Registration Dossier

Administrative data

Link to relevant study record(s)

Reference
Endpoint:
basic toxicokinetics, other
Type of information:
(Q)SAR
Adequacy of study:
key study
Study period:
2019
Reliability:
2 (reliable with restrictions)
Rationale for reliability incl. deficiencies:
accepted calculation method
Justification for type of information:
1. SOFTWARE
OASIS TIMES platform for simulating metabolism
OECD QSAR Toolbox

2. MODEL (incl. version number)
• OASIS TIMES platform for simulating metabolism; with TIMES metabolic simulators used:
- in vitro rat liver S9;
- in vivo rat whole organisms;
- Skin metabolism.
• Toolbox 4.2 for searching analogues of the target chemicals;
• Documented metabolism from research publications and websites;
• Expert evaluation of the simulated metabolism.

3. 4 SMILES OR OTHER IDENTIFIERS USED AS INPUT FOR THE MODEL : the SMILE of 3,5,5 trimethylhexyl 3,5,5 trimethylhexanoate (RN CAS 59219-71-5) (monoconstituant) and SMILES of the representative constituents, with the monoconstitunt, of branched 3,5,5- trimethylhexanoate (UVCB).

4. SCIENTIFIC VALIDITY OF THE (Q)SAR MODEL
The aim of this study is to simulate plausible metabolic pathways of alkyl esters (target chemicals) and to compare the predictions with the documented metabolism and expert evaluations
Simulation of metabolism by the Tissue Metabolism Simulator (TIMES) is focused on the correct reproduction of experimentally documented metabolites. Published data on the metabolism of different organic chemicals, collected mainly from research publications in scientific journals and websites are extracted and stored into electronic databases. Hence, TIMES metabolic simulators (transformation table) represent electronically designed set of structurally generalized, hierarchically arranged biotransformation reactions which can be divided into two main types – enzymatic and non-enzymatic. The enzymatic transformations are Phase I (C-hydroxylation, ester hydrolysis, O-oxidative dealkylation, N-dealkylation, ꞷ-oxidation, etc.) and Phase II (glutathione conjugation, glucuronidation, sulfonation, etc). The main oxidation reactions are accomplished by the catalytic activity of the enzyme cytochrome P-450. The non-enzymatic transformations include hydrolysis of salts, tautomerism and molecular transformations of highly reactive intermediates. These molecular transformations describe the metabolism in skin, liver microsomes/S9 fraction and in vivo (whole organisms) experimental systems. Each transformation in the simulator consists of source and product structural fragments, and inhibiting “masks”. The latter are structural fragments used to increase the specificity of a given metabolic transformation by introducing structural boundaries of the transformed fragment. A probability (P) of occurrence is ascribed to each transformation, which determines its hierarchy in the transformation list. The transformation probabilities are approximated to reproduce documented metabolic transformations, expert knowledge and simulation of experimental endpoints, such as skin sensitization, genotoxicity, etc. Probabilities of occurrence are associated with the formal rates of the metabolic reactions within the timeframe of the test. Thus, the modeling is based on the set of molecular transformations which are hierarchically ordered according to transformation probabilities in order to reproduce the “logic” of the observed metabolism.
Metabolic transformations are characterized by another parameter - their reliabilities. Transformation reliabilities provide information for the degree transformations are supported by available documented metabolism data. This parameter is different from the probabilities of occurrence of the transformations. While probabilities of occurrence are associated with the formal rates of the reactions, the reliabilities of the transformations are associated with the extent they are supported by experimental metabolism data (i.e., experimentally observed molecular transformations).
OECD QSAR Toolbox is a software application intended to be used by governments, chemical industry and other stakeholders in filling gaps in (eco)toxicity data needed for assessing the hazards of chemicals. The Toolbox incorporates theoretical knowledge, experimental data and computational tools from various sources into a logical workflow.
The Toolbox simplifies the use of non-test methods for users with sufficient understanding of (eco)toxicology. The system includes the following main functionalities:
• Search for available experimental data: If chemicals have already been tested and the results are publicly available, users may not need to run any predictions. There are about 50 databases available in Toolbox with almost 80 000 chemicals and above 2 millions experimental data points.
• Profile chemicals: get a summary of relevant properties for your substance(s). The profiler results are used as criteria to find analogues, but they are also useful for preliminary screening or prioritization of substances.
• Find analogues of a substance and associated experimental data for them.
• Simulate metabolism: collect the transformation products observed or predicted for a target substance. Sometimes, metabolites and degradation products can also cause toxicity. Metabolism can be taken into account for a more robust selection of analogues. The modeling of metabolism in Toolbox is based on simplified simulators borrowed from TIMES system.
• Predict: fill a data gap for a substance by using data from analogues with a trend analysis, read-across or existing QSAR models.
• Customize and export the data matrix: a data matrix is the starting point for the work of toxicologists assessing categories.- Defined endpoint: metabolism
- Unambiguous algorithm: yes
- Defined domain of applicability: yes
- Mechanistic interpretation: addition of expert evaluation

5. APPLICABILITY DOMAIN
TIMES : A stepwise approach is used to define the applicability domain of the metabolic models. It is based on the principle of similarity between the query chemicals and the training set chemicals for which the model provides correct predictions. The higher is this similarity the more reliable should be the predictions.
The domain of the TIMES metabolic simulators consists of the following domain layers:
• Parametric layer – includes ranges of variation of log KOW, MW and other physical-chemical parameters for training set chemicals for which the model provides correct predictions;
• Structural layer – this is the structural space formed by atom-centered fragments (ACFs) of the chemicals for which the model has sensitivity of reproducing the documented metabolites greater than or equal to 70%.
- Appropriate measures of goodness-of-fit and robustness and predictivity: A chemical is considered as belonging to the parametric layer of the domain if its log KOW, MW and other physical-chemical parameters are within the ranges specified for these parameters in the parametric layer. Similarly, a chemical belong to structural domain of the metabolic model if its ACFs belong to the structural layer of the model.
In this respect, the applicability domain determines practically the interpolation space of the model. Still, belonging to model domain does not guarantee correct predictions. Just the reliability of the predictions is higher. Similarly, out of domain chemicals not always should have incorrect predictions. Just, reliability of predictions will be lower.

6. ADEQUACY OF THE RESULT
This QSAR evaluation give data to understand metabolic activity of similar ester compound for a knowledge of the toxicokinetic behavior required in Annexe VIII of REACh regulation.
Objective of study:
metabolism
toxicokinetics
Principles of method if other than guideline:
- Software tool(s) used including version:
OASIS TIMES platform for simulating metabolism
OECD QSAR Toolbox
- Model description: see field 'Justification for non-standard information', 'Attached justification' and/or 'Cross-reference'
- Justification of QSAR prediction: see field 'Justification for type of information', 'Attached justification' and/or 'Cross-reference'
- Details on the 3 TIMES simulators:
1- in vitro liver S9 metabolic simulator
The in vitro rodent microsomal/S9 mix metabolic simulator reproduces and predicts the metabolic pathways of xenobiotic chemicals for in vitro experimental systems such as rodent (mostly rat) liver microsomes and S9 fraction. The current in vitro rat liver metabolic simulator represents electronically designed set of 517 structurally generalized, hierarchically arranged biotransformation reactions. The metabolism training set contains experimentally documented in vitro metabolic pathways for 261 parent chemicals of a wide structural diversity, and 978 observed metabolites compiled into a searchable electronic database.
As configured in the TIMES system, options for generation of metabolic maps are set to reproduce no more than three levels or confined by products probability threshold (“weight” ≥ 0.1). This is a practical limitation for confining propagation of simulated metabolism maps based on observation that the documented metabolism usually ends up at third transformation level. Also, the limitation is related with the fact that metabolites generated up to the third level are enough to justify the prediction for in vitro mutagenicity endpoints.

2- In vivo rat (whole organisms) metabolic simulator
The in vivo rodent metabolic simulator reproduces and predicts the metabolic pathways of xenobiotic chemicals in vivo in rodents (mostly rats). The metabolism training set contains experimentally documented in vivo metabolic pathways for 647 structurally different parent chemicals, and 4382 observed metabolites compiled into a searchable electronic database. The current in vivo rat metabolic simulator represents electronically designed set of 622 structurally generalized, hierarchically arranged biotransformation reactions. These molecular transformations describe in vivo metabolism in rats accounting for the whole organism.
As configured in the TIMES system, propagation of in vivo metabolism is confined by a threshold of probabilities to generate metabolites (i.e., probabilities to produce metabolites to be ≥ 0.02).

3- Skin metabolic simulator
The skin metabolism simulator reproduced documented in vitro metabolism of 183 unique parent chemicals having 206 documented metabolic maps. Although in vitro data are used to build this metabolism simulator, given the absence of in vivo metabolism data it is assumed that in vitro data should allow adequately prediction of in vivo metabolism in skin. The current metabolic simulator represents electronically designed set of 276 structurally generalized, hierarchically arranged biotransformation reactions. These molecular transformations describe the metabolism of chemicals in the skin compartment and further used to simulate the enzymatic activation of chemicals when predict skin sensitization.
As configured in the TIMES system, options for generation of metabolic maps are set to reproduce no more than four levels of metabolism or confined by using threshold for probability of generated metabolites (“weight” ≥ 0.003). These limitations are related with the fact that the most probable metabolites associated with skin sensitization effect are generated up to the fourth level of metabolism.
GLP compliance:
no
Specific details on test material used for the study:
The QSAR approah had considered and compared the metabolism of the monoconstituent 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate (RN CAS 59219-71-5), and the representative constituents of branched-nonyl 3,5,5 trimethylhexanoate (UVCB).
Metabolites identified:
yes

4 SMILES were used as input for the models : the SMILE of the 3,5,5 trimethylhexyl 3,5,5 trimethylhexanoate (monoconstituent) and SMILES of representative constituents of branched-nonyl 3,5,5 trimethylhexanoate (UVCB)

 

1.    Metabolism of target chemicals

No documented metabolism of the target chemicals has been found in the scientific literature.

 

2.    Metabolism of structural analogues

Exhaustive search in the literature has been performed to identify analogues of the target chemicals having documented metabolism in in vitro S9, in vivo (whole organism) or skin. As a result, four analogues have been found only being common with the target chemicals with respect to the ester functionality. Structural similarity of the analogues has been ignored to some extent at the cost of the preserving ester functionality which, as shown further, determine the principle molecular transformations of target chemicals. In other words, structural analogues of the target chemicals having documented metabolism have been used to confirm that the main metabolic transformation is ester hydrolysis.      

Documented metabolism of these four aliphatic alkyl esters which are analogues of the target chemicals with respect to the ester functionality has been found in:

Ø In vitro rat liver S9 (one analogue);

Ø In vivo rat (three analogues);

Ø Skin metabolism (one analogue).

The four structural analogues with information for availability of their documented metabolism are presented in Table 2:

Table 2.Documentedmetabolismof structural analogues.

##

Analogues

In vitro ratliver S9 documented metabolism

In vivodocumented metabolism

Skin documented metabolism

1

CAS 540-88-5

tert-butyl acetate

N/A

Yes [6]

N/A

2

CAS 103-23-1

Bis (2-ethylhexyl) adipate

Yes [7]

Yes [7]

N/A

3

CAS 123-86-4

butyl acetate

N/A

Yes [8]

N/A

4

CAS 14062-23-8

4-Biphenylacetic acid, ethyl ester

N/A

N/A

Yes [9]

4-Biphenylacetic acid, ethyl ester (the last analogue in Table 2) is the most dissimilar analogue having commonality to the target compounds only with respect to the presence of an ester group. We have used it in the current analysis, because this is the only aliphatic acid ester analogue for which documented metabolism in skin has been found in literature.

In vitro rat liver S9 metabolism

Simulated in vitro rat liver S9 metabolic map of the structural analogue bis(2-ethylhexyl)adipate (CAS: 103 -23 -1). The major pathway is ester hydrolysis which is obtained with very high probability (P > 0.9). Ester hydrolysis leads to the formation of two metabolitesalcohol and acid. The mono- and dicarboxylic acids (highlighted in green) are documented and simulated correctly by TIMES. It is important to mention that not all simulated metabolites are documented. However, it does not mean that these metabolites are not formed in reality – they are just not documented. Their adequacy is usually analyzed by evaluation of the simulated metabolic “maps” by third-party experts

§ Documented in vitro metabolites are simulated correctly.

§ The major metabolic pathway is enzymatic ester hydrolysis,which results in the formation of acid and alcohol metabolites. 

§ Simulated are some additional oxidation reactions - the secondary hydrolysis product could be oxidized via aldehyde to the corresponding acid.

§ No Phase II metabolites are generated up to three levels of simulation

 

In vivo rat (whole organism) metabolism

Simulated in vivo metabolic maps of three structural analogues are presented in this section.

 

Analogue # 1 -Chemical ID:Name:tert-butylacetate; CAS # 540-88-5

Analogue # 2 -Chemical ID: Name: bis(2-ethylhexyl) adipate; CAS # 103-23-1

Analogue # 3 -Chemical ID:Name: butyl acetate; CAS # 123-86-4

 

§ Documented in vivo metabolites are simulated correctly.

§ The major simulated metabolic transformation is ester hydrolysis which occurs as a first step (with high probability, P > 0.9) and leads to the formation of acid and alcohol.

§ The alcohol metabolite with a primary alcohol group is further oxidized via aldehyde intermediate to the corresponding carboxylic acid.

§ Significant amount ofPhase II metabolites is obtained in vivo as compared with in vitro metabolism.

Skin metabolism

Documented skin metabolism has been found for one structural analogue (4-Biphenylacetic acid, ethyl ester, CAS # 14062-23-8).

The selected analogue (4-Biphenylacetic acid, ethyl ester) having benzyl-type biphenyl fragment next to the alkyl ester functionality is structurally dissimilar to some extent to the targets. Still, ester hydrolysis occurs as primary metabolic transformation forming an acid and alcohol in skin. This metabolic reaction is more facilitated in the presence of benzyl-type biphenyl fragment due to its slight electron-withdrawing effect. This could increase the partial positive charge on the carbonyl carbon atom and the reactivity of the ester carboxyl group towards the SN2-type hydrolysis.

It should be stressed that carboxylesterase enzymatic system which is responsible for the ester hydrolysis in liver is also expressed in skin

Significant amount of Phase II metabolites is obtained by the skin metabolism simulator. However, one should emphasize that the specific enzymatic activity of Phase II UDP-glucuronosyltransferase in skin is reported to be about 10 – 50 % of that in rat liver microsomes. On the other hand, the same UDP-glucuronosyltransferase is significantly more expressed in vivo than in in vitro S9 due to the fact that the whole organism in vivo is multi-organ system involving not only liver but also bile, GI tract, etc. Hence, Phase II glucuronides of a given substrate should be essentially more abundant in vivo followed by in vitro microsomal/S9 system and, ultimately, skin. Nevertheless, the products of esters hydrolysis in skin – acid and alcohol are directly involved in Phase II metabolic reactions, instead of undergoing Phase I aliphatic C-oxidations which occur with the in vitro microsomal/S9 and in vivo.This could be justified by the fact that Phase I oxidizing enzymes such as cytochromes (CYPs), alcohol/aldehyde dehydrogenase, aldehyde oxidase, etc. show lower specific activities in skin than in microsomes and liver. This explains why oxidative reactions such as aliphatic C-oxidations, occurring mainly in liver and following ester hydrolysis are more pronounced in vivo than in skin. The lower activities of Phase I oxidizing enzymes in skin causes an increase in the relative role of Phase II conjugation reactions, although Phase II enzymatic activities in skin are lower as compared to in vivo and in vitro S9 systems.

Similar result is obtained by the TIMES in vivo metabolic simulator.

 

Conclusions:

§ Documented metabolites are simulated correctly by TIMES.

§ Same enzymes controlling metabolism inin vivorat are involved in the metabolism in skin. However, activity of these enzymes is different in all three metabolic systems which provides some specificity of skin metabolism:      

       Activity of Phase I oxidizing enzymes, such as alcohol/aldehyde dehydrogenase and aldehyde oxidase, is lower in skin as compared to S9 and liver.

       Lower activity of Phase I oxidizing enzymes increases the relative importance of Phase II metabolic reactions (such as glucuronidation) in skin which, otherwise, are less expressed in skin as compared to in vivo and in vitro S9.  

§ 

Provided structural analogue is slightly different from the target chemicals due to the presence of a biphenyl-type fragment.

§ Similarity between skin andin viv orat metabolism has been found with respect to simulation ofPhase IImetabolites. Both simulators provide large amount ofPhase II metabolites.

§ The major simulated metabolic transformation is enzymatic ester hydrolysis which occurs with probability P > 0.9 and results in the formation of acid and alcohol.

General conclusions on simulated metabolism of the structural analogues:

 

A.   Documented metabolites are simulated correctly by all three (in vitroS9 mix,in vivowhole organism and skin) TIMES metabolic simulators. The correctly simulated ester hydrolysis pathway is a premise for adequate model predictions.

B.   Same pattern is obtained by all three TIMES metabolic simulators:

a)    Ester hydrolysis is first and major transformation obtained with high probability (P > 0.9).

b)    Primary alcohol is documented and simulated. This product of ester hydrolysis is further oxidized via aldehyde intermediate to carboxylic acid.

C. Activity of Phase II enzymes such as UDP-glucuronosyltranferase is more pronounced for:

a) In vitro microsomal/S9 system as compared to skin;

b) In vivo (whole organism) as compared to in vitro.

D. However, the relative amount of Phase II metabolites is higher in skin as compared with S9 and in vivo at the cost of the lower Phase I (oxidative) enzymatic activities.

E. Phase II metabolic transformations are acting as detoxification reactions in vivo by increasing the water solubility (hydrophilicity) of the substrates, and facilitating their renal and biliary excretion via the resulting Phase II conjugates

 

3.    Simulating metabolism of the target chemicals

In vitro rat liver S9 metabolism

All target chemicals, considered as parent compounds, belong 100% to the parametric and above 80% to the structural domain. Hence, the prediction could be considered as reliable at domain threshold > 80%.

All target chemicals have same metabolic pattern which is similar to the S9 metabolic pattern of the analogue:

·      The first and major metabolic transformation is enzymaticester hydrolysisoccurring with very high probability (P > 0.9);

·      Alcohol product is further oxidized via aldehyde intermediates to corresponding carboxylic acids;

·      The other metabolic product - carboxylic acid undergoes directPhase IIglucuronidation.

·      Predictions could be considered as reliable at domain threshold above 80%.

 

In vivo(whole organisms) metabolism

All target chemicals belong 100% to the parametric and above 75% to the structural domain. Hence, the prediction could be considered as reliable at domain threshold > 75%.

in vivo ester hydrolysis leads to the formation of two metabolic productsalcohol and acid. The alcohol is further oxidized via aldehyde intermediate to the corresponding carboxylicacids. This metabolic pathway is common for both in vitro and in vivo metabolism of this metabolic product.However, as light difference with respect to the second metabolic product (the acid) has been found in in vivo,as reported in literature.Concisely, the acid metabolite contains amethyl substituent at β-position which blocks certain steps in the β-oxidation,typical for straight-chain fatty carboxylic acids in vivo.Alternatively, ω-hydroxylation at the terminal C{sp3}carbon atom is more plausible metabolic pathway as shown for some saturated, branched chain carboxylic acids. Example for such ω-hydroxylation reaction is provided for Phytanicacid(CAS: 14721-66-5) as illustrated below

 

Figure. Metabolic ω-hydroxylation of Phytanic acid

 

ω-Hydroxylation can be further followed by oxidation of the primary alcohol to aldehyde, which is then oxidized to the corresponding carboxylic acid. Ultimately, dicarboxylic acids might be formed.

Hence, the ω-hydroxylation of acids, as ester hydrolysis products, is a slightly different in vivo metabolic pathway as compared to in vitro S9 rat, where the corresponding acid is directly involved in Phase II glucuronidation. Analogically, in skin metabolism (see next section) the acid obtained by the ester hydrolysis is also involved in Phase II reactions.

Conclusions:

Same in vivo metabolic pattern is obtained for the four target chemicals:

•       Ester hydrolysis – first and major metabolic transformation with very high probability (P > 0.9);

•       Two products of ester hydrolysis are obtained – alcohol and acid. Alcohol is further oxidized to corresponding carboxylic acids, whereas acid undergoes hydroxylation at the terminal (ω) carbon, ultimately leading to Phase II conjugates.

•       Phase II reactions are more pronounced in vivo than in in vitro S9.

•       Prediction could be considered as reliable at domain threshold >75%.

 

 Skin metabolism

The predictions are reliable given the fact that all target chemicals belong 100% to parametric and structural domain.

Conclusions:

•       Simulated skin metabolism is similar to the metabolism obtained by the in vitro S9 and in vivo mammalian metabolic simulators:  

•       Ester hydrolysis is the first and major metabolic transformation which occurs with very high probability (P > 0.9).

•       Both ester hydrolysis products further undergo Phase II conjugation reactions.

•       Same enzymes that control in vitro S9 and in vivo (whole organism) metabolism are involved in skin metabolism with activity  : in vivo > in vitro S9 > skin. The corresponding differences in enzymes activity provide specificity of the skin metabolism:

•       Phase I oxidizing enzymes such as alcohol/aldehyde dehydrogenase and aldehyde oxidase show lower activity in skin as compared to liver.

•       Lower activity of Phase I enzymes in skin increases the relative amount of Phase II metabolites as compared to in vivo (whole organism) and in vitro S9 metabolism.

•       The products (an acid and alcohol) of enzymatic hydrolysis in skin undergo direct Phase II metabolic reactions rather than Phase I aliphatic C-oxidation as in in vitro S9 and in vivo metabolism.

•       Kinetics of biotransformations are not provided in the current TIMES metabolic simulators reproducing metabolic transformation statically – within the time frame of the experimental conditions.  

•       The predictions are considered as reliable as all target chemicals belong 100% to parametric and structural domain.

 

Conclusions:
A. No documented metabolism of the target chemicals has been found.
B. Structural analogues of the target chemicals having documented metabolism have been used to confirm the main metabolic transformation - ester hydrolysis.
C. No difference between simulated metabolism pattern of the target chemicals within each of the analyzed (in vivo, in vitro S9, skin) metabolic systems.
D. Commonality of simulated metabolism of the target chemicals with respect to all three metabolic systems is the first and major metabolic transformation, enzymatic ester hydrolysis, which occurs with very high probability (P > 0.9).
E. Metabolic pattern of the acid hydrolysis product (left branch in the maps) in all three metabolic systems could be summarized as follows:
a. Similar metabolism has been found in in vitro microsomes/S9 and skin, where the acid is directly involved in Phase II conjugation reactions.
b. Slightly different transformation has been found for in vivo metabolism, where the acid undergoes hydroxylation at the terminal (ω) carbon, ultimately leading to Phase II conjugates.
F. Metabolism of the alcohol hydrolysis product (right branch in the maps) in all three metabolic systems could be summarized as follows:
a. Similarity has been found for in vitro microsomes/S9 and in vivo metabolism. The only difference is an additional Phase II glucuronidation reaction in vivo which does not occur in vitro S9.
b. Slightly different transformation has been found in skin metabolism due to the lower activity of Phase I oxidizing enzymes as compared to in vitro S9 and in vivo systems. Lower activity of these oxidizing enzymes in skin increases the relative importance of Phase II metabolic reactions.
G. Kinetics of biotransformations are not provided in the current TIMES metabolic simulators reproducing metabolic transformation statically – within the time frame of the experimental conditions. Nevertheless, the different activity of enzymes in the three metabolic systems is not expected to affect significantly the rate of metabolic reactions.
H. According to our experts, no alerts associated with skin sensitization and genotoxicity could be expected in the target chemicals and the obtained metabolites.
I. All predictions could be considered as reliable given the fact that all target chemicals belong 100% to the parametric domain and above 75% to the structural domain.
Executive summary:

The QSAR Approach developped by the Laboratory of Mathematical Chemistry, Bourgas, Bulgaria, consist in simulating metabolism of aliphatic alkyl esters by TIMES.

4 combined elements were used :

•       OASIS TIMES platform for simulating metabolism; with TIMES metabolic simulators used:

•       in vitro rat liver S9;

•       in vivo rat whole organisms;

•       Skin metabolism.

•       Toolbox 4.2 for searching analogues of the target chemicals;

•       Documented metabolism from research publications and websites;

•       Expert evaluation of the simulated metabolism.

SMILES of 4 target substances were used as input for the model : the SMILE of the Hexanoic acid, 3,5,5-trimethyl-, 3,5,5-trimethylhexyl ester (RN CAS 59219-71-5) (monoconstituant) and SMILES of the reprensentative constituents, (with the monoconstituant) of the branched 3,5,5- trimethylhexanoate (UVCB)

3 Simulators were considered :

       In vitro rat liver S9 (one analogue);

       In vivo rat (three analogues);

       Skin metabolism (one analogue).

Documented metabolites are simulated correctly by all three (in vitro S9 mix, in vivo whole organism and skin) TIMES metabolic simulators. The correctly simulated ester hydrolysis pathway is a premise for adequate model predictions.

Same pattern is obtained by all three TIMES metabolic simulators:

a)       Ester hydrolysis is first and major transformation obtained with high probability (P > 0.9).

b)       Primary alcohol is documented and simulated. This product of ester hydrolysis is further oxidized via aldehyde intermediate to carboxylic acid.

Formation of Phase II metabolites is more pronounced in vivo and in skin than in in vitro S9 metabolism.

For the simulating metabolism of the target chemicals, following results were obtained:

A. No documented metabolism of the target chemicals has been found.

B. Structural analogues of the target chemicals having documented metabolism have been used to confirm the main metabolic transformation - ester hydrolysis.

C. No difference between simulated metabolism pattern of the four target chemicals within each of the analyzed (in vivo, in vitro S9, skin) metabolic systems.

D. Commonality of simulated metabolism of the four target chemicals with respect to all three metabolic systems is the first and major metabolic transformation, enzymatic ester hydrolysis, which occurs with very high probability (P > 0.9).

E. Metabolic pattern of the acid hydrolysis product (left branch in the maps) in all three metabolic systems could be summarized as follows:

a. Similar metabolism has been found in in vitro microsomes/S9 and skin, where the acid is directly involved in Phase II conjugation reactions.

b. Slightly different transformation has been found for in vivo metabolism, where the acid undergoes hydroxylation at the terminal (ω) carbon, ultimately leading to Phase II conjugates.

F. Metabolism of the alcohol hydrolysis product (right branch in the maps) in all three metabolic systems could be summarized as follows:

a. Similarity has been found for in vitro microsomes/S9 and in vivo metabolism. The only difference is an additional Phase II glucuronidation reaction in vivo which does not occur in vitro S9.  

b. Slightly different transformation has been found in skin metabolism due to the lower activity of Phase I oxidizing enzymes as compared to in vitro S9 and in vivo systems. Lower activity of these oxidizing enzymes in skin increases the relative importance of Phase II metabolic reactions.  

G. Kinetics of biotransformations are not provided in the current TIMES metabolic simulators reproducing metabolic transformation statically – within the time frame of the experimental conditions. Nevertheless, the different activity of enzymes in the three metabolic systems is not expected to affect significantly the rate of metabolic reactions.

H. According to our experts, no alerts associated with skin sensitization and genotoxicity could be expected in the target chemicals and the obtained metabolites.  

I. All predictions could be considered as reliable given the fact that all target chemicals belong 100% to the parametric domain and above 75% to the structural domain.  

Description of key information

Key value for chemical safety assessment

Bioaccumulation potential:
no bioaccumulation potential

Additional information

Basic toxicokinetics

One study was available to estimate the toxicokinetic behaviour of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate (CAS 59219-71-5) and a bibliographic search complete them :

-       QSAR approach :  Simulating metabolism of aliphatic alkyl esters by TIMES - Laboratory of Mathematical Chemistry, Bourgas, Bulgaria (2019)

In accordance with Annex VIII, Column 1, 8.8.1, of Regulation (EC) 1907/2006 and with ‘Guidance on information requirements and chemical safety assessment Chapter R.7c: Endpoint specific guidance’ (ECHA, 2014), an assessment of the toxicokinetic behaviour of the target substance 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate (CAS 59219-71-5) was conducted to the extent that can be derived from the relevant available information. This comprises a qualitative assessment of the available substance specific data on physico-chemical and toxicological properties according to the Chapter R.7c Guidance document (ECHA, 2014) and taking into account further available information from source substances. The substance 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate (CAS 59219-71-5) is a monoconstituent with a highly branched C9-alcohol moiety and a highly branched C9-acid moiety.

Physico-chemical properties

3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate has a molecular weight 284.48 g/mol. It is a liquid at 20 °C with melting point > - 20 °C with a water solubility of < 0.05 mg/L at 20 °C. The log Pow was estimated to be 7.37 and the vapour pressure was calculated to be 0.0193 Pa at 20 °C.

Absorption

Absorption is a function of the potential for a substance to diffuse across biological membranes. The most useful parameters providing information on this potential are the molecular weight, the octanol/water partition coefficient (log Pow) value and the water solubility. The log Pow value provides information on the relative solubility of the substance in water and lipids (ECHA, 2014).

Oral

In general, molecular weights below 500 and log Pow values between -1 and 4 are favourable for absorption via the gastrointestinal (GI) tract, provided that the substance is sufficiently water soluble (> 1 mg/L). Lipophilic compounds can be taken up by micellar solubilisation by bile salts, but this mechanism may be of particular importance for highly lipophilic compounds (log Pow > 4), in particular for those that are poorly soluble in water (≤ 1 mg/L) as these would otherwise be poorly absorbed. Solids must be dissolved before absorption; the degree depends on the water solubility (Aungst and Shen, 1986; ECHA, 2014).

Some of the physico-chemical characteristics (log Pow and water solubility) of the substance are in a range that indicate poor absorption from the gastrointestinal (GI-) tract following oral ingestion, while the molecular weight and physical state favours uptake. Micellular solubilisation is likely to occur; although it is unclear to what degree it will affect the total absorption rate of the substance.

The indications that the target substance 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate has relatively low oral absorption and/or low acute toxicity due to its physico-chemical characteristics are supported by the available data on acute oral toxicity. In a study in which female mice were administered a single dose of 5 mL/kg bw (equivalent to 4350 mg/kg bw, based on the density) of the target substance, there was no mortality and no adverse clinical signs were noted (Masson, 1986).

In the available subacute (28-day) oral toxicity study, however, adverse effects were observed at the highest dose level of 1000 mg/kg bw/day (Manciaux, 2001). 4/10 females died or were sacrificed prematurely, and the animals were in poor health prior to death. The liver was clearly affected at all dose levels with a dose-related severity and increase in the number of affected animals. The liver effects (increased hepatic enzyme activity, increased liver weight and size, steatosis, hepatocellular hypertrophy) were mainly attributed to the increased load caused by exposure to a fatty ester. The kidney was also a target organ; the observed effects were considered to be specific to male rats (males) or an adaptation to the exposure to fatty esters (females). Although there is no data on the oral absorption percentage of the target substance, it is assumed the absorption will be 100% due to the effects observed in the subchronic oral toxicity study allowing a worst case assumption.

The potential of a substance to be absorbed in the GI-tract may be influenced by several parameters, like: chemical changes taking place in GI-fluids, as a result of metabolism by GI-flora, by enzymes released into the GI-tract or by hydrolysis. These changes will alter the physico-chemical characteristics of the substance and hence predictions based on the physico-chemical characteristics of the parent substance may in some cases no longer apply (ECHA, 2014).

In general, alkyl esters are readily hydrolysed in the GI-tract, blood and liver to the corresponding alcohol and fatty acid by the ubiquitous carboxylesterases. There are indications that the hydrolysis rate in the intestine catalysed by pancreatic lipase is lower for alkyl esters than for triglycerides, which are the natural substrates of this enzyme. The hydrolysis rate of linear esters increases with increasing chain length of either the alcohol or acid. Branching reduces the ester hydrolysis rate, compared with linear esters (Mattson and Volpenhein, 1969, 1972; WHO, 1999). It is therefore unclear what percentage of the ester will be hydrolysed.

Based on the generic information on hydrolysis of alkyl esters, the target 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate is expected to be enzymatically hydrolysed to the highly branched 3,5,5-trimethylhexanoic acid and 3,5,5-trimethylhexan-1-ol at a relatively slow rate.

In general, free fatty acids and alcohols are readily absorbed by the intestinal mucosa. Within the epithelial cells, fatty acids are mainly (re-)esterified with glycerol to triglycerides. Branching is likely to reduce the absorption rate (Greenberger et al., 1966; IOM, 2005; Mattson and Volpenhein, 1962, 1964; OECD, 2006; Sieber, 1974). The fraction of ester that is hydrolysed to the highly branched 3,5,5-trimethylhexanoic acid and alcohol is more likely to be absorbed from the GI-tract than the parent substance. However, the steric hindrance of the target substance and the hydrolysis products may theoretically reduce the absorption rate of these substances. With increasing chain-length the fatty acids are increasingly absorbed via the lymphatic route, and will be metabolised in the liver (Ramirez et al., 2001).

In conclusion, the results of a subacute oral toxicity study of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate suggests that oral absorption will occur. It is not clear if the target substance or the hydrolysis products resulting from the enzymatic hydrolysis in the GI-tract cause the observed effects.

Dermal

The dermal uptake of liquids and substances in solution is higher than that of dry particulates, since dry particulates need to dissolve into the surface moisture of the skin before uptake can begin. Molecular weights below 100 g/mol favour dermal uptake, while for those above 500 g/mol the molecule may be too large. Dermal uptake is anticipated to be low if the water solubility is < 1 mg/L; low to moderate if it is between 1-100 mg/L; and moderate to high if it is between 100-10000 mg/L. Log Pow values in the range of 1 to 4 (values between 2 and 3 are optimal) are favourable for dermal absorption, in particular if the water solubility is high. For substances with a log Pow above 4, the rate of penetration may be limited by the rate of transfer between the stratum corneum and the epidermis, but uptake into the stratum corneum will be high. Log Pow values above 6 reduce the uptake into the stratum corneum and decrease the rate of transfer from the stratum corneum to the epidermis, thus limiting dermal absorption (ECHA, 2014).

The target substance 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate has a molecular weight of 284.48 g/mol, which indicates a potential for dermal absorption. In contrast, the substance has very low water solubility and therefore a low dermal absorption potential might be assumed (ECHA, 2014). The log Pow is > 6, which means that the uptake into the stratum corneum is predicted to be slow and the rate of transfer between the stratum corneum and the epidermis will be slow (ECHA, 2014).

The dermal permeability coefficient (Kp) can be calculated from log Pow and molecular weight (MW) applying the following equation described in US EPA (2012), using the Epi Suite software:

log(Kp) = -2.80 + 0.66 log Pow – 0.0056 MW

The Kp was calculated to be 3.14 cm/h. Using the water solubility (0.00005 mg/cm³), the dermal flux is estimated to be ca. 0.157 µg/cm²/h, indicating a medium- to low dermal absorption potential.

If a substance shows skin irritating or corrosive properties, damage to the skin surface may enhance penetration. The available acute data on the target substance provides very mild or no indications for skin irritating effects in the rabbit (Masson, 1986). In a subacute (14-day) dermal toxicity study severe erythema and cutaneous lesions were observed at the application site from Day 4 (females) and 6 (males), while the microscopic examination showed skin necrosis, acanthosis and ulceration at the application site of 5/5 females and 5/5 males at the highest dose of 860 mg/kg bw/day (Manciaux, 2001). The animals in the highest dose group were sacrificed on Day 9 for animal welfare reasons. Furthermore, systemic effect similar to those observed in the subacute oral toxicity study were observed on body weight (gain), general health of the females and in the liver and kidney. This indicates that the target substance or its hydrolysis products were systemically available via the dermal route. If the substance has been identified as a skin sensitizer then some uptake must have occurred although it may only have been a small fraction of the applied dose (ECHA, 2014). In a human skin sensitisation study the target substance caused very mild skin irritation (4/98 subjects), but no skin sensitising effects in any subject (Harrison, 1997). No skin sensitising effects were noted in a skin sensitisation study performed in the guinea pig using a source substance (Clouzeau, 1991).

Based on the available information dermal absorption of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate will take place.

Inhalation

3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate is a liquid with low vapour pressure (0.0193 Pa at 20 °C), and therefore very low volatility. Therefore, under normal use and handling conditions, inhalation exposure and availability for respiratory absorption of the substance in the form of particles will depend on the aerodynamic particle size (ECHA, 2014). The substance may also be available for inhalatory absorption after inhalation of aerosols, if the substance is sprayed (e.g. as a formulated product). In humans, particles with aerodynamic diameters below 100 μm have the potential to be inhaled. Particles with aerodynamic diameters below 50 μm may reach the thoracic region and those below 15 μm the alveolar region of the respiratory tract. Particles deposited in the nasopharyngeal/thoracic region will mainly be cleared from the airways by the mucocilliary mechanism and swallowed.

It is not clear to which extent the ester and the hydrolysis products will be absorbed in the respiratory tract. For absorption of the hydrolysis products, enzymatic hydrolysis in the airways would be required, and the presence of esterases and lipases in the mucus lining fluid of the respiratory tract would be important. Due to the physiological function of enzymes in the GI-tract for nutrient absorption, esterase and lipase activity in the lung is expected to be lower in comparison to the gastrointestinal tract. Therefore, hydrolysis comparable to that in the gastrointestinal tract and subsequent absorption in the respiratory tract is considered to happen at a lower rate. The molecular weight, log Pow and water solubility indicate that the parent substance may be absorbed across the respiratory tract epithelium by micellar solubilisation to a certain extent. However, low water solubility (<0.05 mg/L) does restrict the diffusion/dissolving into the mucus lining before reaching the epithelium, and it is not clear which percentage of the inhaled aerosol could be absorbed as the ester.

An acute inhalation toxicity study was performed with the read-across (source) substance 2-ethylhexyl oleate (CAS 26399-02-0), in which rats were exposed nose-only to > 5.7 mg/L of an aerosol for 4 hours (Van Huygevoort, 2010). No mortality occurred and no toxicologically relevant effects were observed. Therefore, the target substance is not expected to be acutely toxic by the inhalation route, but no firm conclusion can be drawn on respiratory absorption. Furthermore, the effects observed in the oral and dermal repeated dose toxicity studies with the target substance indicate it may cause effects via the inhalation route as well.

Due to the limited information available a worst case approach is applied, and absorption via inhalation is assumed to be as high as via the oral route

Distribution and Accumulation

Distribution of a compound within the body depends on the physico-chemical properties of the substance; particularly the molecular weight, the lipophilic character and the water solubility. In general, the smaller the molecule, the wider the distribution. If the molecule is lipophilic, it is likely to distribute into cells and the intracellular concentration may be higher than extracellular concentration, particularly in fatty tissues (ECHA, 2014).

As discussed under oral absorption, 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate will undergo some enzymatic hydrolysis in the gastrointestinal tract prior to absorption. The fraction of ester absorbed unchanged will undergo enzymatic hydrolysis by ubiquitous esterases, primarily in the liver (Fukami and Yokoi, 2012). The distribution and accumulation of the hydrolysis products is considered the most relevant.

After being absorbed, linear and simple branched fatty acids are (re-)esterified along with other fatty acids into triglycerides and released in chylomicrons into the lymphatic system. This route of absorption and metabolism of a fatty acid was shown in an in vivo study performed by Sieber (1974). Twenty-four hours after intraduodenal administration of a single dose of [1-14C]-radiolabelled octadecanoic acid to rats, 52.5 ± 26% of the radiolabelled carbon was recovered in the lymph. A large majority (68 - 80%) of the recovered radioactive label was incorporated in triglycerides, 13 - 24% in phospholipids and 0.7 - 1% in cholesterol esters. No octadecanoic acid was recovered. Almost all the radioactivity recovered in the lymph was localized in the chylomicron fraction. Fatty acids of carbon chain length ≤ 12 may be transported directly to the liver via the portal vein as the free acid bound to albumin, instead of being re-esterified. This is supported by the Sieber study (1974), in which, following the same protocol as described above, administration of hexanoic acid lead to only 3.3% recovery from lymphatic fluid. Chylomicrons are transported in the lymph to the thoracic duct and subsequently to the venous system. On contact with the capillaries, enzymatic hydrolysis of chylomicron triacylglycerol fatty acids by lipoprotein lipase takes place. Most of the resulting fatty acids are taken up by adipose tissue and re-esterified into triglycerides for storage. Triacylglycerol fatty acids are also taken up by muscle and oxidized to derive energy or they are released into the systemic circulation and returned to the liver, where they are metabolised, stored or re-enter the circulation (IOM, 2005; Johnson, 1990; Johnson, 2001; Lehninger, 1993; NTP, 1994; Stryer, 1996; WHO, 2001). Methyl substituted fatty acids were show to be metabolised via beta- and omega-oxidation and can be expected to have a similar distribution to linear fatty acids (WHO, 1998). Highly branched fatty acids are expected to be widely distributed, although the distribution may be facilitated via other routes than the chylomicrons, due to the steric hindrance. Absorbed alcohols are mainly oxidised to the corresponding fatty acid and then follow the same metabolism as described above for fatty acids, to form triglycerides. The absorption and metabolism of a fatty alcohol was assessed in an in vivo study performed by Sieber (1974). Twenty-four hours after intraduodenal administration of a single dose of [1-14C]-radiolabelled octadecanol to rats, 56.6 ± 14% of the radiolabelled carbon was recovered in the lymph. More than half (52-73%) of the recovered radioactive label was incorporated in triglycerides, 6-13% in phospholipids, 2-3% in cholesterol esters and 4-10% in unmetabolised octadecanol. Almost all the radioactivity recovered in the lymph was localized in the chylomicron fraction. The results of administration of hexanol resulted in a recovery of 8.5% in the lymph (Sieber, 1974), indicating that alcohols with shorter-length carbon chains are hydrolysed to the corresponding fatty acid and transported directly to the liver via the portal vein as the free acid bound to albumin. Branched fatty alcohols, including the highly branched alcohols, are also converted into the corresponding fatty acids and distributed via the circulation.

Substances which are absorbed through the pulmonary alveolar membrane or through the skin enter the systemic circulation directly before transport to the liver where hydrolysis will generally take place.

Metabolism

The metabolism of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate initially occurs via a stepwise enzymatic hydrolysis of the ester resulting in the corresponding 3,5,5-trimethylhexanoic acid and 3,5,5-trimethylhexan-1-ol. The esterases catalysing the reaction are present in most tissues and organs, with particularly high concentrations in the GI-tract and the liver (Fukami and Yokoi, 2012).

The C9iso-alcohol will mainly be metabolised to the corresponding carboxylic acid via the aldehyde as a transient intermediate (Lehninger, 1993). The stepwise process starts with the oxidation of the alcohol by alcohol dehydrogenase to the corresponding aldehyde, where the rate of oxidation increases with increasing chain-length. Subsequently, the aldehyde is oxidised to carboxylic acid, in a reaction catalysed by aldehyde dehydrogenase. Both the alcohol and the aldehyde may also be conjugated with e.g. glutathione and excreted directly, bypassing further metabolism steps (WHO, 1999).

Fatty acids can be metabolised directly following absorption, following oxidation from an alcohol or following de-esterification of triglycerides. The beta-oxidation pathway for energy generation is the major metabolic pathway for linear and an important pathway for simple branched fatty acids. In this multi-step process, the fatty acids are at first esterified into acyl-CoA derivatives and subsequently transported into cells and mitochondria by specific transport systems. In the next step, the acyl-CoA derivatives are broken down into acetyl-CoA molecules by sequential removal of 2-carbon units from the aliphatic acyl-CoA molecule. Further oxidation via the citric acid cycle leads to the formation of H2O and CO2(Lehninger, 1993). Branched-chain acids can be metabolised via the same beta-oxidation pathway as linear, depending on the steric position of the branch, but at lower rates (WHO, 1999). For fatty acids with complex branching, such as the isononanoic acid, the alpha-oxidation pathway is the major metabolic pathway, as a methyl substituent at the beta-position blocks certain steps in the beta-oxidation (Mukherji, 2003). Generally, a single carbon unit is cleaved off the branched acid in an additional step before the removal of 2-carbon units continues. Depending on the degree of steric hindrance caused by the branching, the fatty acid can also be conjugated (by e.g. glucuronides, sulfates) to more polar products that are excreted in the urine.

The potential metabolites following enzymatic metabolism of the test substance were predicted using the QSAR OECD toolbox (OECD, 2014). This QSAR tool predicts which metabolites of the test substance may result from enzymatic activity in the liver and in the skin, and by intestinal bacteria in the gastrointestinal tract. Eight (8) hepatic metabolites and 6 dermal metabolites were predicted. Primarily, the ester bond is broken both in the liver and in the skin, after which the hydrolysis products may be metabolised further. The resulting liver and skin metabolites are the products of alpha-, beta- or omega-oxidation (= addition of hydroxyl group). In the case of omega-oxidation, it is followed by further oxidation to the aldehyde, which is then oxidised to the corresponding carboxylic acid. For a branched fatty acid, the alpha- and omega pathways are particularly relevant. The ester bond may also remain intact, in which case a hydroxyl group is added to, or substituted with, a methyl group. In general, the hydroxyl groups make the substances more water-soluble and susceptible to metabolism by phase II-enzymes. The metabolites formed in the skin are expected to enter the blood circulation and have the same fate as the hepatic metabolites. Twenty-four (24) metabolites were predicted to result from all kinds of microbiological metabolism. The high number includes many minor variations in the c-chain length and number of carbonyl- and hydroxyl groups; reflecting the many microbial enzymes identified. Not all of these reactions are expected to take place in the human GI-tract. The results of the OECD Toolbox simulation support the information on metabolism routes retrieved in the literature.

There is no indication that 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate is activated to reactive intermediates under the relevant test conditions. The experimental studies performed on genotoxicity (Ames test, gene mutation in mammalian cells in vitro, chromosome aberration assay in mammalian cells in vitro) using the target (3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate) and source substances were consistently negative, with and without metabolic activation (Buskens, 2010a,b; Sire, 2005; Verspeek-Rip, 2010a,b). The result of the skin sensitisation studies performed in humans using the target substance and performed in guinea pigs using a source substance, respectively, were likewise negative (Clouzeau, 1991; Harrison, 1997).

The accumulation of fatty acids takes place only if their intake exceeds the caloric requirements of the organism. The potential of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate, as well as the hydrolysis products 3,5,5 trimethylhexanoic acid and 3,5,5 -trimethylhexan-1-ol, to accumulate is considered to be low.

The QSAR Approach developped by the Laboratory of Mathematical Chemistry, Bourgas, Bulgaria, consist in simulating metabolism of aliphatic alkyl esters by TIMES, 4 combined elements were used :

•       OASIS TIMES (platform for simulating metabolism; with 3 metabolic simulators : in vitro rat liver S9; in vivo rat whole organisms and Skin metabolism).

•       Toolbox 4.2 for searching analogues of the target chemicals;

•       Documented metabolism from research publications and websites;

•       Expert evaluation of the simulated metabolism.

SMILES of target substances were used as input for the model, the were chosen based on the main and similar constituents:

One is the monoconstituent 3,5,5 trimethylhexyl 3,5,5 trimethylhexanoate and the others are, with the monoconstituent, the main compounds of the UVCB branched-nonyl 3,5,5 trimethylhexanoate.

3 Simulators were considered :

       In vitro rat liver S9 (one analogue);

       In vivo rat (three analogues);

       Skin metabolism (one analogue).

Documented metabolites are simulated correctly by all three (in vitro S9 mix, in vivo whole organism and skin) TIMES metabolic simulators. The correctly simulated ester hydrolysis pathway is a premise for adequate model predictions.

Same pattern is obtained by all three TIMES metabolic simulators:

a)       Ester hydrolysis is first and major transformation obtained with high probability (P > 0.9).

b)       Primary alcohol is documented and simulated. This product of ester hydrolysis is further oxidized via aldehyde intermediate to carboxylic acid.

Formation of Phase II metabolites is more pronounced in vivo and in skin than in in vitro S9 metabolism.

For the simulating metabolism of the target chemicals, following results were obtained:

A. No documented metabolism of the target chemicals has been found.

B. Structural analogues of the target chemicals having documented metabolism have been used to confirm the main metabolic transformation - ester hydrolysis.

C. No difference between simulated metabolism pattern of the target chemicals within each of the analyzed (in vivo, in vitro S9, skin) metabolic systems.

D. Commonality of simulated metabolism of the target chemicals with respect to all three metabolic systems is the first and major metabolic transformation, enzymatic ester hydrolysis, which occurs with very high probability (P > 0.9).

E. Metabolic pattern of the acid hydrolysis product (left branch in the maps) in all three metabolic systems could be summarized as follows:

a. Similar metabolism has been found in in vitro microsomes/S9 and skin, where the acid is directly involved in Phase II conjugation reactions.

b. Slightly different transformation has been found for in vivo metabolism, where the acid undergoes hydroxylation at the terminal (ω) carbon, ultimately leading to Phase II conjugates.

F. Metabolism of the alcohol hydrolysis product (right branch in the maps) in all three metabolic systems could be summarized as follows:

a. Similarity has been found for in vitro microsomes/S9 and in vivo metabolism. The only difference is an additional Phase II glucuronidation reaction in vivo which does not occur in vitro S9.  

b. Slightly different transformation has been found in skin metabolism due to the lower activity of Phase I oxidizing enzymes as compared to in vitro S9 and in vivo systems. Lower activity of these oxidizing enzymes in skin increases the relative importance of Phase II metabolic reactions.  

G. Kinetics of biotransformations are not provided in the current TIMES metabolic simulators reproducing metabolic transformation statically – within the time frame of the experimental conditions. Nevertheless, the different activity of enzymes in the three metabolic systems is not expected to affect significantly the rate of metabolic reactions.

H. According to our experts, no alerts associated with skin sensitization and genotoxicity could be expected in the target chemicals and the obtained metabolites.

I. All predictions could be considered as reliable given the fact that all target chemicals belong 100% to the parametric domain and above 75% to the structural domain.  

Excretion

In general, linear fatty acids and fatty acids with simple branching derived directly from the hydrolysis of the ester or from the oxidation of the corresponding alcohol, as well as the fatty acids, will be metabolised for energy generation or stored as lipid in adipose tissue or used for further physiological functions, like incorporation into cell membranes (Lehninger, 1993). Therefore, the fatty acid metabolites are not expected to be excreted to a significant degree via the urine or faeces but to be excreted via exhaled air as CO2 or stored as described above. Experimental data with ethyl oleate (CAS 111-62-6, ethyl ester of oleic acid (ethyl oleate)) support this principle. The absorption, distribution, and excretion of 14C-labelled ethyl oleate was studied in Sprague Dawley rats after a single, oral dose of 1.7 or 3.4 g/kg bw. At sacrifice (72 h post-dose), mesenteric fat was the tissue with the highest concentration of radioactivity. The other organs and tissues had very low concentrations of test material-derived radioactivity. The main route of excretion of radioactivity in the groups was via the expired air as CO2. 12 h after dosing, 40-70% of the administered dose was excreted in expired air (consistent with beta -oxidation of fatty acids). 7-20% of the radioactivity was eliminated via the faeces, and approximately 2% via the urine (Bookstaff et al., 2003).

However,3,5,5-trimethylhexanoic acid is unlikely to be used for energy generation and storage, since saturated aliphatic, branched-chain acids are most likely subjected to omega-oxidation due to steric hindrance by the methyl groups at uneven position, which results in the formation of various diols, hydroxyl acids, ketoacids or dicarbonic acids. In contrast to the products of beta-oxidation, these metabolites may be conjugated to glucuronides or sulphates, which subsequently can be excreted via urine or bile or cleaved in the gut with the possibility of reabsorption (entero-hepatic circulation) (WHO, 1998).

The alcohol component 3,5,5-trimethylhexan-1-ol may be subjected to oxidation the 3,5,5 trimethylhexanoic due to the expected steric hindrance, conjugated with e.g. glutathione to form a more water-soluble molecule and excreted via the urine, bypassing further metabolism steps (WHO, 1999). The fraction of 3,5,5-trimethylhexyl 3,5,5-trimethylhexanoate that is not absorbed in the GI-tract, will be excreted via the faeces.

References

Bookstaff et al. (2003). The safety of the use of ethyl oleate in food is supported by metabolism data in rats and clinical safety data in humans. Regul Toxicol Pharm 37: 133-148.

ECHA (2014). Guidance on information requirements and chemical safety assessment, Chapter R.7c: Endpoint specific guidance.

EPA (2012). Dermwin v2.01, Estimation Programs Interface Suite™ for Microsoft® Windows, v 4.10. United States Environmental Protection Agency, Washington, DC, USA. Downloaded from http://www.epa.gov/opptintr/exposure/pubs/episuitedl.htm. Calculation performed 26 November 2012.

Fukami, T. and Yokoi, T. (2012). The Emerging Role of Human Esterases. Drug Metab Pharmacokinet 27(5): 466-477

Greenberger, N.J., Rodgers, J.B. and Isselbacher, K.J. (1966). Absorption of medium and long chain triglycerides: factors influencing their hydrolysis and transport. J Clin Invest. 45(2):217-27.

IOM (2005). Dietary Reference Intakes for Energy, Carbohydrate, Fiber, Fat, Fatty Acids, Cholesterol, Protein, and Amino Acids (Macronutrients). Institute of the National Academies. The National Academies Press. http://www.nap.edu/openbook.php?record_id=10490&page=R1

Johnson, R.C. et al. (1990). Medium-chain-triglyceride lipid emulsion: metabolism and tissue distribution. Am J Clin Nutr 52(3):502-8.

Johnson W. Jr; Cosmetic Ingredient Review Expert Panel. (2001). Final report on the safety assessment of trilaurin, triarachidin, tribehenin, tricaprin, tricaprylin, trierucin, triheptanoin, triheptylundecanoin, triisononanoin, triisopalmitin, triisostearin, trilinolein, trimyristin, trioctanoin, triolein, tripalmitin, tripalmitolein, triricinolein, tristearin, triundecanoin, glyceryl triacetyl hydroxystearate, glyceryl triacetyl ricinoleate, and glyceryl stearate diacetate. Int J Toxicol. 2001;20 Suppl 4:61-94.

Lehninger, A.L., Nelson, D.L. and Cox, M.M. (1993). Principles of Biochemistry. Second Edition. Worth Publishers, Inc., New York, USA. ISBN 0-87901-500-4.

Mattson, F.H. and Volpenhein, R.A. (1964). The digestion and absorption of triglycerides. J Biol Chem. 239:2772-7.

Mattson, F.H. and Volpenhein, R.A. (1969). Relative rates of hydrolysis by rat pancreatic lipase of esters of C2 - C18 fatty acids with C1 – C18 primary n-alcohols. J Lipid Res Vol(10): 271 – 276

Mattson, F.H. and Volpenhein, R.A. (1972). Hydrolysis of fully esterified alcohols containing from one to eight hydroxyl groups by the lipolytic enzymes of the rat pancreatic juice. Journal of Lipid Research 13: 325-328

Mukherji M. et al. (2003). The chemical biology of branched-chain lipid metabolism. Progress in Lipid Research 42: 359-376.

National Toxicology Program (NTP) (1994) Comparative toxicology studies of Corn Oil, Safflower Oil, and Tricaprylin (CAS Nos. 8001-30-7, 8001-23-8, and 538-23-8) in Male F344/N Rats as vehicles for gavage. http://ntp.niehs.nih.gov/ntp/htdocs/LT_rpts/tr426.pdf (2011-12-19). Report No.: C62215. Owner company: U.S. Department of Health and Human Services, Public Health Services, National Institutes of Health.

OECD (2006). Long Chain Alcohols. SIDS Initial Assessment Report For SIAM 22. Paris, France, 18-21 April 2006. TOME 1: SIAR. http://webnet.oecd.org/Hpv/UI/SIDS_Details.aspx?id=7A14361C-4676-4339-A915-2CFD51F12483

OECD (2014). (Q)SAR Toolbox v3.3. Developed by Laboratory of Mathematical Chemistry, Bulgaria for the Organisation for Economic Co-operation and Development (OECD). Prediction performed 11 May 2015.http://toolbox.oasis-lmc.org/?section=overview

Ramirez, M. et al. (2001). Absorption and distribution of dietary fatty acids from different sources. Early Hum Development 65 Suppl.: S95 – S101

Sieber, S.M., Cohn, V.H., and Wynn, W.T. (1974). The entry of foreign compounds into the thoracic

duct lymph of the rat.Xenobiotica 4(5), 265.

Stryer, L. (1996). Biochemie. 4. Auflage. Heidelberg Berlin Oxford: Spektrum Akademischer Verlag.