- Email: [email protected]
Galactosylated electrospun membranes for hepatocyte sandwich culture
Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
Galactosylated electrospun membranes for hepatocyte sandwich culture Hsiu-Wen Chien a , Juin-Yih Lai b , Wei-Bor Tsai a,∗ a b
Department of Chemical Engineering, National Taiwan University, No. 1, Roosevelt Rd., Sec. 4, Taipei 106, Taiwan R&D Center for Membrane Technology and Department of Chemical Engineering, Chung Yuan Christian University, Chungli, Taoyuan, Taiwan
a r t i c l e
i n f o
Article history: Received 11 October 2013 Received in revised form 7 January 2014 Accepted 24 January 2014 Available online 4 February 2014 Keywords: Electrospun membrane Polyelectrolyte multilayer deposition Sandwich culture Hepatocytes
a b s t r a c t In this work, we developed a galactocylated electrospun polyurethane membrane for sandwich culture of hepatocyte sandwich culture. The electrospun fibrous membranes were bio-functionalized with galactose molecules by a UV-crosslinked layer-by-layer polyelectrolyte multilayer deposition technique. The galactosylated electrospun membranes were employed as a top support membrane for the sandwich culture of HepG2/C3A cells on a collagen substrate. Our results demonstrate that HepG2/C3A cells covered by the galactosylated PU membranes form multi-cellular aggregates and lead to improved albumin secretion ability compared to the control membranes (unmodified PU or poly(ethylene imine)-modified PU). Our study reveals the potential of galactosylated electrospun membranes in the application of liver tissue engineering and the regeneration of liver-tissue substitutes. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Hepatocytes, which are responsible for the major functions of the liver, are commonly isolated and cultured in vitro, e.g. for hepatic tissue engineering studies and drug screening for hepatotoxicology [1]. However, hepatocytes in two-dimensional (2D) culture tend to lose their differentiated functions. Therefore, much effort has been devoted to the development of three-dimensional (3D) microenvironments that simulate the in vivo extracellular matrix (ECM) in order to retain the essential hepatic functions. For example, sandwich culture of hepatocytes between two collagen layers is a well-established in vitro model for re-establishing hepatic polarity and maintaining differentiated functions of hepatocytes [2,3]. Different from the traditional 2D culture, the top support for the sandwich culture serves as a matrix for cell attachment beside the base substrate, creating 3D cellular structure and polarities. Therefore, sandwich cultures provide a microenvironment similar to that found in the liver and promote hepatocyte cell–cell and cell–ECM interactions [2]. Thus, sandwich cultures serve as a tool for investigation of liver physiology, drug metabolism/toxicity testing [3], and hepatocyte-based bioreactors [4]. However, applications of the ECM-based sandwich culture may be limited by low mass transfer owing to the top gelled ECM layer,
∗ Corresponding author. Tel.: +886 2 3366 3996; fax: +886 2 2362 3040. E-mail address: [email protected] (W.-B. Tsai). 0927-7765/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2014.01.040
batch-to-batch variation in the ECM compositions and uncontrollable ECM coatings [5]. To address these issues, synthetic porous membranes have been proposed as an alternative for collagen gels in hepatocyte sandwich culture. Du et al. established a synthetic sandwich culture by overlaying hepatocytes that are cultured on a galactosylated polyethylene terephthalate (PET) film with a porous PET track-etched membrane [5]. They demonstrated that the synthetic sandwich culture could achieve better mass transfer through the top porous PET support in comparison to the collagen gel. The hepatocytes in the sandwich culture with synthetic membranes exhibited a similar process of hepatic polarity formation, better cell–cell interaction and improved differentiated functions compared to hepatocytes in the collagen sandwich culture. These results suggest that better mass transfer through the top support benefits hepatocyte sandwich cultures. Since ECM proteins provide essential biological signals for maintaining cell physiology, we evaluated retaining a collagen substrate layer, while replacing the top-layer collagen gel by a synthetic membrane for hepatocyte sandwich culture. Compared to commercial membrane products, electrospun fibrous membranes appear to be a better choice as the top support of the sandwich culture. Electrospun fibers, fabricated from electrically driven polymer jets, have been employed in various fields such as biomedical engineering and environmental engineering [6,7]. The morphological similarity between the ECM and electrospun fibers suggests an outstanding potential for cell/tissue engineering applications [8,9]. The advantages of electrospinning also include easy fabrication, modulation of fibrous structures by adjusting electrospinning
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
conditions, and a wide variety of materials that can be used. Furthermore, by controlling the collection time, the fiber coverage density of the electrospun fibrous membrane can be controlled, which in turn modulates the mass transfer through the membrane. Therefore, we have evaluated this type of membranes as the top support of the sandwich culture of hepatocytes. In this study, polyurethane (PU) fibrous membranes were fabricated by electrospinning and then biofunctionalized with galactose molecules via a layer-by-layer (LBL) polyelectrolyte deposition technique (Fig. 1). This technique is based on alternating adsorption of positively and negatively charged polyelectrolytes, resulting in a thin coating on a substrate material [10]. One of the advantages of this surface modification technique is that it is not restricted to specific types, sizes and shapes of substrate materials. We previously deposited a tri-layer polyelectrolyte film of poly(ethylene imine) (PEI), poly(acrylic acid)-g-azide (PAA-g-AZ) and PEI on polymeric substrates. After exposure to UV irradiation, the three-layer polyelectrolyte film was crosslinked via a phenylazide-based reaction and covalently conjugated on the substrate material [11]. In this study, galactose molecules were first conjugated onto PEI, and then adsorbed as the outmost layer of a tri-layer polyelectrolyte film on the outer layer of PU electrospun fibers. The galactosylated PU membranes were then evaluated for hepatocyte sandwich culture. 2. Experimental 2.1. Materials Polyurethane (PU, PellethaneTM , 2103-80AE) was obtained from Dow Chemical Company, USA. Poly(ethylene imine) (PEI, Mw ∼ 750 kDa, cat. no. P3143) was received from Sigma-Aldrich. Collagen was purified from bovine skin according to a previously described procedure [12]. All other chemicals were purchased from Sigma-Aldrich unless specified otherwise. Poly(acrylic acid-g-azidoaniline), abbreviated as PAA-g-AZ, was synthesized using a previously published procedure [13]. The content of AZ in PAA-g-AZ, estimated from the ratio of the peaks of the azidophenyl protons at 6.5–7.5 ppm and the methylene protons of the polymer main chain at 1.3–2.5 ppm in the 1 H-nuclear magnetic resonance (1 H-NMR, Avance-500 Hz, Bruker) spectrum was 6.0 mol%. Poly(ethylene imine)-g-galactose (PEI-g-Gal) was synthesized according to a previous protocol [14]. The coupling percentage of lactobionic acids to PEI, estimated from the ratio of the peaks of the protons in the methylene groups and the methane groups of the -galactose at 3.5 to 4.0 ppm and the methylene protons of the PEI main chain at 2.5 to 3.0 ppm in the 1 H-NMR spectrum, was 37.7 mol%. A human hepatoblastoma cell line, HepG2/C3A, was received from Food Industry Research and Development Institute (Hsinchu, Taiwan). The cell culture medium contained ␣MEM (HyClone, USA) supplemented with 10% fetal bovine serum (JRH, Australia), 1% non-essential amino acids (GIBCO, USA), 1.0 mM sodium pyruvate, 2 mg/mL NaHCO3 , 0.5% of fungizone (GIBCO, USA), 0.25% gentamycin (GIBCO, USA) and 0.679% -mercaptoethanol. Phosphate-buffered saline (PBS) was prepared as 137 mM NaCl, 2.7 mM KCl, 10 mM Na2 HPO4 , 1.8 mM KH2 PO4 , pH 7.4. 2.2. Fabrication and surface galactosylation of electrospun PU membranes PU fibrous membranes were fabricated by a conventional electrospinning technique. Briefly, 10% (w/v) PU solution in DMF/THF (1/1, v/v) was electrospun at 11 kV and 0.5 mL/h through a 16G blunt-end needle toward an aluminum collector at a distance of 10 cm. The samples used in sandwich culture as top supports were
577
collected for 1, 3 or 5 min, while the PU membranes for characterization and 2D cell culture were collected for 1 h. Galactose molecules were conjugated on electrospun PU membranes via LBL polyelectrolyte multilayer deposition (Fig. 1). Prior to LBL deposition, PU membranes were treated with radio-frequency glow discharge oxygen plasma (50 W, 0.07 torr, 2 min, 23.5 sccm) to introduce negatively charged functional groups on the surface. The membranes were then immersed in a PEI solution (1 mg/mL) for 30 min, followed by rinses with deionized water. Next, the membranes were immersed in PAA-g-AZ solution (1 mg/mL) in the dark for 30 min. Finally, after rinsing again with deionized water, the PU membranes were immersed in PEI or PEI-g-Gal solution (1 mg/mL) for 30 min, followed by rinsing with deionized water. After drying in air, the membranes were exposed to UV for 1 min to achieve crosslinking. The untreated PU membranes and those modified by PEI/PAA-g-AZ/PEI and PEI/PAA-g-AZ/PEI-g-Gal were abbreviated as PU, PEI and Gal, respectively. 2.3. Characterization of electrospun PU membranes The morphology of electrospun PU fibers was investigated using scanning electron spectroscopy (SEM, JSM-5310, JEOL, Japan). The average diameters of fibers were determined from more than 100 PU electrospun fibers that were randomly selected from the SEM images. The fiber coverage densities of the membranes collected for 1, 3 and 5 min were determined from optical microscopic images of the samples using NIH ImageJ software. The software determined the fiber-covered area and the void area. Fiber coverage density (%) was defined as the ratio of fiber-covered area/total surface area × 100%. For water contact angle measurements, electrospun PU membranes were collected for 1 h. The wettability of the membranes was determined using a goniometer (FTA-125, First Ten Angstroms) with deionized water by the static sessile drop method. Ten droplets (10 L) were measured for each sample. 2.4. Cell culture on electrospun membranes PU, PEI and Gal membranes were sterilized by UV exposure for 1 h prior to cell seeding. HepG2/C3A cells were then seeded on the membranes at a density of 5 × 104 cells/cm2 and cultured for 3 days. For SEM analysis, the samples were fixed by 0.25% glutaraldehyde in PBS for 10 min and 2.5% glutaraldehyde for another hour. The samples were then dehydrated in a graded series of ethanol: 30%, 50%, 70%, 80%, 90%, 95% and 100%, followed by CO2 critical point drying. After gold sputtering, cells were imaged by SEM. 2.5. Sandwich culture with electrospun membranes For 3D culture, the cells were sandwiched between a membrane and collagen substrate (Fig. 1B). Collagen substrates were prepared by adding 35 L ice-cold, neutralized 1 mg/mL collagen solution onto a 10-mm glass coverslip kept at 37 ◦ C for overnight gelation. HepG2/C3A cells were then seeded at 5 × 104 cells/cm2 on the collagen substrate. After 9 h of incubation, the cells were covered by an electrospun fiber membrane for further culture in a 37 ◦ C humidified incubator. The culture media were changed every day. Cell viability was examined by live/dead staining. Briefly, after 2 days of culture, cells were stained with 1 mg/mL fluorescin diacetate (green for live cells) and 1 mg/mL propidium iodide (red for dead cells) for 15 min at 37 ◦ C in the dark. After rinses with PBS, the samples were observed under a confocal microscope (Nikon TE2000-U, Japan).
578
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
Fig. 1. Schematic illustration of hepatocyte sandwich culture between a galactosylated electrospun PU membrane and a collagen substrate.
Cell number was determined by a lactate dehydrogenase (LDH) method [15]. Briefly, cultured cells were lysed by 0.1% (v/v) Triton X-100 (1000 L/well) in PBS at 37 ◦ C for 30 min. 50 L of the cell lysate was mixed with an equal volume of -nicotinamide adenine dinucleotide (NAD) solution (12 mg/mL sodium l-lactate, 1 mg/mL NAD, 0.9 mg/mL diaphorase, 0.1% (w/v) bovine serum albumin, 4 mg/mL sucrose and 0.067 mg/mL iodonitrotetrazolium chloride in PBS) and then incubated at 37 ◦ C for 30 min. The reaction was stopped by addition of 50 L of oxamate solution (16 mg/mL sodium oxamate in PBS). The absorbance was measured at 490 nm using a microtiter plate reader (EL800, Bio-Tek, USA). The cell number was determined from a calibration curve generated by plotting OD vs. cell concentrations. Albumin synthesis of HepG2/C3A cells was detected by immunofluorescent staining. After cell culture, the cells were fixed by 4% paraformaldehyde for 10 min and then permeabilized with 0.1% Triton X-100 in PBS for 20 min. Next, the cells were blocked in 2% bovine serum albumin (BSA) in PBS for 30 min and then incubated with sheep anti-human albumin IgG (1:300; Serotec, USA) at 37 ◦ C for 60 min. After being rinsed with PBST (0.05% Tween 20 in PBS) three times, the samples were incubated with FITC-conjugated anti-sheep IgG antibody (1:300; Sigma, USA) at 37 ◦ C for 60 min. Cell nuclei were counter-stained with 4 ,6 -diamidino-2-phenylindole (DAPI, Invitrogen, USA) for 30 min. The fluorescent images were captured under a confocal microscope. Competitive ELISA was used to quantify the amount of albumin in culture media according to a previously described procedure [16]. 0.6 g/mL albumin was coated on 96-well plates (100 L/well), followed by overnight incubation at 4 ◦ C. The plates were rinsed with PBS twice and then blocked with 2% BSA in PBS at room temperature for 1 h. After rinsing, the albumin-coated plates were ready to use. The culture media were mixed with an equal volume of sheep anti-human albumin IgG in PBS (1:5000) and incubated at room temperature for 15 min. The sampleantibody mixture was then added into the albumin-coated plates (100 L/well) and then incubated at 37 ◦ C for 1 h. After rinsing with PBS, 100 L/well of HRP-conjugated anti-sheep IgG antibody in PBS (1:2500; Santa Cruz) was added and incubated at 37 ◦ C for another hour. After rinsing with 0.5% Tween 20 in PBS, the TMB solution (0.04% 3,3 ,5,5 -tetramethyl-benzidine dihydrochloride (Sigma) in acetic acid) was added (100 l/well) and incubated on a shaker at room temperature for 10 min. Finally, the reaction was stopped by adding 50 L/well of 2 N H2 SO4 . The OD values were read at 450 nm with a microtiter plate reader. A standard curve for the competitive ELISA was obtained with a series of albumin solutions with known concentrations.
2.6. Statistical analysis The data was reported as means ± standard deviation (SD). The statistical analyses between different groups were determined using Student’s t-test. Probabilities of p ≤ 0.05 were considered as significant difference. All statistical analyses were performed using GraphPad Instat 3.0 program (GraphPad Software, USA). 3. Results and discussion 3.1. Characterization of PU electrospun membranes The morphology of the electrospun PU fibers, displayed in the SEM image in Fig. 2, indicates that PU electrospun fibers were fabricated successfully under the given conditions. The diameters of electrospun PU fibers ranged from 0.4 to 1.2 m with an average of 800 ± 162 nm. In this study, LBL polyelectrolyte multilayer deposition combined with UV crosslinking was applied to achieve surface functionalization with galactose molecules. We previously used an equivalent LBL technique to create a functionalized surface for conjugation of anti-fouling materials [11,17]. Surface modification of the electrospun PU membranes was revealed by water contact angle measurement. The water contact angle of the unmodified PU membrane was 104.8◦ ± 1.6◦ . After LBL deposition with PEI as the topmost layer, the water contact angle of the PU membrane
Fig. 2. Representative SEM image of electrospun PU fibers. Scale bar = 10 m. The distribution of fiber diameters was analyzed from SEM images.
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
579
Fig. 3. SEM images of HepG2/C3A cells after 2D cultured on PU, PEI and Gal electrospun membranes for 3 days. The bottom images are enlarged sections of the upper images. Scale bar = 50 m.
was decreased to 67.2◦ ± 3.7◦ (p < 0.001 vs. PU). The PU membrane conjugated with PEI-g-Gal exhibited a water contact angle of 86.9◦ ± 1.3◦ (p < 0.001 vs. PU and PEI). The difference in water contact angles indicated that PU membranes were successfully modified by LBL deposition. 3.2. Cell adhesion on galactosylated electrospun membranes The adhesion of HepG2/C3A cells on the electrospun fibrous membranes after 3 days of culture was evaluated next. Comparing the three types of electrospun membranes, fewer cells attached and spread on the unmodified PU membrane (Fig. 3A), while more cells that resided on the PEI and Gal membranes displayed fully spread morphology (Fig. 3B and C). However, the cells tended to form larger cell aggregates on the Gal membrane (Fig. 3C). This result indicated that the modification of PU membranes by LBL polyelectrolyte multilayer films modulates hepatocyte adhesion and the formation of aggregates. 3.3. Sandwich culture of hepatocytes with overlaid electrospun membranes Overlay of an electrospun membrane on hepatocytes may affect the transport of nutrition, oxygen and wastes, which is thus detrimental to hepatocyte growth. Since the densities of fibers were expected to affect substance transport, different densities of electrospun fibers were prepared by controlling the collection time (1, 3 and 5 min). The surface coverage of fibers was 13%, 24% and 39% for the collection time of 1, 3 and 5 min, respectively (Fig. 4A), which were defined as low, middle and high fiber densities, respectively. HepG2/C3A cells were cultured between a Gal membrane and a collagen substrate for 2 days. Most of the cells were found to be alive (green) under all types of Gal membranes regardless of fiber densities (Fig. 4B). However, cellular albumin secretion, an important function of hepatocytes, was affected by the fiber density. HepG2/C3A cells covered by the middle-density membrane expressed the strongest albumin secretion compared to the highand low-density membranes (Fig. 4C). Therefore, middle-density membranes were used in subsequent experiments. We next compared the sandwich culture of HepG2/C3A cells that were overlaid with PU, PEI or Gal membranes in regard to cell
morphology, proliferation and albumin synthesis. After 2 days of culture, many cells exhibited a spread morphology on the 2D collagen substrate without any membrane coverage (Fig. 5A), while cells covered by PU or PEI membranes displayed an even more spread and polygonal morphology. On the other hand, cells covered by Gal membranes formed aggregates. After 2 days of culture with a seeding density of 5 × 104 cells/cm2 , the cell number on the collagen substrate was 8.1 × 104 cells/cm2 , while the cell number in sandwich culture was 6.8 × 104 , 4.2 × 104 and 3.9 × 104 cells/cm2 for PU, PEI and Gal membranes, respectively (Fig. 5B). The cell number observed under a PU-based membrane was not as high as the cell number observed on the collagen substrate without overlaying a membrane, indicating that cell growth may be restricted when cells were sandwich cultured under a PEI or Gal membrane. Albumin secretion by the cells that were covered by the PU or PEI membrane was almost undetectable by immunofluorescent staining (Fig. 5C), while the cells covered by the Gal membrane displayed significant albumin secretion. Furthermore, the fluorescence intensity observed in Gal-sandwich culture was even stronger than that observed on the 2D collagen substrate. The result of albumin secretion by the cultured cells into the culture medium, that was determined by competitive ELISA, indicates that the HepG2/C3A cells that were overlaid by a Gal membrane possessed better albumin-secretion ability than the cells on the 2D collagen substrate and TCPS (Fig. 5D).
4. Discussion We demonstrate here that electrospun fibrous membranes can act as an advantageous top support for hepatocyte sandwich culture. Electrospun fiber membranes can easily be fabricated with desired densities and conjugated with biomolecules. In this study, electrospun membranes were modified by a UV-crosslinked LBL deposition technique. We showed that PU electrospun membranes could be easily conjugated with galactose moieties in this way. Since this technique is not limited to specific materials and surface morphologies, this method may also be applied to other types of electrospun fibers. The overlay of an electrospun membrane restrained cell proliferation, as shown in Fig. 5B, which may be due to the hindrance to the
580
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
Fig. 4. (A) Optical microscopy images (100×) of electrospun PU fibrous membranes collected at different times. The fibrous density, analyzed from microscopic images by using NIH Image J software, is defined as the area of fibers/total area of the image × 100%. After 2 days of sandwich culture under the Gal membranes, the cells were stained for indicating (B) live (green)/dead (red) and (C) albumin secretion (green). The cell nuclei were counter-stained with DAPI (blue). Scale bar = 50 m. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. HepG2/C3A cells were sandwich-cultured under a PU, PEI or Gal electrospun membrane for 2 days. The cell culture on the 2D collagen substrate was used as a control. (A) Microscopic images of cell morphology and (B) cell numbers. Value = mean ± SD, n = 4. * p < 0.05 and ** p < 0.01 vs. Gal. (C) Immunofluorescent images (green) and (D) amount of albumin secretion after 3 days of culture. Cell nuclei were counter-stained with DAPI (blue). Scale bar = 50 m. SC: HepG2/C3A cells were sandwich cultured between a Gal membrane and a collagen substrate. MC: cells were cultured on the 2D collagen substrate. A tissue culture plate (TCPS) was used as a control. Value = mean ± SD, n = 4. * p < 0.05 and *** p < 0.001 vs. TCPS. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
transport of nutrients and waste, but at the same time the albumin secretion was enhanced by the Gal membrane (Fig. 5C). Enhanced differentiated function enabled by coverage with a top support has been reported previously [5], and may be due to the increase in cellmatrix contacts. The fact that cell viability was not affected by the coverage of the fibrous membrane suggests that the mass transport was not severely hindered by the membrane coverage regardless of fiber densities. However, albumin production was higher for hepatocytes that were covered by the middle-density membrane. The fiber density affects not only mass transport but also cell-substrate contacts. Therefore, our results suggest that proper cell-substrate interactions enhance cell functions. Hepatocytes contain the asialoglycoprotein receptor (ASGPR) that specifically recognizes galactose. Although ASGPR is not a receptor that mediates the adhesion of hepatocytes, many studies reveal that galactose-carrying polymers mediate the adhesion of hepatocytes [14,18,19]. Furthermore, the hepatocytes cultured on galactose-carrying polymers display round morphology, which triggers the formation of multicellular aggregates, leading to enhanced differentiated cell function [20]. Therefore, galactose conjugation has been a popular surface modification strategy for hepatocyte-contacting biomedical devices. Du et al. previously showed that hepatocytes cultured on a 2D galactosylated substrate formed 3D cell spheroids [21]. The authors found that the hepatocytes exhibited strong cell–cell interaction on the galactosylated substrate during cell culture, indicated by the expression of E-cadherin, while the cells on the collagen substrate displayed increased cell-substrate contacts, indicated by increasing phosphorylation of focal adhesion kinase. It has been speculated that the morphogenesis of hepatocytes is governed by the relative strengths of cell–cell and cell–substrate contacts [22]. The formation of multi-cellular spheroids on galactosylated substrates has been attributed to weak interactions between galactose and ASGPR [5]. Therefore, the cell–cell interaction of hepatocytes on galactosylated substrates may dominate the cell-substrate interaction, leading to spheroid cell aggregates. In our studies hepatocytes were cultured on a collagen substrate, which supports the formation 2D hepatocyte monolayer [21]. Overlaying a Gal membrane resulted in the formation of cell aggregates (Fig. 5A), while the cells fully spread by covering an unmodified PU membrane, which supports hepatocyte attachment poorer than the Gal membrane. Our result suggests that the formation of hepatocyte spheroids on galactosylated substrates may not be explained sorely by weak cell-substrate interaction. It is possible that the galactose–ASGPR interaction stimulates cell–cell interaction via strong E-cadherin expression. Our results indicate that the Gal membrane enhanced albumin secretion of hepatocytes in comparison with unmodified and PEImodified PU membranes. The enhancement in albumin expression may come from the formation of cell aggregates. It has been shown that the spheroid structure is critical in maintaining the viability and differentiated functions of hepatocytes in vitro [23,24]. The hepatocytes in spheroids display considerable cell–cell interaction and tight junctions, mimicking the morphology and organization of the hepatocytes in the native liver lobule [25,26]. However, biochemical responses that may be induced by the galactose–ASGPR interaction could not be excluded as stimulation for the better functional properties of the hepatocytes. Nevertheless, up to now there
581
is little knowledge of the downstream signaling pathways induced by the galactose–ASGPR interaction for us to explain our results. 5. Conclusions We have developed a galactosylated electrospun PU membrane for application in the sandwich culture of HepG2/C3A cells. Cells sandwiched by a collagen substrate and the Gal membrane formed multi-cellular aggregate morphology and displayed improved albumin secretion compared with cells on a 2D collagen substrate. The galactose-carrying electrospun membranes are expected to contribute to progress in liver tissue engineering, e.g. in new bioartificial liver-assist devices or in the generation of livertissue substitutes. Acknowledgments The authors thank the National Science Council, Taiwan (97-2628-E-002-028-MY2) for financial support. The authors appreciate the kindness of Dr. Helmut Thissen (CSIRO, Australia) for proofreading the manuscript. References [1] K. Lee, D. Kaplan (Eds.), Tissue Engineering II: Basics of Tissue Engineering and Tissue Applications, Springer, Berlin, 2007, p. 309. [2] F. Berthiaume, P.V. Moghe, M. Toner, M.L. Yarmush, FASEB J. 10 (1996) 1471. [3] T. De Bruyn, S. Chatterjee, S. Fattah, J. Keemink, J. Nicolai, P. Augustijns, P. Annaert, Expert Opin. Drug Metab. Toxicol. 9 (2013) 589. [4] K. Taguchi, M. Matsushita, M. Takahashi, J. Uchino, Artif. Organs 20 (1996) 178. [5] Y. Du, R. Han, F. Wen, S.N.S. San, L. Xia, T. Wohland, H.L. Leo, H. Yu, Biomaterials 29 (2008) 290. [6] Z.M. Huang, Y.Z. Zhang, M. Kotaki, S. Ramakrishna, Compos. Sci. Technol. 63 (2003) 2223. [7] Y. Zhang, C.T. Lim, S. Ramakrishna, Z.M. Huang, J. Mater. Sci. Mater. Med. 16 (2005) 933. [8] W.J. Li, C.T. Laurencin, E.J. Caterson, R.S. Tuan, F.K. Ko, J. Biomed. Mater. Res. 60 (2002) 613. [9] M.M. Stevens, J.H. George, Science 310 (2005) 1135. [10] G. Decher, Science 277 (1997) 1232. [11] W.H. Kuo, M.J. Wang, H.W. Chien, T.C. Wei, C. Lee, W.B. Tsai, Biomacromolecules 12 (2011) 4348. [12] P.Y. Wang, H.H. Chow, W.B. Tsai, H.W. Fang, J. Biomater. Appl. 23 (2009) 347. [13] H.W. Chien, T.Y. Chang, W.B. Tsai, Biomaterials 30 (2009) 2209. [14] H.W. Chien, W.B. Tsai, Acta Biomater. 8 (2012) 3678. [15] J.M. Grunkemeier, W.B. Tsai, M.R. Alexander, D.G. Castner, T.A. Horbett, J. Biomed. Mater. Res. 51 (2000) 669. [16] W.B. Tsai, J.H. Lin, Acta Biomater. 5 (2009) 1442. [17] H.W. Chien, C.C. Tsai, W.B. Tsai, M.J. Wang, W.H. Kuo, T.C. Wei, S.T. Huang, Colloids Surf., B 107 (2013) 152. [18] C.S. Cho, S.J. Seo, I.K. Park, S.H. Kim, T.H. Kim, T. Hoshiba, T. Akaike, Biomaterials 27 (2006) 576. [19] K. Kobayashi, H. Sumitomo, Y. Ina, Polym. J. 17 (1985) 567. [20] S. Tobe, Y. Takei, K. Kobayashi, T. Akaike, Biochem. Biophys. Res. Commun. 184 (1992) 225. [21] Y. Du, R. Han, S. Ng, J. Ni, W. Sun, T. Wohland, S.H. Ong, L. Kuleshova, H. Yu, Tissue Eng. 13 (2007) 1455. [22] E. Martz, H.M. Phillips, M.S. Steinberg, J. Cell Sci. 16 (1974) 401. [23] F.J. Wu, J.R. Friend, C.C. Hsiao, M.J. Zilliox, W.J. Ko, F.B. Cerra, W.S. Hu, Biotechnol. Bioeng. 50 (1996) 404. [24] R. Glicklis, L. Shapiro, R. Agbaria, J.C. Merchuk, S. Cohen, Biotechnol. Bioeng. 67 (2000) 344. [25] L.K. Hansen, C.C. Hsiao, J.R. Friend, F.J. Wu, G.A. Bridge, R.P. Remmel, F.B. Cerra, W.S. Hu, Tissue Eng. 4 (1998) 65. [26] J. Landry, D. Bernier, C. Ouellet, R. Goyette, N. Marceau, J. Cell Biol. 101 (1985) 914.
Contents lists available at ScienceDirect
Colloids and Surfaces B: Biointerfaces journal homepage: www.elsevier.com/locate/colsurfb
Galactosylated electrospun membranes for hepatocyte sandwich culture Hsiu-Wen Chien a , Juin-Yih Lai b , Wei-Bor Tsai a,∗ a b
Department of Chemical Engineering, National Taiwan University, No. 1, Roosevelt Rd., Sec. 4, Taipei 106, Taiwan R&D Center for Membrane Technology and Department of Chemical Engineering, Chung Yuan Christian University, Chungli, Taoyuan, Taiwan
a r t i c l e
i n f o
Article history: Received 11 October 2013 Received in revised form 7 January 2014 Accepted 24 January 2014 Available online 4 February 2014 Keywords: Electrospun membrane Polyelectrolyte multilayer deposition Sandwich culture Hepatocytes
a b s t r a c t In this work, we developed a galactocylated electrospun polyurethane membrane for sandwich culture of hepatocyte sandwich culture. The electrospun fibrous membranes were bio-functionalized with galactose molecules by a UV-crosslinked layer-by-layer polyelectrolyte multilayer deposition technique. The galactosylated electrospun membranes were employed as a top support membrane for the sandwich culture of HepG2/C3A cells on a collagen substrate. Our results demonstrate that HepG2/C3A cells covered by the galactosylated PU membranes form multi-cellular aggregates and lead to improved albumin secretion ability compared to the control membranes (unmodified PU or poly(ethylene imine)-modified PU). Our study reveals the potential of galactosylated electrospun membranes in the application of liver tissue engineering and the regeneration of liver-tissue substitutes. © 2014 Elsevier B.V. All rights reserved.
1. Introduction Hepatocytes, which are responsible for the major functions of the liver, are commonly isolated and cultured in vitro, e.g. for hepatic tissue engineering studies and drug screening for hepatotoxicology [1]. However, hepatocytes in two-dimensional (2D) culture tend to lose their differentiated functions. Therefore, much effort has been devoted to the development of three-dimensional (3D) microenvironments that simulate the in vivo extracellular matrix (ECM) in order to retain the essential hepatic functions. For example, sandwich culture of hepatocytes between two collagen layers is a well-established in vitro model for re-establishing hepatic polarity and maintaining differentiated functions of hepatocytes [2,3]. Different from the traditional 2D culture, the top support for the sandwich culture serves as a matrix for cell attachment beside the base substrate, creating 3D cellular structure and polarities. Therefore, sandwich cultures provide a microenvironment similar to that found in the liver and promote hepatocyte cell–cell and cell–ECM interactions [2]. Thus, sandwich cultures serve as a tool for investigation of liver physiology, drug metabolism/toxicity testing [3], and hepatocyte-based bioreactors [4]. However, applications of the ECM-based sandwich culture may be limited by low mass transfer owing to the top gelled ECM layer,
∗ Corresponding author. Tel.: +886 2 3366 3996; fax: +886 2 2362 3040. E-mail address: [email protected] (W.-B. Tsai). 0927-7765/$ – see front matter © 2014 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.colsurfb.2014.01.040
batch-to-batch variation in the ECM compositions and uncontrollable ECM coatings [5]. To address these issues, synthetic porous membranes have been proposed as an alternative for collagen gels in hepatocyte sandwich culture. Du et al. established a synthetic sandwich culture by overlaying hepatocytes that are cultured on a galactosylated polyethylene terephthalate (PET) film with a porous PET track-etched membrane [5]. They demonstrated that the synthetic sandwich culture could achieve better mass transfer through the top porous PET support in comparison to the collagen gel. The hepatocytes in the sandwich culture with synthetic membranes exhibited a similar process of hepatic polarity formation, better cell–cell interaction and improved differentiated functions compared to hepatocytes in the collagen sandwich culture. These results suggest that better mass transfer through the top support benefits hepatocyte sandwich cultures. Since ECM proteins provide essential biological signals for maintaining cell physiology, we evaluated retaining a collagen substrate layer, while replacing the top-layer collagen gel by a synthetic membrane for hepatocyte sandwich culture. Compared to commercial membrane products, electrospun fibrous membranes appear to be a better choice as the top support of the sandwich culture. Electrospun fibers, fabricated from electrically driven polymer jets, have been employed in various fields such as biomedical engineering and environmental engineering [6,7]. The morphological similarity between the ECM and electrospun fibers suggests an outstanding potential for cell/tissue engineering applications [8,9]. The advantages of electrospinning also include easy fabrication, modulation of fibrous structures by adjusting electrospinning
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
conditions, and a wide variety of materials that can be used. Furthermore, by controlling the collection time, the fiber coverage density of the electrospun fibrous membrane can be controlled, which in turn modulates the mass transfer through the membrane. Therefore, we have evaluated this type of membranes as the top support of the sandwich culture of hepatocytes. In this study, polyurethane (PU) fibrous membranes were fabricated by electrospinning and then biofunctionalized with galactose molecules via a layer-by-layer (LBL) polyelectrolyte deposition technique (Fig. 1). This technique is based on alternating adsorption of positively and negatively charged polyelectrolytes, resulting in a thin coating on a substrate material [10]. One of the advantages of this surface modification technique is that it is not restricted to specific types, sizes and shapes of substrate materials. We previously deposited a tri-layer polyelectrolyte film of poly(ethylene imine) (PEI), poly(acrylic acid)-g-azide (PAA-g-AZ) and PEI on polymeric substrates. After exposure to UV irradiation, the three-layer polyelectrolyte film was crosslinked via a phenylazide-based reaction and covalently conjugated on the substrate material [11]. In this study, galactose molecules were first conjugated onto PEI, and then adsorbed as the outmost layer of a tri-layer polyelectrolyte film on the outer layer of PU electrospun fibers. The galactosylated PU membranes were then evaluated for hepatocyte sandwich culture. 2. Experimental 2.1. Materials Polyurethane (PU, PellethaneTM , 2103-80AE) was obtained from Dow Chemical Company, USA. Poly(ethylene imine) (PEI, Mw ∼ 750 kDa, cat. no. P3143) was received from Sigma-Aldrich. Collagen was purified from bovine skin according to a previously described procedure [12]. All other chemicals were purchased from Sigma-Aldrich unless specified otherwise. Poly(acrylic acid-g-azidoaniline), abbreviated as PAA-g-AZ, was synthesized using a previously published procedure [13]. The content of AZ in PAA-g-AZ, estimated from the ratio of the peaks of the azidophenyl protons at 6.5–7.5 ppm and the methylene protons of the polymer main chain at 1.3–2.5 ppm in the 1 H-nuclear magnetic resonance (1 H-NMR, Avance-500 Hz, Bruker) spectrum was 6.0 mol%. Poly(ethylene imine)-g-galactose (PEI-g-Gal) was synthesized according to a previous protocol [14]. The coupling percentage of lactobionic acids to PEI, estimated from the ratio of the peaks of the protons in the methylene groups and the methane groups of the -galactose at 3.5 to 4.0 ppm and the methylene protons of the PEI main chain at 2.5 to 3.0 ppm in the 1 H-NMR spectrum, was 37.7 mol%. A human hepatoblastoma cell line, HepG2/C3A, was received from Food Industry Research and Development Institute (Hsinchu, Taiwan). The cell culture medium contained ␣MEM (HyClone, USA) supplemented with 10% fetal bovine serum (JRH, Australia), 1% non-essential amino acids (GIBCO, USA), 1.0 mM sodium pyruvate, 2 mg/mL NaHCO3 , 0.5% of fungizone (GIBCO, USA), 0.25% gentamycin (GIBCO, USA) and 0.679% -mercaptoethanol. Phosphate-buffered saline (PBS) was prepared as 137 mM NaCl, 2.7 mM KCl, 10 mM Na2 HPO4 , 1.8 mM KH2 PO4 , pH 7.4. 2.2. Fabrication and surface galactosylation of electrospun PU membranes PU fibrous membranes were fabricated by a conventional electrospinning technique. Briefly, 10% (w/v) PU solution in DMF/THF (1/1, v/v) was electrospun at 11 kV and 0.5 mL/h through a 16G blunt-end needle toward an aluminum collector at a distance of 10 cm. The samples used in sandwich culture as top supports were
577
collected for 1, 3 or 5 min, while the PU membranes for characterization and 2D cell culture were collected for 1 h. Galactose molecules were conjugated on electrospun PU membranes via LBL polyelectrolyte multilayer deposition (Fig. 1). Prior to LBL deposition, PU membranes were treated with radio-frequency glow discharge oxygen plasma (50 W, 0.07 torr, 2 min, 23.5 sccm) to introduce negatively charged functional groups on the surface. The membranes were then immersed in a PEI solution (1 mg/mL) for 30 min, followed by rinses with deionized water. Next, the membranes were immersed in PAA-g-AZ solution (1 mg/mL) in the dark for 30 min. Finally, after rinsing again with deionized water, the PU membranes were immersed in PEI or PEI-g-Gal solution (1 mg/mL) for 30 min, followed by rinsing with deionized water. After drying in air, the membranes were exposed to UV for 1 min to achieve crosslinking. The untreated PU membranes and those modified by PEI/PAA-g-AZ/PEI and PEI/PAA-g-AZ/PEI-g-Gal were abbreviated as PU, PEI and Gal, respectively. 2.3. Characterization of electrospun PU membranes The morphology of electrospun PU fibers was investigated using scanning electron spectroscopy (SEM, JSM-5310, JEOL, Japan). The average diameters of fibers were determined from more than 100 PU electrospun fibers that were randomly selected from the SEM images. The fiber coverage densities of the membranes collected for 1, 3 and 5 min were determined from optical microscopic images of the samples using NIH ImageJ software. The software determined the fiber-covered area and the void area. Fiber coverage density (%) was defined as the ratio of fiber-covered area/total surface area × 100%. For water contact angle measurements, electrospun PU membranes were collected for 1 h. The wettability of the membranes was determined using a goniometer (FTA-125, First Ten Angstroms) with deionized water by the static sessile drop method. Ten droplets (10 L) were measured for each sample. 2.4. Cell culture on electrospun membranes PU, PEI and Gal membranes were sterilized by UV exposure for 1 h prior to cell seeding. HepG2/C3A cells were then seeded on the membranes at a density of 5 × 104 cells/cm2 and cultured for 3 days. For SEM analysis, the samples were fixed by 0.25% glutaraldehyde in PBS for 10 min and 2.5% glutaraldehyde for another hour. The samples were then dehydrated in a graded series of ethanol: 30%, 50%, 70%, 80%, 90%, 95% and 100%, followed by CO2 critical point drying. After gold sputtering, cells were imaged by SEM. 2.5. Sandwich culture with electrospun membranes For 3D culture, the cells were sandwiched between a membrane and collagen substrate (Fig. 1B). Collagen substrates were prepared by adding 35 L ice-cold, neutralized 1 mg/mL collagen solution onto a 10-mm glass coverslip kept at 37 ◦ C for overnight gelation. HepG2/C3A cells were then seeded at 5 × 104 cells/cm2 on the collagen substrate. After 9 h of incubation, the cells were covered by an electrospun fiber membrane for further culture in a 37 ◦ C humidified incubator. The culture media were changed every day. Cell viability was examined by live/dead staining. Briefly, after 2 days of culture, cells were stained with 1 mg/mL fluorescin diacetate (green for live cells) and 1 mg/mL propidium iodide (red for dead cells) for 15 min at 37 ◦ C in the dark. After rinses with PBS, the samples were observed under a confocal microscope (Nikon TE2000-U, Japan).
578
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
Fig. 1. Schematic illustration of hepatocyte sandwich culture between a galactosylated electrospun PU membrane and a collagen substrate.
Cell number was determined by a lactate dehydrogenase (LDH) method [15]. Briefly, cultured cells were lysed by 0.1% (v/v) Triton X-100 (1000 L/well) in PBS at 37 ◦ C for 30 min. 50 L of the cell lysate was mixed with an equal volume of -nicotinamide adenine dinucleotide (NAD) solution (12 mg/mL sodium l-lactate, 1 mg/mL NAD, 0.9 mg/mL diaphorase, 0.1% (w/v) bovine serum albumin, 4 mg/mL sucrose and 0.067 mg/mL iodonitrotetrazolium chloride in PBS) and then incubated at 37 ◦ C for 30 min. The reaction was stopped by addition of 50 L of oxamate solution (16 mg/mL sodium oxamate in PBS). The absorbance was measured at 490 nm using a microtiter plate reader (EL800, Bio-Tek, USA). The cell number was determined from a calibration curve generated by plotting OD vs. cell concentrations. Albumin synthesis of HepG2/C3A cells was detected by immunofluorescent staining. After cell culture, the cells were fixed by 4% paraformaldehyde for 10 min and then permeabilized with 0.1% Triton X-100 in PBS for 20 min. Next, the cells were blocked in 2% bovine serum albumin (BSA) in PBS for 30 min and then incubated with sheep anti-human albumin IgG (1:300; Serotec, USA) at 37 ◦ C for 60 min. After being rinsed with PBST (0.05% Tween 20 in PBS) three times, the samples were incubated with FITC-conjugated anti-sheep IgG antibody (1:300; Sigma, USA) at 37 ◦ C for 60 min. Cell nuclei were counter-stained with 4 ,6 -diamidino-2-phenylindole (DAPI, Invitrogen, USA) for 30 min. The fluorescent images were captured under a confocal microscope. Competitive ELISA was used to quantify the amount of albumin in culture media according to a previously described procedure [16]. 0.6 g/mL albumin was coated on 96-well plates (100 L/well), followed by overnight incubation at 4 ◦ C. The plates were rinsed with PBS twice and then blocked with 2% BSA in PBS at room temperature for 1 h. After rinsing, the albumin-coated plates were ready to use. The culture media were mixed with an equal volume of sheep anti-human albumin IgG in PBS (1:5000) and incubated at room temperature for 15 min. The sampleantibody mixture was then added into the albumin-coated plates (100 L/well) and then incubated at 37 ◦ C for 1 h. After rinsing with PBS, 100 L/well of HRP-conjugated anti-sheep IgG antibody in PBS (1:2500; Santa Cruz) was added and incubated at 37 ◦ C for another hour. After rinsing with 0.5% Tween 20 in PBS, the TMB solution (0.04% 3,3 ,5,5 -tetramethyl-benzidine dihydrochloride (Sigma) in acetic acid) was added (100 l/well) and incubated on a shaker at room temperature for 10 min. Finally, the reaction was stopped by adding 50 L/well of 2 N H2 SO4 . The OD values were read at 450 nm with a microtiter plate reader. A standard curve for the competitive ELISA was obtained with a series of albumin solutions with known concentrations.
2.6. Statistical analysis The data was reported as means ± standard deviation (SD). The statistical analyses between different groups were determined using Student’s t-test. Probabilities of p ≤ 0.05 were considered as significant difference. All statistical analyses were performed using GraphPad Instat 3.0 program (GraphPad Software, USA). 3. Results and discussion 3.1. Characterization of PU electrospun membranes The morphology of the electrospun PU fibers, displayed in the SEM image in Fig. 2, indicates that PU electrospun fibers were fabricated successfully under the given conditions. The diameters of electrospun PU fibers ranged from 0.4 to 1.2 m with an average of 800 ± 162 nm. In this study, LBL polyelectrolyte multilayer deposition combined with UV crosslinking was applied to achieve surface functionalization with galactose molecules. We previously used an equivalent LBL technique to create a functionalized surface for conjugation of anti-fouling materials [11,17]. Surface modification of the electrospun PU membranes was revealed by water contact angle measurement. The water contact angle of the unmodified PU membrane was 104.8◦ ± 1.6◦ . After LBL deposition with PEI as the topmost layer, the water contact angle of the PU membrane
Fig. 2. Representative SEM image of electrospun PU fibers. Scale bar = 10 m. The distribution of fiber diameters was analyzed from SEM images.
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
579
Fig. 3. SEM images of HepG2/C3A cells after 2D cultured on PU, PEI and Gal electrospun membranes for 3 days. The bottom images are enlarged sections of the upper images. Scale bar = 50 m.
was decreased to 67.2◦ ± 3.7◦ (p < 0.001 vs. PU). The PU membrane conjugated with PEI-g-Gal exhibited a water contact angle of 86.9◦ ± 1.3◦ (p < 0.001 vs. PU and PEI). The difference in water contact angles indicated that PU membranes were successfully modified by LBL deposition. 3.2. Cell adhesion on galactosylated electrospun membranes The adhesion of HepG2/C3A cells on the electrospun fibrous membranes after 3 days of culture was evaluated next. Comparing the three types of electrospun membranes, fewer cells attached and spread on the unmodified PU membrane (Fig. 3A), while more cells that resided on the PEI and Gal membranes displayed fully spread morphology (Fig. 3B and C). However, the cells tended to form larger cell aggregates on the Gal membrane (Fig. 3C). This result indicated that the modification of PU membranes by LBL polyelectrolyte multilayer films modulates hepatocyte adhesion and the formation of aggregates. 3.3. Sandwich culture of hepatocytes with overlaid electrospun membranes Overlay of an electrospun membrane on hepatocytes may affect the transport of nutrition, oxygen and wastes, which is thus detrimental to hepatocyte growth. Since the densities of fibers were expected to affect substance transport, different densities of electrospun fibers were prepared by controlling the collection time (1, 3 and 5 min). The surface coverage of fibers was 13%, 24% and 39% for the collection time of 1, 3 and 5 min, respectively (Fig. 4A), which were defined as low, middle and high fiber densities, respectively. HepG2/C3A cells were cultured between a Gal membrane and a collagen substrate for 2 days. Most of the cells were found to be alive (green) under all types of Gal membranes regardless of fiber densities (Fig. 4B). However, cellular albumin secretion, an important function of hepatocytes, was affected by the fiber density. HepG2/C3A cells covered by the middle-density membrane expressed the strongest albumin secretion compared to the highand low-density membranes (Fig. 4C). Therefore, middle-density membranes were used in subsequent experiments. We next compared the sandwich culture of HepG2/C3A cells that were overlaid with PU, PEI or Gal membranes in regard to cell
morphology, proliferation and albumin synthesis. After 2 days of culture, many cells exhibited a spread morphology on the 2D collagen substrate without any membrane coverage (Fig. 5A), while cells covered by PU or PEI membranes displayed an even more spread and polygonal morphology. On the other hand, cells covered by Gal membranes formed aggregates. After 2 days of culture with a seeding density of 5 × 104 cells/cm2 , the cell number on the collagen substrate was 8.1 × 104 cells/cm2 , while the cell number in sandwich culture was 6.8 × 104 , 4.2 × 104 and 3.9 × 104 cells/cm2 for PU, PEI and Gal membranes, respectively (Fig. 5B). The cell number observed under a PU-based membrane was not as high as the cell number observed on the collagen substrate without overlaying a membrane, indicating that cell growth may be restricted when cells were sandwich cultured under a PEI or Gal membrane. Albumin secretion by the cells that were covered by the PU or PEI membrane was almost undetectable by immunofluorescent staining (Fig. 5C), while the cells covered by the Gal membrane displayed significant albumin secretion. Furthermore, the fluorescence intensity observed in Gal-sandwich culture was even stronger than that observed on the 2D collagen substrate. The result of albumin secretion by the cultured cells into the culture medium, that was determined by competitive ELISA, indicates that the HepG2/C3A cells that were overlaid by a Gal membrane possessed better albumin-secretion ability than the cells on the 2D collagen substrate and TCPS (Fig. 5D).
4. Discussion We demonstrate here that electrospun fibrous membranes can act as an advantageous top support for hepatocyte sandwich culture. Electrospun fiber membranes can easily be fabricated with desired densities and conjugated with biomolecules. In this study, electrospun membranes were modified by a UV-crosslinked LBL deposition technique. We showed that PU electrospun membranes could be easily conjugated with galactose moieties in this way. Since this technique is not limited to specific materials and surface morphologies, this method may also be applied to other types of electrospun fibers. The overlay of an electrospun membrane restrained cell proliferation, as shown in Fig. 5B, which may be due to the hindrance to the
580
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
Fig. 4. (A) Optical microscopy images (100×) of electrospun PU fibrous membranes collected at different times. The fibrous density, analyzed from microscopic images by using NIH Image J software, is defined as the area of fibers/total area of the image × 100%. After 2 days of sandwich culture under the Gal membranes, the cells were stained for indicating (B) live (green)/dead (red) and (C) albumin secretion (green). The cell nuclei were counter-stained with DAPI (blue). Scale bar = 50 m. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
Fig. 5. HepG2/C3A cells were sandwich-cultured under a PU, PEI or Gal electrospun membrane for 2 days. The cell culture on the 2D collagen substrate was used as a control. (A) Microscopic images of cell morphology and (B) cell numbers. Value = mean ± SD, n = 4. * p < 0.05 and ** p < 0.01 vs. Gal. (C) Immunofluorescent images (green) and (D) amount of albumin secretion after 3 days of culture. Cell nuclei were counter-stained with DAPI (blue). Scale bar = 50 m. SC: HepG2/C3A cells were sandwich cultured between a Gal membrane and a collagen substrate. MC: cells were cultured on the 2D collagen substrate. A tissue culture plate (TCPS) was used as a control. Value = mean ± SD, n = 4. * p < 0.05 and *** p < 0.001 vs. TCPS. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)
H.-W. Chien et al. / Colloids and Surfaces B: Biointerfaces 116 (2014) 576–581
transport of nutrients and waste, but at the same time the albumin secretion was enhanced by the Gal membrane (Fig. 5C). Enhanced differentiated function enabled by coverage with a top support has been reported previously [5], and may be due to the increase in cellmatrix contacts. The fact that cell viability was not affected by the coverage of the fibrous membrane suggests that the mass transport was not severely hindered by the membrane coverage regardless of fiber densities. However, albumin production was higher for hepatocytes that were covered by the middle-density membrane. The fiber density affects not only mass transport but also cell-substrate contacts. Therefore, our results suggest that proper cell-substrate interactions enhance cell functions. Hepatocytes contain the asialoglycoprotein receptor (ASGPR) that specifically recognizes galactose. Although ASGPR is not a receptor that mediates the adhesion of hepatocytes, many studies reveal that galactose-carrying polymers mediate the adhesion of hepatocytes [14,18,19]. Furthermore, the hepatocytes cultured on galactose-carrying polymers display round morphology, which triggers the formation of multicellular aggregates, leading to enhanced differentiated cell function [20]. Therefore, galactose conjugation has been a popular surface modification strategy for hepatocyte-contacting biomedical devices. Du et al. previously showed that hepatocytes cultured on a 2D galactosylated substrate formed 3D cell spheroids [21]. The authors found that the hepatocytes exhibited strong cell–cell interaction on the galactosylated substrate during cell culture, indicated by the expression of E-cadherin, while the cells on the collagen substrate displayed increased cell-substrate contacts, indicated by increasing phosphorylation of focal adhesion kinase. It has been speculated that the morphogenesis of hepatocytes is governed by the relative strengths of cell–cell and cell–substrate contacts [22]. The formation of multi-cellular spheroids on galactosylated substrates has been attributed to weak interactions between galactose and ASGPR [5]. Therefore, the cell–cell interaction of hepatocytes on galactosylated substrates may dominate the cell-substrate interaction, leading to spheroid cell aggregates. In our studies hepatocytes were cultured on a collagen substrate, which supports the formation 2D hepatocyte monolayer [21]. Overlaying a Gal membrane resulted in the formation of cell aggregates (Fig. 5A), while the cells fully spread by covering an unmodified PU membrane, which supports hepatocyte attachment poorer than the Gal membrane. Our result suggests that the formation of hepatocyte spheroids on galactosylated substrates may not be explained sorely by weak cell-substrate interaction. It is possible that the galactose–ASGPR interaction stimulates cell–cell interaction via strong E-cadherin expression. Our results indicate that the Gal membrane enhanced albumin secretion of hepatocytes in comparison with unmodified and PEImodified PU membranes. The enhancement in albumin expression may come from the formation of cell aggregates. It has been shown that the spheroid structure is critical in maintaining the viability and differentiated functions of hepatocytes in vitro [23,24]. The hepatocytes in spheroids display considerable cell–cell interaction and tight junctions, mimicking the morphology and organization of the hepatocytes in the native liver lobule [25,26]. However, biochemical responses that may be induced by the galactose–ASGPR interaction could not be excluded as stimulation for the better functional properties of the hepatocytes. Nevertheless, up to now there
581
is little knowledge of the downstream signaling pathways induced by the galactose–ASGPR interaction for us to explain our results. 5. Conclusions We have developed a galactosylated electrospun PU membrane for application in the sandwich culture of HepG2/C3A cells. Cells sandwiched by a collagen substrate and the Gal membrane formed multi-cellular aggregate morphology and displayed improved albumin secretion compared with cells on a 2D collagen substrate. The galactose-carrying electrospun membranes are expected to contribute to progress in liver tissue engineering, e.g. in new bioartificial liver-assist devices or in the generation of livertissue substitutes. Acknowledgments The authors thank the National Science Council, Taiwan (97-2628-E-002-028-MY2) for financial support. The authors appreciate the kindness of Dr. Helmut Thissen (CSIRO, Australia) for proofreading the manuscript. References [1] K. Lee, D. Kaplan (Eds.), Tissue Engineering II: Basics of Tissue Engineering and Tissue Applications, Springer, Berlin, 2007, p. 309. [2] F. Berthiaume, P.V. Moghe, M. Toner, M.L. Yarmush, FASEB J. 10 (1996) 1471. [3] T. De Bruyn, S. Chatterjee, S. Fattah, J. Keemink, J. Nicolai, P. Augustijns, P. Annaert, Expert Opin. Drug Metab. Toxicol. 9 (2013) 589. [4] K. Taguchi, M. Matsushita, M. Takahashi, J. Uchino, Artif. Organs 20 (1996) 178. [5] Y. Du, R. Han, F. Wen, S.N.S. San, L. Xia, T. Wohland, H.L. Leo, H. Yu, Biomaterials 29 (2008) 290. [6] Z.M. Huang, Y.Z. Zhang, M. Kotaki, S. Ramakrishna, Compos. Sci. Technol. 63 (2003) 2223. [7] Y. Zhang, C.T. Lim, S. Ramakrishna, Z.M. Huang, J. Mater. Sci. Mater. Med. 16 (2005) 933. [8] W.J. Li, C.T. Laurencin, E.J. Caterson, R.S. Tuan, F.K. Ko, J. Biomed. Mater. Res. 60 (2002) 613. [9] M.M. Stevens, J.H. George, Science 310 (2005) 1135. [10] G. Decher, Science 277 (1997) 1232. [11] W.H. Kuo, M.J. Wang, H.W. Chien, T.C. Wei, C. Lee, W.B. Tsai, Biomacromolecules 12 (2011) 4348. [12] P.Y. Wang, H.H. Chow, W.B. Tsai, H.W. Fang, J. Biomater. Appl. 23 (2009) 347. [13] H.W. Chien, T.Y. Chang, W.B. Tsai, Biomaterials 30 (2009) 2209. [14] H.W. Chien, W.B. Tsai, Acta Biomater. 8 (2012) 3678. [15] J.M. Grunkemeier, W.B. Tsai, M.R. Alexander, D.G. Castner, T.A. Horbett, J. Biomed. Mater. Res. 51 (2000) 669. [16] W.B. Tsai, J.H. Lin, Acta Biomater. 5 (2009) 1442. [17] H.W. Chien, C.C. Tsai, W.B. Tsai, M.J. Wang, W.H. Kuo, T.C. Wei, S.T. Huang, Colloids Surf., B 107 (2013) 152. [18] C.S. Cho, S.J. Seo, I.K. Park, S.H. Kim, T.H. Kim, T. Hoshiba, T. Akaike, Biomaterials 27 (2006) 576. [19] K. Kobayashi, H. Sumitomo, Y. Ina, Polym. J. 17 (1985) 567. [20] S. Tobe, Y. Takei, K. Kobayashi, T. Akaike, Biochem. Biophys. Res. Commun. 184 (1992) 225. [21] Y. Du, R. Han, S. Ng, J. Ni, W. Sun, T. Wohland, S.H. Ong, L. Kuleshova, H. Yu, Tissue Eng. 13 (2007) 1455. [22] E. Martz, H.M. Phillips, M.S. Steinberg, J. Cell Sci. 16 (1974) 401. [23] F.J. Wu, J.R. Friend, C.C. Hsiao, M.J. Zilliox, W.J. Ko, F.B. Cerra, W.S. Hu, Biotechnol. Bioeng. 50 (1996) 404. [24] R. Glicklis, L. Shapiro, R. Agbaria, J.C. Merchuk, S. Cohen, Biotechnol. Bioeng. 67 (2000) 344. [25] L.K. Hansen, C.C. Hsiao, J.R. Friend, F.J. Wu, G.A. Bridge, R.P. Remmel, F.B. Cerra, W.S. Hu, Tissue Eng. 4 (1998) 65. [26] J. Landry, D. Bernier, C. Ouellet, R. Goyette, N. Marceau, J. Cell Biol. 101 (1985) 914.