The assessment of chemical safety is necessary to ensure the protection of human health and the environment. Some chemicals possess potentially hazardous properties, so regulatory schemes such as REACH have been put in place to ensure that appropriate programmes of hazard assessment are conducted to generate data that can be used for classification, labelling/packaging and risk assessment.
Programmes designed to assess potential human health effects must consider possible adverse short- and long-term outcomes following exposure of the body via the skin, eyes, oral or respiratory routes. This will include effects that can occur at the site of contact, such as localized irritation, or whole body effects, such as systemic toxicity, skin sensitization, carcinogenicity and developmental and reproductive toxicity. Programmes for the assessment of environmental effects have to consider potential effects on terrestrial, aquatic and marine ecosystems.
Standardized test methods, such as the OECD Test Guidelines for Testing of Chemicals, have been developed to allow potential effects to be determined experimentally. For many years, methods to assess potential human health effects have required the use of various species of animals. Methods to assess environmental effects have used the target species themselves, such as fish or invertebrates like bees and earthworms.
For ethical reasons, humans are not used in the initial hazard assessment of chemicals, although they may be used in controlled clinical trials of new drugs and cosmetics following extensive preclinical safety investigations, which usually involve the use of animals. Although the use of animals has undoubtedly resulted in the protection of humans and the environment, society in general would prefer that they were not used and toxicologists are striving to develop new, more humane and scientifically reliable ways of generating hazard information.
In 1959, William Russell and Rex Burch introduced the concept of the 3Rs of reduction, refinement and replacement in animal experimentation, which were later termed ‘alternatives.’ The 3Rs are now firmly established as initial considerations whenever a programme of safety evaluation testing becomes necessary. So much importance is placed on them that bodies have been formed to promote best practice when animal use is necessary and to facilitate the development and implementation of new alternatives. These comprise
- Reduction alternatives: Methods for obtaining comparable levels of information from the use of fewer animals in scientific procedures, or for obtaining more information from the same number of animals
- Refinement alternatives: Methods that alleviate or minimize potential pain, suffering and distress, and which enhance animal well-being
- Replacement alternatives: Methods that permit a given purpose to be achieved without conducting experiments or other scientific procedures on animals.
There are also regulatory drivers for the use of alternatives. For instance, it is a requirement of the EU Animal Welfare Directive (2010/63/EU) that the 3Rs are considered ‘systematically’ and that animals are used only when no alternative is available. REACH also provides for the use of suitable alternatives where these exist and mandates the use of fully validated alternatives.
Under the Cosmetics Regulation (1223/2009/EC), the testing of products and ingredients in animals is no longer permitted in the EU and there is a ban on the marketing of finished cosmetics or cosmetics containing ingredients tested on animals after specific time points. This was 11 March 2009 for all human health effects except repeated-dose toxicity, reproductive toxicity and toxicokinetics. For these, the ban has applied since 11 March 2013. Alternative approaches that can potentially be used to replace use of animals include
- In chemico: abiotic measurements of reactivity or other physico-chemical properties of a chemical
- In silico: assessments performed using computational tools
- Ex vivo: studies on cells, tissues or organs taken from an organism
- In vitro: studies using cells or tissues grown or maintained in a laboratory.
To achieve regulatory acceptance, a replacement alternative method must first be shown to work consistently in the process known as validation; then a test guideline has to be produced. This can be a lengthy process, taking a number of years or even decades.
Validations can involve many organizations, and the activities are co-ordinated by bodies such as EURL ECVAM in Europe, ICCVAM in the US and JACVAM in Japan. In Europe, a European Network of Laboratories for the Validation of Alternative Methods (NETVAL) has been created in response to a provision of 2010/63/EU and currently includes 26 organizations.
There have been some successes in terms of the validation, test guideline development and regulatory acceptance of alternative test methods. There are also some methods that have successfully completed validation, but for which test guidelines are still in development. In most cases, full replacement of an animal test is likely to be achieved using a combination of one or more alternative test methods within an integrated approach to testing and assessment (IATA), to address various mechanisms of action or steps in a pathway of toxicity.
An example of where in vitro methods are already well established and accepted by regulators is for the assessment of genotoxicity, for which a range of tests is available to determine the potential of chemicals to cause inheritable damage to DNA, such as the bacterial reverse mutation (Ames) test and the in vitro mammalian chromosome aberration, mammalian cell gene mutation and micronucleus tests.
However, these are not full replacement tests because if positive results are obtained, it is still necessary to conduct confirmatory tests in animals. There are also various in silico tools that can aid in the assessment of genotoxicity.
There are three scientifically validated skin corrosion test methods: one is ex vivo (OECD 430) and two are in vitro (OECD 431 and 435), although 435 is probably more correctly described as an in chemico method. The tests allow classification of skin corrosivity according to the UN Globally Harmonized System for Classification of Chemicals (GHS) and the EU CLP regulations. They have fully replaced the requirement to conduct corrosivity testing in live animals.
The reconstructed human epidermis model uses three-dimensional laboratory grown skin epidermis consisting of human keratinocytes. The same type of test system is used to determine skin irritation potential according to OECD test guideline 439. Using this method, it is possible to distinguish between GHS/EU CLP Category 2 skin irritants and those chemicals that do not require classification for skin irritancy. So for classification purposes, it is no longer necessary to conduct tests for skin corrosion or irritation on animals.
There are similarly three test methods for eye irritation that have been validated and for which OECD test guidelines are available, each having some limitations. Two are ex vivo tests using tissues from animals that have been bred for meat production (OECD 437 and 438). The other (OECD 460) is an in vitro test using Madin-Darby Canine Kidney cells.
These tests can be used to identify chemicals that have the potential to cause serious eye damage (GHS/EU CLP Category 1), whereas the two ex vivo tests also allow for the identification of chemicals that do not require classification. However, none of them allows discrimination between Category 1 and 2 eye irritants, or between Category 2 eye irritants and those chemicals not requiring classification.
Three other test methods have completed validation as partial replacement methods and draft OECD test guidelines are under review. These are the reconstructed human cornea-like epithelium (RhCE) test method, the cytosensor microphysiometer test method and the short-term exposure in vitro test method.
All three should allow identification of chemicals not requiring classification and labelling for eye irritation and serious eye damage, but none will offer a complete replacement of the OECD 405 rabbit eye irritation test. A multi-laboratory, year long project has recently commenced under CEFIC Long Range Research Initiative AIM-T6 to compare the performance of a range of in vitro eye irritation test methods and to propose a testing strategy for full replacement of the test using rabbits.
The Local Lymph Node Assay (LLNA, OECD 429) is currently the most commonly performed method for assessing skin sensitization potential but it requires the use of mice. Testing in guinea pigs is now usually only conducted when the LLNA is not technically possible or for categories of chemicals that are known to produce false positive or negative results in the LLNA.
No single alternative method is available mainly because skin sensitization is a complex endpoint to model. However, the OECD has described an adverse outcome pathway (AOP, a sequence of events from the chemical structure of a target chemical or group of similar chemicals through the molecular initiating event to an in vivo outcome of interest) in its Series on Testing and Assessment, No. 168, and tests have been developed that address key events in the AOP.
Three alternative test methods have successfully completed scientific validation and were adopted as OECD test guidelines on 5 February 2015. One is in chemico (OECD 442C), the Direct Peptide Reactivity Assay, and one is in vitro (OECD 442D), the ARE-Nrf2 Luciferase Assay. Data generated using these methods can be used in a weight-of-evidence (WoE) assessment of skin sensitization potential. A draft OECD test guideline is under review for the third validated method, the Human Cell Line Activation Test (h-CLAT).
The in vitro 3T3 neutral red uptake (NRU) test for phototoxicity (OECD 432) is a validated alternative to the in vivo assay, although no OECD test guideline was ever finalized for the in vivo test. Substances that are phototoxic in vivo after systemic application and distribution to the skin, as well as compounds that could act as phototoxicants after topical application to the skin, can be identified by the test.
However, the pharmaceutical industry regards this assay as oversensitive. Additional in vitro assays, including reconstructed human skin models and the reactive oxygen species assay in combination with the 3T3 assay, will help to further reduce the use of animals for assessment of phototoxic potential.
Several in vitro alternatives have been developed for predicting carcinogenicity. Of these, the in vitro genotoxicity tests address only one mechanism involved in carcinogenicity, the induction of genetic damage. In contrast, in vitro cell transformation assays (CTAs) have been shown to involve a multi-stage process that closely models some stages of in vivo carcinogenesis.
The Syrian Hamster Embryonic (SHE) CTA and the BALB/c 3T3 CTA were selected for the performance assessment in the ECVAM prevalidation study. In 2013, ECVAM approved the use of a CTA based on the Bhas 42 cell line as a predictor of carcinogenic potential. Draft OECD guidelines have been prepared for both, although the SHE assay’s dependence on the harvest of primary cells from animals is one factor that reduces its value.
Such tests will need to be considered as part of a WoE approach or an integrated approach for testing and assessment, which must include genotoxicity tests rather than standalone tests, owing to the complexity of the mechanism for tumour development. Finally, several in vitro assays are now available for endocrine disruption. These are cited in OECD and US OCSPP guidelines for assessing the potential for chemicals to act as endocrine disrupters, including human cell-based whole hormone pathway assays, such as the steroidogensis assay.
In 2009, the US EPA mandated the testing of a list of 67 chemicals in an integrated in vitro and in vivo battery of screening tests to provide an opportunity to evaluate data from a diverse range of available tests. This was seen as the first step to establishing a strategic approach to addressing the enormous gap in data for an estimated 85,000 listed chemicals worldwide.
In May 2015 the agency announced a targeted data collection approach called the Integrated Bioactivity Exposure Ranking, based on high throughput in silico modelling: Toxcast, providing bioactivity assessment and Expocast, providing exposure predictions.
By establishing targeted priorities for testing and continuing investment in the development of new in vitro assays for endocrine disrupter assessment, it is expected that decades of costly and potentially unnecessary testing by industry will be minimized whilst accelerating the accumulation of safety data in the so-called ‘chemical universe.’
Even greater challenges exist in the development and validation of alternative methods to predict systemic toxicity when exposures occur via routes other than oral, and/or by repeated exposures. Against this backdrop, the European Commission launched the largest EU initiative ever undertaken on alternative methods, the €50 million Safety Evaluation Ultimately Replacing Animal Testing (SEURAT-1) under the FP7 programme.
The project commenced in 2011 and is now in its final year. It has focused on establishing the strategy and building blocks needed for the development of new non-animal test systems, leading to pathway based human safety assessments of repeated dose systemic toxicity. The ultimate aim of SEURAT-1 is to prove key concepts underpinning the credible use of combinations of in silico and in vitro methods to support safety assessment decisions without use of animals.
SEURAT-1 has been directed at a cluster of six related research projects spread across 70 European universities, public research institutes and private companies. Activities have included stem cell differentiation to provide human-based, organ-specific target cells to assay toxicity pathways in vitro, the investigation of human biomarkers in cellular models for repeated dose in vitro testing and delivery of an integrated suite of computational tools to predict the effects of long-term exposure to chemicals in humans based on in silico calculations.
Some impressive results were reported at the fifth annual meeting of SEURAT-1 in Barcelona in January and all the tools and methods will be collected in a catalogue to be made available at the end of 2015. Research beyond SEURAT-1 is likely to continue within the EC’s Horizon 2020 research and innovation programme, specifically under the Personalizing Health & Care Programme 33, “New Approaches to Improve Predictive Human Safety Testing.”
A range of in vitro assays has been developed and is being used, either as standalone tests or as part of IATA, to provide alternatives to the use of animals within regulatory frameworks for the safety assessment of chemicals. Inevitably, the initial successes in in vitro modelling have been for the ‘simpler’ endpoints such as skin corrosion and irritation, eye irritation and phototoxicity.
There have also been more recent advances with more complex endpoints, such as skin sensitization, in which in vitro regulatory models have been approved this year, hopefully in time to allow data to be generated for an estimated 20,000 substances for the May 2018 REACH registration deadline. It will be a huge missed opportunity if a significant number of animal tests are performed rather than the in vitro alternatives.
Driven by societal concerns about the use of animals, together with the desire for improvements in animal welfare through the implementation of the 3Rs and for improved science based on identification of AOPs, projects such as SEURAT-1 will provide information that will be extremely important for development of new alternative methods and approaches, some of which may use evolving technologies such as stem cells, microfluidics and 3D tissue modelling, in which ‘mini-organs’ are being produced or tissues even being ‘printed.’
The range of in vitro tests will certainly increase and there may be potential to automate some and incorporate them into high throughput testing programmes, thereby decreasing the time required to obtain results — but, at the same time, generating high volumes of data requiring improved bioinformatics.
It is evident that for the more complex toxicological endpoints, replacing the current use of animals will likely require multiple in vitro tests in combination with more sophisticated in silico tools and a greater understanding of AOPs. Challenges will therefore remain in terms of protracted time scales for the validation, implementation and worldwide regulatory acceptance of the new approaches.