21st Century Safety Science and Non-Animal Approaches at Unilever-Improving human health and environmental risk assessment through an application of Source to Outcome Pathways-based thinking
Published: April 18, 2013
Professor Paul Carmichael is a Science Leader in SEAC. He has worked for Unilever for 9 years, following a previous role in academia at Imperial College London (UK). He has a passion for the modernisation of Toxicology and the introduction of exposure-, mechanism- and pathways-led safety assessment.
Ian Malcomber is an Expertise Leader in SEAC for Environmental & Process Safety and Sustainability. He has worked at Unilever for 10 years following previous roles in the consumer goods industry. An Ecotoxicologist, Ian is actively involved in Unilever and industry activities to advance mechanistic safety assessment approaches.
Dr. Gavin Maxwell is a Science Leader in SEAC. He has worked at Unilever for 8 years, predominantly on the development and application of non-animal risk assessment approaches for skin sensitisation. Before joining Unilever, Gavin studied Immunology and undertook his PhD at the Babraham Institute (University of Cambridge, UK), studying B cell development.
Dr. Oliver Price is a Science Leader in SEAC. He has worked at Unilever for 5 years following a previous role in consultancy supporting the pesticide industry. His primary research interest is in exposure sciences, including the development of tools that better characterize chemical emissions and exposure for use in risk assessments.
Dr. Julia Fentem is a senior leader within Unilever’s R&D organisation and Vice-President of SEAC. She is a biochemical toxicologist, with over 20 years experience in human health risk assessment and non-animal approaches for assessing safety gained while holding positions with an NGO, in government (European Commission) and in industry.
Five years ago, we published an AltTox article ‘Risk Assessment and New Technologies: Opportunities to Assure Safety without Animal Testing and Better Protect Public Health‘ (Fentem & Westmoreland, 2007) outlining Unilever’s research programme in these areas. Whilst assuring consumer safety of novel ingredients without the need for some animal testing remains a formidable challenge (Adler et al., 2011), there have been significant advances in the science of risk assessment since our 2007 article and we remain convinced that, with continued long-term research investment, this goal is ultimately achievable.
In 2007, we described a research strategy that sought to build on successes of the past in the area of alternatives to animal testing and would allow us to protect human health and the environment whilst facilitating innovation and new product development. The fundamental principles of this strategy remain unchanged: (1) defining the actual information required to make a risk-based safety decision and (2) using fundamental understanding of disease processes and adverse effects of chemicals at a mechanistic level to determine how this information could be generated without animal testing (Fentem et al., 2004; Westmoreland et al., 2010).
In 2007, the US National Research Council’s publication ‘Toxicity Testing in the Twenty-First Century (TT21C): A Vision and a Strategy’ (National Research Council [NRC], 2007) outlined an approach to safety assessment that ‘could transform toxicity testing from a system based on whole-animal testing to one founded primarily on in vitro methods that evaluate changes in biologic processes using cells, cell lines, or cellular components, preferably of human origin’. At the core of this Vision is an understanding of key biological pathways (‘toxicity pathways’) which, if sufficiently perturbed by a chemical, would result in an adverse outcome for the exposed human/environment. More recently, this concept of pathways-based approaches to risk assessment has been built upon by the description of ‘Adverse Outcome Pathways’ (AOPs). Each AOP begins with a Molecular Initiating Event (MIE) in which the chemical interacts with a biological target leading to a sequence of events across different levels of biological organisation (subcellular, cellular, suborgan, organ, individual, and population) resulting in an adverse outcome with direct relevance to a given risk assessment context (Ankley et al., 2010). The OECD have recently begun to formalize the use of this universal framework based on AOPs to capture and peer review the mechanistic understanding of specific toxic effects and as a framework for the evaluation of non-animal methods that aim to predict key events in these pathways (Organisation for Economic Co-operation and Development [OECD], 2011). It is essential that aspects of exposure to ingredients are also incorporated into AOP-based approaches for this framework to ultimately be useful in human and environmental risk assessment as described by Crofton, et al. (2011) as ‘Source to Outcome Pathways’.
In addition to the US and OECD activities, there are many other global activities in progress to define and explore pathways-based approaches to safety assessment. Within Europe these include: (1) AXLR8 EU FP7 collaboration between academic, EU Commission, NGO, and industry scientists aiming to accelerate transition to toxicity pathway-based methods for chemical safety assessment, (2) SEURAT-1, an EU FP7 collaboration co-funded by the EU Commission and Cosmetics Europe to drive a first wave of academic research required to replace the need for repeat-dose systemic toxicity animal data, and (3) Projects which look to bring novel and mechanistic approaches to ecotoxicology risk assessment (e.g., EUROECOTOX and CREAM). There are also significant opportunities for high-impact EU-US transatlantic research collaboration on 21st Century safety science and non-animal tools within the EU’s new programme for research and innovation (Horizon 2020; Fentem & Maxwell, 2012). Our own recent efforts in bringing together the important elements and necessary partnerships for 21st Century safety science are illustrated in Unilever’s TT21C.
It is important to understand how the various components of a pathways-based approach to understanding toxicity can be brought together in the context of a risk assessment. Mode of Action (MOA) approaches have been used for chemical classification and read-across for aquatic toxicology for some time (Ankley et al., 2010), and likewise MOA frameworks are also well established in the context of human health assessments (e.g., in analysis of the human relevance of certain types of toxicity observed in experimental animals) (Boobis et al., 2008). However, the new suggestions of the use of AOP/Source to Outcome Pathway-based approaches for assuring safety draw increasingly upon 21st Century scientific disciplines that have not traditionally been key building blocks to (eco)toxicological risk assessments. These new technologies can include tissue engineering, use of ex vivo tissues, data mining and bioinformatics approaches, systems biology, mathematical modelling, analytical science, and imaging. During our research in this area since 2004, we have increasingly found that a major contribution to success has been interdisciplinary teams working and learning together, applying novel modelling and informatics techniques, with specific risk assessment questions at the heart of their research. Several key disciplines have become increasingly important to our work in this area:
Mechanistic Chemistry: The central role of understanding the initial reactions of a foreign chemical with a biological system has been highlighted by the requirement to define MIEs. Whilst chemistry (and in particular analytical chemistry) has always been an important discipline within toxicology, there is now an increasing need for mechanistic chemists and biologists to work closely together to understand reactive chemistry, ligand biology, receptor binding, kinetics, and metabolism (European Partnership for Alternative Approaches to Animal Testing [EPAA], 2010; Gutsell & Russell, 2013).
Exposure Science: Decisions on consumer and environmental safety of novel ingredients are risk-based (made on an assessment of ‘acceptable risk’) and are driven by an understanding of the level of exposure to the ingredient. Traditional toxicological risk assessments have relied on metrics associated with applied dose (e.g., mg/kg/day, µg/cm2) based on information and assumptions about consumer habits and practice. However, new risk assessment paradigms which are pathways-based will demand an understanding of the biologically relevant exposure to an ingredient (e.g., bioavailability in the skin, levels of systemic exposure).
Over the last decade, the development and application of new technologies has enabled rapid advancement of exposure science for use in chemical risk assessment. Exposure science is now considered a distinct discipline, but considerable scientific and technical challenges remain in this area. In 2010, the NRC convened an inter-discipline panel to discuss the current status of exposure science and outline research priorities to revolutionise chemical safety assessment. The review, published in September 2012 entitled ‘Exposure science in the 21st century: a Vision and a Strategy outlines the challenges over the next decade to develop an integrated approach that considers exposures from source to dose, on multiple levels of integration (including time, space, and biologic scale) to multiple stressors, and scaled from molecular systems to individuals, populations, and ecosystems (NRC, 2012). In parallel, the EC Scientific Committees published a discussion paper in 2012 entitled ‘Addressing the new challenges for risk assessment’ (European Commission [EC], 2012), which identified the need to update exposure assessments used to determine human health and environment risk as a research priority. A major driver for change in human health risk assessment approaches is the use of non-animal data. Research priorities include the advancement of exposure based waiving tools, development of systems-based models for estimating internal exposure (e.g., PBPK and biokinetic and biodynamic models), and design of in vitro toxicity testing strategies based on exposure relevant doses. The importance of understanding mode of action to provide a scientifically justified basis of characterising threshold of adverse effects and identifying vulnerable populations was outlined. Such developments will enable scientifically sound read across, a relevant framework for grouping of chemicals, and for the risk assessment of mixtures. The need to develop integrated models of exposure and ecology to improve the realism and relevance of environmental assessments was also identified as a priority.
Mathematical Modelling and Biological Inputs: One of the biggest changes in mindset amongst safety scientists embracing pathways-based approaches to safety assessment has been the partnership between biologists and mathematical modellers. The complex nature of the multistage toxicity pathways that are currently being explored demands that integration of information across these pathways is necessary to allow risk assessment decisions to be made. The modeling of biological systems needs input from the real-life situation. These models evolve when challenged and refined with new experimental data generated in a cyclical fashion, and it is important that laboratory-based safety scientists and mathematical modellers work together as one team to design and carry out bespoke experiments to inform the mathematical models. This reiterative wet-dry cycle is a distinguishing feature of systems biology and is a new way of working for many toxicologists who have traditionally relied on standalone hazard characterisation tests using fixed protocols.
Clinical Data: The new paradigms for consumer safety have also highlighted the importance of safety scientists working closely with clinicians who have experience of studying the adverse outcomes, which these approaches aim to prevent. This was reflected in our original conceptual framework for assuring safety without animals, where we indicated that, (a) understanding the adverse outcomes which we are aiming to prevent in our consumers, and (b) harnessing the new science and technology being applied within the field of clinical research, would be key to the success of a new paradigm for assuring consumer safety (Fentem et al., 2004).
Case Studies at Unilever
To explore how these new technologies may be brought together within Unilever to allow decisions to be made about the risks posed by novel chemicals to consumers, we are continuing to adopt a case study approach as described in our 2007 AltTox article. The feasibility of this approach has been assessed with research partners outside and inside Unilever and we are currently working with 30 US, EU, and China-based academic and contract research groups. This work is regularly presented at conferences and has resulted in almost 100 publications in the scientific literature since 2006. Whilst our focus during the past 10 years was on developing new non-animal scientific capabilities (e.g., mathematical modeling, informatics), Unilever’s focus over the next decade will be on how to interpret non-animal data from applying these new capabilities in an AOP framework for safety and regulatory decision-making on human health (and environmental) safety risks.
Case Study 1: Skin Allergy Risk Assessment
To ensure that our products do not induce skin allergy in our consumers, the aim of this case study is to develop and evaluate a new, scientifically robust approach to enable us to perform risk assessments without the generation of data in animal models. This is the most advanced of our case studies and progress with the overall research programme has been summarized at several points since its initiation (Jowsey et al., 2006; Maxwell et al., 2008; Maxwell et al., 2012; Mackay et al., 2013). The approach taken uses a number of in vitro, in chemico, and in silico approaches derived from a mechanistic understanding of allergic contact dermatitis in humans. This mechanistic understanding was recently reviewed by the OECD in the context of an AOP for skin sensitization initiated by covalent binding to proteins (OECD, 2012).
Figure 1. Flow diagram of the pathways associated with skin sensitization. (Source: OECD. (2012). The Adverse Outcome Pathway for Skin Sensitisation Initiated by Covalent Binding to Proteins Part 1: Scientific Evidence. [Series on Testing and Assessment No.168 ENV/JM/MONO(2012)10/PART1]. Reprinted with permission.)
This AOP clearly maps the disease process from MIE (covalent modification of epidermal protein [haptenation]) to the adverse outcome observed clinically as an eczematous skin reaction. The approach to risk assessment we are currently exploring based on this AOP can be broadly divided into 2 phases:
- Modelling the kinetics and bioavailability of the ingredient to determine a free concentration of the ingredient at the target site within the skin following consumer application. It is this concentration which is available to drive the MIE. The technologies which underpin this part of the risk assessment include the use of physiologically-based compartmental models (Davies et al., 2011). Some of the key input data to these models are from in vitro percutaneous absorption measurements. Such measurements have been used for many years to give estimates of systemic bioavailability of ingredients following dermal application (OECD, 2003), but have been modified to develop methods to study epidermal/dermal disposition of contact allergens in the skin (Pendlington et al., 2008). Several other parameters (e.g., formulation effects, evaporation, diffusion, metabolism, extra/intracellular distribution) are also modelled to give the estimated concentration at the target site in the skin. The final output from this phase of the risk assessment would be an estimate of the concentration of haptenated skin protein for which knowledge of the protein binding potential of the ingredient is used. A number of tools (both in silico and in chemico) can be used to assess the potential reactivity of novel ingredients which we have developed over many years; several of these tools are used by us today in the assessment of new ingredients (Aleksic et al., 2007; Aleksic et al., 2009; Aptula et al., 2009).
- Modelling the likelihood of an adverse outcome in the consumer resulting from the predicted concentration of haptenated skin protein. This research builds on our earlier work to understand how systems biology approaches might be applied to support safety risk assessments of the future where we collaborated with Entelos, Inc. to construct a computer-based mathematical model of the induction of skin sensitization (Maxwell & Mackay, 2008). Whilst the ultimate adverse outcome from this pathway is elicitation of allergic contact dermatitis (ACD), the OECD AOP for skin sensitisation suggests that antigen-specific T-cell proliferation is a key causal event prior to this outcome. At a Unilever workshop with leading experts in 2012, it was suggested that a number of characteristics of the clonally expanded T cell population may be relevant in the establishment of ACD in humans (Kimber et al., 2012). Initially, our mathematical model of the allergic immune response is focusing on one of these characteristics: magnitude (i.e., the number of allergen-reactive T-cells, dependent on the vigour and duration of their proliferation). The model aims to predict, in the first instance, CD8+ T-cell numbers in the blood and skin with the ultimate aim of relating the dose response of this CD8+ T-cell response with the potential to induce skin sensitization. A key component of this research has been to work with clinical partners to initiate new research to generate sensitiser-specific data to test and improve the mathematical model. This includes clinical work in patients diagnosed with ACD or undergoing sensitisation to chemicals (e.g., for treatment of viral warts) to correlate the degree of sensitisation with the number of antigen-specific T-cells.
Ultimately, it is hoped that this modelling approach for skin allergy risk assessments will allow us to benchmark the threshold at which the number of antigen-specific T-cells correlates with clinical adversity. This information will then allow us to make informed decisions about the risk of skin sensitisation which would be predicted in consumer populations following inclusion of a novel ingredient into a Unilever product.
Case Study 2: Systemic Toxicity Risk Assessment (DNA Damage)
In the case study of skin allergy, a well established AOP exists that describes the key events from MIE (haptenation of skin protein) to adverse outcome (ACD). For many other pathways associated with systemic toxicity, sufficient perturbation of these pathways by a novel ingredient can ultimately lead to a variety of adverse outcomes, largely dependent on the absorption, distribution, metabolism, and excretion (ADME) properties of the chemical within the body and the specific nature of dose responses within particular organs. A number of potential toxicity pathways exist, many of which are currently being studied globally in the context of the AOP framework:
- pro-oxidant and free-radical based toxicity
- mitochondrial targets and energy production disruption
- specific receptor agonist/antagonist targets [functional receptors + cell surface receptors; nuclear receptor mediated + promoter regions; hormonal]
- covalent binding of electrophiles (DNA and protein dysfunction)
- immuno-specific mechanisms and cytokine effects
- disruption of calcium homeostasis and signalling
- direct necrosis or apoptosis induction
- impairment of cell proliferation and tissue repair
- interfering with ion transporters and metabolism
We have chosen to explore (in collaboration with the Hamner Institutes in the US) DNA damage responses mediated by the p53 network (Bhattacharya et al., 2011; Adeleye et al., 2013). Whilst we acknowledge that one pathway will not solve all of the issues around assuring safety for systemically available materials, this case study has allowed us to explore how the various aspects of a TT21C-based risk assessment may be brought together. We hope that ultimately the generic lessons from this case study will be applicable to other pathways in the future. P53-mediated DNA damage was chosen because of the central role of DNA damage in the pathway of carcinogenicity and also because the Unilever team have an established expertise in the area of mutation/DNA damage.
Figure 2. Proposed draft AOP for p53-induced DNA damage (Source: Y. Adeleye, SEAC, Unilever, and M. Andersen, Hamner Institutes)
* Endogenous factors: defects in terminal differentiation, resistance to apoptosis & cytotoxicity, defects in growth control, activation of proto-oncogenes and inactivation of tumour suppressor and genomic stability genes.
As with skin allergy (and indeed all risk assessments), the risk assessment process begins with an assessment of consumer exposure to the ingredient, based on knowledge of the product formulation and habits and practice information from consumers. Importantly, in the case of systemic toxicity evaluation, risk assessments will need to change from being based on predicted levels of external exposure to being based on predicted levels of internal/systemic exposure. To allow in vitro-in vivo extrapolation of toxicity pathways data, physiologically-based biokinetic (PBBK) models will play a central role in understanding levels of systemic exposure. PBBK models have been used to predict free concentrations of chemicals in the plasma and tissue following exposure, which requires additional compound-specific information such as microsomal intrinsic clearance, fraction unbound to plasma, aqueous solubility, caco-2 permeability, etc. Systemic exposure estimates are a key parameter for such risk assessments and, in some cases, clinical data may be required to ‘validate’ PBBK estimates. We have recently investigated the use of clinical microdosing studies to give accurate measures of metabolic turnover, absolute oral bioavailability, clearance, and volume of distribution (Duchateau et al., 2012).
The ‘adverse outcome’ which was selected for this initial case study work on the p53 pathway was in vitro micronucleus formation, i.e., an irreversible damage to cellular DNA. Concentration-response information was then generated using high throughput automatic analysis for the case study chemicals to determine the threshold doses at which chemicals begin to induce this adverse outcome. Importantly, to allow in vitro to in vivo extrapolation of results the actual free concentration to which cells are exposed must also be determined analytically rather than relying on the nominal concentration applied to cells. Factors such as evaporation, binding to tissue culture plastic, binding to serum in culture medium, and metabolism can all affect the final free concentration of chemical to which cells are ultimately exposed.
High content in vitro information (both flow cytometry and image analysis data) from cells treated with case study chemicals is then used to determine time and concentration-related effects on key biomarkers within the p53 network (for example levels total of p53, phosphorylated p53, H2AX, ATM, ATR, ChK2, MDM2, Wip1, etc). This information is integrated into ordinary differential equation-based models of the p53 DNA damage network to characterise the dose response for DNA damage pathway activation and ultimately determine an ‘acceptable in vitro concentration’ which does not adversely perturb this pathway. Key to defining this ‘acceptable in vitro concentration’ is a robust understanding of the homeostasis of the pathway being studied. In the case of DNA damage, there are many control and repair responses that may be associated with an initial MIE that may alter DNA, and it is important to differentiate those responses which are adaptive and part of normal biology from those which will lead to irreversible DNA damage.
Ultimately, an understanding of human systemic exposure to a novel ingredient, together with information on concentrations of this ingredient that do not adversely perturb the p53-mediated DNA damage response, will form the basis of a risk assessment approach that brings together all of the aspects of the 2007 NRC Vision and Strategy for toxicity testing in the 21st Century.
Beyond Consumer Safety
The OECD AOP framework which we have described above in the context of new approaches to consumer safety risk assessment provides a unifying framework for broader safety assessments. Indeed the ‘Adverse Outcome Pathway’ and ‘Source to Outcome Pathway’ terminology originated in the field of ecotoxicology (Ankley et al., 2010; Crofton et al., 2011) where it is proposed that a pathways-based approach to ecological risk assessment using new biological, exposure, and computational sciences could help the increasing need for assessors to perform risk assessments with greater speed and accuracy using fewer resources and experimental animals. Within the Safety and Environmental Assurance (SEAC) group at Unilever our toxicologists, ecotoxicologists, and chemists are co-located and work closely together in project teams. We therefore envisage a future where mechanistically-based, exposure-driven risk assessments will underpin and improve the transparency and robustness of decisions across both human and environmental safety. These risk assessments will be driven by new science in the areas of both human and environmental exposure assessment to our ingredients, encompassing the full ‘source to outcome’ aspects of risk assessment and management. Many challenges still exist in bringing this vision to a reality. In particular, a clear understanding is needed (with case study examples) of how a pathways-based approach can be used not only to identify potential hazards associated with chemicals, but to provide information to allow risk assessments to be generated that can be confidently used in safety decision making. A unified framework to approach risk assessments may also address some of the future challenges highlighted recently by the Scientific Committees of the EC. These committees, when looking at the necessary improvement of risk assessment science in view of the needs of risk managers and policy makers, commented that ‘risk assessments should be expressed (whenever appropriate) in terms of value-relevant impacts on humans and ecosystems rather than in terms of the somewhat technical surrogates often used in the routine risk characterizations’ (EC, 2011).
As mentioned earlier, a global approach to harnessing 21st Century science for human and environmental risk assessments will be essential to delivering on these goals. The efforts of researchers within the EU JRC, the US EPA, and other OECD member countries are increasingly focussed on a coherent global strategy in this area (OECD, 2011). Importantly, scientists within China are also engaged with progressing both exposure science and pathways-based research (www.TT21C.org, Price et al., 2012) ensuring that, as the thinking develops in this area, future risk assessment frameworks will be fit-for-purpose for the variety of different consumers and ecosystems which they are intended to protect, and will increasingly be more mechanistic and less reliant on new animal data.
The authors would like to thank Deborah Parkin for her help with the preparation of this article and the many talented scientists at Unilever, SEAC who are currently working on our research programme in this area and have contributed to this research over the past 10 years. We would also like to thank our network of scientific collaborators that currently includes: Academy of Military Medical Sciences (China), ADAS UK, Ltd., ARC Arnot Research and Consulting, Inc., Barts and the London NHS Trust, Brunel University, Chinese Academy of Sciences, Cranfield University, Hamner Institutes for Health Sciences, Lancaster University, Peking University, Stockholm University, Swansea University, University of Bradford, University of Cambridge, University College London, University of Durham, University of Leeds, University of Liverpool, University of Manchester, University of Newcastle upon Tyne, University of Southampton, University of Warwick, University of Waterloo, Utrecht University, Wageningen University, Waterborne Environmental, Inc.
©2013 Carl Westmoreland, Paul Carmichael, Ian Malcomber, Gavin Maxwell, Oliver Price & Julia Fentem
Adler, S. et al. (2011). Alternative (non-animal) methods for cosmetic testing: current status and future prospects-2010. Arch. Toxicol. 85, 367-485.
Aleksic, M. et al. (2007). The role of non-covalent protein binding in skin sensitisation potency of chemicals. Cutan. Ocul. Toxicol. 26, 161-169.
Aleksic, M. et al. (2009). Reactivity Profiling: covalent modification of single nucleophile peptides for skin sensitization risk assessment. Toxicol. Sci. 108, 401–411.
Ankley, G.T. et al. (2010). Adverse outcome pathways: A conceptual framework to support ecotoxicological research and risk assessment. Environ. Toxicol. Chem. 29, 730-741.
Aptula, A.O. (2009). Chemical mechanisms for skin sensitization by aromatic compounds with hydroxy and amino groups. Chem. Res. Toxicol. 22, 1541-1547.
Bhattacharya, S. et al. (2011). Toxicity testing in the 21st Century: Defining new risk assessment approaches based on perturbation of intracellular toxicity pathways. PLoS ONE 6(6), e20887. Retrieved from http://www.plosone.org
Crofton, K. et al. (2011). Pathways of toxicity: Working definitions and examples. In Report of the workshop on using mechanistic information in forming chemical categories (pp. 40-53). (Health and Safety Publications Series on Testing and Assessment No. 138. ENV/JM/MONO(2011)8). Paris, France: OECD.
Davies, M. et al. (2011). Determining epidermal disposition kinetics for use in an integrated non-animal approach to skin sensitization risk assessment. Toxicol. Sci. 119, 308-318.
Duchateau, G. et al. (2012). Absolute oral bioavailability and metabolic turnover of β-sitosterol in healthy subjects. Drug Metab. Dispos. 40, 2026-2030.
European Commission (EC). (2011). Improvement of risk assessment in view of the needs of risk managers and policy makers. Retrieved from http://ec.europa.eu/health/scientific_committees/environmental_risks/docs/scher_o_154.pdf (Accessed March 26, 2013).
EC. (2012, October). SCENIHR (Scientific Committee on Emerging and Newly Identified Health Risks), SCHER (Scientific Committee on Health and Environmental Risks), SCCS (Scientific Committee on Consumer Safety), Preliminary report on addressing the new challenges for risk assessment. Retrieved from http://ec.europa.eu/health/scientific_committees/emerging/docs/scenihr_o_037.pdf (Accessed April 5, 2013).
European Partnership for Alternative Approaches to Animal Testing (EPAA). (2010). Workshop on computational chemistry and systems biology. Harnessing the chemistry of life: Revolutionising toxicology. Retrieved from http://ec.europa.eu/enterprise/epaa/2_activities/2_1_science/ws_comp_chem_final.pdf (Accessed March 26, 2013).
Fentem, J. et al. (2004). The feasibility of replacing animal testing for assessing consumer safety: a suggested future direction. Altern. Lab. Anim. 32, 617-623.
Fentem, J., & Maxwell, G. (2012). Advancing safety science and health research under Horizon 2020 with innovative, non-animal tools. Retrieved from http://www.hsi.org/issues/advancing_science/facts/ep-workshop-oct-2012/fentem.pdf. (Accessed March 26, 2013).
Fentem, J., & Westmoreland, C. (2007). Risk assessment and new technologies: Opportunities to assure safety without animal testing and better protect public health? Retrieved from http://www.alttox.org/ttrc/overarching-challenges/way-forward/fentem-westmoreland/ (Accessed March 26, 2013).
Gutsell, S., & Russell, P. (2013). The role of chemistry in developing understanding of adverse outcome pathways and their application in risk assessment. Toxicol. Res. (submitted for publication).
Jowsey, I.R. et al. (2006). A future approach to measuring relative skin sensitising potency: a proposal. J. Appl. Toxicol. 26, 341-350.
Kimber, I. et al. (2012). Allergic contact dermatitis: a commentary on the relationship between T lymphocytes and skin sensitising potency. Toxicology 291, 18-24.
MacKay, C. et al. (2013). From pathways to people: Applying the adverse outcome pathway (AOP) for skin sensitisation to risk assessment. ALTEX (submitted for publication).
Maxwell, G., & MacKay, C. (2008). Application of a systems biology approach for skin allergy risk assessment. Altern. Lab. Anim. 36(5), 521-556.
Maxwell, G. et al. (2008). Assuring consumer safety without animal testing: A feasibility case study for skin sensitisation. Altern. Lab. Anim. 36, 557–568.
Maxwell, G. et al. (2012). Skin sensitization: Modelling the human adverse response. ALTEX WC8 Proceedings 1/12, 225-229.
National Research Council (NRC). (2007). Toxicity testing in the twenty-first Century: A vision and a strategy. Washington, DC: Author.
NRC. (2012). Exposure science in the 21st Century: A vision and a strategy. Washington, DC:Author.
Organisation for Economic Co-operation and Development (OECD). (2003). OECD guidelines for the testing of chemicals No 428: Skin Absorption – In vitro method. Paris, France: Author.
OECD. (2011). Report of the workshop on using mechanistic information in forming chemical categories. (Health and Safety Publications Series on Testing and Assessment No. 138. ENV/JM/MONO(2011)8). Paris, France: Author.
OECD. (2012). The Adverse Outcome Pathway for skin sensitisation initiated by covalent binding to proteins. Part 2: Use of the AOP to develop chemical categories and integrated assessment and testing approaches. (Series on Testing and Assessment No. 168. ENV/JM/MONO(2012)/PART2). Paris, France: Author.
Pendlington, R.U. et al. (2008). Development of a modified in vitro skin absorption method to study the Epidermal/Dermal disposition of a contact allergen in human skin. Cutan. Ocul. Toxicol. 27, 283-294.
Price, O.R. et al. (2012). Chemicals management and environmental assessment of chemicals in China. Environ. Pollut. 165, 169.
Westmoreland, C. et al. (2010). Assuring safety without animal testing: Unilever’s ongoing research programme to deliver novel ways to assure consumer safety. ALTEX 27(3), 61-65.