In silico toxicology refers to computational methods that predicts toxicological endpoints or mechanisms. It offers a cost-effective, rapid, and animal-free approach to support toxicological hazard assessments. There are increasing numbers of applications where it has been shown that in silico toxicology can be very useful as well as standardized procedures for performing in silico assessments.1,2 However, as part of the development of a chemical product, when is the best time to perform such assessments?

In early drug discovery, large numbers of compounds are routinely assessed and in silico toxicology approaches can support the prioritization of the most promising candidates as well as decisions on how to progress those compounds. An in silico toxicology profile may help identify which follow-on test(s) to perform and in what order. It may also help to support the experimental design of any tests as well as what to focus on as part of an analysis of the results. For example, a 5-day in vivo study may be run as part of the lead identification and optimization process in which the major organ systems are examined. In silico screens may draw the attention to potential toxicological concerns and help a proper analysis of the experimental signals.3,4

In silico methods are also increasingly being used as part of a weight of evidence (WoE) assessments for major toxicological endpoints. For example, the recently updated ICH S1B guideline5 for pharmaceutical carcinogenicity assessment describes the use of 6 WoE factors – (1) target biology, (2) secondary pharmacology, (3) histopathology chronic studies, (4) hormonal effects, (5) genotoxicity, (6) immune modulation.  An assessment of these factors can be performed throughout the R&D process to support product stewardship.6

  • An assessment of the target biology may be performed before any leads have been identified in order to detect possible carcinogenicity risks from biological pathways related to the target and artificial intelligence supports the analysis of pathways and mechanisms from available databases
  • In lead identification and optimization, in silico methods predictive of WoE factors 2-6, alongside any non-GLP in vitro and 5-day in vivo results can form the basis of a WoE assessment
  • In the pre-clinical phase when many of the GLP test results supportive of WoE factors 2, 4, 5, and 6 are generated, in silico methods may help to fill any data gaps and explain false positive signals in the data
  • In parallel to clinical trials, the results from a 6-month repeated dose study will support WoE factor 3, and used in combination with information generated on the other WoE factors supporting a carcinogenicity assessment document (CAD) to be submitted to international regulatory authorities and in silico analysis may also support a mechanistic understanding of available signals helping the interpretation of the findings and their human relevance

There are situations where the performance of in silico methods has been shown to be fit-for-purpose for specific applications and may be used as a direct replacement. For example, for classification and labelling, in silico methods may provide predictions for relevant endpoints for data-poor substances; in the case of acute oral toxicity, it has been shown that using multiple in silico methodologies that predict this endpoint, in combination with an expert review, is sufficiently predictive. A recent paper showed that for a dataset of approximately 2,000 proprietary chemicals with acute toxicity test results, 95% of chemicals were either predicted in the correct Globally Harmonized System of Classification and Labelling of Chemicals (GHS) category or in a more conservative (i.e., it is protective of health).7

In some situations, there is not a sufficient quantity of the test material available to perform an in vitro or in vivo test. For example, when assessing the potential genotoxicity of impurities or extractables & leachables, there is generally not enough material to perform an Ames test or other relevant tests. Time and cost limitations may also prevent the performance of experimental tests. An in silico assessment offers a sufficiently predictive, rapid, and cost-effective alternative approach which can be generated from the chemical structure alone.2,8,9,10

The following figure summarizes the applications above for pharmaceuticals and illustrates the product cycle phases where in silico toxicology is applicable.

Please get in touch if you would like to discuss any of these topics (Glenn Myatt; glenn.myatt@instem.com).

References

  1. Myatt et al., In silico toxicology protocols, Regul. Toxicol. Pharmacol. 96 (2018) 1–17. https://doi.org/10.1016/j.yrtph.2018.04.014 
  2. Myatt et al., Implementation of in silico toxicology protocols within a visual and interactive hazard assessment platform, Comput. Toxicol. 21 (2022) 100201. https://doi.org/10.1016/j.comtox.2021.100201 
  3. Bassan et al., In silico approaches in organ toxicity hazard assessment: current status and future needs in predicting liver toxicity, Comput. Toxicol. 20 (2021) 100187. https://doi.org/10.1016/j.comtox.2021.100187 
  4. Bassan et al., In silico approaches in organ toxicity hazard assessment: current status and future needs for predicting heart, kidney and lung toxicities, Comput. Toxicol. 20 (2021) 100188. https://doi.org/10.1016/j.comtox.2021.100188 
  5. International Council for Harmonisation of technical requirements for pharmaceuticals for human use  harmonised guideline, Testing for carcinogenicity of pharmaceuticals – S1B(R1), 4 August 2022 https://database.ich.org/sites/default/files/S1B-R1_FinalGuideline_2022_0719.pdf
  6. https://go.instem.com/webcast_Implementation_of_a_Weight_of_Evidence_Carcinogenicity_Assessment
  7. Bercu et al., A cross-industry collaboration to assess if acute oral toxicity (Q)SAR models are fit-for-purpose for GHS classification and labelling, Regul. Toxicol. Pharmacol. 120 (2021) 104843. https://doi.org/10.1016/j.yrtph.2020.104843 
  8. ICH, M7 (R1) Assessment and control of DNA reactive (mutagenic) impurities in pharmaceuticals to limit potential carcinogenic risk, 2017. https://database.ich.org/sites/default/files/M7_R1_Guideline.pdf.
  9. Amberg et al., Principles and procedures for implementation of ICH M7 recommended (Q)SAR analyses, Regul. Toxicol. Pharmacol. 77 (2016) 13–24. https://doi.org/10.1016/j.yrtph.2016.02.004 
  10. Amberg et al., Principles and procedures for handling out-of-domain and indeterminate results as part of ICH M7 recommended (Q)SAR analyses, Regul. Toxicol. Pharmacol. 102 (2019) 53–64. https://doi.org/10.1016/j.yrtph.2018.12.007 

Published by Glenn Myatt

Glenn J. Myatt is the co-founder of Leadscope and currently Senior Vice President, In Silico & Translational Science Solutions at Instem with over 30 years’ experience in computational chemistry/toxicology. He holds a Bachelor of Science degree in Computing, a Master of Science degree in Artificial Intelligence and a Ph.D. in Chemoinformatics. He has published 37 papers, 11 book chapters and three books.