ABSTRACT
ZnO dissolution through Fe doping. In another very illustrative example, Zhang et al. [35] were able to demonstrate for 24 metal oxide nanoparticles that it is possible to use conduction band energy levels to delineate their toxicological potential under illumination at the cellular and whole-animal level. The latter study is important because it points toward a predictive toxicological paradigm, at least for one class (metal oxides) of nanomaterials. These examples suggest that high-content and high-throughput screening is a powerful tool for the prediction of nanomaterial hazards. However, the fact that nanomaterials may interfere with commonly used in vitro assays needs to be taken into account [36]. Standardized and validated toxicity and ecotoxicity tests for nanomaterials are therefore needed [37]. In addition to experimental approaches for hazard assessment of nanomaterials, there is a need for in silico methods with which one can develop structure-activity relationships. Indeed, this paradigm allows for predicting of toxicological effects induced by chemicals on the basis of their structural similarity with other chemicals for which toxicological end points have been previously measured. These structure-toxicity relationships can be quantitative or qualitative in nature, and they can predict toxicological effects directly from the physicochemical properties of the entities/chemicals of interest [38]. Therefore, this approach can aid in prioritizing resources in toxicological investigations, while reducing the ethical and monetary costs that are related to animal testing. Quantitative structure-activity relationship (QSAR) models have been successfully applied in the pharmaceutical field and in regulatory toxicology [38] (see Chapter 7 for a further discussion). While QSAR approaches have only recently been used to predict biological effects of nanomaterials, some encouraging initial results have already shown. Puzyn et al. [39] reported on the cytotoxicity of 17 different types of metal oxide nanoparticles to Escherichia coli, and Epa et al. [40] generated quantitative, predictive, and informative models that describe QSAR relationships for cellular uptake and apoptosis induced by metal oxide nanoparticles in several types of cells. Fourches et al. [41] studied the cellular uptake of 109 different nanoparticles with similar cores but diverse surface modifiers. Interestingly, the authors noted that the cellular uptake of nanoparticles possessing the same metal core but different organic molecules on their surface can be predicted by
taking into account the chemical structure of the coating molecules. Therefore, the structural determinants of the biological behavior of nanoparticles can be found both at the core of nanoparticles and at their surface. Indeed, the biocorona is a key determinant of the “identity” of a nanomaterial and should be taken into account when modeling the behavior of nanomaterials in biological systems. To this end, Xia et al. [42] recently developed a biological surface adsorption index to predict the molecular interactions of nanoparticles with proteins on the basis of the competitive adsorbtion of a set of small-molecule probes onto a set of nanoparticles. The measured adsorption coefficients were used to develop descriptors, which, in turn, could be used for in silico modeling to predict the adsorption of small molecules to other nanomaterials. Finally, in a recent study, serum protein corona “fingerprinting” of a library of 105 surfacemodified gold nanoparticles was performed, implicating a set of hyaluronan-binding proteins as mediators of nanoparticle-cell interactions [43]. Notably, the authors developed a multivariate model that used the protein corona “fingerprint” to predict cell association 50% more accurately than a model that uses parameters describing nanoparticle size, aggregation state, and surface charge. 4.3 The Nanomaterial BiocoronaWhen nanomaterials are presented to living systems they interact with biomolecules, including proteins, lipids, and polysaccharides, conferring a “biological identity” to the engineered nanostructures (see Refs. [44, 45] for recent reviews). In the following sections, we discuss the biocorona concept and its implications for the toxicological outcomes of nanomaterial exposure, as well as for clinical applications of nanomaterials, not least for targeted drug delivery. 4.3.1 The Biocorona ConceptThe biocorona is the dynamic layer of biomolecules that adsorbs to nanoparticle surfaces immediately upon their contact with living systems and is presumed to be what organisms or cells “see” and interact with [45]. Thus, as we have suggested in a recent review
[46], and as exemplified by several studies cited in the present chapter, it is the combination of material-intrinsic properties (the “synthetic identity”) and context-dependent properties determined, in part, by the biocorona of a given biological compartment (the “biological identity”) that will determine the interactions of engineered nanomaterials with cells and tissues and subsequent (toxicological) outcomes. Nanomedicines are designed to interact with biological systems at the nanoscale. Thus, to improve biosafety of novel nanomedicines, one must understand the bio-nano interface between these man-made nanostructures and cellular nanostructures or “nanomachines” [47]. Of course, the observation that proteins interact with, and opsonize, particles is not new. Saba and Di Luzio [48] noted, in a study published almost half a century ago on phagocytosis of a variety of particles as a function of opsonization, that “it is now obvious that the removal of colloidal and particulate materials, either inert or viable, is dependent, in part, on such variables as the physico-chemical nature of the particle, particle number, particle size, surface charge, blood flow, opsonic capacity, and species, as well as the functional state of the reticuloendothelial cell.” Indeed, recent studies on nanoparticles have shown that the adsorption of biomolecules onto the surface of nanomaterials is linked to nanomaterial-intrinsic properties, such as size/surface curvature, hydrophobicity, and surface charge [49-51]. In a recent study, the surface topography of colloidal mesoporous silica nanoparticles was shown to play a key role in driving the interactions with proteins [52]. Walkey and Chan [8] compiled data on the plasma-derived protein corona on nanoparticles from 26 published studies and concluded that “the protein corona is complex, that there is no one ‘universal‘ plasma protein corona for all nanomaterials, and that the relative densities of the adsorbed proteins do not, in general, correlate with their relative abundances in plasma” (in other words, there is a degree of specificity). In this context, it needs to be pointed out that there are hundreds of abundant plasma-derived proteins, while depending on their size only a few tens of proteins can adsorb on the surface of one nanoparticle. Thus, there will be a huge variability in the distinct protein coronas around individual nanoparticles, and mass spectrometry-derived compositions of coronas around nanoparticle ensembles describe only the average coverage but not what is
present on each nanoparticle. It was suggested that the protein corona depends on the “synthetic identity” of each nanomaterial [8]. Furthermore, the formation of a biocorona on nanoparticles is a bilateral phenomenon, as proteins that adsorb to nanoparticle surfaces may also alter their behavior as a result of unfolding [53] or fibrillation [54]. Indeed, Mortimer et al. [55] demonstrated that the exposure of cryptic epitopes of albumin upon binding of albumin to layered silicate nanoparticles or “platelets” determines macrophage clearance of these particles via scavenger receptors. Thus, the synthetic and biological identities of nanomaterials are interrelated. Dawson et al. have suggested differentiating between a “hard” (long-lived) corona of slowly exchanging proteins and an outer “soft” corona collection of (weakly interacting and rapidly exchanging) proteins [56, 57]. It has been suggested recently that the difference between the “hard” and “soft” corona might be in first order determined by the dissociation constant of the proteins on the nanoparticle surface [58]. In fact, overall, the literature on the protein corona on nanoparticles is largely based on qualitative data, while more quantitative data on, for instance, dissociation constants, on-and off-rate coefficients, etc., is needed and would further our understanding of the bio-nano interface. Notably, the experimental design, with separation of unbound proteins from protein-nanoparticle conjugates, needs to be considered when decoding the protein corona [58]. Nonetheless, it is generally believed that while the hard corona is formed because of the direct interaction of proteins with the surface of the nanoparticles, instead protein-protein interactions dominate the interactions of the soft corona with the hard corona. It should also be noted that changes in the hard corona may occur when nanoparticles are transferred to a new compartment, for example, upon translocation of nanoparticles into cells, which may influence their toxicity [59].
