Registration Dossier

Toxicological information

Specific investigations: other studies

Currently viewing:

Administrative data

Link to relevant study record(s)

Referenceopen allclose all

Endpoint:
cell culture study
Type of information:
experimental study
Adequacy of study:
supporting study
Study period:
2019-07-31 to 2019-08-16
Reliability:
other: Any kind of reliability rating is not considered to be applicable, since the current in vitro test is not conducted/reported according to standardised guideline
Rationale for reliability incl. deficiencies:
other: Rating according to Klimisch criteria does not apply to mechanistic in vitro studies
Qualifier:
no guideline followed
Principles of method if other than guideline:
THE TOXTRACKER ASSAY
This study was performed to investigate nine cobalt compounds (cobalt oxalate, cobalt resinate, cobalt sulphide, lithium cobalt dioxide, cobalt propinate, cobalt stearate, cobalt dihydroxide, cobalt octoate, cobalt naphthenate, and cobalt hydroxide oxide) for (geno)toxic properties using the ToxTracker reporter assay in the absence and presence of a metabolising system. In addition, compounds will be analysed for their ability to induce hypoxia by assessing the expression of HIF1 alpha target genes using qPCR. The ToxTracker assay is a panel of six validated GFP-based mouse embryonic stem (mES) reporter cell lines that can be used to identify the biological reactivity and potential carcinogenic properties of compounds in a single test. ToxTracker is a mammalian stem cell-based assay that monitors activation of specific cellular signalling pathways for detection of the biological reactivity of compounds. Stem cells are genetically stable and proficient in all cellular pathways required for accurate detection of potentially carcinogenic properties of compounds. Extensive whole-genome transcription profiling has led to identification of a panel of biomarker genes that are preferentially activated upon exposure to different classes of carcinogens and toxicants (Hendricks, G. et al., 2011)*. To allow easy assessment of the activation status of these biomarker genes, green fluorescent (GFP) mES reporter cell lines were generated. These reporters were created using bacterial artificial chromosomes (BAC) that contain the complete biomarker gene including promoter and regulatory elements ensuring physiological regulation of the GFP reporters following transfection into stem cells (Hendricks, G. et al., 2012)*.

ToxTracker consists of a panel of six different mES GFP reporter cell lines representing four distinct biological responses that are associated with carcinogenesis, i.e. general cellular stress (biomarker gene: Btg2), DNA damage (biomarker genes: Bscl2 and Rtkn), oxidative stress (biomarker genes: Srxn1 and Blvrb) and protein damage (biomarker gene: Ddit3). The specificity of the ToxTracker reporter cell lines was extensively validated using different libraries of reference compounds (Kirkland, G. et al, 2016)*. Over 97% of all tested chemicals were classified correctly.

*References:
- Hendriks, G., Atallah, M., Raamsman, M., Morolli, B., van der Putten, H., Jaadar, H., Tijdens, I., Esveldt-van Lange, R., Mullenders, L., van de Water, B., et al. (2011). Sensitive DsRed fluorescence-based reporter cell systems for genotoxicity and oxidative stress
assessment. Mutat Res 709-710, 49–59.
- Hendriks, G., Atallah, M., Morolli, B., Calléja, F., Ras-Verloop, N., Huijskens, I., Raamsman, M., van de Water, B., and Vrieling, H.
(2012). The ToxTracker assay: novel GFP reporter systems that provide mechanistic insight into the genotoxic properties of
chemicals. Toxicol. Sci. 125, 285–298.
- Kirkland, D., Kasper, P., Müller, L., Corvi, R., and Speit, G. (2008). Recommended lists of genotoxic and non-genotoxic chemicals for
assessment of the performance of new or improved genotoxicity tests: a follow-up to an ECVAM workshop. Mutat Res 653, 99–108.
GLP compliance:
no
Type of method:
in vitro
Endpoint addressed:
genetic toxicity
other: mode of action analysis
Specific details on test material used for the study:
STABILITY AND STORAGE CONDITIONS OF TEST MATERIAL
- Storage condition of test material: 4 °C, dark
Details on test animals or test system and environmental conditions:
Not applicable - Since this is an in vitro study there is no information on test animals.
Vehicle:
other: mouse embryonic stem (mES) cell culture medium
Details on exposure:
PREPARATION OF DOSING SOLUTIONS
The test compounds were solved in mouse embryonic stem (mES) cell culture medium.
Analytical verification of doses or concentrations:
yes
Details on analytical verification of doses or concentrations:
The actual maximum cobalt concentration in the mES cell culture medium after 24 hours stirring at 37ºC and filtration, after treatment and in treated cells was quantified via ICP-MS.
Results:
For most compounds, the concentration in medium was >5 μg/mL. For for lithium cobalt dioxide, cobalt naphthenate and cobalt hydroxide oxide the concentration in medium was <2 μg/mL, indicating that these compounds are poorly soluble in mES culture medium. In addition, a concentration of 2.9 µg Co/mL indicated that cobalt stearate is almost insoluble in cell culture medium.
Duration of treatment / exposure:
ToxTracker assay: 24 hours exposure in absence of S9 and presence of 0.25 % S9
Hypoxia assessment: 8 hours
Frequency of treatment:
once
Remarks:
cobalt oxalate (ToxTracker assay): 0, 1.81, 3.63, 7.25, 14.5, and 29 µg/mL
cobalt oxalate (Hypoxia assessment): 0, 7.25 and 29 µg/mL
Remarks:
cobalt resinate (ToxTracker assay): 0, 0.26, 0.53, 1.05, 2.1, and 4.2 µg/mL
cobalt resinate (Hypoxia assessment): 0, 1.05 and 4.2 µg/mL
Remarks:
lithium cobalt dioxide (ToxTracker assay): 0, 0.0026, 0.0051, 0.010, 0.021 and 0.041 µg/mL
lithium cobalt dioxide (Hypoxia assessment): 0, 0.010 and 0.041 µg/mL
Remarks:
cobalt propinate (ToxTracker assay): 0, 1.44, 2.88, 5.75, 11.5 and 23 µg/mL
cobalt propinate (Hypoxia assessment): 0, 5.75 and 23 µg/mL
Remarks:
cobalt stearate (ToxTracker assay): 0, 0.18, 0.36, 0.73, 1.45 and 2.9 µg/mL
cobalt stearate (Hypoxia assessment): 0, 0.73 and 2.9 µg/mL
Remarks:
cobalt dihydroxide (ToxTracker assay): 0, 3.5, 7, 14, 28 and 56 µg/mL
cobalt dihydroxide (Hypoxia assessment): 0, 14 and 56 µg/mL
Remarks:
cobalt octoate (ToxTracker assay): 0, 1.06, 2.13, 4.25, 8.5 and 17 µg/mL
cobalt octoate (Hypoxia assessment): 0, 4.25 and 17 µg/mL
Remarks:
cobalt naphthenate (ToxTracker assay): 0, 0.056, 0.11, 0.22, 0.45 and 0.89 µg/mL
cobalt naphthenate (Hypoxia assessment): 0, 0.22 and 0.89 µg/mL
Remarks:
cobalt hydroxide oxide (ToxTracker assay): 0, 0.081, 0.16, 0.33, 0.65 and 1.3 µg/mL
cobalt hydroxide oxide (Hypoxia assessment): 0, 0.33 and 1.3 µg/mL
No. of animals per sex per dose:
not applicable
Control animals:
other: not applicable
Details on study design:
TEST PROTOCOL
The ToxTracker reporter cells were maintained by culturing them in gelatin-coated dishes in the presence of irradiated primary mouse embryonic fibroblasts (MEFs) in mES cell culture medium. During chemical exposures and reporter analysis the ToxTracker cells were cultured in the absence of fibroblasts in mES cell culture medium.

1) Cytotoxicity testing/dose range finding:
A dose range finder was performed using wild-type mouse embryonic stem (mES) cells (strain B4418). For the dose range finding, a solution of 100 μg/mL of cobalt containing compound was prepared in the mES cell culture medium and stirred for 24 hours at 37ºC. Before treatment of the cell lines, the solutions were filtered with 0.22 μm filters. Wild type mES cells are exposed to 20 different concentrations of the soluble fraction of test substances. Exposure was conducted with and without 0.25 % S9.

Cytotoxicity was estimated by cell count after 24 hour exposure using a flow cytometer and is expressed as percentage of viable cells after 24 hour exposure compared to vehicle control exposed cells. The actual maximum cobalt concentration in the mES cell culture medium after 24 hours stirring at 37ºC and filtration, after treatment and in treated cells was quantified via ICP-MS.

2) ToxTracker assay (3 independent tests):
The six independent mES reporter cell lines (containing biomarker gene Bscl2, Rtkn, Srxn1, Blvrb, Ddit3, or Btg2) were seeded in gelatin-coated 96-well cell culture plates in 200 μL mES cell medium (50.000 cells per well). 24 hours after seeding the cells in the 96-well plates, medium was aspirated and fresh mES cell medium and the diluted chemicals was added to the cells. For each tested compound, five concentrations were tested in 2-fold dilutions. If possible, the highest compound concentration was selected such that it was expected to induce significant cytotoxicity (>40%).

Induction of the green fluorescent (GFP) reporters was determined after 24 hours exposure using a flow cytometer. Only GFP expression in intact single cells was determined. Mean GFP fluorescence was measured and used to calculate GFP reporter induction compared to a vehicle control treatment. Cytotoxicity was estimated by cell count after 24 hours exposure using a flow cytometer and was expressed as percentage of intact cells after 24 hours exposure compared to vehicle exposed controls. For cytotoxicity assessment in the ToxTracker assay, the relative cell survival for the six different reporter cell lines was averaged.

Metabolic activation was included in the ToxTracker assay by addition of S9 liver extract from aroclor1254-induced rats. Cells are exposed to five concentrations of the test compounds in the presence of 0.25% S9 and required co-factors (RegenSysA+B) for 24 hours.

Three independent tests were carried out.

In case auto-fluorescence of the test substances was observed in the dose range finder, wild type mES cells were exposed to the compounds at the same concentrations as used in the ToxTracker. The mean fluorescence caused by the compound was then subtracted from the ToxTracker results of the respective compound.

3) Hypoxia assessment
To assess the induction of cellular hypoxia following exposure to the cobalt containing compounds, induction of a selection of HIF1ɑ target genes was determined using quantitative real-time PCR (qPCR). For the analysis of the expression for HIF1ɑ target genes in response to treatment, wild-type mES were treated with two doses of the test materials solved in mES cell culture medium in the absence of S9 for 8 hours in triplicate. After treatment, cells were lysed in Trizol, RNA was isolated and after cDNA synthesis, a qPCR was performed using primers for Hmox1, Slc2a1, Eno1, Ddit4 and Bnip3, using two technical replicates. HPRT and GAPDH were included as reference genes.

TEST CRITERIA
The ToxTracker assay was considered to have a positive response when a compound induces at least a 2 fold increase in GFP expression in any of the reporters. Activation of the Bscl2-GFP or Rtkn-GFP reporters indicate induction of DNA damage, Srxn1-GFP and Blvrb-GFP indicate induction of cellular oxidative stress and Ddit3-GFP activation is associated with the unfolded protein response. The Btg2-GFP reporter is controlled by the p53 tumor suppressor and is activated by DNA damage but can also be induced by
oxidative stress, hypoxia, metabolic stress and apoptosis. Only GFP inductions at compound concentrations that showed < 75% cytotoxicity are used for the ToxTracker analysis. Data from measurements > 75% cytotoxicity can not be interpreted in a meaningful way and are therefore discarded.

The No Observed Effect Level (NOEL) defines the highest concentrations where none of the ToxTracker reporters showed a >1.5-fold increase in fluorescence. The Lowest Observed Effect Level (LOEL) indicates the lowest concentration were at least one of the ToxTracker reporters showed an induction of ≥2.0. The No Observed Genotoxicity Effect Level (NOGEL) is the highest concentration that did not activate the Bscl2-GFP and/or Rtkn-GFP reporters with a >1.5-fold increase in fluorescence. The Lowest Observed Genotoxicity Effect Level (LOGEL) is the lowest concentration that activated the Bscl2-GFP and/or Rtkn-GFP genotoxicity reporters ≥2.0-fold.

For the analysis of the qPCR data, an induction more than 2-fold compared to the vehicle treated control sample was considered to be positive. A 2 fold increase corresponds to approximately one standard deviation above the untreated control samples.

DATA ANALYSIS
In order to allow comparison of induction levels of the ToxTracker reporter cell lines for large number of compounds Toxplot was developed. Toxplot imports raw GFP reporter data from the flow cytometer, calculates GFP induction levels and cytotoxicity, performs statistical analysis of the data and hierarchical clustering of the tested compounds, and visualises the data in a heatmap allowing convenient interpretation of obtained test results. ToxPlot software uses agglomerative hierarchical clustering to visualize the ToxTracker data. Agglomerative clustering uses the ‘bottom-up’ approach, which puts each observation in its own cluster and pairs of clusters are merged as one moves up the hierarchy. To compare the induction of the six GFP reporters for a collection of compounds, each with different biological reactivities, dose-response relationships and kinetics, Toxplot calculates for each compound the level of GFP induction for every individual reporter at a specified level of cytotoxicity (typically 10%, 25% and 50%). GFP induction levels are calculated by linear regression between two data points around the specified cytotoxicity level. In case the specified level of cytotoxicity can not be reached at the highest tested compound concentration, Toxplot displays the GFP induction level at this top concentration. In the heatmap, Toxplot marks the compounds that do not induce the selected level of cytotoxicity. Because the cytotoxicity for a compound can vary between the ToxTracker cell lines, calculations of the GFP induction levels of the individual reporters by Toxplot can slightly deviate from the GFP induction and cytotoxicity figures.
Examinations:
Please refer to the field "Details on study design" above.
Positive control:
Positive reference treatments with cisplatin (DNA damage; 0, 2.5, and 5 µM; solvent: PBS), diethyl maleate (oxidative stress; 0, 150, and 300 µM; solvent: DMSO), sodium arsenite (oxidative stress; 0, 0.31, 0.63, 1.25, 2.5, and 5 µM; solvent: PBS), tunicamycin (unfolded protein response; 0, 2, and 4 µg/mL; solvent: DMSO), rosuvastatin (cytotoxicity; 0, 7.81, 15.63, 31.25, 62.5, and 125 µM; solvent: DMSO) and aflatoxin B1 (metabolic activation of progenotoxins by S9; 0, 2.5, and 5 µM; solvent: DMSO) were included in all experiments. Solvent concentration was similar in all wells and never exceeded 1% for DMSO.
Details on results:
POSITIVE REFERENCE TREATMENT
The validity of the ToxTracker assay was confirmed. The genotoxic compound cisplatin showed induction of the DNA damage response (Bscl2, Rtkn) and p53-mediated cellular stress (Btg2). Diethyl maleate (DEM) and sodium arsenite induced primarily the oxidative stress related reporters Srxn1 and Blvrb, tunicamycin induced the unfolded/misfolded protein stress response (Ddit3) and rosuvastatin was cytotoxic without activating any of the ToxTracker reporters. The positive control compound aflatoxin B1, selectively induced the Bscl2 and Rtkn reporters when tested in the presence of S9 liver extract. Generally, the controls showed GFP induction levels compliant with historical data and demonstrated the functionality of the mES reporter cell lines. For the positive control treatments specific for this project (sodium arsenite and rosuvastatin) insufficient historical control data was available for a comparison.
Please also refer to the field "Attached background material" below.

SOLUBILITY
For most compounds, the concentration in medium was >5 μg/mL. For for lithium cobalt dioxide, cobalt naphthenate and cobalt hydroxide oxide the concentration in medium was <2 μg/mL, indicating that these compounds are poorly soluble in mES culture medium. In addition, a concentration of 2.9 µg Co/mL indicated that cobalt stearate is almost insoluble in cell culture medium.

CYTOTOXICITY
At the maximum tested concentrations in the absence of a metabolising system cobalt oxalate, cobalt propinate, cobalt dihydroxide, and cobalt octoate induced significant cytotoxicity (>50%) at the following concentrations:
- cobalt oxalate: 22.8 µg/mL
- cobalt propinate: 10.9 µg/mL
- cobalt dihydroxide: 13.5 µg/mL
- cobalt octoate: 14.4 µg/mL

In the presence of a metabolising system, there was no increased cytotoxicity observed for any of the compounds (exception: for cobalt oxolate similar cytotoxicity was observed in the presence of metabolising system). For the other substances little cytotoxicity was observed and the exposure concentration was limited by the solubility of the substances in cell culture medium. The six ToxTracker reporter cell lines showed a comparable cytotoxic response to the test compounds. For this reason, the cell survival graphs in the GFP induction figures show the average cytotoxicity of the six different cell lines (please also refer to the field "Attached background material" below).
No autofluorescence was observed at the tested concentrations of any test compound.

GENOTOXICITY
The Bscl2-GFP reporter is associated with induction of promutagenic DNA lesions and DNA replication inhibition. Induction of the Rtkn-GFP genotoxicity reporter correlates with induction of DNA strand breaks. None of the tested cobalt substances showed any activation of the Bscl2-GFP and/or Rtkn-GFP genotoxicity reporters when tested in the absence or presence of a metabolising system.
Please also refer to the field "Attached background material" below.

OXIDATIVE STRESS
Cobalt oxalate, cobalt propinate, cobalt dihydroxide and cobalt octoate, activated both oxidative stress reporters in absence and presence of S9. Cobalt resinate also activated the Srxn1-GFP reporters, but only a weak induction (>1.5 fold) of Blvrb-GFP which was not sufficient to reach the 2-fold threshold for a positive ToxTracker response. The activation of the oxidative stress reporter correlated with the cobalt concentration in medium. Induction of the Srxn1-GFP reporter is associated with activation of the Nrf2 antioxidant response and activation of the Blvrb-GFP reporter is associated with the Hmox1 antioxidant response.
Please also refer to the field "Attached background material" below.

P53 ACTIVATION
Btg2-GFP, the reporter for p53 activation, was induced upon exposure to cobalt oxalate and cobalt octoate only in the absence of S9. A weak activation of Btg2-GFP (>1.5 fold) was observed for cobalt propinate, but the induction levels did not reach the 2-fold threshold for a positive ToxTracker response. The Btg2-GFP reporter is activated by p53 in response to various types of cellular stress such as protein damage, oxidative stress, DNA damage or cytotoxicity.
Please also refer to the field "Attached background material" below.

PROTEIN DAMAGE
The Ddit3-GFP reporter, associated with protein damage and the unfolded protein response, was weakly activated (>1.5 fold) by the compounds cobalt resinate and cobalt dihydroixde in absence of S9, but the induction levels did not reach the 2-fold threshold for a positive ToxTracker response.
Please also refer to the field "Attached background material" below.

HYPOXIA ASSESSMENT
Cellular hypoxia in the ToxTracker cells activated the Hif1ɑ transcription factor and induced the expression of various target genes. Hmox1, BNIP3 and Ddit4 were all induced more than 2-fold upon exposure to cobalt oxalate, cobalt propinate, cobalt dihydroxide and cobalt octoate. Cobalt resinate activated HMOX, Slc2a1 and Ddit4 but not BNIP3. Slc2a1 activation was also observed for cobalt propinate and cobalt octoate. No activation of Eno1 was observed. The substance that induced hypoxia also activated the oxidative stress reporters in ToxTracker.
Please also refer to the field "Attached background material" below.
Conclusions:
None of the tested compounds was classified as genotoxic in the ToxTracker assay, as no activation of the Bscl2-GFP or Rtkn-GFP reporters was observed. Significant levels of oxidative stress were observed for cobalt oxalate, cobalt resinate, cobalt propinate, cobalt dihydroxide and cobalt octoate. These substances also activated the expression of several hypoxia-associated HIF1ɑ target genes. Only weak activation of the unfolded protein response was observed for cobalt resinate, but not for the other cobalt containing substances.
Endpoint:
cell culture study
Type of information:
experimental study
Adequacy of study:
supporting study
Study period:
2018-12-10 to 2018-12-19
Reliability:
other: Any kind of reliability rating is not considered to be applicable, since the current in vitro test is not conducted/reported according to standardised guideline
Rationale for reliability incl. deficiencies:
other: Rating according to Klimisch criteria does not apply to mechanistic in vitro studies
Qualifier:
no guideline followed
Principles of method if other than guideline:
THE TOXTRACKER ASSAY
This study was performed to test six cobalt compounds (cobalt borate neodecanoate, cobalt fine powder, cobalt sulphide, tricobale tetraoxide, cobalt dichloride hexahydrate, and cobalt (II) 4-oxopent-2-ene-2-olate) for (geno)toxic properties using the ToxTracker assay. In addition, compounds will be analysed for their ability to induce hypoxia. The ToxTracker assay is a panel of six validated GFP-based mouse embryonic stem (mES) reporter cell lines that can be used to identify the biological reactivity and potential carcinogenic properties of compounds in a single test. ToxTracker is a mammalian stem cell-based assay that monitors activation of specific cellular signalling pathways for detection of the biological reactivity of compounds. Stem cells are genetically stable and proficient in all cellular pathways required for accurate detection of potentially carcinogenic properties of compounds. Extensive whole-genome transcription profiling has led to identification of a panel of biomarker genes that are preferentially activated upon exposure to different classes of carcinogens and toxicants (Hendricks, G. et al., 2011)*. To allow easy assessment of the activation status of these biomarker genes, green fluorescent (GFP) mES reporter cell lines were generated. These reporters were created using bacterial artificial chromosomes (BAC) that contain the complete biomarker gene including promoter and regulatory elements ensuring physiological regulation of the GFP reporters following transfection into stem cells (Hendricks, G. et al., 2012)*.

ToxTracker consists of a panel of six different mES GFP reporter cell lines representing four distinct biological responses that are associated with carcinogenesis, i.e. general cellular stress (cellular pathway: p53 signaling; biomarker gene: Btg2), DNA damage (cellular pathway: ATR/Chk1 DNA damaging signaling; biomarker genes: Bscl2 and Rtkn), oxidative stress (cellular pathway: Nrf2 antioxidant response, Nrf2-independent; biomarker genes: Srxn1 and Blvrb) and protein damage (cellular pathway: unfolded protein response; biomarker gene: Ddit3). The specificity of the ToxTracker reporter cell lines was extensively validated using different libraries of reference compounds (Kirkland, G. et al, 2016)*. Over 97% of all tested chemicals were classified correctly.

*References:
- Hendriks, G., Atallah, M., Raamsman, M., Morolli, B., van der Putten, H., Jaadar, H., Tijdens, I., Esveldt-van Lange, R., Mullenders, L., van de Water, B., et al. (2011). Sensitive DsRed fluorescence-based reporter cell systems for genotoxicity and oxidative stress
assessment. Mutat Res 709-710, 49–59.
- Hendriks, G., Atallah, M., Morolli, B., Calléja, F., Ras-Verloop, N., Huijskens, I., Raamsman, M., van de Water, B., and Vrieling, H.
(2012). The ToxTracker assay: novel GFP reporter systems that provide mechanistic insight into the genotoxic properties of
chemicals. Toxicol. Sci. 125, 285–298.
- Kirkland, D., Kasper, P., Müller, L., Corvi, R., and Speit, G. (2008). Recommended lists of genotoxic and non-genotoxic chemicals for
assessment of the performance of new or improved genotoxicity tests: a follow-up to an ECVAM workshop. Mutat Res 653, 99–108.
GLP compliance:
no
Type of method:
in vitro
Endpoint addressed:
genetic toxicity
other: mode of action analysis
Specific details on test material used for the study:
STABILITY AND STORAGE CONDITIONS OF TEST MATERIAL
- Storage condition of test material: 4 °C, dark
Details on test animals or test system and environmental conditions:
Not applicable - Since this is an in vitro study there is no information on test animals.
Vehicle:
other: mouse embryonic stem (mES) cell culture medium
Details on exposure:
PREPARATION OF DOSING SOLUTIONS
The test compounds were solved in mouse embryonic stem (mES) cell culture medium.
Analytical verification of doses or concentrations:
yes
Details on analytical verification of doses or concentrations:
The actual maximum cobalt concentration in the culture medium after 24 hours stirring at 37ºC and filtration was quantified via ICP-MS during the dose range finder.

Results:
The analysis of the cobalt levels in the medium showed that the cobalt levels in the medium for most compounds was around 20 μg/mL or higher. For cobalt sulphide and tricobalt tetraoxide the concentration in medium was much lower (<1 μg/ml), indicating that these two compounds are very poorly soluble in mouse embryonic stem (mES) culture medium.
Duration of treatment / exposure:
ToxTracker assay: 24 hours exposure in absence of S9 and presence of 0.25 % S9
Hypoxia assessment: 8 hours
Frequency of treatment:
once
Remarks:
cobalt borate neodecanoate (ToxTracker assay): 0, 1.2, 2.4, 4.8, 9.5, and 19 µg/mL
cobalt borate neodecanoate (Hypoxia assessment): 4.8 and 19 µg/mL
Remarks:
cobalt fine powder (ToxTracker assay): 0, 1.3, 2.5, 5, 10, and 20 µg/mL
cobalt fine powder (Hypoxia assessment): 5 and 20 µg/mL
Remarks:
cobalt sulphide (ToxTracker assay): 0, 0.06, 0.11, 0.22, 0.45, and 0.89 µg/mL
cobalt sulphide (Hypoxia assessment): 0.22 and 0.89 µg/mL
Remarks:
tricobalt tetraoxide (ToxTracker assay): 0, 0.03, 0.06, 0.12, 0.25, and 0.5 µg/mL
tricobalt tetraoxide (Hypoxia assessment): 0.12 and 0.5 µg/mL
Remarks:
cobalt dichloride hexahydrate (Tox Tracker assay): 0, 1.4, 2.9, 5.8, 11.5, and 23 µg/mL
cobalt dichloride hexahydrate (Hypoxia assessment): 5.8 and 23 µg/mL
Remarks:
cobalt (II) 4-oxopent-2-ene-2-olate (ToxTracker assay): 0, 1.2, 2.4, 4.8, 9.5, and 19 µg/mL
cobalt (II) 4-oxopent-2-ene-2-olate (Hypoxia assessment): 4.8 and 19 µg/mL
No. of animals per sex per dose:
not applicable
Control animals:
other: not applicable
Details on study design:
TEST PROTOCOL
The ToxTracker reporter cells are maintained by culturing them in gelatin-coated dishes in the presence of irradiated primary mouse embryonic fibroblasts (MEFs) in mES cell culture medium. During chemical exposures and reporter analysis the ToxTracker cells are cultured in the absence of fibroblasts in mES cell culture medium.

1) Cytotoxicity testing/dose range finding:
A dose range finder was performed using wild-type mouse embryonic stem (mES) cells (strain B4418). For the dose range finding, a maximum soluble concentration of 100 μg/mL was prepared in the mES cell culture medium, therefore 100 μg/mL of each test compound were stirred for 24 hours at 37ºC. Before treatment of the cell lines, the solutions were filtered with 0.22 μm filters. Wild type mES cells are exposed to 20 different concentrations of the test substances, with a maximum concentration of 100 μg/mL. Exposure was conducted with and without 0.25 % S9.

Cytotoxicity is estimated by cell count after 24 hour exposure using a flow cytometer and is expressed as percentage of viable cells after 24 hour exposure compared to vehicle control exposed cells. The actual maximum cobalt concentration in the mES cell culture medium after 24 hours stirring at 37ºC and filtration was quantified via ICP-MS.

2) ToxTracker assay (3 independent tests):
The six independent mES reporter cell lines (containing biomarker gene Bscl2, Rtkn, Srxn1, Blvrb, Ddit3, or Btg2) are seeded in gelatin-coated 96-well cell culture plates in 200 μL mES cell medium (50.000 cells per well). 24 hours after seeding the cells in the 96-well plates, medium is aspirated and fresh mES cell medium containing 10% fetal calf serum and the diluted chemicals is added to the cells. For each tested compound, five concentrations are tested in 2-fold dilutions. The highest compound concentration will induce significant cytotoxicity (>40%).

Induction of the green fluorescent (GFP) reporters was determined after 24 hours exposure using a flow cytometer. Only GFP expression in intact single cells was determined. Mean GFP fluorescence was measured and used to calculate GFP reporter induction compared to a vehicle control treatment. Cytotoxicity was estimated by cell count after 24 hours exposure using a flow cytometer and was expressed as percentage of intact cells after 24 hours exposure compared to vehicle exposed controls. For cytotoxicity assessment in the ToxTracker assay, the relative cell survival for the six different reporter cell lines was averaged.

Metabolic activation was included in the ToxTracker assay by addition of S9 liver extract from aroclor1254-induced rats. Cells are exposed to five concentrations of the test compounds in the presence of 0.25% S9 and required co-factors (RegenSysA+B) for 24 hours.

Three independent tests were carried out.

In case auto-fluorescence of the test substances was observed in the dose range finder, wild type mES cells were exposed to the compounds at the same concentrations as used in the ToxTracker. The mean fluorescence caused by the compound was then subtracted from the ToxTracker results of the respective compound.

3) Hypoxia assessment
To assess the induction of cellular hypoxia following exposure to the cobalt compounds, induction of a selection of HIF1ɑ target genes was determined using quantitative real-time PCR (qPCR). For the analysis of the expression for HIF1ɑ target genes in response to treatment, wild-type mES were treated with two doses of the test materials solved in mES cell culture medium in the absence of S9 for 8 hours in triplicate. After treatment, cells were lysed in Trizol, RNA was isolated and after cDNA synthesis, a qPCR was performed using primers for Hmox1, Slc2a1, Eno1, Ddit4 and Bnip3, using two technical replicates. HPRT and GAPDH were included as reference genes.

TEST CRITERIA
The ToxTracker assay was considered to have a positive response when a compound induces at least a 2 fold increase in GFP expression in any of the reporters. Activation of the Bscl2-GFP or Rtkn-GFP reporters indicate induction of DNA damage, Srxn1-GFP and Blvrb-GFP indicate induction of cellular oxidative stress and Ddit3-GFP activation is associated with the unfolded protein response. The Btg2-GFP reporter is controlled by the p53 tumor suppressor and is activated by DNA damage but can also be induced by
oxidative stress, hypoxia, metabolic stress and apoptosis. Only GFP inductions at compound concentrations that showed < 75% cytotoxicity are used for the ToxTracker analysis. Data from measurements > 75% cytotoxicity can not be interpreted in a meaningful way and are therefore discarded.

The No Observed Effect Level (NOEL) defines the highest concentrations where none of the ToxTracker reporters showed a >1.5-fold increase in fluorescence. The Lowest Observed Effect Level (LOEL) indicates the lowest concentration were at least one of the ToxTracker reporters showed an induction of ≥2.0. The No Observed Genotoxicity Effect Level (NOGEL) is the highest concentration that did not activate the Bscl2-GFP and/or Rtkn-GFP reporters with a >1.5-fold increase in fluorescence. The Lowest Observed Genotoxicity Effect Level (LOGEL) is the lowest concentration that activated the Bscl2-GFP and/or Rtkn-GFP genotoxicity reporters ≥2.0-fold.

For the analysis of the qPCR data, an induction more than 2-fold compared to the vehicle treated control sample was considered to be positive. A 2 fold increase corresponds to approximately one standard deviation above the untreated control samples.

DATA ANALYSIS
In order to allow comparison of induction levels of the ToxTracker reporter cell lines for large number of compounds Toxplot was developed. Toxplot imports raw GFP reporter data from the flow cytometer, calculates GFP induction levels and cytotoxicity, performs statistical analysis of the data and hierarchical clustering of the tested compounds, and visualises the data in a heatmap allowing convenient interpretation of obtained test results. ToxPlot software uses agglomerative hierarchical clustering to visualize the ToxTracker data. Agglomerative clustering uses the ‘bottom-up’ approach, which puts each observation in its own cluster and pairs of clusters are merged as one moves up the hierarchy. To compare the induction of the six GFP reporters for a collection of compounds, each with different biological reactivities, dose-response relationships and kinetics, Toxplot calculates for each compound the level of GFP induction for every individual reporter at a specified level of cytotoxicity (typically 10%, 25% and 50%). GFP induction levels are calculated by linear regression between two data points around the specified cytotoxicity level. In case the specified level of cytotoxicity can not be reached at the highest tested compound concentration, Toxplot displays the GFP induction level at this top concentration. In the heatmap, Toxplot marks the compounds that do not induce the selected level of cytotoxicity. Because the cytotoxicity for a compound can vary between the ToxTracker cell lines, calculations of the GFP induction levels of the individual reporters by Toxplot can slightly deviate from the GFP induction and cytotoxicity figures.
Examinations:
Please refer to the field "Details on study design" above.
Positive control:
Positive reference treatments with cisplatin (DNA damage; 0, 2.5, and 5 µM; solvent: PBS), diethyl maleate (oxidative stress; 0, 150, and 300 µM; solvent: DMSO), tunicamycin (unfolded protein response; 0, 2, and 4 µg/mL; solvent: DMSO) and aflatoxin B1 (metabolic activation of progenotoxins by S9; 0, 2.5, and 5 µM; solvent: DMSO) were included in all experiments. Solvent concentration was similar in all wells and never exceeded 1% for DMSO.
Details on results:
POSITIVE REFERENCE TREATMENT
The validity of the ToxTracker assay was confirmed. The genotoxic compound cisplatin showed induction of the DNA damage response (Bscl2, Rtkn) and p53-mediated cellular stress (Btg2). Diethyl maleate (DEM) induced primarily the oxidative stress related reporters Srxn1 and Blvrb, tunicamycin induced the unfolded/misfolded protein stress response (Ddit3). The positive control compound aflatoxin B1, induced the Bscl2 and Rtkn reporters when tested in the presence of S9 liver extract. Generally, the controls showed green fluorescent (GFP) induction levels compliant with historical data and demonstrated the functionality of the mouse embryonic stem (mES) reporter cell lines.
Please also refer to the field "Attached background material" below.

SOLUBILITY
For most compounds, the concentration in medium was >20 μg/mL. For for cobalt sulphide and tricobalt tetraoxide the concentration in medium was <1 μg/mL, indicating that these compounds are insoluble in mES culture medium.

CYTOTOXICITY
At the maximum tested concentrations in the absence of a metabolising system, compounds cobalt borate neodecanoate, cobalt fine powder, cobalt dichloride hexahydrate and cobalt (II) 4-oxopent-2-ene-2-olate induced significant levels of cytotoxicity (>50%) at the following concentrations:
- cobalt borate neodecanoate: 13.7 μg/mL
- cobalt fine powder: 10.0 μg/mL
- cobalt dichloride hexahydrate: 11.0 μg/mL
- cobalt (II) 4-oxopent-2-ene-2-olate: 9.4 μg/mL

In the presence of a metabolising system, there was no increased cytotoxicity observed for any of the compounds. The six ToxTracker reporter cell lines showed a comparable cytotoxic response to the test compounds. For this reason, the cell survival graphs in the GFP induction figures show the average cytotoxicity of the six different cell lines (please also refer to the field "Attached background material" below).
No autofluorescence was observed at the tested concentrations of any test compound.

GENOTOXICITY
When tested in the absence or presence of a metabolising system, none of the compounds showed an induction of the Bscl2-GFP and/or Rtkn-GFP genotoxicity reporters. The Bscl2-GFP reporter is associated with induction of promutagenic DNA lesions and DNA replication inhibition. Induction of the Rtkn-GFP genotoxicity reporter correlates with induction of DNA strand breaks.
Please also refer to the field "Attached background material" below.

OXIDATIVE STRESS
Cobalt borate neodecanoate, cobalt fine powder, cobalt dichloride hexahydrate and cobalt (II) 4-oxopent-2-ene-2-olate activated the Srxn1-GFP and Blvrb-GFP oxidative stress reporters significantly, when tested in absence or presence of a metabolising system (induction levels were either 5 - 7 fold, 5 - 8 fold, or 6 - 8 fold lower compared to the positive control diethyl maleate). Induction of the Srxn1-GFP reporter is associated with activation of the Nrf2 antioxidant response and activation of the Blvrb-GFP reporter is associated with the Hmox1 antioxidant response. Only compounds cobalt sulphide and tricobalt tetraoxide did not induce oxidative stress in ToxTracker.
Please also refer to the field "Attached background material" below.

P53 ACTIVATION
The compounds cobalt borate neodecanoate, cobalt fine powder and cobalt (II) 4-oxopent-2-ene-2-olate showed a weak (>1.5 fold) activation of the Btg2-GFP reporter in absence and/or presence of a metabolising system, but induction levels did not reach the 2-fold induction threshold for a positive ToxTracker result. The Btg2-GFP reporter is activated by p53 in response to various types of cellular stress such as protein damage, oxidative stress, DNA damage or cytotoxicity. For cobalt borate neodecanoate, cobalt fine powder and cobalt (II) 4-oxopent-2-ene-2-olate no activation of the Bscl2-GFP or Rtkn-GFP reporters for genotoxicity was observed. Therefore the activation of the Btg2-GFP marker in this project does not seem to be related to DNA damage. As all three compounds activated both markers for oxidative stress as well as HIF1ɑ target genes related to hypoxia, the activation of Btg2-GFP is likely to be related to these processes.
Please also refer to the field "Attached background material" below.

PROTEIN DAMAGE
The Ddit3-GFP reporter, associated with protein damage and the unfolded protein response, was only weakly (>1.5 but < 2-fold) activated in the absence of S9 by compounds cobalt borate neodecanoate, cobalt fine powder and cobalt dichloride hexahydrate.
Please also refer to the field "Attached background material" below.

HYPOXIA ASSESSMENT
Cellular hypoxia in the ToxTracker cells will activate the Hif1ɑ transcription factor and induce expression of various target genes. Cobalt borate neodecatenoate, cobalt fine powder, cobalt dichloride hexahydrate and cobalt (II) 4-oxopent-2-ene-2-olate all increased the expression of the HIF1ɑ target genes Hmox1, Slc2a1 and Ddit4 by more than 2-fold. Cobalt (II) 4-oxopent-2-ene-2-olate also induced the expression of Bnip3 and Eno1 and cobalt dichloride hexahydrate also activated Bnip3, but not Eno1.
Please also refer to the field "Attached background material" below.
Conclusions:
None of the tested compounds was classified as genotoxic in the ToxTracker assay. Activation of the Bscl2-GFP reporter is associated with DNA replication stress and induction of promutagenic DNA adducts. Activation of the Rtkn-GFP ToxTracker reporter is associated with induction of DNA double strand breaks and indicates induction of chromosome damage. Selective activation of Rtkn-GFP but not Bscl2-GFP genotoxicity reporter is often associated with an aneugenic mode-of-action of compounds. Significant levels of oxidative stress were observed for most of the tested compounds with the exception of the compounds cobalt sulphide and tricobalt tetraoxide. Only weak activation of the unfolded protein response was observed for the compounds cobalt borate neodecanoate, cobalt fine powder and cobalt dichloride hexahydrate. Treatment with the compounds cobalt borate neodecatenoate, cobalt fine powder, cobalt dichloride hexahydrate and cobalt (II) 4-oxopent-2-ene-2-olate increased expression of various hypoxia-associated genes.
Endpoint:
cell culture study
Type of information:
experimental study
Adequacy of study:
supporting study
Study period:
not specified
Reliability:
other: Any kind of reliability rating is not considered to be applicable, since the current in vitro test is not conducted/reported according to standardised guideline
Rationale for reliability incl. deficiencies:
other: Rating according to Klimisch criteria does not apply to mechanistic in vitro studies
Qualifier:
no guideline followed
Principles of method if other than guideline:
van den Brule & Lison (2016) and van den Brule & Lison (2018):
Based on the hypothesis that cobalt lung carcinogenesis is an outcome of chronic local tissue damage and sustained inflammation triggered by cobalt compounds, cytotoxicity and induction of pro-inflammatory mediators in lung epithelial cells was examined after in vitro exposure to a range of cobalt substances. The aim is to explore the possibility of grouping cobalt compounds for inhalation carcinogenicity within a read-across approach, based on in vitro data.
GLP compliance:
no
Type of method:
in vitro
Endpoint addressed:
carcinogenicity
other: mode of action analysis
Details on test animals or test system and environmental conditions:
Not applicable - Since this is an in vitro study there is no information on test animals.
Vehicle:
other: exposure medium (F12K + 1 % antibiotic-antimycotic)
Details on exposure:
van den Brule & Lison (2016):
Crystalline silica particles (Min-U-Sil, d50 1.4 μm US silica; positive control) and cobalt compounds (except cobalt dichloride hexahydrate) were directly suspended in exposure medium (F12K + 1 % antibiotic-antimycotic). Min-U-Sil was weighed the day before exposure (± 10 mg) and heated 2 h at 200 °C to remove endotoxins. Cobalt compounds (between 10-100 mg) were weighed on the day of exposure in a glass vial covered in aluminum. Prior to weighing, the glass vial was heated for 2h at 200 °C and left to cool before compounds were added. Cobalt sulphide, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate were suspended in exposure medium at 10 mg/mL and tricobalt tetraoxide at 3.33 mg/ml. Cobalt dichloride hexahydrate was suspended in nanopure H2O at 100 mg/mL and filter-sterilized (0.22 μm) before further dilution. Cobalt dichloride hexahydrate was dissolved in water before filtering to avoid Co precipitation with phosphates in the culture medium and possible depletion of Co from suspension during filtering. Stock suspensions were next serially diluted in exposure medium.

van den Brule & Lison (2018):
Cobalt compounds (except for cobalt dichloride hexahydrate, refer to van den Brule (2016) above) were weighed on the day of exposure in a glass vial covered with aluminum. Prior to weighing, the glass vial was heated for 2 hours at 200°C and left to cool before compounds were added.
Cobalt compounds were then suspended at 10 mM cobalt equivalents in exposure medium (F12K + 1 % antibiotic-antimycotic). Stock suspensions were next serially diluted in exposure medium. Calculations of Co-equivalent molar concentrations were based on the molecular weight of cobalt dichloride hexahydrate considering that it is 100 % pure, and on the percentage of cobalt for other compounds.
Analytical verification of doses or concentrations:
no
Duration of treatment / exposure:
van den Brule & Lison (2016):
HIF-1α quantification: 24 hour exposure (exception cobalt dichloride hexahydrate: 6, 24, 48, and 72 hours)
Cell viability assay: 24, 48, and 72 hour exposure (exception cobalt dichloride hexahydrate: 6, 24, 48, and 72 hours)
Cytokines: 72 hours (exception cobalt dichloride hexahydrate: 6, 24, 48, and 72 hours)

van den Brule & Lison (2018):
HIF-1α quantification: 24 hour exposure
Cell viability assay: 72 hour exposure
Frequency of treatment:
van den Brule & Lison (2016) and van den Brule & Lison (2018):
HIF-1α quantification, cell viability assay, and cytokine quantification: once
Remarks:
10-1000 μM cobalt equivalents (van den Brule & Lison (2016))
Remarks:
250 - 1000 µM cobalt equivalents (van den Brule & Lison (2018))
No. of animals per sex per dose:
van den Brule & Lison (2016) and van den Brule & Lison (2018):
not applicable
Control animals:
other: not applicable (van den Brule & Lison (2016) and van den Brule & Lison (2018))
Details on study design:
van den Brule & Lison (2016):
ENDOTOXIN MEASUREMENT
Cobalt dichloride hexahydrate was dissolved in PBS at concentrations of 640 and 64 μM and tested for endotoxin level with the Pierce LAL chromogenic endotoxin quantitation kit (Thermo Scientific). Cobalt dichloride hexahydrate was also spiked with 0.5 EU LPS/ml (1:1 CoCl2.6H2O 128 μM and endotoxin standard 1 EU LPS/mL).

CELL CULTURE AND EXPOSURE
- cell culture:
The type II alveolar epithelial cell line derived from a human lung carcinoma (A549 from ATCC, CCL-185) was cultured in F12K supplemented with 10 % fetal bovine serum and 1 % antibiotic-antimycotic (= complete medium). Cells were passaged every week in 75 cm² flasks. After washing cells with 10 mL phosphate-buffered saline, 3 mL trypsin 0.25 %-EDTA were added. Flasks were incubated 5 to 10 minutes at 37 °C. Trypsin was then neutralized by adding 7 ml complete medium. Cell suspensions were centrifuged for 10 minutes and resuspended in fresh complete medium. Cell number and viability were determined with Trypan Blue Solution in a Bürker cell under a light microscope.

1.5x10E5 cells/cm² were seeded in 96-well plates (50.000 cells/100 μL/well) or 24-well plates (300.10³ cells/mL/well) with complete medium and grown at 37 °C. After 24 hours, supernatants were discarded and cells were exposed to cobalt compounds (10-1000 μM in Co equivalents) or Min-U-Sil (200 μg/ml) in serum-free medium (100 μL for 96-well plates or 500 μL for 24-well plates).
After exposure (6, 24, 48 and 72 hours), all cell supernatants were collected and frozen at -80°C. Cells were then either tested for viability with the WST-1 assay (96-well plates) or assessed for HIF-1α (24-well plates). For HIF-1α, cells were washed twice with 500 μl PBS and lysed in 150 μL lysis buffer as recommended by the HIF-1α ELISA assay (15 minutes on ice, stored at -80 °C).

van den Brule (2018):
Fourteen cobalt compounds were tested against cobalt dichloride hexahydrate in the current study:
- cobalt carbonate
- cobalt acetylacetonate (Peconal H) Co(II)
- cobalt resinate
- cobalt oxalate
- cobalt (di)hydroxide
- cobalt diacetate tetrahydrate
- Cellcore® LCO (lithium cobalt oxide)
- cobalt dinitrate hexahydrate
- cobalt hydroxide oxide
- cobalt
- cobalt stearate
- cobalt octoate (aka bis (2-ethylhexanoate))
- cobalt propionate
- cobalt borate propionate
- cobalt borate octoate

ENDOTOXIN MEASUREMENT
Endotoxin was measured for the 14 tested compounds with the Pierce LAL chromogenic endotoxin quantitation kit (Thermo Scientific). Cobalt compounds were suspended in PBS to a concentration of 2000 μM cobalt equivalents and then diluted to 1000 μM (maximal concentration tested in the cell exposure experiments). Firstly, it was verified that the pH of the suspensions was between 6 and 8 with pH paper. The direct effect of the suspensions on the absorbance measured at the 405 nm wavelength was also evaluated : 150 μl endotoxin-free H2O and 50 μl stop reagent were added to 50 μl suspensions (1000 μM Co equivalent in PBS) or 50 μl PBS and absorbance was measured. After centrifugation, supernatants were transferred to a new plate and absorbance was measured again.

Cobalt compounds were tested at 1000 μM cobalt equivalents as such (unspiked), and spiked with 0.5 EU LPS/ml (1:1 Co compound 2000 μM + endotoxin standard 1 EU LPS/ml) to verify the absence of quenching by the tested compound (false negative). Briefly, 50 μl H2O, PBS, standards, unspiked or spiked samples were mixed with 50 μl LAL reagent and 100 μl substrate solution in a plate. Negative controls were also included: 50 μl unspiked samples, 150 μl H2O (please also refer to table 1 in the field "Any other information on materials and methods incl. tables" below). Absorbance was measured at 405 nm after addition of stop reagent (Abs A). The plate was centrifuged and 200 μl supernatants were transferred in a new plate. Absorbance was measured again at 405 nm (Abs B). Sample absorbances were calculated by 2 methods. In method A, Abs A of the negative controls was subtracted from the Abs A of spiked and unspiked samples. In method B, the average Abs B of the PBS blank replicates was subtracted from Abs B of spiked and unspiked samples. Method A takes into account particle absorbance by subtracting negative controls and method B by measuring the absorbance on suspension supernatant after centrifugation. For standards, the average absorbance of the H2O blank replicates was subtracted from standard absorbances. For each method, a standard curve was plotted and endotoxin concentration of each sample was interpolated from the linear regression.
Previously, cobalt dichloride hexahydrate was assessed and no endotoxin was detected.

CELL CULTURE AND EXPOSURE
- cell culture:
The type II alveolar epithelial cell line derived from a human lung carcinoma (A549 from ATCC, CCL-185) was cultured in F12K supplemented with 10 % fetal bovine serum and 1 % antibiotic-antimycotic (= complete medium). Cells were passaged every week in 75 cm² flasks. After washing cells with 10 mL phosphate-buffered saline, 3 mL trypsin 0.25 %-EDTA were added. Flasks were incubated 5 to 10 minutes at 37 °C. Trypsin was then neutralized by adding 7 ml complete medium. Cell suspensions were centrifuged for 10 minutes and resuspended in fresh complete medium. Cell number and viability were determined with Trypan Blue Solution in a Bürker cell under a light microscope.

- cell exposure
1.5x10E5 cells/cm² were seeded in 96-well plates (50.000 cells/100 μL/well) or 48-well plates (1.5x105 cells/300 μL/well) with complete medium and grown at 37 °C. After 24 hours, supernatants were discarded and cells were exposed to cobalt compounds (250 - 1000 μM Co equivalents) in serum-free medium (100 μL for 96-well plates or 350 μL for 48-well plates).
After exposure (24 and 72 hours), all cell supernatants were collected and frozen at -80°C. For HIF-1α (24 hour exposure, 48-well plates), cells were washed once with phosphate-buffered saline and lysed in 150 μL lysis buffer as recommended by the HIF-1α ELISA assay (15 minutes on ice, stored at -80°C). For assessing cell viability, cells were washed once with phosphate-buffered saline and tested for viability with the WST-1 assay (72 hour exposure, 96-well plates).
Examinations:
van den Brule & Lison (2016):
CELL VIABILITY ASSAY
After cell exposure in 96-well plates, 100 μl F12K/WST-1 (1:30) were added on cells. Cells were incubated at 37 °C and absorbance (A) measured at 480 nm and 680 nm (reference absorbance). When control (ctl) absorbance (A480 nm - A680 nm) reached 0.7-1.0 OD, 80 μL supernatants were transferred in new 96-well plates (to avoid interferences of cobalt compounds with the absorbance wavelength) and absorbance was measured. The latter absorbance values (A480 nm - A680 nm) were used to calculate cell viability and results were expressed as a ratio to non-exposed cells (ctl).

HIF-1α AND QUANTIFICATION
Levels of HIF-1α were assessed by ELISA in cell lysates and IL-8, IL-6, TNF-α and MCP-1 in cell supernatants according to manufacturer’s instructions (R&D Systems, human CXCL-8/IL-8 DuoSet, DY208, human IL-6 DuoSet, DY206, human TNF-α DuoSet, DY210, human CCL2/MCP-1 DuoSet, DY279, human/mouse total HIF-1 alpha Duoset IC, DYC1935-5). Cell lysates were centrifuged at 4°C (2000 g, 5 min) before testing.

van den Brule & Lison (2018):
CELL VIABILITY ASSAY (3 -4 independent experiments)
After cell exposure in 96-well plates, 100 μL F12K/WST-1 (1:30) were added on cells. Cells were incubated at 37 °C and absorbance (A) measured at 480 nm and 680 nm (reference absorbance). When control (ctl) absorbance (A480 nm - A680 nm) reached 0.7-1.0 OD, 80 μL supernatants were transferred in new 96-well plates (to avoid interferences of undissolved cobalt compounds with the absorbance) and absorbance was measured. The latter absorbance values (A480 nm - A680 nm) were used to calculate cell viability and results were expressed as a ratio to non-exposed cells (control=1).

HIF-1α QUANTIFICATION (3 - 4 independent experiments)
Levels of HIF-1α were assessed by ELISA in cell lysates according to manufacturer’s instructions (R&D Systems, human/mouse total HIF-1 alpha Duoset IC, DYC1935-5). Cell lysates were centrifuged at 4°C (2000 g, 5 min) and testing was performed on the supernatant.

STATISTICS
Differences between control and cobalt compounds were evaluated by a one-way analysis of variance (ANOVA), followed by a Dunnett multiple comparisons test. Statistical significance was considered at P < 0.05.
Positive control:
van den Brule & Lison (2016):
Crystalline silica particles (Min-U-Sil; 200 μg/mL) were selected as a tentative positive control for increasing the secretion of IL-6 and -8 (Hetland, R. B.et al. (2001). Exposures was 6, 24, 48 and 72 hours.

van den Brule & Lison (2018):
not specified

*Reference:
- Hetland RB, Schwarze PE, Johansen BV, Myran T, Uthus N, Refsnes M. Silica-induced cytokine release from A549 cells: importance of surface area versus size. Hum Exp Toxicol 2001; 20: 46 - 55.
Details on results:
van den Brule & Lison (2016):
DETERMINATION OF ENDOTOXIN LEVEL IN COBALT DICHLORIDE HEXAHYDRATE
No trace of endotoxin was detected in cobalt dichloride hexahydrate suspended in PBS (please refer to table 1 in the field "Any other information on results incl. tables" below).

KINETICS STUDY ON A549 CELLS UPON EXPOSURE TO COBALT DICHLORIDE HEXAHYDRATE (inflammation and hypoxia)
Min-U-Sil (positive control) did not increase the secretion of any of the tested cytokines. Cell viability was assessed with the WST-1 assay, cytokine secretion, i. e. inflammatory response, by ELISA on cell supernatants and HIF-1α accumulation, i. e. hypoxia-like response, by ELISA on cell lysates. WST-1 results indicated that 640 μM cobalt dichloride hexahydrate is cytotoxic after 48 and 72 hours. Tested concentrations below 640 μM did not appear to alter cell activity/viability. Intracellular HIF-1α was dose-dependently increased at all time points except at the cytotoxic concentration after 48 and 72 hours. MCP-1 was slightly increased after 48 and 72 hours. This increase was also dose-dependent except at the cytotoxic concentration. IL-8 was decreased or not modified by cobalt dichloride hexahydrate. IL-6 and TNF-α levels were around or below the limit of detection (9.38 and 15.6 pg/mL respectively). Cobalt dichloride hexahydrate did not reproducibly increase MCP-1, IL-6 or TNF-α.
Based on these results, it was decided to assess HIF-1α after 24 hours and to compare other cobalt compounds to a dose range of cobalt dichloride hexahydrate and not to a single dose. Secretion of cytokines was assessed after 72 hours exposure also in comparison to cobalt dichloride hexahydrate.
Please also refer to the field "Attached background material" below.

RESPONSES TO COBALT COMPOUNDS COMPARED TO COBALT DICHLORIDE HEXAHYDRATE
A549 cells were exposed to cobalt compounds (62.5 - 1000 μM cobalt equivalents) during 24, 48 and 72 hours. WST-1 results indicate that none of the 7 cobalt compounds affected cell viability/activity after 24 hours. After 48 hours, cell viability was strongly reduced at the highest concentration (1000 μM Co equivalents) of cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate. Both cobalt sulphide and tricobalt tetraoxide did not modify cell viability. After 72 hours, cell viability was strongly reduced from 500 μM cobalt equivalents of cobalt dichloride hexahydrate, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate. Cobalt metal powder 1 was less potent than the other cobalt metal powder 2 and was cytotoxic only at 1000 μM cobalt equivalents. Again, cobalt sulphide and tricobalt tetraoxide did not modify cell viability. Increased WST1 activity was observed after 48 and 72 h with Cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and Cobalt sulfate heptahydrate at concentrations below 1000 and 500 μM cobalt equivalents, respectively. These compounds might stimulate A549 cell activity at low cobalt concentrations.
HIF-1α was measured after 24 hours exposure in 3 independent experiments leading to very similar results. Cobalt dichloride hexahydrate, cobalt metal powder 2 and cobalt sulfate heptahydrate increased HIF-1α cell content in a dose-dependent manner up to a concentration of 500 μM cobalt equivalents. HIF-1α was still increased after exposure to 1000 μM cobalt equivalents but to a lesser extent. For cobalt metal powder 1 and cobalt monoxide, HIF-1α increase was completely dose-dependent. No modification of HIF-1α was observed with cobalt sulphide and tricobalt tetraoxide. These data indicate an association between cytotoxicity of cobalt compounds and their potential to induce HIF-1α stabilization/hypoxia-like responses. It is suggested that the cobalt content and potentially the released Co2+ is the main determinant of the hypoxia-like response to cobalt compounds, and that expressing doses in cobalt equivalents allows a better grouping of cobalt substances for their capacity to induce a HIF-1α response.
MCP-1, IL-8, IL-6 and TNF-α were measured after 72 hours exposure in 2-3 independent experiments. Cobalt compounds had no reproducible effects on MCP-1 release from A549 cells. Cytotoxic cobalt compounds (at the tested concentrations), i.e. cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate, decreased IL-8 release, while cobalt sulphide and tricobalt tetraoxide did not have any effect. As with cobalt dichloride hexahydrate, levels of IL-6 and TNF-α were near the limits of detection of the assays and were not modified by any cobalt compounds.

Please also refer to the field "Attached background material" below.

GROUPING OF COBALT COMPOUNDS
Data collected by CoRC and in the current study were summarized for the read-across approach for inhalation carcinogenicity. Briefly, it was found that cobalt compounds did not modify the release of pro-inflammatory cytokines (MCP-1, IL-8, IL-6 and TNF-α) by A549 cells, while some cobalt compounds increased HIF-1α cell content. It can be observe that the potential of cobalt compounds to increase intracellular HIF-1α in A549 cells is associated with a lysosomal bioaccessibility (solubilization), cytotoxicity in A549 (as determined in this study) and acute inflammatory activity after inhalation, whereas alveolar bioaccessibility and ROS generation are not related to other effects. These data suggests that cobalt compounds could be classified in 2 groups:
- Group 1: soluble cobalt compounds in lysosomal fluid, cytotoxic to A549 cells and able to increase intracellular HIF-1α, such as Cobalt dichloride hexahydrate, cobalt metals (cobalt metal powder 1 and cobalt metal powder 2), cobalt monoxide and cobalt sulfate heptahydrate tested in this study.
- Group 2: insoluble cobalt compounds, less cytotoxic to A549 cells and unable to increase intracellular HIF-1α, such as cobalt sulphide and tricobalt tetraoxide tested in this study.
Please also refer to the field "Attached background material" below.


van den Brule & Lison (2018):
ENDOTOXIN MEASUREMENT
pH of the suspensions of cobalt compounds was verified to be around pH 7. The absorbance of the particles/suspensions alone was measured at the assay wavelength (405 nm). Cobalt carbonate, cobalt resinate and cobalt oxalate slightly increased the absorbance but this was controlled by measuring the absorbance on supernatants after centrifugation.
Cobalt compound suspensions were then assessed for the presence of endotoxin. Both methods (A and B) revealed no trace of endotoxin in suspensions; spiked samples (+ 0.5 U endotoxin/mL) were positive, excluding the possibility of a false negative.
Please rfer for further results to the field "Attached background material" below.

CELL VIABILITY ASSAY
Findings were made as follows:
- cobalt resinate is more cytotoxic than cobalt dichloride hexahydrate.
- cobalt dihydroxide and cobalt nitrate, follow the same pattern as cobalt dichloride hexahydrate.
- cobalt acetylacetonate, cobalto oxalate, and cobalt acetate are slightly less cytotoxic than cobalt dichloride hexahydrate.
- cobalt carbonate, LCO, and cobalt hydroxide oxide are not cytotoxic up to 1000 μM cobalt equivalents.
- cobalt propionate and cobalt borate propionate (?) follow the same pattern as cobalt dichloride hexahydrate.
- cobalt octoate, and cobalt borate octoate are less cytotoxic than cobalt dichloride hexahydrate.
- cobalt stearate does not appear cytotoxic up to 1000 μM cobalt equivalents, but this was due to the fact that the compound largely stucked to the walls of the culture wells and was, therefore, not available to the cells.
Please also refer to the field "Attached background material" below.

HIF-1α QUANTIFICATION
Findings were made as follows:
- cobalt carbonate and cobalt oxalate dose-dependently increased intracellular HIF-1α.
- cobalt dichloride hexahydrate, cobalt dinitrate, cobalt diacetate, cobalt acetylacetonate and cobalt dihydroxide induced a HIF-
α increase at low doses, which was partially reduced at the highest concentration due to cytotoxicity.
- LCO, and cobalt hydroxide oxide did not impact HIF-1α levels at any of the concentrations tested.
- no HIF-1α response was recorded after exposure to cobalt resinate, as a result of the high cytotoxic activity of this compound.
- cobalt dichloride hexahydrate, cobalt octoate, cobalt borate octoate and cobalt borate propionate dose-dependently increased intracellular HIF-1α.
- cobalt propionate after an increase at low doses, the HIF-1α plateaued at higher concentrations (1000 μM cobalt equivalents) due to cytotoxicity.
- cobalt stearate did not impact HIF-1α levels at any of the concentrations tested.
Please also refer to the field "Attached background material" below.

GROUPING OF COBALT COMPOUNDS
The cobalt compounds tested in A549 cells can be grouped in 3 categories according to 4 main parameters: solubility (form) in the culture medium, solubility in lysosomal fluid, cytotoxicity to A549 cells, and the capacity to stabilize HIF-1α in A549 cells:

- Group 1a (cobalt chloride hexahydrate, cobalt sulfate, cobalt acetylacetonate, cobalt acetate, cobalt nitrate, and cobalt propionate) includes compounds (at least partly) soluble in the culture medium. These compounds are cytotoxic, stabilize HIF-1α, and can be considered to exert a biological activity through cobalt ions.

Group 1b (cobalt oxide, cobalt metal, cobalt carbonate, cobalt oxalate, cobalt dihydroxide, cobalt octoate, cobalt borate propionate, and cobalt borate octoate) includes those compounds that remain as particulate in culture medium, but are readily solubilized in lysosomal fluid. These compounds are generally cytotoxic, stabilize HIF-1α, and can be considered to exert a biological activity through Co ions.
Group 2 (cobalt oxide (II, III), cobalt sulfide, LCO, cobalt hydroxide oxide) includes compounds that remain as particulate in culture medium and are poorly dissolved in lysosomal fluid. These compounds have a low cytotoxicity, do not stabilize HIF-1α. They can be considered not to exert a biological activity through cobalt ions.

- Two compounds were excluded from this grouping:
Like in BEAS-2B cells (Report on the in vitro assessment of early markers of lung inflammation/damage in response to cobalt compounds in BEAS-2B epithelial cells, 25 April 2017), cobalt resinate was more cytotoxic to A549 cells than other cobalt compounds. As previously suggested (Report on the in vitro assessment of early markers of lung inflammation/damage in response to cobalt compounds in BEAS-2B epithelial cells, 25/04/2017), the fact that cobalt resinate did not elicit any HIF-1α response in A549 and BEAS-2B cells supports the view that its cytotoxic activity is not mediated by Co2+ ions. Cobalt resinate should be treated separately in a read-across procedure.
Cobalt stearate could not be reliably tested in vitro because of its hydrophobicity and buoyancy of the particulates in culture medium.

When considering all cobalt compounds, this grouping only partially predicts their in vivo activity (when data is available). Cobalt acetylacetonate, cobalt carbonate and cobalt oxalate appear as false positive (HIF-1α positive but no acute toxicity). However, when considering inorganic cobalt compounds only, the grouping appears to be predictive, with only one exception (cobalt carbonate). Most organic cobalt compounds appear to be outside the grouping framework proposed here.

Please also refer to the field "Attached background material" below.

van den Brule & Lison (2016):

Table 1: Endotoxin level in cobalt dihydrate hexahydrate

 

Absorbance (405 nm)

Endotoxin (U/mL)

result

PBS

-0.00615

0

negative

PBS

-0.00885

0

negative

CoCl2 640 μM

-0.00775

0

negative

CoCl2 640 μM

-0.00745

0

negative

CoCl2 64 μM

-0.00885

0

negative

CoCl2 64 μM

-0.00985

0

negative

CoCl2 64μM + 0.5 U LPS/mL

1.77875

0.5904974

positive

CoCl2 64μM + 0.5 U LPS/mL

1.72695

0.5652107

positive

Conclusions:
van den Brule & Lison (2016):
A549 cells were exposed to cobalt compounds. The results indicate that none of the 7 cobalt compounds affected cell viability/activity after 24 hours. After 48 hours, cell viability was strongly reduced at the highest concentration (1000 μM Co equivalents) of cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate. Both cobalt sulphide and tricobalt tetraoxide did not modify cell viability. After 72 hours, cell viability was strongly reduced from 500 μM cobalt equivalents of cobalt dichloride hexahydrate, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate. Cobalt metal powder 1 was less potent than the other cobalt metal powder 2 and was cytotoxic only at 1000 μM cobalt equivalents. Again, cobalt sulphide and tricobalt tetraoxide did not modify cell viability. Increased WST1 activity was observed after 48 and 72 h with Cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and Cobalt sulfate heptahydrate at concentrations below 1000 and 500 μM cobalt equivalents, respectively. These compounds might stimulate A549 cell activity at low cobalt concentrations.
Cobalt dichloride hexahydrate, cobalt metal powder 2 and cobalt sulfate heptahydrate increased HIF-1α cell content in a dose-dependent manner up to a concentration of 500 μM cobalt equivalents. HIF-1α was still increased after exposure to 1000 μM cobalt equivalents but to a lesser extent. For cobalt metal powder 1 and cobalt monoxide, HIF-1α increase was completely dose-dependent. No modification of HIF-1α was observed with cobalt sulphide and tricobalt tetraoxide.
Cobalt compounds had no reproducible effects on MCP-1 release from A549 cells. Cytotoxic cobalt compounds (at the tested concentrations), i.e. cobalt dichloride hexahydrate, cobalt metal powder 1, cobalt metal powder 2, cobalt monoxide and cobalt sulfate heptahydrate, decreased IL-8 release, while cobalt sulphide and tricobalt tetraoxide did not have any effect. As with cobalt dichloride hexahydrate, levels of IL-6 and TNF-α were near the limits of detection of the assays and were not modified by any cobalt compounds.

van den Brule & Lison (2018):
Cobalt acetylacetonate, cobalt resinate, cobalt oxalate, cobalt dihydroxide, cobalt diacetate and cobalt dinitrate, cobalt octoate, cobalt propionate, cobalt borate propionate and cobalt borate octoate reduced A549 cell viability in a dose-dependent manner after 72 hour exposure while cobalt carbonate, lithium cobalt oxide (LCO), cobalt hydroxide oxide, and cobalt stearate had no effect on cell viability. All cobalt compounds found to be cytotoxic, except cobalt resinate, also dose-dependently stabilized intracellular HIF-1α content after 24 hours. CoCO3 increased HIF-1α expression contrary to LCO, cobalt stearate and cobalt hydroxide oxide. The potential of these compounds to increase intracellular HIF-1α in A549 cells was generally associated with a high lysosomal bioaccessibility of cobalt (solubilization). The high cytotoxicity of cobalt resinate does not appear related to cobalt ions as this compound did not stabilize HIF-1α. Cobalt stearate could not be reliably tested in A549 cells because of its hydrophobicity and buoyancy of the particulates. Based on lysosomal bioavailability, cytotoxicity and HIF-1α stabilization in A549 cells, it is possible to group cobalt compounds tested in 2 groups : groups 1a and 1b can be considered to exert their biological activity through cobalt ions, group 2 does not appear to exert a biological activity through cobalt ions. This grouping did, however, not correlate perfectly with inflammatory activity after acute inhalation.

Description of key information

In order to collect further evidence on the mode of action for cobalt-induced carcinogenicity (see chapter “Summary and discussion of carcinogenicity” or the endpoint summary in IUCLID section 7.7- Carcinogenicity), further in vitro mechanistic studies were conducted. The aim of these studies was to elucidate the role of genotoxicity, oxidative stress and HIF1a activation, which are recognised key events in cells and tissues after cobalt exposure.

In the ToxTracker assay reported by Derr and Brandsma (2019 and Hendriks (2019) mouse embryonic stem (mES) reporter cell lines were exposed to 15 different cobalt substances. The ToxTracker assay is used to identify the biological reactivity and potential carcinogenic properties of these cobalt substances in a single test. The test system monitors activation of specific cellular signalling pathways for detection of the biological reactivity of the cobalt substances. ToxTracker consists of a panel of six different mES GFP reporter cell lines representing four distinct biological responses that are associated with carcinogenesis, i.e. general cellular stress, DNA damage, oxidative stress and the unfolded protein response. The results for the endpoints cytotoxicity, genotoxicity, oxidative stress, p53 activation, protein damage and HIF1a activation are summarised below.

Cytotoxicity: At the maximum tested concentrations in the absence of a metabolising system, compounds CoOx, CoPro, Co(OH)2, CoBNeo, Co , CoCl2 and CoAcAc induced significant levels of cytotoxicity (>50%). In the presence of a metabolising system, there was no increased cytotoxicity observed for any of the compounds. The six ToxTracker reporter cell lines showed a comparable cytotoxic response to the test compounds.

Genotoxicity: When tested in the absence or presence of a metabolising system, none of the compounds showed an induction of the Bscl2-GFP and/or Rtkn-GFP genotoxicity reporters. The Bscl2-GFP reporter is associated with induction of promutagenic DNA lesions and DNA replication inhibition. Induction of the Rtkn-GFP genotoxicity reporter correlates with induction of DNA strand breaks.

Oxidative stress: CoOx, CoPro, Co(OH)2 and Co2Et, CoBNeo, Co, CoCl2 and CoAcAc activated the Srxn1-GFP and Blvrb-GFP oxidative stress reporters significantly, when tested in absence or presence of a metabolising system. CoRes also activated the Srxn1-GFP reporters, but only a weak induction (>1.5 fold) of Blvrb-GFP which was not sufficient to reach the 2-fold threshold for a positive ToxTracker response. The activation of the oxidative stress reporter correlated with the Co concentration in medium. CoS and Co3O4 did not induce oxidative stress in ToxTracker. Induction of the Srxn1-GFP reporter is associated with activation of the Nrf2 antioxidant response and activation of the Blvrb-GFP reporter is associated with the Hmox1 antioxidant response.

P53 activation: Btg2-GFP was induced upon exposure to CoOx, Co(OH)2 and Co2Et only in the absence of S9. The substances CoBNeo, Co, CoPro and CoAcAc showed a weak (>1.5 fold) activation of the Btg2-GFP reporter in absence and/or presence of a metabolising system, but induction levels did not reach the 2-fold induction threshold for a positive ToxTracker result. The Btg2-GFP reporter is activated by p53 in response to various types of cellular stress such as protein damage, oxidative stress, DNA damage or cytotoxicity. For CoOx, Co(OH)2, Co2Et CoBNeo, Co and CoAcAc no activation of the Bscl2-GFP or Rtkn-GFP reporters for genotoxicity was observed. Thus, the activation of the Btg2-GFP marker does not appear to be related to DNA damage. As all six substances activated both markers for oxidative stress as well as HIF1ɑ target genes related to hypoxia, the activation of Btg2-GFP is likely to be related to these processes.

Protein damage: The Ddit3-GFP reporter, associated with protein damage and the unfolded protein response, was only weakly (>1.5 but < 2-fold) activated by CoBNeo, Co, CoRes and CoCl2 - the induction levels did not reach the 2-fold threshold for a positive ToxTracker response.

Hypoxia assessment by qPCR analysis of various HIF1ɑ target genes: Cellular hypoxia in the ToxTracker cells will activate the Hif1ɑ transcription factor and induce expression of various target genes. CoOx, CoPro, Co(OH)2 and Co2E, CoBNeo, Co , CoCl2 and CoAcAc all increased the expression of the HIF1ɑ target genes Hmox1, Slc2a1 and Ddit4 by more than 2-fold. CoRes activated HMOX, Slc2a1 and Ddit4 but not BNIP3. Slc2a1 activation was also observed for CoPro and Co2Et. CoAcAc also induced the expression of Bnip3 and Eno1 and CoCl2 activated Bnip3, but not Eno1.

 

The study by van den Brule and Lison (2018) investigated the cytotoxic activity and potential to stabilize the intracellular hypoxia-inducible factor (HIF)-1a content in alveolar epithelial A549 of 18 different cobalt substances.

Based on the experiments conducted in A549 cells with 18 different cobalt substances, it appears that these can be sub-divided in 3 groups on the basis of the four main endpoints investigated: (i) solubility in the culture medium, (ii) solubility (bioaccessibility) in lysosomal fluid, (iii) cytotoxicity to A549 cells, and (iv) the capacity to stabilize HIF-1a in A549 cells.

Group 1a includes compounds (at least partly) soluble in the culture medium. These compounds are cytotoxic, stabilize HIF-1a and can be considered to exert a biological activity through Co ions. This group comprises the substances: CoCl2, CoSO4, CoAcAc, CoNO3, CoPro.

Group 1b includes those compounds that remain as particulate in culture medium, but are readily solubilized in lysosomal fluid. These compounds are generally cytotoxic, stabilize HIF-1a, and can be considered to exert a biological activity through Co ions. This group comprises the substances: CoO, Co, CoCO3, CoOx, Co(OH)2, Co2Et, CoBPro, CoB2Et.

Group 2 includes compounds that remain as particulate in culture medium and are poorly dissolved in lysosomal fluid. These compounds have a low cytotoxicity, do not stabilize HIF1-a. They can be considered not to exert a biological activity through Co ions. This group comprises the substances: Co3O4, CoS, CoLiO2, CoOOH.

 

Both studies clearly demonstrate that there is a distinct different in the biological responses in vitro after exposure towards different cobalt substances. This information will be considered further in the mode of action analysis and the read-across of the cobalt category substances.

 

Additional information