Abstract

The essence of nanotechnology is the production of nanoparticles (NPs) with unique physicochemical properties allowing worldwide application in new structures, materials, and devices. The consequently increasing human exposure to NPs has raised concerns regarding their health and safety profiles. Titanium dioxide (TiO2) has been reported to induce adverse pulmonary responses in exposed animals. However, the potential more dangerous biological activities of TiO2 NPs compared to their fine-sized counterparts are not fully understood. Therefore, this work is aimed to provide a comprehensive evaluation of the toxic effects induced by TiO2 NPs in in vivo experiments. It is intended to deeply understand the toxicological behaviour of TiO2 NPs and to predict potential human health effects. Moreover, it may be an instrument to extrapolate relevant data for human risk evaluation and management and to identify those critical aspects that deserve great attention in future population and epidemiologic research.

1. Introduction

Nanotechnology is the manipulation of matter on a near-atomic scale to produce new structures, materials, and devises. It has became an important industry in the twenty-first century, and the U.S. National Science Foundation estimated it will grow into a trillion-dollar business, employing millions of workers worldwide, within the next decade [13].

The essence of nanotechnology is the synthesis of engineered nanoparticles (NPs) that exhibit characteristics, such as small size, large surface area to mass ratio, shape, crystallinity, surface charge, reactive surface groups, dissolution rate, state of agglomeration, or dispersal that confer them properties substantially different from those of the bulk particles of the same composition [46]. These properties offer great opportunities for the development of new NP industrial applications increasing their worldwide distribution and enhancing the likelihood of environmental and human exposure [5, 7].

Considerable work has been carried out to advance nanotechnology and its applications. Nevertheless, our understanding of the general and occupational health and safety aspects of NPs is still in its formative phase and greater effort is needed to understand how NPs interact with the human body [8, 9]. In this regard, nanotoxicology and nano-risk have been attracting the increasing attention of toxicologists and regulatory scientists [10] particularly in relation to the unique properties of NPs that may render them potentially more dangerous than their fine-sized counterpart and may cause unexpected adverse health effects to exposed people [6, 11].

Titanium dioxide (TiO2) is an example of a fine, white, crystalline, odorless, low-solubility powder which was considered to exhibit relatively low toxicity [1216]. It is a natural, thermally stable and nonflammable, nonsilicate mineral oxide found primarily in the form of the minerals rutile, anatase, brookite, and as the iron-containing mineral ilmenite [1721]. It has excellent physicochemical properties, such as good fatigue strength, resistance to corrosion, machineability, biocompatibility, whitening and photocatalysis, as well as excellent optical performance and electrical properties [22, 23]. With regard to its potential adverse health effects, several studies have defined TiO2, at least under nonoverload conditions, as biologically inactive and physiologically inert in both humans and animals and thus as little risk to humans [2427]. Pulmonary inflammation, fibrosis, epithelial hyperplasia, and tumorigenesis were reported in animals under conditions of substantial TiO2 particle lung burden due to sufficiently high dose and/or duration of exposure [13, 15, 28, 29]. In 2006, the International Agency for Research on Cancer (IARC), classified and in 2010 reassessed TiO2 as “possibly carcinogenic to humans” on the basis of the sufficient evidence of carcinogenicity in experimental animals and inadequate evidence in humans (Group 2B) [28, 29].

TiO2 is a versatile compound that has broadly been used in nanoparticulate form [21]. According to the National Nanotechnology Initiative of America, nanosized TiO2 particles are among those most widely manufactured on a global scale [30]. TiO2 NPs are widely used in paints, printing ink, rubber, paper, cosmetics, sunscreens, car materials, cleaning air products, industrial photocatalytic processes, and decomposing organic matters in wastewater due to their unique physical, chemical, and biological properties (including the inherent advantages of physical stability, anticorrosion and nanoscale-enhanced photocatalysis) [31]. However, the toxicological profile of TiO2 NPs is not completely understood and several concerns have emerged on the potential undesirable effects of the TiO2 NP properties, in regard to the harmful interactions with biological systems and the environment [11]. The recently recognized occupational carcinogenic potential of the inhaled TiO2 NPs have ulteriourly enhanced these scientific concerns [15]. Therefore, an appropriate assessment of the risks for the general and occupational exposed population requiring TiO2 NP hazard identification and dose-response data seems necessary [32].

In this regard, our previous work [33] reviewed several toxic effects of NPs assessed by in vitro experiments. These studies represent a valid instrument to investigate TiO2 NP, induced cellular changes at biochemical and molecular levels and to determine their underlying mechanistic processes. Moreover, providing interesting information about doses of exposure and endpoint parameters to evaluate, this kind of investigation constitutes a significant step in planning animal research. Unfortunately, several limitations inherent to in vitro assay/cell culture systems have appeared in simulating the complex biological effects of particles administered to experimental animals. These include, but are not limited to, unrealistic particle dose, selection of cell types for simulating the different system microenvironments (single-cell culture systems or coculture systems), particle/cell exposure interactions in culture versus biological fluids, time course of effects represented by 1-, 4-, 24-, or 48-hour incubation in vitro versus acute or chronic exposure in vivo, and appropriate end points for hazard evaluation [34]. These biases do not allow to obtain data that may be representative of human exposure/effects and applicable for a correct population risk assessment.

Therefore, in the present review, we will provide a comprehensive evaluation of the current knowledge regarding the toxic effects induced by TiO2 NPs on organ systems investigated in in vivo studies. This overview is intended to be a useful tool to gain a thorough insight into the toxicological profile of TiO2 NPs and to predict potential human health effects. Moreover, it may reveal itself as a valid instrument to extrapolate relevant data for human risk evaluation and management and to identify those critical aspects that deserve great attention in future population and epidemiologic research.

2. In Vivo Studies

2.1. Respiratory System

Considering the relevance of the respiratory system as a route for TiO2 NP exposure in humans, the scientific community has focused on the toxicological effects of these NPs on animal respiratory models. Such experiments provide the basis to obtain more detail data regarding the hazard of these NPs and to extrapolate evaluations for a correct human risk assessment, in particular, in relation to the characteristics of the particles. Indeed, several animal pulmonary studies were carried out to clarify the role played by TiO2 form, particle size, shape, surface area, and chemistry in determining inflammatory responses, particle lung distribution, and carcinogenic effects (Table 1).

2.1.1. Inflammatory Responses

TiO2 Form
Regarding the TiO2 form, different studies have demonstrated the pulmonary toxicity of acute exposure to anatase in terms of increased bronchoalveolar lavage (BAL) inflammatory parameters [3538], lung tissue structural damage, and inflammatory infiltration [22, 35, 36, 39, 40] although it is modest in both acute and subacute exposure [41]. However, as assessed by Liu et al. [36], only a slight increase in toxicity from anatase exposure was evident relative to rutile and P25 Degussa TiO2 NP treatments.
Only four studies have evaluated the ability of rutile TiO2 to induce toxic effects on lung tissues, reporting conflicting results following acute [4245] or subchronic exposure [46]. While the first two studies demonstrated an increase of inflammatory cells in BAL [43], alterations in gene expression [42], and pathological changes in lung tissue [42, 43], the latter two failed to reveal such changes [4446].
When comparing the lung toxicity induced by rutile NPs with that caused by P25 Degussa NPs, the former were not able to induce alterations of the lung parameters examined suggesting greater toxic effects of the mixture compared to rutile alone [44, 45]. Regarding the toxic effects of P25 Degussa NPs on the lung, other studies have reported that these particles are able to induce increased inflammatory parameters in the BAL of acutely exposed animals [47, 48], though some interspecies differences in inflammatory responses were evident in animals subchronically treated [49]. The species differences detected between hamsters, rats, and mice may reflect the capacity of some animals to rapidly clear particles from the lung, as previously demonstrated in the same species subchronically exposed to pigmentary fine TiO2 particles [50, 51]. Interestingly, Gustafsson et al. [48] demonstrated a dynamic inflammatory response to P25 Degussa NPs characterized by a transient innate immune activation followed by a late-phase, long-lasting recruitment of lymphocytes involved in adaptive immunity.
Other mixtures of anatase and rutile particles were also able to induce increases in BAL fluid inflammatory parameters and lung histopathological alterations in rats subacutely exposed by inhalation [66, 67], while mixtures of anatase and brookite (3 : 1) failed to induce such effects in mice after acute, subacute, or subchronic inhalations [70, 76].
Other studies report similar changes after acute or subacute exposure, such as increased levels of inflammatory mediators in BAL [58, 59, 62, 73], enhanced reactive oxygen species (ROS) production [59] and pathological alterations at the histopathology examination of the lung [57, 73, 75] but lack details regarding the TiO2 form. Interestingly, Baggs et al. [57] demonstrated that chronic exposure to TiO2 NPs induces lesions that are able to regress during a 1-year period following cessation of exposure. Surprisingly, Cho et al. [32], investigating the acute effects of intratracheal instillations of TiO2 NPs, failed to demonstrate any alteration among BAL inflammatory parameters and histological lung characteristics. Finally, Liu et al. [74] demonstrated that intratracheal instillation of not characterized TiO2 NPs damaged alveolar macrophage (AM) phagocytic and chemotactic ability.

Size
Several studies described greater inflammatory effects of NPs compared to their fine counterparts both after acute [38, 53, 54, 62] and chronic exposure [53, 55].
In line with these results on NP size-related effects, a more recent intratracheal instillation study [38] pointed out that the smaller particles lead to greater inflammation in short-term observations. In contrast, no clear relationship was found between pulmonary inflammation and treatments with different agglomeration states of the same primary TiO2 NPs.

Shape
The critical role of shape in TiO2 NP bioactivity is supported by the increased markers of inflammation detected in BAL of anatase nanobelt aspiration-treated mice compared with those determined in animals treated with TiO2 nanospheres [37].

Surface Area
High mass or volume dose of poorly soluble, low-toxicity (PSLT) fine particles in the lungs has been associated with overloading, while ultrafine particles impair lung clearance at lower mass or volume doses [15]. The increased lung retention and inflammatory response of nanosized PSLT particles compared to fine PSLT particles correlate better to the particle surface area dose [81, 82]. Indeed, the larger inflammatory response after TiO2 NP treatment compared to larger particles may be ascribed to an increase in NP surface area. This conclusion has been questioned, and conflicting results are present in literature on this topic. Some studies support the hypothesis that surface area may be the more appropriate dose metric for TiO2 NP pulmonary toxicity studies [12, 60, 61]. When the same mass doses were acutely introduced into rats and mice via intratracheal instillation, TiO2 NPs induced a much greater pulmonary-inflammatory response than fine TiO2 particles [12, 60]. However, when the doses were normalized to the particle-administered surface area, the response in the lung for both nanosized and fine TiO2 particles showed the same dose-response curve. Thus, in pulmonary toxicity studies, surface area, for particles of different sizes but of the same chemistry, proved to be a better dosemetric parameter than particle mass or number.
However, other studies are in conflict with the notion that the inflammatory response is expected to be more severe with higher surface area NPs [39, 44, 45]. In fact, similar BAL cell count alterations were reported in rats after intratracheal administrations of the same dose of fine rutile TiO2 particles, nanosized TiO2 anatase rods and dots, although the latter had a >6-fold increase in surface area compared with the nanorods [39]. In a subsequent study, the same authors [44, 45] confirmed the previous evaluations and concluded that having a larger surface area does not necessarily indicate that NPs will produce greater pulmonary inflammation and cytotoxicity compared to larger particles of a similar composition. In line with the Warheit et al. [39, 44, 45] results, 2- to 5-nm anatase NPs (i.e., with the highest surface area and smallest particle size) were not particularly toxic in the Grassian et al. [41] subacute inhalation study. Furthermore, in Li et al. [35], 3-nm anatase NPs did not elicit more pulmonary toxicity compared to 20-nm NPs, irrespective of their smaller size and greater surface area.
However, an aspect to take in consideration for a correct interpretation of these data is the potential influence of “lung overload” [13]. In the last decade, it has become clear that a breakdown in normal AM-mediated clearance of particles is seen in overload [83]. This is thought to be a consequence of volumetric overload of the AM [84, 85] or a response to the greater particle surface area per mass dose, as in the case of NPs, which is associated with decreased AM clearance and inflammation [13, 86]. In this regard, the importance of the NP surface area overlaps with the role of the dose administered to rats. Most of the studies herein described showed greater TiO2 NP effects at higher exposure doses [35, 36, 39, 40, 4345, 48, 49, 61, 62, 65, 67, 79]. Unfortunately, the large dose range applied in these studies does not allow one to establish a direct relationship between the different treatments and pulmonary effects. Moreover, when a study employes unrealistically high doses of NPs, such as in the case of Chen et al. [22], its relevance is dubious at best. In this light, when analyzing the NP dose role in lung damage, caution should be applied in extrapolating data for human evaluations.

Chemistry
Warheit et al. [44, 45] found that only non-coated particles resulted in BAL alteration and lung histopathological changes compared to their coated counterparts, suggesting a possible influencing role of surface chemistry in nano-TiO2-induced lung toxicity. In this regard, several studies have investigated the influence of particle chemical surface properties, surface coatings, and functionalization on pulmonary responses showing conflicting results [61, 70, 77, 78].
Methylated-hydrophobic TiO2 NPs acutely decreased the percentage of PMN and the protein concentration in BAL of intratracheally administered rats [61], a result in line with those of Oberdörster [87], but in contrast with those of Pott et al. [88] showing that hydrophobic, silanized TiO2 NPs induced acute mortality in intratracheally exposed rats. Other studies reported significant inflammatory reactions after acute [70, 79] and subacute [70, 77, 78] exposure to coated NPs, particularly silica-coated rutile TiO2 NPs [70], rutile TiO2 NPs surface modified with unspecified amount of zirconium, silicon, aluminum, and coated with polyalcohols [77, 78] and rutile Fe-doped TiO2 NPs [79]. However, at present, it is unclear what changes lead to differences in the toxicity induced by surface modified TiO2 NPs. In additional experiments, P25 Degussa TiO2 NPs caused more persistent pulmonary toxicity than carbon black NPs when compared on an equivalent particle surface area basis [68].
On the other hand, several studies investigated the TiO2 NP role as an adjuvant in promoting allergic sensitization or in influencing allergic lung inflammation. In this regard, the results obtained by Larsen et al. [65] and Park et al. [69] supported the adjuvant effect of not characterized and P25 Degussa TiO2 NPs, respectively, through the promotion of a T-helper (Th)-2 immune response assessed by increased levels of interleukin (IL)-4, IL-5, and IL-10 in BAL. These alterations could stimulate allergic reactions, though the underlying mechanisms are not fully understood. Interestingly, in two-day- and two-week-old rats, but not in adult animals, subacute inhalation of P25 Degussa TiO2 NPs produced upregulation of lung neurotrophins which was associated with a greater airway hyperresponsiveness [72]. The age-dependent response to TiO2 NPs suggests that the risk of developing asthma after NP exposure is higher in the earlier stages of lung development. Hussain et al. [80] demonstrated that intrapulmonary doses of TiO2 NPs can aggravate pulmonary inflammation and airway hyperreactivity in a mouse model of diisocyanate-induced asthma. Moreover, mixture of anatase and brookite TiO2 NPs caused airflow limitation in both acute and repeated 4-week-inhalation exposures [76]. In contrasts to these results [65, 69, 72, 76, 80], a recent study [71] demonstrated the role of silica-coated rutile TiO2 NPs as inhibitors of most soluble and cellular mediators of allergic asthma. In fact, in ovalbumin-sensitized mice, the number of inflammatory cells, and the airway hyperreactivity were dramatically reduced after exposure to TiO2 NPs. The conflicting results of the previous studies could be ascribed to the different health status of the animals, the different type of particle employed, or the different route of exposure. Rossi et al. [71] hypothesized that anti-inflammatory Th-2 response caused by allergen sensitization may be suppressed by the competing proinflammatory response elicited by TiO2 exposure. However, future studies are needed to deeply clarify the TiO2 NP role in allergic reactions and the molecular mechanisms involved.

2.1.2. Lung Distribution

Particle size was able to affect the fate of NPs, particularly lung accumulation and distribution in lung compartments, after pulmonary exposure. In fact, several studies described greater NP lung amounts compared to their fine counterparts [12, 5255, 60]. Ferin et al. [53] suggested that TiO2 particle passage in the interstitium could be promoted by the smaller sizes of the particles and the higher concentrations. Though data regarding the accumulation of NPs in the interstitium are too limited to allow a comprehensive evaluation of its consequences, this aspect is of particular interest considering that particle retention in the lung may give rise to potential enhancement of local effects and the increased possibility of systemic redistribution.

Translocation of NPs from pulmonary airways into other pulmonary compartments, such as epithelium/endothelium, connective tissue, the capillary lumen, or into systemic circulation, is the subject of controversial discussion in the literature. Two studies of the same group demonstrated conflicting results in that regard [63, 64]. The first showed that NP distribution was only related to the volume fractions of the corresponding compartments [63], while the second detected a correlation with the time points considered [64]. In fact, it found that NPs were preferentially located in the connective tissue and in the capillary lumen at 1 and 24 h after exposure, respectively. The epithelium and the connective tissues were not deposition sites but only passageways for NPs to reach the capillary circulation. However, whether NPs can translocate from air-filled spaces to the systemic circulation is still not fully understood, and further investigation is needed.

2.1.3. Carcinogenic Effects

Animal and human epidemiological data have led TiO2 to be designated as “possibly carcinogenic” to humans [28, 29, 89]. Surprisingly, however, few studies have investigated the carcinogenicity of TiO2 NPs, and only recently, NIOSH concluded that inhaled ultrafine TiO2 is a potential occupational carcinogen [15]. The most relevant data for assessing the health risk for workers are results from a chronic animal inhalation study performed by Heinrich et al. [56], in which rats exposed by inhalation to P25 Degussa TiO2 NPs showed increased rates of adenocarcinomas compared to controls. Interestingly, in this study, mice exposed to the same NPs, according to the same methodology, did not show differences in tumor rates compared to controls. Aside from demonstrating that TiO2 NPs can induce lung cancer in exposed animals, an interesting finding of this work is the difference in carcinogenicity between rats and mice. The species difference in response to insoluble and low toxicity dust, and the controversial approach to classify such a dust as a potential human carcinogen is subject to debate. Moreover, NIOSH has concluded that TiO2 is not a direct-acting carcinogen, but acts through a secondary genotoxic mechanism primarily related to particle size and surface area [15], as supported by the comprehensive analysis of the data reported by Heinrich et al. [56] and those obtained by Lee et al. with fine-sized TiO2 [90]. However, considering the pressing concern regarding cancerogenic effects of TiO2 NP exposure on general and occupational populations and the limited number of in vivo studies on this topic, additional researches seem highly necessary.

2.2. Nervous System

Regarding TiO2 NP neurotoxicity, different studies have demonstrated the ability of these NPs to translocate into the brain, irrespective of the route of exposure [31, 73, 9194], the form of the TiO2, and the size of the NPs and to accumulate in this organ [31, 9193] inducing numerical and structural changes in the neuronal architecture [31, 92, 93] (Table 2).

Several studies have unequivocally showed that rutile TiO2 NPs, taken up by intranasal exposure, could be translocated into the central nervous system (CNS) via the olfactory pathway in mice and accumulate in the entire brain, mainly in the hippocampus regions [31, 91, 92]. However, to verify the generalization of this phenomenon, further studies are needed in other animal species. Regardless, this neuronal translocation pathway should be taken into account for health risk assessments of TiO2 NPs. In particular, it could be useful to avoid possible toxic effects to the general population but principally to subjects that work with these NPs. Brain accumulation was also demonstrated in mice intratracheally instilled with no characterized TiO2 NPs able to penetrate alveolar-capillary barrier, enter blood circulation, and further penetrate blood-brain barrier inducing tissue damage [73].

The deposition of TiO2 NPs in brain was reported to induce changes in the release and metabolism of neurotransmitters, although with the same conflicting results, maybe due to the different route of exposure [31, 91, 93, 94]. Particularly, the levels of norepinephrine (NE) and 5-hydroxytryptamine (5-HT) increased after intranasal exposure [91] while decreased in response to intragastric administrations of anatase TiO2 NPs [93]. Dopamine (DA), 3,4-dihydroxyphenylacetic acid (DOPAC), homovanillic (HVA), and 5-hydroxyindole acetic acid (5-HIAA) contents were reduced by both types of treatment [91, 94]. Moreover, intranasal instillation of rutile [31] and intragastric administrations of anatase TiO2 NPs [94] increased catalase and acetylcholinesterase (AchE) activity, soluble protein carbonyl, acetylcholine (Ach), glutamic acid (Glu), and NO content. Intraperitoneally injected anatase TiO2 NPs increased NO, while decreased Glu content and AchE activity [92]. Interestingly, the study performed by Hu et al. [94] demonstrated also that the contents of Ca, Mg, Na, K, Fe, and Zn in brain were significantly altered after TiO2 NP exposure. The disturbed homeostasis of trace elements, neurotransmitters, and enzymes in brain may be responsible of the spatial recognition memory impairments reported in treated mice [94]. However, as assessed by a proteomic analysis, proteins of the mouse brain resulted differentially expressed in response to TiO2 NP treatment even if NPs were not detected in the tissue [95].

As observed in vitro [9698], oxidative stress-related damage with a consequent alteration in the balance between oxidative and antioxidative activities was also reported in in vivo studies [31, 73, 92, 93, 95]. Malondiadehyde (MDA) levels increased after intranasal instillations [31, 92], intraabdominal injections [93], and intratracheal instillation in mice [73]. Similarly, Li et al. [73], and Ma et al. [93] demonstrated that the level of superoxide anion ( O 2 ), peroxide (H2O2) [73, 93] and hydroxyl radicals [73] increased in treated animals. TiO2 NPs are able also to induce an inflammatory effect in brain of treated mice, as detected by the increased cytokine release [92, 99]. In the latter study, rutile and P25 Degussa TiO2 NPs intraperitoneally injected in mice after lipopolysaccharide (LPS) increased the mRNA levels of IL-1β and tumor necrosis factor (TNF)-α and also those of IL-1β protein. Interestingly, LPS induction was necessary to cause this phenomenon, suggesting the importance of a trigger element or of a possible synergistic effect in tissue responses to TiO2 NPs.

Finally, several studies demonstrated the embryotoxic role of maternal intravenous injection of no characterized TiO2 NPs [100], or subcutaneous injections of anatase TiO2 NPs [101103]. Following this latter treatment, TiO2 accumulation was detected in the offspring cerebral cortex and olfactory bulb and numerous olfactory bulb cells resulted positive to markers of apoptosis [101]. Moreover, the same prenatal exposure induced an altered expression of genes involved in brain development, cell death, and the response to oxidative stress in the newborn pups [102]. Finally, the influence of prenatal TiO2 NP exposure on the dopaminergic system was demonstrated by the increased DA, DOPAC, HVA and 3-methoxytyramina hydrochloride levels in the prefrontal cortex and in the neostriatum of exposed mice [103].

These data confirm that TiO2 NPs can be transferred from mothers to the fetal brain, inducing toxic effects on fetal brain development carrying a series of nervous system disorders. In this context, a deep knowledge regarding the influence of TiO2 NPs on neuronal cells is becoming urgent. Though additional studies are necessary to support these findings, they should be taken in consideration for the correct risk assessment of TiO2 NP exposure, particularly during a complex biological status like pregnancy and the early stage of life.

2.3. Dermal and Mucosal System

Studies regarding the effects of TiO2 NPs on the dermal system (in particular, the percutaneous absorption of NPs) show more homogeneous findings, demonstrating a clear absence of penetration through the intact epidermal barrier when TiO2 NP formulations are applied on different animal [104, 105] and human [106110] skins (Table 3). Considering the widespread use of TiO2 NPs in cosmetic sunscreens (due to their broad-spectrum UV absorption and high esthetical acceptance) [107], this is an important finding.

Sadrieh et al. [104] analyzed the correlation between different forms of TiO2 and percutaneous absorption in micropig skin, demonstrating that irrespective of the form, after a subchronic exposure, NPs were found in the stratum corneum but not in the deeper epidermal strata. Anatase TiO2 NPs subacutely exposed to hairless rat skin were only detected in the horny layer of the interfollicular epidermis, without penetration into viable cell layers or induction of any cellular changes [111].

Neither the shape [104, 105, 107, 108] nor the surface chemistry [107109] seems to influence NP penetration after acute and subacute exposure.

Pflücker et al. [107] demonstrated that all TiO2 NP types applied on skin areas of human forearm were only detected on the outermost layer of the stratum corneum, irrespective of their surface chemistry, NP size, and shape. The same results were confirmed by Schulz et al. [108]. Similarly, after application of lanceolate TiO2 NPs on the backs of pigs, high concentrations of Ti were found in the stratum corneum and granulosum, but not into the stratum spinosum [105].

Mavon et al. [109] demonstrated that 93% of the TiO2 T805 NPs hydrophobically coated with trimethyloctylsilane applied on human upper arms were found in the stratum corneum, decreasing continuously with depth. No TiO2 penetration into viable skin layers was assessed confirming the in vitro findings of the study.

The studies investigating the Ti penetration and distribution in biopsies obtained from human skin grafts transplanted into severe combined immune deficiency (SCID) mice could only detect Ti in the outermost layers of the stratum corneum of the skin biopsies, irrespective of the various NP formulation applied and the time points considered (1–48 h) [106, 110].

As also confirmed by Menzel et al. [105], Lademann et al. [113] failed to demonstrate a role for hair follicles as a percutaneous absorption pathway. In fact, Ti was detected in the stratum corneum and in the follicle channels but not in the interfollicular space under the stratum corneum or into the viable layers of the forearm skin of human volunteers.

Only a few reports detail Ti penetration into human dermis [112, 115]. For instance, in a pilot study conducted in 1996, Tan et al. [112] demonstrated percutaneous penetration of Ti in patients subchronically applied sunscreen containing microfine TiO2. Although Ti levels in the dermis of exposed subjects were higher than controls, these results were not statistically significant and were not confirmed by subsequent studies. Moreover, Wu et al. [115] reported the ability of TiO2 NPs in different form, anatase, rutile, P25 Degussa, and different sizes to penetrate through the skin, reaching different tissues (e.g., liver, spleen, and heart) where pathological lesions were induced after subchronic dermal exposure of hairless mice. Moreover, increased MDA levels were determined in the skin, which also showed excessive keratinization, thinner dermis and wrinkled epidermidis. These data could not be confirmed when reared pigs were treated with two types of the previously employed NPs for shorter periods of time.

The same interesting points of discussion emerged from this latter study. It seems that the penetrative ability of similar TiO2 NPs into the cutaneous layers depends on the animal species studied and on the period of exposure. Furthermore, pathological skin lesions due to NP application are likely ascribed to oxidative stress, which was also detected on human and murine dermal fibroblasts in vitro [118120].

Intradermal injection of TiO2 NPs aggravated atopic dermatitis (AD) symptoms related to mite allergen in mice assumed to show skin barrier dysfunction/defect through Th-2 immune inflammatory responses and histamine release [114]. These data support the importance of an intact dermal defense system to prevent uptake of TiO2 NPs, which can lead to skin damage, activation of the immune system, and eventually a decrease in the allergy threshold [105].

Ultimately, the importance of the human studies on percutaneous TiO2 absorption is underlined by the fact that these experiments alone are able to provide a relevant approach to direct human risk assessment and are the first step in the complex process of generalizing implications regarding environmental or occupational health.

Regarding skin carcinogenicity, conflicting results were reported [116, 117]. Tumor growth was increased when mice were intraperitoneally injected with TiO2 NPs prior to the subcutaneous implantation of B16F10 melanoma cells [116]. The authors suggest that TiO2 NPs might have the potential to enhance tumor growth in situ through immunomodulation of B- and T-lymphocytes, macrophages, and NK cells as supported also by results obtained in the spleen tissue. In contrast to these findings, Furukawa et al. [117] investigated the safety of coated and uncoated TiO2 NPs in a mouse medium-term skin carcinogenesis bioassay. Both TiO2 NPs applied to mouse skin in the postinhitiation phase did not increase the development of skin nodules histopathologically diagnosed as squamous cell hyperplasia, sebaceous gland hyperplasia, squamous cell papilloma, and keratoacanthoma. According to these data, TiO2 NPs do not possess postinhitiation potential and there is no carcinogenic risk relevant to percutaneous application of TiO2 NP preparations.

2.4. Cardiovascular System

Due to the complexity of the cardiovascular system, studies regarding the potential effects of TiO2 NPs on this system have focused on different functional aspects, such as cardiotoxicity [121, 122], cardiovascular parameters, such as systolic blood pressure (SBP) and heart rate (HR) variability [79], induction of thrombosis [22, 123], and the alteration of vasomotor responses [124128] (Table 4).

TiO2 NPs are able to induce high LDH, creatine kinase (CK), alpha-hydroxybutyrate dehydrogenase (HBDH), and aspartate aminotransferase (AST) activities used as markers of myocardial lesions, irrespective of the form of TiO2, anatase [122] or a mixture of anatase and rutile [129], or the route of exposure, intraperitoneal injections in mice [122], or oral gavage in rats [129]. Moreover, acute cardiotoxicity, in terms of higher serum LDH and alpha-HBDH enzymes, was demonstrated after a single oral gavage of TiO2 NPs administered to mice [121]. Unfortunately, the lack of data regarding the LDH isoform prevents a clear correlation between these increases and cardiac insults. Moreover, the study performed by Wang et al. [121] did not provide information on TiO2 form, and consequently, it does not allow comparisons with the two previous described works [122, 129].

When intratracheally instilled in rats, rutile Fe-doped TiO2 nanorods significantly increased SBP and HR in treated animals, which could be the consequence of the systemic inflammation induced by the same particles [79].

Regarding the TiO2 NP prothrombotic effect, the currently available data are conflicting. For instance, Chen et al. [22] observed thrombosis in the pulmonary vascular system of mice treated via intraperitoneal injection with anatase TiO2 NPs. On the contrary, Bihari et al. [123] failed to confirm thrombosis induction in both the mesenteric and cremasteric circulation of mice given parentally rutile TiO2 NPs. The higher exposure dose used by Chen et al. [22] compared to that of Bihari et al. [123], up to 2592 mg/kg versus 1 mg/kg, respectively, may be responsible for the induction of thrombosis ascribed to the blockage of blood vessels by TiO2 particles. Moreover, the difference in the form of TiO2 used by these two studies may explain some differences in the results. However, the very limited number of in vivo and in vitro studies on this regard does not support such conclusions. Thus, further studies are necessary to confirm these results and to investigate the mechanism(s) leading to thrombus formation.

The vasomotor response of the spinotrapezious muscle arterioles is impaired by P25 Degussa TiO2 NP inhalation [124, 125]. This vascular impairment was due to a dose-dependent reduction in NO endothelium production induced by microvascular oxidative and nitrosative stress [125]. Local ROS production may consume endothelium-derived NO or compromise NO endogenous production leading to vascular dysfunction as confirmed by a partially restored NO production by radical scavenging.

The loss of microvascular vasodilator capacity can significantly influence the normal homeostasis in any organ, causing impairment of tissue perfusion and compromising organ function. In this context, a further study by the same group demonstrated that after TiO2 NP exposure, coronary arteriole response to vasodilators was impaired though with conserved responsiveness to NO [126]. As previously detailed, the role motive of the increased microvascular ROS production was confirmed by the restored impairment after incubation with ROS scavengers [127].

Finally, vasomotor responses of intralobar pulmonary arteries removed from rats intratracheally exposed to Degussa TiO2 NPs were not altered in a study carried out by Courtois et al. [128]. The ex vivo contractile response to prostaglandin F2α and KCl and the relaxant response to Ach were not altered by TiO2 NP treatment [128, 130].

However, without appropriate particle characterization, it is difficult to compare findings between toxicological studies. Future studies are necessary to provide deeper knowledge regarding the toxic effects of TiO2 NPs on the cardiovascular system.

2.5. Liver

The toxic effects of TiO2 NPs on liver function were demonstrated in several studies through the detection of increases of AST, alanine aminotransferase (ALT) [21, 22, 79, 121, 122, 129, 131133], ALP, pseudocholinesterase (PChE), leucine acid peptide (LAP), TP, ALB [21, 40, 122, 132, 133], and globulin (GLB) levels. Decreased ALB/GLB ratios [21, 132] and increased ALT/AST ratios, a more sensitive indicator of hepatic injury [121, 131, 134], have also been detected (Table 5).

These results were obtained irrespective of the form of TiO2, particularly anatase [22, 40, 115, 122, 132135], rutile [79, 115] and P25 Degussa TiO2 NPs [115]; the NP different size [115, 121] and shape [79]; the modified surface chemistry of the particles [79].

Other studies demonstrated TiO2 NP hepatotoxicity effects, in terms of increased ALT levels and ALT/AST ratio both after a single oral administration [121] or repeated intraperitoneal treatments in mice [122], unfortunately, without reporting details regarding the form of TiO2 used.

Moreover, the route of exposure, such as the intragastric [21, 121, 129, 133, 136], intraperitoneal [22, 122, 131, 132], intratracheal [40, 75]; dermal [115], or intraarticular [134], does not seem to determine different toxic effects on liver function, both in terms of type and entity of manifestation. The intraarticular administration was used by Wang et al. [134] to simulate the release of NPs into joint cavities from the nanocoated surface of prostheses considering the good prospects for application of nanomaterials in prosthetic implants.

Interestingly, anatase, rutile, and P25 Degussa TiO2 NPs were able to penetrate the dorsal skin of hairless mice and to accumulate in the liver inducing oxidative stress, and focal necrosis in the parenchyma [115]. The potential role of an oxidative attack in causing liver damage was recently demonstrated in mice intragastrically [136] and intraperitoneally [135] exposed to nanoparticulate anatase TiO2. ROS accumulation, lipid peroxidation, and altered expression of genes involved in antioxidative or detoxification processes were triggered by NP treatment [136].

Interesting results were obtained by Li et al. [137] analysing the interactions between anatase TiO2 NPs and the DNA extracted by the liver of mice intraperitoneally exposed. They reported that nanoanatase TiO2 could be inserted into DNA base pairs, bind to DNA nucleotide, and alter the secondary structure of DNA. Moreover, NPs could cause liver DNA cleavage and hepatocyte apoptosis [137].

Several studies also detected signs of hepatic damage in terms of histopathological changes and hepatocyte ultrastructural alterations that could lead to impaired liver function [21, 22, 115, 121, 132, 134], while only few did not confirm these results [30, 129]. Particularly, hepatic fibrosis hydropic [22, 121] and fatty [133, 134] degeneration of hepatocytes, prominent vasodilatation [21, 132], and focal ischemia were induced by TiO2 NP treatment [132]. Necrotic [22, 121, 133] and apoptotic cells [132, 133, 136] were detected, and infiltration of inflammatory cells was also found [22, 79, 132134]. The inflammatory reaction in response to NP insult was also confirmed by significant increase of both mRNA and protein expression levels of several inflammatory cytokines and mediators in liver of treated mice [132, 133].

Regarding the metabolic homeostasis of mice treated with TiO2 NPs, several studies demonstrated increased contents of triglycerides (TG), total cholesterol (TCHO) [21, 122, 132], high-density lipoprotein cholesterol (HDL-C) [122, 132], and glucose [122] compared to controls, while low-density lipoprotein cholesterol (LDL-C) resulted lower [132].

Moreover, recent metabonomic studies, demonstrated the ability of TiO2 NPs to cause a series of changes in endogenous metabolite levels in the 1H nuclear magnetic Resonance (NMR) spectra of urine and serum [40, 75, 129], both after repeated intragastric doses of a mixture of rutile and anatase TiO2 NPs [129] and after acute intratracheal instillation of no characterised TiO2 NPs [75]. These findings support the role of TiO2 NPs in leading disturbances in the energy and amino acids metabolism through hepatotoxic effects and confirm the changes in liver function previously described.

Regarding the intravenous exposure, Yamashita et al. [100] detected TiO2 NPs in fetal liver after NP injections into pregnant mice, Fabian et al. [138] failed to show any hepatic toxicity in rats treated with injection of a mixture of anatase and rutile TiO2 NPs.

Negative results in terms of alteration in liver parameters were reported also in rats treated with a single intratracheal instillation of TiO2 NPs [30]. Interestingly, in this study, even if functional and histological liver lesions could not be demonstrated, hepatic superoxide dismutase (SOD) activity was reduced, and MDA levels were increased suggesting a possible role of oxidative stress and lipid peroxidation generated by NP exposure, perhaps as an early sign of damage, previous to clear functional parameter alterations. These findings are in line with the in vivo results previously detailed [115, 135, 136] and those obtained in vitro by Shi et al. [139] showing increased ROS and MDA in TiO2 NP-exposed hepatocytes.

In conclusion, though all of the studies reviewed in this section concentrate on hepatotoxic effects of TiO2 NPs, the current knowledge in this field is not yet exhaustive, and further investigation is necessary to fully elucidate the pathogenesis of the liver damage and the potential relationship between liver toxicity and the different characteristics of NPs. Interestingly, the interactions between TiO2 NPs and DNA, both direct or indirect, such as those mediated by oxidative stress, deserve greater attention in order to the understand their potential role in the mechanisms underlining genotoxic and carcinogenic effects. Moreover, no data are present in literature on the in vivo effects of NPs on the gastric and intestinal systems, and only limited studies on cells of the latter system have been performed in vitro. In our opinion, this aspect should be deeply investigated in future research due to its great importance, particularly considering the possibility of NPs penetrating into organisms through the intestines.

2.6. Hematopoietic and Immunological Systems

Only few studies have investigated the influence of TiO2 NP exposure on hematological parameters reporting conflicting results (Table 6). Evidence for NP-related upregulation of systemic inflammation, assessed by increased white cell count, was provided by several studies [43, 79, 129], whereas Duan et al. [21] failed to confirm these findings. Particularly, they demonstrated that the reduced levels of CD3, CD4, CD8-competent T lymphocytes, B, and natural killer (NK) cells and the reduced ratio of CD4 to CD8 detected were a clear sign of a disturbance of the cellular immune function and of the inhibition of the immune response of mice. Differences in the TiO2 form used, rutile [43, 79] versus anatase [21] versus a mixture of the two [129], the shape of the particles, nanorods [43, 79] versus NPs [21, 129], administration route, intratracheal [43, 79] versus intragastric [21, 129], and the length of exposure, acute [43, 79] versus subchronic [21, 129] may account for the discrepancies between the previous reviewed studies.

Moreover, Nemmar et al. [43] detected a reduced number of platelets in vivo due to a possible aggregation confirming the results obtained by the same authors in vitro. Unfortunately, these results were not confirmed when rats were intratracheally exposed to Fe-doped TiO2 nanorods, suggesting a potential role of the particle surface chemistry in influencing hematological effects in vivo [79].

However, all of these data should be interpreted as preliminary results and must be confirmed by future studies.

Other recent works have also investigated the effects of anatase TiO2 NPs on a key organ of the immunological system, the spleen, following both intraperitoneal injections [22, 116, 122, 140] or intragastric administrations [141] (Table 6). These studies revealed that TiO2 NPs are able to accumulate in the spleen in a dose-dependent manner [22, 122, 140] inducing histopathological damage. Apoptotic morphological changes [140], accumulation of neutrophils [22], congestion of the splenic tissue, and lymph node proliferation [140, 141] were detected. Moreover, reduction in total splenocyte, CD4 and CD8 T-lymphocyte number, retardation in B-lymphocyte development, and reduction in LPS stimulated NK cells were reported after exposure to TiO2 NPs [116].

A recent study [71], demonstrated that, in ovalbumin-sensitized mice, silica, coated rutile TiO2 NP inhalation decreased TNF-α and IL-13 expression in spleen cells.

On the other hand, Wang et al. [121] failed to reveal abnormal pathological changes in mouse spleen after a single intragastric administration. Unfortunately, this study lacks details regarding the form of the TiO2 employed.

TiO2 NPs induced increased ROS and MDA in the spleen [140, 141] and decreased levels of SOD antioxidant activity in plasma after intratracheal instillation in rats [30]. Oxidative stress was thought to play an important role in triggering the mitochondrial-mediated apoptotic pathway demonstrated in the spleen tissue [140].

However, the limited number of studies investigating the hematological and immunological effects of TiO2 NPs do not allow comparison and extrapolation of certain conclusions. Further research is necessary to confirm these results and to shed light on the roles of NP characteristics in inducing the alterations described above.

2.7. Renal System

Several studies report alterations, although opposite in some cases, of renal functional parameters, such as increased [40, 75, 121, 131] and reduced [122, 134, 142] blood urea nitrogen (BUN), increased [75, 121, 142] and decreased [134] creatinine (Cr), and reduced uric acid (UA) levels [122, 142] (Table 7). The impact of TiO2 NP treatment on the renal function was ulteriourly confirmed by the altered levels of metabolic products in the urine and serum of intratracheally exposed rats investigated by 1H NMR analysis [40, 75]. However, the clear interpretation of these data and the causative mechanisms are not fully understood. Unfortunately, of these studies, only few characterize the TiO2 form used as anatase [40, 122, 134, 142], while the others lack TiO2 particle characterization [75, 121, 131]. Moreover, the different routes of exposure employed to investigate nephrotoxicity, such as intraperitoneal [122, 131, 142], intra-articular [134], intragastric [121], and intratracheal [75] administrations, do not allow a clear comparison of the results obtained.

On the other hand, different studies did not report TiO2 NP-induced nephrotoxicity neither following intravenous exposure to a mixture of anatase and rutile [138] nor after anatase intraperitoneal injections [22], nor following intratracheal instillation with no characterized TiO2 NP suspension [30]. In this latter study, pathological changes were not observed in kidney tissues. In contrast, renal glomerulus swelling [22] and deposition of proteinic liquid in renal proximal tubules were described [22, 121, 134], also without functional damage [22].

The role of oxidative stress as a possible triggering mechanism for TiO2 NP inducing kidney damage was demonstrated by increased ROS generation [142], enhanced lipid peroxidation [30], as well as decreased SOD, catalase, ascorbate peroxidase, and glutathione peroxidase (GSH-Px) antioxidant activities [30, 142]. Surprisingly, these changes were detected also in a study where functional alterations were not present, suggesting their role as an early sign of damage, previous to clear functional parameter alterations [30].

These fragmented data require confirmation and further investigation, which should focus on the role played by the different characteristics of TiO2 NPs in inducing such alterations.

2.8. Musculoskeletal System

Two studies investigated the biological tolerance of TiO2 NPs when implanted in rat living tissues [143, 144] (Table 8). Interest in this type of research emerges from the promising utilization of NPs in numerous fields of medicine and the not fully understood potential reactions at the tissue-material interface. Inflammation and granulomas were detected in the vicinity of intramuscularly implanted TiO2 NPs as a normal reaction to a “foreign body,” but no signs of intolerance were observed in both studies at 6 or 12 months postimplantation.

Moreover, an inflammatory reaction, characterized by hypertrophy of the sinovial membrane, infiltration of lymphocytes and plasma cells, and proliferation of fibroblasts, was observed in rat knee joints when TiO2 NP suspensions were intraarticularly injected [134]. NPs were also able to stimulate oxidative stress as demonstrated by increases of antioxidants, such as GSH-Px and SOD activities, which were elevated to overcome an increase in ROS production.

TiO2 NPs were able to convert benign nonmetastatic tumor cells into malignant metastatic ones in a manner dependent on the NP surface chemistry and the ability to produce ROS in the target cells [145]. In particular, poorly tumorigenic and nonmetastatic QR-32 fibrosarcoma cells became tumorigenic only after injection into subcutaneous sites previously implanted with hydrophilic ZrO2Al(OH)3-treated TiO2 NPs. Moreover, tumor cell lines derived from the QR-32 cells of both TiO2 NP preimplanted sites acquired a pulmonary and extrapulmonary metastatic phenotype when intravenously injected in mice. Both the 8-hydroxy-2′-deoxyguanosine (8-OHdG) adduct and 4-hydroxy-2-nonenal-positive cells, induced by oxidative stress action, were increased in the hydrophobic ZrO2Al(OH)3 plus steric acid-treated TiO2 NP implantation sites compared to the hydrophilic one.

In conclusion, future research should investigate musculoskeletal tissue-TiO2 NP interactions, especially considering the potential widespread utilization of TiO2 NPs in biomedical products.

2.9. Reproductive System

Only few studies investigated the effects of TiO2 NPs on the male reproductive system [101, 131] (Table 8). Guo et al. [131] demonstrated reductions in sperm density and motility and an increase in sperm abnormality and germ cell apoptosis in mice treated with intraperitoneal injections of TiO2 NPs. Testes and epididymis did not show pathological changes.

When mice were prenatally exposed to anatase TiO2 NPs via subcutaneous injections of dams they showed aggregates of NPs in Leydig cells, Sertoli cells, and spermatids in the testes [101]. Disorganized and disrupted seminiferous tubules, tubule lumens with few mature sperm, as well as decreases in daily sperm production, epididymal sperm motility, and numbers of Sertoli cells were also observed [101, 146].

Such limited knowledge regarding the reproductive toxicological effects of TiO2 NPs and the relevance of this topic makes future investigations a matter of urgency.

2.10. Genotoxicity

TiO2 NP genotoxicity is poorly studied in vivo, and the mechanisms involved have not been clearly defined. To our knowledge, only two studies have investigated in vivo genotoxic effects and report contrasting results (Table 9). One study [147] demonstrated that TiO2 NPs are inert to lung cells, while the other [148] found that NPs induce genotoxicity in organs such as the blood, bone marrow, and embryos.

Specifically, Rehn et al. [147] showed that both hydrophilic P25 Degussa TiO2 NPs and hydrophobic trimethoxyoctylsilane-treated T805 TiO2 NPs intratracheally instilled in rats failed to induce persistent inflammation in BAL. The single lung cell immunohistochemical detection of 8-oxoguanine revealed no differences between both NP-treated groups and controls.

Trouiller et al. [148] exposed adult and pregnant dam mice to water supplemented with solutions containing P25 Degussa TiO2 NPs. Eyespots, a measure of DNA deletion, increased in in utero TiO2 NP-treated mice compared to controls. A dose-dependent increase in DNA double-strand breaks was also detected in bone marrow cells and an increased tail moment was evident in peripheral white blood cells of treated mice. MN frequency and the level of 8-OHdG were higher in peripheral blood erythrocytes and in the livers of TiO2 NP-exposed mice, respectively. Moreover, TNF-α, interferon (INF)-γ and the mouse orthologue of IL-8 were increased after exposure. These studies evaluated both genotoxic effects and inflammatory parameters, because most in vitro TiO2 NP genotoxicity studies have suggested that NPs could damage DNA directly or indirectly via oxidative stress and/or inflammatory responses [148]. The discrepancies between in vivo results reproduce the conflicting in vitro findings [33]. However, though the limited number of in vivo studies does not allow one to extrapolate conclusions, we hypothesize that different TiO2 forms, particle sizes, degrees of aggregation, preparation methods, incubation conditions, doses, and susceptibility between cell types influence responses to TiO2 NPs.

Interestingly, the study carried out by Trouiller et al. [148] is the first showing that DNA deletions in offspring are increased by in utero exposure to TiO2 NPs. This finding is of great importance considering the high susceptibility of embryonic cells to DNA damage and widespread environmental and occupational TiO2 NP exposure.

In conclusion, there are concerns regarding the potential risk of genetic disorders, particularly for people occupationally exposed to high doses of TiO2 NPs. Thus, further research is necessary to fully understand genotoxic TiO2 NP effects and to determine the conditions in which they occur to fully evaluate NP exposure risks.

3. Discussion

Our analysis pointed out interesting and critical aspects, particularly in relation to the interaction between NP-induced effects and remarkable treatment parameters, such as routes of exposure, TiO2 form, NP size, and surface reactivity, that we argue will be the focus of future research. In all the studies reviewed, the commonly proposed pathogenic mechanisms initiated by TiO2 NPs are dominated by inflammation-driven effects, including fibrosis, oxidative stress, and DNA damage, making inflammation a target for toxicological testing [7, 11, 149, 150]. Regarding the respiratory tract, which is considered the most affected organ, lung overload is thought to be responsible for the induction of inflammation, ROS production, and ultimately lung damage. Indeed, it should be carefully evaluated in future studies for both its “volumetric,” “mass,” and “surface area” determinants. Future research will be also necessary to clarify which NP characteristic could play a triggering role in the cascade of events responsible for lung injury. Moreover, a correct evaluation of the risk derived from TiO2 NP respiratory exposure should be performed in consideration of the significant role of this route of exposure for both the general and occupational populations. Particularly, in the latter setting, an appropriate risk assessment focused on the pattern of absorption, the role of the total inhaled amount, direct lung alteration, and/or damages in other distant organs, could greatly influence both the risk communication and the risk management phases, the adoption of valuable methods of environmental and biological monitoring, and the measures of collective and individual protection.

Interestingly, inhaled TiO2 NPs resulted able to be translocate to the CNS via the axons of sensory neurons in the upper respiratory tract [31, 91, 92]. This is an intriguing aspect because it introduces a route for TiO2 NP exposure directly linked to the CNS, in which NPs could elicit their toxic effects. These findings underscore the need for additional studies to further elucidate underlying mechanisms and to characterize the physiological impact, particularly in relation to which NP characteristics could influence this passage. The confirmation of this potential route of exposure may be extremely important for the general and occupational population, because the direct translocation of NPs to the CNS through the olfactory bulb could requires greater attention in public and workplace settings, as well as the urgent adoption of adequate protective measures.

Dermal studies have shown that there is currently little evidence for skin penetration and systemic exposure after dermal applications of TiO2 NPs used in sunscreens [107, 109, 112, 113]. However, the open question has been raised as to whether the usual testing with healthy, intact skin is sufficient [151, 152].

Regarding other organs, the data are too limited to extrapolate general concepts, and additional studies are required, which may offer new details regarding TiO2 NP absorption, systemic uptake, and kinetics, providing a deeper knowledge of the TiO2 NP toxicological profile. In this context, it could be helpful to investigate new, noninvasive, powerful techniques aimed to identify early pathological metabolic and toxicological processes induced by NPs [40, 153155]. Among these methods, metabonomics provide a relatively complete biological summary of the whole body response to NPs, and an appropriate identification of tissue damage biomarkers [40].

Moreover, while TiO2 has been classified as possibly carcinogenic to humans (Group 2B) [28, 29], no definite conclusion is present for TiO2 NPs. The NIOSH recently determined that inhaled ultrafine TiO2 is a potential occupational carcinogen and recommended an exposure limit of 0.3 mg/m3 for ultrafine (including engineered nanoscale) TiO2 as a TWA concentration for up to 10 hr/day during a 40-hour work week [15]. This conclusion was due to weight of scientific data demonstrating TiO2 NP carcinogenic effect on rat lungs [56]. Interestingly, NIOSH has reviewed that the tumor response observed in rats exposed to TiO2 NPs resulted from a secondary genotoxic mechanism involving chronic inflammation and cell proliferation [15]. The central hypothesis is that inflammation leads genotoxic events as well as cell proliferation and tissue remodeling, which are processes that are all required for mutations and progression towards neoplastic lesions [156]. However, the mutagenic or epigenetic mechanisms underlying the TiO2 NP carcinogenic action need ulterior confirms and their role in the multistep process of carcinogenesis, likely as trigger of the initiation, promotion or progression phases of tumor manifestation, should be deeply elucidated [33]. This argument merits much attention in future studies with a particular focus on the conditions in which TiO2 NP genotoxicity and carcinogenicity occur particularly in relation to the potential influence played by different particle production, size, degree of aggregation, preparation method, exposure conditions, dose, and susceptibility of animal species [33, 148, 157, 158]. These results will be of great importance in order to deeply characterize the risk of TiO2 NP exposure and to define suitable exposure limits.

The relevance of the toxicological research in animals lies on the possibility to evaluate the type and entity of the adverse response induced by different doses of TiO2 NPs and to extrapolate threshold levels aimed to protect exposed human populations. The general dose-response model indicates that the higher the concentration of a chemical substance, the greater is the influence on the organism [5]. However, in nanotoxicological research, the dose-response relationship requires special attention considering that the traditional mass dose does not well reflect the biologically effective dose for NPs. Other dose metrics, such as surface area combined with surface reactivity or the particle number, should be evaluated as ulterior, maybe better, descriptors of the potential to cause damage at the site of particle deposition or far from this [32, 82, 159, 160]. In this context, a TiO2 NP standardized characterization including parameters like TiO2 form, NP number, and mass concentration, size, shape, state of agglomeration, charge, crystallinity, solubility, oxidant generation capacity, added functional groups, or impurities and rate of dissolution is the first step to a comprehensive identification of the TiO2 NP hazard [2, 7]. It seems necessary to focus future research on establishing a matrix of relationship between specific NP properties and resultant biological activities. In fact, nanotoxicology literature indicates that the combination of various physicochemical features in any type of NP leads to different interaction with biological systems and consequently to different toxic potentials [34, 61, 161, 162]. A deeper understanding of parameters able to alter the TiO2 NP bioactivity would support a practical approach in the development of “safe” engineered NPs, through the choice of a particular TiO2 form, shape, or surface functionalization [163].

From the studies, we reviewed a great methodological heterogeneity has emerged in terms of type of TiO2 NPs employed, animal species treated, route of exposure, doses applied, endpoint parameters, and techniques of measurement. This causes a biased assessment of the exposure-response relationships, prevents comparison between results and extrapolation of definite conclusions [161]. Moreover, results obtained employing extremely high doses of TiO2 NPs and unrealistic routes of administration, such as bolus instillation or aspiration delivery, should be interpreted with caution as preliminary findings to be validated under more realistic exposure conditions [164]. However, these preliminary data could be useful to plan future experimental projects aimed to reveal early signs of more severe damages induced by higher dose treatments [33, 165].

Currently, concerns regarding nanotechnology in everyday life are emerging, particularly those relative to the risks posed by occupational exposure, in consideration of the increasing number of workers employed in research, manufacture, use, and disposal of nanomaterials [163, 166]. Indeed, although not definite, the potential risk for adverse effects in NP exposed workers demands that industry, labor, and government concert action to a precautionary NP risk management to protect worker health [161, 167172]. Critical steps in reaching this target are an appropriate risk assessment, based on epidemiologic research and exposure evaluation, and the determination of the effective function of risk management programs [163]. Epidemiologic studies are essential in understanding health outcomes associated with exposure to potentially hazardous materials. Such studies will form the basis for quantitative risk estimations to establish levels that protect human health. Currently, only NIOSH has recommended a TiO2 NP exposure limit, however, regulating TiO2 NPs appears to be a challenging task due to the broad diversity of the NP characteristics and their application in a wide range of sectors and products [173]. Moreover, it has to be questioned if a generic threshold limit, without any specification relative to the particle characterization and the context of employment, could have a suitable role in protecting health and safety of occupational exposed subjects.

NP exposure in workplace settings should receive much attention in future research. At present, the relative newness of exposure scenarios, generally occurring in controlled situations, the inconsistencies over the way to identify and classify nanomaterials, questions about metrics and practical instrumentation, and difficulty in gaining access to workplaces mostly prevent such essential assessment [174180].

Additionally, as a part of the occupational exposure and health effect evaluation, future studies should focus at identifying NP-associated biomarkers. In this regard, a deeper knowledge about TiO2 NP mechanisms of disease, and TiO2 NP biokinetics and potential health effects, obtained from in vitro and in vivo experiments, respectively, could be useful to identify indicators of early biological events and to establish appropriate methods for their detection. Experimentally identified biomarkers should be the focus of subsequent validation studies aimed to define their ability in exploring human disease endpoints [6]. Such confirmed biological indicators of exposure and effect could be of great relevance to the development of occupational health surveillance strategies and also to verify the appropriateness of collective and individual defense systems adopted during NP manipulation.

Finally, interpreting and communicating hazard and risk information is an integral part of the TiO2 NP risk management that clearly requires additional research to more thoroughly understand the toxicological interactions of these NPs with biological systems at the molecular, cellular, organ, and whole body level.