Computational tools for the safety assessment of engineered nanomaterials

Abstract

The increasing use of engineered nanomaterials (NMs) in nano-enabled products (NEPs) has raised societal concerns about their possible health and ecological implications. Indeed, despite their clear benefits, NMs may pose environmental, health and safety (EHS) issues. In Europe, the safety of chemicals is subject to the Registration, Evaluation, Authorization and Restriction of Chemicals (REACH) regulation, which requires a thorough Risk Assessment along their life cycles prior to introduction on to the market. However, the heterogeneity of NMs, with respect to physicochemical properties and observed (eco)toxicological effects, makes their case-by-case information gathering for Risk Assessment unsustainable in terms of costs, time, and number of test animals. This has required the development of Integrated Approaches to Testing and Assessment (IATA) in compliance with the 3R (Replacement, Reduction, and Refinement) principles of reducing animal testing to assist industries and regulators in decision making related to the safety of NMs. It is thus essential to make the best use of the available data resources to develop robust in silico methods such as (Quantitative) Structure-Activity Relationships ((Q)SARs) and Grouping for Read-Across models as part of IATA, to inform regulatory Risk Assessment and Safe-by-Design (SbD) decision making. In this context, the EU funded SUstainable Nanotechnology (SUN) project, aimed at combining the bottom-up development of EHS tools, knowledge and data with their top-down integration into a Decision Support System (SUNDS, https://www.sunds.gd/) for risk management of NMs. In this thesis, the SUN project is introduced, and an overview on SUNDS and its structure is provided, focusing on the Human Health Risk Assessment (HHRA) and the Environmental Risk Assessment (ERA) modules, which are evaluated against real-world case studies. Then, the available in silico methods to predict the hazard endpoints are critically reviewed, highlighting their strengths and their limitations and proposing a roadmap for future research in this area, including the adoption of more advanced Machine Learning techniques. Finally, two novel approaches are proposed: the former combines experimental results from simple and fast techniques with multivariate statistical methods to support SbD strategies by highlighting how surface modification can affect the colloidal stability of nanoscale titanium dioxide (TiO2), while the latter uses in a promising way the Subspace Clustering as a tool for the Read-Across and the Classification of NMs, by finding clusters in different Subspaces of the source data, learning a model in these Subspaces, and applying a basic Transfer Learning by projecting the target data on the subspace.