Abstract
Reported bone-implant contact percentages are far below the ideal 100%. We tested a hypothesis that the protein adsorption capability of titanium, which is critical to the process of osseointegration, changes over time before its use. Machined, acid-etched, and sandblasted surfaces were prepared and stored under dark ambient conditions for 3 days, 1 week, or 4 weeks. For all surfaces, protein adsorption decreased as the storage time increased, and their decreasing rates were dependent on titanium topography. After 4 weeks, the amounts of albumin and fibronectin adsorbed by the acid-etched surface were only 20% and 35%, respectively, of that adsorbed by the fresh surface after 2 hours of incubation, and remained substantially low even after 24 hours. This time-dependent degradation in protein adsorption of titanium correlated with its naturally decreasing hydrophilicity, which was not observed for the nickel and chromium surfaces, indicating a titanium-specific biological aging.
INTRODUCTION
The capacity of protein adsorption is a critical component in determining the biocompatibility of any implantable materials and tissue-engineering scaffolds (Woo et al., 2007). Through the interaction of integrins and adsorbed proteins, cells attach to the biomaterial surface, the degree of which may consequently affect the spread, proliferation, and differentiation of the cells (Mata et al., 2003). For instance, fibronectin, a high-molecular-weight glycoprotein found in the blood plasma, directly and indirectly regulates cell-to-material attachment by binding to integrins and to extracellular matrix components such as collagen (Garcia et al., 1998). Fibronectin coating of various biomaterials promotes osteoblastic attachment and proliferation (Sousa et al., 2008). Some proteins also play an important role in the metabolism and function of cells attached to biomaterials. For instance, serum albumin, which constitutes about 60% of human plasma protein, serves as a carrier for molecules of low water solubility, including various hormones and calcium. Albumin-bound lipids regulate cytoplasmic calcium oscillations and stimulate osteoblast proliferation (Tsai et al., 2007).
Dental implants are no exception; upon placement of the implant, titanium surfaces become coated with a thin proteinaceous film. Experimental coating of titanium with fibronectin promotes osteoblast attachment (Degasne et al., 1999), proliferation (van den Dolder et al., 2003), and differentiation (Ku et al., 2005) and eventually enhances the degree of osseointegration (Park et al., 2006). The rate of protein adsorption on implant surfaces is known to differ depending on the chemistry (Yang et al., 2005), structure (Kern et al., 2005), and morphology (Protivinsky et al., 2007) of implant surfaces. Therefore, surface modification of titanium has been a primary focus of implant science and technology in an effort to improve its osteoconductivity. However, the effect of age—in other words, the effect of time after processing the titanium surface—has not been addressed. Titanium surfaces are known to absorb hydrocarbons progressively over time under ambient conditions, which is likely to be responsible, at least in part, for determining the surface energy of titanium (Takeuchi et al., 2005a).
We hypothesized that the protein adsorption capacity of titanium surfaces changes over time before its use. The objective of this study was to determine whether the adsorption rates of albumin and fibronectin to titanium surfaces change over time after processing (i.e., storage time), and if so, how such change is related to time-dependent changes in the surface energy of titanium. We tested 3 different surface topographies of titanium. We also determined whether this phenomenon is specific to titanium.
MATERIALS & METHODS
Titanium Samples and Surface Characterization
Commercially pure grade 2 titanium disks with a diameter of 20 mm were prepared with 3 different surface types: machined, acid-etched, and sandblasted. The machined surface was created by means of a lathe, and the acid-etched surface was prepared by acid-etching of the machined disks with 67% (w/w) sulfuric acid (H2SO4) at 120°C for 75 sec. Al2O3 particles (50 μm) were sandblasted for 1 min at a pressure of 3 kg/m2 and then rinsed with ultrasonic distilled water for 10 min to prepare the sandblasted surface. The surface morphology of the disks was examined with scanning electron microscopy (SEM) (XL30, Philips, Eindhoven, Netherlands) and atomic force microscopy (AFM) (SPM-900J3, Shimadzu, Kyoto, Japan). For AFM studies, the contact mode scanning was performed in an area of 5 μm x 5 μm. In addition to the 3 types of titanium disks, 3 different metal surfaces were prepared, as described previously (Saruwatari et al., 2005), by ‘depositioning’ the cell culture grade polystyrene dish with either titanium, nickel, or chromium by e-beam physical vapor deposition technology (SLONE e-beam evaporator, SLONE Technology Co., Santa Barbara, CA, USA) at 5 Å/sec to a final thickness of 250 nm. The prepared disks or metallic surfaces were used for the experiments immediately or after storage under dark ambient conditions for 3 days, 1 wk, or 4 wks.
Contact Angle Measurement
The surface energy of each of the prepared substrates was evaluated by the contact angle of 1 μL H2O by means of an automatic contact angle measuring device (DCA-VZ, Kyowa Interface Science, Saitama, Japan). Also, side views of H2O spread onto the substrates were digitally photographed for presentation.
Protein Adsorption Measurement
Bovine serum albumin (Pierce Biotechnology, Inc., Rockford, IL, USA) and bovine plasma fibronectin (Sigma-Aldrich, St. Louis, MO, USA) were used as model proteins. A 300-μL quantity of protein solution (1 mg/mL protein/ saline) was pipetted onto and spread over a titanium disk or deposited onto metal surfaces. After several different periods of incubation in sterile humidified conditions at 37°C, the non-adherent protein was removed, and surfaces were washed twice with saline containing 0.9% sodium chloride. Two-hundred-μL aliquots of the initial and removed solutions were mixed with 200 μL microbicinchoninic acid (Pierce Biotechnology, Inc.) and incubated at 37°C for 60 min. The amount of protein was quantified by means of a microplate reader at 562 nm.
Statistical Analysis
All of the experiments described above were performed in triplicate. The effects of different ages and topographies of substrates on the H2O contact angle and the rate of protein adsorption were analyzed by 2-way ANOVA. When needed, a Bonferroni multiple-comparisons test was performed ad hoc. A P value of < 0.05 was considered statistically significant. Possible correlations between the albumin adsorption and H2O contact angle were examined, and regression formulas were determined by least-squares mean approximation.
RESULTS
Lower Rate of Early Albumin Adsorption on Aged Titanium
The 3 different surfaces of titanium disks we tested clearly presented different surface morphologies (Figs. 1A–1C). The root mean-square roughness, peak-to-valley roughness, and inter-irregularities space obtained from AFM analysis were: 0.025 ± 0.006 μm, 0.052 ± 0.010 μm, and 0.590 ± 0.155 μm, respectively, for the machined surface; 0.211 ± 0.071 μm, 0.455 ± 0.074 μm, and 0.891 ± 0.120 μm, respectively, for the acid-etched surface; and 0.206 ± 0.035 μm, 0.423 ± 0.125 μm, 1.856 ± 0.426 μm, respectively, for the sandblasted surface. All values for the acid-etched and sandblasted surfaces were significantly higher than those for the machined surface ( p < 0.001; Bonferroni). The inter-irregularities space of the sandblasted surface was significantly greater than that of the acid-etched surface (p < 0.05).
The rates of albumin adsorption on titanium disks at various storage times are presented as a percentage relative to the total amount incubated for 1 hr (Fig. 1D). A clear trend of a decreasing rate of albumin adsorption with an increase in storage time was observed; 2-way ANOVA showed that the effect of titanium age was significant for all 3 of the surface topographies tested (p < 0.001). Bonferroni multiple-comparisons tests further showed a significant reduction in protein adsorption on the 1- and 4-week-old titanium compared with the freshly prepared titanium within each of the surface topographies ( p < 0.05). The titanium age-related decreasing pattern in albumin adsorption was disproportional among the different topographies tested. The time-induced reduction in albumin adsorption was most marked on the acid-etched surface; the 4-week-old acid-etched surface showed only 30% adsorption of the freshly prepared acid-etched surface. The protein adsorption rate of the freshly acid-etched surface, which was highest among the 3 topography types, decreased rapidly to a level similar to that of the other 2 topography types in 4 wks.
Diminished Capacity of Protein Adsorption on Aged Titanium
Based on the finding of a decelerated rate of early albumin adsorption on aged titanium disks, we examined the percentage of albumin adsorption after longer incubation periods, to determine adsorption capacity. A similar common trend of lower adsorption on the 4-week-old titanium than on the fresh titanium was observed for all the topography types (Fig. 2A). Of note, the amount of albumin adsorption on the 4-week-old titanium did not reach the level of adsorption as that on freshly prepared titanium, even after 24 hrs of incubation, for all the topographies tested, indicating that both the rate and the capacity of albumin adsorption were reduced in aged titanium. The adsorption of fibronectin yielded a trend similar to that of albumin, i.e., substantially reduced adsorption on the 4-week-old disks. The amount of fibronectin adsorbed onto the 4-week-old titanium surfaces did not reach the level achieved by the fresh surfaces after 24 hrs of incubation (Fig. 2B).
Inverse Linear Correlation between H2O Contact Angle and Protein Adsorption
The contact angle of H2O increased with the age of titanium (Figs. 3A, 3B). The superhydrophilic status (contact angle < 5°) of the fresh surfaces of all 3 topography types became hydrophobic (over 60°) after 4 wks of storage. To find a factor responsible for the reduced protein capacity of titanium, we plotted the percentages of early albumin adsorption during 1 hr of incubation against the H2O contact angle (Fig. 3C). A statistically significant regression line between the 2 variables was observed for all 3 surface topography types. Albumin adsorption was negatively correlated with the contact angle with a high coefficient of determination (0.881–0.997; p < 0.001). As shown in the slopes of the regression lines, the albumin adsorption rate decreased with age more rapidly on the acid-etched and sandblasted surfaces than on the machined surface; the decreasing rate of the acid-etched surface was twice that of the machined surface, as indicated by their proportion coefficients. Interestingly, a collective analysis for all the data in the 3 topography types yielded a weaker correlation between albumin adsorption and contact angle (R2 = 0.761).
Titanium-specific Degradation of Protein Adsorption Capacity with Time
The 3 different metal surfaces prepared by physical vapor deposition showed smooth and amorphous surface characteristics and no inter-surface differences in surface topographical parameters (Figs. 4A–4C); the root mean-square surface roughnesses were 3.13 ± 1.02 nm, 3.22 ± 1.22 nm, and 3.15 ± 1.02 nm for the titanium, nickel, and chromium surfaces, respectively. Comparison of the early albumin adsorption rates between the deposited metal surfaces at different storage times showed that the protein adsorption rate was 30% lower on the 4-week-old titanium surface than on the fresh surface. However, this phenomenon was not seen on the nickel and chromium surfaces.
DISCUSSION
This is the first study to introduce the time-related changes in biological capacity for implant materials. In terms of the early albumin adsorption rate, degradation after 4 wks of storage was substantial compared with the level showed by the matching fresh surfaces: a 50% reduction for the machined surface, 70% for the acid-etched surface, and 60% for the sandblasted surface. Similarly, substantial degradation was also observed in the fibronectin adsorption capacity. More importantly, the protein adsorption level of the 4-week-old titanium disks did not reach or even come close to the level of the fresh surfaces after a longer incubation time of 24 hrs. This implies that such an initial difference in biological potential may become permanent and determine the level of subsequent bioactivity of titanium, potentially influencing osseointegration.
It should be noted that the amount of degradation in protein adsorption between the fresh and 4-week-old surfaces was greater for the acid-etched and sandblasted surfaces than for the machined surface. The acid-etched and sandblasted surfaces, which exhibit microtopographic features, are representative of the surface textures currently used for dental implants. Endosseous metallic implants, regardless of whether they are used for dental or orthopedic applications, are basically sold in the market as storable medical devices at the manufacturers’ and users’ levels; quality control of the products related to circulation and inventory periods, or a sense of expiration, has never been addressed. Although the age-inducible changes in osteoconductivity of titanium need to be determined by extending the scope of this study, the present findings may broaden the understanding of the mechanism of bone-implant integration and have a significant impact on many aspects of future orthopedic and dental implant therapy.
The protein adsorption rate of titanium correlated with the contact angle of H2O, which suggests the importance of the hydrophilic status of titanium in determining its protein adsorption ability. However, interestingly, the higher correlations were found within each of the individual topography types than on the data collected from all surface types tested, which indicates that protein adsorption ability is not dependent solely on surface hydrophilicity, but also on other co-elements of surface characterization, such as surface topography. In fact, proportional correlations with a higher reliability and a steeper slope were found for the titanium surface with microroughness, rather than for the titanium surface, which was relatively smooth. This finding indicates that the age of rougher titanium surfaces has a greater effect on protein adsorption ability than does the age of smoother surfaces. The nature and limitation of surface characterizations used in this study are discussed in the Appendix.
As typically shown in studies conducted on the association between substrate hydrophilicity and cellular response, the effect of the surface hydrophilicity of biomaterials on their bio-activity is contentious. For instance, poly(lactide-co-beta-benzyl malolactonate) with improved hydrophilicity promotes NIH3T3 fibroblast attachment and proliferation (He et al., 2004). In contrast, a lower rate of proliferation was found in fibroblasts cultured on polyhydroxyalkanoates with improved hydrophilicity (Wang et al., 2003). Regarding osteoblastic response, more hydrophobic polymer scaffold materials created by altering the composition promoted the healing of bone defects (Jansen et al., 2005). As seen in these studies, it has been understandably difficult to obtain consistency of results and make definitive conclusions on the importance of the hydrophilic status of a material in the material’s biocompatibility. In these studies, surface topography and composition of the material have varied drastically, along with the change of hydrophilicity, making the interpretation of the results difficult. The present study examined the biological potential of titanium with different hydrophilic status, while maintaining the identical surface topographies of the substrates. Additionally, we showed that the age-dependent degradation in protein adsorption ability was applicable only to titanium, and not to nickel and chromium surfaces with comparable surface topography.
Titanium constantly absorbs organic impurities, such as polycarbonyls and hydrocarbons, from the atmosphere, water, and cleaning solutions (Kilpadi et al., 2000; Serro and Saramago, 2003). The fact that the high atomic percentage of carbon is detected unexceptionally on titanium surfaces indicates that such contamination may be unavoidable (Massaro et al., 2002; Buser et al., 2004). Recent reports suggest a link between surface hydrocarbons and the hydrophilic status of titanium: The more hydrocarbons absorbed, the higher the contact angle of H2O (Takeuchi et al., 2005b). Our recent study demonstrated that protein adsorption to titanium surfaces and the progressive removal of surface carbon during ultraviolet light treatment are well-correlated (Aita et al., 2009). Although this result did not provide evidence directly linking protein adsorption and surface carbon, it strongly suggested the need for exploration of the role of surface carbon in determining the bioactivity of titanium surfaces. In addition to the effect of titanium age on its in vivo osteoconductive potential, the role of surface hydrocarbon on titanium in determining its protein adsorption ability, as well as possible dependent and independent effects of hydrocarbon and hydrophilicity, is of great interest.
Scanning electron microscopic images of the 3 types of titanium disks used in this study: machined Mean ± SD adsorption rates of bovine serum albumin Age-dependent reduction of hydrophilicity on titanium surfaces and its correlation with protein adsorption capacity. Scanning electron microscopic images of 3 metallic surfaces deposited onto the cell-culture-grade polystyrene dishes: titanium 
Footnotes
References
Supplementary Material
Please find the following supplemental material available below.
For Open Access articles published under a Creative Commons License, all supplemental material carries the same license as the article it is associated with.
For non-Open Access articles published, all supplemental material carries a non-exclusive license, and permission requests for re-use of supplemental material or any part of supplemental material shall be sent directly to the copyright owner as specified in the copyright notice associated with the article.
