Effect of secondary particle size of nickel oxide nanoparticles on cytotoxicity in A549 cells

Abstract

The effect of nanoparticle type, shape, as well as primary and secondary particle size on toxicity remains poorly characterized. In this study, suspensions of nickel oxide (NiO) nanoparticles with the same primary particle size (< 50 nm) but different secondary particle sizes were prepared, and their cytotoxicity was investigated. A planetary ball mill wet nanopulverizer with zirconium milling balls of decreasing sizes (φ: 0.5, 0.1, and 0.05 mm) yielded NiO nanoparticles of decreasing mean particle size (310.4 ± 6.7, 172.0 ± 2.8, and 102.0 ± 0.5 nm). Stock solutions were diluted to various concentrations in 10% heat-inactivated fetal bovine serum containing minimum essential medium, and shown to have the same primary particle size, but different secondary particle sizes. Tests with A549 cells revealed that cytotoxicity increased with increasing secondary particle size: milling ball diameter φ 0.05 mm (IC₅₀: 148 μg/mL) < φ 0.1 mm (IC₅₀: 83.5 μg/mL) < φ 0.5 mm (IC₅₀: 33.4 μg/mL). Uptake experiments indicated that the intracellular amount of Ni increased with increasing secondary particle size. In summary, the present findings show that differences in secondary particle size affected the cytotoxicity of NiO suspensions, which could be ascribed at least in part to differences in the amount of NiO taken up by the cells.

INTRODUCTION

Nanoparticles are generally defined by a primary particle size < 100 nm (Whatmore, 2006). They have become important components of industrial and cosmetic products, catalysts, paints, and coatings. However, there are concerns about nanoparticles-specific toxicity originating from exposure during the manufacturing process or in consumer products (Schmidt, 2009; Ema et al., 2010). Indeed, in a work-related accident in China (Song et al., 2009), seven female workers aged 18 to 47 years were continually exposed to nanoparticles between January 2007 and April 2008 in a paint manufacturing plant. Nanosilica accumulated in lung tissue was considered the probable cause of lung damage and even death in two of the cases (Song et al., 2011). Tests on nanoparticle safety have been performed in vivo and in vitro. For some nanoparticles, the biological effects depended on chemical composition, size, and physical properties (Dhawan and Sharma, 2010).

Metal oxide nanoparticles are industrially important and have been widely tested in vitro. Yuan et al. (2010) generated SiO₂ nanoparticles of different primary particle sizes and detected cytotoxicity only with particles of 20–80 nm, but not with those ≥ 140 nm. However, primary particles of approximately equal size but different chemical composition have also been reported to differ in cytotoxicity (Karlsson et al., 2009; Xu et al., 2010). In addition, while primary particles are the basic constituents of nanomaterials, secondary particles are aggregates of primary particles that apparently behave as a single particle. According to ISO 26824:2013, agglomerates are also termed secondary particles and the original source particles are termed primary particles. As for the cytotoxic effects of secondary particle sizes, a study evaluating titanium dioxide (TiO₂) nanoparticles with different secondary particle sizes (166 nm and 596 nm), reported that large secondary particles showed stronger cytotoxicity (Okuda-Shimazaki et al., 2010). On the other hand, in vivo tests with rats subjected to endotracheal administration of TiO₂ with equal primary particle size but different secondary particle sizes showed no difference in the induced inflammatory response (Kobayashi et al., 2009).

Nickel oxide (NiO) nanoparticles are industrially important metal oxide nanoparticles. Cytotoxicity is stronger in ultrafine NiO nanoparticles (φ: 20–100 nm) than in fine NiO nanoparticles (φ: 600–2000 nm) (Horie et al., 2009b); however, the effect of equal primary particle size but different secondary particle sizes remains unknown. In the present study, a method for producing NiO suspensions with different secondary particle sizes was attempted and their effect on cytotoxicity was examined.

MATERIALS AND METHODS

Reagents

The following reagents were used: NiO nanoparticles (primary particle size < 50 nm; Sigma-Aldrich, St. Louis, MO, USA), NiCl₂·6H₂O (Wako Pure Chemical Corp., Osaka, Japan), Tween 80 (ICN Biomedical Inc., Costa Mesa, CA, USA), fetal bovine serum (FBS; Intergen, Purchase, NY, USA), non-essential amino acids (NEAA), minimum essential medium (MEM), phenol red-free MEM, 0.05% trypsin-EDTA, phosphate-buffered saline (PBS, pH 7.4; Gibco, Grand Island, NY, USA), Cell Titer 96 AQueous One Solution Reagent (MTS reagent; Promega, Madison, WI, USA), Ni and Ag standard solutions (Wako Pure Chemical Corp.), HNO₃, and HCl (Wako Pure Chemical Corp.). All water used in tests was produced with the Milli-Q Advantage A10 purification system (Millipore, Billerica, MA, USA).

Preparation of NiO nanoparticle suspensions with different secondary particle sizes

Suspensions of NiO nanoparticles were prepared with a planetary ball mill wet nanopulverizer (NP-100; Thinky, Tokyo, Japan). First, 100 mg of NiO nanoparticles were weighed and placed in a zirconium container, together with 2.5 mL water containing 0.1% (w/v) Tween 80 as dispersant. Next, 2.5 g of zirconium balls (0.5, 0.1, and 0.05 mm in diameter) were added, and milling was performed for 2 min in MILL/MIX mode at 2,000 rpm, followed by addition of 7.5 mL water-Tween 80 and mixing in MILL/MIX mode at 400 rpm for 1 min. The zirconium balls were separated from the resulting NiO nanoparticle stock solution (10 mg/mL), which was diluted with 10% FBS-MEM for cytotoxicity testing and other processes.

NiO nanoparticle suspension characterization

The size distribution, mean size, and zeta potential of NiO particles in the suspensions were measured by dynamic light scattering (DLS) and electrophoretic light scattering (laser Doppler method) with an ELSZ-2 analyzer (Otsuka Electronics Co., Ltd., Osaka, Japan), based on three (mean particle size and distribution) or four (zeta potential) iterations.

The concentration of Ni ions was measured in each suspension formed by diluting the stock solution with 10% FBS-MEM. The suspension was centrifuged at 50,000 rpm (approximately 170,000 × g) and 20°C for 1 hr with an angle rotor (P70AT2) in a refrigerated ultracentrifuge (Himac CP65β; Hitachi Koki Co., Ltd., Tokyo, Japan). A 0.5 mL aliquot of supernatant was taken and mixed with 4.5 mL of 5% aqueous HNO₃ to obtain the test solution, which was diluted further with 5% aqueous HNO₃ to an appropriate concentration and passed through a 0.45 μm membrane filter (Sartorius AG, Goettingen, Germany). The Ni ion concentration was measured by inductively coupled plasma mass spectrometry (ICP-MS). As a control, 10% FBS-MEM containing no NiO nanoparticles was used and subjected to the same procedure. These tests were performed with two iterations each (n = 2).

NiO nanoparticle suspension cytotoxicity test

For cytotoxicity testing, A549 human alveolar basal epithelial adenocarcinoma-derived cell line (JCRB Cell Bank, Tokyo, Japan) was incubated in 10% FBS-MEM containing 1% NEAA at 37°C and 5% CO₂. A NiO nanoparticle suspension diluted with 10% FBS-MEM was used as the test liquid. NiCl₂ dissolved in water and then diluted with liquid medium was used to determine Ni ion cytotoxicity. A549 cells were first plated in a 96-well plate (5 × 10³ cells/well), added with test liquid after 24 hr, and incubated for 48 hr. The medium was then removed, and 100 μL of phenol red-free MEM plus 20 μL MTS reagent were added, followed by a reaction at 37°C for 1 hr in a 5% CO₂ atmosphere incubator. The generated formazan was measured at 490 nm with a microplate reader (SH-9000; Corona Electric Co., Ltd., Ibaraki, Japan). Cell viability was calculated relative to that in wells containing growth medium alone and was based on three iterations (n = 3).

Measurement of NiO nanoparticle uptake by cells

To measure NiO nanoparticle uptake, A549 cells were plated in a 6-well plate (2 × 10⁵ cells/well), then exposed to NiO nanoparticles (1–25 μg/mL) after 24 hr, and incubated for 24 and 48 hr. NiCl₂ and medium alone under the same conditions were used for comparison. After incubation, the medium was removed, each well was repeatedly washed with PBS, and cells were detached by trypsinization. The recovered cells were washed repeatedly with PBS and counted. HCl (2%) was added to the recovered cells, the suspension was heated at 60°C for 2 hr, diluted with 5% aqueous HNO₃ to an appropriate concentration, and then filtered through a 0.45 μm membrane (Sartorius AG). The Ni concentration in the filtrate was measured by ICP-MS.

ICP-MS measurement

ICP-MS was performed with an Agilent 7500ce instrument (Agilent Technologies, Inc., Santa Clara, CA, USA) under the following conditions: high-frequency output 1,500 W; plasma gas (Ar) 15 L/min; carrier gas (Ar) 0.7 L/min; makeup gas 0.33 L/min; collision gas (He) 5 mL/min; sampling position 8 mm; spray chamber temperature 2°C; integral time 0.1 sec/element; and three iterations/measurement. The standard solution was serially diluted with 5% aqueous HNO₃ and used as the working solution. The Ag standard solution was diluted with 5% HNO₃ to 5 μg/L and used as the internal standard liquid. Measured mass-to-charge ratios were 60 m/z (Ni) and 107 m/z (Ag).

RESULTS AND DISCUSSION

NiO nanoparticle suspensions with different secondary particle sizes

Stock solutions of NiO nanoparticles with different secondary particle sizes were diluted in 10% FBS-MEM suspensions (50, 100, 200, and 400 μg/mL). The mean particle sizes determined right after preparation are listed in Table 1. The particle size distributions of stock solutions and the 10% FBS-MEM suspension (200 μg/mL) are shown in terms of scattering intensity (Fig. 1, left) and particle counts (Fig. 1, right). In the stock solutions, the NiO mean particle diameter tended to decrease with decreasing zirconium ball diameter in the order 0.5 > 0.1 > 0.05 mm. In the diluted suspensions, the mean particle size was larger and followed a different decreasing order compared to stock solutions. Suspensions prepared with the same zirconium ball size exhibited almost no difference in the average particle size, irrespective of concentration or if left standing at 37°C for 24 hr (Table 1). In both stock solutions and suspensions, the peak shifted toward small particle sizes with decreasing zirconium ball diameter (Fig. 1). Thus, using Tween 80 as a dispersant, it was possible to prepare suspensions having the same primary particle size, but different secondary particle sizes. The results of this study also suggest the possibility of preparing suspensions of various secondary particle sizes for TiO₂ and other metal oxide nanoparticles by using various sizes of grinding balls, indicating the effectiveness of the planetary ball mill wet nanopulverizer. Notably, nanoparticles derived from liquid medium can be detected by DLS analysis even in the absence of metal oxide nanoparticles (Hanagata et al., 2011); however, in the present study, any scattering intensity in 10% FBS-MEM alone was too weak to be detected.

Table 1. Mean diameter (nm) of NiO nanoparticles in stock solutions and 10% FBS-MEM^a.

NiO concentration		Zirconium ball diameter
		φ 0.05 mm			φ 0.1 mm			φ 0.5 mm
10 mg/mLb		102.0	±	0.5	172.0	±	2.8	310.4	±	6.7
400 μg/mLc		154.8	±	1.2	234.7	±	2.2	373.1	±	0.6
	24 hrd	147.2	±	0.5	225.9	±	3.9	367.7	±	11.7
200 μg/mLc		152.6	±	2.5	249.9	±	4.2	411.9	±	13.1
	24 hrd	135.4	±	3.0	219.3	±	0.7	371.2	±	11.4
100 μg/mLc		151.2	±	3.0	258.0	±	1.5	345.1	±	15.4
	24 hrd	132.2	±	2.2	217.1	±	2.0	369.0	±	5.6
50 μg/mLc		135.2	±	1.5	234.5	±	17.8	414.2	±	5.8
	24 hrd	130.4	±	15.2	206.9	±	1.4	348.7	±	11.3

^a Using 0.1% Tween 80 (w/v) as dispersant.

^b Stock solution.

^c 10% FBS-MEM.

^d After 24 hr (37°C).

Fig. 1

Particle size distributions of NiO nanoparticles (left: scattering light distribution; right: particle counts) in stock solution (10 mg/mL) (a, b) and in 10% FBS-MEM (200 μg/mL) (c, d).

The zeta potential of a stock solution prepared with a zirconium ball of φ 0.05 mm was 25.2 ± 0.4 mV (1 mg/mL). In contrast, a suspension formed by diluting the stock solution with 10% FBS-MEM showed a negative zeta potential of –14.1 ± 0.4 mV (200 μg/mL); whereas 10% FBS-MEM alone was –9.6 ± 1.6 mV. In a previous report (Horie et al., 2009a), the zeta potential of metal oxide nanoparticles in liquid medium was also negative, and reflected the trend observed in the present study. The zeta potential of 10% FBS-MEM alone was thought to be mainly due to proteins in the medium, and the zeta potential of NiO nanoparticles in 10% FBS-MEM showed a negative charge because NiO nanoparticles adsorbed those proteins (Horie et al., 2009b).

The Ni ion concentration in the 10% FBS-MEM suspensions is shown in Table 2. The elution rate was 1.0–24% at a Ni ion concentration of 0.82–19 μg/mL right after preparation, and 4.2–46% at a Ni ion concentration of 3.3–36 μg/mL 24 hr later. After 24 hr, no difference was found in the Ni ion concentration or elution ratio, irrespective of zirconium ball diameter. With increasing NiO concentration in the suspension, however, Ni ion seepage into the suspension increased together with its elution rate. The same concentration-dependent increase in ion concentration has been described in prior reports (Horie et al., 2009b), but the cause has remained unclear.

Table 2. Ni ion concentrations and elution ratio in 10% FBS-MEM NiO nanoparticle dispersions.

Zirconium ball diameter	NiO concentration	Right after preparation					After 24 hra
		Ni ion concentration (μg/mL)			Elution ratio (%)		Ni ion concentration (μg/mL)			Elution ratio (%)
φ 0.05 mm	50 μg/mL	2.4	±	0.043	3.1		4.7	±	0.41	6.0
	100 μg/mL	4.9	±	0.23	6.3		8.8	±	0.49	11
	200 μg/mL	8.0	±	0.35	10		18.0	±	1.60	23
	400 μg/mL	19.0	±	1.40	24		36.0	±	2.20	46
φ 0.1 mm	50 μg/mL	1.5	±	0.71	1.9		3.3	±	0.50	4.2
	100 μg/mL	4.1	±	0.89	5.2		6.1	±	0.63	7.8
	200 μg/mL	5.3	±	0.77	6.8		11.0	±	0.95	14
	400 μg/mL	11.0	±	1.00	14		25.0	±	2.20	32
φ 0.5 mm	50 μg/mL	0.82	±	0.084	1.0		3.9	±	0.27	5.0
	100 μg/mL	1.8	±	0.19	2.3		5.7	±	1.5	7.3
	200 μg/mL	4.5	±	0.15	5.7		14.0	±	0.85	18
	400 μg/mL	8.0	±	0.46	10		30.0	±	2.20	38

^a After 24 hr (37°C).

Cytotoxicity of NiO nanoparticle suspensions

Figure 2 shows the cytotoxicity results of NiCl₂ and suspensions prepared with 10% FBS-MEM. As the NiO secondary particle size increased, the cytotoxicity curve shifted toward lower concentrations. The half maximal inhibitory concentration (IC₅₀) values of the suspensions calculated from the cytotoxicity curves were φ 0.05 mm (148 μg/mL), φ 0.1 mm (83.5 μg/mL), and φ 0.5 mm (33.4 μg/mL). Thus, the cytotoxicity of NiO nanoparticle suspensions with the same primary particle size increased with increasing secondary particle size. A similar trend has been observed in the cytotoxicity of TiO₂ nanoparticles with different secondary particle size in the human pulmonary endothelial cell line, NCI-H292 (Okuda-Shimazaki et al., 2010). The calculated Ni ion IC₅₀ value was 43 μg/mL.

Fig. 2

Cytotoxicity of NiO at different secondary particle sizes in 10% FBS-MEM (a), and cytotoxicity of NiCl₂ (b). Values are based on the MTS assay at 48 hr after exposure.

Table 3 shows the measured Ni amount in A549 cells after exposure to NiO nanoparticle suspensions with differing secondary particle sizes. For all suspensions, the intracellular Ni quantity increased with increasing exposure concentration or secondary particle size. In contrast, the intracellular concentration of Ni ions was much lower upon exposure to medium alone (control) or NiCl₂.

Table 3. Amounts of Ni in A549 cells.

	Incubationtime (hr)	Exposure concentrationa		Amount of Ni/cell (pg-Ni/cell)
NiO (φ 0.05 mm)	24	1 μg/mL	0.786 μg-Ni/mL	0.14
		10 μg/mL	7.86 μg-Ni/mL	0.18
		25 μg/mL	19.6 μg-NI/mL	3.8
	48	1 μg/mL	0.786 μg-Ni/mL	0.72
		10 μg/mL	7.86 μg-Ni/mL	1.1
		25 μg/mL	19.6 μg-NI/mL	3.6
NiO (φ 0.1 mm)	24	1 μg/mL	0.786 μg-Ni/mL	0.39
		10 μg/mL	7.86 μg-Ni/mL	4.8
		25 μg/mL	19.6 μg-NI/mL	13
	48	1 μg/mL	0.786 μg-Ni/mL	0.62
		10 μg/mL	7.86 μg-Ni/mL	6.7
		25 μg/mL	19.6 μg-NI/mL	15
NiO (φ 0.5 mm)	24	1 μg/mL	0.786 μg-Ni/mL	1.3
		10 μg/mL	7.86 μg-Ni/mL	16
		25 μg/mL	19.6 μg-NI/mL	45
	48	1 μg/mL	0.786 μg-Ni/mL	1.7
		10 μg/mL	7.86 μg-Ni/mL	16
		25 μg/mL	19.6 μg-NI/mL	70
NiCl2	24	1 μM	0.0587 μg-Ni/mL	0.056
		10 μM	0.587 μg-Ni/mL	0.045
		50 μM	2.93 μg-Ni/mL	0.046
	48	1 μM	0.0587 μg-Ni/mL	0.031
		10 μM	0.587 μg-Ni/mL	0.037
		50 μM	2.93 μg-Ni/mL	0.056
10% FBS-MEM only	24			0.045
	48			0.034

^a The left column describes values as NiO or NiCl₂; the right column describes values as Ni.

If all Ni in NiO suspensions used in the cell uptake tests is assumed to be in ionic form, then the Ni concentration would be 0.786–19.6 μg/mL. However, the measured Ni ion concentration in the medium (Table 2) indicates that the actual ratio of Ni ion elution was at the most one-half, thus suggesting that the concentration of Ni ion in the NiO suspension was at the most equal to that in the NiCl₂ solution (0.0587–2.94 μg/mL). These Ni values indicate that the intracellular Ni amount differed between NiO nanoparticle exposure and Ni ion exposure (Table 3). Therefore, it is likely that the Ni detected in A549 cells exposed to the NiO suspensions was not taken up as Ni ion but as NiO. This finding confirms previous reports that cytotoxicity is generated by intracellular Ni ion emission after NiO nanoparticles are taken into the cell (Horie et al., 2009b). The present study also shows that cytotoxicity increases with increasing secondary particle size, which affects NiO nanoparticle uptake so that larger sizes facilitate intracellular uptake and thereby increase cytotoxicity. However, further study is needed to clarify the reason why larger secondary particle sizes result in a higher uptake of NiO nanoparticles.

In conclusion, the present study describes the preparation of NiO nanoparticle suspensions differing in secondary particle size (primary particle size < 50 nm) using 0.1% Tween 80 aqueous solution and a planetary ball mill nanopulverizer with zirconium milling balls of three different sizes (φ: 0.5, 0.1, and 0.05 mm). The smaller the zirconium ball, the smaller the resulting mean NiO nanoparticle size, which enabled the preparation of stock solutions with different secondary particle sizes. Each stock solution was diluted to several concentrations in 10% FBS-MEM to determine mean particle size and particle size distribution. The suspensions were shown to have the same primary particle size, but different secondary particle sizes. A549 tests with these suspensions showed that cytotoxicity increased with increasing secondary particle size. Assessment of cell NiO uptake confirmed that intracellular Ni increased with increasing secondary particle size. These findings demonstrate that secondary particle size affects cytotoxicity of the NiO nanoparticles suspensions prepared in this study, and this effect is likely the result of a differential NiO uptake by the cell.

ACKNOWLEDGMENTS

This study was supported by grants H23-Kagaku-Ippan-006 and H27-Kagaku-Ippan-008 from the Ministry of Health, Labor, and Welfare of Japan.

Conflict of interest

The authors declare that there is no conflict of interest.

REFERENCES

•  Dhawan,  A. and  Sharma,  V. (2010): Toxicity assessment of nanomaterials: methods and challenges. Anal. Bioanal. Chem., 398, 589-605.
•  Ema,  M.,  Kobayashi,  N.,  Naya,  M.,  Hanai,  S. and  Nakanishi,  J. (2010): Reproductive and developmental toxicity studies of manufactured nanomaterials. Reprod. Toxicol., 30, 343-352.
•  Hanagata,  N.,  Zhuang,  F.,  Connolly,  S.,  Li,  J.,  Ogawa,  N. and  Xu,  M. (2011): Molecular responses of human lung epithelial cells to the toxicity of copper oxide nanoparticles inferred from whole genome expression analysis. ACS Nano, 5, 9326-9338.
•  Horie,  M.,  Nishio,  K.,  Fujita,  K.,  Endoh,  S.,  Miyauchi,  A.,  Saito,  Y.,  Iwahashi,  H.,  Yamamoto,  K.,  Murayama,  H.,  Nakano,  H.,  Nanashima,  N.,  Niki,  E. and  Yoshida,  Y. (2009a): Protein adsorption of ultrafine metal oxide and its influence on cytotoxicity toward cultured cells. Chem. Res. Toxicol., 22, 543-553.
•  Horie,  M.,  Nishio,  K.,  Fujita,  K.,  Kato,  H.,  Nakamura,  A.,  Kinugasa,  S.,  Endoh,  S.,  Miyauchi,  A.,  Yamamoto,  K.,  Murayama,  H.,  Niki,  E.,  Iwahashi,  H.,  Yoshida,  Y. and  Nakanishi,  J. (2009b): Ultrafine NiO particles induce cytotoxicity in vitro by cellular uptake and subsequent Ni(II) release. Chem. Res. Toxicol., 22, 1415-1426.
• ISO 26824:2013 Particle characterization of particulate systems – Vocabulary.
•  Karlsson,  H.L.,  Gustafsson,  J.,  Cronholm,  P. and  Möller,  L. (2009): Size-dependent toxicity of metal oxide particles--a comparison between nano- and micrometer size. Toxicol. Lett., 188, 112-118.
•  Kobayashi,  N.,  Naya,  M.,  Endoh,  S.,  Maru,  J.,  Yamamoto,  K. and  Nakanishi,  J. (2009): Comparative pulmonary toxicity study of nano-TiO₂ particles of different sizes and agglomerations in rats: different short- and long-term post-instillation results. Toxicology, 264, 110-118.
•  Okuda-Shimazaki,  J.,  Takaku,  S.,  Kanehira,  K.,  Sonezaki,  S. and  Taniguchi,  A. (2010): Effects of titanium dioxide nanoparticle aggregate size on gene expression. Int. J. Mol. Sci., 11, 2383-2392.
•  Schmidt,  C.W. (2009): Nanotechnology-related environment, health, and safety research: examining the national strategy. Environ. Health Perspect., 117, A158-A161.
•  Song,  Y.,  Li,  X. and  Du,  X. (2009): Exposure to nanoparticles is related to pleural effusion, pulmonary fibrosis and granuloma. Eur. Respir. J., 34, 559-567.
•  Song,  Y.,  Li,  X.,  Wang,  L.,  Rojanasakul,  Y.,  Castranova,  V.,  Li,  H. and  Ma,  J. (2011): Nanomaterials in humans: identification, characteristics, and potential damage. Toxicol. Pathol., 39, 841-849.
•  Whatmore,  R.W. (2006): Nanotechnology-what is it? Should we be worried? Occup. Med. (Lond.), 56, 295-299.
•  Xu,  M.,  Fujita,  D.,  Kajiwara,  S.,  Minowa,  T.,  Li,  X.,  Takemura,  T.,  Iwai,  H. and  Hanagata,  N. (2010): Contribution of physicochemical characteristics of nano-oxides to cytotoxicity. Biomaterials, 31, 8022-8031.
•  Yuan,  H.,  Gao,  F.,  Zhang,  Z.,  Miao,  L.,  Yu,  R.,  Zhao,  H. and  Lan,  M. (2010): Study on controllable preparation of silica nanoparticles with multi-sizes and their size-dependent cytotoxicity in pheochromocytoma cells and human embryonic kidney cells. J. Health Sci., 56, 632-640.