Sulphur dioxide

Endpoint summary

Key value for chemical safety assessment

Genetic toxicity in vitro

Description of key information

The available data on genetic toxicity allow a conclusive statement on the genetic toxicity for sulfur dioxide and sulfites. Irrespective of the reporting quality of the publications, both positive and 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 sulfur dioxide/sulfites do not show any clastogenic potential. The references discussed under in vitro clastogenicity are rated as not reliable due to experimental and reporting deficiencies and to not show a consistent pattern on the induction of chromosome and genome mutations. A high-quality in vivo study with sodium sulfite via subcutaneous injection in mice did not show an increase of micronuclei formation up to the maximum tolerated dose. This finding is supported by a negative dominant lethal test in rats after single and repeated oral administration (feed) in rats. A number of in vivo clastogenicity studies were assessed as being of limited reliability, since these exhibit reporting and/or other experimental deficiencies and lack biological plausibility. Overall, there is no consistent evidence of induction of genetic toxicity with relevance to humans for sulfites. Endpoint Conclusion: No adverse effect observed (negative).

Genetic toxicity in vivo

Description of key information

The available data on genetic toxicity allow a conclusive statement on the genetic toxicity for sulfur dioxide and sulfites. Irrespective of the reporting quality of the publications, both positive and 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 sulfur dioxide/sulfites do not show any clastogenic potential. The references discussed under in vitro clastogenicity are rated as not reliable due to experimental and reporting deficiencies and to not show a consistent pattern on the induction of chromosome and genome mutations. A high-quality in vivo study with sodium sulfite via subcutaneous injection in mice did not show an increase of micronuclei formation up to the maximum tolerated dose. This finding is supported by a negative dominant lethal test in rats after single and repeated oral administration (feed) in rats. A number of in vivo clastogenicity studies were assessed as being of limited reliability, since these exhibit reporting and/or other experimental deficiencies and lack biological plausibility. Overall, there is no consistent evidence of induction of genetic toxicity with relevance to humans for sulfites. Endpoint Conclusion: No adverse effect observed (negative).

Additional information

Introduction:

Substance-specific genotoxicity data for sulfur dioxide are sparse, largely related to the fact that most test systems (in particular in vitro) do not allow for a reliable exposure to such a gaseous substance. For this reason, extensive read-across is made to the group of sulfite substances, which represent the physical and physiological dissociation products upon contact of sulfur dioxide with aqueous media – note: under most physiological and environmental conditions (i.e. at neutral pH), sulfite and bisulfite (i.e. hydrogensulfite) are present at equimolar concentrations. Further, in consideration of the wealth of information of sulfite substances, older and irrelevant information has no longer been considered for lack of relevance, as explained in the introductory section directly below.

 

Introductory remarks on genetic toxicity:

According to the ECHA Guidance “Guidance on information requirements and Chemical Safety Assessment Chapter R.7a: Endpoint specific guidance” (Version 2.4, February 2014), existing data shall be qualified by its (i) adequacy, (ii) reliability and (iii) relevance. Especially the reliability of data shall be considered when assessing its usefulness for hazard/risk assessment purposes. The guidance suggested the rating according to Klimisch. According to this rating scheme, existing references are rated differently in case of (i) the use of different test guidelines (compared with today's standards) (ii) the inability to characterise the test substance properly (in terms of purity, physical characteristics, etc.) (iii) the use of crude techniques/procedures which have since become refined.

A number of OECD Guidelines on genetic toxicology testing have recently been withdrawn from the portfolio of OECD test guidelines, such as:

477: Sex-Linked Recessive Lethal test in Drosophila melanogaster

478: Genetic Toxicology: Rodent dominant Lethal Test

479: in vitro Sister Chromatid Exchange in Mammalian Cells

480: Saccharomyces cerevisiae, Gene mutation assay

482: DNA Damage and Repair, Unscheduled DNA synthesis in Mammalian Cells.

According to the qualification criteria for existing data stated above, tests conducted according to these guidelines have a minor contribution to the overall assessment, since newer and more up-to-date test guidelines exist. Existing tests conducted according to the more up-to-date guidelines were therefore rated with a higher reliability and were subsequently considered with a higher contribution to the overall assessment of genetic toxicity of the category substance. Further, tests which do not directly address the endpoint genic toxicity were considered of minor relevance, such as DNA damage in bacteria tested according to the rec assay. This assay only measures differential killing and is not a mutation assay.

 

 

Substance specific information for SO₂:

 

In-vitro:

In in-vitro studies, sulfur dioxide was negative in the Salmonella typhimurium mutagenicity test with TA98 and in tests for induction of DNA-single strandbreaks in fetal hamster lung cells or primary rat hepatocytes.

As it is questionable whether under the culture conditions used the cells had sufficient contact with the gas, it seems more reasonable to base the assessment of genotoxicity of sulfur dioxide in vitro on data with sulfites, which are the immediate dissociation product of sulfur dioxide upon contact with aqueous media. For this reason, extensive read-across (see further below) is made to the wealth of reliable information on various sulfite substances.

 

In-vivo:

In vivo studies in mice inhaling sulfur dioxide appeared to suggest clastogenic activity in bone marrow (formation of micronuclei and structural chromosomal aberrations) and DNA damage in a range of different cell types with LOAELs of 5 ppm. Test protocols of these studies were comparable to guideline studies but with the relevant limitation of no positive control group being included.

In contrast, in a recent, state-of-art in vivo micronucleus study in mice conducted according to OECD guideline 474 and GLP rated as reliable, no significant increase of micronuclei in polychromatic erythrocytes of the bone marrow of both male and female mice 24 h after the last exposure up to the highest dose of 30 Molppm was observed, at much higher exposure concentrations than in the study described above (Ziemann, 2008). In a follow-up study (Ziemann, 2008) which was specifically performed to demonstrate that sulfur dioxide was systemically available in the main study, it was shown that Malondialdehyde levels were dose dependently increased in mice exposed to sulfur dioxide with a statistically significant increase in those mice exposed to 10 and 30 Mol ppm sulfur dioxide, thus indicating that sulfur dioxide was systemically available and could reach the bone marrow compartment.[1]The outcome of this follow-up study was further supplemented by a personal written communication by Prof. Dr. Dr. Heinrich as the executive director at the laboratory (Fraunhofer) performing the test at that time. His statement addresses two important parameters that are relevant quality criteria of the study:

(i) Proof that bone marrow was exposed to SO₂: in the study report 17G07023 ("Extended Mammalian Erythrocyte Micronucleus Test with Sulfur Dioxide"), identical to the follow-up study 17G08010 ("Mammalian Erythrocyte Micronucleus Test with Sulfur Dioxide"), a concentration dependent increase of malondialdehyde in erythrocyte lysates of the periphery blood could be demonstrated (Ziemann et al., 2010), which was in males as well as in females at 10 and 30 ppm statistically significant. In the negative control animals (males and females), no malondialdehyde could be detected. This is an indication of specific, SO₂-caused oxidative stress and especially for systemic availability of SO₂and/or its degradation products/metabolites. Since an increase of malondialdehyde in the peripheral blood was evident, it can safely be concluded that bone marrow was exposed to SO₂as well since peripheral blood/plasma and bone marrow are in direct contact.

(ii) signs of general toxicity: please also see the explanation above; the significant increase of malondialdehyde in erythrocyte lysates can be interpreted as a clear sign of systemic oxidative stress and thus as a sign of general toxicity. In principle, this study was designed (concerning dose selection) in direct response to the study by Meng et.al. 2002. In the course of the study 17G07023 not only malondialdehyde but also the formation of sulf- and methaemoglobin were addressed. Due to the fact that the primary focus of the study was the micronucleus test, blood samples were taken only 24h after final SO₂exposure. Hence, no increased sulf- and methaemoglobin concentrations could be observed. Blood sampling prior to examination of the bone marrow may have caused a stimulation of haematopoiesis and thus falsify the micro nucleus test, and were thus omitted. Since methaemoglobin is known to have a half-life of about 1 hour it can be assumed that methaemoglobin generated by SO₂exposure could not have been detected after 24 hours any more due to complete degradation. Sulfhaemoglobin is more stable than methemoglobin but induces massive membrane damage and decrease of formability of the erythrocytes. It can be assumed that after 24 hours, damaged erythrocytes have been completely removed from circulation. Overall, sulf- and methaemoglobin values are unsuitable parameters for the determination of toxicity induced by sulfur dioxide.

 

Read-across concept for sulfites, hydrogensulfites, metabisulfites, dithionites and thiosulfates:

Sulfur dioxide is very soluble in water and forms – as an anhydride – sulfurous acid. Since all physiological processes within e.g. the human body are bound to proceed in aqueous solutions, a comprehensive read-across concept has been developed for sulfur dioxide, sulfites, hydrogensulfites and metabisulfites, based on the pH-dependant equilibrium in aqueous solutions which is summarised in the following equations:[1], [2]

          SO₂+ H₂O <->`H₂SO₃´        H₂SO₃<->H⁺+ HSO₃^-<->2H⁺+SO₃^2-   2HSO₃^-<->H₂O +S₂O₅^2-

Since the nature of the cation (i. e., sodium, potassium, ammonium…) is not assumed to contribute substantially to differences in toxicity and solubility (all compounds are very soluble in water), only the chemical and biological properties of the anion are considered as relevant determinants. Based on the described equilibrium correlations, unrestricted read-across between sulfur dioxide, the groups of sulfites, hydrogensulfites and metabisulfites is considered justified under physiological conditions.

Additionally, it is known that sodium dithionite disproportionates in water to form sodium hydrogen sulfite and sodium thiosulfate (equation II)2, [1], so that this substance can also be considered to be covered by the read-across concept described above. Since it can easily be anticipated that the substance is not stable enough under physiological conditions to fulfil the requirements of study guidelines, instead the products of decomposition must be considered:

      2 S₂O₄^2-+ H₂O→2HSO₃^-+ S₂O₃^2-

Not fully covered by this read-across concept is the substance class of thiosulfates: although the thiosulfates are also well known to disproportionate in aqueous solution to form polythionic acids and SO2(HSO3-), this requires somewhat different, more acidic conditions. Therefore, read-across to sulfites is primarily restricted to appropriate physiological conditions, i. e. oral administration where the gastric passage with the strongly acidic conditions in the stomach will facilitate the chemical disproportionation described above:

      HS₂O₃^-+ H₂S₂O₃→HS₃O₃^-+ SO₂+ H₂O

[1]Hollemann Wiberg, Lehrbuch der Anorganischen Chemie, 101.Auflage

[2]Handbook of Chemistry and Physics, Ed. Lide, DR, 88thedition, CRC Press

 

 

In vitro genetic toxicity:

Gene mutation in bacteria:

Disodium disulfite (aka sodium metabisulfite) was tested unequivocally negative independently by two authors in a bacterial reverse mutation assay (Simmon, 1978). S. Typhimurium tester strains TA 98, TA 100, TA 1535, TA 1537, TA1538 and E. coli WP2 uvrA were exposed to doses up to 10000 µg/plate (up to 10 doses) in the presence and absence of a metabolic activation system. Tests were conducted according to OECD 471 using the plate incorporation method. Cytotoxicity characterised by reduced background lawn was seen in strains TA 100, TA 1535 and E. coli WP2 at the two highest and the highest dose, respectively. In none of the assays performed was any significant increase of revertant colonies observed in any of the tester strains with or without metabolic activation. These references fulfil the requirements for chemicals hazard assessment. Experiments and results are presented in adequate detail, thus both references are considered reliable without restrictions (a guideline was not available and GLP not compulsory at the time of study conduct).

 

Ishidate (1984) also reported test results for (i) potassium metabisulfite (93.0% purity, aka potassium disulfite) (ii) sodium bisulfite, anhydrous (95.0% purity) and (iii) sodium sulfite, anhydrous (95.0% purity). S. Typhimurium tester strains strains TA 92, TA 94, TA 98, TA 100, TA 1535, TA 1537 were exposed at doses of (i) 3 mg/plate (ii) 5 mg/plate (iii) 50 mg/plate in the presence and absence of a metabolic activation system. Tests were conducted according to OECD 471 using the preincubation method. Cytotoxicity characterised by reduced background lawn was assessed in a separate experiment and the maximum dose for the main experiments selected accordingly. In none of the assays performed, a significant increase of revertant colonies was observed in any of the tester strains with or without metabolic activation. The publication is a summary paper, reporting testing of a total of 242 substances. The reference fulfils the basic requirements for scientific publications used in chemicals hazard assessment. Minor reporting or experimental deficiencies: maximum dose for sodium sulfite 10-fold above the recommended maximum dose, cytotoxicity not measured/reported for the main experiments, outcome on mutagenic effects only in binary format- no individual data.

 

In a further testing programme, four sulfites were assessed for their potential to induce gene mutations in the bacterial reverse mutation assay (Engelhardt 1989a, b, c, d, e). The tests were conducted in accordance with OECD TG 471, as amended at the date of study conduct. S. typhimurium strains TA 98, TA 100, TA 1535, TA 1537 were exposed using the preincubation and plate incorporation method to sodium sulfite, sodium dithionite, potassium sulfite, disodium disulfite and potassium metabisulfite as doses of 0, 20, 100, 500, 2500, and 5000 µg/plate. Cells were incubated for 48 hrs with and without metabolic activation system. Slight cytotoxic effects were observed with potassium sulfite for TA 100 at 5000 µg/plate and with disodium disulfite for TA 100 at 2500 and 5000 µg/plate. No increase in the number of revertants was observed for any of the test substances using the plate incorporation or preincubation protocol up to the maximum concentrations with or without metabolic activation. Reference is considered reliable without restrictions.

 

___________________________________________

[1]Meng Z, Zhang L. (1990) Chromosomal aberrations and sister-chromatid exchanges in lymphocytes of workers exposed to sulphur dioxide. Mutat Res 241: 15-20

Meng Z, Zhang L. (1990) Observation of frequencies of lymphocytes with micronuclei in human peripheral blood cultures from workers in a sulphuric acid factory. Environ Mol Mutagen 15: 218-220

Meng Z, Qin G, Zhang B. (2005) DNA damage in mice treated with sulfur dioxide by inhalation.Environ Mol Mutagen 46: 150-155

___________________________________________

 

Sodium thiosulfate was tested in the bacterial reverse mutation assay for the induction of mutations in S. typhimurium TA 1535, TA 1537, TA 1538, TA 98 and TA 100 and E.coli WP2 uvrA(Mortelmans, 1979). Cytotoxicity was assessed in a dose range finding experiment with TA 100 at ten concentrations between 0.3 and 10,000µg/plate, no toxicity was observed in this assay. In the main experiment all five strains were tested using the plate incorporation protocol and incubated for 48 or 72 hours at 0, 33.3, 100, 333.3, 1000, 3333.3, 10,000 µg/plate with and without metabolic activation. Positive control substances were applied and showed a significant increase in revertants. No cytotoxicity or increase in revertant rate was observed in any of the strains with or without metabolic activation.

 

In a short communication Münzer, R. (1980) reports on a bacterial reverse mutation test with sodium bisulfite using strains TA 98, TA 100, TA 1535, TA 1538. Cell were exposed in an acidified medium (pH 5.9) with a single concentration of 1M sodium bisulfite. No results could be obtained due to the marketed bactericidal effect at the culture conditions (data not shown). Due to the lack of any raw data and due to the non guideline compliant culture conditions, this report is considered of no relevance for hazard assessment purposes.

 

The mutagenic potential of disodium disulfite was assessed in S. typhimurium TA 97 (Pagano, 1990). Cells were exposed using the preincubation protocol to 80mM (equivalent to 8.32 g/L) sodium bisulfite at different pH and temperatures. In addition various buffer additives were added to assess whether any mutation frequency is increased or reduced. No increase of the mutation frequency was observed at standard culture conditions (37°C, pH 7, without buffer additives). In the results section of the publications, the mutagenic potential of sodium bisulfite is reported, whereas in the materials section, disodium disulfite is stated as test substance. The strain TA 97 is known for its genetic instability and its limited used for bacterial reverse mutation testing, being the reason for its replacement by the strain TA 97a. Based on the unclear test item used in this experiment, the experimental design and unsuitable strain, this reference is of no relevance for the chemicals hazard assessment.

 

Pagano, D.A. and Zeiger, E. (1987) assessed the mutagenic potential of sodium bisulfite in a range of seventeen S. typhimurium strains. Cells were exposed at concentrations of 0-1.0 M (0-104 g/L) at varying pH values of 5.0 – 8.0, results not reported for all strains. The strains were preincubated for 30 minutes at 37°C, plated and incubated for 48 hrs at 37°C. Authors report a weak increase of revertants at 0.1M sodium bisulfite in one strain in an initial experiment (results only shown in graph, thus no absolute fold-increase can be given). The overall test design is difficult to follow, since it remains unclear at which culture conditions the results were obtained. Authors report revertants in absolute numbers per plate but also as relative percent increase compared with the control culture. Results are only given for a sub-set of strains without indications on toxicity – authors state toxicity in all strains at and above 0.3 M (31.2 g/L) although revertant rates are reported for 1.0 M (104 g/L). Due to the confusing reporting and unsuitable test design with regard to culture conditions and drastic exposure concentrations, this publication is considered of no relevance for chemicals hazard assessment.

 

in vitro mammalian cell gene mutation:

Disodium disulfite was assayed for the ability to induce mutation at the hypoxanthine-guanine phosphoribosyl transferase (hprt) locus (6-thioguanine [6TG] resistance) in mouse lymphoma cells (Stone, 2010). The test was conducted according to OECD 476 and under GLP. In cytotoxicity dose-range finding experiment, 6 concentrations were tested in the absence and presence of S9 ranging from 59.44 to 1902 µg/mL (equivalent to 10 mM at the highest concentration tested). The highest concentration analysed gave 37% and 50% RS in the absence and presence of S9, respectively. In the first main experiment, in the presence of S9, statistically significant increases in mutant frequency were observed at the highest 2 concentrations (1600 and 1902 µg/mL) but not showing a linear trend. This positive finding was not reproduced in the first main experiment without metabolic activation as well in the second main experiment with or without metabolic activation. In order to verify the isolated positive finding, a confirmatory experiment in the presence of S9 was performed. In Experiment III no statistically significant increases in mutant frequency were observed following treatment with disodium disulfite at any concentration tested and there were no significant linear trends, indicating a negative result. It is concluded that disodium disulfite does not induce mutations at the HPRT locus of L5178Y mouse lymphoma cells when tested with and without metabolic activation up to the limit dose of 1902µg/mL (10mM) as foreseen by the test guideline.

 

Mallon (1981) investigated sodium bisulfite in V79 cells for inductions of gene mutations. Cells were treated with 10 and 20 mM in a short-term exposure (15 minutes) or with 1 and 5 mM in a long-term exposure (48 hours). The toxicity of bisulfite was measured by assaying the clonal survival after exposure to the test substance for 15 minutes in PBS. UV-light (GE germicidal lamp G1578) at a fluence of 1.8 J/m² was used as positive control. Cytotoxicity was observed at concentrations above 20 mM. No increase in mutation frequency was observed after short or long-term exposure. The reference exhibits deficiencies in study design, such as non-guideline complaint exposure duration, culture conditions and selection of positive control substance, the genetic stability and sensitivity of the test system was not demonstrated. Due to the aforementioned shortcomings, the reference is considered of not relevance for the chemical hazard assessment.

 

In an in vitro gene mutation test in AS52 (CHO derivative), the induction of point mutations in the XPRT gene was assessed (Meng, 1999, reference is identical with Meng, Z (1997) Molecular analysis of spontaneous and bisulfite-induced gpt mutants in Chinese hamster ovary cells. China Environ. Sci. 17, 171-175). Cells were exposed with 0, 5 and 10mM sodium bisulfite, ethyl methanesulfonate was used as positive control substance. Cells were exposed for 4 hours in triplicate and sub-cultured for the expression for 7 days. Cytotoxicity was expressed as percent survival relative (RS) to those from similarly plated untreated control. Cells treated with sodium bisulfite showed an increase in mutation frequency with a relative survival of 34.5% compared with the untreated control. The increase of MF was within (at 5mM) or slightly above (at 10mM) the range of historical control values by other labs. The absolute relative survival ranged from 70-90%. The high variability in RS in untreated cultures is unusual and indicates difficulties in standard cell-line culture conditions in this lab. Authors state that solid sodium bisulfite was used in the mutation experiments (purity not stated, not checked for pH effects). However sodium bisulfite is only stable in aqueous solution and cannot be isolated in a solid form. Consequently, the test item used in these experiments cannot be verified and the reference is rated as not reliable.

 

in vitro mammalian chromosome aberration/micronucleus test:

In a combined study, Meng (1992, reference is identical with Meng, Z.; Zhang, L (1994). Chromosomal aberrations, sister chromatid exchanges and micronuclei induced in human lymphocytes by sodium bisulfite (sulfur dioxide) Acta Genetica Sinica 21, 1-6) investigated clastogenic, aneugenic and DNA damaging effects in human lymphocytes exposed to sodium bisulfite. Results on DNA damage are discussed below. Peripheral blood lymphocytes were taken from young male and female donors and incubated under addition of a mitogen (PHA), pre-culture duration not stated. A sodium bisulfite stock solution with a pH or 7 was used for all experiments. Mitotic index was calculated on the basis of 2000 lymphocytes. For the analysis of chromosome aberrations, cells from 4 different donors (sex not stated) were incubated for 48 hours at concentrations of 0, 0.05, 0.1, 0.5, 1, 2 mM. Colcemid was added 4 hours prior harvest. Cells were Giemsa stained after hypotonic treatment and drying. 200 cells per group were scored blindly for the presence of isochromatid and chromatid breaks. The dose of 2 mM could not be scored due to excessive toxicity. A dose dependent increase of chromatid breaks was observed, being significant at 0.1mM and above. None of the reported concentrations caused excessive toxicity, i.e. MI > 50% relative to untreated control. Only the maximum dose of 2 mM caused complete toxicity. No explanation was given why further structural aberrations such as fragments, deletions or exchanges were not scored. The results indicate a clastogenic effect of sodium bisulfite at concentrations with moderate cytotoxicity. In the micronucleus experiments, cells were treated under identical conditions compared with the chromosomal aberration experiments, except for the extended incubation period of 72 hours. CytoB was added to the cultures 24 hours prior harvest. Binucleated cells were scored for the presence of micronuclei. Total number of binucleated cells scored per culture was not given. A dose dependent increase in the micronucleus frequency was observed, with a statistical significance at 0.1 mM and above for 2 donors and at 1 mM for the further 2 donors. The mean values for all donors show a statistical significance at 0.5 mM and above. The total frequency of binucleated cells in all cultures was still within the range of untreated control MN frequency in other labs (i.e. 0.7-2.0%), positive controls were not used in the experiments, which would allow a statement on the specificity and sensitivity of the test system.

 

The experimental procedures used in the above experiments are not in accordance with accepted guidelines and the findings raise questions whether a true clastogenic effect was observed. Authors only investigated chromatid breaks in the CA experiments. This type of aberrations is the easiest to be analysed, whereas fragments, deletions and exchanges require more experienced evaluators. This raises questions about the overall experience of the lab with the evaluation of chromosome aberration experiments. The exposure duration of 48 hours for the chromosome aberration and 72 hours for the micronuclei experiments is too long even under conditions of continuous treatment. Usually times representing 1.5 cell cycle lengths (i.e. approx. 24 hours for human lymphocytes) are used in those experiments. No exposure durations of 3-6 hours pulse treatment were used. Further, the mitotic index was assessed in a separate experiment, for which the incubation duration was not given. Consequently, the mitotic indices cannot directly be correlated with the conditions during mutation experiments; hence it remains unclear whether clastogenic effects were caused via unspecific cytotoxicity or a substance specific effect. Authors state that solid sodium bisulfite was used in the mutation experiments (purity not stated, not checked for pH effects). However sodium bisulfite is only stable in aqueous solution and cannot be isolated in a solid form. Based on the above given shortcomings, the positive findings should be considered with great caution.

 

Chinese hamster fibroblast cells (CHL)were tested for the induction of structural aberrations (Ishidate, 1984). Cells were exposed towards (i) potassium metabisulfite (93.0% purity, aka potassium disulfite) (ii) sodium bisulfite, anhydrous (95.0% purity) and (iii) sodium sulfite, anhydrous (95.0% purity)at max. 0.06,0.125 and 0.5 mg/mL, respectively (3 doses, concentrations not reported) for a period of 24 and 48 hours without metabolic activation. Colcemid was added 2 hours before cell harvest and 100 well-spread Giemsa stained metaphases were checked for incidence of polyploid cells as well as of cells with structural chromosomal aberrations such as chromatid or chromosome gaps, breaks, exchanges, ring formations, fragmentations. No data on cytotoxic effects were reported. No increase of structural chromosome aberrations was reported. The publication is a summary paper, reporting clastogenic effects of 242 substances. The resulting reporting detail is very limited, stating the basic experimental parameters and results in a binary format only; only results for 48 hrs exposure duration reported. Authors state that solid sodium bisulfite, anhydrous was used in the mutation experiments (not checked for pH effects). However sodium bisulfite is only stable in aqueous solution and cannot be isolated in a solid form. Based on the above limited reporting detail and the questionable test item, the reference is considered as not reliable.

 

Popescu (1988) tested the clastogenic effect of sodium bisulfite in Syrian hamster foetal cells (HFC) at concentrations of 10, 20 and 40 mM (equivalent to 1.04, 2.08, 4.16 g/L) for an exposure duration of 15 minutes. Chromosomes were prepared 6 and 24 hours after exposure by adding Colcemid 4 hours before cells harvest and Giemsa staining. 200 metaphases were analysed for chromatid or chromosome aberrations (gaps, breaks and exchanges). Chromosome aberration analysis at first (6 hours) and second (24 hours) mitosis showed no significant increases in aberrations over the control.There are deficiencies in performance and reporting of the method: No positive control was used. No data if duplicate cultures were used. Exposure period was too short for the detectable manifestation of chromosomal damage. Such damages require the completion of at least one whole cell cycle – which is not the case after 15 minutes exposure. Chromosome preparation, staining and analysis of the slides were not described in detail. Authors do not provide details on the test item, such as origin, purity, physical state, impurities, pH effects. Based on the above limited reporting detail and the questionable test item, the reference is considered as not reliable.

 

The induction of chromosomal aberrations (CA) and micronuclei formation (MN) by potassium disulfite (99.9% purity, aka potassium metabisulfite) was tested in human peripheral blood lymphocytes (Yavuz-Kocaman, 2008). Cells from 4 healthy donors (2 females, 2 males) were used and incubated for 72 hours prior to exposure. The exposure concentrations were 25, 50, 100 and 200 µg/mL (0.13, 0.26, 0.56, 1.05 mM) for a duration of 24 and 48 hours. In the CA experiments, colchicine was added 2 hours before harvesting; ethyl methanesulfonate was added as positive control. Cells were harvested, fixed and prepared slides were stained (5% Giemsa); 100 metaphases per donor were scored for structural and/or numerical aberrations (excluding gaps). In the MN experiments Cytochalasin B was added at 44 hours of incubation to block cytokinesis. After additional 24 hours incubation at 37°C, cells were harvested, fixed and stained for MN analysis. 2000 binucleated lymphocytes were scored from each donor (8000 binucleated cells were scored per concentration). Potassium disulfite increased the percentage of MN over all concentrations and time points (by 2-3fold) but not in a dose-dependent manner. The mitotic index was increased in a dose dependant manner, cytotoxicity was <50%. Positive findings in the MN assay do not show a clear dose-dependency at both time points. Thus, a clear positive outcome was not demonstrated and in comparison with the findings of the CA assay described below is implausible. Potassium disulfite induced structural CA in a dose and time dependant manner, when compared with the negative control (4-6fold after 24 and 4-10fold after 48 hours).The control frequencies of CA were high (2.5%) compared to normal levels (around 0.5-1.5%). The mitotic index (MI) was reduced by 55 and 59% at 100 and 200µg/mL after 24 hours and by 69% at 200 µg/mL after 48 hours. The assay shows a dose- and time-dependent increase of structural CA. This is an unusual finding, which was also demonstrated by the authors in a parallel in vivo clastogenicity experiment which is discussed further below. Although the CA findings appear robust, there are irregularities in study conduct and reporting. Further the credibility of the testing facility based on unusual or biologically impossible results, as discussed in detail further below is compromised. Based on the questionable results, the results reported in this reference are considered not reliable.

 

For the conduct of an in vitro chromosome aberration test Beckman et al. (1986) used peripheral blood lymphocytes taken from two different non-smoking, healthy individuals (sex not stated). The cells were cultured for 72 hours and exposed to the test substance for 48 hours at a concentration of 0.4 mM sodium hydrogensulfite. Chromosome aberrations were scored in 400 cells. Statistical significance was determined by a pairwise comparison with background cultures. Authors state that a large number of the initial assays failed due to cytotoxic effects. The number of chromosome aberrations was increased, compared with the background frequency. There are a number of relevant deficiencies in performance and reporting of the method, which renders the publication not reliable: Authors state that the test substances induce significant cytotoxicity, however no data was given to prove this; consequently a correlation of the aberration frequency with cytotoxicity is not possible. The control (background) number of chromosome aberrations of 3% is unusually high (12 per 400 cells), which raises questions about the suitability of the test system. The culture conditions are insufficiently described, hence it is unclear how peripheral blood lymphocytes were cultivated without the use of a mitogen to induce cell proliferation. Secondly, it is unclear how the authors counted at least 400 cells per culture without the use of a spindle poison (e.g. colchicine) to arrest the cells in the M-phase with visible chromosomes. Authors state that solid sodium bisulfite was used in the mutation experiments (purity not stated, not checked for pH effects). However sodium bisulfite is only stable in aqueous solution and cannot be isolated in a solid form. Based on the above limited reporting detail and the questionable test item, the reference is considered as not reliable.

 

Abe (1977) tested potassium disulfite in a Chinese hamster cell line for the induction of chromosome aberrations at doses of 0.1, 0.5 and 1 mM. For a given dose at least one culture was made. HBSS was used as vehicle and solvent control.Cells were incubated for26 hours at 37°C in the dark (two cell cycles). 0.25 µg colchicine/ml was added 2 hours prior harvest.Fixed and stained by the fluorescence or Giemsa staining technique and chromosome aberrations were examined on 100 metaphases for each dose, and frequency of aberrations, excluding gaps, was indicated by the number of breaks per cell. Mitotic index (MI) decreased by more than 50% compared with the solvent control value at 0.5 and 1 mM.Neither a significant increase of aberrations nor a dose dependency was observed.The publication is a summary paper, reporting clastogenic effects of 33 substances. The resulting reporting detail is very limited, stating the basic experimental parameters and results in a binary format only. The decrease of MI below 50%, indicate that the maximum tolerated dose was exceeded in the two highest doses, which invalidates the results obtained from these doses. Based on the above limited reporting detail and questionable test results, the reference is considered as not reliable.

 

Human peripheral blood lymphocytes were tested for the induction of chromosome aberrations after sodium hydrogen sulfite exposure (Nordenson, 1984).Thestimulated lymphocytes were cultured for 70-72 hours at 37°C, during the last 48 hours incubation the cells were exposed to 0.375 mM sodium hydrogen sulfite. Cell were arrested by addition of Colcemid, fixed and stained. 200 cells per culture from coded slides were analysed for the presence of chromosomal aberrations. The classes of aberrations scored are unclear – it is possible that some of these abnormalities were gaps, which are normally not considered in the assessment of a positive response. No significant increase was observed in the number of aberrations. The reference is not suitable for risk assessment purposes, since relevant details were not reported/measured, such as only single dose treatment conducted which is not suitable for an assessment of dose-response relationship, cytotoxicity was not determined, only one time point was investigated.

 

Summary entry – investigations on DNA damage in vitro:

Several studies on in vitro DNA damage studies in mammalian cells (involving comet assays) were identified which do not fulfil the relevance, reliability and adequacy criteria as foreseen by the ECHA guidance on information requirements, chapter R.7.7. Most of the references are on mechanistic investigations to examine specific mechanisms of sulfites and are not suitable to address the endpoint of in vitro genetic toxicity. The studies are discussed below in brief for information purposes only and were included in the IUCLID as summary entry.

Chen (1994): Investigated the DNA base substitution in transfected cells followed by isolation and sequencing. Cells were incubated for a maximum of 54 days at sodium bisulfite concentrations up to 50mM. The test design is not suitable for the detection of potentially heritable DNA mutations, since e.g. sensitivity and specificity was not determined, the concentrations and test duration were not compliant with accepted guidelines.

Peden (1982): Published a mechanistic study on in vitro DNA base-substitutions in deletion loops. Sodium bisulfite was used as deamination reagent of the DNA bases at 2 mole/L. The test design is not suitable for the detection of potentially heritable DNA mutations, since e.g. sensitivity and specificity was not determined, the concentrations and test duration were not compliant with accepted guidelines.

MacRae (1979): Investigated the induction of sister chromatid exchanges in CHO cells after sodium bisulfite exposure at concentrations between 0.03 to 7.3 mM after 2 and 24 hrs exposure duration. No information on cytotoxicity available. Sister chromatid exchange detects unspecific DNA damage in mammalian cell in it therefore not a method to assess the mutagenic potential of a substance. Without information on the cytotoxic effects caused by the test item, an interpretation of the SCE frequencies is not possible.

Meng (1992): Investigated the induction of sister chromatid exchanges in human peripheral blood lymphocytes after sodium bisulfite exposure at concentrations between 0.05 to 2 mM after 72 hrs exposure duration. No information on cytotoxicity available. Sister chromatid exchange detects unspecific DNA damage in mammalian cell in it therefore not a method to assess the mutagenic potential of a substance. Without information on the cytotoxic effects caused by the test item, an interpretation of the SCE frequencies is not possible. The exposure duration is not further justified – cells are usually harvested after the completion of one cell cycle.

Doniger (1982):Study investigates substance induced DNA lesions and DNA replication rate in mammalian cells. No increase in DNA lesions was observed via gradient sedimentation. The reference exhibits several reporting deficiencies: a correlation between DNA damage and cytotoxicity was not performed; hence it remains unclear whether any positive findings were secondary to toxicity or due to direct substance interaction. Further, the amount of cells counted for DNA repair was not stated, only one concentration used which does not allow a dose-response analysis.

Popescu 1988: Investigated the induction of sister chromatid exchanges in Syrian hamster foetal cells (HFC) after sodium bisulfite exposure at concentrations of 10, 20 mM after 15 minute exposure duration. No information on cytotoxicity available. Sister chromatid exchange detects unspecific DNA damage in mammalian cell in it therefore not a method to assess the mutagenic potential of a substance per se. Without information on the cytotoxic effects caused by the test item, an interpretation of the SCE frequencies is not possible. It remains unexplained how any sister chromatid exchange could be seen after 15 minutes exposure. The manifestation of such damage requires the completion of a whole cell cycle, which is biologically impossible after 15 minutes exposure.

Yavuz-Kocaman (2008): Investigated the induction of sister chromatid exchanges in human peripheral blood lymphocytes after sodium bisulfite exposure at concentrations 25, 50, 100 and 200 µg/ml after 24 and 48 hrs exposure duration. Sister chromatid exchange detects unspecific DNA damage in mammalian cell in it therefore not a method to assess the mutagenic potential of a substance. The publication is discussed in the sections on in vitro mammalian chromosome aberration and in vivo chromosome aberration, clearly showing that the results should be treated with great caution (see discussion further below).

Beckman (1986): Investigated the induction of sister chromatid exchanges in human peripheral blood lymphocytes after sodium bisulfite exposure at a concentration of 0.4mM after at least 48 hrs exposure duration. No information on cytotoxicity available. Sister chromatid exchange detects unspecific DNA damage in mammalian cell in it therefore not a method to assess the mutagenic potential of a substance. Without information on the cytotoxic effects caused by the test item, an interpretation of the SCE frequencies is not possible. The exposure duration is not further justified – cells are usually harvested after the completion of one cell cycle. A single dose experiment does not allow the determination of a dose-response relationship.

 

Summary entry - unsuitable test systems in vitro:

Several studies were identified which do not fulfil the relevance, reliability and adequacy criteria as foreseen by the ECHA guidance on information requirements, chapter R.7.7. DNA damage in bacteria, induction of SCE in mammalian cells, or tests in yeasts or drosophila are no longer recommended as part of regulatory testing by many agencies worldwide and there are no up-to-date OECD guidelines for their conduct. The majority of the references in the following summary entry are more than 30 years of age and do not comply with today’s standards in genetic toxicity testing. Interpretation of the relevance of both positive and negative results from such tests is therefore unclear and was not used for the current assessment. The studies are discussed below in brief for information purposes only and were included in the IUCLID as summary entry.

A range of references assessed the mutagenic potential of sulfites in bacteria (Kunz, 1983; Mukai, 1970; Münzer, 1980; De Giovanni-Donnelly, 1985, Pagano D.A. et al., 1990). The cells were exposed to excessively high concentrations of sulfite up to 1M at acidic culture conditions (pH 5.7 – 5.2). The incubation duration was 0 to 30 minutes. All references reported significant cytotoxic effects, depending on the exposure duration. The aim of these studies was to assess the bactericidal effects of sulfites but not the mutagenic effects of this substance class. Due to the unsuitable test design with regard to culture conditions, drastic exposure concentrations and duration these references are considered of no relevance for chemicals hazard assessment.

Valencia (1973): Tested the genotoxic effects of sodium sulfite in Drosophila.

Jagiello (1975): Investigated the induction of chromosomal damage in oocytes freshly isolated from mice, ewe and cows. Cells were incubated for 5 and 15 hours for mice and 28 hours for ewe and cow. The number of cells per dose ranged from 6 to 96 – the confidence level at such low cell numbers is significantly impaired and does not allow a reliable evaluation of the clastogenic potential. It remains unclear why such an unusual test system was chosen for these experiments. There is currently no known mutagen which exclusively induces gene mutations in germ cells but not in somatic cells.

Sodium sulfite and sodium bisulfite were tested in a bacterial forward mutation test in micrococcus pyogenes var, M. aureus using a preincubation method (Clark 1953) at a single concentration. A forward mutation assay is not indicative for a specific substance-induced mutagenic response. Use of a single strain, single concentration test system is not suitable to determine a reproducible dose-response relationship.

 

In vivo genetic toxicity tests

in vivo chromosome aberration/micronucleus tests:

The substance sodium sulfite (purity 98.1%) was tested for chromosomal damage (clastogenicity) and for its ability to induce spindle poison effects (aneugenic activity) in NMRI mice according to OECD 474 and under GLP (Schulz, 2008). The substance was administered once subcutaneously to male animals at dose levels of 250, 500 and 1000 mg/kg bw. The vehicle water was used as negative control and cyclophosphamide and vincristine as positive control substances.24 and 48 hours after administration animals of all dose groups and the highest dose group were sacrifices, respectively. A minimum of 2000 fixed and stained polychromatic erythrocytes were analysed for the presence of micronuclei. Bone marrow toxicity was determined by analysing the PCE/NCE ratio. The administration of the test substance led to clinical signs of toxicity at the highest administered dose of 1000 mg/kg body weight. No statistically significant increases in the micronucleated erythrocyte frequency were observed in any evaluated test substance-treated group of animals at either time point. There were no statistically significant decreases in the PCE/NCE ratio in any test-substance treated group. Under the conditions of this study, sodium sulfite did not induce formation of micronuclei in rat bone marrow up to the dose of 1000 mg/kg bw via subcutaneous route of administration up to doses showing signs of systemic toxicity.

In a short communication published by Generoso (1978), the clastogenic effects of sodium bisulfite (purity not stated) was investigated via dominant lethal assay in male and female mice. Male (101 X C3H)F1 mice were administered intraperitoneally 20 times doses of 300 mg/kg bw/day during a 26-day period and 38 times during a 54-day period. Males were paired with two (SEC X C57BL)F1 females after the last injection. Female (C3H X 101)F1 mice were intraperitoneally administered a single dose of 550 mg/kg bw/day and mated with untreated male (101 X C3H)F1 mice. No mortalities were seen in the 550mg/kg dose group, whereas 8 out of 46 and 5 out of 46 died in the 400 and 300 mg/kg bw/day dose group, respectively. Sodium bisulfite did not induce a detectable increase in dominant-lethal mutations in either male or female germ cells of mice. It is concluded that sodium bisulfite does not induce chromosome aberrations in the mouse germ-cell stages tested. The publication is however only a brief short communication. The resulting reporting detail is very limited, stating only the basic experimental parameters and results. Mortalities observed in the repeat-dose experiments clearly indicate that the maximum tolerated dose was exceeded, consequently results from such experiments are considered not reliable. Authors state that solid sodium bisulfite was used in the mutation experiments (purity not stated, no information on toxicity). However sodium bisulfite is only stable in aqueous solution and cannot be isolated in a solid form. Based on the above limited reporting detail, the questionable test item and the inappropriate dose selection, the reference is considered as not reliable.

Yavuz-Kocaman (2008) investigated the induction of chromosomal aberrations in the bone marrow of albino rats. Four animals (2M/2F) per group were given a single intraperitoneal injection of 150, 300, 600 mg/kg bw potassium disulfite (aka potassium metabisulfite, purity not stated) 12 and 24 hours before sacrifice. Colchicine was given 2 hours prior sacrifice via i.p. injection. Urethane was used as positive control substance. Bone marrow of femurs was fixed and stained for microscopic analysis. 100 metaphases per animal (400 per dose group) were scored. Mitotic index was determined by scoring 3000 cells from each animal. The mitotic index showed a dose-dependent cytotoxic effect, dropping form 3.6 in the control to 1.85 at 600 mg/kg bw (51%) after 12 hours – demonstrating no excessive toxicity. A dose and time dependant increase for structural chromosome aberrations was observed (not further specified which type of aberrations were analysed). There were statistically significant increases in structural CA frequency at the top dose (2.5-fold increase) after 12 hrs, and at all 3 doses (2 to 4-fold increases) at 24 hrs. Although this appears to be a clear positive finding, there are reasons to question the reliability and relevance of the results, namely:

·        The intraperitoneal route is not considered a physiologically relevant route of exposure, since this route avoids first-pass metabolism by the liver which is known to effectively eliminate sulfite by oxidation to sulfate via molybdenum cofactor by normal physiological routes (oral, inhalation, dermal). Since redox damage is considered to exhibit a threshold for genotoxic effects, by-passing of anti-redox defence mechanisms would lower the threshold and allow genotoxic effects to be manifest which would not be seen if a physiological route of administration had been used. Clearly negative results in a recent GLP and guideline study with much higher doses of sodium sulfite but also bypassing normal physiological routes of administration (subcutaneous route; Schulz, 2008) are in contradiction to the positive findings of Yavuz-Kocaman (2008).

·        the number of animals per dose group is far too low to ensure statistical robustness; the guideline foresees at least 5 animals per dose and sex.

·        Negative control CA frequencies were very high (around 5.5%). CA frequencies in the bone marrow PCE of control animals (rats and mice) are usually much lower, i.e. at 0-2%. The fact that such high control frequencies were seen in this study brings into question the health of the animals and the competency of the people who scored the slides.

·        As a major criticism, the dose- and time-related dose responses are most unusual: different sampling times were included in this assay to study the effects on different stages of the cell cycle. There are few publications where both time- and dose-related CA responses have been studied with known genotoxins. Some clearly show dose-related increases in cells with CA at one or more sampling times, but this is not always the case as the following papers indicate:

                         i.      McFee and Tice (1990) demonstrated time- and dose-related increases in bone marrow CA for the potent genotoxins mitomycin C and cyclophosphamide, but although DMBA gave dose-related increases in CA at 18 and 26 hr it did not induce CA at 10 hrs.

                       ii.      Tates & Natarajan (1976) observed a clear dose response for CNU-ethanol-induced CA in bone marrow 1 day after ip dosing, but after 2 days there was no clear dose-response.

                      iii.      Aydemir & Bilaloğlu (2003) showed that the anti-cancer drug topotecan induced dose-related CA at both 6 and 24 hr after a single ip dose. However, although gemcitabine induced dose-related increases in bone marrow CA at 24 hr, it did not induce any CA at 6 hr.

·        There are also numerous examples of genotoxins inducing micronuclei (MN) at some sampling times but not at others. For example Sutou et al (1990), reporting on a collaborative study by the MMS subgroup of JEMS, showed that ethylnitrosurea, ethyl methanesulfonate, ARA-C and benzene all induced dose-related increases in bone marrow MN at 6 and 24 hr but not at 48 or 72 hr sampling times after 2 or 4 doses. Thus, potent genotoxins may show dose-related increases in CA at several different sampling times, but in many cases they only show dose-related effects at limited sampling times due to cell-cycle specific effects, death of damaged cells etc. Given the overall genotoxic profile of sulfites (i.e. they do not behave like many potent mutagens that are positive in multiple test systems) it is unlikely that it would be expected to produce dose-related increases in CA at multiple sampling times. It is strange that dose-related increases in numbers of polyploid cells were also seen. For a substance to induce polyploidy, cells need to undergo at least one mitotic division, which in a fast-dividing mammalian cell takes at least 20-24 hours. It is therefore biologically impossible to observe polyploidy as early as 12 hrs after dosing with potassium disulfite. Thus, although there appears to be a significant induction of CA by potassium disulfite in this study, there are concerns, as discussed above, which render this publication not reliable for risk assessment purposes.

 

In the study published by Kayraldiz, A., Topaktas, M. (2007) a similar study design as used by Yavuz-Kocaman was used. Albino rats (3M/3F per group) received a single oral administration and intraperitoneal injection of 250, 500, 750, 1000 mg/kg bw potassium disulfite (aka potassium metabisulfite, purity not stated) 6, 12 and 24 hours before sacrifice. Colchicine was given 2 hours prior sacrifice via i.p. injection. Ethyl carbamate was used as positive control substance. Bone marrow of femurs was fixed and stained for microscopic analysis. 100 metaphases per animal were scored. Mitotic index was determined by scoring 3000 cells from each animal. The mitotic index showed a dose-dependent cytotoxic effect, data for control animals not presented, thus not comparison possible. The drop of the MI between the low and high dose group i.p. administered animals to 39% shows that the MTD was exceeded. A dose and time dependant increase for structural chromosome aberrations was observed. According to the authors frequency of chromosomal aberrations were increased in all concentrations and treatment periods, however results for the negative and positive control group was not included. The language of this reference is difficult to read and lacks precision, so that the experimental procedure and the presentation of the results is confusing. Although this appears to be a clear positive finding, there are reasons to question the reliability and relevance of the results, namely:

·        The number of animals per dose group is too low in order to demonstrate statistical robustness; the guideline foresees at least 5 animals per dose and sex.

·        It appears implausible that animals tolerate the identical dose via two different routes of application. Although the oral absorption is quantitative, a delayed bioavailability is expected. Via i.p. administration, almost the complete dose becomes immediately systemically available and is not metabolised via first pass effect in the liver. Consequently it is expected that the i.p. route is less well tolerated. The same group assessed the MTD via i.p. route in a similar experiment using the same rat strain at 600 mg/kg bw (Yavuz-Kocaman, 2008). A similar MTD of 550 mg/kg bw was determined by another group (Generoso, 1978). It appears implausible that doses exceeding 600img/kg bw are tolerated by animals without serious adverse effects.

·        Results for the negative and positive control animals are not reported. A comparison with the exposed animals is therefore not possible. The same group published results of a similar study in which the CA frequencies were excessively high (around 5.5%), which brings into questions whether the test animals were sufficiently healthy at study initiation. Since such data is lacking in the current publication, the reliability is questionable.

·        Dose- and time-related dose responses are most unusual. Different sampling times were included in this assay to study the effects on different stages of the cell cycle.Effects seen as early as 6 hr would mean cells were exposed (and sensitive to genotoxic effects) in the G2 phase of the cell cycle. There are very few known genotoxins which are active in G2. Most genotoxins act in G1 or S-phase which would be represented by effects at 12 hr. By 24 hr it is likely that many cells would have divided and be in the next cell cycle, so this would mean either that new damage was being induced or that damage was persistent (without being lethal) for more than 1 cell cycle. One would usually expect that cells exhibiting damage early in the cell cycle would not be able to survive through mitosis to the next cell cycle.

·        There are few publications where both time- and dose-related CA responses have been studied with known genotoxins. Some clearly show dose-related increases in cells with CA at one or more sampling times, but this is not always the case as the following papers indicate:

                    i.           McFee and Tice (1990) demonstrated time- and dose-related increases in bone marrow CA for the potent genotoxins mitomycin C and cyclophosphamide, but although DMBA gave dose-related increases in CA at 18 and 26 hr it did not induce CA at 10 hrs.

                       ii.      Tates & Natarajan (1976) observed a clear dose response for CNU-ethanol-induced CA in bone marrow 1 day after ip dosing, but after 2 days there was no clear dose-response.

                      iii.      Aydemir & Bilaloğlu (2003) showed that the anti-cancer drug topotecan induced dose-related CA at both 6 and 24 hr after a single ip dose. However, although gemcitabine induced dose-related increases in bone marrow CA at 24 hr, it did not induce any CA at 6 hr.

·        There are also numerous examples of genotoxins inducing micronuclei (MN) at some sampling times but not at others. For example Sutou et al (1990), reporting on a collaborative study by the MMS subgroup of JEMS, showed that ethylnitrosurea, ethyl methanesulphonate, ARA-C and benzene all induced dose-related increases in bone marrow MN at 6 and 24 hr but not at 48 or 72 hr sampling times after 2 or 4 doses. Thus, potent genotoxins may show dose-related increases in CA at several different sampling times, but in many cases they only show dose-related effects at limited sampling times due to cell-cycle specific effects, death of damaged cells etc. Given the overall toxicokinetic and genotoxic profile of sulfites (i.e. they are rapidly metabolised by sulfite oxidase and do not show positive findings in multiple test systems) it is unlikely that it would be expected to produce dose-related increases in CA at multiple sampling times. Thus, although there appears to be a significant induction of CA by potassium disulfite in this study, there are concerns, as discussed above, which render this publication not reliable for risk assessment purposes due to lack of credibility.

 

In a dominant lethal assay conducted by SRI International (Author unknown, 1979), the induction of dominant lethal mutations in rat after sodium bisulfite administration was investigated. Male Sprague-Dawley rats (53 to 62 days old, bodyweight 247-339 g) were given sodium bisulfite (purity not stated) in diet, ad libitum at doses of 45 (maximum tolerated dose), 15 and 4.5 mg/kg/day over a period of 10 weeks. Animals in the positive control group received triethylenemelamine (TEM). After the 10-week treatment period, 40 male rats from the vehicle control group and 20 male rats from each treatment group were selected and mated with two adult virgin females for seven days. These females were replaced with two new females for an additional 7-day mating period. Each female was sacrificed 15-19 days after the first day of cohabitation. At the end of the 10-week treatment period, body weight gains did not show a dose-response effect. The weight gains ranged from 3% below to 8% above control values. The dominant lethal test produced no consistent responses to suggest that sodium bisulfite is mutagenic to rats. Although the overall reporting quality fulfils the criteria for its use in the chemicals safety assessment, the method is of limited relevance (as detailed in the introductory remarks above). The reference is considered reliable with restriction but only as supporting information in a weight of evidence assessment.

 

In a range of genetic toxicity tests, sodium metabisulfite (purity not stated) was tested in an in vivo cytogenicity and dominant lethal assay in rats (Author unknown, 1972). The reference only contains data on the results (including individual raw data). There is a complete lack of information on (i) the test animals such as strain, source, age, bodyweight at study initiation (ii) housing and feeding conditions (iii) test item characterisation such as purity, impurity, vehicle, stability (iv) evaluation criteria. Although the results of the cytogenicity tests were available in detail, this reference cannot be rated due to the complete lack of administrative information and material and methods descriptions.

 

Carvalho, I.M.C.M.M. et al. (2011) tested sodium disulfite (aka sodium metabisulfite) in a combined in vivo comet and micronucleus assay in CF1 mice after single oral administration. The maximum tolerated dose was determined in a previous experiment in 6 mice at doses of 0.5, 1 and 2 g/kg bw via gavage. In the main study 10 mice per group (five females and five males) were treated for 24 h with a single dose by gavage (0.1 ml/10 g body weight): (a) sodium disulfite (0.5, 1, or 2 g/kg b.w.); (b) negative control group: water; (c) positive control group: cyclophosphamide, 25 mg/kg.

 

This alkaline comet assay was performed in peripheral blood, bone marrow and liver samples obtained 24 hrs post exposure. Comets were stained via silver staining.Damage index (DI) was assessed by visually examining the tail size of each cell, ranging from 0 (no tail) to 4 (maximum-length tails). The damage index was calculated by multiplying the DI with the number of cells scored, i.e. a maximum of 400 could theoretically be reached (100 cells scored x 4 (maximum damage index)).A significant increase was observed on both damage index and damage frequency values, when comparing 1 and 2 g/kg doses to negative control, for all tissues. The positive findings appear plausible. However there are reasons to question these findings: (i) silver staining is not specific for DNA, since silver cations also interact with all negatively charged functional groups in biomolecules, such as proteins, DNA, RNA. It is therefore unclear whether the measured tail was specific to DNA damage. A DNA-specific fluorophore was not used. (ii) DNA damage is usually expressed either in tail length, DNA content in tail or tail moment. Crude visual inspection is prone to subjective interpretation by the experimenter which may result in a significant intra-day and day to day variation. The metric for DNA damage should be measured using (i) micrometer in eyepieces (ii) ruler on photographs or (iii) a computerised automated system, so ensure consistent and reproducible interpretation of DNA tails. Due to the mentioned ambiguities, it is questionable whether a true substance induced DNA damage was observed in the comet assay.

 

Bone marrow and peripheral blood smears were prepared, stained and 2000 polychromatic reticulocytes from the bone marrow or 2000 reticulocytes from the peripheral blood were scored for the presence of micronuclei. Thepolychromatic erythrocytes: normochromatic erythrocytes (PCE/NCE) ratio was scored in 1000 cells for the determination of bone marrow toxicity. Blood and bone marrow cells of mice treated with the higher dose of SMB (2 g/kg) showed significant increase in MN, when compared with negative control, as well as significant reduction in PCE/NCE ratio. No difference in results was observed between sexes. Again, these results appear plausible, there are however serious doubts that these results could be obtained by the described procedure. Results show a dose-dependent increase of MN frequency both in peripheral blood and bone marrow already 24 hours after administration. This finding is most unusual, since micronuclei are formed during mitosis of an erythroblast which are translocated to the immature erythrocytes, which is ultimately released form the bone marrow into the blood stream as reticulocytes. Being a sequential order of events, clastogenic effects must first be observed in the bone marrow before becoming visible in the peripheral blood. It is therefore surprising that in this publication, micronuclei formation was observed in both the bone marrow and peripheral blood at the same time to an equal degree. According to OECD Guideline 474, samples should be drawn (i) not earlier than 24hours after exposure when examining the bone marrow and (ii) not earlier than 36 hours after exposure when examining the peripheral blood. Also, the guideline foresees at least two sampling times to assess the time course of clastogenic events. Although the study design appearsplausible and the positive findings appear robust, there are seriousconcerns as discussed above that render this publication not reliable for hazard and risk assessment purposes.

 

Pal, B.B.; Bhunya, S.P. (1992) investigated the clastogenic effect of sodium disulfite (aka sodium metabisulfite) in a bone marrow chromosome aberration and micronucleus test in mouse. Albino Swiss mice were given doses of 200, 300 and 400 mg/kg bw via subcutaneous, intraperitoneal injection or oral administration. The maximum dose was determined according to the authors via “trial and error method” (not further qualified). A total of 4-6 animals of unknown sex were used per dose group, control groups contained 6 or 10 animals. The animals were given the test item via different routes and different doses as follows:

chromosome aberration test:

1.      i.p. injection of 400 mg/kg bw, animals were sacrificed after 6, 24 and 48 hrs

2.      i.p. injection of 200, 300 mg/kg bw , animals were sacrificed after 24hrs

3.      s.c. and p.o. administration of 400mg/kg bw, animals were sacrificed after 24 hrs

4.      i.p. injection of 400 mg/kg bwdivided into 5 equal parts administered in 24 hr intervals, animals were sacrificed 24 hrs after last injection

micronucleus test:

5.      i.p. injection of 400 mg/kg bwdivided into 2 equal parts administered in 24hr intervals, animals were sacrificed 24hrs after last injection

 

Chromosomal slides were prepared by “colchicine-sodium citrate-acetic acid-alcohol-flame drying-Giemsa schedule”. A total of 75 metaphases were scored per animal. For the micronucleus experiments, the number of MN PCE and NCE was determined by scoring 1000 PCE and NCE per animal. Authors conclude a positive outcome for the chromosomal aberration and micronucleus experiments. However, this interpretation is questioned, namely

·        the aberration frequency is given including gaps, which arenormally not considered in the assessment of a positive response. When excluding gaps, all doses, time points and applications routes do not show any significant increase in aberration frequencies whatsoever, thus in fact leading to a “negative” conclusion. No distinction was made differentiating the type of chromosomal aberrations.

·        the elevated levels of MN PCE at 300 mg/kg bw (0.60 ±0.11) do not show a dose dependency and the values for the 200 and 400 mg/kg bw dose group (0.25% ±0.03, 0.43% ±0.17) are within or very close to the historical control values for a number of other labs (0.02%-0.38%). Due to the low number of animals used in the experiments, the lack of dose-dependency and the lack of a correlate with the chromosome aberration experiments, the elevated MN PCE levels in the 300 mg/kg bw dose group is considered incidental and lacking biological relevance.

·        cytotoxicity was not measured for the chromosomal aberration and micronucleus experiments, likewise the report is lacking records of systemic toxicity in the test animals. It is therefore not possible to correlate clastogenic effects with cytotoxicity.

·        the experimental procedures are very briefly described or completely lacking. According to the authors the maximum tolerated dose was determined using “trial and error method” without specifying further under which exposure conditions, duration, group size this was determined. A description of the experimental preparation for slide preparation and scoring of chromosomes is missing.

·        basic information on animal source, health status, housing and treatment conditions were not given

 

Based on the comments made above the reference is considered not reliable, since it is showing a poor reporting and experimental quality. In addition, the authors’ conclusion of a clear clastogenic finding in the chromosomal aberration experiments must be disputed.

 

Summary entry - DNA damage in vivo:

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 used to rate the comet study for its relevance in chemical safety assessment. The criteria as published by Tice et al. (2000) were used for this screening. In case a reference does not fulfil the criteria stated therein, it was rated as “not rateable” and not further considered for chemical safety assessment purposes.

 

Meng, Z. (2004, identical with Meng, Z. et al. (2005)Damage effects of sulfite sodium on DNA in cells from mice various organs. Food Sci. China 26, 203-205)investigated the DNA damage of a 3:1 sodium sulfite and sodium bisulfite mixture in male mice via comet assay. Three groups of six male mice each received an i.p. dose of a mixture of sodium sulfite and sodium bisulfite (3:1 M/M) (125, 250 or 500 mg/kg bw) in 200 ml of 0.9% sodium chloride daily for 7 days, maximum dose equals the half LD50. No positive control substance was applied. Animals were killed 24 hrs after final administration, assumed target organs (brain, lung, heart, liver, stomach, spleen, thymus and kidney) were removed and single cell suspensions were prepared. Cytotoxicity was determined via Trypan blue exclusion. Cells were plated in low-melting agarose, lysed and a DNA unwinding step in alkaline buffer was performed prior electrophoresis. After electrophoresis, DNA was stained with ethidium bromide. Images of 25 randomly selected cells were analysed from each slide investigated (two slides per mouse). For each group (six male mice), 300 cells were scored. Tail moment was used for the evaluation of the DNA damage. There were no signs of toxicity in either treatment group and viability of target organ cells was > 95% after isolation. The sulfite mixture significantly increased the tail moment of DNA in cells from all organs tested at all doses tested.

 

Summary:

Substance specific information:

In vitro genotoxicity studies in bacteria and mammalian cells are negative. In addition, an in vivo MN test in mice bone marrow yielded a negative result. Systemic availability was shown by a dose dependent increase in malondialdehyde levels.

 

Read-across information:

Conclusion on in vitro genetic toxicity:

There was no evidence whatsoever for any mutagenic activity of sulfites in the bacterial reverse mutation test or the mouse lymphoma assay, up to the maximum concentration limited by cytotoxicity. Consequently, sulfites are considered non-mutagenic in suitable in vitro test systems.

All references investigating in vitro clastogenic and aneugenic effects of sulfites exhibit serious experimental or reporting deficiencies, which render this information unsuitable for the chemical hazard assessment. Consequently, no conclusive statement can be made on the clastogenic and aneugenic effects of sulfites in vitro.

 

Conclusion on in vivo genetic toxicity

The in vivo data for sulfites present a consistent pattern. In studies administering sulfites via subcutaneous injection or the oral route, neither rats nor mice showed an increase of micronuclei in the bone marrow or an increase of dominant lethal mutations. Doses were limited by toxicity. In contrast, some studies involving intraperitoneal and oral administration appear to show a positive response. However, upon careful evaluation, the findings in these studies appear to lack biological plausibility, and also suffer from several methodological deficiencies. By weight-of-evidence, it is therefore concluded that sulfites do not cause clastogenic or aneugenic events in animals.

 

Overall conclusion:

The available data on genetic toxicity allow a conclusive statement on the genetic toxicity for sulfur dioxide and sulfites. Irrespective of the reporting quality of the publications, both positive and negative findings are reported in in vitro as well as in vivo test systems.

Sulfur dioxide: In vitro genotoxicity studies in bacteria and mammalian cells are negative. In addition, an in vivo MN inhalation test at high SO2 levels in mice bone marrow yielded a negative result. Systemic availability was shown by a dose dependent increase in malondialdehyde levels.

Sulfites: Following rigorous relevance and reliability screening, it can be concluded that sulfites do not show any clastogenic potential. The references discussed under in vitro clastogenicity are rated as not reliable due to experimental and reporting deficiencies and to not show a consistent pattern on the induction of chromosome and genome mutations. A high-quality in vivo study with sodium sulfite via subcutaneous injection in mice did not show an increase of micronuclei formation up to the maximum tolerated dose. This finding is supported by a negative dominant lethal test in rats after single and repeated oral administration (feed) in rats. A number of in vivo clastogenicity studies were assessed as being of limited reliability, since these exhibit reporting and/or other experimental deficiencies and lack biological plausibility.

 

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

 In addition to the information directly addressing the endpoint genetic toxicity, the following information should be taken into account when discussing the genetic toxicity of sulfites:

1.      the environmental and physiological half-life is highly dependent on the solvent/vehicle, oxidation status, light and temperature; for example, under environmentally relevant conditions, sulfites are already readily oxidised to sulfate showing a half-life of 16 hrs in seawater and only a few hrs in freshwater.

2.      due to its endogenous occurrence in all mammalian species, almost all tissues show sulfite oxidase activity, with very high activities in liver, kidney and heart, whereas spleen, testes and brain show lower sulfite oxidase activity. A total background serum level in humans of 4.9µmol sulfite/L was measured for both sexes. The overall sulfite oxidase capacity is very high in mammalian species, with a theoretical maximum oxidation rate of 750mmol/kg/day (equivalent to 56g of SO32-/kg/day). The total urinary excretion of sulfate resulting from sulfite metabolism was estimated to be 25mmol out of which the majority was generated from endogenous sulfite (24mmol). This amount is approx. 1000fold the total amount of sulfite being enzymatically processed per day (assuming a total bool volume of 5L per individual).

3.      due to their instability in an aqueous environment and their very rapid oxidation and elimination in animals and humans, it appears highly unlikely that sulfites show adverse systemic effects, when administered via physiological routes. This hypothesis is substantiated by a complete lack of adverse effects in long-term animal studies as well as in direct observations in humans:

·        3-generation reproductive toxicity study, groups of 20 male and 20 female Wistar rats were given doses of up to 955 mg/kg bw/day via diet over a period of 2 years. There were no signs of systemic toxicity, resulting in a NOAEL above the maximum dose of 955 mg/kg bw/day.

·        taking together the results from the three animal studies on sodium and potassium metabisulfite (Tanaka et al., 1979; Til et al., 1972; Feron and Wensfoort, 1972) there was no indication that metabisulfite had any carcinogenic effect

·        4 reliable studies on pulp and paper mill workers were available (Milham and Demers, 1984; Robinson, et al. 1986; Anderson, et al. 1998; Rix, et al. 1997), see section 7.10.2. , based upon which no carcinogenic activity must be expected for the sulfites

 

The use of sulfites as nutritional supplement and preservative in cosmetic formulations has been repeatedly reviewed by national and international authorities, concluding:

“Sodium metabisulfite is listed as GRAS (Generally Recognized as Safe) by the FDA (Food and Drug Administration) as preservatives in certain foods. Sodium metabisulfite is also used up to a concentration of 1% as an antioxidant in hair care products and as a reducing agent in cosmetic formulations (CIR 2003). [...] Sulfur dioxide and sodium metabisulfite are currently not classifiable (Group 3) as to their carcinogenicity to humans (IARC 1992). [...] Conclusions of the OECD SIDS report indicated 2% sodium metabisulfite via feed (20,000 ppm or 1,000 mg/kg/day) for 104 weeks was not carcinogenic in Wistar rats.” (EPA, 2009)

 

“The CIR expert panel concluded that sodium sulfite, potassium sulfite, ammonium sulfite, sodium bisulfite, ammonium bisulfite, sodium metabisulfite and potassium metabisulfite are safe as used in cosmetic formulations.” (CIR, 2003)

“Existing mutagenicity studies support information on different categories of biological endpoints. Gene mutations were analysed in a couple of in vitro tests (Ames-test with Salmonella typhimurium and E. coli, tests with yeasts). Chromosomal aberration and sister chromatid exchange (SCE) were tested in vitro (in human blood lymphocytes) as well as in vivo (micronucleus-test in mice and rats). The induction of unscheduled DNA synthesis (UDS) was analysed in rat hepatocytes after in vivo application. Dominant lethal assays in mice and rats completed the spectrum of different mutagenicity tests. Neither sodium sulfite and sodium metabisulfite nor potassium metabisulfite were found to be genotoxic in the test battery. Some of the in vitro assays performed with sodium bisulfite showed positive results, especially in the Ames test and in the chromosome aberration test. However, none of the in vivo tests have shown a genotoxic potential. To further investigate the above mentioned positive in vitro results found with sodium bisulfite, an UDS in rats and a micronucleus test in mice were added. To be sure that the bisulfite and not the disulfite was tested, an appropriate pH-value was maintained by a buffer system. Both additional tests were negative. Summarising results of the available mutagenicity tests, genotoxic potential of the specified inorganic sulfites, bisulfites and metabisulfites seems to be very unlikely.” (SCCNFP, 2003)

 

References

US EPA 2007. Registration Eligibility Decision- Inorganic Sulfites. Special Review and Reregistration DivisionOffice of Pesticide Programs.

Nair, B.; Elmore, A.R., 2003. Final report and the safty assessment sodium sulfite, potassium sulfite, ammonium sulfite, sodium bisulfite, ammonium bisulfite, sodium metabisulfite and potassium metabisulfite (Cosmetic Ingredient Review). Int. J. Toxicol. 22, 63-88

The Scientific Committee on Cosmetic Products and non-Food Products (SCCNFP) intended for consumers opinion concerning Inorganic Sulfites and Bisulfites, 2003. Adopted at its 23rd plenary meeting of 18 March 2003.

 ________________________________________________________________________

 

[1]Reference cited in OECD SIAR on sulfur dioxide

Justification for classification or non-classification

Justification for non-classification

According to regulation (EC) 1272/2008 (CLP), as amended, substances shall be classified for the endpoint germ cell mutagenicity, in case they may cause mutations in the germ cells of humans that can be transmitted to the progeny. The classification shall be based on the total weight of evidence available, using expert judgement and the relevance of the route of exposure used in the study of the substance compared to the most likely route of human exposure shall also be taken into account.

The reference Ziemann (2008) is considered as the key study for in vivo genetic toxicity and will be used for classification. The overall results are as follows:

Sulfur dioxide did not show a significant or dose-dependent increase in micronucleated cells in the bone marrow of male and female mice via inhalation for 7 days up to the maximum dose of 30Molppm 24 hours after dosing. None of the in vitro genetic toxicity studies given in the technical dossier showed any effect in bacteria or somatic cells following exposure towards sulfur dioxide, thus supporting the negative findings in the in vivo tests as cited above. The available data on genetic toxicity allow a conclusive statement on the genetic toxicity for sulfites. Irrespective of the reporting quality of the publications, both positive and negative findings are reported in in vitro as well as in vivo test systems. There was no evidence whatsoever for any mutagenic activity of sulfites in the bacterial reverse mutation test or the mouse lymphoma assay, up to the maximum concentration limited by cytotoxicity. Consequently, sulfites are considered non-mutagenic in suitable in vitro test systems. Following rigorous relevance and reliability screening, it can be concluded that sulfites do not show any clastogenic potential. A high-quality in vivo study with sodium sulfite via subcutaneous injection in mice did not show an increase of micronuclei formation up to the maximum tolerated dose. This finding is supported by a negative dominant lethal test in rats after single and repeated oral administration (feed) in rats. A number of in vivo clastogenicity studies were assessed as being of limited reliability, since these exhibit reporting and/or other experimental deficiencies and lack biological plausibility.

Overall, there is no consistent evidence of induction of genetic toxicity with relevance to humans for suöfur dioxide/sulfites. Therefore, the classification criteria as laid down in regulation (EC) 1272/2008 are not met and sulfur dioxide/sulfites are not to be classified as germ cell mutagen.