Abstract
At the start of the recent meeting of the Society of Toxicology, one of the more interesting events was a global collaboration coffee meeting organized by International Union of Toxicology (IUTOX) and hosted by IUTOX Secretary-General Emanuela Corsini. The purpose of this virtual coffee morning was to bring together stakeholders from industry, academia, and government to discuss the steps needed to bring us to a point where animals are no longer used as the default model in toxicology. The session was organized with a distinguished group of panelists each giving their own opinions on how best to achieve the goal, followed by an extensive discussion. Panelists included Dr Warren Casey from the National Toxicology Program at NIEHS and the chief of the predictive toxicology branch, Dr Hajime Kojima, the secretary general of the Japanese Center for the Validation of Alternative Methods, Dr Thomas Hartung, director of the Center for Alternatives to Animal Testing at John Hopkins University, Dr Fiona Sewell, program manager for the National Center for the 3Rs, Dr Donna Mendrick, associate director of Regulatory Activities at the National Center for Toxicological Research (NCTR), Dr Elaine Faustman, School of Public Health, University of Washington, and Dr Bob van de Water, Cancer Therapeutics and Drug Safety, Universiteit Leiden.
A number of themes were enumerated throughout the session that ranged from the very practical, such as how to validate alternative methods, to the policy level, such as how to achieve globalization. However, certain themes came through strongly that point to the pathways and the challenges going forward. There was general agreement that the traditional “six pack” of acute toxicity analyses would be replaced for most compounds by 2022. Further it was clear that many of the regulatory agencies were committed to replacing animal testing, but there needed to be agreement into how new testing technologies could and should be developed and furthermore validated. Further it was pointed out that with the development of more complex drug mechanisms, such as biologicals, the requirements upon animal testing have grown with the inclusion of the need to use testing for product quality control. The question of validation and approval of the use of in vitro techniques by regulators was one that clearly concerned many panelists. It was generally agreed that as a scientific community, we needed to accept that neither animal testing nor the current in vitro methodologies were perfect. Clearly, we can see that despite our traditional acceptance of animal data, they do not provide a perfect model. However, we can also understand that novel alternative methods (NAMs) are needed to allow us to optimize the in vitro approach. A number of panelists stated that they could understand how difficult it was to be a regulator who had to rely on purely in vitro data without some clear validation of the technique used. However, all recognized that simple comparison with in vivo data was not the appropriate way to approach validation. It was recognized that in future meetings to discuss the use of NAMs in place of animal models, there would be a need for all components of the scientific community to be involved and thus it should include academics, industry leaders, and regulators.
There was much discussion concerning NAMs including direct testing systems, such as the development of engineered reporter stem cells that could be differentiated to a variety of phenotypes, microphysiological systems that could be particularly useful in the development of Pharmacodynamics and Pharmacokinetics, and the use of big data to generate Artificial Intelligence systems. The question that appears to remain in this context is how such NAMs can come to be accepted as toxicological assessment tools. It was noted that designers of NAMs possibly need to be in contact more with regulators such that they can understand what is wanted/required for the use of an NAM in the approval process. This would lead to a more focused and streamlined technique development. It was also clear that the very best science needs to be involved in NAM development, which may require combining information from different sources. An example of this in the past has been the development of three-dimensional cultures giving more reliable cell data. But even this may be enhanced by including in silico data with culture data to create a virtual four dimensional assessment.
It can be seen that there are a number of phases in the development of applicable NAMs to reduce animal use. First, there is a development phase wherein the best practices are identified and refined. It is important at this phase that one realizes that NAMs are not one for one replacements of existing technologies, but rather are improvements asking new questions and giving novel answers. As an example, one can consider using human-derived cells as a way of asking directly human-relevant toxicology questions. It may be that the involvement of regulators at this critical stage is the most important as they can inform the questions to be asked. It was also expressed quite clearly that if the right questions are asked, then we can consider this an engineering problem that technology can help us answer. The second phase is one of confidence building wherein data from the NAM can be incorporated and compared with known tests. Within this phase, one can also improve harmonization and data acquisition methodologies. Finally, this can lead to acceptance wherein the NAM can reduce the reliance upon animals and produce a more agile validation procedure.
However, the development of integrated techniques that can inform each other is perhaps the best way forward. Next generation risk assessment (NGRA) allows for such integration as has been performed by a number of companies and academic collaborations, including EuToxRisk. Here one can take data from multiple disciplines and integrate them to a position wherein the whole is greater than the sum of the parts. NGRAs, when developed as a mechanism-based integration modality, can use data from informatics, cell and molecular toxicity, absorption distribution metabolism and excretion, bioenergetics, biophysics, computational modeling, data science, and regulatory science. The development of a toolbox of techniques that can answer difficult questions within the NGRA can lead to more efficient and accurate toxicity analysis. One of the most important areas in this toolbox development is the consideration of how well any NAM can be used to address the relevance to human toxicity.
The issue of regulatory consideration impinged on another vital area that concerned the panelists, globalization. It was stated that the industry is often most concerned that NAMs will not be considered acceptable by regulatory agencies. Indeed, what is often considered acceptable in one country is not usable in another. Therefore, there was general agreement that globalized acceptance of NAMs was required for their acceptance as toxicity tests. The principal problem that occurs without true globalized acceptance is that companies are still required to perform animal tests to satisfy those countries that still require them. Thus, there is no real incentive to move away from animal tests. International harmonization is difficult to achieve, and indeed within the panel there were suggestions that both top down (global action first) and bottom approaches (nations first) may have the best chance to succeed. There was a consensus that without approval in the Organization for Economic Cooperation and Development countries, first there was little chance that developing countries would adopt NAMs. It was felt that increased transparency and demonstration of the scientific value of NAMs, beyond just animal replacement, were required. One interesting suggestion was that globalized acceptance of NAMs could be accelerated by the removal of supplanted techniques from the approved OECD lists.
An interesting end to the session was provided by Dr Corsini when she asked the panelists what the three biggest hurdles were to reducing animal testing today. The first was because of the need for international harmonization at the level of regulatory acceptance and because that requires development of high-level technologies that may reduce access to testing. The second was the need to change the mindset surrounding validation. We need to switch from a yes/no approach on techniques that tends to force an oversell on NAMs, such that ultimately, they disappoint. This was felt to be an area where improved education was required. The final concept was that fear about NAMs and the results they might give could squash a product's development. Once again this was an area where education was required. Overall, it was felt that we are moving in the right direction and that with improved communication and investment, one can envisage a future where animal testing has been replaced in toxicity assessment.