Titanium (Ti) is widely used in medicine and dentistry, either in its pure form or in alloys, due to the combination of physical, chemical and biological properties (Scharnweber et al 2002). Titanium is employed in eg, joint prostheses, fracture fixation devices, endovascular stents, pacemaker enclosures, and dental materials (Peters et al 1984; Yamauchi et al 2000; Stoeckel et al 2004; Tschernitschek et al 2005).

The high corrosion resistance of Ti is due to a strongly passivating oxide layer (Hallab et al 2004). Nevertheless, Ti degrades under mechanical stress, abrasion, low pH and low oxygen level (Tschernitschek et al 2005). It is established that Ti-based implants release particles and ions to the surrounding tissue and fluids, as has been demonstrated with hip implants (Mckellop et al 1990; Brien et al 1992), as well as with miniplates used for jaw fractures (Schliephake et al 1993; Torgersen et al 1995), and even unloaded dental implants (Arys et al 1998; Flatebø et al 2006). However, the biological consequences of the degradation products, ions, nano- or micrometer sized particles, remain unclarified.

Based mostly on case reports, it has been suggested that hypersensitivity to Ti is associated with implant loosening (Wooley et al 1996; Matthew and Frame 1998), or skin reactions (Peters et al 1984; Yamauchi et al 2000; Tamai et al 2001; High et al 2006; Thomas et al 2006; Watanabe et al 2006). Diagnosis of contact allergy to metals usually relies on combination of anamnestic information and in vivo tests such as a epicutaneous patch test (Hallab et al 2001; Granchi et al 2006), and – in some instances – ex vivo tests such as a Lymphocyte Stimulation Test (LST) on cells from blood samples (Hallab et al 2001). Hypersensitivity tests have not been established for Ti (Lalor et al 1991; Hallab et al 2001; Klein et al 2004) and Ti is not included in standard commercial routine patch test series. Various titanium compounds, such as Ti-salts (Lalor et al 1991) and TiO₂ micrometer sized particles (Thomas et al 2006) have been tried in hypersensitivity testing, but optimal Ti antigens are not established yet (Bircher and Stern 2001).

The ex vivo test, LST, requires cell culture techniques and bioavailable Ti at physiological culture conditions (Hallab et al 2001). The difficulty with Ti is that this metal is not readily soluble at physiological pH, as nickel salts are (Lalor et al 1991; Hallab et al 2000; Hallab et al 2004). Had relevant and stable Ti-substances been available, the applicability of patch testing (Granchi et al 2006) and LST (Hallab et al 2001) could be improved for Ti.

Nanoparticles, with dimensions less than 100 nm, interact differently with various proteins and subsequently with cells when compared with the conventional size particles (Liu and Webster 2006). Surface functionalization of gold nanoparticles has been used to develop diagnostic and treatment tools (Shenoy et al 2006). Surface modification of nanoparticles can be achieved by protein binding.

Albumin has been shown to be dominant in binding Ti in vivo (Hallab et al 2000). It is the most prevalent protein in plasma (approximately 42 g/l for adults) with important functions such as metal binding, transport, and in providing colloidal osmotic pressure (Vander et al 1998).

By simulating the model of metal-protein binding in vivo, the aim of the present study was to develop Ti-protein antigens based on titanium dioxide nanoparticles that can be useful for testing cellular responses in diagnosis of human Ti hypersensitivity.

The preparation of antigens in the first part of our study was proceeded for a period of 8 days, in orders to asses the effect of the exposure time between nano-TiO₂ and albumin on Ti concentration. Based on obtained results, the study continued with TEM imaging and the determination of stability of nano-TiO₂ antigens in suspension.

The filtered suspensions, with no HSA, gave a low median Ti concentration (by ICP-MS), below 1 μg/l, and unaffected by exposure time. The presence of HSA increased the concentration of Ti by 2–3 orders of magnitude when compared to controls (suspensions without HSA). Longer exposure time raised also the Ti concentrations in suspensions with 0.1 and 1% HSA (Figure 1). Differences between Ti concentrations with 0.1% HSA and 1% HSA were statistically different for all time points, except at day 8.

The TEM images of the unfiltered suspensions revealed that the individual TiO₂ nanoparticles were spherical, with diameters ranging from 20 to more than 100 nm. The TiO₂ nanoparticles formed typical aggregates in both suspensions, with and without albumin (Figures 2a and 2b).

The DLS measurements results showed that TiO₂ nanoparticles formed large aggregates in DMEM without HSA, but the mean aggregate size was about 1/5 in presence of HSA. After filtration, the mean aggregate size was about half of the unfiltered suspensions (Table 1). The evolution of TiO₂ aggregates size in DMEM with 1% HSA revealed a slow growth over 20 h period (regression model: mean size (nm) = 1.46 time (h) + 210; R² = 0.83).

The sedimentation rate results revealed that the TiO₂ aggregates sedimented rapidly (5 h) in absence of HSA, whereas in presence of HSA, the sedimentation rate was close to zero (Figure 3). After filtration, the TiO₂ aggregates were stable in suspension all along a 20 h period (Figure 3). In DMEM without HSA, after filtration, the TiO₂ nanoparticles were not detectable by either Zetasizer or the UV-vis spectrophotometer.

The results presented in this study show that the combination of human serum albumin and Ti in the form of nanoparticles raises the bioavailability of Ti in suspension, most likely through a metal-protein association.

Using nanoparticles to develop a potential diagnostic reagent for hypersensitivity testing is new and clinically appropriate because nanometer sized wear products have been found in tissues surrounding implants (Buscher et al 2005). Titanium dioxide nanoparticles have greater potential toxic effect on cells than the conventional size particles (Chen et al 2006; Long et al 2006).

When compared with the micrometer-sized and easily detectable particles, the nanoparticles have larger specific surfaces and different physiochemical properties, which mean higher surface energy (Long et al 2006). Consequently, their interaction with cell membrane, cell structure, or cell activity may differ from the larger-size particulate matter (Long et al 2006).

Our results suggest that TiO₂ nanoparticles, which aggregate and sediment rapidly in cell culture medium, are more stable in suspensions when they are in presence of HSA. This increased stability could be due to protein adsorption onto the surface of TiO₂ nanoparticles or aggregates (Klinger et al 1997; Oliva et al 2003). Moreover, steric repulsive forces caused by the protein chain reduce the net attractive interactions between particles, and thus alter the aggregation rate (Tirado-Miranda et al 2003; Limbach et al 2005). The presence of albumin slowed down the aggregate settling as shown by the sedimentation rate test, and raised the Ti concentration orders of magnitude even after filtration.

In presence of HSA, the mean aggregate size, as revealed by DLS, was slightly smaller than the pore size of 450 nm and thus most of the aggregates passed through the filter membrane (Table 1). These results are in concordance with results from the quantitative chemical analyses (ICP-MS) showing high final Ti concentration in the samples (Figure 1). Furthermore, the effect of HSA on the reducing the aggregate size of TiO₂ nanoparticles in suspensions is supported by our sedimentation study (Figure 3). Ultrasonication used for 30 sec of TiO₂-albumin suspensions showed also an effect on clustering size, by breaking down the big aggregates.

Few studies have been done on the colloidal stability of nanoparticles in cell culture media. Our results are in accordance with Long et al (2006) who revealed that DMEM affects the rate of aggregation and consequently the size of clusters. In solution with high ionic strength without proteins, such as DMEM medium, the rate of aggregation is high, leading to fast sedimentation (Gomez-Lopera et al 2006; Phenrat et al 2007). However, if HSA is present, the adsorption of the protein onto TiO₂ surfaces is almost independent of the ionic strength (Oliva et al 2003). The binding mechanisms of the TiO₂-albumin interaction need more research.

It can be assumed that the TiO₂ nanoparticles behave in a similar manner in vivo, making aggregates and binding to proteins. This supports the idea that the metal released by implant degradation does not exist as unbound ionic or colloidal forms in serum but bound to serum proteins (Hallab et al 2000).

Since our potential antigens could be prepared already after 2 h of shaking, we recommend using these fresh suspensions, in hypersensitivity testing, instead of those mixed for 8 days.

The biological reactivity of the new possible antigens is now being pursued in Lymphocyte Stimulation Test (LST). More tests needs to be done in order to standardize the method.

The TiO₂ nanoparticle-albumin antigen could be clinically relevant as it mimics degradation products of Ti-based implants. Albumin has an important role in the colloidal stability of TiO₂ nanoparticles in suspension, and thus, on the bioavailability of Ti for cellular tests. The applicability of the substances in clinical testing of hypersensitivity has to be elucidated.