Effect of roughness, wettability and morphology of engineered titanium surfaces on osteoblast-like cell adhesion

Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

Contents lists available at ScienceDirect

Colloids and Surfaces A: Physicochemical and Engineering Aspects journal homepage: www.elsevier.com/locate/colsurfa

Effect of roughness, wettability and morphology of engineered titanium surfaces on osteoblast-like cell adhesion J.I. Rosales-Leal a , M.A. Rodríguez-Valverde b,∗ , G. Mazzaglia a , P.J. Ramón-Torregrosa b , L. Díaz-Rodríguez c , O. García-Martínez c , M. Vallecillo-Capilla a , C. Ruiz c , M.A. Cabrerizo-Vílchez b a b c

Dental Materials, Faculty of Odontology, University of Granada, Campus de Cartuja, E-18071 Granada, Spain Biocolloid and Fluid Physics Group, Department of Applied Physics, University of Granada, Campus de Fuentenueva, E-18071 Granada, Spain Department of Nursing, Physiology Section, School of Health Sciences, University of Granada, Avda. Madrid, s/n E-18071 Granada, Spain

a r t i c l e

i n f o

Article history: Received 26 September 2009 Received in revised form 3 December 2009 Accepted 7 December 2009 Available online 16 December 2009 Keywords: Titanium Surface treatment Topography Wettability Cell culture

a b s t r a c t Texturization of surfaces is usually advantageous in biomaterial engineering. However, the details of the textured surfaces can be more determining on cell adhesion and proliferation, rather than their roughness degree. Titanium is extensively used as a dental implant material in the human body. In this paper, the effect of four surface treatments on commercially pure titanium has been evaluated. These treatments were polishing (pTi); hydrofluoric acid (HF) etching (eTi); Al2 O3 blasting (bTi); Al2 O3 blasting + HF etching (beTi). Roughness and fractal dimensions were obtained from atomic force microscopy. Wettability was measured using water sessile drops. Morphology and surface chemical composition were analyzed with scanning electron microscopy and energy dispersive X-ray (EDX). MG-63 cell cultures were performed at different times (180 min, 24 h, 48 h, 72 h). Lowest roughness was found in pTi samples followed by eTi, bTi and beTi samples. Etching generated surfaces with the highest fractal dimension and negative skewness. Young contact angles were similar except for pTi and bTi surfaces. Silicon and aluminum traces were found in pTi and bTi samples, respectively. Cell adhesion (≤24 h) was greater on bTi and beTi surfaces. After 48 h, cell proliferation, mediated by specific morphologies, was improved in eTi samples followed by beTi surfaces. For the same surface chemistry, cell growth was driven by topography features. © 2009 Elsevier B.V. All rights reserved.

1. Introduction The success of a dental implant is based on the osseointegration that is defined as the direct contact between the bone tissue and the dental implant surface, without fibrous tissue growing at the interface [1,2]. Surface characteristics of the implant play an important role for the evolution of bone tissue of the recipient site, after implantation [3]. Surface properties control the amount and quality of cells adhered on the implant and consequently, the tissue growth. Most determining surface properties for cell adhesion are surface topography and surface chemical reactivity. Accordingly, surface engineering of biomaterials is oriented to modify their surface texture and/or surface chemistry. The topography of a surface substantially affects the macroscopic behaviour of a material [4]. Currently, the influence of surface topography on biological response is a matter of investigation. At cellular level, biological responses, such as the orientation

∗ Corresponding author. Tel.: +34 958 24 00 25; fax: +34 958 24 32 14. E-mail address: [email protected] (M.A. Rodríguez-Valverde). 0927-7757/$ – see front matter © 2009 Elsevier B.V. All rights reserved. doi:10.1016/j.colsurfa.2009.12.017

and migration of cells and the cellular production of organized cytoskeletal arrangements, are directly influenced by the surface topography [5]. There are evidences that a suitable surface roughness, at nano- and microscopic scale, can lead to a successful osseointegration of titanium implants [6]. Osteoblast differentiation, proliferation and matrix production [7] as well as the production of local growth factors and cytokines are affected by surface roughness [8]. Biomolecule adsorption onto implant surfaces “in vivo” is indeed a dynamic process driven by the physico-chemical interactions between adsorbent surface and macromolecule [9]. This precursor process develops a “conditioning film” which will modulate the cellular host response. Surface energy, which is intimately related to wettability [10], is a useful quantity that has often correlated strongly with biological interaction. Hence, implant wettability can become determining for the protein adsorption and consequently, for the cell adhesion [11–13]. For instance, hydrophobic surfaces (i.e. surfaces with low water wettability) presumably decelerate primary interactions with the aqueous biosystem. Thus, it is usually reported that biomaterial surfaces with moderate hydrophilicity, improved cell growth and higher

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

biocompatibility [14]. However, cell adhesion can decrease as the implant wettability is further decreased. This points out to the existence of a range of optimal surface energies [15]. Otherwise, interfacial reactions “in vivo” change relevant physical and chemical surface parameters, such as the surface energy, affecting the long-term stability of implants [16]. Modification of the physico- and physico-chemical surface properties of a biomaterial can improve interaction with cells. In particular, several surface treatments have been applied to optimize the surface topography of titanium implants in bone-contact applications. Most extending treatments are sandblasting and acid etching [17]. The main purpose of these texturization treatments is to achieve greater bone-to-implant contact [18], in order to reduce healing times and accelerate integration into the host tissue. Texturing of dental implants improves the mechanical adhesion to bone but, at the same time, the asperities and grooves may act as preferable sites for protein adsorption. The evaluation of the topography of biomaterial surfaces is important because it usually improves the biological responses found during osteointegration and the long-term response of the bone–implant interface. However, one major difficulty arises in the choice of the topography parameters, which are actually relevant for the interactions between biomaterial surface and surrounding biological medium [19]. Generally, the description of a surface may be performed at three levels, according to the degree of surface information: height/spatial distribution, topology and morphology. The height/spatial distribution provides a statistical description of the surface roughness using amplitude or horizontal lengths and hybrid parameters, both in 2D and 3D analysis [20]. The topology description of a surface is understood as the set of intrinsic properties related to its structure such as connectedness, compactness. The topography of most engineering surfaces is topologically self-affine over a range of scales [21]. The fractal dimension (Df ) measures, in statistical sense, the structural complexity of a selfaffine surface, i.e. the random roughness organization [22]. In other adhesion studies, the fractal dimension correlated better with adhesion than did conventional measurements of surface roughness [23,24]. For this reason, fractal analysis might be more helpful than conventional roughness descriptions in order to elucidate the complex mechanisms occurring at the implant surface in contact with the surrounding biological tissues [19]. Otherwise, the morphological analysis, usually performed by SEM, provides the details and the figure of the surface topography at higher resolution; even although this raw information without post processing is merely qualitative. In surface engineering of biomaterials, it is important to determine when cell spreading is modulated by surface energy, topography or both, at short and long-term times after implantation. In addition, rather than conventional approaches, fractal dimension might be used to quantify the role of the features of implant topography on the cell growth. Accordingly, the aim of this work is to evaluate the effect of four texturization treatments of titanium surfaces on the adhesion and growth of osteoblastlike cells, from the topography (roughness, fractal dimension and morphology) and the water wettability induced by each treatment.

2. Materials and methods 2.1. Titanium samples preparation Commercially pure ASTM grade II titanium (cpTi) cylinders (Manfredi, S. Secondo di Pinerolo, Italy) were suitably cut into small disks of approximately 12 mm in diameter and 2 mm in thickness. The cpTi disks, previously cleaned with distilled water, were engineered as follows:

223

- Group 1 (control): Polished titanium (pTi). cpTi surfaces were metallographically polished using silicon carbide (SiC) papers successively from grade 240, 320, 500, 800, 1200, 2000 to 4000 grit. Next, an ultra-polishing was achieved using the sequence 1–0.3–0.05 ␮m alumina slurries and gauze. The polished disks were cleaned in distilled water by immersion in ultrasonic bath (Selecta, Barcelona, Spain). - Group 2: Etched titanium (eTi). pTi surfaces were etched with a solution of 10% (v/v) hydrofluoric (HF) acid (Panreac, Barcelona, Spain) by complete immersion upon gently agitation for 5 min. - Group 3: Blasted titanium (bTi). pTi surfaces were blasted by alumina (99.78% Al2 O3 ) particles of 110 ␮m diameter projected at an incidence angle of 60◦ and 0.25 MPa pressure for 3 min. - Group 4: Blasted + etched titanium (beTi). pTi surfaces were blasted as in the group 3 but using alumina particles of 250–500 ␮m. Next, these blasted surfaces were etched as described for group 2. After treatments, pTi, eTi, bTi and beTi samples were degreased with a solution of 70% (v/v) acetone (Panreac, Barcelona, Spain) by immersion in ultrasonic bath for 20 min and afterwards, the samples were ultrasonicated in distilled water for 30 min. eTi and beTi samples were passivated with 30% (v/v) nitric (HNO3 ) acid (Panreac, Barcelona, Spain) for 3 min and then again ultrasonicated in diluted acetone for 20 min and distilled water for 30 min. All disks were autoclaved at 121 ◦ C and 1 atm for 30 min after preparation. 2.2. Atomic force microscopy Topographies of the textured titanium surfaces were acquired by an Atomic Force Microscope Nanoscope IV MultiMode in air (Digital Instruments, Santa Barbara, Ca, USA). The microscope was operated in tapping mode on a 2500 ␮m2 scansize with a Si3 N4 V-shape cantilever (stiffness, k = 63 ± 8 N/m). The topography data were sampled in a grid of 256 × 256 points. All images were fitted to a plane and filtered by a Gaussian mask with a cutoff of 2.5 ␮m in order to capture surface features below the average cell size (∼10 ␮m). Three disks were used per group and three topographies were acquired per each disk. 2.2.1. Roughness Usual roughness parameters [20] are described in Table 1. Amplitude parameters, as the arithmetic average roughness (Ra ), the peak roughness (Rp maximum relative height), the valley roughness (Rv , maximum relative depth) and the absolute height (Rmax = Rp + Rv ), usually provide a crude roughness description because they just reveal the amplitude of the topography features without information about their spatial distribution. Instead, the root mean square roughness (Rq ), the skewness (Rs ) and the kurtosis (Rk ) are statistical moments of the height distribution that describe its width, symmetry and form, respectively. Commonly, the root mean square roughness qualitatively describes the same trend than the arithmetic average roughness. Otherwise, hybrid parameters, as the surface area ratio (Rw ), are derivative-based parameters that quantify topography variations with respect to the planar coordinates. The above-mentioned parameters were estimated from the topography data using the software of the own instrument. Although the results of topographical surface characterization depend on the length scale employed [25], 50 ␮m scansizes were selected in order to cover the cell size range (∼10 ␮m). More details about the computation of surface area ratio are described elsewhere [26]. The surface area excess with respect to the pTi surface was also calculated.

224

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

Table 1 Roughness parameters. Parameters

Description a

Rp Rv

Maximum relative height (the highest peak) Maximum relative depth (the deepest valley)

Ra

Arithmetic average roughness

Rq

Root mean square roughness

Mathematical definition

Typical Units

Zmax Zmin

␮m ␮m

   b Ra = Z  xy Rq =

Z2

nm nm

xy

Rmax

Absolute height

Rmax = Rp + Rv

␮m

Rsk

Symmetry of the distribution of heights (skewness)

Rsk =

–

Rku

Form of the distribution of heights (biased kurtosis)

Rw

Surface area ratio (Wenzel factor)

a



With respect to an ideal average plane Z

Lx b

·xy =

1 Lx Ly

Ly dx

0

 3   2 3/2 Z / Z  4 xy  2 xy2 Rku = Z / Z −3 xy   xy    2 2 A dZ dZ Rw ≡

Az

=

1+

dx

+

– –

dy xy

xy

= 0.

(·) dy. 0

2.2.2. Fractal analysis Although recently there has been much effort devoted to the development and manufacturing of nanotextured or structured surfaces [27], the conventional texturization processes provide randomly rough surfaces. Furthermore, these rough surfaces are usually self-affine, i.e. there is an anisotropic disorder because the surface remains statistically invariant under dilations of the x, y, z coordinates but by different scale factors [21]. Fractal dimension provides a scale-independent measure of these strongly disordered systems, regardless of the observation scale. A fractal surface typically has irregularities that “fill” the embedding space (D = 3). Fractal dimension quantifies the disorder in terms of the space-filling ability of the surface, thus a fractal surface must occupy intrinsically more space than a plane surface (i.e. the topological space, D = 2). In fact, a fractal surface usually becomes space-filling but non-uniformly because it occupies only certain regions of the embedding space. Hence, fractal dimension will be related to the compactness degree of the space-filling structure [21]. In this work, fractal dimensions were computed by the wellknown box-counting method [19,28,29]. In this method, the surface topography is covered with boxes of side length d. Once the surface (i.e. the area) is completely covered with N boxes, the following scaling rule of fractal geometry [21] must be fulfilled: N(d) = ˛d−Df

(1)

where ˛ is a geometrical prefactor and Df is the fractal dimension. By continuously changing the magnification scale through changing the size of the boxes (d), the number of boxes covering the surface (N) is counted. The fractal dimension Df is obtained from the slope of the log–log plot of Eq. (1). 2.3. Field-emission scanning electron microscopy The tree titanium disks per group of the Section 2.1 were observed under a Scanning Electron Microscope Leo Gemini 1530 (Carl Zeiss, Oberkochen, Germany). In order to improve the electrical conductivity of titanium, carbon thin films were produced by sputtering on each sample. 2.3.1. Morphology Morphology of each sample was evaluated from 3072 × 2304 pixel images acquired at ×2000 magnification. Further porosity analysis was accomplished on the eTi and beTi samples.

2.3.2. Surface chemical analysis Roughness is often associated with changes in surface chemistry (e.g. plasma deposition). Hence, the texturization treatments of unalloyed titanium described in Section 2.1 can produce unexpected modifications in its surface chemical composition. Accordingly, energy dispersive X-ray (EDX) spectra were obtained at different surface zones of each titanium sample up to a depth of 1 ␮m. This analytical technique serves to detect the chemical elements located at the surface [30]. 2.4. Wettability A liquid drop spreads on a solid surface up to cover a particular area, in prejudice of the surrounding liquid vapour, driven by the solid–liquid intermolecular interactions. This interfacial phenomenon is known as wetting and the affinity of a solid surface to be wetted by a given liquid is referred to as wettability. Wetting, and thus wettability, is strongly influenced by surface roughness [31,32]. Bathomarco et al. found that as the surface area of titanium implants increases, the measured contact angle decreases [32]. In 1936, Wenzel [33] reported that the roughness of a homogenous solid surface affects contact angle measurements (referred to as apparent angles) as follows: cos app = Rw cos Y

(2)

where Rw symbolizes the surface area ratio, also referred to as Wenzel factor,  app is the apparent contact angle associated to the system equilibrium state and  Y is the Young contact angle [31]. The Wenzel law (2) is valid if the characteristic length of asperities is much lower than the size of drop or the distribution of asperities is assumed uniform. This law predicts an increase in the experimental contact angles with growing roughness for  Y > 90◦ and just the opposite trend if  Y < 90◦ . The thermodynamically meaningful angle, i.e. the angle intimately related to surface free energy, is the Young contact angle. This angle would be the equilibrium contact angle of the unattainable smooth surface (i.e. ideal surface). As pointed out by Morra et al. [34], in order to estimate surface free energy from experimentally accessible contact angles on rough surfaces, the concerning Young contact angles must be estimated through Eq. (2). However, the apparent contact angle associated to the true equilibrium state is hardly measurable due to contact angle hysteresis [16,30,35]. Instead, the employment of advancing contact angles for the solid surface free energy determination is a common practice [10], if practically no water film is present behind the liquid drop. Furthermore, Rupp et al. [6] reported val-

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

ues of water receding angles equal to zero for acid-etched titanium surfaces. The main, commonly used, measure of surface chemistry is the surface free energy. However, the surface energy of solid may not be measured directly, at least with relative ease. Instead, there are different approaches for the estimation of surface energy from contact angle measurements (e.g. critical surface tension, the equation of state, acid–base components of solid surfaces, . . .) [31]. Nevertheless, aside from controversies over the validity of these methods, this estimation is an unnecessary complication for practical biomaterial applications. Thereby, since the difference between solid interfacial energies is indeed the thermodynamically relevant quantity, the Young contact angle becomes more meaningful than questionable surface energy values [10]. Since physiological fluids are indeed aqueous saline solutions, pure water is the reference probe liquid used for measuring surface wettability of biomaterials. Water wettability of titanium surfaces was measured through the contact angle of spreading sessile drops, which can be considered as the advancing contact angle, as above discussed. MilliQ water drops (2.5 ␮l of volume) were dispensed using a micropipette (Eppendorf, Hamburg, Germany) at room temperature. Three disks were analyzed per group (see Section 2.1) and three drops were deposited on each disk. Once the drop was dispensed on the substrate, side-view images of drop were acquired and analyzed by Axisymmetric Drop Shape Analysis. More details about this technique can be found elsewhere [36]. Contact angles were directly computed by ad-hoc designed software.

225

in a humidified atmosphere of 95% air and 5% CO2 . The cells were detached from the culture flask with a solution of 0.05% trypsin and 0.02% ethylene diamine tetra-acetic acid (EDTA) (Sigma, St. Louis, MO, USA), and were washed and suspended in complete culture medium with 10% FBS. The cells obtained were inoculated for tests onto samples at 2 × 105 cell/ml in a 24-well plate (Falcon, Becton Dickinson Labware, NJ, USA) in the ratio of 2 ml/well, in which were deposited previously titanium disks. The plates were incubated to 37 ◦ C in CO2 (5%) atmosphere. Assay tests were performed at 180 min, 24, 48 and 72 h for each surface. Cell cultures were repeated four times for each surface and time period. At the end of the culture time, disks were recovered and cleaned with medium solution. Then, cells were detached from the disks with a solution of 0.05% trypsin and 0.02% ethylene diamine tetra-acetic acid (EDTA) (Sigma, St. Louis, MO, USA), and were washed and suspended in complete culture medium with 10% FCS. The number of adhered cells was determined with a counter cytometer (Ortho Diagnostic System, Raritan, Il, USA). Number of cells was quantified per ml (n·104/ml). One additional sample per treatment and time was prepared in order to evaluate the cell morphology. Cell culture followed the method above described. After the incubation period, media were removed and specimens were fixed with 4% glutaraldehyde in PBS (pH 7.2) for 20 min. After dehydration in graded alcohols, samples were immersed in hexametildisilazane for 10 min, air dried and then sputter-coated with gold palladium. Finally, the surface of the specimens was examined by SEM. 3. Results

2.5. Cell culture 3.1. Topography Osteoblast-like cell MG-63 [7,17] were cultured in Dulbecco’s modified Eagle medium (Invitrogen, Carlsbad, CA, USA) with 100 IU/ml penicillin (Roger, Barcelona, Spain), 50 ␮g/ml gentamicin (Braum Medical, Jaén, Spain), 2.5 ␮g/ml anfotericin B (Sigma, St. Louis, MO, USA), 1% glutamine (Sigma, St. Louis, MO, USA), 2% HEPES (Sigma, St. Louis, MO, USA) and supplemented with 10% fetal bovine serum (FBS) (Gibco, Paisley, UK). Cultures were kept at 37 ◦ C

3.1.1. Roughness In Fig. 1, AFM pictures of titanium disks belonging to the four groups (see Section 2.1) are displayed keeping the same scale and scan size. The roughness parameters of the textured titanium surfaces are compiled in Table 2. According to the amplitude parameters (Rp , Rv , Ra and Rmax ), the coarsest surfaces were the

Fig. 1. AFM pictures (256 × 256 points) of the textured titanium surfaces: pTi (a); eTi (b); bTi (c); beTi (d). The scansize area was fixed to 50 × 50 ␮m2 .

226

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

Table 2 Values of roughness parameters of the four engineered titanium surfaces (polished, etched, blasted and blasted-etched surfaces). Surface

Rp (␮m)

pTi eTi bTi beTi

0.47 1.15 3.0 5.1

± ± ± ±

Rv (␮m) −0.60 −1.60 −1.8 −3.3

0.21 0.23 0.9 1.6

± ± ± ±

Ra (nm) 0.19 0.25 0.7 0.9

140 209 861 1370

± ± ± ±

Rq (nm)

17 19 30 78

149 268 1022 1639

blasted ones with the largest particles (beTi), although they were etched later (see Fig. 1d). In decreasing order of any amplitude parameter, the beTi surfaces were followed by bTi, eTi and pTi (see Fig. 1a). Otherwise, from the statistical parameters, the positive biased kurtosis (Rku > 0) and the negative skewness (Rsk < 0) found on the eTi samples predict a fluctuating morphology (see Fig. 1b) with high density of valleys (i.e. microporosity), whereas sandblasting seems to produce spiky and coarse morphologies with negative biased kurtosis and positive skewness (see Fig. 1c and d). This is also illustrated in Fig. 2 with the topography profiles of each sample. The surface area ratio (see Table 3) just softened the trend of amplitude parameters since surface area is a differential parameter which takes into account variations in planar and perpendicular directions. As expected, the greatest values of available surface area were found in the beTi samples and the lowest ones in the pTi samples. However, eTi and bTi surfaces

± ± ± ±

Rmax (␮m)

42 43 45 58

2.3 3.1 5.7 8.2

± ± ± ±

Rsk −0.07 −0.36 0.40 0.38

0.1 0.1 0.3 0.5

Rku ± ± ± ±

0.23 0.12 0.19 0.11

1.4 0.71 −0.75 −0.31

Rw ± ± ± ±

0.7 0.19 0.25 0.21

1.02 1.10 1.15 1.33

± ± ± ±

0.04 0.11 0.16 0.14

exhibited similar values of surface area. The highly fluctuating morphologies induced by etching (Fig. 2) compensated the coarse morphologies blasted with the smallest particles (bTi). 3.1.2. Fractal dimension The values of fractal dimension computed by box-counting method are presented in Table 3. All samples were recognized as fractal surfaces because Eq. (1) was fulfilled up to a limit box size, where fractal behaviour disappeared. No significant differences between pTi, bTi and beTi groups were found, though this did not mean similar morphologies. However, eTi surfaces obtained the highest fractal dimension due to their highly fluctuant morphology (space-filling ability). In spite of common belief, a fractal surface may be less rough and have more space-filling irregularities, i.e. it occupies only certain regions of the embedding space but more compactly [21]. 3.1.3. Morphology Fig. 3 compiles the SEM micrographs of the four groups. These images qualitatively confirmed the results obtained from AFM topographies. The reference surface (pTi) showed a smooth morphology without characteristic features (see Fig. 3a). Instead, acid etching produced angular stepped morphologies with micropores from 1.5 to 2.5 ␮m (see Fig. 3b). This random structure manifested certain hierarchy with double scale roughness, i.e. a nano-rough texture. This indeed revealed the grain-structured titanium. Otherwise, blasting generated very irregular and coarse morphologies (see Fig. 3c) with wide cavities (10–50 ␮m in diameter), as the size of shot alumina particles. Even it was observed the presence of alumina particles incrusted in the titanium surface, as reported [37]. These surface alterations might significantly hinder the cell adhesion. Finally, the beTi surfaces exhibited a coarse morphology though less chaotic than the bTi samples and with micropores like the eTi samples (see Fig. 3d). 3.2. Surface chemical composition

Fig. 2. AFM profiles of the textured titanium surfaces: pTi, eTi, bTi and beTi. The profiles are intentionally shifted for illustrative purposes only. Sandblasting produced spiky and coarse morphologies, where the coarsest surfaces were the blasted ones with the largest particles (beTi) although they were etched later. However, the highly fluctuating morphologies (space-filling ability) of eTi surfaces explain that eTi and bTi surfaces exhibited similar values of surface area. Table 3 Values of surface area, fractal dimension and contact angle of the four engineered titanium surfaces (polished, etched, blasted and blasted-etched surfaces). Surface

Surface area Real areaa (␮m2 )

pTi eTi bTi beTi

2540 2760 2890 3340

± ± ± ±

100 110 110 120

Fractal dimension Area excessb (%) 0 9±2 14 ± 1 32 ± 2

2.33 2.56 2.26 2.24

± ± ± ±

0.09䊉 0.11 0.13䊉 0.12䊉

Contact angle (◦ )

62 53 46 40

± ± ± ±

4 3 2 2

Mean values with symbol (䊉) were statistically similar (p > 0.05). (One-way ANOVA and Tukey’s test). a Apparent area: 2500 ␮m2 . b With respect to ultrapolished surface.

The EDX spectra of all samples (not shown) pointed out that surface chemistry of the textured surfaces was mostly based on titanium dioxide with a constant percentage of metallic titanium behind the protective oxide film (see the ratios of surface chemical species in Table 4). Silicon traces appeared on the pTi samples due to polishing with SiC papers; just as well the bTi samples contained aluminium traces due to the alumina particles trapped into them (as pointed out in Section 3.1.3). Indeed, the carbon from SiC was overshadowed in the spectra by the residual carbon originating Table 4 Values of Young contact angle and surface chemical species ratio of the four engineered titanium surfaces (polished, etched, blasted and blasted-etched surfaces). Surface

Young contact angle (◦ )

pTi eTi bTi beTi

63 57 50 55

± ± ± ±

6 8 10 7

Ti/TiO2

Al/TiO2

Si/TiO2

0.14 0.13 0.14 0.13

0.16 0.00 0.28 0.03

0.34 0.00 0.00 0.00

Mean values with symbol ( ) were statistically dissimilar (p < 0.05). (One-way ANOVA and Tukey’s test).

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

227

Fig. 3. SEM microimages of the textured titanium surfaces: pTi (a); eTi (b); bTi (c); beTi (d) (original magnification ×2000).

from the reflecting film for SEM experiments. Unlike bTi samples, beTi surfaces showed smaller aluminium amounts because the alumina particles were mostly etched. Likewise, no SiC trace was found on eTi samples due to etching. Hence, whereas acid etching just recovered the original surface chemical composition of cpTi, blasting and polishing produced undesired surface doping. 3.3. Wettability The apparent contact angles of water drops on the textured titanium surfaces are compiled in Table 3. These values, measured once drop spreading was finished (1–2 s), corresponded to system metastates very close to advancing mode because the receding angles reported in literature are practically zero [6]. As expected for hydrophilic surfaces (see Eq. (2)), greater surface area ratio corresponds to higher wettability (lower apparent contact angle). Hence, pTi surfaces showed the lowest wettability followed by eTi, bTi and beTi surfaces. It is worth to highlight that this result can be misleading because the effect of roughness overshadows the influence of interfacial energetics. The influence of roughness on contact angle of homogeneous surfaces is a twofold issue. Phenomenologically, advancing contact angles often increase linearly with the concerning surface texture parameter [38], and just the opposite trend is observed for the receding angles. However, in addition to this effect, a drop (meniscus) in contact with a rough and chemically homogeneous surface undergoes the second-order effect [32] predicted by the Wenzel law (2). Although both effects can appear at once, latter effect typically dominates from intermediate values of contact angle hysteresis. Table 4 collects the values of Young contact angle computed from Eq. (2) and the concerning values of surface area ratio (see Table 3) for each surface treatment. The use of Eq. (2) was justified by the larger drop volume (2.5 ␮l) as regards the maximum roughness amplitude (see Table 2). The Young contact angles were statistically similar, except for pTi and bTi surfaces. The bTi samples exhibited the lowest value of Young contact angle due to the

hydrophilic alumina content whereas the pTi samples were anomalously less hydrophilic due to the carbon excess originating from the SiC traces. 3.4. Cell culture Fig. 4 shows cell culture results. Fig. 5a and b show representative cell morphologies. As expected for biocompatible surfaces, the cell adhesion rate increased on all treated titanium surfaces, though specific morphologies developed by each texturing affected differently the main events of cell adhesion. After 180 min and 24 h, cell contact and attachment were more marked on bTi and beTi surfaces than pTi and eTi surfaces. However, after 48 and 72 h, cell spreading and proliferation were enhanced on eTi surfaces, fol-

Fig. 4. Cell adhesion and proliferation on the engineered titanium surfaces as a function of time. Cell cultures were repeated four times for each treatment and time period.

228

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

Fig. 5. Scanning electron micrographs of osteablast-like cells adhered to an acid-etched titanium surface (a) and to a polished titanium surface (b) after 72 h.

lowed by beTi, bTi and pTi surfaces as decreasing surface area. MG63 osteoblastic-like cells showed high adhesion on eTi surfaces, owing to an enhanced filopodial anchorage at the nanoasperities and micropores. Otherwise, on the bTi samples, the cells were adherent but not confluent. 4. Discussion In culturing cells on biomaterial surfaces, surface free energy is an important parameter that guides the first events occurring at the biomaterial/biological interface, such as interaction of water and proteins with biomaterial, and these events guide further response [8,38]. Nevertheless, surface energy estimation from contact angle measurements is a hard task because biomaterial surfaces are always rough and/or heterogeneous. If surface roughness is the primary cause of hysteresis, i.e. Ra ≥ 0.1 ␮m [31], the advancing contact angle is influenced more by the microscopic relief than by interfacial energetics. However, surface energy estimation just requires thermodynamically significant angles. Hence, for rough titanium-based surfaces, the Young contact angles computed by Eq. (2) from advancing contact angles can serve to detect relatively gross changes in surface energy. The good cell response at long times on eTi samples, as regards the rest of textured groups, was explained by their particular morphology (the greatest fractal dimension and negative skewness) and surface chemical homogeneity (no traces). Chemically etched surfaces have long been recognized as fractal, due to the stochastic character of wet etching. Particularly, the morphologies of eTi samples were angular and stepped due to the crystalline grain detachment from the amorphous structure of titanium. Moreover, at dimensions smaller than the grain structure of titanium, a secondary texture was visualized. Hence, eTi samples, even being less rough than bTi and beTi groups, exhibited microporous and nano-rough texture. It is worth to mention that surface roughness in the range from 10 nm to 10 ␮m may influence the interface biology, since it is of the same order in size as cells and large biomolecules. Microporous surfaces enhance osteointegration of titanium implants, indicated by mechanical pull-out tests and histomorphometric analysis [39] or osteoblast spreading tests [40]. As far as biomaterial interactions with proteins are concerned, surface features on the nanometer scale are important. Changes in surface chemistry or roughness, of the same dimension as a protein will affect its adsorption characteristics. Instead, if the rough surface is treated as chemically homogeneous and the size of asperities is larger than the protein dimension, then roughness will simply add available surface area for adsorbing protein. This explains that beTi and bTi samples were more bioadhesive at early stages of cell attachment [6,12,41]. At late stages, the presence of alumina particles and the coarsely blasted morphology, self-affine but

uncomfortable for cell spreading, impaired the cell proliferation as expected [22]. Furthermore, sandblasting can also affect the corrosion resistance of titanium due to the residual surface stresses induced [37]. Fractal dimension arises as a more appropriate index of surface disorder [22,42] rather than the amplitude roughness parameters. However, a single exponent is not enough for the complete description of fractal surfaces. In this work, the partial information provided by fractal dimension was refined by the values of kurtosis and skewness, and confirmed by SEM pictures. Alternatively, lacunarity becomes a counterpart to the fractal dimension because it just describes the texture of a fractal surface [21]. Fractal surfaces having same fractal dimensions can look widely different because of having different lacunarity. This parameter is related to the prefactor ˛ of Eq. (1), known as topothesy. HF acid etching is recommended for titanium-based dental implants rather than sandblasting based in the results of this study. As suggested in the literature, the HF treatment creates a microand nano-level topography that enhances adherent cells [43,44]. Moreover, since surface characterization of biomaterials is crucial to understand the concerning biological events, the use of complementary topography analysis, like fractal analysis, in addition to statistical roughness description are highly recommended. Finally, theories and experimental methods related to wetting phenomena should be carefully employed in biomaterial engineering. 5. Conclusions The conclusions of this work are summarized as follows: 1. The traces of silicon carbide and alumina found in the polished and blasted surfaces modified their surface energy, accordingly. 2. Cell adhesion and proliferation on the textured titanium surfaces with similar surface chemistry (acid-etched and blasted-etched surfaces) were driven by topography features. 3. Cell adhesion (≤24 h) depended on the available surface area (more roughness, more cell adhesion). 4. Cell proliferation (≥48 h) was mediated by specific morphologies: highly fluctuating surfaces (greater fractal dimension) and punching surfaces (negative skewness). Unlike, for surfaces with similar fractal dimension and positive skewness, the cell growth basically depended on their surface area. 5. Acid etching yielded angular stepped morphologies due to the crystalline grain detachment, with horizontal length scale close to the cell size (∼10 ␮m). 6. Blasting treatment generated pit-like surface features with high aspect ratios and chaotic morphologies. These physical surface modifications, aside from the small alumina doping, hindered the cell growth.

J.I. Rosales-Leal et al. / Colloids and Surfaces A: Physicochem. Eng. Aspects 365 (2010) 222–229

Acknowledgements ˜ de EduThis work was supported by the “Ministerio Espanol cación y Ciencia” (project MAT2007-66117 and contract “Ramón y Cajal” RYC-2005-000983), Junta de Andalucía (projects P07-FQM02517 and P08-FQM-4325) and the European Social Fund (ESF). The authors are grateful to Yudi Gómez-Villaescusa for the laboratory assistance.

[23] [24] [25]

[26]

References [27] [1] P.I. Branemark, B.O. Hansonn, R. Adell, U. Breine, J. Lindstrom, O. Hallen, A. Ohman, Osseointegrated titanium implants in the treatment of the edentulous jaw, Scand. J. Plast. Reconstr. Surg. 11 (1977) 171–175. [2] T. Albrektsson, P.I. Branemark, H.A. Hanson, J. Lindstrom, Osseointegrated titanium implants. Requirements for ensuring a long-lasting, direct bone anchorage in man, Acta Orthop. Scand. 52 (1981) 155–170. [3] D. Botticelli, T. Berglundh, D. Buser, J. Lindhe, The jumping distance revisited. An experimental study in the dog, Clin. Oral Implant. Res. 14 (2003) 35–42. [4] H. Assender, V. Bliznyuk, K. Porfyrakis, How surface topography relates to materials’ properties, Science 297 (2002) 973–976. [5] R.G. Flemming, C.J. Murphy, G.A. Abrams, S.L. Goodman, P.F. Nealy, Effects of synthetic micro- and nano-structured surfaces on cell behaviour, Biomaterials 20 (1999) 573–588. [6] F. Rupp, L. Scheideler, D. Rehbein, D. Axmann, J. Geis-Gerstorfer, Roughness induced dynamic changes of wettability of acid etched titanium implant modifications, Biomaterials 25 (2004) 1429–1438. [7] J.Y. Martin, Z. Schwartz, T.W. Hummert, D.M. Schraub, J. Simpson, J. Lankford, D.D. Dan, D.L. Cochran, B.D. Boyan, Effect of titanium surface roughness on proliferation, differentiation, and protein synthesis of human osteoblast-like cells (MG63), J. Biomed. Mater. Res. 29 (1995) 389–401. [8] K. Kieswetter, Z. Schwartz, T.W. Hummert, D.L. Cochran, J. Simpson, D.D. Dean, B.D. Boyan, Surface roughness modulates the local production of growth factors and cytokines by osteoblast-like MG-63 cells, J. Biomed. Mater. Res. A 32 (1996) 55–63. [9] D.A. Puleo, A. Nance, Understanding and controlling the bone-implant interface, Biomaterials 20 (1999) 2311–2321. [10] A.W. Adamson, A.P. Gast, Physical Chemistry of Surfaces, 6th ed., New York, Wiley, 1997. [11] X. Zhu, J. Chen, L. Scheideler, R. Reichl, J. Geis-Gerstorfer, Effects of topography and composition of titanium surface oxides on osteoblast responses, Biomaterials 25 (2004) 4087–4103. [12] L. Hao, J. Lawrence, K.S. Chian, Osteoblast cell adhesion on a laser modified zirconia based bioceramic, J. Mater. Sci. Mater. Med. 16 (2005) 719–726. [13] M.C. Advincula, F.G. Rahemtull, R.C. Advincula, E.T. Ada, J.E. Lemons, S.L. Bellis, Osteoblast adhesion and matrix mineralization on sol–gel derived titanium oxide, Biomaterials 27 (2006) 2201–2212. [14] M.I. Janssen, M.B. Van Leeuwen, T.J. van Coten, J. de Vries, L. Dijkhuizen, H.A. Wösten, Promotion of fibroblast activity by coating with hydrophobins in the beta-sheet end state, Biomaterials 25 (2004) 2731–2739. [15] R.E. Baier, Surface behaviour of biomaterials: the theta surface for biocompatibility, J. Mater. Sci. Mater. Med. 17 (2006) 1057–1062. [16] F. Rupp, L. Scheideler, J. Geis-Gerstorfer, Effect of heterogenic surfaces on contact angle hysteresis: dynamic contact angle analysis in material sciences, Chem. Eng. Technol. 25 (2002) 877–882. [17] M. Bächle, R.J. Kohal, A systematic review of the influence of different titanium surfaces on proliferation, differentiation and protein synthesis of osteoblastlike MG63 cells, Clin. Oral Implant. Res. 15 (2004) 683–692. [18] A. Bagno, C. Di Bello, Surface treatments and roughness properties of Ti-based biomaterials, J. Mater. Sci. Mat. Med. 15 (2004) 935–949. [19] S.W. Ha, A. Gisep, J. Mayer, E. Wintermantel, Topographical characterization and microstructural interface analysis of vacuum-plasmasprayed titanium and hydroxyapatite coatings on carbon fibre-reinforced poly(etheretherketone), J. Mater. Sci. Mat. Med. 8 (1997) 891–896. [20] E.S. Gadelmawla, M.M. Koura, T.M.A. Maksoud, I.M. Elewa, H.H. Soliman, Roughness parameters, J. Mater. Process. Technol. 123 (2002) 133–145. [21] B.B. Mandelbrot, The Fractal Geometry of Nature, Freeman, New York, 1983. [22] K. Anselme, M. Bigerelle, B. Noel, E. Dfresne, D. Judas, A. Iost, P. Hardouin, Qualitative and quantitative study of human osteoblast adhesion on materi-

[28]

[29]

[30]

[31] [32]

[33] [34]

[35]

[36]

[37]

[38]

[39]

[40]

[41]

[42]

[43]

[44]

229

als with various surface roughnesses, J. Biomed. Mater. Res. A 49A (2000) 155– 166. S. Amada, A. Satoh, Fractal analysis of surfaces roughened by grit blasting, J. Adhes. Sci. Technol. 14 (2000) 2741. G. Palasantzas, J.T.M. De Hosson, Influence of surface roughness on the adhesion of elastic films, Phys. Rev. E 67 (2003) 021604. P.F. Chauvy, C. Madore, D. Landolt, Variable length scale analysis of surface topography: characterization of titanium surfaces for biomedical applications, Surf. Coat. Technol. 110 (1998) 48–56. P.J. Ramon-Torregrosa, M.A. Rodríguez-Valverde, A. Amirfazli, M.A. CabrerizoVílchez, Factors affecting the measurement of roughness factor of surfaces and its implications for wetting studies, Colloids Surf. A: Physicochem. Eng. Aspects 323 (2008) 83–94. T.J. Webster, E.S. Ahn, Nanostructured biomaterials for tissue engineering bone, Adv. Biochem. Eng. Biotechnol. 103 (2006) 275–308. C. Douketis, Z. Wang, T.L. Haslett, M. Moskovits, Fractal character of colddeposited silver films determined by low-temperature scanning tunnelling microscopy, Phys. Rev. B 51 (1995) 11022–11031. W. Zahn, A. Zösch, The dependance of fractal dimension on measuring conditions of scanning probe microscopy, Fresenius J. Anal. Chem. 365 (1999) 168–172. M. Morra, C. Cassinelli, Evaluation of surface contamination of titanium dental implants by LV-SEM: comparison with XPS measurements, Surf. Interface Anal. 25 (1997) 983–988. M. Morra, E. Occhiello, F. Garbassi, Knowledge about polymer surfaces from contact angle measurements, Adv. Colloid Interface Sci. 32 (1990) 79–116. R.V. Bathomarco, G. Solorzano, C.N. Elias, R. Prioli, Atomic force microscopy analysis of different surface treatments of Ti dental implant surfaces, Appl. Surf. Sci. 233 (2004) 29–34. R.N. Wenzel, Resistance of solid surfaces to wetting by water, Ind. Eng. Chem. 28 (1936) 988–994. M. Morra, C. della Volpe, S. Siboni, Comment to the paper: enhancing surface free energy and hydrophilicity through chemical modification of microstructured titanium implant surfaces, by F. Rupp, L. Scheideler, N. Olshanska, M. de Wild, M. Wieland, J. Geis-Gerstorfer, J. Biomed. Mater. Res. A 79A (2006) 752–754. M.A. Rodríguez-Valverde, F.J. Ruiz-Cabello, M.A. Cabrerizo-Vilchez, Wetting on axially-patterned heterogeneous surfaces, Adv. Colloid Interface Sci. 138 (2008) 84–100. H.A. Wege, J.A. Holgado-Terriza, J.I. Rosales-Leal, R. Osorio, M. Toledano, M.A. Cabrerizo-Vílchez, Contact angle hysteresis on dentin surfaces measured with ADSA on drops and bubbles, Colloids Surf. A: Physicochem. Eng. Aspects 206 (2002) 469–483. C. Aparicio, F.J. Gil, C. Fonseca, M. Barbosa, J.A. Planell, Corrosion behaviour of commercially pure titanium shot blasted with different materials and sizes of shot particles for dental implant applications, Biomaterials 24 (2003) 263–273. F. Rupp, L. Scheideler, N. Olshanska, M. de Wild, M. Wieland, J. Geis-Gerstorfer, Enhancing surface free energy and hydrophilicity through chemical modification of microstuctured titanium implant surfaces, J. Biomed. Mater. Res. A 76A (2006) 323–334. G. Giavaresi, M. Fini, A. Cigada, R. Chiesa, G. Rondelli, L. Rimondini, P. Torricelli, N.N. Aldini, R. Giardino, Mechanical and histomorphometric evaluations of titanium implants with different surface treatments inserted in sheep cortical bone, Biomaterials 24 (2003) 1538–1594. R.L. Sammons, N. Lumbikanonda, M. Gross, P. Cantzler, Comparison of osteoblast spreading on microstructured dental implant surfaces and cell behaviour in an explant model of osseointegration: a scanning electron microscopic study, Clin. Oral Implant. Res. 16 (2005) 657–666. C.N. Elias, Y. Oshida, J.H. Lima, C.A. Muller, Relationship between surface properties (roughness, wettability and morphology) of titanium and dental implant removal torque, J. Mech. Behav. Biomed. Mater. 1 (3) (2008) 234–242. D. Chappard, I. Degasne, G. Huré, E. Legrand, M. Audran, M.F. Bas, Image analysis measurements of roughness by texture and fractal analysis correlate with contact profilometry, Biomaterials 24 (2003) 1399–1407. J. Guo, R.J. Padilla, W. Ambrose, I.J. De Kok, L.F. Cooper, The effect of hydrofluoric acid treatment of TiO2 grit blasted titanium implants on adherent osteoblast gene expression in vitro and in vivo, Biomaterials 28 (2007) 5418–5425. S.F. Lamolle, M. Monjo, M. Rubert, H.J. Haugen, S.P. Lyngstadaas, J.E. Ellingsen, The effect of hydrofluoric acid treatment of titanium surface on nanostructural and chemical changes and the growth of MC3T3-E1 cells, Biomaterials 30 (2009) 736–742.