Registration Dossier

Administrative data

Key value for chemical safety assessment

Additional information

Introductory statement on read-across

The Precious Metals and Rhenium Consortium (PMRC) has coordinated the REACH registrations for a number of silver substances, i. e. silver metal (including nano forms), disilver oxide, silver nitrate, silver chloride, silver bromide and silver iodide. The data used for the toxicological hazard assessment are not restricted to studies referring to only these individual substances as test items. Instead, a more generic approach for the toxicological assessment of “silver” in general is adopted by also including information/data generated with other inorganic silver substances such as silver acetate, silver sulfate or silver carbonate. The basic assumption for this is that the systemic toxicity of any of these substances is driven by the silver ion (Ag+), which is considered the primary relevant species of silver for the hazard assessment. Further details on the read-across principle are provided in chapter 5 of the CSR.

 

In vitro studies

Bacterial mutagenicity

Silver nanoparticles did not induce gene mutations in an assay with four Salmonella typhimurium (TA1535, TA 1537, TA 95, TA 100) and one E. coli strain when using the pre-incubation method (Kim, J. S. et al., 2012). No statistical increase in revertants was observed in triplicate plating up to concentrations causing cytotoxicity (-S9: 0, 0.977, 1.953, 3.906, 7.813, 15.625, 31.25, 62.5 and 125 µg/plate; +S9: 0, 7.813, 15.625, 31.25, 62.5, 125, 250 and 500 µg/plate) with and without metabolic activation.

The mutagenic potential of silver nanoparticles was also investigated by Kim H. R. et al. (2013) in four Salmonella typhimurium strains (TA1535, TA 1537, TA 95, TA 100) using the plate incorporation method. Bacteria were exposed to five concentrations (100-500µg/plate), showing no increase of revertants either with or without metabolic activation. The authors used commercially available silver nanoparticles, which were not further characterised.

Eliopoulos and Mourelatos (1998) reported a weakly positive finding with silver iodide formulated in a polyacrylamide suspension using the plate incorporation method with four S. typhimurium strains (TA102, TA1535, TA97, TA98) with and without metabolic activation. Concentrations ranged from 10-150µg AgI-polyacrylamide suspension/mL. An increase in revertants was observed in strain TA102 without metabolic activation in two medium doses, and in strain TA97 with metabolic activation in the maximum dose. The authors themselves stated that positive findings were only observed at concentrations toxic to bacteria. Further, the reference does not contain information on the silver iodide purity nor the silver content of the preparation. Based on the above shortcomings, the publication is considered of limited relevance for human health hazard assessment.

Li (2012) investigated the bacterial reverse mutation potential of silver nanoparticles in five Salmonella typhimurium strains (TA98, TA100, TA102, TA1535, TA1537) without metabolic activation. Using the plate incorporation assay up to a concentration of 76.8µg/plate, no increase in revertants was observed. The maximum concentration was limited by toxicity.

Chen et al. (2015) investigated in a combined study silver nanoparticles in three different test systems. The results of the bacterial reverse mutation assay using three Salmonella typhimurium strains (TA98, TA100, TA1535) are reported here. Concentrations ranges from 0.2-1.6µg/plate using the plate incorporation method. The test item (poly(styrene-co-maleic anhydride, SMA)-coated silver nanoparticles) did not show an increase in revertants. The maximum concentrations were limited by cytotoxicity. The reference is considered not reliable, due to several serious shortcomings. The test item description islacking basic information, such as source, physico-chemical characterisation, description of the coating procedure with SMA, analytical information on by-products or impurities. Further, only three strains were used, although the use of five strains is recommended.

In the light of the above publications performed with uncoated silver nanoparticles (Kim, J. S. et al., 2012; Kim H. R. et al., 2013), it was concluded that silver nanoparticles are not mutagenic in the Ames test. Typically, bacterial reverse mutation assays are of limited relevance for metals risk assessment, since the uptake of metal cations by bacteria is considered to be poor, and thus the sensitivity of bacterial test systems for the detection of the mutagenic potential of dissolved metal ions appears to be low (cf. REACH Guidance on information requirements and chemical safety assessment, Chapter R.7a (Version May 2008), Page 390; HERAG fact sheet No 5 Mutagenicity, Chapter 2.1). However, silver is a known bactericidal/bacteriostatic agent with a long history in biocidal applications, such as treatment of burns in patients and in disinfection of water systems. It has been shown that silver cations and silver nanoparticles in fact interact with bacterial cells and become systemically available (R. M. Slawson, 1990; E. Bae, 2011). Consequently, studies on bacterial reverse mutations may be considered relevant for the assessment of genetic toxicity of silver and silver substances.

Several other investigations were not considered since serious shortcomings render these references unsuitable for human health risk assessment purposes.

 

Chromosomal aberrations (CA) in mammalian cells

Kim, J. S. et al (2013) studied chromosomal aberrations in CHO-K1 cells. Cells were exposed in pulse and continuous treatment with and without metabolic activation to silver nanoparticles (24h –S9: 0.488, 0.977 and 1.953µg/mL, 6h, -S9: 0.977, 1.953, 3.906µg/mL, +S9: 7.813, 15.625, 31.25µg/mL), the concentration range being limited by cytotoxicity. For the CHO-K1 cells, the AG-NPs did not elicit any statistically significant increase in the number of cells with chromosomal aberrations when compared to the negative control group at any of the dose levels tested with and without metabolic activation. Furthermore, in both the presence and absence of the S9 mix, the silver nanoparticles did not cause any statistically significant increase in the number of cells with polyploidy or endoreduplication when compared to the negative control group.

Nymark et al. (2013) studied both chromosomal aberrations and induction of micronuclei (discussed further below) of PVP coated silver nanoparticles (85% silver, 15% PVP coating) in human bronchial epithelial BEAS 2B cells. Treatment was conducted only without metabolic activation for 24 and 48 hours (24 hour exposure: 2, 4, 8, 16, 24 and 48 µg/cm² (corresponding to 10, 20, 40, 80, 120, 180 and 240 µg/mL); 48 hour exposure: 0.5, 1, 2, 4, 6, and 8 µg/cm² (corresponding to 2.5, 5, 10, 20, 30 and 40 µg/mL)), and the maximum concentration was limited by cytotoxicity. Silver nanoparticles did not increase the rate of chromatid-type, chromosome-type or total chromosomal aberrations at any of the tested doses or time points, and there was no dose-dependent effect for any type of aberration. In contrast, the positive control mitomycin C (50 ng/ml) induced a significant increase in chromosomal aberrations at both exposure times.

In a poorly documented reference, the induction of chromosomal aberrations by silver nanoparticles was investigated (Hackenberg et al., 2011). Human adipose-tissue-derived mesenchymal stem cells were exposed for 1 hour to concentrations of 0.01, 0.1, 1, 10µg/mL in distilled water, MMS was used as positive control substance. Cytotoxic effects were assessed by Trypan blue exclusion dye analysis. The authors concluded that a significant increase in CAs could be found at concentrations of 0.1µg/ml and higher with a dose-response relationship, indicating a substance-induced mechanism. The reference exhibits serious shortcomings: (i) exposure duration too short, (ii) number of scored metaphases too low, (iii) unclear which aberrations were scored and were used for the genotoxicity assessment, (iv) unclear whether a confirmatory experiment was performed. Thus, the reference was considered insufficient for consideration in a human health hazard assessment.

Chen et al. (2015) investigated in a combined study silver nanoparticles in three different test systems. The results of the in vitro chromosome aberration test in CHO-K1 cells are reported here. The cells were incubated with SMA coated silver nanoparticles at 1.875, 3.75 and 7.5 µg/mL. It is unclear whether the incubation was performed with or without metabolic activation. after 18-21 hrs incubation cells were harvested, incubated in 0.5% KCl, fixed stained and analysed. At least 100 metaphases were scored for each group. Authors report a significant increase in chromosome aberrations (not further specified) already at the lowest concentration and showing a dose dependent increase. The reference is considered not reliable, due to several serious shortcomings. The description of the experimental procedure is confusing, e.g. it is unclear in which order the cells were treated with test item, control substances or cytokinesis block. It is also not described how the authors ensured a test item carry over to the evaluation procedure. The test item description islacking basic information, such as source, physico-chemical characterisation, description of the coating procedure with SMA, analytical information on by-products or impurities. The aberration frequency in the negative control group (8.7% ±0.6) is 8-10-fold too high, compared with historical data from publications or other labs which is usually in the range of 0.9-1.3% and which would have also been expected for the well-established CHO-K1 cell line. It appears that the authors experienced problems with the stability of the cell line for such a test system and that the results of this assay should be viewed with caution. Finally, authors did not report findings of the cytotoxicity measurements, it remains therefore unclear whether the increased aberration frequency was caused by a direct clastogenic effect of the test item or whether this positive finding was caused by excessive cytotoxicity.

 

From the chromosomal aberration assays discussed above, it may be concluded by weight-of-evidence that silver nanoparticles are not clastogenic when tested in mammalian cells up to concentrations inducing cytotoxicity. These findings are however somewhat in contradiction to the equivocal findings in the MN assays (reported below).

 

Induction of micronuclei (MN) in mammalian cells

Nymark et al. (2013) studied both chromosomal aberrations (discussed above) and induction of micronuclei of PVP coated silver nanoparticles (85% silver, 15% PVP coating) in human bronchio-epithelial BEAS 2B cells. Treatment was conducted only without metabolic activation for 48 hours (2, 4, 8, 16, 24, 36 and 48 µg/cm² of silver nanoparticles (corresponding to 10, 20, 40, 80, 120, 180 and 240 µg/mL)), and the maximum concentration was limited by cytotoxicity/morphological alteration of cells. Silver nanoparticles did not increase the number of micronucleated binucleate BEAS 2B cells after the 48-hour exposure at any of the tested doses, and no linear dose-dependence could be observed either. The positive control, mitomycin C (150 ng/mL), significantly increased the number of micronucleated binucleate cells (85.5/1000 binucleate cells, P = 0.03).

In the study by Li et al. (2013), the cytotoxic and genotoxic effects of silver nanoparticles on primary Syrian hamster embryo (SHE) cells were investigated. Cell viability was assessed using methyl tetrazolium (MTT), and genotoxic potential was evaluated in a cytokinesis-block micronucleus (CBMN) assay. Freshly isolated SHE cells were exposed to silver nanoparticles at 10, 20, 40 µg/mL for 24 hours in triplicate. The cytotoxicity was assessed in a separate MTT experiment – no check on the MTT influence of the silver NP was performed. The highest concentration induced cytotoxicity of >60%, which might lead to false positive findings in the micronucleus assay. Silver nanoparticles induced a dose-dependent increase of MN frequency up to a 3-fold in the highest dose. The findings in this study are difficult to assess, since MN induction at the highest dose was measured at a concentration with excessive cytotoxicity. The results of this study are therefore difficult to evaluate and should only be assessed in conjunction with other genotoxicity studies in a weight-of-evidence approach.

In a poorly reported study Kim, H. R. et al. (2013) performed a micronucleus test with and without cytokinesis block using silver nanoparticles. In both assays, cells were treated with an unusual dose selection involving four concentrations (0.01, 0.1, 1, 10µg/mL) for 24 hours. For the assay with cyto-B, the test was performed with and without S9-mix. MN induction was seen in the MN assay and CBMN assay with or without S9 mix. The authors did not report any correlation of the micronuclei findings with cytotoxicity, hence it is not clear whether micronuclei were induced by cytototoxic effects. Based on the lack of cytotoxicity information, this reference is considered of limited relevance for human health risk assessment.

Kawata et al. (2009) performed a micronucleus test with a human hepatoma-derived cell line (HepG2). The cells were exposed to silver nanoparticles at a dose level of 1 mg/L for a duration of 24 hours. Authors observed an increase of MN cells from 2.1% for the control culture to 47.9%. The values reported in this publication appear excessively high: even for the untreated control culture a MN frequency of 2% is highly unusual (a value around 0.4-0.5% would have been expected, Unger et al. 2012). Based on the inadequate reporting and since only one concentration was tested, this reference is considered of limited relevance for human health risk assessment.

Li (2012) investigated the induction of micronuclei in human lymphoblastoid cells (TK6) by silver nanoparticles. The cells were treated with concentrations 10, 15, 20, 25, and 30 µg/mL without metabolic activation in triplicate cultures. The maximum concentration was limited by toxicity, reducing the relative population doubling (RPD) to 40% ±5% (60% ±5% cytotoxicity) which is above the maximum toxicity recommended by the OECD TG 487. The micronucleus analysis was performed via flow cytometry by counting 10000 nuclei, thus the results show a good statistical robustness. Micronucleus frequencies were increased by the silver nanoparticle treatment in a dose-dependent manner, with significant increases in micronuclei measured at both 25 and 30 µg/mL. The 25µg/ml treatment produced a 2.59-fold increase over the vehicle control with a net increase of 1.02% and 30 µg/mL produced a slightly greater response of 3.17-fold over the control with a net increase of 1.60%. Overall, the increase of the MN frequency is weak and the highest MN frequency was evoked at a concentration with excessive cytotoxicity. Since the MN frequency shows a dose-dependency and a significant increase at 25µg/mL at acceptable toxicity, the positive finding can be considered substance-related.

Xu et al. (2012) tested the induction of micronuclei in HeLa cells treated with silver nanoparticles contained in a hydrogel. The preparation contained silver at a concentration of 0.38µg/mL and further unknown ingredients. The authors observed a dose-dependent increase of MN frequency up to the highest concentration of 60mg/mL (it is unclear whether this concentration relates to the preparation or the silver nanoparticles). Authors did not report any toxicity findings, hence it is unclear whether the increase in MN was related to primary effects of the silver nanoparticles or secondary to any toxic effect. Based on the poor reporting quality and the fact that cells were exposed to a silver nanoparticle-containing preparation with unknown composition, this reference is not considered relevant for human health risk assessment purposes.

AshaRani et al. (2009) investigated the induction of micronuclei of starch-coated silver nanoparticles on human glioblastoma cells (U251) and normal human lung fibroblast cells (IMR-90). Cells were treated with dispersed silver nanoparticles at concentrations of 100 and 200 µg/mL for 48 hours without metabolic activation. Metabolic activity was determined in separate experiments after 24, 48 and 72 hrs exposure using the cell titre blue cell viability assay. The results of these experiments were presented only graphically, and are thus difficult to evaluate quantitatively. It was not shown whether the decrease in metabolic activity was related to a cytotoxic mode of action. According to the authors, a significant increase of MN frequency was seen in U251 cells. However, due to (i) the unclear cytotoxicity and (ii) the fact that only two concentrations were tested, this reference is considered of limited relevance for human health hazard assessment purposes.

In an unpublished fully guideline-compliant study (Lloyd, M. 2010) performed according to OECD TG 487 and under GLP, disilver(I) sulfate was tested in an in vitro micronucleus assay using duplicate human lymphocyte cultures prepared from the pooled blood of two male donors in a single experiment. Treatments covering a broad range of concentrations, separated by narrow intervals, were performed both in the absence and presence of metabolic activation (S-9). The test article was formulated in purified water and the highest concentration used in the main experiment, 120 ug/mL (limited by toxicity), was determined following a preliminary cytotoxicity range-finder experiment. Treatments were conducted for 48 hours following mitogen stimulation for 3 hours (with an without S9) and 24 hours (without S9). Micronuclei were analysed at three or four concentrations. Mitomycin C (MMC) and vinblastine (VIN) were employed as clastogenic and aneugenic positive control chemicals, respectively, in the absence of rat liver S-9. Cyclophosphamide (CPA) was employed as a clastogenic positive control chemical in the presence of rat liver S-9. Treatment of cells with disilver(I) sulfate in the absence and presence of metabolic activation (S-9) resulted in frequencies of MNBN cells that were generally similar to (and not significantly different from) those observed in concurrent vehicle controls for all concentrations analysed. The MNBN cell frequency of all disilver(I) sulfate treated cultures fell within the observed normal ranges. It is concluded that disilver(I) sulfate did not induce micronuclei in cultured human peripheral blood lymphocytes in the absence and presence of S-9 when tested up to the limit of cytotoxicity.

The references contained in the summary entry for in vitro micronucleus tests are of 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 and hazard assessment purposes. The information contained therein were included for information purposes only and the deficiencies of the studies are listed below:

Aktepe et al. (2014) reported in a short abstract on the induction of micronuclei in human lymphocytes exposed to silver nanoparticles(not further characterised) at concentrations of 12.5, 50, 250 µg/mL. Authors report an increase in MN frequency in exposed cells, although only one frequency value was reported for the exposed group, although three values should have been expected. Due to the brevity of the abstract, the reference is lacking basic information on experimental conditions, results and discussion. Consequently, the information contained therein are of no relevance for hazard assessment purposes.

Jiang et al. (2013) exposed CHO-K1 cells to BSA coated silver nano particles to concentrations of 1, 5 and 10 µg/mL. Cells were exposed for 24 hrs, fixed, stained and analysed by flow cytometry. Authors report an up to 2.5-fold increase in MN frequency. The reference is lacking basic information on the test item, such as source, physico-chemical characterisation, description of the coating procedure with BSA, analytical information on by-products or impurities. It remains therefore unclear whether any of the findings were elicited by a true test item induced effect, or whether effects were caused by any by-product or impurity. The analysis of MN was conducted using flow-cytometry. The high number of cells measured by this method should lead to statistically robust data with a low SD. However, the results presented here show a high SD. Further, authors did not show any exemplary histogram, showing gating or level of staining. The results therefore remain undocumented. Finally, authors did not check for interference of cell growth with the solid matter. Based on the above given shortcomings, the reference is considered not reliable and of no relevance for hazard assessment purposes.

 

In summary, there are several publications investigating the induction of MN by silver nanoparticles. One reliable publication on PVP coated silver nanoparticles conclusively shows a complete absence of a clastogenic or aneugenic mode of action. Three publications with limited quality indicate a potential for MN induction by uncoated silver nanoparticles. However, these findings are in contradiction to a substantial body of findings in in-vitro chromosome aberration assays reported above, where no induction of aberration or, as reported in one study, no induction of polyploidy was observed. In a guideline and GLP compliant study, the silver salt disilver(I) sulfate did not induce micronuclei in cultured human peripheral blood lymphocytes in the absence and presence of S-9 when tested up to the limit of cytotoxicity.

Overall, in a conservative approach the in-vitro data on the clastogenic and aneugenic potential of uncoated silver nanoparticles appears to be equivocal.

 

Gene mutations in mammalian cells

Mei, N. et al. (2012) tested silver nanoparticles for induction of TK mutations in mouse lymphoma cells. Treatment was only for 4 hours in the absence of metabolic activation. Concentrations ranged from 3-6 µg/ml, with a RTG of <10% at the highest dose. The MFs at the concentrations of 4 and 5 µg/mL of the Ag-NPs were 400 ± 230 x 10-6(mean ± SD) and 453 ± 171 x 10-6, respectively, while the MF for the control group was 89 ± 20 x 10-6, clearly exceeding the Global Evaluation Factor (GEF) of +126 mutants per 106cells as recommended by Moore et al. (2006). In the positive cultures, both large and small colony mutants were induced, but the predominant response was induction of small colony mutants (87%), more indicative of a clastogenic (chromosome damaging) effect. Therefore, silver nanoparticles may induce TK mutations in mouse lymphoma cells with the effect being predominantly clastogenic.

Kim Y. J. et al. (2010) tested silver nanoparticles for the induction of TK mutations in L5178Y mouse lymphoma cells. Treatment was only for 3 hours in the presence and absence of metabolic activation (S9). The concentrations ranged from 313 up to 2500 µg/mL (-S9) and 3750µg/mL (+S9), selection made based on the IC20 value determined in a dose range finding experiment. The treatment of silver nanoparticles did not lead to a significant increase of MF at any of the tested concentrations, with or without metabolic activation.

In an unpublished guideline study performed under GLP, disilver(I) sulfate was assayed for its ability to induce mutation at the TK locus (5-trifluorothymidine [TFT] resistance) in mouse lymphoma cells using a fluctuation protocol (Lloyd, M., 2010). In the absence of S-9 (3 hour treatment), when tested up to toxic concentrations in Experiment 1, an increase in mutant frequency (MF) of approximately 129, which marginally exceeded the Global Evaluation Factor (GEF, 126 mutants per 106viable cells) was observed at 1.5 μg/mL, the highest concentration analysed. However, the MF value exceeded the GEF in only one of the replicate cultures at this concentration. When tested up to toxic concentrations in Experiment 2 (3 hour treatment), increases in MF of approximately 168 and 189 (both slightly exceeding the GEF) were observed at 1.2 and 1.3 μg/mL, respectively, and the individual MF values of the replicate cultures exceeded the GEF at both concentrations. Significant linear trends were observed in both experiments. In the absence of S-9 (24 hour treatment), when tested up to toxic concentrations, no marked increases in MF (which exceeded the GEF) were observed at any concentration analysed in Experiment 2. In the presence of S-9, no marked increases in MF (which exceeded the GEF) were observed at any concentration analysed in Experiments 1 and 2. In the positive cultures, both large and small colony mutants were induced, but the predominant response was induction of small colony mutants, indicative of a clastogenic (chromosome damaging) effect. The exceedance of the sum of the mean control MF plus GEF in the second experiment was observed at marketed cytotoxicity with a %RTG of ≤15%. Since these effects are regarded as primarily clastogenic because of colony sizing, it can be concluded that the positive findings are more a direct result of cytotoxicity rather than a clear mutagenic effect of the test item. Further, no positive findings were reported in the continuous treatment over 24 hours, which is considered the most critical treatment regime.

From the references discussed above it can be concluded that silver nanoparticles as well as a soluble silver salts do not induce point mutations, but show a weak clastogenic mode of action in the TK assay with mouse lymphoma cells, observed by small colony formation at concentrations with high cytotoxicity.

Overall, there was no convincing or consistent evidence of induction of gene mutations in either bacterial or mammalian cells in vitro, but an equivocal evidence for a clastogenic mode of action of silver in mammalian cells.

 

In vitro DNA damage assays

The majority of published experiments on silver substances are investigations of DNA damage of highly variable quality and reliability involving the comet assay as test system. The comet assay is a powerful tool to detect even low levels of DNA damage. However, this assay is also prone to positive findings not caused by the test item but by e. g. inappropriate test design. Minimum quality criteria were therefore used to rate these comet studies for their relevance in chemical safety assessment. The criteria as published by Tice et al. (2000) were used for this screening. In case a reference did not fulfil the criteria stated therein, it was rated as “not rateable” and not further considered for chemical safety assessment purposes.

Nine references, investigating DNA damage via comet assay are available. The following test substances were used:

(i) commercially available silver nanoparticles (≤100nm, Ghosh 2012; 20 and 200nm, Asare 2012; <50nm, Hackenberg 2011, Kim 2013)

(ii) self-synthesised silver nanoparticles by reduction of silver nitrate with sodium borohydrate (Flower 2012)

(iii) PVP coated silver nanoparticles (42.5nm, 85%Ag, 15% residues, Nymark 2013)

(iv) oligonucleotide and lactose modified self-synthesised silver nanoparticles (50nm, Sur 2012)

(v) silver nanoparticles of unknown source and purity (<100nm, Kim 2010).

All references exhibit serious shortcomings in the experimental design and/or the reporting which render them unsuitable for human health risk assessment purposes, such as:

(i) the authors did not correlate DNA damage findings observed in the comet assay with the cytotoxicity at the respective exposure concentration. Thus, it remains unclear whether DNA damage was caused via cytotoxic effects or via direct test item-DNA interaction (Asare 2012, Ghosh 2012, Sur 2012, Cronholm 2013)

(ii) the experimental procedures for cell exposure, staining and scoring were insufficiently or not described, which does not allow an independent evaluation of the results (Hackenberg 2011, Flower 2012, Nymark 2013, AshaRani 2009, Cronholm 2013)

(iii) positive control substances were not used in parallel experiment to assess the sensitivity of the test system (Flower 2012, Cronholm 2013)

(iv) none of the researchers properly considered substance carry-over in the lysis/unwinding/ electrophoresis phase, which might lead to false positive findings due to direct DNA-silver nanoparticle interaction (Kim 2010, Kim 2013)

(v) the substance identity and/or particle size was not conclusively reported. Authors did not demonstrate whether cells were exposed to nanoparticles or whether positive findings were elicited by impurities. This quality criterion was established by the Scientific Committee on Consumer Safety (SCCS) during study evaluation of nano-sized titanium dioxide (“…study is of little value in relation to assessment for nano-form of TiO2 as there is a lack of data on characterisation (particle size distribution) of the tested materials to show that they were nanomaterials. ”, Scientific Committee on Consumer Safety (SCCS) (2013) Opinion on Titanium dioxide (nano form), COLIPA n° S75, document number SCCS/1516/13).

 

In vivo studies in animals

Micronucleus (MN) test

Kim et al. (2011) investigated the induction of micronuclei in the bone marrow of male and female Sprague-Dawley rats exposed via inhalation to silver nanoparticles for 13 weeks at three dose levels (49, 133 and 515 µg/m³); note: 133µg/m³ is the sub-chronic inhalation NOAEC. The frequency of micronucleated polychromatic erythrocytes (MN PCEs) in every 2000 PCEs for the male rats was 0.13, 0.21, and 0.18 percent for the groups exposed to low, middle, and high concentrations of silver nanoparticles, respectively, while that for the control was 0.14 percent. The frequency of MN PCEs in every 2000 PCEs for the female rats was 0.09, 0.08, and 0.13 for the groups exposed to low, middle, and high concentrations of silver nanoparticles, respectively, while that for the control was 0.14 percent. No significant treatment-related increase of MN PCEs was detected in the male and female rats when compared to the corresponding negative controls. A dose-dependent deposition of silver nanoparticles was found in various tissues, indicating that the silver was systemically distributed throughout the test organism. The results suggest that exposure to silver nanoparticles by inhalation for 90 days does not induce genetic toxicity in male and female rat bone marrow in vivo. Systemic exposure was shown by increased silver tissue levels.

In a micronucleus test, 10 male and female Sprague-Dawley rats were exposed orally via gavage to silver nanoparticles at doses of 30, 300 and 1000 mg/kg bw/day over a period of 28-days. No increase of the micronucleated polychromatic erythrocyte frequency was observed (Kim, Y. S. et al., 2008). Systemic exposure was shown by increased silver tissue levels, however direct bone marrow exposure was not analytically verified in this particular study.

Ghosh et al. (2012) investigated the induction of chromosomal aberrations in the bone marrow of male Swiss albino mice following intraperitoneal administration of silver nanoparticles at 10, 20, 40 and 80mg/kg bw. It was found that silver nanoparticles significantly increased the number of chromosome damages, mainly chromatid breaks.

Chen et al. (2015) investigated in a combined study silver nanoparticles in three different test systems. The results of the in vivo micronucleus test in male ICR mice intraperitoneally injected with SMA coated silver nanoparticles. The six week old animals were divided in three groups of 5 animals each, injected with 0.25 and 1 mg/kg SMA coated silver nanoparticles. Positive control animals received cyclophosphamide. A negative control group was presumably included, but details were not provided in the methods description. Peripheral blood samples were drawn via retro-orbital bleeding 48 and 72 hrs after exposure. Cells were stained, and a total of 1000 RET were scored for MN. For the PCE/NCE ratio 1000 cells were scored. No increase in the MN frequency was seen after 48 or 72 hours. The MN frequencies in the negative control and treated animals were within the historical control range of published data or other labs. The reference is considered not reliable, due to several serious shortcomings. The description of the experimental procedure is confusing, e.g. it is how many animals were assigned to how many groups – the method described three groups with 5 animals each, the results report 4 groups with unknown number of animals. The experimental procedure is described very briefly, so that the cell preparation procedure remains unclear. The number of RET scored for presence of MN is too low, compared with the guideline recommendation. The test item description islacking basic information, such as source, physico-chemical characterisation, description of the coating procedure with SMA, analytical information on by-products or impurities.

The increase of micronuclei frequency was investigated by Kovvuru et al. (2015) in C57BL/6J pun/punmice (4 - 6 mice/group, in equal proportions of males and females) following oral administration of 500 mg/kg bw/day PVP coated silver nanoparticles over 5 days. Peripheral blood was taken from treated and control animals via sub-mandibular bleeding 24 hours after one daily dose and 24 hours after five daily doses. Bone marrow samples were collected 24 hours after five daily doses. At least 4000 peripheral blood erythrocytes and 2000 bone marrow erythrocytes were analysed per mouse. According to the authors, PVP-coated silver nanoparticles induced chromosomal damage in bone marrow and peripheral blood. However, the results of this study cannot be evaluated, due to reporting deficiencies and shortcomings in the experimental design. The test item was insufficiently characterised, the presence of PVP-coated silver nanoparticles was not sufficiently demonstrated. The number of animals per group was too low compared with the guideline recommendation, which reduces the statistical robustness of the results. Only one dose was used during testing, which does not allow to investigate a dose-response relationship as recommended by the guideline. In addition, no positive control was used in the study design. The negative control MN frequency was in an unusually very low range (0.025-0.1 %), compared with published historical control data (0.1-0.35%) and the MN frequencies were within the range of published historical negative control data (0.25-0.35%). It appears that authors concluded on a positive finding due to the unusual low MN frequency in the negative control animals. Since a positive control was not used in these experiments, the selectivity of the system was not demonstrated. The statistically significant positive findings in the treated animals might therefore be an accidental finding without biological relevance. Finally, authors did not report any markers of cytotoxicity, thus findings might as well have been caused by elevated cytotoxicity and not a true clastogenic response. Overall, the results were only presented in graphs and raw data was not available. Consequently, the reference is rated not reliable and considered unsuitable for hazard assessment purposes.

 

Chromosome aberration test

El Mahdy et al. (2014) conducted a chromosome aberration study with self-synthesised silver nanoparticles using mature female albino rats. Groups of five rats received doses of 1, 2 and 4 mg/kg bw/day via intraperitoneal injection for 28 days, negative control group was run concurrently. A positive control group was not included. Chromosomal aberration was measured in bone marrow at the end of the treatment period. At least 250 metaphases of each animal were scored, characterising different types of chromosomal aberrations. The reference exhibits serious shortcomings which renders it unsuitable for its use in hazard or risk assessment. Authors did not provide information on: (i) rat strain used in the experiments, (ii) the dose setting or MTD and provide no data on any clinical signs during the course of the study (iii) historical positive or negative control data, thus eligibility of the lab or the rat strain used in the CA experiments were not documented. No positive control group was added in the study, thus the sensitivity of the test system remains undocumented. The test item was given via intraperitoneal injection. Such an non-physiological route of administration is not suitable for the hazard assessment of industrial chemicals and is not compliant with the guideline recommendation (OECD TG 475) which clearly states that “intraperitoneal injection is generally not recommended since it is not an intended route of human exposure”.Finally, the test item description islacking basic information, such as source, physico-chemical characterisation and analytical information on by-products or impurities. It remains therefore unclear whether the cytotoxicity and DNA damage was caused by the test item itself or by any by-product or impurity. Self-synthesised nanoparticles have also not relevance for the risk assessment of industrially manufactured nanomaterials with clearly defined physico-chemical characteristics and analytically verified impurity profiles.Consequently, the reference is rated not reliable and considered unsuitable for hazard assessment purposes.

 

DNA damage (Comet assay) test

The in vivo comet assay study design was checked against the OECD test guideline 489 (26 September 2014), which is based on internationally agreed study designs suggested by various authors in the public domain (e.g. Kirkland 2008, Brendler-Schwaab 2005, Burlinson 2007 and 2012, Smith 2008, Hartmann 2003, McKelvey-Martin 1993, Tice 2000, Singh 1998, Rothfuss 2010). Although the OECD guideline was published after publication of some references summarised below, the comparison with OECD 489 is appropriate, since the majority of the underlying protocols existed at the time of study planning and conduct.

Li, Y. et al. (2013) investigated the induction of DNA damage via comet assay and micronuclei in the bone marrow of B6C3F1 mice. Five male animals were intravenously injected with a single or a repeated dose for 3 consecutive days of 0.5-20 mg/kg bw with (i) PVP-coated and (ii) silicon coated silver nanoparticles. Target organ exposure was analytically verified via ICP-MS analysis. Bone marrow toxicity was only observed in the test group exposed to PVP coated silver nanoparticles. None of the treatments resulted in a significant increase of the percent of micronucleated reticulocyte over the concurrent controls. For the DNA damage analysis, liver samples were collected from mice treated with 25 mg/kg bw PVP- or silicon coated silver nanoparticles, or the vehicle for 3 consecutive days and assayed for DNA damage 3 h after the last treatment. No DNA strand breaks were detected in liver for both PVP- and silicon-coated silver nanoparticles in the standard comet assay while significant induction of oxidative DNA damage was found in the hOGG1-modified comet assay. It remains unclear why the standard alkaline comet assay did not detect any DNA damage whereas the modified comet assay (specific for oxidative DNA strand breaks) was positive.

In a disregarded neutral comet assay, DNA damage was observed in the spleen of male and female Balb/c mice treated via intraperitoneal injection of a single silver nanoparticle dose. The unit of the dose was reported as g Ag/L, due to the lack of body weight information the actual administered dose could not be calculated (Ordzhonikidze et al., 2009).

Dobrzyńska, M. et al. (2014) performed a combined micronucleus test and comet assay in male Wistar rats. Rats were injected intravenously with a single dose of 5 or 10 mg/kg bw of 20 nm silver nanoparticles or with 5 mg/kg bw 200 nm silver nanoparticles. The samples for the comet assay were taken at 24 hours, 1 week and 4 weeks following the exposure. The study showed that silver nanoparticles did not cause cytotoxicity to bone marrow white blood cells. Moreover, the results with the comet assay did not show any significant increase of DNA damage in bone marrow leukocytes of rats. This lack of observations is valid for all time points post single treatment with silver nanoparticles. In silver-nanoparticle-treated animals, a significant increase in the number of micronuclei per 1000 PCE was observed at 24 hours after exposure independent of dose and particle size. One week after the exposure, the significantly enhanced as compared to the negative control levels of micronuclei was noted in all silver nanoparticles exposed groups. Four weeks following the single exposure still the significantly higher frequency of micronuclei was observed only in animals administered to 10 mg/kg silver nanoparticles of 20 nm size. The reference shows inconsistencies in the reporting of the genetic toxicity findings, two tables report scoring results for MN cells, namely MN-PCE and MN-Reticulocytes. Since only one MN experiment was performed, it remains unclear why two tables are reporting results for the identical cell type but with different scoring results.

An unequivocally negative outcome was seen in rats and mice following repeated exposures to silver nanoparticles via inhalation (90-days, with exposures beyond the NOAEC; Kim, Y. S. et al., 2008) and oral administration (28-days, with dosing up to the limit dose, i. e. 1,000 mg/kg bw/d; Kim et al., 2011). The induction of chromosomal aberrations by silver nanoparticles reported by Ghosh et al. (2012) and Dobrzyńska, M. et al. (2014) as well the DNA damages reported by Ordzhonikidze et al. (2009) and Dobrzyńska, M. et al. (2014) were only observed after a non-physiological route of exposure. Thus, the results are considered to lack direct relevance for human health risk assessment purposes for industrial chemicals under REACH. Experiments with PVP- and silicone-coated silver nanoparticles showed a negative outcome in the in vivo MN assay in bone marrow in mice exposure via single and repeated intravenous injection. The standard alkaline comet assay in liver was negative, whereas a modified hOGG1-comet assay indicated oxidative DNA strand breaks – which remained unexplained by the authors.

Awasthi et al. (2015) examined the induction of DNA damage in the liver of male mice after oral administration of self-synthesised silver nanoparticles. Animals (6 males per group) were given (i) a single dose of 50 and 100 mg/kg and (ii) doses of 10 and 20 mg/kg bw/day over 5 weeks via gavage. Animals were sacrificed 3 or 24 hrs after single and 3 hrs after the last of repeated administrations, respectively. Triton X-100 was administered as vehicle control, negative control animals received no treatment, positive control animals received 25 mg/kg cyclophosphamide via oral route. After sacrifice, animals were dissected and single liver cell suspensions were prepared. In addition liver slides were prepared for histopathological examination. Liver function was determined by AST, ALT and ALP measurements in serum. For the comet analysis cells were fixed in low-melting agarose, unwinded and lysed in alkaline buffer. After electrophoresis, slides were stained with ethidium bromide. DNA migration was determined in 100 cells per group via tail length, %DNA in tail and tail moment. Authors observed a dose dependent increase of DNA damage after single and repeated administration. In the histopathological examination treated animals showed hepatic congestion and haemorrhage. All three enzyme levels (AST, ALT and ALP) were increased in the treated animals, indicating cytotoxicity to the liver. According to the authors, silver nanoparticles has the potential to cause genetic damage. However, the results of this study cannot be evaluated, due to shortcomings in the experimental design. First of all, changes in DNA migration observed at doses also show marketed toxic effects in the target tissue (i.e. AST, ALT and ALP increased, haemorrhage in liver). The number of cells analysed (100 cells per group of 6 animals) was significantly too low, compared with the guideline recommendation (150 cells per animal). Conflicting information were presented on how the positive control was administered. Overall, a substance specific induction of DNA damage has not been demonstrated in this study. Finally, the test item description islacking basic information, such as source, physico-chemical characterisation, analytical information on by-products or impurities. It remains therefore unclear whether the cytotoxicity and DNA damage was caused by the test item itself or by any by-product or impurity. Self-synthesised nanoparticles have also not relevance for the risk assessment of industrially manufactured nanomaterials with clearly defined physico-chemical characteristics and analytically verified impurity profiles.Consequently, the reference is rated not reliable and considered unsuitable for hazard assessment purposes.

Al Gurabi et al. (2015) examined the induction of DNA damage in the liver of Swiss albino mice (sex not reported) after intraperitoneal injection of self-synthesised silver nanoparticles. Animals (unclear number of animals per sex and group) were given the test item for 24 and 72 hours at 26, 52, and 78 mg/kg bw silver nanoparticles. The frequency of treatment is unclear, since no further information on the treatment conditions was given. Vehicle control animals received water, no positive control animals were included. Blood samples were drawn and lymphocytes separated from the whole blood. For the comet analysis cells were fixed in low-melting agarose, unwinded and lysed in alkaline buffer. After electrophoresis, slides were stained with ethidium bromide. DNA migration was determined %DNA in tail in 250 cells per treatment group. For the positive control ex-vivo liver cells were treated with 100µM H2O2 for 10 minutes.The authors reported the induction of DNA damage in lymphocytes of mice treated with silver nanoparticles. However, the results of this study cannot be evaluated, due to severe reporting deficiencies and shortcomings in the experimental design. The dosing regimen of the animals is unclear (only treatment for 24 or 72 hrs was reported, without indication of the treatment frequency – it is unclear how to treat animals via i.p. injection over a period of 24 or 72 hrs) as well as the number and sex of animals used during testing. In addition, no valid in vivo positive control was used. Instead ex-vivo liver cells were treated with H2O2, which is not an accepted positive control substance, an ex-vivo treatment of cells is also not a valid method for an in vivo study positive control. Microscopic pictures and fluorescent pictures of comets were of poor quality and therefore unsuitable for assessment. Poor picture quality raises doubts on the percentage tail DNA evaluation.Finally, the test item description islacking basic information, such as source, physico-chemical characterisation, analytical information on by-products or impurities. It remains therefore unclear whether the cytotoxicity and DNA damage was caused by the test item itself or by any by-product or impurity. Self-synthesised nanoparticles have also no relevance for the risk assessment of industrially manufactured nanomaterials with clearly defined physico-chemical characteristics and analytically verified impurity profiles.Consequently, the reference is rated not reliable and considered unsuitable for hazard assessment purposes.

 

In vivo studies in humans

In the study by Aktepe et al. (2015) the genotoxic effect of silver particle exposure among silver jewellery workers was investigated. The silver jewellery workers were assumingly exposed to the silver particles by inhalation. DNA damage in peripheral mononuclear leukocytes was measured by using the alkaline comet assay. According to the authors, exposure to silver particles by inhalation caused significant genotoxicity in mononuclear leukocytes in silver jewellery workers. This publication has a low relevance for human risk assessment, since exposure conditions were not determined. It is unknown to which substances the jewellery workers were expose to and the duration and concentration of exposure is not known. Furthermore, confounding factors were not considered at all, such as sex, smoking habits, alcohol consumption, recreational behaviour, medication. No significant correlation between (assumed) silver exposure and presence of DNA damage could be shown by the authors.

 

Conclusion

The available data on genetic toxicity allow a conclusive statement on the genetic toxicity for silver and silver substances (including nanoparticles). Irrespective of the reporting quality of the publications, initially positive as well as negative findings are reported in in vitro as well as in vivo test systems.

Following rigorous relevance and reliability screening, it can be concluded that silver nanoparticles show a clastogenic potential in mammalian cells in-vitro, but at concentrations with high levels of cytotoxicity. In contrast, high-quality in vivo studies conversely do not report any clastogenic or aneugenic effects after silver nanoparticle exposure via physiologically relevant routes of exposure (i. e. via inhalation or oral routes of administration).

Consequently, the positive in vitro clastogenicity findings are conclusively superseded by the negative in vivo findings. Further, there is convincing evidence that silver nanoparticles do not induce gene mutations either in bacterial or in mammalian cells.

Overall, there is no consistent evidence of induction of genetic toxicity with relevance to humans for silver and silver substances.

 

References

Slawson, RM, H. Lee, H, Trevors, JT (1990) Bacterial interactions with silver, Biology of Metals, 3(3-4), 151-154

Bae, E. et al. (2011) Bacterial uptake of silver nanoparticles in the presence of humic acid and AgNO3, Korean J. Chem. Eng., 28(1), 267-271

Unger, P., Melzig F. (2013) Comparative Study of the Cytotoxicity and Genotoxicity of Alpha- and Beta-Asarone, Sci Pharm. 2012; 80: 663–668

Kirkland, D., G. Speit (2008), Evaluation of the ability of a battery of three in vitro genotoxicity tests to discriminate rodent carcinogens and non-carcinogens III. Appropriate follow-up testing in vivo, Mutation Research, Vol. 654/2, pp. 114-32.

Brendler-Schwaab,S. et al. (2005), The in vivo Comet assay: use and status in genotoxicity testing, Mutagenesis, Vol. 20/4, pp. 245-54.

Burlinson, B. et al. (2007), Fourth International Workgroup on Genotoxicity Testing: result of the in vivo Comet assay workgroup, Mutation Research, Vol. 627/1, pp. 31-5.

Burlinson,B. (2012), The in vitro and in vivo Comet assays, Methods in Molecular Biology, Vol. 817, pp.143-63.

Smith,C.C, et al. (2008), Recommendations for design of the rat Comet assay, Mutagenesis, Vol. 23/3, pp. 233-40.

Hartmann, A.et al. (2003), Recommendations for conducting the in vivo alkaline Comet assay, Mutagenesis, Vol. 18/1, pp. 45-51.

McKelvey-Martin, V.J. et al (1993), The single cell gel electrophoresis assay (Comet assay): a European review, Mutation Research, Vol. 288/1, pp. 47-63.

Tice, R.R.et al.(2000), Single cell gel/Comet assay: guidelines for in vitro and in vivo genetic toxicology testing, Environmental and Molecular Mutagenesis, Vol. 35/3, pp. 206-21.

Singh, N.P.et al. (1988),A simple technique for quantitation of low levels of DNA damage in individual cells, Experimental Cell Research, Vol. 175/1, pp. 184-91.

Rothfuss, A. et al. (2010), Collaborative study on fifteen compounds in the rat-liver Comet assay integrated into 2-and 4-week repeat-dose studies, Mutation Research, Vol., 702/1, pp. 40-69.

 


Short description of key information:
By applying a weight-of-evidence approach, after consideration of the predominantly negative test results of highly reliable in-vitro and in-vivo genotoxicity assays and their route of administration, the overall conclusion is reached that no genotoxicity needs to be expected from exposure to silver substances, regardless of whether they were administered in ionic form or as nanomaterials.

Endpoint Conclusion: No adverse effect observed (negative)

Justification for classification or non-classification

In consideration of the predominantly negative test results in highly reliable genotoxicity assays, no genotoxicity needs to be expected from exposure to silver substances, regardless of whether they were administered in ionic form or as nanomaterials. In consequence, no classification is required.