Iron sinter

Endpoint summary

Key value for chemical safety assessment

Genetic toxicity in vitro

Description of key information

Iron category substances have been tested in several guideline-conform GLP studies on gene mutation in bacteria, gene mutation in mammalian cells, and cytogenicity in mammalian cells. Moreover, substances of the iron category were tested in in vivo assays on DNA damage as well as clastogenicity and aneugenicity. All tests show a negative response, thus iron category substances do not require classification for mutagenic properties.

Endpoint conclusion

Endpoint conclusion:

no adverse effect observed (negative)

Genetic toxicity in vivo

Description of key information

Iron category substances have been tested in several guideline-conform GLP studies on gene mutation in bacteria, gene mutation in mammalian cells, and cytogenicity in mammalian cells. Moreover, substances of the iron category were tested in in vivo assays on DNA damage as well as clastogenicity and aneugenicity. All tests show a negative response, thus iron category substances do not require classification for mutagenic properties.

Endpoint conclusion

Endpoint conclusion:

no adverse effect observed (negative)

Additional information

Additional information

In this dossier, the endpoint germ cell mutagenicity is not addressed by substance-specific information, but instead by a weight of evidence approach based on collected information for all substances of the iron category. Three target substances covered by this read-across (Iron sinter; Iron ores, agglomerates; Mill scale) consist primarily of different iron oxides, as described in the technical dossier section 1.2 Composition. The predominant characteristic of the iron oxides is the inertness being a cause of their chemical stability and very poor reactivity. This is shown by a very low dissolution in water and artificial physiological fluids as well as a very low in vivo bioavailability after oral administration. Further information on the read-across approach is given in the report attached to IUCLID section 13.2. Metallic iron is the fourth target substance covered by the read-across. Iron is a transition-metal and is subject at its surface to passivation by the formation of a passive oxide (i.e. iron oxide) coating. In particular for iron metal and granules, the oxide layer will form a quantitatively continuous layer to envelop the entire particle irrespective of product form. In view of this, it may be assumed that human exposure is secondary to that of iron oxide and that the liberation of ionic iron shows a slower kinetics, compared with soluble iron salts.

 

Germ cell mutagenicity studies with iron category substances

Substances of the iron category were tested in various in vitro and in vivo systems for its ability to induce gene, chromosome, and genome mutations. The systems used include bacterial reverse mutation tests as well as gene mutation assays, chromosome aberration assays, and micronuclei tests in different mammalian cells. Moreover, DNA damage in the gastro-intestinal tract and peripheral blood as well as chromosome aberration and micronucleus formation in peripheral blood and bone marrow were tested in different rat strains. The tests were conducted with iron oxides, triiron phosphide, and metallic iron. For information purposes and for comparative purposes, data on soluble iron salts were included in this dossier.

 

In vitro genetic toxicity tests

Gene mutation in bacteria

Iron oxides

Herbold, B. (1982) performed a bacterial reverse mutation assay in order to evaluate the gene mutation potential of Bayferrox AC 5110 M (Fe3O4). Salmonella typhimurium tester strains TA 98, TA 100, TA 1535, and TA 1537 were exposed for 48 hours to test material concentrations of 8, 40, 200, 1000, and 5000 µg/plate both in absence and presence of induced liver S9 fraction. The assay was performed following the plate incorporation protocol (according to Ames) using quadruplicate cultures. Vehicle (water) and positive (Endoxan, trypaflavine, and 2-aminoanthracene) control cultures were run concurrently. Bayferrox AC 5110 M did not induce any cytotoxic effects in the bacterial reverse mutation test after using concentrations of up to 5000 µg/plate. At a concentration of 5000 µg/plate, precipitation occurred, so that the experiment could not be evaluated at this concentration. The evaluation of the individual test groups did not reveal any biologically relevant differences to the respective negative controls. The positive controls induced a distinct increase in the number of revertant colonies, when compared to the concurrent vehicle control. Thus, the sensitivity of the test system and the metabolic capacity of the S9 mix were demonstrated. In the bacterial reverse mutation test conducted, there was no evidence of mutagenic effects of Bayferrox AC 5110 M in analysable concentrations on the test strains used.

Herbold, B. (1999) evaluated the mutagenic potential of Bayferrox 3950 (ZnFe2O4) in a bacterial reverse mutation assay according to OECD 471 (1983) under GLP. Cultures of S. typhimurium TA 98, TA 100, TA 1535, and TA 1537 were exposed to test material suspensions at concentration levels of 8, 40, 200, 1000, and 5000 µg/plate with and without metabolic activation.  The Salmonella/microsome test, employing doses up to 5000 µg/plate, showed Bayferrox 3950 not to produce bacteriotoxic effects. Substance precipitation occurred at 5000 µg/plate. Evaluation of individual dose groups, with respect to relevant assessment parameters (dose effect, reproducibility), revealed no biologically relevant variations form the respective negative controls. The positive controls increased the mutant counts well over double those of the negative controls, and thus demonstrated the sensitivity of the test system. All validity criteria were met. The study was fully compliant with OECD 471 (1983). Therefore, Bayferrox 3950 was considered to be non-mutagenic without and with S9 mix in the bacterial reversion assay.

Bayferrox 303 T ((Fe, Mn)2O3) was evaluated for its mutagenic potential in a bacterial reverse mutation assay (Herbold, B., 2008). The test was performed according to OECD TG 471 (1997) under GLP. Salmonella tester strains TA 98, TA 100, TA 1535, TA 1537, and TA 102 were exposed to the test material at doses up to and including 5000 µg/plate both in absence and presence of metabolic activation. The initial experiment followed the plate incorporation protocol. In the repeat experiment, the pre-incubation method was used. The Salmonella/microsome test, employing doses of up to 5000 µg/plate, showed Bayferrox 303 T to produce bacteriotoxic effects, starting at 1581 µg/plate. Substance precipitation occurred at 5000 µg/plate. Evaluation of individual dose groups, with respect to relevant assessment parameters (dose effect, reproducibility) revealed no biologically relevant variations from the respective negative controls. The positive control substance induced distinct increases in the revertant colony number and thus demonstrated the activity of the metabolic activation system and the sensitivity of the test system. No indications of mutagenic effects of Bayferrox 303 T could be found at assessable doses of up to 5000 µg/plate in any of the Salmonella typhimurium strains used. All validity criteria were met. The study was fully compliant with OECD 471 (1997). Due to these results Bayferrox 303 T has to be regarded as non-mutagenic.

Hartmann, L. (2019) investigated the mutagenic potential of Iron Oxide Sicovit® Yellow 10 E172 (iron hydroxide oxide nanoparticles) in the bacterial reverse mutation test according OECD TG 471 and under GLP. S. typhimurium TA 1535, TA 1537, TA 98, TA 100 and E. coli WP2 uvrA (pKM101) were exposed using both the plate incorporation and the pre-incubation method at doses of 5, 16, 50, 160, 500, 1600 and 5000 μg/plate. No toxicity (thinning of the background lawn or a reduction in the number of revertants) was found in both experiments. Precipitation (defined for this study as an aggregation of particulates visible to the unaided eye) was observed on the test plates at concentrations of 500 μg/plate and above. Iron Oxide Sicovit® Yellow 10 E172 did not show any evidence of mutagenic activity when tested using the preincubation and plate incorporation method with or without metabolic activation. The sensitivity of the test system was demonstrated. All validity criteria were met. The study was fully valid and compliant with OECD 471 (1997).

Summary entry - gene mutation in bacteria

The references contained in this summary entry represent gene mutation studies in bacteria with very limited value for risk assessment purposes. All references do not fulfil the criteria for quality, reliability and adequacy of experimental data for the fulfilment of data requirements under REACH and hazard assessment purposes (ECHA guidance R4 in conjunction with regulation (EC) 1907/2006, Annexes VII-X). The information contained therein were included for information purposes only.

Weissleder, R. et al. (1989): Superparamagnetic iron oxide particles were tested in bacterial reverse mutation assay. The test material is insufficiently characterised, since information on primary particle size, purity, physical appearance, molecular formula, and CAS no. is missing. Material and method section lacks essential details. The results are restricted to a qualitative statement; quantitative data is missing. No information on sensitivity of the test system or validity of the assay.

Gomaa, I.O. et al. (2013):  Self-synthesised magnetite (Fe3O4) nanoparticles were tested in a bacterial reverse mutation assay. The material tested was a self-synthesised material, which was insufficiently characterised. Only four strains were used, with only three of these strains being recommended standard strains. Only three concentrations were tested and justification for the top concentration tested is missing. Information on replicate number is missing. Sensitivity of the assay is questionable in light of the responses observed for the positive controls used with strain TA 98.

 

Triiron phosphide

In this assessment of the mutagenic potential of Ferrophosphorus (Fe3P), S. typhimurium tester strains TA 1535, TA 1537, TA 98 and TA 100, and E. coli WP2 uvrA (pKM101) were exposed to Ferrophosphorus (Fe3P) suspended in water containing 0.15% agar, which was also used as a vehicle control. Two independent mutation tests were performed in the presence and absence of metabolic activation. The first test was a standard plate incorporation assay; the second included a pre-incubation stage. The bacteria cultures were exposed to Ferrophosphorus suspensions at concentrations of 5, 15, 50, 150, 500, 1500, and 5000 µg/plate. No signs of toxicity were observed towards the tester strains in either mutation test following exposure to Ferrophosphorus (Fe3P). No evidence of mutagenic activity was seen at any concentration of Ferrophosphorus (Fe3P) in either mutation test. The concurrent positive controls demonstrated the sensitivity of the assay and the metabolising activity of the liver preparations. The mean revertant colony counts for the vehicle controls were within or close to the 99% confidence limits of the current historical control range of the laboratory. The assay satisfied all validity criteria and was fully compliant with OECD TG 471 (1997). It is concluded that Ferrophosphorus (Fe3P) showed no evidence of mutagenic activity in this bacterial system under the test conditions employed.

 

Iron

Dunkel, V.C. et al. (1999) investigated the mutagenic potential of electrolytic iron (Fe) in a bacterial reverse mutation assay. Salmonella typhimurium tester strains were exposed to at least five concentration levels up to a concentration of 10 mg/plate. The S. typhimurium strain TA 98, TA 100, TA 102, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method. Moreover, TA 102 was tested in an independent experiment following the pre-incubation protocol. The experiments were performed either in absence of a metabolic activation system, with rat S9 mix, or with hamster S9 mix. Solvent and positive controls were run concurrently. All experiments were performed in triplicate cultures. According to the authors, electrolytic iron did not induce a mutagenic response in any of the strains tested. The publication presented herein shows major methodological and reporting deficiencies. The authors did not state on the vehicle used for electrolytic iron in the bacterial reverse mutation assay. The Salmonella strains TA 98, TA 100, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method only and a confirmatory experiment was not performed. Except for the top concentration, the concentration levels evaluated were not specified. The reporting of the results is restricted to a qualitative statement on the outcome of the assay; quantitative data is missing. The author provided no information on the presence of test material precipitates. Historical control data is not included. The results of the positive controls are not specified. Information on the acceptability criteria is missing. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Dunkel, V.C. et al. (1999) investigated the mutagenic potential of carbonyl iron in a bacterial reverse mutation assay. Salmonella typhimurium tester strains were exposed to at least five concentration levels up to a concentration of 10 mg/plate. The S. typhimurium strain TA 98, TA 100, TA 102, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method. Moreover, TA 102 was tested in an independent experiment following the pre-incubation protocol. The experiments were performed either in absence of a metabolic activation system, with rat S9 mix, or with hamster S9 mix. Solvent and positive controls were run concurrently. All experiments were performed in triplicate cultures. According to the authors, carbonyl iron did not induce a mutagenic response in any of the strains tested. The publication presented herein shows major methodological and reporting deficiencies. The authors did not state on the vehicle used for carbonyl iron in the bacterial reverse mutation assay. The Salmonella strains TA 98, TA 100, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method only and a confirmatory experiment was not performed. Except for the top concentration, the concentration levels evaluated were not specified. The reporting of the results is restricted to a qualitative statement on the outcome of the assay; quantitative data is missing. The author provided no information on the presence of test material precipitates. Historical control data is not included. The results of the positive controls are not specified. Information on the acceptability criteria is missing. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

 

Soluble iron salts

Dunkel, V.C. et al. (1999) investigated the mutagenic potential of ferrous sulphate heptahydrate (FeSO4x7H2O) in a bacterial reverse mutation assay. Salmonella typhimurium tester strains were exposed to at least five concentration levels up to a concentration of 10 mg Fe/plate. The S. typhimurium strain TA 98, TA 100, TA 102, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method. Moreover, TA 102 was tested in an independent experiment following the pre-incubation protocol. The tester strain TA 97a was tested according to a modified pre-incubation method. The experiments were performed either in absence of a metabolic activation system, with rat S9 mix, or with hamster S9 mix. Solvent and positive controls were run concurrently. All experiments were performed in triplicate cultures. According to the authors, ferrous sulphate did not induce a mutagenic response in any of the strains tested under the conditions tested. The publication presented herein shows major methodological and reporting deficiencies. The Salmonella strains TA 98, TA 100, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method only and a confirmatory experiment was not performed. Except for the top concentration, the concentration levels evaluated were not specified. The reporting of the results is restricted to a qualitative statement on the outcome of the assay; quantitative data is missing except for the initial retest using TA 97a. The author provided no information on the presence of test material precipitates. Historical control data is not included. The results of the positive controls are not specified. Information on the acceptability criteria is missing. The pre-incubation method applied, and the concentration used to test TA 97a are insufficiently described. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Dunkel, V.C. et al. (1999) investigated the mutagenic potential of ferric chloride hexahydrate (FeCl3x6H2O) in a bacterial reverse mutation assay. Salmonella typhimurium tester strains were exposed to at least five concentration levels up to a concentration of 10 mg Fe/plate. The S. typhimurium strains TA 98, TA 100, TA 102, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method. Moreover, TA 102 was tested in an independent experiment following the pre-incubation protocol. The experiments were performed either in absence of a metabolic activation system, with rat S9 mix, or with hamster S9 mix. Solvent and positive controls were run concurrently. All experiments were performed in triplicate cultures. According to the authors, ferric chloride did not induce a mutagenic response in any of the strains tested. The publication presented herein shows major methodological and reporting deficiencies. The authors did not state on the vehicle used for ferric chloride in the bacterial reverse mutation assay. The Salmonella strains TA 98, TA 100, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method only and a confirmatory experiment was not performed. Except for the top concentration, the concentration levels evaluated were not specified. The reporting of the results is restricted to a qualitative statement on the outcome of the assay; quantitative data is missing. The author provided no information on the presence of test material precipitates. Historical control data is not included. The results of the positive controls are not specified. Information on the acceptability criteria is missing. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Dunkel, V.C. et al. (1999) investigated the mutagenic potential of ferric phosphate dihydrate (FePO4x2H2O) in a bacterial reverse mutation assay. Salmonella typhimurium tester strains were exposed to at least five concentration levels up to a concentration of 10 mg Fe/plate. The S. typhimurium strains TA 98, TA 100, TA 102, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method. Moreover, TA 102 was tested in an independent experiment following the pre-incubation protocol. The experiments were performed either in absence of a metabolic activation system, with rat S9 mix, or with hamster S9 mix. Solvent and positive controls were run concurrently. All experiments were performed in triplicate cultures. According to the authors, ferric phosphate did not induce a mutagenic response in any of the strains tested.  The publication presented herein shows major methodological and reporting deficiencies. The authors did not state on the vehicle used for ferric phosphate in the bacterial reverse mutation assay. The Salmonella strains TA 98, TA 100, TA 1535, TA 1537, and TA 1538 were tested using the plate incorporation method only and a confirmatory experiment was not performed. Except for the top concentration, the concentration levels evaluated were not specified. The reporting of the results is restricted to a qualitative statement on the outcome of the assay; quantitative data is missing. The author provided no information on the presence of test material precipitates. Historical control data is not included. The results of the positive controls are not specified. Information on the acceptability criteria is missing. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

 

Summary - gene mutation in bacteria

The mutagenic potential of iron oxides was assessed in six different studies (RL-1: 3, RL-2: 1, RL-3 (disregarded): 2) using the bacterial reverse mutation assay. The reliable studies were performed with four different particle types, i.e. zinc ferrite (ZnFe2O4), manganese ferrite ((Fe, MN)2O3), triiron tetraoxide (Fe3O4), and iron hydroxide oxide (FeOOH). All of these iron oxide test materials proved to be exclusively and unequivocally negative in the bacterial reverse mutation assay. Therefore, iron oxides are considered non-mutagenic in the bacterial reverse mutation assay under the conditions applied.

Triiron phosphide (Fe3P) was evaluated for its gene mutation potential in a guideline-conform study (RL-1) under GLP and proved to be unequivocally negative in bacteria. Therefore, triiron phosphide is considered non-mutagenic in the bacterial reverse mutation assay under the conditions of the test.

The available studies on the mutagenic potential of metallic iron were all considered not reliable (RL-3: 2) and were disregarded for the hazard assessment. However, based on the insolubility and the inertness of these substances, a mutagenic potential is not anticipated.

None of the available studies on the gene mutation potential in bacteria with soluble iron salts is considered to be reliable (RL-3: 3). The study summaries on soluble iron salts presented above are provided for information purposes only.

 

Overall, reliable studies performed with iron category substances (different iron oxides and triiron phosphide) were all negative under the conditions tested. Thus, iron category substances are considered to be non-mutagenic in bacteria. Further details on the read-across is provided in the report attached to IUCLID section 13.2.

 

In vitro mammalian cell gene mutation

Iron oxides

Entian, G. (2008) evaluated the potential of Bayferrox 306 (Fe3O4) to induce gene mutations at the Hprt locus in mammalian cells. The test was performed according to OECD TG 476 (1997) under GLP. Bayferrox 306 was tested in the HPRT test at concentrations ranging from 6-36 µg/mL without and with S9 mix.  Under both activation conditions, no relevant cytotoxic effects were induced. However, the test material was tested up to and over the limits of solubility in the medium. Bayferrox 306 induced no biological relevant or biological statistically significant increases in the mutant frequency. The positive control EMS and DBA had a marked mutagenic effect, as was seen by a biologically relevant increase in mutant frequencies as compared to the corresponding untreated controls and thus demonstrated the sensitivity of the test system and the activity of the S9 mix. All validity criteria were met. The study was fully compliant with OECD 476 (1997). Based on these results, Bayferrox 306 is considered to be non-mutagenic in the mammalian cell gene mutation assay, both with and without metabolic activation.

Hargreaves, V. (2019) investigated the mutagenic potential of Iron Oxide Sicovit® Yellow 10 E172 (iron hydroxide oxide nanoparticles) in the in vitro mammalian cell gene mutation test according OECD TG 476 and under GLP. In the cytotoxicity range-finder experiment, six concentrations were tested in the absence and presence of S9 ranging from 63.28 to 2025 μg/mL. Based on the findings of the cytotoxicity range-finder experiment, in the mutation experiment, the cultures were exposed for 3 hours to test material suspensions at concentrations ranging from 0.1955 to 100.1 μg/mL both in the absence and presence of S9. In the main test, the lowest concentration at which precipitate was observed (by eye) at the end of the treatment incubation period in the absence and presence of S9 was retained and higher concentrations were discarded. The highest concentrations analysed were 6.255 μg/mL in the absence of S-9 and 3.128 μg/mL in the presence of S-9 which gave 83% and 109% RS, respectively. Following 3-hour treatment up to precipitating concentrations, no statistically significant increases in MF were observed following treatment with Iron Oxide Sicovit® Yellow 10 E172 at any concentration analysed in the absence and presence of S9 and there were no statistically significant linear trends. The solvent control values were within the acceptable limits in each assay. The positive controls showed distinct and significant increases in mutant frequency and demonstrated sensitivity of the test system. All validity criteria were met and the study was fully compliant with OECD TG 476 (2016). Based on the study results, it is concluded that Iron Oxide Sicovit® Yellow 10 E172 did not induce gene mutations at the hprt locus of L5178Y mouse lymphoma cells when tested up to a precipitating concentration.

 

Triiron phosphide

Ferrophosphorus (Fe3P) was tested for mutagenic potential in an in vitro mammalian cell mutation assay at the thymidine kinase locus Woods, I. (2011). The study consisted of a preliminary toxicity test and two main tests comprising three independent mutagenicity assays. The cells were exposed for either 3 hours or 24 hours in the absence of exogenous metabolic activation (S9 mix) or 3 hours in the presence of S9 mix. Mouse lymphoma L5178Y-3.7.2c cells were to test material suspensions at concentrations of 62.03, 124.06, 248.13, 496.25, 992.5, and 1985 µg/mL. Following 3-hour treatment in the absence and presence of S9 mix, there were no increases in the mean mutant frequencies of any of the test concentrations assessed that exceeded the sum of the mean concurrent vehicle control mutant frequency and the Global Evaluation Factor (GEF), within acceptable levels of toxicity. The maximum concentration assessed for mutant frequency in the 3-hour treatment in both the absence and presence of S9 mix was 1985 μg/mL. No reduction in RTG was observed in the absence or presence of S9 mix. In the 24-hour treatment, the maximum concentration assessed for mutant frequency was 1985 μg/mL. No increase in mutant frequency exceeded the sum of the mean concurrent vehicle control mutant frequency and the GEF. No reduction in RTG was observed. In all tests the concurrent vehicle and positive control were within acceptable ranges. The study satisfied all validity criteria and was fully compliant with OECD TG 476 (1997). It was concluded that Ferrophosphorus (Fe3P) did not demonstrate mutagenic potential in this in vitro cell mutation assay, under the experimental conditions.

 

Iron

Dunkel, V.C. et al. (1999) examined carbonyl iron (Fe) for its gene mutation potential in mammalian cells using the mouse lymphoma assay. Cultures of L5178Y mouse lymphoma cells were exposed for 4 hours to carbonyl iron concentration ranges of 3000 to 5000 µg/mL without S9 mix and 500 to 4000 µg/mL with S9 mix. Duplicate cultures were exposed to the test material, culture medium only, or positive control substances (MMS and DMBA). After a 2-day expression and a 10-12 selection period, the mutant frequencies were determined via an automated system.  According to the authors, carbonyl iron induced a mutagenic effect in L5178Y mouse lymphoma cells both in absence and presence of metabolic activation. Carbonyl iron did not increase relevant cytotoxicity (relative total growth >20%) in absence of metabolic activation. In presence of metabolic activation, the relative total growth was decreased in a concentration-related manner. The top concentration (4000 µg/mL) resulted in excessive cytotoxicity (6% total growth). In cultures exposed to carbonyl iron in absence of metabolic activation, the mutation frequency was a 1.4- to 2.3-fold increase above the concurrent negative control value. In presence of metabolic activation, the increase was 1.5- to 2.3-fold in cultures showing no excessive toxicity (relative total growth ≥10%). In presence of excessive toxicity, the mutant frequency observed was increased 4-fold above the negative control value. The positive controls induced distinct decreases in the mutation frequencies and were clearly mutagenic according to the evaluation criteria set out. Carbonyl iron particles were found attached to and within L5178Y mouse lymphoma cells. The publication presented herein shows major methodological and reporting deficiencies. According to the authors, carbonyl iron induced a mutagenic response based on mutant frequency found to be increased at least a twofold above the concurrent negative control value. However, in absence of excessive cytotoxicity (relative total growth <10%), the mutant frequencies were only marginally above the threshold (2.3x without S9 mix and 2.2x with S9 mix). The cultures, treated with the top concentration in presence of S9 mix, were the only cultures with clearly increased mutant frequency (4-fold). However, the response was observed in presence of excessive toxicity (6% relative total growth) and thus, this response should not be evaluated according to the OECD TG 490. Moreover, according to the current OECD test guideline (OECD TG 490, 2016), positive and negative response should be evaluated considering the Global Evaluation Factor (90x10^-6). However, none of the acceptable cultures exceeded the Global Evaluation Factor. The maximum induced mutant frequencies were 39 and 71x10^-6. Thus, according to the current test guideline, the result is not considered biologically relevant. The concentrations tested in absence of metabolic activation (3-5 mg/mL) exceeded clearly the maximum recommended concentration (2 mg/mL; OECD TG 490, 2016). In presence of metabolic activation, only three concentrations were in accordance with the current test guideline (0.5, 1, and 2 mg/mL). The authors did not state on precipitation of the test material or on pH and osmolality effects on the culture medium. Some values for the absolute cloning efficiency, relative total growth, and average number TFT resistant colonies are missing without further justification. The acceptability criteria are not specified. Historical control data is missing. The cell line was insufficiently characterised, since information on the passage number, doubling time, and karyotype stability is missing). Information on the results of the colony sizing is not provided. Based on these findings the reference is considered to be not reliable [RL-3].

Dunkel, V.C. et al. (1999) examined electrolytic iron for its gene mutation potential in mammalian cells using the mouse lymphoma assay. Cultures of L5178Y mouse lymphoma cells were exposed for 4 hours to electrolytic iron concentration ranges of 3000 to 5000 µg/mL without S9 mix and 1500 to 5000 µg/mL with S9 mix. Duplicate cultures were exposed to the test material, culture medium only, or positive control substances (MMS and DMBA). After a 2-day expression and a 10-12 selection period, the mutant frequencies were determined via an automated system. According to the authors, electrolytic iron did not induce a mutagenic effect in L5178Y mouse lymphoma cells both in absence and presence of metabolic activation. Marked cytotoxicity as indicated by a relative total growth below 20% was not evident. No particles were found attached to or within L5178Y mouse lymphoma cells. The publication presented herein shows major methodological and reporting deficiencies. The concentrations tested in absence of metabolic activation (3-5 mg/mL) exceeded clearly the maximum recommended concentration (2 mg/mL; OECD TG 490, 2016). In presence of metabolic activation, only two concentrations were in accordance with the current test guideline (1.5, and 2 mg/mL). The authors did not state on precipitation of the test material or on pH and osmolality effects on the culture medium. Some values for the absolute cloning efficiency, relative total growth, and average number TFT resistant colonies are missing without further justification. Quantitative data on the results of the colony sizing are not provided and results are restricted to a qualitative statement. The acceptability criteria are not specified. The mutant frequency (23x10^-6) of the negative control cultures without S9 mix undercut the acceptability criterion set out in the OECD TG 490 (35-140x10^-6). The positive reference mutagen, MMS, showed an induced mutation frequency of 88x10^-6, and thus, the response was clearly below the induced mutation frequency of at least 300x10^-6 above the spontaneous background (OECD TG 490, 2016). DMBA induced in presence of S9 a strong response (induced mutation frequency: 328x10^-6) but this was associated with excessive cytotoxicity (7% relative total growth). Historical control data is missing. The cell line was insufficiently characterised, since information on the passage number, doubling time, and karyotype stability is missing. The concentration or volume of S9 mix and S9 in the final culture medium is not specified. The methodologies applied to investigate the particle size distribution and particle uptake are not described. Based on these findings the reference is considered to be not reliable [RL-3].

 

Soluble iron salts

Dunkel, V.C. et al. (1999) examined ferric phosphate dihydrate (FePO4x2H2O) for its gene mutation potential in mammalian cells using the mouse lymphoma assay. Cultures of L5178Y mouse lymphoma cells were exposed for 4 hours to concentration ranges of 229 to 897 µg Fe/mL without S9 mix and 0.449 to 0.897 µg Fe/mL with S9 mix. Duplicate cultures were exposed to the test material, culture medium only, or positive control substances (MMS and DMBA). After a 2-day expression and a 10-12 selection period, the mutant frequencies were determined via an automated system. According to the authors, ferric phosphate did not induce an increase in the number mutants in absence of metabolic activation in L5178Y mouse lymphoma cells. However, in the presence of metabolic activation, there was a concentration-related mutagenic response, with a marked increase in cytotoxicity. The top concentration (0.897 µg Fe/mL) resulted in a mutant frequency, which met the twofold increase cut off criterion in presence of marked toxicity (16.5%). The positive controls induced distinct decreases in the mutation frequencies and were clearly mutagenic according to the evaluation criteria set out. The publication presented herein shows major methodological and reporting deficiencies. In the presence of metabolic activation, the mutant frequency observed ranged from 43 to 117x10^-6 and was increased 1.2- to 2.5-fold above the concurrent negative control value (44x10^-6). Thus, treatment with ferric phosphate with metabolic activation resulted in a slight but concentration-related increase in the mutant frequency. Only treatment with the top concentration (0.897 µg Fe/mL) resulted in a mutant frequency, which met the twofold increase cut off criterion. However, this positive finding was observed only in presence of marked toxicity (16.5%) and should be interpreted with caution, since the effect might be of secondary nature. Moreover, according to the current OECD test guideline (OECD TG 490, 2016), positive and negative response should be evaluated considering the Global Evaluation Factor (positive results: >90x10^-6). However, none of the acceptable cultures exceeded the Global Evaluation Factor. The maximum induced mutant frequency was 71x10^-6. Thus, according to the current test guideline, the result is not considered biologically relevant. The authors did not state on precipitation of the test material or on pH and osmolality effects on the culture medium. The acceptability criteria are not specified. The mutant frequency (29x10^-6) of the negative control culture in absence of metabolic activation undercut the acceptability criterion set out in the OECD TG 490 (35-140x10^-6). The positive control cultures did not show an induced mutant frequency of at least 300x10^-6 (135x10^-6 in absence of S9 and 238x10^-6 in presence of S9), which is required by the current test guideline. Thus, the acceptability criteria set out in OECD TG 490 are only partially met. Historical control data is missing. The cell line was insufficiently characterised, since information on the passage number, doubling time, and karyotype stability is missing. Quantitative data on the results of the colony sizing are not provided and results are restricted to a qualitative statement. The mutant frequency should have been calculated for the negative controls on the basis of four cultures. However, in presence of metabolic activation there were only three values. It is unclear whether the number of cultures was reduced or whether the result was not reported. Based on these findings the reference is considered to be not reliable [RL-3].

Dunkel, V.C. et al. (1999) examined ferric chloride hexahydrate (FeCl3x6H2O) for its gene mutation potential in mammalian cells using the mouse lymphoma assay. Cultures of L5178Y mouse lymphoma cells were exposed for 4 hours to ferric chloride concentration ranges of 309 to 1030 µg Fe/mL without S9 mix and 0.206 to 1.236 µg Fe/mL with S9 mix. Duplicate cultures were exposed to the test material, culture medium only, or positive control substances (EMS and MCA). After a 2-day expression and a 10-12 selection period, the mutant frequencies were determined via an automated system. According to the authors, ferric chloride did not induce an increase in the number mutants in absence of metabolic activation in L5178Y mouse lymphoma cells. However, in the presence of metabolic activation, there was a concentration-related mutagenic response, with a marked increase in cytotoxicity. The positive controls induced distinct decreases in the mutation frequencies and were clearly mutagenic according to the evaluation criteria set out. The publication presented herein shows major methodological and reporting deficiencies. In the presence of metabolic activation, the mutant frequency observed ranged from 45 to 121x10^-6 and was a 1.1- to 2.8-fold increase above the concurrent negative control value (44x10^-6). Thus, treatment with ferric chloride with metabolic activation resulted in a slight but concentration-related increase in the mutant frequency. Importantly, according to the current OECD test guideline (OECD TG 490, 2016), positive and negative response should be evaluated considering the Global Evaluation Factor (positive results: >90x10^-6). However, none of the acceptable cultures exceeded the Global Evaluation Factor. The maximum induced mutant frequency was 76x10^-6. Thus, according to the current test guideline, the result is not considered biologically relevant. The authors did not state on precipitation of the test material or on pH and osmolality effects on the culture medium. The acceptability criteria are not specified. The mutant frequency (26x10^-6) of the negative control culture in absence of metabolic activation undercut the acceptability criterion set out in the OECD TG 490 (35-140x10^-6). In absence of metabolic activation, several cultures undercut the minimum cloning efficiency (65-120%) requested by the test guideline. Thus, the acceptability criteria set out in OECD TG 490 are only partially met. Historical control data is missing. The cell line was insufficiently characterised, since information on the passage number, doubling time, and karyotype stability is missing. Quantitative data on the results of the colony sizing are not provided and results are restricted to a qualitative statement. The authors calculated a mutant frequency of 50x10^-6 at a ferric chloride concentration of 0.412 µg Fe/mL. However, based on the values reported, the mutant frequency should have been about 71x10^-6. Thus, it is unclear whether the values listed in the table were correctly tabulated or result of miscalculation. According to the authors, the absolute cloning efficiency was calculated on the basis of three petri dishes. However, sometimes there were one to four different values reported. It remains unclear where these values come from or why the number of petri dishes was sometimes reduced or increased. The mutant frequency should have been calculated for the negative controls on the basis of four cultures. However, in presence of metabolic activation there were only three values. It is unclear whether the number of cultures was reduced or whether the result was not reported. Based on these findings the reference is considered to be not reliable [RL-3].

Dunkel, V.C. et al. (1999) examined ferrous sulphate heptahydrate (FeSO4x7H2O) for its gene mutation potential in mammalian cells using the mouse lymphoma assay. Cultures of L5178Y mouse lymphoma cells were exposed for 4 hours to ferrous sulphate concentration ranges of 20.1 to 201 µg Fe/mL without S9 mix and 0.804 to 1.508 µg Fe/mL with S9 mix. Duplicate cultures were exposed to the test material, culture medium only, or positive control substances (EMS and DMBA). After a 2-day expression and a 10-12 selection period, the mutant frequencies were determined via an automated system. According to the authors, ferrous sulphate induced a weak mutagenic effect in L5178Y mouse lymphoma cells in absence of metabolic activation and a concentration-related increase in the mutant frequency in presence of metabolic activation. Cytotoxicity was also markedly increased with metabolic activation. The highest concentration tested with S9 mix (1.508 µg Fe/mL) induced excessive toxicity, as indicated by a relative total growth value of 5%, when compared to the negative control. In absence of metabolic activation, the top concentration (201 µg Fe/mL) resulted in marked cytotoxicity as indicated by a relative total growth of 10.5%. In cultures exposed to ferrous sulphate without S9 mix, the mutation frequency was a 1.2- to 3.2-fold increase above the concurrent negative control value. However, the highest fold-increase (3.2-fold) was observed in presence of marked toxicity. With S9 mix, the increase was 2.0- to 2.8-fold in cultures showing no excessive toxicity (relative total growth ≥10%). In presence of excessive toxicity, the mutant frequency observed was a 3.3-fold increase above the negative control value. The positive controls induced distinct decreases in the mutation frequencies and were clearly mutagenic according to the evaluation criteria set out. The publication presented herein shows major methodological and reporting deficiencies. According to the authors, ferrous sulphate induced a mutagenic response based on mutant frequency found to be an increase of at least a twofold above the concurrent negative control value. In the absence of S9 mix, only the top ferrous sulphate concentration resulted in a 3.2 fold-increase above the threshold. However, this response was associated with marked toxicity (10.5%) and should therefore be interpreted with caution, since the effect might be of secondary nature. In presence of metabolic activation, the top ferrous sulphate concentration induced excessive cytotoxicity (5% relative total growth) and should be thus not considered for the evaluation. The remaining three concentration resulted in a slight but concentration-related increase in the mutant frequency (2.0-2.8-fold). Importantly, according to the current OECD test guideline (OECD TG 490, 2016), positive and negative response should be evaluated considering the Global Evaluation Factor (90x10^-6). However, none of the acceptable cultures exceeded the Global Evaluation Factor. The maximum induced mutant frequencies were 55 and 43x10^-6. Thus, according to the current test guideline, the result is not considered biologically relevant. The authors did not state on precipitation of the test material or on pH and osmolality effects on the culture medium. The acceptability criteria are not specified. The mutant frequency (25x10^-6 and 27x10^-6) of the negative control cultures undercut the acceptability criterion set out in the OECD TG 490 (35-140x10^-6). The positive reference mutagen, DMBA, showed an induced mutation frequency of 144x10^-6, and thus, the response was clearly below the induced mutation frequency of at least 300x10^-6 above the spontaneous background (OECD TG 490, 2016). Thus, the acceptability criteria set out in OECD TG 490 are only partially met. Historical control data is missing. The cell line was insufficiently characterised, since information on the passage number, doubling time, and karyotype stability is missing. Quantitative data on the results of the colony sizing are not provided and results are restricted to a qualitative statement. The mutant frequency should have been calculated for the negative controls on the basis of four cultures. However, only two values were presented in both experiments. It is unclear whether the number of cultures was reduced or whether the result was not reported. Based on these findings the reference is considered to be not reliable [RL-3].

 

Summary - In vitro mammalian cell gene mutation

For iron oxides, two guideline-compliant GLP studies (RL-1) on the mutagenic potential in mammalian cells are available. Triiron tetraoxide (Fe3O4) was shown to be unequivocally negative in an HPRT assay using Chinese hamster lung fibroblasts. This finding is consistent with the result of an HPRT assay in mouse lymphoma cells with iron hydroxide oxide (FeOOH). Thus, the iron oxides tested are considered to be non-mutagenic in mammalian cells under the conditions of the tests.

The gene mutation potential of triiron phosphide (Fe3P) in mammalian cells was evaluated in a guideline compliant GLP study (RL-1). Triiron phosphide was shown to be unequivocally negative in the mouse lymphoma assay. Therefore, triiron phosphide is considered to be non-mutagenic in mammalian cells under the conditions of the test.

Two non-reliable (RL-3) studies exist on the gene mutation potential of metallic iron in mammalian cells. These studies, performed by the same group, showed conflicting results in the mouse lymphoma assay. Carbonyl iron (Fe) was found to be mutagenic both in absence and presence of metabolic activation. In contrast, electrolytic iron (Fe) was found to be clearly negative under identical conditions. The discrepancy between the results of these two similar test materials could not be traced back, which is also due to the reporting deficiencies of these studies. Based on the data available, no conclusion can be drawn on the mutagenic potential of metallic iron in mammalian cells.

Soluble iron salts (FePO4, FeCl3, and FeSO4) were assessed for their gene mutation potential in mammalian cells in three non-reliable studies (RL-3). The soluble salts were all tested in the mouse lymphoma assay and showed inconsistent results depending on metabolic activation state. All test materials showed positive results only in the presence of the metabolic activation system. However, the studies summarised above are provided for information purposes only.

 

Overall, reliable studies using different assay types (HPRT and MLA) and test systems consistently showed negative results for several substances of the iron category (iron oxides, triiron phosphide, and electrolytic iron). Thus, substances of the iron category are considered non-mutagenic in mammalian cells. Further details on the read-across is provided in the report attached to IUCLID section 13.2.

 

In vitro clastogenicity and aneugenicity

Chromosome aberration assays:

Iron oxides

Bayferrox 306 (Fe3O4) was tested in an in vitro chromosome aberration assay both in absence and presence of metabolic activation (Thum, M., 2008). The assay was performed according to OECD TG 473 (1997) under GLP. After 4 hours treatment of Chinese hamster V79 cells with Bayferrox 306 concentrations of 6.25, 12.5 and 25 µg/mL were used without and with S9 mix for assessment of the clastogenic potential of Bayferrox 306. In addition, cells were evaluated for chromosomal aberrations after 18 hours treatment with Bayferrox 306 concentrations of 6.25, 12.5 and 25 µg/mL without S9 mix. None of these cultures treated with Bayferrox 306 both with and without metabolic activation showed statistically significant or biologically relevant increases of numbers of metaphases with aberrations. The positive controls induced clear clastogenic effects and demonstrated the sensitivity of the test system and the activity of the S9 mix used. All validity criteria were met. The study was fully compliant with OECD 473 (1997). Based on the results of this test, Bayferrox 306 is considered to be non-clastogenic for mammalian cells in vitro.

Lloyd, M. (2019) evaluated the clastogenic and aneugenic potential of Iron Oxide Sicovit® Yellow 10 E172 (iron hydroxide oxide nanoparticles) by its effects on the frequency of micronuclei in Chinese hamster ovary (CHO) cells treated in the absence and presence of metabolic activation system. In the pulse experiments (3+21 hours), cell cultures were exposed for 3 hours to test material suspensions at concentrations of 0.5859, 1.172, 37.5, 75, and 300 µg/mL in absence of S9 and 2.344, 9.375, 37.5, 75, and 300 µg/mL in presence of S9. Afterwards, the cells were treated with cytoB and harvested after 21 hours. In the continuous treatment experiment (24+0 hours), the cells were exposed for 24 hours to test material suspensions at concentrations of 2.344, 9.375, 18.75, 37.5, and 75 µg/mL in absence of S9. The cultures were simultaneously treated with cytoB and harvest after the exposure period. In the main study, one thousand binucleate cells from each culture (2000 per concentration) were analysed for micronuclei (MNBN cells). Sterile McCoys’s 5A medium was used as vehicle. Vehicle and positive (mitomycin C, cyclophosphamide, noscapine) controls were run concurrently. Treatment of cells with Iron Oxide Sicovit® Yellow 10 E172 for 3+21 hours in the absence and presence of S-9 and for 24+0 hours in the absence of S-9 resulted in frequencies of MNBN cells that were similar to and not significantly different (at the p≤0.05 level), compared to those observed in the concurrent vehicle controls, at all concentrations analysed under each treatment condition. The MNBN cell frequencies in treated cultures fell within the normal ranges at all concentrations analysed under each treatment condition. A weak but statistically significant linear trend (p≤0.05) was observed for the 3+21-hour treatments in the absence of S-9. However, in the absence of any marked increases in MNBN cell frequencies which exceeded the normal range at any concentration analysed, this isolated observation was considered not biologically relevant. The frequency of MNBN cells in vehicle controls fell within the normal ranges. The positive control chemicals induced statistically significant increases in the proportion of MNBN cells. All validity criteria were met and the study was fully compliant with OECD TG 487 (2016). Based on the study results, it is concluded that Iron Oxide Sicovit® Yellow 10 E172 did not induce micronuclei when tested up to and including precipitating concentrations (which did not preclude accurate analysis of micronuclei) for 3+21 hours in the absence and presence of a rat liver metabolic activation system (S-9) and for 24+0 hours in the absence of S-9 under the experimental conditions described.

Koenczoel, M. et al. (2011) assessed the potential of magnetite (Fe3O4) nanoparticles to induce micronuclei in mammalian cells following OECD test guideline 487 (2010). Human alveolar epithelial-like type-II cells (A549) were pulse-treated for 24 hours with magnetite nanoparticle concentrations of 1, 10, 50, and 100 µg/cm². After treatment, the cells were cultured for further 24 hours in presence of cytochalasin B. Afterwards, the cells were fixed and stained with ethidium bromide. Three independent experiments were performed for each sample. A total of 1000 binucleated cells per sample was scored for the presence of micronuclei. The Cytokinesis-Block Proliferation Index (CBPI) was evaluated on the basis of 500 cells to assess cytotoxicity of the test material. Untreated and positive (EMS) control cultures were run concurrently. In addition, the authors examined particle uptake via TEM analysis and intracellular ROS formation using the DCFH-DA assay. According to the authors, A549 cells showed a statistically significantly increased micronucleus frequency when exposed to magnetite nanoparticles at concentrations of 10 and 100 µg/cm² (1.53% and 1.6%, respectively). Untreated cultures showed a micronucleus frequency of 1%. The positive control substance, EMS, induced a statistically significant response (1.83%). The TEM analysis revealed that most of the magnetite nanoparticles where present in membrane-bound vesicles as large aggregates (100-200 µM). Only a few small aggregates were found within the cytoplasm and a single particle was found within the nucleus. The intracellular ROS levels were found to be increased in a concentration-related manner after 24 hours but not after 6 hours. The ROS level was statistically significantly increased above the untreated control only when exposed to 100 but not 200 µg/cm² magnetite nanoparticles. The publication presented herein shows some deficiencies with regard to reporting. The effect observed in the micronucleus assay was marginal since, the strongest response observed was only 1.6-fold induced above the untreated control and the micronucleus frequency was clearly below 2%. Moreover, the response showed no statistically significant concentration-related increase. A justification for the selection of the top concentration is not provided. The description of the test material preparation lacks details, since information on the sonication power/frequency is missing. The cell line is insufficiently characterised, since information on the doubling time, modal number of chromosomes, karyostability, passage number, and mycoplasma contamination is missing. The number of independent experiments performed is conflictingly specified (3 and 4). Acceptability and evaluation criteria are not specified. Information on test material precipitation and effects on the pH and osmolality of the culture medium are not specified. Individual culture data is missing. The results of the CBPI evaluation are not tabulated but only depicted in a diagram. Historical control data is not included.

Triiron tetraoxide (Fe3O4) nanoparticles were tested, by Guichard, Y. et al. (2012), in an in vitro micronucleus assay using Syrian hamster embryo (SHE) cells. The nanoparticles were tested using three different concentrations up to 50 µg/cm² for a 24-hour exposure period. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. At least 1000 cells per slide and experiment were scored for the presence of micronuclei. The experiment was performed three times. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative increase in cell count (RICC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant micronucleus formation was observed in SHE cells exposed to triiron tetraoxide nanoparticles for 24 hours. No relevant cytotoxicity was observed as indicated by RICC values well above 80%, when compared to the negative control cultures. Triiron tetraoxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures The publication presented herein shows significant deficiencies with regard to methodology and reporting. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. Only one treatment schedule (24-hour exposure) was used. Moreover, the treatment schedule used (24 hours; ca. 1.3 cell cycles) does not correspond to the short (3-6 hours) and extended treatment schedules (1.5-2.0 cell cycles) recommended in the guideline in force at that time (OECD TG 487, 2010).  The authors did not include any information on historical control data or the validity of the assay. Thus, several assay acceptability criteria are not met (OECD TG 487, 2010). Secondary cultures of SHE cells are no standard in genotoxicity testing and not recommended by the guideline force at that time. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. An aneugenic reference mutagen was not included. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

The clastogenic and aneugenic potential of magnetite (Fe3O4) nanoparticles in two different mammalian cell lines was evaluated by Kawanishi, M. et al. (2013) using the in vitro micronucleus assay. Human lung carcinoma A549 cells were exposed for six hours to the magnetite nanoparticles at concentrations of 0.02, 0.2, 2.0, 20, and 200 µg/mL (corresponding to 0.003, 0.03, 0.34, 3.4, and 34 µg/cm²). Afterwards, the cells were cultured for 42 hours. Chinese hamster ovary (CHO) AA8 cells were treated for six hours with the test material at concentration levels of 0.2, 2.0, 20, and 200 µg/mL (corresponding to 0.03, 0.34, 3.4, and 34 µg/cm²), and subsequently, cultured for a further 20 hours. After the post-treatment culture period, the different cell cultures were trypsinised, counted, centrifuged, resuspended in hypotonic solution, and fixed. The fixed cells were stained using acridine orange and 1,000 interphase cells per concentrations were analysed for the occurrence of micronuclei by fluorescence microscopy. Vehicle control (0.05% v/v Tween 80 in saline) culture were run concurrently. Cytotoxicity was evaluated by calculation of the growth rate of these cultures. In addition, the authors examined the generation of intracellular ROS. According to the authors, the treatment with magnetite nanoparticles for six hours induced concentration dependent increases in the micronucleus frequency in A549 and CHO AA8 cells. The micronucleus frequency was in both cell lines statistically significantly increased at all concentration levels tested, when compared to the concurrent vehicle control. The maximum micronucleus frequency was 5.2% and 5.0% at the highest test material concentration level tested in A549 and CHO AA8 cells, respectively. In A549 cells, the highest magnetite nanoparticle concentration induced marked cytotoxicity, i.e. 69.2% growth inhibition. Moreover, the level of ROS was increased in both cell lines, when compared to the vehicle control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. The test material is insufficiently characterised, since information on CAS No., purity, and surface treatment is missing. The cell lines used were insufficiently characterised, since information on cell doubling time, karyotype stability, modal number of chromosomes, passage numbers, and mycoplasma contamination are missing. The scoring, acceptability, and evaluation criteria are not specified. The methodology is insufficiently described, since details on the analysis of the slides and whether the slides were coded blindly are missing. The use of replicates and number of independent experiments is not specified. Details on the sonication procedure used for the preparation of the test material is missing. The results of the cytotoxicity experiment were not reported except for the effect (inhibition of cell growth by 69.2%) of the highest test material concentration on A549 cells.  Cytotoxicity was not evaluated using the parameters recommended by the OECD TG 487 (2010) effective at that time, i.e relative increase in cell counts (RICC) or relative population doubling (RPD). The cytotoxicity parameter used is known to underestimate the extent of cytotoxicity when compared with the recommended parameters. The number of cells scored per concentration was lower than recommended (1000 vs. at least 2000 cells). Information on potential effects of the test material on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The authors did not include any information on historical control data or the validity of the assay. A positive control was not included. Thus, the proficiency of the laboratory cannot be assessed. The vehicle control cultures were set up with a 10-fold lower Tween 80 solution. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

Diiron trioxide nanoparticles were tested, by Guichard, Y. et al. (2012), in an in vitro micronucleus assay using Syrian hamster embryo (SHE) cells. Diiron trioxide nanoparticles were tested using three different concentrations up to 50 µg/cm² for a 24-hour exposure period. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. At least 1000 cells per slide and experiment were scored for the presence of micronuclei. The experiment was performed three times. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative increase in cell count (RICC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant micronucleus formation was observed in SHE cells exposed to diiron trioxide nanoparticles for 24 hours. No relevant cytotoxicity was observed as indicated by RICC values well above 80%, when compared to the negative control cultures. Diiron trioxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was statistically significantly increased only after treatment with 10 µg/cm² diiron trioxide nanoparticles for 72 hours. The publication presented herein shows significant deficiencies with regard to methodology and reporting. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. Only one treatment schedule (24-hour exposure) was used. Moreover, the treatment schedule used (24 hours; ca. 1.3 cell cycles) does not correspond to the short (3-6 hours) and extended treatment schedules (1.5-2.0 cell cycles) recommended in the guideline in force at that time (OECD TG 487, 2010).  The authors did not include any information on historical control data or the validity of the assay. Thus, several assay acceptability criteria are not met (OECD TG 487, 2010). Secondary cultures of SHE cells are no standard in genotoxicity testing and not recommended by the guideline force at that time. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. An aneugenic reference mutagen was not included. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

 

Triiron phosphide

Woods, I. (2011) performed a study to assess the ability of Ferrophosphorus (Fe3P) to induce chromosomal aberrations in human lymphocytes cultured in vitro. The study was performed according to OECD TG 473 (1997) under GLP. Human lymphocytes, in whole blood culture, were stimulated to divide by addition of phytohaemagglutinin, and exposed to the test substance both in the absence and presence of S9 mix derived from rat livers. In the first test, the cultures were exposed to for three hours to test material suspensions at concentration levels of 20.33, 794.04 and 1985.1 μg/mL without metabolic activation and 8.13, 794.04 and 1985.1 μg/mL with metabolic activation. The cells were harvested after an 18-hour recovery period. In the second test, the cultures were exposed for 21 hours to concentrations of 3.25, 794.04 and 1985.1 μg/mL in absence of S9 and for 3 hours to 8.13, 794.04 and 1985.1 μg Fe3P/mL with metabolic activation. The recovery period was 0 and 18 hours for cultures treated for 21 and 3 hours, respectively. Solvent and positive control cultures were also included. Two hours before the end of the incubation period, cell division was arrested using Colcemid, the cells harvested and slides prepared, so that metaphase cells could be examined for chromosomal damage. In both the absence and presence of S9 mix, Ferrophosphorus (Fe3P) caused no statistically significant increases in the proportion of metaphase figures containing chromosomal aberrations, at any concentration, when compared with the solvent control, in either test. No statistically significant increases in the proportion of polyploid cells were observed during metaphase analysis, in either test. All positive control compounds caused statistically significant increases in the proportion of aberrant cells, demonstrating the sensitivity of the test system and the efficacy of the S9 mix. The study satisfied all validity criteria and was fully compliant with OECD TG 473 (1997). It is concluded that the test substance Ferrophosphorus (Fe3P) has shown no evidence of causing an increase in the frequency of structural chromosome aberrations in this in vitro cytogenetic test system, under the experimental conditions described.

 

Micronucleus assays:

Iron oxides

Koenczoel, M. et al. (2011) assessed the potential of magnetite (Fe3O4) powder to induce micronuclei in mammalian cells following OECD test guideline 487 (2010). Human alveolar epithelial-like type-II cells (A549) were pulse-treated for 24 hours with magnetite particle suspensions at concentrations of 1, 10, 50, and 100 µg/cm². After treatment, the cells were cultured for further 24 hours in presence of cytochalasin B. Afterwards, the cells were fixed and stained with ethidium bromide. Three independent experiments were performed for each sample. A total of 1000 binucleated cells per sample was scored for the presence of micronuclei. The Cytokinesis-Block Proliferation Index (CBPI) was evaluated on the basis of 500 cells to assess cytotoxicity of the test material. Untreated and positive (EMS) control cultures were run concurrently. In addition, the authors examined particle uptake via TEM analysis and intracellular ROS formation using the DCFH-DA assay. According to the authors, the micronucleus frequency showed a concentration-dependent increase in A549 cells exposed to the magnetite particles for 24 hours. Cell cultures treated with the top concentration showed a statistically significantly increased micronucleus frequency (1.9%), when compared to the negative control cultures (0.6%-1%). The uptake analysis revealed magnetite aggregates within membrane-bound vesicles. Some cells seem to have been destroyed by particles overload. The intracellular ROS levels were found to be increased in a concentration-related manner after 24 hours but not after 6 hours. The ROS level was statistically significantly increased above the untreated control only when exposed to 100 and 200 µg/cm² magnetite particle suspensions. The publication presented herein shows some deficiencies with regard to reporting. In the TEM analysis, some cells appear to have been destroyed by particles overload, which could be indicative of excessively high test concentrations, which in turn might have interfered with the assay readout. A justification for the selection of the top concentration is not provided. The authors did not state on the mathematical method used to analysed concentration-response dependency. The description of the test material preparation lacks details, since information on the sonication power/frequency is missing. The cell line is insufficiently characterised, since information on the doubling time, modal number of chromosomes, karyostability, passage number, and mycoplasma contamination is missing. The number of independent experiments performed is conflictingly specified (3 and 4). Acceptability and evaluation criteria are not specified. Information on test material precipitation and effects on the pH and osmolality of the culture medium are not specified. Individual culture data is missing. The results of the CBPI evaluation are not tabulated but only depicted in a diagram. Historical control data is not included. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

Triiron tetraoxide (Fe3O4) nanoparticles were tested, by Guichard, Y. et al. (2012), in an in vitro micronucleus assay using Syrian hamster embryo (SHE) cells. Triiron tetraoxide nanoparticles were tested using three different concentrations up to 50 µg/cm² for a 24-hour exposure period. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. At least 1000 cells per slide and experiment were scored for the presence of micronuclei. The experiment was performed three times. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative increase in cell count (RICC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant micronucleus formation was observed in SHE cells exposed to triiron tetraoxide nanoparticles for 24 hours. No relevant cytotoxicity was observed as indicated by RICC values well above 80%, when compared to the negative control cultures. Triiron tetraoxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. Only one treatment schedule (24-hour exposure) was used. Moreover, the treatment schedule used (24 hours; ca. 1.3 cell cycles) does not correspond to the short (3-6 hours) and extended treatment schedules (1.5-2.0 cell cycles) recommended in the guideline in force at that time (OECD TG 487, 2010).  The authors did not include any information on historical control data or the validity of the assay. Thus, several assay acceptability criteria are not met (OECD TG 487, 2010). Secondary cultures of SHE cells are no standard in genotoxicity testing and not recommended by the guideline force at that time. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. An aneugenic reference mutagen was not included. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Diiron trioxide (Fe2O3) particles were tested, by Guichard, Y. et al. (2012), in an in vitro micronucleus assay using Syrian hamster embryo (SHE) cells. Diiron trioxide particles were tested using three different concentrations up to 50 µg/cm² for a 24-hour exposure period. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. At least 1000 cells per slide and experiment were scored for the presence of micronuclei. The experiment was performed three times. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative increase in cell count (RICC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant micronucleus formation was observed in SHE cells exposed to diiron trioxide particles for 24 hours. No relevant cytotoxicity was observed as indicated by RICC values well above 80%, when compared to the negative control cultures. Diiron trioxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. Only one treatment schedule (24-hour exposure) was used. Moreover, the treatment schedule used (24 hours; ca. 1.3 cell cycles) does not correspond to the short (3-6 hours) and extended treatment schedules (1.5-2.0 cell cycles) recommended in the guideline in force at that time (OECD TG 487, 2010). The authors did not include any information on historical control data or the validity of the assay. Thus, several assay acceptability criteria are not met (OECD TG 487, 2010). Secondary cultures of SHE cells are no standard in genotoxicity testing and not recommended by the guideline force at that time. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. An aneugenic reference mutagen was not included. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

 

Summary - In vitro clastogenicity and aneugenicity

Three reliable studies (RL-1: 2, RL-2: 1, RL-3 (disregarded): 5) on the cytogenic potential of iron oxides in mammalian cells are available. A guideline-compliant GLP chromosome aberration study (RL-1) showed unequivocally negative results for triiron tetraoxide (Fe3O4), when tested in Chinese hamster lung fibroblasts. Moreover, a guideline-conform study (RL-1) on the clastogenic and aneugenic potential of iron hydroxide oxide (FeOOH) showed exclusively negative results in Chinese hamster ovary cells. Conflicting data were presented by one publication (RL-2) which tested nano and non-nano triiron tetraoxide (Fe3O4) in a micronucleus assay using A549 cells. The tests showed an ambiguous, characterised by a marginal and non-concentration dependent increase of the micronucleus frequency, and a positive response for nano and non-nano triiron tetraoxide, respectively.

A guideline-compliant GLP study (RL-1) assessed the clastogenic potential of triiron phosphide (Fe3P) via a chromosome aberration assay in human peripheral blood lymphocytes. The study proved triiron phosphide to be unequivocally negative in the chromosome aberration assay. Thus, triiron phosphide is considered to be non-clastogenic in the chromosome aberration assay under the conditions of the test.

Reliable information on the clastogenic and aneugenic potential of metallic iron is not available.

 

Overall, all guideline-conform GLP studies (RL-1) evaluating cytogenicity showed consistently negative results for the iron category substances in different assay types (CA and MN) and different test systems both in absence and presence of metabolic activation. These findings are further substantiated by exclusively negative data from reliable in vivo cytogenicity studies on diiron trioxide and a comet assay with triiron tetraoxide. Conflicting data, i.e. one ambiguous and one positive result, were presented only by one largely guideline-compliant study, which was not performed under GLP and which showed a lower reliability (RL-2). Based on the weight of evidence, the substances from the iron category are considered to be non-clastogenic and non-aneugenic in mammalian cells. Further details on the read-across is provided in the report attached to IUCLID section 13.2.

 

In vitro DNA damage

Iron oxides

The DNA damaging potential of α-Fe2O3 particles in human lung (BEAS-2B and IMR-90 cells) cells was evaluated by Bhattacharya, K. (2012). The alkaline comet assay was performed in both cell lines using test material concentrations of 10, 25, 50, and 250 µg/mL. The cells were exposed for 24 hours. Afterwards, the cells were lysed and run under an electrophoretic field. Afterwards, the cells were neutralised, and the DNA damage was quantified by determination of the Olive tail moment. Negative (untreated cells) and positive control (N-ethyl-N-nitrosourea) cultures were run concurrently. The Trypan blue dye exclusion test was performed in order to analyse cell viability and cell death. Potential uptake of the test material was analysed via TEM. Moreover, the authors investigated on ROS generation and potential effects on the mitochondrial membrane potential. According to the authors, the results show clearly that cyto- and genotoxicity caused by α-Fe2O3 particles can only be observed at very high concentrations of ≥ 50 µg/mL. At occupationally relevant concentrations of < 10 µg/mL, no cyto- and genotoxic effects can be observed in both IMR-90 and BEAS-2B cells. The cell viability was concentration-dependently decreased in both cell lines and the proportion of dead BEAS-2B cells was statistically significantly increased at all concentrations tested. Electron microscopic analyses showed the particles to be present in the cells in close proximity to the mitochondria. However, the particles were found neither within the nucleus nor mitochondria. Furthermore, the articles induced intracellular ROS generation and an increase in the mitochondrial membrane potential. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. The cell lines used, i.e. IMR-90 and BEAS-2B are no standard in genotoxicity testing. Furthermore, the cell lines were insufficiently characterised, since information on the passage number, doubling times, mycoplasma contamination, and modal number of chromosomes are missing. The cytotoxicity was evaluated in a separate experiment using the colorimetric Trypan blue exclusion assay, which often underestimates cytotoxicity due false-negative evaluations (slight influx of dye) and the stage of cytotoxicity measured. Furthermore, the scoring of the cells is insufficiently described. The comet assay methodology is only described for the BEAS-2B cells. The authors provided no information on the light conditions during the comet assay. The number of comets analysed is not specified. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The DNA damage was assessed using the Olive tail moment only, whereas the assessment of the tail intensity is recommended by the corresponding in vivo test guideline (OECD, 2016). The response, indicated by the Olive tail moment, appears to be not concentration-related in both strains tested. Moreover, the Olive tail moments were statistically significantly increased only at concentrations showing statistically significantly increased cytotoxicity. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Guichard, Y. et al. (2012) evaluated diiron trioxide (Fe2O3) particles for their DNA damaging potential in Syrian hamster embryo (SHE) cells using the alkaline comet assay. Syrian hamster embryo cell cultures were exposed to diiron trioxide particles for 24 hours at concentrations of 10, 25, and 50 µg/cm². After treatment, the cells were lysed, incubated in alkaline solution, run under an electrophoretic field, and neutralised. The mean proportion of DNA in tail was determined for 100 cells per concentration. The experiment was performed three times under exclusion from direct light exposure. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative cell count (RCC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant DNA damage was found in SHE cells exposed to diiron trioxide particles for 24 hours. The RCC was at all concentrations tested in the comet assay above 90%, and thus, did not indicate relevant cytotoxicity. Diiron trioxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. Secondary cultures of SHE cells are no standard in genotoxicity testing. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Guichard, Y. et al. (2012) evaluated triiron tetraoxide (Fe3O4) particles for their DNA damaging potential in Syrian hamster embryo (SHE) cells using the alkaline comet assay. SHE cell cultures were exposed to triiron tetraoxide particles for 24 hours at concentrations of 10, 25, and 50 µg/cm². After treatment, the cells were lysed, incubated in alkaline solution, run under an electrophoretic field, and neutralised. The mean proportion of DNA in tail was determined for 100 cells per concentration. The experiment was performed three times under exclusion from direct light exposure. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative cell count (RCC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant DNA damage was found in SHE cells exposed to triiron tetraoxide particles for 24 hours. The RCC was at all concentrations tested in the comet assay above 90%, and thus, did not indicate relevant cytotoxicity. Triiron tetraoxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. Secondary cultures of SHE cells are no standard in genotoxicity testing. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Koenczoel, M. et al. (2011) evaluated the DNA damaging potential of magnetite (Fe3O4) particles using the in vitro comet assay. Human alveolar epithelial-like type-II cells (A549) were exposed for 4 hours to magnetite particle suspensions at concentrations of 1, 10, 50, and 100 µg/cm². Untreated and positive (quartz Min-U-Sil) control cultures were run concurrently. A total of 100 randomly chosen nucleoids were scored for the Olive tail moment and tail intensity. Each experiment was performed three times independently. In separate experiments, the authors evaluated the cytotoxicity of the material using the colorimetric WST-1 and nuclear red uptake assays. In addition, the authors examined particle uptake via TEM analysis and intracellular ROS formation using the DCFH-DA assay. According to the authors, the magnetite particles induced a concentration-dependent increase in Olive tail moment and tail intensity. The Olive tail moment and tail intensity were statistically significantly increased at magnetite particle concentrations of 50 µg/cm², when compared to untreated control cultures. In addition, the Olive tail moment was statistically significantly increased at a magnetite particle concentration of 1 µg/cm². The positive control cultures showed a statistically significant increase in both DNA damage parameters. The cytotoxicity assays did not reveal statistically significant cytotoxicity. The uptake analysis revealed magnetite aggregates within membrane-bound vesicles. Some cells seem to have been destroyed by particles overload. The intracellular ROS levels were found to be increased in a concentration-related manner after 24 hours but not after 6 hours. The ROS level was statistically significantly increased above the untreated control only when exposed to 100 and 200 µg/cm² magnetite particle suspensions. The publication presented herein shows significant deficiencies with regard to methodology and reporting. There is currently no validated test guideline for the in vitro comet assay available. The proficiency of the laboratory is not assessable. The laboratory’s proficiency in preparing and analysing comets’ characteristics in comet assays is a prerequisite for reliable data. It was shown that there is huge variability in tail DNA intensities even within negative or positive controls (Tavares, A.M. et al. (2014); Nanogenotox Programme). The authors did not state on the mathematical method used to analysed concentration-response dependency. The description of the test material preparation lacks details, since information on the sonication power/frequency is missing. The cell line is insufficiently characterised, since information on the passage number, and mycoplasma contamination is missing. Scoring, acceptability, and evaluation criteria are not specified. The methodology is only poorly described, since essential information (e.g. lysis, electrophoresis, neutralisation) is missing and is restricted to citations of single references. Information on test material precipitation and effects on the pH and osmolality of the culture medium are not specified. Cytotoxicity was evaluated only in separate experiments with an exposition duration different from that used in the comet assay (4 vs. 24 hours). Individual culture data is missing. Historical control data is missing. The occurrence of hedgehogs is not specified. Based on these findings the reference is considered to be not reliable [RL-3] and is therefore disregarded for the hazard assessment.

Guichard, Y. et al. (2012) evaluated triiron tetraoxide (Fe3O4) nanoparticles for their DNA damaging potential in Syrian hamster embryo (SHE) cells using the alkaline comet assay. SHE cell cultures were exposed to triiron tetraoxide nanoparticles for 24 hours at concentrations of 10, 25, and 50 µg/cm². After treatment, the cells were lysed, incubated in alkaline solution, run under an electrophoretic field, and neutralised. The mean proportion of DNA in tail was determined for 100 cells per concentration. The experiment was performed three times under exclusion from direct light exposure. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative cell count (RCC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant DNA damage was found in SHE cells exposed to triiron tetraoxide nanoparticles for 24 hours. The RCC was at all concentrations tested in the comet assay above 90%, and thus, did not indicate relevant cytotoxicity. Triiron tetraoxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was not statistically significantly different from negative control cultures. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. Secondary cultures of SHE cells are no standard in genotoxicity testing. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. The authors provided no information whether the slides were coded blindly. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

Guichard, Y. et al. (2012) evaluated diiron trioxide nanoparticles for their DNA damaging potential in Syrian hamster embryo (SHE) cells using the alkaline comet assay. SHE cell cultures were exposed to diiron trioxide nanoparticles for 24 hours at concentrations of 10, 25, and 50 µg/cm². After treatment, the cells were lysed, incubated in alkaline solution, run under an electrophoretic field, and neutralised. The mean proportion of DNA in tail was determined for 100 cells per concentration. The experiment was performed three times under exclusion from direct light exposure. Untreated and positive (methyl methanesulfonate) control cultures were run concurrently. In a separate experiment, the cytotoxicity of the test material was assessed by determining the relative cell count (RCC). Moreover, the authors investigated cellular uptake of the test material and intracellular ROS generation. According to the authors, no significant DNA damage was found in SHE cells exposed to diiron trioxide nanoparticles for 24 hours. The RCC was at all concentrations tested in the comet assay above 90%, and thus, did not indicate relevant cytotoxicity. Diiron trioxide was found within in the cells in the form of individual particles and agglomerates. The intracellular ROS content was statistically significantly increased only after treatment with 10 µg/cm² diiron trioxide nanoparticles for 72 hours. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. Secondary cultures of SHE cells are no standard in genotoxicity testing. Furthermore, the cells were insufficiently characterised, since information on karyotype stability and modal number of chromosomes are missing. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The selection of the top concentration is not based on cytotoxic or precipitating (at least not stated) concentrations and a justification for the selection is missing. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

Bhattacharya, K. (2012) assessed the DNA damaging potential of α-Fe2O3 nanoparticles in human lung cells (BEAS-2B and IMR-90 cells) using the alkaline comet assay. The cell cultures were exposed for 24 hours to test material concentrations of 10, 25, 50, and 250 µg/mL. Afterwards, the cells were lysed and run under an electrophoretic field. Afterwards, the cells were neutralised, and the DNA damage was quantified by determination of the Olive tail moment. Negative (untreated cells) and positive control (N-ethyl-N-nitrosourea) cultures were run concurrently. The Trypan blue dye exclusion test was performed in order to analyse cell viability and cell death. Potential uptake of the test material was analysed via TEM. Moreover, the authors investigated on intracellular ROS generation and potential effects on the mitochondrial membrane potential. According to the authors, the results show clearly that cyto- and genotoxicity caused by α-Fe2O3 nanoparticles can only be observed at very high concentrations of ≥ 50 µg/mL. At occupationally relevant concentrations of < 10 µg/mL, no cyto- and genotoxic effects can be observed in both IMR-90 and BEAS-2B cells. Reductions of cell viability at all the exposure concentrations measured were found to be statistically significant when compared to the negative controls. The uptake analysis revealed that the nanoparticles were found in human lung cells as large agglomerates. However, the particles were found neither within the nucleus nor mitochondria. Furthermore, the nanoparticles induced intracellular ROS generation and an increase in the mitochondrial membrane potential. The publication presented herein shows significant deficiencies with regard to methodology and reporting. For the in vitro comet assay is currently no validated test guideline available. Moreover, the assay shows huge intra- and interlaboratory variations with regard to the responses observed. Thus, the demonstration of the proficiency of the laboratory is crucial. However, the authors did not include any information on historical control data or the validity of the assay. The cell lines used, i.e. IMR-90 and BEAS-2B are no standard in genotoxicity testing. Furthermore, the cell lines were insufficiently characterised, since information on the passage number, doubling times, mycoplasma contamination, and modal number of chromosomes are missing. The cytotoxicity was evaluated in a separate experiment using the colorimetric Trypan blue exclusion assay, which often underestimates cytotoxicity due false-negative evaluations (slight influx of dye) and the stage of cytotoxicity measured. Furthermore, the scoring of the cells is insufficiently described. The comet assay methodology is only described for the BEAS-2B cells. The authors provided no information on the light conditions during the comet assay. The number of comets analysed is not specified. The occurrence of hedgehogs is not reported. The scoring, evaluation, and acceptability criteria are not specified. Information on effects on the pH and osmolality of the culture medium and the occurrence of precipitates is not reported. The DNA damage was assessed using the Olive tail moment only, whereas the assessment of the tail intensity is recommended by the corresponding in vivo test guideline (OECD, 2016). The response, indicated by the Olive tail moment, appears to be not concentration-related in both strains tested. Moreover, the Olive tail moments were statistically significantly increased only at concentrations showing statistically significantly increased cytotoxicity. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

Koenczoel, M. et al. (2011) evaluated the DNA damaging potential of magnetite (Fe3O4) nanoparticles using the in vitro comet assay. Human alveolar epithelial-like type-II cells (A549) were exposed for 4 hours to magnetite nanoparticle suspensions at concentrations of 1, 10, 50, and 100 µg/cm². Untreated and positive (quartz Min-U-Sil) control cultures were run concurrently. A total of 100 randomly chosen nucleoids were scored for the Olive tail moment and tail intensity. Each experiment was performed three times independently. In separate experiments, the authors evaluated the cytotoxicity of the material using the colorimetric WST-1 and nuclear red uptake assays. In addition, the authors examined particle uptake via TEM analysis and intracellular ROS formation using the DCFH-DA assay. According to the authors, the magnetite nanoparticles induced a concentration-dependent increase in Olive tail moment and tail intensity. The Olive tail moment was statistically significantly increased at magnetite nanoparticle concentrations of 10 µg/cm² and above, while the tail intensity was statistically significantly at 50 µg/cm² and above, when compared to untreated control cultures. The positive control cultures showed a statistically significant increase in both DNA damage parameters. The cytotoxicity assays did not reveal statistically significant cytotoxicity. The TEM analysis revealed that most of the magnetite nanoparticles where present in membrane-bound vesicles as large aggregates (100-200 µM). Only a few small aggregates were found within the cytoplasm and a single particle was found within the nucleus. The intracellular ROS levels were found to be increased in a concentration-related manner after 24 hours but not after 6 hours. The ROS level was statistically significantly increased above the untreated control only when exposed to 100 but not 200 µg/cm² magnetite nanoparticles. The publication presented herein shows significant deficiencies with regard to methodology and reporting. There is currently no validated test guideline for the in vitro comet assay available. The proficiency of the laboratory is not assessable. The laboratory’s proficiency in preparing and analysing comets’ characteristics in comet assays is a prerequisite for reliable data. It was shown that there is huge variability in tail DNA intensities even within negative or positive controls (Tavares, A.M. et al. (2014); Nanogenotox Programme). The authors did not state on the mathematical method used to analysed concentration-response dependency. The description of the test material preparation lacks details, since information on the sonication power/frequency is missing. The cell line is insufficiently characterised, since information on the passage number, and mycoplasma contamination is missing. Scoring, acceptability, and evaluation criteria are not specified. The methodology is only poorly described, since essential information (e.g. lysis, electrophoresis, neutralisation) is missing and is restricted to citations of single references. The authors did not state on potential nanoparticle remnants during lysis and electrophoresis. Information on test material precipitation and effects on the pH and osmolality of the culture medium are not specified. Cytotoxicity was evaluated only in separate experiments with an exposition duration different from that used in the comet assay (4 vs. 24 hours). Individual culture data is missing. Historical control data is missing. The occurrence of hedgehogs is not specified. Based on these findings the reference is considered to be not reliable [RL-3] and therefore disregarded for the hazard assessment.

 

Summary entry - In vitro DNA damage

The references contained in this summary entry represent in vitro DNA damage experiments in mammalian cells (Comet assay) with very limited value for risk assessment purposes. All references do not fulfil the criteria for quality, reliability and adequacy of experimental data for the fulfilment of data requirements under REACH and hazard assessment purposes (ECHA guidance R4 in conjunction with regulation (EC) 1907/2006, Annexes VII-X). The information contained therein were included for information purposes only.

Protocols employed for the conduct of the in vitro comet assay do not meet minimal quality standards for conduct of the comet assay. Application of the comet assay to intact cells and tissues must carefully control for natural process that can produce DNA fragmentation and false positive assay outcomes. Cytotoxicity (Henderson et al., 2008; Fairbairn et al., 1996), apoptosis (Choucroun et al., 2001; Fairbairn et al., 1996), oxidative and temperature stress and terminal differentiation must all be carefully assessed for their impact upon assay outcomes. Further, the studies do not control for the presence of residual particles during processing and electrophoresis of the cells – presence of particles during these steps might also lead to false positive results. Information can also be obtained from detailed evaluation of the shape of Comet tails and the distribution of DNA within the tail (Lee et al., 2003). Little meaningful information is provided on any of these accepted response parameters. The studies summarised below controlled for none of these sources of artefactual false positives. Interpretation of the relevance of both positive and negative results from this test system cited above is therefore unclear and was not used for the current assessment of in vitro genotoxicity.

Karlsson, H.L. et al. (2009): The DNA damage was evaluated after diiron trioxide (different Fe2O3 particle types) treatment using the alkaline comet assay both in absence and presence of formamidopyrimidine DNA-glycosylase. The proportion of DNA in tail was evaluated for 70 cells (35 per replicate) per sample in each experiment. Each experiment was repeated four times independently. Only two concentration levels were tested which impedes robust evaluation of concentration-response relationship. The methodology is poorly documented and is largely limited to citing of two references. Reporting of results is incomplete. The cytotoxicity was evaluated in a separate experiment under different conditions compared to the genotoxicity experiment.

Magdolenova, Z. et al. (2012): The DNA damage was evaluated after magnetite (different Fe3O4 particle types) treatment using the alkaline comet assay. The proportion of DNA in tail was scored for 100 nucleoids. The experiment was performed in duplicates. The purity of the test material is not specified. The cells were exposed to the test material during lysis and electrophoresis only. The cytotoxicity was not evaluated. The concentrations were given as masses only (20 and 40 µg). Only two concentrations were tested.

Kawanishi, M. et al. (2013): The frequency of DNA double strand breaks was evaluated using the γ-H2AX assay. The formation of γ-H2AX, after magnetite (Fe3O4 nanoparticles) treatment, was analysed in at least 100 cells per concentration using fluorescence microscopy. The test material is insufficiently characterised, since information on CAS No., purity, and surface treatment is missing. No validated test guideline is available for the γ-H2A assay. Only two concentration levels were tested. Information on the cytotoxicity of the test material is not provided. Only 100 cells per concentration were scored.

 

 

Summary - In vitro DNA damage

The references reporting in vitro DNA damage experiments in mammalian cells (Comet assay) are of very limited value for hazard assessment purposes. All references do not fulfil the criteria for quality, reliability and adequacy in accordance with ECHA guidance R4. The test material used in the references was largely characterised insufficiently. Protocols employed for the conduct of the in vitro comet assay do not meet minimal quality standards for conduct of the assay. Application of the comet assay to intact cells and tissues must carefully control for natural process that can produce DNA fragmentation and false positive assay outcomes. Cytotoxicity (Henderson et al., 2008; Fairbairn et al., 1996)*, apoptosis (Choucroun et al., 2001; Fairbairn et al., 1996)*, oxidative and temperature stress and terminal differentiation must all be carefully assessed for their impact upon assay outcomes. Further, the studies do not control for the presence of residual particles during processing and electrophoresis of the cells – presence of particles during these steps might also lead to false positive results. Information can also be obtained from detailed evaluation of the shape of Comet tails and the distribution of DNA within the tail (Lee et al., 2003)*. Little meaningful information is provided on any of these accepted response parameters. The studies summarised above controlled for none of these sources of artefactual false positives. Interpretation of the relevance of both positive and negative results from this test system cited above is therefore unclear and was not used for the current assessment of in vitro genotoxicity of iron category substances. Thus, the studies summarised above are provided for information purposes only.

 

Overall conclusion - in vitro studies

Iron category substances are not expected to be genotoxic, since the in vitro studies on iron category substances were consistently negative in guideline-compliant GLP bacterial reversion assays, gene mutation assay, and cytogenicity assays. The results are predominantly obtained by data on different iron oxides and triiron phosphide. No reliable studies on metallic iron were identified.

Based on the weight of evidence, it is concluded that iron category substances are not mutagenic in vitro.

 

In vivo genetic toxicity tests

In vivo clastogenicity and aneugenicity

Chromosome aberration assay:

Iron oxides

Singh, S.P. et al. (2013) examined the clastogenic potential of diiron trioxide (Fe2O3) in a chromosome aberration assay using bone marrow cells from female Wistar rats. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide particle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. Bone marrow cells from the femur and tibia were sampled 18 and 24 hours after the dosing. The bone marrow cells obtain were centrifuged, fixed, mounted ion microscope slides, dried, and stained with Giemsa. Three slides were prepared for each animal and 100 well spread metaphases per animal were scored for the presence of chromosomal aberrations. The mitotic index (MI) was determined as a measure of cytotoxicity in at least 1000 cells per animal. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. Female Wistar rats treated orally with diiron trioxide (Fe2O3-bulk) particles up to a dose of 2000 mg/kg bw (limit dose according to OECD 475) did not show marked or statistically significant increases in the frequency of bone marrow metaphases with chromosomal aberrations both after 18 and 24 hours post exposure. According to the authors, the responses observed did not show a dose dependency and were well within the normal control ranges. Moreover, neither the frequency of clastogenic effects nor polyploidy was statistically significantly different compared to the vehicle control group. No relevant effect on the mitotic index was observed in animals exposed to diiron trioxide (Fe2O3-bulk) particles both after 18 and 24 hours. In contrast, female rats treated intraperitoneally with cyclophosphamide showed marked and statistically significant increases in the frequency of cells with chromosomal aberration. Moreover, the incidence of polyploid metaphase was statistically significantly increased compared to the vehicle control group. Thus, the sensitivity of the test system was demonstrated. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in the bone marrow cells after 24 and 48 hours in the high-dose group. Moreover, the liver showed a statistically significantly increase Fe level after 24 and 48 hours at dose levels of 2000, 1000 mg/kg bw and 2000 mg/kg bw, respectively. In spleen statistically significant Fe distribution was found only after 72 hours in animals of the high dose group. The brain, whole blood, kidney, and heart did not show statistically significantly elevated Fe levels independent of the experimental condition. The test material was largely excreted via the faeces. Based on the data presented, exposure of the target organ was demonstrated. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The authors did not provided information whether the animals were treated with a metaphase-arresting agent prior to sacrifice. The description of the materials and methods lacks some depth (e.g. incubation times and whether slides were independently coded). The number of cells scored for the proportion of polychromatic erythrocytes in the bone marrow assay is not specified. Information on scoring is restricted to reference to the OECD 475 (1997). The acceptability, and evaluation criteria are not specified. Historical control data is missing. The group mean body weights were not reported neither before nor after treatment. Individual results are missing. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. A justification for the use of females only is missing. The authors stated on the occurrence of aneuploidy the assay. However, the assay is not adequate to conclude on the aneugenic potential of test materials. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

Singh, S.P. et al. (2013) examined the clastogenic potential of diiron trioxide (Fe2O3) nanoparticles in a chromosome aberration assay using bone marrow cells from female Wistar rats. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide particle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. Bone marrow cells from the femur and tibia were sampled 18 and 24 hours after the dosing. The bone marrow cells obtain were centrifuged, fixed, mounted ion microscope slides, dried, and stained with Giemsa. Three slides were prepared for each animals and 100 well spread metaphases per animal were scored for the presence of chromosomal aberrations. The mitotic index (MI) was determined as a measure of cytotoxicity in at least 1000 cells per animal. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. Female Wistar rats treated orally with diiron trioxide (Fe2O3-30 nm) nanoparticles up to a dose of 2000 mg/kg bw (limit dose according to OECD 475) did not show marked or statistically significant increases in the frequency of bone marrow metaphases with chromosomal aberrations both after 18 and 24 hours post exposure. According to the authors, the responses observed did not show a dose dependency and were well within the normal control ranges. Moreover, neither the frequency of clastogenic effects nor polyploidy was statistically significantly different compared to the vehicle control group. No relevant effect on the mitotic index was observed in animals exposed to diiron trioxide (Fe2O3-30 nm) nanoparticles both after 18 and 24 hours. In contrast, female rats treated intraperitoneally with cyclophosphamide showed marked and statistically significant increases in the frequency of cells with chromosomal aberration. Moreover, the incidence of polyploid metaphase was statistically significantly increased compared to the vehicle control group. Thus, the sensitivity of the test system was demonstrated. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in whole blood, liver, heart, kidneys, bone marrow and spleen at all time points examined in animals treated with the highest diiron trioxide (Fe2O3-30 nm) dose. In contrast, the brain Fe level was not statistically significantly different from control animals. The Fe was predominantly excreted via the faeces, but the Fe concentration was statistically significantly elevated in both urine and faeces at all time points examined. Based on the data presented, exposure of the target organ was demonstrated. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The authors did not provided information whether the animals were treated with a metaphase-arresting agent prior to sacrifice. The description of the materials and methods lacks some depth (e.g. incubation times and whether slides were independently coded). The number of cells scored for the proportion of polychromatic erythrocytes in the bone marrow assay is not specified. Information on scoring is restricted to reference to the OECD 475 (1997). The acceptability, and evaluation criteria are not specified. Historical control data is missing. The group mean body weights were not reported neither before nor after treatment. Individual results are missing. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. A justification for the use of females only is missing. The authors stated on the occurrence of aneuploidy the assay. However, the assay is not adequate to conclude on the aneugenic potential of test materials. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

 

Micronucleus assay:

Iron oxides

The clastogenic and aneugenic potential of diiron trioxide (Fe2O3) was evaluated by Sing, S.P. et al. (2013) using the mammalian erythrocyte micronucleus test. The authors performed the micronucleus assay according to OECD 474 (1997) using polychromatic erythrocytes from both peripheral blood and bone marrow from female Wistar rats. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide particle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. For the peripheral blood micronucleus test, blood samples were obtained after 48 and 72 hours after treatment. Subsequently, cells smears were prepared, air dried, fixed, and stained using acridine orange. A total of 2000 polychromatic erythrocytes were scored for the presence of micronuclei. Moreover, cytotoxicity of the test material was determined by calculation of the proportion of immature erythrocytes among total erythrocytes based on 1000 cells. For the bone marrow micronucleus test, bone marrow cells were harvested from femurs obtained 24 and 48 hours after dosing. The cells were spread on slides, air-dried, fixed, and stained with Giemsa. A total of 2000 polychromatic erythrocytes were randomly selected and scored for the presence of micronuclei. The proportion of polychromatic erythrocytes among total erythrocytes was calculated for evaluation of cytotoxicity. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. Animals orally dosed with diiron trioxide (Fe2O3-bulk) particle suspension did not show a relevant shift in the proportion of polychromatic erythrocytes among total erythrocytes, when compared to the vehicle control group, neither in the peripheral blood nor in the bone marrow micronucleus assay. Thus, the test material did not induce obvious cytotoxicity. Female rats treated with diiron trioxide (Fe2O3-bulk) doses of up to 2000 mg/kg bw (limit dose according to OECD 474) showed micronucleus frequencies in polychromatic erythrocytes from both peripheral blood and bone marrow which were not distinctly or statistically significantly different (≤1.3-fold difference) from the values observed in the vehicle control group. Moreover, the micronucleus frequencies observed are well within ranges normally observed in control animals. In contrast, the positive control groups induced distinct (15 to 17-fold higher) and statistically significant increases in the micronucleus frequency compared to the vehicle control group. Thus, the sensitivity of the test system was demonstrated. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in the bone marrow cells after 24 and 48 hours in the high-dose group. Moreover, the liver showed a statistically significantly increase Fe level after 24 and 48 hours at dose levels of 2000, 1000 mg/kg bw and 2000 mg/kg bw, respectively. In spleen statistically significant Fe distribution was found only after 72 hours in animals of the high dose group. The brain, whole blood, kidney, and heart did not show statistically significantly elevated Fe levels independent of the experimental condition. The test material was largely excreted via the faeces. Based on the data presented, exposure of the target organ was demonstrated. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The description of the materials and methods lacks some depth. The number of cells scored for the proportion of polychromatic erythrocytes in the bone marrow assay is not specified. The scoring, acceptability, and evaluation criteria are not specified. Historical control data is missing. The group mean body weights were not reported neither before nor after treatment. Individual results are missing. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. A justification for the use of females only is missing. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

The clastogenic and aneugenic potential of diiron trioxide (Fe2O3) nanoparticles was evaluated by Sing, S.P. et al. (2013) using the mammalian erythrocyte micronucleus test. The authors performed the micronucleus assay according to OECD 474 (1997) using polychromatic erythrocytes from both peripheral blood and bone marrow from female Wistar rats in two separate experiments. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide nanoparticle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. For the peripheral blood micronucleus test, blood samples were obtained 48 and 72 hours after treatment. Subsequently, cell smears were prepared, air-dried, fixed, and stained using acridine orange. A total of 2000 polychromatic erythrocytes were scored for the presence of micronuclei via fluorescence microscopy. Moreover, cytotoxicity of the test material was determined by calculation of the proportion of immature erythrocytes among total erythrocytes based on evaluation of 1000 cells. For the bone marrow micronucleus test, bone marrow cells were harvested from femurs obtained 24 and 48 hours after dosing. The cells were spread on slides, air-dried, fixed, and stained with Giemsa. A total of 2000 polychromatic erythrocytes were randomly selected and scored for the presence of micronuclei by microscopical analysis. The proportion of polychromatic erythrocytes among total erythrocytes was calculated for evaluation of cytotoxicity. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. Animals orally dosed with diiron trioxide (Fe2O3-30 nm) nanoparticle suspension did not show a relevant shift in the proportion of polychromatic erythrocytes among total erythrocytes, when compared to the vehicle control group, neither in the peripheral blood nor in the bone marrow micronucleus assay. Thus, the test material did not induce obvious cytotoxicity. Female rats treated with diiron trioxide (Fe2O3-30 nm) doses of up to 2000 mg/kg bw (limit dose according to OECD 474) showed micronucleus frequencies in polychromatic erythrocytes from both peripheral blood and bone marrow which were not distinctly or statistically significantly different (≤1.7-fold difference) from the values observed in the vehicle control group. Moreover, the micronucleus frequencies observed in control and treatment groups are well within ranges normally observed in control animals. In contrast, the positive control groups induced distinct (15 to 17-fold higher) and statistically significant increases in the micronucleus frequency compared to the vehicle control group. Thus, the sensitivity of the test system was demonstrated. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in whole blood, liver, heart, kidneys, bone marrow and spleen at all time points examined in animals treated with the highest diiron trioxide (Fe2O3-30 nm) dose. In contrast, the brain Fe level was not statistically significantly different from control animals. The Fe was predominantly excreted via the faeces, but the Fe concentration was statistically significantly elevated in both urine and faeces at all time points examined. Based on the data presented, exposure of the target organ was demonstrated. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The description of the materials and methods lacks some depth (e.g. incubation times and concentration of staining solutions). The number of cells scored for the proportion of polychromatic erythrocytes in the bone marrow assay is not specified. The scoring, acceptability, and evaluation criteria are not specified. Historical control data is missing. The group mean body weights were not reported neither before nor after treatment. Individual results are missing. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. A justification for the use of females only is missing. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

 

Summary entry - In vivo clastogenicity and aneugenicity

The references contained in this summary entry represent in vivo clastogenicity/aneugenicity experiments with very limited value for risk assessment purposes. All references do not fulfil the criteria for quality, reliability and adequacy of experimental data for the fulfilment of data requirements under REACH and hazard assessment purposes (ECHA guidance R4 in conjunction with regulation (EC) 1907/2006, Annexes VII-X). The information contained therein were included for information purposes only.

Song, M.-F. et al. (2012): The clastogenic and aneugenic potential of diiron trioxide (Fe2O3)and triiron tetraoxide (Fe3O4) nanoparticles after intraperitoneal injection in female ICR mice was examined using the micronucleus assay. Intraperitoneal injection is considered to be a non-physiological exposure route and is considered to be of limited value for the risk assessment purposes. Only two dose levels were tested which impedes robust evaluation of dose-response relationship. Information on signs toxicity is not provided. The proportion of immature erythrocytes among total erythrocytes is not specified. The number of erythrocytes per animal scored for cytogenetic damage was low (n=1000).

 

 

Summary - In vivo clastogenicity and aneugenicity

Nano and non-nano diiron trioxide (Fe2O3) was evaluated for its cytogenetic potential in four reliable (RL-2) in vivo assays. For both particle types, a chromosome aberration assay in bone marrow erythrocytes and a micronucleus assay using peripheral blood and bone marrow erythrocytes were performed using female Wistar rats. Nano and non-nano diiron trioxide was proved unequivocally negative in all the tests conducted. Thus, nano and non-nano diiron trioxide do not cause local or systemic clastogenic or aneugenic events in rats tested up to the limit dose.

Reliable studies on the cytogenetic potential of other iron category substances do not exist.

 

Overall, the available data showed consistently negative results for two different iron oxide particle types tested in two different assay types (MN and CA) with erythrocytes from both bone marrow and peripheral blood. Thus, iron category substances are considered to be non-clastogenic and non-aneugenic in vivo based on the results obtained for iron oxide and based on the negative results from in vitro guideline-conform GLP studies on the cytogenicty of triiron tetraoxide and iron hydroxide oxide. Further details on the read-across is provided in the report attached to IUCLID section 13.2.

 

In vivo DNA damage

Iron oxides

Ferroxide Black 86 (Fe3O4) was tested for its potential to induce DNA strand breaks in the stomach and duodenum of treated Sprague Dawley rats (Keig-Shevlin, Z., 2020). The study was performed according to OECD TG 489 (2016) under GLP. Male rats received two administrations at 0 (Day 1) and 21 hours (Day 2). The test material was suspended in hydroxypropyl methylcellulose (medium viscosity) 0.5% (w/v) and administered at doses of 500, 1000, and 2000 mg/kg/day. Negative control animals received the vehicle only. Animals exposed to 150 mg/kg ethyl methanesulfonate via a single gavage served as positive control. The stomach and duodenum were sampled on day 2, equivalent to 24 hours. No clinical chemistry changes considered an effect of Ferroxide Black 86 were recorded. On macroscopic examination, dark contents were noted in the stomach, jejunum, ileum, cecum, and colon of animals administered 2000 mg/kg/day; the stomach, jejunum, and ileum of animals administered 1000 mg/kg/day; and the stomach of one animal administered 500 mg/kg/day. On microscopic examination, in the stomach and duodenum, dark material (which was considered to be test article), was noted in the lumen of all animals administered Ferroxide Black 86. There was no dose-related increase in %hedgehogs in stomach and duodenum, thus demonstrating that treatment with Ferroxide Black 86 did not cause excessive DNA damage that could have interfered with comet analysis. In the stomach, animals treated with Ferroxide Black 86 at 500 and 1000 mg/kg/day exhibited group mean and individual animal tail intensity and tail moment values that were similar to the concurrent vehicle control group and which fell within the laboratory’s historical vehicle control 95% reference range. At 2000 mg/kg/day, there was a statistically significant increase (P≤0.05) in group mean tail intensity which also contributed to a statistically significant linear trend. The increase was primarily due to two animals within the group showing elevated tail intensity values (R0305 TI of 9.58 and R0306 TI of 11.23) which were close to or exceeded the historical vehicle control observed maximum tail intensity of 10.69. Although there were no corresponding pathology findings to suggest target tissue toxicity or inflammation, the increases were concomitant with some small increases in %hedgehogs (highly damaged cells). Given the known challenges of working with nanoparticles on site of contact tissues and that additional technical steps were included in order to ensure the tissues were visually free of the particles at the time of tissue processing, it is likely that these increases in tail intensity were due to either mechanical damage due to over processing of these tissues or artifacts due to residual particulates remaining within the tissue (the histopathology data demonstrated residual particles present within the tissue) rather than a true genotoxic effect and therefore the biological relevance is considered to be unlikely. In the duodenum, animals treated with Ferroxide Black 86 at all doses exhibited group mean and individual animal tail intensity and tail moment values that were similar to the concurrent vehicle control group and all tail intensity values fell within the laboratory’s historical vehicle control 95% reference range. There were no statistically significant increases in tail intensity for any of the groups receiving the test article, compared to the concurrent vehicle control group and no evidence of a dose-response. The vehicle control data were within the laboratory’s historical vehicle control data ranges. The positive control induced statistically significant increases in tail intensity in the stomach and duodenum (over the current vehicle control group) that were comparable with the laboratory’s historical positive control data. The assay was therefore accepted as valid. The study was fully compliant with OECD TG 489 (2016). It is concluded that Ferroxide Black 86 did not induce DNA strand breaks in the stomach in male animals when tested up to 2000 mg/kg/day (the regulatory maximum dose level). In the same animals, a statistically significant increase in tail intensity in the duodenum was observed at 2000 mg/kg/day. Although there were no corresponding pathology findings to suggest target tissue toxicity or inflammation, the increases were concomitant with some small increases in %hedgehogs. Given the known challenges of working with nanoparticles on site of contact tissues, it is likely that these increases in tail intensity were due to either mechanical damage due to over processing of these tissues or artifacts due to residual particulates remaining within the tissue rather than a true genotoxic effect and therefore the biological relevance is considered to be unlikely.

Singh, S.P. et al. (2013) investigated the DNA damaging potential of diiron trioxide (Fe2O3) particles in peripheral blood leucocytes of female Wistar rats via the conventional comet assay. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide particle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. Peripheral blood samples from each treatment group were obtained 6, 24, 48, and 72 hours after the dosing. Afterwards, the cells were mounted on slides, lysed, and treated with alkaline solution. The nucleoids were subjected to electrophoresis, neutralised, and the DNA was stained using ethidium bromide. The proportion of DNA in tail was determined for a total of 150 comets, i.e. 50 per replicate slide, per animal using fluorescence microscopy in combination with a comet image analysis system. The trypan blue exclusion assay was performed in order to evaluate potential cytotoxic effects of the test material. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. According to the authors, peripheral blood leukocytes from female rats treated with diiron trioxide showed a cell viability of >90%, as determined via the trypan blue exclusion assay. Female Wistar rats dosed orally with diiron trioxide suspensions up to 2000 mg/kg bw (recommended maximum dose acc. to OECD 489 (2016)) did not show a statistically significant increase in the proportion of DNA in tail after 6, 24, 48, and 72 hours, when compared to the vehicle control group. The values obtained were comparable the control group and the highest increase observed was a 1.2-fold induction of the proportion of DNA in tail. Cyclophosphamide administered via intraperitoneal injection induced distinct (3.7 to 11.5-fold) and statistically significantly increases in the proportion of DNA in tail, when compared to the vehicle control group. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in the bone marrow cells after 24 and 48 hours in the high-dose group. Moreover, the liver showed a statistically significantly increase Fe level after 24 and 48 hours at dose levels of 2000, 1000 mg/kg bw and 2000 mg/kg bw, respectively. In spleen statistically significant Fe distribution was found only after 72 hours in animals of the high dose group. The brain, whole blood, kidney, and heart did not show statistically significantly elevated Fe levels independent of the experimental condition. The test material was largely excreted via the faeces. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The methodology of the cytotoxicity assay is not described, and the results are not tabulated and restricted to one general statement in the main text. The information given on the number of slides prepared/cells analysed is conflicting (150 cells in one part of main text/figure legend vs. 100 cells in second part of the main text). It is unclear whether the authors calculated the mean of the medians from the slide, as recommended by OECD 489 (2016) or the mean of the means. The scoring, acceptability, and evaluation criteria are not specified. The occurrence of hedgehogs is not reported. Historical control data is missing. The body weights of the groups were not reported neither before nor after treatment. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

Red Ferroxide 212P (Fe2O3 nanoparticles) was tested for its potential to induce DNA strand breaks in the stomach and duodenum of treated Sprague Dawley rats (Keig-Shevlin, Z., 2020). The study was performed according to OECD TG 489 (2016) under GLP. Male rats received two administrations at 0 (Day 1) and 21 hours (Day 2). The test material was suspended in hydroxypropyl methylcellulose (medium viscosity) 0.5% (w/v) and administered at doses of 500, 1000, and 2000 mg/kg/day. Negative control animals received the vehicle only. Animals exposed to 200 mg/kg ethylmethanesulfonate via a single gavage served as positive control. The stomach and duodenum were sampled on day 2, equivalent to 24 hours. No clinical chemistry changes considered an effect of Red Ferroxide 212P treatment were recorded. On macroscopic examination, there were no changes which were considered related to Red Ferroxide 212P. On microscopic examination, dark, finely granular material was noted along the mucosal surface of the lumen of the stomach and duodenum which was considered to be test article. However, there were no microscopic changes which were considered to be related to treatment with Red Ferroxide 212P. There were no marked, dose-related increases in %hedgehogs in the stomach or duodenum, thus demonstrating that treatment with Red Ferroxide 212P did not cause excessive DNA damage that could have interfered with comet analysis. In the stomach and duodenum, animals treated with Red Ferroxide 212P at all doses exhibited group mean and individual animal tail intensity and tail moment values that were similar to the concurrent vehicle control group and all tail intensity values fell within the laboratory’s historical vehicle control 95% reference range. There were no statistically significant increases in tail intensity for any of the groups receiving the test article, compared to the concurrent vehicle control group and no evidence of a dose-response. It is concluded that Red Ferroxide 212P did not induce DNA strand breaks in the stomach or duodenum in male animals when tested up to 2000 mg/kg/day (the regulatory maximum dose level). The vehicle control data were within the laboratory’s historical vehicle control data ranges. The positive control induced statistically significant increases in tail intensity in the stomach and duodenum (over the current vehicle control group) that were comparable with the laboratory’s historical positive control data. The assay was therefore accepted as valid. The study was fully compliant with OECD TG 489 (2016).

Singh, S.P. et al. (2013) investigated the DNA damaging potential of diiron trioxide (Fe2O3) nanoparticles in peripheral blood leucocytes of female Wistar rats via the conventional comet assay. Based on the results of a preceding toxicity study, five female rats per group received diiron trioxide nanoparticle suspension at doses of 500, 1000, and 2000 kg/bw day via a single gavage. Vehicle control animals were dosed with MilliQ water. Animals treated with cyclophosphamide (40 mg/kg bw) via intraperitoneal injection served as positive control. Peripheral blood samples from each treatment group were obtained 6, 24, 48, and 72 hours after the dosing. Afterwards, the cells were mounted on slides, lysed, and treated with alkaline solution. The nucleoids were subjected to electrophoresis, neutralised, and the DNA was stained using ethidium bromide. The proportion of DNA in tail was determined for a total of 150 comets, i.e. 50 per replicate slide, per animal using fluorescence microscopy in combination with a comet image analysis system. The trypan blue exclusion assay was performed in order to evaluate potential cytotoxic effects of the test material. In addition, the biodistribution of iron (Fe) was analysed. Urine and faeces samples were collected and pooled at 0-6, 6-24, 24-48, and 48-72 hours after the dosing. Moreover, the Fe levels in the whole blood, brain, liver, kidneys, heart, spleen and bone marrow were examined after 6, 24, 48, and 72 hours. The Fe content in the samples was determined by using atomic absorption spectrophotometry. According to the authors, peripheral blood leukocytes from female rats treated with diiron trioxide nanoparticles showed a cell viability of >90%, as determined via the trypan blue exclusion assay. Female Wistar rats dosed orally with diiron trioxide nanoparticle suspension up to 2000 mg/kg bw (recommended maximum dose acc. to OECD 489 (2016)) did not show a statistically significant increase in the proportion of DNA in tail after 6, 24, 48, and 72 hours, when compared to the vehicle control group. The values obtained were comparable the control group and the highest increase observed was a 1.2-fold induction of the proportion of DNA in tail. Cyclophosphamide administered via intraperitoneal injection induced distinct (3.7 to 11.5-fold) and statistically significantly increases in the proportion of DNA in tail, when compared to the vehicle control group. The biodistribution analysis revealed statistically significantly increased iron (Fe) levels in whole blood, liver, heart, kidneys, bone marrow and spleen at all time points examined in animals treated with the highest diiron trioxide (Fe2O3-30 nm) dose. In contrast, the brain Fe level was not statistically significantly different from control animals. The Fe was predominantly excreted via the faeces, but the Fe concentration was statistically significantly elevated in both urine and faeces at all time points examined. The publication presented herein showed deficiencies with regard to reporting. The description of the preparation of the test material suspension lacks details, since information on sonication frequency/energy and temperature as well as the concentration of the dosing solution are missing. The methodology of the cytotoxicity assay is not described, and the results are not tabulated and restricted to one general statement in the main text. The information given on the number of slides prepared/cells analysed is conflicting (150 cells in one part of main text/figure legend vs. 100 cells in second part of the main text). It is unclear whether the authors calculated the mean of the medians from the slide, as recommended by OECD 489 (2016) or the mean of the means. The scoring, acceptability, and evaluation criteria are not specified. The occurrence of hedgehogs is not reported. Historical control data is missing. The body weights of the groups were not reported neither before nor after treatment. The weight range of the female Wistar rats at study initiation (80-120 g) appears to be unusual low. Based on these findings the reference is considered to be reliable with restrictions [RL-2].

 

Summary entry - In vivo DNA damage

The references contained in this summary entry represent in vivo experiments on genetic toxicity with very limited value for risk assessment purposes. All references do not fulfil the criteria for quality, reliability and adequacy of experimental data for the fulfilment of data requirements under REACH and hazard assessment purposes (ECHA guidance R4 in conjunction with regulation (EC) 1907/2006, Annexes VIIX). The information contained therein were included for information purposes only.

Garry, S. et al. (2003): The DNA damaging potential of diiron trioxide (Fe2O3) particles administered endotracheally in male Sprague Dawley rats was investigated using the alkaline comet assay. Endotracheal administration is considered to be a non-physiological exposure route and is considered to be of limited value for the risk assessment purposes. Only one dose level was tested which precludes evaluation of dose-response relationship. The selection of the concentration tested was not justified and does not correspond to maximum doses recommended by current in vivo genotoxicity test guidelines. The number of groups per animal was low (n=3). Exposure of target organs was not demonstrated.

Song, M.-F. et al. (2012): The accumulation of 8-hydroxy-2’-deoxyguanosine in liver cells of female ICR mice was measured, via HPLC-ECD, after a single intraperitoneal injection of diiron trioxide (Fe2O3) and triiron tetraoxide (Fe3O4) nanoparticles. Intraperitoneal injection is considered to be a non-physiological exposure route and is considered to be of limited value for the risk assessment purposes. Only one dose level was tested which precludes evaluation of dose-response relationship. There is currently no validated guideline for the measurement of 8-OH-dG in DNA available.

 

Summary - In vivo DNA damage

Four in vivo DNA damage assays with iron oxides were considered to be reliable for hazard assessment purposes (RL-1: 2, RL-2: 2, and RL-3 (disregarded): 2). Diiron trioxide (Fe2O3) and triiron tetraoxide (Fe3O4) were evaluated for their site-of-contact genotoxicity after oral administration. Both studies were guideline-compliant GLP studies (RL-1), which showed unequivocally negative results in the comet assay, when examining cells of the stomach and duodenum of male Sprague-Dawley rats. Moreover, the DNA damaging capacity of nano and non-nano diiron trioxide (Fe2O3) was evaluated in peripheral blood erythrocytes of female Wistar rats using the comet assay. Exclusively negative results were obtained in both studies (RL-2) It is therefore concluded that iron oxides do not induce local or systemic DNA damage in animals tested up to the limit dose.

Reliable studies on the DNA damaging potential of other iron oxide category substances are not available.

 

Overall, consistently negative results have been reported on the in vivo DNA damaging potential of different iron oxides in different rat strains and cell types. Thus, iron category substances are considered to have no genotoxic potential with regard to primary DNA damage. Further details on the read-across is provided in the report attached to IUCLID section 13.2.

 

In vivo gene mutation

No data on in vivo gene mutation available.

 

Overall conclusion - in vivo studies

Iron category substances are not expected to be genotoxic or mutagenic, since the in vivo studies on iron category substances were consistently negative in reliable cytogenicity and comet assays. The results are obtained by data on different iron oxide particle types. No reliable studies on metallic iron were identified.

Based on the weight of evidence, it is concluded that iron category substances are not mutagenic in vivo.

 

*References:

Choucroun, P., Gillet, D., Dorange, G., Sawicki, B. and Dewitte, J.D. (2001). Comet assay and early apoptosis. Mutat Res 478: 89-96.

Fairbairn, D.W., Walburger, D.K., Fairbairn, J.J. and O’Neill, KL. (1996). Key morphologic changes and DNA strand breaks in human lymphoid cells: discriminating apoptosis from necrosis. Scanning 18:407-416.

Henderson, L., Wolfreys, A., Fedyk, J., Bourner, C. and Windebank, S. (1998). The ability of the Comet assay to discriminate between genotoxins and cytotoxins. Mutagenesis 13:89-94.

Lee, M., Kwon, J. and Chung M-K (2003). Enhanced prediction of potential rodent carcinogenicity by utilizing Comet assay and apoptotic assay in combination. Mutat Res 541:9-19.

Justification for classification or non-classification

The weight of evidence indicated no potential for iron category substances to induce gene, chromosome, and genome mutations in vitro under the conditions of the tests performed. The evidence is further substantiated by negative in vivo data on primary DNA damage and cytogenicity of different iron oxides.

The classification criteria acc. to regulation (EC) 1272/2008 and its subsequent amendments as germ cell mutagen are not met, thus no classification is required.