A rule for designing safer nanomaterials: do not interfere with the cellular redox equilibrium

• biocompatibility
• mechanistic toxicology
• risk assessment
• safe by design
• computational nanotoxicology

Letter to the Editor

The oxidative stress mechanism is probably the most well developed toxicological pathway for explaining nanoparticle toxicity (Meng et al. 2009) and a number of in vitro models are showing a clear link between exposure to nanoparticles and the generation of oxidative stress (Stone & Donaldson 2006). Under physiological conditions, cells maintain a reduced intracellular state; this process is the result of a balance between the levels of oxidised and reduced species present in the cell. Oxidising substances, like some oxide nanoparticles, can create an imbalance in this state, for example by depleting electrons from biological redox species (e.g. glutathione and cytochrome c). The overall result is a decrease of antioxidants and/or an increase in the production of reactive oxygen species (ROS), a cellular condition which can evolve into inflammatory and cytotoxic responses.

Recently, we have proposed a mechanism-of-action-based model that predicts the ability of oxide nanoparticles to induce oxidative stress (Burello & Worth 2011a, b). To account for this effect, the model characterises the reactivity of the nanoparticles by calculating their energy-band structure and comparing it to the redox potentials of biological reactions that maintain the cellular redox state. When these two energy levels align, electrons can be transferred between the biological material and the oxide nanoparticles, and production of ROS and onset of oxidative stress are eventually observed. Band energies of oxides are calculated from the electronegativity of their constituent cations, from their band-gap energies and points of zero charge (to account for the energy shift according to the Nernstian relation). On the same energy scale, i.e. the absolute vacuum scale, the standard redox potentials of couples active in biological media range from −4.12 to −4.84 eV. The model estimates the position of the valence bands of 70 oxides to be more than 1 eV below the range of redox potentials; however, for some of them, the conduction band is superimposing to that range, indicating that those oxides are able to remove and transfer electrons from the biological material and to potentially induce oxidative stress. This mechanistic approach relies on the principle that redox couples with standard redox potentials near the conduction or valence bands of nanoparticles can exchange electrons with these particles. The electron transfer probability is mainly controlled by energy correlations between the band edges of the oxide nanoparticles and the principal energy levels of the redox couples, °E_red and °E_ox, which represent the most probable energies for the occupied and the unoccupied quantum states in solution.

This model, which was designed for complementing high-throughput screening (HTS) methods, has been validated experimentally and implemented on a HTS platform for determining the toxicity profile of oxide nanoparticles (Zhang et al. 2012). The results of this research study confirmed that it is possible to use our approach to predict the toxicological potential of nanomaterials at both cellular and whole animal levels.

All these information point towards the emergence of an effective rule for designing safer nanomaterials, that is to design nanomaterials that do not interfere with the redox equilibrium of the cell. Although currently there is no clear and structured safe-by-design strategy, at TNO we are developing a number of rules to support the synthesis of safer nanomaterials. The key element is to align functionality and safety, and, in essence, understand how we could modify certain properties which make a nanomaterial appealing for its use – but also possibly hazardous for the environment, health and safety domains, while preserving its functionality. In the case of oxidative stress potential, a number of design features to mitigate toxicity are possible. One could for example play with the size of the nanoparticles, as the variation of this feature is well known to produce a band-gap expansion and therefore a change in the position of the valence and conduction band energy levels (the surface reactivity might however increase due to an increase in the number of surface defects and dandling bonds). If modification of size is not an option, for example because functionality requires that the nanoparticles should have a specific band-gap value for absorption of light at a certain wavelength (e.g. in the case of quantum dots), then one could rely on the use of biocompatible coatings or surface functionalisation, which can mask the surface reactivity of a particle. Coating with antioxidants might also be an option; in this case as ROS are developed on the material's surface they would also be immediately neutralised. Functionalisation with a biocompatible shell material, like silica, is also another way of shielding surface reactivity (and of reducing oxidative stress via the ion leaching mechanism), provided that functionality is not diminished (for example in the case of quantum dots, light adsorption is not diminished if an optically transparent material is used as a shell). If all these approaches might result in a nanomaterial that has lost its functionality, or if direct contact with biological systems is expected to occur only accidentally, controlling the exposure of a nanomaterial by exploiting its aggregation/agglomeration behaviour (e.g. by modifying the hydrophobic/hydrophilic character of the surface or by changing the matrix where the nanoparticles are embedded in) is also another effective option. Eventually, when all options will fail to produce a safer nano-product, researchers will be confronted with the idea of changing the complete structure of the nanomaterial, and move forward by developing new generations of nanomaterials: in this sense the safe-by-design approach will reveal its efficacy to drive innovation.

In conclusion, it is clear that attempts to describe and predict the biological and toxicological effects of nanomaterials are increasingly taking mechanistic considerations into account. While these approaches are still in their initial stages, their further development will allow researchers, in the short term, to streamline and prioritise the hazard assessment on real nanomaterials and, in the long term, to implement safe design strategies in the R&D phase. For example, approaches like the Adverse Outcome Pathway concept (Ankley et al. 2010), where the result of an exposure to a nanomaterial could be described as a sequence of linked events (from a direct molecular initiating event to an adverse outcome at a biological level of organisation), are very useful constructs that help to identify structure–activity relationships and safe-by-design rules, as well as knowledge gaps where further research is required. In the case of safe-by-design, it is clear that changing one design feature does not reflect in a change of one single material's property. When solubility is modified, for example by doping the nanoparticles with additional elements, a change of the particles' electronic structure could also be observed, which might result in an oxidative stress response due to the variation of the band gap or the formation of new energy levels. Similarly, when a coating is applied to prevent ion leaching or surface reactivity, the biological fate of the nanoparticles will also completely change. For this reason a HTS approach would be beneficial to map the entire structure–property relationships space and its relation to toxicity and functionality.

Acknowledgements

We acknowledge the support of the European Commission 7th Framework Programme for the NanoTEST project (Health-2007-1.3-4, Contract no: 201335).

Declaration of interest

The authors report no conflicts of interest in this work.