Endothelium Cell Responses on Pulsed-Anodized NiTi Alloy with HNO₃, NH₄NO₃, H₂SO₄, and (NH₄)₂SO₄ as Electrolytes

Abstract

Pulsed anodization allows to form a Ni-free TiO₂ layer on an almost equiatomic NiTi alloy. The surface is then hydrophilized and impedes the release of Ni ions from the alloy. This study aimed to assess the suitability of electrolytes used in the pulsed anodization process with the view of application as a biomaterial. Toward this end, the NiTi alloy was pulsed-anodized in four different aqueous electrolytes, HNO₃, NH₄NO₃, H₂SO₄, and (NH₄)₂SO₄, after which the endothelium cell behavior was compared. The use of H₂SO₄ as the electrolyte resulted in the formation of a TiO₂ layer with a groove-like structure of several tens of nanometers wide, and this surface inhibited the activity of endothelium cells. The ability to prevent the release of Ni ions was diminished when using the HNO₃ electrolyte, resulted in inferior cell proliferation. The other two electrolytes, NH₄NO₃ and (NH₄)₂SO₄, formed a Ni-free TiO₂ layer with a comparatively smooth surface, which more effectively suppressed the release of Ni ions; thus, the cell attachment as well as proliferation were superior to that of the other two electrolytes. The surface smoothness and Ni suppression were thus concluded to be the factors governing the selection of an electrolyte.

Fig. 6 (a) The average area and (b) circularity of EA.hy926 cells cultivated for 4 h on each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, H₂SO₄, and Ti surfaces. The data were statistically analyzed using ANOVA with a SNK post hoc test (^**p < 0.01, ^*p < 0.05).

1. Introduction

A major disadvantage of self-expanding cardiovascular stents produced from almost equiatomic nickel–titanium (NiTi) with superelastic properties is the occurrence of stent thrombosis.^1–3) This is caused by the delayed endothelization of the NiTi alloy.⁴⁾ Ni, as a constituent element of the alloy, suppresses the activity of endothelial cells (ECs) attached to the surface, thereby preventing endothelialization.⁵⁾ Additionally, an alloy surface with insufficient hydrophilicity often induces the detachment of ECs, leading to the delay of endothelialization. Thus, forming a surface layer with the required corrosion resistance and hydrophilicity is a possible strategy to overcome this drawback. Ideally, the surface layer should be a biocompatible layer with flexibility that allows the self-expanding stent to be tracked. Accounting for these requirements, forming a Ni-free TiO₂ layer with a thickness in the nanometer range is considered a promising way to lower the risk of stent thrombosis.⁶⁾

Our research group proposed the use of pulsed anodization as a process to form a Ni-free TiO₂ layer.^7–9) Pulsed anodization is a simple process that entails connecting the alloy to the anode in an electrolyte and then applying a rectangular-type pulse voltage. Typically, the applied voltage ranges from 0 V to several volts. After applying the pulsed voltage for a few hours, an almost Ni-free TiO₂ layer with a thickness of ca. 50 nm is available. Because thermal treatment is not required in this process, the metallographical change, which may cause the mechanical properties of the alloy to deteriorate, is negligible. The growth principles of the layer are discussed elsewhere.⁷⁾ Briefly, the oxidation of Ti as an anodic reaction proceeds during the period in which the voltage is applied, and the chemical reaction involving the removal of Ni from the surface occurs in the period during which the voltage is turned off. The combination of these two reactions enables an almost Ni-free TiO₂ layer to be formed on the surface of the NiTi alloy. Here, it should be noted that the chemical reaction does not involve only the reaction whereby Ni is removed; other reactions such as the dissolution of the TiO₂ layer itself may also occur simultaneously, and this would also influence the properties of the layer. For instance, Yamasaki et al. demonstrated that the use of a HNO₃ solution as an electrolyte has an oxidizing effect that leads to the formation of a surface layer with a comparatively smooth surface. In contrast, a H₂SO₄ solution resulted in a comparatively rougher surface with a nanometer-scale bumpy structure. Additionally, the corrosion protective property of the H₂SO₄ layer was superior to the anodized layer that was formed using the HNO₃ electrolyte: in fact, the release of Ni ions from the surface in the presence of H₂SO₄ was approximately 1/3 of that of the HNO₃ surface.⁸⁾ These differences in the properties of the surface layer are likely to influence the activity of the ECs thereon; thus, selecting an adequate electrolyte for the pulsed anodization process is an important task. Meanwhile, detailed influence to EC’s activity by these electrolytes have not been assessed experimentally to date.

In this study, a NiTi alloy was pulsed-anodized in four different types of aqueous electrolytes: HNO₃, NH₄NO₃, H₂SO₄, and (NH₄)₂SO₄. After analyzing the surface characteristics in terms of the hydrophilicity and Ni-ion release behavior, ECs were cultured on the surfaces. Then, the cell attachment and proliferation behavior were evaluated carefully. Together with these results, we discuss the remarkable factors that affect the selection of an electrolyte with the aim of improving the surface for ultimate application as a biomaterial.

2. Experimental Procedure

2.1 Process and materials

A NiTi rod (55.05 at% Ni, TOKIN Co.) with a diameter of 15 mm was cut into disks of 2 mm in thickness. These disks were polished to a mirror-like surface using a suspension of colloidal silica with a particle size of ∼40 nm. The polished NiTi disks were ultrasonicated twice with ethanol and distilled water for 5 min, respectively. The reverse side of each disk was completely sealed using Teflon^® tape to limit the surface area for anodization. The electrolytes were aqueous solutions of HNO₃, NH₄NO₃, H₂SO₄, and (NH₄)₂SO₄ adjusted to the concentration of 100 mM. During the anodization, the electrolyte temperature was maintained at 293 K using a water bath. In each anodization experiment, the NiTi disk was connected to the anode, with a Pt plate (20 × 10 × 0.3 mm) forming the cathode at a distance of ∼50 mm from the anode. A rectangular pulse voltage with a pulse width of 100 ms and duty cycle of 50% was employed for the pulse anodization. Here, 3.5 V and 0 V were used as the upper and lower voltages, respectively. The pulsed voltage was applied for 120 min. After completion of the experiment, the alloy was retrieved from the electrolyte, followed by ultrasonication with distilled water for 5 min and storage in ambient atmosphere. Apart from the above, a NiTi disk and a Ti disk (99.5% purity), polished similarly but not anodized, were prepared for comparison.

2.2 Surface characterization

The chemical composition and the chemical state of the anodized NiTi surface were confirmed by X-ray photoelectron spectroscopy (XPS; Phi5000 Versaprobe, Ulvac-Phi). The microstructure of the anodized surface was observed using field emission scanning electron microscopy (FE-SEM; JSM-6701F, JEOL Ltd.) at an acceleration voltage of 5 kV. The hydrophilicity of the anodized surface was evaluated by measuring the water contact angle (DM-CE1, Kyowa Interface Science). The amount of Ni ions that were released from the alloy surface was evaluated in a phosphate-buffered saline (PBS) solution. Here, the NiTi alloy surface in contact with the solution was limited to 0.785 cm² by sealing with Teflon^® tape. The NiTi alloy disks with and without anodization were immersed in 4 mL of PBS solution in a polypropylene tube up to 168 h. The tube was kept at 310 K with simultaneous agitation at 115 rpm. After the prescribed period, the amount of Ni ions in the PBS was determined using graphite furnace atomic absorption spectrometry (GF-AAS; Z-2010, Hitachi).

2.3 Cell tests

EA.hy926 cells, an endothelial cell line derived from human umbilical veins, was used for the cell tests. Prior to the experiments, the cells were cultured in D-MEM medium with 10% fetal bovine serum (FBS; JR Scientific) and 1% antibiotic-antimitotic (100 U·m⁻¹ penicillin, 100 µg·ml⁻¹ streptomycin, and 0.25 µg·ml⁻¹ amphotericin B; GIBCO BRL) at 310 K in 5% CO₂. A NiTi plate was sterilized with 70% ethanol for 30 min, after which it was placed in 24-well polystyrene cell culture plates.

The cell attachment behavior was evaluated by counting the number of cells on the alloy surface after incubating for 4 h following seeding. Here, the amount of cell suspension used for seeding was 500 µL, and the concentration of cells was 1.0 × 10⁵ cells·mL⁻¹. After completing the cultivation period, the attached cells were stained with Trypan blue, and trypsinized, after which they were counted using a hemocytometer. Further, part of the attached cells was fixed in 4% formaldehyde and permeabilized with TritonX-100 for 60 min. The actin filament and nuclei were then stained with phalloidin (Acti-stain™ 488, Cytoekeleton Inc.) and DAPI (Bacstain DAPI solution, DOJINDO), respectively, after which they were observed by a fluorescent microscope (BZ-X710, Keyence) to confirm the cytoskeletal organization.

Apart from the cell attachment test, the cell proliferation behavior was also evaluated using a similar testing process. Here, the cells for seeding were reduced to 1.0 × 10⁴ cells·mL⁻¹ and the incubation period was prolonged to 96 h and 168 h. After completion, the number of cells was counted after they were stained by Trypan blue.

2.4 Statistical analysis

All the experimental results of the in-vitro test were statistically analyzed using analysis of variance (ANOVA) with a Student-Newman-Keuls (SNK) post-hoc test.

3. Results and Discussions

3.1 Characteristics of the pulsed-anodized NiTi surfaces

Elemental analysis using XPS revealed that the topmost surface of the pulsed-anodized NiTi alloy comprised Ti and O as the major elements and Ni as the minor element of less than a few atomic percent, irrespective of the electrolyte used (Table 1). Further, the shape of the corresponding Ti 2p spectrum corresponded with that of TiO₂. The results of the XPS analysis did not show any remarkable dependence on the electrolyte. Thus, it was confirmed that the surface layer produced by the pulsed anodization was an almost Ni-free TiO₂ layer regardless of the electrolyte that was used (HNO₃, NH₄NO₃, H₂SO₄, or (NH₄)₂SO₄). Note that these XPS results were in agreement with the results of our previous study in which we compared the surface characteristics of the pulsed-anodized NiTi alloy using various electrolytes.⁸⁾

Table 1 XPS measurement of elemental concentrations (at%) on the upper surface of each sample.

The morphology of the anodized NiTi surface, observed by FE-SEM, is shown in Fig. 1. Overall, the images of the surfaces anodized in the HNO₃, (NH₄)₂SO₄, and NH₄NO₃ electrolytes showed that the surfaces have a smooth appearance, yet nanometer-sized pores, indicated by black arrow, are dispersed therein. On the other hand, the H₂SO₄ surface had a groove-like structure with grooves of several tens of nanometers, indicated by white arrow, unlike the other three surfaces. With regard to the hydrophilicity (Fig. 2), the water contact angles of all the anodized NiTi surfaces were smaller as compared with the untreated surface. Moreover, the contact angle on the H₂SO₄ surface was slightly smaller than that on the other three surfaces. This is caused by the surface roughness attributable to the groove-like structure, thus unrelated to the difference in the surface free energy.

Fig. 1

FE-SEM images obtained from (a) HNO₃, (b) (NH₄)₂SO₄, (c) NH₄NO₃, and (d) H₂SO₄ surfaces.

Fig. 2

Water contact angle for distilled water on each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, H₂SO₄, and Ti surfaces.

The corrosion resistance of the anodized NiTi alloys was evaluated from the view of the release of Ni ions because these ions cause an allergic reaction and cytotoxicity. Amount of Ni ion release from each sample is shown in Fig. 3. With regard to the anodized surfaces, the extent of Ni-ion release into the PBS solution was significantly lower than that from the untreated surface. Specifically, the amount released from the surface anodized in the HNO₃ electrolyte was ∼1/3 of that of the untreated surface. The amounts released from the surfaces anodized in (NH₄)₂SO₄, NH₄NO₃, and H₂SO₄ electrolytes were similar, that is ∼1/9 of that of the untreated surface. Clearly, pulsed-anodization is an effective process to impede the release of Ni ions from the surface of the NiTi alloy. In addition, the corrosion resistance of the (NH₄)₂SO₄, NH₄NO₃, and H₂SO₄ surfaces is superior to that of the HNO₃ surface. It is worthy commenting on the extent to which the present results compare with those of our previous test.⁸⁾ It should be noted that the Ni ion release test in the previous study was conducted using a shorter period of 24 h. The tendency for the Ni ions to be released seemed to be similar notwithstanding the length of the testing periods, but the difference among the samples in terms of the amount of Ni ions that are released is enhanced notably. These results indicate that the corrosion resistance as a result of pulsed anodization is enhanced in the case of a longer immersion period.

Fig. 3

Amount of Ni ion released into PBS solution from each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, and H₂SO₄ surfaces.

3.2 Attachment of ECs to the pulsed-anodized surfaces

The number of ECs attached to the anodized and untreated surfaces is shown in Fig. 4. The EC concentration on the HNO₃ surface was 3.0 × 10⁴ cells·cm⁻², an increase of ∼50% as compared with the untreated surface, of which the number of ECs almost agreed with that of the Ti surface used as a positive control. On the other hand, the number of ECs on the surfaces anodized in (NH₄)₂SO₄, NH₄NO₃, and H₂SO₄ electrolytes was almost similar to that on the untreated surface, with the number of cells attached to the H₂SO₄ surface smaller than that on the HNO₃ and Ti surfaces. The statistical analysis confirmed this difference to be significant.

Fig. 4

Number of attached EA.hy926 cells cultivated for 4 h on each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, H₂SO₄, and Ti surfaces. Ti was used as a positive control. The data were statistically analyzed using ANOVA with a SNK post hoc test (^*p < 0.05).

The ECs attached to the surfaces anodized in HNO₃, NH₄NO₃, and (NH₄)₂SO₄ electrolytes had a spread cell morphology that extended polygonally, whereas those on the surface anodized in H₂SO₄ electrolyte were less elongated (Fig. 5). Focusing on the cytoskeletal organization in Fig. 5, an actin stress fiber was observed to form in the ECs on the HNO₃, NH₄NO₃, and (NH₄)₂SO₄ surfaces and on the untreated surfaces, which corresponded to that on the surface of Ti (Indicated by red arrow). Yet, this stress fiber was hardly observed to form on the surface anodized in H₂SO₄ electrolyte. The actin stress fiber is well known to be a fibrous protein consisting of the actin monomer bundled by a cross-linking protein known as α-actin.¹⁰⁾ Thus, the absence of the stress fiber on the surface anodized in H₂SO₄ electrolyte suggested that the process of actin bundling was less well developed.

Fig. 5

Fluorescence image of EA.hy926 cells cultivated for 4 h: (a) untreated, (b) HNO₃, (c) (NH₄)₂SO₄, (d) NH₄NO₃, (e) H₂SO₄, and (f) Ti surfaces. Green: Actin Blue: Nucleus.

The observed cell morphology was clarified by calculating the average area and circularity of the ECs attached on each surface using the images of the morphology and ImageJ software (Fig. 6). For each sample, 30 cells in the center of the surface were measured. Here, it should be noted that each cell area and its circularity relate to the cell activity; for instance, a cell with a larger area and lower circularity value is considered to have higher activity.¹¹⁾ Figure 6(a) shows that the cell area of the sample surfaces other than the H₂SO₄ surface approximated 1500 µm². The fluorescence images also showed that the cell area on the H₂SO₄ surface was small and differed significantly from that of the other surfaces. The circularity results (Fig. 6(b)) also showed that the cells on the surface anodized in H₂SO₄ electrolyte were particularly circular. To summarize, the cells on the anodized surface selecting H₂SO₄ electrolyte were small and round, whereas those on the other samples were large and extended polygonally.

Fig. 6

(a) The average area and (b) circularity of EA.hy926 cells cultivated for 4 h on each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, H₂SO₄, and Ti surfaces. The data were statistically analyzed using ANOVA with a SNK post hoc test (^**p < 0.01, ^*p < 0.05).

A hydrophilic surface with adequate roughness is reported to be beneficial for cell attachment as well as activation; however, only a slight positive effect was found in this study.^12–14) Wei et al. investigated an osteoblast-like cell attachment to a Ti surface with the water contact angle in the range 0–108°, and demonstrated that an advantageous effect was only found on the 0° surface.¹⁵⁾ The water contact angles on the pulsed anodized NiTi surface in this study were in the range of 30–108°, of which the values are much higher than 0°, and hence, the advantageous effect resulting from hydrophilization was not highlighted. Further, the surface anodized in H₂SO₄ electrolyte with the roughness of several tens of nanometers served to inactivate the attached ECs. Reportedly, nanoscale roughness increases the cell activity, whereas at the submicron scale the opposite occurs.^16–19) This is because penetration of the filopodia is trapped by such submicron roughness, thereby preventing cell expansion from taking place. The groove-like structure observed on the surface anodized in H₂SO₄ electrolyte had submicron scale roughness, being likely to have a similar effect, thereby decreasing the activity of the ECs.

3.3 Cell proliferation of ECs on the anodized NiTi surfaces

The cell proliferation rate on the surfaces was evaluated by calculating the cell proliferation rate during the cultivation period at 96 h and 168 h (Fig. 7). The number of the attached cells on each surface after cultivation at 96 h and 168 h were normalized using the number of cells on the corresponding surface at 4 h (Fig. 4). During the shorter cultivation period of 96 h, an advantageous effect on EC proliferation was found on the surface anodized in NH₄NO₃, (NH₄)₂SO₄, and H₂SO₄ electrolytes, while the difference in proliferation rate was not found between the untreated surface and the surface anodized in HNO₃ electrolyte. In contrast, for the cultivation period of 168 h, the difference between the anodized and untreated surfaces became obvious; in fact, the rate for the untreated surface was only 4.8%, whereas that for the anodized surfaces exceeded 16.8%, irrespective of the electrolyte that was used. Remarkably, the proliferation rate on the surfaces anodized in NH₄NO₃, (NH₄)₂SO₄, and H₂SO₄ electrolytes reached ∼30%, a rate nearly twice as high as that of the untreated Ti surface.

Fig. 7

The cell proliferation ratio during the cultivation period at 96 h and 168 h on each sample: the untreated, HNO₃, (NH₄)₂SO₄, NH₄NO₃, H₂SO₄, and Ti surfaces. The data were statistically analyzed using ANOVA with a SNK post hoc test (^**p < 0.01).

To discuss the relationship between the proliferation rate and the release of Ni ions in detail, we estimated the Ni concentration in the cell culture medium after cultivating for the prescribed periods. The estimated concentration is obtained by dividing the total Ni-releasing amount by the medium volume. The total amount was calculated from the rate in Fig. 3, and the volume of the cell medium was 500 mL. The estimated Ni concentrations in the medium are as follows: 22.8 µg·mL⁻¹ for the untreated surface, 7.5 µg·mL⁻¹ for the surface anodized in HNO₃ electrolyte, and 3.0 µg·mL⁻¹ for the surfaces anodized in NH₄NO₃, (NH₄)₂SO₄, and H₂SO₄ electrolytes. Shih et al. demonstrated the inhibitory influence on the growth of smooth muscle cells when the concentration of Ni ions in the cell medium exceeded 9 µg·mL⁻¹.²⁰⁾ They explained that this inhibitory effect is caused by the suppression of the transition from the G₁ phase (arrest phase) to the S phase (DNA synthesis phase) in the cell cycle. The estimated Ni concentration for the surfaces anodized in NH₄NO₃, (NH₄)₂SO₄, and H₂SO₄ electrolytes are below Shih’s threshold value; as a result, the corresponding anodized surface supported rapid proliferation similar to an untreated Ti surface. Sugie et al. cultivated fibroblast cells in a Ni-containing medium and evaluated the number of cells in each cultivation period.²¹⁾ The number of cells decreased after 3 days cultivation when the Ni concentration exceeded 25 µg·mL⁻¹, whereas the cells increased moderately when the Ni concentration was 12.5 µg·mL⁻¹, despite the lower than a Ni-free medium. The low proliferation rate of the untreated NiTi surface is thus likely to cause the reduction of ECs owing the cytotoxic effect of the Ni ions. In addition, the improved proliferation on the surface anodized in HNO₃ electrolyte, although inferior to the anodized surface with the other three electrolytes, is due to the suppression of the cell cycle transition by Ni ions.

In summary, we concluded that the selection of an appropriate electrolyte in the pulsed-anodization process is a significant factor responsible for regulating the surface features, thereby determining the attachment as well as the proliferation of ECs. Further study to elucidate the electrolyte effect to the surface characteristics of the pulsed-anodized surface would be indispensable to develop the ideal NiTi alloy for stent application.

4. Conclusion

The selection of an electrolyte for pulsed-anodization technology for an almost equiatomic NiTi alloy is related to the characteristics of the surface morphology and corrosion resistance of the resultant Ni-free TiO₂ layer, which dominates the activity of surface-attached endothelium cells (ECs). For instance, the ECs on the pulsed anodized surface using H₂SO₄ as the electrolyte are comparatively inactive, particularly in the cell attachment phase because the submicron scale groove-like structure inhibits cell expansion. In addition, the facilitation effect of the proliferation of ECs by the pulsed anodization is diminished when HNO₃ is selected as the electrolyte as a result of its inferior ability to offer protection against corrosion. Electrolytes in which the resultant anodized surface is a smooth surface with high protection against corrosion are ideal, because surfaces with these properties are applicable as biomaterial applications. In this regard, the NH₄NO₃ and (NH₄)₂SO₄ electrolytes are more appropriate.

Acknowledgments

The authors thank Mr. Yamane, Mr. Tokuda, and Mr. Shirakawa from the Kitami Institute of Technology for their assistance with XPS analysis, FE-SEM observation, and GF-AAS analysis. The authors also acknowledge the valuable advice of Dr. Sakamoto from Gunma University regarding XPS spectral interpretation.

REFERENCES

• 1)   P.K.  Shah: Circulation 107 (2003) 2175–2177. doi:10.1161/01.CIR.0000069943.41206.BD
• 2)   N.B.  Wood,  S.Z.  Zhao,  A.  Zambanini,  M.  Jackson,  W.  Gedroyc,  S.A.  Thom,  A.D.  Hughes and  X.Y.  Xu: J. Appl. Physiol. 101 (2006) 1412–1418. doi:10.1152/japplphysiol.00051.2006
• 3)   S.  Banerjee,  K.  Sarode,  A.  Mohammad,  O.  Gigliotti,  M.S.  Baig,  S.  Tsai,  N.W.  Shammas,  A.  Prasad,  M.  Abu-Fadel,  A.  Klein,  E.J.  Armstrong,  H.  Jeon-Slaughter,  E.S.  Brilakis and  D.L.  Bhatt: Circulation 9 (2016) e002730. doi:10.1161/CIRCINTERVENTIONS.115.002730
• 4)   S.H.  Im,  D.H.  Im,  S.J.  Park,  Y.  Jung,  D.H.  Kim and  S.H.  Kim: Prog. Mater. Sci. 126 (2022) 100922. doi:10.1016/j.pmatsci.2022.100922
• 5)   M.  Hong,  G.S.  Mintz,  C.W.  Lee,  D.  Park,  K.  Park,  B.  Lee,  Y.  Kim,  J.  Song,  K.  Han,  D.  Kang,  S.  Cheong,  J.  Song,  J.  Kim,  S.  Park and  S.  Park: Circulation 113 (2006) 414–419. doi:10.1161/CIRCULATIONAHA.105.563403
• 6)   M.A.  Surovtseva,  O.V.  Poveschenko,  O.S.  Kuzmin,  I.I.  Kim,  A.S.  Kozhukhov,  N.A.  Bondarenko,  E.V.  Chepeleva,  A.N.  Kolodin,  A.P.  Lykov,  D.V.  Shcheglov,  V.F.  Pichugin and  I.Yu.  Zhuravleva: Appl. Surf. Sci. 578 (2022) 152059. doi:10.1016/j.apsusc.2021.152059
• 7)   N.  Ohtsu,  K.  Yamasaki,  H.  Taniho,  Y.  Konaka and  K.  Tate: Surf. Coat. Technol. 412 (2021) 127039. doi:10.1016/j.surfcoat.2021.127039
• 8)   K.  Yamasaki,  H.  Taniho,  K.  Tate and  N.  Ohtsu: Surf. Coat. Technol. 417 (2021) 127221. doi:10.1016/j.surfcoat.2021.127221
• 9)   H.  Taniho,  K.  Tate,  R.  Kawakami and  N.  Ohtsu: Surf. Interface Anal. 54 (2022) 1070–1077. doi:10.1002/sia.7133
• 10)   Y.  Senju and  H.  Miyata: J. Biochem. 145 (2009) 137–150. doi:10.1093/jb/mvn151
• 11)   M.N.  Andalib,  J.S.  Lee,  L.  Ha,  Y.  Dzenis and  J.Y.  Lim: Acta Biomater. 9 (2013) 7737–7745. doi:10.1016/j.actbio.2013.04.013
• 12)   E.  Lamers,  R.  van Horssen,  J.  te Riet,  F.C.M.J.M.  van Delft,  R.  Luttge,  X.F.  Walboomers and  J.A.  Jansen: Eur. Cells Mater. 20 (2010) 329–343. doi:10.22203/eCM.v020a27
• 13)   R.G.  Thakar,  Q.  Cheng,  S.  Patel,  J.  Chu,  M.  Nasir,  D.  Liepmann,  K.  Komvopoulos and  S.  Li: Biophys. J. 96 (2009) 3423–3432. doi:10.1016/j.bpj.2008.11.074
• 14)   F.  Gentile: J. Biomech. 115 (2021) 110134. doi:10.1016/j.jbiomech.2020.110134
• 15)   J.  Wei,  T.  Igarashi,  N.  Okumori,  T.  Igarashi,  T.  Maetani,  B.  Liu and  M.  Yoshinari: Biomed. Mater. 4 (2009) 045002. doi:10.1088/1748-6041/4/4/045002
• 16)   J.  Park,  S.  Bauer,  K.  Mark and  P.  Schmuk: Nano Lett. 7 (2007) 1686–1691. doi:10.1021/nl070678d
• 17)   A.  Thapa,  D.C.  Miller,  T.J.  Webster and  K.M.  Haberstroh: Biomaterials 24 (2003) 2915–2926. doi:10.1016/S0142-9612(03)00123-6
• 18)   E.A.  Cavalcanti-Adam,  T.  Volberg,  A.  Micoulet,  H.  Kessler,  B.  Geiger and  J.P.  Spatz: Biophys. J. 92 (2007) 2964–2974. doi:10.1529/biophysj.106.089730
• 19)   S.  Miyabe,  T.  Suehiro,  Y.  Fujinaga,  H.  Tsuchiya,  S.  Hiromoto and  S.  Fujimoto: Mater. Trans. 57 (2016) 2065–2071. doi:10.2320/matertrans.MI201516
• 20)   C.  Shih,  S.  Lin,  Y.  Chen,  Y.  Su,  S.  Lai,  G.J.  Wu,  C.  Kwok and  K.  Chung: J. Biomed. Mater. Res. 52 (2000) 395–403. doi:10.1002/1097-4636(200011)52:2%3C395::AID-JBM21%3E3.0.CO%3B2-B
• 21)   T.  Sugie,  H.  Yamamoto,  S.  Natsuno,  M.  Kaga and  K.  Oikawa: Jpn. J. Ped. Dent. 25 (1987) 773–778. doi:10.11411/jspd1963.25.4_773