Strategies of development, assessment and application of in silico models for toxicity assessment of chemical structures of toxicological interest

• Ano de defesa: 2024
• Autor(a) principal: Santos, Carlos Eduardo Matos dos
• Tipo de documento: Tese
• Tipo de acesso: Acesso aberto
• Idioma: eng
• Instituição de defesa: Biblioteca Digitais de Teses e Dissertações da USP
• Palavras-chave em Português: in silico models; Adverse outcome pathways; Impurezas não-genotóxicas; Impurity qualification; Modelos in silico; Mutagenicidade; Mutagenicity; N-nitrosaminas; N-nitrosamines; Non-genotoxic impurities (NGIs); Qualificação de impurezas; Read-across; Vias de desfecho adverso
• Link de acesso: https://www.teses.usp.br/teses/disponiveis/60/60134/tde-22052025-095604/

In silico models have gained considerable prominence in the theme of alternative methods and New Methodological Approaches (NMAs) in toxicology. This work proposes new strategies for the development, evaluation, and application of in silico models in assessing the toxicity potential of chemically relevant structures, considering some of the current regulatory challenges regarding pharmaceutical impurities such as Nnitrosamines, Mutagenic Impurities (MIs), and Non-Genotoxic Impurities (NGIs). To achieve specific objectives related to the development and evaluation strategies suitable for in silico purposes, challenges related to data scarcity, low quality or transparency, lack of alignment with scientific problems, interpretability, predictivity, transparency, and gaps, paradoxes, and lack of precision in regulatory criteria that could hinder the application of in silico models were considered. The current status of each context is assessed, proposing new frameworks, methods, and regulatory criteria revisions, along with reflections on potential pathways to overcome related scientific challenges. The presented results suggest new strategies for scientific and regulatory issues, involving in silico models in decision-making and limit definition for substances of unknown toxicity. Considering specific objectives, it is possible to conclude that the proposed evaluation strategy for N-nitrosamines, called DARAN (Defined Approach for Risk Assessment of New Nitrosamines), may be a quick and viable approach to define acceptable limits for new N-nitrosamines without in vivo carcinogenicity data. The evaluation of the compound 1-methyl-4-nitrosopiperazine exemplifies the adequacy of DARAN, proposing an acceptable limit based on hypotheses with interpretability, predictivity, and transparency. The results also suggest that the CPCA methodology of the European Medicines Agency appears to be protective for a significant portion of known N-nitrosamines, providing limits 30 to 300 times lower than those calculated by linear extrapolation of experimental TD50 values. However, potential limitations are also indicated, where CPCA may underestimate the risk for approximately 18% of known N-nitrosamines. In the context of MI evaluation, the research includes a case study on pregabalin lactam (prega-L), following ICH M7 guidelines, illustrating that in the absence of experimental in vitro mutagenicity data, in silico toxicological evaluation based on objective regulatory criteria can address gaps and discrepancies related to these compounds. For the NGI context, the ELPQ framework (Exposure-Led Process for Qualification) is proposed, representing a significant advancement over previous inconsistencies and paradoxes in ICH Q3A/B guidelines. This approach integrates a detailed exposure matrix and incorporates modern principles of Next-Generation Risk Assessment (NGRA), making the qualification process more systematic. Finally, reflections on integrating in silico models with other emerging methodologies in toxicology are concluded, wherein computational tools can play a central role in establishing connections between toxicological data, building new lines of evidence, and strengthening the interpretability and regulatory acceptance of integrated approaches.