Slags, ilmenite electrothermal smelting

Carcinogenicity

Justification for classification or non-classification

It is inappropriate to base the evaluation of titanium dioxide as a suspect carcinogen solely on the observation that rats develop lung tumours under condition of “lung particle overload”, since such tumours induced in rats by inert poorly soluble particles such as titanium dioxide are widely considered as unreliable predictors of hazard to humans.

Overall, the epidemiological evidence from well-conducted investigations has not shown that exposure to titanium dioxide is not correlated to any detectable carcinogenic potential for humans. Therefore, under the EU system, no classification is warranted given that the mechanism of tumour formation has been clearly identified, with good evidence that this process cannot be extrapolated to man.

Under the GHS system, classification is similarly not warranted given that the tumour responses are seen in a uniquely sensitive species, the rat; tumours in rats only occur at excessive doses (lung overload) and in the presence of chronic lung inflammation and similar responses are not observed in other species including humans. For the reasons presented above, no classification for carcinogenicity is required.

Additional information

UGI (Upgraded Ilmenite) consists primarily of a titanate phase (solid solution) most of which is Ti in an oxidised form. Upon inhalation, a low rate of dissolution in the respiratory tract is assumed, based on the experimental verified inertness of the material in artificial respiratory media. Any material being released from UGI under physiological conditions will be in the form of ionic Ti, which is similarly the case for titanium dioxide: more specifically, the results from in vitro bioaccessibility testing in such fluids demonstrate a similar dissolution pattern of UGI and titanium dioxide (see summary reported under point 7.1.1 basic toxicokinetics), thus read-across from inhalation carcinogenicity data on titanium dioxide is considered feasible without any restrictions.

Both TiO2 and UGI strictly fulfil the criteria for “poorly soluble particles”, which are known to elicit lung carcinogenicity in laboratory rat studies though a mechanism associated with particle overload. However, such lung tumours formed under condition of “lung particle overload” are widely considered as unreliable predictors of hazard to humans. Overall, the epidemiological evidence from well-conducted investigations has shown that exposure to titanium dioxide is not correlated to any detectable carcinogenic potential for humans. Therefore, under the EU system, no classification is warranted given that the mechanism of tumour formation has been clearly identified, with good evidence that this process cannot be extrapolated to man. Under the GHS system, classification is similarly not warranted given that the tumour responses are seen in a uniquely sensitive species, the rat; tumours in rats only occur at excessive doses (lung overload) and in the presence of chronic lung inflammation and similar responses are not observed in other species including humans. For the reasons presented above, no classification for carcinogenicity is required for TiO2 or for UGI.

Chronic Studies: pigment-grade titanium dioxide

In chronic exposure studies, lung tumors were produced in rats under conditions of particle overload. In a 2-year inhalation study in male and female rats, exposures to titanium dioxide articles (rutile type) at concentrations of 10, 50, or 250 mg/m3 produced lung tumors only at the highest concentration (Lee et al., 1985). With one exception, the tumors produced were ultimately characterized as primarily benign pulmonary keratin cysts (Warheit and Frame, 2006). In a study reported by Muhle et al. (1991), titanium dioxide was used as a negative control dust in a two-year inhalation study with toner particles. Male and female rats were exposed (6 hr/day, 5 days/week) to 5 mg/m3 titanium dioxide (rutile form) of 1.1μm MMAD2 with a respirable fraction of 78%. There were no significant increases in lung tumors vs. control rats exposed for up to 24 months by whole body inhalation to titanium dioxide in this study.

Chronic Studies: sub-pigmentary grade titanium dioxide

Heinrich et al. (1995) exposed female rats by whole body inhalation to sub-pigmentary titanium dioxide (80%anatase: 20% rutile) at an average concentration of 10 mg/m3 for 24 months followed by 6 months without exposure. The particle size of the titanium dioxide used ranged from 15 to 40 nm with a MMAD of 0.8μm (agglomerates of sub-pigmentary particles). Statistically significant increases in tumors vs. controls were observed in rats from this study (benign keratinizing cystic squamouscell tumors, adenocarcinomas, squamous-cell carcinomas, and adenomas). Exposure of female mice to sub-pigmentary titanium dioxide under the same conditions as for rats resulted in a significantly decreased lifespan at an inhalation concentration of approximately 10 mg/m3 (Heinrich et al., 1995). However, tumour rates were not statistically increased over prevalence in controls.

Subchronic inhalation studies (reported under the endpoint: repeated dose toxicity)

Subchronic inhalation exposures of rats, mice, and hamsters to either pigmentary or sub-pigmentary titanium dioxide particles at concentrations likely to induce particle overload produced a more severe and persistent pulmonary inflammatory response in rats, as compared with either mice or hamsters. Rats were unique among these three species in the development of progressive fibroproliferative lesions and alveolar epithelial metaplasia (Bermudez et al., 2002 and 2004). Thus, female rats, mice or hamsters were exposed to 10, 50 or 250 mg/m3 concentrations of pigmentary (rutile type) titanium dioxide particles for 6 hours/day, 5 days/week for 13 weeks followed by 4, 13, 26 or 52 weeks of post exposure (46 weeks for hamsters)(Bermudez et al., 2002). Lung and associated lymph node loads of titanium dioxide increased in a concentration-related manner. Retained lung burdens were greatest in mice following exposure, with rats and hamsters displaying similar lung burdens immediately following exposure. Particle retention data indicated that particle overload in the lungs was reached in both rats and mice at the 50 and 250 mg/m3 concentrations. Inflammation was observed in all three species at the two highest concentrations. This inflammation persisted in rats and mice throughout the post exposure recovery period at the highest exposure concentration. In hamsters, inflammatory responses were eventually resolved due to the more rapid clearance of particles from the lung. In rats exposed to the highest concentration (250 mg/m3), pulmonary lesions consisted of epithelial proliferative changes manifested by increased alveolar epithelial cell labelling indices, as evidenced by the results of cell proliferation studies. Associated with these proliferative changes in the rat were increased interstitial particle accumulations and alveolar septal fibrosis. Although rats exposed to 50 mg/m3 developed minimal alveolar cell hypertrophy, accumulation of particle-laden macrophages, and inflammation, no alveolar septal fibrosis or relevant cell turnover at alveolar sites were observed at this lower exposure concentration. Similar changes to those seen in rats were not observed in either mice or hamsters.

In a study with sub-pigmentary titanium dioxide (80% anatase: 20% rutile; average primary particle size = 21 nm), female rats, mice or hamsters were exposed to aerosol concentrations of 0.5, 2.0 or 10 mg/m3 titanium dioxide for 6 hours/day, 5 days/week for 13 weeks followed by 4, 13, 26 or 52 weeks of postexposure (49 weeks for hamsters)(Bermudez et al., 2004). Retained lung burdens increased in a concentration-related manner in all three species. Mice and rats had similar lung burdens at the end of exposures but hamsters were significantly lower. Retardation of particle clearance in rats and mice at the highest exposure concentration (10 mg/m3) indicated that pulmonary particle overload had been achieved. Lesions of the lungs in rats consisted of foci of alveolar epithelial proliferation of metaplastic epithelial cells (alveolar bronchiolization) concomitant with circumscribed areas of heavy, particle-laden macrophages. In rats these changes were manifested by increased alveolar epithelial cell labelling indices, as evidenced by cell proliferation studies. Associated with these foci were areas of interstitial particle accumulation and alveolar septal fibrosis. These lesions observed in the rat became more pronounced with time. Mice developed a less severe inflammatory response without the progressive epithelial and fibroproliferative changes.

Summary of Animal Data

Inhalation exposures to titanium dioxide under conditions causing pulmonary overload have produced primarily benign lung tumours in rats, but no tumours in other common laboratory rodents. A number of conclusions can be reached concerning the lack of relevance of the rat lung tumour data for the carcinogenicity classification of titanium dioxide.

Lung overload in the rat produces excessive toxicity and tumours (reported under the endpoints: repeated dose toxicity, specific investigations and additional toxicological information)

Prolonged inhalation exposures in rats to high levels of titanium dioxide produces delayed alveolar lung clearance and increased retention of particles (Bermudez et al., 2002 and 2004). This condition is described as “lung particle overload” (Mauderly, 1994 and 1996; ILSI, 2000; Donaldson, 2000) and leads to changes in inflammatory cellular mediators, epithelial hyperplasia, pulmonary fibrosis, and, in some cases, adenomas and carcinomas.

The rat is uniquely sensitive to the effects of titanium dioxide lung overload

To summarize the major species differences in lung response to inhaled titanium dioxide particles from both interspecies studies by Bermudez et al. (2002, 2004): 1) the pulmonary clearance of dust was significantly faster in hamsters versus rats or mice; 2) exposures to the higher doses of either pigment-grade or sub-pigmentary titanium dioxide for 3 months produced particle overload in both rats and mice; 3) the pulmonary cellular and tissue responses to particle overload were different in the rats when compared to similarly exposed mice -- i.e., rats developed a greater and sustained lung inflammatory response and a significantly more intensive epithelial and fibroproliferative response. These data suggest that the pulmonary responses of rats, when exposed to particle concentrations that are likely to induce particle overload in the lung, are substantially different from those in mice and hamsters and can lead to the development of chronic adverse lung effects.

Other poorly soluble, low toxicity dusts cause similar impaired pulmonary clearance and persistent inflammation in the rat (ILSI, 2000; Mauderly, 1996; Donaldson, 2000). As an example, male rats were exposed to either titanium dioxide or carbonyl iron particles for 6 hours/day, 5 days/week for 4 weeks at concentrations of 5, 50 and 250 mg/m3 (Warheit et al., 1997). Pulmonary inflammation as well as alveolar macrophage clearance functions, cell proliferation and histopathological changes were measured at several post exposure time intervals through 6 months. At similar lung burden levels, titanium dioxide and carbonyl iron produced sustained pulmonary inflammatory responses measured through 6 months post exposure. Cellular proliferation was increased in the terminal airways and pulmonary parenchymal cells. The impairment of particle clearance mechanisms was accounted for by deficits in the phagocytic and chemotactic potential of alveolar macrophages. The results of this study clearly showed that exposures to high concentrations of two dissimilar types of low toxicity particles produced similar pulmonary effects in the rat.

In studies reported by Nikula et al. (1997, 2001), it has been proposed that the intrapulmonary particle retention patterns and tissue reactions in rats may not be predictive of pulmonary retention patterns and tissue responses in either primates or humans. Male monkeys and rats were exposed for 7 hours/day, 5 days/week for 24 months to diesel exhaust (2 mg/m3), coal dust (2 mg/m3), or diesel exhaust and coal dust combined (1 mg/m3 each) and were subsequently examined histopathologically (Nikula et al., 1997). In all exposed groups, monkeys retained a similar amount or more particulate material in the lungs than did rats. Rats retained a greater proportion of the particulate material in the alveolar ducts and alveoli, whereas monkeys retained a greater proportion of particulate material in the interstitium. Rats, but not monkeys, had significant alveolar epithelial hyperplastic, inflammatory, and septal fibrotic responses to the retained particles. Similar to the findings in monkeys, up to 91% of the retained particulate material in the lungs of coal miners was located in the lung interstitium (Nikula et al., 2001). It was suggested by the authors that these differences in particulate tissue distribution in rats and humans might bring different lung cells into contact with retained particulates or particlecontaining macrophages. This may account for the differences in species responses to inhaled particulates.

Data on coal miners provides the best available human evidence with which to explore the question of lung overload. Using eight studies conducted between 1956 and 1986 from a total of 1,225 miners in the US and UK, Mauderly (1994) converted the lung burden of coal dust into units of specific lung burden and showed that long-term coal miners commonly accumulated dust burdens in the range of 7 to 14 mg per g of lung tissue. This value indicates that the dust burdens in heavily exposed human lungs are in the same range as, or greater than, the heavily exposed experimental animals seen in chronic bioassays. In spite of these high lung burdens, coal dust exposure does not cause a significant increase in lung cancers among miners (IARC, 1996). This reasoning, although quite compelling, does not preclude the possibility that total particle surface area and particle number are also parameters pertinent to biological outcomes.

Lung tumours in rats are most likely caused by a secondary genotoxic effect resulting from inflammation

There is compelling evidence to suggest that the mode of action for lung tumour formation in rats exposed to inert, poorly soluble particles (PSPs), including carbon black, coal dust, and titanium dioxide, involves chronic and persistent inflammatory changes involving cytokine and growth factor release and macrophage recruitment, leading to the eventual production of reactive oxygen species, mutation, and lung tumour development (Driscoll et al., 1996a; Driscoll and Mauer, 1991; Driscoll et al., 1997). Lung inflammation resulting from intratracheal administration of titanium dioxide in rats leads to the production of reactive oxygen species and increased mutation frequencies in alveolar cells isolated from exposed rats. The prevailing scientific consensus is that rat lung tumours induced by PSPs, such as titanium dioxide and carbon black, arise out of a background of chronic and persistent inflammatory changes (ILSI, 2000; Driscoll et al., 1997). The corollary to this being that in the absence of such changes, lung tumours will not occur.