An investigation of the photo-clastogenic potential of ultrafine titanium dioxide particles

Available online at www.sciencedirect.com

Mutation Research 634 (2007) 205–219

An investigation of the photo-clastogenic potential of ultrafine titanium dioxide particles Elizabeth Theogaraj a,1 , Susan Riley a , Laurie Hughes b , Monika Maier c , David Kirkland a,∗ a

c

Covance Laboratories Ltd., Otley Road, Harrogate HG3 1PY, UK b Uniqema, Wilton Centre, Wilton, Redcar TS10 4RF, UK Degussa GmbH, Rodenbacher Chaussee 4, D-63457 Hanau, Germany

Received 12 February 2007; received in revised form 22 May 2007; accepted 10 July 2007 Available online 6 August 2007

Abstract Ultrafine titanium dioxide is widely used in a number of commercial products including sunscreens and cosmetics. There is extensive evidence on the safety of ultrafine titanium dioxide. However, there are some published studies indicating that some forms at least may be photogenotoxic, photocatalytic and/or carcinogenic. In order to clarify the conflicting opinions on the safety of ultrafine titanium dioxide particles, the current studies were performed to investigate the photo-clastogenic potential of eight different classes of ultrafine titanium dioxide particles. The photo-clastogenicity of titanium dioxide was measured in Chinese hamster ovary (CHO) cells in the absence and presence of UV light at a dose of 750 mJ/cm2 . The treatments were short (3 h) followed by a 17-h recovery and achieved concentrations that either induced approximately 50% cytotoxicity or reached 5000 ␮g/ml if non-cytotoxic. None of the titanium dioxide particles tested induced any increase in chromosomal aberration frequencies either in the absence or presence of UV. These studies show that ultrafine titanium dioxide particles do not exhibit photochemical genotoxicity in the model system used. © 2007 Elsevier B.V. All rights reserved. Keywords: Titanium dioxide; Nanoparticles; Photogenotoxicity

1. Introduction Nanotechnology is an emerging field, the benefits of which are widely publicized. Nanomaterials are present in a number of commercially available products including sunscreens, cosmetics and many industrial ∗

Corresponding author. Tel.: +44 1423 848401; fax: +44 1423 848983. E-mail address: [email protected] (D. Kirkland). 1 Current address: F. Hoffmann-LaRoche Ltd., Pharmaceuticals Division, Regulatory Program Management, Building 663, CH-4070 Basel, Switzerland. 1383-5718/$ – see front matter © 2007 Elsevier B.V. All rights reserved. doi:10.1016/j.mrgentox.2007.08.002

applications, but there are uncertainties as to whether the unique properties that support their commercial use may also pose potential occupational health risks [1]. Nanomaterials have a high surface-to-volume ratio, so surface reactivity will be high. These particles may adopt structures that are different from the bulk form of the chemical and, therefore, may exhibit different chemical and physical properties [2]. Ultrafine titanium dioxide is a nanostructure widely used in industry [3]. It occurs in three different crystalline forms (rutile, anatase and brookite). The rutile and anatase grades are of commercial importance, representing 90% and 10% of the market, respectively.

206

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

Owing to its high refractive index, ultrafine titanium dioxide has light-scattering properties, hence its use in protecting against UV exposure [4]. Many marketed sunscreen products contain ultrafine titanium dioxide, surface-treated with either inorganic or organic coatings, which are colorless and reflect and scatter UV more effectively than larger particles. These coatings assist in the dispersion of titanium dioxide particles in cosmetic or sunscreen formulation (hydrophilic or hydrophobic) and can enhance the photostability of sunscreens and prevent reduction of titanium trioxide that could lead to “greying” of the pigment [5]. Doping of the titanium dioxide, i.e., replacing some titanium oxide with a small percentage of a transition metal oxide, can also serve to enhance photostability. Some studies have suggested that ultrafine titanium dioxide can lead to formation of free radicals and oxidative damage to DNA [6]. Micrometer-sized particles of titanium dioxide can penetrate the human stratum corneum [7], which could lead to photocatalysis, DNA damage and hence skin cancer, following exposure to sunlight although the majority of published investigations do not show penetration of ultrafine titanium dioxide beyond the strateum corneum [8,9]. Cai et al. [10] investigated the cytotoxic effects of ultrafine anatase and photo-excited anatase titanium dioxide particles on HeLa cells. Large aggregated titanium dioxide particles were found to be highly cytotoxic to HeLa cells in the absence of UV light, possibly because the particles deposited and covered the cells. When HeLa cells were exposed to titanium dioxide (12–120 ␮g/cm3 ) in the presence of UV light (500-W Hg lamp) a dose-dependent enhancement of killing of HeLa cells was seen following UV exposure, compared with controls, i.e., cells exposed to UV light in the absence of titanium dioxide. The authors proposed two mechanisms by which titanium dioxide induces cytotoxicity in HeLa cells. First, there may be a direct oxidation by photo-excited titanium dioxide on the cells. Second, photo-excited titanium oxide may cause the formation of reactive OH radicals, which can attack the cell membrane and interfere with oxidation–reduction substrates (e.g. glutathione, nicotinamide adenine dinucleotide [NADH] and coenzyme A) that are important for ATP production, thereby leading to cell death [10]. In support of the latter mechanism, previous studies have shown that titanium dioxide in the presence of UV light generates H2 O2 , OH and O2 radicals (cited in Ref. [11]). Uchino et al. [11] investigated the cytotoxicity of various crystal forms and sizes of titanium dioxide following UVA irradiation in CHO cells by measuring the generation of OH radicals. Interestingly, the anatase form

of titanium dioxide caused a dose-dependent increase in the production of OH radicals compared with the rutile form, which produced lower levels of OH radicals. There was also a relationship between OH radical formation and titanium dioxide crystal size: the optimal size for OH radical generation was approximately 30 and 90 nm for the anatase and rutile forms, respectively [11]. Nakagawa et al. [12] reported that four different titanium dioxide particles (two anatase and two rutile) were able to induce DNA damage, as measured by the Comet assay, in L5178Y mouse lymphoma cells in the presence of simulated solar light. Three of the four titanium dioxide particle types tested were much more cytotoxic in the presence of UV than in its absence. The remaining one particle type, despite showing little difference in cytotoxicity in the absence or presence of UV, produced only a borderline comet induction. Reports of titanium dioxide being genotoxic in the absence of UV light are inconsistent [13–15]. A recent study by Falck et al. [16], found that the rutile form of titanium dioxide induced a dose-dependent increase in DNA damage (Comet assay) in human bronchial epithelial (BEAS 2B) cells at dose levels of 1–100 ␮g/ml. As this was only reported in an abstract, many details (e.g. particle size range) are not available. Wang et al. [17] reported that extracts of titanium dioxide prepared in tissue culture medium induced micronuclei, comets and HPRT mutations in human lymphoblastoid WIL2-NS cells. It is difficult to comprehend what might have been extracted from 99% pure titanium dioxide that could cause toxicity and genotoxicity in mammalian cells. The maximum responses for all three endpoints were in the range of two to three times control, but as no historical control data were presented, the biological significance of these findings cannot be accurately assessed. However, despite reports that the rutile form of titanium dioxide is relatively poor at generating reactive oxygen species [11] it would appear to be genotoxic in the dark under some conditions. A study by Rahman et al. [18] demonstrated that ultrafine titanium dioxide (≤20 nm) significantly increased the number of micronuclei (MN) at all concentrations tested (0.5–5.0 ␮g/cm2 ) in Syrian Hamster embryo (SHE) cells. Although the formation of MN increased with exposure from 12–24 h, there was no further increase at 48 and 72 h later, suggesting a saturation effect. As kinetochore analysis revealed no significant increase in kinetochore-positive MN compared with unexposed controls, it suggests that the MN have mainly arisen from a clastogenic effect [18]. However, again details of the form of titanium dioxide particle used

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

(anatase or rutile) and the measures of toxicity are not known. In some long-term carcinogenicity studies in rats, inhaled titanium dioxide and intratracheally instilled titanium dioxide caused benign and malignant lung tumors [19–21]. We have, therefore, tested several different forms of rutile and anatase ultrafine titanium dioxide particles, for photo-clastogenic potential in cultured Chinese hamster cells in accordance with recent recommendations [22,23]. Of the available tests for photogenotoxicity, more chemicals are positive for photoclastogenicity than in other tests, and therefore this approach is considered the most sensitive, efficient and reliable in vitro test for these evaluations. This paper summarizes the photo-clastogenic findings of eight different forms (five surface-coated and three without surface treatment) of ultrafine titanium dioxide particles. 2. Materials and methods 2.1. Test samples The samples indicated in Table 1 were supplied as follows: • Samples A–C were supplied by Degussa AG, Hanau, Germany. • Samples D–H were supplied by various companies, members of the Physical Sunscreen Manufacturers Association, Brussels, Belgium. Coated, doped and uncoated materials were used in this study. Descriptions, particle sizes and coatings are given in Table 1. The terminology used in this paper refers to ultrafine titanium dioxide. There is no full standardization of nomenclature for nanomaterials and in this paper, the particle sizes in Table 1 are the shortest dimension of the primary particles or crystals, which can be measured and identified using a variety of dif-

207

ferent techniques. These individual primary particles/crystals rarely exist independently in dispersions and typically they form aggregates of particles in the dimension of 30–150 nm. The samples tested will have formed aggregates of this dimension. 2.2. Cell culture CHO-WBL cells, supplied by Dr. S. Galloway, West Point, PA, USA, were maintained at Covance Laboratories Limited in tissue–culture flasks containing McCoy’s 5A medium (Gibco, Paisley, Scotland) including 10% (v/v) foetal calf serum (FCS) (Gibco, Paisley, Scotland) and 100 ␮g/ml gentamycin (Gibco, Paisley, Scotland). Stocks of cells were preserved in liquid nitrogen and reconstituted for each experiment to maintain karyotypic stability. The modal chromosome number for the clone of CHO cells used at the testing laboratory was 21. After cell stocks had been reconstituted from frozen samples, cell sheets were removed using trypsin/EDTA solution, and 25-cm2 tissue–culture flasks were seeded with approximately 2.5 × 105 cells per flask in 4.95 ml culture medium. The cells were incubated at 37 ± 1 ◦ C in an atmosphere of 5% (v/v) CO2 in air, and were usually treated on the following day, by which time the cultures were approximately 30–60% confluent. 2.3. UV irradiation An Atlas Suntest® CPS solar simulator light source was used (Heraeus Equipment Limited, Brentwood, UK). The lamp irradiated six 25-cm2 culture flasks at a time within an enclosed and temperature-controlled area. The temperature of the irradiation area was 35 ◦ C and was maintained by a SunCool® Cooling unit attached to the Atlas system. The intensity of UVA and UVB was measured with an Osram Centra UV meter. An ultra-violet glass filter was added to the lamp to remove the UVC component of the light (wavelengths < 290 nm). The filtering effect of the tissue-culture flask plastic and the growth medium covering the cells were measured and taken into account for the calculation of the doses delivered. The ratio of

Table 1 Description of ultrafine titanium dioxide particles tested Sample code

Crystal type

Inorganic coating

Organic coating

Particle size

A B C D E F G H

Anatase (80%), rutile (20%) Anatase (80%), rutile (20%) Anatase (80%), rutile (20%) Rutile (100%) Anatase (100%) Rutile (100%) Rutile (100%) Rutile (100%)

None None, doped di-iron trioxide (2 ± 1%) None Alumina (8–11%) Alumina (37%), silica (12–18%) Alumina (5–6.5%) Alumina (3–8%) Alumina (10.5–12.5%), silica (3.5–5%)

Trimethoxy caprylylsilane None None Simethicone (1–3%) None Dimethicone (1–4%) Stearic acid (5–11% None

Approximately 21 nma Approximately 21 nma Approximately 21 nma 14 nmb 60 nmc 20 nmb 15 nma 20–22 nmb

a b c

Primary particle size determined by transmission electron microscopy (TEM). Primary particle size determined by X-ray diffraction. Characterisation by X-ray disc centrifugation (XDC) giving an aggregate rather than particle size.

208

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

UVA:UVB irradiation delivered to the cells was approximately 5:1.

with 5% Giemsa solution, air-dried and mounted permanently.

2.4. Chromosome aberration tests

2.5. Scoring of aberrations

A preliminary range finder, covering a broad range of doses was performed in the presence of a high dose of UV light (750 mJ/cm3 ) to investigate the phototoxicity of each type of test particle and to determine the dose range to be used in the main study. The PSMA particles (D–H) were dissolved in dimethyl sulphoxide (DMSO), the other particles were suspended as follows:

Slides were coded before microscopic analysis. Two experienced scorers were involved, each scoring half the cells from each replicate culture, so as to control for bias. One hundred cells per replicate (i.e. two hundred metaphases from each treatment condition) were analyzed (where appropriate) for structural chromosome aberrations. Slide analysis was halted if a minimum of 10 aberrant cells (excluding gaps) per slide was observed. The modal chromosome number for the clone of CHO cells used was 21. Consequently, only cells with 19–23 chromosomes were considered acceptable for analysis. Cells with more than 23 chromosomes were noted and recorded separately as hyperdiploid, polyploid or endoreduplicated cells. Classification of structural aberrations was based on the scheme described by ISCN [24].

• A in absolute ethanol and • B and C in physiological saline. After 15 min incubation in the dark (to allow for penetration of the particles into the cells), the cultures were exposed to achieve the dose of UV as described above. After irradiation, cultures were incubated in the dark until 3 h after the start of chemical treatment, after which the culture medium was removed, cultures washed with sterile saline, refed with fresh medium and incubated for a further 17 h. At the time of harvest (20 h after the start of treatment), cells were removed with trypsin/EDTA and counted with a Coulter counter. On the basis of cell-count reductions in the absence of UV, a top concentration that induced close to approximately 50% toxicity (compared to concurrent negative controls) was chosen. In some cases (e.g. particles B and C) the toxicity at the top concentration somewhat exceeded 50%, but as no chromosomal aberrations were induced even at these higher levels of toxicity, this was not considered a problem. In other cases there was less than 50% toxicity, and in those cases a top concentration of approximately 5000 ␮g/ml was selected, in accordance with standard regulatory testing procedures. The concentrations selected in the absence of UV were also tested in the presence of UV. Treatments for the main experiments were performed as above and included negative (solvent) controls, but also 8-methoxypsoralen (8-MOP) and 4-nitroquinoline-1-oxide (NQO) as positive controls. At the time of harvest, the monolayers were again removed using trypsin/EDTA. Although duplicate cultures were established for test chemical and positive control treatments, quadruplicate cultures were established in tissue culture flasks for the negative control treatments. The extra control cultures were to provide additional cells for scoring in the event of an equivocal response with one of the test materials, but in fact they were not needed and only duplicate control cultures were scored. An aliquot of cells was counted to assess toxicity at each concentration. The remaining suspension from each flask was transferred to a plastic centrifuge tube and the cells pelleted by centrifuging at 200 × g for 5 min. The supernatant was removed and the cells re-suspended in 4 ml of 0.075 M KCl at 37 ◦ C for 5 min and fixed in cold methanol/glacial acetic acid (3:1 v/v). Slides were prepared by standard methods, stained

2.6. Analysis of results and statistical analysis For each individual irradiation dose and condition, the proportion of cells for each test treatment was compared with the proportion in concurrent negative controls using Fisher’s exact test. Probability values of p ≤ 0.05 were considered significant. Where chromosome aberration frequencies in treated cultures were statistically different from control and exceeded historical control ranges, they were considered positive.

3. Results The results of the photo-clastogenicity studies with the eight different types of ultrafine titanium dioxide are summarized in Table 2, and given in detail in Tables 3–10. In each experiment the top concentration for testing was either 5000 ␮g/ml (the top concentration required for non-toxic substances of unknown molecular weight) or one that resulted in close to or greater than 50% toxicity. The 45% cytotoxicity induced in the absence of UV for sample A was considered close enough to 50% to be acceptable for a valid test. For particles B and C, toxicity greater than 50% was observed, but as there were no increases in chromosomal aberration frequency, this is not considered a problem. When testing different types of chemicals for photogenotoxicity, treatments are often more cytotoxic in the presence than in the absence of UV light. In our studies there was no marked enhancement of toxicity in the presence of UV light, in fact for seven of the eight particle types toxicity was reduced in the presence of UV light, and this may be due to a photo-protective effect of the titanium dioxide itself. It was therefore possible to use the same concentration ranges of tita-

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

209

Table 2 Summary of photo-clastogenicity data for eight ultrafine titanium dioxide particles Particle code

CA result (±) and toxicity at LEDa or HDTa 750 mJ/cm2 UV

No UV

A B C D E F G H

Result (␮g/ml)

Cytotoxicity (%)

Result (␮g/ml)

Cytotoxicity (%)

– (800) – (800) – (800) – (3000) – (5000) – (5000) – (1950) – (5000)

45

– (800) – (800) – (800) – (3000) – (5000) – (5000) – (1950) – (5000)

25

78 75 57 22 38 58 49

54 71 5 32 0 55 48

a LED = lowest effective dose (␮g/ml) for a positive response. HDT = highest dose tested (␮g/ml) for a negative response. (either 50% cytotoxicity or 5000 ␮g/ml whichever the lowest). CA = chromosome aberration test in CHO cells.

nium dioxide, both in the dark and under UV irradiation conditions. In all experiments, frequencies of chromosomal aberrations in solvent control cultures were low and within the historical negative control range (see Table 11). In addition, for all experiments, the positive control chemical NQO-induced significant increases in aberration frequencies in the absence of UV, and 8-MOP induced significant increases in aberration frequencies in its presence. It can be seen in Tables 2–10 that none of the eight ultrafine titanium dioxide (five surface treated and three without surface treatment) particles induced any increase in chromosomal aberration frequencies either in the absence or presence of UV. 4. Discussion The results of this study demonstrate that none of the eight different forms of uncoated, coated and doped ultrafine titanium dioxide particles was able to induce increases in the frequency of chromosome aberrations either in the absence or the presence of UV. Our data are not consistent with the findings of Nakagawa et al. [12], Falck et al. [16] or Wang et al. [17] and these differences are discussed below. In the studies by Nakagawa et al. [12] the two anatase and two rutile forms of titanium dioxide were able to induce DNA damage (Comets) in L5178Y cells

in the presence but not in the absence of solar simulated light. However, clear increases in comet tail length were only evident at concentrations where cell survival was 70% or less (estimated from graphs). International recommendations for the conduct of the Comet assay advise that concentrations inducing >30% cytotoxicity should be avoided because of the potential for artefacts [25]. Nakagawa et al. [12] also showed that one of the anatase particles tested for photo-clastogenicity produced increased chromosomal aberration frequencies under treatment conditions (in the presence of UV light) that resulted in >50% cytotoxicity. In contrast to our findings, this paper reported much greater levels of cytotoxicity in the presence of UV. It is widely accepted that chromosomal aberrations can be formed as an indirect consequence of high levels of cell killing [26–33]. It is particularly appropriate to comment that clastogenicity may be an indirect consequence of cytotoxicity when normal aberration frequencies are seen at levels of toxicity slightly less than 50%, which was the case with this particular anatase particle. Thus, the biological relevance of the induction of comets and chromosomal aberrations in the presence of UV as published by Nakagawa et al. [12] may be questioned, and the reported effects may be attributed to an indirect consequence of high levels of cytotoxicity. In our studies (Table 2) there was no induction of aberrations even with those particles that induced >50% cytotoxicity (samples B, C and G).

210

Table 3 Detailed data for particle A

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

Table 4 Detailed data for particle B

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

211

212

Table 5 Detailed data for particle C

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

Table 6 Detailed data for particle D

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219 NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range. (*)Excluding gaps, (**)Solubility problems above 3000 ␮g/ml.

213

214

Table 7 Detailed data for particle E

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NR: not representative; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

Table 8 Detailed data for particle F

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

215

216

Table 9 Detailed data for particle G

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

Table 10 Detailed data for particle H

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

NT: not tested; NS: not significant; ND: not done; cell counts not performed. Shading indicates values exceed historical negative control range.

217

218

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219

Table 11 Historical solvent control data for non-irradiated CHO cultures S9 treatment

S9

Category

Structural aberrations including gaps Structural aberrations excluding gaps Numerical aberrations (i.e. outside of 19–23 range)

In the study of Falck et al. [16] there are only data presented on induction of micronuclei in the absence of UV light. As these results were presented in an abstract at a scientific meeting, details such as the homogeneity of the particle suspension, how cytotoxicity was measured, and the extent of toxicity at the concentrations showing increased micronucleus frequencies are not known/available. Micronuclei can be induced by chemicals that disrupt the mitotic spindle [34], so it is possible that the physical presence of large aggregates of titanium dioxide particles affected the division process in these cells leading to chromosome loss rather than chromosome breakage. Our findings of a lack of chromosome breakage after treatment in the absence of UV would be consistent with such an aneugenic effect rather than a clastogenic effect. In another study by Wang et al. [17] treatments with titanium dioxide were performed only in WIL2-NS cells, and again only in the absence of UV light. This study showed that an extract of 99% pure titanium dioxide was found to be highly toxic at 130 ␮g/ml. It is not clear how this could occur. Increases in micronucleated cells, comet tails and HPRT mutations were recorded, but only reached a maximum of two to three times concurrent controls. Since no historical control data were presented for these endpoints in WIL2-NS cells, the biological significance of these findings cannot be accurately assessed. The methodology used in our studies does merit some discussion, as the design deviates twice from OECD guidelines [35]. The deviations were that no treatment in the presence of metabolic activation (S9) was included, nor was a continuous treatment (e.g. 20 h) performed. As the main purpose of photogenotoxicity testing is to make an assessment of the potential of a compound to turn into a photochemical carcinogen upon activation with UV light, whether the UV exposure is short or continuous is of no consequence. The omission of S9 mix in in vitro phototoxicity tests was justified because it has been shown that the addition of material with a high protein content such as S9 in cell cultures can absorb or scatter light in the UV region and thus protect the cells from the possible phototoxic and photogenotoxic effects [23,36]. Also, there are no known examples of photochemical

Total number of cells scored

8547 8547 8693

Aberrant cells scored per 100 cells Mean

Normal range

2.58 1.50 1.66

0–8 0–6 0–6

genotoxicity being induced only after metabolic activation [23]. In addition, it is not expected that titanium dioxide would be metabolized to an electrophilic intermediate that would react with DNA, however, it is not known if the organic and/or inorganic coatings would be metabolized. Earlier investigations suggested that ultrafine titanium dioxide may be a photogenotoxin or a photocatalyst. The studies presented in this paper, which were performed in a rigorous and carefully controlled manner, indicate that eight different rutile and anatase forms of titanium dioxide with different surface treatments (five surface-treated and three without surface treatment) are not activated to photogenotoxins by solar simulated (UVA + UVB) light. Acknowledgements The authors would like to thank the Physical Sunscreen Manufacturing Association for their sponsorship and the technicians in the genetic and molecular toxicology department at Covance Laboratories Ltd., Harrogate, UK for their excellent technical assistance. References [1] P.H.M. Hoet, I. Br¨uske-Hohlfeld, O.V. Salata, Nanoparticles— known and unknown health risks, J. Nanobiotechnol. 12 (2004) 15 (published online). [2] P.J.A. Borm, D. Robbins, S. Haubold, T. Kuhlbusch, H. Fissan, K. Donaldson, R. Schins, V. Stone, W. Kreyling, J. Lademann, J. Krutmann, D. Warheit, E. Oberdorster, The potential risks for nanomaterials: a review carried out for ECETOC, Particle Fibre Toxicol. 3 (11) (2006) 35 (published online). [3] R.C. Rowe (Ed.), Handbook of Pharmaceutical Excipients, 4th ed., American Pharmaceutical Association/The Pharmaceutical Press, Washington, DC, USA/London, UK, 2003, pp. 89–92, 663–664. [4] P. Alexander, Ultrafine titanium dioxide makes the grade, Manuf. Chem. 62 (1991) 21–23. [5] J.P. Hewitt, Titanium dioxide: a different kind of sunshield, Drug Cosmet. Ind. 151 (1992) 26–32. [6] W.G. Wamer, J.J. Yin, R.R. Wei, Oxidative damage to nucleic acids photosensitised by titanium dioxide, Free Radic. Biol. Med. 23 (1997) 851–858.

E. Theogaraj et al. / Mutation Research 634 (2007) 205–219 [7] J. Lademann, H.-J. Weigmann, C. Rickmeyer, H. Barthelmes, H. Schaefer, G. Mueller, W. Sterry, Penetration of titanium dioxide microparticles in a sunscreen formulation into the horny layer and the follicular orifice, Skin Pharmacol. Appl. Skin Physiol. 12 (1999) 247–256. [8] S. Gottbrath, C.C. M¨uller-Goymann, Penetration and visualization of titanium dioxide microparticles in human stratum corneum, effect of different formulations on the penetration of titanium dioxide, SOFW J. 3 (129) (2003) 11–17, ISSN 09427694. [9] A.O. Gamer, E. Leibold, B. van Ravenzwaay, The in vitro absorption of microfine zinc oxide and titanium dioxide through porcine skin, Toxicol. In Vitro 20 (2005) 301–307. [10] R. Cai, K. Hashimoto, K. Itoh, Y. Kubota, A. Fujishima, Photokilling of malignant cells with ultrafine TiO2 powder, Bull. Chem. Soc. Jpn. 64 (1991) 1268–1273. [11] T. Uchino, H. Tokunaga, M. Ando, H. Utsumi, Quantitative determination of OH radical generation and its cytotoxicity induced by induced by TiO2 —UVA treatment, Toxicol. In Vitro 16 (2002) 629–635. [12] Y. Nakagawa, S. Wakuri, K. Sakamoto, N. Tanaka, The photogenotoxicity of titanium dioxide particles, Mutat. Res. 394 (1997) 125–132. [13] J.L. Ivett, B.M. Brown, C. Rodgers, D.E. Anderson, M.A. Resnick, E. Zeiger, Chromosomal aberrations and sister chromatid exchanges in Chinese hamster ovary cells in vitro. IV. Results with 15 chemicals, Environ. Mol. Mutagen. 14 (1989) 165–187. [14] B.C. Myhr, W.J. Caspary, Chemical mutagenesis at the thymidine kinase locus in L5178Y mouse lymphoma cells: results for 31 coded compounds in the National Toxicology Program, Environ. Mol. Mutagen. 18 (1991) 51–83. [15] P.-J. Lu, I.-C. Ho, T.-C. Lee, Induction of sister chromatid exchanges and micronuclei by titanium dioxide in Chinese hamster ovary-K1 cells, Mutat. Res. 414 (1998) 15–20. [16] G.C-M. Falck, H. Scandvik, S. Suhonen, H. Norppa, Induction of DNA damage and micronuclei by nanosized titanium dioxide in BEAS 2B cells, Environ. Mol. Mutagen. 47 (2006) 391–482 (Abstract). [17] J.J. Wang, B.J.S. Sanderson, H. Wang, Cyto- and genotoxicity of ultrafine TiO2 particles in cultured human lymphoblastoid cells, Mutat. Res.: Genet. Toxicol. Environ. Mutagen. 628 (2007) 99–106. [18] Q. Rahman, M. Lohani, E. Dopp, H. Pemsel, L. Jonas, D.G. Weiss, D. Schiffmann, Evidence that ultrafine titanium dioxide induces micronuclei and apoptosis in Syrian hamster embryo fibroblasts, Environ. Health Perspect. 110 (2002) 797–800. [19] U. Heinrich, R. Fuhst, R. Rittinghausen, O. Creutzenberg, B. Bellmann, W. Koch, K. Levsen, Chronic inhalation exposure of Wistar rats and two different strains of mice to diesel exhaust, carbon black, and titanium dioxide, Inhal. Toxicol. 7 (1995) 533–556. [20] F. Pott, M. Roller, Carcinogenicity study with nineteen granular dusts in rats, Eur. J. Oncol. 10 (2005) 249–281. [21] U. Mohr, H. Ernst, M. Roller, F. Pott, Pulmonary tumor types induced in Wistar rats of the so-called “19-dust study”, Exp. Toxicol. Pathol. 58 (2006) 13–20.

219

[22] S.W. Dean, M. Lane, R.H. Dunmore, S.P. Ruddock, C.N. Martin, D.J. Kirkland, N. Loprieno, Development of assays for the detection of photomutagenicity of chemicals during exposure to UV light. I. Assay development, Mutagenesis 6 (1991) 335–341. [23] E. Gocke, L. M¨uller, P.J. Guzzie, S. Brendler-Schwaab, S. Bulera, C.F. Chignell, L.M. Henderson, A. Jacobs, H. Murli, R.D. Snyder, N. Tanaka, Considerations on photochemical genotoxicity: report of the international workshop on genotoxicity test procedures working group, Environ. Mol. Mutagen. 35 (2000) 173–184. [24] F. Mitelman (Ed.), ISCN: An International System for Human Cytogenetic Nomenclature, Recommendations of the International Standing Committee on Human Cytogenetic Nomenclature, Karger, Switzerland, 1995. [25] R.R. Tice, E. Agurell, D. Anderson, B. Burlinson, A. Hartmann, H. Kobayashi, Y. Miyamae, E. Rojas, J.C. Ryu, Y.F. Sasaki, Single cell gel/comet assay: guidelines for in vitro and in vivo genetic toxicology testing, Environ. Mol. Mutagen. 35 (2000) 206–221. [26] D.J. Kirkland, Chromosomal aberration tests in vitro: problems with protocol design and interpretation of results, Mutagenesis 7 (1992) 99–106. [27] C.A. Hilliard, M.J. Armstrong, C.I. Bradt, R.B. Hill, S.K. Greenwood, S.M. Galloway, Chromosome aberrations in vitro related to cytotoxicity of nonmutagenic chemicals and metabolic poisons, Environ. Mol. Mutagen. 31 (1998) 316–326. [28] L. M¨uller, Y. Kikuchi, G. Probst, L. Schechtman, H. Shimada, T. Sofuni, D. Tweats, ICH-harmonised guidances on genotoxicity testing of pharmaceuticals: evolution, reasoning and impact, Mutat. Res. 436 (1999) 195–225. [29] S.M. Galloway, Cytotoxicity and chromosome aberrations in vitro: experience in industry and the case for an upper limit on toxicity in the aberration assay, Environ. Mol. Mutagen. 35 (2000) 191–201. [30] L. M¨uller, T. Sofuni, Appropriate levels of cytotoxicity for genotoxicity tests using mammalian cells in vitro, Environ. Mol. Mutagen. 35 (2000) 202–205. [31] D.J. Kirkland, L. M¨uller, Interpretation of the biological relevance of genotoxicity test results: the importance of thresholds, Mutat. Res. 464 (2000) 137–147. [32] L. M¨uller, P. Kasper, Human biological relevance and the use of threshold-arguments in regulatory genotoxicity assessment: experience with pharmaceuticals, Mutat. Res. 464 (2000) 19–34. [33] V. Thybaud, M. Aardema, J. Clements, K. Dearfield, S. Galloway, M. Hayashi, D. Jacobson-Kram, D. Kirkland, J.T. MacGregor, D. Marzin, W. Ohyama, M. Schuler, H. Suzuki, E. Zeiger, Strategy for genotoxicity testing: hazard identification and risk assessment in relation to in vitro testing, Mutat. Res. 627 (2007) 41–58. [34] J.M. Parry, An evaluation of the use of in vitro tubulin polymerisation, fungal and wheat assays to detect the activity of potential chemical aneugens, Mutat. Res. 287 (1993) 23–28. [35] OECD, OECD Guidelines for the Testing of Chemicals: Genotoxicity. Organisation for Economic Co-operation and Development, Paris 1997. (Revised and new guidelines adopted 1997). [36] European Medicines Evaluation Agency, Note for guidance on photosafety testing, 2002, CPMP/SWP/398/01.