When cells were exposed to MNPs, most nanoparticles were first adhered to the surface, internalized to the cells by endocytosis, and accumulated in digestive vacuoles [³²]. Our TEM images results showed that all three kinds of MNPs were incorporated into BEL-7402 cells after 24-h incubation at 0.5 mg/mL of concentration. The MNPs were distributed on the cell membrane and inside of cell. Some MNPs were observed enclosed by the invaginated cell membrane (Figure 1A), suggesting that endocytosis may involve the MNPs internalization process [²¹]. Lysosomes containing MNPs and swollen mitochondria with lysing cristae were present in MNPs treated cells (Figure 1C-D), coinciding with some results obtained in other MNPs [²¹,³³]. Some cells showed chromatin condensation, typical of apoptotic cell death, and plenty of cytoplasmic vacuoles (Figure 1C-E). Treating with OA-Fe₃O₄ induced less cell damage than that of Fe₃O₄, while the C-Fe causes serious cell damage. Untreated cells had none of these features (Figure 1B).

3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenyltetrazolium bromide and lactate dehydrogenase assays are frequently adopted in assessing nanoparticle toxicity. These assays are used in drug studies, but can lead to aberrant results when using nanoparticles as they can sometimes interfere with the assay components or the readout [³⁴]. Due to its convenience and great sensitivity, recently, the WST-1 assay has become a very popular cytotoxicity assay in the nanotoxicity study [²²]. After 24 h exposure at varying doses of Fe₃O₄, OA-Fe₃O₄, and C-Fe MNPs, BEL-7402 cell viabilities detected by the WST-1 assay resulted in explicit dose-dependent reduction (Figure 2). The viabilities of BEL-7402 cells exposed to all three types of MNPs were above 60% at the concentration of 0.1 mg/mL and below. When the MNPs concentrations increased more than 1 mg/mL, the cell viabilities dropped to below 60%. The viabilities of cells exposed to Fe₃O₄ were lower than that to OA-Fe₃O₄, but higher than to C-Fe at all concentrations, which were correlated with the TEM observations. The cytotoxicity is thus very likely caused by particle overload to cells [³⁵]. It is well known that the surface of BEL-7402 cells is negatively charge. The MNPs absorbed by the cells reduced with the decrease in positively charged surfaces of MNPs due to the electrostatic effects, which could affect the amount of MNPs entering the cells and further affect cytotoxicity. In our results, the surface charge of Fe₃O₄, OA-Fe₃O₄, and C-Fe were 14.4, 4.5, and 23.7 mV, respectively, which were consistent with WST-1 data trend.

The effects of various concentrations (0.05, 0.1, and 1 mg/mL) of the three kinds of MNPs on cell cycle progression and population distribution in BEL-7402 cells were determined by flow cytometry. MNPs-induced effects were detected by comparing the cell cycle profiles between MNP-treated and -untreated cells. Results demonstrated that all three types of MNPs were able to affect cell cycle distribution of BEL-7402 cells (Table 1). After treated with Fe₃O₄ and OA-Fe₃O₄ MNPs at three different concentrations (0.05, 0.1, and 1 mg/mL), the ratio of G0/G1 phase cells increased 3.42%, 18.70%, 28.78% and 4.37%, 3.46%, 15.71% compared with control, respectively. As for C-Fe, 2.85%, 3.21%, and 9.34% G2 phase cell increases were observed. A similar report also showed that single-walled carbon nanotubes also caused a G2 phase arrest in PC12 cells [³⁶]. Therefore, the mechanism of C-Fe MNPs on the cell cycle might be different with that of Fe₃O₄, and OA-Fe₃O₄ MNPs.

Cells with reversibly damaged DNA will accumulate in G1, S, or G2/M phase [³⁶], while cells that carry irreversibly damaged DNA will undergo apoptosis [³⁷,³⁸]. Hence, it is necessary to further analyze the cell apoptosis to fully interpret the toxic effects of MNPs on BEL-7402 cells.

To assess the extent and mode of cell death induced by MNPs, Annexin-V/propidium iodide (PI) staining was performed. Externalization of phosphatidylserine (PS) seems to be a general feature of early stage apoptosis. Annexin V which has a strong Ca²⁺-dependent affinity for PS [³⁹] was used to measure the apoptotic rate of BEL-7402 cells in response to the treatment of MNPs. The BEL-7402 cells were labeled with annexin V-fluorescein isothiocyanate (FITC)/PI. The Annexin V^-/PI^-population was regarded as normal cells, while positive staining just for Annexin V was used as a measure of early apoptosis and Annexin V^-/PI⁺ was related to late apoptosis or necrosis [⁴⁰]. Statistical data were extracted from the dot plots using WinMDI software [³⁷]. As shown in Figure 3, compared with the untreated cells, a significant increase in the ratio of apoptosis cell was observed in Fe₃O₄ and C-Fe MNPs (0.05 mg/mL) treated cells (P < 0.05, the probability values of P < 0.05 were considered as statistics significance). At high concentration (1 mg/mL), all MNPs cause serious cell apoptosis (P < 0.01). Besides, a dose-dependent apoptosis rate was observed in all three types of MNP-treated cells. Moreover, the apoptosis rate of cells exposed to three types of MNPs would be: C-Fe > Fe₃O₄ > OA-Fe₃O₄, in the same concentration.

This apoptosis result is consistent with cytotoxicity trends shown in WST-1 assay. The mechanisms of cytotoxic effects of MNPs on BEL-7402 cells may be implemented through cell cycle arrest and inducing apoptosis.

Apoptosis is a tightly controlled process in which cell death is executed through the activation of specific signaling pathways [⁴¹,⁴²]. Although it is well established that many organelles contribute to apoptosis, extensive research shows that nanoparticles induced cell apoptosis via mitochondria-dependent pathway [⁴³,⁴⁴]. As an indicative of mitochondria involvement in the apoptosis, the apparently swollen mitochondria with lysing cristae were observed by TEM (Figure 1). Therefore, we speculate that BEL-7402 cell apoptosis was induced by MNPs through mitochondria-dependent pathway.

The mitochondrion is an important organelle involved in apoptosis. The loss of MMP is putatively the initial event leading to apoptosis [⁴⁵]. To further elucidate the molecular mechanism of MNPs-induced apoptosis in BEL-7402, we examined loss of MMP using flow cytometry. As illustrated in Figure 4, after 24-h exposure to MNPs (0.05 mg/mL) for 24 h, a significant decrease in MMP was only observed in C-Fe-treated group (P < 0.05), while at high concentration (1 mg/mL), significant decrease of MMP occurred in all three MNPs-treated groups (P < 0.05).

Mitochondrial membrane permeability is regulated through a family of proto-oncogenes. Bax is an important pro-apoptotic protein of the Bcl-2 family members [⁴⁶]. High level of Bax can translocate to the outer mitochondrial membrane (OMM) and insert into the OMM. Then, Bax forms oligomers that are thought to be important in the formation of the mitochondrial permeability transition pore (PTP) [⁴⁷,⁴⁸]. The opening of the mitochondrial PTP can lead to a release of cytochrome C, which is a key event in apoptosis via the mitochondria-mediated pathway [⁴⁹]. We examined expression of Bax and cytochrome C by Western blot. As shown in Figure 5A-B, after 24 h exposure at low concentration (0.05 mg/mL) of MNPs, the expression of Bax protein slightly increased, while without significant differences (P > 0.05). At high concentration (1 mg/mL), Bax protein expression in Fe₃O₄, OA-Fe₃O₄, and C-Fe MNPs-treated groups were about 1.94, 1.89, and 2.43 times compared with the control group, respectively. As for cytochrome C, the protein expression slightly decreased at the low concentration, while without significant differences (P > 0.05). At high concentration, the cytochrome C protein expression of all treated groups decreased dramatically, which is consistent with tendency of MMP. Based on the results mentioned above, we concluded that the three types of MNPs could induce Bax expression, further open PTP, and the PTP opening led to the release of cytochrome C from mitochondria. Once released from the mitochondria, cytochrome C combines with procaspase-9 to form the "apoptosome", which further activates caspase-3 [^50-52]. Caspase-3 has been identified as a key mediator of apoptosis of mammalian cells [⁵³]. Its activity is considered to be an appropriate measure of cytotoxic responsiveness [⁵⁴]. We investigated the activaty of caspases-3 in BEL-7402 after exposure to MNPs for 24 h. As shown in Figure 6, we found that all three types of MNPs can activate caspase-3 in a dose-dependent manner. At low concentration (0.05 mg/mL), the activity of caspase-3 of the experimental groups increased, with significant differences found in Fe₃O₄- and C-Fe-treated groups (P < 0.05). The activity of caspase-3 was significantly increased in all experimental groups at high concentration (1 mg/mL) (P < 0.05).

To sum up, our results indeed suggested that all three types of MNPs can induce apoptosis in BEL-7402 cells through mitochondria-dependent pathway. Moreover, the influence potency of MNPs on the mitochondria-dependent apoptosis would be: C-Fe > Fe₃O₄ > OA-Fe₃O₄, and all in a dose-dependent manner.

RPMI-1640 and fetal bovine serum were purchased from Gibco, Invitrogen Corp., Carlsbad, CA, USA. PI and RNase I were obtained from Sigma, St. Louis, MO, USA. Alexa Fluor^® 488 annexin V/Dead Cell Apoptosis Kit was obtained from Invitrogen, USA. The primary antibodies to Bax, cytochrome C, and β-actin were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA). The goat anti-Mouse IgG-HRP, mouse anti-rabbit IgG-HRP, and Potent ECL kit were purchased from Multisciences, Hanzhou, China. Caspase-3/CPP32 Colorimetric Assay Kit and Mitochondria/Cytosol Fractionation Kit were purchased from BioVision, Mountain View, CA, USA. Total Protein Extraction Kit and BCA Protein Assay Kit were obtained from Applygen Technologies Inc., Beijing, China. The lipophilic cationic dye JC-1 (5, 5, 6, 6-tetrachloro-1, 1, 3, 3-tetraethylbenzimidazol-carbocyanine iodide) was obtained from ChemoMetec, Allerød, Denmark. WST-1 Cell Proliferation and Cytotoxicity Assay Kit was purchased from Beyotime Institute of Biotechnology, Haimen, China. All other reagents are analytical or cultured grade purity.

Human hepatoma BEL-7402 cell line was a gift kindly provided by Medical College of Xiamen University (Xiamen, China). The cells were cultured in RPMI-1640 medium supplemented with 10% heat-inactivated fetal bovine serum. Incubation was carried out at 37°C in a humidified 5% CO₂ incubator. For all experiments, the cells were in the exponential growth phase. The MNPs used in this study were: (1) Fe₃O₄ MNPs, purchased from Aladdin (Shanghai, China); (2) OA-Fe₃O₄ MNPs, purchased from Jinke (Maanshan, China); (3) C-Fe MNPs, purchased from Junye (Shenzhen, China). The purity of three types MNPs are 99.9% and the size distribution of particles are 10-30 nm. Nanoparticle stock suspensions (10 mg/mL) were prepared by UV-sterilization and dispersing a known weight of nanoparticles in RPMI-1640 medium under ultrasonication. The stock suspensions were sonicated for 20 min to distribute the particles, and then dilutions were made in complete media to achieve desired testing concentrations. The test suspensions were sonicated for 20 min before use. Untreated controls were exposed to complete media only, and processed identical to the exposed cells.

Cells (2 × 10⁶) were seeded into 100-cm² petri dishes. Cells were allowed to attach for 24 h and were then treated with each MNPs test suspensions for 24 h in a concentration of 0.5 mg/mL. Then, the cells were collected and fixed with 2.5% glutaraldehyde buffered in 0.1 M PBS overnight at 4°C. The samples were washed with PBS, and post-fixed in 1% osmium tetroxide at 4°C for 1 h. After dehydration in series concentrations of ethanol and infiltration in acetone, cells were embedded in Epon 812, and ultra-thin sections cut with glass knives were stained with uranyl acetate and lead citrate, and viewed under JEM 2100 TEM (JEOL, Tokyo, Japan).

To determine cell toxicity/viability, BEL-7402 cells (0.5 × 10⁴, 100 μL) were plated onto 96-multiwell plates (Costar, Corning, NY, USA) and incubated for 24 h. Then, cells were exposed to various concentrations (0.01-2 mg/mL) of each MNPs test suspensions for 24 h. Afterwards, the old media was discarded and replaced with 100 μL of new complete media. WST-1 solution (10 μL) was added to each well, followed by incubation for 2.5 h. Absorbance at 490 nm (reference at 630 nm) was measured by a spectrophotometric microplate reader (Bio-tek ELX800, BioTek Instruments, Winooski, VT1, USA). A negative control was provided using the culture medium without the nanoparticles. Each of the particle concentrations and the controls was seeded in eight wells. The percentage cell viability was calculated in term of absorbency in cells treated with MNPs relative to that in cells exposed to culture media alone.

A cell cycle assay was carried out by staining the DNA with PI and analyzing the fluorescence using flow cytometry. Following exposure of the BEL-7402 cells to each MNPs for 24 h, any damaged cells can detach from the plate and become suspended in the medium, necessitating medium storage. Briefly, the cells were harvested, washed with PBS, and fixed in ice-cold 70% of ethanol at -20°C before use. After resuspension, cells were washed, and incubated with 100 μL PI (400 μg/mL) and 100 μL RNase I (1 mg/mL) at 37°C for 15 min. Cells were analyzed with an EPICS XL flow cytometer (Beckman Coulter Inc., Fullerton, CA, USA) and the data were consequently evaluated by Mod-Fit (Verity Software, Topsham, ME, USA).

Apoptosis was evaluated using Alexa Fluor^® 488 Annexin V/PI Apoptosis Kit. BEL-7402 cells were treated with two concentrations (0.05 and 1 mg/mL) of each MNPs for 24 h. After exposure, the cells (5-10 × 10⁴) were harvested, washed and resuspended with PBS. The Annexin V/PI staining of cells followed the manufacturer's instructions. Then the samples were analyzed with EPICS XL flow cytometer (Beckman Coulter, USA). The results were expressed as the number of apoptotic cells per thousand cells counted.

The mitochondrial membrane potential was measured by flow cytometry using JC-1 dye. JC-1 changes its fluorescence from green at 535 nm (monomer state) to orange at 590 nm (aggregate state) as it enters in mitochondria of intact cells. When the mitochondrial membrane potential is affected, JC-1 returns to its green monomeric state. All procedures were carried out according to the manufacturer's instructions. The cell treatments were the same with "Detection of apoptosis by annexin V assay" section. In brief, approximately 2 × 10⁶ cells were harvested, washed, resuspended in PBS (1 mL) and stained with 12.5 μL JC-1(200 μg/mL) for 15 min at 37°C in the dark. Both fluorescences emitted by the cells were monitored by flow cytometry and the ratio orange/green fluorescence was calculated. The data was determined by analyzing 10,000 cells using an EPICS XL flow cytometer (Beckman Coulter, Fullerton, CA, USA), and Cell Quest software (Becton Dickinson, San Jose, CA, USA).

The cell treatments were the same with "Detection of apoptosis by annexin V assay" section. Approximately 1 × 10⁷ cells per sample were harvested. The protein samples of Bax and cytochrome C were extracted using Total Protein Extraction Kit and Mitochondria/Cytosol Fractionation Kit, respectively. Protein contents were quantified using the BCA protein assay kit and stored at -70°C. Protein (20 μg per lane) was resolved by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and transferred to nitrocellulose membranes (PVDF, Millipore Corporation, Billerica, MA, USA). The transblotted membrane was washed, blocked, and incubated at 4°C overnight with anti-Bax antibody and anti-cytochrome C antibody, respectively. Immunodetection with the secondary HRP-conjugated antibody and chemiluminescence using Potent ECL Kit were performed according to the manufacturer's protocol. Equal protein loading was verified by probing with anti-β actin antibody. Densitometric analysis for the blots was performed with Bandscan image software.

The activity of caspase-3 was determined using caspase-3/CPP32 Colorimetric Assay Kit. The cell treatments were the same with "Detection of apoptosis by annexin V assay" section. All procedures were carried out according to the manufacturer's instructions. Briefly, 2 × 10⁷ BEL-7402 cells were lysed by the solution provided in the assay kit and the protein concentration was measured using BCA protein assay kit. For caspase-3 activity assay, equal amounts of total cell lysates were mixed with a caspase-specific substrate DEVD-pNA in a 96-well plate in triplicate. After incubation at 37°C for 2 h, the caspase-3-mediated cleavage of DEVD-pNA into free pNA was measured using spectrophotometric microplate reader (Bio-tek ELX800, USA) at 405 nm. The results were expressed as absorbance compared with control.

All results were expressed as mean values ± S.D. Statistical analysis was performed according to the Student's t test. The probability values of P < 0.05 were considered as significant.

After BEL-7402 cells were treated with MNPs (0.05 and 1 mg/mL) for 24 h, cell cycle assay was carried out by PI staining using flow cytometry. *P < 0.05 vs. control; **P < 0.01 vs. control.