Registration Dossier

Diss Factsheets

Administrative data

Description of key information

Acute Oral Toxicity:

The purpose of thein silico study was to predict the Acute Oral Toxicity of the test item Trimethylolpropane trisorbate based on a weight of evidence approach. This prediction was performed using the following QSAR models.

·       Leadscope Model Applier v2.4.1-36

·       US EPA Toxicity Estimation Software Tool (TEST) v4.2.1

·       iSafeRat®Mechanisms of Action tool

Using multiple QSAR models, i.e.,

Using the US EPA TEST v4.2.1 a predicted acute oral rat LD50value of 5754 mg/kg bw was determined.

All the five binary computational models integrated into the Leadscope Model Applier v2.4.1-36 provided negative predictions for Acute Oral toxicity. The query substance was within the applicability domain of Leadscope Model Applier. A negative prediction implies that no structural alerts for acute oral toxicity were matched for the query chemical. This model does not predict a LD50value.

iSafeRat®Mechanisms of Action tool predicts that the stomach will promote rapid hydrolysis of the substance at the ester bonds.

The hydrolysis products will be trimetylolpropane and sorbate, both of which are known to have low toxicity.

The final acute oral toxicity was predicted by applying a consensus method to the results derived from the individual models.

The Acute Oral Toxicity (Oral Rat LD50) value predicted for Trimethylolpropane trisorbate is 5754 mg/kg bw, whilst the oral LD50 values for the two major metabolites are both >10,000 mg/kg bw. These results are further supported by the negative predictions for acute oral toxicity derived using the Leadscope Model Applier (non toxic for all GHS categories) and also the results from iSafeRat® Mechanisms of action tool that suggests that the query substance and its metabolites are expected to be non-toxic. There is no reliable evidence to suggest that, under any circumstances, this substance may present a danger to especially vulnerable populations and in this case classification under GHS is not appropriate.

Based on this prediction, the substance is not classified under the CLP classification criteria, and also not classified under the GHS classification Criteria for Acute Toxicity.

The consensus result is reliable, however, the QSARs used in this report do not currently have the documentation (QMRF and QPRF reports) that is required for reporting a QSAR prediction under REACH sufficient to warrant a Klimisch score of 1, so in this case the result is associated with a Klimisch score of 2.

Key value for chemical safety assessment

Acute toxicity: via oral route

Link to relevant study records

Referenceopen allclose all

Endpoint:
acute toxicity: oral
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
25 July 2019
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with limited documentation / justification
Justification for type of information:
1. SOFTWARE
Leadscope Model Applier

2. MODEL (incl. version number)
Leadscope Model Applier v2.4.1-36

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC=CC=CC(=O)OCC(CC)(COC(=O)C=CC=CC)COC(=O)C=CC=CC

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
See attached Study Report

5. APPLICABILITY DOMAIN
See attached Study Report

6. ADEQUACY OF THE RESULT
If the substance falls within the applicability domain of the model, the result is expected to be reliable.
Qualifier:
equivalent or similar to guideline
Guideline:
OECD Guideline 401 (Acute Oral Toxicity)
Version / remarks:
QSAR Model version: Leadscope Model Applier v2.4.1-36
Deviations:
not applicable
Remarks:
QSAR result
Principles of method if other than guideline:
The following Acute Oral Toxicity models (corresponding to different GHS categories), all integrated into the Leadscope Model Applier, were applied:
a) Acute Rat Oral GHS 1
b) Acute Rat Oral GHS 2
c) Acute Rat Oral GHS 3
d) Acute Rat Oral GHS 4
e) Acute Rat Oral GHS 5

Each model provides a qualitative result (positive or negative). A positive result implies that the query chemical is classified for that given GHS category.
GLP compliance:
no
Remarks:
Not applicable/QSAR result
Test type:
other: QSAR result
Species:
other: Not applicable/QSAR result
Strain:
other: Not applicable/QSAR result
Details on test animals or test system and environmental conditions:
Not applicable / QSAR result
Route of administration:
other: Not applicable/QSAR result
Vehicle:
other: Not applicable/QSAR result
Details on oral exposure:
Not applicable / QSAR result
Doses:
Not applicable / QSAR result
No. of animals per sex per dose:
Not applicable / QSAR result
Control animals:
other: Not applicable/QSAR result
Details on study design:
Not applicable / QSAR result
Statistics:
Please refer to the attached prediction report from Leadscope Model Applier v2.4.1-36.
Preliminary study:
Not applicable / QSAR result
Remarks on result:
other: see 'Any other information on results incl. tables'
Mortality:
Not applicable / QSAR result
Clinical signs:
Not applicable / QSAR result
Body weight:
Not applicable / QSAR result
Gross pathology:
Not applicable / QSAR result
Other findings:
Not applicable / QSAR result

 Model Prediction 
Acute Rat Oral GHS 1 v1.0   Negative (inside applicability domain)
Acute Rat Oral GHS 2 v1.0   Negative (inside applicability domain)
Acute Rat Oral GHS 3 v1.0   Negative (inside applicability domain)
Acute Rat Oral GHS 4 v1.0 Negative (inside applicability domain) 
Acute Rat Oral GHS 5 v1.0 Negative (inside applicability domain) 
Interpretation of results:
GHS criteria not met
Conclusions:
All five binary computational models integrated into the Leadscope Model Applier v2.4.1-36 provided negative predictions for Acute Oral toxicity. The query substance was within the applicability domain of the Leadscope Model Applier. A negative prediction implies that no structural alerts for acute oral toxicity were matched for the query chemical. This model does not predict a LD50 value.
Executive summary:

A negative prediction with the Leadscope Model Applier implies that the test substance is not classified under the CLP classification criteria, and also not classified under the GHS classification Criteria for Acute Toxicity. The test substance falls inside the model's applicability domain however, the model does not currently have the documentation (QMRF and QPRF reports) that is required for reporting a QSAR prediction under REACH, sufficient to warrant a Klimisch score of 1, so in this case the result is associated with a Klimisch score of 2.

Endpoint:
acute toxicity: oral
Type of information:
(Q)SAR
Adequacy of study:
weight of evidence
Study period:
26 July 2019
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with limited documentation / justification
Justification for type of information:
1. SOFTWARE
U.S. EPA T.E.S.T

2. MODEL (incl. version number)
U.S. EPA T.E.S.T v4.2.1

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC=CC=CC(=O)OCC(CC)(COC(=O)C=CC=CC)COC(=O)C=CC=CC

4. SCIENTIFIC VALIDITY OF THE MODEL
See attached the User's Guide for U.S. EPA T.E.S.T v4.2.1

5. APPLICABILITY DOMAIN
See attached the User's Guide for U.S. EPA T.E.S.T v4.2.1

6. ADEQUACY OF THE RESULT
If the substance falls within the applicability domain of the model, the result is expected to be reliable.
Qualifier:
equivalent or similar to guideline
Guideline:
OECD Guideline 401 (Acute Oral Toxicity)
Version / remarks:
US EPA Toxicity Estimation Software Tool (TEST) v4.2.1
Deviations:
not applicable
Remarks:
QSAR result
Principles of method if other than guideline:
U.S. EPA T.E.S.T. model provides a quantitative prediction of the LD50 value for acute oral toxicity to rats by the use of a consensus of Hierarchical Clustering, The Food and Drug Administration (FDA) method, and the Nearest Neighbours approaches.

Hierarchical Clustering method – Weighted average of the predictions derived from several different models is calculated to estimate the toxicity of the query chemical.

FDA method – The toxicity prediction is based on the generation of a new model (at the runtime) with chemicals closest to the query chemical from a structural similarity point of view.

Nearest neighbour method – Toxicity of the query chemical is predicted as the average of the toxicity results for the three most structurally similar training set compounds.
Finally, a consensus result was derived taking into account the individual results from the above-listed models.
GLP compliance:
no
Remarks:
Not applicable/QSAR result
Test type:
other: QSAR result
Species:
other: Not applicable / QSAR result
Strain:
other: Not applicable / QSAR result
Details on test animals or test system and environmental conditions:
Not applicable / QSAR result
Route of administration:
other: Not applicable / QSAR result
Vehicle:
other: Not applicable / QSAR result
Details on oral exposure:
Not applicable / QSAR result
Doses:
Not applicable / QSAR result
No. of animals per sex per dose:
Not applicable / QSAR result
Control animals:
other: Not applicable / QSAR result
Details on study design:
Not applicable / QSAR result
Statistics:
Please refer to the attached prediction report from U.S. EPA T.E.S.T. v4.2.1
Key result
Dose descriptor:
LD50
Remarks:
5754 mg/kg bw
Based on:
other: QSAR result
Remarks on result:
not measured/tested
Remarks:
QSAR result

 Model Prediction for Oral rat LD50
 Hierarchical clustering -Log10(mol/kg)=1.53 
 FDA -Log10(mol/kg)=2.09
 Nearest Neighbour -Log10(mol/kg)=2.42 
 Consensus result derived from the three individual models 5754 mg/kg bw (average of the results in mg/kg bw) from the above three models 
Interpretation of results:
GHS criteria not met
Remarks:
Predicted Acute Oral Toxicity value is >5000 mg/kg bw
Conclusions:
The Acute Oral Toxicity (Oral Rat LD50) value predicted for Trimethylolpropane trisorbate is 5754 mg/kg.
Executive summary:

Using the US EPA TEST v4.2.1 a predicted acute oral rat LD50 value of 5754 mg/kg was determined.

The prediction is reliable, however the model does not currently have the documentation (QMRF and QPRF reports) that is required for reporting a QSAR prediction under REACH, sufficient to warrant a Klimisch score of 1, so in this case the result is associated with a Klimisch score of 2.

Based on this prediction, the substance is not classified under the CLP classification criteria, and also not classified under the GHS classification Criteria for Acute Toxicity.

Endpoint:
acute toxicity: oral
Type of information:
(Q)SAR
Remarks:
Prediction based on Mechanisms of Action
Adequacy of study:
weight of evidence
Study period:
30 July 2019
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
results derived from a valid (Q)SAR model and falling into its applicability domain, with limited documentation / justification
Justification for type of information:
1. SOFTWARE
iSafeRat Mechanisms of Action tool

2. MODEL (incl. version number)
iSafeRat Mechanisms of Action tool

3. SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL
CC=CC=CC(=O)OCC(CC)(COC(=O)C=CC=CC)COC(=O)C=CC=CC

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
Relevant publications on iSafeRat Mechanisms of Action (see under Attached justification below)

5. APPLICABILITY DOMAIN
Relevant publications on iSafeRat Mechanisms of Action (see under Attached justification below)

6. ADEQUACY OF THE RESULT
If the substance falls within the applicability domain of the model, the result is expected to be reliable.
Qualifier:
no guideline followed
Principles of method if other than guideline:
The iSafeRat® Mechanisms of Action tool determines from the chemical structure if any Mechanisms of Actions indicating acute oral toxicity are probable.
GLP compliance:
no
Remarks:
Not applicable / QSAR result
Test type:
other: QSAR
Species:
other: QSAR result
Strain:
other: QSAR result
Details on test animals or test system and environmental conditions:
Not applicable / QSAR result
Route of administration:
other: Not applicable / QSAR result
Vehicle:
other: Not applicable / QSAR result
Details on oral exposure:
Not applicable / QSAR result
Doses:
Not applicable / QSAR result
No. of animals per sex per dose:
Not applicable / QSAR result
Details on study design:
Not applicable / QSAR result
Statistics:
Not applicable / QSAR result
Preliminary study:
Not applicable / QSAR result
Remarks on result:
other: see 'Any other information on results incl. tables'
Mortality:
Not applicable / QSAR result
Clinical signs:
Not applicable / QSAR result
Body weight:
Not applicable / QSAR result
Gross pathology:
Not applicable / QSAR result
Other findings:
Not applicable / QSAR result

Using the MechoA classification scheme (Bauer et al. 2018a; Bauer et al. 2018b), Trimethylolpropane trisorbate (TMP trisorbate) is classified as :MechoA 3.2: Michaël addition with proteins residues, generating adducts for all species.

TMP trisorbate is predicted as acting through MechoA 3.2 because it has α,β-unsaturated ester groups. This moiety has soft electrophile properties which make it highly reactive toward soft nucleophile groups such as thiols, reacting through a Michaël addition. Thus, this substance is expected to bind to proteins, and more specifically to cysteine residues, which possess a thiol group (Aptula and Roberts 2006).

Moreover, it can be expected that TMP trisorbate will be, at least partially, enzymatically hydrolysed, thus generating Trimethylolpropane (TMP) and sorbate ions. This hydrolysis will generate acidity (Bauer 2017; Bicherel et al. 2018). As there are three ester groups in TMP trisorbate, and all may be potentially hydrolysed, one may expect a high level of acidity to be generated if there is significant hydrolysis. This could be a source of toxicity, depending on where this hydrolysis occurs. This toxicity is clearly apparent in aquatic species when comparing the ecotoxicity of diesters with the toxicity of mono-esters (Bauer 2017; Bicherel et al. 2018).

The hydrolysis products, however, are not expected to exhibit toxicity via the oral route, because TMP on one hand is predicted to act through MechoA 1.1 (non-polar narcosis; KREATiS, 2018) and is very hydrophilic, and the sorbate ion, is a commonly used food additive that has been well studied and is rapidly eliminated via the lungs as expired CO2 and gives no significant harmful effects (ECHA 2015).

In an acute oral toxicity test, the main site where ester hydrolysis is expected to occur is the stomach, which is already very acidic and contains esterases, such as gastric lipase, either coming from the mammalian organism itself or from bacteria living in the gastrointestinal tract. Therefore, the generation of acidity through the triple ester hydrolysis of TMP trisorbate is not expected to generate excess toxicity compared to non-polar narcosis, because the stomach has the built-in ability to handle high concentrations of acid. In addition, the hydrolysis products, TMP and sorbate, are expected to have low toxicity, as explained above. However, before its hydrolysis, TMP trisorbate is expected to show protein binding, because of the soft electrophile reactivity (MechoA 3.2) of α,β-unsaturated ester groups, which will bind to cysteine residues of proteins. Therefore, it is not possible to completely exclude the possibility of local toxicity effects to the oesophagus at levels higher than the stomach. The acute oral toxicity (LD50) of TMP is reported as >10,000 mg/kg and for sorbic acid the oral LD50 is reported as 10,500 mg/kg (both values taken from the ECHA disseminated dossiers for the substances).

References:

Aptula AO, Roberts DW. 2006. Mechanistic Applicability Domains for Nonanimal-Based Prediction of Toxicological End Points:  General Principles and Application to Reactive Toxicity.Chem Res Toxicol. 19(8):1097–1105. doi:10.1021/tx0601004.

Bauer F. 2017. Une meilleure caractérisation des mécanismes d’action toxique à partir de la structure moléculaire. [Mulhouse]: Université de Haute-Alsace.

Bauer FJ, Thomas PC, Fouchard SY, Neunlist SJM. 2018a.A new classification algorithm based on mechanisms of action. Comput Toxicol. 5:8–15. doi:10.1016/j.comtox.2017.11.001.

Bauer FJ, Thomas PC, Fouchard SY, Neunlist SJM. 2018b. High-accuracy prediction of mechanisms of action using structural alerts. Comput Toxicol. 7:36–45. doi:10.1016/j.comtox.2018.06.004.

Bicherel P, Bauer F, Thomas PC. 2018. Using mechanisms of toxic action to classify and predict ester ecotoxicity.

ECHA. 2015. Assessment Report - Potassium sorbate. Germany: ECHA Evaluation of active substances. [accessed 2018 Sep 27]. http://dissemination.echa.europa.eu/Biocides/ActiveSubstances/1350-08/1350-08_Assessment_Report.pdf.

Interpretation of results:
GHS criteria not met
Remarks:
The results from iSafeRat® Mechanisms of action tool that suggests that the query substance and its metabolites are not expected to be classified.
Conclusions:
The results from iSafeRat® Mechanisms of action tool that suggests that the query substance and its metabolites are expected to be not classified under CLP/GHS criteria.
Executive summary:

Using the MechoA classification scheme (Bauer et al. 2018a; Bauer et al. 2018b), Trimethylolpropane trisorbate (TMP trisorbate) is classified as :MechoA 3.2: Michaël addition with proteins residues, generating adducts for all species.

Based on this prediction, the substance is not classified under the CLP classification criteria, and also not classified under the GHS classification Criteria for Acute Toxicity.

The identification of the Mechanisms of Action for this substance is reliable. This result is associated with a Klimisch score of 2.

Endpoint conclusion
Endpoint conclusion:
no adverse effect observed
Dose descriptor:
LD50
Value:
5 754 mg/kg bw

Additional information

Justification for classification or non-classification

Based on the combination of the results derived from individual models:

The final Acute Oral Toxicity (Oral Rat LD50) predicted for Trimethylolpropane trisorbate is 5754 mg/kg bw, whilst the oral LD50 values for the two major metabolites are both >10,000 mg/kg bw.

These results are further supported by the negative predictions for acute oral toxicity derived using the Leadscope Model Applier (non toxic for all GHS categories) and also the results from iSafeRat® Mechanisms of action tool that suggests that the query substance and its metabolites are not expected to be toxic.

There is no reliable evidence to suggest that, under any circumstances, this substance may present a danger to especially vulnerable populations and in this case classification under GHS is not appropriate.

Based on this prediction, the substance is not classified under the CLP classification criteria, and also not classified under the GHS classification Criteria for Acute Toxicity.

The consensus result is reliable, however, the QSARs used in this report do not currently have the documentation (QMRF and QPRF reports) that is required for reporting a QSAR prediction under REACH sufficient to warrant a Klimisch score of 1, so in this case the result is associated with a Klimisch score of 2.