Progress in Science 36 (2011) 238–268

Contents lists available at ScienceDirect

Progress in Polymer Science

journal homepage: www.elsevier.com/locate/ppolysci

Design of artificial extracellular matrices for

Byung-Soo Kim a,1, In-Kyu Park b,1, Takashi Hoshiba c, Hu-Lin Jiang d, Yun-Jaie Choi e, Toshihiro Akaike f,∗, Chong-Su Cho e,∗∗ a School of Chemical and Biological Engineering, Seoul National University, Seoul 151-744, South Korea b Department of Biomedical Sciences, Chonnam National University Medical School, Gwangju 501-746, South Korea c Biomaterials Center, National Institute for Materials Science, Tsukuba 305-0044, Japan d College of Veterinary Medicine, Seoul National University, Seoul 151-742, South Korea e Department of Agricultural Biotechnology and Research Institute for Agriculture and Life Sciences, Seoul National University, Seoul 151-921, South Korea f Department of Biomolecular Engineering, Tokyo Institute of Technology, Yokohama 226-8501, Japan article info abstract

Article history: The design of artificial extracellular matrix (ECM) is important in tissue engineering because Received 25 May 2010 artificial ECM regulates cellular behaviors, including proliferation, survival, migration, Received in revised form and differentiation. Artificial ECMs have several functions in tissue engineering, includ- 22 September 2010 ing provision of cell-adhesive substrate, control of three-dimensional tissue structure, and Accepted 7 October 2010 Available online 20 October 2010 presentation of growth factors, cell-adhesion signals, and mechanical signals. Design cri- teria for artificial ECMs vary considerably depending on the type of the engineered tissue. This article reviews the materials and methods that have been used in fabrication of artifi- Keywords: Artificial extracellular matrix cial ECMs for engineering of specific tissues, including liver, , , and skin. This Naturally-derived polymer article also reviews artificial ECMs used for modulation of behaviors for tissue Stem cells engineering applications. Synthetic polymer © 2010 Elsevier Ltd. All rights reserved. Tissue engineering

Contents

1. Introduction ...... 239 1.1. Significance of artificial extracellular matrix in tissue engineering applications ...... 239 1.2. General requirements of artificial ECMs for tissue engineering...... 239 1.3. Classification of artificial ECMs ...... 239 2. Liver ...... 240 2.1. Mechanism of specific interaction between hepatocytes and galactose moieties ...... 240 2.2. Classification of artificial ECMs for liver tissue engineering...... 241 2.2.1. Synthetic ...... 241 2.2.2. Naturally-derived polymers...... 242 2.3. Parameters affecting hepatocellular behavior ...... 245

∗ Corresponding author. Tel.: +81 45 924 5790; fax: +81 45 924 5815. ∗∗ Corresponding author. Tel.: +82 2 880 4868; fax: +82 2 875 2494. E-mail addresses: [email protected] (T. Akaike), [email protected] (C.-S. Cho). 1 These authors contributed equally to this work.

0079-6700/$ – see front matter © 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.progpolymsci.2010.10.001 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 239

2.3.1. Galactose density and microdistribution of galactose ...... 246 2.3.2. Topology of artificial ECMs ...... 247 3. Cartilage ...... 248 3.1. Natural material-based artificial ECMs ...... 248 3.2. Synthetic material-based artificial ECMs ...... 251 3.2.1. Poly(ethylene glycol) (PEG) ...... 252 3.2.2. Injectable hydrogels ...... 252 3.3. Composite artificial ECMs ...... 253 4. Bone ...... 253 4.1. Artificial ECMs for bone tissue engineering ...... 253 4.1.1. Design criteria of artificial ECMs for bone tissue engineering ...... 253 4.1.2. Non-polymeric materials ...... 253 4.1.3. Synthetic polymeric materials ...... 254 4.1.4. Naturally-derived polymers...... 254 4.1.5. Composite materials ...... 254 4.2. Regulation of cellular functions by artificial ECMs ...... 254 4.2.1. Importance of regulation of cellular function by artificial ECMs ...... 254 4.2.2. Regulation of osteoblast functions ...... 254 4.3. Artificial ECMs for osteochondral defects ...... 255 5. Skin ...... 255 5.1. Naturally-derived materials ...... 256 5.2. Synthetic polymeric materials ...... 256 5.3. Future challenges in skin tissue engineering ...... 256 6. Stem cells ...... 256 6.1. Delivery of soluble factors using artificial ECMs...... 256 6.2. Cell adhesion signal presentation by artificial ECMs ...... 258 6.3. Chemistry of artificial ECMs ...... 258 6.4. Architecture of artificial ECMs ...... 259 6.5. Mechanical properties of artificial ECMs ...... 259 6.6. Control of cell shape and size with artificial ECMs ...... 260 6.7. Mechanical signal loading on artificial ECMs ...... 261 7. Concluding remarks ...... 262 Acknowledgements ...... 262 References ...... 262

1. Introduction 1.2. General requirements of artificial ECMs for tissue engineering 1.1. Significance of artificial extracellular matrix in tissue engineering applications Artificial ECMs should be designed to guide cell attachment, proliferation, and differentiation for tissue Tissue engineering, which aims to reconstruct living tis- regeneration [8]. Material and biological requirements sues for replacement of damaged or lost tissues/organs should be considered for each tissue engineering applica- of living organisms, has recently emerged as an exciting tion, even though the exact requirements differ according interdisciplinary area in the life sciences [1]. To achieve to organ [9–11]. For the material side, there are several this aim, it is necessary to use cells together with biosig- requirements for artificial ECMs: artificial ECMs should be nalling molecules and extracellular matrix (ECM) into or biocompatible and non-toxic, and have a desired degrada- onto which cells will develop, organize, and behave as if tion rate, surface properties, processability, porosity, and they are in their native tissue [2]. Tissue engineering typ- mechanical properties. For the biological side, artificial ically uses artificial ECMs to engineer new tissues from ECMs should be non-immunogenic and should not cause cells [3,4]. Artificial ECMs should be designed to bring the foreign body reactions. They should have desired bioactiv- desired cell types in contact with an appropriate environ- ity, nutrient availability, and access to growth factors. ment and to provide mechanical support until the newly formed tissues are structurally stable [5]. Therefore, design 1.3. Classification of artificial ECMs of artificial ECMs is very important in tissue engineering, because artificial ECMs regulate many cellular behaviors, Artificial ECMs can be classified into two types: natural including proliferation and growth, survival, change of polymer-based artificial ECMs and synthetic polymer- cell shape, migration, and differentiation [6]. Also, arti- based ones. Natural polymer-based artificial ECMs consist ficial ECMs have multiple functions, such as serving as of structural and functional proteins, , gly- an adhesive substrate, provision of structure, presen- coproteins, and glycosaminoglycans found in a natural tation and storage of growth factors, and detection of tissue [12], and they have been implanted for construc- signals [7]. tive remodeling of many tissues in preclinical and human 240 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

Table 1 General characteristics of commonly used naturally-derived polymers for tissue engineering [modified from Ref. [9]].

Polymers Biocompatibility Disadvantages Biodegradability Application

Collagen Minimal cytotoxicity, Proteoytic removal of Bulk, Controllable Skin; cartilage; bone; ligaments; Mild foreign body small nonhelical tendons; vessels; nerves; bladder; reaction, Minimal telopepties liver inflammation Hyaluronic acid Minimal foreign body Highly viscous Bulk,1hto1month Skin; cartilage; bone; ligaments; reaction, No solution, Many nerves; vessels; liver inflammation purification steps after chemical modification Alginic acid Minimal foreign body Uncontrollable Bulk, 1 day to 3 months Skin; cartilage; bone; nerves; reaction, No dissolution of muscle; pancreas inflammation Chitosan Minimal foreign body Uncontrollable Bulk, 3 days to 6 months Skin; cartilage; bone; nerves; reaction, No deacetylation and vessels; liver; pancreas inflammation molecular weight Minimal cytotoxicity, Weak mechanical Bulk, Controllable Skin; bone; cartilage; ligaments; Mild foreign body property breast reaction, Minimal inflammation Fibrin Minimal cytotoxicity, Weak mechanical Bulk, Controllable Skin, bone, cartilage; liver,tendons; Mild foreign body property ligaments; vessels reaction, Minimal inflammation Poly(hydroxyalkanoate) Minimal cytotoxicity, Pyrogen removed Bulk, Controllable Skin; bone; tendons; cartilage; Mild foreign body nerves; ligaments; heart; vessels; reaction, Minimal muscle inflammation Silk Minimal cytotoxicity, Inflammation of sericin Bulk, Controllable Skin; ligaments; bone; cartilage; Mild foreign body tympanic membrane; vessels; reaction, Minimal tendons inflammation clinical trials. Naturally-derived polymers have advantages neering is critically important, because hepatocytes are over synthetic ones because they have excellent biological attachment-dependent cells and lose their liver-specific properties, including cell adhesion, mechanical proper- functions without an optimal ECM. This section discusses ties similar to those of natural tissues, biodegradability, mechanisms of the specific interaction between hepatocyte and biocompatibility [12]. General characteristics of the and galactose moieties, classification of artificial ECMs, fac- most commonly used naturally-derived polymers for tis- tors affecting cellular behaviors of hepatocytes, and surface sue engineering are shown in Table 1. A recent review on modification. tissue engineering using natural polymer-based artificial ECMs is available [10]. Synthetic polymer-based artificial ECMs are generally 2.1. Mechanism of specific interaction between fabricated from biocompatible, biodegradable polymers to hepatocytes and galactose moieties avoid chronic foreign body reactions. Synthetic polymers may be obtained by reproduction on a large scale and can Asialoglycoprotein receptors (ASGPR) of hepatocytes be processed into artificial ECMs in which the mechan- were first identified by Pricer et al. [14]; circulating ical properties and degradation time of some synthetic asialoglycoproteins (ASGPS) bind to and are degraded by polymers can be controlled [12]. However, the greatest hepatocytes. Hepatocytes have cell surface receptors that disadvantage of synthetic polymers is the lack of cell- recognize and bind molecules with exposed galactose, recognition sites and the possibility of a pro-inflammatory N-acetyl galactosamine, or glucose residues [15]. After response after implantation. General characteristics of binding to receptors, ASGPS are internalized via coated commonly used synthetic polymers for tissue engineering pits and coated vesicles and subsequently appear in a are shown in Table 2. Recent reviews on synthetic polymers complex arrangement of larger smooth-surface vesicles for tissue engineering scaffold are available [9,11,12]. [16]. Hepatic plasma membrane receptors localized on the sinusoidal face of hepatocytes mediate specific binding and 2. Liver uptake of partially deglycosylated through receptor-mediated endocytosis [17]. Therefore, mimicking The liver plays an important role in a broad spec- the biological environment, the process of incorporation trum of physiological functions, including metabolism, of old proteins by the liver has been imitated; however, storage, and synthesis and release of vitamins, carbohy- hepatocyte transplantation using poly(l-lactic acid) (PLA) drates, proteins, lipids, and cyclic tetrapyrroles [13]. Also, [18], poly(vinyl alcohol) (PVA) [19], poly(glycolic acid) the liver detoxifies and inactivates endogenous and exoge- (PGA) [20], poly(lactic-co-glycolic acid) (PLGA) [21], col- nous substances, and activates precursor molecules [13]. lagen [22], alginate [23], and polyurethane [24] has been Therefore, the design of artificial ECMs for liver tissue engi- achieved in vivo. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 241

Table 2 General characteristics of commonly used synthetic polymers for tissue engineering [modified from Ref. [9]].

Polymers Biocompatibility Disadvantages Biodegradability Application

Poly (lactic acid) Minimal cytotoxicity, Mild Local inflammation, Bulk, 24 months Skin; cartilage; bone foreign body reaction, Random chain hydrolysis ligaments; tendons; Minimal inflammation vessels; nerves; bladder; liver Poly (glycolic acid) Minimal cytotoxicity, Mild Local inflammation, Bulk, 6–12 months Skin; cartilage; bone foreign body reaction, Random chain hydrolysis ligaments; tendons; Minimal inflammation vessels; nerves; bladder; liver Poly (lactic acid-co-glycolic acid) Minimal cytotoxicity, Mild Local inflammation, Bulk, 1–6 months Skin; cartilage; bone foreign body reaction, Random chain hydrolysis ligaments; tendons; Minimal inflammation vessels; nerves; bladder; liver Poly (caprolactone) Minimal cytotoxicity, Mild Hydrophobic Bulk, 3 years Skin; cartilage; bone; foreign body reaction, ligaments; tendons; Minimal inflammation vessels; nerves Poly (ethylene oxide) Mild foreign body reaction, Complex biodegradability Bulk, 1 month-5 years Skin; cartilage; bone; terephthalate-co-butylene No inflammation muscles terephthalate Polyanhydrides Minimal foreign body Limited mechanical Surface erosion, Bone reaction, Minimal property Controllable inflammation, Minimal cytotoxicity Poly (propylene fumarate) Mild foreign body reaction, Weak mechanical property Surface erosion, 1 Bone Minimal inflammation week–16 months Poly (orthoester)s Mild inflammation, Mild Weak mechanical property Bulk ∼ several months Ear; bone; cartilage foreign body reaction Polyphosphazene Minimal foreign body Broad molecular weight Surface erosion, 1 Skin; cartilage; bone reaction, Minimal distribution week–3 years nerves; ligaments inflammation

2.2. Classification of artificial ECMs for liver tissue Akaike and colleagues synthesized galactose-carrying engineering polystyrene (PS) in order to guide hepatocyte adhe- sion. Galactose-carrying PS [poly(N-p-vinylbenzyl-4-O-␤- Two types of artificial ECMs for liver tissue engineering d-galactopyranosyl-d-gluconamide), PVLA] was synthe- will be discussed: one is derived from synthetic polymers, sized according to three steps, as shown in Fig. 1 [29]. and the other is derived from natural sources. Synthesis of PVLA is relatively simple, protection of the hydroxyl groups of lactose is unnecessary, and the yield 2.2.1. Synthetic polymers of each step is high, even though it is expensive to prepare Synthetic polymer applications in liver tissue engineer- compared to other synthetic polymers. The authors inves- ing have attracted attention in recent years due to their tigated the use of PVLA as an artificial ECM for hepatocytes. excellent mechanical properties, processability, low cost, The results indicated that hepatocytes adhered to the PVLA and controllable degradation time. However, the great- through a unique ASGPR-galactose interaction between the est disadvantage of synthetic polymers is the absence of ASGPR of hepatocytes and highly concentrated galactose hepatocyte-recognition sites. Therefore, many efforts have moieties along the polymer chains [30]. The round mor- focused on incorporation of galactose ligands, the cell- phologies of hepatocytes on PVLA that adhered at high adhesion sites for synthetic polymers, into the design of concentrations of PVLA triggered spheroid formation of biomimetic polymers for hepatocytes. hepatocytes in the presence of epidermal Galactose ligand was first coupled to polyacrylamide (EGF), leading to enhanced differentiation functions [31]. (PAAm) by Weigel et al. [25]. PAAm were syn- In addition, due to its amphiphilic character, this polymer thesized by copolymerization of N-succinimidyl acrylate, is stably adsorbed onto the tissue culture dish [32]. acrylamide, and bisacrylamide. Next, 6-aminohexyl ␤-d- Lopina et al. synthesized galactose-modified star galactose was reacted with the activated copolymer poly(ethylene oxide) (PEO) hydrogels by reacting tre- through amide linkage [26]. Hepatocyte adhesion on the sylated PEO hydrogels with 1-amino-1-deoxy galactose galactosylated PAAm gels was verified [27]. The results [33]. They investigated hepatocyte attachment on these indicated that rat hepatocytes bound to galactosylated constructs. Hepatocytes were found to exhibit galactose- polyacrylamide in a Ca2+- and temperature-dependent specific adhesion to the galactose-modified PEO gels, manner via a patch of ASGPR in rat hepatocytes, but not adhering to gels with galactose via the spreading chicken hepatocytes [26,27]. Also, cell binding to the poly- morphologies of hepatocytes at low concentrations of mer was inhibited by asialo-orosomucoid, indicating a immobilized ligands, but not glucose [34]. receptor-mediated mechanism [28]; however, the effect Donati et al. synthesized galactosylated poly(styrene- of galactose-coupled PAAm gel hydration on hepatocyte co-maleic acid) (PSMA) by combining PSMA with 1-amino- ␤ d ␤ d adhesion was not investigated. 1-deoxy- - -galactose (or 1-amino-1-deoxy- - -lactose) 242 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

of the galactosylated PLGA surface in vitro were relatively low compared to those of other culture systems. However, using the perfusion culture system based on galactose- modified macroporous PLGA scaffolds under optimal flow conditions, the rate of albumin secretion was significantly increased due to spheroid formation of hepatocytes in the scaffolds [36]. Mao and colleagues immobilized galactose ligands on arylic acid-graft-copolymerized poly(ethylene tereph- thalate) (PET) film by plasma treatment and evaluated hepatocyte function [37]. The results indicated that hepato- cytes cultured on the galactosylated surface exhibited good albumin and urea synthesis, which were comparable to those of hepatocytes cultured on the -modified PET surface. Also, they prepared the galactosylated PET surface by surface-grafting poly(acrylic acid) on plasma-pretreated PET film to increase the density of the immobilized galac- tose ligands [37]. The results indicated that due to high ligand density, better maintenance of albumin secretion and urea synthesis function was obtained compared to culture on a collagen-coated surface. Further, they pre- pared galactosylated poly(vinylidene difluoride) (PVDF) by coating with galactose-tethered Pluronic and eval- uated hepatocyte function [38]. Results demonstrated that hepatocytes attached to galactosylated PVDF formed multi-cellular spheroids after one day of culture and exhib- ited higher albumin and P4501A1 detoxification than the unmodified PVDF membrane and collagen-coated sur- face. However, due to only hydrophobic interaction in the adsorption of galactose ligands, the stability of the coating after long culture periods is a substantial con- cern. Kang et al. prepared galactose-grafted PS surfaces by grafting of N-p-vinylbenzyl-4-O-␤-d-galactopyranosyl- d-gluconamide to a PS dish using oxygen plasma glow discharge treatment and evaluated hepatocyte adhesion [39]. The results indicated that compared to PVLA, hepa- tocytes adhered more slowly to the galactose-grafted PS surface than to PVLA during the first 2 h of incubation due to the low galactose density of the galactose-grafted PS. Higashiyama et al. prepared fructose- and galactose- modified polyamidoamine dendrimers by imine reaction between polyamidoamine dendrimer and galactose (or Fig. 1. Synthetic scheme of galactose-carrying polystyrene [29]. Copy- fructose) and evaluated hepatocyte function [40]. The right 1985, Nature Publishing Group, London, UK. results indicated that simultaneous modification of the dendrimer with fructose and galactose had a synergistic and testing HepG2 adhesion [35]. The results showed that effect on spheroid formation of hepatocytes, higher urea cell adhesion on galactosylated PSMA or lactosylated PSMA synthesis, and increased albumin gene expression than was approximately five times higher than that of PSMA those cultured on single-ligand modified dendrimers; the itself. ligands introduced in the dendrimer were not verified. Hepatocyte–ECM interaction is a surface phenomenon, Synthetic polymers derivatized with galactose ligands are and is affected by surface properties of ECM; however, summarized in Table 3. bulk properties of the ECM dictate the mechanical prop- erties of the ECM. Therefore, surface modification of the 2.2.2. Naturally-derived polymers ECM with galactose ligand is often necessary for opti- Naturally-derived polymers composed of proteins, mization of ECM for hepatic tissue engineering. Yoon polysaccharides, and nucleic acids have several inher- et al. synthesized galactosylated PLGA film or galacto- ent merits, including bioactivity, the ability to present sylated PLGA sponge by reacting the carboxylic acid of receptor-binding ligands to cells, biodegradability, and sus- PLGA at the terminal group and N-(amino butyl)-O-␤-d- ceptibility to natural remodeling, even though antigenicity, galactopyranosyl-(1 → 4)-d-glucoamide or by reacting the instability/deterioration, complexity of purification, and amino groups of PLGA-PEG and lactobionic acid for hepato- risk of disease transmission should be considered [41]. cyte culture [36], even though the albumin secretion rates This section discusses the liver tissue engineering applica- B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 243

Table 3 Synthetic polymers derivatized with galactose ligands.

Polymers Type of galactose Linkage Characters References

Poly (acryl amide) 6-aminohexyl ␤-d-galactose Amide Firstly coupled temperature [25–28] dependence of endocytosis, No results of effect of hydration on cell adhesion Polystyrene ␤-d-lactose Amide Amphiphilic and micellar, Spheroid [29–31] formation of hepatocytes, Non-degradable Poly (ethylene oxide) 1-amino-1-deoxy-␤-d-galactose SN1 reaction Radiation-crosslinked PEO star [33–34] hydrogel spreading of hepatocytes, Spatial distribution of galactose ligand Poly (styrene-co-maleic acid) 1-amino-1-deoxy-␤-d-galactose Amide Appearance of circular dichroism [35] 1-amino-1-deoxy-␤-d-lactose bands, No mechanism of cell adhesion Poly (lactic acid-co-glycolic acid) N-(aminobutyl)-O- ␤ Amide Introduction of PEG spacer, Spheroid [36] -d-galactopyranosyl-(1 → 4)-d- formation of hepatocytes, Perfusion glucoamide Lactobionic culture system acid Acrylic acid-graft-poly(ethylene 1-O-(6-aminohexyl)-d- Amide UV-induced graft copolymerization, [37] terephthalate) galactopyranoside Spheroid formation of hepatocytes, High albumin secretion and urea synthesis Poly(vinylidene difluoride) ␤-d-lactose Imine reaction Unstable coating due to physical [38] adsorption, Buffering of proteins by PEO Polysyrene ␤-d-lactose Ether Redox initiation, Graft polymerization [39] Polyamidoamine dendrimer Galactose+fructose Imine reaction Uncertain chemistry, Synergistic effect [40] by fructose tions of collagen and fibronectin, which are commonly used tissue engineering. Production of collagen/Poloxamine- in integrin-mediated hepatocyte adhesion, and another methacrylate hybrid networks exhibited the high stiffness group of naturally-derived polymers, including alginate, of the hybrid gels due to the crosslinked Poloxamine hydro- chitosan, gelatin, silk, and xyloglucan, which are regularly gels in addition to enhanced cell adhesion and high survival used in receptor-mediated hepatocyte adhesion. levels [44]. Shape and volume of ammonia-treated colla- Collagen is the most abundant protein in the body, and gen/chitosan sponges were constant over a period of 50 is responsible for maintenance of the structural integrity of days in enzyme solutions until 25 days postseeding [45], vertebrates and several other organisms [42]. More than 20 suggesting the potential of these artificial ECMs for use in different types of collagen have been identified thus far in liver tissue engineering. the human body; the most common collagen is Type I–IV. Fibronectin is a family of multifunctional ECM known Among these, Type I collagen is the single-most abundant to play key roles in cell attachment and migratory cellular protein in the body after cartilage [12]. Characteristics of behaviors, such as embryogenesis, malignancy, homeosta- collagen, such as good biodegradability, low antigenicity, sis, wounding healing, and maintenance of tissue integrity and cell-binding properties have made it a prospect for use [46]. The ability of fibronectin to serve as a substrate in ECMs for liver tissue engineering, even though collagen for cell adhesion is based on the RGD tripeptide [47]. use is limited due to its low mechanical strength, rapid However, due to the high cost, research on the use of biodegradation, and the high cost of pure collagen [12]. fibronectin in tissue engineering, except for the coating Collagen gels have been used for culture of hepatocytes of fibronectin on PS [48] and PLGA surfaces [49], has been for application to the bioartificial liver [43]. The results limited. However, significant research has been conducted indicated that the viability of rat hepatocytes cultured in on the role of fibronectin in regulation of switching from cylindrical collagen hydrogels in vitro remained stable with differentiation to growth [50], tubulin monomer growth stable albumin synthesis throughout the seven day incuba- [51], cell cycle status [52], cell spreading [53], and motil- tion period, and by 10 days culture, aggregated hepatocytes ity [54]. The effect of hepatocyte growth factor (HGF) [55] maintained an albumin production that was two-fold that in hepatocytes was studied for elucidation of its basic of single hepatocytes [43]. Recently, Zhao et al. used molecular biology. Bhadriraju et al. [48] compared hep- a collagen gel to generate engineered hepatic units for atocyte adhesion, growth, and differentiated function on reconstitution of a three-dimensional, vascularized hep- 2.3 kD and 73 kD RGD peptide-coated PS dishes to that on atic tissue in vivo [43]. Staking of hepatic units under the fibronectin-coated ones. Of particular interest, a 2.3 kD RGD subcutaneous space was found to result in significant cell peptide induced a round cell shape, enhanced differenti- engraftment with formation of a large fused hepatic sys- ated function, and inhibited DNA synthesis; however, the tem containing blood vessels, suggesting that this is an 73 kD RGD peptide induced cell spreading, dedifferentia- efficient method for engineering of hepatic tissue in vivo. tion, and enhanced DNA synthesis (similar to the impact To overcome the weak mechanical property of the colla- of fibronectin), suggesting that RGD peptide conformation gen gels, collagen/Poloxamine [44] and collagen/chitosan determines the specificity of cellular response. Fiegel et al. [45] hybrids were prepared as artificial ECMs for liver [49] investigated the effect of three-dimensional culture 244 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

Fig. 2. Reaction scheme of galactosylated alginate [58]. Copyright 2002, Elsevier Inc., Oxford, UK. of rat hepatocytes on fibronectin-coated PLGA scaffolds acids. Free amino groups in chitosan are easily modified under flow cell conditions. The results indicated that cellu- by covalent attachment of molecules using carbodiimide lar functions of hepatocytes, such as growth and albumin chemistry. Park et al. prepared galactosylated chitosan (GC) secretion, were increased significantly by fibronectin coat- by reacting lactobionic acid with amino groups of chitosan ing and flow conditions. using EDC/NHS as coupling agents (Fig. 3) and evaluated the Alginate, the primary structural component of brown subsequent hepatocyte adhesion [59]. The results indicated seaweed, is the monovalent form of alginic acid, and is a that hepatocytes adhered via galactose-specific recogni- linear polymer of ␤ (1 → 4)-linked d-mannuronic acid and tion between GC and ASGPR of hepatocytes, and hepato- (1 → 4)-linked l-guluronic aid [56]. Alginate is widely used cytes that adhered to the surface at high concentrations because it has low toxicity, is easy to modify chemically, of GC exhibited substantial spheroid formation after 24 h and undergoes mild ionotropic gelation in the presence of exposure to EGF. Yang et al. also prepared a hybrid artificial divalent cations; however, degradation of alginate hydro- ECM composed of alginate and GC because the mechanical gel occurs via a slow and unpredictable dissolution process properties of GC were insufficient [60,61]. Hybrid GC and in vivo [57]. Yang et al. [58] prepared galactosylated algi- alginate were found to improve cell adhesion and stability nate via the reaction of carboxylic acids of alginate and of the artificial ECM, which retained differentiated cellu- amine-modified lactobionic acid (Fig. 2) and evaluated lar functions better than alginate itself. Also, pore size of the liver function of hepatocytes in galactosylated algi- the artificial ECM could be controlled by the content and nate microcapsules. The results indicated that binding of molecular weight of GC and the freezing temperature of galactosylated alginate microcapsules with hepatocytes alginate/GC hydrogel. Introduction of heparin in the algi- occurred via the ASGPR patch on hepatocytes and that liver nate/GC artificial ECM enhanced the level of albumin secre- function of hepatocytes was enhanced compared to that in tion from hepatocytes in the presence of HGF due to rapid alginate treated hepatocytes due to spheroid formation of formation of stable spheroids in the alginate/GC/heparin hepatocytes in galactosylated alginate microcapsules. artificial ECM, which is closely related to connexin 32 and Chitosan is a linear polymer of ␤(1 → 4)-linked d- E-cadherin genes in cell-to-cell interactions [62]. glucosamine, and is obtained by deacetylation of chitin; it is Gelatin is a naturally-derived polymer obtained from suitable as a substrate for biomimetic glycopolymers due to collagen and has been used in medical and pharmaceutical the similarity of its structure to glycosaminoglycans found applications because of its biodegradability and biocom- in native tissue [57]; however, it is only soluble in diluted patibility [63]. Hong et al. prepared galactosylated gelatin B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 245

Fig. 3. Synthesis scheme of GC [59]. Copyright 2003, Elsevier Inc., Oxford, UK. by reacting lactobionic acid with amine-incorporated xylose and galactose units in the side chains. Seo et al. [67] gelatin (Fig. 4) and evaluated it for use in hepatocyte culture studied the specific interaction between galactose moieties [64]. The results indicated that the survival time of hep- in the side chain of the xyloglucan and ASGPR of hep- atocytes cultured on galactosylated gelatin artificial ECM atocytes. The results indicated that hepatocyte adhesion was longer than that of hepatocytes cultured on collagen- to xyloglucan-coated PS surfaces was dependent on the coated monolayers, and liver function (e.g. secretion of presence of Ca2+, and spheroidal hepatocytes formed at a albumin, synthesis of urea) was maintained due to the high concentration of xyloglucan in the presence of EGF specific interaction of galactose moieties in galactosylated due to galactose-specific recognition. In addition, alginate gelatin with ASGPR of hepatocytes. microcapsules prepared with xyloglucan as an artificial Silk is naturally produced by spiders and Lepdopterai. ECM enhanced multicellular spheroidal hepatocyte forma- Silk has been used in ligament tissue engineering due to tion due to the specific interaction between the galactose its unique mechanical properties and good biocompati- moieties of xyloglucan and ASGPR of hepatocytes [68]. bility [65]. Gotoh et al. [66] prepared galactosylated silk Naturally-derived polymers used for liver tissue engineer- fibroin by conjugation of lactose with silk using cyanuric ing are summarized in Table 4. chloride as a coupling spacer, and evaluated its ability to facilitate hepatocyte attachment. The results indicated that 2.3. Parameters affecting hepatocellular behavior hepatocyte attachment on galactosylated silk-coated PS dishes increased by approximately eight-fold compared to Long-term and stable hepatocyte culture systems are that of uncoated dishes, and hepatocytes exhibited a round very important for liver tissue engineering and for develop- morphology; however, the investigators did not assess ment of bioartificial liver devices. Morphology, attachment, alterations to liver function caused by three-dimensional growth, differentiation, and survival of hepatocytes are galactosylated silk fibroin. highly dependent on galactose density and microdistribu- Xyloglucan, a polysaccharide derived from tamarind tion of galactose in galactose-carrying polymers, structure seeds, is composed of glucose units in the main chain and of carbohydrate, topology of ECM, coculture, and cell 246 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

Fig. 4. Chemical structure of MGLA [a galactose-containing, modified gelatin (MG) that is modified with lactobionic acid (LA)] [64]. Copyright 2003, John Wiley & Sons, Inc, Hoboken, NJ, USA. source. This section addresses the effects of these parame- a round morphology at higher galactose concentrations ters on hepatocellular behavior. (100 ␮g/ml). The group did not determine an optimal con- centration of round hepatocytes. Ise et al. [69] reported that 2.3.1. Galactose density and microdistribution of hepatocytes adhered to galactose-carrying PS with coating galactose density lower than 20 ng/ml exhibited higher proliferative Ligand density is a primary determinant of hepato- ability than those attached to galactose-carrying PS with cyte attachment, morphology, and function; triantennary coating density higher than 50 ng/ml. Yin et al. [37] also molecules with three terminal galactose residues bind to reported formation of hepatocyte spheroids one day after lectins with higher affinity than biantennary ones with one cell seeding with better albumin secretion and urea synthe- or two terminal galactose residues, and a multi-subunit sis at a high galactose density of 513 nmol/cm2 on the PET receptor of hepatocytes is responsible for binding galactose surface. Kim et al. [70] proposed a mechanism for hepato- residues on desialylated glycoproteins [17]. cyte behavior according to the coating density of the PVLA Kobayashi et al. [30] studied the effects of galactose (Fig. 5). They explained that ASGPR on hepatocyte mem- density in galactose-carrying PS on the morphology and branes are clustered within sites of focal adhesion in a large differentiation of hepatocytes. The results indicated that patch, and that this clustering prevents participation of hepatocytes exhibited higher 3H-thymidine uptake (the integrin receptors during the adhesion process at high den- DNA synthesis indicator) with wider spread morphol- sity. Conversely, hepatocytes allow integrin receptors to ogy at the lower galactose density (0.5 ␮g/ml); however, participate in the adhesion process within the space where hepatocytes showed lower 3H-thymidine uptake with natural ECMs are secreted during cell culture. Thus, inte-

Table 4 Naturally-derived polymers for liver tissue engineering.

Polymers Type of galactose Linkage Characters References

Collagen RGD sequence Amide Weak mechanical property, Fast enzymatic biodegradation, High cost [42–45] Fibronectin RGD sequence Amide High cost, Coating of fibronectin for application [46–47] Alginate Lactobionic acid Amide Gelation, Eassy chemical modification, Unpredictable dissolution of [58] hydrogel in vivo Chitosan Lactobionic acid Amide Easy chemical modification, Difficult of control in deacetylation and [59–62] molecular weight Gelatin Lactobionic acid Amide Minimal inflammation, Clinically approved, Weak mechanical properties [64] Silk Lactose Ether Good mechanical properties, Good biocompatibility, Immune reaction by [66] residual sericin Xyloglucan Galactose – Biocompatible, Thermally-reversible gelation [67–68] B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 247

Fig. 5. Illustration of hepatocyte behavior on PVLA surfaces. ASGPR-dependent adhesion occurs initially within the interface between hepatocytes and the PVLA surface at both (A) high coat density and (B) low coat density of PVLA. However, ASGPR-independent adhesion takes place more rapidly at low coat density than at high coat density during culture. It is expected that ASGPR-independent adhesion would be induced by integrin receptors that have the opportunity to participate in cell adhesion mediated by ECMs secreted from hepatocytes following initial adhesion. Finally, the ASGPR-independent adhesion that occurs at low coat density causes hepatocytes to spread through integrin signaling. The concentrated ASGPR-ligand complexes at high coat density inhibit integrin signaling. [70] Copyright 2003, Elsevier Inc., Oxford, UK. grin receptors play a role in turning round morphologies tion than two-dimensional ones due to the provision of into spread ones and in focal adhesion kinase phosphory- better model systems for physiologic situations. Berthi- lation due to involvement of the integrin-mediated signal aume et al. [72] reported that hepatocytes cultured in pathway. a sandwich configuration of collagen exhibited a dra- The interaction between galactose ligand and ASGPR matic reorganization of the cytoskeleton, adoption of is not only affected by galactose density on the ECM but in vivo-like morphology and polarity, and expression of also by spatial microdistribution of galactose. Cho et al. a wide array of liver-specific functions due to the in [71] studied the effect of orientation on hepatocyte attach- vivo-like configuration; however, this configuration is ment to a PVLA surface prepared by the Langmuir-Blodgett difficult to scale-up and has nutrient transport limita- (LB) method. It was found that hepatocytes that adhered tions. ECM topology in the micro- and nanometer ranges to the LB surface of the polymer even at a very low galac- has been shown to affect cellular behavior, including tose density were round due to the spatial orientation of adhesion, migration, proliferation, and gene expression galactose; these cells were spread on the standard coated [73,74]. Ranucci et al. [75] reported that PLGA foams PVLA surface. Griffith et al. [34] also reported that spa- with subcellular size voids (approximately three microme- tial microdistribution of galactose in the ECM affected ters) induced two-dimensional hepatocyte reorganization, morphology and function of hepatocytes with the acces- whereas foams with subcellular size voids (approximately sibility of galactose clustered in spatial microdomains as 67 micrometers) promoted three-dimensional aggrega- the hepatocyte receptors organized into aggregated struc- tion due to enhanced cell-cell contacts. At intermediate tures within the cell membrane, suggesting that the spatial void sizes (approximately 17 micrometers), both two- microdistribution of the ligand in the ECM and control of dimensional and three-dimensional reorganization was the microenvironmental niche are important for regulation promoted, and hepatocytes that adhered to collagen of cellular behavior and successful tissue engineering. foams with a void size of 82 micrometers exhibited a high degree of spreading with high albumin secre- 2.3.2. Topology of artificial ECMs tion due to three-dimensional intercellular contacts Topology of artificial ECMs is another important [22]. parameter that influences cell morphology, function, and Park et al. [76,77] reported that hepatocytes cultured on physiological responsiveness due to modulation of cell microgrooved glass substrates with a 100 micrometer-high polarity by the topology of the ECM [72].Itiscom- channel between each substrate had significantly better monly accepted that three-dimensional artificial ECMs liver-specific function than those in substrates without more effectively induce differentiated hepatocyte func- microgrooves because hepatocytes were protected from 248 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 detrimental shear and maintenance of oxygen deliv- mechanical stimuli is of primary importance to success- ery. fully produce artificial cartilage for tissue repair. In this Glicklis et al. [78] reported that hepatocytes seeded section, artificial ECMs and suitable cell sources used in within three-dimensional porous alginate sponges with articular cartilage tissue engineering will be introduced, diameters of 100–150 micrometers exhibited more than and growth factors and mechanical stimuli used as cues for 90% hepatocyte aggregation within 24 h post-seeding due inducing differentiation to (Fig. 6) will also be to the non-adherent nature of alginate, and they secreted discussed. large amounts of albumin (60 ␮g/106 cells/day) within a Artificial ECMs ideally should furnish chondrocytes with week, even though they did not proliferate. optimal physical conditions mimicking the natural ECM Recently, a three-dimensional nanofibrous artificial microenvironment of cartilage. Construction of artificial ECM with high porosity and high spatial interconnec- ECMs involves fabrication of a three-dimensional network tivity prepared by electrospinning has been increasingly of artificial ECM similar to that of the original structure and applied for tissue engineering [79,80] because its struc- the provision for structural support and internal space for ture is similar to that of natural collagen fibers [81]; and the residing cells to adhere, proliferate, and differentiate. the nanofibrous artificial ECM can promote cell attach- Therefore, general requirements of three-dimensional arti- ment, proliferation, migration, and differentiation, and it ficial ECMs include high porosity, controlled degradation, can attract more integrin-binding proteins than film artifi- mechanical stiffness and strength, and biocompatibility, cial ECMs [82]. Chua et al. [83] reported that hepatocytes and they also should exhibit adequate surface properties cultured on galactosylated poly(␧-caprolactone-co-ethyl for proper tissue formation of chondrocytes [87]. Higher ethylene phosphate) nanofiber artificial ECMs exhibited porosity allows for the migration and proliferation of cellular function with the enhanced mechanical stability adhering cells and the facilitated exchange of nutrients of hepatocyte spheroids similar to that on the two- and waste products. Controlled biodegradability also is dimensional galactosylated substrate; however, the cell important so as not to hinder the formation of newly regen- morphologies were different. Feng et al. [84] reported that erating tissue within the artificial ECMs. Biocompatibility hepatocytes cultured on galactosylated chitosan nanofi- may facilitate cellular attachment and differentiation. brous artificial ECMs with an average diameter of 160 nm In most cases, biodegradable polymers including syn- formed flat, spheroidal hepatocytes and exhibited bet- thetic or natural polymers have been used in cartilage ter liver-specific function than spheroidal hepatocytes on tissue engineering. These polymers usually are formed galactosylated chitosan films. in sponges, hydrogels, or nanofibers. Recently, composite In summary, ECM for liver tissue engineering should be artificial ECMs composed of a mixture of different poly- designed to have high galactose density for achievement mers were designed to provide chondrocytes residing in of higher affinity in hepatocytes, to have three-dimensional the matrix with the combination of merits that each com- artificial ECMs in order to provide better model systems for ponent contributes. physiologic situations, and to form hepatocyte spheroids for enhancement of liver-specific function. 3.1. Natural material-based artificial ECMs

3. Cartilage Natural materials have been preferred over synthetic materials because of their intrinsic advantages of bio- The connective tissue comprising articular cartilage of compatibility, biodegradability, and an improved capacity the knee is highly specialized for reducing joint friction for cell attachment. They can be divided in two groups: at the interface of two long , and it merely con- protein-based artificial ECMs (e.g. collagen, fibrin, silk) tains chondrocytes in an avascular structure [85]. The and carbohydrate-based artificial ECMs (e.g. hyaluronan, regeneration of damaged articular cartilage remains chal- alginate, agarose, chitosan). Applications of these natural lenging due to its poor intrinsic capacity for repair. Thus materials-based artificial ECMs for in vitro and in vivo car- far, no surgical procedure has been able to reproduce tilage tissue engineering are summarized in Table 5. the biological composition and biomechanical properties Collagen is a natural component of cartilage and is of native cartilage [86]. The specific treatment strategies known to play an important role in cellular adhesion and have been dictated by the nature or size of lesions and differentiation through specific interactions between lig- the preference of the operating surgeon [86]. However, ands on collagen chains and adhering cells [86]. Therefore, the advent of tissue engineering has provided revolu- it is widely used for engineering artificial tissue in a broad tionary potential for treating cartilage-related diseases. It spectrum of organs. Type II collagen also plays an essential is believed that tissue engineering of articular cartilage role in the maintenance of function. Bovine will overcome the current limitations of surgical treat- BMSCs seeded on type II collagen represent the most ment by offering functional regeneration in the defect prominent phenotype of chondrocyte differentiation with region. This technology involves ex vivo culturing of the addition of transforming growth factor (TGF)-␤1ina chondrocytes from autologous or allogenic sources in time-dependent manner [88]. In addition, chondrogenic ECM-based constructs and subsequent implantation into differentiation only was detected in three-dimensional the cartilage defect. Although recent progress has been hydrogels, not in monolayer cultures. In another in vitro made to engineer artificial cartilage via tissue engineering test, it was found that a collagen sponge could provide , the challenges remain significant. Selection of the proper a superior microenvironment for the formation of ECM, artificial ECMs and incorporation of growth factors or such as proteoglycans, when human B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 249

Fig. 6. Components required for chondrocyte tissue engineering [modified from Ref. [86]].

Table 5 Application of natural materials-based scaffolds for cartilage tissue engineering.

Scaffold Component Forms Seeding cells Biofactors Model/target site Results References

Collagen A bilayer Autologous Human/Deep cartilage Hyaline-like cartilage [93] membrane chondrocyte defects formation at the defect Collagen A bilayer Autologous Sheep/chondral defects Reparative tissue [91] membrane chondrocyte formation Fibrin Hydrogel Autologous ADSC Rabbit/chondral defects Hyaline-like cartilage [135] formation at 8-week follow-up complete healing to subchondral bone Alginate/agarose Hydrogel Autologous ADSC In vitro Enhanced chondrogenic [90] differentiation Alginate/agarose Hydrogel Autologous Human/osteochondral Predominant hyaline [106] chondrocyte defects cartilage formation at 2-year follow-up Alginate Hydrogel Allogeneic BMSC Rabbit/Osteochondral Increase in aggrecan and [105] defects type II collagen Chitosan Hydrogel Autologous Whole Sheep/Chondral defects Significantly more hyaline [108] blood repair tissue formation Chitosan Hydrogel Autologous Whole Rabbit/Bilateral trochlear Formation of a more [109] blood defects integrated and hyaline repair tissue complete restoration of GAG levels HA Sponge Autologous BMSC FGF Rabbit/knee Formation of cartilage [101] osteochondral defect tissue with expressed type II collagen HA esterified Sponge Autologous Human/chronic cartilage Continued increase of [103] derivative chondrocyte lesions clinical performance No major adverse events during 3-year follow-up HPMC Hydrogel human nasal Nude mice/subcutaneous Formation of a [110] chondrocytes implantation cartilage-like tissue with GAG and type II collagen Peptide amphiphile Hydrogel hMSC TGF␤-1 Rabbit/full thickness Sustained release of [109] chondral defect TGF␤-1 from peptide hydrogel Promoted regeneration of articular cartilage

HA, Hyaluronic acid; HPMC, hydroxypropyl methylcellulose; PLGA, poly(lactic acid-co-glycolic acid); PEG, poly(ethylene glycol); BMSC, bone-marrow stromal cells; ADSC, adipose-derived stem cells; ESC, embryonic stem cell; FGF, fibroblast growth factor; IGF-1, insulin-like growth factor-I; GAG, gly- cosaminoglycan; hMSC, human mesenchymal stem cells. 250 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 cells were cultured [89]. Artificial ECMs formed from type that of collagen gel [96]. These characteristics also were II collagen were prepared for neo-cartilage synthesis by observed with bone marrow-derived mesenchymal stem crosslinking with genipin, because ECM in cartilage is pre- cells (BMSCs) seeded on silk fibroin artificial ECMs [97]. dominantly composed of type II collagen and networks In particular, BMSCs proliferated more rapidly on the silk of proteoglycans (PG), such as hyaluronic acid (HA) and fibroin artificial ECM, and homogeneous distribution of chondroitin sulfate (CS). Their study assessed the effect of cartilage-like tissue was achieved only with silk artificial glycosaminoglycans (GAGs) added to the culture medium ECM samples among silk, collagen, and cross-linked colla- on proliferation and matrix synthesis of human chondro- gen artificial ECMs. Use of exogenous silk protein for tissue cytes [90]. The results indicated that the addition of CS engineering might be problematic due to its stimulation and HA further up-regulated gene expression of aggrecan of immune responses, such as induction of significant TNF and collagen II in chondrocytes. In an animal experiment release from macrophages [98]. with sheep, cultured autologous chondrocytes seeded in a Hyaluronan is a primary physiological component of porcine collagen I-III bilayer membrane were implanted in ECMs in articular cartilage, which can be chondrogenic chondral defects [91]. Microfracture treatment was used to mesenchymal stem cells. In a transplantation study to enhance the healing response, and it was reported that with autologous BMSCs embedded in a hyaluronic acid combination treatment of microfracture and chondrocyte- sponge, BMSC-loaded artificial ECMs were implanted in seeded collagen membranes further improved the overall full-thickness osteochondral defects of the rabbit knee [99]. cartilage repair. A clinical trial also was performed to Histological findings revealed that newly formed carti- check the efficacy of matrix-induced autologous chondro- lage tissue at the implantation site was very similar to cyte implantation, in which autologous chondrocytes were the surrounding normal tissue, which is much better than seeded on three-dimensional membranes of a bilayer type that of the untreated group. It is usually chemically mod- I-III collagen for implantation in deep cartilage defects [92]. ified for easy fabrication into solid artificial ECMs because The presence of hyaline-like cartilage was observed at the of its highly hydroscopic nature [100]. The most com- implantation site via magnetic resonance, and chondrob- mon form of the modification is the esterification of a lasts and type II collagen also were observed inside the carboxylic acid present at the C6 position of hyaluronic collagen membrane. acid. The benzyl ester of hyaluronic acid has been com- The fibrin artificial ECM is a network of fibrous pro- mercialized as Hayff-11 and tested clinically. Autologous teins naturally polymerized from fibrinogen in the process chondrocytes isolated from patients were seeded on Hyaff- of blood clotting in response to injury, and it forms a 11 and implanted in the chronic cartilage lesions of their natural, artificial ECM mesh. It was reported that this three- knees [101]. There were no major adverse events reported dimensional structure of fibrin can be used as a gel-type during three-year follow-up, and clinical outcomes were artificial ECM to encapsulate cells for delivery [86]. Sev- prominent with the treated group. However, the chem- eral animal experiments were performed using fibrin glue ical modification of hyaluronic acid might compromise artificial ECMs to evaluate the therapeutic effects of autolo- its biocompatible property. In addition, the degrada- gous chondrocytes, mesenchymal stem cells, and bioactive tion of cross-linked hyaluronan caused chondrolysis molecules on cartilage repair. For autologous adipose- [102]. derived stem cells (ADSCs) loaded in a fibrin artificial ECM Alginate is an anionic polysaccharide, which forms a for implantation in the treatment of full-thickness carti- hydrogel instantly in the presence of divalent cations, such lage defects of rabbits, immunostaining, Western blotting, as calcium ions. Alginate beads encapsulating cells are com- reverse transcriptase polymerase chain reaction, and quan- monly produced when cells dispersed in alginate solution titative assessment revealed that articular surface defects, are dropped into a calcium chloride solution. In an in vitro treated with an ADSC-loaded artificial ECM, healed with study using rabbit BMSCs, it was reported that alginate and hyaline-like cartilage; however, very little healing was agarose gels containing BMSCs induced a greater increase observed in the control group. Cartilage markers including in the expression of cartilage-specific markers, such as aggrecan and collagen type II mRNA also were identified in aggrecan and type II collagen, than cells in type I collagen the treated group. However, concerns regarding immuno- gels [103]. Furthermore, when rabbit BMSCs encapsulated logical reactions have been raised; if autologous fibrin is in alginate beads were deployed in the cartilage defects of not used, the method requires additional procedures and rabbit, the beads remained satisfactorily within the defect time necessary for blood collection [93]. regions, which were progressively replaced by the regen- Silk is an artificial ECM that has been studied by sev- erating tissue. Histological findings showed that viable eral groups because of its high mechanical strength in the chondrogenic cells filled in the defects. In a clinical study wet state and its low inflammatory properties [94]. It also following 17 patients (inclusion criteria: an isolated lesion was reported that fibroin, a primary component of silk, of the femoral condyle (grades III and IV)), significant demonstrates good cell attachment and biocompatibility. improvement was observed in patients with lesions larger Wang et al. reported that human chondrocytes seeded in than 3 cm2 when autologous chondrocytes were isolated silk fibroin artificial ECMs could be re-differentiated after and suspended in an alginate-agarose mixture solution and expansion on culturing dishes and deposited in cartilage- subsequently implanted in the lesions of patients [104]. specific ECM for in vitro cartilage tissue engineering [95]. However, alginate gel suffers from instability in physio- In another in vitro experiment, better-defined cartilage tis- logical solution due to a loss in mechanical strength and sue with increased GAG amounts and cell density were integrity induced by the replacement of divalent calcium observed in silk fibroin sponge hydrogels compared to ions with monovalent sodium or potassium ions. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 251

Chitosan also is preferred in articular tissue engineer- ing applications because of its structural similarity to GAG components in cartilage-stimulating chondrogenesis [124] [105]. In a study characterizing cartilage repair in sheep, [111] [112] [113] [121] [118] [119] [122] [123] a chitosan-glycerol phosphate solution was mixed with autologous whole blood from sheep to facilitate the healing process in cartilage defects with blood clots [106]. Inter- estingly, the mixture of chitosan-glycerol phosphate and whole blood showed increased adhesion to walls of the cartilage defect region and significant repair in hyaline car- tilage at a 6-month follow-up. In a subsequent study with a rabbit model, it was reported that beneficial blood clot for- mation could be stabilized by chitosan-glycerol phosphate solution via inhibition of clot retraction [107]. Since chi- tosan possesses primary amino groups at the C2 position of each repeating unit, which is only protonable below pH 6.5, this unique characteristic allows for convenient processing response Synthesis of collagen types I andaggrecan, II, and Gene expression of cartilage-related markers Mature found after 10 weeks In vitro culturing for 4 weeks;cartilage-like formation tissue of Synthesis of higher type II collagen observed 8 weeks after implantation Natural chondrocyte morphology and abundant cartilaginous ECM deposition, the expression of type II collagenaggrecan and Cartilaginous tissue formation in hybrid construct; Intense type II collagen formation at pericullular region. in artificial ECMs under mild conditions. The polycationic Cartilaginous tissue formation Presence of proteoglycan, GAG, and type II collagen property of chitosan can be exploited to form a more stable ionic complex with polyanionic macromolecules includ- ing alginate, GAGs, synthetic poly(acrylic acid), and smaller anionic molecules. Chitosan microspheres were fabricated by drop-wise addition of anionic tripolyphosphate in a chitosan emulsion containing TGF-␤1 for the tissue engi- neering of articular cartilage [108]. Controlled release of a bioactive growth factor was achieved from ionically Goat/Full cartilage defect Significantly better cartilage repair In vitro In vitro Swine/subcutaneous implantation In vitro Nude mice/subcutaneous implantation Nude mice/subcutaneous implantation Nude mice/subcutaneous implantation crosslinked chitosan microspheres, and in vitro culture of Nude mice/subcutaneous implantation porcine chondrocytes on the microspheres demonstrated a significant increase in cell proliferation and ECM produc- -3 ␤ tion. Recently, a novel peptide amphiphile (PA) with a TGF-

␤1-binding peptide domain was explored for cartilage 1 ␤ regeneration applications [109]. This self-assembling PA chondrocyte formed supramolecular nanofibers upon the addition of dexamethasone, ascorbate, TGF charge-shielding ions, which resulted in a hydrogel dis- playing a high density of exposed TGF-␤1-binding peptide for efficient capture and release of TGF-␤1 growth fac- tor. When human mesenchymal stem cells (hMSCs) were suspended in PA solution and injected in a full-thickness, Rabbit chondrocyte Bovine chondrocytes BMSC and ESC TGF- chondrocyte Bovine chondrocyte chondrocyte chondrocyte chondrocytes articular cartilage rabbit model with microfracture, the data demonstrated improved survival and increased chon- drogenic differentiation of hMSCs. Furthermore, PA itself, even without TGF-␤1 growth factor, promoted the regen- eration of articular cartilage in a full-thickness chondral defect, suggesting the possibility that TGF-␤1-binding domains in PA might serve as a reservoir for attracting and hydrogel releasing the endogenous TGF-␤1. hydrogel hydrogel hydrogel

3.2. Synthetic material-based artificial ECMs

Synthetic materials including poly(lactic acid) (PLA), poly(glycolic acid) (PGA), and their copolymer, poly(lactic acid-co-glycolic acid) (PLGA), have been tested for cartilage p(NiPAAm-co-AAc) Thermo-responsive PEGPEGPluronic F-127p(NiPAAm-co-AAc) Crosslinked Hydrogel Thermo-responsive Crosslinked PLGA/collagen Ear cartilage meshPLGA/fibrinMPEG-PLGA/fibrin Sponge Sponge Bovine Rabbit Goat autologous PLGA Sponge Rabbit tissue engineering potential (Table 6). They are thought to have advantages including ease in molding and the ability to design a degradation rate to match tissue growth into the artificial ECM. However, these synthetic materials are not as preferred for cartilage tissue engineering as materials of natural origin, because the acidic byproducts gener- materials-based ated during the degradation process cause an inflammation materials-based Synthetic ScaffoldSynthetic Scaffold component Forms Seeding cells Biofactors Model/target site Results References

reaction, giant cell reaction, and acute chondrocyte death Table 6 Application of synthetic materials or composite-based scaffolds for cartilage tissue engineering. 252 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 due to the abrupt drop in pH in the local microenviron- b-PLA macromer with non-degradable PEG macromers at ment inside the engineered cartilage. The closed cartilage pre-determined mixing ratios [115]. It was found that environment deters rapid clearance of acidic byproducts chondrocytes encapsulated in a PEG hydrogel with a higher from the degrading synthetic artificial ECMs, eventually degradation rate produced more DNA content and type II inducing undesirable side reactions in cartilage. Another collagen, indicating the important role of hydrogel degra- disadvantage of these materials is their poor cell attach- dation in controlling and influencing the deposition and ment. Commonly, these materials were modified to possess distribution of ECM molecules. The photopolymerized PEG an anchorage site for cell adhesion or mixed with other hydrogel also can be functionalized for cellular anchorage naturally-derived materials [110]. by incorporating a cell adhesion peptide, such as Arg-Gly- Asp (RGD), at PEG chain terminals during polymerization 3.2.1. Poly(ethylene glycol) (PEG) hydrogels [116]. Recently, hydrogel-based artificial ECMs have gained greater attention for cartilage tissue engineering applica- 3.2.2. Injectable hydrogels tions because of their similarity to the natural cartilage In situ injectable hydrogel systems have generated environment. Hydrogels contain high content sim- increased interest for cartilage repair applications. They ilar to that of natural cartilage, which serves as a can be injected with encapsulated cells and/or bioactive suitable environment for chondrocytes. These materials materials of interest into the cartilage defect in a minimally are composed of synthetic or natural-based hydrophilic invasive manner and easily can fill the three-dimensional biomaterials cross-linked by physical, ionic, or chemical shape of the defect for facilitated integration with the addi- interactions. They also can be injected transcutaneously tional mechanical property of temperature-dependent sol- into the defect region of the joint, which avoids invasive gel transition. Poly(N-isopropylacrylamide) (PNIPAAm) surgery required for the implantation of prefabricated arti- exhibits reversible phase separation with a lower critical ficial ECMs. PEG is a popular biocompatible hydrophilic solution temperature (LCST) of approximately 32 ◦C. Thus, polymer approved by the FDA, and it has been exten- chondrocytes cells can be dispersed in PNIPAAm solution sively explored for the use of formulating hydrogels to at room temperature (RT) lower than LCST and injected encapsulate bioactive drugs or cells. It is thought that into the cartilage defect for in situ gelation. When iso- crosslinked PEG hydrogel may provide a better environ- lated bovine articular chondrocytes were seeded into the ment for culturing chondrocytes due to its high water loosely crosslinked copolymer hydrogel of NIPAAm and content and mechanical strength, because chondrocytes acrylic acid, p(NIPAAm-co-AAc), and incubated at 37 ◦Cin are surrounded by hydrophilic ECM components in high vitro, the viability of seeded chondrocytes was maintained abundance. In one approach to minimize invasiveness dur- for up to four weeks, and cartilage-like tissue formation ing treatment of the cartilage defect while maintaining was observed in the matrices [117]. In an in vivo study its mechanical and biological properties, photopolymeriza- using this thermo-reversible hydrogel, a larger volume of tion was employed using a PEG macromer polymerizable at cartilage-associated ECM was observed with the addition of both ends and a photoinitiator, followed by exposure of the dexamethasone, ascorbate, and TGF ␤-3 eight weeks after solution to UV light [111]. Photopolymerized PEG hydrogel subcutaneous implantation in nude mice [118]. However, is an attractive material for tissue engineering applications the temperature-dependent phase transition might hinder because of its cell-friendly properties and in situ gelation the thermo-responsive gel from being injected through a upon exposure to UV. In this study, a PEG macromer was long needle or catheter to the desired site in the human mixed with isolated allogenic chondrocytes in the pres- body, because the warmed needle or catheter can even- ence of a photoinitiator. The photopolymerization reaction tually cause premature gelation during injection before can be performed at physiological pH and temperature, reaching the target site. In another approach to inhibit which allows for safe encapsulation of cells and proteins. this pre-mature gelation, a p(NIPAAm-co- vinylimidazole) Chondrocyte viability in this hydrogel was maintained [p(NIPAAm-co-VI)] copolymer hydrogel that responds to with characteristic gene expression for up to three days both temperature and ionic strength was used [119]. Rabbit of in vitro culture. In another study, cell binding peptide chondrocytes were embedded in this composite hydrogel was conjugated to one end of a bifunctional PEG macromer with TGF ␤-1-loaded nanoparticles. Chondrocyte-specific and mixed with PEG diacrylate in the presence of a pho- ECMs were observed eight weeks after subcutaneous toinitiator for crosslinking. When BMSCs and embryonic implantation of the hydrogel in nude mice. stem cell-derived cells (ESCs) were encapsulated in this Pluronic F-127 consists of approximately 70% ethylene hydrogel with a mechanical stimulus, gene expression of oxide and 30% propylene oxide by weight and is avail- cartilage-related markers, such as Sox-9, type II collagen, able commercially [120]. It transforms to hydrogel at RT and aggrecan, was noticed [112]. However, it was deter- at concentrations of 20% or higher. After administration in mined that highly crosslinked PEG hydrogel might hinder vivo, it can be slowly dissolved and cleared by renal and the proliferation and proteoglycan synthesis of encapsu- biliary excretion. Although several trials have been con- lated chondrocytes [113,114]. As the seeded cells grow and ducted using Pluronic F-127, poor mechanical properties form new tissue inside the hydrogel, the scaffold should hinder its application for tissue engineering due to a diffi- degrade accordingly. The effect of the degradation rate of culty maintaining the constructed shape. Thus, composite hydrogel on the engineered cartilage formation was evalu- artificial ECMs that compensate for the poor mechanical ated using hydrogels of different degradation rates, which properties of Pluronic F-127 have been preferred, which were prepared by copolymerizing a degradable PLA-b-PEG- will be described later. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 253

3.3. Composite artificial ECMs 4 weeks. Chondrocytes cultured in this hybrid artificial ECM could be harvested easily from the artificial ECM by A few specific characteristics of single ECM components simply lowering the temperature to RT without any addi- might be insufficient to create the optimal environment tional treatment by proteolytic enzymes, and the harvested to mimic the natural proliferation and differentiation chondrocytes exhibited a round morphology, indicative of of chondrocytes in cartilage. Therefore, in most cases maintenance of the chondrocyte phenotype during culture a combination of multiple components is desired to in the hybrid artificial ECM. address various features required for culturing chondro- cytes. Most synthetic material-based artificial ECMs suffer 4. Bone from poor anchorage of chondrocytes or stem cells, and incorporation of natural ECMs including collagen, fib- 4.1. Artificial ECMs for bone tissue engineering rin, and hyaluronic acid in synthetic artificial ECMs for cell attachment has been a popular approach studied 4.1.1. Design criteria of artificial ECMs for bone tissue extensively. Synthetic materials provide relatively high engineering mechanical strength with a tunable degradation rate, A key component in bone tissue engineering is an arti- whereas their hydrophobicity and lack of cellular anchor- ficial ECM that serves as a template for cell interactions age sites are drawbacks for their application in tissue and the formation of bone ECM to provide structural sup- engineering. Conversely, naturally-derived ECM polymers port for the newly formed tissue. The artificial ECMs for support excellent cellular adhesion and growth due to bone tissue engineering should meet several criteria to their specific cell interaction peptides and hydrophilic- serve this function including possession of mechanical ity, even though their weak mechanical properties make properties similar to those of the bone repair site, biocom- it difficult to use them in load-bearing regions, such patibility, biodegradability at a rate commensurate with as cartilage. Thus, these characteristics of naturally- or remodeling, and porosity of artificial ECMs that allows synthetically-originated materials drive us to attempt a migration and proliferation of osteoblasts and mesenchy- combination of the two materials in order to afford mal cells as well as vascularization. In addition to these higher mechanical strength, tunable degradation, and cel- criteria, artificial ECMs should be able to regulate cellu- lular attachment [121]. When collagen microsponges were lar functions even though several reviews have focused embedded in a PLGA mesh and transplanted subcuta- on artificial ECM properties including mechanical proper- neously into nude mice for chondrocyte tissue engineering, ties [126,127], degradation [128,129], porosity [130], and homogeneous cell distribution, natural chondrocyte mor- osteoconductive activity of the artificial ECM [131,132] for phology, and abundant cartilaginous ECM deposition were bone tissue engineering. In this section, we reviewed stud- observed. As a biomimetic approach, the active cell ies that focused on the regulation of cellular function by anchorage domain of collagen (e.g. RGD) was conjugated artificial ECMs after the introduction of materials used in chemically onto the surface of synthetic material-based bone tissue engineering. In addition to polymeric materi- artificial ECMs [122]. Chondrocytes seeded on RGD- als, non-polymeric materials were also used to fabricate modified PLGA/gelatin microspheres demonstrated the artificial ECMs for bone tissue engineering because bone is best attachment, proliferation, viability, and sulfated GAG a hard tissue. secretion. Another purpose of hybrid artificial ECMs is to incor- porate thermally or chemically responsive components 4.1.2. Non-polymeric materials in natural or synthetic material-based artificial ECMs to Non-polymeric materials including metallic and increase the bioavailability of seeded cells while min- ceramic materials have been widely used for bone tis- imizing their leakage out of the artificial ECMs. Fibrin sue engineering. Generally, the mechanical properties possesses a chemically active gelling property, which typ- of metallic materials (e.g. stainless steel, cobalt-chrome ically occurs during blood coagulation in addition to the based alloys, titanium (Ti) alloys) are superior to those of advantages described previously. Several studies investi- polymeric ones. However, there are several disadvantages, gating a fibrin-based hybrid artificial ECM demonstrated such as the lack of tissue adherence, low rate of degra- its potential for the promotion of homogeneous cell dis- dation, toxicity due to accumulation of metal ions due tribution and cartilaginous tissue formation [110,123,124]. to corrosion, and mismatch of Young’s moduli between PNIPAAm also was used for this purpose, and it is a metallic materials and bone [133–135]. Furthermore, the synthetic thermoresponsive polymer for tissue engineer- Ti alloys are not osteoconductive [136]. Ceramic materials ing with versatile manipulation of its chemical structure such as calcium phosphate also have been used exten- and molecular weight as well as easy conjugation with sively in bone tissue engineering because they are natural other components. Chondrocytes could be seeded easily components of bone tissue [137–140]. Calcium phosphate into the PLA-PNIPAAm hybrid artificial ECMs at RT, at can be used in permanent implants or as biodegradable which the hydrodynamic structure of PNIPAAm was shrunk artificial ECMs due to its osteoconductive and osteoin- to create larger pore sizes for facile cell loading [125]. ductive properties as well as mechanical compatibility Upon incubation at 37 ◦C, the PNIPAAm was rehydrated with native bone. Although calcium phosphate has several to provide a more hydrophilic and chondrocyte-friendly attractive properties, its mechanical properties are poor microenvironment, and its swelling property at physio- [126]. Therefore, its clinical use is limited to non-weight logical temperature in PBS solution was maintained for bearing sites. 254 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

4.1.3. Synthetic polymeric materials improved interactions with osteoblastic cells and pro- Compared to metallic and ceramic materials, polymeric moted osteogenic activity (e.g. osteocalcin expression) in materials are materials in which it is easy to chemi- vitro compared to that of a non-coated Ti-based artificial cally incorporate moieties that regulate cellular functions. ECM [148]. Brittleness of calcium phosphate is a major Also, synthetic polymers can be produced under con- problem for its implementation as an artificial ECM for trolled conditions and, therefore, exhibit predictable and bone tissue engineering. A hydroxyapatite-based artifi- reproducible mechanical and physical properties, such cial ECM was coated with a hydroxyapatite/PCL composite as tensile strength, elastic modulus, and degradation and demonstrated increased compressive strength and rate. Moreover, three-dimensional, biodegradable, syn- elastic modulus compared to an uncoated hydroxyapatite- thetic polymeric systems are of particular interest because based ECM [149]. In addition, a collagen/hydroxyapatite their porosity, hydrophilicity, and degradation time can be composite artificial ECM was developed to support the varied with a high degree of reproducibility. For bone tissue attachment and proliferation of rabbit periosteal cells engineering, there are several commonly used synthetic [150]. Furthermore, a PLGA sponge hybridized with col- polymers, such as PLA, PGA, poly(␧-caprolactone) (PCL), lagen microsponges on which apatite particulates were and PLGA [100,140–143]. To fabricate these polymers deposited is an example of a three-component artificial into three-dimensional artificial ECMs, there are several ECM for bone tissue engineering [151]. techniques including salt-leaching, gas foaming, electro- spinning, and emulsion polymerization [130]. PLGA-based 4.2. Regulation of cellular functions by artificial ECMs artificial ECMs fabricated via consolidation by pressure drop have been used as teeth implants [141]. PLGA/poly 4.2.1. Importance of regulation of cellular function by (vinyl alcohol) blend fabricated via a salt-leaching tech- artificial ECMs nique were used as an artificial ECM for skull bone defect When materials were used for bone regeneration, the treatment [142] even though they have been applied in sole criterion was “to achieve a suitable combination of areas of low mechanical stress in vivo because of their low physical properties to match those of the replaced bone tis- mechanical properties [143]. sue with a minimal toxic response of the host” [152,153]. However, all bone tissue functions cannot be mimicked by 4.1.4. Naturally-derived polymers the materials currently used for bone regeneration because Naturally-derived polymers (e.g. collagen, hyaluronic they are unstable and cannot work well in vivo after long acid, glycosaminoglycans) have the advantages of biocom- periods of implantation. To reconstruct new bone tissue, patibility and biodegradability because they are composed regulation of cellular function is necessary. There are sev- of the structural materials of tissues. Kim et al. used eral approaches, such as the application of growth factors hyaluronic acid-based hydrogels containing bone morpho- and mechanical stimuli that in turn regulate cellular func- genetic protein-2 (BMP-2) and human mesenchymal stem tion. In addition, artificial ECMs play a central role in the cells to treat a rat calvarial defect model [144]. Collagen- regulation of cellular function because the artificial ECMs based artificial ECMs also have been used for bone tissue respond to growth factors and transmit mechanical stim- engineering [145,146]. Type I collagen is a major com- uli. Therefore, the design of artificial ECMs will become ponents of in vivo bone ECM and leads to new bone a primary issue for the regulation of cellular function for from stem/progenitor cells via the developmental cascade. bone tissue engineering applications in the future. For bone Therefore, type I collagen might be a good candidate mate- tissue engineering applications using cells, osteoblasts rial for a biomimetic approach to design artificial ECMs for and mesenchymal stem cells (MSCs) are important cell bone tissue engineering. In addition to hyaluronic acid and sources. In this section, we primarily discussed materi- type I collagen, silk fibroin also has been used as the arti- als for the regulation of osteoblast function. The materials ficial ECM in bone tissue engineering [147]. Salt-leaching for bone tissue engineering using MSCs are reviewed in techniques and freeze-drying techniques are commonly Section 6. used to fabricate these materials. Since they are generally mechanically weak, it might be difficult for natural materi- 4.2.2. Regulation of osteoblast functions als to be used as the bone tissue engineering artificial ECMs To regulate cellular function using artificial matri- by themselves even though they have high biocompatibil- ces, mimicking native ECM is a well established, reliable ity and biodegradability. approach. To mimic native ECM in bone, ECM proteins secreted from osteoblasts or MSCs are used for the 4.1.5. Composite materials regulation of cell function [154–158]. Using these mate- Even though each individual material has advantages rials, it is possible to regulate cellular functions, such as for bone tissue engineering applications, each material proliferation and differentiation. To mimic native bone discussed has disadvantages in various properties includ- ECM more efficiently, fragments of the ECM proteins are ing the brittleness of calcium phosphate and the inferior incorporated within polymeric materials. Type I collagen mechanical properties of natural and synthetic polymers. has been used widely as an artificial ECM, mimicking The combination of different materials to form compos- in vivo ECM in bone tissue engineering applications. The ites has led to overcoming the disadvantages of any one osteoblasts cultured on type I collagen matrix expressed particular material. For example, to develop Ti-based arti- an osteoblastic phenotype at higher levels than those cells ficial ECMs with osteoconductive activity, artificial ECM on plastic surfaces [159]. These cells recognize type I col- surface was coated with a hydroxyapatite layer which lagen by the cell surface receptor, integrin ␣2␤1 [160], B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 255 where the integrin receptor activated intracellular sig- In addition to the physical combination of indepen- naling pathways to express the osteoblastic phenotype. dently engineered cartilage and bone tissues, reconstruc- To confirm this mechanism, synthetic peptides also have tion of osteochondral composites in a multi-phasic artificial been used. When integrins bind to the ECM molecules, ECM has been explored. Scheck et al. developed a biphasic integrins recognize the biologically significant peptide artificial ECM composed of PLA and HA. At the interface region, Arg-Gly-Asp (RGD). Hu et al. developed an RGD between PLA and HA, a thin PGA film was deposited to pre- peptide-modified PLA film [161] on which they demon- vent cell migration between the components [169]. Porcine strated enhanced osteoblast attachment. Also, alkaline articular chondrocytes were seeded in the PLA region of phosphatase activity and calcium deposition increased on this artificial ECM (cartilage layer), while human gingi- this film, indicating that the osteogenic phenotype was val fibroblasts transduced with an adenovirus expressing expressed. In addition to the RGD peptide, a collagen BMP-7 were seeded in the HA region of the artificial ECM peptide motif also was applied to the osteoblast culture (bone layer). Following ectopic implantation in mice, the [162]. The ECM also regulated the activity of cartilage layer contained ECM rich in glycosaminoglycans. by increasing their accessibility to receptors [155,156]. In contrast, blood vessels were observed in bone layer. Therefore, the incorporation of cytokines is one of the Moreover, a mineralized interface was often present at the most useful approaches for regulating cellular function junction. Kon et al. also investigated multilayered gradient [163,164]. artificial ECMs composed of HA and type I collagen [170]. They divided the artificial ECM into three parts (cartilage, transition, and bone regions). According to their results, 4.3. Artificial ECMs for osteochondral defects the compositions of type I collagen and HA varied in each region (cartilage region: 100% type I collagen; transition Osteochondral defects are often associated with region: 40% HA and 60% type I collagen; and bone region: mechanical instability of the joint and, therefore, with 70% HA and 30% type I collagen). In this case, chondrocytes the risk of causing osteoarthritic degenerative changes. were seeded only in the cartilage region and were trans- Grafting of osteochondral units with both chondral and planted in the osteochondral defect model. The regenera- osseous layers represents a promising approach to restore tion of bone and cartilage was significantly enhanced with the biological and mechanical functionalities of the joint. the artificial ECMs compared to that of the control group. To transplant osteochondral units, autologous osteochon- In summary, the incorporation of bioactive molecules is dral grafts (i.e. mosaicplasty technique) are currently used a good approach for regulating cellular function. The opti- clinically; however, this technique suffers from several mization of physical properties (e.g. elasticity, porosity of limitations, such as lack of availability of materials, donor three-dimensional artificial ECMs) also is required to regu- site morbidity, and difficulty in matching the topology of late cellular function precisely for bone tissue engineering the grafts with the injured site. Therefore, osteochondral applications. composites have been developed to overcome these limitations. Although homogenous artificial ECMs were often fabricated for the reconstruction of osteochondral 5. Skin composites [165,166], the materials were designed to pro- duce cartilaginous and bone parts separately [167–170]. Several tissue-engineered skin products are commer- Because cartilage and bone are different tissues, the mate- cially available. The application of tissue-engineered skin is rial design should be optimized accordingly for each tissue. expanding from laboratory use (e.g. skin biology research, To fabricate osteochondral composites, two approaches drug evaluation) to clinical use [171]. Although tissue- have been utilized: (1) cartilage and bone tissues are engineered skins are commercially available, skin tissue independently reconstructed and combined into a single engineering research should progress to include additional composite graft by suturing or adhering them together, applications. Several types of artificial ECMs composed and (2) cartilage and bone tissues are reconstructed in of naturally-derived ECM proteins and synthetic poly- multi-phasic artificial ECMs. meric materials have been used to develop engineered skin For fabricating a single phase of each tissue layer [172–179]. followed by reconstruction in osteochondral composites, Skin tissue is divided into two parts: epidermis and several approaches that were optimized for each tissue dermis. In the epidermis, keratinocytes are the most com- have been utilized. These materials have been discussed mon cell type and form a surface barrier layer. In the in detail previously. For example, periosteal-derived cells dermis, fibroblasts exist in skin ECM and provide strength were cultured in a PLGA and PEG hybrid artificial ECM to and resilience. Most tissue-engineered skins are created by generate reconstructed bone. The cells in the PLGA/PEG expanding skin cells and using them to restore barrier func- artificial ECM deposited mineralized matrix and bone tion or to initiate wound healing. Many tissue engineered specific proteins (e.g. osteocalcin, osteopontin). The engi- skin dermal replacement products contain artificial ECMs neered bone was sutured together with reconstructed as a barrier or substratum for skin cells. The artificial ECM PGA-based engineered cartilage seeded with articular must eventually be discarded or replaced by live skin cells chondrocytes [167]. Similarly, collagen-hydroxyapatite for long-term healing, but artificial ECM can be used tem- hybrid sponges were used as the bone component of osteo- porarily to provide a barrier or a substratum for the cells. chondral composites, and the artificial ECMs were sutured Here, we summarized artificial ECMs used in the skin tissue to the engineered cartilage [168]. engineering field. 256 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

Table 7 Polymers used in skin tissue engineering.

Polymers Name in market Application Reference

Natural-derived polymers Collagen Integra Dermal replacement [172] Collagen Apligraf Epidermal/dermal replacement [173] Collagen Cincinnati skin substrate Epidermal/dermal replacement [174] Fibrin Bioseed-S Epithelial cover [175] Hyaluronic acid Hyalograft-3D Dermal replacement [175] Acellular matrix Alloderm Dermal replacement [176] Polyglactin910 Dermagraft Dermal replacement [177]

Synthetic polymers PCL Not commercially available Mechanical support to type I collagen scaffold [178] PLGA Not commercially available Dermal replacement [179]

5.1. Naturally-derived materials of artificial ECMs, such as material chemistry, architecture and mechanical properties can also modulate stem cells’ For dermal replacement, several naturally-derived ECM behavior, which is summarized in Table 8. proteins have been used as artificial ECMs for skin tis- sue engineering. Type I collagen has been used extensively 6.1. Delivery of soluble factors using artificial ECMs in this application since is a major component of der- mis ECM. Integra (produced by Integra Life Sciences) is an Soluble factors, including growth factors and hormones, example of type I collagen-based artificial ECM for skin play a significant role in the modulation of stem cell behav- tissue engineering [172]. Type I collagen-based artificial ior. Incorporation or local delivery of certain growth factors ECM also has been developed to replace epidermis and using artificial ECMs may increase the accessibility of said dermis [173,174]. Apligraf (produced by Organogenesis) factors to seeded stem cells as well as prevent growth fac- is composed of keratinocytes and fibroblasts with type I tor loss due to diffusion, which would be an effective way collagen-based artificial ECM. In addition to type I colla- of directing stem cell behavior. Artificial ECMs can retain gen, fibrin (BioSeed-S) and hyaluronic acid (Hyalograft-3D) and present growth factors via direct entrapment, affinity have been used as an artificial ECM for skin tissue engineer- and covalent tethering. ing [175]. In addition, acellular dermis from cadaveric skin Growth factors can be directly entrapped in artificial (Alloderm) has been used as an artificial ECM to mimic the ECMs. Moreover, the release of growth factors directly in vivo ECM of skin tissue [176]. entrapped in artificial ECMs or delivery systems is generally controlled via diffusion through or degradation of delivery 5.2. Synthetic polymeric materials systems, or even a combination of these two mechanisms [180,181]. One example of stem cell control by the growth Synthetic polymeric materials have also been used for factor delivery is that vascular endothelial growth factor skin tissue engineering applications. Bioresorbable PLGA (VEGF) delivery using a hydrogel can induce vascular differ- (polyglactin 910) is used in Dermagraft which is a commer- entiation of human embryonic stem cells (ESCs) entrapped cially available skin substitute [177]. Furthermore, other in the hydrogel [182]. Another example is that insulin- synthetic polymeric materials, such as PCL and PLGA, have like growth factor-1 and transforming growth factor-␤1 been used for a long time as artificial ECM in skin tissue (TGF-␤1) released from gelatin microparticles entrapped in engineering [178,179]. Table 7 summarizes the character- oligo[poly(ethylene glycol)fumarate] hydrogel can induce istics of polymers used in skin tissue engineering. chondrogenic differentiation of mesenchymal stem cells (MSCs) entrapped in the hydrogel [183]. 5.3. Future challenges in skin tissue engineering An affinity-based delivery system can release growth factors retained in a delivery system via non-covalent Although tissue-engineered skins are commercially interactions. Heparin-conjugated artificial ECMs have been available, there are several challenging issues including developed for the local release of growth factors with the improvement of safety and angiogenesis. In addition affinity to heparin. The delivery system is based on the to these issues, reconstruction of more complex structures inherent capacity of heparin to bind to growth factors such in skin is expected because in addition to epidermal and as fibroblast growth factor 2 (FGF2), VEGF, platelet-derived dermal tissues there are hairs, sweat glands, and touch growth factor (PDGF), TGF-␤ and bone morphogenetic receptors in native skin tissues. Therefore, it is necessary to protein (BMP) via electrostatic interactions between the develop engineered skin with these structures in the future. negatively-charged sulfate groups of heparin and the positively-charged amino acid residues of the growth fac- 6. Stem cells tor proteins [184,185]. Such a system can deliver FGF2, which allows long-term self-renewal [186] and pluripo- Artificial ECMs can modulate stem cells’ behavior, tency maintenance [187] of human ESCs. Heparin was also including adhesion, proliferation and differentiation. Sol- conjugated to poly(lactic-co-glycolic acid) (PLGA) scaffolds uble factors, cell adhesion signals and mechanical signals for BMP-2 delivery, creating a system capable of potentiat- can be incorporated into artificial ECMs for the control of ing the osteogenic efficacy of BMP-2 [188]. Local delivery stem cell behavior. The chemical and physical properties of BMP-2 using this system induced in situ osteogenic dif- B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 257

Table 8 Modulation of stem cells’ behavior with interactions with signals from artificial ECMs.

Parameter Artificial ECM Stem cell Stem cell’s behavior References

Soluble factor VEGF Dextran hydrogel Human ESCs Vascular cell differentiation [182] IGF-1/TGF-␤1 Gelatin/PEG-fumarate Rabbit MSCs Chondrogenic differentiation [183] hydrogel BMP-2 Heparin-conjugated PLGA Human MSCs Osteogenic differentiation [189] scaffold EGF PMMA-g-PEO Human MSCs Promotion of survival [190]

Cell adhesion factor RGD peptide PEG diacrylate hydrogel Human MSCs Osteogenic differentiation [192] RGD peptide PEG-fumarate hydrogel Rat MSCs Osteogenic differentiation [193] RGD peptide PEG diacrylate hydrogel Human ESCs Chondrogenic differentiation [195] RGD peptide P(NIPAAm-co-AAc) Human ESCs Maintenance of pluripotency [197] IKVAV peptide (laminin epitope) Self-assembling peptide Mouse neural Neuronal differentiation [215] nanofibers progenitor cells Fibronectin PEG diacrylate hydrogel Mouse MSCs Osteogenic differentiation [198] Fibronectin PET Human cord blood Promotion of proliferation [199] CD34+ cells Material chemistry Hyaluronic acid Hyaluronic acid/PEG diacrylate Goat MSCs Chondrogenic differentiation [200] hydrogel Hyaluronic acid Hyaluronic acid hydrogel Sheep MSCs Ligament protein expression [201] Hyaluronic acid Hyaluronic acid hydrogel Human ESCs Maintenance of pluripotency [202] Collagen type II Collagen type II Human MSCs Chondrogenic differentiation [203] Apatite Apatite-coated PLGA scaffold Human ESCs Osteogenic differentiation [205] Apatite Apatite-coated PLGA scaffold Human MSCs Osteogenic differentiation [208] Hydroxypatite Hydroxypatite/Self-assembling Mouse ESCs Osteogenic differentiation [209] peptide hydrogel Material architecture Nanofiber Polyamide nanofiber scaffold Mouse ESCs Promotion of proliferation [213] Nanofiber Titanium oxide nanotube Human MSCs Osteogenic differentiation [214] scaffold 3D hydrogel PEG hydrogel Mouse ESCs Chondrogenic differentiation [216]

Material mechanics Soft matrix Polyacrylamide Human MSCs Neuronal differentiation [229] Rigid matrix Polyacrylamide Human MSCs Osteogenic differentiation [229] Soft gel PEG-silica gel Human MSCs Neuronal differentiation [236] Rigid gel PEG-silica gel Human MSCs Osteogenic differentiation [236]

Cell shape and size Spread and large Fibronectin-coated polystyrene Human MSCs Osteogenic differentiation [240] Round and small Fibronectin-coated polystyrene Human MSCs Adipogenic differentiation [240] Round Tissue adhesive-coated Mouse embryonic Undifferentiation [241] polystyrene mesenchymal cells Elongated Tissue adhesive-coated Mouse embryonic Smooth muscle cell differentiation [241] polystyrene mesenchymal cells Flower shape Fibronectin-coated surface Human MSCs Adipogenic differentiation [242] Star shape Fibronectin-coated surface Human MSCs Osteogenic differentiation [242]

Mechanical signal Shear force Polyurethane Mouse ESCs Vascular cell differentiation [245] Cyclic strain Collagen type I-coated surface Mouse embryonic Smooth muscle cell differentiation [246] mesenchymal cells Multidimensional strain Collagen type I gel Bovine MSCs Ligament cell differentiation [248] Cyclic compression PEG diacrylate hydrogel Goat MSCs Chondrogenic differentiation [112] Cyclic strain Poly(lactide-co-caprolactone) Mouse ESCs Cardiomyogenic differentiation [249] scaffold ferentiation of human bone marrow MSCs transplanted in on an epidermal growth factor (EGF)-tethered poly(methyl undifferentiated state [189]. methacrylate)-graft-poly(ethylene oxide) (PMMA-g-PEO) Growth factors can be retained on artificial ECMs via surface or a PMMA-g-PEO surface with EGF supplemented covalent tethering. Recently, a study showed that a growth to the culture medium. EGF is known to promote MSC factor tethered to an artificial ECM exhibits higher activity proliferation and migration without affecting pluripotency than soluble growth factor [190]. The authors hypothesized [191]. Surprisingly, tethered EGF was superior to EGF in that the cellular response to tethered growth factor would solution for the promotion of MSCs spreading and survival be different due to cellular binding to the growth factor in the presence of Fas ligand, a potent cell death factor. The on a culture surface, which would inhibit internalization of authors concluded that EGF-tethered artificial ECM may the growth factor into the cytoplasm. MSCs were cultured offer a protective advantage over MSCs seeded on artificial 258 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

ECM in vivo if the artificial ECM causes an acute inflamma- MSCs cultured on the fibronectin-conjugated PEG diacry- tory response detrimental to the MSCs. late hydrogel also adhered to the hydrogel and underwent osteogenic differentiation, as indicated by matrix miner- 6.2. Cell adhesion signal presentation by artificial ECMs alization [198]. Compared to free fibronectin, fibronectin conjugated to artificial ECM promotes cell proliferation Conjugation of cell adhesion peptides to artificial ECMs [199]. On the other hand, conjugation of fibronectin to can promote stem cell differentiation. Previously, RGD polyethylene terephthalate scaffolds results in higher pro- peptides were covalently conjugated to PEG diacrylate liferation of human umbilical cord blood CD34+ cells hydrogels using conventional N-hydroxysuccinimidyl- compared to free, soluble fibronectin. This result could have ester (NHS) chemistry. It was found that functionalization been due to the prevention of fibronectin detachment from of PEG diacrylate hydrogels promotes the osteogenic dif- the scaffolds by covalent conjugation, compared to the high ferentiation of human MSCs [192]. Another study showed frequency of fibronectin desorption that occurs with free, that RGD peptide incorporation into oligo(PEG-fumarate) soluble fibronectin. hydrogels promotes osteogenic differentiation of MSCs [193]. The osteogenic differentiation of MSCs on peptide- 6.3. Chemistry of artificial ECMs modified oligo(PEG-fumarate) hydrogels is dependent on the concentration of incorporated peptide: the ALP activi- Hyaluronic acid is a polysaccharide found in many ties of MSCs in hydrogels modified with 2.0 and 1.0 ␮mol tissues and is known to influence stem cells’ behavior. peptide/g were significantly greater compared to MSCs in Hyaluronic acid has been shown to induce chondrogenic hydrogels modified with 0.1 ␮mol peptides/g or no pep- differentiation of MSCs. Entrapment of hyaluronic acid tide. The behavior of bone marrow stromal osteoblasts in photo-crosslinkable PEG diacrylate hydrogel containing cultured in hydrogels modified with two types of pep- cultured MSCs enhances cartilage-specific gene expres- tides, rat osteopontin-derived peptide and RGD, were sion [200]. In addition, MSCs cultured in hyaluronic acid compared previously [194]. The results showed that hydro- hydrogels express a variety of ligament proteins [201]. gel modified with osteopontin-derived peptide exhibits Hyaluronic acid also influences ESC behavior; human ESCs significantly higher cell migration than RGD-modified encapsulated in hyaluronic acid hydrogels maintain their hydrogels. RGD peptide-conjugated PEG diacrylate hydro- pluripotency and undifferentiated state, whereas human gels also promote the chondrogenic differentiation of ESCs encapsulated in hydrogels of dextran, a similarly human ESCs [195]. Compared to human ESCs in PEG structured polysaccharide, undergo spontaneous differ- diacrylate hydrogels, cells in RGD peptide-conjugated PEG entiation [202]. Furthermore, simple addition of soluble diacrylate hydrogels exhibit enhanced gene expression of growth factors to these hydrogels triggers lineage-specific cartilage-specific markers (i.e., aggrecan and collagen type differentiation of the entrapped ESCs. II). Collagen was also shown to favorably modulate stem The incorporation of cell adhesion peptide into arti- cells’ behavior. Human MSCs were mixed with collagen ficial ECMs may favorably influence stem cell behavior. type II isolated from bovine cartilage and then cultured in Artificial ECM comprising a semi-interpenetrating poly- pellet form [203]. The cells on the collagen type II scaffold mer network, or polymer hydrogel, was designed to mimic underwent chondrogenic differentiation into cartilage- the native ECM in terms of mechanical properties and specific ECMs. Chondrogenesis of MSCs is most likely cell adhesion properties. Poly(N-isopropylacrylamide-co- mediated by interaction with collagen type II through acrylic acid) [p(NIPAAm-co-AAc)] was loosely cross-linked integrin ␤1, since blocking of this integrin impeded chon- with an acrylated peptide, Gln-Pro-Gln-Gly-Leu-Ala-Lys- drogenesis. Implantation of type II collagen-cell constructs NH2 (QPQGLAK-NH2), the sequence of which was designed resulted in maintenance of chondrogenesis, inhibition of to be cleaved by matrix metalloproteinase-13 and other endochondral ossification and prevention of vessel inva- collagenases [196]. The mechanical properties of the sion in vivo. p(NIPAAm-co-AAc) hydrogel were controlled by vary- Hydroxyapatite can promote bone formation by stem ing the density of the network cross-linker. To provide cells. Hydroxyapatite is known to enhance osteoblastic dif- cell adhesion properties to the hydrogel, the hydrogel ferentiation as well as osteoblast growth [204]. A study was further interpenetrated by polyacrylic acid-graft-Ac- previously demonstrated that PLGA scaffolds incorporat- CGGNGEPRGDTYRAY-NH2 [p(AAc)-g-RGD] linear polymer ing apatite on their surfaces can stimulate the regeneration chains, which provide an active RGD motif [197]. Human of bone tissue in vivo by osteogenic cells derived from ESCs cultured in this hydrogel showed increased adherence human ESCs [205]. Efficient apatite incorporation into and viability, as well as the expression of markers of undif- PLGA scaffolds can be achieved through either gas foam- ferentiated human ESCs. In contrast, human ESCs cultured ing of PLGA-hydroxyapatite scaffolds [206] or biomimetic on gelatin-adsorbed polystyrene exhibited morphologies apatite-coating [207]. PLGA-hydroxyapatite composite of spontaneously differentiating cells. scaffold exposed to high amounts of hydroxyapatite can In addition to cell adhesion peptide conjugation, cell be fabricated by the gas-forming or particulate-leaching adhesion proteins such as fibronectin have been incor- method without the use of organic solvents. The sur- porated into artificial ECMs made of synthetic polymers. face of the PLGA-hydroxyapatite composite scaffold can The surfaces of photo-crosslinkable PEG diacrylate scaf- further be coated with bone-like apatite by incubation folds were modified by covalently conjugating fibronectin in simulated body fluid [207]. The biomimetic apatite- for the promotion of cell adhesion properties [198]. Mouse coating process was accelerated by the introduction of B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 259 hydroxyapatite nanoparticle nucleation sites to the PLGA- nanofibers promote neuronal differentiation of neural pro- hydroxyapatite composite scaffold. Bone-like apatite on genitor cells [215]. the scaffold surface significantly enhances the in vivo Stem cell differentiation is also significantly affected by osteogenic potential of the scaffold due to improved osteo- the dimensions of the culture system, which may be due conductivity [207]. Apatite-coated PLGA films enhance to differences in cellular interactions between cells cul- proliferation, alkaline phosphatase activity and osteocalcin tured in 2D and 3D systems. In 2D culture systems, cells production in human MSCs [208]. Further, the inclusion of adhere to one side of the culture surface. In contrast, cells hydroxyapatite particles in self-assembling peptide hydro- cultured in 3D hydrogels or scaffolds may attach to all gel promotes osteogenic differentiation of mouse ESCs sides of the hydrogels or scaffolds, to other cells, or even [209]. to secreted ECMs. ESC-derived EBs cultured in monolayer in the presence of chondrogenic soluble factors show lim- 6.4. Architecture of artificial ECMs ited expression of chondrogenic markers (aggrecan and Sox 9) [195,216]. However, when the EBs were cultured in 3D The macroscale, microscale and nanoscale structures PEG hydrogels, the soluble factors dramatically enhanced of artificial ECMs can determine the fate of stem cells. chondrogenic marker expression. Therefore, a 3D culture The macroscale and microscale ECM structures affect environment more closely mimics chondrogenesis in vivo cell migration, nutrient and waste diffusion and bulk and may be responsible for enhanced chondrogenic differ- mechanical properties. Generally, artificial ECMs are entiation. macroscopically designed to have physical properties Nanofabrication techniques have been used to study (mechanical properties, pore size, porosity, and degrada- stem cells’ behavior in response to nanopatterned artifi- tion rate) for the regeneration of specific tissues or organs. cial ECMs. The nanopatterned surface of artificial ECMs can The resultant artificial ECMs may lack the nanoscale fea- modulate diverse cell behaviors in terms of cell adhesion, tures required for precise control of cell behavior and tissue cell orientation, cell migration, cytoskeletal change and formation. Therefore, the nanoscale design of artificial gene expression [210]. Nanoscale topographical features ECMs is critical, since the nanoscale structure of the arti- that modulate cell behavior include scale (5 nm to microm- ficial ECMs can provide a substantial cellular interface for eter scale), order type (e.g., ridge, step, groove, pillar and effective interactions with cells [210] as well as the modu- pit) and symmetry (e.g., orthogonal or hexagonal packing of lation of cell behavior, including cell adhesion, biophysical nanopit) [217–223]. Increasing the nanoscale roughness of force transmission, cell shape, cytoskeletal structure and pore walls of artificial ECMs has been shown to enhance cell differentiation [211,212]. adhesion, proliferation and expression of ECM components Stem cells interact typically with nanoscaled ECMs [224]. Modulation of the nanoscale topography of cells may in vivo, and this interaction may influence stem cell behav- be mediated by changes in the adsorption and conforma- iors. Natural ECMs are composed of a complex network tion of cell adhesion proteins, the type of integrin that binds of ECM molecules with nanoscale and microscale dimen- to cell adhesion proteins and, consequently, integrin signal- sions, such as a nanofibrillar collagen fibril network. To ing [222]. Nanoscale topography may also modulate focal investigate whether or not the nanostructure of the artifi- adhesion of the cells as well as cytoskeletal organization, cial ECMs influences stem cells’ behavior, 3D nanofibrillar, both of which affect intracellular signaling [217]. artificial ECMs were produced by depositing electrospun nanofibers composed of polyamide onto the surfaces of 6.5. Mechanical properties of artificial ECMs glass coverslips [213]. Culture on the nanofibrillar matrix greatly enhanced the proliferation and self-renewal of The cellular microenvironments within tissues or mouse ESCs compared to culture without nanofibers [213]. organs can be physically diverse. The mechanical prop- The dimensions of the nanotubular matrix significantly erties of the artificial ECMs on which cells reside can influence stem cell behavior. Human MSCs were cul- also vary as much as 300-fold from soft brain tissue tured on nanotubular-shaped titanium oxide matrices, and to rigid, calcified bone [225]. Such variations in the osteogenic differentiation of human MSCs was induced by mechanical properties of ECMs within various tissues controlling the diameter of the nanotubes in the absence are obvious during development [226,227], and such of osteogenic soluble factors [214]. As a result, nanotubes properties may play important roles in tissue or organ with large diameters (70–100 nm) were found to promote development [228]. Muscle cells prefer a compliant matrix cell elongation and osteogenic differentiation of MSCs, for optimal contraction and relaxation, whereas bone whereas nanotubes with small diameters (30 nm) elicit cell cells prefer a rigid matrix on which they can miner- adhesion only without noticeable differentiation. alize. While the effects of growth factors on stem cell Artificial nanofiber ECMs can be functionalized using commitment have been well recognized, the importance bioactive factors to enhance stem cell differentiation. The of mechanical properties of matrices as a regulator of addition of hydroxyapatite particles into self-assembling stem cell differentiation is just now being appreciated peptide nanofiber matrices promotes osteogenic dif- [229]. ferentiation of cultured mouse ESCs [214]. Further, MSCs can sense the mechanical properties of an arti- incorporation of specific peptide sequences into self- ficial ECM and transduce this information as changes in assembling peptide nanofibers can promote neuronal cytoskeletal organization or as stem cell commitment. differentiation of progenitor cells. Finally, incorporation Cells can feel the resistance to deformation of the artifi- of laminin epitope IKVAV into self-assembling peptide cial ECM by adhering to and pulling against the artificial 260 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

ECM. Physical forces are generated by cells adhered to the substrates containing fibronectin-coated islands of various matrix due to the sliding of myosin bundles along actin sizes were fabricated using polydimethylsiloxane (PDMS) filaments. Actin structures are linked to focal adhesions, stamps. Human MSCs cultured on larger islands tended which provide a pathway for the transmission of force to flatten, spread and then differentiate into osteoblasts, from the cell to the matrix inside [230]. The tension gen- whereas cells cultured on smaller islands tended to be erated from the cell-matrix interaction induces changes round and unspread and differentiate into adipocytes. The in cytoskeletal organization [231] as well as cellular con- shape-dependent control of differentiation was mediated traction [232]. The biophysical cues are then converted by RhoA activity and cytoskeletal tension. Surprisingly, into biochemical signals via signaling molecules that act as controlling RhoA activity did not require the use of soluble mechano-transducers [233]. Finally, the biochemical sig- differentiation factors for adipogenic or osteogenic differ- nals commit the cell to a specific lineage [229]. entiation. The mechanical properties of a matrix or culture sur- Cell shape also modulates differentiation of embry- face can modulate stem cell differentiation. Engler et al. onic mesenchymal cells [241]. The role of embryonic found that matrix elasticity directs the differentiation of mesenchymal cell shape in smooth muscle differentia- human MSCs into cells the tissue type of which matches tion has been investigated by plating undifferentiated the stiffness of the culture environment [229]. Specifically, embryonic mesenchymal cells on microsurfaces. Micro- matrix elasticity was varied by controlling the cross-linking surfaces with diameters of 10 and 20 ␮m were prepared of polyacrylamide gels [234]. To promote cell-adhesive by coating multi-perforated polycarbonate membranes properties, the gels were coated with collagen type I, with holes of identical diameter using tissue adhesive. which is known to support both myogenic and osteogenic Embryonic MSCs cultured on 10 ␮m diameter microsur- differentiation [235]. MSC differentiation correlated with faces conserved their original round cell shape, whereas the in vivo matrix elasticity. Soft matrices that mimic cells cultured on 20 ␮m diameter microsurfaces were elon- brain induced neurogenic differentiation, stiffer matrices gated. The round-shaped cells remained undifferentiated that mimic muscle induced myogenic differentiation, and and expressed ␣-fetoprotein, which is produced by fetal comparatively rigid matrices that mimic bone induced cells but not by differentiated cells. In comparison, the elon- osteogenic differentiation (Fig. 7). Various mechanical sig- gated cells differentiated into smooth muscle, as indicated nals were generated by the cells pulling on the matrices by the expression of smooth muscle-specific proteins and with various levels of elasticity. Moreover, the inhibition membrane potential-dependent calcium ion currents. The of nonmuscle myosin II, a cytoskeletal motor, blocks the cell shape was the critical factor in directing differentiation modulation of cell differentiation by variable matrix elas- into smooth muscle, regardless of the supplementation of ticity. TGF-␤1. The relationship between MSC differentiation and When cells are of identical size, cell shape becomes sig- matrix elasticity in 2D culture systems [229] has also been nificant in modulating stem cell differentiation. A study by demonstrated in 3D culture systems [236]. 2D culture sys- Chen et al. demonstrated that cell shape and size direct tems for the modulation of stem cell differentiation may stem cell fate [240] and that cell shape also affects cell by incompatible with 3D tissue fabrication methods that size. However, their study did not isolate the effect of use stem cells. Pek et al. demonstrated that mechanical either cell size or shape from each other. Kilian et al. properties modulate the differentiation of MSCs cultured investigated the effect of cell shape only on stem cell dif- in 3D gels [236]. They previously developed a thixotropic ferentiation by varying cell shape and keeping cell size PEG-silica nanocomposite gel for 3D cell culturing [237]. constant [242]. A microcontact printing technique using This gel was used as a 3D matrix to investigate the effects PDMS stamps was used to produce a patterned culture of of mechanical signaling on MSC differentiation in a 3D human MSCs with various cell shapes and with identical matrix. The stiffness of the gel was varied by controlling cell sizes in the presence of both adipogenic and osteogenic the amount of fumed silica [238]. When human MSCs were soluble factors. Cells that were flower-shaped differenti- cultured in the gels with varying degrees of mechanical ated into adipocytes, whereas star-shaped cells resulted stiffness, the highest expression levels of neural (ENO2), in osteogenic differentiation, possibly due to a difference myogenic (MYOG) and osteogenic (Runx2, OC) transcrip- in cytoskeletal contractility. A previous study showed that tion factors were obtained using gels with low, medium contractile cytoskeletal movement changes the cell from and high degrees of stiffness, respectively [236]. In addi- a flower shape to a star shape [243]. The concave regions tion, the immobilization of RDG ligand to the gel further between the points of the star are highly contractile regions enhanced osteogenic factor expression, due possibly to the where the cell spans across a non-adhesive area [242]. Star- RGD ligand-integrin complex acting as a mechanical signal shaped cells have larger focal adhesion and cytoskeletal receptor that allows the transmission of signals through a tension than cells that are flower-shaped. It has been sug- signaling pathway [239]. gested that local curvatures that increase the cytoskeletal tension of the cell also promote osteogenic differentiation 6.6. Control of cell shape and size with artificial ECMs rather than adipogenic differentiation [242]. It was also suggested that cells with enhanced contractility undergo Soft lithography techniques allow for spatial patterning osteogenic differentiation due to the increased activation of cell adhesion molecules, which can be utilized to control of c-Jun N-terminal kinase (JNK) and extracellular related cell size and shape. Chen and colleagues showed that cell kinase (ERK1/2) in conjunction with elevated wingless type shape and size direct stem cell fate [240]. Micropatterned (Wnt) signaling. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 261

Fig. 7. Elasticity of artificial ECM directs MSC differentiation. Soft matrix leads to neuronal differentiation (␤3 small arrows: Tubulin-positive), stiffer matrix leads to myogenic differentiation (mid-size arrow: MyoD-positive), and comparatively rigid matrix leads to osteogenic differentiation (large arrow: CBF␣1-positive). The scale bars denote 5 ␮m. [229] Copyright 2006, Elsevier Inc., Oxford, UK.

6.7. Mechanical signal loading on artificial ECMs various forms of mechanical forces. Cells of the heart, arter- ies, bladder and intestines are subjected to cyclic strain. Stem cell differentiation into certain types of cells Further, chondrocytes are subjected to cyclic compression is influenced by the mechanical microenvironment. For in vivo. It is thought that the same forces that govern tis- example, cells in tissues or organs in vivo are subjected to sue development in vivo also enhance tissue development 262 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 in vitro [244]. Thus, the mechanical signals could play sig- been shown to stimulate the cardiomyogenic differentia- nificant roles in the lineage commitment of stem cells. For tion of stem cells. Bhang et al. showed that combining a example, shear force [245] or cyclic strain [246] induces mechanical signal (cyclic strain) and soluble signal (TGF- vascular cell differentiation of ESCs. It is speculated that ␤1) promotes cardiomyogenic marker expression in MSCs mechanical signals are transmitted to the cells via cell more effectively than soluble signal alone [251]. The surface receptors at focal adhesion sites, which serve as expression of cardiac-specific markers (cardiac ␣-actin, mechanical transducers between the cell-adhesion surface tropomyosin, cardiac troponin-I and cardiac myosin heavy and cytoskeleton of the cell. The transmitted mechanical chain) was more upregulated in MSCs cultured with TGF- signals are converted into biochemical responses through a ␤1 and subjected to cyclic strain than in MSCs cultured mechanism termed as mechanotransduction [247], result- statically with TGF-␤1. Incomplete cardiomyogenic dif- ing in activation of specific genes that modulate stem cell ferentiation of MSCs could result in the persistence of differentiation. undifferentiated stem cells that undergo unanticipated Mechanical stimulation alone, without exogenous differentiation upon implantation. On the other hand, a growth factors, can induce the differentiation of bone mar- combination of two or more types of signals could enhance row mesenchymal progenitor cells into a ligament cell the efficiency of stem cell differentiation. lineage [248]. Collagen gel was used as an artificial ECM for the transduction of mechanical signals to cells in vitro. The 7. Concluding remarks application of ligament-like multidimensional mechanical strains (translational and rotational strain) to undifferen- Artificial ECMs can regulate the cellular function to engi- tiated cells embedded in a collagen gel over a period of 21 neer functional tissues. For engineering functional tissues, days resulted in increased expression of ligament fibroblast signals and parameters that can modulate cell behav- markers, including collagen types I and III and tenascin-C, ior can be engineered into artificial ECMs. The signals as well as induced cell alignment and collagen fiber orien- and parameters include soluble signals, cell-adhesion sig- tation, all features characteristic of ligament cells. nals, mechanical signals, and chemistry, architecture, and Mechanical compression could promote the chondro- mechanical properties of artificial ECMs. Design criteria for genic differentiation of stem cells. Hydrogels can be used artificial ECMs vary considerably depending on the tissue as an artificial ECM for the transduction of compressive of interest. Artificial ECMs that can present appropriate mechanical signals to cells. Exposure of bone marrow- signals and parameters can coordinate the interactions derived MSCs cultured in PEG diacrylate hydrogels to cyclic between the artificial ECMs and cells to regenerate a compressive strain in vitro increases the gene expres- desired tissue. The artificial ECMs developed for tissue sion of cartilage-related markers such as Sox-9, type II engineering may in turn provide a novel experimental sys- ␤ collagen and aggrecan regardless of TGF- 1 supplemen- tem to elucidate the mechanisms by which native ECMs tation [112]. On the other hand, mechanical compres- regulate tissue development and stem cell behavior. sion increases expression of cartilage-specific genes in human ESC-derived cells encapsulated in tyrosinearginine- Acknowledgements glycine-aspartate-serine (YRGDS)-modified PEG-acrylate (YRGDS-PEG-acrylate) hydrogels only in the presence of This work was supported by a grant (ROA-2010- TGF-␤1 [112]. 0018428) from NRL program of Korea Science and Elastic, artificial ECMs permit the transduction of Engineering Foundation (KOSEF), the Ministry of Educa- mechanical cyclic strain signals for cardiomyogenesis. Car- tion, Science and Technology, Republic of Korea and a grant diomyocytes in the body are subjected to cyclic strain (2010-0020352) from the National Research Foundation of induced by the heart beating. A study has demonstrated Korea. that cyclic strain promotes cardiomyogenesis of embryonic stem cell-derived cardiomyocytes (ESCCs) [249]. The appli- cation of cyclic strain to ESCCs cultured on elastic polymer References [poly(lactide-co-caprolactone), PLCL] scaffolds in vitro ele- [1] Patrick CW, Mikos AG, McIntire LV, editors. Frontiers in tissue engi- vated cardiac gene expression compared to unstrained neering. Oxford: Pergamon; 1998. controls. Elastic PLCL scaffolds were also used as a vehi- [2] Rosso F, Giordano A, Barbarisi M, Barbarisi A. From cell-ECM inter- cle for cell transplantation into infarcted rat myocardium. actions to tissue engineering. J Cell Physiol 2004;199:174–80. [3] Langer R, Vacanti JP. Tissue engineering. Science 1993;260:920–6. PLCL scaffolds were seeded with ESCCs and implanted on [4] Abatangelo G, Brun P, Radice M, Cortivo R, Auth MKH. Tissue engi- the surface of infracted rat heart. As the heart beats, cyclic neering. In: Barbucci R, editor. Integrated biomaterial science. New strain signals were transmitted to ESCCs seeded on the York: Kluwer Academic; 2001. p. 885–945. elastic scaffolds. Importantly, cardiac gene expression was [5] Putnam AJ, Mooney DJ. Tissue engineering using synthetic extra- cellular matrices. Nat Med 1996;2:824–6. upregulated in the elastic patches compared to unstrained [6] Daley WP, Peters SB, Larsen M. Extracellular matrix dynam- control patches (non-elastic PLGA scaffolds seeded with ics in development and regenerative medicine. J Cell Sci ESCCs), suggesting for the presence of cardiomyocyte- 2008;121:255–64. [7] Rozario T, DeSimone DW. The extracellular matrix in development specific microstructures including myofibrillar bundles and morphogenesis: a dynamic view. Dev Biol 2010;341:126–40. and Z-lines. [8] Hollister SJ. Porous scaffold design for tissue engineering. Nat Mater Mechanical signals in combination with soluble sig- 2005;4:518–24. [9] Moroni L, de Wijn JR, van Blitterswijk CA. Integrating novel tech- nals synergistically stimulate the differentiation of stem nologies to fabricate smart scaffolds. J Biomater Sci Polym Ed cells. Either cyclic strain [249] or TGF-␤1 [250] alone has 2008;19:543–72. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 263

[10] Badylak SF, Freytes DO, Gilbert TW. Extracellular matrix as a bio- gluconamide) by quartz-crystal microbalance. Colloids Surf B logical scaffold material: Structure and function. Acta Biomater Biointerfaces 2005;42:137–40. 2009;5:1–13. [33] Lopina ST, Wu G, Merrill EW, Griffith-Cima L. Hepatocyte culture [11] Mano JF, Silva GA, Azevedo HS, Malafaya PB, Sousa RA, Silva SS, Boe- on carbohydrate-modified star polyethylene oxide hydrogels. Bio- sel LF, Oliveira JM, Santos TC, Marques AP, ves NM, Reis RL. Natural materials 1996;17:559–69. origin biodegradable systems in tissue engineering and regener- [34] Griffith LG, Lopina S. Microdistribution of substratum-bound lig- ative medicine: present status and some moving trends. J R Soc ands affects cell function: hepatocyte spreading on PEO-tethered Interface 2007;4:999–1030. galactose. Biomaterials 1998;19:979–86. [12] Rosso F, Marino G, Giordano A, Barbarisi M, Parmeggiani D, Barbarisi [35] Donati I, Gamini A, Vetere A, Campa C, Paoletti S. Synthesis, A. Smart materials as scaffolds for tissue engineering. J Cell Physiol characterization, and preliminary biological study of glycoconju- 2005;203:465–70. gates of poly(styrene-co-maleic acid). Biomacromolecules 2002;3: [13] Zakim D, Boyer TD, editors. Hepatology: a textbook of liver disease. 805–12. Philadelphia: W.B. Saunders Co; 1990. [36] Yoon JJ, Nam YS, Kim JH, Park TG. Surface immobilization of galac- [14] Pricer Jr WE, Ashwell G. The binding of desialylated glycoproteins tose onto aliphatic biodegradable polymers for hepatocyte culture. by plasma membranes of rat liver. J Biol Chem 1971;246:4825–33. Biotechnol Bioeng 2002;78:1–10. [15] Newfeld E, Ashwell G. Carbohydrate recognizable systems for [37] Yin C, Ying L, Zhang PC, Zhuo RX, Kang ET, Leong KW, Mao HQ. receptor-mediated pinocytosis. In: Lennartz W, editor. Biochem- High density of immobilized galactose ligand enhances hepato- istry of and proteoglycans. New York: Plenum Press; cyte attachment and function. J Biomed Mater Res A 2003;67: 1979. p. 241–66. 1093–104. [16] Wall DA, Wilson G, Hubbard AL. The galactose-specific recognition [38] Lu HF, Lim WS, Wang J, Tang ZQ, Zhang PC, Leong KW, Chia system of mammalian liver: the route of ligand internalization in SM, Yu H, Mao HQ. Galactosylated PVDF membrane promotes rat hepatocytes. Cell 1980;21:79–93. hepatocyte attachment and functional maintenance. Biomaterials [17] Ashwell G, Harford J. Carbohydrate-specific receptors of the liver. 2003;24:4893–903. Annu Rev Biochem 1982;51:531–54. [39] Kang IK, Lee DW, Lee SK, Akaike T. Grafting of lactose-carrying [18] Uyama S, Kaufmann PM, Takeda T, Vacanti JP. Delivery of whole styrene onto polystrene dishes using plasma glow discharge liver-equivalent hepatocyte mass using polymer devices and hep- and their interaction with hepatocytes. J Mater Sci Mater Med atotrophic stimulation. Transplantation 1993;55:932–5. 2003;14:611–6. [19] Kneser U, Kaufmann PM, Fiegel HC, Pollok JM, Kluth D, Herbst [40] Higashiyama S, Noda M, Kawase M, Yagi K. Mixed-ligand modifica- H, Rogiers X. Long-term differentiated function of heterotopically tion of polyamidoamine dendrimers to develop an effective scaffold transplanted hepatocytes on three-dimensional polymer matrices. for maintenance of hepatocyte spheroids. J Biomed Mater Res A J Biomed Mater Res 1999;47:494–503. 2003;64:475–82. [20] Uyama S, Kaufmann PM, Kneser U, Fiegel HC, Pollok JM, [41] Nair LS, Laurencin CT. Biodegradable polymers as biomaterials. Prog Kluth D, Vacanti JP, Rogiers X. Hepatocyte transplantation using Polym Sci 2007;32:762–98. biodegradable matrices in ascorbic acid-deficient rats: compari- [42] Sell SA, McClure MJ, Garg K, Wolfe PS, Bowlin GL. Electrospinning son with heterotopically transplanted liver grafts. Transplantation of collagen/biopolymers for regenerative medicine and cardiovas- 2001;71:1226–31. cular tissue engineering. Adv Drug Deliv Rev 2009;61:1007–19. [21] Hasirci V, Berthiaume F, Bondre SP, Gresser JD, Trantolo DJ, Toner [43] Zhao Y, Xu Y, Zhang B, Wu X, Xu F, Liang W, Du X, Li R. M, Wise DL. Expression of liver-specific functions by rat hepato- In vivo generation of thick, vascularized hepatic tissue from col- cytes seeded in treated poly(lactic-co-glycolic) acid biodegradable lagen hydrogel-based hepatic units. Tissue Eng Part C Methods foams. Tissue Eng 2001;7:385–94. 2010;16:653–9. [22] Ranucci CS, Kumar A, Batra SP, Moghe PV. Control of hepatocyte [44] Sosnik A, Sefton MV. Semi-synthetic collagen/poloxamine matrices function on collagen foams: sizing matrix pores toward selective for tissue engineering. Biomaterials 2005;26:7425–35. induction of 2-D and 3-D cellular morphogenesis. Biomaterials [45] Wang X, Yan Y, Xiong Z, Lin F, Wu R, Zhang R, Lu Q. Preparation 2000;21:783–93. and evaluation of ammonia-treated collagen/chitosan matrices for [23] Elkayam T, Amitay-Shaprut S, Dvir-Ginzberg M, Harel T, Cohen S. liver tissue engineering. J Biomed Mater Res B Appl Biomater Enhancing the drug metabolism activities of C3A-a human hepato- 2005;75:91–8. cyte cell line-by tissue engineering within alginate scaffolds. Tissue [46] Gutman A, Kornblihtt AR. Identification of a third region of cell- Eng 2006;12:1357–68. specific alternative splicing in human fibronectin mRNA. Proc Natl [24] Nakazawa K, Ijima H, Fukuda J, Sakiyama R, Yamashita Y, Shi- Acad Sci USA 1987;84:7179–82. mada M, Shirabe K, Tsujita E, Sugimachi K, Funatsu K. Development [47] Ruoslahti E, Pierschbacher MD. New perspectives in cell adhesion: of a hybrid artificial liver using polyurethane foam/hepatocyte RGD and integrins. Science 1987;238:491–7. spheroid culture in a preclinical pig experiment. Int J Artif Organs [48] Bhadriraju K, Hansen LK. Hepatocyte adhesion, growth and dif- 2002;25:51–60. ferentiated function on RGD-containing proteins. Biomaterials [25] Weigel PH, Schmell E, Lee YC, Roseman S. Specific adhesion of rat 2000;21:267–72. hepatocytes to beta-galactosides linked to polyacrylamide gels. J [49] Fiegel HC, Havers J, Kneser U, Smith MK, Moeller T, Kluth Biol Chem 1978;253:330–3. D, Mooney DJ, Rogiers X, Kaufmann PM. Influence of flow [26] Schnaar RL, Weigel PH, Kuhlenschmidt MS, Lee YC, Roseman conditions and matrix coatings on growth and differentiation S. Adhesion of chicken hepatocytes to polyacrylamide gels of three-dimensionally cultured rat hepatocytes. Tissue Eng derivatized with N-acetylglucosamine. J Biol Chem 1978;253: 2004;10:165–74. 7940–51. [50] Mooney D, Hansen L, Vacanti J, Langer R, Farmer S, Ingber D. [27] Weigel PH. Rat hepatocytes bind to synthetic galactoside sur- Switching from differentiation to growth in hepatocytes: control faces via a patch of asialoglycoprotein receptors. J Cell Biol by extracellular matrix. J Cell Physiol 1992;151:497–505. 1980;87:855–61. [51] Mooney DJ, Hansen LK, Langer R, Vacanti JP, Ingber DE. Extracel- [28] Weigel PH, Oka JA. Temperature dependence of endocytosis medi- lular matrix controls tubulin monomer levels in hepatocytes by ated by the asialoglycoprotein receptor in isolated rat hepatocytes. regulating protein turnover. Mol Biol Cell 1994;5:1281–8. Evidence for two potentially rate-limiting steps. J Biol Chem [52] Hansen LK, Mooney DJ, Vacanti JP, Ingber DE. Integrin binding 1981;256:2615–7. and cell spreading on extracellular matrix act at different points [29] Kobayashi K, Ina Y. Synthesis and function of polystyrene deriva- in the cell cycle to promote hepatocyte growth. Mol Biol Cell tives having pendant oligosaccharides. Polym J 1985;17:567–75. 1994;5:967–75. [30] Kobayashi A, Goto M, Kobayashi K, Akaike T. Receptor-mediated [53] Mooney DJ, Langer R, Ingber DE. Cytoskeletal filament assembly and regulation of differentiation and proliferation of hepatocytes by the control of cell spreading and function by extracellular matrix. J synthetic polymer model of asialoglycoprotein. J Biomater Sci Cell Sci 1995;108:2311–20. Polym Ed 1994;6:325–42. [54] Powers MJ, Griffith-Cima L. Motility behavior of hepatocytes on [31] Tobe S, Takei Y, Kobayashi K, Akaike T. Receptor-mediated for- extracellular matrix substrata during aggregation. Biotechnol Bio- mation of multilayer aggregates of primary cultured adult rat eng 1996;50:392–403. hepatocytes on lactose-substituted polystyrene. Biochem Biophys [55] Hoshiba T, Wakejima M, Cho CS, Shiota G, Akaike T. Different Res Commun 1992;184:225–30. regulation of hepatocyte behaviors between natural extracellu- [32] Takei R, Seo SJ, Cho CS, Okahata Y, Akaike T. Adsorption behaviors lar matrices and synthetic extracellular matrices by hepatocyte of poly(N-p-vinylbenzyl-4-O-beta-D-galactopyranosyl-[1 → 4]-D- growth factor. J Biomed Mater Res A 2008;85:228–35. 264 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

[56] Percival E, McDowell RH. Algal Polysaccharides. In: Dey PM, editor. [81] Dzenis Y. Material science. Spinning continuous fibers for nan- Methods in plant biochemistry carbohydrates. London: Academic otechnology. Science 2004;304:1917–9. Press; 1990. p. 523–47. [82] Mo XM, Xu CY, Kotaki M, Ramakrishna S. Electrospun P(LLA-CL) [57] Izydorczyk M, Cui SW, Wang Q. Polysaccharide gums: structures, nanofiber: a biomimetic extracellular matrix for smooth muscle cell functional properties, and applications. In: Cui SW, editor. Food car- and endothelial cell proliferation. Biomaterials 2004;25:1883–90. bohydrates: chemistry, physical properties, and applications. Boca [83] Chua KN, Lim WS, Zhang P, Lu H, Wen J, Ramakrishna S, Leong Raton, FL: CRC Press; 2005. p. 263–307. KW, Mao HQ. Stable immobilization of rat hepatocyte spheroids on [58] Yang J, Goto M, Ise H, Cho CS, Akaike T. Galactosylated alginate as a galactosylated nanofiber scaffold. Biomaterials 2005;26:2537–47. scaffold for hepatocytes entrapment. Biomaterials 2002;23:471–9. [84] Feng ZQ, Chu X, Huang NP, Wang T, Wang Y, Shi X, Ding Y, Gu ZZ. The [59] Park IK, Yang J, Jeong HJ, Bom HS, Harada I, Akaike T, Kim SI, Cho effect of nanofibrous galactosylated chitosan scaffolds on the for- CS. Galactosylated chitosan as a synthetic extracellular matrix for mation of rat primary hepatocyte aggregates and the maintenance hepatocytes attachment. Biomaterials 2003;24:2331–7. of liver function. Biomaterials 2009;30:2753–63. [60] Yang J, Chung TW, Nagaoka M, Goto M, Cho CS, Akaike [85] Tuzlakoglu K, Reis RL. Biodegradable polymeric fiber structures in T. Hepatocyte-specific porous polymer-scaffolds of algi- tissue engineering. Tissue Eng Part B Rev 2009;15:17–27. nate/galactosylated chitosan sponge for liver-tissue engineering. [86] Getgood A, Brooks R, Fortier L, Rushton N. Articular cartilage tissue Biotechnol Lett 2001;23:1385–9. engineering: today’s research, tomorrow’s practice? J Bone Joint [61] Chung TW, Yang J, Akaike T, Cho KY, Nah JW, Kim SI, Cho CS. Prepa- Surg Br 2009;91:565–76. ration of alginate/galactosylated chitosan scaffold for hepatocyte [87] Wang L, Detamore MS. Tissue engineering the mandibular condyle. attachment. Biomaterials 2002;23:2827–34. Tissue Eng 2007;13:1955–71. [62] Seo SJ, Choi YJ, Akaike T, Higuchi A, Cho CS. Alginate/galactosylated [88] Awad HA, Wickham MQ, Leddy HA, Gimble JM, Guilak F. Chondro- chitosan/heparin scaffold as a new synthetic extracellular matrix genic differentiation of adipose-derived adult stem cells in agarose, for hepatocytes. Tissue Eng 2006;12:33–44. alginate, and gelatin scaffolds. Biomaterials 2004;25:3211–22. [63] Tabata Y. Biomaterial technology for tissue engineering application. [89] Gruber HE, Hoelscher GL, Leslie K, Ingram JA, Hanley Jr EN. Three- J R Soc Interface 2009;6:S311–24. dimensional culture of human disc cells within agarose or a [64] Hong SR, Lee YM, Akaike T. Evaluation of a galactose-carrying collagen sponge: assessment of proteoglycan production. Bioma- gelatin sponge for hepatocytes culture and transplantation. J terials 2006;27:371–6. Biomed Mater Res A 2003;67:733–41. [90] Wu CH, Ko CS, Huang JW, Huang HJ, Chu IM. Effects of exogenous [65] Altman GH, Diaz F, Jakuba C, Calabro T, Horan RL, Chen J, Lu glycosaminoglycans on human chondrocytes cultivated on type II H, Richmond J, Kaplan DL. Silk-based biomaterials. Biomaterials collagen scaffolds. J Mater Sci Mater Med 2010;21:725–9. 2003;24:401–16. [91] Dorotka R, Bindreiter U, Macfelda K, Windberger U, Nehrer [66] Gotoh Y, Niimi S, Hayakawa T, Miyashita T. Preparation of lactose- S. Marrow stimulation and chondrocyte transplantation using silk fibroin conjugates and their application as a scaffold for a collagen matrix for cartilage repair. Cartilage hepatocyte attachment. Biomaterials 2004;25:1131–40. 2005;13:655–64. [67] Seo SJ, Park IK, Yoo MK, Shirakawa M, Akaike T, Cho CS. Xyloglu- [92] Cherubino P, Grassi FA, Bulgheroni P, Ronga M. Autologous can as a synthetic extracellular matrix for hepatocyte attachment. chondrocyte implantation using a bilayer collagen membrane: a J Biomater Sci Polym Ed 2004;15:1375–87. preliminary report. J Orthop Surg (Hong Kong) 2003;11:10–5. [68] Seo SJ, Akaike T, Choi YJ, Shirakawa M, Kang IK, Cho CS. Algi- [93] Kawabe N, Yoshinao M. The repair of full-thickness articular car- nate microcapsules prepared with xyloglucan as a synthetic tilage defects Immune responses to reparative tissue formed by extracellular matrix for hepatocyte attachment. Biomaterials allogeneic growth plate chondrocyte implants. Clin Orthop Relat 2005;26:3607–15. Res 1991;268:279–93. [69] Ise H, Sugihara N, Negishi N, Nikaido T, Akaike T. Low [94] Vepari C, Kaplan DL. Silk as a biomaterial. Prog Polym Sci asialoglycoprotein receptor expression as markers for highly pro- 2007;32:991–1007. liferative potential hepatocytes. Biochem Biophys Res Commun [95] Wang Y, Blasioli DJ, Kim HJ, Kim HS, Kaplan DL. Cartilage tissue 2001;285:172–82. engineering with silk scaffolds and human articular chondrocytes. [70] Kim SH, Kim JH, Akaike T. Regulation of cell adhesion signaling Biomaterials 2006;27:4434–42. by synthetic glycopolymer matrix in primary cultured hepatocyte. [96] Aoki H, Tomita N, Morita Y, Hattori K, Harada Y, Sonobe M, Wakitani FEBS Lett 2003;553:433–9. S, Tamada Y. Culture of chondrocytes in fibroin-hydrogel sponge. [71] Cho CS, Goto M, Kobayashi A, Kobayashi K, Akaike T. Effect of Biomed Mater Eng 2003;13:309–16. ligand orientation on hepatocyte attachment onto the poly(N- [97] Hofmann S, Knecht S, Langer R, Kaplan DL, Vunjak-Novakovic p-vinylbenzyl-o-beta-D-galactopyranosyl-D-gluconamide) as G, Merkle HP, Meinel L. Cartilage-like tissue engineering a model ligand of asialoglycoprotein. J Biomater Sci Polym Ed using silk scaffolds and mesenchymal stem cells. Tissue Eng 1996;7:1097–104. 2006;12:2729–38. [72] Berthiaume F, Moghe PV, Toner M, Yarmush ML. Effect of extracel- [98] Panilaitis B, Altman GH, Chen J, Jin HJ, Karageorgiou V, Kaplan DL. lular matrix topology on cell structure, function, and physiological Macrophage responses to silk. Biomaterials 2003;24:3079–85. responsiveness: hepatocytes cultured in a sandwich configuration. [99] Kayakabe M, Tsutsumi S, Watanabe H, Kato Y, Takagishi K. Trans- FASEB J 1996;10:1471–84. plantation of autologous rabbit BM-derived mesenchymal stromal [73] Chen CS, Mrksich M, Huang S, Whitesides GM, Ingber DE. Geometric cells embedded in hyaluronic acid gel sponge into osteochondral control of cell life and death. Science 1997;276:1425–8. defects of the knee. Cytotherapy 2006;8:343–53. [74] Patel N, Padera R, Sanders GH, Cannizzaro SM, Davies MC, Langer R, [100] Tortelli F, Cancedda R. Three-dimensional cultures of osteogenic Roberts CJ, Tendler SJ, Williams PM, Shakesheff KM. Spatially con- and chondrogenic cells: a tissue engineering approach to mimic trolled cell engineering on biodegradable polymer surfaces. FASEB bone and cartilage in vitro. Eur Cell Mater 2009;17:1–14. J 1998;12:1447–54. [101] Nehrer S, Domayer S, Dorotka R, Schatz K, Bindreiter U, Kotz [75] Ranucci CS, Moghe PV. Polymer substrate topography actively reg- R. Three-year clinical outcome after chondrocyte transplanta- ulates the multicellular organization and liver-specific functions of tion using a hyaluronan matrix for cartilage repair. Eur J Radiol cultured hepatocytes. Tissue Eng 1999;5:407–20. 2006;57:3–8. [76] Park J, Berthiaume F, Toner M, Yarmush ML, Tilles AW. Microfabri- [102] Knudson W, Casey B, Nishida Y, Eger W, Kuettner KE, Knudson cated grooved substrates as platforms for bioartificial liver reactors. CB. Hyaluronan oligosaccharides perturb cartilage matrix home- Biotechnol Bioeng 2005;90:632–44. ostasis and induce chondrocytic chondrolysis. Arthritis Rheum [77] Park J, Li Y, Berthiaume F, Toner M, Yarmush ML, Tilles AW. Radial 2000;43:1165–74. flow hepatocyte bioreactor using stacked microfabricated grooved [103] Diduch DR, Jordan LC, Mierisch CM, Balian G. Marrow stromal substrates. Biotechnol Bioeng 2008;99:455–67. cells embedded in alginate for repair of osteochondral defects. [78] Glicklis R, Shapiro L, Agbaria R, Merchuk JC, Cohen S. Hepato- Arthroscopy 2000;16:571–7. cyte behavior within three-dimensional porous alginate scaffolds. [104] Selmi TA, Verdonk P, Chambat P, Dubrana F, Potel JF, Barnouin L, Biotechnol Bioeng 2000;67:344–53. Neyret P. Autologous chondrocyte implantation in a novel alginate- [79] Sill TJ, von Recum HA. Electrospinning: applications in drug delivery agarose hydrogel: outcome at two years. J Bone Joint Surg Br and tissue engineering. Biomaterials 2008;29:1989–2006. 2008;90:597–604. [80] Baker BM, Handorf AM, Ionescu LC, Li WJ, Mauck RL. New directions [105] Suh JK, Matthew HW. Application of chitosan-based polysaccharide in nanofibrous scaffolds for soft tissue engineering and regenera- biomaterials in cartilage tissue engineering: a review. Biomaterials tion. Expert Rev Med Devices 2009;6:515–32. 2000;21:2589–98. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 265

[106] Hoemann CD, Hurtig M, Rossomacha E, Sun J, Chevrier A, Shive [128] Roy TD, Simon JL, Ricci JL, Rekow ED, Thompson VP, Parsons JR. MS, Buschmann MD. Chitosan-glycerol phosphate/blood implants Performance of degradable composite bone repair products made improve repair in ovine microfracture defects. J via three-dimensional fabrication techniques. J Biomed Mater Res Bone Joint Surg Am 2005;87:2671–86. A 2003;66:283–91. [107] Hoemann CD, Sun J, McKee MD, Chevrier A, Rossomacha E, Rivard [129] Gogolewski S. Bioresorbable polymers in trauma and bone surgery. GE, Hurtig M, Buschmann MD. Chitosan-glycerol phosphate/blood Injury 2000;31:28–32. implants elicit hyaline cartilage repair integrated with porous [130] Karageorgiou V, Kaplan D. Porosity of 3D biomaterial scaffolds and subchondral bone in microdrilled rabbit defects. Osteoarthritis Car- osteogenesis. Biomaterials 2005;26:5474–91. tilage 2007;15:78–89. [131] LeGeros RZ. Properties of osteoconductive biomaterials: calcium [108] Kim SE, Park JH, Cho YW, Chung H, Jeong SY, Lee EB, Kwon phosphates. Clin Orthop Relat Res 2002;395:81–98. IC. Porous chitosan scaffold containing microspheres loaded with [132] Porter JR, Ruckh TT, Popat KC. Bone tissue engineering: a review transforming growth factor-beta1: implications for cartilage tissue in bone biomimetics and drug delivery strategies. Biotechnol Prog engineering. J Control Release 2003;91:365–74. 2009;25:1539–60. [109] Shah RN, Shah NA, Del Rosario Lim MM, Hsieh C, Nuber G, Stupp SI. [133] Hulbert SF, Young FA, Mathews RS, Klawitter JJ, Talbert CD, Supramolecular design of self-assembling nanofibers for cartilage Stelling FH. Potential of ceramic materials as permanently regeneration. Proc Natl Acad Sci USA 2010;107:3293–8. implantable skeletal prostheses. J Biomed Mater Res 1970;4: [110] Munirah S, Kim SH, Ruszymah BH, Khang G. The use of fib- 433–56. rin and poly(lactic-co-glycolic acid) hybrid scaffold for articular [134] Cadosch D, Chan E, Gautschi OP, Filgueira L. Metal is not inert: role cartilage tissue engineering: an in vivo analysis. Eur Cell Mater of metal ions released by biocorrosion in aseptic loosening–current 2008;15:41–52. concepts. J Biomed Mater Res A 2009;91:1252–62. [111] Nicodemus GD, Villanueva I, Bryant SJ. Mechanical stimulation [135] Robertson DM, Pierre L, Chahal R. Preliminary observations of TMJ condylar chondrocytes encapsulated in PEG hydrogels. J of bone ingrowth into porous materials. J Biomed Mater Res Biomed Mater Res A 2007;83:323–31. 1976;10:335–44. [112] Terraciano V, Hwang N, Moroni L, Park HB, Zhang Z, Mizrahi J, [136] Huang J, Best SM, Bonfield W, Buckland T. Development and char- Seliktar D, Elisseeff J. Differential response of adult and embry- acterization of titanium-containing hydroxyapatite for medical onic mesenchymal progenitor cells to mechanical compression in applications. Acta Biomater 2010;6:241–9. hydrogels. Stem Cells 2007;25:2730–8. [137] Yoshikawa H, Myoui A. Bone tissue engineering with porous [113] Schmidt O, Mizrahi J, Elisseeff J, Seliktar D. Immobilized fibrinogen hydroxyapatite ceramics. J Artif Organs 2005;8:131–6. in PEG hydrogels does not improve chondrocyte-mediated matrix [138] Carson JS, Bostrom MP. Synthetic bone scaffolds and fracture repair. deposition in response to mechanical stimulation. Biotechnol Bio- Injury 2007;38:S33–7. eng 2006;95:1061–9. [139] Barralet JE, Grover L, Gaunt T, Wright AJ, Gibson IR. Preparation of [114] Nicodemus GD, Bryant SJ. The role of hydrogel structure and macroporous calcium phosphate cement tissue engineering scaf- dynamic loading on chondrocyte gene expression and matrix for- fold. Biomaterials 2002;23:3063–72. mation. J Biomech 2008;41:1528–36. [140] Dong J, Kojima H, Uemura T, Kikuchi M, Tateishi T, Tanaka J. In vivo [115] Bryant SJ, Anseth KS. Controlling the spatial distribution of ECM evaluation of a novel porous hydroxyapatite to sustain osteogenesis components in degradable PEG hydrogels for tissue engineering of transplanted bone marrow-derived osteoblastic cells. J Biomed cartilage. J Biomed Mater Res A 2003;64:70–9. Mater Res 2001;57:208–16. [116] Bryant SJ, Nicodemus GD, Villanueva I. Designing 3D photopolymer [141] Maspero FA, Ruffieux K, Muller B, Wintermantel E. Resorbable hydrogels to regulate biomechanical cues and tissue growth for defect analog PLGA scaffolds using CO2 as solvent: structural char- cartilage tissue engineering. Pharm Res 2008;25:2379–86. acterization. J Biomed Mater Res 2002;62:89–98. [117] Hacker MC, Klouda L, Ma BB, Kretlow JD, Mikos AG. Syn- [142] Oh SH, Kang SG, Kim ES, Cho SH, Lee JH. Fabrication and char- thesis and characterization of injectable, thermally and chem- acterization of hydrophilic poly(lactic-co-glycolic acid)/poly(vinyl ically gelable, amphiphilic poly(N-isopropylacrylamide)-based alcohol) blend cell scaffolds by melt-molding particulate-leaching macromers. Biomacromolecules 2008;9:1558–70. method. Biomaterials 2003;24:4011–21. [118] Na K, Park JH, Kim SW, Sun BK, Woo DG, Chung HM, Park KH. Deliv- [143] An YH, Woolf SK, Friedman RJ. Pre-clinical in vivo evalu- ery of dexamethasone, ascorbate, and growth factor (TGF beta-3) ation of orthopaedic bioabsorbable devices. Biomaterials in thermo-reversible hydrogel constructs embedded with rabbit 2000;21:2635–52. chondrocytes. Biomaterials 2006;27:5951–7. [144] Kim J, Kim IS, Cho TH, Lee KB, Hwang SJ, Tae G, Noh I, Lee SH, Park [119] Park KH, Lee DH, Na K. Transplantation of poly(N- Y, Sun K. Bone regeneration using hyaluronic acid-based hydrogel isopropylacrylamide-co-vinylimidazole) hydrogel constructs with bone morphogenic protein-2 and human mesenchymal stem composed of rabbit chondrocytes and growth factor-loaded cells. Biomaterials 2007;28:1830–7. nanoparticles for neocartilage formation. Biotechnol Lett [145] Glowacki J, Mizuno S. Collagen scaffolds for tissue engineering. 2009;31:337–46. Biopolymers 2008;89:338–44. [120] Saim AB, Cao Y, Weng Y, Chang CN, Vacanti MA, Vacanti CA, [146] Pedraza CE, Marelli B, Chicatun F, McKee MD, Nazhat SN. An Eavey RD. Engineering autogenous cartilage in the shape of a helix in vitro assessment of a cell-containing collagenous extracellular using an injectable hydrogel scaffold. Laryngoscope 2000;110: matrix-like scaffold for bone tissue engineering. Tissue Eng Part A 1694–7. 2010;16:781–93. [121] Dai W, Kawazoe N, Lin X, Dong J, Chen G. The influence of struc- [147] Nazarov R, Jin HJ, Kaplan DL. Porous 3-D scaffolds from regenerated tural design of PLGA/collagen hybrid scaffolds in cartilage tissue silk fibroin. Biomacromolecules 2004;5:718–26. engineering. Biomaterials 2010;31:2141–52. [148] Ramires PA, Romito A, Cosentino F, Milella E. The influence of [122] Tan H, Huang D, Lao L, Gao C. RGD modified PLGA/gelatin micro- titania/hydroxyapatite composite coatings on in vitro osteoblasts spheres as microcarriers for chondrocyte delivery. J Biomed Mater behaviour. Biomaterials 2001;22:1467–74. Res B Appl Biomater 2009;91:228–38. [149] Kim HW, Knowles JC, Kim HE. Hydroxyapatite/poly(epsilon- [123] Lind M, Larsen A, Clausen C, Osther K, Everland H. Cartilage repair caprolactone) composite coatings on hydroxyapatite porous bone with chondrocytes in fibrin hydrogel and MPEG polylactide scaf- scaffold for drug delivery. Biomaterials 2004;25:1279–87. fold: an in vivo study in goats. Knee Surg Sports Traumatol Arthrosc [150] Lickorish D, Ramshaw JA, Werkmeister JA, Glattauer V, Howlett CR. 2008;16:690–8. Collagen-hydroxyapatite composite prepared by biomimetic pro- [124] Sha’ban M, Yoon SJ, Ko YK, Ha HJ, Kim SH, So JW, Idrus RB, Khang G. cess. J Biomed Mater Res A 2004;68:19–27. Fibrin promotes proliferation and matrix production of interverte- [151] Chen G, Ushida T, Tateishi T. Poly(DL-lactic-co-glycolic acid) sponge bral disc cells cultured in three-dimensional poly(lactic-co-glycolic hybridized with collagen microsponges and deposited apatite par- acid) scaffold. J Biomater Sci Polym Ed 2008;19:1219–37. ticulates. J Biomed Mater Res 2001;57:8–14. [125] Huang X, Zhang Y, Donahue HJ, Lowe TL. Porous thermoresponsive- [152] Hench LL, Polak JM. Third-generation biomedical materials. Science co-biodegradable hydrogels as tissue-engineering scaffolds for 2002;295:1014–7. 3-dimensional in vitro culture of chondrocytes. Tissue Eng [153] Navarro M, Michiardi A, Castano O, Planell JA. Biomaterials in 2007;13:2645–52. orthopaedics. J R Soc Interface 2008;5:1137–58. [126] Babis GC, Soucacos PN. Bone scaffolds: the role of mechanical sta- [154] Datta N, Pham QP, Sharma U, Sikavitsas VI, Jansen JA, Mikos AG. bility and instrumentation. Injury 2005;36:S38–44. In vitro generated extracellular matrix and fluid shear stress syn- [127] Meaney DF. Mechanical properties of implantable biomaterials. ergistically enhance 3D osteoblastic differentiation. Proc Natl Acad Clin Podiatr Med Surg 1995;12:363–84. Sci USA 2006;103:2488–93. 266 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

[155] Chen XD, Dusevich V, Feng JQ, Manolagas SC, Jilka RL. Extracel- [177] Marston WA, Hanft J, Norwood P, Pollak R. The efficacy and safety lular matrix made by bone marrow cells facilitates expansion of Dermagraft in improving the healing of chronic diabetic foot of marrow-derived mesenchymal progenitor cells and pre- ulcers: results of a prospective randomized trial. Diabetes Care vents their differentiation into osteoblasts. J Bone Miner Res 2003;26:1701–5. 2007;22:1943–56. [178] Chen G, Sato T, Ohgushi H, Ushida T, Tateishi T, Tanaka J. Culturing of [156] Hoshiba T, Kawazoe N, Tateishi T, Chen G. Development of stepwise skin fibroblasts in a thin PLGA-collagen hybrid mesh. Biomaterials osteogenesis-mimicking matrices for the regulation of mesenchy- 2005;26:2559–66. mal stem cell functions. J Biol Chem 2009;284:31164–73. [179] Powell HM, Boyce ST. Engineered human skin fabricated using [157] Hoshiba T, Kawazoe N, Tateishi T, Chen G. Development of extracel- electrospun collagen-PCL blends: morphogenesis and mechanical lular matrices mimicking stepwise adipogenesis of mesenchymal properties. Tissue Eng Part A 2009;15:2177–87. stem cells. Adv Mater 2010;22:3042–7. [180] Richardson TP, Peters MC, Ennett AB, Mooney DJ. Polymeric system [158] Pham MT, Reuther H, Maitz MF. Native extracellular matrix coating for dual growth factor delivery. Nat Biotechnol 2001;19:1029–34. on Ti surfaces. J Biomed Mater Res A 2003;66:310–6. [181] Sokolsky-Papkov M, Agashi K, Olaye A, Shakesheff K, Domb AJ. Poly- [159] Lynch MP, Stein JL, Stein GS, Lian JB. The influence of type I collagen mer carriers for drug delivery in tissue engineering. Adv Drug Deliv on the development and maintenance of the osteoblast phenotype Rev 2007;59:187–206. in primary and passaged rat calvarial osteoblasts: modification of [182] Ferreira LS, Gerecht S, Fuller J, Shieh HF, Vunjak-Novakovic G, expression of genes supporting cell growth, adhesion, and extra- Langer R. Bioactive hydrogel scaffolds for controllable vascu- cellular matrix mineralization. Exp Cell Res 1995;216:35–45. lar differentiation of human embryonic stem cells. Biomaterials [160] Mizuno M, Fujisawa R, Kuboki Y. Type I collagen-induced osteoblas- 2007;28:2706–17. tic differentiation of bone-marrow cells mediated by collagen- [183] Park H, Temenoff JS, Tabata Y, Caplan AI, Raphael RM, Jansen JA, alpha2beta1 integrin interaction. J Cell Physiol 2000;184:207–13. Mikos AG. Effect of dual growth factor delivery on chondrogenic [161] Hu Y, Winn SR, Krajbich I, Hollinger JO. Porous polymer scaf- differentiation of rabbit marrow mesenchymal stem cells encap- folds surface-modified with arginine-glycine-aspartic acid enhance sulated in injectable hydrogel composites. J Biomed Mater Res A bone cell attachment and differentiation in vitro. J Biomed Mater 2009;88:889–97. Res A 2003;64:583–90. [184] Jeon O, Ryu SH, Chung JH, Kim BS. Control of basic fibroblast growth [162] McCann TJ, Mason WT, Meikle MC, McDonald F. A collagen peptide factor release from fibrin gel with heparin and concentrations of motif activates tyrosine kinase-dependent calcium signalling path- fibrinogen and thrombin. J Control Release 2005;105:249–59. ways in human osteoblast-like cells. Matrix Biol 1997;16:273–83. [185] Takada T, Katagiri T, Ifuku M, Morimura N, Kobayashi M, Hasegawa [163] Shen H, Hu X, Yang F, Bei J, Wang S. The bioactivity of rhBMP- K, Ogamo A, Kamijo R. Sulfated polysaccharides enhance the bio- 2 immobilized poly(lactide-co-glycolide) scaffolds. Biomaterials logical activities of bone morphogenetic proteins. J Biol Chem 2009;30:3150–7. 2003;278:43229–35. [164] Roeker S, Bohm S, Diederichs S, Bode F, Quade A, Korzhikov V, [186] Levenstein ME, Ludwig TE, Xu RH, Llanas RA, VanDenHeuvel- van Griensven M, Tennikova TB, Kasper C. A study on the influ- Kramer K, Manning D, Thomson JA. Basic fibroblast growth factor ence of biocompatible composites with bioactive ligands toward support of human embryonic stem cell self-renewal. Stem Cells their effect on cell adhesion and growth for the application in bone 2006;24:568–74. tissue engineering. J Biomed Mater Res B Appl Biomater 2009;91: [187] Xu RH, Peck RM, Li DS, Feng X, Ludwig T, Thomson JA. Basic FGF and 153–62. suppression of BMP signaling sustain undifferentiated proliferation [165] Alhadlaq A, Elisseeff JH, Hong L, Williams CG, Caplan AI, Sharma of human ES cells. Nat Methods 2005;2:185–90. B, Kopher RA, Tomkoria S, Lennon DP, Lopez A, Mao JJ. Adult stem [188] Jeon O, Song SJ, Kang SW, Putnam AJ, Kim BS. Enhancement of cell driven genesis of human-shaped articular condyle. Ann Biomed ectopic bone formation by bone morphogenetic protein-2 released Eng 2004;32:911–23. from a heparin-conjugated poly(L-lactic-co-glycolic acid) scaffold. [166] Cao T, Ho KH, Teoh SH. Scaffold design and in vitro study of Biomaterials 2007;28:2763–71. osteochondral coculture in a three-dimensional porous polycapro- [189] Kang SW, La WG, Kang JM, Park JH, Kim BS. Bone morphogenetic lactone scaffold fabricated by fused deposition modeling. Tissue protein-2 enhances bone regeneration mediated by transplan- Eng 2003;9:S103–12. tation of osteogenically undifferentiated bone marrow-derived [167] Schaefer D, Martin I, Shastri P, Padera RF, Langer R, Freed LE, mesenchymal stem cells. Biotechnol Lett 2008;30:1163–8. Vunjak-Novakovic G. In vitro generation of osteochondral compos- [190] Fan VH, Tamama K, Au A, Littrell R, Richardson LB, Wright JW, ites. Biomaterials 2000;21:2599–606. Wells A, Griffith LG. Tethered epidermal growth factor provides [168] Schaefer D, Martin I, Jundt G, Seidel J, Heberer M, Grodzinsky A, a survival advantage to mesenchymal stem cells. Stem Cells Bergin I, Vunjak-Novakovic G, Freed LE. Tissue-engineered compos- 2007;25:1241–51. ites for the repair of large osteochondral defects. Arthritis Rheum [191] Tamama K, Fan VH, Griffith LG, Blair HC, Wells A. Epidermal growth 2002;46:2524–34. factor as a candidate for ex vivo expansion of bone marrow-derived [169] Schek RM, Taboas JM, Segvich SJ, Hollister SJ, Krebsbach PH. mesenchymal stem cells. Stem Cells 2006;24:686–95. Engineered osteochondral grafts using biphasic composite solid [192] Nuttelman CR, Tripodi MC, Anseth KS. Synthetic hydrogel niches free-form fabricated scaffolds. Tissue Eng 2004;10:1376–85. that promote hMSC viability. Matrix Biol 2005;24:208–18. [170] Kon E, Delcogliano M, Filardo G, Fini M, Giavaresi G, Francioli S, [193] Shin H, Temenoff JS, Bowden GC, Zygourakis K, Farach-Carson MC, Martin I, Pressato D, Arcangeli E, Quarto R, Sandri M, Marcacci Yaszemski MJ, Mikos AG. Osteogenic differentiation of rat bone M. Orderly osteochondral regeneration in a sheep model using marrow stromal cells cultured on Arg-Gly-Asp modified hydrogels a novel nano-composite multilayered biomaterial. J Orthop Res without dexamethasone and beta-glycerol phosphate. Biomateri- 2010;28:116–24. als 2005;26:3645–54. [171] MacNeil S. Progress and opportunities for tissue-engineered skin. [194] Shin H, Zygourakis K, Farach-Carson MC, Yaszemski MJ, Mikos Nature 2007;445:874–80. AG. Attachment, proliferation, and migration of marrow stromal [172] Stern R, McPherson M, Longaker MT. Histologic study of artificial osteoblasts cultured on biomimetic hydrogels modified with an skin used in the treatment of full-thickness thermal injury. J Burn osteopontin-derived peptide. Biomaterials 2004;25:895–906. Care Rehabil 1990;11:7–13. [195] Hwang NS, Varghese S, Zhang Z, Elisseeff J. Chondrogenic [173] Bello YM, Falabella AF. The role of graftskin (Apligraf) in difficult- differentiation of human embryonic stem cell-derived cells to-heal venous leg ulcers. J Wound Care 2002;11:182–3. in arginine-glycine-aspartate-modified hydrogels. Tissue Eng [174] Boyce ST, Kagan RJ, Greenhalgh DG, Warner P, Yakuboff KP, Palmieri 2006;12:2695–706. T, Warden GD. Cultured skin substitutes reduce requirements for [196] Kim S, Healy KE. Synthesis and characterization of injectable harvesting of skin autograft for closure of excised, full-thickness poly(N-isopropylacrylamide-co-acrylic acid) hydrogels with burns. J Trauma 2006;60:821–9. proteolytically degradable cross-links. Biomacromolecules [175] Dainiak MB, Allan IU, Savina IN, Cornelio L, James ES, James SL, 2003;4:1214–23. Mikhalovsky SV, Jungvid H, Galaev IY. Gelatin-fibrinogen cryogel [197] Li YJ, Chung EH, Rodriguez RT, Firpo MT, Healy KE. Hydrogels as dermal matrices for wound repair: preparation, optimisation and artificial matrices for human embryonic stem cell self-renewal. J in vitro study. Biomaterials 2010;31:67–76. Biomed Mater Res A 2006;79:1–5. [176] Wainwright D, Madden M, Luterman A, Hunt J, Monafo W, Heim- [198] Lu Y, Mapili G, Suhali G, Chen S, Roy K. A digital micro-mirror bach D, Kagan R, Sittig K, Dimick A, Herndon D. Clinical evaluation of device-based system for the microfabrication of complex, spa- an acellular allograft dermal matrix in full-thickness burns. J Burn tially patterned tissue engineering scaffolds. J Biomed Mater Res Care Rehabil 1996;17:124–36. A 2006;77:396–405. B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268 267

[199] Feng Q, Chai C, Jiang XS, Leong KW, Mao HQ. Expansion of engrafting [223] Yim EK, Reano RM, Pang SW, Yee AF, Chen CS, Leong KW. human hematopoietic stem/progenitor cells in three-dimensional Nanopattern-induced changes in morphology and motility of scaffolds with surface-immobilized fibronectin. J Biomed Mater Res smooth muscle cells. Biomaterials 2005;26:5405–13. A 2006;78:781–91. [224] Pattison MA, Wurster S, Webster TJ, Haberstroh KM. Three- [200] Sharma B, Williams CG, Khan M, Manson P, Elisseeff JH. In vivo dimensional, nano-structured PLGA scaffolds for bladder chondrogenesis of mesenchymal stem cells in a photopolymerized tissue replacement applications. Biomaterials 2005;26: hydrogel. Plast Reconstr Surg 2007;119:112–20. 2491–500. [201] Cristino S, Grassi F, Toneguzzi S, Piacentini A, Grigolo B, Santi S, Ric- [225] Reilly GC, Engler AJ. Intrinsic extracellular matrix properties regu- cio M, Tognana E, Facchini A, Lisignoli G. Analysis of mesenchymal late stem cell differentiation. J Biomech 2010;43:55–62. stem cells grown on a three-dimensional HYAFF 11-based proto- [226] Krieg M, Arboleda-Estudillo Y, Puech PH, Kafer J, Graner F, Muller type ligament scaffold. J Biomed Mater Res A 2005;73:275–83. DJ, Heisenberg CP. Tensile forces govern germ-layer organization [202] Gerecht S, Burdick JA, Ferreira LS, Townsend SA, Langer R, Vunjak- in zebrafish. Nat Cell Biol 2008;10:429–36. Novakovic G. Hyaluronic acid hydrogel for controlled self-renewal [227] Rozario T, Dzamba B, Weber GF, Davidson LA, DeSimone DW. and differentiation of human embryonic stem cells. Proc Natl Acad The physical state of fibronectin matrix differentially regu- Sci USA 2007;104:11298–303. lates morphogenetic movements in vivo. Dev Biol 2009;327: [203] Chang CF, Lee MW, Kuo PY, Wang YJ, Tu YH, Hung SC. Three- 386–98. dimensional collagen fiber remodeling by mesenchymal stem cells [228] Freytes DO, Wan LQ, Vunjak-Novakovic G. Geometry and force con- requires the integrin-matrix interaction. J Biomed Mater Res A trol of cell function. J Cell Biochem 2009;108:1047–58. 2007;80:466–74. [229] Engler AJ, Sen S, Sweeney HL, Discher DE. Matrix elasticity directs [204] Marra KG, Szem JW, Kumta PN, DiMilla PA, Weiss LE. In vitro stem cell lineage specification. Cell 2006;126:677–89. analysis of biodegradable polymer blend/hydroxyapatite compos- [230] Tamada M, Sheetz MP, Sawada Y. Activation of a signaling cascade ites for bone tissue engineering. J Biomed Mater Res 1999;47: by cytoskeleton stretch. Dev Cell 2004;7:709–18. 324–35. [231] Giannone G, Sheetz MP. Substrate rigidity and force define form [205] Kim S, Kim SS, Lee SH, Eun Ahn S, Gwak SJ, Song JH, through tyrosine phosphatase and kinase pathways. Trends Cell Kim BS, Chung HM. In vivo bone formation from human Biol 2006;16:213–23. embryonic stem cell-derived osteogenic cells in poly(d,l-lactic- [232] Ahmed I, Ponery AS, Nur EKA, Kamal J, Meshel AS, Sheetz MP, co-glycolic acid)/hydroxyapatite composite scaffolds. Biomaterials Schindler M, Meiners S. Morphology, cytoskeletal organization, 2008;29:1043–53. and myosin dynamics of mouse embryonic fibroblasts cultured on [206] Kim SS, Sun Park M, Jeon O, Yong Choi C, Kim BS. Poly(lactide- nanofibrillar surfaces. Mol Cell Biochem 2007;301:241–9. co-glycolide)/hydroxyapatite composite scaffolds for bone tissue [233] Bershadsky AD, Balaban NQ, Geiger B. Adhesion-dependent cell engineering. Biomaterials 2006;27:1399–409. mechanosensitivity. Annu Rev Cell Dev Biol 2003;19:677–95. [207] Kim SS, Park MS, Gwak SJ, Choi CY, Kim BS. Accelerated bonelike [234] Pelham Jr RJ, Wang Y. Cell locomotion and focal adhesions apatite growth on porous polymer/ceramic composite scaffolds in are regulated by substrate flexibility. Proc Natl Acad Sci USA vitro. Tissue Eng 2006;12:2997–3006. 1997;94:13661–5. [208] Murphy WL, Hsiong S, Richardson TP, Simmons CA, Mooney DJ. [235] Garcia AJ, Reyes CD. Bio-adhesive surfaces to promote osteoblast Effects of a bone-like mineral film on phenotype of adult human differentiation and bone formation. J Dent Res 2005;84:407–13. mesenchymal stem cells in vitro. Biomaterials 2005;26:303–10. [236] Pek YS, Wan AC, Ying JY. The effect of matrix stiffness on mesenchy- [209] Garreta E, Gasset D, Semino C, Borros S. Fabrication of a three- mal stem cell differentiation in a 3D thixotropic gel. Biomaterials dimensional nanostructured biomaterial for tissue engineering of 2010;31:385–91. bone. Biomol Eng 2007;24:75–80. [237] Pek YS, Wan AC, Shekaran A, Zhuo L, Ying JY. A thixotropic [210] Stevens MM, George JH. Exploring and engineering the cell surface nanocomposite gel for three-dimensional cell culture. Nat Nan- interface. Science 2005;310:1135–8. otechnol 2008;3:671–5. [211] Griffith LG, Swartz MA. Capturing complex 3D tissue physiology in [238] Khan SA, Zoeller NJ. Dynamic rheological behavior of flocculated vitro. Nat Rev Mol Cell Biol 2006;7:211–24. fumed silica suspensions. J Rheol 1993;37:1225–35. [212] Saha K, Pollock JF, Schaffer DV, Healy KE. Designing synthetic [239] Ingber DE. Tensegrity: the architectural basis of cellular mechan- materials to control stem cell phenotype. Curr Opin Chem Biol otransduction. Annu Rev Physiol 1997;59:575–99. 2007;11:381–7. [240] McBeath R, Pirone DM, Nelson CM, Bhadriraju K, Chen CS. Cell [213] Nur EKA, Ahmed I, Kamal J, Schindler M, Meiners S. Three- shape, cytoskeletal tension, and RhoA regulate stem cell lineage dimensional nanofibrillar surfaces promote self-renewal in mouse commitment. Dev Cell 2004;6:483–95. embryonic stem cells. Stem Cells 2006;24:426–33. [241] Yang Y, Relan NK, Przywara DA, Schuger L. Embryonic mesenchy- [214] Oh S, Brammer KS, Li YS, Teng D, Engler AJ, Chien S, Jin S. Stem cell mal cells share the potential for smooth muscle differentiation: fate dictated solely by altered nanotube dimension. Proc Natl Acad myogenesis is controlled by the cell’s shape. Development Sci USA 2009;106:2130–5. 1999;126:3027–33. [215] Silva GA, Czeisler C, Niece KL, Beniash E, Harrington DA, Kessler [242] Kilian KA, Bugarija B, Lahn BT, Mrksich M. Geometric cues for direct- JA, Stupp SI. Selective differentiation of neural progenitor cells by ing the differentiation of mesenchymal stem cells. Proc Natl Acad high-epitope density nanofibers. Science 2004;303:1352–5. Sci USA 2010;107:4872–7. [216] Hwang NS, Kim MS, Sampattavanich S, Baek JH, Zhang Z, Elisseeff [243] James J, Goluch ED, Hu H, Liu C, Mrksich M. Subcellular J. Effects of three-dimensional culture and growth factors on the curvature at the perimeter of micropatterned cells influences chondrogenic differentiation of murine embryonic stem cells. Stem lamellipodial distribution and cell polarity. Cell Motil Cytoskeleton Cells 2006;24:284–91. 2008;65:841–52. [217] Curtis AS, Gadegaard N, Dalby MJ, Riehle MO, Wilkinson CD, [244] Burdick JA, Vunjak-Novakovic G. Engineered microenvironments Aitchison G. Cells react to nanoscale order and symmetry in their for controlled stem cell differentiation. Tissue Eng Part A surroundings. IEEE Trans Nanobiosci 2004;3:61–5. 2009;15:205–19. [218] Teixeira AI, Nealey PF, Murphy CJ. Responses of human keratocytes [245] Huang H, Nakayama Y, Qin K, Yamamoto K, Ando J, Yamashita J, Itoh to micro- and nanostructured substrates. J Biomed Mater Res A H, Kanda K, Yaku H, Okamoto Y, Nemoto Y. Differentiation from 2004;71:369–76. embryonic stem cells to vascular wall cells under in vitro pulsatile [219] Price RL, Ellison K, Haberstroh KM, Webster TJ. Nanometer surface flow loading. J Artif Organs 2005;8:110–8. roughness increases select osteoblast adhesion on carbon nanofiber [246] Riha GM, Wang X, Wang H, Chai H, Mu H, Lin PH, Lumsden AB, compacts. J Biomed Mater Res A 2004;70:129–38. Yao Q, Chen C. Cyclic strain induces vascular smooth muscle cell [220] Andersson AS, Backhed F, von Euler A, Richter-Dahlfors A, differentiation from murine embryonic mesenchymal progenitor Sutherland D, Kasemo B. Nanoscale features influence epithe- cells. Surgery 2007;141:394–402. lial cell morphology and production. Biomaterials [247] Orr AW, Helmke BP, Blackman BR, Schwartz MA. Mechanisms of 2003;24:3427–36. mechanotransduction. Dev Cell 2006;10:11–20. [221] Zinger O, Zhao G, Schwartz Z, Simpson J, Wieland M, Landolt [248] Altman GH, Horan RL, Martin I, Farhadi J, Stark PR, Volloch V, Rich- D, Boyan B. Differential regulation of osteoblasts by substrate mond JC, Vunjak-Novakovic G, Kaplan DL. Cell differentiation by microstructural features. Biomaterials 2005;26:1837–47. mechanical stress. FASEB J 2002;16:270–2. [222] Webster TJ, Schadler LS, Siegel RW, Bizios R. Mechanisms of [249] Gwak SJ, Bhang SH, Kim IK, Kim SS, Cho SW, Jeon O, Yoo KJ, Put- enhanced osteoblast adhesion on nanophase alumina involve vit- nam AJ, Kim BS. The effect of cyclic strain on embryonic stem ronectin. Tissue Eng 2001;7:291–301. cell-derived cardiomyocytes. Biomaterials 2008;29:844–56. 268 B.-S. Kim et al. / Progress in Polymer Science 36 (2011) 238–268

[250] Li TS, Hayashi M, Ito H, Furutani A, Murata T, Matsuzaki [251] Bhang SH, Gwak SJ, Lee TJ, Kim SS, Park HH, Park MH, Lee DH, Lee SH, M, Hamano K. Regeneration of infarcted myocardium by Kim BS. Cyclic mechanical strain promotes transforming-growth- intramyocardial implantation of ex vivo transforming growth factor-beta1-mediated cardiomyogenic marker expression in factor-beta-preprogrammed bone marrow stem cells. Circulation bone-marrow-derived mesenchymal stem cells in vitro. Biotechnol 2005;111:2438–45. Appl Biochem 2010;55:191–7.