- Email: [email protected]
Comparative in vitro study on a ultra-high roughness and dense titanium coating
ARTICLE IN PRESS
Biomaterials 26 (2005) 4948–4955 www.elsevier.com/locate/biomaterials
Comparative in vitro study on a ultra-high roughness and dense titanium coating Veronica Borsaria, Gianluca Giavaresia,b,, Milena Finia,b, Paola Torricellia,b, Matilde Tschona, Roberto Chiesac, Loris Chiusolid, Armando Salitoe, Andreas Volpertf, Roberto Giardinoa,b a Department of Experimental Surgery, Research Institute ‘Codivilla-Putti’, Rizzoli Orthopaedic Institute, Bologna, Italy C.R.I.S.M.A., Interuniversity Research Centre for Advanced System, Department of Clinical Medicine and Applied Biotechnology ‘‘D. Campanacci’’, University of Bologna, Bologna, Italy c Department of Chemistry, Materials and Materials Engineering ‘‘G. Natta’’, Polytechnic of Milano, Milano, Italy d SAMO S.p.A., Via Matteotti 37, 40057 Cadriano di Granarolo Emilia, Bologna, Italy e Sonnenweg 15, 5610 Wohlen/AG, Switzerland f Sulzer Metco AG, Rigackerstrasse 16, 5610 Wohlen, Switzerland
b
Received 19 August 2004; accepted 5 January 2005
Abstract A new implant surface has been developed with the purpose of avoiding as much stress shielding as possible, and thus prolong the prosthesis lifespan. The aim of this study was to investigate the in vitro effect of this new ultra-high roughness and dense Titanium (Ti) surface (PG60, Ra ¼ 74 mm) in comparison with medium (TI01, Ra ¼ 18 mm) and high (TI60, Ra ¼ 40 mm) roughness and open porous coatings; all the coatings were obtained by vacuum plasma spraying. MG63 osteoblast-like cells were seeded on the tested materials and polystyrene, as control, for 3 and 7 days. Cells proliferated on the material surfaces similarly to the control. Alkaline phosphatase activity had lower values for TI60 than TI01 (po0.0005) and PG60 (po0.005). Osteocalcin levels measured on TI60 were significantly (po0.0005) lower in comparison with TI01 and PG60 at 7 days. Procollagen-I synthesis reduced with increasing roughness and the lowest data was found for PG60. While at 3 days Transforming Growth Factor b1 levels augmented with increasing roughness, at 7 days TI60, the high roughness surface, was significantly lower than PG60 (po0.005) and TI01 (po0.001). All tested materials showed significantly higher Interleukin-6 levels than those of polystyrene at both experimental times. Nitric Oxide activity on TI01 was significantly (po0.05) higher than on TI60 and polystyrene. In conclusion, the new ultra-high roughness and dense coating PG60 provided a good biological response, even though, at least in vitro, it behaved similarly to the coatings already used in orthopaedics. r 2005 Elsevier Ltd. All rights reserved. Keywords: Osteoblast; Titanium; Plasma spraying; Surface roughness; Cell proliferation; Cytokines
1. Introduction
Corresponding author. Servizio di Chirurgia Sperimentale, Istituto di Ricerca Codivilla-Putti, Istituti Ortopedici Rizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. Tel.: +39 051 6366787; fax: +39 051 6366580. E-mail address: [email protected] (G. Giavaresi).
0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.01.010
Because of the increase in life expectancy, the number of total joint arthroplasties has risen and the longevity of orthopaedic prosthesis should be improved. Despite great advances in cement and biological fixations over the last decade, the life course of a joint prosthesis is usually reduced by various conditions. Since prostheses react to stresses differently compared to the original
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
bone, no longer receiving its usual loading, the stress distribution on the surrounding bone causes a loosening process (stress shielding), which is the main culprit of prosthesis failure [1]. Attempts to prolong the prosthesis lifetime have been made by improving and accelerating osteointegration, through the development of innovative implant surfaces with the purpose of avoiding stress shielding as much as possible. As the interaction between host tissue and implant takes place on the material surface, the bone-implant interaction and the conditions which allow optimal osteointegration are dependent on the chemical and physical properties of the implant surface. Titanium (Ti) and Ti alloys are widely used for orthopaedic implants thanks to their good mechanical properties and biocompatibility [2]. For biological fixation, the use of metal and ceramic-coated prostheses ensures a secure mechanical engagement in the bone. In fact, the aim of the circumferential coating is to maximize proximal metaphyseal fit, thus reducing stress shielding and preventing wear debris from causing osteolysis due to the specific defense mechanism of phagocytosis that stimulates macrophages/monocytes to secrete mediators of bone resorption such as IL-6, IL-1, TNF-a and eicosanoids [3]. Several in vitro studies have showed that also osteoblasts phagocytose particles and upregulate the release of cytokines and PGE2, thus reducing type-I Collagen synthesis, and playing a critical role in pathological bone resorption, through both osteoclast activation and reduction of osteoblastic bone formation [4–7]. Surface topography and roughness are considered to be very important for osteointegration. Surface roughness in particular seems to have a direct effect on osteoblast attachment and subsequent proliferation and differentiation [2,8–18]. Osteoblast-like cells adhere more readily to rough surfaces and appear more differentiated on rougher surfaces, with regards to morphology, extracellular matrix production, Alkaline Phosphatase activity and Osteocalcin production, and response to systemic hormones such as 1,25-(OH)2D3 [19–21]. It has been demonstrated that roughness degree also influences the synthesis of two local factors, TGFb1 and PGE2, which can act on the osteoblastic cells as autocrine regulators, and modulate the activity of bone resorbing cells via paracrine mechanisms [10,21–23]. Since the most appropriate surface coating has not yet been identified, new coating techniques are continuously being developed and investigated, thus producing physicochemical and morphological modifications. Vacuum plasma spraying (VPS) technology offers significant advantages for depositing oxygen-sensitive materials, selecting the roughness and porosity of an implant surface, giving a well-bonded coating with optimized structure and surface morphology, and finally for developing firm and stable long-term fixation between
4949
the implant and the living tissue [24–26]. A new ultrahigh roughness and dense Ti VPS coating, whose application in orthopaedics has never been proposed, with an improved bonding to the Ti6Al4 V implant and stability, has been characterized in a previous study in comparison to other commercial coatings already used [27] showing good cell adhesion and differences in cell morphology depending on the degree of roughness. The aim of this study was to investigate closely the in vitro effect of this new ultra-high roughness Ti surface on MG63 cell proliferation, differentiation, synthetic activity and production of local factors.
2. Materials and methods 2.1. Material samples Disks of Ti6Al4 V ELI alloy (11.30 mm in diameter and 2 mm in thickness) were used as a substrate for all surface treatments, while Ti grade 4 (ISO 5832-2) was used as powder to produce all the different coatings. Ti6Al4 V disk substrates, performed by lathe machining (SAMO SpA, Bologna, Italy), were grit-blasted with angular Al2O3 powder and then coated on one side by means of VPS technology (SUMEsPLANT, Sultzer Metco, Wohlen, Switzerland). Three Ti coatings (Table 1) were produced by the precise control of the partial melting point, velocity and
Table 1 Surface coating parameters of thickness, porosity and roughness obtained by mechanical profilometer (Mahr, Perthometer S2, translation system PGK, tip MFW 250-6) Parameters
Thickness (mm) Porosity (%) Ra (mm) Rz (mm) Rt (mm)
Coatings Medium roughness and porous
High roughness and porous
Ultrahigh roughness and dense
TI01 335.0 25.4 18.22 109 120
TI60 420.0 26.7 39.64 215 248
PG60 500.0 4.5 73.45 351 351
TI01: Angular blocky surface with open porosity produced by melted and semi-melted Ti VPS plasma sprayed particles of medium size; TI60: Macro angular blocky surface with open porosity produced by melted and semi-melted Ti VPS plasma sprayed particles of large size; PG60: Mix of Ti splats and semi-melted spherical Ti particles without open porosity; Ra : arithmetic mean of the area between the roughness profile and its mean line; Rz : arithmetic mean of the five highest peaks plus the depth of the five deepest valleys over the evaluation length; Rt : length of the highest peak plus the depth of the deepest valley over the evaluation length.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
4950
size of the Ti particles: medium and high roughness and opened porous coatings (TI01 and TI60, respectively), and an ultra-high roughness and dense coating (PG60). The parameters considered and changed during the deposition process were: Ti powder grain size, system vessel pressure, plasmogenic gas flow, current, voltage, powder flow and specimens distance. The coating surface morphology was examined by scanning electron microscopy (JEOL LTD, JSM-5410) on two disks for each surface treatment (Fig. 1a, c, e). Chemical and crystallographic analysis carried out by X-ray diffraction (XRD) technique (Philips PW 1710). XRD analysis on TI01, TI60, PG60 materials, regardless of the surface
morphology and roughness, highlighted a chemical composition of a pure Ti. All samples were first cleaned with isopropyl alcohol and then ultrasonically with de-ionised water, and finally dried in a clean room under laminar flow ventilation. All samples were packaged in blisters and sterilized by g-rays in air with a standard dose of 25 kGy. 2.2. Cell culture Human osteoblast-like cells (MG63) were cultivated in DMEM containing 10% FCS, 100 IU/ml penicillin and 100 mg/ml streptomycin solution at 37 1C in an
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 1. Scanning electron micrographs of all tested materials before (a, c, e) and after (b, d, f) cell culture at a magnification of 150 (a, c, e—scale bar: 200 mm) and of 500 (b, d, f—scale bar: 10 mm): (a and b) TI01; (c and d) TI60; (e and f) PG60.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
atmosphere of 5% CO2 and 95% humidity. When cells were confluent, they were released with 0.05% (w/v) trypsin and 0.02% (w/v) EDTA, counted (Coulter Counter Z1, Beckman Coulter Inc., Miami, FL,USA) and used at a concentration of 0.5 104 cells/ml for the experiment. Ten specimens from each group were placed in 24-well plates and a cell suspension (0.5 104 cells in 100 ml) was directly seeded over each sample. The same amount of osteoblast-like cells was plated into the same number of empty polystyrene wells as controls (polystyrene). After a 2-h incubation to allow cell adhesion to the substrate, 900 ml of DMEM (supplemented with 50 mg/ml Ascorbic acid and b-Glycerophosphate 108 M) was added to 10 wells for each material. Cultures were maintained under the same conditions as described above for two experimental times: four samples of each material were incubated for 3 days and the other six for 7 days, when cells reached confluence on polystyrene, and the medium was replaced with fresh medium on the third day. After 3 and 7 days the supernatant was collected from four wells and centrifuged to remove particulates, if any. Aliquots of the supernatant were dispensed in Eppendorf tubes for storage at 80 1C and assayed for type I Procollagen (CICP, Prolagen-C enzyme Immunoassay kit, Metra Biosystem, CA, USA), Interleukin-6 (IL-6, Human IL-6 Immunoassay kit, Biosource International, CA, USA) and Transforming Growth Factor b1 (TGFb1, Quantikine human TGF-b1, Immunoassay, R&D Systems, MN, USA), Alkaline Phosphatase activity (ALP, Sigma Kinetic method kit, St. Louis, MO, USA), Nitric Oxide (NO, Sigma colorimetric assay, St. Louis, MO, USA), Osteocalcin (OC, Novocalcin enzyme Immunoassay kit, Metra Biosystem, CA, USA), and Lactate Dehydrogenase (LD-L, Sigma Diagnostics, St. Louis, MO, USA). The WST-1 test (Cell Proliferation Reagent WST-1, Roche, IN, USA) was carried out on the cells cultivated on four samples of each group for each experimental time to assess cell proliferation and viability: 100 ml of WST-1 solution and 900 ml of medium (final dilution: 1:10) were added to the cell monolayers, and the multiwell plates were incubated at 37 1C for a further 4 h. Supernatants were quantified spectrophotometrically at 450 nm with a reference wavelength of 640 nm. Results are reported as optical density (OD). At the end of the experimental time, 7 days, the last two specimens of each material were processed for analysis with a scanning electron microscope (SEM): cells grown on the material were fixed in 2.5% glutaraldehyde in a pH 7.4, 0.01 M phosphate buffer for 1 h, rinsed in the buffer and finally dehydrated in a graded series of ethanol. The samples were then washed four times in liquid CO2 and submitted to the critical point dehydration in CO2 atmosphere (80 ATM at 33 1C). Finally, the specimens were mounted on
4951
aluminium stubs for SEM analysis, earthed with carbon tape and then coated with a thin layer (25 A˚) of Au–Pd. The specimens were examined at 5 kV with a Jeol 840A SEM and the images were digitalized. 2.3. Statistical analysis Statistical analysis was performed using the SPSS v.10.1 software (SPSS Inc., Chicago, IL, USA). Data are reported as mean7SD at a significance level of po0.05. After testing data for normal distribution, homogeneity of variance, correlations among the parameters, a multivariate two-way (material: TI01, TI60, PG60, polystyrene; and experimental time: 3 and 7 days) ANOVA test was performed to assess significant interaction effects between biochemical data and selected factors. When such interactions were found, a univariate ANOVA was performed to investigate the effects of the factors on the data by means of hypotheses expressed as linear matrix according to the SPSS syntax. When no interaction was found, a one-way ANOVA (for each factor that showed a significant effect) was performed. The analysis was performed in two steps by: (A) comparing all the tested materials versus polystyrene; and (B) multiple comparisons among TI01, TI60 and PG60.
3. Results Table 2 reports the results for MG63 cultures in contact with the tested materials for 3 and 7 days. The statistical analysis, performed by considering the interaction of ‘material’ and ‘experimental time’ factor effects on cell viability (WST-1) and synthetic activity results of MG63 cultures, showed significant interactions for WST-1 (F ¼ 78:68; po0.0005), OC (F ¼ 4:27; po0.05), TGF-b1 (F ¼ 16:12; po0.0005) and IL-6 (F ¼ 19:73; po0.0005). On the contrary, no significant interactions were seen for ALP, CICP and NO parameters. When the effect of each factor was analysed independently for these last parameters, significant effects of both the ‘material’ and ‘experimental time’ factors were found: ALP (material: F ¼ 5:98; po0.005; experimental time: F ¼ 92:32; po0.0005); CICP (material: F ¼ 11:92; po0.0005; experimental time: F ¼ 224:61; po0.0005) and NO (material: F ¼ 3:98; po0.05; experimental time: F ¼ 68:80; po0.0005). Since experimental time effect was highly significant (po0.0005) in all tested parameters, no further considerations were made on the differences between experimental times, and only the one-way ANOVA by considering the factor ‘material’ was performed for ALP, CICP and NO. Finally, LDH activity was not detected in the supernatant of all tested materials.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
4952
Table 2 Cell viability, synthetic activity and local factors results for the MG63 cultures in contact with tested materials for 3 and 7 days (mean7SD, n ¼ 4) TI01 WST1 (OD) ALP (UI/L) OC (ng/ml) CICP (ng/ml) TGF-beta1(pg/ml) IL-6 (pg/ml) NO (mM)
3 7 3 7 3 7 3 7 3 7 3 7 3 7
days days days days days days days days days days days days days days
1.22870.016 1.71370.027 13.5071.13 10.8970.46 10.2070.50* 9.0070.40* 27.8075.96 66.9074.50 464738*,a 504723 2073* 209755*,a 2973 17.871.4
TI60 1.24970.025 1.95870.030*,b 11.6370.87 8.4770.82 10.4070.40 7.4070.50*,a 21.3075.40 57.6075.00 572725 419729*,a 2173* 128729* 2672 15.171.1
PG60
Polystyrene a
1.27470.004 1.72470.010 14.0971.21 9.6270.59 10.0070.30* 9.0070.40* 20.7075.50 44.9078.40 651760 497731 1772* 127728* 2772 17.971.0
1.25670.017 1.74570.006 13.4672.04 8.7170.91 11.0671.40 10.0070.30 34.3076.10 62.9076.80 62778 53879 471 871 20.7776.45 16.5372.81
Univariate ANOVA test between tested materials for each experimental times according to the established comparisons ((A) all the tested materials versus Polystyrene; (B) multiple comparisons among TI01, TI60 and PG60). WST-1 : (A) 7 days: *, TI60 versus Polystyrene (po0.0005); (B) 3 days: a, PG60 Versus TI01 (po0.005); 7 days: b, TI60 versus TI01 and PG60 (po0.0005). OC: (A) 3 days: *, TI01 and PG60 versus Polystyrene (po0.05);7 days: *, all tested materials versus Polystyrene (po0.05); (B)7 days: a, TI60 versus TI01 and PG60 (po0.0005). TGF-b1: (A) 3 days: *, TI01 versus Polystyrene (po0.0005); 7 days: *, TI60 versus Polystyrene (po0.0005). (B) 3 days: a, TI01 versus TI60 and PG60 (po0.05); 7 days: a, TI60 versus TI01 (po0.001) and PG60 (po0.005). IL-6: (A) 3 days: *, all tested materials versus Polystyrene (po0.0005); 7 days: *, all tested materials versus Polystyrene (po0.001). (B) 7 days: a, TI01 versus TI60 and PG60 (po0.05).
3.1. Cell proliferation and viability All tested materials behaved similarly in comparison with polystyrene; significantly higher results in WST-1 were seen for PG60 (4%, po0.005) in comparison with TI01 at 3 days. At 7 days, TI60 showed significantly higher WST-1 values (12%, po0.0005) in comparison with all the other surfaces.
3.2. Synthetic activity The trend of ALP activity was similar for all tested materials, having lower values for TI60 than TI01 (po0.0005) and PG60 (po0.005) without significant differences of all tested materials when compared with polystyrene. The levels of OC detected in the supernatant of all tested materials were significantly (po0.05) lower than that of polystyrene at both experimental times, except for TI60 at 3 days (ns). While OC levels measured on the different materials were similar at 3 days, TI60 was significantly (po0.0005) lower in comparison with TI01 and PG60 at 7 days. The CICP synthesis reduced with increasing roughness at both experimental times and the lowest data was seen for PG60, the ultra-high roughness surface, which was significantly different from TI01 (po0.0005) and polystyrene (po0.0005). In addition, TI60 showed
significantly lower CICP values (po0.05) in comparison with polystyrene. 3.3. Local factors As for TGF-b1, but only at 3 days, the significantly lowest value was found for TI01. At 7 days, TGF-b1levels of TI60, the high roughness surface, was significantly lower than PG60 (po0.005) and TI01 (po0.001). TGF-b1-levels of TI01 at 3 days and of TI60 at 7 days were significantly lower than those of polystyrene (po0.0005). All tested materials showed significantly higher IL-6 levels than those of polystyrene at both experimental times. While the levels of IL-6 assayed on the different materials were similar at 3 days, significant highest IL-6 synthesis was found for TI01 (po0.05) in comparison with TI60 and PG60 at 7 days. Finally, NO activity of all tested materials decreased from 3 to 7 days. When all data were considered together, all materials showed higher NO activity compared with polystyrene, and TI01 was significantly (po0.05) higher in comparison with TI60 and polystyrene. 3.4. Cell morphology After 7 days of culture, the MG63 cells seeded on the samples were well attached, in close contact with the
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
surface, and they looked rounded with no cytoskeletal processes (Fig. 1b, d, f). In particular, on TI01 and TI60 (Fig. 1b,d), the cell size (20–80 mm) was not comparable with the dimensions of the peaks and depths of the materials (porosity); on TI01 (Fig. 1b) cells were placed on the top (and probably on the bottom) of the projections of the material, because the contact area was very narrow, while on TI60 (Fig. 1d), where the contact area was larger, cells were also positioned along the side of the protrusions. On PG60 (Fig. 1f) the contact area was much wider and cells were located on the surface as if it was smooth.
4. Discussion The study evaluated and compared the biological response of MG63 osteoblast-like cells to Ti surfaces with different roughness levels (Ra 18C73 mm). To the authors’ knowledge, information about in vitro behaviour of materials with so high Ra values is limited. The ultra-high roughness surface PG60, is a novel coating whose roughness has been increased three times in comparison to TI01, which has been used for more than 15 years, but with a porosity level reduced to about 5%, thus giving a dense and potentially stronger coating. The high roughness coating, TI60, showed an increase in roughness parameters of nearly double those of TI01, while maintaining the porosity at about the same level of TI01, and has been recently used for prosthesis coatings. The present findings only partially support the hypothesis that surface roughness influences the proliferation and phenotypic expression of MG63 osteoblast-like cells, in agreement with the results reported in the literature [12], even though some considerations have to be made in order to understand the paradoxical behaviour of ultra-high roughness and dense coating PG60. Rough surfaces have been reported to produce better in vivo osteointegration than smoother surfaces, suggesting that surface modulates the bone response in terms of osteoblast adhesion, proliferation, differentiation and extra-cellular matrix deposition and calcification [23]. However, no consensus has been reached about the optimal surface roughness level of bone implants or whether there is a limit to improving roughness. In fact, in vivo studies by Vercaigne et al. on different high surface roughnesses (Ra 17C38 mm) showed that increases in Ra did not result in improvement of bone response, supposing that the increase of roughness leads to a higher Ti ion release with a negative effect on osteogenesis [28]. Conversely, more uniform results have been achieved with in vitro studies using MG63 cells as a model of relatively immature osteoblasts. Several authors reported that MG63 cells in contact with Ti and Ti alloys (Ra 0.6C7 mm) exhibit roughness-dependent increased
4953
expression of osteoblastic phenotype, as shown by reduced cell proliferation and increased ALP activity, and OC production as well as that of TGF-b1 and PGE2 [9,10,20,22,23]. Moreover, it has been found that surface topography positively influences the response of MG63 cells to regulatory stimuli such as 1,25(OH)2D3 above all on cell proliferation, on smooth Ti surfaces, and that OC production increases with rough Ti surfaces [9,12,19–21]. Finally, the observations by Schwartz et al. [21] suggested that the roughness-dependent regulation of osteoblast proliferation, differentiation and local factor production is related to the activation of integrin receptors by substrate, thus regulating Phosphokinase C (PKC) and A (PKA) through Pohspholipase C (PLC) and A2 (PLA2) pathway [29]. Although it is usually reported that surface roughness reduces cell proliferation [18], current results indicated that MG63 viability was not roughness-dependent; especially at 7 days. Proliferation was good on all the tested surfaces as seen by WST-1 levels that were not different in comparison to controls, and just the high roughness and porous TI60 coating showed the highest value. Previous investigations into PG60 revealed that surface morphology had a marked effect on osteoblast behaviour, by determining morphologic irregularities and reducing adhesion [27]. This phenomenon may be explained by considering the theory of Curtis and Wilkinson, in which the width and depth of the grooves on a surface, as well as the number of adjacent grooves, are considered the determining factors in establishing a positive reaction and orientation of the cells on substrates [30–32]. However, when the grooves or ridges are wider than the cell, the effect on orientation is not very marked [31]. The peaks and depths of the grooves of the tested coatings become much wider than the dimensions of the cell with increasing roughness, from TI01 to PG60, as observed by the metallographic analysis results in the previous study [27]. MG63 cells might have sensed the topography as smooth, as confirmed by the cell morphology results (Fig. 1), thus determining a different behaviour in terms of cell proliferation, that was not rough dependent. In particular, the big spherical Ti particles of PG60 could not have been recognized as discontinuities by MG63 cells, which ‘‘understood’’ the surface as smooth as polystyrene with a huge widening of the contact surface available for adhesion [27]. With regards to the experimental times of 3 and 7 days, they were chosen respectively, since they were believed to represent the minimum culture times to reach confluence on the polystyrene wells (3 days) and on the tested materials (7 days) at the seeding cell concentration of 0.5 104 cells/ml. However, cells did not reach the confluence on the tested materials at 7 days, because of the widening of the contact area in
ARTICLE IN PRESS 4954
V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
comparison to smooth polystyrene. In the current authors’ opinion, further investigations with longer experimental times such as 14 and 21 days could provide useful information about the condition that cell confluence was reached on high and ultra-high roughness coatings. The data concerning early differentiation (ALP), at both experimental times, and late differentiation (OC), at 7 days, are inversely related to cell viability, independently of roughness. In fact, from medium to high roughness level there is a reduction in cell differentiation, on the contrary to what is reported in the literature, while from high to ultra high roughness cell differentiation increases. However, the ALP activity results are similar to polystyrene, and just TI60, where cell proliferation was more pronounced, presented the lowest values. The decrease in OC levels measured on the tested surface in comparison to polystyrene could be explained by the fact that cells were not confluent on the materials after 7 days of culture, even though Schmidt et al. [33] observed a progressive reduction in OC synthesis over time on Ti coated materials. Procollagen type I synthesis results were inversely related to cell proliferation and to the roughness level, as found by Fini et al. [18]: low roughness promoted CICP production while high roughness had a negative influence. Some authors have demonstrated that the suppression of procollagen a1[I] gene expression followed by reduced type I collagen synthesis, is produced by phagocytosis of particulate debris by cells, that would lead to the activation of nuclear transcription factor-kappaB (NF-kB) and the subsequent upregulation of the release of proinflammatory cytokines (i.e. IL6) and PGE2 [3,34,35]. Surfaces with such a high roughness level and discontinuities could release particulate debris in vivo and this hypothesis could be also verified by histological analysis, but the present results indicate that this release occurred also in vitro. In fact, the IL-6 results increased from 3 to 7 days of culture and were higher on all tested surface coatings than polystyrene at both experimental times. The parallel increase in IL-6 synthesis and the reduction in CICP levels would support this hypothesis, thus confirming the observations of other authors on particle debris effect on MG63 functions [3,34]. It has been reported that greater surface roughness enhances the production and secretion of TGF-b1 and some authors suggested that the contact of the cells to material surface could stimulate a mechanism of control of cell functions and differentiation [10,18,19]. In particular, Batzer et al. [19] found that greater surface roughness enhances production and secretion of autocrine and paracrine mediators such as PGE2 in vivo, thus affecting bone formation distal to the implant and mediating the effect of the implant surface on the cells. The same increase in PGE2 was found with the present
tested materials at 3 days, while the highest PGE2 production was seen in TI01 and PG60 at 7 days (data not shown). The present findings differed from those reported in the literature because, even though the increase in TGF-b1 production seems to be proportional to the roughness level of the tested materials, TGF-b1 levels are similar to polystyrene (TI60 and PG60) or lower (TI01). Regarding PG60, the synthesis of TGF-b1 was not influenced by roughness because cells might not recognize such a high roughness level, as previously observed for cell proliferation. TGF-b1 results of TI01 and TI60 do not agree with those in the literature probably because the relationship between roughness and cell response observed by other authors is not appropriate when roughness values are so high. Also concerning NO synthesis, which increased by 16% on average on the tested materials when compared to polystyrene, two hypotheses can be made: a particledependent mechanism and a PGE2-dependent one, that would increase the release of NO and, as a consequence, lead to the suppression of osteoblast functions and the in vivo localized bone destruction around implants [36].
5. Conclusions The present in vitro evaluation performed on MG63 cells confirmed the biocompatibility of the tested coatings (TI01, TI60 and PG60), as expected, and also verified their effect on cell functions, which might be useful for future in vivo investigations into osteointegration. The new VPS ultra-high roughness and dense coating PG60 provided a good biological response, even though, at least in vitro, it behaved similarly to the coatings already used in orthopaedics. The effect of the coating stability and ultra-high roughness level after surgical implantation and during dynamic bone healing and remodelling has yet to be established [37]. Finally, the present findings highlighted that when the roughness level is greater than the cell dimensions it does not enhance cell response.
References [1] Yao J, Glant TT, Lark MW, Mikecz K, Jacobs JJ, Hutchinson NI, Hoerrner LA, Kuettner KE, Galante JO. The potential role of fibroblasts in periprosthetic osteolysis: fibroblast response to titanium particles. J Bone Miner Res 1995;10(9):1417–27. [2] Martin JY, Schwartz Z, Hummert TW, Schraub DM, Simpson Jr J, Lankfond J, Dean DD, Cochran DL, Boyan BD. Effect of titanium surface roughness on proliferation, differentiation and protein synthesis of human osteoblast-like cells (MG63). J Biomed Mater Res 1995;29:389–401. [3] Dean DD, Schwartz Z, Liu Y, Blanchard CR, Agrawal CM, Mabrey JD, Sylvia VL, Lohmann CH, Boyan BD. The effect of ultra-high molecular weight polyethylene wear debris on MG63 osteosarcoma cells in vitro. J Bone Jt Surg Am 1999;81(4):452–61.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955 [4] Kurihara N, Civin C, Roodman GD. Osteotropic factor responsiveness of highly purified populations of early and late precursors for human multinucleated cells expressing the osteoclast phenotype. J Bone Miner Res 1991;6(3):257–61. [5] Kurokouchi K, Kambe F, Yasukawa K, Izumi R, Ishiguro N, Iwata H, Seo H. TNF-alpha increases expression of IL-6 and ICAM-1 genes through activation of NF-kappaB in osteoblastlike ROS17/2.8 cells. J Bone Miner Res 1998;13(8):1290–9. [6] Galien R, Evans HF, Garcia T. Involvement of CCAAT/ enhancer-binding protein and nuclear factor-kappa B binding sites in interleukin-6 promoter inhibition by estrogens. Mol Endocrinol 1996;10(6):713–22. [7] Takei H, Pioletti DP, Kwon SY, Sung KL. Combined effect of titanium particles and TNF-alpha on the production of IL-6 by osteoblast-like cells. J Biomed Mater Res 2000;52(2):382–7. [8] Wen X, Wang X, Zhang N. Microrough surface of metallic biomaterials: a literature review. Biomed Mater Eng 1996;6(3):173–89. [9] Boyan BD, Batzer R, Kieswetter K, Liu Y, Cochran DL, Szmuckler-Moncler S, Dean DD, Schwartz Z. Titanium surface roughness alters responsiveness of MG-63 osteoblast-like to 1a,25-(OH)2D3. J Biomed Mater Res 1998;39:77–85. [10] Lincks J, Boyan BD, Blanchard CR, Lohmann CH, Liu Y, Cochran DL, Dean DD, Schwartz Z. Response of MG63 osteoblast-like cells to titanium and titanium alloy is dependent on surface roughness and composition. Biomaterials 1998;19(23):2219–32. [11] Ong JL, Hoppe CA, Cardenas HL, Cavin R, Carnes DL, Sogal A, Raikar GN. Osteoblast precursor cell activity on HA surfaces of different treatments. J Biomed Mater Res 1998;39(2):176–83. [12] Lohmann CH, Bonewald LF, Sisk MA, Sylvia VL, Cochran DL, Dean DD, Boyan BD, Schwartz Z. Maturation state determines the response of osteogenic cells to surface roughness and 1,25dihydroxyvitamin D3. J Bone Miner Res 2000;15(6):1169–80. [13] Deligianni DD, Katsala ND, Koutsoukos PG, Missirlis YF. Effect of surface roughness of hydroxyapatite on human bone marrow cells adhesion, proliferation, differentiation and detachment strength. Biomaterials 2001;22:87–96. [14] Perizzolo D, Lacefield WR, Brunette DM. Interaction between topography and coating in the formation of bone nodules in culture for hydroxyapatite- and titanium-coated micromachined surfaces. J Biomed Mater Res 2001;56:494–503. [15] Korovessis PG, Deligianni DD. Role of surface roughness of titanium versus hydroxyapatite on human bone marrow cells response. J Spinal Disord Tech 2002;15(2):175–83. [16] Montanaro L, Arciola CR, Campoccia D, Cervellati M. In vitro effects on MG63 osteoblast-like cells following contact with two roughness-differing fluorohydroxyapatite-coated titanium alloys. Biomaterials 2002;23:3651–9. [17] Campoccia D, Arciola CR, Cervellati M, Maltarello MC, Montanaro L. In vitro behaviour of bonemarrow derived mesenchymal cells cultured on fluorohydroxyapatite-coated substrata with different roughness. Biomaterials 2003;24:587–96. [18] Fini M, Giardino R, Borsari V, Torricelli P, Rimondini R, Giavaresi G, Nicoli Aldini N. In vitro behaviour of osteoblasts cultured on orthopaedic biomaterials with different roughness, uncoated and flourohydroxyapatite-coated, relative to the in vivo osteointegration rate. Int J Artif Organs 2003;26(6):520–8. [19] Batzer R, Liu Y, Cochran DL, Szmuckler-Moncler S, Dean DD, Boyan BD, Schwartz Z. Prostaglandins mediate the effects of titanium surface roughness on MG63 osteoblast-like cells and alter cell responsiveness to 1 alpha,25-(OH)2D3. J Biomed Mater Res 1998;41(3):489–96. [20] Lohmann CH, Sagun Jr. R, Sylvia VL, Cochran DL, Dean DD, Boyan BD, Schwartz Z. Surface roughness modulates the response of MG63 osteoblast-like cells to 1,25-(OH)(2)D(3) through regulation of phospholipase A(2) activity and activation of protein kinase A. J Biomed Mater Res 1999;47(2):139–51.
4955
[21] Schwartz Z, Lohmann CH, Vocke AK, Sylvia VL, Cochran DL, Dean DD, Boyan BD. Osteoblast response to titanium surface roughness and 1alpha,25-(OH)(2)D(3) is mediated through the mitogen-activated protein kinase (MAPK) pathway. J Biomed Mater Res 2001;56(3):417–26. [22] Orsini G, Assenza B, Scarano A, Piattelli M, Piattelli A. Surface analysis of machined versus sandblasted and acid-etched titanium implants. Int J Oral Maxillofac Implants 2000;15(6):779–84. [23] Boyan BD, Bonewald LF, Paschalis EP, Lohmann CH, Rosser J, Cochran DL, Dean DD, Schwartz Z, Boskey AL. Osteoblastmediated mineral deposition in culture is dependent on surface microtopography. Calcif Tissue Int 2002;71(6):519–29. [24] Brossa F, Cigada A, Chiesa R, Paracchini L, Consonni C. Adhesion properties of plasma sprayed hydroxyapatite coatings for orthopaedic prostheses. Biomed Mater Eng 1993;3:127–36. [25] Winkler-Gniewek W, Stallforth H, Ungethu¨m M, Gruner H. Structure and properties of VPS coatings in medical technology. First Plasma technik symposium, Lucerne/Switzerland, May 18th to 20th 1988. Vol 3, p. 97–102. [26] Salito A, Barbezat G, Filmer H, Hochstrasser J, Nicoll AR. Plasma spray systems for the deposition of materials for biomedical applications. In: Ravaglioli A, Krajewski A, editors. Proceeding of the international congress on bioceramic and the human body. Faenza, Italy, 2–5 April 1991. IRTEC-CNR. [27] Borsari V, Giavaresi G, Fini M, Torricelli P, Salito A, Chiesa R, Chiusoli L, Volpert A, Rimondini L, Giardino R. Physical characterization of different roughness titanium surfaces, with and without hydroxyapatite coating, and their effect on human osteoblast-like cells. J Biomed Mater Res, in press. [28] Vercaigne S, Wolke JC, Naert I, Jansen JA. The effect of titanium plasma sprayed implants on trabecular bone healing in the goat. Biomaterials 1998;19:1093–9. [29] Boyan BD, Sylvia VL, Liu Y, Sagun R, Cochran DL, Lohmann CH, Dean DD, Schwartz Z. Surface roughness mediates its effects on osteoblasts via protein kinase A and phospholipase A2. Biomaterials 1999;20(23–24):2305–10. [30] Anselme K, Bigerelle M, Noe¨l B, Iost A, Hardouin P. Effect of grooved titanium substratum on human osteoblastic cell growth. J Biomed Mater Res 2002;60:529–40. [31] Curtis A, Wilkinson C. Topographical control of cells. Biomaterials 1997;18:1573–83. [32] Curtis A, Clark P. The effect of topographic and mechanical properties of materials on cell behaviour. CRC Rev Biocompat 1990;5:343–62. [33] Schmidt C, Steinbach G, Decking R, Claes LE, Ignatius AA. IL-6 and PGE2 release by human osteoblasts on implant materials. Biomaterials 2003;24(23):4191–6. [34] Vermes C, Roebuck KA, Chandrasekaran R, Dobai JG, Jacobs JJ, Glant TT. Particulate wear debris activates protein tyrosine kinases and nuclear factor kappaB, which down-regulates type I collagen synthesis in human osteoblasts. J Bone Miner Res 2000;15(9):1756–65. [35] Nakashima Y, Sun DH, Trindade MC, Maloney WJ, Goodman SB, Schurman DJ, Smith RL. Signaling pathways for tumor necrosis factor-alpha and interleukin-6 expression in human macrophages exposed to titanium-alloy particulate debris in vitro. J Bone Jt Surg Am 1999;81(5):603–15. [36] Lohmann CH, Dean DD, Bonewald LF, Schwartz Z, Boyan BD. Nitric oxide and prostaglandin E2 production in response to ultra-high molecular weight polyethylene particles depends on osteoblast maturation state. J Bone Jt Surg Am 2002;84A(3):411–9. [37] Fini M, Giardino R. In vitro and in vivo tests for the biological evaluation of candidate orthopaedic materials: benefits and limits. J Appl Biomater Biomech 2003;1:155–63.
Biomaterials 26 (2005) 4948–4955 www.elsevier.com/locate/biomaterials
Comparative in vitro study on a ultra-high roughness and dense titanium coating Veronica Borsaria, Gianluca Giavaresia,b,, Milena Finia,b, Paola Torricellia,b, Matilde Tschona, Roberto Chiesac, Loris Chiusolid, Armando Salitoe, Andreas Volpertf, Roberto Giardinoa,b a Department of Experimental Surgery, Research Institute ‘Codivilla-Putti’, Rizzoli Orthopaedic Institute, Bologna, Italy C.R.I.S.M.A., Interuniversity Research Centre for Advanced System, Department of Clinical Medicine and Applied Biotechnology ‘‘D. Campanacci’’, University of Bologna, Bologna, Italy c Department of Chemistry, Materials and Materials Engineering ‘‘G. Natta’’, Polytechnic of Milano, Milano, Italy d SAMO S.p.A., Via Matteotti 37, 40057 Cadriano di Granarolo Emilia, Bologna, Italy e Sonnenweg 15, 5610 Wohlen/AG, Switzerland f Sulzer Metco AG, Rigackerstrasse 16, 5610 Wohlen, Switzerland
b
Received 19 August 2004; accepted 5 January 2005
Abstract A new implant surface has been developed with the purpose of avoiding as much stress shielding as possible, and thus prolong the prosthesis lifespan. The aim of this study was to investigate the in vitro effect of this new ultra-high roughness and dense Titanium (Ti) surface (PG60, Ra ¼ 74 mm) in comparison with medium (TI01, Ra ¼ 18 mm) and high (TI60, Ra ¼ 40 mm) roughness and open porous coatings; all the coatings were obtained by vacuum plasma spraying. MG63 osteoblast-like cells were seeded on the tested materials and polystyrene, as control, for 3 and 7 days. Cells proliferated on the material surfaces similarly to the control. Alkaline phosphatase activity had lower values for TI60 than TI01 (po0.0005) and PG60 (po0.005). Osteocalcin levels measured on TI60 were significantly (po0.0005) lower in comparison with TI01 and PG60 at 7 days. Procollagen-I synthesis reduced with increasing roughness and the lowest data was found for PG60. While at 3 days Transforming Growth Factor b1 levels augmented with increasing roughness, at 7 days TI60, the high roughness surface, was significantly lower than PG60 (po0.005) and TI01 (po0.001). All tested materials showed significantly higher Interleukin-6 levels than those of polystyrene at both experimental times. Nitric Oxide activity on TI01 was significantly (po0.05) higher than on TI60 and polystyrene. In conclusion, the new ultra-high roughness and dense coating PG60 provided a good biological response, even though, at least in vitro, it behaved similarly to the coatings already used in orthopaedics. r 2005 Elsevier Ltd. All rights reserved. Keywords: Osteoblast; Titanium; Plasma spraying; Surface roughness; Cell proliferation; Cytokines
1. Introduction
Corresponding author. Servizio di Chirurgia Sperimentale, Istituto di Ricerca Codivilla-Putti, Istituti Ortopedici Rizzoli, Via di Barbiano, 1/10, 40136 Bologna, Italy. Tel.: +39 051 6366787; fax: +39 051 6366580. E-mail address: [email protected] (G. Giavaresi).
0142-9612/$ - see front matter r 2005 Elsevier Ltd. All rights reserved. doi:10.1016/j.biomaterials.2005.01.010
Because of the increase in life expectancy, the number of total joint arthroplasties has risen and the longevity of orthopaedic prosthesis should be improved. Despite great advances in cement and biological fixations over the last decade, the life course of a joint prosthesis is usually reduced by various conditions. Since prostheses react to stresses differently compared to the original
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
bone, no longer receiving its usual loading, the stress distribution on the surrounding bone causes a loosening process (stress shielding), which is the main culprit of prosthesis failure [1]. Attempts to prolong the prosthesis lifetime have been made by improving and accelerating osteointegration, through the development of innovative implant surfaces with the purpose of avoiding stress shielding as much as possible. As the interaction between host tissue and implant takes place on the material surface, the bone-implant interaction and the conditions which allow optimal osteointegration are dependent on the chemical and physical properties of the implant surface. Titanium (Ti) and Ti alloys are widely used for orthopaedic implants thanks to their good mechanical properties and biocompatibility [2]. For biological fixation, the use of metal and ceramic-coated prostheses ensures a secure mechanical engagement in the bone. In fact, the aim of the circumferential coating is to maximize proximal metaphyseal fit, thus reducing stress shielding and preventing wear debris from causing osteolysis due to the specific defense mechanism of phagocytosis that stimulates macrophages/monocytes to secrete mediators of bone resorption such as IL-6, IL-1, TNF-a and eicosanoids [3]. Several in vitro studies have showed that also osteoblasts phagocytose particles and upregulate the release of cytokines and PGE2, thus reducing type-I Collagen synthesis, and playing a critical role in pathological bone resorption, through both osteoclast activation and reduction of osteoblastic bone formation [4–7]. Surface topography and roughness are considered to be very important for osteointegration. Surface roughness in particular seems to have a direct effect on osteoblast attachment and subsequent proliferation and differentiation [2,8–18]. Osteoblast-like cells adhere more readily to rough surfaces and appear more differentiated on rougher surfaces, with regards to morphology, extracellular matrix production, Alkaline Phosphatase activity and Osteocalcin production, and response to systemic hormones such as 1,25-(OH)2D3 [19–21]. It has been demonstrated that roughness degree also influences the synthesis of two local factors, TGFb1 and PGE2, which can act on the osteoblastic cells as autocrine regulators, and modulate the activity of bone resorbing cells via paracrine mechanisms [10,21–23]. Since the most appropriate surface coating has not yet been identified, new coating techniques are continuously being developed and investigated, thus producing physicochemical and morphological modifications. Vacuum plasma spraying (VPS) technology offers significant advantages for depositing oxygen-sensitive materials, selecting the roughness and porosity of an implant surface, giving a well-bonded coating with optimized structure and surface morphology, and finally for developing firm and stable long-term fixation between
4949
the implant and the living tissue [24–26]. A new ultrahigh roughness and dense Ti VPS coating, whose application in orthopaedics has never been proposed, with an improved bonding to the Ti6Al4 V implant and stability, has been characterized in a previous study in comparison to other commercial coatings already used [27] showing good cell adhesion and differences in cell morphology depending on the degree of roughness. The aim of this study was to investigate closely the in vitro effect of this new ultra-high roughness Ti surface on MG63 cell proliferation, differentiation, synthetic activity and production of local factors.
2. Materials and methods 2.1. Material samples Disks of Ti6Al4 V ELI alloy (11.30 mm in diameter and 2 mm in thickness) were used as a substrate for all surface treatments, while Ti grade 4 (ISO 5832-2) was used as powder to produce all the different coatings. Ti6Al4 V disk substrates, performed by lathe machining (SAMO SpA, Bologna, Italy), were grit-blasted with angular Al2O3 powder and then coated on one side by means of VPS technology (SUMEsPLANT, Sultzer Metco, Wohlen, Switzerland). Three Ti coatings (Table 1) were produced by the precise control of the partial melting point, velocity and
Table 1 Surface coating parameters of thickness, porosity and roughness obtained by mechanical profilometer (Mahr, Perthometer S2, translation system PGK, tip MFW 250-6) Parameters
Thickness (mm) Porosity (%) Ra (mm) Rz (mm) Rt (mm)
Coatings Medium roughness and porous
High roughness and porous
Ultrahigh roughness and dense
TI01 335.0 25.4 18.22 109 120
TI60 420.0 26.7 39.64 215 248
PG60 500.0 4.5 73.45 351 351
TI01: Angular blocky surface with open porosity produced by melted and semi-melted Ti VPS plasma sprayed particles of medium size; TI60: Macro angular blocky surface with open porosity produced by melted and semi-melted Ti VPS plasma sprayed particles of large size; PG60: Mix of Ti splats and semi-melted spherical Ti particles without open porosity; Ra : arithmetic mean of the area between the roughness profile and its mean line; Rz : arithmetic mean of the five highest peaks plus the depth of the five deepest valleys over the evaluation length; Rt : length of the highest peak plus the depth of the deepest valley over the evaluation length.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
4950
size of the Ti particles: medium and high roughness and opened porous coatings (TI01 and TI60, respectively), and an ultra-high roughness and dense coating (PG60). The parameters considered and changed during the deposition process were: Ti powder grain size, system vessel pressure, plasmogenic gas flow, current, voltage, powder flow and specimens distance. The coating surface morphology was examined by scanning electron microscopy (JEOL LTD, JSM-5410) on two disks for each surface treatment (Fig. 1a, c, e). Chemical and crystallographic analysis carried out by X-ray diffraction (XRD) technique (Philips PW 1710). XRD analysis on TI01, TI60, PG60 materials, regardless of the surface
morphology and roughness, highlighted a chemical composition of a pure Ti. All samples were first cleaned with isopropyl alcohol and then ultrasonically with de-ionised water, and finally dried in a clean room under laminar flow ventilation. All samples were packaged in blisters and sterilized by g-rays in air with a standard dose of 25 kGy. 2.2. Cell culture Human osteoblast-like cells (MG63) were cultivated in DMEM containing 10% FCS, 100 IU/ml penicillin and 100 mg/ml streptomycin solution at 37 1C in an
(a)
(b)
(c)
(d)
(e)
(f)
Fig. 1. Scanning electron micrographs of all tested materials before (a, c, e) and after (b, d, f) cell culture at a magnification of 150 (a, c, e—scale bar: 200 mm) and of 500 (b, d, f—scale bar: 10 mm): (a and b) TI01; (c and d) TI60; (e and f) PG60.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
atmosphere of 5% CO2 and 95% humidity. When cells were confluent, they were released with 0.05% (w/v) trypsin and 0.02% (w/v) EDTA, counted (Coulter Counter Z1, Beckman Coulter Inc., Miami, FL,USA) and used at a concentration of 0.5 104 cells/ml for the experiment. Ten specimens from each group were placed in 24-well plates and a cell suspension (0.5 104 cells in 100 ml) was directly seeded over each sample. The same amount of osteoblast-like cells was plated into the same number of empty polystyrene wells as controls (polystyrene). After a 2-h incubation to allow cell adhesion to the substrate, 900 ml of DMEM (supplemented with 50 mg/ml Ascorbic acid and b-Glycerophosphate 108 M) was added to 10 wells for each material. Cultures were maintained under the same conditions as described above for two experimental times: four samples of each material were incubated for 3 days and the other six for 7 days, when cells reached confluence on polystyrene, and the medium was replaced with fresh medium on the third day. After 3 and 7 days the supernatant was collected from four wells and centrifuged to remove particulates, if any. Aliquots of the supernatant were dispensed in Eppendorf tubes for storage at 80 1C and assayed for type I Procollagen (CICP, Prolagen-C enzyme Immunoassay kit, Metra Biosystem, CA, USA), Interleukin-6 (IL-6, Human IL-6 Immunoassay kit, Biosource International, CA, USA) and Transforming Growth Factor b1 (TGFb1, Quantikine human TGF-b1, Immunoassay, R&D Systems, MN, USA), Alkaline Phosphatase activity (ALP, Sigma Kinetic method kit, St. Louis, MO, USA), Nitric Oxide (NO, Sigma colorimetric assay, St. Louis, MO, USA), Osteocalcin (OC, Novocalcin enzyme Immunoassay kit, Metra Biosystem, CA, USA), and Lactate Dehydrogenase (LD-L, Sigma Diagnostics, St. Louis, MO, USA). The WST-1 test (Cell Proliferation Reagent WST-1, Roche, IN, USA) was carried out on the cells cultivated on four samples of each group for each experimental time to assess cell proliferation and viability: 100 ml of WST-1 solution and 900 ml of medium (final dilution: 1:10) were added to the cell monolayers, and the multiwell plates were incubated at 37 1C for a further 4 h. Supernatants were quantified spectrophotometrically at 450 nm with a reference wavelength of 640 nm. Results are reported as optical density (OD). At the end of the experimental time, 7 days, the last two specimens of each material were processed for analysis with a scanning electron microscope (SEM): cells grown on the material were fixed in 2.5% glutaraldehyde in a pH 7.4, 0.01 M phosphate buffer for 1 h, rinsed in the buffer and finally dehydrated in a graded series of ethanol. The samples were then washed four times in liquid CO2 and submitted to the critical point dehydration in CO2 atmosphere (80 ATM at 33 1C). Finally, the specimens were mounted on
4951
aluminium stubs for SEM analysis, earthed with carbon tape and then coated with a thin layer (25 A˚) of Au–Pd. The specimens were examined at 5 kV with a Jeol 840A SEM and the images were digitalized. 2.3. Statistical analysis Statistical analysis was performed using the SPSS v.10.1 software (SPSS Inc., Chicago, IL, USA). Data are reported as mean7SD at a significance level of po0.05. After testing data for normal distribution, homogeneity of variance, correlations among the parameters, a multivariate two-way (material: TI01, TI60, PG60, polystyrene; and experimental time: 3 and 7 days) ANOVA test was performed to assess significant interaction effects between biochemical data and selected factors. When such interactions were found, a univariate ANOVA was performed to investigate the effects of the factors on the data by means of hypotheses expressed as linear matrix according to the SPSS syntax. When no interaction was found, a one-way ANOVA (for each factor that showed a significant effect) was performed. The analysis was performed in two steps by: (A) comparing all the tested materials versus polystyrene; and (B) multiple comparisons among TI01, TI60 and PG60.
3. Results Table 2 reports the results for MG63 cultures in contact with the tested materials for 3 and 7 days. The statistical analysis, performed by considering the interaction of ‘material’ and ‘experimental time’ factor effects on cell viability (WST-1) and synthetic activity results of MG63 cultures, showed significant interactions for WST-1 (F ¼ 78:68; po0.0005), OC (F ¼ 4:27; po0.05), TGF-b1 (F ¼ 16:12; po0.0005) and IL-6 (F ¼ 19:73; po0.0005). On the contrary, no significant interactions were seen for ALP, CICP and NO parameters. When the effect of each factor was analysed independently for these last parameters, significant effects of both the ‘material’ and ‘experimental time’ factors were found: ALP (material: F ¼ 5:98; po0.005; experimental time: F ¼ 92:32; po0.0005); CICP (material: F ¼ 11:92; po0.0005; experimental time: F ¼ 224:61; po0.0005) and NO (material: F ¼ 3:98; po0.05; experimental time: F ¼ 68:80; po0.0005). Since experimental time effect was highly significant (po0.0005) in all tested parameters, no further considerations were made on the differences between experimental times, and only the one-way ANOVA by considering the factor ‘material’ was performed for ALP, CICP and NO. Finally, LDH activity was not detected in the supernatant of all tested materials.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
4952
Table 2 Cell viability, synthetic activity and local factors results for the MG63 cultures in contact with tested materials for 3 and 7 days (mean7SD, n ¼ 4) TI01 WST1 (OD) ALP (UI/L) OC (ng/ml) CICP (ng/ml) TGF-beta1(pg/ml) IL-6 (pg/ml) NO (mM)
3 7 3 7 3 7 3 7 3 7 3 7 3 7
days days days days days days days days days days days days days days
1.22870.016 1.71370.027 13.5071.13 10.8970.46 10.2070.50* 9.0070.40* 27.8075.96 66.9074.50 464738*,a 504723 2073* 209755*,a 2973 17.871.4
TI60 1.24970.025 1.95870.030*,b 11.6370.87 8.4770.82 10.4070.40 7.4070.50*,a 21.3075.40 57.6075.00 572725 419729*,a 2173* 128729* 2672 15.171.1
PG60
Polystyrene a
1.27470.004 1.72470.010 14.0971.21 9.6270.59 10.0070.30* 9.0070.40* 20.7075.50 44.9078.40 651760 497731 1772* 127728* 2772 17.971.0
1.25670.017 1.74570.006 13.4672.04 8.7170.91 11.0671.40 10.0070.30 34.3076.10 62.9076.80 62778 53879 471 871 20.7776.45 16.5372.81
Univariate ANOVA test between tested materials for each experimental times according to the established comparisons ((A) all the tested materials versus Polystyrene; (B) multiple comparisons among TI01, TI60 and PG60). WST-1 : (A) 7 days: *, TI60 versus Polystyrene (po0.0005); (B) 3 days: a, PG60 Versus TI01 (po0.005); 7 days: b, TI60 versus TI01 and PG60 (po0.0005). OC: (A) 3 days: *, TI01 and PG60 versus Polystyrene (po0.05);7 days: *, all tested materials versus Polystyrene (po0.05); (B)7 days: a, TI60 versus TI01 and PG60 (po0.0005). TGF-b1: (A) 3 days: *, TI01 versus Polystyrene (po0.0005); 7 days: *, TI60 versus Polystyrene (po0.0005). (B) 3 days: a, TI01 versus TI60 and PG60 (po0.05); 7 days: a, TI60 versus TI01 (po0.001) and PG60 (po0.005). IL-6: (A) 3 days: *, all tested materials versus Polystyrene (po0.0005); 7 days: *, all tested materials versus Polystyrene (po0.001). (B) 7 days: a, TI01 versus TI60 and PG60 (po0.05).
3.1. Cell proliferation and viability All tested materials behaved similarly in comparison with polystyrene; significantly higher results in WST-1 were seen for PG60 (4%, po0.005) in comparison with TI01 at 3 days. At 7 days, TI60 showed significantly higher WST-1 values (12%, po0.0005) in comparison with all the other surfaces.
3.2. Synthetic activity The trend of ALP activity was similar for all tested materials, having lower values for TI60 than TI01 (po0.0005) and PG60 (po0.005) without significant differences of all tested materials when compared with polystyrene. The levels of OC detected in the supernatant of all tested materials were significantly (po0.05) lower than that of polystyrene at both experimental times, except for TI60 at 3 days (ns). While OC levels measured on the different materials were similar at 3 days, TI60 was significantly (po0.0005) lower in comparison with TI01 and PG60 at 7 days. The CICP synthesis reduced with increasing roughness at both experimental times and the lowest data was seen for PG60, the ultra-high roughness surface, which was significantly different from TI01 (po0.0005) and polystyrene (po0.0005). In addition, TI60 showed
significantly lower CICP values (po0.05) in comparison with polystyrene. 3.3. Local factors As for TGF-b1, but only at 3 days, the significantly lowest value was found for TI01. At 7 days, TGF-b1levels of TI60, the high roughness surface, was significantly lower than PG60 (po0.005) and TI01 (po0.001). TGF-b1-levels of TI01 at 3 days and of TI60 at 7 days were significantly lower than those of polystyrene (po0.0005). All tested materials showed significantly higher IL-6 levels than those of polystyrene at both experimental times. While the levels of IL-6 assayed on the different materials were similar at 3 days, significant highest IL-6 synthesis was found for TI01 (po0.05) in comparison with TI60 and PG60 at 7 days. Finally, NO activity of all tested materials decreased from 3 to 7 days. When all data were considered together, all materials showed higher NO activity compared with polystyrene, and TI01 was significantly (po0.05) higher in comparison with TI60 and polystyrene. 3.4. Cell morphology After 7 days of culture, the MG63 cells seeded on the samples were well attached, in close contact with the
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
surface, and they looked rounded with no cytoskeletal processes (Fig. 1b, d, f). In particular, on TI01 and TI60 (Fig. 1b,d), the cell size (20–80 mm) was not comparable with the dimensions of the peaks and depths of the materials (porosity); on TI01 (Fig. 1b) cells were placed on the top (and probably on the bottom) of the projections of the material, because the contact area was very narrow, while on TI60 (Fig. 1d), where the contact area was larger, cells were also positioned along the side of the protrusions. On PG60 (Fig. 1f) the contact area was much wider and cells were located on the surface as if it was smooth.
4. Discussion The study evaluated and compared the biological response of MG63 osteoblast-like cells to Ti surfaces with different roughness levels (Ra 18C73 mm). To the authors’ knowledge, information about in vitro behaviour of materials with so high Ra values is limited. The ultra-high roughness surface PG60, is a novel coating whose roughness has been increased three times in comparison to TI01, which has been used for more than 15 years, but with a porosity level reduced to about 5%, thus giving a dense and potentially stronger coating. The high roughness coating, TI60, showed an increase in roughness parameters of nearly double those of TI01, while maintaining the porosity at about the same level of TI01, and has been recently used for prosthesis coatings. The present findings only partially support the hypothesis that surface roughness influences the proliferation and phenotypic expression of MG63 osteoblast-like cells, in agreement with the results reported in the literature [12], even though some considerations have to be made in order to understand the paradoxical behaviour of ultra-high roughness and dense coating PG60. Rough surfaces have been reported to produce better in vivo osteointegration than smoother surfaces, suggesting that surface modulates the bone response in terms of osteoblast adhesion, proliferation, differentiation and extra-cellular matrix deposition and calcification [23]. However, no consensus has been reached about the optimal surface roughness level of bone implants or whether there is a limit to improving roughness. In fact, in vivo studies by Vercaigne et al. on different high surface roughnesses (Ra 17C38 mm) showed that increases in Ra did not result in improvement of bone response, supposing that the increase of roughness leads to a higher Ti ion release with a negative effect on osteogenesis [28]. Conversely, more uniform results have been achieved with in vitro studies using MG63 cells as a model of relatively immature osteoblasts. Several authors reported that MG63 cells in contact with Ti and Ti alloys (Ra 0.6C7 mm) exhibit roughness-dependent increased
4953
expression of osteoblastic phenotype, as shown by reduced cell proliferation and increased ALP activity, and OC production as well as that of TGF-b1 and PGE2 [9,10,20,22,23]. Moreover, it has been found that surface topography positively influences the response of MG63 cells to regulatory stimuli such as 1,25(OH)2D3 above all on cell proliferation, on smooth Ti surfaces, and that OC production increases with rough Ti surfaces [9,12,19–21]. Finally, the observations by Schwartz et al. [21] suggested that the roughness-dependent regulation of osteoblast proliferation, differentiation and local factor production is related to the activation of integrin receptors by substrate, thus regulating Phosphokinase C (PKC) and A (PKA) through Pohspholipase C (PLC) and A2 (PLA2) pathway [29]. Although it is usually reported that surface roughness reduces cell proliferation [18], current results indicated that MG63 viability was not roughness-dependent; especially at 7 days. Proliferation was good on all the tested surfaces as seen by WST-1 levels that were not different in comparison to controls, and just the high roughness and porous TI60 coating showed the highest value. Previous investigations into PG60 revealed that surface morphology had a marked effect on osteoblast behaviour, by determining morphologic irregularities and reducing adhesion [27]. This phenomenon may be explained by considering the theory of Curtis and Wilkinson, in which the width and depth of the grooves on a surface, as well as the number of adjacent grooves, are considered the determining factors in establishing a positive reaction and orientation of the cells on substrates [30–32]. However, when the grooves or ridges are wider than the cell, the effect on orientation is not very marked [31]. The peaks and depths of the grooves of the tested coatings become much wider than the dimensions of the cell with increasing roughness, from TI01 to PG60, as observed by the metallographic analysis results in the previous study [27]. MG63 cells might have sensed the topography as smooth, as confirmed by the cell morphology results (Fig. 1), thus determining a different behaviour in terms of cell proliferation, that was not rough dependent. In particular, the big spherical Ti particles of PG60 could not have been recognized as discontinuities by MG63 cells, which ‘‘understood’’ the surface as smooth as polystyrene with a huge widening of the contact surface available for adhesion [27]. With regards to the experimental times of 3 and 7 days, they were chosen respectively, since they were believed to represent the minimum culture times to reach confluence on the polystyrene wells (3 days) and on the tested materials (7 days) at the seeding cell concentration of 0.5 104 cells/ml. However, cells did not reach the confluence on the tested materials at 7 days, because of the widening of the contact area in
ARTICLE IN PRESS 4954
V. Borsari et al. / Biomaterials 26 (2005) 4948–4955
comparison to smooth polystyrene. In the current authors’ opinion, further investigations with longer experimental times such as 14 and 21 days could provide useful information about the condition that cell confluence was reached on high and ultra-high roughness coatings. The data concerning early differentiation (ALP), at both experimental times, and late differentiation (OC), at 7 days, are inversely related to cell viability, independently of roughness. In fact, from medium to high roughness level there is a reduction in cell differentiation, on the contrary to what is reported in the literature, while from high to ultra high roughness cell differentiation increases. However, the ALP activity results are similar to polystyrene, and just TI60, where cell proliferation was more pronounced, presented the lowest values. The decrease in OC levels measured on the tested surface in comparison to polystyrene could be explained by the fact that cells were not confluent on the materials after 7 days of culture, even though Schmidt et al. [33] observed a progressive reduction in OC synthesis over time on Ti coated materials. Procollagen type I synthesis results were inversely related to cell proliferation and to the roughness level, as found by Fini et al. [18]: low roughness promoted CICP production while high roughness had a negative influence. Some authors have demonstrated that the suppression of procollagen a1[I] gene expression followed by reduced type I collagen synthesis, is produced by phagocytosis of particulate debris by cells, that would lead to the activation of nuclear transcription factor-kappaB (NF-kB) and the subsequent upregulation of the release of proinflammatory cytokines (i.e. IL6) and PGE2 [3,34,35]. Surfaces with such a high roughness level and discontinuities could release particulate debris in vivo and this hypothesis could be also verified by histological analysis, but the present results indicate that this release occurred also in vitro. In fact, the IL-6 results increased from 3 to 7 days of culture and were higher on all tested surface coatings than polystyrene at both experimental times. The parallel increase in IL-6 synthesis and the reduction in CICP levels would support this hypothesis, thus confirming the observations of other authors on particle debris effect on MG63 functions [3,34]. It has been reported that greater surface roughness enhances the production and secretion of TGF-b1 and some authors suggested that the contact of the cells to material surface could stimulate a mechanism of control of cell functions and differentiation [10,18,19]. In particular, Batzer et al. [19] found that greater surface roughness enhances production and secretion of autocrine and paracrine mediators such as PGE2 in vivo, thus affecting bone formation distal to the implant and mediating the effect of the implant surface on the cells. The same increase in PGE2 was found with the present
tested materials at 3 days, while the highest PGE2 production was seen in TI01 and PG60 at 7 days (data not shown). The present findings differed from those reported in the literature because, even though the increase in TGF-b1 production seems to be proportional to the roughness level of the tested materials, TGF-b1 levels are similar to polystyrene (TI60 and PG60) or lower (TI01). Regarding PG60, the synthesis of TGF-b1 was not influenced by roughness because cells might not recognize such a high roughness level, as previously observed for cell proliferation. TGF-b1 results of TI01 and TI60 do not agree with those in the literature probably because the relationship between roughness and cell response observed by other authors is not appropriate when roughness values are so high. Also concerning NO synthesis, which increased by 16% on average on the tested materials when compared to polystyrene, two hypotheses can be made: a particledependent mechanism and a PGE2-dependent one, that would increase the release of NO and, as a consequence, lead to the suppression of osteoblast functions and the in vivo localized bone destruction around implants [36].
5. Conclusions The present in vitro evaluation performed on MG63 cells confirmed the biocompatibility of the tested coatings (TI01, TI60 and PG60), as expected, and also verified their effect on cell functions, which might be useful for future in vivo investigations into osteointegration. The new VPS ultra-high roughness and dense coating PG60 provided a good biological response, even though, at least in vitro, it behaved similarly to the coatings already used in orthopaedics. The effect of the coating stability and ultra-high roughness level after surgical implantation and during dynamic bone healing and remodelling has yet to be established [37]. Finally, the present findings highlighted that when the roughness level is greater than the cell dimensions it does not enhance cell response.
References [1] Yao J, Glant TT, Lark MW, Mikecz K, Jacobs JJ, Hutchinson NI, Hoerrner LA, Kuettner KE, Galante JO. The potential role of fibroblasts in periprosthetic osteolysis: fibroblast response to titanium particles. J Bone Miner Res 1995;10(9):1417–27. [2] Martin JY, Schwartz Z, Hummert TW, Schraub DM, Simpson Jr J, Lankfond J, Dean DD, Cochran DL, Boyan BD. Effect of titanium surface roughness on proliferation, differentiation and protein synthesis of human osteoblast-like cells (MG63). J Biomed Mater Res 1995;29:389–401. [3] Dean DD, Schwartz Z, Liu Y, Blanchard CR, Agrawal CM, Mabrey JD, Sylvia VL, Lohmann CH, Boyan BD. The effect of ultra-high molecular weight polyethylene wear debris on MG63 osteosarcoma cells in vitro. J Bone Jt Surg Am 1999;81(4):452–61.
ARTICLE IN PRESS V. Borsari et al. / Biomaterials 26 (2005) 4948–4955 [4] Kurihara N, Civin C, Roodman GD. Osteotropic factor responsiveness of highly purified populations of early and late precursors for human multinucleated cells expressing the osteoclast phenotype. J Bone Miner Res 1991;6(3):257–61. [5] Kurokouchi K, Kambe F, Yasukawa K, Izumi R, Ishiguro N, Iwata H, Seo H. TNF-alpha increases expression of IL-6 and ICAM-1 genes through activation of NF-kappaB in osteoblastlike ROS17/2.8 cells. J Bone Miner Res 1998;13(8):1290–9. [6] Galien R, Evans HF, Garcia T. Involvement of CCAAT/ enhancer-binding protein and nuclear factor-kappa B binding sites in interleukin-6 promoter inhibition by estrogens. Mol Endocrinol 1996;10(6):713–22. [7] Takei H, Pioletti DP, Kwon SY, Sung KL. Combined effect of titanium particles and TNF-alpha on the production of IL-6 by osteoblast-like cells. J Biomed Mater Res 2000;52(2):382–7. [8] Wen X, Wang X, Zhang N. Microrough surface of metallic biomaterials: a literature review. Biomed Mater Eng 1996;6(3):173–89. [9] Boyan BD, Batzer R, Kieswetter K, Liu Y, Cochran DL, Szmuckler-Moncler S, Dean DD, Schwartz Z. Titanium surface roughness alters responsiveness of MG-63 osteoblast-like to 1a,25-(OH)2D3. J Biomed Mater Res 1998;39:77–85. [10] Lincks J, Boyan BD, Blanchard CR, Lohmann CH, Liu Y, Cochran DL, Dean DD, Schwartz Z. Response of MG63 osteoblast-like cells to titanium and titanium alloy is dependent on surface roughness and composition. Biomaterials 1998;19(23):2219–32. [11] Ong JL, Hoppe CA, Cardenas HL, Cavin R, Carnes DL, Sogal A, Raikar GN. Osteoblast precursor cell activity on HA surfaces of different treatments. J Biomed Mater Res 1998;39(2):176–83. [12] Lohmann CH, Bonewald LF, Sisk MA, Sylvia VL, Cochran DL, Dean DD, Boyan BD, Schwartz Z. Maturation state determines the response of osteogenic cells to surface roughness and 1,25dihydroxyvitamin D3. J Bone Miner Res 2000;15(6):1169–80. [13] Deligianni DD, Katsala ND, Koutsoukos PG, Missirlis YF. Effect of surface roughness of hydroxyapatite on human bone marrow cells adhesion, proliferation, differentiation and detachment strength. Biomaterials 2001;22:87–96. [14] Perizzolo D, Lacefield WR, Brunette DM. Interaction between topography and coating in the formation of bone nodules in culture for hydroxyapatite- and titanium-coated micromachined surfaces. J Biomed Mater Res 2001;56:494–503. [15] Korovessis PG, Deligianni DD. Role of surface roughness of titanium versus hydroxyapatite on human bone marrow cells response. J Spinal Disord Tech 2002;15(2):175–83. [16] Montanaro L, Arciola CR, Campoccia D, Cervellati M. In vitro effects on MG63 osteoblast-like cells following contact with two roughness-differing fluorohydroxyapatite-coated titanium alloys. Biomaterials 2002;23:3651–9. [17] Campoccia D, Arciola CR, Cervellati M, Maltarello MC, Montanaro L. In vitro behaviour of bonemarrow derived mesenchymal cells cultured on fluorohydroxyapatite-coated substrata with different roughness. Biomaterials 2003;24:587–96. [18] Fini M, Giardino R, Borsari V, Torricelli P, Rimondini R, Giavaresi G, Nicoli Aldini N. In vitro behaviour of osteoblasts cultured on orthopaedic biomaterials with different roughness, uncoated and flourohydroxyapatite-coated, relative to the in vivo osteointegration rate. Int J Artif Organs 2003;26(6):520–8. [19] Batzer R, Liu Y, Cochran DL, Szmuckler-Moncler S, Dean DD, Boyan BD, Schwartz Z. Prostaglandins mediate the effects of titanium surface roughness on MG63 osteoblast-like cells and alter cell responsiveness to 1 alpha,25-(OH)2D3. J Biomed Mater Res 1998;41(3):489–96. [20] Lohmann CH, Sagun Jr. R, Sylvia VL, Cochran DL, Dean DD, Boyan BD, Schwartz Z. Surface roughness modulates the response of MG63 osteoblast-like cells to 1,25-(OH)(2)D(3) through regulation of phospholipase A(2) activity and activation of protein kinase A. J Biomed Mater Res 1999;47(2):139–51.
4955
[21] Schwartz Z, Lohmann CH, Vocke AK, Sylvia VL, Cochran DL, Dean DD, Boyan BD. Osteoblast response to titanium surface roughness and 1alpha,25-(OH)(2)D(3) is mediated through the mitogen-activated protein kinase (MAPK) pathway. J Biomed Mater Res 2001;56(3):417–26. [22] Orsini G, Assenza B, Scarano A, Piattelli M, Piattelli A. Surface analysis of machined versus sandblasted and acid-etched titanium implants. Int J Oral Maxillofac Implants 2000;15(6):779–84. [23] Boyan BD, Bonewald LF, Paschalis EP, Lohmann CH, Rosser J, Cochran DL, Dean DD, Schwartz Z, Boskey AL. Osteoblastmediated mineral deposition in culture is dependent on surface microtopography. Calcif Tissue Int 2002;71(6):519–29. [24] Brossa F, Cigada A, Chiesa R, Paracchini L, Consonni C. Adhesion properties of plasma sprayed hydroxyapatite coatings for orthopaedic prostheses. Biomed Mater Eng 1993;3:127–36. [25] Winkler-Gniewek W, Stallforth H, Ungethu¨m M, Gruner H. Structure and properties of VPS coatings in medical technology. First Plasma technik symposium, Lucerne/Switzerland, May 18th to 20th 1988. Vol 3, p. 97–102. [26] Salito A, Barbezat G, Filmer H, Hochstrasser J, Nicoll AR. Plasma spray systems for the deposition of materials for biomedical applications. In: Ravaglioli A, Krajewski A, editors. Proceeding of the international congress on bioceramic and the human body. Faenza, Italy, 2–5 April 1991. IRTEC-CNR. [27] Borsari V, Giavaresi G, Fini M, Torricelli P, Salito A, Chiesa R, Chiusoli L, Volpert A, Rimondini L, Giardino R. Physical characterization of different roughness titanium surfaces, with and without hydroxyapatite coating, and their effect on human osteoblast-like cells. J Biomed Mater Res, in press. [28] Vercaigne S, Wolke JC, Naert I, Jansen JA. The effect of titanium plasma sprayed implants on trabecular bone healing in the goat. Biomaterials 1998;19:1093–9. [29] Boyan BD, Sylvia VL, Liu Y, Sagun R, Cochran DL, Lohmann CH, Dean DD, Schwartz Z. Surface roughness mediates its effects on osteoblasts via protein kinase A and phospholipase A2. Biomaterials 1999;20(23–24):2305–10. [30] Anselme K, Bigerelle M, Noe¨l B, Iost A, Hardouin P. Effect of grooved titanium substratum on human osteoblastic cell growth. J Biomed Mater Res 2002;60:529–40. [31] Curtis A, Wilkinson C. Topographical control of cells. Biomaterials 1997;18:1573–83. [32] Curtis A, Clark P. The effect of topographic and mechanical properties of materials on cell behaviour. CRC Rev Biocompat 1990;5:343–62. [33] Schmidt C, Steinbach G, Decking R, Claes LE, Ignatius AA. IL-6 and PGE2 release by human osteoblasts on implant materials. Biomaterials 2003;24(23):4191–6. [34] Vermes C, Roebuck KA, Chandrasekaran R, Dobai JG, Jacobs JJ, Glant TT. Particulate wear debris activates protein tyrosine kinases and nuclear factor kappaB, which down-regulates type I collagen synthesis in human osteoblasts. J Bone Miner Res 2000;15(9):1756–65. [35] Nakashima Y, Sun DH, Trindade MC, Maloney WJ, Goodman SB, Schurman DJ, Smith RL. Signaling pathways for tumor necrosis factor-alpha and interleukin-6 expression in human macrophages exposed to titanium-alloy particulate debris in vitro. J Bone Jt Surg Am 1999;81(5):603–15. [36] Lohmann CH, Dean DD, Bonewald LF, Schwartz Z, Boyan BD. Nitric oxide and prostaglandin E2 production in response to ultra-high molecular weight polyethylene particles depends on osteoblast maturation state. J Bone Jt Surg Am 2002;84A(3):411–9. [37] Fini M, Giardino R. In vitro and in vivo tests for the biological evaluation of candidate orthopaedic materials: benefits and limits. J Appl Biomater Biomech 2003;1:155–63.