Skip to main content

Identification of the mechanisms that drive the toxicity of TiO2 particulates: the contribution of physicochemical characteristics


This review focuses on outlining the toxicity of titanium dioxide (TiO2) particulates in vitro and in vivo, in order to understand their ability to detrimentally impact on human health. Evaluating the hazards associated with TiO2 particles is vital as it enables risk assessments to be conducted, by combining this information with knowledge on the likely exposure levels of humans. This review has concentrated on the toxicity of TiO2, due to the fact that the greatest number of studies by far have evaluated the toxicity of TiO2, in comparison to other metal oxide particulates. This derives from historical reasons (whereby the size dependency of particulate toxicity was first realised for TiO2) and due to its widespread application within consumer products (such as sunscreens). The pulmonary and dermal hazards of TiO2 have been a particular focus of the available studies, due to the past use of TiO2 as a (negative) control when assessing the pulmonary toxicity of particulates, and due to its incorporation within consumer products such as sunscreens. Mechanistic processes that are critical to TiO2 particulate toxicity will also be discussed and it is apparent that, in the main, the oxidant driven inflammatory, genotoxic and cytotoxic consequences associated with TiO2 exposure, are inherently linked, and are evident both in vivo and in vitro. The attributes of TiO2 that have been identified as being most likely to drive the observed toxicity include particle size (and therefore surface area), crystallinity (and photocatalytic activity), surface chemistry, and particle aggregation/agglomeration tendency. The experimental set up also influences toxicological outcomes, so that the species (or model) used, route of exposure, experiment duration, particle concentration and light conditions are all able to influence the findings of investigations. In addition, the applicability of the observed findings for particular TiO2 forms, to TiO2 particulates in general, requires consideration. At this time it is inappropriate to consider the findings for one TiO2 form as being representative for TiO2 particulates as a whole, due to the vast number of available TiO2 particulate forms and large variety of potential tissue and cell targets that may be affected by exposure. Thus emphasising that the physicochemical characteristics are fundamental to their toxicity.


The field of nanotechnology is expanding at a tremendous rate due to the realisation that the properties exhibited by materials at a 'nano' scale are often exceedingly different to those demonstrated by bulk forms of the same material. Nanomaterials (defined as having at least one dimension less than 100 nm [1]), of various types and quantities, are therefore attractive for exploitation within diverse products, which harness the novel properties exhibited by materials with nano dimensions. As a result, an improved understanding of the potential risks (comprising of exposure and hazard assessments) of such materials is required, and in particular, determination of nanomaterial characteristics that may detrimentally affect human health (see for example, Maynard et al. [2]). This knowledge will be useful in managing risk in the future, by allowing the implementation of specific control measures for minimising exposure to such materials, perhaps through the introduction of regulations, or through the use of alternative materials. This would therefore allow safety to be built into the design of nanomaterials and their applications, to allow their safe integration into products. This hazard review, relating to the toxicity of titanium dioxide (TiO2) particulates was adapted from a series of reviews conducted as part of the Engineered Nanoparticles: Review of Health and Environmental Safety' (ENRHES) project, funded by the European Commission FP7 funding programme The project aimed to conduct a comprehensive, and critical review of the available health and environmental safety data for a variety of nanomaterials, in order to determine the current level of understanding (relating to nanomaterial toxicity to humans), and to identify the current gaps in knowledge, thereby allowing elucidation of the research that should be conducted with highest priority in the future. In addition, the hazard information generated within this review is to be combined with a review of the available human and environmental exposure data and an evaluation of industrial activity in this area, in order to provide the basis for a risk assessment, based on current knowledge.

Metal oxide nanoparticles (NPs) can be composed of a variety of diverse materials, including titanium, zinc, cerium, aluminium and iron oxides. The size of such particles is integral to their exploitation, but size is also responsible for prompting concern surrounding their potential toxicity. When considering metal oxide particulates, the greatest number of available studies have focussed on revealing the toxicity of TiO2, due to its widespread exploitation. These studies will therefore constitute the focus of this review. TiO2 exposure, via the pulmonary route, has been considered with particular interest due to its use as a (negative) control particulate when assessing the pathogenicity of other particulate materials such as alpha-quartz (see below), but also due to concern regarding occupational exposures. Assessing TiO2 particle toxicity to the skin has also formed the focus of a number of studies, due to their inclusion within sunscreens and cosmetic products. The ingestion of TiO2 is also of relevance due to its incorporation into foods and medicines.

Concern about the potential toxicity of nanoparticles in general, and TiO2 in particular, originally emanates from the studies conducted by Ferin et al. [3, 4] and Oberdorster et al. [5]. These investigators first demonstrated that pulmonary inflammation, particle retention and translocation of ultrafine (equivalent to NPs, in terms of size) TiO2 was enhanced, compared to that of its larger, fine equivalent. Previously, TiO2 has been largely used in pulmonary toxicology studies as a negative control when assessing the toxicity of pathogenic particulates such as alpha-quartz. As a result, this had to be re-considered based on the apparent size dependency of TiO2 toxicity. These studies paved the way for the consideration of particle size, in relation to the toxic response that is manifested by particulates. More recently, studies have continued to focus on the importance of TiO2 particle size (and surface area) to particle toxicity, but have now expanded considerations to include determining the contribution of the particle crystal phase, surface chemistry, and photoactivity to their toxic potential.

In vivo assessment of TiO2toxicity

In general, in vivo investigations that evaluate the toxicity of TiO2 particles concentrate on a particular route of exposure, which is driven by the anticipated application of TiO2, for example the utilisation of TiO2 containing sunscreens has prompted investigations into the dermal toxicity of TiO2.

Pulmonary exposure to TiO2 particulates

The greatest number of available studies, by far address the consequences of the exposure of the lungs to TiO2. In particular, the size dependence of any effects has been prioritised within available investigations. In addition, consideration of the contribution of the experimental set up to the observed toxicity has also been a focus of a number of studies. A summary of the pulmonary toxicity of TiO2 can be found in table 1.

Table 1 The pulmonary toxicity exhibited by TiO2

As mentioned previously, a study conducted by Ferin et al. [3] was one of the first to demonstrate that particle size was fundamental to the pulmonary toxicity of TiO2. In this study, rats were exposed via inhalation, to 21 nm (23.5 mg/m3) and 250 nm (23.0 mg/m3) TiO2 for 12 weeks, and examination of the consequences of TiO2 exposure evaluated over a 64 week post-exposure period. Alternatively, rats were administered TiO2 of various sizes (12, 21, 230 and 250 nm in diameter) via a single intratracheal instillation (up to 1000 μg/rat) and toxicological investigations made 24 hours post-exposure. Following intratracheal or inhalation exposure, nanoparticulate TiO2 induced a greater pulmonary inflammatory response (characterised by neutrophil infiltration), than its microparticulate (250 nm) counterpart, which did not elicit any changes in the inflammatory status of the lung. The smaller sized TiO2 particles were also found to remain within the lung for longer periods (501 days) following inhalation, than the larger TiO2 particles (174 days). This demonstrates that the clearance of smaller particles from the lung was slower. In fact, the prolonged retention of smaller TiO2particles in the lung was suggested to derive from the finding that they were able to translocate to the pulmonary interstitium more efficiently than the larger TiO2 particles. The authors proposed that this phenomenon was facilitated by the fact that smaller particles were not efficiently taken up by alveolar macrophages, which thereby allowed for their prolonged interaction with alveolar epithelial cells. It was found that an increased mass dose (which translates to an increased number of particles, for smaller particles) promoted the movement of particles within the pulmonary system. It was therefore observed that particle size, and the number of particles administered (which was related to the delivered dose) impacted on the translocation process, and therefore TiO2 toxicity. However, since the publication of this study it has been demonstrated that the surface area of particles is influential to their toxicity. As such, smaller particles have a greater surface area than their larger counterparts, and as a consequence exhibit greater toxicity, when administered at an equal mass dose.

The size dependency of TiO2 (and carbon black (CB)) toxicity was investigated by Renwick et al. [6]. To achieve this, particle mediated inflammatory responses were investigated, and the ability of the particles to impact on clearance mechanisms within the rat lung determined. Rats were exposed to particles via intratracheal instillation, at 125 μg/rat or the relatively high dose of 500 μg/rat, and toxicological investigations conducted at 24 hours post exposure. At the highest dose only, nanoparticulate (29 nm), but not microparticulate (250 nm) TiO2 stimulated the recruitment of neutrophils into the lungs, epithelial damage, increased permeability of the lung epithelium, and cytotoxicity, which were measured within the bronchoalveolar lavage fluid (BALF). The NPs were also able to diminish the phagocytic ability of isolated rat alveolar macrophages (from exposed animals), and increase the ability of these macrophages to migrate towards the chemotactic signal, C5a. The authors suggested that the consequences of such a stimulation would be to promote the retention of macrophages, and therefore their particle burden, within the lung. The NPs were therefore demonstrated to elicit a greater pulmonary inflammatory response than their larger counterparts.

In a different approach, Bermudez et al. [7] determined if the choice of species was able to influence the pulmonary response to nanoparticulate TiO2 (21 nm). This was achieved through the exposure of rats, mice and hamsters, to TiO2 NPs, via inhalation, for 13 weeks (6 hours/day, 5 days/week) at concentrations of 0.5, 2, or 10 mg/m3. The pulmonary response was assessed up to 52 weeks post exposure, and included assessment of inflammation, cytotoxicity, lung cell proliferation and histopathological analysis. It was demonstrated that a pulmonary inflammatory response was stimulated by TiO2 within mice and rats, but was absent in hamsters. The nature of the response was also observed to vary within the different species. Specifically, a greater neutrophillic response, which decreased with time, was apparent in rats, with progressive epithelial fibroproliferative changes also apparent. In mice, there was neutrophil and macrophage component of the inflammatory response, and these cell populations remained elevated throughout the observation time. The severity of the response was ranked in the following order, by the authors; rat>mouse>hamster. Consequently, the rat was found to be the most sensitive species to the effects of TiO2. The limited toxicity apparent within hamsters was thought to derive from the low lung burden of particles, as hamsters had the greatest propensity to efficiently clear particles from the lung. However, for mice and rats, the initial retention of particles was similar, and decreased with time, demonstrating that the pulmonary clearance kinetics varied amongst the different species. The study highlights that differences in the response, and therefore sensitivity of different species are worthy of consideration within investigations, and the model which mimics the human situation most accurately requires assessment. It is also of note that the effects observed were dose dependent, and so the toxicity of NPs was enhanced by increasing the concentration. The findings demonstrated within this study were comparable with those of a companion study conducted by Bermudez et al. [8]. In this study rats, mice and hamsters were exposed to fine TiO2 particles (1 μm), via inhalation, at concentrations of 10, 50 or 250 mg/m3, with the experimental design identical to that described previously [7]. Again, species differences were witnessed, so that some species were most susceptible to particle exposure (rat>mouse>hamster), in addition to evidence that the pulmonary toxicity observed was concentration dependent. However, it is difficult to distinguish if size dependent effects were observed within these separate studies. This is derived from the utilisation of a higher dose of particles within the microparticulate study, when compared to those encountered within the NP study. However, as no effects were elicited by fine TiO2 particles when administered at a concentration of 10 mg/m3 it is reasonable to assume that the NP sample is more toxic, when administered at the same mass dose, as toxic effects were observed at this dose for NPs. In fact, the concentration of particles used were exceptionally high in both studies, and thus their physiological relevance is questionable (see later). However, the authors aimed to achieve a particle overload situation (see later), so that the administration of high particle concentrations was undertaken to investigate the efficiency of clearance processes in the different species, and thus the experimental set up can therefore be justified.

The ability of the experimental set up to influence the pulmonary response to TiO2 was also investigated by Warheit et al. [9]. It was determined if the exposure method, or TiO2 formulation (specifically concentrating on surface coating modification) were able to impact on their pulmonary toxicity within rats. For inhalation, rats were exposed to particles (ranging from 290 to 440 nm in diameter) at concentrations of up to 1300 mg/m3, for a duration of 4 weeks. In general, following inhalation of TiO2, there was an accumulation of particle containing macrophages within the lungs. Following intratracheal instillation (up to 10 mg/kg) it was observed that some TiO2 preparations were able to stimulate a transient, pulmonary inflammatory response that was typified by the infiltration of neutrophils, and lactate dehydrogenase (LDH) release, and that this response was resolved within one week post-exposure. It is relevant that the response that transpired following particle exposure was very much dependent on the TiO2 formulation in question. Specifically, the samples that contained the highest alumina or amorphous silica content elicited the greatest adverse pulmonary response. Overall, both exposure methods were associated with minimal adverse effects, but this may be derived from the size of particles utilised, as all samples were out-with nano dimensions (i.e. >100 nm). It must also be noted that the concentrations of particles administered in this experiment are deemed to be excessive, and therefore the relevancy of the findings are questionable. The utilisation of high particle concentrations is a common problem within this field of research, as they are not expected to be encountered within humans, and thus this issue will be discussed in greater detail later.

Similarly, Grassian et al. [10] investigated what particle attributes drove the toxicity of TiO2 NPs (5 and 21 nm) within mice, subsequent to inhalation (0.7 or 7 mg/m3, for 4 hours with toxicological observations made immediately and 24 hours post exposure) or nasal instillation (up to 12.7 μg/mouse). An elevated macrophage population was associated with the inhalation of particles, and these cells were observed to internalise particles. An infiltration of neutrophils, was associated with the nasal instillation of TiO2. In both exposure scenarios, the response induced by 21 nm particles was greater than that exerted by their 5 nm counterparts. Consequently, the surface area of particles was not deemed to be the sole determinant of TiO2 toxicity, which was unexpected. Instead, the crystallinity of the samples was suggested to influence the toxicity of TiO2 (see later), as 5 nm particles were pure anatase, whereas the 21 nm particles were a mixture of rutile and anatase forms of TiO2. It is also relevant that the nature of the response varied for the different exposure methods, specifically inhalation exposure promoted a macrophage driven response, and instillation exposure triggered the development of a neutrophil driven response. This insinuates that the experimental set up influences the results of toxicity tests, and may be related to differences within their distribution within the lung following exposure, which requires further investigation.

In line with these findings, Warheit et al. [11] aimed to determine if the crystalline form of TiO2 was able to influence its pulmonary toxicity within rats, subsequent to intratracheal administration (1 or 5 mg/kg). Two rutile NP samples, a 'mixed' NP (80% anatase, 20% rutile) sample, and a microparticulate rutile (negative particulate control), in addition to α-quartz sample (included as a positive particulate control) were considered. All TiO2 particle samples were highly agglomerated (>2000 nm), following their dispersion in phosphate buffered saline (PBS), despite the fact that their primary particle size was less than 100 nm. The pulmonary response to the particle panel was evaluated up to 3 months post exposure. Although quartz had the greatest inflammogenicity, it was evident that several TiO2 particle types elicited a short-term, pulmonary inflammatory response that was characterised by the infiltration of neutrophils, and was greatest, and more sustained for the mixed NP sample. Furthermore the only samples capable of eliciting a release of LDH (a measure of cytotoxicity) and protein (indicative of increased vascular permeability) into BALF were quartz and the mixed TiO2 sample. The properties responsible for driving TiO2 toxicity were then considered and it was suggested that differences in particle surface area accounted, in part, for the responses exhibited by particles. Accordingly, the mixed sample had the greatest surface area (53 m2/g), and therefore greatest reactivity, so that surface properties were suggested to drive TiO2 toxicity. However as the rutile NP samples had surface areas of 18.2 m2/g and 35.7 m2/g, and the TiO2 microparticles had a surface area of 5.6 m2/g, it was expected that the rutile NP particle samples would exhibit an intermediate level of toxicity (between the mixed, and microparticulate samples), but this did not transpire within the findings. Therefore, additional factors were speculated to contribute to the toxicity of TiO2 particulates, besides their surface area, such as crystallinity. It was suggested that the rutile NP samples were less toxic than those that were predominantly anatase (see later). In addition, the authors suggested that post production processing, that removed chloride from the particle surface, and enhanced agglomeration of particles, may also contribute to the reduced toxicity of the rutile NP particles. Again, the particle concentrations administered were excessively high, and unlikely to be encountered by humans, and therefore justification of such concentrations is required by investigators, in order to determine their relevancy.

The nature of the pathological response associated with the pulmonary exposure to TiO2 has been studied by a number of investigators. Chen et al. [12] exposed mice, via intratracheal instillation (0.1 and 0.5 mg/mouse), to nano (19-21 nm) and micro (180-250 nm) forms of TiO2. Pulmonary toxicity was determined 3 days, 1 week or 2 weeks post exposure. Histological assessment illustrated that morphological alterations within the lung were evident on exposure to nanoTiO2, which were emphysema-like in nature (including, for example alveolar enlargement). Lesions were more pronounced in areas where particles preferentially accumulated, and increased in severity, with increasing time and dose. An inflammatory response, recognised by the infiltration of macrophages (that were particle laden), upregulation of cytokines (including monocyte chemotactic protein (MCP)-1, interleukin (IL)-1, tumour necrosis alpha (TNFα) and several neutrophil chemoattractants) and complement activation was also observed. Despite the large doses used, no pathological lesions were found in response to microparticulate TiO2. This study again illustrates the size dependent toxicity exhibited by TiO2, and the tissue lesions, and biochemical markers associated with the particle toxicity.

In a different approach, Heinrich et al. [13] exposed rats to P25 TiO2 NPs (size reported by others to be 21 nm (see for example Ferin et al [3]) via inhalation for 2 years. Particle exposure was followed by exposure to clean air for 6 months. In addition, mice were exposed to TiO2 for 13.5 months then clean air for 9.5 months. TiO2 was able to increase mortality within rats and mice, so that their lifetime was significantly shortened. Lung tumours were observed in both species, as a consequence of the chronic exposure to TiO2. The high doses utilised in this experiment undermines the relevancy of the findings In fact, they are likely to represent particle 'overload' conditions, where the lung is overwhelmed by particle presence to enhance the toxicity that is manifested. Overload can be described as the deposition of particles that exceeds the ability of macrophages to remove them. As a consequence, particles are more likely to be retained within the lung, and interact with the different lung cell populations to stimulate toxicity, such as tumour development. This study therefore highlights how the experimental set up has the ability to influence the findings (see later). Consequently, TiO2 cannot be regarded as carcinogenic based on the findings from this study. Therefore, it is anticipated that the administration of excessively high concentrations of particles was responsible for promoting tumour development within this model, and not an inherent genotoxic property exhibited by TiO2 particles.

Ahn et al. [14] investigated what processes were responsible for particulate mediated stimulation of excessive mucus secretion within humans. To achieve this, TiO2 (4 mg/kg, 0.29 μm diameter) was exposed to rats via intratracheal instillation, and toxicological observations made from 4 to 72 hours post exposure. TiO2 exposure stimulated an increase in goblet cell hyperplasia, which is, in part, attributed to an increase in muc5 gene expression and IL-13 production. Therefore, it could be speculated that particle mediated increases in mucus secretion can contribute to the aggravation of chronic airway disease symptoms within humans, and therefore warrants further investigation.

The ability of TiO2 exposure to influence the pulmonary toxicity of known inflammogenic agents has been investigated by Inoue et al. [15]. The impact of TiO2 (15, 50, 100 nm) intratracheal instillation (8 mg/kg) on lipopolysaccharide (LPS) mediated pulmonary inflammation in mice was determined. Combined TiO2 and LPS treatment was able to exacerbate LPS mediated inflammation (indicated by keratinocyte chemoattractant (KC), IL-1β and MCP-1 production). Circulating cytokines were also increased, as well as coagulatory factors such as fibrinogen. As observed previously, there was a size dependency to the findings was also observed, whereby smaller particles were most toxic.

In a different approach, Nurkiewicz et al. [16] determined the impact of TiO2 particles (1 μm) or ROFA particulate matter on the systemic microvasculature, following intratracheal instillation of rats (0.1 or 0.25 mg/rat, for 24 hours). The authors therefore encompassed the possibility that systemic effects were a component of the pulmonary response to particulates. Particle exposure stimulated a neutrophil influx into the lungs, but no cytotoxic response was evident. Particles were able to induce an impairment of endothelium dependent arteriolar dilation. The response was suggested to be related to increased neutrophil adhesion to the vessels, myleoperoxidase (MPO) deposition, and oxidative stress within the vessel wall. These findings are of concern, as they indicate that an inflammatory response may be stimulated within the vessel. However, the response was independent of the level of pulmonary inflammation, and was not thought to be reliant on the migration of particles from the lung. This study highlighted the systemic (inflammatory mediated) responses that are associated with particle exposure, but at this time the mechanisms underlying this response is unknown.

Consideration of the fate of TiO2, subsequent to pulmonary exposure has been apparent within a number of investigations, with a particular focus on the transfer of particles to the brain. Wang et al. [17] investigated the distribution of rutile (80 nm) and anatase (155 nm) TiO2 particles within the mouse brain, following nasal instillation exposure (500 μg/mouse, every other day, for a total of 30 days) and determined if any neurotoxicity was associated with exposure. Both forms of TiO2 were able to access the brain, with accumulation within the cerebral cortex, thalamus and hippocampus evident, and postulated to occur via the olfactory bulb. This route of uptake however, was unlikely to be mediated via penetration into the cardiovascular system and via the blood. Instead, TiO2 delivery to the brain was speculated to occur via neuronal transport. Accumulation of TiO2 resulted in morphological alterations and loss of neurones in the hippocampus, which was accounted for by the higher distribution of TiO2 within this brain region. In addition, it was suggested that TiO2 elicited oxidative stress within the brain due to elevated catalase activity, decreased superoxide dismutase (SOD) activity and evidence of increased lipid peroxidation and protein oxidation. Therefore neuronal mediated translocation of TiO2 to the brain, following nasal instillation, was observed, with the hippocampus illustrated as being the main target of accumulation and toxicity. Wang et al. [18] expanded upon these findings and found that the phenomenon was time dependent (maximal at 30 days), and that an inflammatory response (indicated by IL-1β, and TNFα) within the brain was also stimulated by TiO2 exposure. The response was measured at day 2, 10, 20, and 30. It was apparent that repeated exposures, over a period of 30 days, were required to enable the accumulation of TiO2 within the brain. It is therefore of interest that the neuronal transport of NP containing substances between the nose and CNS could be exploited, in order to bypass the blood brain barrier, but that the stimulation of an inflammatory response may be associated with this phenomenon.

The results of the outlined studies imply that the toxic potential of TiO2 was primarily dictated by particle size, and crystallinity; whereby decreasing particle size, and anatase forms of TiO2 were observed to enhance particle pulmonary toxicity. In addition, it is suggested that the experimental model (including species used, exposure method and dose administered) was able to impact on the toxicity of TiO2, which complicates making comparisons between different investigations. The pulmonary response to TiO2 has been demonstrated to be inflammogenic; with a contribution from neutrophils and macrophages (which are elevated and thought to contribute to particle removal). In addition, epithelial damage, oxidative stress and cytotoxicity are also common findings. In addition, chronic exposure to TiO2 also has the ability to promote tumour development, illustrating their carcinogenic potential, but this is only likely to manifest when administered in particle overload conditions.

Intraperitoneal exposure to TiO2

Chen et al. [19] investigated the acute toxicity of nano-TiO2 (80-100 nm) subsequent to the intraperitoneal injection of mice. The doses used were exceptionally high (ranging from 324 to 2592 mg/kg), and it is therefore unsurprising that mortality was associated with exposure. TiO2 was observed to block pulmonary vessels, leading to thrombosis, with pathology also evident within the liver, spleen and kidneys. This study highlighted the need to consider the use of realistic doses of particles, as the observed effects are likely to derive from the high dose of particles administered, and not the inherent toxicity of particles.

Dermal exposure to TiO2

Considering the impact of TiO2 dermal exposure is especially relevant due to its inclusion within sunscreens and cosmetics that are directly applied to skin. As a consequence, the dermal penetration and toxicity of TiO2 particles, has been a focus of a number of investigations. The skin, and in particular, the stratum corneum, is a primary barrier against the penetration of particles contained within dermally exposed substances, and therefore its efficiency will also be discussed.

Mavon et al. [20] determined the distribution of TiO2 (20 nm) within the skin in vitro and in vivo. Five hours following the direct topical application of TiO2 (2 mg/cm2) to human skin or to human skin explants tape stripping (placing and removing adhesive tape onto and from the skin repeatedly and subsequently analysing the components striped from the skin surface) was used to determine the dermal penetration of TiO2. In vivo, and in vitro, the majority of the TiO2 was contained within the stratum corneum, with minimal distribution within the epidermis. Therefore in both preparations, TiO2 presence decreased with increasing depth of the skin, so that the penetration of TiO2 past the stratum corneum into viable skin layers was minimal.

Similarly, Schulz et al. [21] determined the influence of particle size, coating and shape on TiO2 skin penetration. A number of TiO2 containing sunscreen formulations were tested that had different particle surface characteristics; trimethyloctylsilane coated (20 nm), or aluminium oxide (Al2O3) and silica (SiO2) coated (10-15 nm). Formulations were topically exposed to human skin at a concentration of 4 mg/cm2, for 6 hours, and skin biopsies taken. All particle types were solely located on the outermost surface of the stratum corneum, and did not penetrate deeper to subcutaneous, epidermis or dermis layers.

In a different approach, Kiss et al. [22] evaluated the barrier function of skin, within human foreskin grafts transplanted onto severe immunodeficient mice. A commercially available TiO2 microparticulate containing sunscreen was administered (2 mg/cm2) via an occlusive bandage to skin grafts for 24 hours, and the penetration of TiO2 determined within skin biopsies. It was found that TiO2 particles did not penetrate through the stratum corneum of human skin transplants. The stratum corneum was therefore deemed to be an adequate, effective barrier against TiO2 penetration in intact human skin. However, the authors noted that as it is likely that TiO2 exposure occurs when skin barrier functions may be impaired (such as sunburn), and as a consequence, TiO2 may come into direct contact with underlying skin cells, a concept which has yet to be investigated.

The findings from the available studies demonstrated that the penetration of TiO2 is negligible within (healthy) skin. This is important, as NPs appear to be unable to reach the living cells present within the deeper skin layers, and thus their propensity for toxicity is anticipated to be minimal. Therefore, it is relevant, within future studies, to consider the fate of particles within skin models that take into consideration that sunscreens are often applied to burnt, damaged and diseased skin, where the barrier function of the stratum corneum will inevitably be impaired [22, 23]. This has been discussed within several investigations but not actually studied.

Oral exposure to TiO2

Only one study could be identified that addressed the consequences of oral exposure of TiO2. Wang et al. [24] investigated the distribution and acute toxicity of nanoparticulate (25 and 80 nm) and microparticulate (155 nm) forms of TiO2 in mice, following oral exposure (5 g/kg). TiO2 particles of all sizes, 2 weeks post exposure, were distributed to the liver, spleen, lungs, kidneys, thus providing evidence that they could be transported to other sites subsequent to exposure, due to their translocation into the blood. Within the liver, NPs initiated an inflammatory response, with liver damage indicated by a rise in serum transanimases, and hepatic necrosis was revealed in histopathological investigations. Markers of cardiac damage were also observed to be elevated by TiO2 NPs, within the serum. Limited toxicity was associated with microparticle exposure. This study therefore provided interesting results. Specifically the potential for TiO2 particle transfer into the blood from the GIT was demonstrated, which would necessitate that they are able to pass through the gut wall. It also revealed that the liver was a primary target for the toxicity of TiO2. However, the use of an exceptionally high dose of particles undermines the relevancy of the results, as such exposure levels are unexpected in humans.

ADME Profile of TiO2

Determining the kinetics of TiO2 within the body, subsequent to exposure (via the lungs, gut and skin) is necessary to identify potential targets of toxicity. The delivery of TiO2 particles from the exposure site to secondary target organs, such as the liver or brain requires their transfer into blood, or transport within neurones. Therefore their likelihood of accessing different organs and tissues within the body is also necessary to direct appropriate in vitro investigations at relevant target sites. However, a number of barriers (at the exposure site and those apparent within secondary targets) are in place to prevent against uptake, and it is necessary to determine if they are overcome by TiO2, to determine its systemic availability.

Only one example of particle translocation into the blood from the gut could be identified. Specifically, TiO2 NPs were observed to translocate into the blood, following oral exposure, and thereafter distribute to secondary targets, including the liver, spleen, lungs, and kidneys [24]. Studies investigating the dermal penetration of TiO2 suggest that the transport of nanoparticles past the stratum corneum is negligible, and therefore it is unlikely that the particles will access the circulation via this route. The limited penetration of TiO2 NPs within skin has also been replicated within in vitro models (see later). Wang et al. [17, 18] demonstrated the transfer of TiO2 NPs to the brain, following intranasal exposure, which occurred via their transport within neurones.

No evidence of TiO2 metabolism, or elimination from the body could be identified in the literature at this time.

The translocation of TiO2 particles, subsequent to pulmonary and oral exposure should encompass the possibility that distal sites are affected; including the CNS, liver, and cardiovascular system. The translocation is likely to be due to their passage into blood, but may also be mediated by neuronal transport. Both these processes would allow for their transport to targets sites (such the brain), and within different regions of a particular target (such as different brain regions). In contrast, the penetration of particles within the skin is negligible, and so particles are unlikely to become systemically available, following exposure via this route. The elimination of particles from the body is also a necessary consideration, which would provide information regarding the longevity of particles within the body, and therefore their propensity for damage.

Distribution of TiO2 following intravenous exposure

Following injection, it is necessary to determine particle distribution to identify potential targets for toxicity. There are limited available studies that purport to investigate this. In addition, it is assumed that if NPs gain access to the circulation following exposure via the lungs, skin or GIT, a similar distribution pattern is expected, albeit it to a lesser extent.

Fabian et al. [25] determined the tissue distribution of TiO2 NPs (20-30 nm) within rats following intravenous injection (5 mg/kg), 1, 14 and 28 days post exposure, TiO2 was cleared from the blood and primarily accumulated within the liver, but was also apparent within the spleen, lungs and kidneys. The level of TiO2 within the liver was maintained over the observation time, however levels decreased with time within the other organs. No serum cytokine or enzyme changes were observed, which insinuated that no toxicity was associated with TiO2 exposure, however further investigations, including histopathological analysis would be necessary to confirm this.

The only study identified suggested that following intravenous administration, TiO2 particles are cleared from the blood and are able to accumulate primarily within the liver, but also the lungs and spleen. This is likely to derive from their uptake by the resident macrophage populations, contained within the reticulendothelial system. Further investigations are necessary to determine if an inflammatory or toxic response is associated with such an accumulation.

In vitro investigations of TiO2toxicity

Lung models

Investigations that assess the pulmonary toxicity of TiO2 particles in vitro have used a variety of models including macrophages (due to their contribution to clearance), epithelial cells (due to their abundance, and expected interaction with particles) and explanted tissue.

Park et al. [26] investigated the cytotoxicity of TiO2 NPs (21 nm) to BEAS-2B lung epithelial cells, at concentrations ranging from 5 to 40 μg/ml, for up to 96 hours. A dose and time dependent decrease in cell viability was observed. Caspase-3 was activated by TiO2, with chromosome condensation also apparent, which was suggestive that an apoptotic mechanism of cell death was involved. Reactive oxygen species (ROS) production, glutathione (GSH) depletion, and increased hemeoxygenase (HO-1) expression was evident, thereby implying that an oxidative mechanism drove the cytotoxic response. IL-8, IL-1, IL-6, TNFα, and CXCL2 (neutrophil chemoattractant) cytokine gene expression was increased, insinuating that an inflammatory response was also stimulated by TiO2. The findings therefore indicated that an oxidant driven response was integral to the toxicity of TiO2, which had inflammatory and cytotoxic consequences.

Churg et al. [27] determined the impact of TiO2 exposure (up to 7 days), on the development of a fibrotic response, within the airway wall on rat tracheal explants. Both microparticulate (0.12 μm) and nanoparticulate (0.021 μm) forms of TiO2 were investigated. No changes in gene expression were observed until day 5. NPs stimulated growth factor expression (such as platelet-derived growth factor (PDGF)) and enhanced pro-collagen expression. The authors suggested that the NPs were more fibrogenic, which is anticipated to contribute to airway obstruction in vivo. However, TGFα and TGFβ expression was increased for microparticles only, but no fibrotic-like morphological changes occurred, and so it was suggested that these mediators were not responsible for driving the fibrotic response mediated by NPs.

Gurr et al. [28] investigated the oxidative damage exhibited by TiO2 (10, 20 or >200 nm) within bronchial BEAS-2B cells. The oxidative potential of TiO2 was confirmed by the finding that lipid peroxidation (indicated by increased malondialdehyde (MDA)) and ROS (hydrogen peroxide) and reactive nitrogen species (NO.) production were enhanced by nanoparticulate forms of TiO2. DNA adducts were only formed following exposure of cells to TiO2 NPs. The oxidative and genotoxic potential of nanoparticulate forms of TiO2 was therefore demonstrated to be superior to that of their larger counterparts.

Simon-Deckers et al. [29] investigated the toxicity of a panel of particles, including TiO2 (in anatase and rutile forms, and a variety of sizes) and Al2O3, to A549 cells (a human carcinoma derived alveolar epithelial type 2 cell line). In general, the cytotoxicity of the particles was low, despite the fact that they were internalised by cells, and contained within cytoplasmic vacuoles or lysosomes. However, it is noteworthy that it is possible for particles to be internalised by cells without having a detrimental impact on normal cell function. The crystal phase of TiO2 was observed to influence the cytotoxicity exhibited by particles. Specifically, the greater the anatase content of the sample, the greater the ability to induce cell death. In addition, the size of NPs was suggested to contribute to their toxicity, as TiO2 NPs were more toxic than their larger counterparts. In addition, the chemical composition of NPs was also thought to impact on their toxicity, as 12 nm TiO2 NPs were more toxic than 11 nm Al2O3, despite their similarity in size. Consequently, the size, composition and crystal phase of particles were all suggested to influence NP toxicity, despite the fact that the toxicity of NPs used was generally low. Therefore, it is acknowledged that investigating size dependent effects of particles is confounded by differences in other particles attributes such as the sample phase. Consequently determining the properties of TiO2 that drive toxicity is difficult, and therefore isolating a particle characteristic that is solely responsible for mediating adverse effects is not likely to be attainable.

Kim et al. [30] assessed TiO2 (1 μm) mediated cytotoxicity to alveolar macrophages (obtained from rats, via bronchoalveolar lavage), when exposed to concentrations ranging from 0.5 to 5 mg/ml, for up to 5 hours. TiO2 elicited a dose dependent decrease in cell viability, which was suggested to be mediated by suppressed adenosine triphosphate (ATP) generation, and also reliant on the interaction of TiO2 with scavenger receptors on the cell surface. It is relevant that the response of silica particles (1.6 μm) was also considered. Silica had a greater cytotoxic potential than TiO2 and the mechanisms underlying the response were different.

As responses of organs and tissues are likely to involve interactions between different cell types, it is necessary to incorporate this within in vitro tests. Barlow et al. [31] investigated the ability of L2 epithelial cells to modulate macrophage migration in vitro, on exposure to nano and microparticulate forms of TiO2 (29 nm and 250 nm, respectively) and CB (260 nm and 14 nm diameter respectively). TiO2 and CB induced a dose dependent decrease in epithelial cell viability, which was greatest for NPs (when exposed at concentrations ranging from 62.5 to 2000 μg/ml for 24 hours). Conditioned medium (obtained from epithelial cells treated with particles) was able to increase macrophage migration, but only for ultrafine carbon black NPs. Therefore, TiO2 NPs were unable to stimulate the release of chemotaxins from epithelial cells, which the authors suggest could be due to their smaller surface area, when compared to that of ufCB.

From the investigations discussed, it is evident that TiO2 particles are able to detrimentally affect both lung derived macrophage and epithelial cells. Therefore, in general, particles mediated oxidative, inflammatory and genotoxic effects that eventually culminated in cytotoxicity. However, in some situations there was no evidence of toxicity, mediated by TiO2. These differences may be derived from the model used, or the particle under investigation. It is of interest that the mechanisms driving the toxicity of TiO2 in vitro (particularly inflammation and oxidative stress) are also witnessed in vivo (see earlier).

Dermal models

The penetration of TiO2 particles within the skin has been a focus of a number of in vitro investigations due to the exploitation of particles within sunscreens and cosmetic products. Gamer et al. [32] evaluated the penetration of sunscreen formulations containing micron-sized TiO2 (up to 400 μg/cm2, for 3 to 24 hours) within porcine skin explants. It was observed that the location of both particle types was restricted to the stratum corneum following their topical application. Specifically, particles were deposited on the skin surface, within the outmost layer of the stratum corneum, as evidenced by their removal via washing or tape stripping analysis. There was therefore no evidence that particles penetrated into the deeper stratum corneum layers, epidermis or dermis, and thus the barrier function of the skin was successful in limiting the passage of particles.

Similarly, Pflucker et al. [33] illustrated that TiO2 NPs (20-50 nm) did not penetrate porcine skin in vitro. A TiO2 emulsion (4 mg/cm2) was topically applied to excised pig skin for 24 hours. Skin biopsies were taken following exposure, and the penetration of particles assessed using tape stripping. Again TiO2 was exclusively located within the outermost layer of the stratum corneum, with no distribution to the underlying living cell layers evident. Consequently, this study suggests that the intact stratum corneum is an effective barrier to TiO2, and that the penetration of particles within the skin is negligible.

In addition, Dussert et al. [34] investigated the distribution of TiO2 and ZnO NP containing sunscreen formulations within human skin explants. Again, sunscreens were topically applied to skin but no penetration of particles, or intracellular delivery was observed past the stratum corneum skin surface.

The functional implications of the exposure of the various skin cell populations have also been determined within in vitro investigations. From the data available for TiO2 particles, such studies are potentially more relevant to compromised/damaged/diseased skin than normal healthy skin, due to the negligible penetration of particles within normal skin. There is therefore a need to consider the distribution of particles within damaged or sunburnt skin, which would be likely to promote the penetration of particles, due to damage to the stratum corneum. There is little or no data currently available to assess the impact of such damage on penetration or hazard.

Kiss et al. [22] exposed HaCaT keratinocytes, human dermal fibroblast cells, SZ95 sebaceous gland cells and primary human melanocytes to TiO2 (9 nm diameter), at concentrations of 0.15 to 15 μg/cm2, for up to 4 days. Particles were internalised, and evident in the cytoplasm and perinuclear region of fibroblasts and meloncytes. Particle uptake was also associated within an increase in intracellular calcium in these cell types. However, no particle uptake, or alterations in calcium signalling were observed within keratinocytes or sebocytes. A dose and time dependent decrease in cell proliferation was evident within all cell types, and an increase in cell death (via apoptosis) within fibroblasts was apparent. The direct contact of cells with TiO2 particles is therefore of concern, as it can disturb normal cell functions within the different skin cell populations, but it is noteworthy that all cell types were not affected similarly, and thus cell specific effects are encountered on exposure to TiO2 particles. Kiss et al. [22] also demonstrated that the stratum corneum was an effective barrier against micronized TiO2 in vivo (see earlier). However it would have been of greater interest to determine if NPs are able to penetrate skin, as the in vitro component of the study illustrated that when in direct contact with skin cells TiO2 NPs are able to elicit toxicity.

Jin et al. [35] determined the cytotoxicity of TiO2 (20-100 nm) to mouse L929 fibroblasts at concentrations of 3-600 μg/ml, for up to 48 hours. There was a time and dose dependent decrease in cell viability induced by TiO2. An increase in ROS production and decreases in GSH and SOD activity were observed, thereby implying that oxidative stress was a feature of the response. TiO2 was also demonstrated to be internalised by phagocytosis, and was contained within lysosomes. Analysis of cell morphology illustrated that cell morphology was detrimentally affected by TiO2, and that cell adhesion and proliferation was prevented, confirming the decrease in cell survival.

The studies conducted demonstrated that the toxicity of TiO2 to skin cells is of concern, with oxidative, genotoxic and cytotoxic consequences evident, and likely to act in concert. However, in order to execute these effects, particles would be required to penetrate past the stratum corneum, to reach the underlying cell populations, which is unlikely to occur in healthy skin (see earlier). As noted previously, the penetration of particles within damaged skin should therefore be considered a priority in future experiments.

GIT models

Only one study could be identified that evaluated the impact of TiO2 exposure on the GIT, in vitro. Zhang et al. [36] investigated the contribution of photoexcitation, to the cytotoxicity of TiO2 (21 nm) within human colon carcinoma Ls-174-t cells. Cells were exposed to TiO2 for 24 hours, at concentrations up to 1000 μg/ml. Limited cytotoxicity of cells was observed, following TiO2 exposure, in the absence of ultraviolet (UV)A irradiation. In contrast, a high level of TiO2 induced cell death was observed in the presence of UVA light, which was a dose and time dependent phenomenon. Therefore, the toxicity of TiO2 had a photocatalytic component. Consequently, TiO2 particles and light irradiation therapy was suggested by the authors to be a suitable candidate for the treatment of cancer, and warrants further investigation. However, this may be limited by the ability of UV light to penetrate the human body, and the specificity of the response, and so the exploitation of this for tumour treatment will inevitably be restricted.

Liver models

Hussain et al. [37] compared the impact of a range of NPs, including iron oxide (Fe3O4, 30 nm), and TiO2 (40 nm) on BRL 3A liver cells, at concentrations up to 250 μg/ml, following a 24 hour exposure. Both particle types were able to decrease cell viability, but only at the highest concentration tested.

Linnainmaa et al. [38] assessed the toxicity of TiO2, in NP (rutile and anatase (20 nm)), and microparticulate (170 nm) forms) to rat liver epithelial cells, in the presence and absence of UVA light. No cytotoxicity or chromosomal damage was observed within cells exposed to all TiO2 particle types in the presence or absence of UVA.

The two studies that investigated the toxicity of TiO2 particles to the liver, suggest that TiO2 exhibits a low level of toxicity to liver cells, but this would require further investigation to be confirmed, due to the limited number of liver cell targets considered within the available studies. The consequences of liver exposure, to TiO2 are of particular interest due to evidence of preferential NP accumulation within this organ (see earlier).

Cardiovascular models

The interaction of TiO2 with endothelial cells that line blood vessels has been a focus of investigations, to determine the consequences of particle transport with the blood. In addition, the impact of TiO2 on cardiomyocyte function has been investigated.

Peters et al. [39] evaluated the impact of a NP panel (including TiO2, 70 nm) on HDMEC endothelial microvascular cell viability and function. Cells were exposed to TiO2 at concentrations ranging from 0.5 to 50 μg/ml, for up to 72 hours. NPs were internalised by cells into cytoplasmic vacuoles, but minimal toxicity was associated with this, therefore highlighting that particle uptake does not necessarily impact on cell function. Consequently, TiO2 was relatively non-toxic, with no cytotoxicity, and minimal IL-8 release stimulated by exposure.

Courtois et al. [40] determined the impact of TiO2 (15 nm or 0.4 μm diameter) exposure on nitric oxide (NO) mediated relaxation of pulmonary arteries in vitro. Pulmonary arteries were exposed to particles for 24 hours (at concentrations up to 200 μg/ml), in the presence or absence of vascular relaxants. It was found that urban particulate matter was able to impair NO-dependent relaxation within intrapulmonary arteries in vitro and in vivo (following intratracheal exposure). In contrast, manufactured NPs, including TiO2 did not exhibit this effect. Therefore this study suggested that the pulmonary circulation was not affected by TiO2 exposure, and that cells exhibited differed in their sensitivity to particles.

Helfenstein et al. [41] observed that TiO2 NPs (up to 125 μg/ml, 20-30 nm diameter) were able to affect cardiomyocyte electrophysiology, enhance ROS production, and reduce myofibril organisation. A panel of particles was tested, of which diesel exhaust particles (DEPs) were demonstrated to exhibit the greatest toxic potency, and SWCNTs were found to have limited toxicity in comparison. The findings from this study suggest that cardiac cells may be damaged by TiO2 or DEP exposure, and so were unable to function normally in vitro.

A limited number of studies have been conducted to determine the impact of TiO2 particles within in vitro models of the cardiovascular system. However, the findings suggest that TiO2 may promote an inflammatory response within blood vessels which has the propensity to stimulate or aggravate cardiovascular disease. Consequently, the ability of TiO2 to contribute to inflammatory mediated disease requires further consideration in vivo. In addition, the ability of TiO2 to negatively impact of cardiomyocyte function has the ability to disturb normal cardiac electrophysiology. Again, further studies, in vivo, would be required to verify this observation.

CNS models

Long et al. [42] established the contribution of oxidative stress to the neurotoxicity of TiO2 NPs (30 nm diameter, but particles agglomerated in culture medium). BV2 microglia cells were exposed to TiO2 for periods of up to 120 mins, at concentrations ranging from 5 to 120 ppm, and ROS production determined. The oxidative response mediated by TiO2 within microglia had two components. Firstly, a rapid increase in ROS production was observed at 1-5 minutes, and termed an oxidative burst. This was followed by a greater, sustained ROS release from 60-120 minutes. It is also relevant, that despite having a primary particle size of 30 nm, the particles were observed to aggregate within the cell culture medium, which increased with increasing concentration, and this appeared to promote ROS production, perhaps due to the greater uptake of larger structures by phagocytosis. Internalised particles were found within the cytoplasm. Mitochondria located in the vicinity of the internalised aggregates were found to be swollen and disrupted, which was postulated by the authors to promote an apoptotic or necrotic response.

Similarly, Long et al. [43] investigated the neurotoxicity of nanoparticulate TiO2 (up to 120 ppm). It was illustrated that particle aggregates were phagocytosed by BV2 microglial cells, and contained within membrane bound vesicles. ROS production by cells was associated with particle exposure, and stimulated the upregulation of genes involved with inflammation, apoptosis, and cell cycling, and a down-regulation in energy metabolism. It was therefore evident that increased ROS production as a consequence of TiO2 exposure also triggered cytotoxicity via apoptosis. However, TiO2 exposure was non-toxic to N27 neuronal cells, following a 72 hour exposure, despite being internalised. These results reinforce the finding that particle internalisation does not necessarily equate to the initiation of a toxic response within cells, and that cell types differ in their susceptibility to particle toxicity. In contrast there was evidence of neurone loss within primary cultures of rat striatum within 6 hours of exposure, which was suggested to occur via microglia mediated ROS production.

Information regarding the neurotoxicity of TiO2 is severely lacking, and restricted to the work conducted by one research group. However, it is evident that the response is likely to be dictated by the cell type under investigation. Again, an oxidant driven response appears to be integral to particle toxicity. At present, there is insufficient evidence to make generalisations regarding TiO2 mediated neurotoxicity.

Kidney models

Only one study was found that assessed the toxicity of TiO2 NPs to the kidneys, in vitro. L'azou et al. [44] investigated the effect of TiO2 (15 or 50 nm) on IP15 mesangial, and LLC-PK1 proximal tubular epithelial renal cells (at concentrations up to 512 μg/ml, for 24 hours). The different cell types exhibited different sensitivities to the toxicity of TiO2. No cytotoxicity was observed within IP15 cells on exposure to TiO2. However, TiO2 elicited a cytotoxic response within LLC-PK1 cells, suggesting that this cell type was more sensitive to TiO2 toxicity. This was postulated to derive from their high endocytic activity, and therefore greater internalisation of particles. The cytotoxic response was also observed to be size dependent, with a greater response exhibited by smaller particles. TiO2 (15 nm) was also able to induce morphological changes within both cell types (namely cell shrinkage, and detachment), and was internalised into cytoplasmic vacuoles. TiO2 was unable to induce an increase in ROS production. However, 13 nm carbon black NPs were consistently recognised as being more toxic, in terms of cytotoxicity, and ROS production, which was evident in both cell types. Toxicity to renal cells was therefore observed to be particle size, particle composition, and cell type specific.

The findings from the only study identified as studying the toxicity of TiO2 particles to the kidneys, suggest that the response of the individual cell populations vary. However, further investigations would be required to identify if the finding was applicable to other particle types, and if the same response is evident in vivo. Therefore it is difficult to draw definitive conclusions about particle toxicity to the kidney, due to insufficient data being available.

Immune system models

It is acknowledged that macrophages are primarily responsible for the clearance of particles at sites of exposure, and accountable for the accumulation of particles within different target sites, which is anticipated to derive from their role within the reticuloendothelial system. The consequences of particle uptake by macrophages is therefore of interest, particularly the initiation of an oxidative driven or inflammatory response, and impact on their phagocytic function. In addition, the ability of particles to affect other immune cell populations, such as lymphocytes, is of interest as particles would be expected to encounter these cells within the blood.

Renwick et al. [45] determined the impact of particle exposure on the phagocytic activity of macrophages. It was observed that TiO2 (and CB) particles were able to impair J774.2 macrophage cell phagocytic activity (indicated by the uptake of 2 μm latex beads), in a concentration dependent manner. TiO2 NPs (29 nm) were more effective at inhibiting macrophage phagocytosis than their microparticulate (250 nm) counterparts, and the same response was exhibited by CB particles. It is likely, that the impairment of phagocytosis occurs as a consequence of the fact that cells are overwhelmed by the TiO2 particle burden, and are therefore unable to further internalise the latex particles. This was suggested to be the case, as an increase in particle-laden macrophages correlated with a decrease in phagocytic cells. The impairment in phagocytic function of macrophages would be anticipated to impact on the clearance of particles, thus increasing their retention, and thereby prolonging their interaction with cells, and enhancing their propensity for eliciting damage.

Afaq et al. [46] demonstrated that alveolar macrophages were increased in rats on intratracheal exposure (2 mg/rat, with observations made for a period of up to 16 days) to TiO2 NPs (<30 nm), which was maximal at 8 days post exposure. The exposure of alveolar macrophages to TiO2 was associated with lipid peroxidation, glutathione depletion and enhanced hydrogen peroxide production, illustrating the oxidative response that developed. Exposure was also associated with cytotoxicity.

Kang et al. [47] determined the mechanisms underlying TiO2 (25 nm) NP toxicity within lymphocytes, at concentrations ranging from 20 to 100 μg/ml, for up to 48 hours. A particular focus of the study was on the oxidative potential of TiO2. A dose and time dependent decrease in cell viability was observed. TiO2 NPs also exhibited a genotoxic effect, whereby the frequency of micronuclei formation increased, which insinuated chromosomal damage occurred on exposure of cells to TiO2. DNA damage was also observed, using the Comet assay. Importantly, the pre-treatment of cells with the antioxidant N-acetyl cysteine (NAC), decreased TiO2 mediated DNA damage, thereby inferring that the response was ROS mediated. ROS production was subsequently confirmed to be increased by TiO2 exposure. The DNA damage inflicted by TiO2 resulted in the activation of a protective response in the form of increased levels of p53 protein, highlighting an attempt of the cells the repair the damage mediated by TiO2. TiO2 mediated increases in ROS production, and therefore DNA damage, was thought to be central to its toxicity within lymphocytes.

Investigation of the impact of TiO2 particle exposure on immune cell function is lacking. Alveolar macrophages are recognised as being integral to the removal of particles from sites of deposition, and are, in the main, responsible for the accumulation of particles within different target sites. It is therefore of concern that particles appear to impair macrophage phagocytosis, and this should be considered further in future experiments. It is also suggested that particles initiate an oxidative driven response within immune cells that may have inflammatory and/or cytotoxic consequences.

The biological mechanisms driving TiO2 nanoparticle toxicity

TiO2 mediated inflammatory responses

Inflammatory responses have been illustrated as being a prominent feature of a number of studies that investigate the toxicity of nanoparticles such as TiO2. In vivo, the infiltration of neutrophils has been repeatedly demonstrated to characterise the inflammatory response to TiO2 NPs [3, 6], in addition to elevated, particle laden macrophage numbers [10, 12]. This is likely to be mediated by increases in inflammatory mediators, such as IL-8, and TNFα that have been observed in vivo [12] and in vitro [18]. The inflammogenic nature of TiO2 has been observed within a number of target sites including the lung, brain, and cardiovascular system, and so is not an organ specific observation. Inflammation has been demonstrated to be recurrently associated with TiO2exposure, with its manifestation evident within different models, in response to different TiO2 samples. Therefore, in general the inflammatory response that manifest is not organ, or particle sample specific.

TiO2 mediated oxidative responses

A focus of a number of investigations has been to identify the contribution of oxidative stress to the inflammatory and cytotoxic responses elicited by TiO2. Studies previously discussed found that oxidative stress was a prominent feature of the response to TiO2 NPs. These include evidence of increased ROS production, depletion of cellular antioxidants, increase in oxidative products (such as lipid peroxidation) or evidence that toxicity is diminished on pre-treatment with antioxidants. These observations have been made in a variety of cell types, including for example; lung epithelial cells [26], fibroblasts [35], and microglia [42]. A summary of the oxidative responses mediated by TiO2 can be found in table 2.

Table 2 The oxidative potential of TiO2

It is acknowledged that the development of a moderate level of oxidative stress is connected to the initiation of inflammatory responses through the activation of ROS sensitive signalling cascades [48]. Consequently Kang et al. [49] compared the ability of nano (21 nm) and microparticulate (1 μm) forms of TiO2 (0.5-200 μg/ml) to initiate an oxidative response within RAW 264.7 macrophages, following exposure for up to 24 hours. No alterations in cell viability were observed on exposure of cells to both forms of TiO2. This finding is of importance, as when assessing the processes that underlie particle toxicity, sub-lethal concentrations are required. ROS production was increased by TiO2, and was greatest in magnitude for the NPs. The ROS sensitive mitogen-activated protein kinase (MAPK) signalling pathway was activated by TiO2 exposure (indicated by ERK1/2 phosphorylation), which was suggested to initiate the increase in pro-inflammatory mediator production (TNFα, MIP-2). Within the endpoints measured, the response exhibited by NPs was consistently greater than that of their larger counterparts, when administered at an equivalent mass basis. The oxidative stress paradigm therefore held true within this study; specifically that particle mediated ROS production (at a moderate level) stimulates signalling cascades that promote the activation of transcription factors that, in turn, initiate an inflammatory response. Accordingly, the ability of particles to promote ROS production is thought to be central to their ability to regulate inflammatory responses.

In line with these findings, Xia et al. [50] investigated whether TiO2 (11 nm), ZnO (13 nm), and CeO2 (8 nm) oxidative stress development or particle dissolution contributed to their toxicity. RAW 264.7 macrophages and BEAS-2B epithelial cells were exposed to the NP panel for up to 15 hours, at concentrations up to 60 μg/ml. However, only ZnO particles were capable of inducing cytotoxicity, within both cell types. In addition, ZnO stimulated an increase in ROS production, which in turn, increased HO-1 expression, and activated the JNK signalling pathway, which paralleled the release of IL-8 and TNFα. Increased calcium release was also mediated by ZnO, and was associated with mitochondrial damage. These findings therefore imply that ZnO elicits an oxidant driven response that was responsible for the initiation of an inflammatory and cytotoxic response, which was absent for TiO2 and CeO2. This is an important finding, as contrary to evidence that TiO2 can exhibit an oxidative mediated response, this does not always transpire, and therefore the experimental model and TiO2 sample may be influential to the observed response. On the contrary, CeO2 was able to stimulate a cytoprotective response, whereby pre-treatment of cells with CeO2 protected against diesel exhaust particle mediated cell damage (which is known to be oxidant driven). All particle types were internalised by an endocytic mechanism, however, only ZnO accumulated within lysosomes, which was suggested to drive their ability to inflict oxidative injury, and promote particle dissolution. ZnO also induced ultrastructural alterations, including nuclear fragmentation, apoptotic body formation and mitochondrial disappearance. ZnO was therefore recognised as being the most toxic particle within the panel, however this was thought to be accounted for, in part, by their dissolution and release of Zn2+ ions. The toxicity of the metal oxide particles was therefore suggested to be driven by their oxidant properties, which was dependent on their composition, dissolution and intracellular fate. The composition of NPs has been demonstrated to be integral to their toxicity within this study, and thus not all metal oxide particulates are expected to behave similarly.

Lu et al. [51] had a slightly different focus and determined the impact of TiO2 (0.2-3 mg/ml) on protein tyrosine nitration (using bovine serum albumin (BSA) as a model protein). Protein nitration was associated with TiO2 exposure, in the presence of UV light, with anatase TiO2 samples being more photoactive than rutile forms. Antioxidants, such as GSH and vitamin E were able to prevent against the protein nitration observed. To determine if proteins contained within mouse skin were targets for TiO2 mediated nitrosation, similar tests were conducted using mouse skin homogenate. Again, anatase forms of TiO2 were observed to exhibit a greater propensity to nitrosate proteins. This study therefore highlighted the photocatalytic activity of TiO2, and the superior toxicity of anatase forms, and demonstrated the ability of TiO2 to induce nitrative stress.

Oxidative responses exhibited by TiO2 particles have been a focus of a number of studies, and its recurrence has prompted the suggestion that oxidative stress drives the inflammatory and cytotoxic responses evident. It is relevant that the level of oxidative stress, which is inevitably related to the duration or concentration of particles administered, drives the nature of the response that follows particle exposure. Specifically, at moderate levels of oxidative stress, inflammatory responses may be stimulated due to the activation of ROS sensitive signalling pathways [48]. At higher levels of oxidative stress, cytotoxicity is evident, as cells are damaged by an overwhelming burden of ROS [48].

Uptake of TiO2 into cells

The clearance of TiO2 particles by phagocytic cells has been a particular focus of studies, due to their ability to remove particles from the exposure sites, circulation or secondary target sites. However, a number of non-phagocytic cell populations also have the propensity to internalise particles. Considering the internalisation of particles by cells is of relevance as particle uptake may enhance their toxicity due to their interference with normal cellular physiology and function. However, it is also possible that particles act from outside the cell to elicit toxicity that is mediated by interactions of particles with the cell surface. Furthermore, particles can be internalised by cells and not impact on cell function, which is likely to be driven by the intracellular location in which they accumulate or the extent of uptake.

Stearns et al. [52] demonstrated that 50 nm TiO2 (40 μg/ml) was internalised by alveolar A549 epithelial cells into membrane bound vesicles after 3, 6 and 24 hour exposures. However, it was observed that the uptake of particles by these cells was limited to their aggregated form. TiO2 internalisation was suggested to be mediated via phagocytosis, due to the fact that plasma membrane projections surrounded and engulfed the particles prior to uptake.

A focus of a number studies has been to reveal the contribution of macrophage populations to the clearance of particles, particularly within the lung. However, Geiser et al. [53] argued that alveolar macrophages were not primarily responsible for the uptake and clearance of NPs within the lung in vivo. Rats were exposed to TiO2, via inhalation (0.1 mg/m3), and alveolar macrophages were isolated, and the internalisation of particles assessed, 1 or 24 hours post exposure. Large TiO2 particles (3-6 μm) were more effectively cleared by alveolar macrophages, than their smaller (20 nm) counterparts. Therefore TiO2 NPs were able to bypass the important alveolar macrophage mediated clearance mechanism within the lung, due to their small size. These findings could be expected due to the known size limitations of uptake processes such as phagocytosis, which is thought to be restricted to substances that are 1 to 5 μm in size [54]. The ability of TiO2 NPs to evade phagocytosis was confirmed by the findings that they were not enclosed by a vesicular membrane equivalent to that surrounding the larger TiO2 particles. Therefore it was suggested that phagocytosis was not responsible for NP uptake, and instead it was suggested that a sporadic, non-specific mechanism of uptake enabled NP uptake by macrophages. Alternatively it was suggested that NPs were internalised unintentionally when macrophages phagocytosed other material. The findings of Geiser et al. [53] are also able to provide an explanation for the finding that only TiO2 agglomerates, and not individual NPs, were internalised by A549 cells in vitro [52]. This was also recognised by Stearns et al. [52] who highlighted that there was limited evidence of the uptake of individual particles by cells. As Stearns et al. [52] utilised epithelial cells, and Geiser et al. [53] used macrophages which are professional phagocytes, it is relevant that the processes that drive particle uptake may be cell specific, and reliant on their function.

Rothen-Rutishauser et al. [55] utilised an in vitro airway wall model (triple cell co-culture), containing macrophages, epithelial, and dendritic cells to determine the translocation of particles between the different cell types, and to assess the intracellular fate of particles. Membrane-bound aggregates (>0.2 μm) of TiO2 were evident within all cell types. However, smaller aggregates (<0.2 μm) were apparent within the cell cytoplasm, but were not membrane bound, suggesting that the entry mechanism of particles was size dependent. However, it is relevant that the behaviour of gold, and polystyrene particles was also assessed, and it was found that their intracellular location differed, which implied that different NPs are internalised by different mechanisms or follow different intracellular trafficking processes.

The uptake of TiO2 by a variety of cell types has been demonstrated on numerous occasions. The consequences of particle internalisation are anticipated to be oxidative and cytotoxic in nature. It is anticipated that particle physicochemical properties may influence their internalisation by cells. Accordingly, the size and surface charge of particles has the ability to influence their uptake by cells. In addition, particle size and the cell type under investigation also have the potential to determine the mechanism of uptake and intracellular fate of particles.

Genotoxicity of TiO2

Evidence of genotoxicity has been previously encountered within a number of studies; micronuclei development is associated with TiO2 exposure, which is indicative of chromosomal damage, DNA damage has also been observed in response to TiO2 particulate exposure.

Rahman et al. [56] investigated the potential for TiO2 to elicit DNA damage within SHE fibroblasts. Micronuclei were evident within cells exposed to nano (<20 nm), but not micro (>200 nm) TiO2 particles (exposed at concentrations up to 10 μg/cm2, for up to 72 hours), which insinuates that chromosomal damage has occurred. The NPs also triggered the induction of apoptosis within cells, which is recognised as a common response to DNA damage.

Karlsson et al. [57] evaluated the ability of a variety of metal oxide NPs (copper oxide (CuO), ZnO, TiO2, Fe3O4) to induce oxidative stress and DNA damage within A549 lung epithelial cells, at concentrations of up to 80 μg/ml for a period of up to 18 hours. CuO and ZnO were able to decrease cell viability, but TiO2 and Fe3O4 were not capable of eliciting such an effect. In addition, all metal oxide particles, except Fe3O4 were able to elicit DNA damage, as determined by the Comet assay. The CuO NPs were consistently demonstrated to be the most toxic particle within the panel, which was thought to be drive by their ability to induce oxidative stress, and in turn DNA damage, which prompted cell death. It was suggested that the release of Cu2+ ions may contribute to the response, but are not solely accountable for the toxicity of CuO NPs. Importantly, this study illustrated that not all metal oxides behave similarly, and that their dissolution was able to contribute to their toxicity.

Dunford et al. [58] investigated the DNA damage inflicted by a panel of TiO2 containing, commercially available sunscreens on fibroblasts, when exposed for a period of up to 60 minutes. It was demonstrated that, on sunlight illumination, the ability of TiO2 to cause DNA damage was enhanced in plasmid DNA and cells. Anatase forms of TiO2 were more effective at inducing damage than rutile forms. Mannitol (an antioxidant) was able to protect against DNA damage, therefore implying that TiO2 mediated ROS production is central to its genotoxic potential. Consequently, absorption of UV light by TiO2 is thought to stimulate ROS generation, which, in turn, leads to DNA damage.

Nakagawa et al. [59] investigated the genotoxicity of a panel of TiO2 NPs (in rutile and anatase forms, ranging from 0.021 to 0.255 μm in diameter). A number of different in vitro genotoxicity tests were utilised; a microbial mutation assay (Salmonella typhimurium), the Comet assay to detect DNA damage (mouse lymphoma L5178Y cells) and a chromosome aberration assay (Chinese hamster CHL/IU cells). TiO2 was used at concentrations up to 3200 μg/ml, with an exposure duration of up to 24 hours, and experiments were conducted in the absence or presence of UV light. Without UV light, TiO2 NPs induced no, or very limited genotoxicity. However in the presence of UV light TiO2 elicited DNA damage and chromosome aberrations (but no gene mutations) that was greatest for anatase forms. The phototoxicity of TiO2 was therefore realised in this study.

Theogaraj et al. [60] investigated the UV dependence of the genotoxic potential of a panel of TiO2 NPs (in rutile and anatase forms, all of which had a diameter of <60 nm). CHO cells, were exposed to TiO2 at concentrations up to 5000 μg/ml for a period of 3 hours. In contrast to previously discussed investigations, no chromosomal alterations were observed within exposed cells, in the presence or absence of UV light.

In a different approach, Driscoll et al. [61] determined if there was a relationship between the inflammatory and genotoxic potential of several particles; namely ufCB (14 nm), TiO2 (0.18 μm) and α quartz (0.9 μm). Rats were exposed to particles via intratracheal instillation, at a dose of 10 or 100 mg/kg, and analysis conducted 15 months post exposure. All particle types induced the infiltration of neutrophils into the lungs, indicating that particles could induce an inflammatory response. Hypoxanthine-guanine phosphoribosyl transferase (hrpt) gene mutation frequency was increased within alveolar type 2 cells, on exposure to particles. This response was replicated in vitro, following the exposure of RLE-6TN alveolar epithelial cells to BAL cells isolated from particle treated animals. It was thus suggested that BAL cell derived ROS were responsible for the mutagenic effects that transpired, and they were suggested to derive from neutrophil activity. Therefore the particle mediated neutrophillic driven inflammatory response within the rat lung, was postulated to increase the frequency of mutations within alveolar epithelial cells, which was a dose and material dependent finding. Accordingly, quartz was consistently found to elicit the greatest genotoxic response, followed by carbon black and then titanium dioxide, and this was related to their inflammogenic potential. Therefore, the particles themselves are not thought to be inherently genotoxic, but the inflammatory response, and in particular neutrophil presence (and therefore increased oxidant burden) instigated is anticipated to mediate this effect.

Concern regarding the genotoxic potential of TiO2, has also emanated from findings that have illustrated that tumours develop in vivo following a chronic exposure regime ([13] see earlier). However, as stated previously this is very much dependent on the exposure conditions, and in particular the exceptionally high doses that were administered within this model.

The ability of TiO2 NPs to inflict DNA damage has been observed on numerous occasions, and is thought to be driven by particle mediated ROS production, with cell death often stimulated as a protective response. In addition, the manifestation of genotoxic events, as a secondary consequence of an inflammatory response is also evident, and requires further investigation. The appearance of TiO2 mediated genotoxicity appears to be influenced by the crystal form of the sample, particle dissolution and light conditions. The finding that chronic exposure to TiO2 can induce tumours in vivo (under extreme exposure conditions), also implies that the genotoxicity exhibited by TiO2 is not limited to in vitro investigations.

Reproductive toxicology of TiO2

Evaluation of TiO2 NP effects on the reproductive system is limited to one in vitro study.

Komatsu et al. [62] determined the potential for TiO2 NPs, DEP and ufCB to impair the function of male mouse reproductive system. This study evaluated the direct effect of NPs on testis-constituent cells, and examined the effect of TiO2 on mouse Leydig TM3 cells, the testosterone-producing cells of the testis. TiO2 NPs (25-70 nm) at concentrations 1 to 1000 μg/ml were examined and uptake into Leydig cells was detected using transmission electron microscopy (TEM) or field emission type scanning electron microscopy/energy-dispersive X-ray spectroscopy (FE-SEM/EDS). TiO2 was more cytotoxic to Leydig cells than the other carbon based particles used in the study. The proliferation of Leydig cells was suppressed transiently by treatment with TiO2. TiO2 NPs were taken up by Leydig cells, and in turn affected cell viability, proliferation and gene expression.

Only one available study examined the effects of TiO2 on male reproductive physiology. The literature highlights toxicity of TiO2 NPs to male Leydig; however investigations are limited in number and sample size, and therefore further studies would be required to confirm such a finding. No literature examining TiO2 NP effects on organs or cell types in the female reproductive system were found.

Linking the physicochemical attributes of TiO2 particulates to their pathogenicity or toxicity

As TiO2 particulates exhibit great diversity, with regards to their size (and therefore surface area), composition, and crystal form, it is necessary to outline what properties are most influential in driving the toxicity of TiO2. However it is also relevant to highlight that the experimental set up is also able to influence the findings, including the choice of species or cell type, method of exposure, and particle dispersal (see later)

Size dependency

Particle dimensions are recognised as being fundamental to their toxicity. This derives from the fact that NPs have been consistently demonstrated to be capable of eliciting more pronounced toxicity than their larger (microparticulate) counterparts. The size dependency of TiO2 toxicity has been frequently demonstrated [3, 12, 24, 28, 45, 47], and appears to be applicable to a variety of TiO2 forms, and occurs regardless of the model used (table 3). However it is relevant that a high degree of particle aggregation and agglomeration is associated with TiO2 administration, and so exposure to particles is unlikely to occur in a 'nano' form. However, what seems to be critical to the toxic potential of particle samples is the size of particles that make up the agglomerates. It is relevant that the terms agglomeration and aggregation are often used interchangeably within the field of nanotoxicology. However, some authors have suggested that NP aggregation and agglomeration are distinct phenomena with agglomerates formed by clusters of particles that are held together by electrostatic interactions, whereas aggregates are formed from covalently fused or sintered particles that are not easily separated [63].

Table 3 A summary of the size dependent toxicity of TiO2 particulates

As discussed, the superior toxicity of nanoparticulate forms of TiO2 has been repeatedly documented. However, a study conducted by Dick et al. [64] illustrated that compared to other NPs such as nickel, cobalt or carbon black, TiO2 nanoparticles were less potent at inducing an inflammatory response within the lung, despite their similarity in surface area. This was assumed to be related to their lower capacity to generate free radicals, and therefore perhaps particle composition and surface chemistry are also important contributors to toxic responses [64].

Bermudez et al. [7] exposed rats to TiO2 via inhalation, to investigate their pulmonary toxicity. Despite the fact that the NPs had a primary particle size of 21 nm, the aerosol generated contained particle aggregates (1.37 μm). This is a common experience in the exposure of animals or cell culture models to nanomaterials, and is likely to be encountered within the exposure of humans. However, despite the formation of aggregates, the NPs still exerted toxicity. Similarly, aggregates of TiO2 NPs (1.44 μm) have been demonstrated to be more toxic than similarly sized aggregates of their larger counterparts [3]. However, Grassian, et al. [10] demonstrated that the properties of agglomerates of particles that formed during aerosol generation were dependent on the primary particle size. Specifically, 21 nm particle based agglomerates were less dense than their 5 nm counterparts whose agglomerates contained particles that were more tightly packed. It was found that 21 nm TiO2 was more toxic than 5 nm particles following inhalation and instillation exposure. The agglomeration state of particles was therefore suggested to dictate their toxic potential. Consequently, although 21 nm particles were larger, it is anticipated that they would form agglomerates that would more easily deagglomerate due to the weaker interactions that held the particles together. As a result, 21 nm particles would be available as smaller structures to stimulate an enhanced toxic response. Furthermore the aggregates formed were larger within inhalation preparations than intratracheal suspensions, and so this may explain why intratracheal exposures produced a greater toxic response.

The greater toxicity of NPs, compared to their larger counterparts, is anticipated to be driven by the greater surface area of smaller particles, when administered at an equivalent mass dose. In fact, the surface area of particles has been previously suggested to dictate the toxic potential of particles [6567]. In line with this hypothesis, Sager et al. [68] addressed which dose metric was most influential in driving the toxicity of TiO2; the surface area or mass of particles administered. NPs (suspended in BALF) were observed to agglomerate to a diameter of 200-300 nm (despite having a primary particle size of 21 nm). This finding emphasises that particle exposure often does not occur in an individual particle form. Microparticle (1 μm) and NP particles were able to induce an inflammatory response (neutrophil infiltration, LDH activity, TNFα, MIP and IL-1β release), subsequent to intratracheal administration of rats. It was of interest that a higher mass of microparticulate TiO2 was required to obtain the same inflammogenic response as nanoparticle TiO2. Therefore, the inflammatory and cytotoxic potential of NPs was greater than that of microparticles, when administered at an equivalent mass dose. However, when the dose of particles administered was equivalent, in terms of surface area delivered (which necessitates the administration of a much higher mass of microparticles), both particle types behaved similarly. These findings suggests that the size, and surface area of particles is integral to their toxicity. Similarly, Sager et al. [69] determined whether the surface area of ufCB and TiO2 NPs drives their inflammatory potential. Particles were administered to rats via intratracheal instillation, and toxicological observations made up to 42 days post exposure. At equivalent surface area doses, both particle types induced an inflammatory response that was characterised by the infiltration of neutrophils at day 1 which was of a similar magnitude. However the response decreased with time for ufCB, but for TiO2 the response was sustained in nature. This pattern of response was similar for BALF protein levels. Consequently, the composition of particles may also contribute to their toxicity. This observation remains debatable as others have disputed this finding [11].

The size (and therefore surface area) of TiO2 nanoparticles is known to be fundamental to their toxicity. The aggregation and agglomeration state of nanoparticles is also likely to be influential to their toxicity. Consequently, it is often the case that although the primary particle size is stipulated by investigators; cells and animals are not exposed to the particles in this form, [27, 32, 33, 35, 42, 43, 52, 7072]. However it is often observed that agglomerates/aggregates of nanoparticles are more toxic than similarly sized agglomerates/aggregates of their larger counterparts. Therefore, the propensity of particles to aggregate has prompted investigators to prevent against its occurrence through the use of sonication, or inclusion of dispersants such as within particle suspensions, but this also needs to take into account what is relevant to human exposure.

Influence of TiO2 particulate surface chemistry/modification/coatings to their toxicity

As for other nanoparticles, the surface of TiO2 particulates can be altered through the attachment of surface moieties, a process which is termed particle functionalisation. The surface of TiO2 particles can also be modified through coating with aluminium oxide, or silica, which has generally been encountered within sunscreens to enhance protection from UV radiation [9].

Warheit et al. [9] investigated the impact of surface treatments (alumina or silica) on TiO2 toxicity, within the lungs (see earlier). It was demonstrated that those exhibiting the greatest toxicity, were those that contained the highest aluminium oxide and/or silica content on the particle surface. However, overall it was observed that TiO2 particles induced low pulmonary toxicity, despite the ability of surface coatings to influence this.

Oberdorster [73] demonstrated that the surface properties of TiO2 were able to influence their toxicity. This was demonstrated via the intratracheal exposure of rats (500 μg/animal, for 24 hours) to TiO2 (20 nm) that remained uncoated (hydrophilic) or received a silane coating (hydrophobic). The samples varied, with regards to their inflammatory potency within the lungs, so that the uncoated particles induced a lower response, than their coated counterparts.

Hohr et al. [74] investigated the impact of surface properties on TiO2 toxicity. Microparticulate (180 nm diameter), uncoated and coated (via methylation) NPs (20-30 nm diameter) of TiO2 were exposed to rats, via intratracheal instillation (1 or 6 mg/rat), and analysis conducted 16 hours post exposure. Administration of all particle types stimulated the infiltration of neutrophils, with the effect most pronounced with administration of NPs. It is also of interest that the coated (hydrophobic) NPs tended to stimulate neutrophil recruitment to a lesser extent than the uncoated (hydrophillic) particles. Protein, LDH, TNFα and MIP-2 levels in BAL also exhibited a similar pattern, with regards to the potency of the different samples, but overall the impact of surface methylation on TiO2 toxicity was negligible. However, it was concluded that the toxicity of TiO2 was driven by the surface area of particles, and not their surface coating. Therefore, a stronger inflammatory response was associated with NPs compared with their larger counterparts and so the effect of methylation was negligible.

Singh et al. [75] determined the impact of TiO2 size (microparticles: 40-300 nm, NPs: 20-80 nm), surface modification (using methylation, to make particles more hydrophobic) and radical generating potential on their uptake and toxicity within A549 lung epithelial cells. TiO2 particles, regardless of their size or methylation were phagocytosed as clusters of particles. Internalisation of NPs via clathrin mediated endocytosis, was only apparent for small particle clusters that were less than 30 nm, illustrating that this may be a size dependent phenomenon. ROS generation, and IL-8 production exhibited by cells exposed to NPs was greater than that of microparticles, and the findings were unaffected by the methylation of the particle surface. An important observation was that both particle types were present in an aggregated form, and that the NPs were in a smaller size range (<500 nm) than their larger counterparts (2000-5000 nm). Therefore the superior toxicity of NPs held true, despite their tendency to aggregate. Therefore, even when in an aggregated form the toxicity of TiO2 particles was assumed to be primarily driven by their surface area.

Thevenot et al. [76] determined the impact of TiO2 NPs, with various surface modifications (-COOH, -OH or -NH2 functional groups), on the survival of a variety of cancer cell lines. Exposure concentrations were high (up to 10 mg/ml) and occurred for a duration of 3 to 24 hours. The cancer cell lines used were B16F10 and BF16F1 melanoma, Lewis lung carcinoma, JHU prostate cancer, and 3T3 fibroblast cell lines. The different cell lines showed different sensitivities to the cytotoxicity of TiO2 NPs. 3T3, B16F10 and BF16F1 cells were unresponsive to all types of TiO2. However, TiO2 decreased the viability of JHU and LLC cells in a dose dependent manner. In addition, by altering the surface chemistry of TiO2 particles, the toxicity of TiO2 could be modified. In general, NH2, and OH surface modified TiO2 exhibited greater toxicity than COOH modified TiO2. This study therefore illustrated that TiO2 toxicity was very much dependent on the cell type in question, the surface chemistry of TiO2 NPs, and concentration of particles used.

Rehn et al. [70] illustrated that unmodified, or silane coated forms of TiO2 were non toxic to the lungs. Rats were exposed to the different forms of TiO2 (20 nm) via intratracheal instillation (up to 1.2 mg/rat) and the inflammatory and genotoxic effects within the lung were determined, at 3, 21 and 90 days post exposure. All forms of TiO2 were able to stimulate a modest increase in neutrophils and macrophages within the lungs (at day 3), but this was not persistent in nature, and was not associated with TNFα production. In addition, no genotoxicity was evident with TiO2 exposure. In contrast, quartz induced a persistent and progressive inflammatory response, with a genotoxic element to the response also observed. However, the only genotoxic endpoint that was assessed was the detection of 8 oxy-guanine, so perhaps a more comprehensive genotoxic study should be considered for TiO2.

The modification of the surface of TiO2 particles is able to influence its toxicity, however, this is likely to be dependent on the modification, and cell type in question.


TiO2 exists in two main crystal phases, termed rutile and anatase. These TiO2 forms vary with regards to their crystalline structure and surface properties, which is responsible for differences within their toxicity [77]. Anatase has been demonstrated to be the most toxic form of TiO2, with a number of previously mentioned studies supporting this conclusion (see for example [17]). The photoactivity of TiO2 is also dictated by the crystal phase, and therefore surface characteristics of particles (such as oxygen vacancies on the particle surface), with anatase forms having a greater capacity to generate ROS on exposure to light (see later). A summary of the contribution of particle of crystallinity to TiO2 toxicity is demonstrated in table 4.

Table 4 The importance of crystallinity to TiO2 toxicity

Sayes et al. [71] assessed the in vitro toxicity of anatase (10 nm), rutile (5 nm) or anatase/rutile (3 nm) TiO2 samples to A549 epithelial cells and HDF fibroblasts, at concentrations ranging from 3 μg/ml to 30 mg/ml, for up to 48 hours. The level of cytotoxicity exhibited by the different particle types varied. Accordingly, anatase particles were the most cytotoxic, while the rutile TiO2 particles were the least toxic. This response was paralleled within the release of IL-8 from cells. TiO2 particle stimulated ROS generation and photoactivity was also greatest for pure anatase samples. Overall, the results implied that the anatase samples had the greatest toxic potency. Consequently, the authors suggested that it was the phase of TiO2 that was integral to its toxic potential. Therefore, despite the relatively large surface area of rutile TiO2 NPs (112 m2/g), their surface chemistry is thought to be less reactive than that of anatase samples (that had a larger surface area of 153 m2/g), and so rutile particles are less toxic. The findings also suggested that the ability of TiO2 particles to generate ROS, governs their cytotoxic and inflammatory potential, which is dictated by the crystal structure of TiO2, and hence toxicity is not solely driven by surface area. However, rutile NPs were observed to agglomerate to the greatest extent, and so this may also explain why this particle type was observed to be less toxic. Although the findings demonstrate that particle crystallinity is important to TiO2 toxicity, the relevance of the findings is questionable, as in general, toxicity was only observed at higher concentrations (greater than 300 μg/ml), which are not deemed to be physiologically relevant (see later).

Pan et al. [72] investigated the toxicity of a panel of TiO2 particles to primary human dermal fibroblasts. Uncoated rutile TiO2 (15 nm), polymer coated rutile TiO2 or uncoated anatase TiO2 (200 nm) particles were exposed to cells for up to 11 days, at concentrations up to 0.8 mg/ml. Uncoated rutile TiO2 caused alterations in cell morphology, so that cells had a smaller cell area, became elongated, and detached from the culture surface. Anatase TiO2 caused a greater magnitude of damage to cells, with severe morphological changes observed including breakage of actin filaments, and plasma membrane rupture within exposed cells. Rutile TiO2 did not impact on cell viability, but did slow cell proliferation rate but were internalised by cells, and were contained within cytoplasmic vesicles. Anatase TiO2 was internalised by cells, and could access the nucleus. H2O2 production by cells was greatest with anatase particles, implying that their ability to generate ROS was greatest within the TiO2 particle panel. Anatase particles were therefore more potent that rutile particles at inducing cell damage. Rutile particles were then coated in order to prevent against particle adhesion to the plasma membrane, and thus limit their potential for internalisation. Coated particles were not evident on the cell surface, or within the cell interior. Consequently, a polymer coating abolished particle toxicity, which was suggested to derive from the lack of particle adherence to the cell surface, and thus limited internalisation, so that they were unable to impact on normal cell function. This study therefore demonstrates that particle coatings, and therefore surface chemistry, are able to modify particle toxicity. However, this requires further investigation, and may be dependent on the surface coating and cell type in question.

The phase composition of TiO2 has been proposed to be influential in dictating the toxic potency of particles. This is likely to derive from their substantially different surface chemistry, with anatase forms exhibiting the greatest photocatalytic and biological activity. The importance of crystallinity, to the toxicity of particles has also been demonstrated for SiO2 (see for example Kaewamatawong et al. [78]).

Furthermore the ability of UVA or visible light to increase the toxic potency of TiO2 particles, through increased ROS production, has been a focus of a number of previously discussed studies [see for example [36, 51, 58]], but does not always transpire [38, 60]. This phenomenon is anticipated to be exploited, particularly within the treatment of cancer [36]. The crystal form of TiO2 has been suggested to be responsible for driving the photoactivity of particles, due to the importance of surface properties [77].

Quality of experiments conducted

The described studies vary greatly with regards to their experimental set up. It has been highlighted that the size of particles, their crystallinity, route of exposure, particle concentration, experiment duration and species used can all impact on TiO2 particle toxicity. Therefore, deciphering what particle attributes are most influential to TiO2 toxicity is challenging. For example, determining if the crystal phase of TiO2 is primarily responsible for driving particle toxicity is confounded by the fact that the different TiO2 samples under investigation will not just vary with regards to their crystal phase but also, for example, their size and surface area. Consequently, it is difficult to isolate specific particle properties that are responsible for TiO2 toxicity, and it is likely that a variety of factors act in concert to dictate the toxic response observed. In line with this, it is also vital for investigators to carry out relevant physicochemical characterisation of their particle sample. It is suggested that prioritising the assessment of particle size, surface area, composition (including surface chemistry), crystal structure, and aggregation tendency would be of greatest relevance, due to their apparent importance to particle toxicity. If this is routinely carried out, this will facilitate making comparisons between different investigations to dissect out what particle attributes are driving the toxicity of TiO2.

A frequent observation within existing studies, is that excessively high concentrations of particles have been administered to animals or cells. It is therefore essential that investigators consider the relevancy of particle concentrations that are utilised, and justify their use if they are excessively high, as otherwise the relevancy of experiments is questioned. This is important as when administered at a large concentration, overload effects dominate, and the inherent toxicity of particles cannot be distinguished. For example, in vivo, this derives from the fact that an excessive particle burden cannot be efficiency cleared by defence mechanisms, and so the toxicity of particles is enhanced. In addition, the use of high particle concentrations is particularly concerning when assessing the toxicity of TiO2 to secondary target sites, such as the liver, as these targets are anticipated to be exposed to smaller concentrations of particles, due to the limited translocation of particles into the circulation from exposure sites. However, in saying this, wide concentration ranges can provide valuable information regarding the dose dependency of particle toxicity, and can also allow for the generation of threshold doses. However, for ethical reasons, the use of high concentrations should be restricted for in vivo work.

Gaps in current knowledge have been identified within the review, and thus there are a number of areas that should be considered with priority in future investigations. There is a paucity of data relating to the systemic transfer of TiO2 particles following exposure via the lungs, skin and gut, and this should be a focus of future experiments. Studies have focused on the dermal and pulmonary toxicity of particles, but there is an absence of data on the consequences of exposure to the GIT, and within damaged/diseased skin. Other relevant target organs include the liver, kidney, cardiovascular system and brain which is required due to the fact that NPs are likely to become systemically or neuronally available. The liver could be highlighted as a priority due to the propensity of particles to accumulate in this organ.


Due to historical reasons, a focus on the size (and surface area) dependence of TiO2 particulate toxicity has been repeatedly investigated, and confirmation that particle toxicity increases as particle size decreases has been consistent within wide ranging investigations. However, it has become evident that other physicochemical factors are able to contribute to TiO2 toxicity; including particle aggregation, crystal phase, and surface modification. The exposure method, dose administered, species used, cell type under investigation and light conditions also have the potential to impact on the toxicity of TiO2 particles, indicating that the experimental set up is also very influential to the toxicological observations. Therefore, making generalisations about TiO2 particulate behaviour should be approached with caution, and the findings from a limited number of studies should not be considered to be representative for TiO2 particles as a whole. This is important as particle characteristics and experimental set up influence the toxicological observations made.

The toxicity of TiO2 has been demonstrated to have inflammogenic, oxidative, and genotoxic consequences; with these endpoints considered to be inherently linked. It is of interest that the biological mechanisms identified as being responsible for driving the toxicity of TiO2 particulates is replicated within in vivo and in vitro settings. Cytotoxicity is also a common end point that is evaluated within studies, although the relevance of this is questionable (in terms of human exposure levels), except when establishing sub-lethal concentrations for subsequent studies that allow the identification of mechanistic processes that are responsible for toxicity. The ability of particles to exert toxicity at a variety of target sites is reliant on their transfer into blood, and this should therefore be a focus of future experimentation, as at this time, the systemic availability of TiO2 particles following exposure is uncertain. Accordingly, investigations into the toxicity of TiO2 via specific routes of delivery, or at particular cell and organ targets, are often insufficient in number to make definite conclusions about particle behaviour. In addition, the quality (including the concentrations used, experimental model), and relevancy of conducted experiments is an important consideration, which is of vital importance when considering the risk associated with TiO2 exposure.


  1. BSi Report: PAS 136 Terminology for nanomaterials. 2007. []

    Google Scholar 

  2. Maynard AD, Aitken RJ, Butz T, Colvin V, Donaldson K, Oberdörster G, Philbert MA, Ryan J, Seaton A, Stone V, Tinkle SS, Tran L, Walker NJ, Warheit DB: Safe Handling in nanotechnology. Nature 2006, 444: 267–269. 10.1038/444267a

    Article  CAS  PubMed  Google Scholar 

  3. Ferin J, Oberdorster G, Penney DP: Pulmonary retention of ultrafine and fine particles in rats. Am J Respir Cell Mol Biol 1992, 6: 535–542.

    Article  CAS  PubMed  Google Scholar 

  4. Ferin J, Oberdoster G, Penney DP, Soderholm SC, Gelein R, Piper HC: Increased pulmonary toxicity of ultrafine particles? I. Particle clearance, translocation, morphology. J Aerosol Sci 1990, 21: 381–384. 10.1016/0021-8502(90)90064-5

    Article  Google Scholar 

  5. Oberdorster G, Ferin J, Finkelstein J, Wade P, Corson N: Increased pulmonary toxicity of ultrafine particles? II. Lung lavage studies. J Aerosol Sci 1990, 21: 384–387. 10.1016/0021-8502(90)90065-6

    Article  Google Scholar 

  6. Renwick LC, Brown D, Clouter A, Donaldson K: Increased inflammation and altered macrophage chemotactic responses caused by two ultrafine particle types. Occup Environ Med 2004, 61: 442–447. 10.1136/oem.2003.008227

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  7. Bermudez E, Mangum JB, Wong BA, Asgharian B, Hext PM, Warheit DB, Everitt JI: Pulmonary responses of mice, rats, and hamsters to subchronic inhalation of ultrafine titanium dioxide particles. Toxicol Sci 2004, 77: 347–357. 10.1093/toxsci/kfh019

    Article  CAS  PubMed  Google Scholar 

  8. Bermudez E, Mangum JB, Asgharian B, Wong BA, Reverdy EE, Janszen DB, Hext PM, Warheit DB, Everitt JI: Long term pulmonary responses of three laboratory rodent species to subchronic inhalation of pigmentary titanium dioxide particles. Toxicol Sci 2002, 70: 86–97. 10.1093/toxsci/70.1.86

    Article  CAS  PubMed  Google Scholar 

  9. Warheit DB, Brock WJ, Lee KP, Webb TR, Reed KL: Comparative pulmonary toxicity inhalation and instillation studies with different TiO 2 particle formulations: impact of surface treatments on particle toxicity. Toxicol Sci 2005, 88: 514–524. 10.1093/toxsci/kfi331

    Article  CAS  PubMed  Google Scholar 

  10. Grassian VH, Adamcakova-Dodd A, Pettibone JM, O'shaughnessy PT, Thorne PS: Inflammatory response of mice to manufactured titanium dioxide nanoparticles: comparison of size effects through different exposure routes. Nanotoxicology 2007, 1: 211–226. 10.1080/17435390701694295

    Article  CAS  Google Scholar 

  11. Warheit DB, Webb TR, Reed KL, Frerichs S, Sayes CM: Pulmonary toxicity study in rats with three forms of ultrafine-TiO 2 particles: differential responses related to surface properties. Toxicology 2007, 230: 90–104. 10.1016/j.tox.2006.11.002

    Article  CAS  PubMed  Google Scholar 

  12. Chen HW, Su SF, Chien CT, Lin WH, Yu SL, Chou CC, Chen JJ, Yang PC: Titanium dioxide nanoparticles induce emphysema-like lung injury in mice. FASEB J 2006, 20: 2393–2395. 10.1096/fj.06-6485fje

    Article  CAS  PubMed  Google Scholar 

  13. Heinrich U, Fuhst R, Rittinghausen S, Creutzenberg O, Bellmann B, Koch W, Levsen K: Chronic inhalation exposure of wistar rats and two different strains of mice to diesel engine exhaust, carbon black, and titanium dioxide. Inhal Toxicol 1995, 7: 533–556. 10.3109/08958379509015211

    Article  CAS  Google Scholar 

  14. Ahn MH, Kang CM, Park CS, Park SJ, Rhim T, Yoon PO, Chang HS, Kim SH, Kyono H, Kim KC: Titanium dioxide particle-induced goblet cell hyperplasia: association with mast cells and IL-13. Respir Res 2005, 6: 34–43. 10.1186/1465-9921-6-34

    Article  PubMed Central  PubMed  Google Scholar 

  15. Inoue K, Takano H, Ohnuki M, Yanagisawa R, Sakurai M, Shimada A, Mizushima K, Yoshikawa T: Size effects of nanomaterials on lung inflammation and coagulatory disturbance. Int J Immunopath Pharmacol 2008, 21: 197–206.

    CAS  Google Scholar 

  16. Nurkiewicz TR, Porter DW, Barger M, Millecchia L, Rao KM, Marvar PJ, Hubbs AF, Castranova V, Boegehold MA: Systemic microvascular dysfunction and inflammation after pulmonary particulate matter exposure. Environ Health Perspect 2006, 114: 412–419.

    Article  PubMed Central  PubMed  Google Scholar 

  17. Wang J, Chen C, Liu Y, Jiao F, Li W, Lao F, Li Y, Li B, Ge C, Zhou G, Gao Y, Zhao Y, Chai Z: Potential neurological lesion after nasal instillation of TiO 2 nanoparticles in the anatase and rutile crystal phases. Toxicol Lett 2008, 183: 72–80. 10.1016/j.toxlet.2008.10.001

    Article  CAS  PubMed  Google Scholar 

  18. Wang J, Liu Y, Jiao F, Lao F, Li W, Gu Y, Li Y, Ge C, Zhou G, Li B, Zhao Y, Chai Z, Chen C: Time-dependent translocation and potential impairment on central nervous system by intranasally instilled TiO 2 nanoparticles. Toxicol 2008, 254: 82–90. 10.1016/j.tox.2008.09.014

    Article  CAS  Google Scholar 

  19. Chen J, Dong X, Zhao J, Tang G: In vivo , acute toxicity of titanium dioxide nanoparticles to mice after intraperitioneal injection. J App Toxicol 2008, 29: 330–337. 10.1002/jat.1414

    Article  Google Scholar 

  20. Mavon A, Miquel C, Lejeune O, Payre B, Moretto P: In vitro percutaneous absorption and in vivo stratum corneum distribution of an organic and a mineral sunscreen. Skin Pharmacol Physiol 2007, 20: 10–20. 10.1159/000096167

    Article  CAS  PubMed  Google Scholar 

  21. Schulz J, Hohenberg H, Pflucker F, Gartner E, Will T, Pfeiffer S, Wepf R, Wendel V, Gers-Barlag H, Wittern KP: Distribution of sunscreens on skin. Advan Drug Del Revi 2002, 54: S157-S163. 10.1016/S0169-409X(02)00120-5

    Article  CAS  Google Scholar 

  22. Kiss B, Biro T, Czifra G, Toth BI, Kertesz Z, Szikszai Z, Kiss AZ, Juhasz I, Zouboulis CC, Hunyadi J: Investigation of micronized titanium dioxide penetration in human skin xenografts and its effect on cellular functions of human skin-derived cells. Exp Dermatol 2008, 17: 659–667. 10.1111/j.1600-0625.2007.00683.x

    Article  CAS  PubMed  Google Scholar 

  23. Borm PJ, Robbins D, Haubold S, Kuhlbusch T, Fissan H, Donaldson K, Schins R, Stone V, Kreyling W, Lademann J, Krutmann J, Warheit D, Oberdorster E: The potential risks of nanomaterials: a review carried out for ECETOC. Part FibreToxicol 2006, 3: 11–46.

    Google Scholar 

  24. Wang J, Zhou G, Chen C, Yu H, Wang T, Ma Y, Jia G, Gao Y, Li B, Sun J, Li Y, Jiao F, Zhao Y, Chai Z: Acute toxicity and biodistribution of different sized titanium dioxide particles in mice after oral administration. Toxicol Lett 2007, 168: 176–185. 10.1016/j.toxlet.2006.12.001

    Article  CAS  PubMed  Google Scholar 

  25. Fabian E, Landsiedel R, Ma-Hock L, Wiench K, Wohlleben W, Van RB: Tissue distribution and toxicity of intravenously administered titanium dioxide nanoparticles in rats. Arch Toxicol 2008, 82: 151–157. 10.1007/s00204-007-0253-y

    Article  CAS  PubMed  Google Scholar 

  26. Park EJ, Yi J, Chung KH, Ryu DY, Choi J, Park K: Oxidative stress and apoptosis induced by titanium dioxide nanoparticles in cultured BEAS-2B cells. Toxicol Lett 2008, 180: 222–229. 10.1016/j.toxlet.2008.06.869

    Article  CAS  PubMed  Google Scholar 

  27. Churg A, Gilks B, Dai J: Induction of fibrogenic mediators by fine and ultrafine titanium dioxide in rat tracheal explants. Am J Physiol 1999, 277: L975-L982.

    CAS  PubMed  Google Scholar 

  28. Gurr JR, Wang AS, Chen CH, Jan KY: Ultrafine titanium dioxide particles in the absence of photoactivation can induce oxidative damage to human bronchial epithelial cells. Toxicology 2005, 213: 66–73. 10.1016/j.tox.2005.05.007

    Article  CAS  PubMed  Google Scholar 

  29. Simon-Deckers A, Gouget B, Mayne-L'hermite M, Herlin-Boime N, Reynaud C, Carriere M: In vitro investigation of oxide nanoparticle and carbon nanotube toxicity and intracellular accumulation in A549 human pneumocytes. Toxicology 2008, 253: 137–146. 10.1016/j.tox.2008.09.007

    Article  CAS  PubMed  Google Scholar 

  30. Kim JK, Lee WK, Lee EJ, Cho YJ, Lee KH, Kim HS, Chung Y, Kim KA, Lim Y: Mechanism of silica- and titanium dioxide-induced cytotoxicity in alveolar macrophages. J Toxicol Environl Health Part A 1999, 58: 437–450. 10.1080/009841099157160

    Article  CAS  Google Scholar 

  31. Barlow PG, Clouter-Baker A, Donaldson K, Maccallum J, Stone V: Carbon black nanoparticles induce type II epithelial cells to release chemotaxins for alveolar macrophages. Part Fibre Toxicol 2005, 2: 11–25. 10.1186/1743-8977-2-11

    Article  PubMed Central  PubMed  Google Scholar 

  32. Gamer AO, Leibold E, Van RB: The in vitro absorption of microfine zinc oxide and titanium dioxide through porcine skin. Toxicol In Vitro 2006, 20: 301–307. 10.1016/j.tiv.2005.08.008

    Article  CAS  PubMed  Google Scholar 

  33. Pflucker F, Hohenberg H, Holzle E, Will T, Pfeiffer S, Wepf R, Diembeck W, Wenck H, Gers-Barlag H: The outermost stratum corneum layer is an effective barrier against dermal uptake of topically applied micronized titanium dioxide. Int J Cosmet Sci 1999, 21: 399–411. 10.1023/A:1005450606035

    Article  CAS  PubMed  Google Scholar 

  34. Dussert AS, Gooris E, Hemmerle J: Characterization of the mineral content of a physical sunscreen emulsion and its distribution onto human stratum corneum. Int J Cos Sci 1997, 19: 119–129. 10.1111/j.1467-2494.1997.tb00175.x

    Article  CAS  Google Scholar 

  35. Jin CY, Zhu BS, Wang XF, Lu QH: Cytotoxicity of titanium dioxide nanoparticles in mouse fibroblast cells. Chemical Research in Toxicology 2008, 21: 1871–1877. 10.1021/tx800179f

    Article  CAS  PubMed  Google Scholar 

  36. Zhang AP, Sun YP: Photocatalytic killing effect of TiO 2 nanoparticles on Ls-174-t human colon carcinoma cells. World J Gastroenterol 2004, 10: 3191–3193.

    CAS  PubMed  Google Scholar 

  37. Hussain SM, Hess KL, Gearhart JM, Geiss KT, Schlager JJ: In vitro toxicity of nanoparticles in BRL 3A rat liver cells. Toxicol In Vitro 2005, 19: 975–983. 10.1016/j.tiv.2005.06.034

    Article  CAS  PubMed  Google Scholar 

  38. Linnainmaa K, Kivipensas P, Vainio H: Toxicity and cytogenetic studies of ultrafine titanium dioxide in cultured rat liver epithelial cells. Toxicol In Vitro 1997, 11: 329–335. 10.1016/S0887-2333(97)00000-3

    Article  CAS  PubMed  Google Scholar 

  39. Peters K, Unger RE, Kirkpatrick CJ, Gatti AM, Monari E: Effects of nano-scaled particles on endothelial cell function in vitro : studies on viability, proliferation and inflammation. J Mat Sci Mat Med 2004, 15: 321–325. 10.1023/B:JMSM.0000021095.36878.1b

    Article  CAS  Google Scholar 

  40. Courtois A, Andujar P, Ladeiro Y, Baudrimont I, Delannoy E, Leblais V, Begueret H, Galland MA, Brochard P, Marano F, Marthan R, Muller B: Impairment of NO-dependent relaxation in intralobar pulmonary arteries: comparison of urban particulate matter and manufactured nanoparticles. Environ Health Perspect 2008, 116: 1294–1299.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  41. Helfenstein M, Miragoli M, Rohr S, Muller L, Wick P, Mohr M, Gehr P, Rothen-Rutishauser B: Effects of combustion-derived ultrafine particles and manufactured nanoparticles on heart cells in vitro . Toxicology 2008, 253: 70–78. 10.1016/j.tox.2008.08.018

    Article  CAS  PubMed  Google Scholar 

  42. Long TC, Saleh N, Tilton RD, Lowry GV, Veronesi B: Titanium dioxide (P25) produces reactive oxygen species in immortalized brain microglia (BV2): implications for nanoparticle neurotoxicity. Environ Sci Technol 2006, 40: 4346–4352. 10.1021/es060589n

    Article  CAS  PubMed  Google Scholar 

  43. Long TC, Tajuba J, Sama P, Saleh N, Swartz C, Parker J, Hester S, Lowry GV, Veronesi B: Nanosize titanium dioxide stimulates reactive oxygen species in brain microglia and damages neurons in vitro . Environ Health Perspect 2007, 115: 1631–1637.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  44. L'azou B, Jorly J, On D, Sellier E, Moisan F, Fleury-Feith J, Cambar J, Brochard P, Ohayon-Court C: In vitro effects of nanoparticles on renal cells. Part Fibre Toxicol 2008, 5: 22–36. 10.1186/1743-8977-5-22

    Article  PubMed Central  PubMed  Google Scholar 

  45. Renwick LC, Donaldson K, Clouter A: Impairment of alveolar macrophage phagocytosis by ultrafine particles. TAAP 2001, 172: 119–127.

    CAS  Google Scholar 

  46. Afaq F, Abidi P, Matin R, Rahman Q: Cytotoxicity, pro-oxidant effects and antioxidant depletion in rat lung alveolar macrophages exposed to ultrafine titanium dioxide. J App Toxicol 1998, 18: 307–312. Publisher Full Text 10.1002/(SICI)1099-1263(1998090)18:5<307::AID-JAT508>3.0.CO;2-K

    Article  CAS  Google Scholar 

  47. Kang SJ, Kim BM, Lee YJ, Chung HW: Titanium dioxide nanoparticles trigger p53-mediated damage response in peripheral blood lymphocytes. Environ Mol Mutagen 2008, 49: 399–405. 10.1002/em.20399

    Article  CAS  PubMed  Google Scholar 

  48. Nel A, Xia T, Madler L, Li N: Toxic potential of materials at the nanolevel. Science 2006, 311: 622–627. 10.1126/science.1114397

    Article  CAS  PubMed  Google Scholar 

  49. Kang JL, Moon C, Lee HS, Lee HW, Park EM, Kim HS, Castranova V: Comparison of the biological activity between ultrafine and fine titanium dioxide particles in RAW 264.7 cells associated with oxidative stress. J Toxicol Environ Health Part A 2008, 71: 478–485. 10.1080/15287390801906675

    Article  CAS  PubMed  Google Scholar 

  50. Xia T, Kovochich M, Liong M, Madler L, Gilbert B, Shi H, Yeh JI, Zink JI, Nel AE: Comparison of the mechanism of toxicity of zinc oxide and cerium oxide nanoparticles based on dissolution and oxidative stress properties. ACS Nano 2008, 2: 2121–2134. 10.1021/nn800511k

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  51. Lu N, Zhu Z, Zhao X, Tao R, Yang X, Gao Z: Nano titanium dioxide photocatalytic protein tyrosine nitration: a potential hazard of TiO 2 on skin. Biochem Biophys Res Commun 2008, 370: 675–680. 10.1016/j.bbrc.2008.04.010

    Article  CAS  PubMed  Google Scholar 

  52. Stearns RC, Paulauskis JD, Godleski JJ: Endocytosis of ultrafine particles by A549 cells. Am J Respir Cell Mol Biol 2001, 24: 108–115.

    Article  CAS  PubMed  Google Scholar 

  53. Geiser M, Casaulta M, Kupferschmid B, Schulz H, Semmler-Behnke M, Kreyling W: The role of macrophages in the clearance of inhaled ultrafine titanium dioxide particles. Am J Respiy Cell Molr Biol 2008, 38: 371–376. 10.1165/rcmb.2007-0138OC

    Article  CAS  Google Scholar 

  54. Patel LN, Zaro JL, Shen W-C: Cell Penetrating peptides: intracellular pathways. Pharm Res 2007, 24: 1977–1992. 10.1007/s11095-007-9303-7

    Article  CAS  PubMed  Google Scholar 

  55. Rothen-Rutishauser B, Muhlfeld C, Blank F, Musso C, Gehr P: Translocation of particles and inflammatory responses after exposure to fine particles and nanoparticles in an epithelial airway model. Part Fibre Toxicol 2007, 4: 9–18. 10.1186/1743-8977-4-9

    Article  PubMed Central  PubMed  Google Scholar 

  56. Rahman Q, Lohani M, Dopp E, Pemsel H, Jonas L, Weiss DG, Schiffmann D: Evidence that ultrafine titanium dioxide induces micronuclei and apoptosis in Syrian hamster embryo fibroblasts. Environ Health Perspect 2002, 110: 797–800.

    Article  CAS  PubMed Central  PubMed  Google Scholar 

  57. Karlsson HL, Cronholm P, Gustafsson J, Moller L: Copper oxide nanoparticles are highly toxic: a comparison between metal oxide nanoparticles and carbon nanotubes. Chem Res Toxicol 2008, 21: 1726–1732. 10.1021/tx800064j

    Article  CAS  PubMed  Google Scholar 

  58. Dunford R, Salinaro A, Cai L, Serpone N, Horikoshi S, Hidaka H, Knowland J: Chemical oxidation and DNA damage catalysed by inorganic sunscreen ingredients. FEBS Lett 1997, 418: 87–90. 10.1016/S0014-5793(97)01356-2

    Article  CAS  PubMed  Google Scholar 

  59. Nakagawa Y, Wakuri S, Sakamoto K, Tanaka N: The photogenotoxicity of titanium dioxide particles. Mutat Res 1997, 394: 125–132.

    Article  CAS  PubMed  Google Scholar 

  60. Theogaraj E, Riley S, Hughes L, Maier M, Kirkland D: An investigation of the photo-clastogenic potential of ultrafine titanium dioxide particles. Mutat Res 2007, 634: 205–219.

    Article  CAS  PubMed  Google Scholar 

  61. Driscoll KE, Deyo LC, Carter JM, Howard BW, Hassenbein DG, Bertram TA: Effects of particle exposure and particle-elicited inflammatory cells on mutation in rat alveolar epithelial cells. Carcinogenesis 1997, 18: 423–430. 10.1093/carcin/18.2.423

    Article  CAS  PubMed  Google Scholar 

  62. Komatsu T, Tabata M, Kubo-Irie M, Shimizu T, Suzuki KI, Nihei Y, Takeda K: The effects of nanoparticles on mouse testis Leydig cells in vitro . Toxicology in Vitro 2008, 22: 1825–1831. 10.1016/j.tiv.2008.08.009

    Article  CAS  PubMed  Google Scholar 

  63. Oberdorster G, Stone V, Donaldson K: Toxicology of nanoparticles: a historical perspective. Nanotoxicology 2007, 1: 2–25. 10.1080/17435390701314761

    Article  CAS  Google Scholar 

  64. Dick CA, Brown DM, Donaldson K, Stone V: The role of free radicals in the toxic and inflammatory effects of four different ultrafine particle types. Inhal Toxicol 2003, 15: 39–52. 10.1080/08958370304454

    Article  CAS  PubMed  Google Scholar 

  65. Duffin R, Tran CL, Clouter A, Brown DM, MacNee W, Stone V, Donaldson K: The importance of surface area and specific reactivity in the acute pulmonary inflammatory response to particles. Ann Occup Hyg 2002, 46: 242–245.

    Article  Google Scholar 

  66. Stoeger T, Reinhard C, Takenaka S, Schroeppel A, Karg E, Ritter B, Heyder J, Schulz H: Instillation of six different ultrafine carbon particles indicates a surface area threshold dose for acute lung inflammation in mice. Environ Health Perspect 2006, 114: 328–333.

    Article  PubMed Central  PubMed  Google Scholar 

  67. Brown DM, Wilson MR, MacNee W, Stone V, Donaldson K: Size-dependent proinflammatory effects of ultrafine polystyrene particles: a role for surface area and oxidative stress in the enhanced activity of ultrafines. TAAP 2001, 175: 191–199.

    CAS  Google Scholar 

  68. Sager TM, Kommineni C, Castranova V: Pulmonary response to intratracheal instillation of ultrafine versus fine titanium dioxide: role of particle surface area. Part Fibre Toxicol 2008, 5: 17–32. 10.1186/1743-8977-5-17

    Article  PubMed Central  PubMed  Google Scholar 

  69. Sager TM, Castranova V: Surface area of particle administered versus mass in determining the pulmonary toxicity of ultrafine and fine carbon black: comparison to ultrafine titanium dioxide. Part Fibre Toxicol 2009, 6: 15–46. 10.1186/1743-8977-6-15

    Article  PubMed Central  PubMed  Google Scholar 

  70. Rehn B, Seiler F, Rehn S, Bruch J, Maier M: Investigations on the inflammatory and genotoxic lung effects of two types of titanium dioxide: untreated and surface treated. TAAP 2003, 189: 84–95.

    CAS  Google Scholar 

  71. Sayes CM, Wahi R, Kurian PA, Liu Y, West JL, Ausman KD, Warheit DB, Colvin VL: Correlating nanoscale titania structure with toxicity: a cytotoxicity and inflammatory response study with human dermal fibroblasts and human lung epithelial cells. Toxicol Sci 2006, 92: 174–185. 10.1093/toxsci/kfj197

    Article  CAS  PubMed  Google Scholar 

  72. Pan Z, Lee W, Slutsky L, Clark RA, Pernodet N, Rafailovich MH: Adverse effects of titanium dioxide nanoparticles on human dermal fibroblasts and how to protect cells. Small 2009, 5: 511–520. 10.1002/smll.200800798

    Article  CAS  PubMed  Google Scholar 

  73. Oberdorster G: Pulmonary effects of inhaled ultrafine particles. Int Arch Occup Environ Health 2001, 74: 1–8. 10.1007/s004200000185

    Article  CAS  PubMed  Google Scholar 

  74. Hohr D, Steinfartz Y, Schins RP, Knaapen AM, Martra G, Fubini B, Borm PJ: The surface area rather than the surface coating determines the acute inflammatory response after instillation of fine and ultrafine TiO 2 in the rat. Int J Hyg EnvironHealth 2002, 205: 239–244. 10.1078/1438-4639-00123

    Article  Google Scholar 

  75. Singh S, Sh T, Duffin R, Albrecht C, van BD, Hohr D, Fubini B, Martra G, Fenoglio I, Borm PJ, Schins RP: Endocytosis, oxidative stress and IL-8 expression in human lung epithelial cells upon treatment with fine and ultrafine TiO 2 : role of the specific surface area and of surface methylation of the particles. TAAP 2007, 222: 141–151.

    CAS  Google Scholar 

  76. Thevenot P, Cho J, Wavhal D, Timmons RB, Tang L: Surface chemistry influences cancer killing effect of TiO 2 nanoparticles. Nanomed 2008, 4: 226–236.

    Article  CAS  Google Scholar 

  77. Shi L, Chen J, Feng Z, Chen T, Lian Y, Wang X, Li C: Photoluminescence characteristics of TiO 2 and their relationship to the photoassisted reaction of water/methanol mixture. J Physi Chem 2007, 111: 693–699.

    CAS  Google Scholar 

  78. Kaewamatawong T, Shimada A, Okajima M, Inoue H, Morita T, Inoue K, Takano H: Acute and subacute pulmonary toxicity of low dose of ultrafine colloidal silica particles in mice after intratracheal instillation. Toxicol Pathol 2006, 34: 958–965. 10.1080/01926230601094552

    Article  CAS  PubMed  Google Scholar 

Download references


This work was funded as part of the European Commission FP7 funded project ENRHES. The authors would like to thank all participants of the ENRHES project for useful discussions.

Author information

Authors and Affiliations


Corresponding author

Correspondence to Helinor J Johnston.

Additional information

Competing interests

The authors declare that they have no competing interests.

Authors' contributions

HJ and VS contributed to all sections. GH contributed to the reproductive toxicity section. FC aided in the editing of the paper. SP contributed to the literature search. SH was involved in the scientific management of the project. All authors read and approved the final manuscript.

Rights and permissions

This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Reprints and permissions

About this article

Cite this article

Johnston, H.J., Hutchison, G.R., Christensen, F.M. et al. Identification of the mechanisms that drive the toxicity of TiO2 particulates: the contribution of physicochemical characteristics. Part Fibre Toxicol 6, 33 (2009).

Download citation

  • Received:

  • Accepted:

  • Published:

  • DOI: