3-dodecyl-1-(2,2,6,6-tetramethyl-4-piperidyl)pyrrolidine-2,5-dione

Bioaccumulation: aquatic / sediment

Link to relevant study record(s)

Reference

Endpoint:

bioaccumulation in aquatic species: fish

Type of information:

(Q)SAR

Adequacy of study:

key study

Study period:

November 2017

Reliability:

1 (reliable without restriction)

Rationale for reliability incl. deficiencies:

results derived from a valid (Q)SAR model and falling into its applicability domain, with adequate and reliable documentation / justification

Remarks:

Resutls from two relevant supporting models corroborate key model result.

Justification for type of information:

See attached QPRF and QMRF section "Attached justification".

Guideline:

other: ECHA Guidance on Information Requirements and Chemical Safety Assessment. Chapter R.6: QSARs and grouping of chemicals

Version / remarks:

2008

Guideline:

other: ECHA Practical guide How to use and report (Q)SARs

Version / remarks:

2016

Principles of method if other than guideline:

Model selection for BCF calculation was based on a recent publication by Gissi et al. (2015), evaluating and comparing several relevant QSAR models for BCF prediction under REACH based on a validated experimental data set of 851 compounds containing a subset of ionisable compounds (at least 131). For model selection it was considered in addition that the submission substance is ionisable (alkaline functionality). As a consequence, the following models were selected:
Key model: US EPA T.E.S.T v. 4.2.1, Consensus Method (see attached QMRF and QPRF);
Supporting model (1): US EPA EPI SuiteTM module BCFBAF (Meylan et al.)
Supporting model (2): log Kow based equations - Connell equation (recalculated from Connell and Hawker, 1988) recommended for compounds with log Kow values above 6.
For full details on a) rational for model selection; b) details on selected models with regard to comparative statistical evaluation by Gissi et al. (2015) and c) full literature references, see IUCLID section “Any other information on materials and methods incl. tables”!

GLP compliance:

no

Specific details on test material used for the study:

• CAS number: 79720-19-7
• EC number: 279-242-6
• IUPAC name / common name: 3-dodecyl-1-(2,2,6,6-tetramethylpiperidin-4-yl)pyrrolidine-2,5-dionet (IUPAC-name) / Hostavin 3055
• Structural formula: C25-H46-N2-O2.

Structure codes:
o Canonical SMILES neutral from: CCCCCCCCCCCCC1CC(=O)N(C1=O)C2CC(NC(C2)(C)C)(C)C (used for all model predictions)
o SMILES protonated form: CCCCCCCCCCCCC1CC(=O)N(C1=O)C2CC([NH2+]C(C2)(C)C)(C)C (used for BCFBAF Meylan et al. model, only)

• Partition coefficient n-octanol water of the neutral molecule, log Kow: 7.1 at pH 11 and 25°C (OECD 117; Siemens, 2011)
• Dissociation constant: The submission substance contains a cyclic amine residue (secondary amine) which is of alkaline functionality. The following pKa values were calculated with scientifically well-established models:
o QSAR (ACDLabs ): 9.5 (SD ± 0.4)
o QSAR (ChemAxon ): 10.39
o pKa used: arithmetic mean value, i.e. pKa 9.945
References:
- ACD ToxSuite 2.95.1; Danish (Q)SAR Database, Division of Diet, Disease Prevention and Toxicology, National Food Institute, Technical University of Denmark, http://qsar.food.dtu.dk
- ChemAxon, Calculator Plugin 17.14.0 (2017-07-03), https://www.chemaxon.com/

Details on sampling:

Not applicabale

Details on preparation of test solutions, spiked fish food or sediment:

Not applicabale

Test organisms (species):

other: mostly lower trophic level fresh water fish species

Details on test organisms:

For the key model US EPA T.E.S.T. v. 4.2.1. the bioconcentration factor data set was compiled by researchers at the Mario Negri Istituto Di Ricerche Farmacologiche and based on Dimitrov et al. (2005). Fish experimental BCF values (two databases combined; experimental data obtained according to OECD 305 protocol; fish species: Cyprinus Carpio and salmonids) were obtained according to official protocols. Furthermore, as explained in the Literature (Zhao et al. 2008), all structures were checked one-by-one within the EC funded project CAESAR, by at least two scientists. Further high quality data were added from the Arnot & Gobas database as well as the EURAS Gold Standard Database.
The following relevant references are given in the User’s Guide on T.E.S.T. v. 4.2.:
- Dimitrov, S.; Dimitrova, N.; Parkerton, T.; Combers, M.; Bonnell, M.; Mekenyan, O. Base-line model for identifying the bioaccumulation potential of chemicals. SAR QSAR Environ. Res. 2005, 16, 531-554
- Arnot, J. A.; Gobas, F. A. P. C. A review of bioconcentration factor (BCF) and bioaccumulation factor (BAF) assessments for organic chemicals in aquatic organisms. Environ. Rev. 2006, 14, 257-297.
- EURAS. Establishing a bioconcentration factor (BCF) Gold Standard Database. http://www.euras.be/eng/project.asp?ProjectId=92 (accessed 5/20/09).
- Zhao, C. B., E.; Chana, A.; Roncaglioni, A.; Benfenati, E. A new hybrid system of QSAR models for predicting bioconcentration factors (BCF). Chemosphere 2008, 73, 1701-1707.

Route of exposure:

aqueous

Justification for method:

aqueous exposure method used for following reason: Recommended according to ECHA guidance R.7c (2017) if experimentally feasible

Test type:

other: Predominantly (94%) flow through according to Arnot and Gobas (2006) BCF database considering acceptable (category 1) data, only

Water / sediment media type:

other: Predominantly (80%) fresh water according to Arnot and Gobas (2006) BCF database considering acceptable (category 1) data, only

Hardness:

Not applicable

Test temperature:

Not applicable

pH:

Not applicable

Dissolved oxygen:

Not applicable

TOC:

Not applicable

Salinity:

Not applicable

Conductivity:

Not applicable

Details on test conditions:

Not applicable

Nominal and measured concentrations:

Not applicable

Reference substance (positive control):

not required

Details on estimation of bioconcentration:

For key model calculation via US EPA T.E.S.T. v.4.2.1. BCF Consensus Method, see attached QMRF and QPRF!
For supporting estimations via US EPA BCFBAF program v. 3.01 as well as BCF-calculation via the modified Connell equation, see below section "Any other information on materials and methods including tables"!

Lipid content:

>= 5 - <= 6 %

Remarks on result:

other: generally assumed lipid content for lower trophic fish

Key result

Type:

BCF

Value:

>= 26 - <= 447 L/kg

Basis:

whole body w.w.

Calculation basis:

other: QSAR

Remarks:

Actual result of BCF Consensus Method: 23 L/kg (range of 5 single models: 5.78-93.54 L/kg); upper limit externally calculated taking into account deviations of calculated from experimental values for similar compounds.

Remarks on result:

other: Key model: US EPA T.E.S.T. v. 4.2.1. BCF Consensus Method

Type:

BCF

Value:

<= 370 L/kg

Basis:

whole body w.w.

Remarks:

BCF at environmental pH 4 to 9 (value conservatively calculated for pH 9)

Calculation basis:

other: QSAR

Remarks:

BCF derived from weighing BCF results for neutral and ionized molecule according to the degree of ionization at pH 9

Remarks on result:

other: Supporting model: BCFBAF model v. 3.01

Type:

BCF

Value:

<= 211 L/kg

Basis:

whole body w.w.

Remarks:

BCF at environmental pH 4 to 9 (value conservatively calculated for pH 9)

Calculation basis:

other: QSAR

Remarks:

BCF calculated based on SMILES for neutral molecule and log Dow at pH 9 (log Dow 6.11, calculated from (1) pKa: 9.945; (2) log Kow neutral form: 7.1)

Remarks on result:

other: Supporting model: BCFBAF model v. 3.01

Type:

BCF

Value:

< 543 L/kg

Basis:

whole body w.w.

Remarks:

BCF calculated for neutral fraction , only (0.1% at pH 7; 10% at pH 9)

Calculation basis:

other: QSAR

Remarks:

BCF calculated for neutral molecule based on experimentally determined log Kow 7.1 for the neutral form (pH 11)

Remarks on result:

other: Supporting model: Modified Connell equation

Details on kinetic parameters:

Not applicable

Metabolites:

Not applicable

Results with reference substance (positive control):

Not applicable

Details on results:

KEY MODEL RESULT: US EPA T.E.S.T. (version 4.2.1) BCF-Calculation
The T.E.S.T. Consensus Method estimates BCF based on the results of 5 sub-models (single models), taking into account the applicability domain of each method. At least predictions of 2 sub-models are required, otherwise the predicted value is deemed unreliable and not used. Sub-model results are integrated in the Consensus Method for BCF by averaging. Details on US EPA T.E.S.T. BCF model is given in the available QSAR Model Reporting Format (QMRF_BCF_TEST_v4-2-1) attached to this dossier.
The following results per sub-model were obtained and used within the T.E.S.T. Consensus method for BCF:
• Hierarchical clustering: BCF 72.68 L/kg wet weight
• Single model: BCF 42.41 L/kg wet weight
• Group contribution: BCF 6.64 L/kg wet weight
• FDA: BCF 5.78 L/kg wet weight
• Nearest neighbour: BCF 93.54 L/kg wet weight

The BCF based on the estimate of the T.E.S.T. Consensus Method for the submission substance is given as a range to account for some uncertainty in the estimate apparent from analysis of calculated and experimental values for similar compounds: taking as the lower value the actual estimate according to the Consensus Method (BCF 26 L/kg) and as the upper value the sum of (a) the highest absolute deviation between experimental and calculated value from analysis of similar substances (1.24 log units) and (b) the actual estimate (1.41 log units), resulting in 2.65 log units, i.e. BCF 447 L/kg wet weight.
This must be regarded as a conservative approach taking into account the low deviations observed for the 4 similar amines contained in test (2) and training (2) set. Estimates for these amines are of special relevance because of their ionizing properties characteristic also for the submission substance.

FINAL RESULT relevant for PBT assessment and exposure and risk assessment:
BCF (US EPA T.E.S.T version 4.2.1): 26 to 447 L/kg whole body wet weight (range of 5 single models: 5.78-93.54 L/kg)
For further information please see corresponding methodological background on the model QMRF as well as QPRF attached to this dossier.

RESULTS SUPPORTING MODELS BCFBAF v3.01 (Meylan et al.) AND MODIFIED CONNELL EQUATION
See section "Any other information on results incl. tables".

 Supporting Model 1: BCFBAF calculation – results

Interim result (1): BCF for the neutral form of Hostavin 3055

The output of BCFBAF was as follows:

·      Log Kow used by BCF estimates: 7.10 (user entered)

·      Equation used to make BCF estimate:

·      Log BCF = -0.49 log Kow + 7.554 + Correction

·      Correction(s):                                    Value

·      Alkyl chains (8+ -CH2- groups)            -0.596

·      Estimated Log BCF = 3.479 (BCF = 3010 L/kg wet-wt)

 Interim Result (2): BCF for the protonated form of Hostavin 3055 (based on SMILES for protonated form)

·      Log Kow (estimated) : 5.35

·      Log Kow (experimental): not available from database

·      Log Kow used by BCF estimates: 5.35

·      Equation used to make BCF estimate:

·      Log BCF = 1.85 (Ionic; 11 or more -CH2- groups)

·      Estimated Log BCF = 1.850 (BCF = 70.79 L/kg wet-wt)

Result from estimation based on the neutral form of Hostavin 3055 and log Dow

Instead of the log Kow as optional user input, a value for log Dow is used in this estimate. Dow is defined as the apparent Kow for weak electrolytes. According to ECETOC (2014), log Dow can be estimated via the following relationship between log Kow and pKa (bases):

log Dow = log Kow – log[1 + 10^(pKa - pH)]                                           (1)

Based on the experimental log Kow of 7.1, the estimated pKa of 9.945 and an assumed environmental pH of 9.0, the following result for log Dow is obtained:

log Dow = 6.11 (rounded)

A slightly higher but rounded to 2 decimals equal value is obtained via the relationship Dow = f_n x Kow_neutral + f_d x Kow_ion (see equations 2 and 3 below for definition of f_n and f_d), which does not neglect partitioning of the ionized fraction like equation (1) above. Kow_ion can be estimated via the empirical relationship Kow_ion = Kow_neutral – 3.5 (Trapp and Horobin, 2005).

The output of BCFBAF was as follows:

·      Log Kow used by BCF estimates: 6.11 (user entered)

·      Equation used to make BCF estimate:

·      Log BCF = 0.6598 log Kow - 0.333 + Correction

·      Correction(s):                                   Value

·      Alkyl chains (8+ -CH2- groups)       -1.374

·      Estimated Log BCF = 2.324 (BCF = 210.9 L/kg wet-wt)

BCF derived from weighing interim results (1) and (2) according to the degree of ionization at environmentally relevant pH

As outlined under "Specific details on test material", the estimated pKa (basic) for the submission substance is 9.945.

The environmentally relevant pH range as given in ECHA guidance R.7a (ECHA, 2016) is pH 4 to 9. The charged fraction for a base is highest at acidic and lowest at alkaline pH. As such, to cover the whole environmental pH range, pH 9 is taken as the relevant pH as a conservative approach.

The neutral substance fraction (f_n) present at a given pH can be calculated according to the Henderson-Hasselbach equation as follows:

 

f_n= 1/(1+10^(pKa-pH)) = 0.1019                                                                     (2),

and the dissociated (protonated) fraction (f_d) is calculated as

f_d= 1-f_n= 0.8981                                                                                           (3)

 

The environmentally relevant BCF (BCF_env) is calculated based on the BCFBAF results for the neutral (interim result1) and the ionized molecule (interim result 2) considering these degrees of dissociation as follows:

BCF_env= BCF_neutral x f_n + BCF_ionized x f_d = 370.39 L/kg wet weight

log BCF_env= 2.57

Conclusion on results supporting model BCFBAF

The Meylan model contained within BCFBAF v. 3.01 is recommended according to ECHA guidance document R.7c (ECHA, 2017), Appendix R.7.10-3 for ionized substances. While the ionized (dissociated) fraction of the submission substance is high (ca. 90% at pH 9), at least at this upper limit for the environmentally relevant pH range according to ECHA guidance R.7a (ECHA, 2016) there is still a relevant neutral fraction to be considered (while at pH 7, 99.9% would be ionized). Therefore, it seemed to be most realistic to follow two approaches:

1)   To weigh results for BCF from the Meylan et al. model a) for the neutral molecule and b) the ionized form according to their fractions calculated from pKa and pH 9 as a conservative limit value for environmental pH.

2)   To estimate BCF based on the SMILES code for the neutral molecule and the apparent distribution coefficient n-octanol / water (log Dow) estimated from experimental log Kow, pKa and pH 9 as a conservative limit value for environmental pH.

Approach 1) resulted in a BCF of 370.39 L/kg wet-weigth, approach 2) in a somewhat lower value of 210.9 L/kg wet-weight. Considering the difference in these two approaches, these values agree well. Both estimates are within the range estimated with the key model T.E.S.T Consensus method (BCF: 26 – 447 L/kg wet-weight). Overall, the results of the supporting model therefore clearly support the result of the key model. It must further be considered that these values were determined for the upper limit of the environmentally relevant pH range 4-9. Correspondingly, BCF values will be lower at lower pH considering the higher dissociation degree.

References:

ECETOC, European Centre for Ecotoxicology and Toxicology of Chemicals (2014)

Technical Report No. 123. Environmental risk assessment of ionisable compounds

Brussels, Belgium

Trapp, S.; Horobin, R.W. (2005)

A predictive model for the selective accumulation of chemicals in tumor cells

European Biophysics Journal, 34, 959-966

Validity criteria fulfilled:

yes

Remarks:

with regard to OECD principles for (Q)SAR validation as well as ECHA guidance document R.6 (2008)

Conclusions:

The results of the key model (US EPA T.E.S.T. v. 4.2.1 BCF Consensus Method) and the supporting models (supporting model 1: BCFBAF v3.01; supporting model 2: Modified Connell Equation) conclusively demonstrate that:
• The submission substance is not bioaccumulative (BCF clearly below 2000 L/kg);
• The upper limit of the estimated BCF (447 L/kg) based on T.E.S.T. Consensus Method results (key model) will be sufficiently conservative as the relevant figure for chemical risk assessment.

Executive summary:

To conclude on a bioconcentration potential (BCF), three different models have been applied:

Key model: US EPA T.E.S.T. version 4.2.1 Consensus Method on BCF

The estimate for BCF according to T.E.S.T. is adequate for PBT/vPvB assessment (first tier) and risk assessment (first tier) for the submission substance because: (a) the model fulfils the OECD principles for QSAR models (algorithm and experimental data used for model building and validation are freely available) and (b) the submission substance is in the applicability domain of the model and evaluation of sufficiently similar compounds with experimental data increases the confidence in the estimated value for the submission substance according to ECHA guidance chapter R.6 (ECHA, 2008), section R.6.1.5.3.

All single model results integrated within the Consensus Method are far below the regulatory threshold for B (BCF 2000):

·      Hierarchical clustering:           BCF 72.68 L/kg wet weight

·      Single model:                         BCF 42.41 L/kg wet weight

·      Group contribution:                BCF 6.64 L/kg wet weight

·       FDA:                                    BCF 5.78 L/kg wet weight

·      Nearest neighbour:                 BCF 93.54 L/kg wet weight

·   Consensus method:             BCF 26 L/kg wet weight (average over 5 single models); log BCF= 1.41

Based on available data for similar compounds, some uncertainty in the estimate should be accounted for (external to the calculation routine of the model):

Considering the highest absolute deviation between experimental and calculated value (Consensus Model) for the reported similar compounds of test and training set (-1.24 log units, i.e. underestimation of BCF), and adding this to the predicted value for the submission substance (1.24 + 1.41) yields an upper threshold log BCF value of 2.65, corresponding to a BCF of 447 L/kg (rounded). This must be regarded as a conservative approach taking into account the low deviations observed for the 4 similar amines contained in test (2) and training (2) set. Estimates for these amines are of special relevance because of their ionising properties that are characteristic also for the submission substance.

Result Key model:

BCF (US EPA T.E.S.T version 4.2.1): 26 to 447 L/kg whole body wet weight (range of 5 single models: 5.78-93.54 L/kg)

Supporting model results corroborate the result obtained from the key model:

Supporting model 1:

Meylan et al. (1999) BCF model included in BCFBAF v. 3.01 (part of EPI Suite TM v4.1; EPA, 2011) – result:

(1) BCF≤211 L/kg wet weight (SMILES for neutral molecule and log Dow at pH 9);

(2) BCF≤370 L/kg at pH 4-9 (BCF derived from weighing BCF results for neutral and ionized molecule according to the degree of ionization at pH 9);

Both estimates are within the range estimated by the T.E.S.T Consensus method (BCF: 26 – 447 L/kg wet-weight). The results from the Meylan et al. (1999) BCF model therefore clearly support the result of the key model.

 

Supporting model 2:

BCF estimated according to the modified Connell equation for the neutral molecule, based on experimentally determined log Kow for non-dissociated species– result:

BCF (neutral molecule) 543 L/kg wet weight.

According to the result from the modified Connell equation for the neutral form, the BCF accounting for predominance of the dissociated from (99.9% at pH 7) must be <<543 L/kg. As such, this is a strong support for the key model result.

 

Taken together, it is safe to conclude that

·      The submission substance is not bioaccumulative (BCF clearly below 2000 L/kg);

·      Taking the upper limit of the estimated BCF (447 L/kg) as the relevant figure for chemical risk assessment will be sufficiently conservative.

Description of key information

BCF fish (US EPA T.E.S.T version 4.2.1): 26 to 447 L/kg whole body weight

Key value for chemical safety assessment

BCF (aquatic species):

447 L/kg ww

Additional information

To conclude on a bioconcentration potential (BCF), three different models have been applied:

Key model: US EPA T.E.S.T. version 4.2.1 Consensus Method on BCF

The estimate for BCF according to T.E.S.T. is adequate for PBT/vPvB assessment (first tier) and risk assessment (first tier) for the submission substance because: (a) the model fulfils the

OECD principles for QSAR models (algorithm and experimental data used for model building and validation are freely available; see QMRF attached to registration dossier) and (b) the submission substance is in the applicability domain of the model and evaluation of sufficiently similar compounds with experimental data increases the confidence in the estimated value for the submission substance according to ECHA guidance chapter R.6 (ECHA, 2008), section R.6.1.5.3 (see QPRF attached to registration dossier).

All single model results integrated within the Consensus Method are far below the regulatory threshold for B (BCF 2000):

·      Hierarchical clustering:           BCF 72.68 L/kg wet weight

·      Single model:                         BCF 42.41 L/kg wet weight

·      Group contribution:                BCF 6.64 L/kg wet weight

·       FDA:                                    BCF 5.78 L/kg wet weight

·      Nearest neighbour:                 BCF 93.54 L/kg wet weight

·      Consensus method:             BCF 26 L/kg wet weight (average over 5 single models); log BCF= 1.41

Based on available data for similar compounds, some uncertainty in the estimate should be accounted for (external to the calculation routine of the model):

Considering the highest absolute deviation between experimental and calculated value (Consensus Model) for the reported similar compounds of test and training set (-1.24 log units, i.e.

underestimation of BCF), and adding this to the predicted value for the submission substance (1.24 + 1.41) yields an upper threshold log BCF value of 2.65, corresponding to a BCF of 447

L/kg (rounded). This must be regarded as a conservative approach taking into account the low deviations observed for the 4 similar amines contained in test (2) and training (2) set.

Estimates for these amines are of special relevance because of their ionising properties that are characteristic also for the submission substance.

Result Key model:

BCF (US EPA T.E.S.T version 4.2.1): 26 to 447 L/kg whole body wet weight (range of 5 single models: 5.78-93.54 L/kg)

Supporting model results corroborate the result obtained from the key model:

Supporting model 1:

Meylan et al. (1999) BCF model included in BCFBAF v. 3.01 (part of EPI Suite TM v4.1; EPA, 2011) – results from two separate approaches for considering ionisability of the substance:

(1) BCF≤211 L/kg wet weight (SMILES for neutral molecule and log Dow at pH 9);

(2) BCF≤370 L/kg at pH 4-9 (BCF derived from weighing BCF results for neutral and ionized molecule according to the degree of ionization at pH 9);

Both estimates are within the range estimated by the T.E.S.T Consensus method (BCF: 26 – 447 L/kg wet-weight). The results from the Meylan et al. (1999) BCF model therefore clearly support

the result of the key model.

 

Supporting model 2:

BCF estimated according to the modified Connell equation for the neutral molecule, based on experimentally determined log Kow for non-dissociated species– result:

BCF (neutral molecule) = 543 L/kg wet weight.

According to the result from the modified Connell equation for the neutral form, the BCF accounting for predominance of the dissociated from (99.9% at pH 7) must be <<543 L/kg. As such,

this is a strong support for the key model result.

 

Taken together, it is safe to conclude that

·      The submission substance is not bioaccumulative (BCF clearly below 2000 L/kg);

·      Taking the upper limit of the estimated BCF (447 L/kg) as the relevant figure for chemical risk assessment will be sufficiently conservative.