Chemical Hazard Assessment for Low-Solubility, Non-Nanoscale1) Silver (CAS # 7440-22-4)

Template Copyright 2012 © Clean Production Action

Content Copyright 2016 © Coming Clean Inc.

Chemical Hazard Assessment for low-solubility, non-nanoscale1) silver (CAS # 7440-22-4)

Modified from GreenScreen® Version 1.2[1]

Confirm application of the Disclosure and Assessment Rules and Best Practice[2]: (List disclosure threshold and any deviations)

because this review is based on generic silver, not a particular manufacturer’s product, the de minimus rule is not applicable.

Chemical Name (CAS #): Silver, CAS # 7440-22-4, restricted to the non-nanoscale particulate silver (> ~100 nm and up to respirable particle sizes).

Suitable analogs or moieties of chemicals used in this assessment (CAS #’s):

Silver Chloride (> ~100nm) (CAS # 7783-90-6) and Silver (II) Oxide (> ~100nm) (CAS # 1301-96-8)

Chemical Structure(s):

*Note: Include chemical structure(s) of all surrogates, analogs (and /or moieties) used in the assessment.

Ag (see description of properties below)

AgCl (silver chloride, non-nano)

AgO (silver oxide, non-nano)

Justification for Chemical Surrogates:

To the extent that the mammalian toxicity of conventional silver is a function of dissolved, soluble silver, toxicity should be similar among various silver compounds due to the ubiquitous presence of chloride ion in physiological systems. The dissolved, bioaccessable silver was determined for silver metal, disilver oxide, and silver nitrate, incubated with various artificial physiological media (ECHA, 2012). For these three compounds, with solubility characteristics spanning insoluble silver metal to soluble silver nitrate, the dissolved concentrations of silver were very similar and independent of the original silver compound (silver metal, disilver oxide, silver nitrate). The authors hypothesized that the complex ionic environment and the likely formation of poorly soluble silver chloride leads to very similar equilibrium concentrations of dissolved silver, independent of the originating substance (ECHA, 2012).

Likewise, the ecotoxicity hazard of conventional silver depends upon silver bioavailability, which is recognized to be a function of water chemistry. The Ecotoxicity Hazard Score derived in this Green Screen is considered to be theoretical and based on the results of standardized tests that do not necessarily represent the dynamic environmental conditions experienced in the field. The Biotic Ligand Model is a metal bioavailability model that was developed to incorporate metal speciation and the protective effects of competing cations into predictions of metal toxicity (WHO, 2002). Use of the Biotic Ligand Model may therefore be advisable in predicting a more pragmatic estimate of ecotoxicity hazard than what is conservatively provided in this Green Screen.

Poorly soluble forms of conventional silver were selected for this assessment for the following reasons: 1) To enable a direct comparison with poorly soluble forms of nanoscale silver in a companion assessment and 2) due to the rapid dissociation of silver nitrate in aqueous media, it can deliver soluble silver at a high rate, potentially causing excessive toxicity before equilibrium conditions prevail .

Data for ionic silver, highly soluble silver compounds (e.g. silver nitrate), and moderately soluble organic silver salts (e.g. silver acetate) was not used in this report to fulfill datagaps, primarily because soluble silver was determined to be outside of the scope of this GreenScreen Assessment. NSF acknowledges however that since more soluble forms of silver generally result in greater toxicity due to greater ion release, it would not be unreasonable to expand the scope in a future review to include data on the more soluble forms for fulfilling data gaps for metallic silver. Where applicable, this GreenScreen report does indicate where data on the soluble salts were available and not considered.

Notes related to production specific attributes[3]:

For Inorganic Chemicals and relevant particulate organics (if not relevant, list NA)

Define Properties:

The following inorganic chemical characteristics were examined in each study and reported where available as part of the assessment of study quality and relevance:

1. Particle size: mean or median > ~100 nm and up to the respirable range if exposure is by inhalation.
2. Structure
3. Mobility (e.g. Water solubility, volatility).
4. Bioavailability
5. Chemical composition
6. Purity
7. Whether any characterization was conducted in the relevant experimental media.

Identify Applications/Functional Uses:

(e.g., Cleaning product, TV casing)

1. Textile applications as an antimicrobial fabric protector (e.g. silver coated fibers).
   

GreenScreen Benchmark Score and Hazard Summary Table:[4],[5],[6],[7] Conventional Silver was assigned a Benchmark Score of 1 based on combined very high persistence coupled with very high aquatic toxicity, as determined in standardized tests. Note, this particular combination of hazards is a trigger for Benchmark 1, and this score could not improve with fewer data gaps. It is possible, however, that application of the Biotic Ligand Model to account for metal bioavailability under environmental conditions, could improve the very High Ecotoxicity score due to the protective effects of competing cations. For more information, refer to the U.S.EPA website regarding use of the Biotic Ligand Model and WHO (2002).

As a word of caution, a data gap should not be interpreted as implying hazard or safety, but rather there was insufficient data to characterize the hazard as low, moderate, or high.

Note: Hazard levels (Very High (vH), High (H), Moderate (M), Low (L), Very Low (vL)) in italics reflect estimated values, authoritative B lists, screening lists, weak analogues, and lower confidence. Hazard levels in BOLD font are used with good quality data, authoritative A lists, or strong analogues. Group II Human Health endpoints differ from Group II* Human Health endpoints in that they have four hazard scores (i.e., vH, H, M and L) instead of three (i.e., H, M and L), and are based on single exposures instead of repeated exposures.

Environmental Transformation Products and Ratings[8]:

Identify feasible and relevant environmental transformation products (i.e., dissociation products, transformation products, valence states) and/or moieties of concern[9]

Introduction:

Silver is an EPA registered active ingredient in numerous pesticide products, including antimicrobial treated textiles. NSF recently performed a GreenScreen™ assessment to characterize the hazards of inorganic, low solubility nanosilver in textiles. In the present report, NSF performs a GreenScreen™ assessment on conventional (low-solubility, non-nano) silver for comparison.

Hazard Classification Summary Section:

For all hazard endpoints:

• Search all GreenScreen specified lists. Report relevant results either in each hazard endpoint section or attach to the end of the report.
• Always indicate if suitable analogs or models were used.
• Attach modeling results (See Appendix C).
• Include all references either in each hazard endpoint section or at the end of the report.

Group I Human Health Effects (Group I Human)

Carcinogenicity (C) Score (H, M or L): DG

Conventional silver was assigned a score of Data Gap for carcinogenicity based on lack of data.

• Authoritative and Screening Lists
• Authoritative:
• US EPA - IRIS Carcinogens - (1986) Group D - Not classifiable as to human carcinogenicity
• Screening: Not present on any screening lists

• No cancer studies of silver or silver compounds were found for inhalation, oral, or dermal exposure in humans or animals (ATSDR, 1990).
• EPA’s Cancer Classification for Silver is Group D – Not Classifiable as to Human Carcinogenicity. The basis for the classification is inadequate evidence. EPA notes that while local sarcomas have been induced after implantation of films and disks of silver, the interpretation of these findings has been questioned due to the phenomenon of solid-state carcinogenesis in which even insoluble solids such as plastic have been shown to result in local fibrosarcomas (IRIS, 2003). It should be noted this designation is not specific to any particular chemical or physical form of silver.
• The Japanese NITE classification for carcinogenicity is “classification not possible”. This was based on descriptions that carcinogenicity was not observed in the test in which powder was intramuscularly injected to rats as reported in Patty’s Industrial Hygiene (5th, 2001), there was no carcinogenic evidence to humans as reported in Patty’s Industrial Hygiene (5th, 2001) and HSDB (2003), and there is also no information of classification evaluation from organizations such as IARC (NITE, 2006).

Mutagenicity/Genotoxicity (M) Score (H, M or L): M

Conventional silver was assigned a score of Moderate (low confidence) for mutagenicity based on limited or marginal evidence of mutagenicity in in vitro genotoxicity studies using bacteria and nonhuman mammalian cell cultures. As stated in the 1990 ATSDR profile for silver “existing data on mutagenicity are inconsistent, but data on genotoxicity suggest that the silver ion is genotoxic. From the results of in vitro genotoxicity studies using bacteria and nonhuman mammalian cell cultures it is evident that the silver ion does bind with DNA in solution in vitro, and that it can interact with DNA in ways that cause DNA strand breaks and affect the fidelity of DNA replication. However, silver has not been found to be mutagenic in bacteria. The low confidence is assigned as the score is based on results only observed within in vitro studies.

Authoritative and Screening Lists

• Authoritative: Not on any authoritative lists
• Screening: Not on any screening lists

• The Japanese NITE classification for germ cell mutagenicity is “classification not possible” based on lack of data (NITE, 2006).

In vivo - mammalian

• Groups of Crl:CD-1 (ICR) BR mice/dose were administered silver chloride (>99.5% a.i.) in a homogenous suspension of corn oil by i.p. injection at 31.25, 62.4, and 125 mg/kg, in accordance with OECD guideline 474. In a previous range finding study, 125 mg/kg was determined to be the maximum tolerated dose. Bone marrow was harvested at 24 hours from 6 animals/dose. At 48 hours, 12 additional animals were sacrificed, 6 for vehicle controls, and 6 for the 125 mg/kg dose level. Three additional animals were also dosed at 125 mg/kg to ensure survival of 6 animals for bone marrow extraction. Corn oil was the solvent control, and cyclophosphamide was the positive control. The test article induced signs of clinical toxicity as rough hair coat and/or hunched posture at 62.5 and 125 mg/kg. The test article did not induce statistically significant increases in micronucleated polychromatic erythrocytes (PCEs) when compared to vehicle control responses. There was a statistically significant decrease in PCE:NCE (normochromatic erythrocytes) ratio for the 125 mg/kg dose group, thus confirming that the test article reached and was cytotoxic to the bone marrow (Erexson, 2004).
• Balb/c mice were given 2.5 g of 13 nm silver nanoparticles or 2-3.5 µm silver microparticles, directly into the stomach, and the livers were examined 3 days later for histopathological analysis. [Authors did not report the number of mice or the number of exposures, the latter is presumably one]. Nano- and micro-silver exposed mice livers demonstrated lymphocyte infiltration, suggestive of inflammation. From a microarray analysis of the RNA from the livers, the expression of genes related to apoptosis and inflammation were confirmed to be altered, and these gene expression changes may lead to phenotypical changes resulting in increased apoptosis and inflammation. However, there was almost no reduction in mitochondrial activity for either the nano- or micro-silver. While DNA contents were decreased up to 18% for nano-particle exposed livers, and up to 10% for micro-particle exposed livers, the nano-group had no dose-dependency, the macro-group had only weak dose-dependency, and neither group expressed increases in glutathione production which is normally associated with increased oxidative stress. The authors also reported in vitro exposure of human hepatoma cells (Huh-7) to the silver nano- and micro-particles, and noted the mitochondrial activity and glutathione production was not appreciably affected with DNA contents decreased by 15% in the nano-particle treated cells, and 10% in the micro-particle treated cells [further details were not provided] (Cha et al., 2008). [This is not a recognized assay, genotoxic significance is not established].

In vitro and/or non-mammalian

• A Comet assay and analysis for bioaccumulation was performed in the polychaete, Nereis diversicolor. Groups of 3-5 worms were exposed to nominal concentrations of nanosilver (<100 nm, 99.5% metals basis, coated by 0.2 wt% PVP), micro-silver (2-3.5 µm, ≥99.9% trace metal analysis), and ionic silver (AgNO3) at 0, 1, 5, 10, 25, and 50 µg/Ag/g dry weight (dw) sediment for 10 days. There was one control worm per group, plus one positive control worm with cells extracted and exposed to UV light, and PVP-controls to examine if the coating was responsible for the genotoxicity of the nano-sized particles. DNA damage, measured as tail moment and tail DNA intensities, was dependent on dose and silver-form. Damage was significantly higher at 25 and 50 µg/g dw in nano- and micro-Ag treatments, and at 50 µg/g dw for the ionic silver, compared to controls. The presence of highly crystalline material was observed in nanosilver, suggesting the presence of large silver particles (aggregates, 20-200 nm, average 162 nm). For micro-silver, 5-10% of non-crystalline material was observed, suggesting it was not as pure as described by the manufacturer (i.e. <99.9% purity), and had both micro- and nano-sized particles (8nm – 3 µm). Reported silver body burdens for the nano-, micro-, and ionic-silver treatments were 8.56, 6.92, and 9.86 µg/g dw, respectively. These values correspond to BAF factors of 0.17, 0.14, and 0.20, respectively (Cong et al., 2011). [This Comet assay did not include controls for apoptosis, the relevance of the measured DNA strand breaks to heritable genetic damage is uncertain, and the relevance of this species to human health is not known].
• Lead (99.9% purity, 10 µm), Bismuth (99.9%, 10 µm), Indium (99.9%, 45 µm), Silver (99.9%, 10-20 µm) , and Antimony (99.9%, 10 µm) were tested for genotoxicity using a reverse mutation assay and a chromosome aberration test. Test substances were suspended in DMSO. Micro- silver in DMSO, was negative in the reverse mutation assay in Salmonella typhimurium strains TA100, TA1535, TA98, and TA1537, and in Escherichia coli WP2uvrA/pKM101, with and without metabolic activation at doses ranging from 313 µg/plate to 5,000 µg/plate, based on less than a two-fold increase in the mean number of revertant colonies compared to the negative controls. Microbial toxicity was not observed in any of the tester strains with or without S9 mix, although precipitates were found at silver doses of 2,500 µg/plate and higher. In the chromosome aberration test, the test substances were suspended in 1% sodium carboxymethylcellulose (CMC-Na), and exposed to Chinese hamster liver cells (CHL/IU cell line). Evaluations were performed for growth inhibition and chromosome aberrations, both structural and numerical. Micro-scale Silver was dosed at up to 5,000 µg/mL, with and without S9 mix. Results were negative based on < 5% of either type of aberration at any dose, in the presence and absence of S9. The IC50 for growth inhibition was >5,000 µg/mL with and without S9, suggesting the Silver was not cytotoxic to the CHL cells (Asakura et al., 2009). [Results for the lead, bismuth, indium and antimony are not summarized here because they are outside the scope of this paper].
• Kanematsu et al. (1980) tested 127 metal compounds, including silver chloride, silver nitrate, and silver sulfate, for DNA damage in Bacillus subtilis. Criteria for a positive finding consisted of a more pronounced inhibition of cellular growth with the recombination-repair-deficient (rec-) than with wild bacteria (rec+), indicative of DNA damage. Samples were exposed at 0.005-0.5M of each test substance and exposed per a cold incubation method. All three silver forms were reported as negative is this assay (Kanematsu et al., 1980). [Specific results and statistical analyses were not reported].
• Mutagenicity data for Silver sulfadiazine (AgSu) was found, however, not considered within the scope of this GreenScreen assessment (McCoy and Rosenkranz, 1978).
• Robison et al. (1982) examined DNA strand breaks in Chinese hamster ovary (CHO) cells. The cells were exposed in vitro to crystalline (1-4 µm) nickel sulfide, cobalt sulfide, cadmium sulfide, silver sulfide, copper sulfide and trinickel disulfide at 10 µg/mL for 24 hours. DNA strand breaks were determined by analysis of the number average molecular weight of DNA compared to controls. All the insoluble crystalline sulfides induced considerable reductions in the MW of the DNA. Authors refer to a secondary reference (Costa et al., draft paper) which postulates that these compounds phagocytosed by the CHO cells in response to the DNA breakage. Authors further suggest the strand breaks

Reproductive Toxicity (R) Score (H, M, or L): DG

Conventional silver was assigned a score of Data Gap for reproductive toxicity based on lack of mammalian data. While one study reports no decrease in fertility in male rats it was poorly reported and therefore cannot be used to determine a hazard score with adequate confidence.

• Authoritative and Screening Lists
• Authoritative: Not on any authoritative lists
• Screening: Not on any screening lists
  
• ATSDR (1990) reported no studies of reproductive toxicity from exposure to silver or silver compounds were found for inhalation, oral, or dermal exposure.
• The Japanese NITE classification for reproductive toxicity is “classification not possible” based on no data (NITE, 2006).
• There was no decrease of fertility in male rats exposed for life to drinking water containing 635-660 mg silver/day as silver chloride (Olcott, 1948).

Developmental Toxicity incl. Developmental Neurotoxicity (D) Score (H, M or L): DG