In simple terms, a read-across is a solution to the ethical dilemma on how to avoid unnecessary testing in animals with unknown compounds. But does it provide the solution? To answer this and more, the Symposium Session “Strategic Development of Read-Across within the EU-ToxRisk Project and Beyond” at the SOT 58th Annual Meeting and ToxExpo brought together experts in this developing field of read-across. The session was chaired by Dr. Marcel Leist, head of the Department of In Vitro Toxicology and Biomedicine at the University of Konstanz, Germany, and co-chaired by Dr. Russell Thomas from the US Environmental Protection Agency. Dr. Leist introduced the nomenclature of new approach methods (NAM), traditionally called alternatives to animal testing, which include in vitro methods, in silico methods (QSAR), and now the read-across approach. He went on to elaborate that the read-across involves using the information from the analogue substances (structurally similar compounds), which are the source to predict properties of the unknown substance. Dr. Leist provided the example of butyl paraben; if it were an unknown compound, then information from propyl paraben and pentyl paraben can be used to conduct a read-across. He believed three important questions would be answered by the end of the session: (1) How successfully can NAM be used to reduce uncertainties and to improve the quality of read-across? (2) What is the relationship between QSAR and read-across? (3) How can we make read-across successful? He spoke about European Chemicals Agency Read Across Assessment Framework (RAAF), which provides only an assessment framework, not a guidance, which is missing, and how EU-ToxRisk is working to bring together a read-across guidance. With this brief introduction, Dr. Leist opened the session for other distinguished speakers to discuss this in depth.
Dr. Bob van de Water, head of the Cancer Therapeutics and Drug Safety group at Leiden University, who is also the coordinator of the EU-ToxRisk project, took the stage to highlight the status of the project, which is funded with approximately 30 M through the European Commission’s Horizon 2020 scheme for six years and involves more than 38 European partners. At the start of the project, the idea was to build a roadmap on what is needed to implement the NAM testing and set up a timeline to achieve this. Dr. van de Water described a few case studies wherein toxicokinetic information and exposure scenarios were used from a known molecule to predict the toxicodynamic information of the unknown molecules and the different test systems available. He emphasized the need to discuss these case studies with regulators and affirm whether the NAM can be used. He listed out the test systems that the EU-ToxRisk project is utilizing, including in vitro and in silico methods from HEPG2 cell lines to CRISPR/cas mediated platforms to computational models and QSAR. He quoted examples of case studies ranging from repeat-dose toxicity studies for microvesicular liver steatosis to ab initio liver toxicity studies. Finally, he explained how EU-ToxRisk is working on a joint case study with Syngenta wherein Syngenta’s compounds will be used to develop and study test systems. The EU-ToxRisk group wants to officially evaluate whether their test systems are fit for purpose for read-across.
Dr. Bob van de Water’s presentation raised the question of how the NAMs can be utilized in risk assessment, and this was answered in detail by Dr. Sylvia Escher from the Fraunhofer Institute for Toxicology and Experimental Medicine, Hannover, Germany. She explained the concept of the analogue and the category approach in the read-across process and how the NAM can be utilized to bridge the gap using integrated approaches to testing and assessment (IATA). A few of the case studies listed by Dr. van de Water were discussed in depth by Dr. Escher, emphasizing the difference between targeted and untargeted testing. The most vital piece of information that was provided in her talk was how the read-across prediction was validated by performing an in vivo study. Dr. Escher explained a few more cases and posed the question whether the NAM approach is fit for regulatory purpose.
After a very engaging technical session, Dr. Hennicke Kamp from the Experimental Toxicology group of BASF SE lightened the mood by joking that he came from the “dark side” because he is a chemist. His talk was about the international acceptance of read-across based on NAMs. Dr. Kamp emphasized that the current guidance or frameworks like IATA, RAAF, and others are not the only solution to the problem. The acceptance of a read-across submission is less than 10%, and the reasons for nonacceptance by the regulators were highlighted by Dr. Kamp, like unclear substance identity and difficulty to ascertain structural similarity, lack of sufficient evidence, read-across using inappropriate data, and the lack of scientific possibility. He also quoted a case study with phenoxy herbicides. The general feedback from the regulators was to get the science right, convince them that you’ve got it right, and address the specific requirements of the regulatory process. For regulators, if adequacy and reliability of information cannot be assessed, then the information is nonexistent. Dr. Kamp reiterated the stand of EU-ToxRisk to come up with a guidance document.
This was followed by a talk on big data and machine learning to predict chemical toxicity by Dr. Thomas Hartung from the Johns Hopkins University Center for Alternatives to Animal Testing (CAAT). He elaborated on how modern technologies have made sense of big data and the momentum that read-across has gained because of this. He quoted the importance of big data and machine learning in circumstances wherein scarce data are available—for example, during the Elk River chemical spill in January 2014, wherein only the LD50 rat data were available. Dr. Hartung supported the fact quoted by Dr. Kamp about low acceptance of read-across by the regulatory by citing that 75% of read-across submissions were done between 2010 and 2013 and only a few were accepted. Dr. Leuchfeld, a student of Dr. Hartung, extracted the data from European Chemicals Agency’s dataset into a structured, machine-readable database of 70 million structures. He also informed the audience about the new web-based tool REACHAcross, developed with Underwriters Laboratories, which supports structure-based read-across for acute oral and dermal toxicity, eye and skin irritation, mutagenicity, and skin sensitization. Next, he spoke about data fusion wherein a single model was used for nine hazards using chemicophysical information. Dr. Hartung claimed that 91% sensitivity was obtained with these predictions, which outperformed repeat animal studies.
The audience heard about what the regulators accept and what they do not by many of the speakers, and among the audience was Dr. Suzanne C. Fitzpatrick from the US Food and Drug Administration (US FDA) Center for Food Safety and Applied Nutrition. Dr. Fitzpatrick was the final speaker of the session, who explained the view from the regulatory perspective. She spoke about the US FDA Predictive Toxicology Roadmap. She also spoke about the goals of the cooperative research and development agreement (CRADA) between the US FDA and other private organizations and the Interagency Coordinating Committee on the Validation of Alternative Methods (ICCVAM) work group on read-across. The most important aspect that Dr. Fitzpatrick highlighted was the use of toxicology equivalent factors involving biological endpoints rather than chemical to arrive at a better understanding of the read-across approach.
To summarize this symposium session, the EU-ToxRisk is a “flagship” program that is working toward bringing out a guidance on the read-across approach using new approach methods. Some read-across case studies have been validated by in vivo data, but more such studies are required. Also discussed was how big data and machine learning can support this approach and finally how to convince the regulators to accept this to minimize use of animals in research as well as reduce risk to human health.
This blog was prepared by an SOT Reporter. SOT Reporters are SOT members who volunteer to write about sessions and events they attend during the SOT Annual Meeting and ToxExpo. If you are interested in participating in the SOT Reporter program in the future, please email Giuliana Macaluso.