The Relationships among Structure, Activity, and Toxicity of Engineered Nanoparticles

Abstract

Particles within the nanometer size regime (1–100 nm) exhibit properties across quantum and classical mechanics. An example of the quantum mechanical nature of some particles includes their ability to bend light and change color appearance in suspension. An example of the classical mechanical nature of some particles includes their tendency to agglomerate in suspension. Both of these phenomenons are extensively studied in the literature and are the subject of many research projects that examine the utility of nanoparticles in biomedical and environmental applications. However, these unique properties have also been shown to induce unintentional toxicological effects in various biological and ecological systems. In this paper, the applications and implications of engineered nanoparticles in aqueous suspensions will be reviewed and discussed relevant to the particle’s structure on the nanometer size scale and its subsequent biological activity at the cellular level.

1. Introduction

Engineered nanostructures are ubiquitous in the world around us. Personal interactions with these materials are inevitable, most notably through worker exposure in the chemical industry or consumer exposure through biomedical products. Therefore, assessments of potential risks associated with contact, including toxicological evaluations, are of paramount importance in the safe development of engineered nanostructures. One important factor within the regulatory discussion is the need for accurate and relevant nanomaterial characterization of the physicochemical properties for toxicological evaluations. The toxicity of nanomaterials can be ranked according induced toxicities and inflammation at different concentrations over different time points. Chemical composition and surface coating of the nanomaterial can be related to toxicological responses. Data suggest that users of certain nanomaterials may rely on ex vivo measures of concentration, composition, agglomeration, leached metal ions, and reactive species production, as well as other features such as crystallite phase, surface chemistry, and particle surface charge as indicators for safety assessment.

2. Structure-activity relationships and their relevance to particle science

The model engineered-nanoparticle has three distinct structural features that strongly contribute to its potential biological activity. These common features of a nanoparticle include highly crystalline, monodisperse, and large surface area materials. These features are exploited to gain a unique material-type for a particular application; however, this exploitation comes at a cost in that highly crystalline particles often produce reactive oxygen species when suspended in aqueous solutions, small monodisperse particles distribute within an in vivo model rapidly and distally, and particles with large surface area interact with native biomolecules which in turn disrupt normal biological processes.

Traditionally, structure-activity relationships (SARs) relate the chemical structure of a compound to qualitative biological activity. Toxicologists worldwide face a daunting task in screening hundreds of chemicals suspended in different media for innumerable toxicological endpoints. While in vivo (animal) and in vitro (cell culture) testing are still considered the backbones of obtaining a comprehensive toxicological and risk assessment framework, these traditional methods are both time consuming and expensive. A systematic and organized approach such as computer-based modeling methods that relates experimental data to measurable biological responses is the need of the hour. These resulting “models” will hence be a quick and effective tool in predicting and characterizing a chemical’s toxicity mechanism and consequently integrating physicochemical and biological properties into a strong risk assessment framework for regulatory considerations. Thus, SARs is a multidisciplinary concept and is a powerful link that ties three major fields, namely chemistry, biology and statistics together.

The application of SARs to the field of nanotoxicology in particular seems promising. A host of engineered-nanomaterials is already being incorporated in several consumer products, resulting in large-scale production of these new, innovative classes of materials. Given the complex life cycle of these nano-enabled products, it is essential to predict toxicological effects in a fast and inexpensive manner in order to reduce the risk and improve the benefits of this new technology. Mathematical models can, therefore, be employed to rank, relate, and predict toxicities based on physicochemical nanomaterial properties. This approach has the potential to reduce the need for, or extent of animal testing. Furthermore, these models generate a common conceptual framework for all stakeholders affected by toxicological regulations. Consequently, scientifically rational and informed decisions can be made which leads to a better environmental health protection.

3. Structure, activity, and toxicity of engineered nanoparticles

When designing a nanotoxicology study, a three-step process must be utilized to determine the interaction of an engineered-nanoparticle with tissue or cell. First, the structure of nanoparticle must be determined. Second, the chemical reactivity of the nanoparticle must be understood. Third, the particles’ route of exposure, reaction with the biological target, and resultant cellular dysfunction must be reported. These experiments should be conducted over a range of increasing concentrations as well over a well-defined time period. Table 1 summarizes some of the most common techniques to measure this three-step process.

Table 1 Commonly Studied Physicochemical Properties of Engineered Nanomaterials

Nanomaterial Property	Description	Common Instrument or Measurement Technique
Particle structure	• Size and size distribution • Surface area • Topology • Aggregation/agglomeration state • Shape • Crystal structure and allotropy	• Brunauer, Emmett, and Teller (BET) physisorption for surface area analyses • Nanoparticle Tracking Analysis (NTA) for size and size distribution • Dynamic light scattering (DLS) for agglomeration and aggregation analyses • Transmission and scanning electron microscopy (TEM & SEM) for primary particle size determination • High resolution transmission and scanning electron microscopy (HR-TEM & HR-SEM) for nano clusters • Atomic force microscopy (AFM) for topology • X-ray diffraction (XRD) and electron diffraction (ED) for crystallinity information
Chemical activity	• Stability • Sample purity • Surface characterization • Chemical composition (includes both composition of the particle’s core and its surface) • Therapeutic • Imaging • Functionality • Photo luminescence • Surface charge • Solubility	• UV-Visible spectroscopy (UV-Vis) • Ion selective electrodes (ISE) to measure leaked ions • X-ray photoemission spectroscopy (XPS) to measure electronic state of the elements that exist within a material • Inductively coupled plasma mass spectroscopy (ICP-MS) for trace metal analysis • Fourier-transform infrared spectroscopy (FTIR) to analyze chemical bonding • Differential thermal analysis (DTA) • Energy Dispersive X-Ray spectroscopy (EDS) for elemental analysis and chemical characterization • Zeta potential and isoelectric point for surface charge and stability measurements • Electron spin resonance (ESR) for studying materials with unpaired electrons • Spectrofluorometer for photoluminiscence and quantum yield measurements for fluorescent particles
Cytotoxicity	• Decrease in viability • Alterations in metabolism • Oxidative stress due to reactive species • Trigger apoptosis • Membrane damage (nuclear, cytoplasmic, other)	• Glutathione and oxidation detection • Colorimetric indicators for decrease in metabolism (XTT, MTT) • Live and dead cell counts via fluorescent dyes • Leaky membrane LDH release • Protein, cytokine, and chemokine expressions (ELISAs, western blotting) • Damage and repair to DNA (TUNEL)

3.1 Structure

3.1.1 Surface area

Brunauer, Emmett, and Teller (BET)—the original authors of the 1938 article describing this technique-physisorption is a method of measuring surface area and porosity of a material by physically adsorbing molecules of nitrogen gas (N₂) onto the surface of a solid. BET is one of the most widely used methods of determining surface area. It is accomplished by designating particular points at which relative pressure measurements are to be taken—pressure of a tube containing sample measured against saturation pressure tube. BET Surface Area measurements are used in the pharmaceutical, ceramics, carbon black, fuel, and electronics industry, as well as being a key physical characterization method for nanoparticles. The BET Surface Area of various building materials were measured by passing Nitrogen gas over a “degassed” solid at the boiling temperature of liquid nitrogen. The building materials were kept as close to their native state as possible however “coupons” had to be made in order to fit them through a 1/4 inch diameter opening on the sample tube. The instrument should be installed on a stable tabletop and gas tanks should be secured. Also do not kink the gas lines or damage the integrity of the lines lead to the machine. The operator is responsible for proper cleaning of the sample tubes for each use. The operator should report mass values to a level as precise as possible. The operator is responsible for identifying a degas procedure. The operator should ensure that no test sample is sucked up into the degas or analysis ports when removing sample tubes after degassing and analysis. Unused gas ports should be covered or plugged when not in use. Liquid Nitrogen can cause cryoburns, wear proper PPE when filling up liquid nitrogen Dewar for use in analysis. Allow sample tubes to warm to room temp after completion of the analysis run. Degas temperature can get quit high (upwards of 350°C), allow sample tubes to cool down before touching and removal of the heating jackets.

There are thermodynamic consequences of high surface areas. Difference in free energy between bulk material and nanoparticle contribute to the difficulty in predicting unknown effects of a nanoparticle based on known effects of larger micrometer sized materials. Specific to nanocrystals, reduced melting points, large surface areas, and increased lattice contraction give nanoparticles the heightened ability to produce reactive oxygen and nitrogen species on the particle’s surface. While crystals on the larger size scale are also known to produce reactive species, some nanocrystals have different crystallographic structures when compared to their bulk phase counterparts.

3.1.2 Dynamic light scattering

Dynamic light scattering (DLS) is also referred to as Photon Correlation Spectroscopy (PCS). The technique is used for measuring particle size and size distribution when in liquid suspension. DLS measures the Brownian motion of the particles. Brownian motion is the random movement of suspended particles due to bombardment between the particle and the solvent molecules. As the solvent molecules collide with the particles, they exert a force, which is capable of moving the particles in a random direction. This random movement is more exaggerated in a small particle than for a large particle of the same composition in the same time period. This movement is known as the translational diffusion and is captured in the Einstein-Stokes equation as D along with the hydrodynamic radius of the particle.   

$$d(H)=kT/3\pi nD$$(1)

where D is the translational diffusion coefficient, k is Boltzmann’s constant, T is temperature, and n is the viscosity of the solvent. The Stokes equation is only applicable for spherical or spheroidal particle-types.

Measuring the size of particles when suspended in aqueous solvent is an important characterization metric for nanotoxicology studies. It provides the researcher with a rapid means of generating a size profile (or hydrodynamic radius) of nanoparticles. DLS is also used to measure proteins in a biological buffer. The technique does have some limitations. For example, in a polydisperse sample, the larger particles will scatter more light, thus skewing the intensity of the size profile. The intensity of scattered light is proportional to the diameter of the particle to the 6th power. This phenomenon is called Rayleigh Scattering. It is difficult to determine the size of individual particles using light scattering when aggregates or agglomerates are present in the suspension—a polydisperse sample mixture may read as if it were void of the smaller fraction of nanoparticles. It is important to use control samples without the nanoparticle to determine the sizing profile of components in the suspension medium.

3.1.3 Transmission electron microscopy and energy dispersive spectroscopy

The use of transmission electron microscopy (TEM) is a necessary characterization tool in nanotoxicology. TEM takes advantage of the extremely small wavelength of electrons (0.2 nm) to increase the resolution and magnification of nanometer sized particles as well as cellular organelles. In transmission electron microscopy, the sample, generally referred to as the specimen, is exposed to an electron beam. The transmitted electrons are viewed on a phosphorescent screen, camera, or film. Depending on the specimen characteristics, the TEM can provide information such as size, shape, aggregation, and crystalline structure. This technique may be used post-exposure for both in vitro and in vivo tissue samples to determine the intracellular localization of particles and cell structure changes (Reynolds, 1963). Table 2 summarises the comparisons of microscopy methods used in nanotoxicology.

Table 2 Comparison of Microscopy Methods used in Nanotoxicology

	Molecular structure	Sizing limits	Aggregation State (in liquids)	Composition determination	Generation of artifacts	Special features
Optical	No	> 250 nm	No	Staining	Low	Real time imaging
SEM	No	> 30 nm	No	EDX	High	Topography of complex shapes
TEM, dried	Yes	> 0.5 nm	No	EELS	High	Cross-sectional details
TEM, cryo	No	1.5 nm	Yes	None	Medium	Preserves the native state of the sample

TEM provides information about a nanomaterial in its native state or in a complex mixture. However, there are some limitations to this technique. Biological specimens are immediately damaged upon exposure due to the high energy of the electron beam. Proper biological specimen preparation determines the quality of the specimen. Preparation is a multiple step process consisting of fixation, dehydration, embedding, and sectioning. Post-staining the specimen with a combination of uranyl acetate and lead citrate will enhance contrast and allow better visualization of cell structure. The combination of TEM with composition determination is an even more powerful technique. Complementary methodologies such as energy dispersive spectroscopy (EDS), electron diffraction (ED), and high resolution transmission electron microscopy (HRTEM) are useful. Energy dispersive spectroscopy as a micro-analytical tool can yield both quantitative and qualitative results. When emitted electrons strike the atoms of interest in the sample, inelastic reactions generate emitted electrons and X-rays that are then detected via spectroscopy. The technique is very sensitive and, therefore, special attention must be placed on sterility during specimen preparation as to not add artifacts to the sample.

3.1.4 Crystallinity

X-ray diffraction (XRD) is one of the most commonly used methods for confirming crystalline product. Limitations of the technique include large sample amount needed (100 mg powder sample). While small grain sizes can be determined, the technique requires long collection times. The chemical composition influences the crystalline structure, as does synthesis (and sometimes storage) method.

Two equations are used in this technique: Bragg’s law and Scherrer equation. Bragg’s law is used to determine the angles for coherent and incoherent scattering from a crystal lattice.   

$$\text{n}\lambda =2d\hspace{0.17em}\text{sin}\hspace{0.17em}\theta$$(2)

where n is an integer, λ is the wavelength of incident wave, d is the spacing between the planes in the atomic lattice, and θ is the angle between the incident ray and the scattering planes.

Scherrer equation is used to determine the size and shape of the nanocrystalline material, also known as crystallite.   

$$D(\mathring{\text{A}})=0.89\lambda /\beta \hspace{0.17em}\text{cos}{\theta }_{\text{B}}$$(3)

where D is the crystallite size, λ is the X-ray wavelength (1.54 Å for Cu Ka radiation), θ_B is the Bragg angle, and β = (B²−b²)^1/2 where B and b are taken as the FWHM (full width at half max peak) of standard and sample.

Electron diffraction (ED) is the phenomenon that results in an interference pattern when a beam of electrons bombard a sample. It is a commonly used technique to study the crystal structure of solids, especially in material characterization techniques such as transmission and scanning electron microscopy. Based on the diffraction patterns, the crystallinity of the sample under investigation can be determined. Once this pattern has been attained, it is compared with that of a standard material. This technique may come into play when a particle has multiple crystalline states (Sayes et al., 2006a). HRTEM is also used to study the crystallographic structure at the atomic scale.

Recent studies have related the crystal structure of a material to its photocatalytic and cytotoxicological activities (Sayes et al., 2006b). Different crystalline phases of nanoscale titania (TiO₂) particles contribute to their differential cytotoxic responses in cultured cells. The most inactive catalytic material (TiO₂ in the rutile crystal phase) is less cytotoxic than similarly sized TiO₂ particles in the anatase crystal phase. Anatase TiO₂ also exhibits high photoactivity. This correlation is a manifestation of a fundamental structure-activity relationship in nanoscale titania—nanoparticle structures optimized to produce reactive species under UV illumination also are more effective at disrupting cellular functioning. Simple ex vivo tests of nanoscale titania photoactivity could prove useful as a comparative screen for cytotoxicity in this important class of materials.

3.2 Activity

3.2.1 Zeta potential

The zeta potential of a particle is used as a measure of the particle’s surface charge. In actuality, the zeta potential is a measure of the electric double layer and is generated from oriented solute molecules and ions surrounding the nanoparticle surface. The electric double layer is dependent on it environment and will change intensity if pH or ionic strength is altered. For example, all nanoparticles have an isoelectric point. The particle’s isoelectric point is the pH of the suspension when the particle’s zeta potential is zero net charge (Berg et al., 2009).

Zeta potential is also used as a measurement of colloidal stability (Harush-Frenkel et al., 2007; Zhang, 2009). Values above ±30 mV are considered a stable system while values below ±30 mV are deemed unstable suspensions and aggregate readily. The benchmark time scale for this assessment is generally 30 days. This aggregation is due to the lack of charge-charge repulsion between individual nanoparticles.

3.2.2 Photocatalytic degradation

The surface of a nanoparticle has been identified as a source of leached ions, molecules, and reactive species. There are a variety of techniques and an array of colorimetric probes that are used to measure the reactivity of the surface of a nanoparticle that yields these species into the surrounding particle matrix. Vitamin C and Congo Red are probes used to measure the photocatalytic degradation of the particle’s surface. When identifying and quantifying the reactivity of a particle’s surface, it is important to recognize that no single technique can be used for all types of nanoparticles. Results in the literature indicate that the mechanism for color change in these tests is a result of charge transfer complexes between the probe and the “active sites” on the particle surface (Warheit et al., 2007a).

3.2.3 Spectroscopy

One of the most common analytical techniques in nano-science is spectroscopy. Spectroscopy is defined as the measurement and interpretation of changes in the electromagnetic spectra arising from either absorption or emission of radiant energy by a sample. Spectroscopic techniques are especially useful for particle analyses. Information about the chemical composition, structure, surface functionality, and optical/electronic properties of the sample can be obtained. For example, mass spectroscopy is used to determine the masses of small electrically charged particles and can be utilized to measure properties proteins absorbed onto the particle surface. Raman spectroscopy is used to study vibrational, rotational, and other low-frequency modes in a system and can determine the type and degree of functionalization on the particle surface. Absorption spectroscopy is used to quantify the amount of photons a substance absorbs and can be utilized to measure the size of nanoparticles. Fluorescence spectroscopy is used to analyze the different frequencies of light emitted by a substance which is then used to determine the structure of the vibrational levels of that substance.

3.3 Toxicity

3.3.1 Concentration

Of all the physicochemical properties associated with nanomaterials, the feature that is most easily correlated to toxicity is concentration or dose. However, to measure the concentration of a sample is non-trivial. Three techniques are usually employed to determine concentration of a nanoparticle in a suspension. Gravimetric tests can be used, but must assume particle composition. Methods based on chemical reactant reactions can be utilized, but can produce erroneous results. Transmission electron microscopy of cryogenically frozen samples in suspension is the best measure of particle concentration. There are a variety of factors that influence the accuracy of measurements of concentration, such as the purity of sample and quantity of material available for analyses. Special considerations must be taken when a nanoparticle is surface functionalized with an organic coating. For example, surfactant coatings are not stable, and much like a protein absorbed onto the particle’s surface, the surfactant will dissociate and be replaced with other molecules present in the suspension.

3.3.2 Exposure to nanomaterial samples

Real-world exposure to engineered nanomaterials is very likely to be a mixture of multiple substances, such as airborne particulate matter. Epidemiological data had suggested that airborne particulate matter (a mixture of many different substances) was more toxic than the sum of its parts. This was attributed to the “synergistic” effect due to the mixture of critical components within airborne particulate matter. In the nanotoxicology environment, human exposure to mixed nanomaterials is likely because multiple types of nanomaterials are often used simultaneously. For example, both metal oxide nanoparticles (e.g., Fe₂O₃) and graphitic carbon nanoparticles (e.g., carbon black) are likely to be used in various industrial applications. Therefore, it is imperative to understand the toxicological effect of exposure to mixed nanomaterials as well as the effects of the individual components alone. Recent publications have highlighted the importance of studying multiple nanoparticle exposure scenarios (Berg et al., 2010).

3.3.3 Hemolytic potential

The hemolytic potential of a nanoparticle sample gives vital information on the effects of the particle on a biological membrane. The assay has been used in the past century with a variety of particulates including silicate powders, asbestos, and more recently a multitude of nanoparticles. Hemolysis (i.e. rupturing of erythrocytes or red blood cells) is an ex vivo characterization method that is complementary to assessing cell death. The ability of nanoparticles to react with a biological membrane is an important indicator especially for particles incorporated into medicinal devices for therapy or diagnosis (Warheit et al., 2007a) (Harington et al., 1971; Nolan et al., 1981; Warheit et al., 2007b).

Particulates, such as nanomaterials, interact with the erythrocyte and exert their cytotoxic effect via a multitude of hypothesized mechanisms depending on the characteristics of the particle such as size, composition, surface functionalization, and zeta potential. The mechanical interaction between the particle and the membrane can cause rupture releasing the intracellular hemoglobin into the ambient solution thus turning the solution pink to red in color. In addition to mechanical interactions, nanoparticles may react with the membrane of the erythrocyte indirectly via actively oxidizing the membrane molecular structure thereby increasing the membrane permeability. The particles may alter the ionic composition of the suspension medium through particle surface degradation, which could then increase the amount of hemoglobin in surrounding medium, as well.

3.3.4 Oxidative stress indicators

The presence of excessive reactive oxygen species (ROS) can cause cellular oxidative stress, which may lead to sub-cellular damage such as DNA, RNA, protein, mitochondrial, and membrane or other lipid degradation (Sayes et al., 2004; Fortner et al., 2005; Lyon et al., 2005; Sayes et al., 2005; Sayes et al., 2006). Exposure to nanoparticles has been shown to produce excess ROS. Consequently, oxidative stress damage has become a critical assessment tool in characterization the physical, chemical, and toxicological properties of engineered nanostructures.

In nanotoxicology, DNA oxidative stress is used as an indicator for potential cytotoxicity (Cattley and Glover, 1993; Hwang and Kim, 2007). Research has shown that different types of DNA damage and repair after oxidative damage can provide insights into mechanisms of action (Cooke et al., 2003). DNA damage is a result of a break or change in the chemical sequence of DNA in the nucleus of a cell. Each type of DNA lesion has different affects on the cells. For example, some DNA lesions result in mutations and can affect DNA replication and transcription; other types of DNA damage are repairable (Cooke et al., 2003). There are numerous biomarkers that measure oxidative stress endpoints in cells, urine, blood, and other tissues. The assays capable of determining DNA oxidative stress, damage, or repair after damage usually measure levels of 8-hydroxyguanosine (8-OHG) and 8-hydroxydeoxyguanosine (8-OHdG) nucleosides (Cattley and Glover, 1993; Hwang and Kim, 2007). The Comet assay measures general DNA damage via gel electrophoresis of cell lysate, followed by fluorescence signal quantification.

Protein oxidation is the process of oxidative damage to polypeptides and amino acids present in cells. Examples of assays that measure oxidative protein damage are protein carbonyl content (PCC) and 3-nitrotyrosine. Techniques used to evaluate the oxidant-antioxidant biomarkers are enzyme-linked immunosorbent assays (ELISA), ion exchange chromatography, immunoblotting, and electron paramagnetic resonance imaging. Several mechanisms are known that describe how electophiles and pro-oxidants cause protein damage. Alterations in protein function can occur following formation of covalent bonds between electrophiles and nucleophilic amino acids, oxidation of nucleophilic amino acids, or production of reactive nucleophiles.

Oxidative stress can also be caused by the accumulation of nanoparticles within the mitochondria (Nicholls and Budd, 2000) (Oberdorster et al., 2005; Li et al., 2008) (Sayes et al., 2004). It has been demonstrated that nanoparticles of various size and chemical composition are preferentially transferred to and accumulate in the mitochondria (Li et al., 2008). ROS from the nanoparticle surface at the site of the mitochondria or ROS produced by the mitochondria because of nanoparticle exposure leads to cell damage or death. In structurally intact mitochondria, the production of ROS is balanced by an extensive antioxidant defense system that works to detoxify the oxygen radical generation (Nicholls and Budd, 2000). This protection is generally provided via glutathione and superoxide dismutase. Heightened cellular production of ROS occurs after damage to the mitochondria; thus, oxidative stress is often detected by measuring mitochondrial function.

Lipid peroxidation and membrane damage are used as indicators of oxidative stress and cellular damage. The major concern in this type of oxidative damage is disruption the ion channel flow. Lipid peroxides are unstable and degrade to molecules such as reactive carbonyl compounds. Polyunsaturated fatty acid peroxides generate malondialdehyde (MDA) and 4-hydroxyalkenals (HAE). MDA and HAE are colorimetric/flourometric indicators for this type of oxidation. Cytoplasmic and other cellular component membranes are easily targeted due to the amount of lipids found within the membrane.

Fluorescent dyes are very good indicators to detect oxidative stress in cells and tissues (Wang and Joseph, 1999; Lin et al., 2006). These ROS-sepcific dyes are categorized according to which type of ROS being probed. In general, there are multiple fluorescence probes for the detection of superoxide radical (O₂^•−), hydrogen peroxide (H₂O₂), singlet oxygen (1O₂), hydroxyl radical (HO^•), or peroxy radical (ROO^•), as well as others (see Table 3). As with any colorimetric or fluorescent indictors, however, special attention must be placed on whether or not the nanoparticle auto fluoresces at the same wavelength or interferes with dye to avoid false positives or false negatives in results. There is a wide range of applications for these dyes within the realm of detecting ROS. For example, some probes are mainly used in an indirect approach to probe for antioxidant capacity, while others probe for oxidation in particular sub-cellular compartments and membranes (Gomes and Fernandes, 2005).

Table 3 Examples of fluorescent ROS dyes

Fluorescent probe	Excitation/emission wavelengths (nm)	ROS species detected
DCFDA (2,7-dichlorodihydrofluorescein)	498/522	H2O2 ROO• HO•
DHE (Dihydroethidium)	520/610	O2•−
B-TOH (BOPIDY-α-tocopherol)	514/565	ROO• RO• HO•
HPF (Hydroxyphenyl fluorescein)	485/535	HO•
CM-H2DCFDA (chloro methyl dihydro derivative of fluorescein)	485/538	H2O2 ROO• HO•

In addition to fluorescence spectroscopy, confocal microscopy can be used as a tool to probe stress (Uggeri, 2004). Oxidative stress can induce a dose-related increase or decrease in the level of fluorescence depending on the fluorescent probe of choice. In confocal microscopy, fluorescent probes are used to stain cell structures. The fluorescent probes dihydrorhodamine 123, dihydroethidium, 2’,7’ dihydro-dichlorofluorescein (H2DCF), and dichlorodihydrofluorescein diacetate (H2DCFDA) are some of the most commonly used detectors of reactive oxygen species (ROS) in spectroscopy and microscopy (Uggeri, 2004).

3.3.5 In vitro toxicology

The successful development of in vitro assays as predictive screens for assessing particle toxicity is an important goal during early product development. Ideally, the screen would be utilized prior to more substantive in vivo toxicity testing. If properly validated, the advantages of these early screening tests would be evident. In vitro testing is simpler, faster, and less expensive than in vivo assays. Successful development of in vitro toxicity models should reduce the use of animals for hazard identification of potential toxicants as well as to improve the ability to evaluate new materials in small quantities. Although implementation of in vitro toxicity screens would be beneficial, the accuracy and predictability of these tests with particulates have shown high correlative evidence when compared to results obtained in vivo models.

An interne national panel of scientists recommended five grand challenges associated with the safe handling of nanotechnology. The panel has recommended strategic research strategies to support sustainable nanotechnologies by maximizing benefits and minimizing environmental and health risks (Maynard et al., 2006). One of the five key grand challenges cited was to develop and validate alternatives to in vivo toxicity testing of engineered nanomaterials, recognizing that there would be several limitations inherent to in vitro assays and cell culture systems when simulating complex biological effects after exposure to nanoparticles. These limitations include particle dose, selection of cell-type for simulating the organ system of choice, characterization of exposure, effects over time, and appropriateness of endpoints for hazard evaluation (Sayes et al., 2007).

In recent study, a number of variables which strongly impact the ability of in vitro screens to accurately reflect in vivo pulmonary toxicity of several particle types in rats were systematically evaluated (Sayes et al., 2007). The variables tested included particle dose, time course, duration of treatment exposure, and pulmonary cell types. In vivo effects in the lungs of rats were utilized as the barometer for comparison. Under the conditions of this study, the results of in vivo and in vitro cytotoxicity and inflammatory cell measurements demonstrated little correlation when single cell cultures were used, but resulted in a higher level of correlation when co-cultures of epithelial cells and macrophages were used together. There is a need for better culture systems and for more research to designate the most relevant biological endpoints to test for in vitro studies (Romoser et al., 2011; Romoser et al., 2012; Berg et al., 2013).

3.3.6 Cellular uptake

Interaction of engineered nanoparticles occurs on a size scale previously seen with the entrance of viral pathogens across the cytoplasmic membrane (Oberdorster et al., 2004; Larese et al., 2009). The uptake of nanoparticles provides a unique way of accessing intracellular compartments through intentional (i.e. drug delivery) or unintentional (i.e. occupational or consumer exposure). In either scenario, nanoparticles are capable of intracellular localization through a variety of endocytic mechanisms ranging from the clathrin-mediated endocytosis (CME) to non-specific macropinocytosis or simple passive diffusion (Qaddoumi et al., 2003; Geiser et al., 2005; Rothen-Rutishauser et al., 2006; Harush-Frenkel et al., 2007; Harush-Frenkel et al., 2008; Jiang et al., 2008). Many physicochemical characteristics such as nanoparticle size, surface coating, and surface charge have been postulated as factors influencing the endocytosis of nanoparticles. Many endocytic mechanisms are structurally size limited by the protein scaffolding lining of the vesicle on the cytoplasmic membrane. Calculations based on thermodynamic limitations and cell-ligand interactions have calculated the most efficient receptor-mediated nanoparticle endocytosis occurs when the particle is a spherical is shape and 30 nm in diameter (Freund and Lin 2004; Gao et al., 2005). When the particles aggregate, however, the resultant agglomerate is treated as larger particle. In these circumstances, phagocytosis by immune-modulator macrophages is responsible for clearance of agglomerates.

While primary particle size and agglomeration state play the primary role in directing the specific route in nanoparticle endocytosis, they are not the only factors. Surface charge also plays a role in the interactions between nanoparticles and cell membranes (Harush-Frenkel et al., 2008; Zhang, 2009). For example, negatively charged particle surfaces, such as particles coated with hydrophilic acids, and positively charged particle surfaces, such as particles coated with polymers, are more capable of incorporation into human cells than particles with little to no surface charge. The charge of a nanoparticle will also dictate the pathway through which the particle is internalized (Harush-Frenkel et al., 2008; Mayor and Pagano, 2007; Doherty and McMahon, 2009; Gould and Lippincott-Schwartz, 2009; Zhang, 2009). For example, while uptake of both positively and negatively charged particles are energy and F-actin polymerization dependent, they have been shown to undergo different pathways. Positively charged nanoparticles enter cells through macropinocytosis and CME (Qaddoumi et al., 2003; Chung et al., 2007). Negatively charged nanoparticles enter cells through dynamin-independent mechanisms (Qaddoumi et al., 2003; Schroeder et al., 2007; Partha et al., 2008; Villanueva et al., 2009). While associations between charge and mechanisms of cellular internalization have been made, further research into this topic is needed.

In addition to nanoparticle charge, a modified surface coating influences the uptake by cells. Another factor dictating nanoparticle cellular uptake and potential toxicity is the adsorption of proteins onto the surface of the particle (Chen et al., 2005; Chen et al., 2006; Huang et al., 2007; Lundqvist et al., 2008; Schellenberger et al., 2008). Proteins have been shown to coat particles when placed in cell culture media, serum, and biological fluids. The extent of the protein coating is dependent upon physical factors such as size and shape and chemical factors such as charge and other surface functionalization. Antibody conjugation onto the surface of a particle increases linearly with respect to the radius. This association between size and protein binding is due to the high rate of curvature exhibited by the surface of a small nanoparticle which sterically hinders the binding of additional antibodies. Protein adsorption may be dramatically altered as any single physicochemical property is changed—thus prohibiting many mathematical models to be used in this field. It is this protein adsorption theory, which necessitates the importance of proper characterization of nanomaterials, as well as puts forth complicating task of linking nanomaterial physicochemical properties with their biological endpoint.

4. The phases of nanomaterial characterization within a study design

The most thorough nanotoxicology studies include a thoughtful experimental design for the material characterization. The nanoparticle sample should be characterized and re-characterized for physical and chemical properties throughout the course of the designed study. When changing the matrix surrounding the material or simply measuring changes over time, a nanoparticle sample should be analyzed and labeled, as material characterization is the primary phase, the secondary phase, or the tertiary phase (Fig. 1).

Fig. 1

The Physicochemical Characterization Required for Toxicological Evaluations.

Primary phase physical and chemical characteristics refer to those nanoscale properties immediately after the material is synthesized using liquid-phase synthesis or aerosolized synthesis. Furthermore, the characterization should be performed on both lab-synthesized and purchased material. Those properties relevant to toxicity testing include a diverse range of particle features, including: chemical composition, surface functionalization, surface area, size, size distribution, shape, crystallinity, and purity. Secondary phase characterization relevant to toxicity testing includes aggregation/agglomeration/coagulation, concentration, activity/reactivity, presence of reactive species, surface modifications, and leached ions or molecules. Properties in the tertiary phase that are relevant to toxicity testing include cellular uptake and tissue distribution, adsorption of molecules onto the surface, aggregation, degradation and decomposition, and dissociated ions or molecules.

There are specific characteristics regarding nanoparticle production that are most relevant to toxicologists. These issues include quantity of material produced and availability for exposure, particle composition and amount of impurities, and presence and identity of a surface coating. When performing toxicological evaluations on nanoparticles, these properties should be addressed.

One of the most frequently asked questions in nanotoxicology is: to what extent do particles aggregate (Sager et al., 2007; Murdock et al., 2008)? Particles are always attracted to each other. Forces such as van der Waals interactions, type of solvent, and Hamaker constant are all contributors for particles to aggregate. It is noteworthy, however, that aggregated particles, may not have significantly different surface areas compared to individual particles. Nanoaggregates are often referred to as “soft” aggregates. Chemical changes are needed for aggregates of nanoparticles to become permanently fixed to each other. Factors such as direct covalent attachment at grain boundaries require chemical reaction between particles. When nanoaggregates become fixed, they are often are referred to as “hard” aggregates. It is this type of aggregation (or agglomeration) that leads to much lower surface areas. Nanoparticles can be stabilized against aggregation. For example, ionic strength alters repulsion between particles by several orders of magnitude). In ionic stabilization, strong repulsive electrostatic forces are present and are dependent on the type of solvent used in the system. The strength of these electrostatic forces is affected by surface charge density. If the density of the surface charge is large in magnitude and equivalent in sign, then the repelling forces on a particle-to-particle basis can operate in relatively long ranges. Another example of prohibiting aggregation is steric stabilization. In this case, the coatings on surfaces of the nanoparticles define the extent of stabilization. Understating the interactions of the coatings with the solvent essential and ultimately can reduce van der Waals interactions.

These principles apply to nanoparticles. When characterizing a nanoparticle for any type of assessment, it is important to consider an engineered nanomaterial as a particle obeying quantum mechanics (Fig. 2). When the particle is aggregated into larger agglomerate, then the particle may be obeying classical mechanics. Nonetheless, particles on the nanometer size scale are different from particles on the micrometer size scale (Nel et al., 2006).

Fig. 2

The Properties Exploited in Specific Particle Size Ranges.

5. Conclusion on utility of structure-activity relationships for nanoparticles

The role of engineered nanomaterials in a variety of fields in science and engineering requires an adequate methodology for evaluating their potential impact on safety and health. Recent publications point to an increase of various biological responses, such as inflammatory or immune responses, after exposure to nanomaterials. These biological responses have been observed and reported on from either in vitro (cell culture based) or in vivo (whole animal) model systems; however, not enough effort has been placed on relating the physicochemical features of each engineered nanomaterial to the observed biological responses. Developing experimental models and methodologies that not only establish relationships between different physical structures and biological activity but also can reliably predict biological responses to nanomaterials is critical. This effort, however, requires a large and organized effort within the nanoscience community. Many data sets measuring multiple endpoints are required in order to accurately predict biological responses. The same is true for the physicochemical characterization.

The structure of a nanoparticle can be used as one of many predictors of biological responses. However, it must be noted that structure is not the only physicochemical feature that contributes to a material’s potential toxicity. For example, not all nanoparticles are more toxic than their larger micrometer sized counterparts. Size, alone, is not a measure of toxicity. Surface coatings of particles have been shown to more likely influence toxicity—no matter the size or structure of the material. And particle aggregation and subsequent de-aggregation complicates the structure-activity relationship even more since this process is dynamic depending on time, suspension matrix, and temperature.

Using both analytical chemistry and biochemistry approaches, a systematic methodology is under development that will assist researchers and regulators in evaluating potential responses of a specific biological system to a specific nanomaterial exposure. While the methodology development is well underway, there is still an urgent need for new tools, standardized practices, and meaningful interpretations that help address the challenges presented by nanomaterials and nano-enabled products. Due to their complexity, nanomaterials do not fit into established structure-activity models of chemical behavior. Nonetheless, it is important to identify the trends among different nanoparticles with varying measurable physicochemical characteristics for hazard identification, exposure analyses, risk assessments, and regulatory and policy decisions. The result, ideally, would be a framework that would produce meaningful groupings of particles in an effort to generalize some of the common features and resultant common responses that so many particles on the nanometer size scale seem to induce. For instance, particles less than 10 nm in diameter not only have different optical properties, they also have different translocation tendencies in cellular systems. Metal-based nanoparticles, a.k.a. metal colloids, tend to leach metal ions at different rates and at different pH’s. Crystalline materials produce more oxidant species than amorphous materials. While there are always exceptions to generalizations, these types of observations can be made when considering the literature as a whole.

Author’s short biography

Christie M. Sayes

Dr. Christie M. Sayes is the Program Manager for Nanotoxicology RTI International. She was formerly a professor of toxicology at Texas A&M University. She has more than a decade of experience in the fields of nanotechnology and nanotoxicology. She has authored publications, including original research, invited reviews and book chapters. She is a member of the Society of Toxicology, the American Chemical Society, and the Society of Environmental Toxicology and Chemistry. She serves on the Editorial Boards of Nanotoxicology, Toxicological Sciences, and Toxicology Letters. Recently, she was elected onto the Executive Committee of the North Carolina Chapter Society of Toxicology.

References

•  Berg  J.M.,  Romoser  A.,  Banerjee  N.,  Zebda  R.,  Sayes  C.M., The relationship between pH and zeta potential of ∼30 nm metal oxide nanoparticle suspensions relevant to in vitro toxicological evaluations. Nanotoxicology, 3 (2009) 276–283.
•  Berg  J.M.,  Ho  S.,  Hwang  W.,  Zebda  R.,  Cummins  K.,  Soriaga  M.P.,  Taylor  R.,  Guo  B.,  Sayes  C.M., Internalization of carbon black and maghemite iron oxide nanoparticle mixtures leads to oxidant production, Chemical Research in Toxicology, 23 (2010) 1874–1882.
•  Berg  J.M.,  Romoser  A.A.,  Figueroa  D.E.,  West  C.S.,  Sayes  C.M., Comparative cytological responses of lung epithelial and pleural mesothelial cells following in vitro exposure to nanoscale SiO₂, Toxicology in Vitro, 27 (2013) 24–33.
•  Cattley  R.C.,  Glover  S.E., Elevated 8-hydroxydeoxyguanosine in hepatic DNA of rats following exposure to peroxisome proliferators: Relationship to carcinogenesis and nuclear localization, Carcinogenesis, 14 (1993) 2495–2499.
•  Chen  C.,  Kwak  E.-S.,  Stein  B.,  Kao  C.C.,  Dragnea  B., Packaging of gold particles in viral capsids, Journal of Nanoscience and Nanotechnology, 5 (2005) 2029–2033.
•  Chen  C.,  Daniel  M.-C.,  Quinkert  Z.T.,  De  M.,  Stein  B.,  Bowman  V.D.,  Chipman  P.R.,  Rotello  V.M.,  Kao  C.C.,  Dragnea  B., Nanoparticle-templated assembly of viral protein cages, Nano Letters, 6 (2006) 611–615.
•  Chung  T.-H.,  Wu  S.-H.,  Yao  M.,  Lu  C.-W.,  Lin  Y.-S.,  Hung  Y.,  Mou  C.-Y.,  Chen  Y.-C.,  Huang  D.-M., The effect of surface charge on the uptake and biological function of mesoporous silica nanoparticles in 3t3-l1 cells and human mesenchymal stem cells, Biomaterials, 28 (2007) 2959–2966.
•  Cooke  M.S.,  Evans  M.D.,  Dizdaroglu  M.,  Lunec  J., Oxidative DNA damage: Mechanisms, mutation, and disease, The FASEB Journal, 17 (2003) 1195–1214.
•  Doherty  G.J.,  McMahon  H.T., Mechanisms of Endocytosis, Annu. Rev. Biochem., 78 (2009) 31.1–31.46.
•  Fortner  J.D.,  Lyon  D.Y.,  Sayes  C.M.,  Boyd  A.M.,  Falkner  J.C.,  Hotze  E.M.,  Alemany  L.B.,  Tao  Y.J.,  Guo  W.,  Ausman  K.D.,  Colvin  V.L.,  Hughes  J.B., C60 in water: Nanocrystal formation and microbial response, Environmental Science & Technology, 39 (2005) 4307–4316.
•  Freund  L.B.,  Lin  Y., The role of binder mobility in spontaneous adhesive contact and implications for cell adhesion, Journal of the Mechanics and Physics of Solids, 52 (2004) 2455–2472.
•  Gao  H.,  Shi  W.,  Freund  L.B., Mechanics of receptor-mediated endocytosis, Proceedings of the National Academy of Sciences of the United States of America, 102 (2005) 9469–9474.
•  Geiser  M.,  Rothen-Rutishauser  B.,  Kapp  N.,  Schurch  S.,  Kreyling  W.,  Schulz  H.,  Semmler  M.,  Im Hof  V.,  Heyder  J.,  Gehr  P., Ultrafine particles cross cellular membranes by non-phagocytic mechanisms in lungs and in cultured cells, Environ Health Perspect, 113 (2005) 1555–1560.
•  Gomes  A.,  Fernandes  E.,  Lima  J.L.F.C., Fluorescence probes used for detection of reactive oxygen species, Journal of Biochemical and Biophysical Methods, 65 (2005) 45–80.
•  Gould  G.W.,  Lippincott-Schwartz  J., New Roles for Endosomes: From Vesicular Carriers to Multi-Purpose Platforms, Nature Reviews/Molecular Cell Biology, 10 (2009) 287–292.
•  Harington  J.S.,  Miller  K.,  Macnab  G., Hemolysis by asbestos, Environ Res, 4 (1971) 95–117.
•  Harush-Frenkel  O.,  Rozentur  E.,  Benita  S.,  Altschuler  Y., Surface charge of nanoparticles determines their endocytic and transcytotic pathway in polarized MDCK cells, Biomacromolecules, 9 (2008) 435–443.
•  Harush-Frenkel  O.,  Debotton  N.,  Benita  S.,  Altschuler  Y., Targeting of nanoparticles to the clathrin-mediated endocytic pathway, Biochem Biophys Res Commun, 353 (2007) 26–32.
•  Huang  X.,  Bronstein  L.M.,  Retrum  J.,  Dufort  C.,  Tsvetkova  I.,  Aniagyei  S.,  Stein  B.,  Stucky  G.,  Mckenna  B.,  Remmes  N.,  Baxter  D.,  Kao  C.C.,  Dragnea  B., Self-assembled virus-like particles with magnetic cores, Nano Lett, 7 (2007) 2407–2416.
•  Hwang  E.S.,  Kim  G.H., Biomarkers for oxidative stress status of DNA, lipids, and proteins in vitro and in vivo cancer research, Toxicology, 229 (2007) 1–10.
•  Jiang  W.,  Kim  B.Y.,  Rutka  J.T.,  Chan  W.C., Nanoparticle-mediated cellular response is size-dependent, Nat Nanotechnol, 3 (2008) 145–150.
•  Larese  F.F.,  D’agostin  F.,  Crosera  M.,  Adami  G.,  Renzi  N.,  Bovenzi  M.,  Maina  G., Human skin penetration of silver nanoparticles through intact and damaged skin, Toxicology, 255 (2009) 33–37.
•  Li  J.J.,  Zou  L.,  Hartono  D.,  Ong  C.N.,  Bay  B.H.,  Lanry Yung  L.Y., Gold nanoparticles induce oxidative damage in lung fibroblasts in vitro, Advanced Materials, 20 (2008) 138–142.
•  Lin  W.,  Huang  Y.W.,  Zhou  X.D.,  Ma  Y., In vitro toxicity of silica nanoparticles in human lung cancer cells, Toxicol Appl Pharmacol, 217 (2006) 252–259.
•  Lundqvist  M.,  Stigler  J.,  Elia  G.,  Lynch  I.,  Cedervall  T.,  Dawson  K.A., Nanoparticle size and surface properties determine the protein corona with possible implications for biological impacts, Proc Natl Acad Sci USA, 105 (2008) 14265–14270.
•  Lyon  D.Y.,  Fortner  J.D.,  Sayes  C.M.,  Colvin  V.L.,  Hughe  J.B., Bacterial cell association and antimicrobial activity of a c60 water suspension, Environ Toxicol Chem, 24 (2005) 2757–2762.
•  Maynard  A.D.,  Aitken  R.J.,  Butz  T.,  Colvin  V.,  Donaldson  K.,  Oberdorster  G.,  Philbert  M.A.,  Ryan  J.,  Seaton  A.,  Stone  V.,  Tinkle  S.S.,  Tran  L.,  Walker  N.J.,  Warheit  D.B., Safe handling of nanotechnology, Nature, 444 (2006) 267–269.
•  Mayor  S.,  Pagano  R.E., Pathways of clathrin-independent endocytosis, Nat Rev Mol Cell Biol, 8 (2007) 603–612.
•  Murdock  R.C.,  Braydich-Stolle  L.,  Schrand  A.M.,  Schlager  J.J.,  Hussain  S.M., Characterization of nanomaterial dispersion in solution prior to in vitro exposure using dynamic light scattering technique, Toxicological Sciences, 101 (2008) 239–253.
•  Nel  A.,  Xia  T.,  Madler  L.,  Li  N., Toxic potential of materials at the nanolevel, Science, 311 (2006) 622–627.
•  Nicholls  D.G.,  Budd  S.L., Mitochondria and neuronal survival, Physiol Rev, 80 (2000) 315–360.
•  Nolan  R.P.,  Langer  A.M.,  Harington  J.S.,  Oster  G.,  Selikoff  I.J., Quartz hemolysis as related to its surface functionalities, Environ Res, 26 (1981) 503–520.
•  Oberdorster  G.,  Sharp  Z.,  Atudorei  V.,  Elder  A.,  Gelein  R.,  Kreyling  W.,  Cox  C., Translocation of inhaled ultrafine particles to the brain, Inhal Toxicol, 16 (2004) 437–445.
•  Oberdorster  G.,  Oberdorster  E.,  Oberdorster  J., Nanotoxicology: An emerging discipline evolving from studies of ultrafine particles, Environ Health Perspect, 113 (2005) 823–839.
•  Partha  R.,  Mitchell  L.R.,  Lyon  J.L.,  Joshi  P.P.,  Conyers  J.L., Buckysomes: Fullerene-based nanocarriers for hydrophobic molecule delivery, ACS Nano, 2 (2008) 1950–1958.
•  Qaddoumi  M.G.,  Gukasyan  H.J.,  Davda  J.,  Labhasetwar  V.,  Kim  K.J.,  Lee  V.H., Clathrin and caveolin-1 expression in primary pigmented rabbit conjunctival epithelial cells: Role in PLGA nanoparticle endocytosis, Molecular Vision, 9 (2003) 559–568.
•  Reynolds  E.S., The use of lead citrate at high ph as an electron-opaque stain in electron microscopy, J Cell Biol, 17 (1963) 208–212.
•  Romoser  A.A.,  Chen  P.L.,  Berg  J.M.,  Seabury  C.,  Ivanov  I.,  Criscitiello  M.F.,  Sayes  C.M., Quantum dots trigger immunomodulation of the NFκB pathway in human skin cells, Molecular Immunology, 48 (2011) 1349–1359.
•  Romoser  A.A.,  Figueroa  D.E.,  Sooresh  A.,  Scribner  K.,  Chen  P.L.,  Porter  W.,  Criscitiello  M.F.,  Sayes  C.M., Distinct immunomodulatory effects of a panel of nanomaterials in human dermal fibroblasts, Toxicology Letters, 210 (2012) 293–301.
•  Rothen-Rutishauser  B.M.,  Schurch  S.,  Haenni  B.,  Kapp  N.,  Gehr  P., Interaction of fine particles and nanoparticles with red blood cells visualized with advanced microscopic techniques, Environ Sci Technol, 40 (2006) 4353–4359.
•  Sager  T.M.,  Porter  D.W.,  Robinson  V.A.,  Lindsley  W.G.,  Schwegler-Berry  D.E.,  Castranova  V., Improved method to disperse nanoparticles for in vitro and in vivo investigation of toxicity, Nanotoxicology, 1 (2007) 118–129.
•  Sayes  C.M.,  Fortner  J.D.,  Guo  W.,  Lyon  D.,  Boyd  A.M.,  Ausman  K.D.,  Tao  Y.J.,  Sitharaman  B.,  Wilson  L.J.,  Hughes  J.B.,  West  J.L.,  Colvin  V.L., The differential cytotoxicity of water-soluble fullerenes, Nano Letters, 4 (2004) 1881–1887.
•  Sayes  C.M.,  Gobin  A.M.,  Ausman  K.D.,  Mendez  J.,  West  J.L.,  Colvin  V.L., Nano-c60 cytotoxicity is due to lipid peroxidation, Biomaterials, 26 (2005) 7587–7595.
•  Sayes  C.M.,  Wahi  R.,  Kurian  P.A.,  Liu  Y.,  West  J.L.,  Ausman  K.D.,  Warheit  D.B.,  Colvin  V.L., Correlating nanoscale titania structure with toxicity: A cytotoxicity and inflammatory response study with human dermal fibroblasts and human lung epithelial cells, Toxicological Sciences, 92 (2006a) 174–185.
•  Sayes  C.M.,  Liang  F.,  Hudson  J.L.,  Mendez  J.,  Guo  W.,  Beach  J.M.,  Moore  V.C.,  Doyle  C.D.,  West  J.L.,  Billups  W.E.,  Ausman  K.D.,  Colvin  V.L., Functionalization density dependence of single-walled carbon nanotubes cytotoxicity in vitro, Toxicology Letters, 161 (2006b) 135–142.
•  Sayes  C.M.,  Reed  K.L.,  Warheit  D.B., Assessing toxicity of fine and nanoparticles: Comparing in vitro measurements to in vivo pulmonary toxicity profiles, Toxicological Sciences, 97 (2007) 163–180.
•  Schellenberger  E.,  Schnorr  J.,  Reutelingsperger  C.,  Ungethüm  L.,  Meyer  W.,  Taupitz  M.,  Hamm  B., Linking proteins with anionic nanoparticles via protamine: Ultrasmall protein-coupled probes for magnetic resonance imaging of apoptosis, Small, 4 (2008) 225–230.
•  Schroeder  J.E.,  Shweky  I.,  Shmeeda  H.,  Banin  U.,  Gabizon  A., Folate-mediated tumor cell uptake of quantum dots entrapped in lipid nanoparticles, Journal of Controlled Release, 124 (2007) 28–34.
•  Uggeri  J.,  Gatti  R.,  Belletti  S.,  Scandroglio  R.,  Corradini  R.,  Rotoli  B.,  Orlandini  G., Calcein-am is a detector of intracellular oxidative activity, Histochemistry and Cell Biology, 122 (2004) 499–505.
•  Villanueva  A.,  Cañete  M.,  Roca  A.G.,  Calero  M.,  Veintemillas-Verdaguer  S.,  Serna  C.J.,  del Puerto  Morales M.,  Miranda  R., The influence of surface functionalization on the enhanced internalization of magnetic nanoparticles in cancer cells, Nanotechnology, 20 (2009) 115103.
•  Wang  H.,  Joseph  J.A., Quantifying cellular oxidative stress by dichlorofluorescein assay using microplate reader, Free Radic Biol Med, 27 (1999) 612–616.
•  Warheit  D.B.,  Webb  T.R.,  Colvin  V.L.,  Reed  K.L.,  Sayes  C.M., Pulmonary bioassay studies with nanoscale and fine-quartz particles in rats: Toxicity is not dependent upon particle size but on surface characteristics, Toxicological Sciences, 95 (2007a) 270–280.
•  Warheit  D.B.,  Webb  T.R.,  Reed  K.L.,  Frerichs  S.,  Sayes  C.M., Pulmonary toxicity study in rats with three forms of ultrafine-tio2 particles: Differential responses related to surface properties, Toxicology, 230 (2007b) 90–104.
•  Zhang  L.W.,  Monteiro-Riviere  N.A., Mechanisms of quantum dot nanoparticle cellular uptake, Toxicological Sciences, 110 (2009) 138–155.