Oral toxicological study of titanium dioxide nanoparticles with a crystallite diameter of 6 nm in rats

Background Though titanium dioxide (TiO2) is generally considered to have a low impact in the human body, the safety of TiO2 containing nanosized particles (NPs) has attracted attention. We found that the toxicity of silver NPs markedly varied depending on their particle size, as silver NPs with a diameter of 10 nm exhibited fatal toxicity in female BALB/c mice, unlike those with diameters of 60 and 100 nm. Therefore, the toxicological effects of the smallest available TiO2 NPs with a crystallite size of 6 nm were examined in male and female F344/DuCrlCrlj rats by repeated oral administration of 10, 100, and 1000 mg/kg bw/day (5/sex/group) for 28 days and of 100, 300, and 1000 mg/kg bw/day (10/sex/group) for 90 days. Results In both 28- and 90-day studies, no mortality was observed in any group, and no treatment-related adverse effects were observed in body weight, urinalysis, hematology, serum biochemistry, or organ weight. Histopathological examination revealed TiO2 particles as depositions of yellowish-brown material. The particles observed in the gastrointestinal lumen were also found in the nasal cavity, epithelium, and stromal tissue in the 28-day study. In addition, they were observed in Peyer's patches in the ileum, cervical lymph nodes, mediastinal lymph nodes, bronchus-associated lymphoid tissue, and trachea in the 90-day study. Notably, no adverse biological responses, such as inflammation or tissue injury, were observed around the deposits. Titanium concentration analysis in the liver, kidneys, and spleen revealed that TiO2 NPs were barely absorbed and accumulated in these tissues. Immunohistochemical analysis of colonic crypts showed no extension of the proliferative cell zone or preneoplastic cytoplasmic/nuclear translocation of β-catenin either in the male or female 1000 mg/kg bw/day group. Regarding genotoxicity, no significant increase in micronucleated or γ-H2AX positive hepatocytes was observed. Additionally, the induction of γ-H2AX was not observed at the deposition sites of yellowish-brown materials. Conclusions No effects were observed after repeated oral administration of TiO2 with a crystallite size of 6 nm at up to 1000 mg/kg bw/day regarding general toxicity, accumulation of titanium in the liver, kidneys, and spleen, abnormality of colonic crypts, and induction of DNA strand breaks and chromosomal aberrations. Supplementary Information The online version contains supplementary material available at 10.1186/s12989-023-00533-x.

concerns have been raised on TiO 2 containing nanosized particles (NPs; at least one external dimension measuring ≤ 100 nm). In animal experiments, the noobserved-adverse-effect levels (NOAEL) on general toxicity endpoints for a 90-day feeding study of both food-grade (E171, 100% anatase, particle size 150 nm) [2] and nanosized TiO 2 (P25, 80% anatase and 20% rutile, primary particle size 14-21 nm, average secondary particle size 142.9 ± 43.97 nm) [3] in Sprague Dawley (SD) rats were a maximum dose of 1000 mg/kg bw/day. The NOAEL of both pigment-grade and ultrafine/nanoscale TiO 2 with different secondary particle sizes (50th percentile particle size [D 50 ] = 42-213 nm in dosing solution) in developmental toxicity studies was also a maximum dose of 1000 mg/kg bw/day [4]. However, owing to a concern for genotoxicity and other potential adverse effects, such as immunotoxicity, inflammation, neurotoxicity, and induction of aberrant crypt foci (ACF) in the colon, the European Food Safety Authority (EFSA) Panel on Food Additives and Flavourings (FAF) concluded that TiO 2 (E171) is no longer considered safe as a food additive [5].
Another concern is that the toxicity of NPs varies greatly with particle size. Our previous study demonstrated that intraperitoneal injection of silver NPs with a particle size of 10 nm in diameter (0.2 mg/mouse) exhibited severe toxicity with liver congestion and singlecell necrosis in female BALB/c mice, resulting in moribund or death within 24 h, whereas those of 60 nm and 100 nm in diameter did not exhibit such toxic effects [6]. This suggests that nanomaterials with very small particle sizes (~ 10 nm) may have more severe toxic effects than those with larger particle sizes. As for TiO 2 , it has been reported that intragastric administration of TiO 2 NP (5-6 nm) to CD-1 (ICR) mice for 90 days induced a decrease in body weight gain [7,8]. However, there have been no reports of subchronic toxicity studies on TiO 2 with a primary particle size of ≤ 10 nm in rats. Therefore, in this study, we examined the subacute and subchronic oral toxicity of TiO 2 nanoparticles with a crystallite diameter of 6 nm in male and female F344/DuCrlCrlj rats. Furthermore, we assessed the accumulation in major organs, that is, the liver, kidneys, and spleen, and the potential for the induction of DNA damage and colorectal abnormalities.

Characterization of TiO 2 NPs in the dosing suspension
When the TiO 2 NPs were suspended in ultrapure water at a concentration of 100 mg/ml to prepare a dosing suspension at a maximum dose of 1000 mg/kg bw/day, sedimentation separation was immediately observed, suggesting that the test substance was highly aggregated in water. As the specific surface area of TiO 2 NPs with a 6 nm crystallite size used in this study was 280 m 2 /g, such a large surface area can promote interparticule attraction and facilitate aggregation. Additionally, the isoelectric point of TiO 2 is close to neutral, which may cause large agglomerations when suspended in water. When 0.2% Na 2 HPO 4 (FUJIFILM Wako Pure Chemical, Osaka, Japan) was used as a dispersant, a uniform emulsion-like suspension was obtained. The median of secondary particle size was 206.3 ± 44.2 nm (Fig. 1) based on the particle number distribution from triplicate measurement of three independently prepared dosing suspensions (nine measurements in total). Although the secondary particle size of most of the fractions of the TiO 2 NPs in the suspension was > 100 nm, a small fraction with particles ˂ 100 nm in diameter (4.9 ± 11.1%) was occasionally observed, which possibly represented the free aggregates that were temporally released from agglomerates. Phosphorous is an essential nutrient and an unavoidable constituent; hence, JECFA adopted the Fig. 1 Particle size distribution of titanium dioxide nanoparticles suspended in 0.2% Na 2 HPO 4 . The left panel shows the particle count distribution, while the right panel displays the cumulative particle percentage. The data are presented as mean with a 95% confidence interval maximum tolerable daily intake (MTDI) rather than the acceptable daily intake to evaluate the safety of dibasic sodium phosphate as a food additive. Phosphorus intake at a dose of 10 mL/kg bw/day of 0.2% Na 2 HPO 4 was calculated to be 4.36 mg/kg bw/day, which was lower than that of the MTDI (70 mg/kg bw/day, as phosphorus) [9], suggesting that the administration of this concentration of Na 2 HPO 4 has no toxic effects. Therefore, a 0.2% Na 2 HPO 4 suspension of 100 mg/ml AMT-100 was used as the dosing solution.

28-Day subacute oral toxicity study
All animals survived during the study period. No changes in body weight were observed in either sex in response to treatment ( Fig. 2A), and food intake was comparable among all the groups (Fig. 2B).
The results of the hematological analysis (Table 1) showed that a significant decrease in white blood cell count (WBC) was observed in the 10 mg/kg bw/ day group of males, and a significant decrease in mean corpuscular hemoglobin (MCH) was observed in the 1000 mg/kg bw/day group of females. Biochemical analysis of the serum (Table 2) showed a significant increase in triglycerides (TG) in the 1000 mg/kg bw/day group of females and a significant decrease in creatinine (CRE) in the 100 and 1000 mg/kg bw/day groups of males.
For organ weights (Table 3), a significant increase was observed in the relative weight of the spleen in the 10 mg/kg bw/day of males and the absolute and relative weights of the thyroid in the 10 mg/kg bw/day of males. Additionally, the relative weight of the pituitary and prostate glands in the 100 mg/kg bw/day group of males and the relative weight of the thyroid gland in the 1000 mg/kg bw/day group of females was significantly increased.
Histopathological examination (Table 4) revealed yellowish brown materials in the lumen, epithelial mucosa, and submucosal tissue of the olfactory epithelium in the nasal cavity of males and females treated with 1000 mg/ kg bw/day. However, reactive changes such as inflammation or tissue injury were not observed (Fig. 3A, lower and upper left panels). Immunostaining for CD68, a macrophage marker, confirmed that macrophage infiltration was not induced at the deposition site (Fig. 3A, lowerright panel). In contrast, an inflammatory response was observed in the respiratory epithelium in areas without any visible particles (Fig. 3A, upper right panel). In the gastrointestinal tract, yellowish-brown materials were observed in the lumen; however, no inflammation was observed. Unlike the particles in the nasal cavity, those in the submucosal tissue were not clearly distinguished (Fig. 3B). Other sporadic histopathological changes were also observed, but they were considered incidental and of no toxicological significance.

90-day subchronic oral toxicity study
All animals survived during the study period. No changes in body weight were observed in either sex in response to treatment (Fig. 4A), and food intake was comparable among all groups (Fig. 4B). No significant changes were observed in the urinalysis at week 13 (Table 5).
Hematological analysis showed significant changes in leukocyte fractions [increase in neutrophils (NEUT), monocytes (MONO), and basophils (BASO), and a decrease in lymphocytes (LYMPH)] in the male 1000 mg/  (Table 6). However, no differences were observed in the WBC count and absolute numbers of each leukocyte type. A significant decrease in MCH and an increase in the absolute number of eosinophils (EO) were observed in the female 1000 mg/kg bw/day group. A significant decrease in the fraction and absolute number of BASO was observed in the female 100 mg/kg bw/day group.
Biochemical analysis of the serum showed a significant decrease in TG in the 300 and 1000 mg/kg bw/day male groups (Table 7). Additionally, high Na levels were observed in the 1000 mg/kg bw/day group of males, and low Cl levels were found in the 100 and 300 mg/kg bw/ day group of males and the 1000 mg/kg bw/day group of females. A significant increase in the relative kidney weight was observed in the male 100 mg/kg bw/day group (Table 8).
Histopathological examination revealed yellowishbrown materials in the tissues of the nasal cavity (Fig. 5A), trachea (Fig. 5B), lungs (bronchus-associated lymphoid tissue; BALT) ( Fig. 5C), cervical (Fig. 5D) and mediastinal (Fig. 5E) lymph nodes, and ileum (Peyer's patches) (Fig. 5F) in treated males and females. They exhibited a dose-dependent trend, and visible particles in Peyer's patches were observed only in males and females treated with 1000 mg/kg bw/day. No reactive changes, such as inflammation or tissue injury, were associated with the deposition of yellowish-brown material. Minimal or mild hypertrophy of mucous cells was observed in the respiratory epithelium and the nasopharyngeal duct of the nasal cavity (Fig. 5G). Minimal hypertrophy of mucous cells in the respiratory epithelium was also observed in the two control males. A few other histopathological changes were also observed, but all were incidental, and no dosing-related lesions were observed (Table 9).

Titanium levels in the organs
In both the 28-day and 90-day studies, trace amounts of titanium (Ti) were detected in the liver, kidneys, and spleen of all rats, including the control groups. In comparison with the control groups, a significant increase in Ti concentration in the liver was observed in the 1000 mg/kg bw/day group of females in the 28-day study (Fig. 6A). However, the difference in total Ti in the liver between the 1000 mg/kg bw/day group and the control group was 0.033 μg, which was the equivalent of 0.41 ppm of the daily Ti dose in the final week (Additional file 1: Table S1). There were no significant differences in the kidneys and spleen when compared with those in the control group. In the 90-day study, one of the females in the 1000 mg/kg bw/day group exhibited an approximately tenfold increase in Ti concentrations in the liver and spleen when compared with that of other animals; however, there were no significant differences when compared with those in the control groups (Fig. 6B) as well as total Ti levels in these organs (Additional file 1: Table S2).

Assessment of the possibility of induced abnormalities in the colon
Increased expression and nuclear/cytoplasmic translocation of β-catenin have been reported to be indicators of a direct premalignant lesion in the colon, dysplastic ACF, rather than hyperplastic ACF [10]; hence, immunohistochemical examination of β-catenin was performed using colon specimens from both 28and 90-day studies. Neither increased expression nor nuclear/cytoplasmic translocation of β-catenin was observed in the colon of the male and female 1000 mg/ kg bw/day groups in either study (Fig. 7A, B). Changes in cell proliferation were also assessed by immunohistochemistry of Ki-67, and no increased proliferation of colonic epithelial cells was evident in the male and female 1000 mg/kg bw/day groups in both studies (Fig. 7C, D).   20:23 Values are mean ± S.D *P < 0.05; **P < 0.01, compared with the 0 mg/kg bw/day group † n=4 due to loss of the organ of one animal at necropsy  Assessment of DNA-damaging potential EFSA concluded that TiO 2 particles have the potential to induce DNA strand breaks and chromosomal damage; therefore, the concern for genotoxicity of TiO 2 cannot be ruled out [5]. Thus, we examined the DNA-damaging and clastogenic potentials of TiO 2 NPs through the induction of micronucleus and γ-H2AX foci (Fig. 8A) in liver samples from the 28-day study. The liver is a major target Fig. 3 Representative images of yellowish-brown materials found at the submucosa of the nasal cavity (A) and gastrointestinal tract (B) of male and female F344/DuCrlCrlj rats orally administered titanium dioxide for 28 days. The submucosal particles are indicated using arrows. Scale bar for submucosal particles is 20 μm, while that for others is 50 μm organ for chemical carcinogens, and a repeated-dose liver micronucleus assay with 2-or 4-week repeateddose administration was reported to enable the evaluation of micronucleated hepatocytes in mature rats [11]. No increase in micronucleated hepatocytes was observed in either the male or female 1000 mg/kg group (Fig. 8B). γ-H2AX, a phosphorylated form of the histone variant H2AX at Ser139 [12], is strongly induced at DNA doublestrand break sites and is a useful biomarker for screening for chemical carcinogenicity in the bladder [13][14][15]. Consistent with the liver micronucleus study, there was no increase in γ-H2AX-positive hepatocytes in either the male or female 1000 mg/kg groups (Fig. 8C). In addition to the liver and bone marrow (Additional file 1: Figure S1), no γ-H2AX induction was observed at the sites of deposition of yellowish-brown materials in the nasal cavity, BALT, trachea, Peyer's patches, and cervical and mediastinal lymph nodes in the 90-day study (Fig. 8D).

Discussion
In this study, we carried out repeated oral administration of TiO 2 NPs with a crystallite size of 6 nm at a maximum dose of 1000 mg/kg bw/day for 28 or 90 days. Owing to the aggregative nature of TiO 2 NPs, animals were not exposed to free TiO 2 NPs with a diameter of 6 nm but to agglomerated large particles, and nanosized secondary particles constituted only a very small fraction of the dosing suspension. No significant changes were observed in the general condition, body weight gain, food intake, or urinalysis results in the 90-day study. In view of the importance of the safety of TiO 2 as a food additive, a number of studies have been conducted on the repeated oral administration of TiO 2 NPs (summarized in [5]). Some of the studies reported that repeated oral administration of TiO 2 NPs caused a decrease in body weight gain in male [7] and female [8] CD-1 (ICR) mice, male SD [16], and Wistar albino rats [17], but this finding was not observed in the present study. This discrepancy may be due to the differences in the physicochemical properties of TiO 2 NPs, dispersion status, experimental conditions, and species/strain. Although some significant hematological, serum biochemical, and organ weight changes were observed in the 28-and 90-day studies, we considered that these changes were of little toxicological relevance for the following reasons: The significant decrease in MCH observed in the 1000 mg/kg bw/day group of females in both the 28-and 90-day studies was considered of little toxicological significance because the change was minor, and there were no changes in the other erythrocyte markers suggestive of anemia in red blood cell count (RBC), hemoglobin (HGB), mean corpuscular volume (MCV), and mean corpuscular hemoglobin concentration (MCHC). The changes in TG were inconsistent in the 28-day and 90-day studies, with an increase in the TG level in the 1000 mg/kg bw/day group of females in the 28-day study, but a decrease in the ≥ 300 mg/kg bw/day group of males in the 90-day study. The lack of changes in other parameters, which was suggestive of abnormal lipid metabolism or liver damage in both studies, led us to conclude that the changes in TG levels observed in these studies were of little toxicological significance. Several studies have also observed changes in TG levels [16,18,19]; however, the EFSA FAF Panel considered that the changes in TG level reported in these studies were likely to be incidental because the reduction was observed only in one sex and/or with no clear dose response [5]. In the 90-day study, the significant changes in leukocyte fractions in the 1000 mg/kg bw/day group of males were of little toxicological significance because no significant differences in total leukocyte counts and absolute numbers of each leukocyte type were observed. The significant increase in the absolute number of EO observed in the 1000 mg/kg bw/day group of females was considered to have no toxicological significance, as it was a minor change, and no fractional changes were observed. Other findings, including the significant increase in Na in the 1000 mg/kg bw/ day male group and a significant decrease in Cl in the 100 and 300 mg/kg bw/day male groups and the 1000 mg/kg bw/day female group, were considered to be of little toxicological significance because they were minor changes and were not accompanied by changes in other parameters including renal dysfunction or histopathological renal changes. Other minor or non-dose-responsive changes were observed; however, they were considered toxicologically insignificant.
Histopathological examination revealed abundant small deposits of yellowish-brown material in the gastrointestinal lumen of rats that received TiO 2 NPs, indicating that similar yellowish-brown granular materials observed in other tissues also were TiO 2 NPs. The yellowish-brown material found in the nasal cavity was likely due to the reflux of the dosing suspension. Because higher concentrations of the TiO 2 NPs suspension may  cause aggregation, the dosing volume in this study was set at 10 mL/kg bw/day. This dose was higher than that used in our standard study; hence, it might have caused unintended reflux [20]. In the olfactory epithelium, yellowish-brown granular material was observed not only in the lumen but also in the mucosa and submucosa. In the 90-day study, yellowish-brown materials were found in the trachea, BALT, cervical and mediastinal lymph nodes, and Peyer's patches of the ileum in addition to the nasal cavity, suggesting that TiO 2 particles may accumulate with long-term exposure. The material found in the nasal cavity, the trachea, and BALT was considered to originate from the reflux of the dosing solution [20]. The material found in the cervical and mediastinal lymph nodes was considered translocated from the neighboring respiratory tract. The presence of yellowish-brown material in the Peyer's patches of the ileum indicated that TiO 2 NPs were taken up by the gut-associated lymphoid tissue, which is consistent with previous findings [21]. Our microscopic examination did not find TiO 2 NPs in other tissues, including mesenteric lymph nodes suggested that the absorbed TiO 2 NPs persisted in the neighboring lymphoid tissues. Although previous studies have reported that the intestinal incorporation of TiO 2 NPs [22] and their presence in the mucosa of the small intestine, liver, and spleen by transmission electron microscopy [23], routine histopathological analysis did not reveal the presence of TiO 2 NPs in these organs in the present study. It has been reported that pulmonary deposited TiO 2 NPs were translocated to the liver [24,25]. In this study, although TiO 2 NPs were unintentionally incorporated from the respiratory tract owing to reflux, it is likely that the exposure from this bypass was not sufficient for the detection of Ti in the liver. With regard to the toxicological effects of the accumulated TiO 2 NPs in the organs, the significant increase and increasing tendency in the incidence of mucous cell hypertrophy in the nasal cavity in the 1000 mg/kg bw/ day group of females and males suggested that they were treatment-related changes; however, the deposition of yellowish-brown materials and mucous cell hypertrophy in the respiratory epithelium and nasopharyngeal duct of the nasal cavity were not co-localized. Moreover, the degree of histological changes was mild and was also seen in the two control males. Hence, the biological effects were considered minor. Notably, no reactive changes such as inflammatory reactions or tissue injury, were observed at the site of deposition, which is consistent with previous findings [26], suggesting that there were no adverse effects on the organs.
For Ti levels in major organs, a significant increase in liver Ti levels was observed in the 1000 mg/kg bw/day group of females when compared with that of the control   been reported in the liver: TiO 2 NPs with a size of 24 nm induced edema, fatty degeneration, and necrosis in SD rats [28]; those with a size of 5-6 nm induced lymphocyte infiltration and necrobiosis in male CD-1(ICR) mice [7]; food-grade TiO 2 (E171) induced necroinflammatory foci in male NFR mice [29]. No biochemical or histopathological changes suggestive of liver toxicity were observed in this study, which may be attributed to the fact that no treatment-related increase in Ti content was observed in the liver. Although the absorption of TiO 2 NPs is low, studies have reported that they can accumulate in various organs, and their persistence in the body is a concern due to their long systemic half-life, which is estimated to be approximately 200-450 days [5] after intravenous administration to rats. After oral administration of 250 mg/ kg bw/day of TiO 2 NPs with primary sizes of 5-10 nm or 20-25 nm (they formed aggregates of approximately 100 nm and more, or larger than 100 nm, respectively, in aqueous solution) to male SD rats for 28 days, Ti was detected in the brain, lungs, heart, liver, kidneys, spleen, small intestine, testes, and blood with the 5-10 nm NPs, but only in the liver, kidneys, spleen, and small intestine with the 20-25 nm NPs [30]. The translocation of NPs into organs as well as renal excretion can be affected by their secondary particle sizes, e.g., NPs with a hydrodynamic diameter of less than 5.5 nm excreted rapidly and efficiently in urine [31]. In the present study, 1000 mg/kg bw/day of TiO 2 NPs with a primary size of 6 nm did not accumulate in the liver, kidneys, or spleen, which may be attributed to the fact that only a small fraction of nanosized secondary particles (50-100 nm) was present in the dosing suspension. However, there are conflicting reports that the 90-day repeated oral administration of TiO 2 NPs with a hydrodynamic size of 37.4 ± 0.4 nm in water also exhibited no accumulation in the liver, kidney, spleen, and brain [32], and TiO 2 NPs with a far greater hydrodynamic diameter of 352 nm were also found to be mainly excreted in urine [33]. Therefore, further research is required to elucidate the relationship between the physicochemical properties and toxicokinetics of TiO 2 NPs. Furthermore, it should be noted that NPs with extremely small crystallite diameters have markedly higher specific surface areas, they may have different physiological effects than that of the non-nano-sized crystallites, though they were aggregated into non-nano-sized secondary particles. It has been reported that TiO 2 induced ACF in rat colon [34]. Although we did not monitor the ACF, immunohistochemical analysis of β-catenin and Ki-67 did not reveal crypt abnormalities.
Several genotoxicity studies have reported positive results in micronucleus/chromosome aberration tests and comet assays in vivo, and the genotoxic potential of TiO 2 NPs is of great concern to EFSA, particularly with regard to the induction of DNA strand breaks and chromosomal aberrations [5]. Therefore, we examined the potential genotoxic effects of TiO 2 NPs in the liver and in sites where yellowish-brown materials were deposited, using γ-H2AX as a marker for DNA damage. In this study, no increase in micronucleated or γ-H2AX hepatocytes was observed in the liver of males and females dosed with 1000 mg/kg bw/day, which does not support the DNA-damaging potential of TiO 2 NPs in the liver. Although these assays are appropriate for evaluating the induction of DNA double-strand breaks and consequent structural chromosomal damage, they Values are mean ± S.D **P < 0.01, compared with the 0 mg/kg bw/day group  are not defined as genotoxicity tests by toxicity testing guidelines, and the absence of appropriate concurrent positive controls in the present study precludes us from concluding that TiO 2 NPs do not induce DNA damage or chromosomal abnormalities. Additionally, histopathological examination and quantification of Ti in the liver suggested that the liver was not exposed to orally administered TiO 2 NPs in this study. However, our results were consistent with a previous report that TiO 2 NPs translocated to the liver by pulmonary Table 9 Histopathological findings of F344/DuCrlCrlj rats treated with titanium dioxide nanoparticles for 90 days *P < 0.05; **P < 0.01, compared with the 0 mg/kg bw/day group † n = 9 due to the organ was not found at specimen preparation exposure did not induce DNA damage, as assessed by the comet assay, in contrast to carbon black NPs [25]. Moreover, the absence of γ-H2AX positive cells in the tissues where yellowish-brown material was deposited indicated that TiO 2 NPs did not induce DNA damage in these tissues. Taken together, our results revealed no evidence supporting the potential induction of genotoxicity by oral administration of TiO 2 NPs in this study.

Conclusion
Our study revealed that there were no toxic changes, including general toxicity, induction of colonic abnormalities, DNA-damaging potential, and accumulation of Ti in the liver, kidney, or spleen following the oral administration of anatase TiO 2 NPs with a crystallite size of 6 nm for 28 or 90 days. The NOAEL in both 28-and 90-day studies was 1000 mg/kg bw/day. Our results provide further evidence for evaluating the safety of oral exposure to TiO 2 that may contain very small crystallites.

Test material and preparation for administration
Anatase-type nanosized TiO 2 with a crystallite diameter of 6 nm (AMT-100, purity 93%, specific surface area of 280 m 2 /g, according to the manufacturer) was provided from Tayca (Osaka, Japan). As per the manufacturer's communication, the primary particle size distribution of AMT-100 was ranged from approximately 3.5-7 nm, with a modal diameter of 5 nm accounting for over 75% frequency, as measured by transmission electron microscopy. The highest dose was set at 1000 mg/kg bw/day, in accordance with the Guidelines for the Designation of Food Additives and Revision of Standards for Use (Ministry of Health, Labour and Welfare, Japan). For the 28-day subacute toxicity study, medium and low doses were set at 100 and 10 mg/kg bw/day, respectively, to obtain a wide dose range because no toxicity information was available for the test material. In the 90-day subchronic toxicity study, the medium and low doses were set at 300 and 100 mg/kg bw/day, respectively. 0.2% Na 2 HPO 4 (FUJIFILM Wako Pure Chemical, Osaka, Japan) was used as a dispersant, as previously described [35], and its administration volume was set at 10 ml/kg bw to intend to prevent aggregation. The particle size in suspension (secondary particle size) was determined by dynamic light scattering using a Zeta-potential and Particle size Analyzer ELSZ-2 (Otsuka Electronics, Hirakata, Osaka, Japan). The stability of the particle size in the suspension was confirmed at 3 h after preparation. The TiO 2 suspension was freshly prepared daily, mixed well for each dose, and administered within 2 h of preparation.

Experimental animals
Five-week-old specific pathogen-free male and female F344/DuCrlCrlj rats were purchased from the Charles River, Japan. CRF-1 basic diet (Oriental Yeast, Tokyo, Japan) and water were provided ad libitum. After a 1-week acclimation period, the rats were subjected to experiments at 6 weeks of age. The animals were housed in a barrier system at a temperature of 23 ± 1 °C, humidity of 50 ± 5%, ventilation frequency of 20 times/h, and 12-h light/dark cycle. The animals were housed in polycarbonate box cages lined with soft chips (Sankyo Lab Service, Tokyo, Japan) as bedding, and the cages and bedding were changed twice per week. Animal experiments were conducted after review and approval by the Animal Experiment Committee of the National Institute of Health Sciences and in compliance with the "Regulations Concerning the Proper Conduct of Animal Experiments" established by the institute.

Animal treatment
The subacute (28-day) and subchronic (90-day) studies were conducted in accordance with the Organisation for Economic Co-operation and Development Guidelines for the Testing of Chemicals 407 and 408, respectively. The animals were divided into four groups (5 and 10 animals per group for 28-day and 90-day studies, respectively) based on their body weight on the day before the first dose. The animals were housed five animals/cage in the 28-day study and three, three, and four animals/cage in the 90-day study. The dosing suspension of TiO 2 NPs was administered once daily by oral gavage using a flexible polytetrafluoroethylene tube (length: 85 mm, outer diameter: 1.46 mm, head: 2.4 mm, Fuchigami Instruments, Muko, Kyoto, Japan) at volume based on the latest body weight. During the experimental period, the general condition and mortality of the animals were observed daily and their body weight and food intake were measured once a week. After the administration period, the animals were fasted overnight and laparotomized under isoflurane inhalation anesthesia. Blood samples were collected from the abdominal aorta of all animals, and the animals were euthanized by exsanguination.

Urinalysis
In the 13th week of the 90-day study, five out of 10 animals in each group of males and females were placed individually in metabolic cages (Natsume Seisakusho, Tokyo, Japan) and kept for 4 h under fasting, and urine samples were spontaneously collected. AUTION Sticks 10EA (Arkray, Kyoto, Japan) were dipped in fresh urine and glucose, protein, bilirubin, urobilinogen, pH, specific gravity, occult blood, ketone body, nitrile, leukocytes, and color were measured using the Urine Analysis System AE-4020 (Arkray).

Hematology and serum biochemistry
An aliquot of collected blood was transferred to a Venoject II vacuum collection tube (Terumo, Tokyo, Japan).

Organ weights and histopathology
During necropsy, the brain, thymus, lungs, heart, spleen, liver, adrenal glands, kidneys and gonads were weighed. Endogenous peroxidase activity was devitalized using 3% H 2 O 2 /methanol for 10 min. After blocking with 10% normal goat serum (Nichirei Biosciences, Tokyo, Japan), the slides were incubated with primary antibodies at 4 °C overnight, washed three times with phosphate-buffered saline (PBS) and incubated with appropriate Histofine Simple Stain MAX-PO(R) horseradish peroxidase-conjugated secondary antibodies (Nichirei Biosciences). Next, the slides were washed three times with PBS, followed by incubation with 0.05% 3,3'-diaminobenzidine tetrahydrochloride/0.0286% hydrogen peroxide in PBS for color development. Finally, the slides were immersed in hematoxylin for counterstaining and then permanently mounted. To measure the number of γ-H2AXpositive hepatocytes per liver area, the area of the liver parenchymal region in the left lateral lobe and medial lobe sections was measured, excluding major vacuoles, bile ducts, blood vessels, and stromal regions. The number of γ-H2AX positive hepatocytes within that area was counted under a microscope. The mean ± standard deviation of the observed area was 156.6 ± 11.2 mm 2 in males and 121.1 ± 13.3 mm 2 in females.

Determination of titanium content
Titanium (Ti) concentrations in the liver, kidneys, and spleen of all animals were measured using inductively coupled plasma mass spectrometry (ICP-MS) at Japan Food Research Laboratories (Tokyo, Japan). Microwave digestion with 2 mL water and 3 mL nitric acid (Kanto Chemical) was used for sample preparation. An Ultra-WAVE microwave oven (Milestone General, Kawasaki, Kanagawa, Japan) was used for digestion under the following program: 10 min at 120 °C with a maximum power of 800 W, 15 min at 250 °C with a maximum power of 1500 W, and 15 min at 250 °C with a maximum power of 1500 W. The digested materials were transferred to polypropylene containers and concentrated to approximately 2 mL at 110 °C on a hot plate, and 0.025 mL of hydrofluoric acid (FUJIFILM Wako Pure Chemical) was added. After cooling, 2 mL of the internal standard solution was added and the test solution was made up to 50 mL with water. The internal standard solution was prepared by mixing 2500 µL of gallium, indium, tellurium, and thallium mixed standard solution (SCP Sciences, Quebec, Canada) and 125 µL of rhodium standard solution (Sigma Aldrich, St. Louis, MO, USA) diluted with water, nitric acid, and acetic acid. L-cysteine was added, and the volume was fixed at 1000 mL. ICP-MS measurements were performed using an Agilent 8800 Triple Quadrupole ICP-MS instrument (Agilent Technologies, Santa Clara, CA, USA). The analysis conditions were as follows: radio frequency power, 1550 W; carrier gas, Ar; flow rate, 15 L/min; collision gas, O 2 ; internal standard, Rh; and the target elements of ICP-MS analyses were 48 Ti 16 O + and 103 Rh.

Hepatocellular micronucleus test
The hepatocellular micronucleus test was performed as previously described [11], using formalin-fixed liver tissues of all animals in the control and high-dose groups in the 28-day study. Briefly, ten 3 mm-cubes of diced and fixed liver tissue were incubated in 50% potassium hydroxide (FUJIFILM Wako Pure Chemical) at room temperature for 16 h and washed with water. The cubes were homogenized using a Potter-Elvehjem grinder, filtered through a cell strainer with a pore size of 100 μm (Corning), and resuspended in water. The suspension was centrifuged at 50×g for 5 min and washed three times with 10% neutral-buffered formalin (FUJIFILM Wako Pure Chemical). The suspension was mixed, stained with an equal volume of 2 × SYBR Gold (Thermo Fisher Scientific, Waltham, MA, USA), dropped onto glass slides, and covered with a glass coverslip. At least 2000 hepatocytes were analyzed for each individual by fluorescence microscopy using BZ-X710 (Keyence, Osaka, Japan) equipped with a 40 × optical lens and a GFP filter.

Statistical analysis
Significant differences in body weight, hematological, and serum biochemical test results, organ weights, and organ Ti content were evaluated using Dunnett's test. A p value ˂ 0.05, calculated using two-sided tests indicated statistical significance in all analyses. Significant differences in the incidence of histopathological findings were evaluated using the Fisher's exact probability test. Graph-Pad Prism 8 (GraphPad Software, San Diego, CA, USA) was used for statistical analyses.
Additional file 1. Table S1. Titanium content in the liver, kidneys, and spleen of F344/DuCrlCrlj rats treated with titanium dioxide nanoparticles for 28 days. Table S2. Titanium content in the liver, kidneys, and spleen of F344/DuCrlCrlj rats treated with titanium dioxide nanoparticles for 90 days. Figure S1. Representative images of immunohistochemical detection of γ-H2AX in the liver (A) and bone marrow (B). No increase in γ-H2AX positive cells was noticed in either organ. Scale bars are 50 μm.