Bio-corona: the biological identity of NPs

When nanomaterials are introduced into a living system, they interact with biomolecules (e.g., proteins, lipids, polysaccharides), conferring a “biological identity” to the engineered nanostructures. Indeed, as suggested recently, it is the combination of material-intrinsic properties (the “synthetic identity”) and context-dependent properties determined by the biomolecules of a given biological compartment (the “biological identity”) that will determine the interactions of engineered nanomaterials with cells and tissues and subsequent outcomes.^{Reference Fadeel, Feliu, Vogt, Abdelmonem and Parak9} It is important to differentiate between a “hard” (long-lived) corona of slowly exchanging proteins and an outer “soft” corona collection of (weakly interacting and rapidly exchanging) proteins.^{Reference Walczyk, Bombelli, Monopoli, Lynch and Dawson10} It is generally accepted that while the hard corona is formed because of the direct interaction of proteins with the surface of the NPs, protein–protein interactions instead dominate the interactions of the soft corona with the hard corona.^{Reference Monopoli, Åberg, Salvati and Dawson11} However, changes in the hard corona may occur when NPs are transferred to a new biological compartment. Interestingly, using polystyrene particles, Lundqvist et al. found that while a significant evolution of the corona occurs upon transfer from the first (human plasma) to the second biological solution (cytoplasmic fluid), the final corona contains a “fingerprint” of its history, suggesting that it would be possible to map the fate of NPs based on the analysis of the adsorbed proteins.^{Reference Lundqvist, Stigler, Cedervall, Berggård, Flanagan, Lynch, Elia and Dawson12}

In a recent extension of this work, it was shown that a significant portion of the original corona (derived from the serum in which particles are presented to the cells) is preserved as NPs are trafficked through the cells.^{Reference Bertoli, Davies, Monopoli, Moloney, Gunko, Salvati and Dawson13} In the latter study, the authors showed that the internalized silica-coated magnetite NPs could be isolated from the organelles recovered from the cells following cellular uptake of the NPs, allowing the study of the most tightly NP-bound biomolecules, analogous to the hard corona that has been characterized in extracellular environments. Furthermore, in a recent study, Albanese et al. demonstrated that the protein corona composition on gold NPs undergoes dynamic changes in a conditioned cell culture medium as a result of the secretion of proteins from cells.^{Reference Albanese, Walkey, Olsen, Guo, Emili and Chan14} This evolution of the “biological identity” affected cell uptake and retention of NPs.

The bio-corona of proteins—and lipids—can modify the cellular binding and uptake of NPs. The question, however, is whether the bio-corona is capable of mediating specific cellular uptake, or whether the presence of a corona of proteins, irrespective of their identity, suffices to promote (or prevent) cellular uptake. Lunov et al. evaluated the effect of serum proteins on cellular uptake of carboxyl- or amine-modified polystyrene NPs using THP-1 monocyte-like cells.^{Reference Lunov, Syrovets, Loos, Beil, Delacher, Tron, Nienhaus, Musyanovych, Mailänder, Landfester and Simmet15} They demonstrated that the amount of internalized NPs, the uptake kinetics, and its mechanism were critically dependent on particle adsorption of serum proteins. Similar results were obtained by Lesniak et al., who reported that silica NPs incubated with A549 lung carcinoma cells in the absence of serum have a stronger adhesion to the cell membrane and higher internalization efficiency when compared with NPs with a preformed surface corona.^{Reference Lesniak, Fenaroli, Monopoli, Åberg, Dawson and Salvati16}

Kapralov et al. found that the adsorption of lung surfactant proteins and lipids promoted cellular uptake of single-walled carbon nanotubes (CNTs) by RAW264.7 macrophage-like cells.^{Reference Kapralov, Feng, Amoscato, Yanamala, Balasubramanian, Winnica, Kisin, Kotchey, Gou, Sparvero, Ray, Mallampalli, Klein-Seetharaman, Fadeel, Star, Shvedova and Kagan17} Ehrenberg et al. reported that the capacity of polystyrene NP surfaces to adsorb protein is indicative of their tendency to associate with cells.^{Reference Ehrenberg, Friedman, Finkelstein, Oberdörster and McGrath18} However, removal of the most abundant proteins from the cell culture medium did not result in a differential association with endothelial cells, suggesting that the cellular uptake is not dependent on the specific identity of adsorbed proteins. Caracciolo et al., on the other hand, provided evidence that NPs can acquire a selective targeting capability through the adsorption of a protein corona.^{Reference Caracciolo, Cardarelli, Pozzi, Salomone, Maccari, Bardi, Capriotti, Cavaliere, Papi and Laganà19} They demonstrated that lipid/DNA NPs, upon interaction with human plasma, became coated with vitronectin which, in turn, promoted uptake by cancer cells expressing the corresponding integrin receptor.

In a recent, comprehensive study, Walkey et al. characterized the serum protein corona composition for a library of 105 surface-modified gold NPs.^{Reference Walkey, Olsen, Song, Liu, Guo, Olsen, Cohen, Emili and Chan20} The authors were able to develop a model that uses the protein corona composition to predict cell association (quantified by inductively coupled plasma-atomic emission spectroscopy) more accurately than a model that uses parameters describing NP size, aggregation state, and surface charge. However, the model derived using gold NPs could not be applied to accurately predict the cell association of silver NPs.^{Reference Walkey, Olsen, Song, Liu, Guo, Olsen, Cohen, Emili and Chan20} Nonetheless, the study is important, as it suggests that structure–activity relationships could be developed to relate NP properties (i.e., the synthetic + biological identity) to biological responses in order to predict the biological behavior of different classes of NPs.

The bio-corona has also been observed to modulate toxicity of NPs. Dutta et al. found that pre-coating amorphous silica NPs with a nonionic surfactant (Pluronic F127) dramatically reduced the adsorption of serum proteins and toxicity in the RAW264.7 macrophage-like cell line.^{Reference Dutta, Sundaram, Teeguarden, Riley, Fifield, Jacobs, Addleman, Kaysen, Moudgil and Weber21} Tenzer et al. showed that the formation of a protein corona on NPs affected cellular uptake and endothelial cell death.^{Reference Tenzer, Docter, Kuharev, Musyanovych, Fetz, Hecht, Schlenk, Fischer, Kiouptsi, Reinhardt, Landfester, Schild, Maskos, Knauer and Stauber22} We noted that SiO₂ NP-induced cytotoxicity toward the MCF-7 breast cancer cell line is mitigated in the presence of serum, potentially through masking of reactive surface groups by serum proteins, whereas ZnO NPs, which undergo rapid dissolution, were cytotoxic both in the presence and in the absence of serum.^{Reference Shi, Karlsson, Johansson, Gogvadze, Xiao, Li, Burks, Garcia-Bennett, Uheida, Muhammed, Mathur, Morgenstern, Kagan and Fadeel23}

Importantly, a complete “stripping” of the protein corona may occur in certain environments such as in phagolysosomes due to highly acidic pH causing the native synthetic identity of the NPs to be revealed.^{Reference Cho, Duffin, Thielbeer, Bradley, Megson, MacNee, Poland, Tran and Donaldson24} Wang et al. recently showed, using positively charged polystyrene NPs, that the adsorbed corona is retained on the NPs as they enter cells and are trafficked to the lysosomes.^{Reference Wang, Yu, Monopoli, Sandin, Mahon, Salvati and Dawson25} There, the corona is degraded, and this is followed by lysosomal damage, leading to the release of lysosomal proteases (cathepsins) into the cytosol, and apoptosis. This suggests that the protein corona may shield NPs until they reach the interior of the cell, and also underscores that lysosomal damage is a key issue in nanotoxicology. Indeed, it has been noted that lysosomes—dubbed “suicide bags” by the Nobel laureate Christian de Duve, who discovered these organelles—may play a special role in cytotoxic responses to diverse types of NPs.^{Reference Andón and Fadeel26}

Metal oxide NPs and inflammation

We recently reviewed the literature on the in vitro and in vivo toxicity of a wide range of different types of metal and metal-oxide NPs.^{Reference Karlsson, Toprak, Fadeel, Nordberg, Fowler and Nordberg27} Here, we focus on inflammatory responses to metal oxide NPs.

Cho et al. evaluated the inflammogenic potential of a panel of eight different NPs upon instillation into rat lungs; all exposures were carried out at equal-surface-area doses.^{Reference Cho, Duffin, Poland, Howie, MacNee, Bradley, Megson and Donaldson28} Only CeO₂, NiO, ZnO, and CuO NPs were shown to trigger inflammation at the doses used. Interestingly, each of these NPs induced a different pattern or “footprint” of inflammation both acutely (24 h) and chronically (4 weeks). Acutely, patterns of neutrophil and eosinophil infiltrates differed after exposure to CeO₂, NiO, ZnO, and CuO NPs. The chronic inflammatory responses also differed, with neutrophilic, neutrophilic/lymphocytic, eosinophilic/fibrotic/granulomatous, and fibrotic/granulomatous inflammation being caused, respectively, by CeO₂, NiO, ZnO, and CuO NPs.^{Reference Cho, Duffin, Poland, Howie, MacNee, Bradley, Megson and Donaldson28} As suggested by the authors, the effects could arise from the NPs themselves, from soluble ions released from them, or from both. Notably, no single physicochemical characteristic, including zeta potential (ζP), clearly predicted the inflammogenic potency of the NPs studied. The study thus suggests that metal oxide NPs cannot be viewed as a single hazard entity, and that NPs should be considered separately. The authors also cautioned that conventional in vitro assays might not be able to identify the immunopathological mechanisms for the NPs studied, as the inflammatory response is an orchestrated response that involves several different cell types and soluble mediators.^{Reference Cho, Duffin, Poland, Howie, MacNee, Bradley, Megson and Donaldson28}

Screening of NPs using in vitro assays, however, is an urgent priority; it is not ethically or economically feasible to test all NPs in vivo. To this end, validated in vitro assays are needed. Donaldson and co-workers^{Reference Cho, Duffin, Bradley, Megson, MacNee, Lee, Jeong and Donaldson29} recently performed a study of nine different NPs in order to assess whether commonly used in vitro assays are able to predict lung inflammation in a rat instillation model. They found that most in vitro assays were good at detecting high-solubility NPs that act via soluble ions (e.g., ZnO and CuO NPs), but noted that in vitro assays were of limited use in detecting toxicity of low-solubility NPs, with the exception of the hemolysis assay (destruction of red blood cells). In another study, based on a panel of 15 metal/metal oxide NPs, zeta potential and solubility were identified as two parameters that may have utility in preliminary assessment (screening) of the potential pulmonary inflammation hazard of NPs.^{Reference Cho, Duffin, Thielbeer, Bradley, Megson, MacNee, Poland, Tran and Donaldson24} Specifically, the authors suggested that the zeta potential (ζP) of low-solubility NPs correlates with their inflammogenic potential, while solubility to toxic ions correlates with toxicity for high-solubility NPs. It was suggested that in the phagolysosome under acidic conditions, a high positive ζP may lead to damage of phagolysosomal membranes (see Figure 1).^{Reference Cho, Duffin, Thielbeer, Bradley, Megson, MacNee, Poland, Tran and Donaldson24} The bio-corona of proteins and lipids might alter the surface charge of NPs, but the NPs may revert back to their intrinsic ζP in lysosomes following the removal (“stripping”) of the bio-corona, as discussed previously.

Figure 1.

Hypothetical mechanism of lung inflammation caused by metal/metal oxide nanoparticles (NPs). Positively charged NPs are likely to bind lung lining fluid biomolecules (proteins, lipids) on deposition leading to uptake into macrophages. Based on their comprehensive study of a panel of 15 NPs, Cho et al.^{Reference Cho, Duffin, Thielbeer, Bradley, Megson, MacNee, Poland, Tran and Donaldson24} suggested that acidic pH and lysosomal enzymes in lysosomes would digest these bio-coronas, revealing the zeta potential (ζP) of the low-solubility NPs, which can destabilize lysosomes, resulting in inflammation. On the other hand, high-solubility NPs in the acidic lysosomal environment are thought to cause lysosomal destabilization—and inflammation—dependent on the released metal ions. Reprinted with permission from Reference 24. © 2012 Oxford University Press.

In a comprehensive study using 24 metal oxide NPs, Zhang et al. demonstrated that it is possible to use conduction band energy levels (bandgap) to determine the hazard potential of NPs.^{Reference Zhang, Ji, Xia, Meng, Low-Kam, Liu, Pokhrel, Lin, Wang, Liao, Wang, Li, Rallo, Damoiseaux, Telesca, Mädler, Cohen, Zink and Nel30} Hence, among the materials, the overlap of conduction band energy (E _c) levels with the cellular redox potential was strongly correlated to the ability of Co₃O₄, Cr₂O₃, Ni₂O₃, Mn₂O₃, and CoO NPs to induce oxidative stress and inflammation (i.e., the generation of acute neutrophilic inflammation and cytokine responses in the lungs of CB57 Bl/6 mice). For CuO and ZnO NPs, the acute pulmonary inflammation could not be predicted by E _c levels, but for these NPs, the adverse effects could be explained by their solubility.^{Reference Zhang, Ji, Xia, Meng, Low-Kam, Liu, Pokhrel, Lin, Wang, Liao, Wang, Li, Rallo, Damoiseaux, Telesca, Mädler, Cohen, Zink and Nel30} The study by Zhang et al. is important, as it points toward a predictive nanotoxicological paradigm. This work also implies that the “synthetic identity” of metal oxide NPs is a critical determinant of toxicity and that—despite the possible/probable interaction with various proteins both in vitro and in vivo—the “biological identity” of the NPs did not obfuscate the hazard prediction potential.

Systems biology approaches using omics-based technologies (such as transcriptomics or proteomics) may aid in the delineation of pathways of toxicity of NPs. To this end, detailed in vitro studies of two metal oxide NPs, TiO₂ and ZnO, using primary human monocyte-derived macrophages and dendritic cells (antigen-presenting cells of the immune system), as well as the Jurkat human T cell line, were recently performed.^{Reference Tuomela, Autio, Buerki-Thurnherr, Arslan, Kunzmann, Andersson-Willman, Wick, Mathur, Scheynius, Krug, Fadeel and Lahesmaa31} The TiO₂ NPs were found to be non-cytotoxic to all three cell types at the doses tested, whereas the ZnO NPs displayed a dose-dependent cytotoxicity. Subsequently, global gene expression profiling was conducted to identify transcriptional responses underlying the cytotoxicity of the ZnO NPs. Interestingly, the only genes shown to be upregulated in all three cell types were genes encoding metallothioneins, whereas a notable proportion of the genes were regulated in a cell type-specific manner.^{Reference Tuomela, Autio, Buerki-Thurnherr, Arslan, Kunzmann, Andersson-Willman, Wick, Mathur, Scheynius, Krug, Fadeel and Lahesmaa31} Thus, metallothioneins could be considered as potential biomarkers of the cytotoxic effects of ZnO NPs. Further studies using a panel of surface-modified ZnO NPs with a different degree of dissolution demonstrated that the cellular response to ZnO was largely due to leaching of ions, in line with other previous studies. However, the data revealed different responses at the transcriptional level, suggesting that cell type-specific responses to the same metal oxide NPs exist and also that primary immune cells and immune-derived cell lines may differ considerably in terms of the underlying signaling pathways of toxicity.^{Reference Tuomela, Autio, Buerki-Thurnherr, Arslan, Kunzmann, Andersson-Willman, Wick, Mathur, Scheynius, Krug, Fadeel and Lahesmaa31} Further exploration of the transcriptomics data could suggest novel end-points for toxicity assays for NPs. To facilitate the use of the microarray data from this study and other related transcriptomics studies, the NanoMiner web resource has been created.^{Reference Kong, Tuomela, Hahne, Ahlfors, Yli-Harja, Fadeel, Lahesmaa and Autio32}

Biotransformation: the shifting identities of NPs

Metal/metal oxide NPs may undergo profound chemical transformations in biological environments, and NPs may also form secondarily in a biological system.^{Reference Liu, Wang, Liu, Kane and Hurt33,Reference Wang, von dem Bussche, Kabadi, Kane and Hurt34} Conversely, the dissolution of engineered NPs may drive the toxicity of certain NPs. Indeed, while the bio-corona seems to play a role in cellular recognition/uptake, the toxicity of oxide NPs seems to correlate with their physicochemical identity, a good example being parameters such as zeta potential under acid conditions for low-solubility NPs and solubility (dissolution rate) for high-solubility NPs, as discussed previously. The dissolution of oxide NPs within acidic compartments in the cell may lead to adverse cellular effects. This phenomenon, sometimes referred to as a Trojan horse-type uptake mechanism because it circumvents the plasma membrane barrier and allows toxic ions to “sneak” into cells, has been shown, for instance, for oxides of Zn, Fe, Mn, and Co.^{Reference Stark35}

Gliga et al. recently reported that the cytotoxicity of silver NPs in the BEAS-2B lung cell line correlated with the released fraction of silver, but the released silver in cell culture medium did not induce any cytotoxicity. These results imply that intracellular Ag release was responsible for the toxicity.^{Reference Gliga, Skoglund, Odnevall Wallinder, Fadeel and Karlsson36} The study supports the view that the NPs act as carriers of toxic ions into cells (i.e., as Trojan horses).

Several studies have shown that the dissolution of ZnO NPs is critical to their cytotoxic response. Xia et al. reported that ZnO NPs induced toxicity in both RAW 264.7 and BEAS-2B cell lines, leading to the generation of reactive oxygen species (ROS) and cell death.^{Reference Xia, Kovochich, Liong, Mädler, Gilbert, Shi, Yeh, Zink and Nel37} Using inductively coupled plasma-mass spectrometry and fluorescent-labeled ZnO, the authors concluded that dissolution of the ZnO NPs could occur in culture medium and endosomes. In subsequent studies, the authors achieved purposeful reduction of ZnO NP cytotoxicity by iron doping.^{Reference George, Pokhrel, Xia, Gilbert, Ji, Schowalter, Rosenauer, Damoiseaux, Bradley, Mädler and Nel38} Similarly, decreased dissolution of ZnO NPs by iron doping yielded NPs with reduced toxicity in the rodent lung and zebrafish embryos.^{Reference Xia, Zhao, Sager, George, Pokhrel, Li, Schoenfeld, Meng, Lin, Wang, Wang, Ji, Zink, Mädler, Castranova, Lin and Nel39} Buerki-Thurnherr et al. reported that ZnO NP-induced cell death of the Jurkat leukemic T cell line is largely dependent on the extracellular release of large amounts of Zn²⁺ ions, with their rapid uptake by the cells and the induction of a caspase- and ROS-independent apoptosis pathway.^{Reference Buerki-Thurnherr, Xiao, Diener, Arslan, Hirsch, Maeder-Althaus, Grieder, Wampfler, Mathur, Wick and Krug40} Notably, apoptosis was not affected by overexpression of Bcl-2, a key regulator of mitochondria-dependent cell death. Taken together, intra- or extracellular dissolution may drive cytotoxicity of ZnO NPs, apparently with or without the involvement of oxidative stress.

In a recent study, Li et al. demonstrated that rare-earth oxide (REO) NPs caused cellular and pulmonary damage through a process of intracellular biotransformation in lysosomes.^{Reference Li, Ji, Chang, Dunphy, Cai, Meng, Zhang, Sun, Wang, Dong, Lin, Wang, Liao, Brinker, Nel and Xia41} Using a panel of REO NPs (Y₂O₃, La₂O₃, CeO₂, Nd₂O₃, Sm₂O₃, Eu₂O₃, Gd₂O₃, Dy₂O₃, Er₂O₃, and Yb₂O₃), the authors showed that REO NPs in acidifying lysosomes undergo dissolution and phosphate complexation resulting in organelle damage due to “stripping” of phosphates from the surrounding lipid bilayer. This leads to biotransformation of the NPs into “urchin-shaped” structures, with organelle damage, release of lysosomal cathepsins, inflammasome activation, and IL-1β release from macrophages.^{Reference Li, Ji, Chang, Dunphy, Cai, Meng, Zhang, Sun, Wang, Dong, Lin, Wang, Liao, Brinker, Nel and Xia41} These mechanisms were not described for other oxide NPs (e.g., TiO₂ NPs), which do not undergo biotransformation (see Figure 2). Thus, macrophage uptake with pH-dependent NP dissolution and biotransformation emerges as a novel mechanism of REO NP hazard. The authors also showed that prior phosphate coating at a neutral pH decreases the hazard potential of the REO NPs.^{Reference Li, Ji, Chang, Dunphy, Cai, Meng, Zhang, Sun, Wang, Dong, Lin, Wang, Liao, Brinker, Nel and Xia41}

Figure 2.

Biotransformation of rare-earth oxide nanoparticles (NPs) in THP-1 cells. Li et al.^{Reference Li, Ji, Chang, Dunphy, Cai, Meng, Zhang, Sun, Wang, Dong, Lin, Wang, Liao, Brinker, Nel and Xia41} determined cellular uptake by confocal microscopy and transmission electron microscopy (TEM). For confocal studies, the cells were incubated with FITC-labeled NPs (green) for 12 h, before staining with Hoechst 33342 dye (blue) and Alexa fluor 594-labeled LAMP1 antibody (red) to identify nuclei and lysosomes, respectively. Energy-dispersive x-ray (EDX) was used to characterize the particle transformation (i.e., NP dissolution and phosphate complexation). An asterisk in the EDX spectra represents background elements. Note the spiky, “urchin”-like appearance of the La₂O₃ NPs. The authors also showed that the intracellular biotransformation of the NPs caused lysosomal damage, leading to inflammation. Reprinted with permission from Reference 41. © 2014 American Chemical Society.

Finally, although outside the scope of the present review, it is worth noting that enzymatic biodegradation of carbon-based nanomaterials, including CNTs and graphene oxide, by cells of the innate immune system, is also a form of biotransformation; hence, while complete “digestion” of CNTs may serve to mitigate the (pulmonary) toxicity of these materials, partial biodegradation could result in the production of toxic metabolites, including polyaromatic hydrocarbons, which could mediate toxicity.^{Reference Kotchey, Hasan, Kapralov, Ha, Kim, Shvedova, Kagan and Star42}