Annexe 3b. Detailed evaluation of in vivo papers with green and amber ratings
Published 11 October 2024
© Crown copyright 2024
This publication is licensed under the terms of the Open Government Licence v3.0 except where otherwise stated. To view this licence, visit nationalarchives.gov.uk/doc/open-government-licence/version/3 or write to the Information Policy Team, The National Archives, Kew, London TW9 4DU, or email: psi@nationalarchives.gov.uk.
Where we have identified any third party copyright information you will need to obtain permission from the copyright holders concerned.
This publication is available at https://www.gov.uk/government/publications/statement-on-the-com-assessment-of-in-vitro-and-in-vivo-genotoxicity-of-titanium-dioxide/annexe-3b-detailed-evaluation-of-in-vivo-papers-with-green-and-amber-ratings
‘Green’ papers – MN assay
Donner and others (2016) reported the results of 6 studies evaluating the potential of 3 pigment grade (PG), and 3 nanoscale (ultrafine (UF)) TiO2 forms to induce MN by analysis of micronucleated reticulocytes (MN-RETs) in rat peripheral blood cells, according to Organisation for Economic Co-operation and Development (OECD) TG474.
The 3 nanoscale (ultrafine; UF) samples contained anatase and rutile crystal structures (UF-1; 89%/11%; primary size 43 nanometres (nm)), anatase (UF-2; 100%; 42nm) or rutile (UF-3; 100%; 47nm). The pigment grade TiO2 samples were anatase (PG-1; 100%; primary size 153nm) and rutile (PG-2; 100%; 195nm and PG-3; 100%; 213nm). Secondary sizes were unclear. Particles were dispersed in deionised water by probe sonication.
The studies were conducted in 2 different laboratories on Wistar (UF-3, PG-2 and PG-3) or Sprague Dawley (SD) (UF-1, UF-2 and PG-1) rats. Five animals/sex/dose (7 in the highest dose group) received a single oral dose of 500, 1,000 or 2,000 mg/kg body weight (bw) of one of the materials. Peripheral blood was collected at 48 and 72 hours post-dosing for MN evaluation and analysis of titanium. In total, 20,000 reticulocytes (RETs) per animal were analysed. One PG and one UF material each were evaluated for potential systemic exposure/uptake from the gastrointestinal tract by analysis of TiO2 in blood and liver.
The vehicle control (sterile water) was administered as a single dose by oral gavage. The positive control (cyclophosphamide (CP); 10 mg/kg bw/day) was administered either by oral gavage (UF-1, UF-2 and PG-1) or by intraperitoneal (ip) injection (UF-3, PG-2 and PG-3). Cytotoxicity was assessed by the %RETs.
No clinical signs of toxicity were observed and there was no change in the %RETs. The negative and positive control groups both exhibited responses consistent with historical control data. The negative controls ranged from 0.06 to 0.08% and 0.07 to 0.1% RETs in males and 0.07 to 0.1% and 0.05 to 0.1% RETs in females, at 48 and 72 hours, respectively. In the PG-1 study, the positive controls showed MN induction levels 7.9-fold higher than background levels in males and 7.8-fold higher in females at 48 hours. For PG-2, positive control levels were 19.7-fold higher than background in males and 15.3-fold higher in females at 48 hours. For PG-3, positive controls were 13.4-fold higher than background in males and 14.1-fold higher in females at 48 hours. At 72 hours, the positive control group was either not evaluated (UF-2, PG-2, PG-3) or the % RETs in this group was not significantly higher than background (UF-1, UF-3, PG-1).
Following treatment, no biologically relevant increases in MN-RET frequency was observed in any TiO2-exposed group, and no biologically relevant decreases in %RETs was seen among total erythrocytes.
There were no significant increases in TiO2 in blood compared to controls (48 or 72 hour) or liver (72 hour) following exposures up to the OECD 474 guideline limit dose of 2,000 mg/kg bw TiO2, indicating that there was little or no absorption of the test material from the gastrointestinal (GI) tract into the blood circulation. The authors noted that the observed lack of genotoxic effects for all six test materials might have been due to a lack of exposure owing to the inability of the test material to migrate from the GI tract into the blood.
COM opinion
Whilst considered to be a good study, with a comprehensive assessment of primary particle characteristics, preparation of test material suspensions and secondary characteristics were not described clearly. Also, the positive control was not significant for some of the 72-hour timepoints, although concurrent positive control data for all timepoints is not mandatory. All 48-hour positive control treatments responded as expected and the vehicle control data were within expected ranges, thus demonstrating laboratory proficiency. Despite such reservations, the study was given a red, amber, green (RAG) rating of green.
The authors claim that lack of absorption from the GI tract (based on terminal toxicokinetic (TK) sampling) was the reason for the lack of genotoxicity, but TK sampling timepoints were not measured robustly enough to be considered definitive. There was a potential lack of adequate bone marrow exposure for hazard identification, however the authors tested up to the maximum recommended dose and the route of exposure was physiologically relevant for dietary exposure of TiO2 from a risk assessment perspective. Overall, the study was considered to be negative. This study was cited in the the European Food Safety Authority (EFSA) review (EFSA, 2021).
Sadiq and others (2012)
Sadiq and others (2102) conducted MN and Pig-a mutation assays to evaluate the genotoxicity of TiO2 anatase NPs in mice.
The TiO2-NPs had a narrow size distribution. The particles had a slight ellipsoidal shape, with the minor axes (smallest diameter) of 12.1 ± 3.2nm. TiO2-NPs were prepared in PBS and sonicated. NP agglomerations consisting of a few hundred NPs had a size distribution of around 130nm in the treatment solution and around 170nm in cell culture medium.
Groups of 5 male B6C3F1 mice were treated via intravenous (iv) injection for 3 consecutive days with 0.5, 5.0, and 50 mg/kg bw TiO2-NPs. Blood was sampled one day before the treatment and on day 4, and at weeks 1, 2, 4, and 6 after the beginning of the treatment. %MN-RET frequencies were measured on day 4 only.
Additional animals were treated iv with 3 daily doses of 50 mg/kg bw TiO2-NPs for the measurement of titanium levels in bone marrow, 4, 24, and 48 hours after the last treatment.
The negative control was phosphate-buffered saline (PBS) and the positive control was N-ethyl-N-nitrosourea (ENU; 140 mg/kg bw) administered once via ip. In total, 20,000 CD71-positive RETs were counted for each animal. Cytotoxicity was assessed using %RETs.
The titanium levels in bone marrow were significantly increased over the control at all three sampling times, with fold changes ranging from 12.1 to 14.2, suggesting that the TiO2-NPs reached the bone marrow, the target tissue for the genotoxicity assays.
The negative control was approximately 0.3% MN-RETs and the positive control was 8 to 9-fold higher than background.
Following treatment, no differences in %MN-RET frequencies were observed between TiO2-NP-treated and negative control animals.
The authors concluded that the 10nm TiO2-NPs tested were not clastogenic or aneugenic in the MN assay at the dose levels studied. They were, however, cytotoxic to mouse bone marrow (approximately 35% reduction in RETs at 5 mg/kg and 55% reduction at 50 mg/kg TiO2 on day 4). Thus, although TiO2-NPs can reach the mouse bone marrow and are capable of inducing cytotoxicity, they were not demonstrated to be genotoxic.
COM opinion
This study was considered to be of good quality with appropriate dosing and the study was given a RAG rating of green. Whilst the iv route was not physiologically relevant, this did ensure target organ (bone marrow) exposure and overall the study was considered to be negative. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: MN assay
Chakrabarti and others (2019)
Chakrabarti and others (2019) evaluated the cytotoxic potential of TiO2-NPs both in vitro and in vivo. The in vivo genotoxicity tests conducted included MN and chromosomal aberration (CA) assays carried out according to OECD TG 474 and 475, respectively, and a Comet assay (see section below for CA data). The Comet assay was given a RAG status of red so is not included in this review.
Scanning electron microscope (SEM) analysis of TiO2-NPs (type not given) revealed a spherical, smooth, homogenous, and uniform structure. The average particle diameter was 58.25 ± 8.11nm. No information on the secondary size was provided, nor on the method of dispersal, with the authors simply saying NPs were suspended in 500 microliters (µl) water.
Groups of 5 male and 5 female Swiss albino mice were orally exposed to either vehicle only (water) or 200 or 500 mg/kg bw/day of TiO2-NPs suspended in water, for 90 days. Lungs, heart, liver, bone marrow, and kidneys were collected at termination (day 91).
Cyclophosphamide (positive control; 40 mg/kg bw/day) was administered as a single dose by ip on day 88 and animals were sacrificed 48 hours later. The MN test was conducted on bone marrow, with 2,000 polychromatic erythrocytes (PCE) per animal being scored. Polychromatic erythrocyte/ Normochromatic erythrocyte (PCE:NCE) ratio was used to assess cytotoxicity.
No clinical signs were observed. However, on gross examination at necropsy, bleeding was observed in the abdominal and pelvic cavities of animals dosed at 500 mg/kg bw/day.
The negative control was 0.14% MNPCE and the positive control was 24-fold higher than background. The PCE:NCE ratio was similar in all groups (the positive and negative control groups and the 200 and 500 mg/kg TiO2 treatment groups).
Following treatment, %MNPCE was 4-fold higher than background at 500 mg/kg (11.33 ± 1.21 MN PCE/2000 PCE, 0.57% MN PCE) but remained unaltered at the 200 mg/kg bw dose. Cell cycle analysis by flow cytometry showed dose-related oxidative damage / apoptosis (measured as accumulation in G2/M) in liver and kidney cells.
COM opinion
The MN assay was considered appropriate, although only 2,000 PCE/animal were analysed; also there was no evidence of bone marrow exposure and no confirmation of cellular uptake. Due to these observations, the study was given a RAG rating of amber.
Overall, this study was considered to be positive. However, there was evidence of gross toxicity at the highest dose tested and the authors concluded there was evidence of dose-related oxidative damage. This study was cited in the EFSA review (EFSA, 2021)
El-Ghor and others (2014)
El-Ghor and others (2014) investigated the effects of co-administration of the free radical scavenger chlorophyllin (CHL) on the clastogenicity, genotoxicity, and mutagenicity of TiO2 in mice as determined by the MN assay and alkaline Comet assay.
The primary size of TiO2 NPs was <100nm and secondary size was 45.6 ± 12.9 when suspended in water. Small agglomerates were formed in aqueous solution. Transmission electron microscopy (TEM) indicated that most of the TiO2-NPs had polyhedral morphologies.
Groups of 5 male Swiss Webster mice were administered nano-sized TiO2 (500, 1,000, or 2,000 mg/kg bw/day) by ip for 5 consecutive days and sacrificed 24 hours after the last treatment. The MN assay was conducted on bone marrow cells. The alkaline Comet assay was performed on bone marrow, liver, and brain tissues. Biochemical evaluation of hepatic 3,4-methylenedioxyamphetamine (MDA) and glutathione (GSH), and superoxide dismutase (SOD), cationic amino acid transporter (CAT) and GPx activities was done in animals treated with 500 and 2000 mg/kg bw TiO2.
Water was used as the negative control and CP (25 mg/kg bw) as the positive control. Cytotoxicity was assessed by assessing the PCE/NCE ratio per 1,000 cells. In total, 2,000 PCEs per animal were scored to determine the number of micronucleated poly-chromatic erythrocytes (MNPCEs).
A cytotoxic effect of TiO2 was indicated by a decrease in the PCE/NCE ratio at 500, 1,000 and 2,000 mg/kg (0.71, 0.57 and 0.44, respectively), compared with the negative controls. The negative control was 0.52% and the positive control was increased 7-fold over background.
Following treatment, a significant dose-related increase in MNPCEs was seen at 500, 1,000 and 2,000 mg/kg bw TiO2 (7.8-, 9.9- and 12.4-fold increase over controls respectively).
Treatment with TiO2 significantly increased the MDA level in a dose-dependent manner compared with the negative control and induced a significant dose-dependent decrease in the GSH level, both suggestive of oxidative stress. A statistically significant decrease was also seen in SOD, CAT, and GPx levels, again in a dose-dependent manner.
The authors concluded that these results demonstrate dose-dependent clastogenicity, genotoxicity, and mutagenicity in the tested organs, with the highest damage in bone marrow cells, and argue that the observed TiO2-induced genotoxicity could be attributed to the accumulation of reactive oxidative species (ROS). Moreover, the authors considered the indirect deoxyribonucleic acid (DNA) damage via oxidative stress to be confirmed by the reported high p53 mutations, elevated MDA (marker of lipid peroxidation), and decreased antioxidant defence systems.
COM opinion
In this study it was unclear what type of TiO2 sample was used and only 2,000 PCEs per animal were scored. Also, the ip route is not recommended for this type of study and the background micronucleus frequencies were relatively high. There was also a relatively high level of cytotoxicity at the highest dose tested (56% reduction in PCE/NCE ratio relative to control). Due to these observations, the study was given a RAG rating of amber.
The level of positive response, and the fact the response was completely ablated by CHL, is unusual. Overall, this study was considered to be positive, which the authors suggested was associated with oxidative damage. This study was cited in the EFSA review (EFSA, 2021).
Relier and others (2017)
Relier and others (2017) investigated TiO2 NP-induced genotoxicity in lung overload (in which high levels of exposure cause impairment of particle clearance from the deep lung) and non-overload conditions in Sprague-Dawley (SD) rats, as measured by MN, Comet, Pig-a mutation and γ-H2AX assays (see sections below for Comet, Pig-a mutation data and γ-H2AX data.
The test material was TiO2 NPs (AEROXIDE TiO2 P25, also named NM-105 in the OECD Nanomaterial Testing Sponsorship Program). The authors confirmed NPs primary size of 25.6 ± 15nm and a secondary size of 100nm. Suspensions of TiO2 NPs were prepared by sonication and diluted in phosphate-buffered saline (PBS). Particle agglomeration was analysed by centrifugal liquid sedimentation and stability of the final suspensions (zeta potential) was measured.
Sprague-Dawley (SD) rats (12 per group) were administered 3 endotracheal instillations at 4-day intervals, so that exposure was spread over 8 days, providing total NP doses of 0.0, 0.5, 2.5 or 10 mg/kg bw. Six animals from each group were sacrificed 2 hours after the last treatment to investigate immediate NP-induced toxicity (data not shown) and the remaining 6 were sacrificed 35 days after last treatment to detect mutations induced in bone marrow erythroblasts. Lung, peripheral blood and liver were collected.
NP-induced chromosomal damage was assessed in blood with the MN test. In total, 2,000 PCEs were assessed.
The PCE/NCE ratio was also calculated. The negative control was PBS and the positive controls were methyl methanesulphonate (MMS; 150 mg/kg bw final dose administered in three gavages at a one-day interval just before sacrifice).
No change was seen in the PCE/NCE ratio. The negative control MN frequency was 0.1% and the positive control induced a 6-fold higher level of MN than background. Tested doses of TiO2 P25 NPs did not induce GSH changes in lung, blood or liver, indicating lack of oxidative stress at the time points studied.
Following TiO2 treatment, no changes in MN levels were seen after 2 hours. After 35 days, small statistically significant increases in MN were seen at 0.5, 2.5 and 10 mg/kg bw (2-, 1.8- and 1.9-fold higher than control, respectively). The authors state that such small non-dose related increases may not be considered biologically relevant.
COM opinion
The study was considered acceptable, although dosing was by intratracheal instillation which is not a physiologically relevant route of exposure, and there was no evidence of exposure of the bone marrow. Only 2,000 PCE were scored. Due to these observations, the study was given a RAG rating of amber.
Overall, the study was considered to be negative. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: CA assay
Chakrabarti and others (2019)
Chakrabarti and others (2019) evaluated the cytotoxic potential of TiO2-NPs both in vitro and in vivo. The in vivo genotoxic endpoints were estimated by means of MN and CA assays carried out according to OECD TG 474 and 475, respectively as well as the Comet assay (see above for the study methodology and for MN data). The Comet assay was given a RAG status of red so is not included in this review.
No clinical signs were observed. However, on gross examination at necropsy, bleeding was observed in the abdominal and pelvic cavities of animals dosed at 500 mg/kg bw/day. The negative control was 0.76% and the positive control was 12.7-fold higher than background.
Following treatment, there was no significant difference in the percentage of CAs observed between the 200 mg/kg bw group and the negative control, but there was a significant difference in the 500 mg/kg bw group (2.5-fold higher; 1.9%).
COM opinion
The CA assay was considered appropriate, although suitability of the repeat-dose protocol is questionable (due to potential loss of damaged mitotic cells) and positive results were seen only at the highest dose. Also, there was no reference to concurrent toxicity assessment and there was evidence of gross toxicity at the highest dose tested. It is unclear if CA frequency includes or excludes gaps; no details are provided on the types of aberrations found which makes interpretation problematic. There is also no evidence of bone marrow exposure, and cellular uptake is not confirmed. Due to these observations, the study was given a RAG rating of amber.
Overall, this study was considered to be positive, however, cell cycle analysis is suggestive of dose-related oxidative damage and/or apoptosis. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: Comet assay
Relier and others (2017)
Relier and others (2017) investigated TiO2 P25 NP-induced genotoxicity in lung overload and non-overload conditions in SD rats as measured by MN, Comet, Pig-a mutation and γ-H2AX assays (see below for sections on the study methodology, [MN data], Pig-a mutation data and γ-H2AX data).
NP-induced DNA damage was assessed in lung, peripheral blood, and liver cells using the Comet assay. A total of 150 cells per rat were assessed for tail DNA percentages.
Cytotoxicity was assessed on bronchoalveolar lavage (BAL) and serum through lactate dehydrogenase (LDH) release, inflammatory cytokine markers and haematology parameters. The negative control was PBS and the positive control was MMS (150 mg/kg bw final dose administered in three oral gavages at one-day intervals just before sacrifice). TiO2 uptake was measured in the lung and liver.
Following treatment, increases in LDH levels and macrophage accumulation were observed at 2.5 and 10 mg/kg bw. Increases in inflammatory cytokines measured in BAL and plasma, and increases in neutrophil counts, were noted at 10 mg/kg bw. The %tail intensity in the negative control was 0.3% and 0.7% in liver and blood, respectively, after 2 hours, and 0.6% and 0.4% after 35 days. The positive control was 57-, 13- and 17-fold higher in lung, liver and blood, respectively. Tested doses of TiO2 P25 NPs did not induce GSH changes in lung, blood or liver, indicating lack of oxidative stress at the time points studied.
Following treatment, the lowest tested doses had no toxicity or genotoxicity effects in the lung. In blood, no lymphocyte DNA damage, could be detected. Following treatment with higher doses, in the lung % tail intensity was increased at 2.5 and 10 mg/kg bw (3.2- and 4.7-increase compared to controls, respectively) after 35 hours. In liver, an increase in DNA damage was seen at both time points at 2.5 and 10 mg/kg bw (2 hours; 3.4- and 3.7-fold increase, respectively; 35 days; 3.8-fold increase at both doses). In blood, no DNA damage was seen after 2 hours but a slight increase was seen after 35 hours at the highest two doses (2.0- and 2.3-fold increase, respectively).
TiO2 was found above the basal level (supplied in diet) in the lung and liver. At 2 days, TiO2 was found in the lung in all dose groups in a dose-dependent manner (0.5, 2.4, and 9.2 mg TiO2 /g lung; almost 100% of administered doses) and in the liver only at the highest dose (10% of the administered dose). After 3 months, TiO2 continued to be found in the lung in a dose-dependent manner.
COM opinion
The study was considered acceptable, although dosing was by intratracheal instillation, which is not a physiologically relevant route of exposure, but there was evidence of target organ exposure in the lung and liver. The authors did not include a 24-hour sampling timepoint as recommended by OECD TG 489. Therefore, the study was given a RAG rating of amber.
Overall, the study was considered to be positive, however the observed genotoxicity was associated with cytotoxicity and inflammatory changes. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: Pig-a mutation assay
Sadiq and others (2012)
Sadiq and others (2102) conducted MN and Pig-a mutation assays to evaluate the genotoxicity of 10nm TiO2 anatase NPs in mice (see section above for the study methodology and MN data).
Groups of5 male B6C3F1 mice were treated by iv for 3 consecutive days with 0.5, 5.0, and 50 mg/kg bw TiO2-NPs. Blood was sampled one day before the treatment and on Day 4, and at weeks 1, 2, 4, and 6 after the beginning of the treatment. Pig-a mutant frequencies were determined at day −1 and weeks 1, 2, 4 and 6 in 1 x 106 RBC and 3 x 105 RET/animal.
A reduction in %RETs was observed in TiO2-NP-treated animals on day 4, suggesting treatment-related cytotoxicity with an associated rebound erythropoiesis in week 1 and normalisation by day 28. The negative control was 0 to 1.2 × 10−6 and 0 to 4 × 10−6 in red blood cell (RBC) (CD24−) and RET (CD24−). The positive control was 254 to 305-fold and 216 to 868-fold higher in RBC (CD24−) and RET (CD24−), based on a maximum ENU response in week 2 (304.80 × 10−6 mutRBC and 864.00 × 10−6</sup< mutRET).
Following treatment, no increase was observed in RBC and RET frequencies in TiO2-NP-treated animals compared with controls.
The authors concluded that the 10nm TiO2-NPs tested were not mutagenic in the Pig-a mutation assay; they were, however, cytotoxic to mouse bone marrow. Thus, although TiO2-NPs can reach the mouse bone marrow and are capable of inducing cytotoxicity, they were not demonstrated to be genotoxic.
COM opinion
This study was considered to be of acceptable quality although only 3 daily doses were used and only 5 animals per dose were used which is not consistent with OECD test guidelines. Due to these observations, the study was given a RAG rating of amber.
Overall, the study was considered to be negative. This study was cited in the EFSA review (EFSA, 2021).
Suzuki and others (2016)
Suzuki and others (2016) investigated the genotoxicity of TiO2 NP suspensions in male gpt Delta transgenic C57BL/6J mice using MN, Comet, Pig-a mutation and gpt mutation assays (see the section below for gpt mutation data). The MN and Comet assays were given a RAG status of red so are not included in this review.
Titanium dioxide, Aeroxide P25® (TiO2-P25; 20% rutile and 80% anatase) had an average particle size of 21nm, The test material was suspended by sonication in disodium phosphate (DSP) The Z-average diameter of the TiO2-P25 particles in suspension was about 150nm.
Four male gpt Delta transgenic C57BL/6J mice were administered TiO2-P25 (2, 10 or 50 mg/kg bw per week), via iv, for 4 consecutive weeks. Mice were sacrificed 9 days after the final injection and blood collected.
Disodium phosphate was used as a negative control and ENU (70 mg/kg) as the positive control (single ip dose sampled 30 days after dosing). A total of 1,000,000 TER-119-positive cells were analysed to determine the frequency of CD24-negative RBCs.
No data on cytotoxicity was presented. The RBC mutation frequency in the negative control group was 0.4 ± 0.55 x 10-6 (CD24-/CD71-/TER-119+ cells). The mutation frequency in the positive control group was 127.5 times higher than background.
Following treatment, the Pig-a mutant frequency was not significantly different in TiO2-treated groups at any dose compared with the DSP-treated control group.
COM opinion
The study protocol was generally acceptable, although group size was lower than recommended in OECD test guidelines, weekly dosing is not optimal, and there is no mention of mutant cell enrichment in the methodology (albeit the minimum number of cells as recommended in OECD TG 470 were analysed), and no cytotoxicity measurements were described. Due to these observations, the study was given a RAG rating of amber.
The vehicle and positive controls gave expected results, and despite the iv route the results are clearly negative. This study was cited in the EFSA review (EFSA, 2021).
Relier and others (2017)
Relier and others (2017) investigated TiO2 P25 NP-induced genotoxicity in lung overload and non-overload conditions in SD rats as measured by MN, Comet, Pig-a mutation and γ-H2AX assays (see sections below for the [study methodology and MN data[(#NP), Comet data and γ-H2AX data).
NP-induced Pig-a mutation was assessed in the blood of rats sacrificed only at 35 days (3 rats per dose group). The number of mutants was recorded relative to a cell population of over 108 total RBC and 106 RETs.
Cytotoxicity was assessed using %RETs. The negative control was PBS and the positive control was N-methyl-N-nitrosourea (MNU; 60 mg/kg bw), given in one ip injection 35 days before sacrifice.
Following treatment, no change in cytotoxicity was observed at any dose tested. The negative control produced 0.5 x10-6 mutant RBCs and 1.3 x10-6 mutant RETs. The positive control showed a 52-fold increase in mutant RBCs and a 12.3-fold increase in mutant RETs. Tested doses of TiO2 P25 NPs did not induce GSH changes in lung, blood or liver, indicating lack of oxidative stress at the time points studied.
Following treatment, no increase in the frequency of mutant RBC and RETs was observed.
COM opinion
The study was considered acceptable, although dosing was by intratracheal instillation which is not a physiologically relevant route of exposure, and there was no evidence of bone marrow exposure. The intermittent dosing regimen used in the study is not optimal. Due to these observations, the study was given a RAG rating of amber.
Overall, the study was considered to be negative. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: gpt and Spi- mutation assay
Suzuki and others (2016)
Suzuki and others (2016) investigated the genotoxicity of TiO2 NP suspensions in male gpt Delta transgenic C57BL/6J mice using MN, Comet, Pig-a mutation and gpt mutation assays (see section below for the study methodology and for Pig-a mutation data). The MN and Comet assays were given a RAG status of red so are not included in this review.
Four male gpt Delta transgenic C57BL/6J mice were administered once weekly iv injections of TiO2-P25 (2, 10 or 50 mg/kg bw per week) for 4 consecutive weeks. Mice were euthanized on day 9 after the final injection of TiO2-P25; portions of the middle lobe of the liver were removed and stored at −80◦C until genomic DNA isolation. High-molecular-weight genomic DNA was extracted from the liver by the standard method.
The negative control produced background mutation frequencies of 0.8x10-6 for gpt and 7.86x10-6 for Spi-. The positive control induced an 11.2-fold increase in gpt mutants and a 41.5-fold increase in Spi- mutants.
Compared with the DSP-treated controls, no significant increase in liver gpt mutation frequency (point mutations) or Spi- mutant frequency (deletion mutations) was observed in mice administered TiO2-P25 at any dose, suggesting that TiO2-P25 had no genotoxic effect.
COM opinion
The study protocol was generally acceptable, although group size was lower than recommended in OECD test guidelines, weekly dosing is not optimal, and there is no mention of cytotoxicity measurements. Due to these observations, the study was given a RAG rating of amber.
The vehicle and positive controls gave expected results, and despite the iv exposure route the results are clearly negative. This study was cited in the EFSA review (EFSA, 2021).
‘Amber’ papers: γ-H2AX assay
Relier and others (2017)
Relier and others (2017) investigated TiO2 P25 NP-induced genotoxicity in lung overload and non-overload conditions in SD rats as measured by MN, Comet, Pig-a mutation and γ-H2AX assays (see sections below for the study methodology and MN data, Comet data and Pig-a mutation data).
NP-induced DNA double-strand breaks (DSBs) were assessed in lung, blood lymphocytes, and liver cells with c-H2AX immunostaining. The number of foci per nucleus among 100 cells was counted by fluorescent microscopy and cells were grouped into 4 categories: 0 foci, 1–4 foci, and 4–10, >10 for analysis. Mean number of foci per cell per rat were also calculated.
Cytotoxicity was assessed on BAL and serum through LDH release, inflammatory cytokine markers and haematology parameters. The negative control was PBS and the positive control was MMS (150 mg/kg bw final dose administered in 3 gavages at a one-day interval just before sacrifice). TiO2 uptake was measured in the lung and liver.
Following treatment, increases in LDH levels and macrophage accumulation were observed at 2.5 and 10 mg/kg bw. Increases in inflammatory cytokines measured in BAL and plasma, and increases in neutrophil counts, were noted at 10 mg/kg bw. Tested doses of TiO2 P25 NPs did not induce GSH changes in lung, blood or liver, indicating lack of oxidative stress at the time points studied. The negative controls showed 2-3 mean foci per cell per rat in lung, blood and liver after 2 hours and 35 days. The positive control values were 2-fold higher than background.
The number of cells in each of the four categories of foci per nucleus was not different from the vehicle exposed control group, with the exception of the lung in which an increase of DSB was found immediately after exposure to the highest dose. This result was confirmed through analysis of mean foci per cell, which was 2 foci/cell for vehicle, low and mid dose, and 4 foci/cell for the high dose. These results suggest that a high dose of TiO2 NPs may induce DNA DSBs in the lung.
TiO2 was found above the basal level (supplied in diet) in the lung and liver. At 2 days, TiO2 was found in the lung in all dose groups in a dose-dependent manner (0.5, 2.4, and 9.2 mg TiO2/g lung; almost 100% of administered doses) and in the liver only at the highest dose (10% of the administered dose). After 3 months, TiO2 continued to be found in the lung in a dose-dependent manner.
COM opinion
The study was considered acceptable, although dosing was by intratracheal instillation, which is not a physiologically relevant route, but there was evidence of target organ exposure in the lung and liver. The positive control was only 2-fold higher than background. Due to these observations, the study was given a RAG rating of amber.
Overall, the study was considered to be positive in the lung 2 hours after dosing at the highest dose, which was also associated with inflammation and particle overload. This study was cited in the EFSA review (EFSA, 2021).
Summary
Six papers were identified, following screening of papers cited in the EFSA opinion (EFSA, 2021) as described in the methodology section and an assessment of the newer literature (2021 to 2023, Annexe 1), to be of sufficient quality to warrant further assessment.
Regarding the in vivo genotoxicity of TiO2, a total of 11 studies covering 5 genotoxicity assays, namely the MN (green=2; amber=3), CA (amber=1), Comet (amber=1), Pig-a (amber=3) and gpt and Spi- mutation (amber=1) assays, all of which are recognised by the OECD and other international regulatory bodies, were considered to be of sufficient quality. Data from the γ-H2AX assay (amber=1) were also included to give mechanistic information.
Studies were assessed as red, amber or green based on study design, using criteria outlined in Table 3. An overall summary of the data are presented in
Several of the papers also outlined non-regulatory experiments on the role of oxidative stress and DNA interactions which may aid insight into mechanisms of action.
Exposure to the test material in the studies was via the oral route, by iv injection, ip injection, or by endotracheal instillation.
Positive results for in vivo genotoxicity were obtained in the MN assay by both oral and ip exposure, in the CA assay after oral exposure, and in the comet assay following endotracheal instillation. All 3 iv studies were negative, but limitations in the study designs were noted.
Only 2 studies were deemed to have used robust methodology. One of these utilised the MN assay to detect MN-RET in peripheral blood cells following oral dosing. No biologically relevant increases in MN-RET frequency were observed in any TiO2 exposed group and no biologically relevant decreases in %RETs was seen. The observed lack of genotoxic effects was attributed by the authors to a lack of exposure due to the inability of the test material to migrate from the GI tract into the blood.
In the other ‘green’ study (Sadiq and others 2021), no differences in %MN-RET frequencies were observed between TiO2-NP-treated and negative control animals following iv injection. The authors concluded that the 10nm TiO2 -NPs tested were not clastogenic or aneugenic in the MN assay at the dose levels studied. They were, however, cytotoxic to mouse bone marrow. Thus, although the TiO2-NPs were demonstrated to reach the mouse bone marrow and be capable of inducing cytotoxicity, they were not shown to be genotoxic.
Results from the three oral studies are considered the most physiologically relevant for dietary exposure to TiO2. These utilised the MN assay (2 studies) and the CA assay (one study). Positive results were obtained in the CA assay, but only at the highest dose, and in one MN assay, again only at the highest dose. These positive changes were associated with cytotoxicity, oxidative damage and/or inflammation. As noted above, the only robust oral study (which utilised the MN assay) yielded negative results.
Table 5. Summary of the ‘green’ MN results
| Test material | Primary size | Concentration mg/kg bw | Species, strain or sex | Route and duration of administration | Endpoint | Result | Reference |
|---|---|---|---|---|---|---|---|
| PG and UF TiO2 |
PG 153 to 213nm UF 42-47 |
500 to 1,000 | 5 male and females Wistar or CD rats | Oral (no further details) Single dose |
MN | Neg | Donner and others (2016) |
| TiO2 NPs (anatase) | Ellipsoidal, with minor axes 12.1 ± 3.2nm Agglomerations had size distribution of c around 130 to 170nm |
0.5 to 50 | 5 male B6C3F1 mice |
iv injection 3 consecutive daily doses |
MN-RET | Neg | Sadiq and others (2012) |
Table 6. Summary of the ‘amber’ MN results
| Test material | Primary size | Concentration mg/kg bw | Species, strain or sex | Route and duration of administration | Endpoint | Result | Reference |
|---|---|---|---|---|---|---|---|
| TiO2-NPs (no further info) | 58.25 ± 8.11nm | 200 to 500 | 5 male and female Swiss albino mice | Oral (no further details) 90 days | MN | Pos (associated with gross toxicity and oxidative damage) | Chakrabarti and others (2019) |
| TiO2 NPs | <100nm | 500 to 2,000 | 5 male Swiss Webster mice | ip 5 days | MN | Pos (associated with oxidative damage and a relatively high level of cytotoxicity) | El-Ghor and others (2014) |
| AEROXIDE TiO2 P25 (NM-105) | 25.6 ± 15nm | 0.5 to 10 | 12 male SD rats | Endotracheal instillation 3 instillations over 8 days |
MN | Neg | Relier and others (2017) |
Table 7. Summary of the ‘amber’ CA results
| Test material | Primary size | Dose mg/kg bw | Species/strain/sex | Route and duration of administration | Endpoint | Result | Reference |
|---|---|---|---|---|---|---|---|
| TiO2-NPs (no further info) | 58.25 ± 8.11nm | 200 to 500 | 5 male and female Swiss albino mice | Oral (no further details) 90 days | CA | Pos (associated with gross toxicity and oxidative damage) | Chakrabarti and others (2019) |
Table 8. Summary of the ‘amber’ Comet assay results
| Test material | Primary size | Dose mg/kg bw | Species, strain or sex | Route and duration of administration | Endpoint | Result | Reference |
|---|---|---|---|---|---|---|---|
| AEROXIDE TiO2 P25 (NM-105) |
25.6 ± 15nm | 0.5 to 10 | 12 male SD rats | Endotracheal instillation 3 instillations over 8 days |
Comet | Pos (associated with cytotoxicity and inflammatory changes) | Relier and others (2017) |
Table 9. Summary of the ‘amber’ Pig-a mutation, gpt mutation and γ-H2AX assay results
| Test material | Primary size | Dose mg/kg bw | Species, strain or sex | Route and duration of administration | Endpoint | Result | Reference |
|---|---|---|---|---|---|---|---|
| TiO2 NPs (anatase) | 12.1 ± 3.2nm | 0.5 to 50 | 5 male B6C3F1 mice | iv 3 days | Pig-a mutation | Neg | Sadiq and others (2102) |
| Aeroxide P25® (TiO2-P25; 20% rutile and 80% anatase) | 21nm | 2 to 50 | 4 male gpt Delta transgenic C57BL/6J mice | iv 4 weeks | Pig-a mutation | Neg | Suzuki and others (2016) |
| AEROXIDE TiO2 P25 (NM-105) | 25.6 ± 15nm | 0.5 to 10 | 12 male SD rats | Endotracheal instillation 3 instillations over 8 days |
Pig-a mutation | Neg | Relier and others (2017) |
| Aeroxide P25® (TiO2-P25; 20% rutile and 80% anatase) | 21nm | 2 to 50 | 4 male gpt Delta transgenic C57BL/6J mice | iv 4 weeks | gpt mutation | Neg | Suzuki and others (2016) |
| AEROXIDE TiO2 P25 (NM-105) | 25.6 ± 15nm | 0.5 to 10 | 12 male SD rats | Endotracheal instillation 3 instillations over 8 days |
γ-H2AX | Pos lung (associated with cytotoxicity and inflammatory changes) | Relier and others (2017) |
Abbreviations
| Abbreviation | Meaning |
|---|---|
| BAL | bronchoalveolar lavage |
| CA | chromosomal aberrations |
| CAT | cationic amino acid transporter |
| CHL | chlorophyllin |
| COM | Committee on the Mutagenicity |
| CP | cyclophosphamide |
| DNA | deoxyribonucleic acid |
| DSBs | double-strand breaks |
| DSP | disodium phosphate |
| EFSA | European Food Safety Authority |
| ENU | N-ethyl-N-nitrosourea |
| GSH | glutathione |
| i.p. | intraperitoneal |
| i.v. | intravenous |
| LDH | lactate dehydrogenase |
| MDA | 3,4-Methylenedioxyamphetamine |
| MMS | methyl methanesulphonate |
| MN | micronucleus |
| MNPCEs | micronucleated poly-chromatic erythrocytes |
| MN-RETs | micronucleated reticulocytes |
| MNU | N-methyl-N-nitrosourea |
| NCE | normochromatic erythrocyte |
| NP | nanoparticle |
| OECD | Organisation for Economic Co-operation and Development |
| PBS | phosphate-buffered saline |
| PCE | polychromatic erythrocytes |
| PCE/NCE | polychromatic erythrocytes/normochromatic erythrocyte |
| Pig-a | phosphatidylinositol glycan class A gene |
| RAG | red, amber, green |
| RBC | red blood cell |
| RET | reticulocyte |
| ROS | reactive oxygen species |
| SD | Sprague Dawley (rats) |
| SEM | scanning electron microscope |
| SOD | superoxide dismutase |
| TEM | transmission Electron Microscopy |
| TiO2 | titanium dioxide (E171) |
| TG | test guideline |
| TK | toxicokinetic |
References
- Chakrabarti S, Goyary D, Karmaka S and Chattopadhyay P. ‘Exploration of cytotoxic and genotoxic endpoints following sub-chronic oral exposure to titanium dioxide nanoparticles’ Toxicology and Industrial Health 2019: volume 35, issue 9, pages 577 to 592
- Donner EM, Myhre A, Brown SC, Boatman R and Warheit DB. ‘In vivo micronucleus studies with 6 titanium dioxide materials (3 pigment-grade & 3 nanoscale) in orally-exposed rats’ Regulatory Toxicology and Pharmacology 2016: volume 74, pages 64 to 74
- EFSA Panel on Food Additives and Flavourings (FAF) (2021). ‘Safety assessment of titanium dioxide (E171) as a food additive’ EFSA Journal 2021: volume 19
- El-Ghor A, Noshy M, Galal A and Mohamed H (2014). ‘Normalization of nano-sized TiO2-induced clastogenicity, genotoxicity and mutagenicity by chlorophyllin administration in mice brain, liver, and bone marrow cells’ Toxicological Sciences 142
- Relier C, Dubreuil M, Lozano Garcia O, Cordelli E, Mejia J, Eleuteri P, Robidel F, Loret, T, Pacchierotti F, Lucas S, Lacroix G and Trouiller B. (2017). ‘Study of TiO2 P25 nanoparticles genotoxicity on lung, blood, and liver cells in lung overload and non-overload conditions after repeated respiratory exposure in rats’ Toxicological Sciences 2017: volume 156, issue 2, pages 527 to 537
- Sadiq R, Bhalli JA, Yan J, Woodruff RS, Pearce MG, Li Y, Mustafa T, Watanabe F, Pack LM, Biris AS, Khan QM and Chen T. ‘Genotoxicity of TiO2 anatase nanoparticles in B6C3F1 male mice evaluated using Pig-a and flow cytometric micronucleus assays’ Mutation Research/Genetic Toxicology and Environmental Mutagenesis 2012; volume 745, issue 1, pages 65 to 72
- Suzuki T, Miura N, Hojo R, Yanagiba Y, Suda M, Hasegawa T, Miyagawa, M and Wang R-S. (2016). ‘Genotoxicity assessment of intravenously injected titanium dioxide nanoparticles in gpt delta transgenic mice’ Mutation Research/Genetic Toxicology and Environmental Mutagenesis 2016: volume 802, pages 30 to 37