*2.4.6. Specificities of food-grade TiO<sup>2</sup>*

Food-grade TiO2 powders are finally characterized by a low specific surface area (around 10 m2 /g), a pure crystalline anatase phase (sometimes traces of rutile), a low isoelectric point (around 4.1 in water) related to the phosphate found at its surface, a mean size of 140 nm with a distribution spanning from 30 to 300 nm and a fraction of nanoparticles comprised between 17 and 36%. For toxicological studies, including toxicity assessment by oral exposure, another kind of TiO2 , called P25, is commonly used as it is considered as a reference material [34]. This compound is characterized by 100% NPs, a mean size of 23 nm, a specific surface area of 50 m<sup>2</sup> /g, a mixture of anatase and rutile grains (85/15) and an isoelectric point at pH 6.5 [18, 19]. In **Figure 3**, some physical and chemical properties of E171 and P25 samples, extracted from two studies [18, 19], are reported.

The P25 samples clearly distinguish from E171 samples by all parameters taken into account. A peculiar sample of E171, rich in rutile phase, is observed as well. E171 TiO<sup>2</sup> being strongly different from the reference material P25, we thus concluded that P25 does not appear to be the most suitable reference material for toxicity studies by ingestion [18]. It is, moreover, not the most relevant material to represent the nanoparticle fraction of E171.

**Figure 3.** Physical and chemical parameters describing E71 and P25 forms of TiO<sup>2</sup> , namely the content in NPs, isoelectric point (IEP), the mean size of the distribution deduced by transmission electron microscopy (mean S) and the span of the particle size distribution (span). Data come from references [8, 19].

#### **2.5. Fate of TiO2 after ingestion**

leading to a mean hydrodynamic diameter of 150 nm [4]. In solutions added with salts (NaCl

a particle size of agglomerates remaining between 360 and 390 nm for at least 2 h. The same experiment conducted with P25 sample showed rapid and extensive aggregation of the par-

), E171 particles dispersed by sonication presented a moderate stability, with

powders are finally characterized by a low specific surface area (around

, called P25, is commonly used as it is considered as a reference material [34]. This

/g,

being strongly

, namely the content in NPs, isoelectric

/g), a pure crystalline anatase phase (sometimes traces of rutile), a low isoelectric point (around 4.1 in water) related to the phosphate found at its surface, a mean size of 140 nm with a distribution spanning from 30 to 300 nm and a fraction of nanoparticles comprised between 17 and 36%. For toxicological studies, including toxicity assessment by oral exposure, another

compound is characterized by 100% NPs, a mean size of 23 nm, a specific surface area of 50 m<sup>2</sup>

a mixture of anatase and rutile grains (85/15) and an isoelectric point at pH 6.5 [18, 19]. In **Figure 3**, some physical and chemical properties of E171 and P25 samples, extracted from two

The P25 samples clearly distinguish from E171 samples by all parameters taken into account.

different from the reference material P25, we thus concluded that P25 does not appear to be the most suitable reference material for toxicity studies by ingestion [18]. It is, moreover, not

point (IEP), the mean size of the distribution deduced by transmission electron microscopy (mean S) and the span of the

A peculiar sample of E171, rich in rutile phase, is observed as well. E171 TiO<sup>2</sup>

the most relevant material to represent the nanoparticle fraction of E171.

**Figure 3.** Physical and chemical parameters describing E71 and P25 forms of TiO<sup>2</sup>

particle size distribution (span). Data come from references [8, 19].

and NaHCO3

12 Application of Titanium Dioxide

Food-grade TiO2

kind of TiO2

*2.4.6. Specificities of food-grade TiO<sup>2</sup>*

studies [18, 19], are reported.

ticles [4].

10 m2

Among the different routes of exposure to TiO<sup>2</sup> , the oral uptake route remains the less documented. Once ingested, TiO<sup>2</sup> particles pass through the digestive tract, starting with the port of entry, the oral cavity followed by the gastrointestinal tract, comprising oesophagus, stomach, small and large intestines and rectum (**Figure 4**).

During the transit through the digestive fluids, TiO<sup>2</sup> particles were not metabolized and were found to be mainly agglomerated, mediated by proteins and electrolytes [35, 36], but according to some studies, a small fraction is still in the nanosized range [35, 37, 38]. The low absorption of TiO2 and reversely the high percentage of titanium dioxide excreted from the body in faeces [39, 40] were believed to be the proof of any adverse effect. However, the recent data on the intestinal compartment call this belief into question. Indeed, the intestinal barrier, which involves epithelium, mucus and microbiota in its luminal side (**Figure 4**), provides a physical, chemical and biological line of defence for the host, probably through an orchestrated manner [41, 42]. Taken together or independently, these three partners exhibit some alterations due to the presence of TiO2 particles, which are briefly reported from the microbiota to the epithelium.

#### *2.5.1. TiO2 in interaction with the intestinal microbiota*

The effects of TiO<sup>2</sup> on the gut microbiota composition and metabolic activity in animal models or humans are largely unknown, whereas the intestinal microbiota contributes actively to the maintenance of host homeostasis. Indeed, it plays a key role in the gut, fulfilling protection, maturation and production functions. In particular, it acts as a barrier against pathogens, preventing their implantation, and participates in xenobiotic metabolism [43].

**Figure 4.** Schematic representation of the fate of TiO2 within the digestive tract illustrating potential mechanisms by which ingested nanoparticles interact with the intestinal barrier; (1) mucus; (2) microbiota; (3–5) epithelium with (3) internalization and active transport to Peyer's patch lymphoid follicles by M-cells, (4) transcellular transport and (5) paracellular transport through intercellular tight junctions (intercellular space between adherent epithelial cells). The unrestricted migration through foci of damaged epithelium is not represented. For the sake of clarity, scheme is not to scale.

Studies reported to date were mainly focused on the antibacterial activity of TiO2 nanoparticles in *in vitro* pure cultures using *Escherichia coli* as the bacterial representative [44–46]. Such an activity is generally associated with the photocatalytic effects of TiO<sup>2</sup> , although increasing experimental evidence also demonstrated TiO2 -mediated cell alterations without UV illumination [50, 51]. Taylor et al. [47] investigated the *in vitro* exposure of a gut microbial community from a healthy donor to three different types of metal oxide nanoparticles, including TiO<sup>2</sup> , in a model colon. Such exposure-induced changes in the phenotypic traits of the gut community, including short-chain fatty acid production (particularly for butyric acid), cell hydrophobicity, sugar content of extracellular polymers, cell size and electrophoretic mobility. In a further study, Waller et al. [48] evaluated the impact of food-grade TiO2 (vs industrial-grade TiO2 ) on the composition and phenotype of a human gut microbiota. An inhibition of the control-induced shift in microbial composition from Proteobacteria to Firmicutes phyla was observed. TiO<sup>2</sup> exposure also resulted in a lower value of the colonic pH (∼pH 4) as compared to the control (>5). Additionally, similar trends in microbial community hydrophobicity and electrophoretic mobility were obtained between control and food-grade exposures. Interestingly, different microbial responses were observed with the industrial-grade form, underlying the significance of physical and chemical properties of TiO2 in intestinal homeostasis.

#### *2.5.2. TiO2 in interaction with the intestinal mucus*

Mucus is the viscoelastic gel that lines and protects the intestinal epithelium. It is secreted continuously along the whole intestine by specialized goblet cells in the epithelium (**Figure 4**), and is present in larger amounts in the colon than elsewhere. Mucus was long considered to act as a 'simple' physical barrier, but it is now known to have other key functions essential for the preservation of intestinal homeostasis [49–51], including (i) lubrication of the epithelium, facilitating the progress of material along the digestive tract, (ii) maintenance of a stable microenvironment at the epithelial surface, (iii) protection of the epithelium through the presence of immune system molecules and (iv) provision of an ecological niche for the intestinal microbiota.

Interactions between TiO<sup>2</sup> and intestinal mucus are far from being understood. Variable capacities for absorption and transport of TiO2 nanoparticles have been described *in vitro* [52], depending on whether epithelial cells are cultured alone or in the presence of mucus-secreting goblet cells. In fact, Caco-2 cells in monoculture only displayed low levels of intracellular nano-TiO2 accumulation after 24-h exposure, whereas the same treatment in Caco-2/HT29-MTX mucus-producing co-culture led to 50 times higher levels of accumulation [52]. In ex vivo studies on porcine buccal mucosa [36, 38], TiO2 nanoparticles, regardless of their size and hydrophilicity/hydrophobicity, were able to permeate mucus and penetrate underlying tissues.

#### *2.5.3. TiO2 in interaction with the intestinal epithelium*

Epithelium is in charge of nutrients and water absorption while restricting the access for potentially noxious substances to the internal organs. Thus, it constitutes a selective—and dynamic—barrier, mediating transport of compounds through the transcellular pathway (i.e., across the cells) and/or the paracellular pathway (i.e., between the cells). It is polarized into an apical and basolateral surface with the apical surface covered with microvilli to increase the absorptive surface area. There are at least three pathways enabling uptake/translocation of TiO2 nanoparticles (**Figure 4**): first, they can disrupt the cell junctions (paracellular route), second they can be internalized by the cells (transcellular route, e.g., endocytosis) and finally they can exert a toxic effect on the cells or alter their function, resulting in cell death [53]. In addition, many studies underlined the involvement of the M-cell-rich layer of Peyer's patches which are epithelial cells specialized for the transcytosis of macromolecules and particles [40, 53–56] (**Figure 4**). However, this mechanism of translocation is still under debate since contradictory results were obtained for *in vitro* cells [36, 40, 53, 55–57].

*In vitro* studies, mainly on Caco-2 cells, converge on the possible TiO2 -mediated disruption of the epithelial barrier. Indeed, subtle or more substantial alterations were depicted, including cytotoxicity [58], alteration of the brush-border microvilli [26, 53], upregulation of nutrient transporters and efflux pumps [59], production of reactive oxygen species [59, 60], misbalance of redox repair systems [59], increase in epithelial permeability [60] and uptake/translocation of TiO2 nanoparticles [53, 55, 60], at a different extent according to the type of TiO<sup>2</sup> nanoparticles (size and crystal phase) and experimental conditions used.

In line with the findings of Faust et al. [26], recent piece of evidence suggests some adverse effects of oral exposure to E171 on the intestinal mucosa barrier with a putative additional impact on intestinal diseases and colorectal cancer [61–63]. Proquin et al. [63] showed *in vitro* that E171-induced ROS formation and DNA damage through its micro-sized and/or nanosized fractions in Caco-2 and HCT116 cells. In rodents, Bettini et al. [61] found TiO2 particles present in Peyer's patches along the small intestine as well as in the colonic mucosa of rats orally given E171 at human relevant levels. No significant change in epithelial paracellular permeability was observed.
