Bone bonding ability—how to measure it?

Abstract

Bone bonding ability can be examined using in vitro assays. However, conventional approaches use static conditions simulating body fluids without biologically relevant compounds (proteins/growth factors). These features could have a major impact on surface properties, thus we investigated the effect of perfusion and simulated body fluid (SBF) composition of medium on the formation of apatite layer on nickel-titanium alloy (NiTi) with modified surfaces. Apatite layer formation in SBF was investigated under three conditions: static, dynamic (perfusion) and dynamic in serum supplemented SBF (S-SBF). It was found that perfusion and protein supplemented media altered results significantly when compared with standard approaches. Adequate media that contained both organic and inorganic phases provided a more accurate model for bone bonding ability testing. It was also confirmed that both plasma modified and ground to a mirror finish (and cleaned) encourage formation of apatite layers in all test conditions, which suggests the positive interaction of the material in bone tissue environment.

---

Introduction

The ability of materials to integrate within hard tissues can be assessed by in vitro tests in simulated body fluids (SBF). These tests were originally developed by Kokubo^1–3 by using a fluid that mimics the ionic composition of human plasma. Bone bonding ability (i.e., the potential to bind with hard tissue) can be monitored by the formation of apatite layers on the surface of synthetic materials. It is well established that the materials that encourage apatite layer formation are generally integrated with human tissue^.4–6 However, there is still debate whether such methodology can be used universally for all types of materials. In particular, in the last decade several revised compositions of SBF have been proposed that mimic the body environment much better than SBF composition used in the original assay by Kokubo.^1,7 In fact apatite was not formed on the surface of β-TCP when using SBF for predicting bone bonding ability, however, the results of in vivo tests showed that this material integrated extensively with bone in the body.^5,6 Due to relatively complex preparation of media and lack of carbonate content Bohner and Lemaitre⁷ proposed modified versions of SBF, and demonstrated similar efficacy in predicting bone bonding ability of materials. This new medium also contains only inorganic ions that are prepared by mixing two solutions to prevent premature precipitation prior to experiment, which was its main advantage. Chrzanowski^8–10 proposed the feasibility of rapid surface analysis (<5 min) for determining the apatite layer formation using precise chemical and topographical techniques such as XPS and AFM. It was demonstrated that this approach provides a good correlation with cell-based assays. Importantly, Chrzanowski gave a new direction for assessing bone bonding ability by introducing organic compounds such as proteins into SBF media. His study demonstrated that organic content of media interacts immediately with the surface and plays a significant role in surface reactivity.¹¹

It is pertinent to note that body fluids contain inorganic ions as well as biologically relevant compounds (e.g., proteins, growth factors). Therefore, it is critical to include the effects of homeostasis and perfusion in SBF test to mimic precisely the physiological environment. Proteins interact with the surface immediately after placing the implant in the body and create a thin film that promotes cell responses.¹² It is therefore important using media that contains relevant proteins for predicting bone bonding ability of an implant. Similarly, it is necessary to take into account the impact of dynamic conditions in the body.

The objective of this study was to evaluate the effects of media composition and test conditions on the formation of apatite layer and material/media interactions for assessing bone bonding ability of a material. We selected NiTi as a test material because this alloy possesses a limited activity in the body. However, it has been demonstrated that modified surfaces (polished, plasma treated) are osteoinductive—capable to stimulate osteoblastic cell differentiation and unregulated expression of osteogenic genes.

Experimental

Superelastic NiTi alloy (Johnson Matthey Inc. UK) was used in this study. Samples were cut into square shapes of 8 × 8 mm and the surface was ground to a mirror finish using SiC paper (grit 500–2400). Prior to the surface modification by plasma sputtering, the samples were precisely cleaned using ultrasonication in propan-2-ol, washing with deionised water (18 MΩ cm⁻¹) each step for 5 min, immersing in nitric acid (63%) for 10 min, and finally ultrasonicating in deionised water for 5 min. Samples were dried with pressurised air and kept in a closed container to avoid contamination.¹³ These cleaned NiTi samples (coded as NT) were also used as control.

Surface modification

Plasma sputtering. NiTi samples (NT) were sputtered using rf-plasma argon glow discharge (Plasma Cleaner PC-150, SBT Inc. US) at set up conditions of: forward power = 20 W, reflected power = 4 W, DC bias = −140 V, base pressure = 0.008 Torr and working pressure (argon) = 0.4 Torr, and time of treatment of 300 s. The samples were coded as PC.

Surface characterisation

Surface chemistry. The surface chemistry of samples before and after modification was examined using X-ray photoelectron spectroscopy (XPS, SPECS-XPS, Germany) utilizing an Al Kα monochromated X-ray source. Survey scan for each sample was obtained at 100 eV passing energy between 0 and 1400 eV. Five scans for each sample were collected. High resolution spectra for the single elements were collected at 20 to 50 eV pass energy. The elemental composition (atomic concentration) was calculated from high-resolution scans using CasaXPS (Casa software Ltd) after subtracting Shirley background. Each measurement was performed at least in triplicate to ensure repeatability of data.

Wetting ability. Static contact angle (sessile drop) measurements were performed using a Krüss contact angle analyzer (DSA10-MK2, Germany). Droplets of approximately 0.75 μL of ultra-pure water (polar liquid; 18 MΩ cm-1), diiodomethane, glycerol (nonpolar liquid) and formamide (nonpolar liquid) were placed on the sample surface using automated syringe. The drop profile was recorded at 1 s intervals for 45 s, and triplicate measurements were carried out. In this study, the Fowkes equation was used to calculate the surface free energy (SFE). Fractional polarity (FP) was also calculated, which provides an indication of the proportion of polar components that contribute towards the total SFE.¹¹

Surface topography and roughness. The topography of NiTi samples were analysed using AFM (AsylumResearch, MFP-3D). The samples were scanned in non-contact mode using Si tip. The roughness Ra was obtained from lines profiles on images of 45 × 45 μm.

In vitro bioactivity tests. Simulated body fluid (SBF) was prepared according to Bohner's method for the bioactivity test.⁷ The tests were conducted in static and dynamic conditions. The SBF was supplemented with 5% fetal bovine serum (FBS Australian origin, Supplier PAA, Australia) (S-SBF) to investigate the effects of proteins on calcium phosphates precipitation (i.e., bioactivity). The supplemented SBF with FBS was used only under dynamic conditions.

In static conditions the samples were immersed in individual vials containing 5 mL of SBF (without any organic component) and incubated at 37 °C with 90% relative humidity. Three samples of each group were removed from the vials at day 1, 3, 7 and day 14. Samples were washed with 1 mL of deionised water three times and air dried.

In a dynamic condition, a custom-built perfusion reactor was used in the bioactivity test (Fig. 1). SBF was perfused at a rate of 1 mL min⁻¹ in a closed loop. The reactor and reservoir were kept in the incubator at 37 °C and 90% relative humidity. The SBF in the reservoir was replaced every three days.

Fig. 1 Schematic diagram of dynamic system.

Tests were conducted without controlling CO₂ content because our tests do not involve cells and control of CO₂ was not critical. Similarly other bioactivity tests are proposed to be run without CO₂ control.^1,7

The bioactivity tests evaluation. The bioactivity (determined by the precipitation of apatite layer on the surface of samples) was assessed by several analytical techniques. Scanning electron microscopy (SEM) was used to observe the morphology, chemical composition of the precipitated crystals and ions distribution was mapped using energy-dispersive X-ray spectroscopy (EDX), and the structure of precipitated crystals was analysed using X-ray diffraction (XRD) analysis.

Zeiss EVO 50 SEM/EDX spectrometer was used to analyse the samples. EDX maps were recorded at different magnification and at least four different locations at the sample. Ca/P ratio was calculated as an average from all the locations.

TF-XRD Shimadzu XRD-6000: S6000 with Cu-Kα radiation (λ = 1.54 Å) at a generator voltage of 45 kV and a current of 40 mA was used. The data collection was performed by the 2θ scan method with 1° as incident beam angle and a scan speed of 2° min⁻¹ at 0.02° step angle.

3. Results and discussion

3.1. Surface characterization

Surface chemistry. XPS measurements showed that plasma treatment (PC) reduced the concentration of nickel and carbon (contaminations) when compared with cleaned NiTi samples (Table 1). Plasma treated samples showed higher amounts of nitrogen, which could come from the sputter system that was purged with nitrogen. Based on XPS results of elemental concentration the Ni/Ti ratio of NT samples was reduced from 0.15 to 0.079 after plasma treatment.

Table 1 Chemical composition (atomic%) on the surface based on XPS analysis for NT and PC

Sample	Ni (at.%)	Ti (at.%)	O (at.%)	N (at.%)	C (at.%)	Ca (at.%)
NT	2.84 ± 0.12	18.63 ± 0.18	48.89 ± 0.33	0.70 ± 0.09	27.22 ± 0.42	1.73 ± 0.10
PC	1.53 ± 0.10	19.38 ± 0.19	53.92 ± 0.39	2.06 ± 0.17	23.15 ± 0.50	n/a

---

Chemical analysis evidenced a small quantity of calcium on NT samples. The presence of calcium was not expected, calcium is regarded as a typical laboratory contaminant and it is believed that it appeared during polishing and washing steps. Calcium contaminations were previously reported and associated with laboratory contaminations.

The analysis of high resolutions spectra in Fig. 2 show that Ni2p line had two main peaks at binding energies of 852.7 and 856.2 eV. These peaks correspond to metallic nickel and nickel oxides (Ni(OH)₂ or Ni₂O₃).^10,12 Titanium spectra for NT sample (Fig. 2b) have main peaks at 454.4 eV, and the second double-peak at 458.9 eV and 464.7 eV. These energies correspond to metallic titanium and titanium dioxide.

Fig. 2 XPS detailed spectra of (a) Ni2p and (b) Ti2p for NT and PC.

Surface roughness analysis in Fig. 3a shows that in general the surfaces are smooth and the estimated mean roughness is 4 nm. Plasma treatment exhibited negligible impact on the surface roughness. As depicted in Fig. 3b, grinding process resulted in creating some scratch marks (e.g., depth <10 nm) on the surface of both types of samples.

Fig. 3 AFM images of the (a) cleaned-NT and (b) plasma cleaned- PC.

Surface free energy. Plasma treatment resulted in significant increase in the total surface free energy (γt) when compared to NT samples (p < 0.01). This surface modification enhanced the polar component (γp) and reduced the contribution of dispersive component (γd) on the surface of NiTi. Subsequently as the data shows in Table 2, fractional polarity (FP) increased for PC treated samples. Therefore, it can be concluded that plasma treatment promoted surface activity interactions with body fluids. It has been reported that FP of 0.3 promotes the adhesion of fibroblast cells. However, the impact of FP on promoting bonding and formation of apatite-like layers is still not clear.

Table 2 Average values of total surface free energy, γt, its polar, γp, and dispersive, γd, components and the calculated fractional polarity, FP

SFE (mN m^−1)	γd	γp	γt	FP
NT	29.80 ± 0.33	14.68 ± 0.19	44.48 ± 0.50	0.33 ± 0.005
PC	15.64 ± 0.24	42.14 ± 0.32	57.78 ± 0.56	0.73 ± 0.007

---

3.2 Bioactivity test using SBF

Identification of apatite precipitation. The formation of apatite on NiTi surface was confirmed by the presence of peaks at 2θ = 26° and 32° in XRD profiles shown in Fig. 4. The peak at 2θ = 42.3° corresponds to NiTi structure. The results in Fig. 4a and b demonstrate the presence of apatite on NiTi, and plasma treated samples in the static condition, respectively, at day 14 and 7. As shown in Fig. 4c and d in the dynamic condition, for both NiTi and PC samples the apatite peaks were detected after 3 days of SBF perfusion, which showed accelerated rates of precipitation. However, when tests were conducted in dynamic conditions using serum supplemented SBF (Fig. 4e and f) no apatite peaks were detected even after 14 days, which was similar for both NiTi and PC samples. The absence of apatite peaks indicates that the precipitated layer was not composed of crystalline apatite. This result demonstrates that the content of inorganic phase that is naturally present in the body has a significant impact on the formation of apatite and its structure and crystallinity.

Fig. 4 XRD spectra at various time points: a) control (NT) b) PC in static SBF; c) control (NT), d) PC in dynamic SBF; e) control (NT), f) PC in dynamic S-SBF. * indicates position of apatite peaks.

SEM analysis. SEM imaging in Fig. 5 confirms the presence of precipitation on the surface of samples for both groups under static condition (Fig. 5). After one day of incubation of NiTi samples in SBF agglomerates of apatite started to form; after three days the apatite layer covered the entire sample surface. After 14 days of incubation the layer had substantial thickness and completely covered the surface. The rate of apatite precipitation on plasma-treated samples was higher and at day one a large number of globules of apatite precipitates were observed as depicted in Fig. 6. These results are in agreement with the XRD results.

Fig. 5 SEM images for NT and PC samples after incubation in SBF under static conditions.

Fig. 6 SEM images for NT and PC samples after incubation in SBF under dynamic conditions.

Under dynamic conditions, shown in Fig. 7, the precipitation of apatite layer was substantially accelerated and for both types of samples a thick layer of apatite was observed at day three. The precipitation layers were prone to cracking under vacuum and even delaminated after the third day of incubation due to their substantial thickness.

Fig. 7 SEM images for NT and PC samples after incubation in FBS supplemented SBF under dynamic conditions.

The profile of apatite layer formation was influenced by the type of medium. The rate of apatite layer formation was decreased when SBF was supplemented with FBS-S-SBF. The total surface coverage of precipitated apatite was only observed for PC samples after 14 days as depicted in Fig. 7. XRD did not detect apatite even after 14 days but SEM images showed the presence of apatite crystals predominantly in globular form. The results of AFM analysis in Fig. 8 are in agreement with the data acquired from SEM and XRD; the apatetic globules were larger in size on NT samples than those on PC samples.

Fig. 8 AFM image (2 × 2 μm) of NT and PC samples incubated in S-SBF.

Mean Ca/P ratio in Table 3 is higher than that of stoichiometric hydroxyapatite (1.67) likely attributed to the type of CaP formed on the surface of treated NiTi alloy. Results show a range of Ca/P ratio from 1.37–2.26. There was no significant difference in Ca/P ratio between NiTi and PC for all three incubation conditions; static, dynamic and dynamic S-SBF (p > 0.05). However, by comparing the ratios across incubation conditions, Ca/P ratios for dynamic condition in SBF were significantly higher than both static conditions in SBF and dynamic conditions in S-SBF groups (p < 0.05).

Table 3 Mean Ca/P ratio from EDX compositional analysis of four areas on each sample

Ca/P ratio	Static		Dynamic		Dynamic S-SBF
	NT	PC	NT	PC	NT	PC
Day 1	1.56	1.64	1.44	1.37	1.37	1.47
Day 3	1.66	1.68	1.73	1.81	1.56	1.42
Day 7	1.59	1.72	2.26	2.21	1.43	1.48
Day 14	1.91	1.92	2.24	2.26	1.49	1.68

---

Stoichiometric hydroxyapatite has Ca/P ratio of 1.67. Hydroxyapatite is the most stable form of CaP and a favourable bioactive material for bone replacement due to its similar chemistry to the bone mineral phase. Several samples possessed Ca/P ratio of 1.67, however, their XRD profiles were not comparable with hydroxyapatite. For example samples with the greatest ratio of Ca and P had very clear apatite peaks (PC dynamic 7 and 14 days), while the samples with ratios close to 1.67 did not present any apatite peak in XRD spectra (PC dynamic S-SBF). Generally, peaks similar to that of hydroxyapatite were detected in XRD profile of samples having Ca/P ratio above 1.67. The higher Ca/P observed in this study could be due to the substitution of carbonate for the phosphate group in the precipitated apatite.¹⁴

Significant morphological differences of the precipitates were observed for all 3 test conditions (Fig. 9). Precipitates under static SBF conditions had the form of dense, small and uniform globules. Standard SBF and dynamic perfusion conditions resulted in the formation of a more developed structure; many elongated, flake-like crystals were formed, while the incubation in S-SBF resulted in the formation of CaP precipitates that were smooth and more rounded.

Fig. 9 SEM image showing morphological difference in Ca/P growth on samples.

4. Discussion

Plasma modification improves cell responses for various materials.^15,16 This strategy has been broadly used as a standard method for the surface modification of some biomaterials. The results achieved so far are promising for enhancing cell-material interactions and suggest that this approach can be beneficial for less active materials such as nickel-titanium alloys.

Nickel-titanium alloy is one of the materials used in biomedical applications, which has limited bioactivity due to the high nickel content. Several techniques were used to improve biological relevance of the material and recent studies have demonstrated that cell responses to the alloy were significantly improved by surface plasma modifications.¹⁰ These modifications addressed one of the outstanding problems of nickel-titanium alloy, which is its low bone bonding ability.

This characteristic is particularly critical for orthopaedics (i.e., spinal applications) of the alloy (e.g., spinal cages/discs). Chrzanowski has demonstrated that positive cell responses and upregulation of osteocalcin and osteopontin expressions was achieved by the plasma modifications.^9,10 The surfaces showed osteogenic character and the results are in good agreement with the apatite layer formation in SBF,⁸ which is attributed to bone bonding ability of materials. However, these tests were conducted using a standard protocol:⁸ static conditions in simulated body fluid developed by Kokubo.^1,2 Such a model replicates only the inorganic components of body media while it has been already shown that surfaces react with proteins within a few minutes forming a layer that can totally change interactions (precipitations) with ions from the body fluid.^9,17 Therefore, to mimic the natural body environment it is critical to introduce dynamic (perfusion) conditions and supplement traditional simulated body fluids with organic components such as proteins. Both dynamic and modified media make the test more biologically relevant and result in a new standard method in bone bonding ability testing.

In this study we investigated the effects of bone bonding ability testing conditions, i.e., perfusion and body media composition on formation of apatite layer to validate a new standard in these tests. Secondly, the study was conducted to demonstrate the advantages of plasma modifications on apatite layer formation. Two groups of samples were used: (i) ground to a mirror finish and, (ii) plasma modified nickel-titanium. It has been previously shown that both surfaces have osteogenic character and demonstrate similar high bioactivity.¹⁸

The formation of apatite layers and its compositions in different conditions (static vs. dynamic) has demonstrated that perfusion condition accelerated the precipitation of calcium phosphates. The flow of the medium allowed for increasing the deposition and mineralisation of precipitates on the surface. In dynamic conditions, the precipitates had higher Ca/P ratio and apatite peaks were detected even at the third day of incubation regardless of the type of substrate. However, in standard-static-conditions they were observed at day 7 for plasma treated samples and at day 14 for untreated samples. It is believed that the continuous exchange of media, thus delivery of ions present in dynamic conditions, contributed to the accelerated formation of apatite layer. Most likely the physical interactions of fluid and ‘pushing’ the media to the surface could have an impact on the apatite layer formation. We demonstrated that perfusion is a key factor in the assessment of apatite layer formation and provides data in conditions that represent more closely the body environment. It is anticipated that both nickel-titanium samples used in the study exhibit similar levels of activity. These results support our hypothesis: perfusion of the medium in the bone bonding ability test has an impact on the results. The combination of perfusion and correct medium composition provides a model for the bioactivity testing that is closer to the real scenario.

Further analyses were conducted under dynamic conditions using protein supplemented medium. The formation of apatite layers was comparable for both types of samples. However, the results of SEM analysis show that the rate of formation of this layer was decreased when compared with both static and dynamic conditions using standard SBF. Hence, the adsorption of proteins present in the serum had a negative impact on initially formed CaP, by reducing the surface availability and affinity to precipitate bonelike apatite.¹⁹ The main composition of serum is albumin which is negatively charged at physiological pH.^20,21 Thus Ca²⁺ ions bind to the albumin²² and decrease the rate of CaP precipitation. This phenomena may also be attributed to the rapid absorption of protein in serum on the surface of NiTi^23,24 that impede Ca²⁺ and PO₄²⁻ ion binding sites on the surface of the samples, hence nucleation of apatite crystals.²⁰ Importantly, no apatite peaks were detected in XRD spectra, which indicates that apatite was not formed on the surface. It can be anticipated that the Ca/P layer was predominately amorphous. The results are not in full agreement with our previous studies^8–10,25 because osteogenic character was not confirmed here. However, precipitates were clearly observed for both types of the samples at similar levels and it confirmed our hypothesis: organic components of medium used in the bone bonding ability tests change drastically the test results. It is still postulated that these results are more relevant to predict material behaviour in vivo but care must be taken because no apatite was observed in these conditions but samples were reported to trigger osteoblastic differentiation and mineralization.^10,12,18 This effect could indicate that such tests could produce false negative results and further measurement such as protein adsorption and cell responses are still required.

The type of CaP formed on the surface can be identified by conducting the morphology analysis. The flaky and porous structure of CaP was formed under dynamic SBF conditions and it is similar to the morphology of octacalcium phosphate (OCP). However, the smooth and globular CaP as observed in both static and dynamic S-SBF incubated conditions tended to be more needle-shape apatite-like crystals.²⁶ Based on EDX results, apatite formed in dynamic conditions using supplemented SBF showed lower Ca/P ratios because the Ca²⁺ ions could bind to proteins present in the media.²¹ It can be speculated that due to lower Ca²⁺ ions available for precipitation, calcium deficient hydroxyapatites was formed. The chemical and structural analysis of precipitates revealed that apatite diffraction peaks are detected only when the ratio of Ca/P was above 1.67, which refers to stoichiometric hydroxyapatite. This finding suggests that carbonate groups are substituted for phosphate groups during apatite formation. The higher Ca/P ratio of precipitated apatite has been reported in previous studies.^14,27

Bone bonding ability tests serve as in vitro analysis to reduce the number of animals for in vivo tests.^1,2 Differences in CaP film formation observed in the standard method for both tested materials were not observed in the modified methods, which is in agreement with our previous in vitro studies. The addition of serum and presence of shear force mimics bodily conditions, thus apatite forming ability in such conditions can be considered as a new approach for predicting bone bonding ability. However, osseointegration, bone tissue growth at bone-implant interface without the formation of fibrous tissue,^28,29 is a complex process and this proposed test is an alternative strategy that acts as a screening method prior to cell/tissue/animal based tests to minimise the required in vitro and in vivo tests. We demonstrated that various factors affect bioactivity and a new method developed for testing bone bonding ability mimics the human body environment and provide insights for materials integration with bone prior to in vivo studies. We found that plasma treated samples have been found to improve apatite film formation when a standard approach⁷ was used. However, when dynamic conditions and FBS supplemented media were used, negligible effects of treatment on the ability to form CaP layer were noted. These tests demonstrated that the apatite layers were readily formed on both types of samples and it is anticipated that both materials integrate within surrounding hard tissues. This finding is in full agreement with previous results, where insignificant differences between both types of samples were reported and expressions of osteogenic gens were at the same level.³⁰ Hence, tests with the use of dynamic conditions and serum supplemented media correlate well with cell-based experiments.

This study demonstrates that lower nickel content and more favourable surface free energy improved the apatite layer formation only for standard method (SBF without organic components in static condition). It could be stated that changes in surface energy and small differences in chemistry did not play a significant role in the layer formation in modified techniques due to the extensive interactions of surface with proteins that alter apatite layer formation. This data was in agreement with previous work.

5. Conclusions

The Perfusion of simulated body fluid is an important factor in the assessment of apatite layer formation. Typically in the dynamic condition, interactions between medium components and surfaces were enhanced, which resulted in accelerating the deposition of apatite film. Furthermore, it was found that organic components of medium used in bone bonding ability tests change drastically the results due to the protein/drug/peptides interactions with the surface of biomaterials. The media that contains both organic and inorganic phases provides a more accurate model for bioactivity testing. It was also confirmed that both plasma modified and ground to a mirror finish (and cleaned) encouraged formation of apatite layers in all test conditions, which allows to predict the positive integration of material with bone.

Despite the high rate of the layer precipitations on the control ground nickel-titanium and plasma treated when using perfusion and supplement medium no apatite peak was detected by XRD. The apatite might have formed but its level was below detection of XRD. It was postulated that perfusion and media that contain organic compounds are more relevant to predict material behaviour in vivo because they capture early protein/peptides interactions with surfaces, which are key in cell-surface communication. The lack of apatite formation suggests that such tests could produce false negative results and further aspects such as protein adsorption and cell responses are still required.

References

1. T. Kokubo and H. Takadama, Biomaterials, How useful is SBF in predicting in vivo bone bioactivity?, 2006, 27(15), 2907–2915.
2. T. Kokubo, Acta Materialia, Apatite formation on surfaces of ceramics, metals and polymers in body environment, 1998, 46(7), 2519–2527.
3. T. Kokubo, Materials Science and Engineering: C, Design of bioactive bone substitutes based on biomineralization process, 2005, 25(2), 97–104.
4. K. Nishio, Journal of Biomedical Materials Research, et al., The effect of alkali- and heat-treated titanium and apatite-formed titanium on osteoblastic differentiation of bone marrow cells, 2000, 52(4), 652–661.
5. J. Isaac, Journal of Biomedical Materials Research Part A, et al., : Bone-like tissue formation on a biomimetic titanium surface in an explant model of osteoconduction, 2009, 89A(3), 585–593.
6. A. Fukuda, Acta Biomaterialia, et al., Bone bonding bioactivity of Ti metal and Ti–Zr–Nb–Ta alloys with Ca ions incorporated on their surfaces by simple chemical and heat treatments, 2011, 7(3), 1379–1386.
7. M. Bohner and J. Lemaitre, Biomaterials, Can bioactivity be tested in vitro with SBF solution?, 2009, 30(12), 2175–2179.
8. W. Chrzanowski, Acta Biomaterialia, et al., : Effect of surface treatment on the bioactivity of nickel–titanium, 2008, 4(6), 1969–1984.
9. W. Chrzanowski, Journal of Biomedical Materials Research Part A, et al., In vitro studies on the influence of surface modification of Ni–Ti alloy on human bone cells, 2010, 93A(4), 1596–1608.
10. W. Chrzanowski, Journal of Biomaterials Applications, et al., : Biocompatible, Smooth, Plasma-Treated Nickel–Titanium Surface – An Adequate Platform for Cell Growth, 2012, 26(6), 707–731.
11. W. Chrzanowski, Journal of Biomedical Materials Research Part A, et al., In vitro studies on the influence of surface modification of Ni–Ti alloy on human bone cells., 2010, 93(4), 1596–1608.
12. W. Chrzanowski, Acta Biomaterialia, et al., : Effect of surface treatment on the bioactivity of nickel-titanium, 2008, 4(6), 1969–1984.
13. W. Chrzanowski, Materials Science and Engineering: C, et al., Impaired bacterial attachment to light activated Ni–Ti alloy, 2010, 30(2), 225–234.
14. R. Rohanizadeh, Journal of Biomedical Materials Research, et al., : Apatite precipitation after incubation of biphasic calcium-phosphate ceramic in various solutions: Influence of seed species and proteins, 1998, 42(4), 530–539.
15. P. K. Chu, Surface and Coatings Technology, Enhancement of surface properties of biomaterials using plasma-based technologies, 2007, 201(19–20), 8076–8082.
16. L. Tan, R. A. Dodd and W. C. Crone, Biomaterials, Corrosion and wear-corrosion behavior of NiTi modified by plasma source ion implantation, 2003, 24(22), 3931–3939.
17. W. Chrzanowski, Archives of Materials Science and Engineering, et al., Study on bioactivity of NiTinol after surface treatment, 2008, 21(1), 3.
18. W. Chrzanowski, Engineering surface properties of nickel-titanium alloy for improved osteointegration. 1 ed. Vol. 1., 2012Gliwice: Printing House of the Silesian University of Technology. 161.
19. S. Areva, Journal of Biomedical Materials Research Part A, et al., Use of sol–gel-derived titania coating for direct soft tissue attachment, 2004, 70A(2), 169–178.
20. Y. Wang, Materials Science and Engineering: C, et al.,In vitro behavior of fluoridated hydroxyapatite coatings in organic-containing simulated body fluid, 2007, 27(2), 244–250.
21. F. Sussman and H. Weinstein, Proceedings of the National Academy of Sciences, On the ion selectivity in Ca-binding proteins: the cyclo(-L-Pro-Gly-)3 peptide as a model, 1989, 86(20), 7880–7884.
22. H. A. Saroff and M. S. Lewis, The Journal of Physical Chemistry, THE BINDING OF CALCIUM IONS TO SERUM ALBUMIN, 1963, 67(6), 1211–1216.
23. A. S. Svetlana, Bio-Medical Materials and Engineering, Surface, corrosion and biocompatibility aspects of Nitinol as an implant material, 2002, 12(1), 69–109.
24. S. Shabalovskaya, J. Anderegg and J. Van Humbeeck, Acta Biomaterialia, Critical overview of Nitinol surfaces and their modifications for medical applications, 2008, 4(3), 447–467.
25. K. Page, Journal of Materials Science, et al., Study of the adhesion of &lt;i&gt;Staphylococcus aureus&lt;/i&gt; to coated glass substrates, 2011, 46(19), 6355–6363.
26. C. Gobel, Journal of Materials Chemistry, et al., Phase formation and morphology of calcium phosphate-gelatine-composites grown by double diffusion technique: the influence of fluoride, 2004, 14(14), 2225–2230.
27. R. Rohanizadeh, M. Trécant-Viana and G. Daculsi, Calcified Tissue International, Ultrastructural Study of Apatite Precipitation in Implanted Calcium Phosphate Ceramic: Influence of the Implantation Site, 1999, 64(5), 430–436.
28. D. A. Puleo and A. Nanci, Biomaterials, Understanding and controlling the bone-implant interface, 1999, 20(23-24), 2311–2321.
29. J. Davies, Journal of Dental Education, Understanding peri-implant endosseous healing, 2003, 67(8), 932–949.
30. W. Chrzanowski, Journal of Biomaterials Applications, et al., Biocompatible, smooth, plasma-treated nickel–titanium surface – an adequate platform for cell growth, 2011.

---