**2. Nanoparticle cytotoxicity analyses and their limitations**

The cytotoxicity study is an essential and crucial step in testing novel substances/nanomaterials in the context of biological and medical applications. Assays based on tetrazolium salts, like MTT (2-(4,5-dimethyl-2-thiazolyl)-3,5-diphenyl-2H-tetrazolium bromide), nitroblue tetrazolium (NBT) and the second-generation tetrazolium salts, such as XTT (sodium 2,3-bis(2 methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)-carbonyl]-2H-tetrazolium inner salt), MTS (5-[3-(carboxymethoxy)phenyl]-3-(4,5-dimethyl-2-thiazolyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt) and WST-1 (sodium 5-(2,4-disulfophenyl)-2-(4-iodophenyl)-3-(4-nitrophenyl)- 2H-tetrazolium inner salt) are basic tools for cytotoxicity determination, but nanomaterial testing is associated with certain challenges. The type of nanomaterials, manufacturing conditions, colloidal dispersion, chemical purity and photocatalytic activity of NPs may determine the usage of most traditional assays. Interactions between nanoparticles and molecules (i.e., reactants) used in well-established assays significantly affect the results and are one of the reasons of result variations [18, 19]. In assays based on colorimetric and fluorescence measurements, it has been found that nanomaterials, such as carbon nanotubes (CNTs), graphene/ graphene oxide nanosheets, TiO<sup>2</sup> nanoparticles or boron nitride, interact with chromophore molecules, which may lead to false-positive results [19–21]. In other cases, the total surface area of NPs was sufficient to adsorb the reagent or fluorescent molecules, especially those with aromatic rings, which in turn led to false-negative results [21]. These results suggest using alternative cytotoxicity assays based on tetrazolium salts, e.g., XTT, WST-1, INT or other assays that complement the analysis, e.g., Alamar Blue (AB), neutral red uptake (NRU) assay, LDH leakage assay, flow cytometry, cell death analysis (using trypan blue or annexin V/propidium iodide), protein concentration measurements using the Bradford assay, measurements of mitochondrial membrane permeability (MMP) or loss of glutathione (e.g., GSH) and the activation of proinflammatory cytokines (e.g., IL-6, IL-8 and/or TNF-α) [15, 19, 21].

while nanoscale devices smaller than 20 nm can cross blood vessels and cumulate in tissues [2]. Particles with larger surface area display tendency to agglomerate in the liquid, interact with molecules, such as proteins and DNA and may cause oxidation and DNA damage [3, 4, 14].

It is known that the shape (aspect ratio) also determines cellular uptake efficiency and may affect cell viability. Additionally, surface chemistry of nanomaterials modulates the response of biological systems and distribution in the organism. Surface functionalisation (with Fe3

gold nanoparticles; type of bonding on the surface, e.g., covalent, noncovalent; dispersing agents, e.g., PEG) is a crucial factor that can significantly change the toxicity of NPs and prevent NPs from aggregating; it can also change their fate in biological systems [2, 3, 11]. For

Na) increased their biocompatibility in comparison with unfunctionalised or carboxylic acid–functionalized (─COOH) MWCNTs [15]. Cellular uptake depends on different factors, such as nanomaterial and cell type, but also on environmental properties and the complexity of culture media. These specific conditions determine the aggregation process, which makes the interpretation of data on nanoparticle biodistribution or uptake difficult [3]. NP agglomerates affect and limit the direct extrapolation of *in vitro* data (providing a basis for understanding the mechanism of NP cytotoxicity and their uptake at the cellular level) to *in vivo* exposure [11, 16]. Intercellular localization of NPs and their interaction with cell components, such as the membrane, mitochondria, lysosomes and/or nucleus, are essential [11]. NPs can affect cell and organelle membranes, induce oxidative stress (ROS), DNA damage and mutagenesis and evoke apoptosis and protein up-/downregulation [11]; they can also modulate immune response [3, 17].

example, functionalization of MWCNTs with sodium sulfonic acid salt (─SO<sup>3</sup>

**2. Nanoparticle cytotoxicity analyses and their limitations**

graphene oxide nanosheets, TiO<sup>2</sup>

The cytotoxicity study is an essential and crucial step in testing novel substances/nanomaterials in the context of biological and medical applications. Assays based on tetrazolium salts, like MTT (2-(4,5-dimethyl-2-thiazolyl)-3,5-diphenyl-2H-tetrazolium bromide), nitroblue tetrazolium (NBT) and the second-generation tetrazolium salts, such as XTT (sodium 2,3-bis(2 methoxy-4-nitro-5-sulfophenyl)-5-[(phenylamino)-carbonyl]-2H-tetrazolium inner salt), MTS (5-[3-(carboxymethoxy)phenyl]-3-(4,5-dimethyl-2-thiazolyl)-2-(4-sulfophenyl)-2H-tetrazolium inner salt) and WST-1 (sodium 5-(2,4-disulfophenyl)-2-(4-iodophenyl)-3-(4-nitrophenyl)- 2H-tetrazolium inner salt) are basic tools for cytotoxicity determination, but nanomaterial testing is associated with certain challenges. The type of nanomaterials, manufacturing conditions, colloidal dispersion, chemical purity and photocatalytic activity of NPs may determine the usage of most traditional assays. Interactions between nanoparticles and molecules (i.e., reactants) used in well-established assays significantly affect the results and are one of the reasons of result variations [18, 19]. In assays based on colorimetric and fluorescence measurements, it has been found that nanomaterials, such as carbon nanotubes (CNTs), graphene/

molecules, which may lead to false-positive results [19–21]. In other cases, the total surface area of NPs was sufficient to adsorb the reagent or fluorescent molecules, especially those with aromatic rings, which in turn led to false-negative results [21]. These results suggest

nanoparticles or boron nitride, interact with chromophore

SO<sup>3</sup>

180 Cytotoxicity

O4 ,

Na or -phenyl-

A number of studies have been carried out to verify the effect of NPs on assay reagents [22]. Wörle-Knirsch and co-workers [23] indicated that CNT analysed using the MTT assay caused false-positive results due to the strong interaction between CNT and the insoluble formazan crystals [19, 23]. In the aforementioned study, SWCNTs were analysed on A549 (human alveolar epithelial cell line), ECV304 (endothelial cells derived from umbilical cord) and NR8383 (rat alveolar macrophage cell line) cell cultures and the results obtained in the MTT assay were verified by WST-1, LDH, mitochondrial membrane potential (MMP) and annexin V/PI analysis. The MTT assay indicated that SWCNTs affected cell viability, reducing it almost by 60% after a 24-h incubation period. Moreover, the decreased cell viability did not recover after longer incubation or higher concentrations of nanotubes. The results of the MTT assay were verified using WST-1, and no reduction in viability was detected. LDH and MMP assays confirmed WST-1 results, and flow cytometry using annexin V/propidium iodide showed lack of necrosis and/or apoptosis. Wörle-Knirsch et al. [23] concluded that nanotoxicological assays needed standardising with regard to the tested nanomaterial

Lupu and Popescu [24] used the MTT assay to evaluate TiO<sup>2</sup> toxicity. The effect of TiO<sup>2</sup> nanoparticles on living models is crucial due to their utilisation in food, beauty care and pharmacology industries. Additionally, TiO<sup>2</sup> nanoparticles are known to exhibit photocatalytic properties: the ability to catalyse redox reactions of molecules adsorbed on the surface during light exposition (λ < 385 nm). Photocatalytic reaction may occur by direct charge transfer of electrons (e− ) and holes (h<sup>+</sup> ) generated by light on the surface of titania nanoparticles. The reaction may be also mediated by reactive oxygen species (ROS), e.g., hydroxyl radicals (˙ OH) or superoxide anions (˙ O2 − ) formed at the interface of TiO<sup>2</sup> and water. Lupu and Popescu [24] clearly demonstrated that TiO<sup>2</sup> nanoparticles induced transformation of noncellular MTT to formazan. Formazan formation was found to be proportional to titania NPs, and this process was enhanced by daylight exposure. Moreover, the results obtained in the experiment without cellular model were validated using three cell lines—V79-4, HeLa and B16. The results demonstrated false viability that increased up to 14% for TiO<sup>2</sup> concentrations higher than 50 μgml−1 [24]. In addition, the TiO<sup>2</sup> -MTT reaction was analysed in PBS environment. The reaction rate (formazan production rate) was proportional to TiO<sup>2</sup> and UV radiations (at 312 and 365 nm wavelengths) and inversely proportional to initial concentration of MTT. Moreover, reaction efficiency was enhanced by the presence of Na<sup>2</sup> HPO<sup>4</sup> (phosphate concentration of 0.005 M for maximum efficiency), which is the basic component of PBS [25].

Casey et al. [26, 27] proved that single-walled carbon nanotubes (HiPco®) interacted with indicator dyes applied in Coomassie Blue, AlamarBlue™, neutral red uptake, MTT and WST-1 assays. In all cases, nanotubes interacted with dyes, which resulted in the reduction of the associated absorption/fluorescent emission. A spectroscopic study demonstrated that SWCNTs interacted with Coomassie and reduced the absorbance for all concentrations tested (0.003–0.800 mgml−1). As regards the AlamarBlue™ analysis (fluorescent measurements of all single-walled carbon nanotube solutions), quenching was monitored as a function of SWCNT concentration and plotted as an emission ratio at 595 nm by 540 nm excitation for the AB assay (5% solution of AB in culture medium) against SWCNT concentration. Another assay measuring NR uptake also showed SWCNT's ability to quench NR emission and was described as a function of SWCNT concentration. The MTT assay used in the cited study was found to interact with CNT. The reduction in MTT was associated with SWCNT concentration (absorption reduction was higher with increasing SWCNT concentration). For the WST-1 assay, it was concluded that the reduction in WST-1 absorbance was dependent on nanotube concentrations above 0.0125 mgml−1. Casey et al. concluded that Coomassie, AB, NRU, MTT and WST-1 assays were not appropriate for the cytotoxicity analysis of carbon nanotubes [26, 27].

Limitations of MTT in cytotoxicity studies on graphene and graphene-related materials have been demonstrated in numerous publications. CCK-8 (tetrazolium-8-[2-(2-methoxy-4-nitrophenyl)- 3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium] monosodium salt) assays are an attractive alternative for the MTT test. Evaluation of graphene adsorption was based on cell-free adsorptive experiments and demonstrated a gradual reduction of MTT to 93% during 2-h incubation, whereas CCK-8 was significantly reduced to 73% after exposure to graphene for 2 h. The intensity of graphene adsorption to MTT and CCK-8 was time-dependent. The quantity of the CCK-8 reagent absorbed by graphene was higher than that of MTT. It was reported that the π-π conjugated system of the CCK-8 molecule was much stronger than that of MTT due to three benzene rings and one five-membered heterocycle. MTT contains only two benzene rings and two five-membered heterocycles. Another reason for that process is that benzene ring groups strongly affect the adsorption. It was also noted that graphene can suppress the fluorescence effect caused by electron transfer from the dye molecule to the graphene surface. Although MTT and CCK-8 reagents are not fluorescent dyes, both of them display a positive electron on the molecules, similar to some known fluorescent dyes. Thus, it is possible that electron transfer occurs during the incubation with graphene and interferes with the dye molecule that contacts the enzymes. Jiao et al. noted that CCK-8 molecules can be more significantly disturbed by graphene than by the MTT reagent. Additionally, optical properties of graphene may also result in the loss of light signals used for detection in assays *in vitro* [28].

Cytotoxicity can also be determined using the lactate dehydrogenase (LDH) assay. The LDH assay is performed to exclude interactions between nanomaterials and fluorophore molecules [19]. The LDH assay, similar to the MTT assay, is a colorimetric method; thus, it can also interact with nanoparticles (e.g., CNT). Formazan crystals can be absorbed on the surface of MWNT (multi-walled nanotubes) through a strong π-π stacking interaction. The analysis of Ali-Boucetta et al. [19] proved that media containing the released LDH showed the same absorbance (at 490 nm) as MWNT:F127 (multi-walled nanotubes dispersed in the presence of Pluronic 127) dispersion in culture media. Ali-Boucetta et al. [19] proposed LDH assay modification that would eliminate the potential risk of interference of assay components with NPs (modified method vs. traditional procedure is presented in **Figure 1**).

the presence of Cu-40 and AG-35 in a dose-dependent manner. The effect of TiO<sup>2</sup>

results with caution because of metal-catalysed oxidation [29].

**Figure 1.** Schematic of the original (A) and modified LDH assay (B) [19].

NPs was not significant. In conclusion, these authors underlined the necessity to interpret the

Wang et al. [30] proposed modifying the LDH assay that would correct the erroneous results caused by potential interference of nanotubes with reporter chromophore, resulting in its adsorption on nanoparticle surface. The idea of this modification is based on the incubation of LDH derived from a known number of cells (e.g., DH82 macrophage cells) or a purified LDH standard (lactic dehydrogenase enzyme purified from rabbit muscle) with a precise amount (at different concentrations ranging from 5 to 100 μgml−1) of SWCNT or SWCNT-ox (carbon nanohorns). This additional procedure enables the quantification of the effects of NPs on the LDH level. The results obtained by Wang and co-workers clearly demonstrated that LDH concentrations decreased with increasing CNT concentration (at a wavelength of 490 nm). On the other hand, the 580 nm peak was elevated at the increased maximum absorbing wavelength. Based on the observation and regression analyses performed by Wang et al. [30], it was suggested that


General Cytotoxicity and Its Application in Nanomaterial Analysis

http://dx.doi.org/10.5772/intechopen.72578

183

In the experiment of Han et al. [29], copper (Cu-40), silver (Ag-35 and Ag-40) and titanium dioxide (TiO2 -25) were used to validate the popular assay. It was found that LDH was inactivated in General Cytotoxicity and Its Application in Nanomaterial Analysis http://dx.doi.org/10.5772/intechopen.72578 183

**Figure 1.** Schematic of the original (A) and modified LDH assay (B) [19].

(0.003–0.800 mgml−1). As regards the AlamarBlue™ analysis (fluorescent measurements of all single-walled carbon nanotube solutions), quenching was monitored as a function of SWCNT concentration and plotted as an emission ratio at 595 nm by 540 nm excitation for the AB assay (5% solution of AB in culture medium) against SWCNT concentration. Another assay measuring NR uptake also showed SWCNT's ability to quench NR emission and was described as a function of SWCNT concentration. The MTT assay used in the cited study was found to interact with CNT. The reduction in MTT was associated with SWCNT concentration (absorption reduction was higher with increasing SWCNT concentration). For the WST-1 assay, it was concluded that the reduction in WST-1 absorbance was dependent on nanotube concentrations above 0.0125 mgml−1. Casey et al. concluded that Coomassie, AB, NRU, MTT and WST-1

assays were not appropriate for the cytotoxicity analysis of carbon nanotubes [26, 27].

the loss of light signals used for detection in assays *in vitro* [28].

(modified method vs. traditional procedure is presented in **Figure 1**).

ide (TiO2

182 Cytotoxicity

Limitations of MTT in cytotoxicity studies on graphene and graphene-related materials have been demonstrated in numerous publications. CCK-8 (tetrazolium-8-[2-(2-methoxy-4-nitrophenyl)- 3-(4-nitrophenyl)-5-(2,4-disulfophenyl)-2H-tetrazolium] monosodium salt) assays are an attractive alternative for the MTT test. Evaluation of graphene adsorption was based on cell-free adsorptive experiments and demonstrated a gradual reduction of MTT to 93% during 2-h incubation, whereas CCK-8 was significantly reduced to 73% after exposure to graphene for 2 h. The intensity of graphene adsorption to MTT and CCK-8 was time-dependent. The quantity of the CCK-8 reagent absorbed by graphene was higher than that of MTT. It was reported that the π-π conjugated system of the CCK-8 molecule was much stronger than that of MTT due to three benzene rings and one five-membered heterocycle. MTT contains only two benzene rings and two five-membered heterocycles. Another reason for that process is that benzene ring groups strongly affect the adsorption. It was also noted that graphene can suppress the fluorescence effect caused by electron transfer from the dye molecule to the graphene surface. Although MTT and CCK-8 reagents are not fluorescent dyes, both of them display a positive electron on the molecules, similar to some known fluorescent dyes. Thus, it is possible that electron transfer occurs during the incubation with graphene and interferes with the dye molecule that contacts the enzymes. Jiao et al. noted that CCK-8 molecules can be more significantly disturbed by graphene than by the MTT reagent. Additionally, optical properties of graphene may also result in

Cytotoxicity can also be determined using the lactate dehydrogenase (LDH) assay. The LDH assay is performed to exclude interactions between nanomaterials and fluorophore molecules [19]. The LDH assay, similar to the MTT assay, is a colorimetric method; thus, it can also interact with nanoparticles (e.g., CNT). Formazan crystals can be absorbed on the surface of MWNT (multi-walled nanotubes) through a strong π-π stacking interaction. The analysis of Ali-Boucetta et al. [19] proved that media containing the released LDH showed the same absorbance (at 490 nm) as MWNT:F127 (multi-walled nanotubes dispersed in the presence of Pluronic 127) dispersion in culture media. Ali-Boucetta et al. [19] proposed LDH assay modification that would eliminate the potential risk of interference of assay components with NPs

In the experiment of Han et al. [29], copper (Cu-40), silver (Ag-35 and Ag-40) and titanium diox-


the presence of Cu-40 and AG-35 in a dose-dependent manner. The effect of TiO<sup>2</sup> -25 and Ag-40 NPs was not significant. In conclusion, these authors underlined the necessity to interpret the results with caution because of metal-catalysed oxidation [29].

Wang et al. [30] proposed modifying the LDH assay that would correct the erroneous results caused by potential interference of nanotubes with reporter chromophore, resulting in its adsorption on nanoparticle surface. The idea of this modification is based on the incubation of LDH derived from a known number of cells (e.g., DH82 macrophage cells) or a purified LDH standard (lactic dehydrogenase enzyme purified from rabbit muscle) with a precise amount (at different concentrations ranging from 5 to 100 μgml−1) of SWCNT or SWCNT-ox (carbon nanohorns). This additional procedure enables the quantification of the effects of NPs on the LDH level. The results obtained by Wang and co-workers clearly demonstrated that LDH concentrations decreased with increasing CNT concentration (at a wavelength of 490 nm). On the other hand, the 580 nm peak was elevated at the increased maximum absorbing wavelength. Based on the observation and regression analyses performed by Wang et al. [30], it was suggested that LDH assay results should be verified by calibration curves in the presence of different SWCNT concentrations (in the range of 5–100 μgml−1) at two wavelengths, 580 and 490 nm, for each LDH assay. This procedure more accurately determines cellular toxicity values [30].

NPs at higher concentrations tend to form aggregates (agglomerates) under artificial conditions of *in vitro* cell cultures [16]. Many experiments found that NPs that form aggregates were not as cytotoxic as the same NPs at lower concentrations. Lower concentrations of NPs resulted in better internalisation and biodistribution in the circulatory system and organs [3]. Aggregation process is caused by magnetic attraction forces (types 1, 2 and 4), hydrophobic-hydrophobic interactions (for type 1) or hydrogen bonding between hydroxyl groups [39]. Different types of nanomaterials exhibit different tendency to form aggregates in PBS and culture media. CNTs have a strong tendency to agglomerate due to van der Waals interactions [40]. Metal oxides display higher tendency to form agglomerates in comparison to MWCNT. Metal oxides differ in size but were of similar size in PBS environment; thus, it was concluded that surface chemistry and/or the environment had a more significant effect on the aggregate formation process [14]. The size of aggregates may be dependent on the concentration and they tend to be slightly larger in culture media than in PBS. Moreover, monovalent and divalent cations may affect aggregate formation. Adsorption of media components, serum proteins and Ca2+ on nanoparticle surface determines NP aggregations and size distribution [14]. Agglomeration leads to cytotoxicity reduction, because of lower availability of inorganic NPs in contact with cells. In addition, the size of aggregates prevents their cellular internalisation [39]. Studies based on silica nanoparticles indicated that minimization of NP aggregation could be obtained by introducing an optimum balance of inert (e.g., methyl phosphonate) and active (e.g., hydroxyl and aldehyde) functional groups to the surface [41].

General Cytotoxicity and Its Application in Nanomaterial Analysis

http://dx.doi.org/10.5772/intechopen.72578

185

The protein layer of several nanometres on particle surface is called protein corona and it can be divided into a peripheral *soft* corona (SC)—dynamic protein exchanges with the surrounding medium—and a *hard* corona (HC)—a layer of more or less temporal constant composition (**Figure 2**, **Table 2**) [34, 42, 43]. In blood plasma, the surface of nanoparticles mainly adsorbs proteins, but some minor traces of lipids have also been found in the corona structure. Adsorption of proteins on the nanoparticle surface is the result of protein-nanoparticle binding affinities and protein-protein interactions. *Hard* corona interacts directly with the nanomaterial surface. *Soft* corona proteins interact with the hard corona via weak proteinprotein interactions. Interestingly, the corona on the NP surface does not completely mask the nanomaterial surface or its functional groups [43]. The formation of protein corona and its thickness is a parameter that is also dependent on protein concentration, temperature, dura-

tion of particle-protein interaction, serum concentration and shear stress [34, 44].

and the targets on the cell surface [44, 47].

Protein corona formation strongly affects cellular uptake mechanism, cell-nanoparticle interactions, intracellular location as well as cellular response (e.g., biocompatibility) [34, 35, 44]. The protein corona on the NP surface is hypothesised to hinder interactions of nanoparticle ligands

The study of Mirshafiee et al. [44] found that the protein corona formed on BCN-NPs (NPs functionalized with bicyclononyne) incubated in medium with 10% serum and 100% serum consisted of abundant proteins, such as chain A, a novel allosteric mechanism in haemoglobin, fetuin, haemoglobin foetal subunit beta or apolipoprotein A-II precursor. It was also reported that ≥88% of proteins in BCN-NP coronas had a molecular weight below 30 kDa. Even relatively low molecular weight proteins created corona that significantly reduced NP

Smith et al. [31] presented a simple protocol modification of the LDH analysis, which included membered additional conditional-specific controls. This modification enables accurate simultaneous measurement of the effects of death and growth inhibition. The additional step provides quantitative information that can be useful in applications such as drug discoveries [31]. Another approach in LDH assay analysis was proposed in the experiments of Chan et al. [32]. Modification of the LDH protocol allows to detect necrosis, including secondary necrosis [32].

In addition, calcein AM (CAM), Live/Death, neutral red, CellTiter®, Aqueous One (96 AQ), Alamar Blue (AB), CellTiter-Blue® (CTB), CytoToxOne™, and flow cytometry were used to determine their utility in nanoparticle toxicity evaluation. In the cited study, it was found that the results of the assay that depended on direct staining of cells were difficult to interpret, because of dye interactions with NPs. The 96 AQ assay proved optimal for NP analysis. The results were not significantly altered by interactions between the test factor and reagents in the assay [16].

Herzog et al. [33] suggested the clonogenic assay to determine real cytotoxic effect on cell cultures due to the false results (positive or negative) that may occur in NP testing. The clonogenic assay (colony formation assay) is based on the ability of a single cell to form a colony. The latter study was based on the ability of A549, BEAS-2B (normal human bronchial epithelial cells) and HaCaT (normal human keratinocytes) cells to form colonies after 7 (for HaCaT cells) and 10 (for A549 and BEAS-2B) days of incubation with SWCNT (HiPco®). The EC50 comparison showed that the A549 cell line was more resistant than the other two lines. On the other hand, the analysis based on colony size showed that A549 was more sensitive than HaCaT cells. Although the clonogenic assay provided more accurate results than colorimetric tests, it did not become popular because it was too time-consuming for rapid toxicity screening [19, 33].
