A facile approach to improve light induced cell sheet harvesting through nanostructure optimization

Abstract

In the present study, the effects of nanostructure on the light-induced cell detachment property of anatase TiO₂ films are investigated and discussed. Anatase dense films, nanodots films and porous films are prepared through sol–gel and phase separation-induced self-assembly based methods. It is found that all the films possessed good cell adhesion but there is a decrease in cell proliferation in the dense film. The porous film exhibits improved cell detachment performance in both single cell and cell sheet detachment tests. It only takes about 3 min of ultraviolet illumination for cell sheets to detach from the porous film, while those for nanodots film and dense film are 5 min and 10 min, respectively. The cell sheets harvested show good performance and thus can be further utilized in cell sheet tissue engineering. The beneficial effects of film nanostructure are ascribed to the differences of initial adsorption status of extracellular proteins, as well as their changes after ultraviolet illumination. The present work demonstrates that optimization of film nanostructures could be an effective approach to improve light induced cell sheet harvesting.

---

Introduction

Cell sheet technology is an attractive approach in tissue engineering because of the directly transplantable cell sheets or three-dimensional layers, which are composed of different individual cell sheets, become possible.¹ In order to harvest these high-activity cell sheets, thermo-responsive,² electricity induced,³ magnetism induced⁴ and pH change induced⁵ methods have been developed. However, there is still much room to improve the detachment efficiency and cell viability of the detached cell sheets. TiO₂ has been reported to possess excellent biocompatibility, non-toxicity and chemical stability, thereby classified as a bioactive material.^6–9 Previously, it is demonstrated that TiO₂ nanodots film could be utilized to achieve cell detachment through 365 nm ultraviolet (UV365) illumination.¹⁰ Mouse calvaria-derived, pre-osteoblastic MC3T3-E1 cells could easily developed into cell sheets, and detach from the films with extracellular matrix spontaneously with UV365 illumination. The reason for such light induced cell sheet detachment was ascribed to the light responses of extracellular matrix proteins.¹⁰

Moreover, many studies have reported that surface morphology can greatly influence protein adsorption and cell behaviour.^11–13 Influencing of the tissue response through surface morphology becomes one of the most important criteria for the development of high performance biomaterials, since these criteria are decisive for the cell–material interaction. Designing materials surfaces possessing expected nanostructure to guide cell migration, shape cell morphology, and direct cell differentiation has drawn enormous attention in biomedical researches.¹⁴ There are many reports about surface topography influence the cell adhesion, proliferation and differentiation.^15,16 However, few works have been done on cell detachment. Since anatase nanodots films have shown good performance of light-induced cell detachment, it is interesting to explore the effect of the nanostructure on TiO₂ films on cell detachment performance. The present study is aimed to investigate the influence of nanostructure of anatase TiO₂ films on light-induced cell detachment. Nanodots film, dense film and porous film were chosen. Moreover, in order to investigate whether this method is also effective to other cell lines, mouse NIH3T3 cells were used instead of the MC3T3 cells used in the previous work.¹⁰ Also, the reason of such influences were characterized and discussed in a protein adsorption angle. The present work could be utilized to optimize light-induced cell detachment behavior and know more about on light-induced cell detachment behavior.

Results and discussion

Morphology and crystalline phase of TiO₂ films

Surface morphology of TiO₂ films was shown in Fig. 1. For nanodots film, the diameter of the TiO₂ nanodots ranged from 40 to 110 nm, and the dot density was about 6 × 10¹⁰/cm²; while the surface of TiO₂ dense film was smooth and dense. As for porous film, it was observed that the film was actually composed of TiO₂ grains, which formed pores approximately ranging from 10 to 100 nm.

Fig. 1 SEM micrographs of TiO₂ films with different nanostructure. (a) Nanodots film, (b) dense film, (c) porous film.

Fig. 2a exhibited the X-ray diffraction patterns of TiO₂ films. The peaks at 25.35°, 37.83° and 48.08° corresponded well to the (101), (004) and (200) planes of anatase. Moreover, in Fig. 2b, the anatase phases of TiO₂ could be further identified through Raman spectra. As showed in Fig. 2b, TiO₂ porous film, TiO₂ dense film, TiO₂ nanodots film after heat treatment at 500 °C showed a strong Raman shifts at 144 cm⁻¹ which could be ascribed to E_g mode of anatase phase (the shift at 144 cm⁻¹ is the strongest one for the anatase phase).^17,18 The results of Raman spectra, in general, were in well consistent with XRD analysis.

Fig. 2 Crystalline phase of TiO₂ films with different nanostructure. (a) XRD patterns; (b) Raman spectra.

Cell adhesion and proliferation on TiO₂ films

Cell adhesion and proliferation assay was shown in Fig. 3. The results revealed that after 6 h, 12 h, 24 h and 72 h culture, the cell adhesion on all the three films was similar at the first 6 h. However, after 12 h, cells on the TiO₂ porous film showed the highest adhesion and proliferation rate of all the three films, while those on TiO₂ dense film showed the lowest. The cell morphology shown in Fig. 4, which was observed after 24 h culture showed similar results with CCK-8 characterization. For nanodots and porous films, irregular spread cells with longer filopodias were observed (Fig. 4a); while for dense film, the cells were rather shrank with less filopodias. Han's and Samarasekera's work^19,20 reported that nanostructures with a range of tens of nanometer are beneficial for cell attachment and proliferation, our work is well coincident with that.

Fig. 3 Cell adhesion and proliferation assay on TiO₂ films by CCK-8. The data are expressed as the mean ± SD. Significant differences were considered at P < 0.05. OD value means the optical density of absorbance at 450 nm.

Fig. 4 Effects of film nanostructure on the morphology of NIH3T3 cells. (a) Nanodots film; (b) dense film; (c) porous film.

Single cell detachment on TiO₂ films

After 24 h of culture, the samples were illuminated with UV365 light from the back. As shown in Fig. 5a, initially, cells attached on the films at different density, the porous film showed the largest, while the dense film showed the lowest. With increasing time of UV365 illumination, fewer and fewer cells remained on the surface. Obviously, nanostructure of the films greatly affected the number of residual cells. The dense film showed more residual cells than the nanodots film; while the porous film showed the smallest amount of residual cells. That means, the porous films showed the best light induced cell detachment performance among the three films with different nanostructures, and dense film was the worst among the three.

Fig. 5 Light induced cell detachment behaviors of single cells. (a) DAPI staining results of residual cells on TiO₂ nanostructured films. Scale bars represent 20 μm; (b) kinetic analysis of cell detachment against UV365 illumination time.

The cell counting kit-8 (CCK-8) assay of residual cell gave more comprehensive results on residual cells, as shown in Fig. 5b. It was observed that the porous film and nanodots film showed almost the same amount of residual cells after 20 min UV365 illumination, while the amount of the residual cells on the dense film was much higher after the same time of UV365 illumination. In order to further analyze the effects of different nanostructure on the light induced cell detachment, the first-order kinetic model proposed by Bourne and Jennings was utilized to evaluate the detaching rate.^21,22

According to Bourne and Jennings, the amount of residual cells on the substrate could be evaluated through the following equations:

y = y₀ exp^(−kx) (1)

It is considered that the cell detachment could be also described through eqn (1). Therefore, y means the residual cell amounts on substrates at any given time and y₀ means the amount of cells on substrates at zero time; k is the detaching rate constant of cells, x is the UV365 illumination time. These kinetic parameters in eqn (1) were estimated statistically through a data-fitting procedure based on a nonlinear least-squares regression method.

Fig. 5b showed the experimental detaching data of cells and the fitted profiles. In the three cases, the estimated curves were in good agreement with the experimental data (R = 0.997, R = 0.959, R = 0.996 respectively). It was observed that the detaching rate constants k of different films were as following: −0.123 for the nanodots film, −0.061 for the dense film and −0.160 for the porous film. Clearly, porous film showed the fastest detaching rate, followed by nanodots film, and the dense film showed the slowest rate.

Cell sheet detachment assay and live-dead staining

After cultured for 4 d, cell sheets were formed on the substrate. With the UV365 illumination, cell sheets consisted of cells and ECM could detach from the films spontaneously. Again, different nanostructures significantly affected the detachment of the cell sheets. It took 3 min for the porous film, 5 min for the nanodots film and 10 min for the dense film. As shown in the live-dead staining results in Fig. 6, all the cell sheets obtained showed green and there was very few cells showing red compare to the negative control, which was treated by ethanol. This result declared that the cell sheets harvested through UV365 illumination remained alive and maintained high activity. That is in well agreement with our previous work on osteoblast cells.^10,23 It also means that such low-density, short-time UV365 illumination is safe for cells. It is expected that such cell sheet could be further utilized in cell sheet tissue engineering.

Fig. 6 Viability of detached cell sheets from TiO₂ films with different nanostructure.

Effects of nanostructure on protein adsorption

When cells attach to materials, proteins are considered as intermediates. Hence, the adsorption statuses of proteins are considered to be the keys to study such light induced cell detachment. In this work, bovine serum albumin (BSA) was chosen as a model to investigate the effects of film nanostructure on proteins. As shown in Fig. 7, after BSA adsorption, some sunken areas in nanodots film and pores in the porous film were filled up with BSA. That may make the surfaces of all the films became rather smooth. Similar protein adsorption could be quite prevalent in in vitro and in vivo situations. Also, it is reported that the dimension of BSA molecule is around 4 nm × 4 nm × 14 nm when BSA is in its natural conformation,²⁴ that is quite similar to the dimension of the pores in the porous films and the sunken areas of the nanodots film. That means, besides topographical variations, the conformation variation of protein molecules after adsorption could also play important roles in their different cellular responses.

Fig. 7 SEM micrographs of BSA adsorbed TiO₂ films with different nanostructure. (a) Nanodots film; (b) dense film; (c) porous film.

In order to further investigate the differences of protein molecules adsorbed on TiO₂ films, circular dichroism (CD) spectra assay was used to study the secondary structure of protein molecules adsorbed on the films before and after UV365 illumination. In Fig. 8, the curves of BSA adsorbed were shown. Based on these curves, the negative minima in ultraviolet region at 208 nm is the characteristic of an α-helix structure of protein.^25,26 The CD data were deconvoluted on the Dichroweb server, SELCON3 program module was chosen for computational analysis and the Set 4 database was chosen as reference.^27,28 For the curves, the most obvious difference was that after 5 minutes UV365 illumination, the signal intensity of CD spectra weakened, which means the amounts of adsorbed BSA decreased. For the secondary structure characteristics tabulated in Table 1, it was clear that BSA adsorbed on nanodots film and porous film showed similar percentage on secondary structure components, i.e. α-helix, β-sheet, β-turn and random-coil. That means they may have similar conformation; there did show different percentage for these components on dense film. After UV365 illumination, BSA adsorbed on nanodots film and porous film showed changes in secondary structure. However, the magnitudes were quite different, porous film and nanodots film showed much larger structural changes after UV365 illumination, and that adsorbed on dense film showed less changes. Moreover, after UV365 illumination, the percentages of secondary structural components of BSA on different films became rather similar.

Fig. 8 CD spectra of BSA adsorbed on TiO₂ films before (left) and after (right) UV365 illumination.

Table 1 Secondary structure component of BSA adsorbed on TiO₂ films (%)

Sample	UV365	α-Helix	β-Sheet	β-Turn	Random coil
Nanodots film	Before	31.6	19.0	20.2	29.1
	After	26.1	22.1	19.2	32.6
	Δ	5.5	3.1	1.0	3.5
Dense film	Before	25.5	22.8	22.3	29.5
	After	22.5	22.3	21.0	34.2
	Δ	3.0	0.5	1.3	4.7
Porous film	Before	32.1	19.3	19.8	28.8
	After	24.4	22.4	22.6	30.6
	Δ	7.7	3.1	2.8	1.8

---

Obviously, the conformation variations of BSA adsorbed on different films showed significant differences. Such results imply that quasi-three-dimensional nanostructures could be beneficial for cellular responses. The distribution of protein-binding sites in such structure may promote protein molecules to be adsorbed in more natural conformation, and that will facilitate the attachment of cells. Moreover, light induced hydroxyl groups also take effects in quasi-three-dimensional mode. These differences resulted from nanostructure will induce more significant conformation changes of adsorbed protein molecules and eventually more significant cellular responses.

In fact, many works have reported that nanostructures are crucial in tailoring protein adsorption, cell attachment and proliferation on the surface of biomaterials.^29,30 While this work further indicates nanostructure also makes great contribution in tailoring cell detachment or harvest.

Since light induced cell detachment exhibits advantages on no materials residual, as well as little impact on microenvironment, the faster detachment could further minimize the possible detrimental effects caused by the detachment process. That is very helpful in subsequent utilization of cell sheets in tissue engineering, which will be carried out in our future works. Also, it is believed that such structural optimization strategy is also helpful in other cell detachment platforms.

Conclusion

The present study is aimed to investigate the influence of the nanostructure of anatase TiO₂ films on light-induced cell detachment. Anatase TiO₂ films with different nanostructure were prepared through sol–gel based methods. It is concluded that:

(1) It is found all the TiO₂ films showed good biocompatibility, cells could well attach and proliferate on the surface the films. The porous film and nanodots film showed better effects on cell proliferation.

(2) The nanostructure of the films significantly affects the light induced cell detachment property. Among all the three nanostructure tested, the porous film showed the best performance. That is mainly ascribed to the effect of porous structure on protein adsorption.

(3) The reason for the effects of nanostructure on light induced cell detachment is ascribed to the different initial adsorption conformation of protein molecules and the light responses of TiO₂. Proper nanostructure facilitates adsorption of protein molecules in rather natural conformation, and then more obvious changes in response to UV365 illumination on the films.

Experimental

Preparation of TiO₂ nanodot films

TiO₂ films with different nanostructures were prepared via sol–gel and phase-separation-induced self-assembly based method.³¹ For nanodots film, tetrabutyltitanate (TBOT, Aladdin Chemical Reagent, CP, >98%), acetylacetone (AcAc, Sinopharm Chemical Reagent, AR, >99%), deionized water with the mole ratio of 1 : 0.3 : 1 were dissolved in ethanol (Sinopharm Chemical Reagent, AR, >99.7%), then polyvinylpyrrolidone (PVP, K30, Sinopharm Chemical Reagent, AR, >99%) was added to obtain a homogeneous precursor sol (sol 1); for dense film, a similar precursor sol without PVP addition was prepared (sol 2) and then used for film preparation; for porous film, monoethanolamine (MEA Sinopharm Chemical Reagent, AR, >99%) was added to sol 1 with a MEA/TBOT ratio of 0.4 to be sol 3. Sol 1, sol 2 and sol 3 were spun on quartz glass (10 mm × 10 mm) substrates at 8000 rpm for 40 s with a spin-coater (Chemat Technology, KW-4A), respectively. After that, the films were heat treated at 500 °C for 2 hours.

Characterization of the TiO₂ nanodots films

Surface morphology of the films was characterized through a field-emission scanning electron microscopy (FESEM, Hitachi, SU47). X-ray diffraction (XRD) patterns of the samples were collected by a Philips X-ray diffractometer (PANalytical, X'Pert PRO) to examine the crystalline phase of the nanostructure TiO₂ films. Raman spectra were taken on the samples using OMNIC Dispersive Raman (Thermo Fisher Scientific, DXR532) with a DXR laser operating at 532 nm with incident power of 10 mW.

Cell culture

Mouse NIH3T3 cells (208F, ATCC) were used as model cells. Subconfluent NIH3T3 on polystyrene (PS) dishes were trypsinized with 0.25% trypsin/1 mM EDTA (Gibco) and were subcultured on TiO₂ films with Dulbecco's modified eagle medium (DMEM, Gibco) supplemented with 10% fetal bovine serum (FBS, PAA, Australia), 1% sodium pyruvate (Gibco), 1% antibiotic solution containing 10 000 units per ml penicillin, 10 000 mg ml⁻¹ streptomycin (Gibco), and 1% DMEM non-essential amino acids (Gibco) under a humidified atmosphere of 5.0% CO₂ at 37 °C.

Cell adhesion and proliferation

Cell counting kit-8 (CCK-8) assay was used to measure cell adhesion and proliferation.³² In brief, cells cultured at a density of 5 × 10⁴ cells per cm² were seeded on nanodots films, dense films and porous films in 24 wells. CCK-8 solution was added in the wells as the concentration ratio of 1 : 10 with culture solution. After reacted for 3 h in the incubator, the optical density (OD) value of supernatant liquid was measured at 450 nm with a microplate reader (Multiskan MK3) after cultured for different time (6 h, 12 h, 24 h, 72 h).

Cell morphology

Morphologies of NIH3T3 cells attaching to the coatings were observed by SEM. After cultured 24 h, the specimens were first fixed with 2.5% glutaraldehyde in PBS (0.1 M, pH 7.0) for more than 4 h, and washed three times in PBS for 15 min, respectively. Then they were post-fixed with 1% OsO₄ in PBS for 1 h and washed three times in PBS for 15 min, followed by dehydration by a graded series of ethanol (50, 70, 80, 90 and 95%) for 15 min and by 100% ethanol twice for 20 min. Afterwards, the specimens were transferred to the mixture of ethanol and iso-amyl acetate (v/v = 1 : 1) for 30 min and pure iso-amyl acetate for 1 h, followed by final dehydration in Hitachi Model Hcp-2 critical point dryer with liquid CO₂. In order to evaluate the morphology of cells, more than 20 individual cells were observed on each film.

Single cell detachment assay

Single cell detachment was carried out as described previously.¹⁰ Briefly, cells at a density of 5 × 10⁴ cells per cm² were cultured on TiO₂ nanostructure films with different surface topographies for 24 h. A UV light with a wavelength of 365 nm and a power of 2 mW cm⁻² were used. After illumination for different time, cells were rinsed with PBS gently. The CCK-8 assay was applied to evaluate residual cells on the surface of the films so that the detachment ratio could be calculated. In order to display the results more intuitively, after the intended time (5, 10, 15, 20 min) of UV365 illumination, the residual cells on the substrate were firstly rinsed with PBS for three times after fixing with glutaraldehyde for 2 h at room temperature, and then stained by 1 μg ml⁻¹ 4,6-diamidino-2-phenylindole (DAPI, Sigma) for cell nuclear which was shown as an indication of total cell number. The automated counts was used to evaluate residual cells on the surface, namely at least 24 random images per sample from three independent cultures were acquired by a laser scanning confocal microscope (Olympus, Fluoview FV1000). As a control, cells without UV illumination were also test, so that the detachment ratio could be calculated.

Cell sheet detachment assay and live-dead staining

NIH3T3 cells were seeded on different TiO₂ nanostructure films at a density of 5 × 10⁴ cells per cm² and confluent cell sheets could be formed after 4 d culture. After UV365 irradiation for intend time, an intact cell sheets could detach from the surface of the films spontaneously.

The viability of cell sheets was evaluated using a live-dead staining method,³³ based on a simultaneous determination of live and dead cells with calcein-AM and PI (DOJINDO, Japan). It distinguished live and dead cells in the way that live cells would be colored by calcein-AM with green while dead cells would be labeled by PI with red. The detached cell sheets were first rinsed three times with PBS and then incubated with appropriate calcein-AM for 5 min in an incubator. When time was up, cell sheets were rinsed three times with PBS. Then PI was added and incubated for 20 min. After rinsing another three times with PBS, cell sheets were observed using a confocal laser scanning microscope (FLUOVIEW FV1000, Olympus). As a negative control, a piece of cell sheet was treated in ethanol for 30 min, and the same procedure was followed as mentioned above.

Protein adsorption and characterization

Morphologies of BSA attaching to the coatings were observed by SEM. BSA was dissolved in PBS with a concentration of 15 mg ml⁻¹. After incubation at room temperature in a sealed environment for 12 h, the samples were washed thoroughly with deionized water twice then dried in room temperature. CD spectra were measured using a MOS-450 spectropolarimeter for wavelengths from 190 to 250 nm. After three samples of anatase TiO₂ film-coated quartz substrate spectra were recorded these samples were immersed in the bovine serum albumin solution and incubated at 37 °C for 24 h. The spectra were summation of the spectra for the substrate and protein, so the BSA spectra were obtained by subtracting the substrate spectra. The scanning rate was 50 nm min⁻¹, and the spectral width was 0.5 nm. For each sample, five spectra were accumulated and averaged to get the final CD spectrum. The CD data were deconvoluted on the Dichroweb server. The SELCON3 program module was chosen for computational analysis and the Set 4 database as reference.^27,28

Statistical analysis

All values were expressed as means ± standard deviation. Statistical analysis was carried out by a one-way analysis of variance (one-way ANOVA) and Scheffe's post hoc test with the SPSS software for multiple comparison tests. Differences were considered statistically significant when P < 0.05.

Acknowledgements

This work is financially supported by National Basic Research Program of China (973 Program, 2012CB933600), National Natural Science Foundation of China (31579062, 51372217, 51472216, 51272228), The Key Science Technology Innovation Team of Zhejiang Province (2013TD02), Zhejiang Provincial Natural Science Foundation (LY15E020004), Fundamental Research Funds for the Central Universities (2014XZZX005).

References

1. N. Matsuda, T. Shimizu, M. Yamato and T. Okano, Adv. Mater., 2007, 19, 3089.
2. J. B. Lin, B. C. Isenberg, Y. K. Shen, K. Schorsch, O. V. Sazonova and J. Y. Wong, Colloids Surf., B, 2012, 99, 108.
3. R. Inaba, A. Khademhosseini, H. Suzuki and J. Fukuda, Biomaterials, 2009, 30, 3573.
4. A. Ito, K. Ino, T. Kobayashi and H. Honda, Biomaterials, 2005, 26, 6185.
5. Y.-H. Chen, Y.-C. Chung, I. Wang and T.-H. Young, Biomaterials, 2012, 33, 1336.
6. S. J. Chen, H. Y. Yu and B. C. Yang, J. Biomed. Mater. Res., Part A, 2013, 101, 64.
7. A. Kar, K. S. Raja and M. Misra, Surf. Coat. Technol., 2006, 201, 3723.
8. T. Kasuga, H. Kondo and M. Nogami, J. Cryst. Growth, 2002, 235, 235.
9. Y. T. Sul, C. Johansson, E. Byon and T. Albrektsson, Biomaterials, 2005, 26, 6720.
10. Y. Hong, M. F. Yu, W. J. Weng, K. Cheng, H. M. Wang and J. Lin, Biomaterials, 2013, 34, 11.
11. Y. Y. Wang, H. Deng, C. X. Huangfu, Z. W. Lu, X. X. Wang, X. Y. Zeng, H. B. He and H. Rao, Surf. Interface Anal., 2015, 47, 245.
12. H. H. Huang, C. P. Wu, Y. S. Sun, W. E. Yang, M. C. Lin and T. H. Lee, Surf. Coat. Technol., 2014, 259, 206.
13. W. H. Yang, X. F. Xi, X. K. Shen, P. Liu, Y. Hu and K. Y. Cai, J. Biomed. Mater. Res., Part A, 2014, 102, 3598.
14. E. Schnell, K. Klinkhammer, S. Balzer, G. Brook, D. Klee, P. Dalton and J. Mey, Biomaterials, 2007, 28, 3012.
15. J. He, W. Zhou, X. J. Zhou, X. X. Zhong, X. L. Zhang, P. B. Wan, B. S. Zhu and W. T. Chen, J. Mater. Sci.: Mater. Med., 2008, 19, 3465.
16. Y. Ito, Biomaterials, 1999, 20, 2333.
17. J. Yan, G. Wu, N. Guan, L. Li, Z. Li and X. Cao, Phys. Chem. Chem. Phys., 2013, 15, 10978.
18. J. Zhang, M. Li, Z. Feng, J. Chen and C. Li, J. Phys. Chem. B, 2006, 110, 927.
19. C. Samarasekera, B. Tan and K. Venkatakrishnan, J. Biomed. Nanotechnol., 2015, 11, 623.
20. Z. J. Han, A. E. Rider, M. Ishaq, S. Kumar, A. Kondyurin, M. M. M. Bilek, I. Levchenko and K. Ostrikov, RSC Adv., 2013, 3, 11058.
21. M. C. Bourne and W. G. Jennings, J. Am. Oil Chem. Soc., 1963, 40, 517.
22. M. C. Bourne and W. G. Jennings, J. Am. Oil Chem. Soc., 1965, 42, 546.
23. K. Cheng, X. Wang, W. Weng, J. Lin and H. Wang, Thin Solid Films, 2015, 584, 18.
24. T. Makinodan, N. Gengozian and R. E. Canning, J. Immunol., 1960, 85, 439.
25. N. J. Greenfield, Trends Anal. Chem., 1999, 18, 236.
26. M. Zhang, X. H. Li, Y. D. Gong, N. M. Zhao and X. F. Zhang, Biomaterials, 2002, 23, 2641.
27. L. Whitmore and B. A. Wallace, Nucleic Acids Res., 2004, 32, W668.
28. N. Sreerama and R. W. Woody, Anal. Biochem., 2000, 287, 252.
29. L. W. Lv, Y. S. Liu, P. Zhang, X. Zhang, J. Z. Liu, T. Chen, P. L. Su, H. Y. Li and Y. S. Zhou, Biomaterials, 2015, 39, 193.
30. J. Park, S. Bauer, K. von der Mark and P. Schmuki, Nano Lett., 2007, 7, 1686.
31. M. Luo, K. Cheng, W. J. Weng, C. L. Song, P. Y. Du, G. Shen, G. Xu and G. R. Han, Nanotechnology, 2009, 20, 215605.
32. V. Guarani, G. Deflorian, C. A. Franco, M. Kruger, L. K. Phng, K. Bentley, L. Toussaint, F. Dequiedt, R. Mostoslavsky, M. H. H. Schmidt, B. Zimmermann, R. P. Brandes, M. Mione, C. H. Westphal, T. Braun, A. M. Zeiher, H. Gerhardt, S. Dimmeler and M. Potente, Nature, 2011, 473, 234.
33. T. Sasagawa, T. Shimizu, S. Sekiya, Y. Haraguchi, M. Yamato and Y. Sawa, Biomaterials, 2010, 31, 1646.

---

Footnote

† Contributed equally.

---