Student Award Winner in the Undergraduate’s Degree Categoryfor the Society for Biomaterials 34th Annual Meeting,Seattle, Washington, April 21–24, 2010

Characterization of poly(ethylene glycol) gels with addedcollagen for neural tissue engineering

Rebecca Scott,* Laura Marquardt,* Rebecca Kuntz Willits

Department of Biomedical Engineering, Saint Louis University, Saint Louis, Missouri

Received 5 January 2010; accepted 8 January 2010

Published online 12 March 2010 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jbm.a.32775

Abstract: Over the past decade, it has been increasingly rec-

ognized that both chemical and mechanical properties of

scaffolds influence neural cell behavior, ranging from growth

to differentiation to migration. However, mechanical proper-

ties are difficult to control for in the design of scaffolds for

nerve regeneration, as properties change over time for most

biologically derived scaffolds. The focus of this project was

to examine how the mechanical properties of a nondegrad-

able scaffold, poly(ethylene glycol) (PEG) gels, influenced

nerve cell behavior. Low concentration PEG gels, of 3, 4, or

5% PEG, with added collagen to alter chemical properties

were examined for both their mechanical properties and their

ability to support nerve expression and extension. Stiffness

(G*) significantly increased with increased PEG concentra-

tion. The addition of chemically conjugated collagen signifi-

cantly decreased the stiffness compared to plain gels. This

phenomenon was confirmed to be an effect of the conjugate,

and not the protein itself, as G* of gels containing conjugate,

but no protein, was not significantly different than G* of gels

with conjugated protein. PC12 cell neurite expression

increased with decreasing PEG and increasing collagen con-

centration. At its best, the expression approached the value

on collagen-coated tissue culture plastic, which is a substan-

tial improvement over previous studies on PEG. Neurite

extension of dorsal root ganglia was also improved on these

same gels over gels with either higher PEG concentration or

lower collagen amount. Overall, these results suggest that

exploration of lower stiffness materials is necessary to

improve neurite growth and extension in three-dimensional

synthetic scaffolds. VC 2010 Wiley Periodicals, Inc. J Biomed Mater

Res Part A: 93A:817–823, 2010.

Key Words: poly(ethylene glycol), protein conjugation, colla-

gen, neurite growth, stiffness

INTRODUCTION

Complete severance of a nerve, caused by injury or disease,can result in impaired muscle function, sensation, or pain.1

Although the possibility of nerve regeneration exists, espe-cially in the peripheral nervous system, nerves rarely regen-erate correctly without the aid of surgical intervention. Cur-rent methods of neuronal regeneration include surgicalrepair or grafting and nerve conduits. Alternatively, func-tional regeneration may be possible through tissue engi-neering, where a three-dimensional scaffold would supportand guide neuronal growth through various mechanical andchemical cues.

Both natural and synthetic materials have been or arecurrently under investigation for use as tissue-engineeredbiomaterials for neuronal regeneration.2–5 Natural materials,

such as extracellular matrix (ECM) proteins, have beenfound to influence neurite growth in both two-dimensionaland three-dimensional studies.6 Specifically, collagen I, hashad a key role in nerve guides, as it supports nerve growthas a substrate,7 gel,8 and nerve graft.9 Collagen coatingshave been found effective for both neurite attachment anddifferentiation.10 In addition, this laboratory has previouslyexamined neurite growth within three-dimensional gels,demonstrating that neurite length was influenced by me-chanical properties11 and chemical properties12 within colla-gen gels. However, three-dimensional scaffolds composedprimarily of ECM materials can be degraded by the body,which would alter their ability to support nerve growthlong term. A scaffold with a synthetic polymer base is anexciting alternative to guide neuronal growth, as it can be

*These authors contributed equally to this work.

Correspondence to: R. K. Willits; e-mail: willitsr@slu.edu

Contract grant sponsor: NSF CBET Award; contract grant number: 0755389

Contract grant sponsor: NSF REU Site Award; contract grant number: 0849621

Contract grant sponsor: Sigma Xi GIAR

VC 2010 WILEY PERIODICALS, INC. 817

designed with a more controlled time frame for degradation.In addition, the properties of synthetic scaffolds can bereadily altered to produce materials with specific mechani-cal and chemical properties.

In several laboratories, poly(ethylene glycol) (PEG) gelsare being investigated as scaffolds for use in neural applica-tions due to the biocompatible and nonimmunogenic natureof PEG.13,14 PEG, currently FDA approved for several medicaldevices,14 has been investigated for use in drug delivery devi-ces15 and tissue-engineering scaffolds.16,17 PEG scaffolds canprovide mechanical support and a swollen, three-dimensionalenvironment for encapsulating cells.18 These gels can bemanipulated via the concentration of PEG,11 with previousstudies demonstrating that increased PEG concentrations ledto corresponding increases in the mechanical properties ofthe gel.19,20 In addition, PEG is a readily chemically modi-fied21–23 and, therefore, the chemical nature of the scaffoldcan be altered through the conjugation of ECM proteins21 orpeptide sequences20 into the gel to influence cell growth.24

The conjugation of proteins, like collagen I, directly tothe PEG chains within the gel allows for better control ofthe spatial properties of the gel. For example, by conjugatingthe collagen directly into the gel, the protein can be homo-geneously distributed within the gel. In addition, as collagenI directly interacts with cell surface receptors and influencescell attachment, proliferation, and expression,9 the incorpo-ration of this protein will provide chemical cues necessaryfor cell growth.20,21,23 Gunn et al. found that a neural cellmodel, PC12 cells, seeded on top of PEG gels with addedRGDS, an amino acid sequence that serves as a cell recogni-tion site for the binding of ECM, were able to extend neu-rites in 11.4% 6 1.2% of cells, with better extension ongels with lower mechanical properties.20 It is noteworthy,however, to mention that 11.4% neurite expression forPC12 cells is significantly decreased from the typical 80%expression seen on two-dimensional tissue culture plastic.25

To extend upon this idea, the focus of this article was toexamine low concentrations of PEG gels to determine howthe decreased mechanical stiffness would influence PC12and dorsal root ganglia (DRG) neurite extension. Further,collagen was added at 10, 100, and 500 lg/mL to examinehow the chemical properties of the PEG gels would furtherinfluence neurite extension. The results demonstrated thatdecreasing the mechanical stiffness increased PC12 neuriteexpression, with overall neurite expression approaching thevalue seen on collagen-coated tissue culture plastic. A simi-lar trend was also exhibited by partially dissociated DRGseeded on top of these gels, with increased neurite exten-sion on gels with decreased stiffness and increased collagen.

MATERIALS AND METHODS

MaterialsPEG-DA was synthesized as previously described26 andcharacterized using nuclear magnetic resonance (NMR). ThePEG-collagen conjugate was fabricated by covalently bindingN-hydroxysuccinimidyl group (NHS) on the acryl-PEG-NHSmolecule to an amine group on collagen.27 Calculatedamounts (4 NHS:1 collagen) of acryl-PEG-NHS (Nektar) and

collagen I, extracted from rat tails,28 were mixed overnightat RT in PBS. Excess salts and any unreacted acryl-PEG-NHSwere removed via Amicon Ultra-15 Centrifugal Filter (Milli-pore) and the conjugate was stored at �20�C. Irgacure2959 was obtained from CIBA; F12K, serum, Hoescht33258, and formalin were obtained from Sigma, nervegrowth factor (NGF, 2.5s) was obtained from Millipore.

Mechanical testing of PEG-collagen conjugate gelsPEG gels were made with (3, 4, 5% w/w) PEG-DA, Irgacure2959, and DI H2O. PEG-conjugate gels were made with vary-ing concentrations of PEG-DA (3, 4, 5% w/w), PEG-proteinconjugate (0.1, 1, 10, or 100 lg/mL), Irgacure, and DI H2O.Solutions were crosslinked between two glass sheets via UVlight (3.2 mW/cm2, 365 nm), stored in DI H2O overnight,and cut into 25 mm rounds. The mechanical stiffness of eachgel (G*) was measured via oscillatory shear rheometry, usinga 25 mm parallel plate configuration, a frequency sweep of1–100 rad/s, and constant strain of 5%; all tests were donewithin the linear viscoelastic regime for these materials.ANOVA was utilized (p < 0.05) for statistical significance.

Mechanical testing of PEG-NHS conjugate gelsThe addition of protein to the PEG gel could cause furtherchanges in the mechanical nature of the gel as proteins likecollagen can exhibit their own mechanical properties. Toverify whether the protein contributed to the mechanicalstiffness of the PEG-protein conjugate gels, PEG-NHS conju-gate gels were mechanically tested. Acryl-PEG-NHS wasadded to the PEG gels and mechanically tested in the samemanner as the acryl-PEG-collagen conjugate. The amount ofacryl-PEG-NHS added to the gel was based on the molarequivalent to what was conjugated to each equivalentamount of protein assuming complete reaction. For example,if 1 lg of collagen conjugate (�1.24 � 10�12 moles) wasadded per mL of gel, then 1.24 � 10�12 moles of Acryl-PEG-NHS was added for this testing.

Swelling of PEG-collagen conjugate gelsPEG-collagen conjugate solutions were prepared asdescribed previously and hydrated to equilibrium for 24 h.The gels were weighed (Wwet), lyophilized to dryness, andthen weighed again (Wdry). Swelling capacity of the gels wascalculated using:

%C ¼ Wwet �Wdry

Wwet� 100 (1)

Spatial distribution of protein conjugateTo label the protein, calculated amounts of PEG-NHS andcollagen I were mixed with fluorescent dye. The AlexaFluor546 Protein Labeling Kit (Invitrogen) was used according tothe instructions provided by the manufacturer. AlexaFluor546 has its excitation maximum at 558 nm and its emissionmaximum at 573 nm. PEG-protein conjugate solutions wereprepared using PEG-DA (5% w/w), fluorescently labeledprotein conjugate (10, 100 lg/mL), Irgacure, and F12K.

818 SCOTT, MARQUARDT, AND WILLITS PEG GELS FOR NEURAL TISSUE ENGINEERING

Conjugate gels were crosslinked in glass bottomed 6-wellplates under UV light, and stored in F12K. Gels were visual-ized using a Zeiss LSM 5 Pascal confocal microscope with aPlan-Neofluor 40� objective. Scans were completed with axy area of 852.6 mm2 and three stacks, each 10 mm (0.25mm per step) in the z direction, were imaged at top, middle,and bottom at three separate locations of each gel. Eachimage was taken at the same exposure setting to ensuresimilar darkness values; plain 5% PEG gels were used ascontrols to set the exposure level. Matlab was used tonumerically determine the average amount of fluorescenceexhibited per volume in each gel. ANOVA was utilized (p <

0.05) for statistical significance.

PC12 neurite expression on PEG collagen orPEG-collagen conjugate gelsPEG gels were made with PEG-DA (3, 4, 5%), Irgacure, andF12K. Collagen or PEG-collagen conjugate was added intogels at concentrations of 0, 10, 100, or 500 lg/mL. All solu-tions, except 500 lg/mL collagen conjugate, were cross-linked for 360 s under UV light. Addition of the 500 lg/mLconjugate did not permit gelation, likely due to the highnumber of acryl-PEG-collagen molecules inhibiting the chainlengthening process. PC12 cells were obtained from ATCCand cultured in T75 flasks with F12K media supplementedwith 15% horse serum and 2.5% fetal bovine serum. Mediawas changed every 2–3 days and kept within 2–5 passages.PC12 cells were seeded onto the gels at 5 � 103 cells/cm2

and cultured in supplemented medium with 50 ng/mL NGFand 10 lg/mL Hoescht 33342 for 48 h. An additional 25 ngof NGF was added to the PC12 cells after 24 h. Followingculture, cells were imaged live on a Zeiss inverted micro-scope. The total number of cell groupings, clusters contain-ing three or more cells in contact, was counted in eachimage.25 The number of those cell groupings that expressedneurites of at least one cell diameter long were counted andcompared with the total number of cell groups, yielding apercent expression.25 Percent expression of PC12 cells onPEG gels containing no collagen was used as a control, towhich sample results were compared. ANOVA was utilized(p < 0.05) for statistical significance.

Neurite extension of partially dissociated DRG onPEG collagen or PEG-collagen conjugate gelsPEG gels were made as described earlier for PC12 cells.DRG were extracted from embryonic day 9 chicks and par-tially dissociated by addition of 1� trypsin for 10 min.Approximately 3 partially dissociated DRG were seeded pergel and cultured in F12K with 20% FBS and 50 ng/mL NGFin the system for 48 h. An additional 25 ng of NGF was sup-plied at 24 h. DRG were imaged on a Zeiss inverted micro-scope and radial extension was measured using Axiovisionsoftware. The percent of DRG extending compared to totalnumber of DRG examined was also examined. Statistical sig-nificance was determined using ANOVA (p < 0.05) and neu-rite extension was compared to gels of no added collagen.

RESULTS

Mechanical testing of PEG gelsRheological testing was utilized to determine G*, a measureof mechanical stiffness. As expected for these materials, thestorage modulus (G0) was greater than the loss modulus(G00), so all materials were considered a gel. Examination ofG*, at 10 rad/s, demonstrated significant differencesbetween 3, 4, and 5% PEG gels, with respective averagestiffness of 69.23 6 4.27 Pa, 479.49 6 9.79 Pa, and 992.246 36.07 Pa. Testing of the PEG-collagen conjugate gelsresulted in decreased G* values as compared to plain PEGgels, at 10 rad/s (Fig. 1). A general trend was seen in allPEG-collagen conjugate gels, whereas the concentration ofthe conjugate increased, the mechanical stiffness of the geldecreased. PEG gels with 100 lg/mL PEG-collagen conjugatewere found to exhibit a significantly decreased average stiff-ness versus plain PEG gels. In addition, significant decreaseswere also seen for 4% PEG with 1 or 10 lg/mL PEG-colla-gen conjugate and 5% PEG with 10 lg/mL PEG collagenconjugate, as compared to plain PEG gels. PEG-NHS gels fol-lowed the same general trends as PEG-collagen conjugategels, where the mechanical stiffness of the gel decreasedwith decreasing concentrations of PEG-DA and increasingamounts of PEG-NHS (Table I). PEG gels with 10 or 100 lg/mLPEG-NHS exhibited a significantly decreased average stiff-ness when compared to plain PEG gels. PEG-NHS gels didnot exhibit a significant difference in mechanical stiffnesswhen compared to the PEG-collagen conjugate gels.

Swelling of PEG gelsSwelling capacity tests found the PEG gels exhibited anincreased swelling capacity with decreased PEG concentra-tion, where 3% gels swelled 97.65% 6 0.12%, 4% gelsswelled 96.99% 6 0.07%, and 5% gels swelled 94.72% 60.52% (Fig. 2). Analysis of the physical properties of theplain PEG gels found that a significant difference in swelling

FIGURE 1. Stiffness (as represented by G*) of the PEG-collagen conju-

gate gels. The G* was determined for each gel at 10 rad/s and aver-

aged. Error bars represent standard error and * represents statistical

difference from plain gels. N � 4.

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | 1 JUN 2010 VOL 93A, ISSUE 3 819

capacity was seen between 3, 4, and 5% PEG gels. As theconcentration of PEG-collagen conjugate increased, a signifi-cant difference among 3, 4, and 5% gels was found, withthe exception of 4 and 5% PEG gels with 0.1 or 1 lg/mLPEG-collagen conjugate, which were not found to be signifi-cantly different from each other (Fig. 2). No significant dif-ference in swelling capacity was seen in 5% gels with theaddition of any concentration of the PEG-collagen conjugate,when compared to the plain PEG gels.

Spatial distribution of protein conjugateExamination of the images obtained from fluorescently labeledPEG-collagen conjugate gels showed stained collagen moleculespositioned homogeneously throughout the gel. As expected,the average fluorescence level increased as the concentrationof PEG-collagen conjugate increased, where gels containing 10and 100 lg/mL PEG-collagen conjugate exhibited an averagefluorescence level of 0.019 6 0.002 and 0.101 6 0.008 percm3, respectively. Furthermore, the average fluorescence levelfound at the top, middle, and bottom of the gels was not foundto be significantly different (Table II).

PC12 neurite expressionPercent neurite expression of PC12 cell was analyzed onPEG, PEG gels with collagen mixed in, and PEG-collagen con-jugate gels. Solutions composed of 5% PEG and 500 lg/mLof collagen conjugate did not crosslink, therefore, no expres-sion data was determined. Percent neurite expression

increased as the concentration of PEG-DA decreased and theconcentration of PEG-collagen conjugate increased (Fig. 3).At 100 lg/mL PEG-collagen conjugate gels, average neuriteexpressions of 81.4% 6 4.6%, 26.76% 6 5.25%, and14.94% 6 4.07% for 3, 4, and 5% PEG gels were exhibited,indicating increased neurite expression with decreased PEGconcentration. Increased collagen concentration also affectedneurite expression, seen in 3% PEG gels where 0, 10, 100,and 500 lg/mL collagen mixed gels expressed 2.71% 61.34%, 48.67% 6 6.96%, 56.25% 6 8.41%, and 80.33% 65.39%, respectively. The greatest percent expression wasseen on 3% PEG gels with 100 lg/mL PEG-collagen conju-gate, where 81.4% 6 4.6% of cells expressed neurites. Inaddition, percent neurite expression was significantlyincreased on 3% PEG gels with 100 lg/mL PEG-collagenconjugate, as compared to expression seen on 3% PEG gelswith 100 lg/mL collagen mixed into the gel.

Partially dissociated DRG extensionRadial neurite extension was examined on plain PEG, PEGgels with collagen mixed in, and PEG-collagen conjugategels. Plain 3% gels showed significantly increased extension,605.76 6 40.43 lm, over plain 4 and 5% PEG gels, 269.876 19.21 lm and 294.61 6 20.02 lm, respectively (Fig. 4).Gels consisting of 3% PEG with 10 lg/mL collagen, 100 lg/mL collagen, or 100 lg/mL PEG-collagen conjugate [Fig.4(A)] were found to have significantly increased neuriteextension versus plain PEG gels. For 4% PEG gels [Fig.4(B)], gels with 10 lg/mL PEG-collagen conjugate or 500lg/mL collagen had significantly increased neurite exten-sion as compared to plain gels. The only difference between5% plain gels and gels with added collagen was with 500lg/mL of collagen [Fig. 4(C)], where the extension was sig-nificantly decreased. In addition, the number of partially dis-sociated DRG exhibiting neurite extension compared to thetotal number of DRG examined was analyzed. In general,

TABLE I. Mechanical Stiffness (G*) of PEG-Collagen Conjugate and PEG-NHS Conjugate Gels

Conjugate(lg/mL)

G* (Pa)

5% PEG-DA 4% PEG-DA 3% PEG-DA

PEG-NHS PEG-Collagen PEG-NHS PEG-Collagen PEG-NHS PEG-Collagen

0.1 891.36 6 66.70 907.62 6 24.03 400.98 6 57.33 431.32 6 23.09 74.89 6 3.89 72.75 6 9.511 845.21 6 93.61 898.38 6 25.79 358.46 6 44.91 353.17 6 44.23 56.83 6 8.86 59.96 6 3.6210 835.86 6 54.85 839.77 6 54.91 293.51 6 26.89 292.52 6 9.41 51.83 6 3.38 52.41 6 8.24100 794.00 6 55.38 826.29 6 13.44 223.30 6 26.57 232.11 6 22.72 42.13 6 2.49 45.71 6 2.76

FIGURE 2. Swelling capacity of PEG-collagen conjugate gels. Error

bars represent standard error and * represents statistical difference

from plain gels. N ¼ 5.

TABLE II. Average Fluorescence Levels of PEG-Collagen

Conjugate Gels

Locationin Gel

5% PEG-DA

100 lg/mLPEG-Collagen

10 lg/mLPEG-Collagen

0 lg/mLPEG-Collagen

Top 0.129 6 0.013 0.023 6 0.004 0.002 6 0.002Middle 0.112 6 0.016 0.020 6 0.003 0.004 6 0.002Bottom 0.087 6 0.014 0.020 6 0.004 0.002 6 0.002

820 SCOTT, MARQUARDT, AND WILLITS PEG GELS FOR NEURAL TISSUE ENGINEERING

more DRG exhibited neurite extension on 3% gels over 4and 5% gel gels (Table III).

DISCUSSION

In the development of tissue-engineered scaffolds, the useof a synthetic scaffold allows for better control of the me-chanical and chemical properties over gels of extracellularmatrix. In studying PEG scaffolds for nerve regeneration,previous studies have determined the effects of higher con-centration PEG gels, where the gels support limited neuriteextension.20 Therefore, this study investigated the effect oflower concentration PEG scaffolds on neurite expressionand extension, varying the chemical properties through theaddition of collagen, and the mechanical properties by vary-ing the PEG-DA concentration.

Mechanical evaluation of the PEG gels demonstrated thatas the concentration of PEG-DA increased, gel stiffnessincreased. This result is consistent with previous tests onPEG gels in tensile testing, where the increase of PEG-DAconcentration led to increases in elastic modulus.14,16,29

Although 5% PEG gels typically exhibit a mechanical stiff-ness an order of magnitude greater than that of collagengels, the mechanical stiffness exhibited by 3% PEG gels isapproximately the same order of magnitude as the mechani-cal stiffness of 2 mg/mL collagen gels.11 Using these lowerconcentration PEG gels, therefore, allowed for investigationof stiffness over approximately an order of magnitude.

The addition of protein conjugate to the PEG gels furtherreduced the stiffness of the gel, even at the lowest concen-

tration of added protein (0.1 lg/mL). A decrease is notunexpected, as the protein conjugate terminates the growingPEG chain, which would, in turn, lower the stiffness. Toexamine how the PEG gel changed, the storage modulus wascompared to the average molecular weight between entan-glements (Me) using the equation

G0 ¼ pRT

Me(2)

where p is the polymer density, T is the temperature, and Ris the gas constant.30 As predicted by the model, the aver-age molecular weight between entanglements increased asG0 decreased (Table IV). An increased average molecularweight between chains predicts fewer entanglementsbetween chains and a smaller G0. The decrease in G0 can beaccounted for in two ways. First, the addition of the conju-gate effectively caps PEG chains, which in turn decreasesthe mechanical stiffness. Second, decreased concentrationsof PEG-DA provided fewer available chains for elongationduring crosslinking, thereby decreasing the mechanical stiff-ness of the gel. Further analysis of Me can be performed tocompare mesh size (n) of the gel using the equation

n ¼ lC0:5n n0:5V0:33

2;S (3)

where Cn is the characteristic ratio of the polymer, l is thebond length within the repeat unit, n is the number ofbonds between crosslinks, and V2,S is the volume fraction ofpolymer in the swollen gel,31 calculated from swelling

FIGURE 3. Percent neurite expression of PC12 cells in (A) 3%, (B) 4%, and (C) 5% PEG gels with added collagen or PEG-collagen conjugate. No

expression was found on any 5% plain gels. Error bars represent standard error and * represents statistical difference from plain gels. Neurite

expression was examined in N � 30 images.

FIGURE 4. Radial neurite extension of partially dissociated DRG (A) 3%, (B) 4%, and (C) 5% PEG gels with added collagen or PEG-collagen conju-

gate. Error bars represent standard error and * represents statistical difference from plain gels. Neurite extension was examined in N � 40 DRG.

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | 1 JUN 2010 VOL 93A, ISSUE 3 821

capacity. Watkins and Anseth previously reported that asthe concentration of PEG within a gel increased, the meshsize of the gel decreased,32 agreeing with the trend seen inthis study, where the mesh size of the gels increased withdecreased concentrations of PEG (Table IV).

To investigate the interplay between PEG gel propertiesand cell behavior, two-dimensional cell studies examiningpercent neurite expression of PC12 cells and radial neuriteextension of partially dissociated DRG were performed.Increased PC12 neurite expression was evident for cellsseeded on 3, 4, and 5% PEG gels with collagen and PEG-col-lagen conjugate, with percent expression approaching tissueculture plastic with decreasing concentration of PEG andincreasing concentration of collagen (% expression on TCplastic from previous research25). Both the collagen concen-tration as well as the PEG concentration influenced theoverall neurite expression. At 100 lg/mL, a significantincrease in PC12 neurite expression was determined for 3%PEG gels with conjugated collagen compared to gels mixedwith collagen, indicating a possible difference in how thecollagen is presented to the cells. Other research hasreported findings where the presentation of the protein orpeptide had significant effects on cell behavior. Heilshornet al. reported that altering the location of lysine residues,incorporated at specific crosslinking sites, on CS5 peptidesequence resulted in varied amounts of endothelial attach-ment.33 In another study, Murphy et al. immobilized a spe-cific protein from the tenth domain of fibronectin, by cova-lently fusing the protein to a phosphonate ligand ontri(ethylene glycol)-terminated self-assembled monolayers, amethod which allowed control of density, patterning, andorientation of the proteins at presented to cells.34 Althoughincreased expression was noted at 100 lg/mL in 3% PEG, itwas not seen at other concentrations, indicating it is not theprimary mechanism of improving the neurite expression.However, presentation of peptides and proteins within mat-rices is important to fully understand protein–cell interac-tions and allows for versatility in tissue engineering design.

In addition to protein interactions, cell behavior is alsoaffected by the mechanical properties of the environment inwhich they are growing. In this study, neurite expressionwas increased for 3% PEG gels at any concentration of con-jugate over both 4 and 5% PEG gels, further indicating thatlower concentration gels better promote neurite expression.The amount of neurite expression exhibited by the PEG gelsin this study was enhanced over previous studies using

15% PEG gels with PEG-RGDS conjugate, where only 11.4%neurite expression was noted.20 In addition, this lab haspreviously investigated PC12 neurite extension on 10% PEGgels with 500 lg/mL added collagen, with 26.94% 6 3.61%expression (unpublished data). Therefore, the results of thiswork suggest that gels with reduced stiffness and increasedcollagen concentrations perform better as substrates forPC12 cells.

The influence of stiffness was further characterized viaDRG extension, where 3% gels supported increased neuriteextension of DRG (105.6%–124.5% increase) over both 4and 5% gels. Although extension of DRG is likely affected byboth PEG and collagen concentrations, collagen addition hadmore impact on extension at low stiffness than at high stiff-ness and the percent of DRG extending was only slightlyincreased on collagen-containing gels over plain gels withina PEG concentration. These findings are similar to theresults seen in previous two-dimensional studies, whereDRG extensions on PEG gels were similar after 2 days whenaccounting for fibrinogen concentration and stiffness.35 Inaddition, Sarig-Nadir and Seliktar determined that DRGseeded on PEG:fibrinogen gels had an average outgrowth of�400 lm at 2 days culture, similar to the range of growthseen in this study at higher stiffness; all of the gels theytested were >10% PEG.35 These trends need to be furtherconfirmed in three-dimensional studies.

The use of low concentration PEG gels is promising forneural engineering applications as the PC12 neurite expres-sion on these gels was significantly increased over higherconcentration PEG gels and approached that of tissue cul-ture plastic. Further, significant increases in extension fromDRG were found on the gels with reduced stiffness. Thisresult encourages further exploration of these low concen-tration gels, either with or without added ECM components.In hopes of designing a three-dimensional scaffold for usein neuronal regeneration, the balance of mechanical andchemical properties need to be further related to neurite

TABLE III. Percent (%) of Partially Dissociated DRG Exhibiting Neurite Extension Gels Containing PEG-Collagen

Conjugate and Collagen Mixed In

[PEG-DA]

Collagen Concentration (lg/mL)

0

10 100 500

Mixed Conjugated Mixed Conjugated Mixed

3% 97.6 100.0 98.0 96.3 80.0 88.54% 60.6 88.7 60.0 48.2 76.5 83.95% 64.5 77.6 52.6 90.2 50.6 79.7

TABLE IV. Storage Modulus, Molecular Weight Between

Entanglements, and Mesh Size of PEG Gels

[PEG-DA] G0 (Pa) Me � 106 (g/mol) n (lm)

5% 992.24 3.02 0.194% 479.97 6.25 0.303% 63.96 46.90 0.92

822 SCOTT, MARQUARDT, AND WILLITS PEG GELS FOR NEURAL TISSUE ENGINEERING

extension, specifically using a primary nerve cell in vitroand a growth model in vivo.

REFERENCES1. Dodla MC, Bellamkonda RV. Differences between the effect of ani-

sotropic and isotropic laminin and nerve growth factor presenting

scaffolds on nerve regeneration across long peripheral nerve

gaps. Biomaterials 2008;29:33–46.

2. Archibald S, Krarup C, Shefner J, Li S, Madison R. A collagen-

based nerve guide conduit for peripheral nerve repair: An electro-

physiological study of nerve regeneration in rodents and nonhu-

man primates. J Comp Neurol 1991;306:685–696.

3. Bellamkonda R, Ranieri J, Aebischer P. Laminin oligopeptide

derivatized aragose gels allow three-dimensional neurite exten-

sion in vitro. J Neurosci Res 1995;41:501–509.

4. Bellamkonda R, Ranieri JP, Bouche N, Aebischer P. Hydrogel-

based three-dimensional matrix for neural cells. J Biomed Mater

Res 1995;29:663–671.

5. Yu XJ, Dillon GP, Bellamkonda RV. A laminin and nerve growth

factor-laden three-dimensional scaffold for enhanced neurite

extension. Tissue Eng 1999;5:291–304.

6. Pietri T, Eder O, Breau MA, Topilko P, Blanche M, Brakebusch C,

Fassler R, Thiery J-P, Dufour S. Conditional {beta}1-integrin gene

deletion in neural crest cells causes severe developmental altera-

tions of the peripheral nervous system. Development 2004;131:

3871–3883.

7. Borkenhagen M, Clemence JF, Sigrist H, Aebischer P. Three-

dimensional extracellular matrix engineering in the nervous sys-

tem. J Biomed Mater Res 1998;40:392–400.

8. Lamoureux P, Zheng J, Buxbaum RE, Heidemann SR. A cytome-

chanical investigation of neurite growth on different culture surfa-

ces. J Cell Biol 1992;118:655–661.

9. Tonge DA, Golding JP, Edbladh M, Kroon M, Ekstrom PE,

Edstrom A. Effects of extracellular matrix components on axonal

outgrowth from peripheral nerves of adult animals in vitro. Exp

Neurol 1997;146:81–90.

10. Ivins JK, Yurchenco PD, Lander AD. Regulation of neurite out-

growth by integrin activation. J Neurosci 2000;20:6551–6560.

11. Willits RK, Skornia SL. Effect of collagen gel stiffness on neurite

extension. J Biomater Sci Polym Ed 2004;15:1521–1531.

12. Wood MD, Willits RK. Applied electric field enhances DRG neurite

growth: Influence of stimulation media, surface coating and

growth supplements. J Neural Eng 2009;6:046003.

13. Merrill E, Salzman E. Polyethylene oxide as a biomaterial. ASAIO

J 1983;6:60–64.

14. Drury JL, Mooney DJ. Hydrogels for tissue engineering: Scaffold

design variables and applications. Biomaterials 2003;24:4337–

4351.

15. Peppas NA, Keys KB, Torres-Lugo M, Lowman AM. Poly(ethylene

glycol)-containing hydrogels in drug delivery. J Control Release

1999;62:81–87.

16. Mann BK, Gobin AS, Tsai AT, Schmedlen RH, West JL. Smooth

muscle cell growth in photopolymerized hydrogels with cell adhe-

sive and proteolytically degradable domains: Synthetic ECM ana-

logs for tissue engineering. Biomaterials 2001;22:3045–3051.

17. Bryant SJ, Anseth KS, Lee DA, Bader DL. Crosslinking density

influences the morphology of chondrocytes photoencapsulated in

PEG hydrogels during the application of compressive strain.

J Orthop Res 2004;22:1143–1149.

18. Bryant SJ, Anseth KS. Hydrogel properties influence ECM produc-

tion by chondrocytes photoencapsulated in poly(ethylene glycol)

hydrogels. J Biomed Mater Res 2002;59:63–72.

19. Skornia SL, Bledsoe JG, Kelso B, Willits RK. Mechanical proper-

ties of layered poly(ethylene glycol) gels. J Appl Biomater Bio-

mech 2007;5:176–183.

20. Gunn JW, Turner SD, Mann BK. Adhesive and mechanical proper-

ties of hydrogels influence neurite extension. J Biomed Mater Res

A 2005;72:91–97.

21. Bi JJ, Downs JC, Jacob JT. Tethered protein/peptide-surface-

modified hydrogels. J Biomater Sci Polym Ed 2004;15:905–916.

22. Dai W, Belt J, Saltzman W. Cell-binding peptides conjugated to

poly(ethylene glycol) promote neural cell aggregation. Bio/Tech-

nology 1994;12:797–801.

23. Rhee W, Rosenblatt J, Castro M, Schroeder J, Rao PR, Harner

CFH, Berg RA. In vivo stability of poly(ethylene glycol) collagen

composites. In: Harris Z, editor. ACS Symposium #680: Poly(Eth-

ylene Glycol): Chemistry and Biological Applications. Washington,

DC: American Chemical Society; 1997. p 420–440.

24. Zalipsky S, Seltzer R, Menon-Rudolph S. Evaluation of a new rea-

gent for covalent attachment of polyethylene glycol to proteins.

Biotechnol Appl Biochem 1992;15:100–114.

25. Johnson P, Skornia S, Stabenfeldt S, Willits R. Maintaining bioac-

tivity of NGF for controlled release from PLGA using PEG.

J Biomed Mater Res A 2008;86:420–427.

26. Buxton AN, Zhu J, Marchant R, West JL, Yoo JU, Johnstone B.

Design and characterization of poly(ethylene glycol) photopoly-

merizable semi-interpenetrating networks for chondrogenesis of

human mesenchymal stem cells. Tissue Eng 2007;13:2549–2560.

27. Belcheva N, Baldwin SP, Saltzman WM. Synthesis and characteri-

zation of polymer-(multi)-peptide conjugates for control of spe-

cific cell aggregation. J Biomater Sci Polym Ed 1998;9:207–226.

28. Silver FH, Trelstad RL. Type I collagen in solution. Structure and

properties of fibril fragments. J Biol Chem 1980;255:9427–9433.

29. Zilinski JL, Kao WYJ. Tissue adhesiveness and host response of

in situ photopolymerizable interpenetrating networks containing

methylprednisolone acetate. J Biomed Mater Res A 2004;68:

392–400.

30. Aklonis J, MacKnight W. Introduction to Polymer Viscoelasticity.

New York: John Wiley & Sons Inc.; 1983.

31. Mellott MB, Searcy K, Pishko MV. Release of protein from highly

cross-linked hydrogels of poly(ethylene glycol) diacrylate fabri-

cated by UV polymerization. Biomaterials 2001;22:929–941.

32. Watkins AW, Anseth KS. Investigation of molecular transport and

distributions in poly(ethylene glycol) hydrogels with confocal

laser scanning microscopy. Macromolecules 2005;38:1326–1334.

33. Heilshorn SC, Liu JC, Tirrell DA. Cell-binding domain context

affects cell behavior on engineered proteins. Biomacromolecules

2004;6:318–323.

34. Murphy WL, Mercurius KO, Koide S, Mrksich M. Substrates for

cell adhesion prepared via active site-directed immobilization of a

protein domain. Langmuir 2003;20:1026–1030.

35. Sarig-Nadir O, Seliktar D. Compositional alterations of fibrin-

based materials for regulating in vitro neural outgrowth. Tissue

Eng Part A: Tissue Eng 2008;14:401.

ORIGINAL ARTICLE

JOURNAL OF BIOMEDICAL MATERIALS RESEARCH A | 1 JUN 2010 VOL 93A, ISSUE 3 823