethanesulfonic acid (MES), methacrylic acid, ion exchange resin (IRA 400 OH form), and phosphatebuffered saline were purchased from Sigma Aldrich (St. Louis, MO). Ethanol was purchased from Decon Labs (King of Prussia, PA), acetonitrile and diethyl ether from EMD Biosciences (Gibbstown, NJ), ethylene glycol from VWR (West Chester, PA), tetraethylene glycol dimethacrylate from Polysciences (Warrington, PA), 2-hydrolxyethyl methacrylate, t-butyl bromoacetate, and N-cyclohexyl-2-aminoethanesulfonic acid (CHES) from TCI America (Portland, OR), and N,N,N’,N’-Tetramethylethylenediamine (TEMED) from Bio-Rad Laboratories (Hercules, CA). N-Methyldiethanolamine, triethanolamine, methanesulfonic acid, and trifluoroacetic acid were purchased from Acros Organics (Morris Plains, NJ). Dulbecco’s Modified Eagle Medium, fetal bovine serum, nonessential amino acids, and penicillin-streptomycin were purchased from Invitrogen Corp (Carlsbad, CA). Cyclo(Arginine-Glycine-Asparginine-D-Tyronsine-Lysine) peptide (cRGD) was purchased from Peptides International (Louisville, KY). Hydrogen peroxide and sodium chloride salt were purchased from J.T. Baker (Phillipsburg, NJ), and ImmunoPureÒ o-Phenylenediamine dihydrochloride was purchased from Pierce (Rockford, IL). Horseradish peroxidase (HRP)-conjugated antifibrinogen was purchased from US Biological (Swampscott, MA). COS-7 cells (African Green Monkey fibroblast cells) were purchased from the American Tissue Culture Collection (Manassas, VA). All water used had been purified to 18.2 mU on a Millipore Simplicity water purification system. 2.2. Synthesis of carboxybetaine monomer and carboxybetainebased crosslinker 2.2.1. Synthesis of the CBMA monomer 2-Carboxy-N,N,-dimethyl-N-(20 -(methacryloyloxy)ethyl) ethanaminium inner salt (carboxybetaine methacrylate, CBMA) was synthesized as previously reported [21]. 2.2.2. N-Methyl-N-di(ethanolamine methacrylate) (1) For a schematic of the synthesis of the CBMA crosslinker, see Scheme 2. N-Methyldiethanolamine (11.9 g, 0.1 mol), toluene (100 ml), hydroquinone (2.0 g), and methacrylic acid (21.5 g, 0.25 mol) were added to a 500 ml reaction flask fitted with a stirrer, condenser, and Dean-Star trap. Methanesulfonic acid (14.4 g, 0.15 mol) was added slowly and the mixture was heated to reflux. After the theoretical water of reaction was collected azeotropically, the solution was cooled to room temperature. The mixture was then neutralized with 25 wt% aqueous sodium hydroxide and the organic phase was washed with 10% brine solution and dried over anhydrous magnesium sulfate. The solution was filtered, and the filtrate was further treated with activated carbon and basic alumina. The resulting solution was removed under vacuum to give a colorless oil with 73% yield. 1H NMR

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

963

Scheme 2. Chemical synthesis of carboxybetaine dimethacrylate (CBMAX) (3).

(CDCl3), d: 6.10(s, 2H), 5.56(s, 2H), 4.25(t, J ¼ 6.0 Hz, 4H), 2.78(t, J ¼ 6.0 Hz, 4H), 2.39(s, 3H), 1.94(s, 6H). 2.2.3. N-Methyl-N-di(2-methacryloyloxy-ethyl)-N-1-(tbutyloxycarbonylmethyl) ammonium bromide (2) Compound (1) (12.75 g, 50 mmol), t-butyl bromoacetate(11.70 g, 60 mmol), acetonitrile(100 ml) were mixed in a nitrogen-filled flask. The mixture was stirred at 60  C for 2 days. The solvent was evaporated under vacuum and the residue was washed with ether and dried to obtain a white solid with 90% yield. 1H NMR (CDCl3), d: 6.15(s, 2H), 5.67(s, 2H), 4.80(s, 2H), 4.73(t, J ¼ 6.0 Hz, 4H), 4.47(m, 4H), 3.75(s, 3H), 1.95(s, 6H), 1.48(s, 9H). 2.2.4. 1-Carboxy-N-methyl-N-di(2-methacryloyloxy-ethyl) methanaminium inner salt (3) The tert-butyl ester moiety of compound (2) (13.5 g, 30 mmol) was removed by treatment with trifluoroacetic acid (TFA, 30 ml) in dichloromethane (120 ml) for 2 days at room temperature. The solvent was removed under vacuum and replaced with acetonitrile (120 ml). The solution was neutralized over an ion exchange resin (IRA 400 OH form), subsequently concentrated and precipitated into ether, and finally vacuum dried to obtain a white solid with quantitative yield. 1H NMR (D2O), d: 6.05(s, 2H), 5.66(s, 2H), 4.56(t, 4H), 4.20(m, 2H), 3.95(m, 4H), 3.24(s, 3H), 1.83(s, 6H).

2.3. Hydrogel preparation Monomer solutions were prepared in 1 M NaCl, at a monomer concentration of 65% by weight. To these solutions, the crosslinkers (N’,N’-methylenebis(acrylamide)) (MBAA) or CBMAX was added at amounts ranging from 2 to 23% (molar percent of monomer). CBMAX formulations above this were prepared up to 75% by maintaining a constant total molar amount of monomer and crosslinker, and adjusting their relative molar amounts. 100% CBMAX was simply prepared by dissolving the desired amount of CBMAX in 1 M NaCl. The solutions were mixed by sonication. In some cases (above w10% MBAA), the crosslinker did not dissolve completely. A 40% solution of ammonium persulfate and 15% sodium metabisulfite were added to the solution to initiate polymerization. The solutions were polymerized between glass microscope slides separated by 0.5 or 2 mm-thick polytetrafluoroethylene (PTFE) spacers at 60  C. The gels were then removed from the slides and immersed in phosphate-buffered saline (PBS) to hydrate. This

hydration water was changed daily for 5 days to remove unreacted chemicals and excess salt. Biopsy punches were used to punch hydrated hydrogels into 5 mm-diameter disks. HEMA hydrogels were prepared by mixing 0.78 ml HEMA monomer in 1.5 ml of a mixed solvent comprised of 1 part ethanol, 1.5 parts ethylene glycol, and 1.5 parts water. The crosslinker, tetraethylene glycol dimethacrylate (TEGDMA), was then added (60 ml), and the resulting solution was sonicated to mix well. Finally, 24 ml 40% ammonium persulfate and 7.5 ml TEMED were added. This hydrogel was formed as above. 2.4. Swelling properties of hydrogels Hydrogels were allowed to swell to equilibrium in PBS for five days. Disks with 0.5-cm diameter were cut from the swollen gel cast at 0.5 mm thickness. The disks were weighed and then dehydrated under vacuum at 50  C and 30 in. Hg vacuum for 3 days. Dried hydrogel disks were measured with a caliper and weighed. The volume fraction (F) of polymer within a swollen hydrogel is given by:




42 ¼

D0 D

3 (1)

where Do and D are the diameters of dried and swollen disks, respectively. The volume fraction of polymer in the relaxed (unswollen) hydrogel, 4o, was determined from the monomer/ crosslinker solutions: The volume of the solution was compared to the volume of water added. The difference in volume was taken as the volume of monomer, which corresponds to the volume of polymer after polymerization [36]. The equilibrium water content values were determined as:

mw  md EWC ¼ 100ð%Þ* mw

(2)

where mw is the mass of the wet hydrogel and md is the mass of the dry hydrogel. All samples were measured in triplicate. 2.5. Protein adsorption Evaluated by enzyme-linked Immunosorbent Assay (ELISA) To measure fibrinogen adsorption, hydrogel disks of 0.5 cm diameter (0.5 mm thickness when cast) were incubated with 1 mg/

964

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

ml fibrinogen in a well plate for 15 min at room temperature, followed by many washes with PBS buffer over 4 h. The hydrogels were then removed from the last PBS wash and transferred to new wells. They were next incubated with a 1:500-dilution of horseradish peroxidase (HRP)-conjugated anti-fibrinogen in PBS for 10 min, followed by another 5 washes with the same buffer. Again, the hydrogels were washed repeatedly with PBS buffer for 4 h, removed from the last wash, and transferred to new wells. Finally, 500 ml 1 mg/ml o-Phenylenediamine hydrochloride (OPD) in 0.1 M citrate-phosphate buffer, pH 5.0, containing 0.03% hydrogen peroxide, was added to each hydrogel at 30-s intervals. The samples incubated in the OPD solution for 30 min away from light. The supernatant was removed from each hydrogel disk, transferred to a cuvette, and its absorbance at 492 nm was measured. All samples were measured in triplicate. 2.6. Cell adhesion to hydrogels Three hydrogel disks of 0.5 cm diameter (0.5 mm thickness when cast) were placed individually in the wells of a 48-well plate with 500 ul PBS solution. To sterilize the hydrogels, they were irradiated with UV light for 30 min and refrigerated overnight in 1x penicillin-streptomycin in PBS. COS-7 cells (p ¼ 7) were seeded onto the hydrogels at a concentration of 104 cells/ml in supplemented Dulbecco’s Modified Eagle Medium. Cells were allowed to grow for 72 h at 37  C, 5% CO2, and 100% humidity, after which time the hydrogels were photographed at 10x magnification on a Nikon Eclipse TE2000-U microscope. Photographs were taken at five predetermined areas on the surface of the hydrogel, for a total of fifteen images per hydrogel formulation, and the number of adherent cells from each image was totaled and normalized to the number of cells adhered to pHEMA hydrogels. 2.7. RGD functionalization via EDC/sulfoNHS chemistry CBMA/MBAA and CBMA/CBMAX gels were functionalized with cRGD. Three hydrogel disks of 0.5 cm diameter (0.5 mm thickness when cast) were placed in the wells of a 24-well plate. All gels of each formulation were split into two groups. The gels of each group were placed in the same well, with care taken to ensure that the surface of each hydrogel was uncovered. The hydrogels were incubated in 500 ul MES buffer solution (pH ¼ 5.5, 10 mM MES, 100 mM NaCl) overnight. The MES was then removed from one group of each formulation and replaced with 500 ml of an MESbased buffer solution containing 5 mM sulfoNHS and 100 mM EDC, to activate the surface, for 2 h at room temperature. As a control, the second group of each formulation was incubated with an MESbased buffer solution of 5 mM sulfoNHS, for the same amount of time. The EDC/sulfoNHS and sulfoNHS solutions were then removed from the wells, and the hydrogel disks were washed three times with MES buffer. To all wells was next added 500 ml 1.4 mM cRGD in CHES buffer (pH ¼ 9, 50 mM CHES, 100 mM NaCl), and the reaction was allowed to proceed at room temperature. After 3 h, the hydrogels were washed three times with PBS, allowing 10 min per wash. Finally, hydrogels were transferred to individual wells of a 48-well polystyrene tissue culture plate, with care taken that the functionalized surface remained facing upwards. The hydrogels were sterilized with penicillin-streptomycin overnight as described above, and cell adhesion was performed the next day, also as described above. 2.8. Mechanical strength of hydrogels At least five 0.5 cm diameter disks of each formulation (2 mm thickness when cast) were compressed to failure at a rate of

1 mm/min using an Instron 5543A mechanical tester (Instron Corp., Norwood, MA) with a 10 N load cell. The Young’s modulus was calculated from the initial 10% strain. 2.9. Calculation of physical properties of CBMA hydrogels Hydrogel hydration properties (swollen and relaxed volume fractions) and stress-strain curves can be used to calculate the crosslinker density and mesh size of the different 2-arm CBMAXand MBAA-crosslinked hydrogel formulations. To calculate the crosslinker density (ne/V) of hydrogels, the following equation [36] may be used:

hv iRTa  a2  ss ¼ e V ð42 =40 Þ2=3

(3)

where ss is the stress at a particular strain, in units of Pa, a is the deformation ratio, or the ratio of the deformed length to the original length of a crosslinked hydrogel under compression. R is the universal gas constant, T is absolute temperature, 42 is the volume fraction of polymer at equilibrium (in the fully swollen hydrogel), and 4o is the volume fraction of polymer in the relaxed state (unswollen, but not dehydrated, hydrogel). A plot of ss vs, (a - a2) is linear at low strain, and the crosslink density (ne/V) can be extracted from the slope (ne/V)(RT)(42/4o)2/3. The crosslink density in moles/unit volume can be converted to distance between crosslinks, or mesh size, through equation (4):

x ¼


NA ve 1=3 V

(4)

where NA is Avagadro’s number. Equation (4) inverts moles/volume to volume per mole, converts moles to number of crosslinks, and takes a cube root to obtain distance per crosslink. This value is effectively pore size of the hydrogel. 3. Results and discussion 3.1. Hydration properties of CBMA hydrogels with CBMAX The high water content of hydrogels makes them attractive for biological applications, because this water content often mirrors the water content of biological tissues [7,37]. Equations (1) and (2) allowed us to calculate the swelling properties of CBMA hydrogels; the mass and dimensions of hydrogel disks before and after dehydration were used to calculate the equilibrium water content and volume fraction of gels, respectively. Fig. 1 shows the equilibrium water content (a) and volume fraction (b) measured for MBAAcrosslinked hydrogels up to the solubility limit at 17% crosslinker and for CBMA-crosslinked hydrogels from 2% CBMAX to 100% CBMAX (open and closed bars, respectively). As expected, an increase in crosslinker content results in a decrease of hydration and swelling. At 100% CBMAX, hydration has fallen to about 60% equilibrium water content, a value that is still remarkably high for such a large amount of crosslinking. As these figures show, there appears to be no difference in water content when the CBMA is crosslinked with CBMAX or with MBAA. It is worth noting, however, that Goda et al. [34], also found identical equilibrium water content in PC-based hydrogels made with MBAA or PC crosslinkers. They additionally reported a qualitative difference in the water; hydrogels made from zwitterionic monomers and zwitterionic crosslinkers had more structured water than hydrogels made from zwitterionic monomers and MBAA. This is consistent with hydration studies based on molecular simulations that have come from our previous studies [31,32], showing more

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

a

0.4

CBMAX

Volume Fraction

0.35

965

MBAA

0.3 0.25 0.2 0.15 0.1 0.05 0

EWC (%)

b

100 90 80 70 60 50 40 30 20 10 0

2%

4%

6%

9%

13% 17% 23% 29% 33% 50% 75% 100% Crosslinker Content CBMAX MBAA

2%

4%

6%

9%

13% 17% 23% 29% 33% 50% 75% 100% Crosslinker Content

Fig. 1. Hydration properties of CBMAX- and MBAA-crosslinked CBMA hydrogels. Volume fraction (a) is the amount, by volume, of polymer in a swollen hydrogel, and equilibrium water content (EWC) (b) is the amount, by weight, of water in the swollen hydrogel. Closed bars represent CBMAX-crosslinked CBMA hydrogels, while open bars represent MBAAcrosslinked CBMA hydrogels.

3.2. Nonfouling properties of CBMA hydrogels with CBMAX CBMA hydrogels made with 4% CBMAX, 4% MBAA, 17% CBMAX, and 17% MBAA, and CBMAX hydrogels, were chosen to be tested for nonfouling. The low crosslinker content was chosen because it is the lowest that is practically handled, while the 17% crosslinker hydrogels were chosen because 17% is the upper limit of MBAA incorporation; this composition was expected to most clearly demonstrate any difference between the fouling properties of these two hydrogels. Fibroblast (COS-7) cells were seeded on antibioticsterilized hydrogels and allowed to grow for three days in supplemented growth medium. On the third day, cell adhesion was quantified visually: fifteen microscopy images of each hydrogel formulation were collected, and the absolute numbers of cells adhered to the gels were counted. The number of cells adhere to hydrogels of each formulation are compared quantitatively in Fig. 2, as a function of crosslinker content, and normalized to the number of cells adhere to pHEMA hydrogels. Solid bars represent CBMAX-crosslinked hydrogels, while open bars represent MBAA-crosslinked hydrogels. The rightmost bar represents the control, pHEMA hydrogel. All CBMA hydrogels show a significant decrease (w80e90%) in cell adhesion relative to low fouling pHEMA, especially in supplemented growth medium. pHEMA itself exhibits only a small percent of the cell adhesion on TCPS [17]. At low crosslinker content, both CBMAX- and MBAA-crosslinked hydrogels have very little fouling, about 10e15% that of pHEMA

hydrogels. They are the same within error. At higher crosslinker content, however, there is a noticeable divergence: CBMAX-crosslinked hydrogels remain very nonfouling (about 10% that of pHEMA), while the fouling of MBAA-crosslinked hydrogels has worsened (to about 20% that of pHEMA). Furthermore, even 100% CBMAX hydrogels demonstrate the same low fouling levels of the lightly crosslinked CBMA hydrogels (still around 10% that of pHEMA hydrogels). Thus, MBAA incorporation corrupts the nonfouling properties of CBMA hydrogels, possibly by disrupting the ordered water around the zwitterionic polymer chains, whereas the CBMA crosslinker provides continuity of chemistry and hydration of the

Relative Cell Adhesion

structured water around zwitterionic materials than around neutral and hydrophilic materials.

1.2 1

CBMAX MBAA

0.8 0.6 0.4 0.2 0

4%

17% 100% Crosslinker Content

HEMA

Fig. 2. Relative cell adhesion to nonfouling CBMAX- and MBAA-crosslinked CBMA hydrogels. Cell adhesion of CBMA hydrogels normalized to the cell adhesion of pHEMA hydrogels.

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

molecular level, and thereby preserves the nonfouling properties within CBMA hydrogels. These results were supported by measuring the non-specific protein adsorption via ELISA. 3.3. RGD-functionalization of CBMA hydrogels Similar to the nonfouling study just described, CBMA hydrogels made 17% CBMAX and 17% MBAA hydrogels were chosen to demonstrate their ability for functionalization. The hydrogels were post-functionalized with cRGD, a cell-binding motif found on all integrins [38e40], using traditional EDC/sulfoNHS chemistry. Identical chemistry without EDC was performed on control hydrogels. COS-7 cells were then seeded on the EDC/sulfoNHScRGD-treated hydrogels, and on the sulfoNHS-cRGD-treated control hydrogels. The cell cultures were allowed to grow for 3 days, after which time cell proliferation was quantified with light microscopy. Fifteen pictures were taken of each hydrogel formulation, and the absolute numbers of cells adhered to the gels were counted. Comparative images of cells on functionalized and nonfunctionalized hydrogel surfaces are provided in the Supporting Information. As seen in Table 1, both functionalized CBMAX- and MBAAcrosslinked hydrogels exhibit an increase in cell adhesion. Nonspecifically adhered proteins from the supplemented medium will foul a surface and facilitate non-specific cell adhesion. As shown by ELISA, CBMAX-crosslinked hydrogels resist non-specific protein adsorption much more effectively than do MBAA-crosslinked hydrogels, and this is reflected in the lower overall number of cells adhered to the CBMAX-crosslinked hydrogels. The higher absolute numbers of adhered cells on 17% MBAA-crosslinked hydrogels, however, reflects the higher non-specific cell adhesion, or fouling, that results from the presence of the MBAA. Both hydrogels are functionalized with cRGD to induce specific cell adhesion and demonstrate controlled biocompatibility, but CBMAX-crosslinked hydrogels display lower levels of background non-specific cell adhesion. 3.4. Mechanical properties of CBMA hydrogels Mechanical properties are a major challenge for highly hydrated hydrogels. In general, the more water in the hydrogel, the weaker the structure. In order to function effectively as a mimic for biological tissue and to provide a conducive environment for cell growth, hydrogels should be “soft”. Practically, however, mechanical strength is required for handling. Furthermore, substrate stiffness plays an important role in cell fate and stem cell differentiation [41e43]. It is important, therefore, to study the mechanical properties of hydrogels to understand their potential interaction with cells. Hydrogel disks were compressed to failure, and the extracted compressive Young’s Moduli are shown in Fig. 3 as a function of crosslinker composition (CBMAX versus MBAA, represented by closed diamonds and open squares, respectively) and content (by %

Table 1 Cell adhesion on 17% CBMAX- and MBAA-crosslinked hydrogels functionalized with cRGD. Non-specific protein adsorption from ELISA reflects non-specific cell adhesion on pre-functionalized hydrogels. Protein Adsorption

Cell Adhesion

Crosslinker (17%)

Non-specific Protein Adsorption (ELISA, absorbance)

Cells Adhered Before Functionalization (abs. no./Frame)

Cells Adhered After Functionalization (abs. no./Frame)

CBMAX MBAA

0.07  0.03 0.79  0.02

21 10  2

16  3 25  3

Compressive modulus (MPa)

966

10 9 8 7 6 5 4 3 2 1 0

y = 0.0797x + 0.6789 R² = 0.9616

CBMAX MBAA

0

y = 0.0342x + 0.1189 R² = 0.8562 20 40 60 80 Crosslinker Content (%)

100

Fig. 3. Mechanical properties of CBMAX- and MBAA-crosslinked CBMA hydrogels. Hydrogels were compressed to failure, and the modulus reported was taken from the first 10% strain. Closed diamonds represent CBMAX-crosslinked CBMA hydrogels, while open squares represent MBAA-crosslinked CBMA hydrogels.

crosslinker). These values are also tabulated in Table 2. At the crosslinker concentrations for which we can directly compare CBMAX- and MBAA-crosslinked hydrogels (4%, 9%, and 17%), the CBMAX hydrogels demonstrated improved mechanical properties. The compressive modulus of CBMAX-crosslinked hydrogels is higher than that of MBAA-crosslinked hydrogels at all concentrations accessible to both crosslinkers, and the break stress is higher at all concentrations as well. Higher crosslinker content is accessible with CBMAX. At 100% CBMAX, the compressive modulus reaches 8 MPa, an impressive value for a hydrogel with relatively high water content (60%), and one that places this material well within the range of articular cartilage [44]. This also compares favorably to pHEMA hydrogels (compressive modulus of 0.6 MPa [45]) and even to PEG hydrogels (compressive modulus up to w3 MPa with 50% crosslinker [9]). Because we found no difference in the equilibrium water content of CBMA hydrogels with CBMAX versus CBMA hydrogels with MBAA crosslinker, we cannot attribute these improved mechanical properties to decreased hydration, but rather to a difference in the quality of the water structure. In studying the water structure in PC-hydrogels, Goda et al. [34] also found identical equilibrium water content along with improved mechanical properties, relative to PC-based hydrogels made with MBAA crosslinker. They attributed the improved mechanical properties of their zwitterionic crosslinked hydrogels to the presence of more ordered water. Also striking from the compression data is the linear nature of the relationship between the amount of crosslinker and the compressive modulus of the CBMA-CBMAX hydrogels. When compressive modulus is plotted as a function of crosslinker content

Table 2 Mechanical properties of CBMAX- and MBAA-crosslinked CBMA hydrogels. Modulus, break strain, and break stress, were extracted from stress-strain curves of CBMA hydrogels under compression. Crosslinker content (%)

4 9 17 33 50 75 100

Modulus (MPa)

w Break strain (%)

w Break stress (MPa)

MBAA

CBMAX

MBAA

CBMAX

MBAA

CBMAX

0.20  0.03 0.57  0.03 0.65  0.06 e e e e

0.6  0.1 1.11  0.06 1.79  0.09 4.1  0.6 5.5  0.3 6.44  0.02 8.2  0.5

40 45 30 e e e e

30 25 30 25 20 15 45

0.06 0.3 0.3 e e e e

0.1 0.4 0.6 0.6 1 1 2.5

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

from 4% crosslinker to 100% CBMAX (Fig. 3), the relationship is linear, and a regression of 0.9616. This underlines the compatibility of the gel system: the crosslinker is uniformly incorporated into the growing linear polymer chains at low crosslinker compositions, and, at the other end of the scale, the monomer is uniformly incorporated into the highly crosslinked network at high crosslink compositions. Thus, the hydrogel structure is more homogeneous. The MBAA crosslinker, on the other hand, showed poorer integration into the growing polymer chains, as evidenced by the less linear (R2 ¼ 0.8562) relationship between crosslinker composition and compressive modulus. Increasing the crosslinker content from 9% to 17% only marginally improved the mechanical properties of the gel, due to limited solubility even at this low concentration, and to poor incorporation of the dissolved crosslinker into the growing linear CBMA polymers. The microscopic structural defects that appear in inhomogeneous hydrogel networks compromise the mechanical integrity of the macroscopic hydrogel. 3.5. Physical properties of CBMA hydrogels Using the hydration data and stress-stain curves collected above, equations (3) and (4) allow us to calculate some physical properties of CBMA hydrogels when crosslinked with MBAA or CBMAX. Specifically, the stress-strain curves are manipulated to stress vs. (a-a2) curves, where a is the ratio of the deformed length to the original length. At low strains, this relationship is linear, and the crosslinking density can be extracted from the slope of this line (see equation (3)). Then, converting the number of crosslinks per volume to the volume per single crosslinker yields the distance between two crosslinks. These calculated values of crosslinker density and mesh size, as a function of crosslinking composition and content, are shown in Fig. 4 a and b, respectively. It can be seen

Crosslinker density (µmol/mm3)

a

Mesh size (nm)

b

1 0.9 0.8 0.7 0.6 0.5 0.4 0.3 0.2 0.1 0

CBMAX

0

12

20

40 60 80 Crosslinker Content (%)

CBMAX

10

MBAA

100

MBAA

8 6 4 2 0

0

20

40 60 80 Crosslinker Content (%)

100

Fig. 4. Physical properties of CBMAX- and MBAA-crosslinked CBMA hydrogels. Crosslinking density (a) and Mesh size (b) were calculated from mechanical and swelling properties. Closed diamonds represent CBMAX-crosslinked CBMA hydrogels, while open squares represent MBAA-crosslinked CBMA hydrogels.

967

that at the same nominal crosslinker content, the crosslinking densities of MBAA-crosslinked hydrogels is lower than that of CBMAX-crosslinked hydrogels, reinforcing the hypothesis that MBAA is less compatible with the CBMA monomer and thus the acrylamides of the MBAA incorporate poorly into the growing methacrylate polymer chains. On the other hand, the higher crosslinker density of the CBMAX-crosslinked hydrogels suggests better compatibility and more uniform incorporation of the CBMAX methacrylates into the growing methacrylate polymer. The mesh sizes calculated from the crosslinker density values tell a similar story, and further support the idea that the difference in the monomer-crosslinker compatibility plays a major role in the difference in the mechanical properties of the CBMA-CBMAX and CBMA-MBAA systems. Additionally, due to its excellent solubility and copolymerization with the CBMA monomer, the CBMAX-crosslinked hydrogels have access to a wider range of formulations, which provides greater diversity of physical properties. The different crosslinking densities, and thus different pore sizes, that arise in a controlled manner with CBMAX-crosslinked hydrogels afford yet another means of biological manipulation. Hydrogels are prized for their biomimetic structure: high water content and pores that allow passage of biomolecules. Pores ranging in size from <2 nm for small molecules such as sugars, growth factors, etc., to pores >10 nm for biomacromolecules such as large proteins and antibodies [46] are all accessible to hydrogels. By controlling the crosslinker content, we can control the pore size of the hydrogels without compromising nonfouling properties. 4. Conclusions We have introduced a crosslinker that has excellent compatibility with zwitterionic hydrogel systems. Chemically, the CBMAX crosslinker is very similar to the CBMA monomer. This has two decisive effects. Firstly, the zwitterionic nature of the crosslinker means that, in addition to the improved solubility, the crosslinker does not interrupt the restructuring of water that occurs around the zwitterionic monomer sidechains. Compared to hydrogels crosslinked with MBAA, CBMAX-crosslinked hydrogels exhibited better nonfouling at all compositions accessible to MBAA-crosslinked hydrogels, and ultimately demonstrated only about 10% of the nonspecific protein adsorption of pHEMA hydrogels. The second key effect of the chemical similarity between the CBMA monomer and CBMAX crosslinker is high polymerization compatibility. CBMAXcrosslinked hydrogels showed apparent stoichiometric incorporation of monomers and crosslinker. On the other hand, as MBAAcrosslinked hydrogels reached the solubility limit of MBAA, the relationship between crosslinker content and compressive modulus began to fail. Thus, CBMAX-crosslinked hydrogels exhibited greatly improved mechanical properties relative to MBAA-crosslinked hydrogels at the same nominal crosslinker content. The water solubility of the CBMAX crosslinker, additionally, allows access to a much wider range of formulations than were previously possible. Hydrogels made by polymerizing only the crosslinker (in the absence of monomer) were found to have high water content (60%), excellent nonfouling properties (about 90% lower non-specific cell adhesion than the nonfouling pHEMA control hydrogel), and high mechanical strength (compressive modulus of w8 MPa). Furthermore, these hydrogels composed of the CBMAX crosslinker, like those made with the CBMA monomer, are functionalizable by means of simple EDC/sulfoNHS chemistry. Acknowledgement This work was supported by the Office of Naval Research (N000140711036).

968

L.R. Carr et al. / Biomaterials 32 (2011) 961e968

Appendix. Supplementary material The supplementary data associated with this article can be found in the on-line version at doi:10.1016/j.biomaterials.2010. 09.067 References [1] Wichterle O, Lim D. Hydrophilic gels for biological use. Nature 1960;165:117e8. [2] Mathur AD, Collinsworth AM, Reichert WM, Kraus WE, Truskey GA. Endothelial, cardiac muscle and skeletal muscle exhibit different viscous and elastic properties as determined by atomic force microscopy. J Biomech 2001;34:1545e53. [3] Ratner BD, Miller IF. Transport through crosslinked poly(2-hydroxyethyl methacrylate) hydrogel membranes. J Biomed Mater Res 1973;7:353e67. [4] Hoffman AS. Hydrogels for biomedical applications. Adv Drug Deliv Rev 2002;43:3e12. [5] Peppas NA, Hilt JZ, Khamitzhanova G, Langer R. Hydrogels in biology and medicine: from molecular principles to bionanotechnology. Adv Mater 2006;18:1345e60. [6] Slaughter BV, Khurshid SS, Fisher OZ, Khademhosseini A, Peppas NA. Hydrogels in regenerative medicine. Adv Mater 2009;21:1e23. [7] Castillo EJ, Koening JL, Anderson JM, Lo J. Protein adsorption on hydrogels 2. Reversible and irreversible interactions between lysozyme and soft contactlens surfaces. Biomaterials 1985;6:338e45. [8] Refojo MF, Yasuda H. Hydrogels from 2-hydroxyethyl methacrylate and propylene glycol monoacrylate. J Appl Polym Sci Symp 1965;9:2425e35. [9] Rakovsky A, Marbach D, Lotan N, Lanir Y. Poly(etheylene glycol)-based hydrogels as cartilage substitutes: synthesis and mechanical characteristics. J Appl Polym Sci Symp 2009;112:390e401. [10] Ostuni E, Chapman RG, Holmlin RE, Takayama S, Whitesides GM. A survey of structure-property relationships of surfaces that resist the adsorption of protein. Langmuir 2001;17:5605e20. [11] Herold DA, Keil K, Bruns DE. Oxidation of polyethylene glycols by alcohol dehydrogenase. Biochem Pharmacol 1989;38:73e6. [12] Harder P, Grunze M, Dahint R, Whitesides GM, Laibinis PE. Molecular conformation in oligo(ethylene glycol)-terminated self-assembled monolayers on gold and silver surfaces determines their ability to resist protein adsorption. J Phys Chem B 1998;102:426e36. [13] Shen MC, Martinson L, Wagner MS, Castner DG, Ratner BD, Horbett TA. PEO-like plasma polymerized tetraglyme surface interactions with leukocytes and proteins: in vitro and in vivo studies. J Biomater Sci Polym Ed 2002;13:367e90. [14] Li LY, Chen SF, Jiang SY. Protein interactions with oligo(ethylene glycol) selfassembled monolayers: OEG stability, surface packing density, and protein adsorption. J Biomater Sci Polym Ed 2007;18:1415e27. [15] Gaberc-Porekar V, Zore I, Podobnik B, Menart V. Obstacles and pitfalls in the PEGylation of therapeutic proteins. Curr Opin Drug Discov Devel 2008;11:242e50. [16] Zhang Z, Chao T, Jiang S. Physical, chemical, and chemical-physical double network of zwitterionic hydrogels. J Phys Chem 2008;112:5327e32. [17] Chen S, Jiang S. An avenue to new nonfouling materials. Adv Mater 2007;30:335e8. [18] Chang Y, Chen SF, Yu QM, Zhang Z, Bernards M, Jiang SY. Development of biocompatible interpenetrating polymer networks containing a sulfobetainebased polymer and a segmented polyurethane for protein resistance. Biomacromolecules 2007;8(1):122e7. [19] Chang Y, Chen SF, Zhang Z, Jiang S. Highly protein-resistant coatings from well-defined diblock copolymers containing sulfobetaines. Langmuir 2006;22 (5):2222e6. [20] Cheng G, Chen S, Bryers J, Jiang S. Inhibition of bacterial adhesion and biofilm formation on zwitterionic surfaces. Biomaterials 2007;28:4192e9. [21] Zhang Z, Chao T, Chen S, Jiang S. Superlow fouling sulfobetaine and carboxybetaine polymers on glass slides. Langmuir 2006;22(24):10072e7.

[22] Zhang Z, Chen S, Chang Y, Jiang S. Surface grafed sulfobetaine polymers via atom transfer radical polymerization as superlow fouling coatings. J Phys Chem B 2006;110(22):10799e804. [23] Zhang Z, Chen S, Jiang S. Dual-functional biomimetic materials: nonfouling poly(carboxybetaine) with active functional groups for protein immobilization. Biomacromolecules 2006;7(12):3311e5. [24] Chen S, Yu FC, Yu QM, He Y, Jiang S. Strong resistance of a thin crystalline layer of balanced charged groups to protein adsorption. Langmuir 2006;22 (19):8186e91. [25] Ladd J, Chen S, Zhang Z, Hower J, Jiang S. Zwitterionic polymers exhibiting high resistance to non-specific protein adsorption from human serum and plasma. Biomacromolecules 2008;9(5):1357e61. [26] Tsai WB, Shi Q, Grunkemeier JM, McFarland C, Horbett TA. Platelet adhesion to radiofrequency glow-discharge-deposited fluorocarbon polymers preadsorbed with selectively depleted plasmas show the primary role of fibrinogen. J Biomater Sci Polym Ed 2004;15(7):817e40. [27] Yang W, Xue H, Li W, Zhang J, Jiang S. Pursuing “zero” protein adsorption of poly(carboxybetaine) from undiluted blood serum and plasma. Langmuir 2009;25(19):11911e6. [28] Yang W, Zhang L, Wang S, White AD, Jiang S. Functionalizable and ultra stable nanoparticles coated with zwitterionic poly(carboxybetaine) in undiluted blood serum. Biomaterials 2009;30:5617e21. [29] Cheng G, Xite H, Zhang Z, Chen S, Jiang S. Switchable biocompatible polymer surface with self-sterilizing and non-fouling/ biocompatible capabilities. Angew Chem Int Ed Engl 2008;47(46):8831e4. [30] Cheng G, Li G, Xue H, Chen S, Bryers JD, Jiang S. Zwitterionic carboxybetaine polymer surfaces and their resistance to long-term biofilm formation. Biomaterials 2009;30:5234e40. [31] He Y, Hower JC, Chen S, Bernards MT, Chang Y, Jiang S. Molecular simulation studies of protein interactions with zwitterionic phosphorylcholine selfassembled monolayers in the presence of water. Langmuir 2008;24:10358e64. [32] He Y, Chang Y, Hower JC, Chen S, Jiang S. Origin of repulsive force and structure/dynamics of interfacial water in OEG-protein interactions: a molecular simulation study. Phys Chem Chem Phys 2008;10:5539e44. [33] Huglin MB, Rego JM. Influence of temperature on swelling and mechanical properties of a sulfobetaine hydrogel. Polymer 1991;32:3354e8. [34] Goda T, Watanabe J, Takai M, Ishihara K. Water structure and improved mechanical properties of phospholipid polymer hydrogel with phosphorylcholine centered intermolecular cross-linker. Polymer 2006;47:1390e6. [35] Goda T, Matsuno R, Konno T, Takai M, Ishihara K. Protein adsorption resistance and oxygen permeability of chemically crosslinked phospholipid polymer hydrogel for ophthalmologic biomaterials. J Biomed Mater Res B Appl Biomater 2009;89:184e90. [36] Park K, Shalaby WSW, Park H. Biodegradable hydrogels for drug delivery. Lancaster, PA, USA: Technomic Publishing Company; 1993 ([Chapter 4]). [37] Wichterle O, Lim D.U.S. Pat. 2,976,576 (March 28, 1961). [38] Pierschbacher MD, Ruoslahti ER, Sundelin J, Lind P, Peterson PA. The cell attachment domain of fibronectin. J Biol Chem 1982;257:9593e7. [39] Pierschbacher MD, Ruoslahti ER. Cell attachment activity of fibronectin can be duplicated by small synthetic fragments of the molecule. Nature 1984;309:30e3. [40] Perlin L, MacNeil S, Rimmer S. Production and performance of biomaterials containing RGD peptides. Soft Matter 2008;4:2331e49. [41] Engler A, Sen S, Sweeney HL, Discher D. Substrate stiffness and cell fate: Matrix elasticity directs stem cell lineage specification. Cell 2006;126:677e89. [42] Even-Ram S, Artym V, Yamada KM. Matrix control of stem cell fate. Cell 2006;126:645e7. [43] Leipzig ND, Shoichet MS. The effect of substrate stiffness on adult neural stem cell behavior. Biomaterials 2009;30:6867e78. [44] Shepherd DET, Seedhom BB. The “instantaneous” compressive modulus of human articular cartilage in joints of the lower limb. Rheumatology 1999;38:124e32. [45] Han YA, Lee EM, Ji BC. Mechanical properties of semi-interpenetrating polymer network hydrogels based on poly(2-hydroxyethyl methacrylate) copolymer and chitosan. Fibers Polym 2008;9(4):393e9. [46] Bhatia SR, Khattak SF, Roberts SC. Polyelectrolytes for cell encapsulation. Curr Opin Colloid Interface Sci 2005;10:45e51.