Application of AOPs to the development of Defined Approaches (DA) and Integrated Approaches to Testing and Assessment (IATA) Gavin Maxwell, Unilever, SEAC, Bedford, UK +44-1234-264-888; [email protected] ; tt21c.org
Application of AOPs to the development of Defined Approaches (DA) and Integrated Approaches to
Testing and Assessment (IATA)
Gavin Maxwell, Unilever, SEAC, Bedford, UK
+44-1234-264-888; [email protected]; tt21c.org
• ACD – Allergic Contact Dermatitis
• ADME – Absorption Distribution Metabolism Excretion
• DA – Defined Approach
• DC – Dendritic Cell
• DNCB – 2,4-Dinitrochlorobenzene
• IATA – Integrated Approaches to Testing and Assessment
• DPRA – Direct Peptide Reactivity Assay
• GARD – Genomic Allergen Rapid Detection
• GPMT – Guinea Pig Maximization Test
• hCLAT – human Cell Line Activation Test
Abbreviations• HRIPT – human repeat insult patch test
• hTCPA – human T Cell Proliferation Assay
• LLNA – mouse Local Lymph Node Assay
• MHC – Major Histocompatibility Complex
• NAM – New Approach Methodologies
• Q(SAR) – Quantitative Structure Activity Relationship
• TIMES-SS – TImes MEtabolism Simulator platform for predicting Skin Sensitization
• TK – Toxicokinetic
• TD – Toxicodynamic
• TcR – T cell receptor
Outline
• Skin Sensitisation AOP, NAMs & DAs
• DA example : ‘qAOP’ TKTD model of skin sensitization
• DA example 2: Skin Allergy Risk Assessment (SARA) model
• IATA example: NexGen Skin Allergy Risk Assessment
• Learnings and insights
• Questions?
Why Skin Sensitisation?• first human health AOP to be defined (2012)
• useful case study for considering how OECD has developed and applied Integrated Approaches to Testing and Assessment (IATA) (2014-16) and Defined Approach (DA) (2017-present) concepts
• lots of published examples demonstrating the value of the AOP for driving the application of Novel Approach Methodologies (NAMs)
Covalent Protein Binding leading to Skin Sensitisation AOP: https://aopwiki.org/aops/40
Key Event 1 (KE1) KE2 KE3 KE4 Adverse Outcome (AO)
Predictive Chemistry
For example:• DEREK-NEXUS• OECD QSAR Toolbox• TIMES• ToxTree
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
Protein Reactivity*
OECD TG 442C
Includes:• ADRA• DPRA
* Other non-OECD test guideline methods also available
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
Keratinocyte Activation*
OECD TG 442D ARE Nrf2 luciferase test
Includes:• KeratinoSens™• LuSens
* Other non-OECD test guideline methods also available; e.g. SENS-IS®, SenCeeTox®
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
Dendritic Cell Activation*
OECD TG 442E
Includes:• h-CLAT• IL-8 Luc Assay• U-Sens™
* Other non-OECD test guideline methods also available; e.g. GARD™
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
T Cell Proliferation
For Example:• Artificial Lymph Node
Tissue model• Human T cell
proliferation assays (hTCPA)
Skin Sens NAMs (New Approach Methodologies)
in silico NAM
in chemico/vitro NAM
in vivo evidence
Skin Sensitisation
OECD TG 429: mouse local lymph node assay (LLNA)
& variants TG442A & 442B
OECD TG 406: Buehler & Guinea Pig Maximisation
Test (GPMT)
Human evidence e.g. Human Repeat Insult
Patch Test (HRIPT)
Integrated Approach to Testing and Assessment (IATA)• are pragmatic, science-based approaches for chemical hazard characterisation
that rely on an integrated analysis of existing information coupled with the generation of new information using testing strategies
• follow an iterative approach to answer a defined question in a specific regulatory context, taking into account the acceptable level of uncertainty associated with the decision context
• range from more flexible, non-formalised judgment based approaches (e.g. grouping and read-across) to more structured, prescriptive, rule based approaches [e.g. integrated testing strategy (ITS)]
• can include a combination of methods and can be informed by integrating results from one or many methodological approaches
www.oecd.org/chemicalsafety/risk-assessment/iata-integrated-approaches-to-testing-and-assessment.htm
IATA general workflow
Problem Formulation
Gather existing information
Weight of Evidence Assessment: adequate information for decision-making?
Generate additional information
DecisionAOP Yes
No Yes
OECD GD No. 260: Using AOPs to develop novel testing and assessment approaches
www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=ENV/JM/MONO(2016)67/&doclanguage=en
Problem Formulation
Gather existing information
Weight of Evidence Assessment: adequate information for decision-making?
Generate additional information
DecisionAOP Yes
No Yes
Organise and structure information using AOP as a frame
Use AOP to identify and/or develop targeted testing, testing strategy or assay development
Defined Approach (DA)• consists of a fixed data interpretation procedure (DIP) used to
interpret data generated with a defined set of information sources, that can either be used alone or together with other information sources, to satisfy a regulatory need
Relevant OECD guidance documents :
• No. 255: Reporting of Defined Approaches to be used within Integrated Approaches to Testing and Assessment: www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2016)28&doclanguage=en
• No. 256: Reporting of Defined Approaches and Individual Information Sources to be Used within Integrated Approaches to Testing and Assessment (IATA) for Skin Sensitisation www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2016)29&doclanguage=en ;
Annex 1: www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2016)29/ann1&doclanguage=en ; Annex 2: www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2016)29/ann2&doclanguage=en
Case Study submitter
SkinBioavail.
Phys-ChemInfo.
(Q)SAR Protein Binding
Keratinocyte activation
DCactivation
T cell proliferation
AdverseEvent
Other
BASF DPRA - OECDTG 442C
KeratinoSens - OECD TG 442D, LuSens
h-CLAT, m-MUSST
Dupont Yes OECD TG 442C, Glutathione
depletion assay
OECD TG 442D h-CLAT, U937 LLNA - OECD TG429
GPMT - OECD TG 406
e.g. Skin irritation,
Ames
EC-JRC TIMES-SS, Dragon
Givaudan Yes TIMES-SS Cor1C420-assay OECD TG 442D
ICCVAM Yes OECD Toolbox h-CLAT
Kao Corp. OECD TG 442C h-CLAT
Kao Corp. DEREK Nexus OECD TG 442C h-CLAT
L’Oréal Yes Various OECD TG 442C OECD TG 442D, ARE-Nrf2 assay
U-Sens, PGE2 assay
P&G Kasting Skin PBPKmodel
Yes TIMES-SS OECD TG 442C OECD TG 442D h-CLAT, U937 LLNA - TG 429
RIVM Various OECD TG 442C OECD TG 442D, HaCaTgene signature
h-CLAT
Shiseido Yes SH test AREc32 assay h-CLAT
Unilever modified OECD TG 428
modified OECD TG 428
TG No. 256 DA case studies: sensitiser potency
1. Bauch C, Kolle SN, Ramirez T, Eltze T, Fabian E, Mehling A, Teubner W, van Ravenzwaay B, Landsiedel R. Putting the partstogether: combining in vitro methods to test for skin sensitizing potentials. Regul Toxicol Pharmacol. 2012. 63:489-504
2. Patlewicz G, Kuseva C, Kesova A, Popova I, Zhechev T, Pavlov T, Roberts DW, Mekenyan O. Towards AOP Application –implementation of an integrated approach to testing and assessment (IATA) into a pipeline tool for skin sensitisation. RegulToxicol Pharmacol. 2014. 69:529-545
4. Natsch, A. Emter, R., Gfeller, H., Haupt, T. and G. Ellis. 2015. Predicting skin sensitizer potency based on in vitro data fromKeratinoSens and kinetic peptide binding: Global vs. domain-based assessment. Toxicological Science 143:319-332 .
6. & 7. Takanouchi et al. Test battery with the human cell line activation test, direct peptide reactivity assay and DEREK basedon a 139 chemical data set for predicting skin sensitizing potential and potency of chemicals, J Appl Toxicol. 2015. 35. 1318-32
8. Gomes, C., Noçairi, H., Thomas, M., Collin, J.F., Ibanez, F., Saporta, G., 2012. Stacking prediction for a binary outcome.COMPSTAT, 20th International Conference on Computational Statistics, Limassol, Cyprus. 2012: 271-282.
9. Jaworska. J, Dancik Y. Kern P. Gerberick F., Natsch A, A Bayesian integrated testing strategy to assess skin sensitizationpotency: from theory to practice, 2013 J. App. Toxicology 33. 1353–1364.
10. Van der Veen JW, Rorije E, Emter R, Natsch A, van Loveren H, Ezendam J. (2014) Evaluating the performance of integratedapproaches for hazard identification of skin sensitizing chemicals. Regul Toxicol Pharmacol. 69. 371-9.
11. Morihiko hirota et al., Evaluation of combinations of in vitro sensitization test descriptors for the artificial neural network-based risk assessment model of skin sensitization, J. Applied Toxicology. 2015. 35. 1333-47
12. MacKay, C., M. Davies, V. Summerfield and G. Maxwell. From Pathways to People: Applying the Adverse Outcome Pathway(AOP) for Skin Sensitization to Risk Assessment. ALTEX. 2013. 30. 473-486.
OECD TG No. 256 Annex A case study references Annex 1: www.oecd.org/officialdocuments/publicdisplaydocumentpdf/?cote=env/jm/mono(2016)29/ann1&doclanguage=en
Outline
• Skin Sensitisation AOP, NAMs & DAs
• DA example : ‘qAOP’ TKTD model of skin sensitization
• DA example 2: Skin Allergy Risk Assessment (SARA) model
• IATA example:NGRA Skin Allergy Risk Assessment
• Learnings and insights
• Questions?
Objective: model should be simplest representation of the AOP capable of reproducingthe induction of contact allergy to DNCB to enable prediction of a safe level of skinexposure (MacKay et al. 2013. http://altweb.jhsph.edu/altex/30_4/MacKay.pdf )
DA example : Quantitative AOP (qAOP)/TKTD model
DA example 1: Quantitative AOP (qAOP)/TKTD model
Objective: model should be simplest representation of the AOP capable of reproducingthe induction of contact allergy to DNCB to enable prediction of a safe level of skinexposure (MacKay et al. 2013. http://altweb.jhsph.edu/altex/30_4/MacKay.pdf )
Modelling covalent binding toskin proteins1. diffusion and partitioning
into the stratum corneum and skin
2. sensitiser clearance by dermal capillaries
3. covalent modification of protein nucleophiles by hapten.
DA example : Quantitative AOP (qAOP)/TKTD model
DA input: DNCB skin bioavailability data
- Ex vivo human skin penetration experiment – OECD TG 428 to include multiple time points and chemical free/bound measurements
- Pendlington et al. (2008). Cutan. Ocul. Toxicol. 27, 283-94.- Skin PK model: Davies et al. (2011). Toxicol. Sci. 119, 308–18.- Fitting to experimental data provides parameters for DNCB TK Model
Bespoke method:• Radiolabelled
DNCB in solution incubated with tape-stripped ex vivo human skin for up-to 24hrs
• Reaction diffusion system modelled and reaction rate determined by fitting to experimental data
DA input: DNCB protein binding data
Modelling T cell activation and proliferation4. proteasome processing of protein
nucleophiles to form small peptides and transport to the endoplasmic reticulum (ER)
5. binding of peptides and hapten-peptide complexes to Class I MHC and transport to plasma membrane
6. binding of pMHC and hapten-pMHC to CD8+ T cell receptors
7. activation of naïve specific CD8+ T cells
DA example : Quantitative AOP (qAOP)/TKTD model
proteasome
damaged proteins
peptides
DA input: proteasome processing rate parameter
Clin. Exp. Immunol. (1983) 53, 709-715.
Quantitative relationships between sensitizing dose of DNCB and reactivity in normal subjectsP. S. Friedmann, C. Moss, S. Schuster & J. M. Simpson
- 165 healthy human volunteers divided into five dose groups
- Single exposure to between 62.5 - 1000 µg DNCB applied to 7.1cm2
volar forearm in 100μL acetone vehicle
- Sensitisation assessed four weeks later by DNCB challenge
- Sensitisation varied between 8 - 100% across dose groups
Benchmarking DA output: historical human data
DA simulation of single 16.5µgcm-2 DNCB exposure
Thresholding DA output to predict adverse response
• Threshold defined (using literature data) that predicts the rate of human naïve CD8+ TcR signaling that would be sufficient to confer allergic status
• Threshold defined (using literature data) that predicts the rate of human naïve CD8+ TcR signaling that would be sufficient to confer allergic status
• Reverse Dosimetry used to back-calculate the doses of DNCB that would cross the TcR signal rate threshold/cause contact allergy
Thresholding DA output to predict adverse response
Outline
• Skin Sensitisation AOP, NAMs & DAs
• DA example : ‘qAOP’ TKTD model of skin sensitization
• DA example 2: Skin Allergy Risk Assessment (SARA) model
• IATA example: NGRA Skin Allergy Risk Assessment
• Learnings and insights
• Questions?
DA 2: Skin Allergy Risk Assessment (SARA) model
Protein Reactivity
OECD TG 442C: DPRA
ARE Nrf2 Activation
OECD TG 442D: KeratinoSens™
DC Activation
OECD TG 442E:
• hCLAT• U-Sens™
Historical AO data
• TG 429: LLNA• HRIPT
Objective: model should allow a weight-of-evidence prediction to be made of thehuman skin sensitisation potency using available in vitro and historical in vivo data(Reynolds et al. 2019 https://doi.org/10.1016/j.comtox.2018.10.004)
Model inputs:
DA 2: Skin Allergy Risk Assessment (SARA) model
Protein Reactivity
OECD TG 442C: DPRA Cys/Lys
depletion
ARE Nrf2 Activation
OECD TG 442D: KeratinoSens™
EC1.5
DC Activation
OECD TG 442E: • hCLAT CD54
EC200; CD86 EC150
• U-Sens™ EC150
Historical AO data
• TG 429: LLNA EC3• HRIPT dose, cohort
size, number sensitised
Approach:• ‘SARA’ Bayesian multi-level, hierachical regression model was developed to
estimate the human population skin sensitisation induction threshold, under theconditions of an HRIPT, for 30 chemicals using publicly available data.
• SARA model is probabilistic which allows the uncertainty in the populationthreshold prediction to be explicitly quantified
MIE/KE1 KE2 KE3 AO
DA 2: Skin Allergy Risk Assessment (SARA) model
Estimated human population threshold distributions using: all data (purple); HRIPTonly (green); LLNA only (blue); in vitro data only (red); no chemical specific data (grey)
Estimated human population threshold distributions using: all data (purple); HRIPTonly (green); LLNA only (blue); in vitro data only (red); no chemical specific data (grey)
Estimated human population threshold distributions using: all data (purple); HRIPTonly (green); LLNA only (blue); in vitro data only (red); no chemical specific data (grey)
Estimated human population threshold distributions using: all data (purple); HRIPTonly (green); LLNA only (blue); in vitro data only (red); no chemical specific data (grey)
DA 2: Skin Allergy Risk Assessment (SARA) model
Estimated human population threshold distributions using: all data (purple); HRIPTonly (green); LLNA only (blue); in vitro data only (red); no chemical specific data (grey)
Outline
• Skin Sensitisation AOP, NAMs & DAs
• DA example : ‘qAOP’ TKTD model of skin sensitization
• DA example 2: Skin Allergy Risk Assessment (SARA) model
• IATA example : NGRA Skin Allergy Risk Assessment
• Learnings and insights
• Questions?
IATA general workflow
Problem Formulation
Gather existing information
Weight of Evidence Assessment: adequate information for decision-making?
Generate additional information
DecisionAOP Yes
No Yes
IATA general workflow: applied to NexGen Skin Allergy Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
[Note: assume MDBGN is a new ingredient and no in vivo data or read-across candidates exist]
IATA general workflow: applied to Skin Sensitisation Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
Parameter Face cream
Amount of product used per day (g/day) 1.54
Retention factor 1
Skin surface area (cm2) 565
Amount of product in contact with skin (mg) 1536
Percentage ingredient in product (%) 0.2
Amount of ingredient in contact with skin (µg) 3072
Local dermal exposure (µg/cm2) 5
* Deterministic worse case (90th percentile) for Europe based upon SCCS notes of guidance (SCCS/1602/18)
IATA general workflow: applied to Skin Sensitisation Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
• Exposure predicted: 5µg/cm2 skin
KE1/MIE evidence
• TIMES: pre/pro-SN2
• ToxTree: MA/SN2
• OECD QSAR toolbox: SN2
• Expert view: Pro-MA
• DPRA: 100% depletion*
KE2/KE3 evidence
• KeratinoSens™
– EC1.5 = 7.8*
• hCLAT
– CD54 EC200 = 9.42*
– CD86 EC150 = 10.8*
• USens™
– CD86 EC150 = 3*
*Hoffmann et al. 2018: https://doi.org/10.1080/10408444.2018.1429385
IATA general workflow: applied to Skin Sensitisation Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
• Exposure predicted: 5µg/cm2 skin
• Evidence of sensitiser potential, apply SARA model (Reynolds et al. 2019)
Due to differences between the HRIPT
exposure and consumer use of a
face cream the dose response was adjusted by a factor of 10 fold
Vehicle Frequency / duration
Skin site condition
Dotted line represents product exposure (0.2% in a face cream)
IATA general workflow: applied to Skin Sensitisation Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
• Exposure predicted: 5µg/cm2 skin
• Evidence of sensitiser potential, apply SARA model (Reynolds et al. 2019) to predict human population threshold
Dotted line represents product exposure (0.2% in a face cream)
• If we use all in vitro data SARA prediction, the most likely rate of sensitization for use of a face cream containing 0.2% MDBGN is 1 in 200
• We conclude this exposure is high risk
IATA general workflow: applied to Skin Sensitisation Risk Assessment
• Is 0.2% Methyldibromoglutanitrile (MDBGN) safe for use in a face cream?
• Exposure predicted: 5µg/cm2 skin
• Evidence of sensitiser potential, apply SARA model (Reynolds et al. 2019)
Dotted line represents product exposure (0.2% in a face cream)
• If we use all in vitro data SARA prediction, the most likely rate of sensitization for use of a face cream containing 0.2% MDBGN is 1 in 200
• We conclude this exposure is high risk
– Note: inclusion of historical LLNA data would not alter the outcome
Learnings and insights
• AOP provides a framework for data weighting and integration
• Quantitative modelling can help refine mechanistic understanding by evaluating the value of each piece of information for decision-making
• Bayesian statistical approaches enable a weight-of-evidence prediction to be made even when input data are incomplete or conflicting
• DAs can enhance our ability to make weight-of-evidence risk assessment decisions but do not alter our need for good quality input data and upfront hypothesis generation