Biomedical Microdevices

, Volume 12, Issue 3, pp 555–568

Hydrogel cell patterning incorporating photocaged RGDS peptides


  • Catherine A. Goubko
    • Department of Chemical and Biological EngineeringUniversity of Ottawa
  • Swapan Majumdar
    • Ottawa Health Research Institute, Loeb Building, Chronic Diseases ProgramThe Ottawa Hospital
  • Ajoy Basak
    • Ottawa Health Research Institute, Loeb Building, Chronic Diseases ProgramThe Ottawa Hospital
    • Department of Chemical and Biological EngineeringUniversity of Ottawa
    • Ottawa-Carleton Institute for Biomedical EngineeringUniversity of Ottawa

DOI: 10.1007/s10544-010-9412-7

Cite this article as:
Goubko, C.A., Majumdar, S., Basak, A. et al. Biomed Microdevices (2010) 12: 555. doi:10.1007/s10544-010-9412-7


As we aim towards enhancing our knowledge of complex cell behaviors and developing intricate cell-based devices and improved therapeutics, it becomes imperative that we be able to control and manipulate the spatial localization of cells. Here we have developed a novel strategy to pattern cells using a hyaluronic acid hydrogel material and photocaged RGDS (Arg-Gly-Asp-Ser) peptides. In this report, we discuss the chemical synthesis and photoactive properties of the caged peptides as well as the subsequent binding of these peptides to our hydrogel base. We further demonstrate the ability of this modified hydrogel material to pattern fibroblast cells on the micron scale using near-UV light exposure through a patterned photomask to selectively switch areas of the hydrogel surface from cell non-adhesive to cell adhesive. The cells are found to adhere and proliferate along the developed line patterns for at least 2.5 days, demonstrating significantly enhanced pattern longevity in comparison with previously reported studies.


Cell adhesionExtracellular matrixMicropatterningHyaluronic acidPeptide

1 Introduction

The nature of the microenvironment surrounding cells has a substantial impact on cell behavior (Liu and Chen 2005; Rosso et al. 2004; Sands and Mooney 2007; Yamada and Nelson 2007). In the body, complex tissue architectures exist with carefully crafted microenvironments to guide cell behavior. As a result, techniques are needed to gain in vitro design control over the cell microenvironment to allow for the creation of improved cell-based devices and to assist researchers in advancing our current knowledge of cell biology. In traditional culture methods, cells are seeded randomly on a surface. In contrast, cell patterning systems can allow for control over the degree of contact that cells make with an underlying material substrate and with neighboring cells (Chen et al. 1997; Goubko et al. 2009; Nelson and Chen 2002). These parameters can then be manipulated to optimize or explore cell functions while the high degree of control attained can allow for improved experimental reproducibility and lower variability in cell-based devices. As such, the potential applications for cell patterning systems are numerous. For example, cell patterning has demonstrated that the multicellular organization of cells determines their growth pattern (Nelson et al. 2005) and that cell shape can influence whether cells grow or die (Chen et al. 1997). Spatially defined neural networks have been patterned and will ultimately lend to a greater understanding of the organization of the nervous system and its function in information processing (Morin et al. 2006; Vogt et al. 2005a, b). In the field of tissue engineering, epithelial cells have been patterned and organized atop human lens capsules towards the development of a retinal implant (Lee et al. 2004), and patterned materials have been developed to maintain chondrocyte phenotype for cartilage regeneration (Petersen et al. 2002). Numerous other examples can be found in the literature and the future will undoubtedly bring more fascinating and far-reaching applications.

The majority of cell patterning methods developed to date have relied heavily on either photolithography or soft lithography techniques (Falconnet et al. 2006; Kleinfeld et al. 1988; Liu and Chen 2005; Shin 2007; Singhvi et al. 1994). While commonly used, these techniques suffer drawbacks in practice. For example, photolithography—originally developed for the semi-conductor industry—utilizes toxic solvents which can hinder the use of easily denatured biomolecules and requires specialized clean rooms and expensive equipment. This makes it inaccessible to most life scientists. Soft lithography, while much more amicable to the biomolecules used for cell patterning, often depends on adsorption as opposed to covalent binding of biomolecules to a surface in creating cell patterns, which can limit pattern longevity. In this study, we report a novel cell patterning platform based on a hydrogel which makes use of extracellular matrix (ECM) materials in order to more closely mimic the natural cell microenvironment and employs methods accessible to life scientists. Patterning is performed with light, and thus has the potential to eventually obtain resolutions near the single cell level (Kikuchi et al. 2008). Additionally, the biomolecules involved in our pattern formation are covalently bound to the base of the pattern to maximize longevity.

An overview of our cell patterning methodology can be seen in Fig. 1. The base of the design is a non-adhesive crosslinked hyaluronic acid (HA) hydrogel. RGDS peptides were bound to the hydrogel to create a cell-adhesive layer. This sequence is naturally found in the ECM as a recognition site within the cell-adhesive protein fibronectin and binds to integrin receptors located on the surfaces of cells (Hersel et al. 2003). In order to allow for pattern creation, a photo-labile caging group, 2-nitrobenzyl, was covalently bound to the RGDS peptides to disrupt integrin recognition and thus cell binding. Therefore, the photocaged-RGDS peptides immobilized to the HA base create a surface that is cell non-adhesive. The 2-nitrobenzyl photocage can be selectively removed from RGDS upon near-UV light exposure through a patterned photomask to produce adhesive RGDS regions on an otherwise cell non-adhesive background. In this manner, a cell pattern can be formed as dictated by the photomask used.
Fig. 1

Strategy to create a patterned surface: (a) Adhesive RGDS peptides are caged and bound to a non-adhesive HA hydrogel to produce a cell non-adhesive surface (b) Caging groups are removed upon exposure to UV light through a photomask (c) leaving patterned exposed regions of RGDS (d) to which cells can bind

In this work, the photocaged RGDS peptide was synthesized such that the 2-nitrobenzyl group was attached to the nitrogen atom of the backbone amide group between the Arg and Gly residues (designated as R[G]DS) as indicated by the symbol [G]. Such binding of a photocage group to a peptide backbone and its biochemical use have been recently demonstrated in the literature (Nandy et al. 2007; Rhee et al. 2008; Tatsu et al. 2002). In selecting where along the peptide to bind the cage, it was noted that substituting the Gly residue with Ala, where the only structural difference is a single methyl group, has been shown to knock out RGD binding activity (Cherny et al. 1993; Hersel et al. 2003). Therefore, we hypothesized that the introduction of a bulky group, such as 2-nitrobenzyl, would serve to severely disrupt the adhesive properties of the RGDS sequence. Furthermore, the stability of the bond that could be formed at this location on the peptide appeared superior to other possible sites such as the carboxylic acid group on the Asp residue (Bourgault et al. 2007) or the guanidinium group on the Arg residue (Wood et al. 1998).

During the course of this study, a couple of other research groups have independently reported the development of photoresponsive RGD-based peptides in communication papers with Petersen et al. being the first to do so (Ohmuro-Matsuyama and Tatsu 2008; Petersen et al. 2008). Both works were able to demonstrate significantly decreased cell adherence to caged RGD-modified surfaces versus RGD- or uncaged RGD- modified surfaces. In addition, Petersen et al. showed the formation of a preliminary cell pattern. However, cells showed significant off-pattern binding only 6 h after seeding. The strategy of Petersen et al. relied on an oligo(ethylene glycol) (OEG) linker to both tether the caged RGD-based peptide onto a solid silica-based surface and act as a cell non-adhesive background. Similarly, Ohmuro-Matsuyama and Tatsu used a poly(ethylene glycol) (PEG) linker to create a non-adhesive background on a poly-L-lysine coated culture dish (Ohmuro-Matsuyama and Tatsu 2008). Our strategy, in contrast, is based on a hydrogel of crosslinked ECM molecules in place of a solid synthetic surface to which peptides were bound with a zero-length crosslinker. Hydrogels based on natural materials have demonstrated great potential for biological and medical applications due to their biocompatibility (Peppas et al. 2006). As such, the development of a technique to spatially localize cells on such a material has increased potential for future tissue engineering applications.

This paper outlines the development of our cell patterning strategy including the development of the HA hydrogel base, along with the synthesis and analysis of the R[G]DS peptides. Finally, we demonstrate that cell patterns can be formed on our novel hydrogel surface which demonstrates significantly increased pattern longevity from those previously reported employing similar photocaged peptides.

2 Materials and methods

2.1 Materials

Peptide coupling reagents, i.e. N,N′-diisopropylcarbodiimide (DIC), 4-dimethylaminopyridine (DMAP), 2-(1H-7-Azabenzotriazol-1-yl)-1,1,3,3-tetramethyl uranium hexafluorophosphate methanaminium (HATU), and N,N-diisopropylethylamine (DIPEA) were purchased from Applied Biosystems (Framingham, MA). For peptide synthesis, protected amino acids, Boc-Arg(Pbf)-OH, H-Gly-OMe·HCl, Boc-Asp(OBut)-OH and H-Ser(But)-OBut were bought from ChemImpex International (Wooddale, IL) and Bachem Inc. (Torrance, CA). All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO) unless otherwise indicated. All chemicals were used as received.

2.2 Preparation of HA hydrogels

A final fermentation-derived hyaluronic acid sodium salt solution of 4 mg/mL concentration in sterile double distilled water (dd-water) was used to make HA gels. To prepare the crosslinked HA gel substrate, a solution of adipic acid dihydrazide (ADH, 15 mg) in 0.5 mL of sterile dd-water was filtered through a 0.45 µm filter and added to 10 mL of aqueous HA solution. The resulting solution was adjusted to pH 3.5 by slowly adding 1 M HCl. This was followed by the addition of 16.5 mg of filtered 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC) in 1 mL of sterile dd-water with vigorous agitation. To the wells of a 12-well polystyrene tissue culture plate, 1.4 mL of the resulting reaction solution was transferred to let set. To prevent premature crosslinking, the reaction solution was cooled prior to EDC addition. Gelation was allowed to occur overnight at room temperature. The resulting gels were then washed in sterile phosphate-buffered saline (PBS) solution at pH 7.4 (Invitrogen) for 24 h with light agitation, followed by a 24 h wash with dd-water. This initial HA layer was air dried for several days and two more layers of gel were deposited atop the first in the same fashion as that outlined above. Prior to use, gels were rehydrated to equilibrate in sterile PBS.

2.3 Peptide synthesis

The two tetrapeptides, caged RGDS (1) and native RGDS (2) (chemical structures shown in Fig. 2(a)), were synthesized stepwise via liquid phase and solid phase peptide chemistry methods, respectively. Figure 3 outlines the synthesis of the caged RGDS peptide (1). The preparation of native RGDS peptide (steps not shown) was accomplished by normal HATU/DIPEA mediated Fmoc based solid phase chemistry as described earlier (Basak et al. 2007). For the caged peptide (1), a 2-nitrobenzyl group was linked to the backbone amide nitrogen between the Arg and Gly residues. To achieve this, N-α-2′-nitrobenzyl-glycine methyl ester (3) was first prepared from o-nitro benzaldehyde by coupling it with glycine methyl ester in alkaline solution in the presence of sodium bicarbonate followed by reduction with sodium borohydride, as shown in Fig. 2(b). Specifically, o-nitro benzaldehyde (3.02 g, 0.02 mol) in methanol (50 ml) was added to 50 ml cold Gly-OMe·HCl (0.02 mol) NaHCO3 solution (2%, w/v) under agitation, and the reaction mixture continued to be stirred for an additional 1 h while the temperature was kept below 4°C. Subsequently, NaBH4 (0.800 g, 0.021 mol) was slowly added to the reaction while maintaining the temperature between 0–5°C. After stirring for another 2 h, the resultant solution was dried under vacuum to evaporate methanol, and the residue was extracted with ether (3 × 50 ml) at low temperature. The combined organic layer was washed with brine solution, dried over anhydrous Na2SO4 and fully evaporated under vacuum to obtain (3) as a dry brown gummy residue (1.2 g, not optimized). The caged tetrapeptide R[G]DS was then synthesized using this intermediate according to the scheme shown in Fig. 3. To a stirred, cold mixture of Boc-Arg(Pbf)-OH (1 mmol) and (3) (1.1 mmol) in dichloromethane (15 ml), DIC (182 mg, 1.2 mmol) was added followed by DIPEA (200 mg, 1.5 mmol) and a catalytic amount of DMAP. After completion of the reaction, as revealed by thin layer chromatography (TLC), dichloromethane was removed using a rotary evaporator, and ethyl acetate (25 ml) was added. The white precipitate (i.e. byproduct from DIC) formed was filtered off and washed with ethyl acetate (20 ml). The combined organic layer was evaporated and dried under vacuum. The white powder thus obtained was dissolved in dioxane (20 ml) and cooled to 0°C in an ice bucket. A 0.2 M aqueous NaOH (10 ml) solution was added dropwise for hydrolysis of the ester bond. After completion of the reaction (∼10 min) as revealed by TLC, dioxane was removed by rotary evaporation. The obtained residue was re-dissolved in 50 ml water and pH adjusted to 3.0 by 2 M aqueous KHSO4. The reaction mixture was extracted with ethyl acetate (3 × 20 ml) in a separatory funnel. The combined organic layer was dried over anhydrous Na2SO4 and evaporated to dryness under vacuum to afford Boc-Arg(Pbf)-[Gly]-OH (4) as a thick yellow powder with molecular m/z of 718 (MALDI-tof MS), consistent with its chemical structure. Similarly, Boc-Asp(OBut)-OH and H-Ser(But)-OtBu were coupled together under the same reagent conditions as mentioned above to produce the protected dipeptide Boc-Asp(OBut)-Ser(But)-OtBu (5). The Boc group was then selectively deprotected using 1 M HCl in dioxane for 1 h, as checked by MALDI-tof mass spectrometry (m/z 388 in agreement with the chemical structure) to produce H-Asp(OBut)-Ser(But)-OtBu (6).
Fig. 2

(a) Chemical structures of caged and native tetrapeptides used in the present study and (b) Chemical steps involved in the synthesis of caged Gly residue
Fig. 3

Complete schematic diagram showing various steps involved in the synthesis of caged RGDS tetrapeptide using liquid phase peptide chemistry

Coupling of the two dipeptides, Boc-Arg(Pbf)-[Gly]-OH (4) (amino acid within the second bracket indicates the caged residue) and H-Asp(OBut)-Ser(But)-OtBu (6), was then achieved using the same coupling procedure as described above leading to the formation of protected caged tetrapeptide (7). Finally, all protecting groups were cleaved by treating (7) at room temperature for 3 h with a mixture of trifluoroacetic acid (TFA)/triisopropylsilane (TIPS)/water (95 / 2.5 / 2.5). Subsequently, the TFA was removed under vacuum, and the crude peptide was precipitated by cold ether at ∼0°C. The precipitate was collected by centrifugation and the product was lyophilized after dissolving in 0.1% TFA in water. The crude caged peptide (1) was then purified by RP-HPLC and fully identified by MALDI-tof mass spectrometry (m/z 567) and proton NMR spectroscopy.

2.4 Photolysis

Near-UV irradiation from a 365 nm Longwave UV Lamp (Black-Ray B-100 Longwave UV lamp, 100 W, UVP, Upland, CA) was used to remove the 2-nitrobenzyl caging group from the R[G]DS peptide. Samples containing R[G]DS were irradiated on a platform 4 inches from the light source. Gels containing bound R[G]DS were irradiated for 10–12 min. For patterning experiments, a line-patterned mask was placed on the gel surface during UV exposure.

2.5 1H NMR spectrum of caged and uncaged RGDS peptides

1H NMR characterizations of the R[G]DS peptide before and after exposure to near-UV irradiation were conducted to examine the uncaging event. Two mg of the caged peptide was dissolved in deuterium oxide (D2O) and analyzed by 1H NMR spectrum employing a Bruker AVANCE 500 MHz Wide Bore spectrometer. The R[G]DS peptide solution was subsequently exposed to near-UV light for 1 h and the 1H-NMR spectrum was recorded again. The 1H NMR spectrum of caged R[G]DS peptide (Supplemental Fig. 1) indicates that there are possibly two different R[G]DS conformational states based on the appearance of two sets of doublets in δ 4–5.5 ppm region with significantly higher coupling constants (J ∼ 16–18 Hz) for the two benzylic geminal CH2-protons. This is also supported by the observed splitting profiles of aromatic protons. 1H NMR for caged R[G]DS peptide (500 MHz, D2O): δ 8.23 (0.5H, d, J = 8.2 Hz), 8.11 (0.5H, d, J = 8.2 Hz), 7.79 (0.5H, t, J = 7.5 Hz), 7.71 (1H, t, J = 7.5 Hz), 7.64 (1H, t, J = 7.5 Hz), 7.59 (0.5H, t, J = 7.5 Hz), 7.51 (0.5H, d, J = 7.8 Hz), 7.47 (0.5H, d, J = 7.8 Hz), 5.27 (0.5H, d, J = 17.2) and 5.16 (0.5H, benzylic, d, J = 15.8), 4.98 (1H, one Gly αH), 4.52–4.43 (m, 3H, second Gly αH, Asp and Ser αH), 4.21 (0.5H, d, J = 18 Hz) and 4.16 (0.5H, benzylic, d, J = 16.6 Hz), 3.96–3.83 (m, 2H, Ser βH, + HOD formed due to solvent exchange), 3.21–3.16 (m, 2H, Arg δH), 3.09–3.05 (m, αH), 2.96–2.90 (m, αH), 2.87–2.69 (m, 2H, Asp βH), 1.96–1.88 (m, 2H, βH of Arg) and 1.71–1.60 (m, 2H, γH of Arg). The proton integrations of some of the signals have been found to be half of a single proton suggesting that the above peptide exists in two conformational forms in D2O solvent. The large coupling constants with J ∼ 15–18 Hz for benzylic protons are consistent with the geminal nature of the protons. It may be pointed out that some of the proton assignments, particularly those of α-protons of amino acids, are tentative and may be interchanged.

Upon UV irradiation, the aromatic protons consisting of two sets of doublets at δ 8.23 and 8.11 ppm, two sets of triplets at δ 7.64 and 7.59 ppm, and another two sets of doublets at δ 7.51 and δ 7.47 all disappeared completely as expected (Supplemental Fig. 2), indicating the loss of the aromatic nitrobenzyl group. Instead, additional aromatic peaks appeared at δ 7.85–7.72 for 2H and δ 7.35–7.25 for another 2H due to the formation of 2-nitroso benzaldehyde following UV-irradiation. The formation of the latter product was further confirmed by the appearance of a singlet at δ 8.25 for aldehydic proton (CHO). The peaks at δ 5.27 ppm (0.5H, d, J = 17.2 Hz), 5.16 (0.5H, d, J = 15.8), 4.21 ppm (0.5H, d, J = 18 Hz), and 4.16 (0.5H, d, J = 16.6 Hz) correspond to benzylic protons of two conformational states of caged peptide which also disappeared as expected following the irradiation, indicating a regeneration of RGDS peptide.

2.6 Photolysis reaction rate investigation

To further study the R[G]DS photolysis reaction, the kinetics of the disappearance of R[G]DS upon UV irradiation was investigated. To this end, 1 mg/mL solutions of the caged peptide were prepared in PBS (pH 7.4) and exposed to near-UV light for varying times. The irradiated solutions were analyzed by RP-HPLC to determine, as a function of exposure time, the percentage of R[G]DS reacted based on the peptide’s predetermined retention time. The chromatographic separation of the reaction products was performed with a Waters 2695 Separation Module using a reversed phase column (Waters XBridge BEH 130 C18, 3.5 µm) employing a mobile phase in gradient elution mode with a flow rate of 1 mL/min at room temperature. Initially, a mobile phase consisting of 95% water with 0.1% TFA (Pierce) and 5% acetonitrile (Fisher) with 0.1% TFA was delivered, and after 5 min the solvent ratios were linearly decreased to 5% water (0.1% TFA): 95% acetonitrile (0.1% TFA) over a period of 35 min. UV absorbance was recorded for wavelengths ranging from 190 nm to 500 nm using a Waters 2996 Photodiode Array Detector.

2.7 Peptide binding to the hydrogel

To bind peptides (both RGDS and R[G]DS) to the HA gel, a solution consisting of 25 mg/mL EDC and 15 mg/mL N-hydroxysuccinimide (NHS) (Pierce) in 0.1 M 2-(N-morpholino) ethanesulfonic acid (MES) (Fisher) buffer was prepared. Subsequently, 1 mL of this solution was added to each HA gel that was previously transferred to and crosslinked in wells of a 12-well plate. The reaction was carried out at room temperature for 15 min to activate the carboxyl groups in HA for peptide binding. The gels were then washed once in PBS (pH 7.4) to get rid of unreacted chemical residuals. The desired peptide solution was subsequently added to the gels and allowed to react for 2 h. Finally, the resulting gels were washed in PBS overnight. In order to visualize the immobilized peptide, fluorescent bovine serum albumin (BSA) was used as a model biomolecule to demonstrate the immobilization of the protein to the HA gel. This reaction was carried out using 0.3 mg of FITC-BSA in bulk solution per gel as per the above procedure and the resulting gels were subsequently washed for 22 h to remove unbound protein.

2.8 Cell adhesion tests in vitro

To test the relative cell adhesiveness of the peptide modified HA surfaces, 3T3 fibroblasts were used as model adherent cells. The cells were routinely maintained in medium containing Dulbecco’s Modified Eagle’s Medium (DMEM) (Invitrogen) supplemented with 10% fetal bovine serum (FBS) (Invitrogen), 100 U/mL Penicillin (Invitrogen), and 0.1 mg/mL Streptomycin (Invitrogen), and kept in T-75 flasks at 37°C in a humidified environment containing 5% CO2. Prior to cell seeding on the experimental surfaces, cells were trypsinized, centrifuged to a pellet, re-suspended in culture medium and counted using a hemocytometer. The cells were subsequently seeded onto the experimental surfaces at a density of 1 × 104 cells/cm2. Cells plated onto the experimental surfaces were initially maintained for 24 h, after which the surfaces were gently (and carefully) washed twice with sterile PBS (pH 7.4) to remove non-adherent cells. The adherent cells were then detached from the hydrogel surface using enzymatic digestion and collected. The surfaces were subsequently washed several times with culture medium to collect the rest of the cells, and the collected cells were pooled and centrifuged to a pellet. Live/dead cell counts were performed using trypan blue staining followed by a hemocytometer count.

For cell cultures on patterned HA surfaces, cells were plated onto the patterned surfaces in a similar fashion as described above except that the non-adherent cells were not washed away until pre-determined time points when the cultures were examined using an inverted phase contrast microscope (Olympus 1 × 81 F), and the observations were documented using Image-Pro Plus (Media Cybernetics, Silver Spring, MD).

2.9 Cell staining

CellTracker™ Red CMTPX (Invitrogen) was used to label the fibroblasts in this study according to the vendor’s protocol. A 10 mM stock solution of the dye in DMSO was prepared. This solution was subsequently dissolved in DMEM to produce a 2.5 µM working solution. One milliliter of the working solution was added to wells containing adherent cells in 12-well plates and allowed to incubate for 45 min at 37°C. The staining solution was then replaced with pre-warmed fresh media and allowed to incubate for another 30 min at 37°C. Cells were then washed in PBS (pH 7.4) followed by incubation in fresh media. The stained cells were used in cell adhesion and cell patterning studies, as outlined in Section 2.8, and visualized with fluorescence microscopy.

3 Results and discussion

3.1 Hydrogel development

HA hydrogels were developed to form the base of the cell pattern. HA was an ideal choice because (a) it is a component of the ECM in the natural cell microenvironment known to regulate cell proliferation and adhesion (Alberts et al. 2002; Collins and Birkinshaw 2007; Stern et al. 2006; Yamane et al. 2005), (b) in its native state, HA is naturally cell non-adhesive so it could prevent the background binding of cells located off-pattern (Morra and Cassineli 1999), and (c) it contains easily functionalized carboxylic acid groups for crosslinking to form stable hydrogels as well as for the covalent addition of biomolecules (Hu et al. 2000; Lee and Spicer 2000; Park et al. 2003). To form a stable HA gel, HA molecules were crosslinked in such a way that the material remained non-adherent to cells.

First, HA gel was prepared. Gelation was determined by inverting the tube in which the gels formed to verify that the material would maintain its shape. Gels were formed by joining HA chains via ADH crosslinkers which formed covalent bonds to carboxylic acid groups on the HA polymer chains, as shown in Fig. 4(a). Similar crosslinking approaches using ADH in HA have been described elsewhere (Cui et al. 2006; Hou et al. 2006; Vercruysse et al. 1997). To further confirm the incorporation of ADH into the HA polymer chains (i.e. crosslinking reaction), solid-state 13C NMR analysis of the hydrogel after crosslinking was carried out. The spectrum of the crosslinked sample as well as that of the native HA before crosslinking (data not shown) matched results previously reported by Pouyani et al. (Pouyani et al. 1994), suggesting the success of the crosslinking reaction.
Fig. 4

(a) Structure of ADH-crosslinked HA polymers with the functional groups involved in coupling the two molecules highlighted in red (b) Single layer of HA gel 1 day post seeding of 3T3 fibroblasts onto the surface (image obtained by phase contrast) (c) Layered HA gel 1 day post seeding of 3T3 fibroblasts onto the surface (image obtained by phase contrast). Note the cracks in the gel formed during the rehydration process in the preparation of the layered gel (d) Fluorescence image of HA hydrogel bound to FITC-BSA facilitated by EDC and (e) control HA hydrogel exposed to FITC-BSA in the absence of EDC after 22 h of washing (f) Phase contrast image of 3T3 fibroblasts adhering to a layered HA gel bound with RGDS 1 day post seeding (g) Phase contrast image 3T3 fibroblasts adhering to a polystyrene tissue culture plate 1 day post seeding Scale bars = 100 μm and images were obtained using an Olympus 1 × 81 F microscope

To evaluate these HA gels for cell-adhesiveness, 3T3 fibroblasts were first seeded onto single-layered gel surfaces. One day after the initial cell seeding, the gels were gently washed and subsequently observed. As seen in Fig. 4(b), a significant number of fibroblast cells remained adherent to the gel; however, it was also evident that the cells could not spread well onto the hydrogel. Fibroblasts were round in shape and had grouped together to form distinct clusters. Filopodia could be seen extending from the clusters anchoring the cells to the underlying HA gel, preventing the cells from removal upon washing.

It is known that increasing the fractional surface coverage of HA can reduce cell adhesion (Morra and Cassineli 1999). Therefore, the concentration of HA in the hydrogel was increased in order for the pattern base to better repel cells. To achieve this, the original hydrogel layer was dried and new hydrogel layers were deposited onto its surface to create a dried, layered gel which was subsequently rehydrated. Such gels were visibly less swollen than the original single-layered gels, and as a result, conceivably contain a higher concentration of HA molecules. As seen in Fig. 4(c), when seeded with 3T3 fibroblasts, these layered gels exhibited significantly improved cell repulsion capabilities since virtually no cells were found to adhere 1 day post seeding. This layered gel configuration was used for the rest of the studies in this report.

3.2 Binding of peptide to the gel

In developing the cell patterns on the HA hydrogel, both native RGDS and caged R[G]DS peptides were covalently bound to the gel via carboxyl groups on HA. Merely adsorbing RGD peptides onto surfaces leads to poor cell attachment (Hersel et al. 2003), and such peptides are more easily dislodged, negatively impacting pattern longevity. We confirmed that peptides could be covalently bound to the gel surface using fluorescent (FITC-labeled) bovine serum albumin (BSA) as a model molecule (for ease of detection). The gel shown in Fig. 4(d) was treated with both FITC-BSA and a peptide coupling agent, EDC, and displayed strong fluorescence over the entire gel surface. In comparison, the control gel in Fig. 4(e) exposed only to FITC-BSA exhibited significantly less fluorescence intensity. An additional control in which FITC-BSA was omitted in the reaction exhibited no autofluorescence (result not shown). Since the fluorescence images were obtained using identical exposure times, the sharp contrast in fluorescence demonstrated that the gels exposed to EDC were able to covalently immobilize a significant amount of peptides while the controls were not, resulting in the adsorbed peptides mostly being washed away in the extensive washing steps.

Similarly, RGDS peptides were bound to the HA gel. Figure 4(f) shows the resulting morphology of fibroblasts grown on the HA hydrogel bound with RGDS 24 h after seeding. It is evident that the immobilization of RGDS drastically increased the adhesiveness of the HA hydrogel (cf. Figs 4(c) and (f)). Cells assumed a well-spread and flattened morphology with numerous extensions, exhibiting a similar morphology to that of cells grown on the control polystyrene tissue culture plate, as shown in Fig. 4(g).

3.3 Peptide characterization

In developing the cell patterned surfaces, it was important to synthesize the caged peptide, R[G]DS, with a high purity since any RGDS contaminants could potentially bind cells thus allowing for off-pattern binding. However, our first attempts to synthesize R[G]DS using a solid phase peptide synthesis approach resulted in a relatively low yield, with the occurrence of significant side reactions. The reasons for these side reactions are still under investigation. Alternatively, liquid phase synthesis was attempted with success. We were able to prepare high purity R[G]DS in sufficient quantities. The synthetic peptides, both RGDS and R[G]DS, were fully characterized by MALDI-tof mass spectrometry as well as 1H NMR spectroscopy (see Section 2).

3.3.1 Molecular modeling

A 3D energy minimized molecular modeling structure for each peptide was developed using Hyperchem software (Hypercube Inc., Gainesville, FL). These theoretically derived structures (Fig. 5) suggested that the introduction of a nitrobenzyl group at the backbone amide nitrogen atom of the Gly residue led to a significant change in the geometry of the molecule. Thus, the caged RGDS peptide appeared to possess a more compact structure compared to regular RGDS tetrapeptide which exhibited a β-turn structure at the R-G amide bond. These significantly different structural features provide, at least in part, the rationale for this patterning approach and perhaps explain the cell patterning results as shown below.
Fig. 5

Hyperchem generated theoretical energy minimized molecular model structures of RGDS and caged R[G]DS peptides

3.3.2 Photolysis reaction

To investigate the kinetics of the R[G]DS photolysis reaction, the rate of disappearance of the caged peptide upon UV exposure was evaluated using RP-HPLC. Figure 6(a) is a plot of absorbance versus wavelength for various irradiation times corresponding to the R[G]DS peptide elution peak on the RP-HPLC chromatogram. The R[G]DS peptide exhibited an absorption maxima at 266 nm, but absorbed light with a wavelength up to approximately 370 nm. This observation agrees well with that reported by Zhang et al. who found that 2-nitrobenzyl caged 5-fluorouracel prodrugs had maximum absorbances of around 265 nm in acetonitrile/water systems (Zhang et al. 2005). As expected, the absorbance at this elution time decreased with increasing UV irradiation time due to the removal of the strongly absorbing 2-nitrobenzyl group from the R[G]DS peptide. Despite the maximum absorption found at 266 nm, uncaging reactions were carried out with a 365 nm light source to ensure the safety of our biomolecule-based system. It has been established in the literature that, in general, exposure to irradiation above 350 nm is relatively safe for work in biological systems (Furuta and Noguchi 2004; Sigrist et al. 1995). Figure 6(b) plots the percentage of R[G]DS reacted as a function of irradiation time, as determined from measuring the area under the R[G]DS peak in the HPLC chromatograms. It can be seen that after 10–12 min of irradiation, 60–70% of the R[G]DS had reacted, and over 90% of the caged peptide disappeared after 30 min. This is encouraging since it indicates that the majority of the R[G]DS will react upon UV irradiation at 365 nm in 30 min. 1H NMR data supported the regeneration of the RGDS peptide (see Section 2). A plot of ln[C]/[C]o versus time (where [C] represents the concentration of R[G]DS peptide and [C]o the initial concentration) showed a linear relationship, which indicated that first order reaction kinetics could be used to describe the 2-nitrobenzyl photolysis as seen previously (Kim and Diamond 2006). Furthermore, from Fig. 6(c) the apparent first order uncaging reaction rate constant was calculated to be 1.6 × 10−3 s−1, similar to values reported by Kim and Diamond for 2-nitrobenzyl ester compounds, which ranged from 1–9 X10−4 s−1. Differences in the values may be attributed to differences in the solvent systems and light intensity used in the experiments (Kim and Diamond 2006).
Fig. 6

(a) The absorbance of the R[G]DS elution peak versus wavelength following UV irradiation for the indicated times from an RP-HPLC chromatogram (b) The percentage of R[G]DS reacted with UV irradiation time (c) Plot of ln[C]/[C]o versus irradiation time, t. Error bars represent one standard deviation of the mean

3.4 Cell binding studies

In order to examine all of the various surfaces that will ultimately make up the cell patterning platform, several homogenous surfaces were created for cell seeding to test their relative surface adhesiveness. These surfaces include a pure, layered crosslinked HA gel, HA gel bound to RGDS, HA gel bound to R[G]DS, HA gel bound to R[G]DS exposed to near-UV light for 12 min, and a control polystyrene tissue culture surface. All of the gels were created with equimolar quantities of the different peptides in bulk solution (0.3 mg of R[G]DS peptide) and seeded with stained 3T3 fibroblasts. From Figs. 7, it can be seen that virtually no cells adhered to the surface of the pure HA gel or to the surface of the HA gel bound with the caged peptide, R[G]DS. This cell non-adhesive background is crucial for cell patterning to take place—the surface of the gel prior to exposure to near-UV light must be completely cell repellant to prevent against off-pattern binding. On the other hand, the HA gel bound to RGDS and the HA gel bound to R[G]DS uncaged with near-UV light both supported cell adhesion. Moreover, both surfaces supported cells with similar morphologies. Cells appeared well-spread and morphologically comparable to those grown on the control tissue culture plate. This suggests that the uncaging of R[G]DS took place upon near-UV exposure and regenerated RGDS peptides on the HA gel surface, converting a cell non-adhesive surface to a cell adhesive one to allow for cell attachment. The cell counts based on Fig. 7(f) showed similar numbers on the HA-RGDS surface as on the tissue culture plate. This suggests a high degree of biocompatibility for the developed HA material. The biocompatibility of the material is also supported by the fact that very few dead cells were counted on any of the surfaces. It is interesting to note that fewer cells were adherent to the HA-R[G]DS surface exposed to near-UV light, suggesting that not all of the R[G]DS peptides became uncaged during near-UV light exposure. Virtually no cells were counted on the caged R[G]DS-HA surfaces and on HA gel surfaces (see Fig 7(f)) which suggests that positioning the caging group on the peptide backbone between the Arg and Gly residues of the RGDS sequence does indeed inhibit cell binding to the sequence, most likely through disrupting integrin recognition of the RGDS sequence. This is consistent with our prediction. In fact, X-ray structure analysis performed to study the interactions of an integrin receptor (i.e. αVβ3) and a bound RGD ligand (i.e. c(-RGDf[NMe]V-)) showed that the Gly residue was directly positioned on the integrin surface, suggesting its direct involvement in hydrophobic interactions with the integrin (Gottschalk and Kessler 2002; Marinelli et al. 2003). In addition, it has been further suggested that this hydrophobic interaction may add to the stability of the RGD-integrin binding complex (Gottschalk and Kessler 2002). We can thus speculate that the introduction of a bulky group such as 2-nitrobenzyl on the peptide backbone by Gly would, to some degree, disrupt the ability of the Gly residue to interact with the integrin surface. It has also been suggested that close contact between the polar amide groups adjacent to Gly and the integrin receptor is essential for integrin mediated cell attachment (Dechantsreiter et al. 1999). Since our caging group is covalently linked to one of such amide groups, it would certainly serve to disrupt any such receptor binding at this location. Another explanation for the inhibition of cell binding to our R[G]DS peptide could be that the introduction of the caging group drastically changes the conformation of the RGDS peptide. This was noted in our modeling studies. In order to further demonstrate that cell binding is RGD-dependant, cells were seeded on HA-R[G]DS surfaces after UV exposure in the presence of free RGDS peptides. Preliminary results (not shown) demonstrate that 1 mg/mL of free RGDS peptides present during cell seeding can almost completely inhibit 3T3 binding to the surface. This would seem to indicate that binding is integrin dependant—as the free RGDS peptides would serve to bind and block cell integrin receptors preventing their participation in surface binding—and not mediated by other mechanisms involving non-specific protein adsorption or unpredicted UV modifications to the surface after exposure. These findings ultimately suggest that the caged R[G]DS-HA surface could indeed be converted from cell non-adhesive to adhesive upon exposure to near-UV light, which supported the notion that this system could be used for cell patterning.
Fig. 7

3T3 Fibroblasts 24 h post seeding on (a) pure layered ADH-crosslinked HA hydrogel, (b) HA gel bound to RGDS peptide (HA-RGDS) (c) HA gel bound to a caged R[G]DS peptide (HA-R[G]DS) (d) HA gel bound to a caged R[G]DS peptide exposed to UV light (HA-R[G]DS-UV) and (e) a control in which cells were grown on a polystyrene 12-well plate (no gel) (f) Number of live and dead cells harvested after 24 h of culture on the indicated surfaces with error bars representing one standard deviation from the mean with a sample size of three gels and Scale bars = 100 μm

3.5 Cell patterning

To produce cell patterns on the HA hydrogel surfaces, photomasks with line patterns were placed overtop R[G]DS-HA gels, and near-UV light was shone onto the surfaces to generate patterns of adhesive RGDS. Gels were then sterilized with ethanol and washed overnight. Subsequently, 3T3 fibroblasts were seeded onto the patterned gels. Figure 8(a)–(c) show fluorescent micrographs of the fibroblasts stained with CellTracker™ Red 1 day post seeding on the pattern as well as the mask used. Cells can be seen adherent to the surface in distinct lines matching the pattern on the photomask. Figure 8(d)–(g) is a compilation of phase contrast images showing fibroblasts adhering to line patterns from 12 h to 2.5 days post seeding. Over the 2.5 days post seeding, it was noted that cells appeared to spread and grow, increasing in density along the patterned regions. Significantly, the cell pattern was well maintained as few cells could be seen growing off-pattern even after 2.5 days in culture. It was quite interesting that even during proliferation, cells remained aligned to the underlying pattern.
Fig. 8

(a, b) Fluorescent micrographs depicting line patterns of 3T3 fibroblasts stained with CellTracker™ Red and grown on a HA hydrogel bound with patterned caged and uncaged R[G]DS peptides 1 day post seeding (c) Mask used to generate the line pattern. Phase contrast images of line patterns of 3T3 fibroblasts grown on a HA hydrogel bound with patterned caged and uncaged R[G]DS peptides at (d) 12 h, (e) 1 day and (f) 2.5 days post seeding (g) Mask used to generate the line patterns Scale bars = 100 μm

One potential concern with this current patterning strategy based on light exposure is that ambient light and light exposure from microscopy techniques may serve to uncage the R[G]DS peptide and degrade the pattern over time. However, this did not seem to be an issue over the 2.5 days the pattern was monitored in our study as the cells did not appear to move significantly off-pattern despite routine examination of the surface using phase contrast and fluorescence microscopy.

In comparison with previous studies using caged RGD peptides for cell patterning, we were able to achieve increased pattern longevity; this study demonstrated pattern stability for at least 2.5 days in culture. In contrast, a study by Petersen et al., also employing 3T3 fibroblasts as in the current study, reported significant off-pattern deviations just 6 h post seeding (Petersen et al. 2008). While there may be many other reasons for the significantly improved cell pattern longevity achieved in this study, we attribute our success to two chief reasons: 1) the use of a strongly non-adhesive HA base to provide a non-adhesive background, and 2) the use of a zero-length crosslinker to bind the R[G]DS to the gel surface. As mentioned, previous works with caged RGD relied on flexible OEG or PEG chains to link the peptides to a solid surface and provide a non-adhesive background (Ohmuro-Matsuyama and Tatsu 2008; Petersen et al. 2008). Flexible chains acting as spacers are often used to link RGD peptides to surfaces to enhance access of cell integrin receptors to the peptide sequence and thus enhance cell adhesion (Beer et al. 1992; Nishi et al. 2007). It was hypothesized that the use of a zero-length crosslinker would limit the flexibility of the caged RGDS peptide to undergo such conformational changes which could otherwise potentially permit partial binding to integrins on cell surfaces, and therefore the zero-length crosslinker may assist in maintaining the cell non-adhesive background essential to cell patterning.

Our design is not only functional, but also quite versatile. The potential exists to pattern a wide variety of cell types since the RGD peptide sequence employed is one of the most widely recognized sequences to support cell-surface adhesion. There is also the potential to form a wide variety of patterns through the use of different photomasks. Furthermore, it is felt that our strategy could eventually allow for patterning at the single cell level since resolution is only theoretically limited by the density of peptide coverage on the underlying surface, which can be readily manipulated. Numerous potential applications exist for our unique patterning approach. The spatial control that it affords over cell positioning on a biocompatible surface makes it an excellent platform for more controlled studies in cell biology. Since our patterned design is based on a natural hydrogel material, we envision great potential for future applications in tissue engineering. HA has been incorporated into numerous tissue engineering scaffolds (Allison and Grande-Allen 2006), and HA polymers modified with RGD peptides have also found several tissue engineering applications (Cui et al. 2006; Glass et al. 1996; Shu et al. 2004). Our design is especially interesting because it has been shown to not only support the adhesion of cells onto a pattern, but new cell growth was also directed along the patterned lines. This characteristic behavior could be harnessed in the future to direct the growth of cells along desired routes in vivo. Our current design therefore holds much potential and future refinement in resolution promises to bring new and exciting applications.

4 Conclusions

We have developed a novel cell patterning platform in patterning a hydrogel surface with a photocaged RGDS tetrapeptide. Our surface and pattern base consist of cell non-adhesive hyaluronic acid, a material naturally found in the extracellular matrix. An RGDS peptide attached to a 2-nitrobenzyl cage via the peptide backbone amide nitrogen atom was synthesized, fully characterized and bound to the hydrogel base. We were able to demonstrate that this peptide was efficiently bound to the gel via covalent binding. Moreover, the generated surface was non-adhesive to cells, but could be switched to become cell adhesive upon exposure to near-UV light. In this way, a pattern of cell adhesive regions on an otherwise cell repellant hydrogel surface was created by shining near-UV light through a patterned photomask. We have demonstrated the creation of line patterns of cells using this novel method. Cells were also shown to proliferate along the created line patterns, adhering to the patterns for at least 2.5 days, demonstrating increased longevity compared to previous reports.


The authors would like to acknowledge financial support from the Natural Sciences and Engineering Research Council of Canada (NSERC) (XC) and the Canadian Institutes of Health Research (CIHR) (AB). Catherine A. Goubko is supported by an NSERC Canada Graduate Scholarship and Dr. Swapan Majumdar is supported by a CIHR-HOPE fellowship. SM is thankful to Tripura University, India for leave of absence.

Supplementary material

10544_2010_9412_MOESM1_ESM.doc (10.4 mb)
Supplemental Fig. 1S1H NMR spectrum of caged-R[G]DS tetrapeptide in D2O with expanded portions within the insets (DOC 10628 kb)
10544_2010_9412_MOESM2_ESM.doc (10.4 mb)
Supplemental Fig. 2S1H NMR spectrum of caged R[G]DS tetrapeptide in D2O following UV-irradiation with expanded portions within the insets (DOC 10663 kb)

Copyright information

© Springer Science+Business Media, LLC 2010