In vitro engineering of transitional tissue by patterning and functional control of cells in fibrin gel

Abstract

Tendon insertions show specific three-dimensional (3D) cell and matrix configurations. To fabricate this transitional tissue in vitro, the development of a cell manipulation technique that can control cell and matrix patterning in the 3D construct is crucial. Bone marrow-derived stromal cells and osteoblasts were respectively cultured in fibrin gel using a custom-made device that can apply uniaxial continuous tensile strain to the gel at different strain rates. Cells in the strained fibrin gel showed a specific orientation, which was parallel to the strain direction, because of the structural change of fibrin fibres within the strained gel. The direction of cell proliferation in the strained gel was also restricted to the same direction. Subsequently, linearly aligned cell sets, similar to the cell patterning of tendon tissue, were obtained in this culture system. Notably, linearly aligned extracellular matrix and mineral patterning, which are crucial for the fabrication of tendon insertions, were also confirmed in the strained fibrin gel. A cell proliferation assay indicated that the strained gel with a higher strain rate enhanced cell proliferation in the gel. In contrast, a Real-Time PCR study indicated that the strained gel with a lower strain rate enhanced cell differentiation in the gel. Thus, our new cell culture system enabled us to fabricate a 3D construct containing patterned cells and matrices in a controllable manner. This would be an effective tool for reproducing tendon insertions in vitro.

---

Introduction

In the case of tumors, bone necrosis caused by X-ray radiations, or injuries requiring orthopedic or craniofacial surgery,^1,2 regional segmentation of tissues, including bone, muscles, nerves, and blood vessels, is carried out. For bone reconstruction in these regions, biomaterials, including titanium³ or sintered ceramic materials,⁴ are often applied to rebuild the original morphology of the defected bone.⁵ However, this type of bone reconstruction has a limited ability for generating complicated bone motility since the reconstructed materials are not connected to skeletal muscles. Recent progress in tissue engineering using biomaterials and biologic materials (e.g., cells and soluble factors) to generate biological tissue has introduced a number of methods for regenerating bone^6,7 or muscle tissue,^8,9 respectively. However, effective connections between bone and muscle tissue still remains a critical issue.¹⁰

Tendon tissue is crucial not only for connecting bone and muscle tissue, but also for generating force from muscle tissue to bone for motility.¹¹ Since tendon or ligament tissue possesses a limited capacity for regeneration,¹² a number of therapies including tissue grafts (e.g., autografts, allografts, and xenografts) and artificial prostheses has been carried out. However, only limited amounts of autografts from a patient are available, and it may result in donor site morbidity.¹³ Allografts from cadavers may cause a harmful response from the immune system of a patient.¹⁴ Permanent artificial prostheses often lack material durability, which may lead to mechanical failures in the future.¹⁵ Therefore, alternative tendon repair options need to be explored.

Anatomically, transitional tissue is divided into four groups: tendons, fibrocartilage, mineralised fibrocartilage, and bone.¹⁶ Therefore, mineralisation control of the cells comprising three-dimensional (3D) configurations is crucial for reproducing tendon insertions. Histologically, tendon tissues show specific cell and matrix orientations in their 3D configurations.¹⁷ Therefore, controlling cell patterning as well as matrix and mineral patterning in the 3D configurations would be crucial for reproducing this connective tissue. To date, two-dimensional (2D) cell patterning, e.g., cell alignment control on micro-patterned surfaces fabricated by micro electro mechanical systems (MEMS)¹⁸ or template-assisted electrohydrodynamic atomisation spraying, has been introduced.^19,20 These methods are successfully used to create cell arrays for biological analysis or to improve the biomaterial characteristics so that the materials have higher affinity to the host tissue. However, these methods of 2D cell patterning can be applied only to the surface of a material, not in the internal region of a 3D construct. In consideration of 3D cell patterning, a 3D cell printer was developed to seed cells at appropriate 3D positions in a configuration.^21,22 By this method, the initial positions of cells in a 3D construct can be controlled, but the cells may show random positioning in the construct because of their proliferation and migration during the cell culture period. Moreover, patterning of the extracellular matrix (ECM) and minerals are not achieved in the 3D construct by this method. Our previous study indicated that applying continuous uniaxial tensile strain to fibrin gel oriented internal fibrin fibrils in a specific direction that was parallel to the strained direction.²³ In addition, these aligned fibrils led to cell alignment in the constructs, which was parallel to the strain direction, because of the physical attenuation triggered by fibrin alignments inhibiting random cell migration. Thus, 3D constructs containing uniaxially oriented cell sets have already been formed successfully. However, matrix patterning and cell functions, including cell proliferation and differentiation, in this strained fibrin gel system are not yet fully understood. In this study, we cultured bone marrow-derived stromal cells (BMSCs) and osteoblasts respectively in fibrin gel under various applications of strain and evaluated the cell and matrix patterning. Cell proliferation and differentiation under the different strain rates were also investigated to determine the optimal conditions for engineering transitional tissue in vitro.

Experimental section

Preparation of strained fibrin gel containing cells

Pre-osteoblasts (MC3T3-E1; Riken, Japan) and mouse BMSCs (from bulb/c mouse bone marrow) containing a fibrinogen solution (4 mg/ml, Sigma-Aldrich, MO, USA) were respectively mixed with a thrombin solution (2.5 U/ml, Sigma-Aldrich) in a 1 : 1 ratio and poured into a cylindrical silicone mold (length: 10 mm, diameter: 6 mm). The mixture solution was kept in an incubator at 37 °C for 30 min. The resulting fibrin gel was subjected to continuous uniaxial strain by using a custom-made device (Fig. 1a). The length of the gel extension (L) from the initial gel length (L₀) was defined as the applied strain (strain ratio = L/L₀ × 100). Our previous study indicated that the strained fibrin gel formed by using this device contained uniaxially aligned fibrin bundles (Fig. 1b).²³ Fibrin gel containing cells without application of strain was defined as the control.

Fig. 1 (a) Fibrin gels were subjected to continuous tensile strain using a custom-made device. Length of fibrin gel extension (L) from initial gel length (L₀) was defined as applied strain (L/L₀ × 100%). (b) Bundle-like structure was formed in strained fibrin gel and showed specific alignment parallel to strain direction (bar: 20 μm). Arrows in figure indicate strain direction.

Osteoblasts in fibrin gel were cultured with Dulbecco's Modified Eagle's Medium (DMEM) containing 10% fetal bovine serum (FBS). BMSCs derived from mouse bone marrow in fibrin gels were cultured with Minimum Essential Medium Alpha (MEMα) containing 20% FBS, β-glycerophosphate disodium salt hydrate (1 × 10⁻² M, Sigma-Aldrich), ascorbic acid (50 μg/ml, Sigma-Aldrich), dexamethasone (10⁻⁶ M, Sigma-Aldrich), 10 ng/ml recombinant human bone morphogenetic protein-2 (BMP-2; PeproTech, NJ, USA), and 10 mM of calcium chloride solution for controlling the calcium concentration in the medium (dif-MEM). An anti-fibrinolytic agent, aprotinin (5 mg/ml, Sigma-Aldrich), was added to both mediums.

Cell and matrix patterning in strained fibrin gel

The cells contained in the control gel and 50%-strained fibrin gels were observed using a light microscope (TE2000, Nikon, Japan). The angles formed between arbitrary cells and the strain direction in the gel were quantified with Image J (NIH, MD, USA). Fibrin gel containing osteoblasts was embedded in paraffin, which was then cut into 5-μm thick sections. For histological evaluation, the sections were deparaffinised then incubated in PBS containing 0.1% triton-X and 1% bovine serum albumin for 20 min. After being washed twice, the sections were incubated with anti-mouse type I collagen (Chemicon, MA, USA) for 40 min, then incubated with a secondary antibody conjugated with Alexa Fluoro 488 (Invitrogen, CA, USA) for 40 min, followed by nuclear staining with Hoechst33342 (Invitrogen). Mineralisation in the fibrin gel containing BMSCs was investigated by von Kossa staining at day 50. Microbeam X-ray diffraction analysis (M18XHF22-SR, Mac Science, Japan)²⁴ and scanning electron microscopy (SEM; JSM-6390, JEOL, Japan) combined with an energy dispersive X-ray spectrometer (EDS) were used to characterise the minerals in the fibrin gel.²⁵ For microbeam X-ray diffraction analysis, the relative intensity of (002) to (211) of the deposited minerals in the fibrin gel was investigated to evaluate the orientations of the crystals. For SEM observation, fibrin gels were fixed with 4% paraformaldehyde, post-fixed with 1% OsO₄, and dehydrated in a series of ethanol concentrations of up to 100%. The gels were critical-point dried with CO₂ for 45 min, mounted, and then sputter-coated with gold.

Cell function control in strained fibrin gel

The number of cells in the fibrin gels was adjusted to 5.0 × 10³ cells/gel. The gels, with different strain rates (0–50% strain) and containing cells, were washed with PBS and trypsinised for 30 min at days 2, 5, and 8 to dissolve the fibrin gel. The number of cells in the gels at different strain rates was counted using a hemocytometer. To analyse the phenotypic difference of osteoblasts, the total RNA of the osteoblasts in the fibrin gel at different strain rates was extracted using Trizol reagent (Invitrogen) with RNeasy Mini Kit (QIAGEN, Germany). After DNase I treatment (Ambion, TX, USA), cDNAs were synthesised from 2 mg of the total RNA using Super Script III reverse transcriptase (Invitrogen). The cDNAs of osteopontin (Opn: Mm00436767), osteocalcin (Oc: Mm00649782), and glyceraldehyde 3-phosphate dehydrogenase (Gapdh: NM_008084) were amplified by a polymerase chain reaction using a TaqMan probe (Applied Biosystems, CA, USA). The expression of each cDNA was analysed with a 7300 Real Time PCR system (Applied Biosystems).

A confirmation study to prove the effect of strain rate on cell functions in strained fibrin gel was carried out. BMSCs were initially cultured in fibrin gel with 50% strain. Then, at day 21 or 28, the strain rate of that fibrin gel was reduced to 0%. Cells cultured in fibrin gel with 50% strain for 50 days were prepared as a control. At day 50, the fibrin gel under each of these conditions was embedded in paraffin after fixation with paraformaldehyde. The paraffin was then cut into 5-μm thick sections for von Kossa staining. The area of mineral deposition in the stained samples was analysed using image analysis software (Photoshop CS, Adobe Systems, CA, USA).²⁶

Statistical analysis

Statistical analysis of the obtained data was accomplished using one factor analysis of variance. Student's t-test was used for comparison at a 95% confidence interval. All data points and error bars in the graphs represent the mean value and standard deviation (SD), respectively, calculated from four independent experiments.

Results

Three-dimensional dynamics of osteoblasts in strained fibrin gel

Osteoblasts in the control gel without uniaxial tensile strain showed random orientations. In contrast, the cells in the strained gel displayed a specific alignment that was parallel to the strain direction in the gel (Fig. 2a–c). Furthermore, the direction of cell proliferation was identical to that of cell alignment. Consequently, the oriented cells formed a number of linear cell groups that were aligned parallel to the strain direction in the strained gel.

Fig. 2 Cells and ECM dynamics in fibrin gel. Light microscopic images of cells (indicated by white arrows) in (a) control fibrin gel and in (b) 50%-strained fibrin gel (bar: 200 μm). (c) Alignment of cells in strained gel was determined using image-analysis software Image-J. Fluorescent immunostaining of type I collagen (green) and nuclear staining (red) in (d) control fibrin gel and in (e) strained fibrin gel (bar: 250 μm). Mineral depositions detected by von Kossa staining of (f) control gel and (g) strained gel (bar: 50 μm). Red arrows indicate strain direction.

Matrix and mineral patterning in strained fibrin gel

Type I collagen, the main matrix protein of osteoblasts, showed random deposition according to the random distribution of cells in the control gel (Fig. 2d). In contrast, in the strained fibrin gel, type I collagen deposition showed a specific orientation that was parallel to the strain direction (Fig. 2e). Merged images with nuclear-stained cells indicated that the matrix deposition was identical to the cell positions in the gel. When BMSCs were cultured with the osteogenic differentiation medium (dif-MEMα), mineralisation derived from the cells was observed in the fibrin gel. The mineralisation in the control gel distributed radially from the cells. The stain colour, indicating mineral deposition, lightened depending on the distance from the cells (Fig. 2f). Similar to the matrix deposition, mineral deposition was localised according to the cell position and showed specific alignment in the strained gel (Fig. 2g).

X-ray diffraction (XRD) peak profiles revealed that the obtained mineral in the fibrin gel was hydroxyapatite (HAp) (Fig. 3a). Additionally, specific orientation of the HAp crystals parallel to the strain direction was confirmed from the relative intensity of (002) to (211) in the XRD profiles (Fig. 3b). SEM images indicated that the mineralised matrix vesicles were concatenated serially parallel to the strain direction in the strained gel compared to those in the control gel (Fig. 3c–f). EDS analysis revealed that the mineralised matrix vesicles contained both Ca and P. The concentrations of P at the mineralised vesicle regions were much higher than those of Ca (Fig. 3g).

Fig. 3 (a) XRD pattern of deposited minerals in strained fibrin gel embedded in paraffin (upper) and of embedding material (paraffin, lower). (b) Peak intensity ratio of (002)/(211) of deposited minerals obtained in strained fibrin gel (*: p < 0.05). SEM images of mineralised matrix vesicles with different magnification in control gel (c, d) and in strained fibrin gel (e, f) at day 50 (c, e; bar: 10 μm; d, f; bar: 2 μm). Arrows indicate strain direction. (g) EDS profile of mineralised matrix vesicles in strained fibrin gel.

Cell functions in strained fibrin gel with different strain rates

To investigate the alternations of cellular functions in strained fibrin gel, osteoblasts were cultured in gels with different strain rates. At day 8, the gel subjected to a higher strain rate had enhanced cell proliferation compared to the gel subjected to a lower strain rate (Fig. 4a). The mRNA expressions of Opn and Oc, osteogenic differentiation markers, were investigated at day 4. Both Opn and Oc expressions decreased with the increase of the strain rate from 0 to 50% (Fig. 4b–c). These results suggested that cell functions in the strained fibrin gel were regulated by the alteration of strain rate. To confirm this, cell-derived mineralisation in the gel in which strain rate was changed during the culture period was investigated. Mineralisation caused by cell differentiation was detected only in the sample in which the strain rate was reduced from 50 to 0% at day 21, as shown in Fig. 5 (III). In contrast, mineral deposition was not detected in the gel in which the strain rate was reduced from 50 to 0% at day 28 (II) or in the gel with 50% strain maintained for 50 days (I).

Fig. 4 (a) Osteoblast (MC3T3-E1) proliferation in fibrin gel subjected to different strain rates. mRNA expression of (b) osteopontin (Opn) and (c) osteocalcin (Oc) of osteoblasts in fibrin gel at day 4.

Fig. 5 Mineralised area in strained fibrin gel cultured with different strain conditions. (I) 50% strain for 50 days, (II) 50% strain for 28 days reduced to 0% strain for another 22 days, and (III) 50% strain for 21 days reduced to 0% strain for another 29 days (*: p < 0.05).

Discussion

New approaches for repairing tendon defects have been introduced in tissue engineering studies. For example, functional connection between bone and muscle tissue by using knitted poly-lactic-co-glycolic loaded with BMSCs²⁷ or cell-collagen scaffolds combined with mechanical stimulation²⁸ has been reported. In these in vivo approaches, the authors indicated the potential of polymer materials combined with cells for connective tissue reconstruction. However, the cell, matrix, and mineral organisations in these constructs, which were similar to those in natural tendon tissue, were not well documented. Since tendon tissue shows well-oriented patterning of cells and matrices and the gradient composition of both organics and inorganics in the tendon tissue configuration,²⁹ to control cell functions, including cell patterning, proliferation, and differentiation, in 3D constructs is extremely important for obtaining cellular-based materials with more functionality for connective tissue regeneration. Therefore, in vitro manipulations of cells in gel material were conducted in this study.

Fibrin is a fibrous protein that forms fibrin clots to prevent further bleeding from an injured blood vessel. Because fibrinogen and thrombin, the sources of fibrin, can be obtained from the peripheral blood of any individual,³⁰ fibrin gel is one of the ideal biomaterials for regenerating tissue when applied to individuals. Therefore, fibrin gel was selected as a cell substrate in this study.

Osteoblasts and BMSCs showed a specific cell alignment that was parallel to the strain direction in the fibrin gel when the gel containing cells was subjected to a continuous tensile strain by using a custom-made device. The direction of cell proliferation was also determined to be the same as that of cell alignment because of the restriction by the aligned fibrin bundles formed by aggregation of the fibrin fibrils in the strained gel. Finally, a 3D construct composed of gel with linearly aligned cell sets was obtained. This result was identical to our previously reported results, in which different cell types, including myoblasts and endothelial cells, were used in the strained gel.²³ Importantly, the cells in the strained gel maintained their patterning during the culture period. Fluorescent immunostaining of type I collagen revealed that the deposited type I collagen showed specific alignments identical to the cell positions in the strained fibrin gel. Moreover, the mineral deposition generated by the cells also showed specific alignment that was identical to the cell and matrix positions in the strained fibrin gel. Thus, specific alignments of the matrix and minerals in the 3D configuration were obtained by using this system. In the control gel, matrix and mineral deposition distributed radially over the culturing period. Previous studies indicated that magnetically aligned fibrils in fibrin gel enhanced the specific alignment of the cells. However, linearly aligned cell sets or matrix patterning were not observed in the system.³¹ From these points, it was suggested that the specific alignments of the matrix and minerals obtained in the strained gel were also caused by the physical restriction of matrix diffusion because of the dense aligned fibrin bundles formed by the continuous tensile strain to the gel. Indeed, the aligned fibrin bundles consisting of tightly aggregated fibrin fibrils showed no spaces or pores within the bundles in the SEM images.

One of the mineralisation processes in biological hard tissue development is considered to begin at matrix vesicles, which contain alkaline phosphatase and bone matrix proteins.³² Mineralised matrix vesicles observed in the gel by using SEM indicated that a similar mechanism of biomineralisation was adopted in this fibrin gel system. Interestingly, mineralised vesicles identified as HAp by XRD also indicated the directional connections of neighbouring mineralised vesicles. Therefore, it was suggested that the alignment of fibrin bundles in the strained gel also played a significant role in the orientation of HAp crystals. EDS analysis indicated that these mineralised matrix vesicles contained both Ca and P, but the concentrations of P were much larger than those of Ca. This is because the matrix vesicles consist of phospholipids similar to the cellular membrane.

The cells were exposed to extension and compressive stress within the gel by the application of continuous tensile strain. This unique physical environment for cells might trigger cell specialization. Therefore, the alternation of cellular functions in strained fibrin gel with a variety of strain rates was investigated. The results indicated that a higher strain rate enhanced cell proliferation as well as suppressed cell differentiation in the gel. In contrast, a lower strain rate enhanced cell differentiation and attenuated cell proliferation. These results suggested that cell functions in strained fibrin gel could be regulated by the gel strain rate. To confirm this idea, cell-derived mineralisation in gels with different strain rates at different culture periods was investigated. Note that the amount of mineralisation was significantly enhanced when the strain rate of the gel was diminished at day 21, whereas no mineral deposition was detected in the samples in which a strain rate of 50% was maintained for 50 days or in which the strain rate was reduced at day 28. Thus, we demonstrated that the fibrin gel system used in this study enabled us to control cell functions by altering the strain rate of the gel. Recent studies have demonstrated that mechanical stimulation or physical environments have pivotal roles in the management of cell functions, including cell proliferation and differentiation.^33–35 For example, mesenchymal stem cells cultured on gel substrates with different rigidities indicated different cell phenotypes, including neurons, myoblasts, or osteoblasts, depending on the rigidity.³³ Kong et al. revealed that more rigid gel enhanced osteoblast proliferation, whereas less rigid gel enhanced osteoblast differentiation.³⁵ Our previous study indicated that the mechanical property of strained fibrin gel increased with the increase of the strain rate.²³ From these points, fibrin gel rigidity altered by the strain application would be one of the important factors for regulating the functions of cells in strained gel in this study.

Conclusion

Three-dimensional cell and matrix patterning, which is similar to transitional tendon tissue, was formed in 3D fibrin gel by applying continuous tensile strain to fibrin gel containing cells. Aligned mineral depositions, showing identical patterning with cell and matrix positions, were also confirmed in the strained fibrin gel. Osteoblastic cell functions, including cell proliferation and differentiation, were regulated by the alteration of gel strain rate in this special cell culture condition. Thus, we developed a new system enabling a 3D construct containing patterned cells and matrices to be obtained in a controllable manner. This would be an effective tool for reproducing tendon insertion in vitro.

Acknowledgements

This study was supported by a Grant-in-Aid for Scientific Research from the Ministry of Education, Culture, Sports, Science and Technology of Japan (MEXT; 18680039, 19659503) and Japan-Korea cooperative research program from Japan Society of the Promoting Science (JSPS) and Korea Science and Engineering Foundation (KOSEF).

References

1. J. C. Reichert, S. Saifzadeh, M. E. Wullschleger, D. R. Epari, M. A. Schütz, G. N. Duda, H. Schell, M. van Griensven, H. Redl and D. W. Hutmacher, Biomaterials, 2009, 30, 2149.
2. J. Nishida and T. Shimamura, Med. Sci. Monit., 2008, 14, RA107.
3. A. Bagno and C. Di Bello, J. Mater. Sci.: Mater. Med., 2004, 15, 935.
4. H. Ohgushi, N. Kotobuki, H. Funaoka, H. Machida, M. Hirose, Y. Tanaka and Y. Takakura, Biomaterials, 2005, 26, 4654.
5. C. I. Bonzani, J. H. George and M. M. Stevens, Curr. Opin. Chem. Biol., 2006, 10, 568.
6. K. J. Burg, S. Porter and J. F. Kellam, Biomaterials, 2000, 21, 2347.
7. M. Yamamoto, Y. Takahashi and Y. Tabata, Tissue Eng., 2006, 12, 1305.
8. A. D. Bach, J. P. Beier, J. Stern-Staeter and R. E. Horch, J. Cell. Mol. Med., 2004, 8, 413.
9. M. N. Giraud, E. Ayuni, S. Cook, M. Siepe, T. P. Carrel and H. T. Tevaearai, Artif. Organs, 2008, 32, 692.
10. L. M. Larkin, S. Calve, T. Y. Kostrominova and E. M. Arruda, Tissue Eng., 2006, 12, 3149.
11. N. Juncosa-Melvin, K. S. Matlin, R. W. Holdcraft, V. S. Nirmalanandhan and D. L. Butler, Tissue Eng., 2007, 13, 1219.
12. S. L. Woo, S. D. Abramowitch, R. Kilger and R. Liang, J. Biomech., 2006, 39, 1.
13. D. M. Doroski, K. S. Brink and J. S. Temenoff, Biomaterials, 2007, 28, 187.
14. T. Tischer, S. Vogt, S. Aryee, E. Steinhauser, C. Adamczyk, S. Milz, V. Martinek and A. B. Imhoff, Arch. Orthop. Trauma Surg., 2007, 127, 735.
15. P. Kannus, Scand. J. Med. Sci. Sports, 2000, 10, 312.
16. B. Wopenka, A. Kent, J. D. Pasteris, Y. Yoon and S. Thomopoulos, Appl. Spectrosc., 2008, 62, 1285.
17. Y. Liu, H. S. Ramanath and D. A. Wang, Trends Biotechnol., 2008, 26, 201.
18. X. Jiang, R. Ferrigno, M. Mrksich and G. M. Whitesides, J. Am. Chem. Soc., 2003, 125, 2366.
19. Z. Ahmad, E. S. Thian, J. Huang, M. J. Edirisinghe, S. M. Best, S. N. Jayasinghe, W. Bonfield, R. A. Brook and N. Rushton, J. Mater. Sci.: Mater. Med., 2008, 19, 3093.
20. X. Li, G. Koller, J. Huang, L. D. Silvio, T. Renton, M. Esat, W. Bonfield and M. Edirisinghe, J. R. Soc. Interface, 2010, 7, 189.
21. T. Xu, J. Jin, C. Gregory, J. J. Hickman and T. Boland, Biomaterials, 2005, 26, 93.
22. M. Nakamura, A. Kobayashi, F. Takagi, A. Watanabe, Y. Hiruma, K. Ohuchi, Y. Iwasaki, M. Horie, I. Morita and S. Takatani, Tissue Eng., 2005, 11, 1658.
23. T. Matsumoto, J. Sasaki, E. Alsberg, H. Egusa, H. Yatani and T. Sohmura, PLoS One, 2007, 2, e1211.
24. T. Nakano, K. Kaibara, Y. Tabata, N. Nagata, S. Enomoto, E. Marukawa and Y. Umakoshi, Bone, 2002, 31, 479.
25. R. Yoh, T. Matsumoto, J. Sasaki and T. Sohmura, J. Biomed. Mater. Res. A, 2008, 87, 222.
26. H. Egusa, F. E. Schweizer, C. C. Wang, Y. Matsuka and I. Nishimura, J. Biol. Chem., 2005, 280, 23691.
27. H. W. Ouyang, J. C. Goh, A. Thambyah, S. H. Teoh and E. H. Lee, Tissue Eng., 2003, 9, 431.
28. V. S. Nirmalanandhan, M. Rao, J. T. Shearn, N. Juncosa-Melvin, C. Gooch and D. L. Butler, J. Biomech., 2008, 41, 822.
29. H. Sano, Y. Saijo and S. Kokubun, J. Shoulder Elbow Surg., 2006, 15, 743.
30. T. Hasegawa, K. Okada, Y. Takano, Y. Hiraishi and Y. Okita, J.Thorac. Cardiovasc. Surg., 2007, 133, 1268.
31. N. Dubey, P. C. Letourneau and R. T. Tranquillo, Biomaterials, 2001, 22, 1065.
32. T. Borrás and N. Comes, Exp. Eye Res., 2009, 88, 738.
33. A. J. Engler, S. Sen, H. L. Sweeney and D. E. Discher, Cell, 2006, 126, 677.
34. T. Kanno, T. Takahashi, T. Tsujisawa, W. Ariyoshi and T. Nishihara, J. Cell. Biochem., 2007, 101, 1266.
35. H. J. Kong, J. Liu, K. Riddle, T. Matsumoto, K. Leach and D. J. Mooney, Nat. Mater., 2005, 4, 460.

---