Journal of Bioscience and Bioengineering VOL. 120 No. 5, 577e581, 2015 www.elsevier.com/locate/jbiosc
Fabrication of uniformly cell-laden porous scaffolds using a gas-in-liquid templating technique Takayuki Takei,1, * Ryuta Aokawa,2 Takamasa Shigemitsu,1 Koei Kawakami,2 and Masahiro Yoshida1 Department of Chemical Engineering, Graduate School of Science and Engineering, Kagoshima University, 1-21-40 Korimoto, Kagoshima 890-0065, Japan1 and Department of Chemical Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0385, Japan2 Received 19 December 2014; accepted 25 March 2015 Available online 23 April 2015
Design of porous scaffolds in tissue engineering field was challenging. Uniform immobilization of cells in the scaffolds with high porosity was essential for homogeneous tissue formation. The present study was aimed at fabricating uniformly cell-laden porous scaffolds with porosity >74% using the gas-in-liquid foam templating technique. To this end, we used gelatin, microbial transglutaminase and argon gas as a scaffold material, cross-linker of the protein and porogen of scaffold, respectively. We confirmed that a porosity of >74% could be achieved by increasing the gas volume delivered to a gelatin solution. Pore size in the scaffold could be controlled by stirring speed, stirring time and the pore size of the filter through which the gas passed. The foaming technique enabled us to uniformly immobilize a human hepatoblastoma cell line in scaffold. Engraftment efficiency of the cell line entrapped within the scaffold in nude mice was higher than that of cells in free-form. These results showed that the uniformly cell-laden porous scaffolds were promising for tissue engineering. Ó 2015, The Society for Biotechnology, Japan. All rights reserved. [Key words: Scaffold; Gas foaming; Gelatin; Alginate; Tissue engineering]
The objective of tissue engineering is to develop biological substitutes for damaged tissues but the design of scaffolds in this field is challenging. Requirements of scaffolds are to promote cellular adhesion, proliferation, and differentiation and to enhance engraftment of transplanted cells at targeted sites. In a general procedure to fabricate the biological substitutes, cells are first seeded in scaffolds at low density and transplanted into bodies. Penetration of blood vessels into the scaffolds is then promoted to supply sufficient oxygen and nutrients to the cells, resulting in cell proliferation in the scaffolds and following creation of the biological substitutes with high cell density (1). Scaffolds are classified into two types: porous sponges and non-porous hydrogels. The former are superior to the latter with regard to accelerated permeability of oxygen/nutrients and cell proliferation because of their porous structure (2). Especially, enhancement of permeability of oxygen and nutrients are important to keep cells alive until blood vessels penetrate into the scaffolds. Uniform immobilization of cells in the porous scaffolds is essential for homogeneous tissue formation. Many methods for fabricating porous scaffolds have been described, such as solvent casting/particulate leaching (3e5), foam templating (6e9), and freeze-drying techniques (10,11). The conventional scaffolds are prepared using toxic compounds (e.g. organic solvents and chemical cross-linkers) or under harsh conditions (e.g., high temperature
* Corresponding author. Tel./fax: þ81 99 285 3283. E-mail addresses: [email protected] (T. Takei), breeze21.1220@ hotmail.co.jp (R. Aokawa), [email protected] (T. Shigemitsu), jjdjq118@ ybb.ne.jp (K. Kawakami), [email protected] (M. Yoshida).
and drying). Therefore, cells are immobilized in the scaffolds after removing the toxic compounds or after the harsh process. It is quite difficult to uniformly immobilize cells within the pre-formed scaffolds by the commonly-used cell seeding method in which the scaffolds are overlaid with cell suspensions (a large proportion of the cells locate on the surface of the scaffolds in this method). Some researchers succeeded in uniform cell immobilization within porous scaffolds by incorporating cells within the scaffold matrix during the scaffold preparation process (2,12e15). Among the reports, a soft sacrificial particle leaching technique is the most likely candidate for achieving high porosity (2,12,14,15). The typical procedure of the technique for preparing porous scaffolds is as follows. Gelatin gel beads as pore templates are first prepared and then, added to a sodium alginate solution containing cells (2). The alginate solution was then formed into a gel by immersion into a calcium chloride solution. Finally, pores were formed by melting the gelatin beads by warming to 37 C. Because this process can proceed under mild conditions without using toxic chemicals, cells can be incorporated within the scaffold matrix during the scaffold preparation process, resulting in uniform cell immobilization in scaffolds. However, the maximum porosity is limited to 74% (random close-packing limit), which is accomplished when the beads are packed in their most compact arrangement. Higher porosity is needed to accelerate permeation of oxygen and nutrients, particularly for the regeneration of tissues with high oxygen and nutrient demands, such as the liver (2). In this study, we focused on preparing uniformly cell-laden porous scaffolds with a porosity >74% by the gas-in-liquid foam templating technique. The technique developed by Barbetta et al. (16,17) is able to achieve a high porosity using a non-toxic gas. In
1389-1723/$ e see front matter Ó 2015, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2015.03.017
578
TAKEI ET AL.
the previous reports, the above authors did not achieve uniform cell immobilization in their scaffolds because they used toxic chemical cross-linkers or adopted a drying process. To prepare uniformly cell-laden scaffolds by the technique, we used gelatin as a scaffold material and microbial transglutaminase (MTG) to cross-link it. Gelatin is commonly used in the tissue engineering field because of its good cellular adhesion, biocompatibility and biodegradability (18). For use as a scaffold material, covalent cross-linking of gelatin is essential to improve its thermal stability. MTG catalyzes an acyl transfer reaction between the g-carboxyamide groups of glutamine residues in proteins and primary amines (e.g., the amino groups of lysine residues) (19). Gelatin molecules can be covalently crosslinked in an aqueous environment by the enzyme without damaging mammalian cells (20). Thus, the combination of gelatin and MTG would enable us to uniformly immobilize viable cells in the porous scaffolds prepared by the gas-in-liquid foam templating technique. We first attempted control of porosity and pore size in scaffolds prepared by the gas-in-liquid foam templating technique. Subsequently, we confirmed that viable cells could be uniformly immobilized into the scaffolds. Finally, to obtain basic knowledge concerning compatibility of mammalian cells with the scaffolds, the functions of the cells in the scaffolds was examined in vivo. MATERIALS AND METHODS Materials Porcine gelatin (type A) was purchased from SigmaeAldrich Co. (St. Louis, MO, USA). MTG and sodium alginate with a molecular weight of 70,000 Da (61% guluronic acid residues, Kimica I-1G) were kindly donated by Ajinomoto Co., Inc. (Tokyo, Japan) and Kimica Co. (Tokyo, Japan), respectively. Male nude mice (BALB/cA Jcl-nu, 6 weeks old) were obtained from Kyudo Co. Ltd. (Saga, Japan). A human hepatoblastoma cell line (HepG2, RCB1648) was obtained from Riken Cell Bank (Tsukuba, Japan). The cells (passage number: 35e40) were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin. Preparation of foam A cylindrical glass apparatus (diameter 37 mm) equipped with a porous glass septum (pore sizes 5e10 or 20e30 mm, Vidtec, Fukuoka, Japan) placed at its base was used for the preparation of foam (Fig. 1). Gelatin (12% (w/v)) and sodium alginate (0.5% (w/v)) were dissolved in Ca2þ- and Mg2þ-free phosphate buffered-saline (PBS(), pH 7.4). The solution (7 ml) was poured into the glass apparatus, thermostatically controlled at 40 C. Argon gas was delivered to the polymer solution through the glass septum, resulting in the generation of gas bubbles in the solution. The gas flow rate was adjusted to 5 ml/ min using a syringe pump (gas volume delivered to the solution: 10.5, 21.0 or 36.8 ml). To maintain foam homogeneity during gas insufflation, the foam was stirred using an overhead impeller designed to avoid the incorporation of air (stirring rate: 100 or 250 rpm). Immediately after its formation, the foam was transferred to plastic vials (diameter 13 mm, height 3 mm, volume 0.4 ml) and incubated at 4 C for gelation of the foam through thermally-induced physical cross-linking of gelatin. After removal from the vials, the foam was immersed in a 65 units/ml MTG solution containing 100 mM CaCl2 and 10 mM HEPES (pH 7.4) for 30 min at room temperature to cross-link the gelatin by MTG and the alginate
J. BIOSCI. BIOENG., TABLE 1. Experimental conditions. Volume of Condition Theoretical Actual porosity (%) porosity delivered gas (ml) (%) A B C D E F
60 75 84 60 60 84
57 6 73 1 85 1 e e e
10.5 21.0 36.8 10.5 10.5 36.8
Pore size of glass septum (mm) 5e10 5e10 5e10 5e10 5e10 20e30
Stirring Stirring time (s) rate (rpm) 126 252 442 442 442 442
100 100 100 100 250 100
by Ca2þ. After low vacuum (7.5 104 Pa)/aeration cycles repeated five times to infiltrate the MTG solution into the foam, it was incubated in the solution for 120 min to allow further cross-linking. We confirmed negligible shrinkage of the foam during the cross-linking. Detailed experimental conditions are given in Table 1. The disk-shaped foam (0.4 ml) was then freeze-dried and the structure was observed using scanning electron microscopy (SEM, SS-550, Shimadzu, Kyoto, Japan). Pore size was determined from the area of more than 95 randomly selected pores using the free software package Image J (NIH, Bethesda, MD, USA) under the expedient assumption that the pore cross-sections were exact circles. Theoretical and actual porosities were determined by the following equations: Theoretical porosity (%) ¼ gas volume 100/(gas volume þ polymer solution volume) (1) Actual porosity (%) ¼ (1 polymer solution volume/foam height crosssectional area of cylindrical glass apparatus) 100
(2)
where the polymer solution volume and cross-sectional area of the cylindrical glass apparatus were 7 ml and 10.8 cm2, respectively. Cell damage by stirring and low vacuum/aeration cycles Foam containing HepG2 was prepared using a 12% (w/v) gelatin and 0.5% (w/v) sodium alginate 6 solution, suspending HepG2 cells at a density of 1.0 10 cells/ml, according to the procedure described above (pore size of glass septum 5e10 mm; gas volume 36.8 ml; stirring rate 250 rpm; and stirring time 442 s). The stirring condition was the severest for cells in all foam preparation conditions. After gelation of the diskshaped foam by cooling at 4 C, the low vacuum (7.5 104 Pa)/aeration cycle in PBS () was repeated five times at 15 C. The 10 disk-shaped foams (total volume: 4 ml) were then dissolved in warm PBS () (37 C) and the viability of the cells released in the buffer was examined by Trypan blue exclusion using a hemacytometer. This experiment was performed in triplicate. Cell damage by gelatin cross-linking process HepG2 cells were seeded on a 96-well culture dish at a cell density of 2.0 104 cells/well (0.1 ml medium/well). After cultivation for 24 h at 37 C, each well was rinsed with PBS () and then 0.1 ml of MTG aqueous solution (65 units/ml MTG, 100 mM CaCl2, 10 mM HEPES and 1 g/L glucose in PBS()) was added to the wells. After 3 h of incubation at room temperature, aliquots were removed and the cells were rinsed twice with cell culture media. Subsequently, culture media containing 10% (v/v) WST-8 reagent (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) were added to each well and incubated for 1 h at 37 C. Absorbance of the medium at 450 nm was measured using a spectrophotometer. Control cells were treated with an aqueous solution containing no MTG (100 mM CaCl2, 10 mM HEPES, 1 g/L glucose in PBS()). Viability of MTG-treated HepG2 cells was calculated from the measured absorbances of MTG-treated and -untreated cells. Preparation of cell-laden foam HepG2 cells were suspended in 12% (w/v) gelatin and 0.5% (w/v) sodium alginate solution at a cell density of 1.5 106 cells/ml. Disk-shaped cell-laden foam was prepared using the protocol described above (condition C in Table 1) and then, cross-linked by MTG and Ca2þ. To determine the distribution of cells in the foam, 3 foams were fixed in 10% (w/v) formaldehyde immediately after foam preparation, embedded in paraffin, sectioned (3 cross-sections per scaffold) and stained with hematoxylin and eosin. The cross-section (2 mm wide) was divided into 3 parts (upper side, middle and lower side). All cells in the 3 parts were manually counted. From the cell number, we determined the distribution of cells in the scaffold. This experiment was performed in triplicate.
FIG. 1. Schematic depiction of foaming process.
In vivo evaluation of albumin productivity of HepG2 cells in foam HepG2 cell-laden foam (volume 0.4 ml; cell density 1.0 105 cells/foam) was prepared under condition C shown in Table 1. The foam was cultured in 1 ml DMEM with 10% FBS for 24 h and rinsed with saline. The foam was subcutaneously transplanted into male nude mice under anesthesia with pentobarbital. One week after transplantation, blood was obtained from postcava under anesthesia with pentobarbital. Plasma samples were obtained by centrifugation of the blood samples. The amount of human albumin produced by transplanted HepG2 cell in the samples was determined using a human albumin ELISA kit (Bethyl Laboratories, Montgomery, TX, USA). In a second condition, HepG2 cells suspended in saline (0.4 ml, 1.0 105 cells/0.4 ml saline) were subcutaneously injected into mice (control group). All
VOL. 120, 2015
UNIFORMLY CELL-LADEN POROUS SCAFFOLD
animal experiments were conducted in accordance with the recommendations of Kyushu University’s “Guide for the Care and Use of Laboratory Animals”. Statistical analysis Data are presented as mean standard deviation. Statistical differences between two groups were analyzed using the two-tailed Student’s unpaired t-test. Statistical differences among more than three groups were analyzed using the one-way analysis of variance (ANOVA) with Bonferroni analysis.
RESULTS AND DISCUSSION The key requirement for fabricating foam with a high porosity by the gas-in-liquid foam templating technique is to prepare a stable foam. To this end, we selected argon gas. Physical phenomena that cause destabilization of foam are creaming and Ostwald ripening. Creaming is the condensation of the dispersed phase through migration (upward or downward) of the phase within the continuous phase, resulting in eventual demulsification. As shown by Stokes law, the migration rate (condensation rate) of the dispersed phase increases in proportion to the difference in density between the dispersed and continuous phases. Thus, it is essential to minimize this difference to prepare a stable foam. The density of argon gas is greater than those of other candidate gases, such as air or nitrogen (16,17). Furthermore, argon inhibits the destabilization of foam induced through Ostwald ripening more than other candidate gases, such as carbon dioxide. Ostwald ripening in the foam is the growth of larger bubbles at the expense of smaller ones due to the difference in Laplace pressure between these bubbles. The solubility of a gas in a polymer solution increases in proportion to the pressure. The pressure is higher in smaller bubbles than in larger bubbles, resulting in a higher gas concentration in polymer solution surrounding small bubbles than that surrounding large ones. This difference in gas concentration causes the transfer of gas through a polymer solution from small bubbles to large ones, leading to growth of large bubbles and eventual disappearance of small ones. Argon gas can reduce this phenomenon because of its low water solubility (16). The use of a fast gelling polymeric system is also necessary to prepare foams with a high porosity because foam destabilization proceeds with time. Gelatin can fulfill the requirement. An aqueous gelatin solution quickly forms hydrogels upon cooling ( 95). (B) Pore size distributions within foam. þ p < 0.05 vs condition A, þþ p < 0.05 vs condition D, þþþ p < 0.05 vs condition C.
and intracellular signaling, but poor nutrient supply, gas diffusion and waste removal (21). On the other hand, large pores have opposite effects (21). It has been experimentally demonstrated that the optimum pore size of scaffolds depends on tissue type (e.g., 20e125 mm for skin regeneration, 45e150 mm for liver tissue regeneration, 100e400 mm for bone regeneration, 250e500 mm for cartilage regeneration) (21,22). Here, we attempted to control pore size (Fig. 3). Under condition A, the pore size was 546 231 mm. Pore size decreased with increasing porosity and the size under condition C was 180 75 mm. This decrease is expected to be caused by the longer stirring time. To confirm the effect of stirring time, we extended the time under condition A to 442 s (after insufflating of gas, stirring was continued for a further 316 s) (condition D). As expected, the pore size decreased to 383 116 mm (p < 0.05 vs condition A). The effect of stirring rate on pore size was also examined by increasing the rate under condition D (100 rpm) to 250 rpm (condition E). Pore size decreased to 294 148 mm (p < 0.05 vs condition D). Finally, the pore size of the glass septum under condition C (5e10 mm) was changed to 20e30 mm (condition F). It was confirmed that the pore size of the foam increased to 360 159 mm (p < 0.05 vs condition C). These results show that the pore size of the foam can be controlled by stirring time, stirring rate and the pore size of the glass septum.
HepG2-laden foam We preliminarily examined the influence of the foam preparation process on viability of cells entrapped in the foam. It was quite difficult to determine the viability of HepG2 cells entrapped in foam after cross-linking of gelatin by MTG. Therefore, we separately investigated the influence of (i) stirring and application of the low vacuum/aeration and (ii) treatment with MTG on cell viability (detailed procedures are described in Cell damage by stirring and low vacuum/aeration cycles and Cell damage by gelatin cross-linking process in Materials and methods). The viabilities after each process were 90.2 1.7% and 103.4 12.9%, respectively. This examination suggests that HepG2 are not markedly damaged during the process. In subsequent experiments, we fabricated foam with high porosity (approximately 85%) under condition C. As described in Introduction, uniform cell immobilization within scaffolds is essential for generating homogeneous tissue. Here, we confirmed whether cells could be immobilized uniformly within the foam. Fig. 4A and B show a cross-section of a disk-shaped foam, 3 mm in height. Cell numbers in three regions (upper side, middle and lower side in Fig. 4A) were counted to confirm the cell distribution within the foam. As expected, HepG2 cells were entrapped uniformly within the foam (Fig. 4C). Furthermore, we confirmed by in vitro cell culture that the cells in the foam retained albumin productivity (Fig. S2).
FIG. 4. (A) Image of a cross-section of the foam. (B) Magnified image of panel A. (C) Cell distribution in the three areas shown in panel A. Arrows show HepG2 cells in foam. Scale bars are 300 mm (A) and 100 mm (B).
VOL. 120, 2015
UNIFORMLY CELL-LADEN POROUS SCAFFOLD
2.
3. 4.
5.
6.
7.
FIG. 5. Human albumin concentration in plasma of nude mice transplanted with a HepG2 cell-laden foam or the cell suspension in saline 1 week after transplantation (cell-laden foam group: n ¼ 3, free-cell group: n ¼ 5). *p < 0.05 vs free-cell group.
One of the roles of scaffolds is to enhance engraftment of transplanted cells at transplantation sites. To examine the ability of the present foam, 1 week after transplantation of a HepG2 cellladen foam into nude mice, the concentration of human albumin produced by HepG2 cells was measured in the plasma (noting that no human albumin is normally found in the blood of the mice). In a control condition, free HepG2 cells (suspended in saline) were injected at the same site. The human albumin concentration in mice transplanted with the foam was higher than that in mice injected with free cells (Fig. 5). These results demonstrate that the foam promoted the engraftment of transplanted HepG2 cells. In the present study, we fabricated uniformly cell-laden porous scaffolds of porosity >74% using the gas-in-liquid foam templating technique. The pore size of the foam could be controlled by stirring rate, stirring time and the pore size of the glass septum used to introduce argon gas. The engraftment efficiency of HepG2 cells entrapped within the foam in nude mice was higher than that of cells introduced in free-form. Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jbiosc.2015.03.017. ACKNOWLEDGMENTS The authors gratefully acknowledge Dr. Yoshiyuki Kumazawa at Ajinomoto Co., Inc. for provision of MTG. This work was partially supported by a Grant-in-Aid for Young Scientists (B), 25820383, 2013, from Japan Society for the Promotion of Science (JSPS).
8. 9.
10.
11. 12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
References 22. 1. Ajioka, I., Nishio, R., Ikekita, M., Akaike, T., Sasaki, M., Enami, J., and Watanabe, Y.: Establishment of heterotropic liver tissue mass with direct link to the host liver following implantation of hepatocytes transfected with
581
vascular endothelial growth factor gene in mice, Tissue Eng., 7, 335e344 (2001). Hwang, C. M., Sant, S., Masaeli, M., Kachouie, N. N., Zamanian, B., Lee, S. H., and Khademhosseini, A.: Fabrication of three-dimensional porous cell-laden hydrogel for tissue engineering, Biofabrication, 2, 035003 (2010). Wei, G. and Ma, P. X.: Nanostructured biomaterials for regeneration, Adv. Funct. Mater., 18, 3568e3582 (2008). Lu, L. C., Peter, S. J., Lyman, M. D., Lai, H. L., Leite, S. M., Tamada, J. A., Vacanti, J. P., Langer, R., and Mikos, A. G.: In vitro degradation of porous poly(L-lactic acid) foams, Biomaterials, 21, 1595e1605 (2000). Freed, L. E., Vunjaknovakovic, G., Biron, R. J., Eagles, D. B., Lesnoy, D. C., Barlow, S. K., and Langer, R.: Biodegradable polymer scaffolds for tissue engineering, Biotechnology, 12, 689e693 (1994). Barbetta, A., Gumiero, A., Pecci, R., Bedini, R., and Dentini, M.: Gas-in-liquid foam templating as a method for the production of highly porous scaffolds, Biomacromolecules, 10, 3188e3192 (2009). Li, B., Yoshii, T., Hafeman, A. E., Nyman, J. S., Wenke, J. C., and Guelcher, S. A.: The effects of rhBMP-2 released from biodegradable polyurethane/microsphere composite scaffolds on new bone formation in rat femora, Biomaterials, 30, 6768e6779 (2009). Muthutantri, A., Huang, J., and Edirisinghe, M.: Novel method of preparing hydroxyapatite foams, J. Mater. Sci-Mater. Med., 19, 1485e1490 (2008). Danielsson, C., Ruault, S., Simonet, M., Neuenschwander, P., and Frey, P.: Polyesterurethane foam scaffold for smooth muscle cell tissue engineering, Biomaterials, 27, 1410e1415 (2006). Sultana, N. and Wang, M.: Fabrication of HA/PHBV composite scaffolds through the emulsion freezing/freeze-drying process and characterisation of the scaffolds, J. Mater. Sci-Mater. Med., 19, 2555e2561 (2008). Whang, K., Goldstick, T. K., and Healy, K. E.: A biodegradable polymer scaffold for delivery of osteotropic factors, Biomaterials, 21, 2545e2551 (2000). Gong, Y., Su, K., Lau, T. T., Zhou, R., and Wang, D. A.: Microcavitary hydrogelmediating phase transfer cell culture for cartilage tissue engineering, Tissue Eng. Part A, 16, 3611e3622 (2010). Chen, P. Y., Yang, K. C., Wu, C. C., Yu, J. H., Lin, F. H., and Sun, J. S.: Fabrication of large perfusable macroporous cell-laden hydrogel scaffolds using microbial transglutaminase, Acta Biomater., 10, 912e920 (2014). Lau, T. T., Ho, L. W., and Wang, D. A.: Hepatogenesis of murine induced pluripotent stem cells in 3D micro-cavitary hydrogel system for liver regeneration, Biomaterials, 34, 6659e6669 (2013). Liu, J., Xu, H. H. K., Zhou, H., Weir, M. D., Chen, Q., and Trotman, C. A.: Human umbilical cord stem cell encapsulation in novel macroporous and injectable fibrin for muscle tissue engineering, Acta Biomater., 9, 4688e4697 (2013). Barbetta, A., Rizzitelli, G., Bedini, R., Pecci, R., and Dentini, M.: Porous gelatin hydrogels by gas-in-liquid foam templating, Soft Matter., 6, 1785e1792 (2010). Barbetta, A., Carrino, A., Costantini, M., and Dentini, M.: Polysaccharide based scaffolds obtained by freezing the external phase of gas-in-liquid foams, Soft Matter., 6, 5213e5224 (2010). Sakai, S., Yamaguchi, S., Takei, T., and Kawakami, K.: Oxidized alginatecrosslinked alginate/gelatin hydrogel fibers for fabricating tubular constructs with layered smooth muscle cells and endothelial cells in collagen gel, Biomacromolecules, 9, 2036e2041 (2008). Moriyama, K., Sung, K., Goto, M., and Kamiya, N.: Immobilization of alkaline phosphatase on magnetic particles by site-specific and covalent cross-linking catalyzed by microbial transglutaminase, J. Biosci. Bioeng., 111, 650e653 (2011). Yung, C. W., Wu, L. Q., Tullman, J. A., Payne, G. F., Bentley, W. E., and Barbari, T. A.: Transglutaminase crosslinked gelatin as a tissue engineering scaffold, J. Biomed. Mater. Res. A, 83A, 1039e1046 (2007). Dehghani, F. and Annabi, N.: Engineering porous scaffolds using gas-based techniques, Curr. Opin. Biotechnol., 22, 661e666 (2011). Whang, K., Healy, K. E., Elenz, D. R., Nam, E. K., Tsai, D. C., Thomas, C. H., Nuber, G. W., Glorieux, F. H., Travers, R., and Sprague, S. M.: Engineering bone regeneration with bioabsorbable scaffolds with novel microarchitecture, Tissue Eng., 5, 35e51 (1999).
Fabrication of uniformly cell-laden porous scaffolds using a gas-in-liquid templating technique Takayuki Takei,1, * Ryuta Aokawa,2 Takamasa Shigemitsu,1 Koei Kawakami,2 and Masahiro Yoshida1 Department of Chemical Engineering, Graduate School of Science and Engineering, Kagoshima University, 1-21-40 Korimoto, Kagoshima 890-0065, Japan1 and Department of Chemical Engineering, Graduate School of Engineering, Kyushu University, 744 Motooka, Nishi-ku, Fukuoka 819-0385, Japan2 Received 19 December 2014; accepted 25 March 2015 Available online 23 April 2015
Design of porous scaffolds in tissue engineering field was challenging. Uniform immobilization of cells in the scaffolds with high porosity was essential for homogeneous tissue formation. The present study was aimed at fabricating uniformly cell-laden porous scaffolds with porosity >74% using the gas-in-liquid foam templating technique. To this end, we used gelatin, microbial transglutaminase and argon gas as a scaffold material, cross-linker of the protein and porogen of scaffold, respectively. We confirmed that a porosity of >74% could be achieved by increasing the gas volume delivered to a gelatin solution. Pore size in the scaffold could be controlled by stirring speed, stirring time and the pore size of the filter through which the gas passed. The foaming technique enabled us to uniformly immobilize a human hepatoblastoma cell line in scaffold. Engraftment efficiency of the cell line entrapped within the scaffold in nude mice was higher than that of cells in free-form. These results showed that the uniformly cell-laden porous scaffolds were promising for tissue engineering. Ó 2015, The Society for Biotechnology, Japan. All rights reserved. [Key words: Scaffold; Gas foaming; Gelatin; Alginate; Tissue engineering]
The objective of tissue engineering is to develop biological substitutes for damaged tissues but the design of scaffolds in this field is challenging. Requirements of scaffolds are to promote cellular adhesion, proliferation, and differentiation and to enhance engraftment of transplanted cells at targeted sites. In a general procedure to fabricate the biological substitutes, cells are first seeded in scaffolds at low density and transplanted into bodies. Penetration of blood vessels into the scaffolds is then promoted to supply sufficient oxygen and nutrients to the cells, resulting in cell proliferation in the scaffolds and following creation of the biological substitutes with high cell density (1). Scaffolds are classified into two types: porous sponges and non-porous hydrogels. The former are superior to the latter with regard to accelerated permeability of oxygen/nutrients and cell proliferation because of their porous structure (2). Especially, enhancement of permeability of oxygen and nutrients are important to keep cells alive until blood vessels penetrate into the scaffolds. Uniform immobilization of cells in the porous scaffolds is essential for homogeneous tissue formation. Many methods for fabricating porous scaffolds have been described, such as solvent casting/particulate leaching (3e5), foam templating (6e9), and freeze-drying techniques (10,11). The conventional scaffolds are prepared using toxic compounds (e.g. organic solvents and chemical cross-linkers) or under harsh conditions (e.g., high temperature
* Corresponding author. Tel./fax: þ81 99 285 3283. E-mail addresses: [email protected] (T. Takei), breeze21.1220@ hotmail.co.jp (R. Aokawa), [email protected] (T. Shigemitsu), jjdjq118@ ybb.ne.jp (K. Kawakami), [email protected] (M. Yoshida).
and drying). Therefore, cells are immobilized in the scaffolds after removing the toxic compounds or after the harsh process. It is quite difficult to uniformly immobilize cells within the pre-formed scaffolds by the commonly-used cell seeding method in which the scaffolds are overlaid with cell suspensions (a large proportion of the cells locate on the surface of the scaffolds in this method). Some researchers succeeded in uniform cell immobilization within porous scaffolds by incorporating cells within the scaffold matrix during the scaffold preparation process (2,12e15). Among the reports, a soft sacrificial particle leaching technique is the most likely candidate for achieving high porosity (2,12,14,15). The typical procedure of the technique for preparing porous scaffolds is as follows. Gelatin gel beads as pore templates are first prepared and then, added to a sodium alginate solution containing cells (2). The alginate solution was then formed into a gel by immersion into a calcium chloride solution. Finally, pores were formed by melting the gelatin beads by warming to 37 C. Because this process can proceed under mild conditions without using toxic chemicals, cells can be incorporated within the scaffold matrix during the scaffold preparation process, resulting in uniform cell immobilization in scaffolds. However, the maximum porosity is limited to 74% (random close-packing limit), which is accomplished when the beads are packed in their most compact arrangement. Higher porosity is needed to accelerate permeation of oxygen and nutrients, particularly for the regeneration of tissues with high oxygen and nutrient demands, such as the liver (2). In this study, we focused on preparing uniformly cell-laden porous scaffolds with a porosity >74% by the gas-in-liquid foam templating technique. The technique developed by Barbetta et al. (16,17) is able to achieve a high porosity using a non-toxic gas. In
1389-1723/$ e see front matter Ó 2015, The Society for Biotechnology, Japan. All rights reserved. http://dx.doi.org/10.1016/j.jbiosc.2015.03.017
578
TAKEI ET AL.
the previous reports, the above authors did not achieve uniform cell immobilization in their scaffolds because they used toxic chemical cross-linkers or adopted a drying process. To prepare uniformly cell-laden scaffolds by the technique, we used gelatin as a scaffold material and microbial transglutaminase (MTG) to cross-link it. Gelatin is commonly used in the tissue engineering field because of its good cellular adhesion, biocompatibility and biodegradability (18). For use as a scaffold material, covalent cross-linking of gelatin is essential to improve its thermal stability. MTG catalyzes an acyl transfer reaction between the g-carboxyamide groups of glutamine residues in proteins and primary amines (e.g., the amino groups of lysine residues) (19). Gelatin molecules can be covalently crosslinked in an aqueous environment by the enzyme without damaging mammalian cells (20). Thus, the combination of gelatin and MTG would enable us to uniformly immobilize viable cells in the porous scaffolds prepared by the gas-in-liquid foam templating technique. We first attempted control of porosity and pore size in scaffolds prepared by the gas-in-liquid foam templating technique. Subsequently, we confirmed that viable cells could be uniformly immobilized into the scaffolds. Finally, to obtain basic knowledge concerning compatibility of mammalian cells with the scaffolds, the functions of the cells in the scaffolds was examined in vivo. MATERIALS AND METHODS Materials Porcine gelatin (type A) was purchased from SigmaeAldrich Co. (St. Louis, MO, USA). MTG and sodium alginate with a molecular weight of 70,000 Da (61% guluronic acid residues, Kimica I-1G) were kindly donated by Ajinomoto Co., Inc. (Tokyo, Japan) and Kimica Co. (Tokyo, Japan), respectively. Male nude mice (BALB/cA Jcl-nu, 6 weeks old) were obtained from Kyudo Co. Ltd. (Saga, Japan). A human hepatoblastoma cell line (HepG2, RCB1648) was obtained from Riken Cell Bank (Tsukuba, Japan). The cells (passage number: 35e40) were maintained in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% (v/v) fetal bovine serum (FBS), 100 U/ml penicillin and 100 mg/ml streptomycin. Preparation of foam A cylindrical glass apparatus (diameter 37 mm) equipped with a porous glass septum (pore sizes 5e10 or 20e30 mm, Vidtec, Fukuoka, Japan) placed at its base was used for the preparation of foam (Fig. 1). Gelatin (12% (w/v)) and sodium alginate (0.5% (w/v)) were dissolved in Ca2þ- and Mg2þ-free phosphate buffered-saline (PBS(), pH 7.4). The solution (7 ml) was poured into the glass apparatus, thermostatically controlled at 40 C. Argon gas was delivered to the polymer solution through the glass septum, resulting in the generation of gas bubbles in the solution. The gas flow rate was adjusted to 5 ml/ min using a syringe pump (gas volume delivered to the solution: 10.5, 21.0 or 36.8 ml). To maintain foam homogeneity during gas insufflation, the foam was stirred using an overhead impeller designed to avoid the incorporation of air (stirring rate: 100 or 250 rpm). Immediately after its formation, the foam was transferred to plastic vials (diameter 13 mm, height 3 mm, volume 0.4 ml) and incubated at 4 C for gelation of the foam through thermally-induced physical cross-linking of gelatin. After removal from the vials, the foam was immersed in a 65 units/ml MTG solution containing 100 mM CaCl2 and 10 mM HEPES (pH 7.4) for 30 min at room temperature to cross-link the gelatin by MTG and the alginate
J. BIOSCI. BIOENG., TABLE 1. Experimental conditions. Volume of Condition Theoretical Actual porosity (%) porosity delivered gas (ml) (%) A B C D E F
60 75 84 60 60 84
57 6 73 1 85 1 e e e
10.5 21.0 36.8 10.5 10.5 36.8
Pore size of glass septum (mm) 5e10 5e10 5e10 5e10 5e10 20e30
Stirring Stirring time (s) rate (rpm) 126 252 442 442 442 442
100 100 100 100 250 100
by Ca2þ. After low vacuum (7.5 104 Pa)/aeration cycles repeated five times to infiltrate the MTG solution into the foam, it was incubated in the solution for 120 min to allow further cross-linking. We confirmed negligible shrinkage of the foam during the cross-linking. Detailed experimental conditions are given in Table 1. The disk-shaped foam (0.4 ml) was then freeze-dried and the structure was observed using scanning electron microscopy (SEM, SS-550, Shimadzu, Kyoto, Japan). Pore size was determined from the area of more than 95 randomly selected pores using the free software package Image J (NIH, Bethesda, MD, USA) under the expedient assumption that the pore cross-sections were exact circles. Theoretical and actual porosities were determined by the following equations: Theoretical porosity (%) ¼ gas volume 100/(gas volume þ polymer solution volume) (1) Actual porosity (%) ¼ (1 polymer solution volume/foam height crosssectional area of cylindrical glass apparatus) 100
(2)
where the polymer solution volume and cross-sectional area of the cylindrical glass apparatus were 7 ml and 10.8 cm2, respectively. Cell damage by stirring and low vacuum/aeration cycles Foam containing HepG2 was prepared using a 12% (w/v) gelatin and 0.5% (w/v) sodium alginate 6 solution, suspending HepG2 cells at a density of 1.0 10 cells/ml, according to the procedure described above (pore size of glass septum 5e10 mm; gas volume 36.8 ml; stirring rate 250 rpm; and stirring time 442 s). The stirring condition was the severest for cells in all foam preparation conditions. After gelation of the diskshaped foam by cooling at 4 C, the low vacuum (7.5 104 Pa)/aeration cycle in PBS () was repeated five times at 15 C. The 10 disk-shaped foams (total volume: 4 ml) were then dissolved in warm PBS () (37 C) and the viability of the cells released in the buffer was examined by Trypan blue exclusion using a hemacytometer. This experiment was performed in triplicate. Cell damage by gelatin cross-linking process HepG2 cells were seeded on a 96-well culture dish at a cell density of 2.0 104 cells/well (0.1 ml medium/well). After cultivation for 24 h at 37 C, each well was rinsed with PBS () and then 0.1 ml of MTG aqueous solution (65 units/ml MTG, 100 mM CaCl2, 10 mM HEPES and 1 g/L glucose in PBS()) was added to the wells. After 3 h of incubation at room temperature, aliquots were removed and the cells were rinsed twice with cell culture media. Subsequently, culture media containing 10% (v/v) WST-8 reagent (Dojindo Molecular Technologies, Inc., Kumamoto, Japan) were added to each well and incubated for 1 h at 37 C. Absorbance of the medium at 450 nm was measured using a spectrophotometer. Control cells were treated with an aqueous solution containing no MTG (100 mM CaCl2, 10 mM HEPES, 1 g/L glucose in PBS()). Viability of MTG-treated HepG2 cells was calculated from the measured absorbances of MTG-treated and -untreated cells. Preparation of cell-laden foam HepG2 cells were suspended in 12% (w/v) gelatin and 0.5% (w/v) sodium alginate solution at a cell density of 1.5 106 cells/ml. Disk-shaped cell-laden foam was prepared using the protocol described above (condition C in Table 1) and then, cross-linked by MTG and Ca2þ. To determine the distribution of cells in the foam, 3 foams were fixed in 10% (w/v) formaldehyde immediately after foam preparation, embedded in paraffin, sectioned (3 cross-sections per scaffold) and stained with hematoxylin and eosin. The cross-section (2 mm wide) was divided into 3 parts (upper side, middle and lower side). All cells in the 3 parts were manually counted. From the cell number, we determined the distribution of cells in the scaffold. This experiment was performed in triplicate.
FIG. 1. Schematic depiction of foaming process.
In vivo evaluation of albumin productivity of HepG2 cells in foam HepG2 cell-laden foam (volume 0.4 ml; cell density 1.0 105 cells/foam) was prepared under condition C shown in Table 1. The foam was cultured in 1 ml DMEM with 10% FBS for 24 h and rinsed with saline. The foam was subcutaneously transplanted into male nude mice under anesthesia with pentobarbital. One week after transplantation, blood was obtained from postcava under anesthesia with pentobarbital. Plasma samples were obtained by centrifugation of the blood samples. The amount of human albumin produced by transplanted HepG2 cell in the samples was determined using a human albumin ELISA kit (Bethyl Laboratories, Montgomery, TX, USA). In a second condition, HepG2 cells suspended in saline (0.4 ml, 1.0 105 cells/0.4 ml saline) were subcutaneously injected into mice (control group). All
VOL. 120, 2015
UNIFORMLY CELL-LADEN POROUS SCAFFOLD
animal experiments were conducted in accordance with the recommendations of Kyushu University’s “Guide for the Care and Use of Laboratory Animals”. Statistical analysis Data are presented as mean standard deviation. Statistical differences between two groups were analyzed using the two-tailed Student’s unpaired t-test. Statistical differences among more than three groups were analyzed using the one-way analysis of variance (ANOVA) with Bonferroni analysis.
RESULTS AND DISCUSSION The key requirement for fabricating foam with a high porosity by the gas-in-liquid foam templating technique is to prepare a stable foam. To this end, we selected argon gas. Physical phenomena that cause destabilization of foam are creaming and Ostwald ripening. Creaming is the condensation of the dispersed phase through migration (upward or downward) of the phase within the continuous phase, resulting in eventual demulsification. As shown by Stokes law, the migration rate (condensation rate) of the dispersed phase increases in proportion to the difference in density between the dispersed and continuous phases. Thus, it is essential to minimize this difference to prepare a stable foam. The density of argon gas is greater than those of other candidate gases, such as air or nitrogen (16,17). Furthermore, argon inhibits the destabilization of foam induced through Ostwald ripening more than other candidate gases, such as carbon dioxide. Ostwald ripening in the foam is the growth of larger bubbles at the expense of smaller ones due to the difference in Laplace pressure between these bubbles. The solubility of a gas in a polymer solution increases in proportion to the pressure. The pressure is higher in smaller bubbles than in larger bubbles, resulting in a higher gas concentration in polymer solution surrounding small bubbles than that surrounding large ones. This difference in gas concentration causes the transfer of gas through a polymer solution from small bubbles to large ones, leading to growth of large bubbles and eventual disappearance of small ones. Argon gas can reduce this phenomenon because of its low water solubility (16). The use of a fast gelling polymeric system is also necessary to prepare foams with a high porosity because foam destabilization proceeds with time. Gelatin can fulfill the requirement. An aqueous gelatin solution quickly forms hydrogels upon cooling ( 95). (B) Pore size distributions within foam. þ p < 0.05 vs condition A, þþ p < 0.05 vs condition D, þþþ p < 0.05 vs condition C.
and intracellular signaling, but poor nutrient supply, gas diffusion and waste removal (21). On the other hand, large pores have opposite effects (21). It has been experimentally demonstrated that the optimum pore size of scaffolds depends on tissue type (e.g., 20e125 mm for skin regeneration, 45e150 mm for liver tissue regeneration, 100e400 mm for bone regeneration, 250e500 mm for cartilage regeneration) (21,22). Here, we attempted to control pore size (Fig. 3). Under condition A, the pore size was 546 231 mm. Pore size decreased with increasing porosity and the size under condition C was 180 75 mm. This decrease is expected to be caused by the longer stirring time. To confirm the effect of stirring time, we extended the time under condition A to 442 s (after insufflating of gas, stirring was continued for a further 316 s) (condition D). As expected, the pore size decreased to 383 116 mm (p < 0.05 vs condition A). The effect of stirring rate on pore size was also examined by increasing the rate under condition D (100 rpm) to 250 rpm (condition E). Pore size decreased to 294 148 mm (p < 0.05 vs condition D). Finally, the pore size of the glass septum under condition C (5e10 mm) was changed to 20e30 mm (condition F). It was confirmed that the pore size of the foam increased to 360 159 mm (p < 0.05 vs condition C). These results show that the pore size of the foam can be controlled by stirring time, stirring rate and the pore size of the glass septum.
HepG2-laden foam We preliminarily examined the influence of the foam preparation process on viability of cells entrapped in the foam. It was quite difficult to determine the viability of HepG2 cells entrapped in foam after cross-linking of gelatin by MTG. Therefore, we separately investigated the influence of (i) stirring and application of the low vacuum/aeration and (ii) treatment with MTG on cell viability (detailed procedures are described in Cell damage by stirring and low vacuum/aeration cycles and Cell damage by gelatin cross-linking process in Materials and methods). The viabilities after each process were 90.2 1.7% and 103.4 12.9%, respectively. This examination suggests that HepG2 are not markedly damaged during the process. In subsequent experiments, we fabricated foam with high porosity (approximately 85%) under condition C. As described in Introduction, uniform cell immobilization within scaffolds is essential for generating homogeneous tissue. Here, we confirmed whether cells could be immobilized uniformly within the foam. Fig. 4A and B show a cross-section of a disk-shaped foam, 3 mm in height. Cell numbers in three regions (upper side, middle and lower side in Fig. 4A) were counted to confirm the cell distribution within the foam. As expected, HepG2 cells were entrapped uniformly within the foam (Fig. 4C). Furthermore, we confirmed by in vitro cell culture that the cells in the foam retained albumin productivity (Fig. S2).
FIG. 4. (A) Image of a cross-section of the foam. (B) Magnified image of panel A. (C) Cell distribution in the three areas shown in panel A. Arrows show HepG2 cells in foam. Scale bars are 300 mm (A) and 100 mm (B).
VOL. 120, 2015
UNIFORMLY CELL-LADEN POROUS SCAFFOLD
2.
3. 4.
5.
6.
7.
FIG. 5. Human albumin concentration in plasma of nude mice transplanted with a HepG2 cell-laden foam or the cell suspension in saline 1 week after transplantation (cell-laden foam group: n ¼ 3, free-cell group: n ¼ 5). *p < 0.05 vs free-cell group.
One of the roles of scaffolds is to enhance engraftment of transplanted cells at transplantation sites. To examine the ability of the present foam, 1 week after transplantation of a HepG2 cellladen foam into nude mice, the concentration of human albumin produced by HepG2 cells was measured in the plasma (noting that no human albumin is normally found in the blood of the mice). In a control condition, free HepG2 cells (suspended in saline) were injected at the same site. The human albumin concentration in mice transplanted with the foam was higher than that in mice injected with free cells (Fig. 5). These results demonstrate that the foam promoted the engraftment of transplanted HepG2 cells. In the present study, we fabricated uniformly cell-laden porous scaffolds of porosity >74% using the gas-in-liquid foam templating technique. The pore size of the foam could be controlled by stirring rate, stirring time and the pore size of the glass septum used to introduce argon gas. The engraftment efficiency of HepG2 cells entrapped within the foam in nude mice was higher than that of cells introduced in free-form. Supplementary data related to this article can be found at http:// dx.doi.org/10.1016/j.jbiosc.2015.03.017. ACKNOWLEDGMENTS The authors gratefully acknowledge Dr. Yoshiyuki Kumazawa at Ajinomoto Co., Inc. for provision of MTG. This work was partially supported by a Grant-in-Aid for Young Scientists (B), 25820383, 2013, from Japan Society for the Promotion of Science (JSPS).
8. 9.
10.
11. 12.
13.
14.
15.
16. 17.
18.
19.
20.
21.
References 22. 1. Ajioka, I., Nishio, R., Ikekita, M., Akaike, T., Sasaki, M., Enami, J., and Watanabe, Y.: Establishment of heterotropic liver tissue mass with direct link to the host liver following implantation of hepatocytes transfected with
581
vascular endothelial growth factor gene in mice, Tissue Eng., 7, 335e344 (2001). Hwang, C. M., Sant, S., Masaeli, M., Kachouie, N. N., Zamanian, B., Lee, S. H., and Khademhosseini, A.: Fabrication of three-dimensional porous cell-laden hydrogel for tissue engineering, Biofabrication, 2, 035003 (2010). Wei, G. and Ma, P. X.: Nanostructured biomaterials for regeneration, Adv. Funct. Mater., 18, 3568e3582 (2008). Lu, L. C., Peter, S. J., Lyman, M. D., Lai, H. L., Leite, S. M., Tamada, J. A., Vacanti, J. P., Langer, R., and Mikos, A. G.: In vitro degradation of porous poly(L-lactic acid) foams, Biomaterials, 21, 1595e1605 (2000). Freed, L. E., Vunjaknovakovic, G., Biron, R. J., Eagles, D. B., Lesnoy, D. C., Barlow, S. K., and Langer, R.: Biodegradable polymer scaffolds for tissue engineering, Biotechnology, 12, 689e693 (1994). Barbetta, A., Gumiero, A., Pecci, R., Bedini, R., and Dentini, M.: Gas-in-liquid foam templating as a method for the production of highly porous scaffolds, Biomacromolecules, 10, 3188e3192 (2009). Li, B., Yoshii, T., Hafeman, A. E., Nyman, J. S., Wenke, J. C., and Guelcher, S. A.: The effects of rhBMP-2 released from biodegradable polyurethane/microsphere composite scaffolds on new bone formation in rat femora, Biomaterials, 30, 6768e6779 (2009). Muthutantri, A., Huang, J., and Edirisinghe, M.: Novel method of preparing hydroxyapatite foams, J. Mater. Sci-Mater. Med., 19, 1485e1490 (2008). Danielsson, C., Ruault, S., Simonet, M., Neuenschwander, P., and Frey, P.: Polyesterurethane foam scaffold for smooth muscle cell tissue engineering, Biomaterials, 27, 1410e1415 (2006). Sultana, N. and Wang, M.: Fabrication of HA/PHBV composite scaffolds through the emulsion freezing/freeze-drying process and characterisation of the scaffolds, J. Mater. Sci-Mater. Med., 19, 2555e2561 (2008). Whang, K., Goldstick, T. K., and Healy, K. E.: A biodegradable polymer scaffold for delivery of osteotropic factors, Biomaterials, 21, 2545e2551 (2000). Gong, Y., Su, K., Lau, T. T., Zhou, R., and Wang, D. A.: Microcavitary hydrogelmediating phase transfer cell culture for cartilage tissue engineering, Tissue Eng. Part A, 16, 3611e3622 (2010). Chen, P. Y., Yang, K. C., Wu, C. C., Yu, J. H., Lin, F. H., and Sun, J. S.: Fabrication of large perfusable macroporous cell-laden hydrogel scaffolds using microbial transglutaminase, Acta Biomater., 10, 912e920 (2014). Lau, T. T., Ho, L. W., and Wang, D. A.: Hepatogenesis of murine induced pluripotent stem cells in 3D micro-cavitary hydrogel system for liver regeneration, Biomaterials, 34, 6659e6669 (2013). Liu, J., Xu, H. H. K., Zhou, H., Weir, M. D., Chen, Q., and Trotman, C. A.: Human umbilical cord stem cell encapsulation in novel macroporous and injectable fibrin for muscle tissue engineering, Acta Biomater., 9, 4688e4697 (2013). Barbetta, A., Rizzitelli, G., Bedini, R., Pecci, R., and Dentini, M.: Porous gelatin hydrogels by gas-in-liquid foam templating, Soft Matter., 6, 1785e1792 (2010). Barbetta, A., Carrino, A., Costantini, M., and Dentini, M.: Polysaccharide based scaffolds obtained by freezing the external phase of gas-in-liquid foams, Soft Matter., 6, 5213e5224 (2010). Sakai, S., Yamaguchi, S., Takei, T., and Kawakami, K.: Oxidized alginatecrosslinked alginate/gelatin hydrogel fibers for fabricating tubular constructs with layered smooth muscle cells and endothelial cells in collagen gel, Biomacromolecules, 9, 2036e2041 (2008). Moriyama, K., Sung, K., Goto, M., and Kamiya, N.: Immobilization of alkaline phosphatase on magnetic particles by site-specific and covalent cross-linking catalyzed by microbial transglutaminase, J. Biosci. Bioeng., 111, 650e653 (2011). Yung, C. W., Wu, L. Q., Tullman, J. A., Payne, G. F., Bentley, W. E., and Barbari, T. A.: Transglutaminase crosslinked gelatin as a tissue engineering scaffold, J. Biomed. Mater. Res. A, 83A, 1039e1046 (2007). Dehghani, F. and Annabi, N.: Engineering porous scaffolds using gas-based techniques, Curr. Opin. Biotechnol., 22, 661e666 (2011). Whang, K., Healy, K. E., Elenz, D. R., Nam, E. K., Tsai, D. C., Thomas, C. H., Nuber, G. W., Glorieux, F. H., Travers, R., and Sprague, S. M.: Engineering bone regeneration with bioabsorbable scaffolds with novel microarchitecture, Tissue Eng., 5, 35e51 (1999).
