• dosigui
• nano-dosimetry
• nanotechnology

With the term nanomaterial (NM) is possible to define every particle, crystal, fiber, film or composite with at least one dimension smaller than 1 μm. It is important to understand that the nanometric scale, combined with a huge surface area of the resulting constituents, leads to unique properties. On the other hand, one of the consequences of this particular nature is the possibility that NMs exhibit toxicity. Indeed, material properties (both physical and chemical) may radically change when their size is scaled down to the nano-dimension.
Nanoparticles (NPs) are discrete nano-scaled (10e−9m) assemblies of atoms, having all their three dimensions smaller than 1 μm, as defined by Mohanraj et al., (2007). They can be prepared from a variety of materials such as proteins, polysaccharides and synthetic polymers and are very peculiar and studied because the number of atoms in the particles is quite small, and a large percentage of these are located on the surface (or close to it). This significantly modifies the NP’s atomic, electronic, and magnetic structures, physical and chemical properties, and reactivity with respect to the bulk material.
Given that NPs as well as NMs in general provide a particularly useful platform, in recent years a specific scientific and technological area - the so-called nanotechnology (NT), devoted to the production in the nanometer size range - has arisen. Since structures in this scale retain unique properties, this is a promising field with potentially wide-ranging applications, especially in the biomedical area . When studying medical applications, nanotoxicology assessment is crucial. In fact, also for materials whose toxicity has already been thoroughly studied at larger scales, it is not possible to assume the toxic behaviour of a NM referring to its characteristics in other scales (bulk material).
Thus, it is evident that the rapid pace of introduction of new NMs into commerce - increasing human exposure through consumer products - makes accurate dosimetry particularly important, in order to be able to understand the potential health hazards related to these novel materials. To this aim, prior to the assessment of nanotoxicity in vivo, the first step consists in in vitro studies on cell cultures. These experiments allow to obtain a first toxicological screening of NMs, based on the estimation of cell-delivered doses. In vitro tests can be performed on cell cultures composed of cells derived from human cell lines, without having to resort on animals, which , in any case, cannot represent a down-scaled model of the human body. To ensure a meaningful quantification of the dose in vitro, it is necessary to establish standardized methodologies, enabling reliable and repeatable experimental measurements and improving dose-response assessment for NMs.
Given this background, experimental measurements of cellular dose represent one of the main challenge of in vitro nanotoxicology, in terms of both technical difficulties and economic costs. Hence, mathematical modelling of the behaviour of certain particles in in vitro experiments and their compu- tational implementation (the so-called in silico models) are very helpful for predicting the gravitational settling, diffusion and dissolution kinetics (or, in general, particokinetics) of NPs in liquids, as well as their deposition onto cells cultured in monolayer.
Three models for in vitro NP dosimetry were studied in this work:
• One-dimensional Distorted Grid (DG) model
• In vitro Sedimentation, Diffusion and Dosimetry (ISDD) model
• In vitro Sedimentation, Diffusion, Dissolution and Dosimetry (ISD3) model
These represent the state-of-the-art for computational nanodosimetry, but the problem with them is that running several simulations and comparing the results from different models is not an easy task. The user is required to have a good knowledge of the Mathworksà MATLAB computing environment, but, even with good skills in the MATLAB suite, running more than one simulation is a laborious task and it is easy to get into errors, since any action must be implemented via code scripts.
To tackle these problems, a completely new approach to promote the study of NP behaviour in vitro through the DG and ISD3 models is proposed in this work. Such in silico-in vitro pipeline is step-by-step described in the thesis manuscript, which is composed by six chapters, organized as follows:
1. Nanotechnology: overview of the use of nanotechnology and the state-of-the-art of research in this field, and of the issues related to the safety of these new technologies;
2. Standard approaches for dose and toxicity assessment: how these problems are most commonly approached in current research, starting with an introduction to the nanotoxicology problem and to why in vitro and in silico models could be a reliable alternative to animal studies;
3. Computational dosimetry: state of the art: introduction to particokinetics in vitro and to the nanodosimetry with in silico models (ISDD, DG, and ISD3);
4. NP dosimetry with DosiGUI: description of the graphical user interface (DosiGUI) purposely de- veloped to help the user in running simulations, comparing results and visualising output graphs of the models described in chapter 4. DosiGUI is a standalone open-source application that does not require any other software to run (i.e. MATLAB, that is a licensed software), making the use of computational nanodosimetry models easier and widely accessible.
5. DosiGUI validation: finally, to promote the routine use of NP dosimetry models for accurate pre- dictions of effective dose in in vitro systems, an in vitro-in silico pipeline leveraged on the GUI was developed. The pipeline includes methods for validating the models through the use of reference surfaces. Following this pipeline, DosiGUI’s performance was evaluated for three insoluble engi- neered NPs (ENPs) by fitting predictions on data generated from experiments performed with a purely reflective or maximally adsorptive bottom.
6. Identification of stickiness parameters: after the goodness of model predictions for the three different ENPs was assessed, a “stickiness index” (i.e., a parameter which describes the affinity between NPs and cells) was determined for each one of them through the estimation of empirical constants with experiments consisting in measurements of the quantity of ENP mass adsorbed by HepG2 cells over time. Finally, as a proof-of-concept, the pipeline was applied to estimate the effective dose of the three ENPs perceived by HepG2 cells in a standard exposure scenario.
The robustness of the approach was demonstrated for three ENPs (NM-105, NM-212, NM-220), which are reported to have a negligible solubility. Starting from the physicochem- ical characterization reported by Wohlleben’s group, the reliability of simulations performed using DosiGUI was validated experimentally reproducing and predicting the effective dose for two reference boundaries with a totally adsorptive and a purely reflective surface, respectively. This allows identifying the most suitable model to simulate the dynamics of each ENP. For the three ENPs studied in this work, the DG model was the most suitable.
Following this first validation step, parameters describing the adsorption kinetics of ENPs on HepG2 cells were identified, so as to set a realistic stickiness index for simulations. The results confirmed that the boundary stickiness is an ENP-specific feature. In fact, the values of kinetic constants for the three ENPs differ significantly from each other, with NM-220 showing the highest affinity for HepG2. Thus, setting the right stickiness index is a crucial step for accurate effective dose estimation.
Having characterized the adsorptive behaviour of HepG2 cells, DosiGUI was employed to predict the effective dose of each ENP in a standard in vitro test configuration. As expected, only a fraction of the administered amount of ENPs is computed to interact with the monolayer, with a “saturation” effect emerging at high nominal doses. This is because the adsorption of sedimenting NPs is a surface occupancy-driven mechanism. The mechanism is explicitly modelled in DG, which may be the reason why its predictions better correlate with experimental data for insoluble ENPs. As expected, effective dose estimations are higher for ENPs characterized by higher stickiness indices.
In conclusion, this study provides an important contribution to more accurate dose-response character- ization of NPs and to the improvement of safety and risk assessment resulting from the exposure of human tissues and organs to commonly employed ENPs. DosiGUI should facilitate the incorporation of in silico tools in nanotoxicology, encouraging more data sharing, cross-laboratory comparisons and an exhaustive characterization of cell stickiness for a wide spectrum of phenotypes, which could also be integrated as an essential part of existing ENP databases. Finally, the in silico models embedded within DosiGUI could be extended to three-dimensional (3D) configurations for effective dose estimations in in vitro cell aggregates having higher levels of complexity, such as spheroids and organoids. Besides improvements in nano-dosimetry, 3D models could be used to simulate the dynamics of NPs in more physiologically relevant scenarios and for a better in vitro-to-in vivo extrapolation.