| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
,
,
,1
* Department of Bioengineering, University of California, San Diego, La Jolla, California, USA;
Harvard-Massachusetts Institute of Technology Division of Health Sciences and Technology, and
Department of Electrical Engineering and Computer Science,
Division of Medicine, Brigham and Women’s Hospital, Boston, Massachusetts, USA; Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; and
|| Department of Bioengineering, Rice University, Houston, Texas, USA
1Correspondence: Massachusetts Institute of Technology, 77 Massachusetts Ave., E19–502D, Cambridge, MA 02139, USA. E-mail: sbhatia{at}mit.edu
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
5ß1 integrin of cocultured hepatocytes. Three-dimensional photopatterned constructs were visualized by digital volumetric imaging and cultured in a continuous flow bioreactor for 12 d where they performed favorably in comparison to unpatterned, unperfused constructs. These studies will have impact in the field of liver biology as well as provide enabling tools for tissue engineering of other organs.—Liu Tsang, V., Chen, A. A., Cho, L. M., Jadin, K. D., Sah, R. L., DeLong, S., West, J. L., Bhatia, S. N. Fabrication of 3D hepatic tissues by additive photopatterning of cellular hydrogels.
Key Words: tissue engineering • liver • hepatic • PEG • microscale
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
, encapsulation of hepatocyte aggregates (2
, 3)
, and biodegradable polymer scaffolds (4
5
6)
. Transplantation of hepatocytes (preaggregated or attached to protein-coated microcarriers) has been shown to provide initial function, but their potential for prolonged efficacy has not been demonstrated (7)
. The high metabolic rate of hepatocytes presents transport limitations as diffusive nutrient transport is only sufficient over short length scales. As a result, methods have been developed to generate scaffolds with architectural features that allow convective perfusion; however, acellular scaffolds can be difficult to populate with cells that are really nonproliferative and nonmotile in vitro, like hepatocytes (8)
. Proposed solutions to this challenge include dynamic seeding of constructs by perfusion with cell suspensions or fabrication of channels in vascular-mimetic patterns (9
10
11)
. Others have incorporated growth factors into scaffolds seeded with hepatocytes to promote engraftment and angiogenesis, resulting in improvements in microvascular density and hepatocyte engraftment over controls (12
, 13)
. Despite these innovations, to date hepatocytes transplanted into rats on biodegradable polymer matrices are still found to be inferior to liver grafts of equivalent liver mass in compensating for metabolic deficiencies (14)
. Thus, methods of building tissue engineered liver constructs with intricate architectures that provide adequate transport of oxygen and nutrients as well as an environment that promotes hepatocyte viability and liver-specific function are still lacking.
In recent years, rapid prototyping manufacturing technologies have been applied toward the fabrication of three-dimensional scaffolds with tunable micro- and macroscale features (15
16
17
18
19
20
21
22
23
24
25
26)
. Synthetic hydrogels as a class of biomaterials have also been of increasing interest in tissue engineering because they can be tailored for specific applications (adhesive regions, biodegradable linkages, etc.) and can be used to entrap cells at tissue-like densities. The transport properties of hydrogels can also be customized by adjusting polymer chain lengths and density. In particular, poly(ethylene glycol) (PEG)-based hydrogels have been widely utilized because of their biocompatibility, hydrophilicity, and ability to be chemically tailored (27)
. They have been explored to immobilize various cell types, including chondrocytes (28
29
30)
, vascular smooth muscle cells (31)
, osteoblasts (32)
, neural cells (33)
, and fibroblasts (34
35
36)
, which can attach, grow, and produce matrix, as well as progenitor cells, which can differentiate into adipose tissue (37)
or cartilage tissue (38)
. PEG-based hydrogels have also been customized by incorporation of adhesion domains of ECM proteins to promote cell adhesion, growth factors to modulate cell function, and both hydrolytic and protease-sensitive linkages (31
, 35
, 39
40
41
42
43
44
45)
.
Photoencapsulation in PEG hydrogels combined with multilayer fabrication methods, therefore, represents a potential opportunity for hepatic tissue engineering (46)
. In this study, we combine the use of PEG hydrogels with photolithographic methods to fabricate hepatic tissues that enable high-density encapsulation while specifying scaffold architecture that enables nutrient delivery by convective flow. We first optimize conditions to incorporate hepatocytes in PEG hydrogels by stabilizing primary hepatocytes in culture and incorporating adhesive peptides to ligate hepatocyte integrins. Next, we photopattern single-layer hydrogel structures that incorporate hepatocytes at high density to reduce diffusive nutrient transport limitations. We subsequently fabricate and image multilayer 3D tissue structures with defined scaffold architecture. Finally, we perfuse these 3D tissue structures in a continuous-flow bioreactor and demonstrate improved viability and liver-specific function of the encapsulated hepatocytes over unpatterned controls.
|
MATERIALS AND METHODS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Detailed procedures for isolation and purification of hepatocytes were previously described by Dunn et al. (48)
. Routinely, 150–250 million cells were isolated with viability between 85% and 95%, as judged by trypan blue exclusion. Nonparenchymal cells, as judged by their size (<10 µm in diameter) and morphology (nonpolygonal or stellate), were less than 1%. Culture medium was Dulbecco’s modified Eagle medium (DMEM; Life Technologies, Inc., Grand Island, NY, USA) supplemented with 10% FBS (Sigma, St. Louis, MO, USA), 0.5 U/mL insulin, 1.5 µg/mL glucagon, 7.5 µg/mL hydrocortisone, 100 U/mL penicillin, and 100 µg/mL streptomycin. NIH 3T3-J2 cells were the gift of Howard Green, Harvard Medical School. Cells grown to preconfluence were passaged by trypsinization in 0.01% trypsin (ICN Biomedicals, Costa Mesa, CA, USA)/0.01% EDTA (Boehringer Mannheim, Indianapolis, IN, USA) solution in PBS for 5 min, diluted, and then inoculated into a fresh tissue culture flask. Cells were passaged at preconfluency no more than 16 times. Cells were cultured in 175 cm2 flasks (Fisher Scientific, Springfield, NJ, USA) at 5% CO2, balance moist air. Culture medium consisted of DMEM (Life Technologies, Inc.) with high glucose (Glc), supplemented with 5% bovine calf serum (BCS, JRH Biosciences, Lenexa, KS, USA) and 100 U/mL penicillin and 100 µg/mL streptomycin. Prior to use, the cells were growth-arrested by incubating in medium with 2 mg/mL Mitomycin C for 2 h so that they would not overgrow the coculture dishes.
Hepatocyte preconditioning and retrieval for flow cytometry and adhesion experiments
For integrin characterization and adhesion experiments, hepatocytes were cultured in one of three conditions: double gel, Matrigel, and coculture with fibroblasts. In addition, Day 0 hepatocytes were obtained from the pool of freshly isolated hepatocytes. Hepatocytes cultured in the double gel, or "sandwich configuration", were cultured between two 0.5 mm layers of collagen gel (49)
. Matrigel gel layers of 0.5 mm thickness consisted of Matrigel Basement Membrane Matrix (BD Discovery Labs, Bedford, MA, USA). Hepatocytes in the coculture condition were seeded on adsorbed collagen and allowed to adhere and spread, and an equal ratio of growth-arrested NIH 3T3-J2 fibroblasts were added on day 1. All cultures were maintained in the hepatocyte growth medium, which was replaced daily.
Cells were removed from culture after 1, 4, and 7 d of culture. Hepatocytes seeded in the double gel configuration were released by incubating the dishes with 5 ml of collagenase (Sigma, 2.68 mg/mL in 5mM CaCl2 in KRB buffer) for 30–40 min at 37°C. BD Cell Recovery solution (BD Discovery Labs) was used to dissolve the Matrigel substrate by incubation for 1 h on ice. Cells in the coculture condition were released by washing with serum-free medium, treating with 0.05% trypsin/EDTA for 2 min to separate cells from each other and release fibroblasts, washing with medium containing serum to inactivate the trypsin, and finally incubating with 5 ml collagenase solution to release the cells from the dish (50)
. Large clumps of cells were removed by passing the cell suspensions through 70 µm mesh cell strainers (BD Biosciences) and washed with culture medium by centrifugation (50 g for 5 min at 4°C).
Cell preparation for hydrogel photopatterning
For 3D tissue fabrication, hepatocytes were maintained in coculture with fibroblasts. P-150 culture dishes were each plated with 7.5 million hepatocytes. Coculture dishes were cultured with an equal number of fibroblasts at the time of seeding. Removal of cells from culture was accomplished by 0.25% trypsin/EDTA with shaking. Some cells were fluorescently labeled with chloromethylfluorescein diacetate (CMFDA; C-2925, Molecular Probes, Eugene, OR, USA), chloromethylbenzoylaminotetramethyl rhodamine (CMTMR; C-2927, Molecular Probes), or 7-amino-4-chloromethylcoumarin (CMAC, C-2110, Molecular Probes) for identification at (ex/em) 492/517, 514/565, or 353/466 nm wavelengths, respectively.
Cell number within the patterned hydrogel construct was determined by calculating the vol of each layer based on the geometry of the masks and by specifying the concentration of hepatocytes in the prepolymer solution. The vol of the hydrogel construct was determined to be 
50 µl. Therefore, the unpatterned gel (same overall thickness and diameter, but no photopatterning) was fabricated by tuning the hepatocyte concentration in the prepolymer solution to achieve the same number of hepatocytes (4x105 cells).
Flow cytometry for integrin characterization
Cells (0.25 x 106) were added to each FACS tube and washed with 1 ml FACS buffer (1% FBS, 0.1% sodium azide in PBS) by centrifugation. Primary antibodies used for flow cytometry were as follows: Armenian hamster IgG 1, 2, 5, hamster IgG isotype control (BD Pharmingen, San Diego, CA, USA), Armenian hamster IgM ß1and hamster IgM istoype control (Pharmingen), mouse IgG 4 and mouse IgG isotype control (Pharmingen), mouse 3 (Ralph 3.1 developed by Louis Reichardt, obtained from the Developmental Studies Hybridoma Bank, developed under the auspices of the NICHD and maintained by the University of Iowa, Department of Biological Sciences), mouse 6ß1 (Biosource, Caramillo, CA, USA), rabbit 9 (clone 1057, kindly provided by Dean Sheppard, Department of Medicine, UCSF), V (Chemicon, Temecula, CA, USA) and rabbit IgG isotype control (Southern Biotech, Birmingham, AL, USA). Diluted primary antibody (Ab; 100 µl; 1:100 in FACS buffer) was added to each tube of cells and allowed to incubate on ice for 15 min. Because the 9 and V antibodies ligate the cytoplasmic domain of the integrin subunit, the cells were first permeabilized in a 50% ethanol solution (1 ml cell culture medium, 1 ml FBS, 6 ml 70% ethanol) by incubating on ice for 30 min and washed with FACS buffer before incubation with primary Ab, according to a protocol previously reported (51)
. All cells were washed by adding 1 ml FACS buffer and centrifuging at 500 rpm for 5 min at 4°C.
Secondary antibodies used were mouse antihamster IgG FITC (Pharmingen), mouse antihamster IgM FITC (Pharmingen), goat anti-mouse IgG FITC (Jackson Labs), and goat anti-rabbit IgG FITC (Sigma). The appropriate secondary antibodies (100 µl, diluted in FACS buffer, 1:100 for hIgG, mIgG, and rIgG, and 1:200 for hIgM) were added to cells and allowed to incubate on ice for 15 min. The cells were washed in 1 ml FACS buffer and then resuspended in 300 µl of FACS buffer. The analysis was performed using a FACScan flow cytometer (Becton Dickinson, Fullerton, CA, USA). Cells (5000 or 10,000) were analyzed for each sample.
Integrin expression was analyzed using FACS for hepatocytes from two rats. In the first experiment, all integrin antibodies were analyzed for all culture conditions. Integrin subunits 2 and 4 were not expressed in any conditions at any time points, and therefore were not used in the second experiment. Expression of 1, 3, 5, 6, 9, and V were studied in both experimental runs.
Matrix adhesion studies
Cells were cultured in and released from common culture conditions (double gel, coculture with fibroblasts, Matrigel) and then added to matrix-coated culture plates to determine whether preculturing of cells affects adhesion and spreading onto certain surfaces. Human laminin, fibronectin, vitronectin, and collagen I coated plates were obtained from BD Discovery Labs. BSA-coated wells were used as a negative control. The cells were shaken gently every 15 min for 1 h and were allowed to adhere for 4 h. The wells were then washed three times with PBS to remove nonadherent cells. Tris-EDTA buffer (Sigma) was added to all wells, which were stored at –20°C until the time of assay. The PicoGreen DNA assay (Molecular Probes, Eugene, OR, USA) was used to quantify the number of cells attached in each well. Cells were subjected to three freeze/thaw cycles followed by ultrasonication, prepared in multiwell format with calf thymus DNA, and interrogated on a SpectraMAX GeminiXS spectrofluorometer (Molecular Devices, Sunnyvale, CA, USA) with ex/em of 485/530 nm. An average amount of 10 pg DNA per hepatocyte was used to estimate cell number.
Peptide film adhesion studies
The peptides RGDS, RGES, and YIGSR were obtained from Sigma. KQAGDV, GF[Hyp]GER, and cyclic (-RGD[D-Phe]K-) were synthesized by Global Peptide Corp. Conjugation of peptides to acryloyl-PEG-NHS and subsequent reaction characterization were performed using a protocol previously described (31
, 35)
. The resulting monomers consisted of PEG chains flanked on one side by the desired peptide and on the other by an active acrylate group.
Peptide hydrogel solutions were created by combining desired concentrations of the PEG monoacrylate-Peptide with poly(ethylene glycol) diacrylate (PEGDA) (MW 3400, Nektar Transforming Therapeutics, Huntsville, AL, USA) and dissolving in PBS to form a 20% w/v solution. The photoinitiator 2,2-dimethoxy-2-phenyl-acetophenone (Sigma) was added to the polymer solution, resulting in a 0.1% w/v final concentration.
Hydrogel films were created using by photocrosslinking the polymer solution in a chamber consisting of a TeflonTM base, silicone spacer (250 µm thickness with 27 mm circle punched out), and a 34 mm circular glass coverslip treated with 3-(Trimethoxysilyl) propyl methacrylate to have methacrylate groups exposed for covalently binding the hydrogel film (46)
. The peptide-PEGDA solution was placed inside the chamber and exposed to 365 nm UV light (10 mW/cm2) through the glass for 45 s to cross-link the hydrogel. The 34 mm glass coverslips with covalently bound hydrogel films were then transferred to 6-well plates for further experimentation. 0.25 x 106 freshly isolated hepatocytes or hepatocytes released from culture in double collagen gel, Matrigel, or coculture with fibroblasts were seeded onto hydrogel films in the absence of serum and incubated at 37°C for 6 h. The substrates were then washed twice with PBS to remove nonadherent cells. Cells in six fields per well and three wells per condition were counted using light microscopy.
Hydrogel chemistry and photopatterning
PEGDA (3400 MW, Nektar Therapeutics, Huntsville, AL, USA) was dissolved in PBS at 40% w/v to form a 2x prepolymer solution. Acryloyl-PEG-RGDS (52)
was combined with the prepolymer solution to achieve a 40 µmol/mL concentration of RGDS in the 2x solution. 2-hydroxy-1-[4-(2-hydroxyethoxy)phenyl]-2-methyl-1-propanone (Irgacure 2959, Ciba Specialty Chemicals, Tarrytown, NY, USA) was dissolved in 1-vinyl-2-pyrolidone (100 mg/mL) (Sigma) to form the photoinitiator stock solution, and this solution was added to the prepolymer solution to form 0.2% w/v of initiator. Cells suspended in medium were added to prepolymer solution in a 1:1 vol ratio, resulting in a final prepolymer solution containing 5–15 million/mL hepatocytes, 20 µmol/mL RGDS, 20% w/v PEGDA, and 0.1% w/v photoinitiator.
Patterning of the hydrogel was achieved using the apparatus described in previously (46)
. Clean 34 mm cover glass circles (Fisher Scientific) were treated with a 2% v/v solution of 3-(trimethoxysilyl) propyl methacrylate (Aldrich, Milwaukee, WI, USA) in 95% ethanol (pH 5 with acetic acid) for 2 min, rinsed with 100% ethanol, and then baked at 100°C for at least 5 min to create free methacrylate groups on the glass. The hydrogel solution was sandwiched between the glass and the TeflonTM, with the height of the chamber determined by the thickness of the silicone spacer used. An emulsion mask (designed using Corel Draw 11.0 and printed using a commercial Linotronic-Hercules 3300 dpi high-resolution line printer) was placed against the glass circle and chamber was exposed to 100 mW/cm2 of 365 nm UV light (as measured before passing through the glass slides and mask; 64 mW/cm2 of light passes through the glass and emulsion mask layers to reach the polymer solution) for 90–110 s using an EXFO Lite UV spot curing system equipped with a 365 nm light filter and a collimating lens (EXFO, Mississuaga, ON, Canada). Addition of layers was accomplished by washing the uncrosslinked solution with PBS and repeating the process using a thicker spacer and new mask. Crosshairs were included as markers on the masks and glass wafers to allow for alignment of successive layers. After completion of patterning, the hydrogel containing cells was washed with PBS and placed in a Petri dish containing culture medium. Alternatively, hydrogels were also fabricated as one thick layer using a mask with no internal pattern to create "bulk" hydrogel constructs for comparison studies.
Bioreactor design
A perfusion circuit, designed and assembled as described previously (53)
, provided a constant supply of nutrients to the hepatic constructs. Hepatic constructs were perfused in a bioreactor chamber consisting of a simple flow-through chamber with a stainless steel porous platen to support the sample while allowing perfusion through the chamber (54)
. Flow circuits were set up in parallel so that two separate bioreactor chambers could be perfused simultaneously. Negative pressure-driven flow was continuous (0.5 ml/min) using a programmable push-pull syringe pump (Harvard Apparatus, Holliston, MA, USA). Culture medium drawn from the reservoir was equilibrated with 5% carbon dioxide and air balance in a gas exchanger made with gas-permeable silastic tubing before entering the perfusion chambers. All circuit components except for the syringe pump were housed in a PID-controlled incubator maintained at 37°C. The tubing and connectors were sterilized by autoclave prior to use.
Hepatocyte viability and function
Cell viability was visualized qualitatively using the MTT assay, in which the tetrazolium ring of the MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide; Sigma) is cleaved by mitochondrial dehydrogenase enzymes of living cells to form a purple precipitate. The hydrogels containing cells were incubated for 3 h with a 0.5 mg/mL solution of MTT in DMEM without phenol red. Cells that have formed the precipitate appear dark purple under light microscopy.
Albumin synthesis, a surrogate marker of protein synthesis for hepatocytes, was used to measure the function of the hepatocytes within hydrogels. Samples of culture medium were collected and assayed for bulk albumin using an HRP-TMB ELISA assay (Ultra-TMB-S, Research Diagnostics, Flanders, NJ, USA). An intracellular albumin Ab assay was used to visualize the albumin within the hydrogels (55)
. Urea synthesis was also measured as a marker of liver specific function using a protocol previously described (55)
.
Microscopy and digital volumetric imaging
Hydrogels and cells were observed and recorded using a Nikon TE300 inverted microscope and digital camera (Photometrics CoolSnap HQ; Roper Scientific, Tucson, AZ, USA) capable of epifluorescence and Hoffman modulation contrast microscopy along with the MetaMorph Image Analysis System (Universal Imaging, Westchester, PA, USA). Alternatively, color photos were taken with a Nikon CoolPix 4300 digital camera that could attach to the microscope. Laser scanning confocal microscopy (Bio-Rad MRC-1024UV, Hercules, CA, USA) was used to image hydrogels containing fluorescently labeled cells.
For digital volumetric imaging (DVI; ref. 56
), samples were fixed in 0.5% gluteraldehyde at room temperature for 30 min, embedded in resin (Nanoblast Miniset Embedding Kit, Polysciences, Warrington, PA, USA) with black India ink opacifier (Higgins), and imaged at 4x magnification. Images were visualized using RESView 3.0 software (Microscience Group, Inc., Corte Madera, CA, USA) as 3D renderings, and exported into Matlab 7.0 software (The Mathworks Inc., Natick, MA, USA) to create image rotations.
Statistics and data analysis
Error bars represent SD, with n = 3 for toxicity, matrix adhesion, and peptide-film adhesion experiments. Bioreactor experiments allowed duplicates of hydrogel perfusion in each of the two runs. Statistical significance was determined using one-way ANOVA followed by Tukey’s Multiple Comparison Test. Unless otherwise stated, statistical significance was determined with P < 0.05. The flow cytometry experiments were repeated twice, and the results presented were collected from one full experiment, with trends representative of all experiments.
|
RESULTS AND DISCUSSION |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
schematically depicts the fabrication method developed in this study for additive photopatterning of 3D tissue constructs. We specifically apply this technique to the fabrication of hepatic tissue constructs by tailoring both hydrogel chemistry and architecture. Hepatocytes are fragile, adhesion-dependent cells; therefore, we also developed methods to stabilize cells after isolation and ligate cell surface integrins in order to promote survival and retention of liver-specific functions over two weeks (Supplemental Fig. 1). To explore the capability of the technique, we photopattern both single-layer hydrogel structures that reduce nutrient transport limitations and multilayer constructs with fluidic channels that enable perfusion in a continuous-flow bioreactor. 3D hepatic tissues were characterized by digital volumetric imaging and by monitoring albumin and urea synthesis as markers of liver-specific function.
|
Hepatocyte integrin expression and adhesion to extracellular matrix
To address the issue of tailoring PEG hydrogels for hepatocytes by incorporating adhesion peptides, we probed the integrin expression profile of hepatocytes to better understand the types of adhesive ligands that may be effective. In addition to examining freshly isolated hepatocytes, we were also interested in changes that occur in integrin expression under various cell culture models that would be used to stabilize the cells prior to encapsulation. Flow cytometry was used to determine the integrin expression profile of primary hepatocytes that were freshly isolated or cultured for up to one week. Results for freshly isolated hepatocytes indicated that 1, 5, and ß1 were expressed on the cell surface (data not shown). 3 and 6 integrin subunits were barely detectable under these experimental conditions. The subunits 2 and 4 were not expressed by freshly isolated rat hepatocytes.
The shifts of the arithmetic means are plotted for integrin subunits 1, 3, 5, and integrin 6ß1 in Supplemental Fig. 2 to compare the expression of hepatocytes that were freshly isolated and those that had been cultured in coculture with fibroblasts, in a collagen double gel sandwich configuration, and on Matrigel over a period of a week (all three methods have been described to stabilize the phenotype of hepatocytes in vitro). Results indicated that 1 integrin subunit expression was maintained across culture conditions, whereas changes in expression level were observed for subunit 5 and integrin 6ß1. Although 5 was detected in freshly isolated hepatocytes, the expression level declines in cultured cells within the first day. Hepatocytes cocultured with fibroblasts could recover their 5 subunit expression with time, while those cultured in double gel maintained the lowered expression level and Matrigel cultures experienced a further decline. The up-regulation of integrin expression on cocultured cells may be a result of the presence of fibronectin, associated with 5ß1, which is produced by fibroblasts. Conversely, hepatocytes that were cultured on Matrigel, a matrix composed mostly of laminin, continued to lose expression of the 5 integrin subunit over 7 d. This may be explained by the fact that hepatocytes cultured on Matrigel form spheroids, resulting in increased cell-cell interactions and decreased cell-matrix interactions. Surprisingly, the levels of 6ß1 integrin increased in cultured cells, even though expression was barely detectable in freshly isolated cells. Notably, although 6ß1 is not typically found in healthy adult liver, hepatocytes have been reported to up-regulate expression of both this integrin and laminin production in some disease states (57)
.
Based on the flow cytometry results, we hypothesized that hepatocytes stabilized in culture might be altered in their capacity to adhere to full-length ECM or their adhesive peptides. Adhesion of hepatocytes on fibronectin, collagen I, laminin, and vitronectin were therefore evaluated (Supplemental Fig. 3). After incubation for 4 h in the absence of serum, freshly isolated hepatocytes attached and spread to fibronectin, collagen, and laminin substrates, but did not adhere to vitronectin. Hepatocytes that had been cultured on Matrigel, in contrast, were not able to adhere to matrix substrates to any large degree. Hepatocytes released from collagen gel could attach to matrix, and some cell spreading was observed. The levels of integrin expression as compared among various culture conditions after 7 d was consistent with their ability to adhere to matrix substrates: coculture > double gel > Matrigel for all conditions. The flow cytometry results for 1, 5, and 6ß1 all yielded similar results when comparing different culture conditions: cocultured hepatocytes expressed the highest levels of integrins, cells cultured on Matrigel expressed the lowest levels, and double gel cultured cells expressed medium levels of integrins (higher than Matrigel, lower than coculture). Therefore, these semiquantitative results indicate that cocultivation with fibroblasts may serve the dual purpose of inducing liver-specific functions and preserving maximal levels of functional integrin expression. This indicates that stabilization of hepatocytes in coculture may better prepare the cells for incorporation into biomaterials by increasing their affinity to ligands as compared to other culture models.
Hepatocyte adhesion on peptide films
Next, we embarked on studies to translate insights from cell adhesion to full length ECM to selection of synthetic peptides to incorporate into inert PEG hydrogels to improve hepatocyte viability and function. The use of small peptides for promoting interaction with biomaterials is desirable because they are relatively stable compared with whole matrix molecules, are easily conjugated to biomaterials, are easily synthesized, and have low immunogenic activity (58)
. The peptides studied included RGDS (Arg-Gly-Asp-Ser, derived from fibronectin), YIGSR (Tyr-Ile-Gly-Ser-Arg, derived from laminin), GF(Hyp)GER (Gly-Phe-HydroxyPro-Gly-Glu-Arg, derived from collagen), KQAGDV (Lys-Gln-Ala-Gly-Asp-Val, derived from fibrinogen), and the negative control RGES (Arg-Gly-Glu-Ser). Peptides were incorporated into PEGDA hydrogel films, and adhesion on these surfaces was compared to adhesion on PEGDA only gels (negative control) as well as adsorbed collagen I (positive control).
Adhesion of freshly isolated and precultured hepatocytes was investigated. Overall, adhesion of freshly isolated hepatocytes was minimal for peptide-doped hydrogels as compared to adhesion on adsorbed ECM. Hepatocytes did not spread appreciably on any of the PEG-peptide films. In contrast, fibroblasts seeded on RGDS films attached and spread in a dose-dependent manner (not shown), verifying that the peptides were indeed present and available for binding. Of the peptides used in this study, only RGDS was able to confer hepatocyte adhesion but not spreading for the concentrations tested, similar to previous reports (Fig. 2
A) (59
60
61
62)
. A negative control of a similar peptide, RGES, was also examined to rule out the non-specific binding to the short peptide.
|
Hepatocytes cultured in double collagen gel, on Matrigel, and in coculture with fibroblasts for 1 wk were seeded onto 20 µmol/mL RGDS-doped PEGDA films to determine adhesion capabilities of precultured hepatocytes. Results were similar to those on full-length ECM in that cells that had been cultured on Matrigel demonstrated the least adhesion to RGDS-PEG films, while those that had been cocultured with fibroblasts demonstrated the highest levels of adhesion (Fig. 2B
). Based on these results, a combination of coculture of hepatocytes with fibroblasts followed by incorporation into a PEG hydrogel containing 20 µmol/mL of RGDS were used as a starting point for fabrication of a hepatic construct.
Hepatocyte adhesion to YIGSR has been shown previously when grafted onto polymer surfaces (63)
, but we were unable to demonstrate attachment in the range of peptide concentrations tested. Although hepatocytes did not spread on RGD-containing surfaces, it has been shown previously that rounded cells on matrix surfaces (low density) preserve liver-specific function, while hepatocytes that spread on high-density matrix substrates lose phenotypic stability over time (61
, 62)
. The incorporation of RGD peptides in gels with hepatocytes is not unprecedented—hepatocyte spheroids cultured in thermoreversible gels containing RGD have been reported to maintain viability and liver-specific functionality at a higher level than in the same biomaterial without RGD peptides (64)
. In addition, transformed, SV-40 immortilized murine hepatocytes exhibit enhanced albumin production when cultured in PEG hydrogels incorporating RGD peptides compared to PEG hydrogels without peptide (65)
.
A lack of adhesion in our peptide-doped hydrogel adhesion study is not an absolute indicator that hepatocytes cannot adhere to these peptides; integrin binding to peptides may be affected by the way in which the peptides are presented (adjacent amino acids, spacer chains, concentration of peptides for clustering, synergy domains, etc.). It has been reported that integrin binding to RGD requires a PHSRN synergy domain (66)
. While groups have shown that the PHSRN synergy site is unnecessary for 5ß1 integrin that is in the activated state (67
, 68)
, Garcia and coworkers have maintained that synergy plays a role in cell adhesion to RGD (69
, 70)
. Alternatively, others have shown that the affinity of cell binding can be modulated by varying the presentation of RGD from various tether lengths (71
, 72)
or in nanoscale clusters (40
, 73)
. A combination of peptides may also provide a more effective approach to cell adhesion (74)
. It has also been suggested that incorporation of protein domains may enhance adhesion as compared to the minimalist peptide sequences because they retain more characteristics of the natural matrix molecules (75)
. Other, carbohydrate-based domains may also improve on hepatocyte adhesion. For example, incorporation of galactose-derived sugars onto various surfaces have been reported to confer hepatocyte adhesion and spreading via an asialoglycoprotein receptor (76
77
78)
, and may offer an alternative to or work in conjunction with adhesion peptides. Finally, it is also possible to incorporate certain whole matrix molecules such as fibronectin into biomaterials to achieve cell attachment (79
80
81)
.
Hepatocytes encapsulated in RGD-doped PEG hydrogels
Based on the results described above, hepatocytes were stabilized in coculture after isolation and then photoencapsulated into RGDS-doped PEG hydrogel discs. The incorporation of 20 µmol/mL RGDS increased the liver-specific function by 5-fold over PEG controls over 3 days, as assessed by urea secretion (Fig. 2C
) and albumin production (not shown).
Next, we investigated the potential for photopatterning of the hydrogels to improve nutrient transport and hepatocyte viability in single layer constructs by assessing mitochondrial activity by MTT 24 h after photoencapsulation (Fig. 3
). Hepatocytes were photoencapsulated at density of 7.5 x 106 cells/ml in hydrogel discs of 1.5 mm thickness and 10 mm diameter and compared to constructs of the same density and outer dimensions (1.5 mm thickness, 10 mm diameter) but with photopatterned features of 500 µm. Hydrogel discs exhibited central pallor in MTT staining consistent with a necrotic core, whereas photopatterned constructs retained dark staining throughout the construct (Figs. 3A, B)
. Densitometry analysis of photomicrographs revealed that the 55% reduction of optical density (OD) in the core vs. edge of the unpatterned hydrogel plug is mitigated by photopatterning (Fig. 3C
). Thus, photopatterning of hydrogels facilitates the transport of oxygen and nutrients required to meet the high metabolic demands of encapsulated hepatocytes.
|
Using the parameters identified for homogenous MTT staining in photopatterned constructs, three single-layer patterns were constructed and viability was confirmed by loading with fluorescent membrane integrity dyes (Fig. 3D
). These data set the stage for fabrication of multilayer constructs by assembling the individual layers.
Multilayer 3D hepatic tissue construct
The design of the multilayer hepatic construct consisted of a three-layered hexagonal branching structure that served to demonstrate the utility of the photopatterning technology, minimize transport limitations, and mimic the branching architecture of the liver in vivo. The patterning masks utilized to generate this construct were designed to provide single UV exposures to all regions of cells despite the need for three photopatterning steps (Fig. 4
A). Fluorescently labeled hepatocytes were incorporated into the hydrogel fabrication process for visualization by digital volumetric imaging (Fig. 4B-D
). The 3D rendering (Supplemental Movie) demonstrates the incorporation of cells in three clearly identifiable layers of the tissue construct.
|
Scale-up without transport limitations represents a challenging issue in hepatic tissue engineering, as hepatocytes are highly metabolic and require a nearby nutrient supply source for viability and phenotypic stability (82
, 83)
. Therefore, to evaluate the liver-specific function of the multilayer construct, the tissue was transferred to a perfusion bioreactor. The channels created in the multilayer structure allow for convective flow of culture medium through the construct and relatively short distance for diffusive transport to all encapsulated cells. After 1 or 3 d in the bioreactor, samples were removed and again imaged using MTT as a marker of metabolic activity. Figure 5
depicts hydrogel constructs from days 1 and 3, demonstrating that viable cells are visible throughout all three layers of the construct. Culture medium collected over 12 d in perfusion culture was assayed for albumin and urea synthesis, which serve as representative markers of synthetic and metabolic function in the liver. Hepatic constructs cultured in the bioreactor produced higher levels of both albumin and urea than unpatterned constructs of equal cell number and external scaffold dimensions (Fig. 5D, E
). The levels of albumin production are comparable with those reported for hepatocyte spheroids in rotating cultures (84)
but are lower than albumin produced by hepatocytes stabilized in 2D coculture (55)
. Urea synthesis, however, is on par with levels measured in stabilized 2D cultures (55)
. The levels of function achieved in the patterned, perfused constructs can be attributed to the deliberate incorporation of a combination of microenvironmental factors, including cell-cell interactions, adhesive peptides and improved nutrient transport. Hydrogels can also be tailored to incorporate additional bioactive ligands (85
, 86)
, degradable linkages (87
, 88)
, pore sizes that promote protein release (89)
or other supporting cell types (e.g., endothelial cells) organized in defined layers. Strategic manipulation and optimization of these parameters promises to allow further enhancement of hepatocyte functions. Collectively, our results indicate that 3D hepatic tissue constructs can be fabricated using multilayer photopatterning and maintained for two weeks in a perfusion bioreactor, setting the stage for further in vitro and in vivo studies.
|
|
CONCLUSIONS |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
. Therefore, use of multilayer photopatterning to further specify microscale tissue architecture enables progress toward the goal of so-called "organ printing". From a fundamental perspective, combining multilayer photopatterning with techniques to organize cells within hydrogels will allow investigation of structure/function relationships in a 3D tissue context (91
, 92)
.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
Received for publication August 9, 2006. Accepted for publication October 11, 2006.
|
REFERENCES |
|---|
|
TOPABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTS AND DISCUSSIONCONCLUSIONSREFERENCES |
|---|
