at SciVerse ScienceDirect

Biomaterials 32 (2011) 9649e9657

Contents lists available

Biomaterials

journal homepage: www.elsevier .com/locate/biomater ia ls

Two-tier hydrogel degradation to boost endothelial cell morphogenesis

Karolina Chwalek, Kandice R. Levental, Mikhail V. Tsurkan, Andrea Zieris, Uwe Freudenberg,Carsten Werner*

Leibniz Institute of Polymer Research Dresden (IPF), Max Bergmann Center of Biomaterials Dresden (MBC) & Technische Universität Dresden,Center for Regenerative Therapies Dresden (CRTD), Hohe Str. 6, 01069 Dresden, Germany

a r t i c l e i n f o

Article history:Received 6 July 2011Accepted 26 August 2011Available online 19 September 2011

Keywords:AngiogenesisHydrogelDegradationEndothelial cellMatrix metalloproteinase

* Corresponding author.E-mail address: werner@ipfdd.de (C. Werner).

0142-9612/$ e see front matter � 2011 Elsevier Ltd.doi:10.1016/j.biomaterials.2011.08.078

a b s t r a c t

Cell-responsive degradation of biofunctional scaffold materials is required in many tissue engineeringstrategies and commonly achieved by the incorporation of protease-sensitive oligopeptide units. Inextension of this approach, we combined protease-sensitive and -insensitive cleavage sites for the far-reaching control over degradation rates of starPEG-heparin hydrogel networks with orthogonallymodulated elasticity, RGD presentation and VEGF delivery. Enzymatic cleavage was massively acceleratedwhen the accessibility of the gels for proteases was increased through non-enzymatic cleavage of esterbonds. The impact of gel susceptibility to degradation was explored for the 3-dimensional ingrowth ofhuman endothelial cells. Gels with accelerated degradation and VEGF release resulted in stronglyenhanced endothelial cell invasion in vitro as well as blood vessel density in the chicken chorioallantoicmembrane assay in vivo. Thus, combination of protease-sensitive and -insensitive cleavage sites canamplify the degradation of bioresponsive gel materials in ways that boost endothelial cellmorphogenesis.

� 2011 Elsevier Ltd. All rights reserved.

1. Introduction

Tissue regeneration is based on the ability of cells to expand,degrade, invade, and dynamically remodel the surrounding extra-cellular matrix (ECM) [1] using a sophisticated enzymaticmachinery [2]. ECM rearrangement is especially important whencells become hypoxic and the development of new blood vessels isrequired. Efficient remodeling then becomes a critical prerequisitedetermining the maintenance of a tissue. Biomaterials promotingtissue healing and regeneration need to mimic and modulate keycharacteristics of naturally occurring ECM types [3] with respect tovarious different biochemical and biophysical signals [4]. Degra-dation rates adjusted to the desired migratory activity of particularcell types define a very important requirement for such materials.The degradation of polymer hydrogels can be engineered throughlinkages between the building blocks that undergo cleavage uponaction of specific stimuli such as light, pH changes or enzymaticactivity. Ester bonds are among the most common non-enzymatically cleavable bonds in materials. Being stable over thepH range of 3.0e7.0 many esters slowly become hydrolyzed atphysiological conditions (pH 7.4, 37 �C) [5e7] at rates that can be

All rights reserved.

tuned by varying of its molecular environment [8,9], thus allowingsimple customization of the degradation properties. Enzymaticdegradation is usually integrated into materials through theinclusion of peptide sequences cleaved by specific cell-releasedenzymes. Among those, matrix metalloproteinases (MMPs) areattracting special attention because of their role in the cell-mediated remodeling of ECM during wound healing and tissueregeneration [10], andMMP-susceptible peptides were successfullyapplied in the design of a number of bioresponsive materials[11e13].

Herein, we aimed at developing multibiofunctional hydrogelswith far-going modulation of their degradation rates at otherwiseinvariant characteristics. For that purpose, MMP-cleavable peptideswere combined with linkers of differing hydrolytic sensitivity(Fig. 1) in the formation of star-shaped poly(ethylene glycol)(starPEG) e heparin hydrogels that offer independently tunablephysical and biomolecular properties [14]. The inclusion ofa specific linear oligopeptide into the hydrogel network waspreviously shown to render the material MMP-degradable [15],allowing for the localized cellular invasion into the hydrogel [16]. Tomodulate the degradation of the peptide-containing biohybrid gelmaterials we now applied two different chemical linkages betweenthe MMP-cleavable peptide and the starPEG units, i.e. an amidebond which is stable over the considered pH range and an esterbond which slowly degrades under physiological conditions. Using

mailto:werner@ipfdd.dewww.sciencedirect.com/science/journal/01429612http://www.elsevier.com/locate/biomaterialshttp://dx.doi.org/10.1016/j.biomaterials.2011.08.078http://dx.doi.org/10.1016/j.biomaterials.2011.08.078http://dx.doi.org/10.1016/j.biomaterials.2011.08.078

Fig. 1. Schematic representation of degradable starPEG-heparin hydrogels containing either ester (EG) or amide (AG) conjugated MMP-cleavable peptides.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e96579650

these two types of gel materials with varied elasticity, RGD conju-gation and VEGF delivery we were able to explore the impact ofdifferences in the rate of gel degradation on the morphogenesis ofendothelial cells (ECs) in vitro and on the angiogenesis in thechicken chorioallantoic membrane (CAM) assay in vivo.

2. Materials and methods

2.1. Preparation of acryloyl terminated 4-armed PEG

Synthesis of the acryloyl terminated 4-armed PEG was performed as describedbefore [15]. In brief, 0.1 g (1 � 10�5 mol) of commercially available hydroxyl-terminated 4-armed PEG (PEG-(OH)4 MW ¼ 10 kDa) was dissolved in 0.5 ml ofCH2Cl2, next 90 ml (6 � 10�5 mol) of acryloyl chloride was added following by slowaddition of 115 ml of TEA. The reactionwas stirred overnight under N2. About 300mgof dry Na2CO3 was added to the reaction mixture and stirred for 2 h. The reactionmixture was filtrated and precipitated from 150 ml of diethyl ether twice. Thecollected white precipitate was dried under vacuum overnight and kept at e 20 �C.1H NMR (500 Mz CDCl3): d 6.42 (dd, J ¼ 18.0, 1.4 Hz, 4H), 6.15 (dd, J ¼ 18.0, 1.4 Hz,4H), 5.85 (dd, J ¼ 18.0, 1.4 Hz, 4H), 4.31 (t, J ¼ 1 Hz, 8H), 3.79e3.49 (m, 1200H), 3.41(s, 8H). The conversion rate of the terminal hydroxyl groups on PEG was calculatedfrom the ratio between the signals of the vinyl group residues at d 5.8e6.5 and thePEG core at d 3.41 (100% corresponds to a ratio of 2:1).

2.2. Preparation of maleimide terminated 4-armed PEG

0.1 g (1 � 10�5 mol) of amino-terminated 4-armed PEG (PEG-(NH2)4MW ¼ 10 kDa) was dissolved in 0.5 ml of CH2Cl2, next 112 mg of (4.2 � 10�5 mol) 3-Maleimidopropionic acid N-hydroxysuccinimide ester was added. The reaction wasstirred overnight under N2, following double precipitation from 150 ml of diethylether. The collected white precipitate was dried under vacuum overnight and keptat �20 �C. 1H NMR (500Mz CDCl3): d 6.42 (s 8H), 6.27 (s, 4H), 3.85 (t, J ¼ 7.5, 8H)3.79e3.49 (m, 1216H), 3.41 (m, 16H), 2.52 (t, J ¼ 7.5 Hz, 8H). The conversion rate of

terminal aminogroupsof PEGwas calculated fromthe ratiobetween the signals of themaleimide group residues at d 6.42 and the PEG core at d 3.41 (100% corresponds toa ratio of 1:2).

2.3. Preparation of MMP-cleavable peptide

The MMP-cleavable sequence GPQG[YIWGQ was included as a bioactivemodule into peptide NH2-GGPQGIWGQGGCG-CONH2 (IWGC) which was synthe-sized using solid-phase methods on an Activo P11 (Activotec, UK) peptide synthe-sizer by standard Fmoc-chemistry with a C-terminal capping protection strategy.Activation was achieved by O-(Benzotriazol-1-yl)-N,N,N0 ,N0-tetramethyluroniumtetrafluoroborate (TBTU), and 1-hydroxybenzotriazole (HOBt) in DMF. Deprotectionof the amino acid side chains and cleavage from the resinwas performed by reactionwith a mixture of trifluroacetic acid (85% v/v), phenol (5% v/v), dithiothreitol (2.5% v/v), tri-isopropyl silane (2.5% v/v) and water (5% v/v) for 2.5 h at room temperature.The crude peptide was precipitated in cold anhydrous diethyl ether, collected byvacuum filtration, extensively washed with diethyl ether and dichloromethane, anddried under vacuum. Final purification was achieved by preparative reversed-phaseHPLC (Agilent Technologies 1200 Series; linear gradient H2O/AcN on XBridge BEH300 C-18 column 10 mM particle size, 19 � 250 mm) equipped with a UV/Visdetector/spectrophotometer. Purity of the peptide and the accuracy of the synthesiswere evaluated by a single peak in analytical reversed-phase HPLC (Agilent Tech-nologies 1100 Series; on XBridge BEH 300 C-18 column 5 mM particle size,2.1 � 250 mm, Waters, USA) and accurate molecular ion mass in ESI-MS (Marinerspectrometer, Applied Biosystems, Germany).

2.4. Preparation of starPEG-MMP conjugates

Both maleimide- or acryloyl-functionalized PEG were converted into theirstarPEG-MMP conjugates by simplemixing of stoichiometric amounts of the peptideand functionalized starPEG (10%weight/volume) in phosphate buffer pH¼ 7. A smallamount of tris(2-carboxyethyl)phosphine chloride (TCEP) was added to the reactionmixture in order to prevent oxidation of the cysteine residue. The reaction mixturewas stirred overnight and purified by dialysis (nitrocellulose membrane with

K. Chwalek et al. / Biomaterials 32 (2011) 9649e9657 9651

MWCO ¼ 2 kDa) against water. White fluffy remnants were collected after lyophi-lizing the dialyzed solution. The purity of the formed starPEG-MMP conjugates wasevaluated by single peak in analytical reverse phase HPLC as explained in detailbefore [15,16].

2.5. Preparation of starPEG-heparin gels and modification with RGD

Biodegradable starPEG-heparin gels were prepared as previously reported fornon-cleavable starPEG-heparin gels using the starPEG-MMP conjugates and EDC-s-NHS activated heparin [15]. Briefly, hydrogels were prepared as planar layers ofapproximately 200 mm thickness. The liquid gel mixture was placed between anaminofunctionalized and a hydrophobic glass cover slip and allowed to polymerizeovernight. After removal of the hydrophobic cover slip, solid gels were washed andswollen in phosphate buffered saline (PBS) for 24 h. For the in vitro and in vivostudies, the hydrogels were functionalized with cyclicRGD (Peptides International,Louisville, KY, USA) using the reaction of EDC/s-NHS-activated carboxylic acidsgroups of heparin with 5 or 20 mg/ml cyclicRGD. All samples were extensivelywashed with PBS before use in cell culture experiments.

2.6. Rheological measurements

Storage moduli of the hydrogels were determined as previously described [14].Briefly, oscillating measurements on swollen gel disks were carried out on a rota-tional rheometer (Ares LN2, TA Instruments, Eschborn, Germany) fitted witha 25 mm parallel plate geometry. Frequency sweeps were carried out at 25 �C ina shear frequency range of 102e10�1 rad s�1 with the strain amplitude of 2%. Bothstorage and loss modulus were measured as a function of the shear frequency. Meanvalues of the storage modulus were calculated. Experiments were performed intriplicate and repeated at least three times.

2.7. Enzymatic degradation studies

5 ml gel droplets were used for degradation experiments. After gelation over-night, the hydrogel drops were washed 3 times with PBS and swollen in PBS over-night. To determine the kinetics of degradation, gel drops were placed in plastic UVcuvettes (PlastiBrand, Germany) with either 2 ml of PBS or 1.0U/ml collagenase IV(Biochrom AG, Germany). The UV absorption was determined at 278 nm usinga spectrophotometer (Beckman Coulter, DU800, USA) and recorded for all samplesevery 15 min for a time period of 2500 min in total. The samples during themeasurements were thermostated at 37 �C.

2.8. Mass swelling degradation studies

In order to visualize the gel materials, Alexa-488 (Invitrogen, USA) labeledheparinwas used for the gel formation. 50 ml gel samples were formed in 0.5 ml vialsfor the degradation experiments. After overnight polymerization the hydrogelsamples were washed 5 times (by 15 min) with PBS and swollen in PBS overnight.The PBS solution was removed by pipetting and the vials with the gel samples wereweighted, themeasuredmasses were used as an initial mass by subtraction themassof the empty vials. To determine the kinetics of degradation, the gel samples wereincubated at 37 �C with 400 ml of 10.0 U/ml collagenase IV (Biochrom AG, Germany)solution in PBS. Fresh solutions of collagenasewere prepared at least each 12 h. Massloss of the samples was determined during degradation by precision weighing(Sartorius, Germany). In parallel to the mass determination all samples were imagedin the dark with front illumination by UV lamp (366 nm; Benda NU-8-KL, Germany).

2.9. VEGF uptake and release

Surface-bound gels (n ¼ 3) were placed in custom-made incubation chamberswith minimized contact of the solution to additional surfaces [17]. 200 ml of VEGF(1 mg/ml; Peprotech, Hamburg, Germany) solution were added per cm2. Immobili-zation was performed overnight at 22 �C. The VEGF solution was subsequentlyremoved, followed by washing with PBS twice. Each of the solutions was collectedand assayed in duplicates using an ELISA Quantikine kit (R&D Systems, Minneapolis,USA). Immobilization under conditions described above resulted in w198 ng VEGFper cm2 gel surface area. Following the loading procedure, VEGF was released fromthe gels at 22 �C into 250 ml/cm2 of serum-free endothelial cell growth medium(ECGM; Promocell GmbH, Heidelberg, Germany) supplemented with 0.02% sodiumazide (Fluka) � 0.5U collagenase IV (Biochrom AG, Berlin, Germany). Samples takenat intervals were stored at �80 �C until analyzed by ELISA. An equal volume of freshmedium was added at each time point.

2.10. Cell culture experiments

Before cell seeding, the hydrogels were equilibrated in cell culture medium for1 h at 37 �C. Human umbilical vein endothelial cells (HUVECs), isolated according tothe procedure proposed by Weis et al. [18] were used. The cells were cultured up to80% confluence in ECGM containing 2% calf serum at 37 �C and 5% CO2. Cells from

passage 2e6were used in these studies. Cell morphology, attachment and spreadingwere monitored daily with light microscopy (Olympus IX50, 10� objective).

2.11. Cell adhesion

Cells were seeded at a density of 1 � 104/cm2 in serum-free medium. After30 min samples were fixed with 2% paraformaldehyde (PFA) and actin was visual-ized with AlexaFluor 633-labeled Phalloidin (Invitrogen, Germany). Samples wereinvestigated using a confocal laser scanning microscope (SP5, Leica Microsystems,Germany, 20�/0.70 objective). Projected cell area was measured with ImageJ (NIH)software.

2.12. Viability/proliferation WST-1 assay

Cells were seeded at a density of 1 � 104/cm2. After 7 days cell viability/prolif-erationwas estimated using a colorimetricWST-1 assay (Roche, Germany) accordingto the manufacturer’s instruction. Briefly, cells were incubated for 30 minwithWST-1 reagent diluted 1:10 (v:v) in serum-free medium, followed by a reading ofabsorbance at 450 nm.

2.13. 3-dimensional migration studies

Hydrogels were created with 0.5% addition of heparin-Alexa-488. Before the cellseeding gels were incubated overnight with 1 mg of recombinant human VEGF165(Peprotech, Germany) in PBS at room temperature. Cells were seeded at a density of4 � 104/cm2 and cultured for 24 h in serum-free medium, after which the sampleswere fixed with 2% PFA and actin was visualized with AlexaFluor 633-labelledPhalloidin (Invitrogen, Germany). Samples were investigated using a confocallaser scanning microscope (SP5, Leica Microsystems, Germany, 20�/0.70 objective).

2.14. Zymography

For the zymography 2 � 105 cells were seeded in 6-well tissue culture treatedplate. Reaching 80% of confluency, cells were incubatedwith serum-free medium for24 h and VEGF 1 ng/ml was added. After 6 h cell culture medium was collected andconcentrated with Amicon Ultra Filters, 10 K (Millipore, Ireland). Protein content ofthe cell culture medium samples was determined with BCA kit (Pierce, Germany).Aliquots containing 28.5 mg of protein were loaded per each lane. Zymography wasperformed under non-reducing conditions on 10% polyacrylamide gels, copoly-merized with porcine skin gelatin 1 mg/ml (SigmaeAldrich, Germany). After 2 h ofelectrophoresis at 120 V, the gel was incubated at 37 �C overnight in freshlyprepared development buffer (0.05 M TriseHCl pH 8.8, 5 mM CaCl2, 0.02% NaN3).Following morning the gel was stained according to the standard procedure withCoomassie Briliant Blue (SigmaeAldrich, Germany). To assay gelatin lysis, images ofthe gel were taken using transmitted light with the Lumi Imager (Roche, Germany)and densitometric analysis was performed with ImageJ (NIH) software. The semi-quantitative results were expressed as a ratio of untreated sample.

2.15. CAM assay

Chicken eggs (Erzeugergemeinschaft Pharma-Ei GmbH, Germany) were ob-tained at embryonic development day (EDD) 0. After wiping with 70% ethanol theywere incubated at 37 �C, high humidity for 3 days with automatic turning. On EDD3eggs were opened to sterile weighting boats. The ex-ovo cultures were maintainedin a humified incubator at 37 �C. On EDD8 samples were applied on the outer regionsof the chorioallantoic membrane and cultures were brought back to the incubator.Matrigel (BD, Germany) was used as a positive control. On EDD 12 100 ml Indian Inkwas injected into the vasculature of the embryos using a syringe with 30G needlefollowing with immediate imaging with stereomicroscope (Leica Microsystems,S8AP0, 25�). The proangiogenic response was evaluated by analyzing the conver-gence of blood vessels toward the graft as depicted below (see Fig. 2). Blood vesselsgrowing within a distance of 1 mm from the samples were counted, normalized tothe number of vessels in the untreated CAM, and expressed as a relative vasculari-zation index. The results were expressed as a ratio of the untreated sample. Pictureswith a lower magnification as used for quantification are shownwithin the paper topoint out qualitative differences in vascularization.

2.16. Image analysis and statistics

All images were analyzed using the ImageJ software (NIH). Experiments wereperformed two times or more with at least triplicate samples in each of theexperiments. Statistical analysis was performed with GraphPad Instat software(GraphPad, CA, USA). Analysis of variance (ANOVA) was the primary analysis methodusing the Turkey post-hoc test or KruskaleWallis ANOVA on rankswith Dunn’s post-hoc modification depending on the results of the normality test. In addition, two-tailed Students’ t-test was employed, when appropriate. Levels of significancewere determined when * ¼ p < 0.05, ** ¼ p < 0.01, *** ¼ p < 0.001.

Fig. 2. Evaluation of proangiogenic response by macroscopic scoring of vessels. Thenumber of blood vessels growing within a distance of 1 mm from the samples wasdetermined.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e96579652

3. Results and discussion

3.1. Design of degradable starPEG-heparin hydrogels

In this work we sought to investigate the impact of the cleavagerate of MMP-cleavable materials on cell-mediated remodeling.MMP-susceptible peptides were incorporated into starPEG-heparinhydrogels through hydrolytically labile ester bonds or more stableamide bonds, respectively (Fig. 1). The selected MMP-responsivepeptide (GPQGY[IWGQ) is recognized and cleaved by severalproteases, including MMP1, MMP2, MMP3, MMP7, MMP8 andMMP9 [19]. The ester-linked PEG-peptide conjugate (PE-IWGC)was formed by substitution of all four arms of a hydroxyl-terminated starPEG (Mw ¼ 10000) with acrylate groups, followedby coupling with the thiol group of the cysteine in the peptide-containing the MMP-cleavable sequence (IWGC) [15]. In order toform the amide-linked PEG-peptide conjugate (PA-IWGC), a similaramino-terminated star PEG (Mw¼ 10000) was functionalized withmaleimide groups. The PEG-maleimide conjugatewas then coupledto the thiol group of the IWGC peptide as the PEG-acrylateconjugate.

3.2. Hydrogel degradation and growth factor release

Gel materials (AG stands for amide-linked gel, and EG stands forester-linked gels) were prepared by converting the N-terminalamino groups of the starPEG-peptide conjugates with carboxylicgroups of heparin by carbodiimide chemistry. Utilizing differentcrosslinking degrees permitted the design of a set of hydrogels witha range of storage moduli (2, 3.5, and 7.5 kPa). As shown in Fig. 3A,this allows to recapitulate similar physical properties in both EGand AG materials. The principal design concept of our hydrogelsystem permits the independent modulation of physical andbiomolecular characteristics [14] which offers most valuableoptions through the application of secondary biofunctionalizationschemes of heparin, including the covalent binding of adhesivepeptides and the non-covalent binding of a variety of growthfactors.

EG and AG type materials with a similar storage modulus of2 kPa were evaluated for biodegradation and release of VEGF(Fig. 3B and C). To examine biodegradation the gels were incubatedin a solution of bacterial collagenase IV (CLS) which has beenshown to be a reliable model of MMP activity [20]. Both materialsdegraded almost completely within 15 h, however the EG materialdegraded more rapidly than the AG material. In the absence ofenzyme, only minor degradation of both materials occurred,although the initial degradation of EG material was higher than forAG materials. However, EG materials were found to persist for atleast several weeks in absence of enzyme, which is in agreementwith results from others who had successfully applied ester-linkedPEGePEG [21,22] or PEG-heparin [6,23,24] gels in long termexperiments.

The release of VEGF from the materials, a decisive parameter forthe functionality of the gels to stimulate endothelial cells, wasfound to correlate well with the erosion of the compared gel types(Fig. 3C): In PBS the release of VEGF was slow (1.6e4.5 ng/cm2) andcomparable with previous data of non-cleavable starPEG-heparingels [14,25] whereas collagenase-induced degradation of the gelssignificantly increased the release of hydrogel-associated VEGF (to56e68 ng/cm2). After 180min 54 ng VEGF/cm2 and 24 ng VEGF/cm2

were released for EG and AG materials, respectively. This ratio ofw2.3 corresponds to the twofold higher degradation rate of the EGmaterials.

To analyze the influence of the degree of crosslinking on thedegradation rate, EG and AG materials of similar storage moduli 2,3.5, and 7.5 kPa were degraded in the presence of CLS. The initialdegradation rate was linear, as shown before in similar systems[11], and evidently modulated by the crosslinking degree for bothEG and AG materials (Fig. 3D).

To dissect the degradation of ester linkages from the MMP-degradation we furthermore prepared gels containing a peptideof a similar amino acid content with scrambled sequence, which isinsensitive to MMPs. EG-scr materials showed after a small initialswelling a rather slow decomposition over 75 h to 80e85% of theinitial mass. In contrast, EG materials showed a sharp drop in themass which led to complete degradation within 24 h and AGmaterials showed a slower degradation leading to about two thirdsof mass loss within 24 h and a similarly complete decompositionwithin 75 h (Fig. 3B and E).

While degradation of ester-linked, MMP-cleavable PEGePEGgels was reported to occur across the whole volume of thematerial (obvious from the initially increased swelling of thedegrading sample) [11,12,26] both EG and AG materials do notexhibit this behavior but a primarily surface erosion drivendegradation process. Considering the mesh size of the hydrogels(estimated from rubber elasticity theory, see [14] for details, withaverage mesh sizes of 13 � 2.5 nm for EG or AG type A gels,11 � 1.5 nm for EG and AG type B, and 8 � 1 nm for EG and AGtype C gels) MMPs with molecular weights of 67e92 kDa [19],which were shown to be secreted by endothelial cells in culture(see Suppl. Fig. 1), can be assumed to be restricted from enteringthe EG and AG materials. (The impact of the mesh size on theuptake of BSA (Mw 68 kDa) into comparable starPEG-heparinhydrogels was previously investigated [14].) However, thedecomposition rate of the EG materials exceeds the superpositionof the rates of the separated enzymatic (¼ AG) and non-enzymatic (¼ EG-scr) degradation processes. We thereforeconclude that the non-enzymatic degradation of ester bonds inthe EG materials increases the accessibility for MMPs andconsequently accelerates MMP-degradation.

In consequence, starPEG-heparin hydrogels can be customizedto adjust the degradation rates, including systems that exclusivelyundergo surface erosion (AG) and others for which this process is

Fig. 3. Characterization of cell-responsive starPEG-heparin hydrogels: (A) Mechanical properties of the hydrogels: A-mesh size 13 � 2.5 nm, B-mesh size 11 � 1.5 nm, C-mesh size8�1 nm, (B) Decomposition kinetics of the 2 kPa, monitored by UV-spectroscopy of the released peptide, (C) Cumulative VEGF release from the type 2 kPa EG and AGhydrogels � CLS monitored by ELISA, (D) Enzymatic decomposition (þCLS) of the EG and AG hydrogels of different stiffness: A-2kPa, B-3.5 kPa, C-7.5 kPa, monitored by UV-spectroscopy of released peptide, (E) Enzymatic decomposition (þCLS) of the EG hydrogel with scrambled peptide sequence (EG-scr), EG and AG hydrogels, monitored by weighing.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e9657 9653

amplified by the non-enzymatic cleavage of ester bonds (EG). Thevariability of the degradation rate at otherwise constant proper-ties (elasticity, adhesion ligand conjugation and growth factordelivery) defines a valuable advantage when choosing materialsto support particular tissue regeneration schemes. With thisextension, our platform of cell-instructive biohybrid matricesprovides a unique base to investigate the impact of the matrixsusceptibility on the cellular response in vitro and in vivo. Thiswas exemplarily studied herein for the morphogenesis of humanendothelial cells (ECs).

3.3. Spreading and long term viability of primary endothelial cells

EC morphogenesis is directed by both the physical andbiochemical properties of the cell substratum, in particular thematerial’s viscoelasticity and the availability of adhesive sites andrelevant growth factors [27,28]. First, we evaluated the initial cellspreading on different gels of different stiffness at invariant RGD

functionalization [14]. The cells efficiently adhered and rapidlyspread on the gels and the projected area of the ECs was similaron EG and AG materials, independent of their stiffness (Fig. 3A).Proliferation assays after a prolonged culture time of 7 daysshowed that the cells responded similarly to both gel typesindependently of the material stiffness in a range 2e7.5 kPa.However, for the 2 kPa EG material only 50% of cells could bedetected after this time period due to the relatively fast degra-dation of the gel (Figs. 4B and 3D). Thus, the gel propertiesinduced a largely similar initial cell response proving the feasi-bility of the above-mentioned idea to create materials differing inthe degradation characteristics only. ECs are in general thought tobe influenced by the susceptibility of the underlying matrix[29,30] since this process is known to play an important role invessel formation [31]. Therefore, EC morphology was monitoredover a time period of 24 h to explore the proangiogenic potentialof our cell-responsive hydrogels with varied degradationdynamics.

Fig. 4. Endothelial cell adhesion and growth on the starPEG-heparin hydrogels (A) Initial cell spreading analyzed by projected cell area. (B) Cell viability/proliferation measured bycolorimetric WST-1 assay after 7 days of culture.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e96579654

3.4. Endothelial cell morphogenesis and hydrogel invasion

Primary ECs are able to keep their phenotype during thein vitro culture and have the ability to form tube-like structureswithin 24 h after seeding on appropriate substrates [32]. More-over, tube formation is promoted when EC adhesiveness isreduced by blocking integrins with antibodies or by reducingadhesive ligand concentration [33]. Also, the anchorage of adhe-sion ligands to polymer surfaces was shown to influence capillaryformation through the force balance of integrin-ligand-interactions [34,35]. Since cell response is regulated by both thebiochemical and biophysical properties of a substrate [36e38], weextended the cell morphology study to EG and AG materialsvarying in crosslinking degree (3.5 vs. 7.5 kPa), RGD-liganddensity (5 vs. 20 mg/ml) and presentation of growth factors(loading with 0 vs. 1 mg VEGF per gel sample). In contrast to celladhesion and proliferation, cell morphology and migration weredirectly influenced by the degradation of the gel materials. Cellsgrowing on the faster degrading EG materials were more spindle-shaped in morphology than those on the slower degrading AGmaterials (Fig. 5A and B). Moreover, EG materials better sup-ported the formation of cord-like structures and the three-dimensional (3D) expansion of the ECs (Fig. 5A and B, arrowsfor cord-like structures). This might be due to the fact, that forfast degrading EG materials the cellular microenvironmentbecomes more rapidly adapted by cell-secreted ECM to betterpromote morphogenesis.

Increasing the RGD functionalization of the gels resulted ina change in the morphology of the cells, making them morespread or even provoked the formation of a confluent layer.Additionally, loading of the gels with VEGF increased thefrequency of cord-like structure formation and reduced theoverall spreading of the cells (Fig. 5A and B). Various studies haveshown that VEGF is a potent chemoattractant for ECs, influencestheir migration and morphogenetic events, such as tubularsprouting and network formation [39]. Zisch et al. found that PEGmatrices lacking VEGF showed less cell invasion than VEGF-incorporated matrices [40]. In line with these findings, weobserved a VEGF-stimulated cellular invasion into the hydrogelsdue to chemotaxis. Quantification of the invasion depth showedthat the 3D migration speed of cells plated on the fasterdegrading EG materials was modulated by both biochemical andmechanical properties of the gel scaffold. The invasion depthdecreased when the RGD concentration increased (Fig. 5C), whichis in line with results published by Lutolf et al. [11]. Moreover,there was a similar trend if the stiffness of the gel increased from

3.5 to 7.5 kPa (Fig. 5D). ECs did not invade the AG gels within24 h (Fig. 5D), even in the presence of VEGF and low adhesiveligand concentration, which further proved the observed slowererosion and higher stability of AG materials. Nevertheless,migratory behavior was observed after 48 h in AG gels with lowcrosslinking degree (Fig. 5E), suggesting that the initial degrada-tion rate of the materials is a crucial determinant in EC invasioninto the starPEG-heparin hydrogels. Thus, the degradation char-acteristics tremendously affect the EC arrangement, morphology,and migration rate.

From the comparison of the two gel types we conclude thatthe initial cell penetration in EG materials is accelerated due tothe combined effect of non-enzymatic and enzymatic surfacedegradation which is in line with the degradation behaviorobserved for this material in the presence of CLS (see Fig. 3).Additional effects may result from the faster release of VEGF andthe change of the mechanical resistance of the more rapidlydegrading materials. In turn, AG materials can be beneficial wheremechanical integrity of a scaffold is more important than fastercell/tissue ingrowth.

The importance of proteolysis in EC morphogenesis and angio-genesis is underpinned by the observation that VEGF induces thesynthesis of plasminogen activator, collagenases, and other prote-ases [41,42]. Zymography of the collected cell culture mediaconfirmed the presence of various proteinases within the super-natants and showed a slight upregulation after stimulation by VEGF(e.g. inactive and active forms of gelatinases MMP9 and MMP2, seeSuppl. Fig. 1). The detection of MMPs in the conditioned cell culturemedium shows that the lack of invasion into AG gels after 24 h hasto be attributed to the lower susceptibility of the material but not tothe lack of enzymes.

3.5. Angiogenesis in vivo e chicken CAM assay

Having demonstrated the decisive influence of the gel degra-dation on EC migration in vitro the proangiogenic potential of thebiodegradablematerial was further evaluated in vivo using the CAMassay (Fig. 6). In view of the results discussed above we putemphasis on VEGF-preloaded hydrogels with differing degradationcharacteristics to stimulate angiogenesis. In comparison with theuntreated CAM, slow degrading AG starPEG-heparin hydrogelsalone did not induce any angiogenic response, while fast degradingEG materials caused an increase in the density of blood vesselssurrounding the implants. On the contrary, VEGF-loaded gels ofboth types evoked an angiogenic response and increased thevascular density of the CAM. The more rapid release of VEGF from

Fig. 5. Endothelial cell morphology and migration within the starPEG-heparin hydrogels. Representative planar and cross-sectional images illustrating cell morphology and 3D cellmigration within (A) EG materials and (B) AG materials after 24 h of culture, arrows indicate cord-like structures, gray line indicates the thickness of the cell monolayer. Quantifiedinvasion depth of endothelial cells into (C) the EG hydrogels at 5 and 20 mg/ml RGD and (D) EG and AG materials at 3.5 and 7.5 kPa stiffness, with and without VEGF, after 1 day ofculture. (E) Quantified invasion depth of endothelial cells into AG hydrogels at 2 kPa stiffness, after 2 days of culture.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e9657 9655

the fast degrading EG materials correlates well with the lesspronounced increase of vascularization compared to the slowdegrading AGmaterials, thus there is a clear link between the VEGFrelease (Fig. 3C) and the pro-angiogeneic response in the CAM assay(Fig. 6A and B). Importantly, the most effective starPEG-heparin geltype, for which the degradationwas boosted by the combination oftwo different degradation mechanisms, was found to be as efficientin promoting angiogenesis as Matrigel, which was used here as

a positive control. Considering the drawbacks of the widely usedMatrigel, in particular the unavoidable batch to batch variation, theill-defined composition and the presence of undesired compo-nents, our well-defined biohybrid materials can be seen as a veryattractive, highly reproducible alternative for many in vitro appli-cations. Further on, the reported results suggest that fast degradingstarPEG-heparin hydrogels can be applied to stimulate the forma-tion of new blood vessels in vivo.

Fig. 6. Chicken embryo chorioallantoic membrane (CAM) angiogenesis assay (A) Representative examples of CAM, as observed under a stereomicroscope at 25� magnification.(B) Scored angiogenic response to the applied starPEG-heparin hydrogels shown in arbitrary units.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e96579656

4. Conclusions

Bioresponsive starPEG-heparin hydrogels allow for the far-reaching modulation of erosion and cytokine release bycombining enzymatic and non-enzymatic degradation mecha-nisms. With this extension, the versatile hydrogel system

becomes suitable to evoke tissue regeneration on cellular level.We demonstrate that the multifactorial adjustment of thedefined polymer platform can induce a strong proangiogenicresponse in vitro and in vivo, opening new perspectives for theuse of biomaterials in cardiovascular regenerative therapies andbeyond.

K. Chwalek et al. / Biomaterials 32 (2011) 9649e9657 9657

Acknowledgments

We thank Mareike Roth for valuable technical assistance.U.F., M. T. and C.W. were supported by the Deutsche For-schungsgemeinschaft (DFG) through grants WE 2539-7/1, SFB 655and FOR/EXC999, and by the Leibniz Association. K.R.L., M. T. andC.W. were supported by the Seventh Framework Programme of theEuropean Union through the Integrated Project ANGIOSCAFF. C.W.was supported by the Bundesministerium für Bildung, Forschungund Technologie (BMBF) through grant 01 GN 0946.

Appendix. Supplementary material

Supplementarymaterial associatedwith this article can be found,in the online version, at doi:10.1016/j.biomaterials.2011.08.078.

References

[1] Ross R. Cell biology of atherosclerosis. Annu Rev Physiol 1995;57:791e804.[2] Chapman HA. Plasminogen activators, integrins, and the coordinated regula-

tion of cell adhesion and migration. Curr Opin Cell Biol 1997;9(5):714e24.[3] Choi CK, Breckenridge MT, Chen CS. Engineered materials and the cellular

microenvironment: a strengthening interface between cell biology andbioengineering. Trends Cell Biol 2010;20(12):705e14.

[4] Ulijn RV. Bioresponsive hydrogels. Mater Today 2007;10(4):40e8.[5] Kim J, Kim IS, Cho TH, Lee KB, Hwang SJ, Tae G, et al. Bone regeneration using

hyaluronic acid-based hydrogel with bone morphogenic protein-2 and humanmesenchymal stem cells. Biomaterials 2007;28(10):1830e7.

[6] Kim M, Lee JY, Jones CN, Revzin A, Tae G. Heparin-based hydrogel as a matrixfor encapsulation and cultivation of primary hepatocytes. Biomaterials 2010;31(13):3596e603.

[7] Hiemstra C, van der Aa LJ, Zhong Z, Dijkstra PJ, Feijen J. Novel in situ forming,degradable dextran hydrogels by Michael addition chemistry: synthesis,rheology, and degradation. Macromolecules 2007;40(4):1165e73.

[8] Jo YS, Gantz J, Hubbell JA, Lutolf MP. Tailoring hydrogel degradation and drugrelease via neighboring amino acid controlled ester hydrolysis. Soft Matter2009;5(2):440e6.

[9] Rodriguez-Galan A, Franco L, Puiggali J. Degradable poly(ester amide)s forbiomedical applications. Polymers 2010;3(1):65e99.

[10] Salo T, Makela M, Kylmaniemi M, Autio-Harmainen H, Larjava H. Expressionof matrix metalloproteinase-2 and -9 during early human wound healing. LabInvest 1994;70(2):176e82.

[11] Lutolf MP, Lauer-Fields JL, Schmoekel HG, Metters AT, Weber FE, Fields GB,et al. Synthetic matrix metalloproteinase-sensitive hydrogels for theconduction of tissue regeneration: engineering cell-invasion characteristics.Proc Natl Acad Sci U S A 2003;100(9):5413e8.

[12] Raeber GP, Lutolf MP, Hubbell JA. Molecularly engineered PEG hydrogels:a novel model system for proteolytically mediated cell migration. Biophys J2005;89(2):1374e88.

[13] Kim S, Chung EH, Gilbert M, Healy KE. Synthetic MMP-13 degradable ECMsbased on poly(N-isopropylacrylamide-co-acrylic acid) semi-interpenetratingpolymer networks. I. degradation and cell migration. J Biomed Mater Res A2005;75(1):73e88.

[14] Freudenberg U, Hermann A, Welzel PB, Stirl K, Schwarz SC, Grimmer M, et al.A star-PEG-heparin hydrogel platform to aid cell replacement therapies forneurodegenerative diseases. Biomaterials 2009;30(28):5049e60.

[15] Tsurkan MV, Levental KR, Freudenberg U, Werner C. Enzymatically degradableheparin-polyethylene glycol gels with controlled mechanical properties.Chem Commun (Camb) 2010;46(7):1141e3.

[16] Tsurkan MV, Chwalek K, Levental KR, Freudenberg U, Werner C. ModularstarPEG-heparin gels with bifunctional peptide linkers. Macromol RapidCommun 2010;31(17):1529e33.

[17] Zieris A, Prokoph S, Welzel PB, Grimmer M, Levental KR, Panyanuwat W, et al.Analytical approaches to uptake and release of hydrogel-associated FGF-2.J Mater Sci Mater Med 2010;21(3):915e23.

[18] Weis JR, Sun B, Rodgers GM. Improved method of human umbilical arterialendothelial cell culture. Thromb Res 1991;61(2):171e3.

[19] Nagase H, Fields GB. Human matrix metalloproteinase specificity studiesusing collagen sequence-based synthetic peptides. Biopolymers 1996;40(4):399e416.

[20] Seliktar D, Zisch AH, Lutolf MP, Wrana JL, Hubbell JA. MMP-2 sensitive, VEGF-bearing bioactive hydrogels for promotion of vascular healing. J Biomed MaterRes A 2004;68(4):704e16.

[21] Lee S-H, Miller JS, Moon JJ, West JL. Proteolytically degradable hydrogels witha fluorogenic substrate for studies of cellular proteolytic activity and migra-tion. Biotechnol Prog 2005;21(6):1736e41.

[22] Anderson SB, Lin C-C, Kuntzler DV, Anseth KS. The performance of humanmesenchymal stem cells encapsulated in cell-degradable polymer-peptidehydrogels. Biomaterials 2011;32(14):3564e74.

[23] Tae G, Kim Y-J, Choi W-I, Kim M, Stayton PS, Hoffman AS. Formation of a novelheparin-based hydrogel in the presence of heparin-binding biomolecules.Biomacromolecules 2007;8(6):1979e86.

[24] McGonigle JS, Tae G, Stayton PS, Hoffman AS, Scatena M. Heparin-regulateddelivery of osteoprotegerin promotes vascularization of implanted hydrogels.J Biomater Sci Polym Ed 2008;19(8):1021e34.

[25] Zieris A, Prokoph S, Levental KR, Welzel PB, Grimmer M, Freudenberg U, et al.FGF-2 and VEGF functionalization of starPEG-heparin hydrogels to modulatebiomolecular and physical cues of angiogenesis. Biomaterials 2010;31(31):7985e94.

[26] Metters A, Hubbell J. Network formation and degradation behavior ofhydrogels formed by Michael-type addition reactions. Biomacromolecules2004;6(1):290e301.

[27] Vernon RB, Angello JC, Iruela-Arispe ML, Lane TF, Sage EH. Reorganization ofbasement membrane matrices by cellular traction promotes the formation ofcellular networks in vitro. Lab Invest 1992;66(5):536e47.

[28] Ingber DE, Folkman J. Mechanochemical switching between growth anddifferentiation during fibroblast growth factor-stimulated angiogenesisin vitro: role of extracellular matrix. J Cell Biol 1989;109(1):317e30.

[29] Hanjaya-Putra D, Bose V, Shen YI, Yee J, Khetan S, Fox-Talbot K, et al.Controlled activation of morphogenesis to generate a functional humanmicrovasculature in a synthetic matrix. Blood 2011;118(3):804e15.

[30] Hanjaya-Putra D, Yee J, Ceci D, Truitt R, Yee D, Gerecht S. Vascular endothelialgrowth factor and substrate mechanics regulate in vitro tubulogenesis ofendothelial progenitor cells. J Cell Mol Med 2010;14(10):2436e47.

[31] Schnaper HW, Grant DS, Stetler-Stevenson WG, Fridman R, D’Orazi G,Murphy AN, et al. Type IV collagenase(s) and TIMPs modulate endothelial cellmorphogenesis in vitro. J Cell Physiol 1993;156(2):235e46.

[32] Kubota Y, Kleinman HK, Martin GR, Lawley TJ. Role of laminin and basementmembrane in the morphological differentiation of human endothelial cellsinto capillary-like structures. J Cell Biol 1988;107(4):1589e98.

[33] Gamble JR, Matthias LJ, Meyer G, Kaur P, Russ G, Faull R, et al. Regulation ofin vitro capillary tube formation by anti-integrin antibodies. J Cell Biol 1993;121(4):931e43.

[34] Pompe T, Markowski M, Werner C. Modulated fibronectin anchorage atpolymer substrates controls angiogenesis. Tissue Eng 2004;10(5e6):841e8.

[35] Herklotz M, Werner C, Pompe T. The impact of primary and secondary ligandcoupling on extracellular matrix characteristics and formation of endothelialcapillaries. Biomaterials 2009;30(1):35e44.

[36] Neff JA, Tresco PA, Caldwell KD. Surface modification for controlled studies ofcell-ligand interactions. Biomaterials 1999;20(23e24):2377e93.

[37] Maheshwari G, Brown G, Lauffenburger DA, Wells A, Griffith LG. Cell adhesionand motility depend on nanoscale RGD clustering. J Cell Sci 2000;113(10):1677e86.

[38] Yeung T, Georges PC, Flanagan LA, Marg B, Ortiz M, Funaki M, et al. Effects ofsubstrate stiffness on cell morphology, cytoskeletal structure, and adhesion.Cell Motil Cytoskeleton 2005;60(1):24e34.

[39] Abedi H, Zachary I. Vascular endothelial growth factor stimulates tyrosinephosphorylation and recruitment to new focal adhesions of focaladhesion kinase and paxillin in endothelial cells. J Biol Chem 1997;272(24):15442e51.

[40] Zisch AH, Lutolf MP, Ehrbar M, Raeber GP, Rizzi SC, Davies N, et al. Cell-demanded release of VEGF from synthetic, biointeractive cell ingrowthmatrices for vascularized tissue growth. FASEB J 2003;17(15):2260e2.

[41] Unemori EN, Ferrara N, Bauer EA, Amento EP. Vascular endothelial growthfactor induces interstitial collagenase expression in human endothelial cells.J Cell Physiol 1992;153(3):557e62.

[42] Pepper MS, Ferrara N, Orci L, Montesano R. Vascular endothelial growth factor(VEGF) induces plasminogen activators and plasminogen activator inhibitor-1in microvascular endothelial cells. Biochem Biophys Res Commun 1991;181(2):902e6.

http://dx.doi.org/10.1016/j.biomaterials.2011.08.078

Two-tier hydrogel degradation to boost endothelial cell morphogenesis1 Introduction2 Materials and methods2.1 Preparation of acryloyl terminated 4-armed PEG2.2 Preparation of maleimide terminated 4-armed PEG2.3 Preparation of MMP-cleavable peptide2.4 Preparation of starPEG-MMP conjugates2.5 Preparation of starPEG-heparin gels and modification with RGD2.6 Rheological measurements2.7 Enzymatic degradation studies2.8 Mass swelling degradation studies2.9 VEGF uptake and release2.10 Cell culture experiments2.11 Cell adhesion2.12 Viability/proliferation WST-1 assay2.13 3-dimensional migration studies2.14 Zymography2.15 CAM assay2.16 Image analysis and statistics

3 Results and discussion3.1 Design of degradable starPEG-heparin hydrogels3.2 Hydrogel degradation and growth factor release3.3 Spreading and long term viability of primary endothelial cells3.4 Endothelial cell morphogenesis and hydrogel invasion3.5 Angiogenesis in vivo – chicken CAM assay

4 Conclusions Acknowledgments Appendix Supplementary material References