Hepatocyte Cocultures with Endothelial Cells and Fibroblasts on

Feb 18, 2014 - ABSTRACT: The maintenance of hepatocyte phenotype and functions remains as a great challenge in the generation of functional liver tiss...
0 downloads 0 Views 2MB Size
Article pubs.acs.org/Biomac

Hepatocyte Cocultures with Endothelial Cells and Fibroblasts on Micropatterned Fibrous Mats to Promote Liver-Specific Functions and Capillary Formation Capabilities Yaowen Liu, Huinan Li, Shili Yan, Jiaojun Wei, and Xiaohong Li* Key Laboratory of Advanced Technologies of Materials, Ministry of Education, School of Materials Science and Engineering, Southwest Jiaotong University, Chengdu 610031, People’s Republic of China ABSTRACT: The maintenance of hepatocyte phenotype and functions remains as a great challenge in the generation of functional liver tissue and in vitro model for drug metabolism studies. The use of hepatocyte coculture systems plays essential roles in the establishment of cell−cell and cell−extracellular matrix communications similar to native liver tissues. In the current study, micropatterned electrospun fibrous mats were created to load hepatocytes, fibroblasts, and endothelial cells (ECs), which were precisely assembled to establish their spatially controlled coculture for mimicking the in vivo structure of hepatic lobules. Hepatocytes formed compact polyhedral spheroids with an average diameter of 80−100 μm, reorganized actin filaments in the cell−cell contact regions, and well-developed bile canaliculi. Compared with hepatocytes cultured alone, the coculture of hepatocytes with either fibroblasts or ECs led to significantly higher albumin secretion, urea synthesis and cytochrome P-450 expression, which were dramatically improved by the coculture of hepatocytes with both fibroblasts and ECs. The cocultured ECs well spread on patterned regions with little organized filamentous actin, and significantly higher densities and deeper penetration into patterned scaffolds were determined for ECs after coculture with fibroblasts and hepatocytes compared with those after cultured alone or coculture with either fibroblasts or hepatocytes. A Matrigel overlay assay showed that the capabilities of ECs to form capillary tubes were significantly enhanced by micropatterned coculture with fibroblasts and hepatocytes. Thus, the coculture of hepatocytes, fibroblasts, and ECs on micropatterned fibrous mats helps both hepatocytes in the maintenance of hepatic functions and ECs in the formation of capillary-like structures. It is suggested that the micropatterned coculture model described here not only provides functional hepatic tissues for predictions of drug metabolism profiles, but also will enable investigations on more complex and physiological cell−cell communications.



INTRODUCTION The liver is playing critical functions as regulating glucose metabolism, producing plasma proteins and controlling the detoxification of endogenous and exogenous compounds. Freshly isolated hepatocytes are currently regarded as the most superior cell source for generation of functional liver tissues and in vitro model for use in prediction of drug metabolism profiles.1 However, primary hepatocytes do not substantially proliferate and rapidly lose their phenotype and liver-specific functions after isolation from the liver. Various methods for hepatocyte culture have been proposed to maintain the viability through optimization of cell sources, and medium components.2 The use of collagen, laminin, and fibronectin as substrates and the conjugation of galactose ligands and other cell adhesion peptides on substrates have been adopted to establish chemical interactions with hepatocytes and guide their assembly into functional tissues. Qiu et al. covalently coupled galactosylated chitosan on poly(εcaprolactone) sponge, promoting hepatocyte adhesion, spheroid formation, and long-term maintenance of liver-specific functions such as albumin secretion.3 These studies have made © 2014 American Chemical Society

some progress in maintaining hepatic functions in vitro, but have not exhibited any replication of the exact microenvironment for hepatocytes to proliferate and realize functions. For example, the loss of their specific morphologies and biological functions is believed to be associated with insufficient interactions with neighboring cells and extracellular matrices (ECMs).4 Hepatic lobule is the basic unit of liver tissue and consists of primary hepatocytes, endothelial cells (ECs), fibroblasts, hepatic stellate cells, and kupffer cells. Within a lobule, hepatocytes are arranged into a single cord-like structure and are separated by adjacent sinusoids. Hepatocytes and ECs together account for more than 80% of the liver mass, and this cord-like structure allows hepatocytes to be in close proximity with ECs and interposed by sparse ECMs.5 In addition, fibroblasts have been clearly shown to support hepatocytes in maintaining their differentiated functions for a long period. But Received: December 30, 2013 Revised: February 17, 2014 Published: February 18, 2014 1044

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Figure 1. (a) Digital images of photomasks printed by E-beam mask lithography system, (b) patterned collectors, and (c) pattered fibrous mats for loading ECs (a1−c1), fibroblasts (a2−c2), and hepatocytes (a3−c3). Bars represent 2 mm. (d) SEM images of patterned fibrous mats with distinct ridge and groove regions, and (e, f) fibrous morphologies in the ridge regions with different magnifications.

fibroblasts are not in physical contact with hepatocytes in native liver tissues, and the space of disse resembles morphologically between fibroblasts and hepatocytes.6 As physiological and pathological events of the liver highly depend on active contacts among these cell types, it is necessary to construct an in vitro culture model to establish cell−cell and cell−ECM communications similar to native liver structure. Attempts have been made to create an exact mimicry of the sinusoidal microstructure of liver and to preserve liver functions by the use of hepatocyte coculture systems, including threedimensional (3D) coculture in bioactive hydrogel matrices, layered hepatocytes, and feeder cells, and heterotypic spheroid formation in microwells.7−9 Nahmias et al. described a coculture system containing mature hepatocytes and ECs on Matrigel, where hepatocytes were recruited to endothelial vascular structures by EC-derived hepatocyte growth factor and then formed a sinusoid-like structure.10 Yamada et al. developed a hydrogel fiber-based scaffold for forming 3D restiform hepatic micro-organoids consisting of primary rat hepatocytes and fibroblasts with a length of up to 1 mm and a diameter of 50 μm, mimicking the hepatic cord structures found in liver. After 90 days of culture, a significant enhancement of hepatic functions, including albumin secretion and urea synthesis as well as expression of hepatocyte-specific genes, was observed because of heterotypic and homotypic cell−cell interactions, compared with conventional monolayer culture and single cultivation in hydrogel fibers.11 Chen et al. established the coculture of hepatocytes with both ECs and fibroblasts in poly(ethylene glycol) diacrylate polymer scaffolds, indicating that the coencapsulation of ECs with hepatocytes/ fibroblasts clusters was the most beneficial to rat hepatocellular function.12 Although such in vitro systems do many improvements in maintaining hepatocyte phenotype and functions, it is still difficult to create heterotypic micro-organoids with precisely ordered multiple cell types. The increase in the heterotypic interface using a micropatterned coculture is supposed to enhance liver-specific functions. Micropatterning approaches, such as photolithography, microcontact printing, micromolding, inkjet printing,

and dip-pen spotting, have been used to create spatial variations in charge, hydrophilicity, and topology to mediate cell patterning.13 Particularly formation of hepatic spheroids using nonadhesive surfaces or patterned ECMs has been one of the most widely applied techniques for precisely positioning multiple cell types into micrometer scale materials and efficiently achieving cell−cell interactions.14 Kohji et al. fabricated a polydimethylsiloxane (PDMS) chip to pattern circular holes with diameters of 500 μm coated with collagen in which hepatocytes were first adhered to patterned collagencoated regions, followed by peeling off this chip to allow fibroblasts adhered only in non-collagen-coated regions. In this coculture system, hepatocytes surrounded by fibroblasts improved liver-specific functions as compared to unorganized cocultures.15 However, the lack of porous structure, the use of nondegradable materials, and the two-dimensional (2D) feature might be problematic upon the formation of 3D hepatic cordlike microtissues in vitro. Electrospun fibers as an attractive scaffold for tissue engineering have a porous structure similar to protein fibers within native ECMs, and are capable of supporting the attachment and proliferation of a variety of cell types.16 A few studies have focused on hepatocyte culture on electrospun fibrous mats.17,18 Chua et al. indicated that hepatocytes cultured on electrospun fibers of galactosylated poly(ε-caprolactone-co-ethyl ethylene phosphate) formed aggregates of 20−100 μm, indicating advantages in the stable immobilization of hepatocyte spheroids by engulfing the fibers compared with a film substrate of the same polymer matrices.18 To our knowledge, no work focuses the coculture on micropatterned electrospun fibrous mats to form hepatic micro-organoids that closely mimic the in vivo structure of hepatic lobules. In our previous study, a glass substrate patterned with an electrically conductive circuit was designed by lithography as a collector for electrospinning, and patterned fibrous mats were constructed with distinct ridge and groove regions. The cell growth and collagen deposition appeared to be confined to precise locations, sizes, and shapes of patterned scaffolds without any adverse effect on the cell viability and ECM secretions.19 In the current study, the coculture of 1045

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

hepatocytes with fibroblasts and ECs were developed by precise assembly of cell-loaded patterned electrospun mats to reassemble into an in vivo-like structure in which ECs were intercalated between hepatocytes and fibroblasts. Fibroblasts were located in patterned regions separately from those of hepatocytes, since the random coculture of hepatocytes and fibroblasts restricted hepatocytes move, and hampered the formation of spheroids.15 The viability, spheroid formation, and liver-specific functions of hepatocytes, including urea synthesis, albumin secretion, and cytochrome P450 expression were evaluated. The proliferation and formation of capillary structures by ECs were also determined after coculture on patterned fibrous mats.



Voltage Power Supply Company, Tianjing, China). The collected fibers with a patterned distribution were used for hepatocyte loadings. Additionally, patterned PELA fibrous mats with different pattern features were also prepared as above for loading fibroblasts and ECs separately. Characterization of Patterned Fibrous Mats. The patterning features of electrospun fibrous mats was observed by an optical microscope (Nikon Eclipse TS100, Japan). The micropatterns of fibrous mats and morphologies of electrospun fibers were investigated by a scanning electron microscope (SEM, FEI Quanta 200, The Netherlands) equipped with a field-emission gun (20 kV) and a Robinson detector after 2 min of gold coating to minimize the charging effect. The fiber diameter was measured from SEM images as described previously.19 To investigate the feasibility of assembly into micropatterned scaffolds, fibrous mats were incubated into 1 mg/mL of FITC, rhodamine B and DAPI separately. The three patterned fibrous mats were rightly stacked by fitting the bulges of one patterned mat nearly into the dents of another mat, ensuring a close contact with each other. After washed with distilled water, the scaffold was observed by a confocal laser scanning microscope (CLSM, Leica TCS-SP2, Germany) under the excitation and emission wavelengths of 490 and 520 nm for FITC, 550 and 620 nm for rhodamine B, and 340 and 490 nm for DAPI, respectively. The fluorescence signals were merged by ImageJ software to confirm the assembly of micropatterned scaffolds. Coculture of Hepatocytes, Fibroblasts, and ECs on Micropatterned Scaffolds. Swiss mouse embryo fibroblasts NIH3T3 and human umbilical vein ECs were from American Type Culture Collection (Rockville, MD). Primary hepatocytes were isolated from livers of adult SD rats using collagenase perfusion procedure as described previously,22 and the hepatocyte viability was determined to be greater than 90% according to trypan blue exclusion assay. Male SD rats weighing from 120 to 150 g were from Sichuan Dashuo Biotech Inc. (Chengdu, China), and all animal protocols were approved by the University Animal Care and Use Committee. The hepatocytes, fibroblasts, and ECs were incubated in Dulbecco’s modified Eagle’s medium (DMEM, Gibco BRL, Rockville, MD), supplemented with 10% of fetal bovine serum (FBS, Gibco BRL, Grand Island, NY). Patterned fibrous mats were seeded with 1 mL of cell suspension at a cell density of 2 × 105, 3.0 × 105 and 1.5 × 105 cells/cm2 for hepatocytes, fibroblasts and ECs, respectively. After culture for 1 day to make cells diffuse into and adhere onto the fibrous mats, the three patterned fibrous mats were assembled as above. As expected, the coculture of above cells was established with controlled cellular distribution in the respective regions. To assist cell seeding and define the patterns of fibrous mats, a circular torus of 14 cm in diameter was placed right above the stacked fibrous mats. The assembled constructs were further cultured in DMEM supplemented with 10% FBS, and the culture media were refreshed every 2−3 days. Characterization of Cell Growth on Patterned Fibrous Mats. The cellular distribution on fibrous mats was determined after immunohistochemical (IHC) staining on intracellular albumin produced by hepatocytes, collagen I secretion by fibroblasts, and collagen IV secretion by ECs. Briefly, cell-loaded fibrous mats were washed three times with PBS, fixed with 4% paraformaldehyde for 30 min at 4 °C, and then incubated in 0.1% Triton X-100 in phosphatebuffered saline (PBS) for 20 min at room temperature. After rinsing with PBS, cell-loaded fibrous mats were incubated with goat antirat albumin antibody at 4 °C for 24 h. After washing five times with PBS, cells were incubated with mouse antigoat IgG-FITC at 37 °C for 30 min. The collagen I produced by cocultured fibroblasts was stained using rabbit antimouse collagen I antibody as the primary antibody and goat antirabbit IgG-AMCA as the secondary antibody. IHC staining of collagen IV from cocultured ECs was performed by mouse antihuman collagen IV antibody and goat antimouse IgG-TRITC. After washing with distilled water, CLSM images were observed as above under the excitation and emission wavelengths of 490 and 520 nm for FITC labeling, 550 and 575 nm for TRITC, and 350 and 450 nm for AMCA, respectively. The fluorescence signals were merged for each sample by ImageJ software to confirm the coculture of three types of cells on micropatterned scaffolds. To clarify the vertical distribution

EXPERIMENTAL SECTION

Materials. Poly(ethylene glycol)-poly(DL-lactide) (PELA, Mw = 42.3 kDa, Mw/Mn = 1.23) was prepared by bulk ring-opening polymerization of lactide using stannous chloride as the initiator.20 Lactosylated poly(DL-lactide) (lac-PLA, Mw = 7.6 kDa, Mw/Mn = 1.32) was prepared by bulk ring-opening polymerization of D,L-lactide using pentaerythritol as the core and stannous octoate as the initiator, followed by conjugation of lactobionic acid.21 Pentaerythritol, N,N′dicyclohexylcarbodiimide (DCC), and N-hydroxysuccinimide (NHS) were from Chemical Regents Company of China (Shanghai, China). Poly(ethylene glycol) (Mw = 6 kDa), lactobionic acid, stannous octoate, 4,6-diamidino-2-phenylindole dihydrochloride (DAPI), rhodamine B, fluorescein isothiocyanate (FITC), tetramethylrhodamine isothiocyanate (TRITC)-phalloidin, and collagenase IV were procured from Sigma Aldrich (St. Louis, MO). Calcein-AM, propidium iodide, and fluorescein diacetate (FDA) were purchased from Molecular Probes (Carlsbad, CA). Goat antirat albumin antibody was received from Abcam (Cambridge, U.K.), and goat antirabbit IgG-aminomethylcoumarin acetate (AMCA) were from Innova Biosciences Ltd. (Innova, U.K.). Mouse antihuman CD31 antibody, rabbit antimouse collagen I antibody, mouse antihuman collagen IV antibody, goat antimouse IgG-TRITC, and mouse antigoat IgG-FITC were obtained from Biosynthesis Biotechnology Co., Ltd. (Beijing, China). All chemical reagents used were of analytical grade and from Chengdu Kelong Reagent Co. (Chengdu, China), unless otherwise indicated. Construction of Patterned Collector. The patterned collector was constructed on a glass template patterned with an electrically conductive circuit, as described previously.19 Briefly, the designed template contained several squares of 15 mm in diameter, each having blank and patterned areas containing parallel strips of 100 and 200 μm wide. The strips were drawn by tanner L-edit software and printed with high resolution by E-beam mask lithography system (Mark 40, CHA Industries, Fremont, CA) to obtain a photomask (Figure 1a). An insulating glass substrate with dimensions of 15 × 15 cm2 was deposited with a silver layer by DC sputtering (Sunicoat 594L, Sunic, Korea), coated with a layer of photoresist (MicroChem Inc., Newton, MA), then covered with the photomask, and exposed by a lithography machine (Suss Mircotec MA6, Germany). The exposed regions became soluble, and the glass substrate was rinsed to remove the photoresist, followed by etching away silver in the exposed area. After etching and removing the rest of the photoresist, the glass substrate was obtained with micropatterned silver circuit as a collector for electrospinning process (Figure 1b). Preparation of Patterned Fibrous Mats. Electrospun patterned fibrous mats were obtained as described previously with some modifications.19 Briefly, PELA and Lac-PLA were blended at the weight ratio of 1/1, and dissolved in chloroform at 15 wt %. The polymer solution was added into a 2 mL syringe, attached with a clinicshaped metal capillary. A syringe pump (Zhejiang University Medical Instrument Company, Hangzhou, China) was used to feed the polymer solution to maintain a steady flow at 0.6 mL/h from the capillary outlet. The distance between the capillary tip and patterned collector was about 15 cm, and the electrospinning voltage was controlled within 20 kV using a high voltage statitron (Tianjing High 1046

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

of cells in patterned fibrous mats, CLSM images were taken by z-stack scanning with a step size of 5 μm. The images were stacked and reconstructed in a 3D projection by ImageJ software to measure the cell distributions within micropatterned fibrous scaffolds as described previously.23 The organization of actin filaments (F-actin) in hepatocyte, fibroblasts, and ECs on fibrous mats were stained with TRITCphalloidin. Briefly, cell-loaded fibrous mats were incubated with TRITC-phalloidin of 10 μg/mL for 20 min at 37 °C, followed by an extensive wash with PBS. After counterstaining with DAPI, the samples were observed by CLSM under the excitation and emission wavelengths of 550 and 575 nm for TRITC labeling, and 360 and 460 nm for DAPI, respectively. The cell-loaded fibrous mats were dehydrated through a series of graded ethanol solutions and then freeze-dried. Dry constructs were sputter coated with gold and observed by SEM as described above. Characterization of Cell Viability on Patterned Fibrous Mats. For the detection of cell viability, cell-loaded scaffolds were incubated with 50 mM calcein-AM and 25 mg/mL propidium iodide in culture media for 50 min at 37 °C, followed by fixing with 4% paraformaldehyde for 30 min at room temperature. CLSM images of cells were viewed under the excitation and emission wavelengths of 490 and 515 nm for calcein-AM staining, and 535 and 617 nm for propidium iodide, respectively. The diameter and amount of hepatocyte spheroids in each image were processed by Image-Pro 6.0 imaging software, as described previously.16 In order to have a better perception of cell migration within patterned scaffolds and organization into capillary-like structures, CLSM images were taken by z-stack scanning with a step size of 5 μm. These images were postprocessed using the image processing software ITK-SNAP to identify the spatial distribution of capillary-like structures as described previously.24 Assessment of Hepatocyte Functions on Patterned Fibrous Mats. Albumin and urea secretions were analyzed by measuring the concentration of albumin and urea in the media conditioned. Briefly, culture media were removed for analysis at predetermined intervals, and replaced with fresh media. The retrieved media were centrifuged to collect the supernatant in prechilled tubes, and stored at −20 °C. The albumin secretion was measured by enzyme-linked immunosorbent assay (ELISA) using commercial rat albumin Quantitation Kit (Nanjing Jiancheng Bioengineering Institute, Nanjing, China). Urea levels were measured with commercially available kits (Nanjing Jiancheng Bioengineering Institute, Nanjing, China) based on its specific reaction with diacetyl monoxime. Absorbance was measured with a microplate reader (Elx-800, Bio-Tek Instrument Inc., Winooski, VT), and standard curves were generated using purified rat albumin or urea dissolved in culture media. The activity of cytochrome P450 1A2 (CYP1A2) was determined using the luminescent P450-Glo CYP1A2 cell-based assay kit (Promega, Madison, WI) following the manufacturer’s protocol. Briefly, the retrieved culture media were transferred into a 96-well plate, and an equal volume of luciferin detection reagent was added to initiate a luminescent reaction. The plate was incubated at room temperature for 20 min, and the luminescence was read using a microplate reader as above. The biliary excretion of hepatocytes was determined by FDA staining as described previously.25 Briefly, cell-loaded fibrous mats were incubated in culture media containing 3 mg/mL of FDA at 37 °C for 45 min. After extensive wash with PBS, cell-loaded fibrous mats were observed by CLSM under the excitation and emission wavelengths of 488 and 533 nm, respectively. These images was processed by Image-Pro 6.0 imaging software to quantify the fluorescein localization in the intercellular sacs between hepatocytes following procedures described previously.25 Briefly, a confocal image was imported into Image-Pro 6.0 and then segmented by using the “Segmentation” tool. In order to reduce the background noises, two suitable threshold values of pixel were separately chosen in the segmented image to ensure that all intracellular and total area were covered by FDA staining. The areas of fluorescence signals were quantified by using the “Count/Size” tool, and the biliary excretion of a FDA-staining image was indicated by the ratio of the area of excreted

fluorescein in the intracellular sacs to the total area covered by FDAcontaining cells. For each sample, five original images were randomly chosen to get statistically meaningful results. Characterization of EC Growth on Patterned Fibrous Mats. To examine the EC growth on patterned fibrous mats, ECs were immunostained for CD31 and quantified through image analysis. Briefly, after culture for 7 days, cell-loaded fibrous mats were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 in PBS. After rinsing with PBS, cell-loaded fibrous mats were incubated with mouse antihuman CD31 antibody at 4 °C for 24 h, and then with goat antimouse IgG-TRITC at 37 °C for 30 min. After counterstained with DAPI for 5 min, fibrous samples were observed by CLSM under excitation and emission wavelengths of 550 and 575 nm for TRITC labeling, and 360 and 460 nm for DAPI, respectively. The fluorescence signals of DAPI and TRITC were merged for each sample by ImageJ software. To evaluate EC distribution into the micropatterned fibrous scaffold, CLSM images were taken by z-stack scanning with a step size of 5 μm. The images were stacked and reconstructed in a 3D projection by ImageJ software to measure the vertical distributions of CD31 stained cells within micropatterned fibrous scaffolds.23 Additionally, a 3D image reconstruction was used to measure the amount of CD31-positive cells by using Image-Pro software as described previously.26 Briefly, since all of the CLSM images were taken with the same exposure time, a global intensity threshold of the half intensity level of the maximum (the absolute intensity value of 255 for all images) was applied uniformly to create a binary map for each image. In threshold images only areas contiguous over more than four pixels were counted. The cell density was evaluated with the total cell number in each microfeature, divided by the periodic area, and then averaged through 10 random fields of view per sample.27 Characterization of Capillary Formation Capability by ECs on Patterned Fibrous Mats. A Matrigel overlay method was used to induce capillary formation by ECs in the coculture systems, as described previously.28 Briefly, ECs were stained with Cell-Tracker CM-DiI (Invitrogen, Carlsbad, CA) according to the manufacturer’s instructions. After 7 days of coculture, the medium was aspirated and cell-loaded fibrous mats were overlaid with 200 μL Matrigel (BD Biosciences, Bedford, MA). Matrigel was allowed to polymerize for 1 h in an incubator before the addition of culture media. After Matrigel overlay for 3 d, fibrous samples were observed by CLSM under the excitation and emission wavelengths of 553 and 570 nm, respectively. Images were acquired and processed using ImageJ software. To quantify the tube formation, four fields were chosen randomly and imaged for each sample, and three samples were tested for each culture method. The capillary morphogenesis index (CMI) was defined as the percentage of the areas covered by CM-DiI-positive ECs compared with the top surface area of the Matrigel layer. Statistical Analysis. The values are expressed as means ± standard deviation (SD). Whenever appropriate a two-tailed Student’s t test was used to discern the statistical difference between groups. A probability value (p) of less than 0.05 was considered to be statistically significant.



RESULTS Characterization of Patterned Electrospun Fibrous Mats. In the current study, hepatocytes, ECs and fibroblasts were proposed to be cocultured on patterned fibrous mats. Electrospun PELA fibers have shown advantages as tissue engineering scaffolds, and the inoculation of hydrophilic poly(ethylene glycol) segments into poly(DL-lactide) improved the hydrophilicity of electrospun fibers to create a better environment for cell attachment.29 Hepatocytes containing asialoglycoprotein receptors on the surface can selectively adhere to galactose ligands, inducing the formation of hepatocyte aggregates and a higher level of liver-specific functions.30 Thus, in the current study, blend electrospun fibers of PELA and Lac-PLA were used for hepatocyte loadings, while fibroblasts and ECs were seeded on patterned PELA fibrous mats. As shown in Figure 1a, a photomask containing 1047

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

distribution on fibrous mats was determined by IHC staining after coculture for 7 days. Fluorescence staining with albumin, collagen I, and collagen IV was used to visualize the cocultures of hepatocytes, fibroblasts, and ECs in the patterned regions, respectively. Figure 3a shows the merged fluorescence signals, indicating that hepatocytes, fibroblasts, and ECs were regularly arranged in the patterned strip regions, and no across condition has been found among the three types of cells. To clarify the distribution of cells within the fibrous scaffold, CLSM images were taken by z-stack scanning with a step size of 5 μm. After merging the images in each layer, side views of the merged images were reconstructed to show the vertical distribution of cells in patterned fibrous mats. As shown in Figure 3b, most of the fibroblasts and ECs infiltrated into the patterned regions of fibrous mats, while hepatocytes indicated less significant penetration into and distributed mostly on the patterned fibrous mats. Cell Morphologies on Patterned Fibrous Mats. The cell morphologies of hepatocytes, fibroblasts, and ECs were characterized from IHC staining, the organization of actin filaments and SEM observation. Figure 3c shows the magnified images of cells in patterned regions, indicating that cocultured cells performed biological functions in albumin and ECM secretions. Hepatocytes formed aggregates at the entire hepatocyte region, and fibroblasts and ECs appeared significant proliferation in a stripe pattern. Figure 3d shows SEM images of cells cocultured on patterned fibrous mats. Hepatocytes formed aggregates to maintain their 3D morphologies like in vivo, exhibiting compact polyhedral spheroidal morphology with a rough surface. The hepatocyte spheroids consisted of indistinct individual cells with tight cell−cell contacts. As shown in Figure 3d, both fibroblasts and ECs had a high cell−substrate contact area. Fibroblasts were tightly attached on and spread over multiple fibers and stretched very well. Fibroblast pseudopods indicated that cells aligned along the directionality of electrospun fibers in patterned regions with an elongated morphology. ECs were tightly attached on and spread over multiple fibers, and formed widespread cell membranes on fibrous mat. The F-actin organization of cells cocultured on patterned fibrous mats was evaluated via staining with TRITC-phalloidin. The formation of hepatocyte spheroids led to the rearrangement of F-actin. As shown in Figure 3e, the cytoskeleton underwent a significant rearrangement with actin stress fiber formation, and the distribution of dense F-actin indicated that hepatocyte aggregates were effectively formed on fibrous substrates and had strong cell−cell contacts. Fibroblasts were anchored on fibrous substrate in a bipolar manner and elongated into a single stretch, while most ECs on fibers were round and contained little or no organized filamentous actin. Hepatocyte Viability on Patterned Fibrous Mats. The cell viability was stained with calcein-AM and propidium iodide, and Figure 4a summarizes CLSM images of cells after coculture on patterned scaffolds. Cells seeded on micropatterned fibrous mats showed good viability and indicated no apparent cell death for each type of cells during 15 days of coculture. Fibroblasts and ECs showed a significant proliferation during incubation and hepatocytes migrated toward one another along the aligned fibers to form cell clusters continuously throughout the culture period. Figure 4b shows magnified images of patterned regions for hepatocyte loading after 7 days of culture alone or coculture with fibroblasts and ECs, indicating that

parallel strips of 100 and 200 μm wide and gaps between the strips of 200 μm was designed. Silver was deposited on a glass substrate, and a photolithography process was used to remove silver layer in the exposed area as described previously.19 A patterned silver circuit was obtained on the substrate as a collector, indicating that the patterning features of the photomask were well maintained (Figure 1b). During the electrospinning process, fibers were preferentially deposited on the silver circuit, driven by Coulombic interactions between the positive charges on the fibers and negative charges on the collector.19 As shown in Figure 1c, patterned fibrous mats were obtained with a stripped structure similar to that of the collector, and few fibers were found with gaps between the strips. Figure 1d shows SEM morphologies of patterned fibrous mats, indicating distinct ridge and groove regions. The width of ridge and groove strips was close to that of silver strips and gaps between them, respectively. As shown in Figure 1e, electrospun fibers deposited on the ridge regions indicated an apparent alignment along the direction of silver strips. The diameter and pore size were determined manually from SEM images (Figure 1f) using ImageJ software, indicating an average diameter of 1.23 ± 0.58 μm and pore size of 20.5 ± 7.3 μm. Cell Distribution on Patterned Fibrous Mats. In the current study, patterned scaffolds for cell cocultures were constructed through a precise assembly of patterned fibers. To investigate the feasibility of scaffold construction, fibrous mats were incubated into different fluorescent solutions before assembly. As show in Figure 2, fibrous mats with red, green, and

Figure 2. (a) CLSM images of micropatterned fibers containing rhodamine B (red), (b) FITC (green), and (c) DAPI (blue). (d) Merged CLSM image of micropatterned fibrous scaffolds after precise assembly of fibers containing rhodamine B, FITC, and DAPI by fitting the bulges of one patterned mat nearly into the dents of another mat. Bars represent 500 μm.

blue fluorescence were well patterned for loading hepatocyte, fibroblast, and ECs, respectively. The micropatterned fibrous mats were precisely assembled by fitting the bulges of one patterned mat nearly into the dents of another mat. Figure 2d shows the distribution profiles of patterned fibers, indicating a close contact of patterned fibrous mats with each other. Thus, the assembly of patterned fibers was an efficient method to construct patterned scaffolds for cell coculture with controlled cellular distribution in the respective regions. The cellular 1048

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Figure 3. (a) CLSM image of immunofluorescent staining of collagen I secretion (blue) by fibroblasts, collagen IV (red) by ECs, and albumin (green) by hepatocytes after coculture on micropatterned fibrous scaffolds for 7 days. (b) Image reconstructed by side view of CLSM images taken by z-stack scanning with a step size of 5 μm from the top of a fibrous mat. (c) Magnified CLSM images of immunofluorescent staining of collagen I secretion by fibroblasts, collagen IV by ECs, and albumin by hepatocytes after coculture on micropatterned fibrous scaffolds for 7 days. (d) SEM images of hepatocyte spheroids, fibroblasts, and ECs after coculture on micropatterned fibrous scaffolds for 7 days. (e) CLSM images of F-actin of hepatocytes, fibroblasts, and ECs stained with TRITC-phalloidin and counterstained with DAPI after coculture on micropatterned fibrous scaffolds for 7 days.

both fibroblasts and ECs was about 10-fold higher than that of hepatocytes cultured alone for 15 days. Figure 5c summarizes the cytochrome P-450 expression of hepatocytes cocultured on patterned fibrous mats, indicating no significant decrease during incubation for 15 days. The P-450 expression of hepatocytes cultured alone showed a slight decrease to about 50.4 pmol/ cell/day after 15 days of incubation. The coculture of hepatocytes with either fibroblasts or ECs for 15 days led to about 2.5-fold higher P-450 expression than hepatocytes cultured alone. The P450 expressions of hepatocyte after coculture with both fibroblasts with ECs for 15 days, at about 150 pmol/cell/day, were significantly higher than those of hepatocytes cocultured with either of these cells or cultured alone (p < 0.05). Bile canaliculi are important for native liver to excrete metabolites and toxins from the body.31 In the current study, the biliary excretory function of hepatocyte spheroids was investigated by incubating with FDA, which entered cells via a passive diffusion and was hydrolyzed by intracellular esterases into fluorescein before excretion by bile canaliculi. As shown in Figure 6a, strong fluorescence signals were detected along the border of hepatocytes in all groups, indicating that these hepatocytes maintained their ability to uptake chemicals and efflux bile acid. Hepatocytes cultured alone exhibited subtle concentrated localization of fluorescein in the apical domain between adjacent hepatocytes, while fluorescein in the hepatocyte spheroids after coculture with fibroblasts and ECs was evenly distributed in the intracellular space and developed a structurally intact bile canaliculus system. The fluorescein localized in the intercellular sacs between hepatocytes was

hepatocytes spontaneously formed spheroids on the patterned strips without apparent cell death of inner spheroids. Previous work has shown that the size of aggregates was crucial to hepatocyte functions, and the oxygen diffusion was not rate limiting and central hepatocytes did not become hypoxic when the aggregate diameter was on the order of 100 μm or less.17 Figure 4c shows the size distribution of hepatocyte spheroids after 7 days of coculture on patterned scaffolds. The culture of hepatocytes on patterned fibrous scaffolds led to part of them formed spheroids with the size of around 50 μm. After coculture on patterned fibrous scaffolds, the majority of hepatocytes were incorporated into spheroids with an average diameter of 80−100 μm. Hepatocyte Functionality on Patterned Fibrous Mats. The liver-specific functions were determined on hepatocyte spheroids after coculture with fibroblasts and ECs with respect to albumin secretion, urea synthesis, cytochrome P-450 expression and bile canaliculi formation. As shown in Figure 5a,b, although the synthesized albumin and urea progressively decreased throughout the incubation, cocultured hepatocytes sustained albumin secretion and urea synthesis during 15 days of incubation. Hepatocytes cultured alone rapidly lost their functional characteristics, and the levels of albumin and urea were significantly lower at every time point compared to those of cocultured hepatocytes. The albumin secretion and urea synthesis of hepatocytes cultured alone for 15 days were about 5.4 and 5.2 μg/106 cell/day, respectively. The coculture of hepatocytes with fibroblasts and ECs resulted in significantly higher amount of albumin secretion and urea synthesis (p < 0.05). Particularly, albumin secretion in such a coculture with 1049

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Figure 4. (a) CLSM images of calcein-AM (green) and propidium iodide (red) stained hepatocytes, fibroblasts, and ECs after coculture on micropatterned fibrous scaffolds for 3, 7, and 15 days. Bars represent 200 μm. (b) CLSM images and (c) the size distribution of hepatocyte spheroids after coculture with fibroblasts (Hep-Fib), ECs (Hep-EC), and both fibroblasts and ECs (Hep-Fib-EC) on micropatterned fibrous scaffolds for 7 days, compared with hepatocytes cultured alone (Hep). Bars represent 50 μm.

patterned fibrous mats. As shown in Figure 7b, ECs penetrated into patterned fibrous mats, and a deeper penetration was detected for ECs after coculture with fibroblasts and hepatocytes. The fluorescence intensities of ECs were quantified, and Figure 7c summarizes the densities of ECs after different culture conditions. Significantly higher EC densities were detected for ECs after coculture with fibroblasts and hepatocytes than those of ECs cultured alone or coculture with either fibroblasts or hepatocytes (p < 0.05). EC Functionality on Patterned Fibrous Mats. Patterned electrospun fibrous mats provide the possibility for ECs to invade into the scaffolds and form capillary-like structures.19 Figure 8a shows magnified images of calcein-AM stained ECs after 15 days of coculture, indicating different behaviors of capillary formation. ECs cultured alone did not show a capillary-like structure. ECs formed more capillary structures with branching after coculture with fibroblasts and hepatocytes than those cocultured with either fibroblasts or hepatocytes. The 3D reconstitution of EC images in cell-loaded scaffolds provided further information on the formation of capillary-like structures, as described by Santos et al.34 As shown in Figure 8b, tubular structures formed by ECs after coculture with fibroblasts and hepatocytes were present at different depths along the z-axis. Thus, it demonstrated the abilities of ECs to penetrate into the patterned fibrous scaffolds and to organize into capillary-like structures.

quantified by image processing to compare the extent of biliary excretion of hepatocytes after coculture under different conditions. As shown in Figure 6b, the hepatocyte coculture significantly enhanced the formation of bile canaliculi (p < 0.05), and the coculture of hepatocytes with both fibroblasts and ECs indicated significantly stronger biliary excretory function than other groups (p < 0.05). EC Viability on Patterned Fibrous Mats. CD31 was expressed in regions of EC−EC contact, and the expression and localization of CD31 indicates the maintenance of an endothelial phenotype and the formation of intercellular junctions between ECs.32 In native liver tissues sinusoidal ECs continuously expressed CD31 from the portal space to the centrolobular vein.33 After coculture for 15 days, IHC staining of CD31 was performed to demonstrate the proliferation and penetration of ECs into patterned fibrous scaffolds. Figure 7a shows the staining results of ECs cultured alone or after coculture with fibroblasts and hepatocytes, indicating that ECs in general retained their cobblestone-like morphology. The staining of ECs cultured alone showed to some extent the opening of intercellular contacts, and the cocultured ECs formed an interconnected endothelial network. To better understand the 3D distribution and proliferation of ECs, the cells densities were quantified on 3D projected fluorescent images from CLSM observations by z-stack scanning with a step size of 5 μm. The side views of the fluorescence images were reconstructed to show the vertical distribution of cells in 1050

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Figure 6. (a) CLSM images of FDA-stained hepatocytes after coculture with fibroblasts (Hep-Fib), ECs (Hep-EC), and both fibroblasts and ECs (Hep-Fib-EC) on micropatterned fibrous scaffolds for 7 days, compared with hepatocytes cultured alone (Hep). Bars represent 200 μm. (b) Percent area of excreted fluorescein in intracellular sacs to the total area covered by FDA-containing cells through quantifying CLSM images of hepatocytes cultured alone or after coculture on micropatterned fibrous scaffolds for 7 days (n = 5); *p < 0.05.

Figure 5. (a) Albumin secretion, (b) urea synthesis, and (c) cytochrome P-450 expression of hepatocytes after coculture with fibroblasts (Hep-Fib), ECs (Hep-EC), and both fibroblasts and ECs (Hep-Fib-EC) on micropatterned fibrous scaffolds for 15 days, compared with hepatocytes cultured alone (Hep; n = 4); *p < 0.05.

providing complex liver functions, but hepatocytes lose many of their characteristic features rapidly and generally fail to proliferate after isolated and cultured on traditional culture dishes. For example, the cell border becomes indistinct and the cytoskeleton undergoes rearrangement.36 In vivo liver is a structurally and functionally heterocellular construct constituted by parenchymal and nonparenchymal cells. Therefore, the coculture of hepatocyte indicates advantages in the establishment of cell−cell and cell−ECM communications to maintain hepatocyte phenotype and functions in vitro,8 and among them the interactions between hepatocyte and fibroblasts and ECs play critical roles in the tissue morphogenesis and development.5 For example, ECs participate in metabolic activities of hepatocytes, playing a pivotal role in the balance of lipids, cholesterol and vitamins, and fibroblasts could support the hepatic functions by continuously expressing or regulating soluble factors.37 In addition, fibroblasts are not in physical contact with hepatocytes and separated from hepatocytes in native liver tissues, in which the space of disse is located between hepatocytes and fibroblasts.6 Therefore, in the current study, a novel approach was developed to create heterotypic micro-organoids of hepatic lobules, using micropatterned fibrous scaffolds for coculture of hepatocyte, ECs and fibroblasts. The precise assembly of cell-loaded patterned

Many in vitro assays have been developed to study the capillary tube formation and usually involve the culture of ECs on or within fibrin, collagen, or basement membrane-like Matrigel.35 A Matrigel overlay method was used in the current study to investigate the capabilities of ECs after coculture on patterned scaffolds to form capillary-like networks as described previously.28 As shown in Figure 8c, ECs migrated from the fibrous scaffolds and formed a complex network of capillary tubes in Matrigel, and the tubes could be one or several cells in diameter. CLSM images were processed to quantify the tube formation, and Figure 8d summarizes the results. ECs cultured alone formed capillary-like formations with CMI of 30.8 ± 4.3%. A significantly higher CMI was detected for ECs after coculture with either fibroblasts or hepatocytes (p < 0.05), at 42.6 ± 5.2 and 48.9 ± 5.4%, respectively. ECs cocultured with both fibroblasts and hepatocytes showed abundant formation of capillary-like and the CMI indicated a dramatic increase to 63.4 ± 6.8% compared with other culture methods (p < 0.05).



DISCUSSION There is an increasing demand to establish a liver tissue model in vitro for understanding the mechanisms of liver diseases and for developing new drugs. Hepatocyte is a sole cell source for 1051

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Figure 7. (a) CLSM images of CD31-positive ECs (red), counterstained with DAPI (blue), after coculture with fibroblasts (EC-Fib), hepatocytes (EC-Hep), and both fibroblasts and hepatocytes (EC-Fib-Hep) on micropatterned fibrous scaffolds for 15 days, compared with ECs cultured alone (EC). Bars represent 10 μm. (b) Images reconstructed by side view of CD31 stained images taken by z-stack scanning with a step size of 5 mm from the top of a fibrous mat. (c) Quantitative analysis of the fluorescence intensities of CD31 stained images to show the densities of ECs cultured alone or after coculture on micropatterned fibrous scaffolds (n = 5); *p < 0.05.

Figure 8. (a) CLSM images of calcein-AM and propidium iodide stained ECs after coculture with fibroblasts (EC-Fib), hepatocytes (EC-Hep), and both fibroblasts and hepatocytes (EC-Fib-Hep) on micropatterned fibrous scaffolds for 15 days, compared with ECs cultured alone (EC). White arrows indicate capillary-like structures. Inset shows a magnified image of typical capillary-like structure. Bars represent 200 μm. (b) Reconstructed image after postprocessing CLSM images taken by z-stack scanning with a step size of 5 μm from the top of a fibrous mat to identify the spatial distribution of capillary-like structures. (c) CLSM images of Matrigel-induced capillary morphogenesis after CM-DiI staining on ECs cultured alone or after coculture. Bars represent 50 μm. (d) Capillary morphogenesis index of ECs cultured alone or after coculture, defined as the percentage of the areas covered by CM-DiI-positive ECs compared with the top surface area of the Matrigel layer (n = 3); *p < 0.05.

1052

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

the effect of coculture systems on the EC proliferation and the capabilities of tube-like structure formation was initially investigated. Phalloidin staining of cocultured ECs revealed little or no organized filamentous actin (Figure 3e), indicating a profile similar to mature capillary structure.28 The calcein-AM staining indicated that ECs migrated into patterned scaffolds and formed capillary-like structures with different depths along the z-axis (Figure 8a,b). IHC staining of CD31 revealed significantly higher EC densities after coculture with fibroblasts and hepatocytes than other culture methods (Figure 7). The Matrigel overlay assay showed that ECs was able to form a complex network of capillary tubes, and the quantitative analysis revealed that the capabilities of capillary-like formation for ECs was enhanced by the micropatterned coculture with fibroblasts and hepatocytes (Figure 8d). Therefore, the micropatterned fibrous mats were useful for generating a stable in vitro hepatic coculture model and resembling heterotypic micro-organoids of hepatic lobules, with enhanced hepatocyte functions, EC proliferation, and capabilities of tube-like structure formation.

electrospun mats was proposed to help both hepatocytes in the maintenance of their phenotypic morphologies and ECs in the formation of tube-like structures. Also of note is that fibroblasts NIH3T3 and human umbilical vein ECs were chosen in the current study, because they were much easier to obtain and culture than primary cells isolated form livers. It is supposed that the micropatterned coculture model should achieve a better performance of both hepatocytes and ECs by the use of liver-derived stromal cells, due to a more precise imitation of cell sources and distributions in hepatic lobules. The coculture on micropatterned substrates not only created heterotypic micro-organoids with precisely ordered multiple cell types, but also significantly improved liver-specific functions as compared to unorganized cocultures.15 Khetani et al. fabricated micropatterned areas of ECM proteins for selectively loading hepatocytes, followed by seeding fibroblasts into the remaining bare areas. The micropatterned coculture of hepatocytes and fibroblasts on multiwell plates led to significant higher albumin secretion and urea synthesis than hepatocytes cultured alone.38 However, compared with multiwell plates and PDMS stencils consisting patterned islands or circular holes,39 significant cell infiltration was detected in the micropatterned fibrous scaffolds (Figure 3b), which should be beneficial for the formation of hepatocyte spheroids and EC tubes. Although the coculture of hepatocytes with ECs and fibroblasts did not play an important role in the formation of hepatocyte spheroids (Figure 4c), dramatic enhancement and maintenance of liver specific functions was determined with significantly higher albumin secretion, urea synthesis, and P450 expression compared with coculture of hepatocytes with either ECs or fibroblasts (Figure 5). The establishment of cell polarity and functional activity is one of the key features of hepatocyte culture, such as bile canaliculi formation between hepatocytes in concert with changes in cytoskeleton distribution.25 F-actin staining indicated that the actin filaments of hepatocytes cocultured on patterned fibrous mats were reorganized to the cell−cell contact regions (Figure 3e), resembling the F-actin distribution in vivo.40 FDA excretion analysis indicated that hepatic ultrastructures, bile canaliculi were well-developed, and hepatocytes formed multiple junctions within spheroids (Figure 6), suggesting a liver-mimicking structure in the preservation of polarized phenotype of cocultured hepatocytes.41 Compared with hepatocytes cultured alone, the coculture of hepatocytes with fibroblasts and ECs was able to actively form tight junctions among hepatocytes and secrete into bile canaliculi (Figure 6), maintaining healthy morphologies of bile canaliculi. Therefore, the organization of cocultured hepatic ultrastructures could facilitate the diffusive delivery of culture media to the center of hepatic spheroids, which is an important parameter to affect cell morphology, function, and physiological responsiveness.42 Hepatic lobule, the basic unit of liver tissue, enables highly differentiated functions of hepatocytes by abundant blood supplies from liver microvessels, termed liver sinusoids, which also exhibit fenestrae and function as a scavenger system in the liver by removing waste macromolecules. In addition, most approaches in tissue engineering have relied on postimplantation neovascularization from the host, but it is clearly insufficient for large and metabolically demanding organs to meet the demand for blood vessel ingrowth.43 It is thus necessary in the reconstruction of liver structures in vitro to induce capillary morphogenesis of ECs.44 In the current study,



CONCLUSIONS The coculture of hepatocytes, fibroblasts, and ECs was established with controlled cellular distributions through precise assembly of cell-loaded micropatterned fibrous mats. Hepatocytes formed spheroids with significant rearrangement of F-actin, while ECs formed widespread cell membranes on fibrous mat and contained little or no organized filamentous actin. Compared with hepatocytes cultured alone, the coculture of hepatocytes with either fibroblasts or ECs led to significantly higher albumin secretion, urea synthesis, cytochrome P-450 expression and formation of bile canaliculi. These hepatic functions were dramatically improved by coculture of hepatocytes with both fibroblasts and ECs. Significantly higher EC densities, deeper penetration into patterned scaffolds, and stronger capabilities of capillary formation were determined for ECs after coculture with fibroblasts and hepatocytes compared with those cultured alone or after coculture with either fibroblasts or hepatocytes. The coculture of hepatocytes, fibroblasts, and ECs on micropatterned fibrous mats resembled heterotypic micro-organoids of hepatic lobules, with enhanced hepatic functions, EC proliferation, and capillary formation capabilities.



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Tel.: +8628-87634068. Fax: +8628-87634649. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by National Natural Science Foundation of China (51073130 and 21274117), Specialized Research Fund for the Doctoral Program of Higher Education (20120184110004), and National Scientific and Technical Supporting Programs (2012BAI17B06).



REFERENCES

(1) Hewitt, N. J.; Lechón, M. J.; Houston, J. B.; Hallifax, D.; Brown, H. S.; Maurel, P.; Kenna, J. G.; Gustavsson, L.; Lohmann, C.; Skonberg, C.; Guillouzo, A.; Tuschl, G.; Li, A. P.; LeCluyse, E.; 1053

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054

Biomacromolecules

Article

Groothuis, G. M.; Hengstler, J. G. Drug Metab. Rev. 2007, 39, 159− 234. (2) Cho, C. H.; Berthiaume, F.; Tilles, A. W.; Yarmush, M. L. Biotechnol. Bioeng. 2008, 101, 345−356. (3) Qiu, Y.; Mao, Z. W.; Zhao, Y. M.; Zhang, J. C.; Guo, Q.; Gou, Z. R.; Gao, C. Y. Macromol. Res. 2012, 20, 283−291. (4) Oda, H.; Yoshida, Y.; Kawamura, A.; Kakinuma, A. Biochem. Pharmacol. 2008, 75, 1209−1217. (5) Braet, F.; Wisse, E. Comp. Hepatol. 2002, 1, 1. (6) Hui, E. E.; Bhatia, S. N. Proc. Natl. Acad. Sci. U.S.A. 2007, 104, 5722−5726. (7) Kim, M.; Lee, J. Y.; Jones, C. N.; Revzin, A.; Tae, G. Biomaterials 2010, 31, 3596−3603. (8) Kim, K.; Ohashi, K.; Utoh, R.; Kano, K.; Okano, T. Biomaterials 2012, 33, 1406−1413. (9) Wong, S. F.; No, D. Y.; Choi, Y. Y.; Kim, D. S.; Chung, B. G.; Lee, S. H. Biomaterials 2011, 32, 8087−8096. (10) Nahmias, Y.; Schwartz, R. E.; Hu, W. S.; Verfaillie, C. M.; Odde, D. J. Tissue Eng. 2006, 12, 1627−1638. (11) Yamada, M.; Utoh, R.; Ohashi, K.; Tatsumi, K.; Yamato, M.; Okano, T.; Seki, M. Biomaterials 2012, 33, 8304−8315. (12) Chen, A. A.; Thomas, D. K.; Ong, L. L.; Schwartz, R. E.; Golub, T. R.; Bhatia, S. N. Proc. Natl. Acad. Sci. U.S.A. 2011, 108, 11842− 11847. (13) Kang, I. K.; Kim, G. J.; Kwon, O. H.; Ito, Y. Biomaterials 2004, 25, 4225−4232. (14) Fukuda, J.; Nakazawa, K. Biomicrofluidics 2011, 5, 022205. (15) Nakazawa, K.; Shinmura, Y.; Higuchi, A.; Sakai, Y. J. Bioprocess. Biotech. 2011, S3, 002. (16) Greiner, A.; Wendorff, J. H. Angew. Chem., Int. Ed. 2007, 46, 5670−5703. (17) Feng, Z. Q.; Chu, X. H.; Huang, N. P.; Leach, M. K.; Wang, G.; Wang, Y. C; Ding, Y. T.; Gu, Z. Z. Biomaterials 2010, 31, 3604−3612. (18) Chua, K. N.; Lim, W. S.; Zhang, P.; Lu, H.; Wen, J.; Ramakrishna, S.; Leong, K. W.; Mao, H. Q. Biomaterials 2005, 26, 2537−2547. (19) Liu, Y. W.; Zhang, L.; Li, H. N.; Yan, S. L.; Yu, J. S.; Weng, J.; Li, X. H. Langmuir 2012, 28, 17134−17142. (20) Li, X. H.; Zhang, Y. H.; Jia, W. X.; Deng, X. M.; Huang, Z. T. J. Controlled Release 2000, 68, 41−52. (21) Kim, S. H.; Han, Y. K.; Ahn, K. D.; Kim, Y. H.; Chan, T. Makromol. Chem. 1993, 194, 3229−3236. (22) Seglen, P. O. Methods Cell Biol. 1976, 13, 29−83. (23) Vaquette, C.; Cooper-White, J. J. Acta Biomater. 2011, 7, 2544− 2557. (24) Yushkevich, P. A.; Piven, J.; Hazlett, H. C.; Smith, R. G.; Ho, S.; Gee, J. C.; Gerig, G. Neuroimage 2006, 31, 1116−1128. (25) Du, Y.; Han, R.; Wen, F.; Ngsansan, S.; Xia, L.; Wohland, T.; Leo, H. L.; Yu, H. Biomaterials 2008, 29, 290−301. (26) Liang, H.; Hou, H.; Yi, W.; Yang, G.; Gu, C.; Lau, W. B.; Gao, E.; Ma, X.; Lu, Z.; Wei, X.; Pei, J.; Yi, D. Eur. Heart J. 2013, 34, 1681− 1690. (27) Zheng, Y.; Henderson, P. W.; Choi, N. W.; Bonassar, L. J.; Spector, J. A.; Stroock, A. D. Biomaterials 2011, 3, 5391−5401. (28) Connolly, J. O.; Simpson, N.; Hewlett, L.; Hall, A. Mol. Biol. Cell 2002, 13, 2474−2785. (29) Cui, W. G.; Li, X. H.; Zhou, S. B.; Weng, J. Polym. Degrad. Stab. 2008, 93, 731−738. (30) Yin, C.; Ying, L.; Zhang, P. C.; Zhuo, R. X.; Kang, E. T.; Leong, K. W.; Mao, H. Q. J. Biomed. Mater. Res. A 2003, 67, 1093−1104. (31) Liu, X.; Chism, J. P.; LeCluyse, E. L.; Brouwer, K. R.; Brouwer, K. L. Drug Metab. Dispos. 1999, 27, 637−644. (32) Dejana, E. J. Clin. Invest. 1997, 100, 7−10. (33) Pusztaszeri, M. P.; Seelentag, W.; Bosman, F. T. J. Histochem. Cytochem. 2006, 54, 385−395. (34) Santos, M. I.; Tuzlakoglu, K.; Fuchs, S.; Gomes, M. E.; Peters, K.; Unger, R. E.; Piskin, E.; R. Reis, L.; Kirkpatrick, C. J. Biomaterials 2008, 29, 4306−4313.

(35) Grant, D. S.; Tashiro, K.; Segui-Real, B.; Yamada, Y.; Martin, G. R.; Kleinman, H. K.. Cell 1989, 58, 933−943. (36) Bhatia, S. N.; Balis, U. J.; Yarmush, M. L.; Toner, M. FASEB J. 1999, 13, 1883−1900. (37) Chia, S. M.; Lin, P. C.; Yu, H. Biotechnol. Bioeng. 2005, 89, 565− 573. (38) Khetani, S. R.; Bhatia, S. N. Nat. Biotechnol. 2008, 26, 120−126. (39) Zinchenko, Y. S.; Schrum, L. W.; Clemens, M.; Coger, R. N. Tissue Eng. 2006, 12, 751−761. (40) Hamilton, G. A.; Jolley, S. L.; Gilbert, D.; Coon, D. J.; Barros, S.; LeCluyse, E. L. Cell Tissue Res. 2001, 306, 85−99. (41) Sudo, R.; Ikeda, S.; Sugimoto, S.; Harada, K.; Hirata, K.; Tanishita, K.; Mochizuki, Y.; Mitaka, T. J. Cell Physiol. 2004, 199, 252−261. (42) Berthiaume, F.; Moghe, P. V.; Toner, M.; Yarmush, M. L. FASEB J. 1996, 10, 1471−1484. (43) Kaully, T.; Kaufman-Francis, K.; Lesman, A.; Levenberg, S. Tissue Eng. B 2009, 15, 159−169. (44) Yang, J.; Yamato, M.; Shimizu, T.; Sekine, H.; Ohashi, K.; Kanzaki, M.; Ohki, T.; Nishida, K.; Okano, T. Biomaterials 2007, 28, 5033−5043.

1054

dx.doi.org/10.1021/bm401926k | Biomacromolecules 2014, 15, 1044−1054