Biomacromolecules 2010, 11, 1387–1397
1387
Functional Assessment of Cross-Linked Porous Gelatin Hydrogels for Bioengineered Cell Sheet Carriers Jui-Yang Lai*,†,‡,§ and Ya-Ting Li† Institute of Biochemical and Biomedical Engineering, Biomedical Engineering Research Center, and Molecular Medicine Research Center, Chang Gung University, Taoyuan, Taiwan 33302, Republic of China Received February 25, 2010; Revised Manuscript Received March 19, 2010
An efficient carrier for corneal endothelial cell therapy should deliver and retain the cell sheet transplants at the site of injury without causing adverse effects. Here we introduced a simple stirring process combined with freezedrying (SFD1) method for the development of gelatin hydrogels with enlarged pore structure that can improve the aqueous humor circulation. Samples fabricated by air-drying (AD) or freeze-drying method were used for comparison. After cross-linking with 1 mM 1-ethyl-3-(3-dimethyl aminopropyl) carbodiimide (EDC), the discs were investigated to assess their functionality. The simultaneous presence of ice crystals and gas bubbles resulted in large pore size (461 ( 85 µm) and high porosity (48.0 ( 1.9%) of SFD1 carriers. Among all of the samples studied, the SFD1 hydrogels showed the most appropriate swelling characteristics without squeezing effect on the anterior segment tissues of the eye. The enlarged pore structure also allowed carriers to contain the highest fraction of mobile water and exhibit the lowest resistance to the glucose permeation. In comparison with AD samples, the SFD1 materials had better cytocompatibility and biocompatibility and more effectively prevented a drastic change of intraocular pressure. Rheological measurements showed that the SFD1 hydrogels behaved like an elastic solid and had a tough (rigid and deformable) texture. As a temporary supporter, the biodegradable gelatin hydrogel could facilitate cell sheet transfer and avoid long-term residence of foreign carriers in the body. Our findings suggest that the gelatin discs with enlarged pore structure have potential as cell sheet carriers for intraocular delivery and corneal tissue engineering.
Introduction Worldwide, more than 10 million people are suffering from blindness because of the various corneal conditions such as pseudophakic or aphakic bullous keratopathy, Fuch’s endothelial dystrophy, keratitis, and repeat graft.1 Corneal transplantation remains the most commonly used treatment, and it has a higher success rate than transplantation of other organs. Approximately 38 000 transplants are performed annually in the United States, and nearly half of all of these grafts are done for visual loss because of corneal endothelial dysfunction.2 Therefore, corneal endothelial cell (CEC) therapy is considered to be a promising concept to overcome the limitations of corneal replacement with an allograft. Direct injection of isolated CEC suspensions into the anterior chamber of rabbit eyes has been used for repopulation of the damaged corneal endothelium, but the transplantation method has a major drawback such as difficult control of the distribution and size of cell grafts.3 Additionally, cells in suspension undergo apoptosis because of their lower ocular immune-privileged properties relative to cells in well-organized sheets. Along with the rapid advance in tissue engineering, several biomaterials such as gelatin, collagen-coated dextran, Descemet’s membrane, poly L-lysine/fibronectin-coated methyl methacrylate/N-vinyl pyrrolidone copolymer, amniotic membrane, and collagen have been investigated for CEC cultivation and transplantation.4 However, the permanent residence of culture * To whom correspondence should be addressed. Tel: +886-3-211-8800, ext. 3598. Fax: +886-3-211-8668. E-mail:
[email protected]. † Institute of Biochemical and Biomedical Engineering, Chang Gung University. ‡ Biomedical Engineering Research Center, Chang Gung University. § Molecular Medicine Research Center, Chang Gung University.
substrate materials in the body may cause the poor graft-host integration and high risk of foreign body reaction. To overcome the problems associated with traditional tissue engineering approaches, we, as others, have presented a technique to fabricate bioengineered human CEC sheets on thermoresponsive poly(N-isopropylacrylamide)(PNIPAAm)-basedculturesupports.5,6 Without the use of biomaterial substrates, the thermally detached cell sheets with proper structure and function can feasibly be used as tissue equivalents for replacing compromised corneal endothelium.7,8 It is known that the anterior chamber is filled with a watery fluid (i.e., aqueous humor), which leads to unstable attachment of donor cell/tissue grafts onto corneal posterior surfaces. To deliver and retain the transplants at the site of injury is essential for achieving high tissue regeneration efficiency. In the author’s previous work, the bioadhesive gelatin discs were developed as temporary carriers for intraocular delivery of CEC sheets.9 Gelatin is a protein-based biomaterial and has received attention for medical and pharmaceutical applications.10-13 In ophthalmic tissue engineering, the gelatin carriers have been studied for delivery of various corneal cells, including basal epithelial cells,14 fibroblast precursor cells,15 and endothelial cells.9 The advantage of using gelatin materials as cell sheet delivery systems for intraocular grafting is not only that it is useful for maintaining tissue-like architecture and cellular organization of sheet transplants but also that the deformable hydrogel carriers can facilitate establishing a minimally invasive surgery.8,16 Although the application of gelatin-CEC sheet constructs partially restores corneal clarity in a xenogeneic model,8 the gelatin discs with dense structure in the anterior chamber may interrupt the transport of nutrients and upset the balance of intraocular pressure (IOP) (Figure 1a). This study
10.1021/bm100213f 2010 American Chemical Society Published on Web 03/31/2010
1388
Biomacromolecules, Vol. 11, No. 5, 2010
Lai and Li
Figure 1. Schematic representation of intraocular delivery of cell sheet grafts using gelatin carriers with (a) dense structure and (b) porous structure.
aims to address the concerns of existing gelatin-based delivery systems by developing highly efficient cell sheet carriers. We hypothesize that the introduction of porous structure to the gelatin discs will dramatically improve aqueous humor circulation, thereby minimizing adverse effects of carrier materials on the reconstructed corneal tissues (Figure 1b). There are currently a variety of techniques for fabricating materials with porous structures such as fiber bonding,17 gas foaming,18 phase separation,19 and electrospinning.20 Nevertheless, these methods involving additives are limited because of the possible residual toxicity of organic solvents and foaming agents. Gelatin is well-known to be an aqueous polymer. A combined procedure, using water as porogen solvent, followed by freeze-drying, has been proposed to produce porous gelatin hydrogels.21 As the frozen ice crystal is sublimated in a freezedrying process, numerous cavities are created within the bulk of the biopolymer matrix. In terms of tissue engineering applications, the pore size and distribution of biomaterials are of particular importance for regulating nutrient uptake to and waste removal from the cultured cells. The porous structure of freeze-dried gelatin hydrogels can be controlled by the size of ice crystals formed during freezing at different temperatures.21,22 To fabricate highly porous gelatin matrices, an overrun process method based on the injection of air bubbles and ice recrystallization has been developed.23 Although it is beneficial to mass transfer, the increase in pore size and porosity may result in decreased mechanical properties and accelerated degradation rates of biomaterials. Cross-linking is a necessary step to prepare stable porous gelatin hydrogels under physiological conditions while preserving their multifunctionality for cell sheet delivery use. Gelatin has been chemically modified with various crosslinkers, including formaldehyde, glutaraldehyde, epoxy compounds, carbodiimides, and genipin.24,25 Among them, 1-ethyl3-(3-dimethyl aminopropyl) carbodiimide (EDC) is particularly preferable because it can induce cross-linking of biomaterials without taking part in the linkages but simply change to watersoluble urea derivatives that have very low cytotoxicity.26 In this article, the gelatin hydrogels with different pore sizes and porosities were respectively fabricated by adjusting the freezing temperature to -20 and -196 °C. Furthermore, we designed a simple stirring process combined with a freeze-drying method to endow the gelatin carriers with enlarged pore structures that impart better delivery capabilities. Prior to freezing at -20 °C, the aqueous gelatin solution was vigorously stirred at 25 °C (i.e., below the gelling temperature) to entrap gas bubbles, thereby producing enlarged pores in the hydrogel matrix. The gelatin samples prepared by air-drying method served as controls. After cross-linking with 1 mM EDC, various
gelatin carriers were investigated by determinations of swelling ratio, mechanical strength, glucose permeability, biodegradability, cytocompatibility, biocompatibility, and cell-adhesive ability. To the best of our knowledge, this is the first report to demonstrate the potential advantage of using cross-linked porous gelatin hydrogels as cell sheet delivery systems. The information from functional assessment is necessary to evaluate whether the proposed porous structure will further improve the safety and efficacy of gelatin carriers applied in CEC transplantation therapy.
Experimental Section Materials. Gelatin (type A; 300 Bloom), EDC, and matrix metalloproteinase-2 (MMP-2, EC 3.4.24.24) were purchased from SigmaAldrich. Balanced salt solution (BSS, pH 7.4) was obtained from Alcon. Dulbecco’s modified Eagle’s medium (DMEM) was purchased from Gibco-BRL. Fetal bovine serum (FBS) and the antibiotic/antimycotic (A/A) solution were obtained from Biological Industries. Temperatureresponsive culture dishes (35 mm in diameter, UpCell Surface) were purchased from CellSeed. Preparation of Cross-Linked Porous Gelatin Carriers. An aqueous solution of 10 wt % gelatin was prepared by dissolution of gelatin powder in double-distilled water at 40 °C. The 5 mL of resulting solution was then poured in a polystyrene planar mold (5 × 5 cm2, 1.5 cm depth), subjected to freezing at -20 (FD1 group) or -196 °C (FD2 group) for 24 h, and lyophilized at -55 °C for 2 days. On the other hand, the 5 mL of gelatin solution was cooled to 25 °C and stirred with a rate of 350 rpm for 20 min. After casting into the mold, the solution was subjected to freezing at -20 °C (SFD1 group) for 24 h and lyophilized at -55 °C for 2 days. As a control, the gelatin solution (35 mL) prepared at 40 °C was poured in the mold and air-dried for three days at 25 °C (AD group). All fabricated gelatin hydrogel sheets were cross-linked by immersing in an ethanol/water mixture (8:2 v/v, pH 4.75) of 1 mM EDC at 25 °C for 12 h. The samples were thoroughly washed with double-distilled water to remove excess EDC and urea byproduct. Subsequently, the cross-linked gelatin hydrogels were subjected to freezing at -20 °C for 24 h and lyophilized again. Using a 7 mm diameter corneal trephine device, the hydrogel sheets were cut out to obtain gelatin discs (∼700 µm in thickness). Characterization of Cross-Linked Porous Gelatin Carriers. Specimens were prepared for scanning electron microscopy (SEM) as previously described.27,28 The surface and cross-section morphologies of gelatin discs were examined using SEM (Hitachi). Twenty different pores were randomly selected, and the average pore diameters were calculated. The solvent replacement method was used for porosity measurements.29 Each gelatin disk was first dried to constant weight (Wi) in vacuo. The test samples were immersed in absolute ethanol overnight,
Porous Gelatin Hydrogel for Cell Sheet Engineering blotted with tissue paper to remove excess ethanol on the surface, and weighed (Wf) immediately. The porosity (%) was calculated to be ((Wf - Wi)/VF) × 100, where V is the volume of the hydrogel disk and F is the density of absolute ethanol. Results were averaged on five independent runs. The ninhydrin assay was used to determine the amount of free amino groups of each gelatin disk before (Ci) and after (Cf) cross-linking.30 The extent of cross-linking of gelatin discs was calculated to be crosslinking index (%) ) ((Ci - Cf)/Ci) × 100. Results were the average of six independent measurements. Swelling Tests. The swelling ratio of each cross-linked gelatin sample was determined by immersing the disk in BSS at 34 °C (physiological temperature of the cornea). The swelling ratio was defined as ((Ws - Wi)/Wi), where Ws is the weight of swollen gelatin hydrogel and Wi is its initial dry weight. Additionally, the gross appearances of swollen discs were photographed by digital camera (Nikon). Results were averaged on four independent runs. Rheological Measurements. Dynamic shear oscillation measurements at small strain were used to characterize the rheological properties of cross-linked gelatin hydrogel discs. Each sample was immersed in BSS until fully swollen and loaded into the rheometer. The measurements at oscillatory shear deformation were carried out using a rheometer (TA Instruments) equipped with a Peltier temperature control and a cone-plate geometry (diameter ) 40 mm, angle ) 2°). Mechanical spectra were recorded in a constant strain mode, with a low deformation of 0.05 maintained over the frequency range of 0.01-10 Hz (rad/s) at 34 °C. All experiments were performed in triplicate. Determination of Freezable Water Content. Differential scanning calorimetry (DSC) measurements were used to examine the states of water in the cross-linked gelatin discs. The samples were placed in a DSC cell (TA Instruments), cooled to -20 °C to freeze swollen hydrogels, and heated to 20 °C at a heating rate of 5 °C/min under a nitrogen gas flow. The amount of freezable water was evaluated from DSC endothermic icemelting profile of frozen hydrogel. The enthalpy of melting (∆Hm) obtained by integration and normalization is in units of joules per gram of swollen hydrogel. Temperatures and enthalpies of melting of the samples were calibrated using pure water as the standard. The latent heat of water is 333.5 J/g of pure water. The gram of freezable water per gram of swollen gelatin hydrogel (WfH/Ws) was calculated to be ∆Hm/333.5. Results were averaged on six independent runs. Glucose Permeation Studies. Glucose permeation studies were performed at 34 °C using a horizontal glass diffusion cell (PermeGear) having two stirred chambers with sampling ports. The donor chamber was filled with a 6.9 µmol/mL (the glucose concentration of aqueous humor in rabbit31) glucose solution in BSS (3 mL) and receptor chamber with BSS (3 mL). After immersion in BSS until fully swollen, the crosslinked gelatin samples were placed between two chambers. During the measurements, all solutions were stirred continuously to provide uniform solute distribution and to reduce boundary layering of glucose. At specific time intervals, the receptor chamber was sampled and analyzed using a glucose assay kit following the manufacturer’s instructions. Photometric readings at 540 nm were measured with a spectrophotometer (Thermo Scientific). Results were averaged on five independent runs. In Vitro Degradation Tests. To measure the extent of degradation, each cross-linked gelatin disk was first dried to constant weight (Wi) in vacuo. The test samples were immersed at 34 °C in BSS containing 50 ng/mL MMP-2. The dry weight of gelatin discs after degradation (Wd) was determined, and the percentage of weight remaining (%) was calculated to be (Wd/Wi) × 100. Results were the average of five independent measurements. Gel permeation chromatography (GPC) was used to characterize the degradation products. (See the Supporting Information.) Mean peak molecular weight (Mw) in daltons and corresponding polydispersity index (PDI) were calculated. Biocompatibility Studies. All animal procedures were approved by the Institutional Review Board and were performed in accordance with the ARVO Statement for the Use of Animals in Ophthalmic and Vision
Biomacromolecules, Vol. 11, No. 5, 2010
1389
Research. Twelve adult New Zealand white rabbits (National Laboratory Animal Breeding and Research Center, Taipei, Taiwan) weighing from 3.0 to 3.5 kg were used for biocompatibility testing. Surgical operation was performed in the single eye of animals with the normal fellow eye. In the two test groups (AD and SFD1) of animals (six rabbits/ group), the gelatin implants were inserted in the anterior chamber of the eye. Surgery was conducted according to the method previously described.32,33 Under the surgical microscope (Carl Zeiss), the cornea was penetrated near the limbus by using a slit knife. Then, the corneal/ limbal incision was enlarged to 7.5 mm with corneal scissors to allow gelatin implantation. To determine the implant-tissue interaction in the anterior chamber, ophthalmic evaluations were performed before and immediately after surgical insertion of test samples. Subsequently, we examined the bilateral eyes of 12 rabbits at postoperative 3, 6, and 12 h, then daily for 2 weeks, and thereafter twice weekly for 8 weeks. The corneal morphology was analyzed by gross photography (Nikon). The CEC density in rabbit eyes was measured by specular microscopy (Topcon). The IOP was measured using a Schiotz tonometer (AMANN). For each IOP determination, five readings were taken on each eye, and the mean was calculated. The IOP values of contralateral normal eyes were used as baseline readings. Data were expressed as the difference from baseline values at each time point. Animals were euthanized with CO2 gas at the end of experiments. To evaluate the biodegradability of gelatin implants, the aqueous humor from each rabbit eye was immediately aspirated using a 30-gauge needle without touching the iris, lens, and corneal endothelium. Aqueous humor specimens were analyzed for hydroxyproline.34 In brief, the hydrolyzed samples were then mixed with a buffered chloramine-T reagent, and the oxidation was allowed to proceed for 25 min at room temperature. After the addition of Ehrlich’s aldehyde reagent to each sample, the absorbance was read at 550 nm by using a spectrophotometer (ThermoLabsystems) and compared with a standard calibration curve to quantify the amount of hydroxyproline. All experiments were conducted in six replicates. The excised rabbit cornea and iris were processed for histological examinations. The samples were fixed in 4% paraformaldehyde in PBS, dehydrated in a graded series of ethanol solutions, embedded in paraffin, and cut into 5 µm sections. Thin sections were stained with hematoxylin and eosin (H&E) and examined under light microscope (Carl Zeiss). Cell Sheet Transfer Studies. L929 murine fibroblasts (BCRC no. 60091) purchased from the Bioresource Collection and Research Center (Hsinchu, Taiwan) were used for preparation of bioengineered cell sheets. The cells were plated on temperature-responsive culture dishes at a density of 4 × 104 cells/cm2 and were cultivated in DMEM supplemented with 10% FBS and 1% A/A solution. Cultures were incubated in a humidified atmosphere of 5% CO2 at 37 °C. The medium was changed twice weekly until cells had reached confluence. To harvest cell sheets, we reduced the incubation temperature to 20 °C. After cell sheet detachment from PNIPAAm-grafted surfaces, the sterilized gelatin carriers were immediately placed on the apical surface of cell layers to create gelatin-cell sheet constructs. Then, the samples were mounted in OCT embedding medium, frozen, and cut into 5 µm sections at -20 °C. After fixation with 4% paraformaldehyde, the sections were stained with H&E. To visualize cell nuclei, sections were counterstained with Hoechst 33258 (Invitrogen) and observed under fluorescence microscope (Carl Zeiss). Statistics. Results were expressed as mean ( standard deviation. Comparative studies of means were performed using one-way analysis of variance (ANOVA). Significance was accepted with p < 0.05.
Results and Discussion Characterization of Cross-Linked Porous Gelatin Carriers. Figure 2 shows the cross-sectional and surface SEM images and porosity of various gelatin discs. The samples prepared by a simple stirring process combined with freeze-drying method
1390
Biomacromolecules, Vol. 11, No. 5, 2010
Lai and Li
Figure 2. Scanning electron microscopic images of various gelatin discs. (a) AD, (b) SFD1, (c) FD1, and (d) FD2 groups. CS: cross-section; S: surface. 1: before EDC cross-linking; 2: after EDC cross-linking. Scale bars: 100 µm. (e) Porosity of various gelatin discs. An asterisk indicates statistically significant differences (*p < 0.05; n ) 5) between the non-cross-linked (EDCb) and cross-linked (EDCa) groups for each type of gelatin disk.
possessed enlarged pore structure without dense surface skin formation. The simultaneous presence of ice crystals and gas bubbles resulted in large pore size (461 ( 85 µm) and high porosity (48.0 ( 1.9%) of cross-linked SFD1 carriers. The order of increasing cross-linking index for the discs was SFD1 > FD1 > FD2 > AD, indicating that the material structure may play an important role in the cross-linking reaction. Detailed results and discussion are provided in the Supporting Information. Swelling Tests. It is known that the swelling ratio is one of the most important parameters for evaluating hydrogels and is correlated to the biomaterial structure. The mechanism of hydration of gelatin is a capillary phenomenon of water molecules penetrating the tiny interstices of triple-helical fibrils in the gelatin matrix.9 In this study, the water absorption capability of various gelatin discs was determined by a gravimetric method (Figure 3a). In the AD groups, the swelling ratio reached a plateau level of ∼2.1 within 1 h of incubation in BSS at 34 °C. In contrast, significantly higher water uptake rates were observed in the SFD1, FD1, and FD2 groups. At each time point from 1 to 12 h, the order of increasing swelling ratio for the discs was FD2 > FD1 > SFD1, indicating that the
extent of cross-linking had a profound influence on the water absorption capability of gelatin samples. The osmotic driving force of the network chains toward infinite dilution will cause the hydrogel to imbibe additional water after the polar and hydrophobic sites of macromolecules have interacted with water molecules.36 The formation of covalent cross-links between adjacent polymer chains can lead to an elastic network retraction force against additional swelling. A higher number of crosslinks in the SFD1 samples would further support the theory that cross-links reduce the swelling of gelatin matrices. It was noted that although the AD carriers had the lowest cross-linking index, these samples showed the relatively low swelling ratio at 6 h. Risbud et al. have shown that bulk water is unable to penetrate the matrix because of the very low porosity of air-dried hydrogel membranes.35 Our findings suggest that the porous structure of gelatin carrier may be responsible for its high water absorption and high water retention properties. An ideal condition for intraocular grafting of CEC sheets is that the hydrogel carriers should swell rapidly to a size large enough to facilitate the attachment of bioengineered cell monolayers onto corneal posterior surfaces. The residence of
Porous Gelatin Hydrogel for Cell Sheet Engineering
Biomacromolecules, Vol. 11, No. 5, 2010
1391
Figure 3. (a) Time course of swelling ratio of various gelatin discs after incubation in BSS at 34 °C. An asterisk indicates statistically significant differences (*p < 0.05; n ) 4) for the mean value of swelling ratio compared with the value at the previous time point. Typical photographs of gelatin samples (b) AD, (c) SFD1, (d) FD1, and (e) FD2 are shown before testing and after incubation for distinct time periods. Incubation time point: hour (h). A scale of graph paper is 1 mm.
temporary carriers in the anterior chamber for at least overnight (i.e., 18 h) is required to maintain the localization of cell sheets at the damaged sites. After immersion for 1 h in BSS, all of the gelatin discs with initial dimension of 7 mm in diameter and 700 µm in thickness exhibited rapid swelling (Figure 3b-e). In the AD groups, the samples disintegrated into small pieces at 18 h; therefore, the swelling tests were not performed entirely. The three types of porous gelatin carriers remained structurally intact at the end of swelling, although the FD2 samples were partially dissolved. These results imply that the carriers prepared by air-drying or freeze-drying at -196 °C may have an unsatisfactory swelling effect on cell sheet delivery use. The central anterior chamber depth of rabbit eyes is ∼2 mm.37 At 6 h, the maximum thickness of swollen hydrogels in the SFD1, FD1, FD2, and AD groups was 2.0, 3.3, 3.6, and 4.0 mm, respectively, which indicates that the dimension of test samples increased with decreasing cross-linking degree. Once implanted, all of the gelatin carriers except for SFD1 sample will have a squeezing effect on the anterior segment tissues of the eye during hydrogel swelling. Furthermore, the maximum volume of SFD1 samples was ∼0.1 cm3, which was smaller than the anterior chamber volume (0.3 cm3) of rabbit eyes.38 Our findings suggest that the gelatin discs prepared by a simple stirring process combined with freeze-drying method are more suitable for cell sheet carriers. Rheological Measurements. The storage modulus (G′) represents the elastic behavior of the sample, whereas the loss modulus (G′′) represents the viscous behavior of the specimen. The order of increasing G′ for the hydrogels was AD > SFD1 > FD1 > FD2 (Figure 4), indicating that the dense gelatin
Figure 4. Mechanical spectra of swollen gelatin hydrogel discs measured at 34 °C.
materials are stiffer than their porous counterparts. Van Den Bulcke et al. studied the viscoelastic properties of cross-linked methacrylamide-modified gelatin films and have shown that the hydrogels with high G′ values are formed because of the structuring of helices.39 Here the AD samples were prepared by air-drying for 3 days. When the gelatin solution undergoes a coil-helix transition on cooling from 40 to 25 °C, a strong physical network is formed. However, a quick freeze-drying process may cause obstruction in the helix formation, thereby reducing the number of the helices and producing a weaker hydrogel.
1392
Biomacromolecules, Vol. 11, No. 5, 2010
Lai and Li
Figure 6. Time course of concentration of glucose permeated through various gelatin hydrogel discs at 34 °C. An asterisk indicates statistically significant differences (*p < 0.05; n ) 5) as compared with the AD groups.
Figure 5. (a) Typical DSC thermograms of swollen gelatin hydrogel discs. (b) Freezable water content (WfH/Ws) of various gelatin samples. An asterisk indicates statistically significant differences (*p < 0.05; n ) 6) as compared with the AD groups. (c) Schematic structure of nonfreezable bound water in gelatin hydrogel.
In the whole range of frequencies investigated, G′ of EDC cross-linked gelatin hydrogels was higher than G′′, and no crossover point between G′ and G′′ was observed. This indicates that the sample behaves like an elastic solid. For cross-linked biopolymer with the formation of an elastic gel network, G′ and G′′ maintain a constant value and are parallel to each other at higher degrees of cross-linking.40 Our results suggest that among the porous gelatin carriers, the SFD1 samples can possess relatively high mechanical strength to support surgical handling. Determination of Freezable Water Content. The nature of water in hydrogel materials is important to control the solute diffusion mechanism.41 There are three main types of waters in swollen hydrogels: nonfreezable bound water, freezable bound water, and free water.42 For the AD samples, the lower temperature peak (about -1 °C) was attributed to freezable bound water, and the higher temperature peak (about 3 °C) was attributed to free water (Figure 5a). In contrast, only one melting peak was observed in the curves for SFD1, FD1, and FD2 samples at ∼0 °C. A study from Wan et al. demonstrated that sharp peaks at ∼2 °C were found for porous fibrous membranes made from acrylonitrile-based polymers, whereas a visible shoulder appeared for the corresponding dense membrane.43 In
accordance with these data, we found that freezable bound water did not exist in porous gelatin discs, or the amount of this type of water was very small. One possible explanation is that the water molecules filled in the space between network chains and the center of pores are almost freely mobile. However, for gelatin samples with dense structure, the movement of water molecules was restricted by the limited free volume. The solute permeability of hydrogels strongly depends on the fraction of water that is available for solute diffusion (i.e., freezable water). The content of freezable water is indicative of the total amount of freezable bound and free water in swollen gelatin hydrogels. As shown in Figure 5b, the freezable water content in the AD, SFD1, FD1, and FD2 groups was 0.59 ( 0.03, 0.88 ( 0.03, 0.83 ( 0.01, and 0.77 ( 0.01, respectively. This indicates that the mobile fraction of water significantly increased with increasing porosity of gelatin materials (p < 0.05). In addition to pore structure, the extent of cross-linking is another important factor affecting freezable water content of gelatin hydrogels. The water molecules in swollen discs may be interacted with free carboxylic acid and amino groups of gelatin by hydrogen bonding. When the two hydrogen atoms of water molecules are firmly fixed to the amino acid residues of gelatin, such water is referred to as nonfreezable bound water (Figure 5c). With increasing cross-linking index, the number of remaining free carboxylic acid and amino groups significantly decreased. Therefore, the SFD1 samples had the least amount of nonfreezable bound water, whose mobility is completely retarded. These findings suggest that the gelatin carriers prepared by a simple stirring process combined with freeze-drying method contain a high fraction of mobile water and may be beneficial to solute diffusion. Glucose Permeation Studies. The aqueous humor is an intraocular fluid responsible for the supply of nutrients to and removal of metabolic wastes from the avascular cornea. Because the implantation of gelatin hydrogels in the anterior chamber may affect mass transport, it is necessary to measure nutrient permeability through carrier materials for predicting their success. At each time point, the amount of permeated glucose was significantly higher in the SFD1 groups, compared with those of the AD, FD1, and FD2 groups (p < 0.05) (Figure 6). Additionally, we noticed that the initial relatively rapid permeation rate decreased with time as the SFD1 samples became saturated with the solute. Yanagawa et al. have shown that the permeability of dextran is strongly influenced by the complexity
Porous Gelatin Hydrogel for Cell Sheet Engineering
Biomacromolecules, Vol. 11, No. 5, 2010
Figure 7. Time course of weight remaining of various gelatin discs after incubation at 34 °C in BSS containing MMP-2. An asterisk indicates statistically significant differences (*p < 0.05; n ) 5) for the mean value of weight remaining compared with the value at the previous time point.
and irregularity of porous structure in the poly(2-hydroxyethyl methacrylate) membranes.44 The present data also suggest that glucose permeation is drastically suppressed either by the dense structure of air-dried materials or by the surface skin layer of freeze-dried samples. The gelatin carriers with enlarged pore structure have the lowest resistance to glucose permeation. Thereby, it is expected that this type of material in the anterior chamber will minimize any adverse effect on corneal physiology. In Vitro Degradation Tests. Under inflammation conditions, the gelatinase in the aqueous humor may be activated.45 MMP-2 becomes overexpressed with intraocular inflammatory diseases such as cataract, glaucoma, and uveitis,46 which often occur in patients requiring penetrating keratoplasty for treatment of corneal endothelial dysfunction. Additionally, the initial response to implantation of biomaterials is the acute and subacute phases of inflammation caused by surgical trauma.47 The activated MMP-2 may enhance the degradation of gelatin discs after their insertion in the anterior chamber. Therefore, in vitro carrier degradation by MMP-2 was evaluated (Figure 7). The weight remaining after 1 day in the AD and FD2 groups was 41.7 ( 1.7 and 39.8 ( 1.9%, respectively. These values were significantly lower than those of the SFD1 (58.2 ( 1.1%) and FD1 (46.4 ( 2.0%) groups (p < 0.05). Our results suggest that the resistance against gelatinase digestion is highly correlated with the extent of cross-linking of chemically modified porous gelatin samples. However, when the degradation time exceeds 3 days, the residual weight for all four groups was reduced to a level below 10% (p > 0.05). One possible explanation is that high porosity increases the access of enzyme to the active sites of the polymer chains, thereby accelerating the degradation of gelatin matrices. The juxtacanalicular meshwork is likely to be a major site of outflow resistance of aqueous humor.48 A greater resistance is attributable to proteins or glycoproteins in aqueous humor
1393
that obstruct the flow. The low-molecular-weight fraction ( 0.05) (Figure 8e). Figure 9a shows the IOP profile of AD and SFD1 groups. Mean baseline IOP values ranged from 16.5 to 19.7 mmHg. A sharp fall in the IOP was observed immediately after insertion of the gelatin discs in the anterior chamber, presumably because
Table 1. GPC Mw Values (Da) and PDI of Gelatin Hydrogels after Enzymatic Degradation for Different Time Intervals 1 day
3 days
7 days
14 days
28 days
sample
Mw
PDI
Mw
PDI
Mw
PDI
Mw
PDI
Mw
PDI
AD SFD1 FD1 FD2
38 891 198 442 108 934 64 599
3.39 8.15 6.29 3.73
13 665 16 376 13 314 14 099
1.66 1.28 1.68 4.92
12 591 7986 9141 10 110
1.44 1.10 1.23 1.38
6798 4390 5388 6717
1.85 1.44 1.49 1.87
3565 3529 3459 3920
1.40 2.07 1.30 1.31
1394
Biomacromolecules, Vol. 11, No. 5, 2010
Lai and Li
Figure 8. Typical photographs of rabbit eyes in the (a) AD and (b) SFD1 groups at different time periods after implantation of gelatin discs in the anterior chamber. Follow-up time point: postoperation (post); hour (h); day (d); week (w). Scale bars: 5 mm. Typical specular microscopic images of rabbit corneal endothelium are shown (c) before surgery and 8 weeks after insertion of (d) AD and (e) SFD1 samples in the anterior chamber. CD: cell density (cells/mm2). Values are mean ( standard deviation (n ) 3).
of the loss of a large amount of the aqueous humor through the corneal/limbal incision.32 At postoperative 3 h, the presence of swollen hydrogels was able to maintain the shape of the anterior chamber. However, a significantly higher increase in IOP was noted in the AD groups as compared with the SFD1 groups (p < 0.05), indicating that the dense gelatin materials may result in a greater aqueous humor outflow resistance in the trabecular meshwork. The IOP changes due to SFD1 implants reverted to the baseline level within 6 h. In contrast, in rabbits receiving AD samples, the IOP was maintained at a high level (i.e., about 35 mmHg) even at 6 h. During the follow-up period from 1 to 5 days, the IOP gradually decreased below baseline values. One possible explanation is that the swelling of AD implants causes a squeezing effect on the tissue (Swelling Tests section), thereby enlarging the corneal/limbal incision and producing a substantial loss of aqueous humor. Our data demonstrate that the material structure of gelatin carriers may affect the balance of IOP. To evaluate the biodegradability of gelatin implants in the anterior chamber, we analyzed the aqueous humor from each rabbit eye for hydroxyproline, which is an amino acid marker of collagen. At 8 weeks postoperatively, the amount of hydroxyproline did not show a significant difference between the control (9.9 ( 0.3 µg/mL of aqueous humor) and SFD1 (10.3 ( 0.3 µg/mL of aqueous humor) groups (p > 0.05) (Figure 9b). Furthermore, these values were significantly lower than those of the AD (35.1 ( 0.2 µg/mL of aqueous humor) groups (p < 0.05). The total hydroxyproline content is 12.5 g per 100 g protein in collagen.34 In this study, the initial weight of a gelatin disk was ∼30 mg, indicating that the amount of hydroxyproline was 12.5 mg/mL of aqueous humor immediately after surgery. Although the differences in amount of hydroxyproline between
AD and SFD1 groups are noted, both samples have a very high in vivo degradability of >99.7% at the end of follow-up. The histological findings for various groups at 8 weeks postoperatively are shown in Figure 9c-e. In the control groups, no inflammation reaction was observed. H&E staining demonstrated normal cornea and iris. The cornea comprises three layers: the multilayered epithelium, the thick stroma consisting of keratocytes and orthogonally arranged collagen lamellae, and the endothelium. In the AD groups, significant infiltration of inflammatory cells was noted on the anterior and posterior surfaces of cornea. Additionally, the exposure to dense gelatin materials caused corneal neovascularization and stromal edema. In contrast, no adverse inflammatory reaction was found after implantation of SFD1 samples, indicating good biocompatibility. To guide successful corneal reconstruction, the biomaterial carriers must exhibit excellent biocompatibility and biostability. In general, biocompatibility is governed mainly by the interface between foreign materials and host living tissues.32,33 Here the implant-tissue interaction was determined by corneal morphological studies, IOP measurements, and histological examinations. The results suggest that the implantation of AD samples may induce undesirable host reactions and some adverse biological effects. Sweeney et al. have shown that after insertion into the feline cornea, the membrane implants with pore diameter
