Dependence of Nanoparticle Toxicity on Their Physical and Chemical Properties

1The International Organization for Standardization define nanoparticles (NPs) as structures whose sizes in one, two, or three dimensions are within the range from 1 to 100 nm. Apart from size, NPs may be classified in terms of their physical parameters, e.g., electrical charge; chemical characteristics, such as the composition of the NP core or shell; shape (tubes, films, rods, etc.); and origin: natural NPs (NPs contained in volcanic dust, viral particles, etc.) and artificial NPs, which are the focus of this review.

2Nanoparticles have become widely used in electronics, agriculture, textile production, medicine, and many other industries and sciences. NP toxicity for living organisms, however, is the main factor limiting their use in treatment and diagnosis of diseases. At present, researchers often face the problem of balance between the positive therapeutic effect of NPs and side effects related to their toxicity. In this respect, the choice of an adequate experimental model for estimating toxicity between in vitro (cell lines) and in vivo (experimental animals) ones is of paramount importance. The NP toxic effects on individual cell components and individual tissues are easier to analyze in in vitro models, whereas in vivo experiments make it possible to estimate the NP toxicity for individual organs or the body as a whole. In addition, the possible toxic effect of NPs depends on their concentration, duration of their interaction with living matter, their stability in biological fluids, and the capacity for accumulation in tissues and organs. Development of safe, biocompatible NPs that can be used for diagnosis and treatment of human diseases can only be based on complete understanding of the interactions between all factors and mechanisms underlying NP toxicity.

3In medicine, NPs can be used for diagnostic or therapeutic purposes. In diagnosis, they can serve as fluorescent labels for detection of biomolecules and pathogens and as contrast agents in magnetic resonance and other studies. In addition, NPs can be used for targeted delivery of drugs, including protein and polynucleotide substances; in photodynamic therapy and thermal destruction of tumors, and in prosthetic repair. Some types of NPs are already successfully used in clinic for drug delivery and tumor cell imaging.

4Examples of the use of gold NPs have been accumulating recently. They have proved to be efficient carriers of chemotherapeutics and other drugs. Gold NPs are highly biocompatible; however, although gold as a substance is inert towards biological objects, it cannot be argued that the same is true for gold NPs, since there are no conclusive data yet on the absence of delayed toxic effects. In addition to gold NPs, those based on micelles, liposomes, and polymers with attached “capture molecules” are already used as drug carriers. Single- and multiwalled nanotubes are good examples of NPs used for drug delivery. They are suitable for attaching various functional groups and molecules for targeted delivery, and their unique shape allows them to selectively penetrate through biological barriers. The use of NPs as vehicles for drugs enhances the specificity of delivery and decreases the minimum amount of NPs necessary for attaining and maintaining the therapeutic effect, thereby reducing the eventual toxicity. This is especially important in the case of highly toxic and short-lived chemo- and radiotherapeutic agents.

5Quantum dots (QDs) constitute another group of NPs with a high potential for clinical use. QDs are semiconductor nanocrystals from 2 to 10 nm in size. Their capacity for fluorescence in different spectral regions, including the infrared one, makes them suitable for labeling and imaging cells, cell structures, or pathogenic biological agents, as well as various processes in cells, tissues, and body as a whole, which has important diagnostic implications. NPs based on superparamagnetic iron oxide are efficiently used as contrast agents in magnetic resonance tomography (MRT) for imaging liver, bone marrow, and lymph node tissues. There is also an example where radioactively labeled single-walled carbon nanotubes functionalized with phospholipids were used for labeling integrin-containing tumors and their subsequent detection by means of positron emission tomography in experiments on mice.

6Nanoparticles have also been used in designing biosensors, including those based on carbon nanotubes for measuring the glucose level, detecting specific DNA fragments and regions, and identifying bacterial cells.

7Silver (or silver-containing) NPs exert antimicrobial and cytostatic effects; for this reason, they are widely used in medicine, e.g., for treating bandages, surgical instruments, prostheses, and contraceptives. Silver NPs have been reported to serve as effective and safe preservation agents in the cosmetic industry.

8However, NPs may still be highly toxic, even if the safety of using many of their chemical constituents in medicine has been proved. The toxic effect may be caused by their unique physical and chemical properties, which underlie specific mechanisms of interaction with living systems. In general, this determines the importance of studying the causes and mechanisms of the potential toxic effect of NPs.

9The toxicity of NPs is largely determined by their physical and chemical characteristics, such as their size, shape, specific surface area, surface charge, catalytic activity, and the presence or absence of a shell and active groups on the surface.

10The small size of NPs allows them to penetrate through epithelial and endothelial barriers into the lymph and blood to be carried by the bloodstream and lymph stream to different organs and tissues, including the brain, heart, liver, kidneys, spleen, bone marrow, and nervous system, and either be transported into cells by transcytosis mechanisms or simply diffuse into them through the cell membrane. Nanomaterials can also increase access to the blood stream through ingestion. Some nanomaterials can penetrate the skin and even greater microparticles can penetrate skin when it is flexed. Nanoparticles, because of their small size, can extravasate through the endothelium in inflammatory sites, epithelium (e.g., intestinal tract and liver), tumors or penetrate microcapillaries. Experiments modeling the toxic effects of NPs on the body have shown that NPs cause thrombosis by enhancing platelet aggregation, inflammation of the upper and lower respiratory tracts, neurodegenerative disorders, stroke, myocardial infarction, and other disorders. Note that NPs may enter not only organs, tissues, and cells, but also cell organelles, e.g., mitochondria and nuclei; this may drastically alter cell metabolism and cause DNA lesions, mutations, and cell death.

11The toxicity of QDs has been shown to be directly related to the leakage of free ions of metals contained in their cores, such as cadmium, lead, and arsenic, upon oxidation by environmental agents. QDs may be absorbed by mitochondria and cause morphological changes and dysfunction of the organelles. Entry of cadmium-based QDs into cells and formation of free Cd2+ ions causes oxidative stress.

12Recent studies have shown that contact of lung tissue with NPs about 50 nm in size leads to perforation of the membranes of type I alveolar cells and the resultant entry of the NPs into the cells. This, in turn, causes cell necrosis, as evidenced by the release of lactate dehydrogenase. There is evidence that QD penetration increases the cell membrane fluidity. On the other hand, the formation of reactive oxygen species (ROS) induced by peroxidation of membrane lipids may lead to the loss of membrane flexibility, which, as well as an abnormally high fluidity, inevitably results in cell death.

13Interaction of NPs with the cytoskeleton may also damage it. For example, TiO2 NPs induce conformational changes in tubulin and inhibit its polymerization, which disturbs intracellular transport, cell division, and cell migration. In human umbilical vein endothelial cells (HUVECs), damage of the cytoskeleton hinders the maturation of coordination adhesive complexes which link the cytoskeleton to the extracellular matrix, thereby disturbing the formation of the vascular network.

14In addition, the NP cytotoxicity may interfere with cell differentiation and protein synthesis, as well as activate proinflammatory genes and synthesis of inflammatory mediators. It should be specially noted that normal protective mechanisms do not affect NPs; macrophage uptake of large PEGylated nanoparticles is more efficient than uptake of small ones, which leads to accumulation of NPs in the body. Superparamagnetic iron oxide NPs have been demonstrated to disturb or entirely suppress osteogenic differentiation of stem cells and activate the synthesis of signal molecules, tumor antigens, etc.. In addition, interaction of NPs with the cell enhances the expression of the genes responsible for the formation of lysosomes, disturbs their functioning, and inhibits protein synthesis. A study on the toxic effects of NPs of different compositions on lung epithelial cells and human tumor cell lines has shown that NPs stimulate the synthesis of inflammation mediators, e.g., interleukin 8. According to Park, who studied the expression of proinflammatory cytokines in vitro and in vivo, the expressions of interleukin 1 beta (IL-1β) and tumor necrosis factor alpha (TNFα) are enhanced in response to silicon NPs.

15Oxidation, as well as action of various enzymes on the shell and surface of NPs, results in their degradation and release of free radicals. In addition to the toxic effect of free radicals expressed as oxidation and inactivation of enzymes, mutagenesis, and disturbance of chemical reactions leading to cell death, degradation of NPs leads to alteration or loss of their own functionality (e.g., the loss of the magnetic moment and the changes in the fluorescence spectrum and transport or other functions).

16In summary, the most common mechanisms of NP cytotoxicity are the following:NPs may cause oxidation via formation of ROS and other free radicals;NPs may damage cell membranes by perforating them;NPs damage components of the cytoskeleton, disturbing intracellular transport and cell division;NPs disturb transcription and damage DNA, thus accelerating mutagenesis;NPs damage mitochondria and disturb their metabolism, which leads to cell energy imbalance;NPs interfere with the formation of lysosomes, thereby hampering autophagy and degradation of macromolecules and triggering the apoptosis;NPs cause structural changes in membrane proteins and disturb the transport of substances into and out of cells, including intercellular transport;NPs activate the synthesis of inflammatory mediators by disturbing the normal mechanisms of cell metabolism, as well as tissue and organ metabolism (Fig. 1).

17NPs may cause oxidation via formation of ROS and other free radicals;

18NPs may damage cell membranes by perforating them;

19NPs damage components of the cytoskeleton, disturbing intracellular transport and cell division;

20NPs disturb transcription and damage DNA, thus accelerating mutagenesis;

21NPs damage mitochondria and disturb their metabolism, which leads to cell energy imbalance;

22NPs interfere with the formation of lysosomes, thereby hampering autophagy and degradation of macromolecules and triggering the apoptosis;

23NPs cause structural changes in membrane proteins and disturb the transport of substances into and out of cells, including intercellular transport;

24NPs activate the synthesis of inflammatory mediators by disturbing the normal mechanisms of cell metabolism, as well as tissue and organ metabolism (Fig. 1).

25Although there are numerous mechanisms of NP toxicity, it is necessary to determine and classify the type and mechanism of each particular toxic effect of NPs as dependent on their physical and chemical properties.

26The toxicity of NPs is considered to depend on their physical and chemical characteristics, including the size, shape, surface charge, chemical compositions of the core and shell, and stability. In particular, Oh et al., using the data meta-analysis of 307 papers describing 1741 cell viability-related data samples, recently analyzed the CdSe quantum dot toxicity. It has been shown that the QD nanotoxicity is closely correlated with their surface properties (including shell, ligand, and surface modifications), diameter, toxicity assay type used, and the exposure time. Which of these factors is the most important is determined by the specific experimental task and model; therefore, we will now consider each factor separately.

27The NP size and surface area play an important role, largely determining the unique mechanism of NP interaction with living systems. NPs are characterized by a very large specific surface area, which determines their high reaction capacity and catalytic activity. The sizes of NPs (from 1 to 100 nm) are comparable with the size of protein globules (2–10 nm), diameter of DNA helix (2 nm), and thickness of cell membranes (10 nm), which allows them to easily enter cells and cell organelles. For example, Huo et al. have demonstrated that gold NPs no larger than 6 nm effectively enter the cell nucleus, whereas large NPs (10 or 16 nm) only penetrate through the cell membrane and are found only in the cytoplasm. This means that NPs several nanometers in size are more toxic than 10 nm or larger NPs, which cannot enter the nucleus. Pan et al. have traced the dependence of the toxicity of gold NPs on their size in the range from 0.8 to 15 nm. The NPs 15 nm in size have been found to be 60 times less toxic than 1.4-nm NPs for fibroblasts, epithelial cells, macrophages, and melanoma cells. It is also noteworthy that 1.4-nm NPs cause cell necrosis (within 12 h after their addition to the cell culture medium), whereas 1.2-nm NPs predominantly cause apoptosis. These data suggest not only that NPs can enter the nucleus, but also that the correspondence of the geometric size of NPs (1.4 nm) to that of the major groove of DNA allows them to effectively interact with the negatively charged sugar–phosphate DNA backbone and block the transcription.

28In addition, the NP size largely determines how the NPs interact with the transport and defense systems of cells and the body. This interaction, in turn, affects the kinetics of their distribution and accumulation in the body. The review paper by presents both theoretical considerations and numerous experimental data demonstrating that NPs smaller than 5 nm usually overcome cell barriers nonspecifically, e.g., via translocation, whereas larger particles enter the cells by phagocytosis, macropinocytosis, and specific and nonspecific transport mechanisms. An NP size of about 25 nm is believed to be optimal for pinocytosis, although this also strongly depends on the cell size and type. In vivo experiments have shown that NPs smaller than 10 nm are rapidly distributed among all organs and tissues upon intravenous administration, whereas most larger NPs (50–250 nm) are found in the liver, spleen, and blood. This suggests that large NPs are recognized by specific defense systems of the body and absorbed by the system of mononuclear phagocytes, which prevents them from entering other tissues. In addition, Talamini et al. claimed that the NP size and shape influence the kinetics of accumulation and excretion of gold NPs in filter organs, and only star-like gold NPs are able to accumulate in the lung. They have also shown that the changes in the NP geometry do not improve the NP passage of the blood–brain barrier.

29The large specific surface area ensures effective adsorption of NPs on the cell surface. This was shown in a study on the hemolytic activity of 100- to 600-nm mesoporous silicon particles towards human erythrocytes. The particles 100 nm in size were effectively adsorbed on the erythrocyte surface without causing cell destruction or any morphological changes in the cells, whereas 600-nm particles deformed the membrane and entered the cells, resulting in erythrocyte destruction (hemolysis).

30The characteristic shapes of NPs are spheres, ellipsoids, cylinders, sheets, cubes, and rods. NP toxicity strongly depends on their shape. This has been shown for numerous NPs of different shapes and chemical compositions. For example, spherical NPs are more prone to endocytosis than nanotubes and nanofibers. Single-walled carbon nanotubes have been found to more effectively block calcium channels compared to spherical fullerenes.

31Comparison of the effects of hydroxyapatite NPs with different shapes (needle-like, plate-like, rod-like, and spherical) on cultured BEAS-2B cells have shown that plate-like and needle-like NPs cause death of a larger proportion of cells than spherical and rod-like NPs. This is partly accounted for by the capacity of plate-like and needle-like NPs for damaging cells and tissue upon direct contact. Hu et al. obtained interesting data when studying the damage of mammalian cells by graphene oxide nanosheets. The toxicity of these NPs was determined by their shape allowing them to physically damage the cell membrane. However, their toxicity was found to decrease with an increase in the fetal calf serum concentration in the culture medium. This was explained by a high capacity of graphene oxide NPs for adsorbing protein molecules, which cover the NP surface, thereby changing the shape of the NPs and partly preventing the damage of cell membranes.

32Although the toxicity of NPs strongly depends on their size and shape, the influence of other factors, such as the NP chemical composition and crystal structure, should not be disregarded. Comparison of the effects of 20-nm silicon dioxide (SiO2) and zinc oxide (ZnO) NPs on mouse fibroblasts has shown that they differ in the mechanisms of toxicity. ZnO NPs cause oxidative stress, whereas SiO2 NPs alter the DNA structure.

33The toxicity of NPs is indeed largely determined by their chemical composition. It has been shown that degradation of NPs can occur, and its extent depends on the environment conditions, e.g., pH or ionic strength. The most common cause of the toxic effect of NPs interacting with cells is leakage of metal ions from the NP core. The toxicity also depends on the composition of the core of NPs. Some metal ions, such as Ag and Cd, are in fact toxic and, therefore, cause damage of the cells. Other metal ions, such as Fe and Zn, are biologically useful, but, at high concentrations, they could damage cellular pathways and, hence, cause high toxicity. However, this effect can be decreased, e.g., by coating NP cores with thick polymer shells, silica layers, or gold shells instead of short ligands or by using nontoxic compounds for NP synthesis. On the other hand, the composition of the core could be altered by addition of other metals. This can result in enhanced chemical stability against NP degradation and metal ion leakage into the body.

34The toxicity of NPs also depends on their crystal structure. The relationship between crystal structure and toxicity has been studied using a human bronchial epithelium cell line and titanium oxide NPs with different types of crystal lattice. It has been demonstrated that NPs with a rutile-like crystal structure (prism-shaped TiO2 crystals) cause oxidative damage of DNA, lipid peroxidation, and formation of micronuclei, which indicates abnormal chromosome segregation during mitosis, whereas NPs with anatase-like crystal structure (octahedral TiO2 crystals) of the same size are nontoxic. It should be noted that the NP crystal structure may vary depending on the environment, e.g., upon interaction with water, biological fluids, or other dispersion media. There is evidence that the crystal lattice of ZnS NPs is rearranged into a more ordered structure upon contact with water.

35The surface charge of NPs plays an important role in their toxicity, because it largely determines the interactions of NPs with biological systems.

36NP surfaces and their charges could be modified by grafting differently charged polymers. PEG (polyethylene glycol) or folic acid is often used to improve the NP intracellular uptake and ability to target specific cells. The synthesis of biocompatible TiO2 nanoparticles containing functional NH2 or SH groups has also been reported. Other substances, such as methotrexate, polyethyleneimine, and dextran, had also been used to modify NP surfaces and their charge.

37A high toxicity of positively charged NPs is explained by their ability to easily enter cells, in contrast to negatively charged and neutral NPs. This is accounted for by electrostatic attraction between the negatively charged cell membrane glycoproteins and positively charged NPs. Comparison of the cytotoxic effects of negatively and positively charged polystyrene NPs on HeLa and NIH/3T3 cells has shown that the latter NPs are more toxic. This is not only because positively charged NPs more effectively penetrate through the membrane, but also because they are more strongly bound to the negatively charged DNA, causing its damage and, as a result, prolongation of the G0/G1 phase of the cell cycle. Negatively charged NPs have no effect on the cell cycle. Similar results have been obtained for positively and negatively charged gold NPs, positive NPs being absorbed by cells in larger amounts and more rapidly than negative ones and being more toxic.

38Positively charged NPs have an enhanced capacity for opsonization, i.e., adsorption of proteins facilitating phagocytosis, including antibodies and complement components, from blood and biological fluids. The adsorbed proteins, referred to as the protein crown, may affect the surface properties of NPs. For example, they may alter the surface charge, aggregation characteristics, and/or hydrodynamic diameter of NPs. In addition, adsorption of proteins on the NP surface leads to their conformational changes, which may decrease or completely inhibit the functional activities of the adsorbed proteins. The protein crown mainly consists of major serum proteins, such as albumin, fibrinogen, and immunoglobulin G, as well as other effector, signal, and functional molecules. Binding to NPs alters the protein structure, which leads to the loss of their enzymatic activity, disturbance of biological processes, and precipitation of ordered polymeric structures, e.g., amyloid fibrils. This may lead to various diseases, such as amyloidosis. In vitro experiments have demonstrated that QDs coated with a hydrophilic polymer accelerate the formation of fibrils of human β2 microglobulin, which are then arranged into multilayered structures on the particle surface; this results in a local increase in the protein concentration on the NP surface, precipitation, and formation of oligomers.

39Xu et al. developed a method for changing the NP charge from negative to positive via various modifications of the surface. For example, polymer NPs were modified with a pH-sensitive polymer so that, being negatively charged in a neutral medium, they acquired a positive charge in an acid medium, at pH 5–6. This technique makes it possible to substantially increase the rate of NP uptake by cells, which could be used for drug delivery to tumor cells. Estimation of the cytotoxicity of surface-modified cerium oxide NPs for H9C2, HEK293, A549, and MCF-7 cells has shown that basically different biological and toxic effects can be obtained by using different polymers to make the NPs positively or negatively charged or neutral. Specifically, positively charged and neutral NPs are absorbed by all cell types at the same rate, whereas negatively charged ones predominantly accumulate in tumor cells. Thus, modification of the NP charge allows their localization and toxicity to be controlled, which could be used for developing effective systems for delivery of chemotherapeutic drugs to tumors.

40Application of a shell onto the surface of NPs is necessary for changing their optical, magnetic, and electrical properties; it is used for improving NP biocompatibility and solubility in water and biological fluids by decreasing their aggregation capacity, increasing their stability, etc. Thus, the shell decreases the toxicity of NPs and provides them with the capacity for selective interaction with different types of cells and biological molecules. In addition, the shell considerably influences the NP pharmacokinetics, changing the patterns of NP distribution and accumulation in the body.

41As noted above, NP toxicity is largely related to the formation of free radicals. However, the shell can considerably mitigate or eliminate this negative effect, as well as stabilize NPs, increase their resistance to environmental factors, decrease the release of toxic substances from them, or make them tissue-specific. For example, Cho et al. modified polymer NPs by coating them with lectins. The modified NPs selectively bound with tumor cells presenting sialic acid molecules on the surface, which made the NPs suitable for specifically labeling cancer cells.

42The NP surface can be modified with both organic and inorganic compounds, e.g., polyethylene glycol, polyglycolic acid, polylactic acid, lipids, proteins, low molecular weight compounds, and silicon. This variety of modifiers makes it possible to form complex systems on the NP surface for changing the NP properties and for their specific transport and accumulation.

43Nanoparticles coated with shells of synthetic polymers are used for delivery of antigens, thus serving as adjuvants boosting the immune response. This allows obtaining vaccines against the antigens that are targets of strong natural nonspecific cellular immunity.

44The shell is often used for improving solubilization and decreasing the toxicity of QDs, because their metal cores are hydrophobic and mainly consist of toxic heavy metals, such as cadmium, tellurium, and mercury. The shell increases the stability of the QD core and prevents its desalination and oxidative or photolytic degradation. This, in turn, decreases the leakage of metal ions outside of the QD core and, hence, the toxicity of QDs.

45During the past two decades, the use of NPs has tremendously extended and led to the foundation of nanotoxicology, a new science studying the potential toxic effects of NPs on biological and ecological systems. The general goal of nanotoxicology is to develop the rules of synthesis of safe NPs. This calls for a comprehensive, systemic approach to analysis of the toxic properties of NPs and their effects on cells, tissues, organs, and the body as a whole.

46There are two routine approaches to the study of the effects of various substances on living systems, which are also applicable to NP toxic effects: in vitro experiments on model cell lines and in vivo experiments on laboratory animals. We do not consider here the third possible approach to estimating NP toxicity, computer simulation, because the pathways and mechanisms of the toxic effects of NPs are not known well enough for a computer model to predict the consequences of interactions between NPs and living matter for a wide range of NPs with sufficient reliability.

47Both cell culture and animal experimental models for studying NP toxicity have their specific advantages and disadvantages. The former allow deeper insight into the molecular mechanisms of toxicity and identification of the primary targets of NPs; however, the patterns of the distribution of NPs in the body and their transport to different tissues and cells are not taken into consideration. The study of NP toxicity in animal experiments allows the delayed effects of NP action in vivo to be estimated. However, the general pattern of toxicity manifestations becomes so complicated that it is impossible to determine which of them is the primary cause of the observed effect and which are its consequences.

48Many studies of NP toxicity are carried out in primary cell cultures serving as models of various types of human and animal tissues. In some cases, tumor cell lines are used, e.g., for estimating the toxic effects of NPs used in cancer chemotherapy. The type of cells is selected according to the potential route by which NPs enter the body. This may be oral uptake (mainly by ingestion), transdermal uptake (through the skin surface), inhalation uptake of NPs contained in the breathing air, or intentional NP injection in clinic. Intestinal epithelium cells (Caco-2, HT29, and SW480) are often used in experimental models for studying the toxicity of ingested NPs (Table 1). In these models, the kinetics of NP uptake by cells and the viability of cells upon the NP uptake are studied.

49The NPs that serve as carriers of drugs or contrast agents, or those used for imaging, are administered by injection. The toxicity of these NPs is studied in primary blood cell cultures. Most commonly, hemolysis, platelet activation, and platelet aggregation are estimated. In addition to primary blood cell cultures, cultured HUVECs, mesenchymal stem cells, mononuclear blood cells, and various tumor cell lines (HeLa, MCF-7, PC3, C4-2, and SKBR-3) are used (Table 2).

50The toxicity of inhaled NPs is studied using the cell lines modeling different tissues of the respiratory system, e.g., A549 and C10 cells of pulmonary origin, alveolar macrophages (RAW 264.7), various epithelial cells and fibroblasts (BEAS-2B, NHBE, 16-HBE, SAEC), as well as human monocytes (THP-1) (Table 3).

51The toxicity of NPs that enter the body transdermally is usually studied in keratinocytes, fibroblasts, and, more rarely, sebocytes (cells of sebaceous glands) (Table 4).

52Although the majority of in vitro nanotoxicity studies are carried out on cell monocultures, studies using two other approaches are increasingly often reported in the literature. One of them is co-culturing of several types of cells; the other is the use of 3D cultures. The rationale for these approaches is the need for more realistic models of mammalian tissues and organs. For example, co-cultured Caco-2 epithelial colorectal adenocarcinoma cells and Raji cells (a lymphoblast cell line) have served as a model of the human intestinal epithelium in experiments on the toxicity of silver NPs. A co-culture of three cell lines derived from lung epithelial cells, human blood macrophages, and dendritic cells has been used as an experimental model in a study on the toxic effects of inhaled NPs. A model of skin consisting of co-cultured fibroblasts and keratinocytes has been suggested.

53It is known that the cell phenotype, as well as cell functions and metabolic processes, is largely determined by the complex system of cell interactions with other cells and the surrounding extracellular matrix. Therefore, many important characteristics of cells with an adhesive type of growth in a monolayer culture substantially differ from those of the same cells in the living tissue; hence, conclusions from many experiments on the NP toxic effects on cells growing in a monolayer are somewhat incorrect. Experimental 3D models of tissues and organs have been used for analysis of NP toxicity and penetration into cells in several published studies. For example, there are 3D models based on polymer hydrogels and models constructed in special perfusion chambers containing a semipermeable membrane to which the cells are attached. Li et al. and Lee et al. used multicellular spheroids about 100 μm in size to obtain a 3D model of the liver and compare the toxicities of CdTe and Au NPs in experiments on this model and a monolayer culture of liver cells. The results obtained using the 3D model were more closely correlated with the data obtained in experiments on animals, which indicates a considerable potential of this approach for adequate and informative testing of NP toxicity.

54In addition to the study of multilayered and 3D cell cultures, the behavior of NPs in the living body is being extensively studied. Since these studies are focused on the biomedical applications of NPs, the NP toxicity for living organisms remains an important issue. Although NPs are highly promising for various clinical applications, they are potentially hazardous. This hazard cannot be estimated correctly in vitro, following from the comparison of the in vivo and in vitro effects of NPs.

55Titanium dioxide (TiO2) particles are among the most widely used NPs, in particular, in environment protection measures. Therefore, it was exceptionally important to estimate their toxicity in the case of a 100% bioavailability, namely, in experiments with their intravenous injection to experimental animals. This study has been performed by Fabian et al.. Experimental animals (rats) were injected with a suspension of TiO2 NPs at a dose of 5 mg/kg, and their biodistribution, as well as the general condition of the animals, was monitored. The results have shown that the animals exhibit no signs of ailment or disorder, nor is inflammation or another manifestation of a toxic effect observed, within 28 days. This suggests that TiO2 NPs are relatively harmless.

56Silver NPs are another example of NPs potentially useful in medicine, owing to their antimicrobial activity. Their toxicity and biodistribution were analyzed in an experiment where CD-1 mice were intravenously injected with 10 mg/kg of silver NPs of different sizes (10, 40, and 100 nm) coated with different shells. Although each type of NPs was found to cause toxic damage of tissues, larger particles were less toxic, probably, due to their lower penetration capacity. Asare et al. estimated the genotoxicity of silver and titanium NPs administered at a dose of 5 mg/kg. They have found that silver NPs cause DNA strand breaks and oxidation of purine bases in the tissues examined. Gold nanoparticles have a similar effect. They have been shown to be toxic for mice, causing weight loss, decrease in the hematocrit, and reduction of the red blood cell count.

57Targeted drug delivery is one of the most important applications of NPs. In this case, it is also paramount to know their toxic properties, because the positive effect of their use should prevail over the negative one. Kwon et al. have developed antioxidant NPs from the polymeric prodrug of vanillin. Their study has shown that the NPs have no toxic effect on the body, specifically the liver, at doses lower than 2.5 mg/kg. Similar results have been obtained for gelatin NPs modified with polyethylene glycol, which are planned to be used for targeted delivery of ibuprofen sodium salt. The NPs have proved to be nontoxic at the dose that is necessary for effective drug delivery (1 mg/kg), which has been confirmed by measuring the inflammatory cytokine levels in the animals studied, as well as histological analysis of their organs.

58Quantum dots are among the NPs that are most promising for medical applications (Fig. 2). However, they are potentially hazardous for human health, because they exhibit various toxic effects in both in vitro and in vivo experiments.

59Toxic effects of QDs in vivo are usually studied in experiments on mice and rats. A study on the toxicity of cadmium-based QDs for mice showed that QDs were distributed throughout the body as soon as 15 min after injection to the caudal vein, after which they accumulated in the liver, kidneys, spleen, red bone marrow, and lymph nodes. Two years after the injection, fluorescence was mainly retained in lymph nodes; in other organs, no QDs were detected. It should be also noted that the fluorescence spectrum was shifted to the blue spectral region because of the destruction of the QD shell and changes in the shape, size, and surface charge of the QDs. This, however, occurred rather slowly, because the QDs were found to be nontoxic after their injection at the doses at which pure cadmium ions would have had a lethal effect. Similar results were obtained by Yang et al.. Zhang et al. showed that CdTe QDs predominantly accumulated in the liver, decreasing the amount of antioxidants in it and inducing oxidative stress in liver cells.

60Cadmium and tellurium ions tend to accumulate in various organs and tissues upon degradation and decay of the cores of CdTe/ZnS QDs. Experiments on mice have shown that cadmium predominantly accumulates in the liver, kidneys, and spleen, whereas tellurium accumulates almost exclusively in the kidneys. Ballou et al. found that cadmium-containing QDs coated with polymer shells of polyacrylic acid or different derivatives of polyethylene glycol had no lethal effect on experimental mice and remained fluorescent for 4 months. СdSe/ZnS NPs also had no detectable pathological effect on mice; however, the absence of distinct signs of pathology still does not mean that the QDs are absolutely nontoxic.

61Hu et al. found that lead-containing QDs had no toxic effect on mice for 4 weeks; however, this was most probably because the QDs studied were coated with a polyethylene glycol shell.

62Since heavy metals contained in QDs are a factor of their toxicity, several research groups suggested that heavy-metal-free NPs be synthesized. For example, Pons et al. synthesized CuInS2/ZnS QDs fluorescing in the near-infrared spectral region (at a wavelength of about 800 nm) and supposed that this composition would make the QDs nontoxic for experimental animals. Comparison of the effects of CuInS2/ZnS and CdTeSe/CdZnS QDs on regional lymph nodes in mice showed that the lymph nodes were only slightly, if at all, enlarged upon injection of the QDs not containing heavy metals, whereas injection of the CdTeSe/CdZnS QDs induced a distinct immune response in them. QDs in which silicon was substituted for heavy metals also had no toxic effect on mice.

63Even QDs containing heavy metals are often found to be nontoxic. One of the possible explanations is that QDs are coated with the protein crown upon entering the living body; this crown shields their surface and protects cells against damage. Usually, the proteins that are included in the NP molecular corona are major serum proteins, such as albumin, immunoglobulin G (IgG), fibrinogen, and apolipoproteins. Molecular corona also can influence on the interaction of NPs with cells. Zyuzin et al. have demonstrated that, in human endothelial cells, the NP protein corona decreases the NP nonspecific binding to the cell membrane, increases the residence time of NP in early endosomes, and reduces the amount of internalized NPs.

64However, even in the absence of direct signs of intoxication in experimental animals, it remains unclear whether the use of QDs in medicine is safe for humans. In some cases, the QD toxicity was not detected in mice because the NPs were neutralized by the liver and accumulated in it; in other cases, QDs coated with phospholipid micelles exhibited reduced toxicity owing to the shell. Despite the extensive in vivo studies on QD toxicity, their use in biomedicine remains an open question. One of the main reasons is that all the delayed effects of QDs cannot be monitored in experimental animals, because their lifespan is as short as a few years, which is insufficient for complete elimination or degradation of NPs.

65The potential toxicity of NPs is the main problem of their use in medicine. Therefore, not only positive results of the use of NPs, but also the possible unpredictable negative consequences of their action on the human body, should be scrutinized. The toxicity of NPs is related to their distribution in the bloodstream and lymph stream and their capacities for penetrating into almost all cells, tissues, and organs and interacting with various macromolecules and altering their structure, thereby interfering with intracellular processes and the functioning of whole organs. The NP toxicity strongly depends on their physical and chemical properties, such as the shape, size, electric charge, and chemical compositions of the core and shell. Many types of NPs are not recognized by the protective systems of cells and the body, which decreases the rate of their degradation and may lead to considerable accumulation of NPs in organs and tissues, even to highly toxic and lethal concentrations. However, a number of approaches to designing NPs with a decreased toxicity compared to the traditional NPs are already available. Advanced methods for studying the NP toxicity make it possible to analyze different pathways and mechanisms of toxicity at the molecular level, as well as reliably predict the possible negative effect at the body level.

66Thus, it is obvious that designing NPs that have small or no negative effects is impossible unless all qualitative and quantitative physical and chemical properties of NPs are systematically taken into consideration and a relevant experimental model for estimating their influence on biological systems is available.