Titanium Dioxide Nanoparticles Induce Endoplasmic Reticulum Stress-Mediated Autophagic Cell Death via Mitochondria-Associated Endoplasmic Reticulum Membrane Disruption in Normal Lung Cells

Nanomaterials are used in diverse fields including food, cosmetic, and medical industries. Titanium dioxide nanoparticles (TiO2-NP) are widely used, but their effects on biological systems and mechanism of toxicity have not been elucidated fully. Here, we report the toxicological mechanism of TiO2-NP in cell organelles. Human bronchial epithelial cells (16HBE14o-) were exposed to 50 and 100 μg/mL TiO2-NP for 24 and 48 h. Our results showed that TiO2-NP induced endoplasmic reticulum (ER) stress in the cells and disrupted the mitochondria-associated endoplasmic reticulum membranes (MAMs) and calcium ion balance, thereby increasing autophagy. In contrast, an inhibitor of ER stress, tauroursodeoxycholic acid (TUDCA), mitigated the cellular toxic response, suggesting that TiO2-NP promoted toxicity via ER stress. This novel mechanism of TiO2-NP toxicity in human bronchial epithelial cells suggests that further exhaustive research on the harmful effects of these nanoparticles in relevant organisms is needed for their safe application.


Introduction
Titanium dioxide nanoparticles (TiO 2 -NP) have widespread applications in various fields and are used in sunscreens, cosmetics, food products, toothpastes, and medical reagents [1][2][3][4]. In recent years, many studies have focused on the biomedical application of TiO 2 -NP in areas such as cancer therapy, drug delivery systems, cell imaging, genetic engineering, biosensors, and biological experiments [5][6][7].
However, with the increasing developments in the application of TiO 2 -NP, concerns regarding their toxicity to humans also increase. Many studies have reported that TiO 2 -NP elicit a toxic response in in vitro and in vivo systems. Bhattacharya et al. reported that TiO 2 -NP of <100 nm in diameter were able to generate free radicals and elevate DNA adduct formation  in human lung fibroblasts [8]. In addition, in A549 cells, the anatase TiO 2 -NP induced mitochondrial injury in a dose-dependent manner owing to reactive oxygen species (ROS) generation [9]. Oesch and Landsiedel reviewed the genotoxicity of TiO 2 -NP using various test results [10]. Moreover, Sager et al. reported that P-25 TiO 2 -NP suspension (anatase: rutile = 80:20, 21 nm) induces an inflammation response in rats [11]. Oberdorster et al. [12] reported a similar result that 21-nm TiO 2 -NP had inflammatory effects on the alveolar interstitium in the lungs. Ferin et al. detected polymorphonuclear (PMN) leukocytes in lavage cells in rat lung after inhalation of~20-nm TiO 2 -NP [13]. Although there are many toxicity results, the detailed molecular mechanism of TiO 2 -NP toxicity is not clear.
The endoplasmic reticulum (ER) is an organelle that regulates protein secretion, cell surface development, and maintenance of the calcium ion (Ca 2+ ) concentration of cells [4]. Thus, disruption of ER homeostasis leads to protein misfolding and ER stress, which affect both the quality control and translation of protein. The membranes of the ER and mitochondria are enriched with Ca 2+ -binding chaperones called mitochondria-associated ER membranes (MAMs), which preserve and regulate cellular homeostasis in different environments [14]. Studies have shown that ER stress is linked closely to changes in the composition of MAMs, deregulated Ca 2+ transport, and cell death [15]. Furthermore, ER stress is associated with protein degradation via autophagy, which at abnormal levels, leads to cytotoxic processing or mechanisms such as apoptosis [16].
In this study, we demonstrated that ER stress-mediated MAM disruption, autophagy, and mitochondrial dysfunction might play a key role in the TiO 2 -NP-induced toxic responses in human bronchial epithelial cells.

Characterization of TiO 2 nanoparticles
The TiO 2 nanoparticles (TiO 2 -NP, P-25; anatase:rutile, 8:2) were purchased from Degussa Korea (Inchon, Korea). The structure and morphology of the TiO 2 -NP were characterized by transmission electron microscopy (TEM) with an accelerating voltage of 100 kV. The TEM samples were dispersed in methanol, and a drop of the suspension was placed on formvar-carbon film on a square 300-mesh copper grid, followed by drying the grid at room temperature for 1 h. We conducted X-ray diffraction (XRD) using the X'pert PW1827 diffractometer (Philips, Netherlands) to confirm the crystal structure of the TiO 2 -NP [17]. The goniometer was motorized and moved through a scanning range of θ-2θ. The diffractometer was operated at 40 kV and 40 mA in the range of 20-80°. The steps were performed in increments of 0.05°, and counts were collected for 5 s at each step [18]. For dynamic light scattering (DLS) measurements, 4 mL of a 0.2 mg/mL suspension of TiO 2 -NPs in distilled water was sonicated for 30 s. The hydrodynamic sizes and zeta potentials of the particle suspension were measured at room temperature using an Electrophoretic Light Scattering Spectrophotometer (ELS-8000, Photal, Osaka, Japan), with an accumulation time of 70 times and an equilibration time of 60 s.

Suspension of TiO 2 nanoparticles
We chose a suspension protocol that has been proven to yield the best dispersion of the nanomaterials in previous research [19]. For suspension in culture medium, TiO 2 -NP powder was dispersed in phosphate-buffered saline (PBS) at 10 mg/mL and sonicated for 10 min using an Ultrasonic cleaner (5510-DTH, Branson, MI, USA). After sonication, to prepare the end-point concentrations, Dulbecco's modified Eagle's medium (DMEM)-F12 medium (Gibco, NY, USA) was transferred to test tubes and diluted with the TiO 2 -NP stock solution.

Measurement of reactive oxygen species (ROS)
The generation of ROS in the cells following TiO 2 -NP treatment was determined using 2,7-dichlorodihydrofluorescein diacetate (H 2 -DCFDA), an indicator of intracellular ROS [20]. The process involves elimination of the acetate groups (non-fluorescent) by intracellular esterases with the generation of oxidation in the cells [21]. H 2 -DCFDA stock solution (10 mM) in DMSO was diluted in medium to attain a 5 μM working concentration. After incubation for 1 h, the cells were rinsed with PBS and fixed with 4% paraformaldehyde for 10 min at room temperature. The fluorescence was detected using a confocal laser scanning microscope (LSM710, Carl Zeiss, Germany) and fluorescence-activated cells sorting (FACS, BD Bioscience, New Jersey, USA) [17].

Immunofluorescence staining
The TiO 2 -NP-treated cells were fixed with 4% paraformaldehyde (PFA) for 10 min and then treated with 1% Triton X-100 for 15 min. After fixation, the samples were washed and blocked with 3% bovine serum albumin in 1× tris-buffered saline (TBS) for 1 h at 37°C. After washing, samples were incubated overnight with the primary antibodies (IP3R, VDAC, and LC3) diluted 1:200 in blocking solution. After washing with 1× Tween 20-TBS, the fluorescence of the HRPconjugated secondary antibodies was measured (Invitrogen, CA, USA). After rinsing, samples were mounted with DAKO Cytomation Faramount Aqueous Mounting solution (DAKO, CA, USA) and imaged using confocal laser scanning microscopy (CLSM, LSM710, Carl Zeiss, Germany) [17]. The stained LC3-fluorescent puncta were counted using the Image J program supported by National Institutes of Health (MD, USA).

Measurement of mitochondrial intensity and adenosine triphosphate (ATP) levels
After incubation of the cultures with TiO 2 -NP for 24 and 48 h, the mitochondrial intensity was detected by staining with 100 nM Mito-tracker (Invitrogen, CA, USA). The Mito tracker images were analyzed using the Mytoe program, which is software for analyzing mitochondrial dynamics from fluorescence microscope images [22]. To monitor the levels of ATP, normal lung cells were seeded on a 96-well plate and treated with TiO 2 -NP for 24 and 48 h, followed by incubation with 100 μL of CellTiter-Glo reagent (Promega, TN, USA) for 10 min. The luminescent signal was detected using a luminometer (Berthold, Germany).

Statistical analysis
The results are presented as the mean ± standard error of the mean (SEM) with n = 3. The statistical analyses were carried out using one-way and two-way ANOVA. The differences were considered statically significant with p values of p < 0.05, p < 0.01, and p < 0.001.

Results
Characterization of TiO 2 nanoparticles The particle shape and size of the TiO 2 -NP were analyzed using TEM (Fig 1A). In addition, the XRD pattern showed that the TiO 2 -NP contained a mixture of the anatase and rutile forms (S1 Fig). The hydrodynamic diameter and zeta potential of the TiO 2 -NP were determined using DLS. The TiO 2 -NP had an induced hydrodynamic size of approximately 250 nm in water or in medium, confirming that they are able to aggregate in suspension. The zeta potential results showed that the TiO 2 -NP had negatively charged surfaces in the medium (Table 1). Zeta potential values that are greater than +25 or less than −25 mV typically indicate good stability [23]. Therefore, the zeta potential of TiO 2 -NP in water or medium indicated that the TiO 2 -NP were aggregated in suspension. To investigate the cellular effects of the TiO 2 -NP, cell proliferation and viability were determined during 72 h in real time using the xCELLigence system. The real-time viability of TiO 2 -NP-treated cells was decreased as a function of TiO 2 -NP concentration and time ( Fig 1B). We also confirmed this result using the MTT and WST-1 assays. As shown Fig Fig 1D). Further, the number of stained cells increased in the TiO 2 -NP-treated cells. Our results show that TiO 2 -NP is able to generate intracellular ROS in human bronchial epithelial cells.

TiO 2 nanoparticles induce ER stress in normal human lung cells
To determine the ER stress response in cells following TiO 2 -NP treatment, we incubated human bronchial epithelial cells with various doses of TiO 2 -NP at different time points. The expression of the ER stress-related proteins, including Grp78/Bip, IRE-1α, phospho-IRE-1α, and CHOP, were analyzed by western blot. The results indicated that the expression levels of these proteins increased in TiO 2 -NP-treated cells. The results indicated that the expression levels of these proteins were increased in TiO 2 -NP-treated cells as determined by densitometric analysis (Fig 2A). These results showed that TiO 2 -NP exposure increases ER stress in cells.
Phosphorylation of the IRE-1α protein is reported to be a marker of its activation [24]. Therefore, we calculated the ratio of phosphorylated IRE-1α to total IRE-1α. In cells treated with 100 μg/mL TiO 2 -NP for 48 h, the IRE-1α phosphorylation significantly increased. Calnexin is an integral protein of the ER, which is known to correlate with the condition of the ER [25].
Our results demonstrated that treatment with TiO 2 -NP decreased calnexin expression (Fig A  and B in S3 Fig). These results show that TiO 2 -NP induce ER stress and affect the condition of the ER in human bronchial epithelial cells. TiO 2 nanoparticles disrupt mitochondria by damaging MAMs and cellular calcium homeostasis Disruption of calcium homeostasis is related directly to the induction of ER stress [26]. To investigate Ca 2+ homeostasis following TiO 2 -NP treatment, the integrity of mitochondria-associated ER membranes (MAMs), which regulate the Ca 2+ flow between the ER and mitochondria, was evaluated. Disruption of MAMs was detected by measuring the expression levels of several cellular Ca 2+ -regulating proteins such as inositol triphosphate receptor (IP3R; a glycoprotein complex in the ER membrane), voltage-dependent anion-selective channel protein 1  (VDAC1; outer mitochondrial membrane protein that mediates Ca 2+ flow), and 75-kDa glucose regulated protein (Grp75; a heat shock 70 family protein that regulates both IP3R and VDAC1) (Fig 3A and 3B and S4 Fig). To determine the quantitative changes in mitochondrial Ca 2+ levels, cells were stained with Rhod-2 AM, a mitochondrial Ca 2+ indicator. CLSM images revealed that the fluorescence decreased in TiO 2 -NP-treated cells. (Fig 3C, upper panel). Changes in the mitochondrial status also were assessed using CLSM imaging with the Mitotracker staining and TEM imaging. The Mito-tracker staining images showed that the mitochondria morphology was altered by TiO 2 -NP treatment; the alterations of the mitochondria were digitalized by the Mytoe program (Fig 3C, middle panel and Fig 3D). The Mytoe analysis revealed that the intensity of the Mito tracker staining decreased significantly in TiO 2 -NPtreated cells (Fig 3D). The TEM images also clearly demonstrated morphological changes in the mitochondria, such as swelling and disruption of the outer and inner mitochondrial membranes (Fig 3C, bottom panel). To determine the disruption of mitochondrial function, we measured the expression levels of the respiration chain-related proteins, NADH dehydrogenase (ND6, complex subunit I), succinate dehydrogenase complex subunit A (complex subunit II, SDHA), cytochrome C (Cyto C, complex subunit III), and cyto C oxidase IV (COX 1V, complex IV subunit). The western blot results showed that the protein levels decreased in TiO 2 -NP-treated cells (Fig 3E). To monitor mitochondrial biogenesis, the ATP levels were measured after treatment of cells with TiO 2 -NP. The cellular ATP levels decreased significantly (Fig 3F). Our results show that induced ER stress could demolish MAMs and mitochondrial calcium homeostasis in cells. In addition, induced ER stress could disrupt mitochondrial morphology and biogenesis.

TiO 2 nanoparticles induce autophagy in normal human lung cells
ER stress-induced autophagy correlates with cell death [27]. To determine if autophagy was induced in TiO 2 -NP-treated cells, the expression levels of LC3, sequestosome-1 (SQSTM1/ p62), and BECN1 were examined by western blot assay and densitometric analysis. LC3 is one of the key regulators of autophagy, which is measured quantitatively by the ratio of LC3 II to LC3 I [28]. Our results showed that the LC3 II/LC3 I ratio increased significantly in the TiO 2 -NP-treated cells. Additionally, the expression of SQSTM1/p62 and BECN1 also increased significantly in the TiO 2 -NP-treated cells (Fig 4A and 4B). Furthermore, the immunofluorescence assay revealed that TiO 2 -NP treatment increased the number of LC3 fluorescence puncta (red) (Fig 4C). We counted the LC3 puncta using the Image J software (Fig 4D). Taken together, our results clearly demonstrate that the TiO 2 -NP induces autophagy in human bronchial epithelial cells.

Inhibition of ER stress decreases TiO 2 nanoparticles cytotoxicity
Tauroursodeoxycholic acid (TUDCA) is a well-known inhibitor of ER stress [29]. To determine the potential effects of the inhibition of ER stress on mitochondria function disruption, autophagy, and cell toxicity, normal lung cells were treated with TUDCA (1mM) for 24h before exposure to TiO 2 -NP. Compared with the expression levels of actin, the internal standard, cotreatment with TiO 2 -NP and TUDCA did not induce any significant changes in the protein expression levels of phospho-IRE-1α, Hsp60, BECN, and p62 ( Fig 5A). Furthermore, treatment with TUDCA and TiO 2 -NP did not cause any significant changes in the LC3 levels ( Fig 5B). TEM images demonstrated no significant changes in the mitochondrial or ER morphology in cells pretreated with TUDCA (Fig 5B). To confirm whether ER stress is associated with the biogenesis of mitochondria, we compared the ATP levels of TiO 2 -NP-treated cells with TUDCA pretreatment with those without pretreatment. Our results indicated that TiO 2 -NP alone decreased ATP levels. However, TUDCA pretreatment prevented this effect (Fig 5C and 5D). The levels of ER marker, calnexin remained unaltered in TUDCA co-treated cells (S5 Fig). Together, our results indicate that the cells with blocked ER stress do not exhibit any autophagy and mitochondria disruption following treatment with TiO 2 -NP.

Discussion
TiO 2 -NP are widely used owing to their unique characteristics, which include extended photostability, strong oxidizing energy, relatively low descriptive toxicity, and easy availability. Due to these properties, TiO 2 -NP are being used in diverse biomedical areas such as cell imaging, biosensors, genetic engineering in cancer therapy, drug delivery, and development of medical implants [5]. However, prior to its extensive use in humans, the mechanisms underlying the potential toxic effects of TiO 2 -NP should be carefully identified.
The ROS generated owing to nanomaterials, including metal oxide-based nanoparticles, TiO 2 -NP, carbon fullerenes, and carbon nanotubes, are associated closely with cell damage such as inflammation, oxidative stress, and cell organelle disruption [30][31][32][33]. Many studies have reported that TiO 2 -NP induce toxic response by increasing the generation of ROS, which disrupts redox homeostasis [34,35]. Our H 2 -DCFDA staining data revealed the generation of ROS following TiO 2 -NP treatment in human bronchial epithelial cells (Fig 1 and S1 Fig). Therefore, we hypothesized that the elevated ROS levels affected the ER, which is one of the organelles that plays an important role in the cellular quality control systems and sensitivity to oxidative stress. In addition, this action may be a key mechanism of TiO 2 -NP toxicity. The ER is susceptible to diverse stressors such as misfolded proteins, environmental triggers, metabolic disturbances, and oxidative stress [36]. ER stress elevates Grp78/Bip, the glucose-regulated protein and chaperone [37]. The chaperone interacts with the misfolded proteins and promotes refolding, thereby playing a critical role in regulating three ER transmembrane proteins, protein kinase-like ER kinase (PERK), IRE-1α, and activating transcription factor 6 (ATF6) [38]. Diverse nanoparticles, including zinc oxide, silver nanoparticles, and iron oxide nanoparticles, were shown to elevate ER stress events [39,40,41]. However, only a few studies have shown that TiO 2 -NP could induce ER stress events in mammalian cells. Our data show that TiO 2 -NP could induce IRE-1α phosphorylation, elevate Grp78/Bip and CHOP, and activate the ER stress pathway in human bronchial epithelial cells. In addition, ER stress is strongly associated with the disruption of cellular Ca 2+ homeostasis. The ER and mitochondria cooperatively regulate the cellular homeostatic network and act as physiological barriers against abnormal intracellular Ca 2+ changes [42,43]. Ca 2+ translocates from the ER to the mitochondria through the MAMs. van Vliet et al. suggested that MAMs regulate energy metabolism, mitochondria biogenesis, and diverse signaling pathways involving ER stress and autophagy [15,29]. To evaluate the effect of TiO 2 -NP on the MAMs, we evaluated specific proteins that have been identified in MAMs, including the IP3R, which plays a crucial role in Ca 2+ handling in the ER, VDAC, and Each bar represents mean ± SEM, n = 5. Significant differences are indicated by **p < 0.01, ***p < 0.001 (comparison between TiO 2 -NP treatment concentrations, 50 and 100 μg) and ## p < 0.01, ### p < 0.001 (comparison between TiO 2 -NP treatment time, 24 and 48 h). (E) To detect mitochondria dysfunction, the proteins related to the respiration chain in the mitochondria, including ND6 (complex subunit I), SDHA (complex subunit II), Cyto c (complex subunit III), and COXIV (complex subunit IV), were analyzed using western blot assay. (F) ATP levels in the mitochondria after treatment with TiO 2 -NP. Significant differences are indicated by *p < 0.05, **p < 0.01 (comparison between TiO 2 -NP treatment concentrations, 50 and 100 μg) and # p < 0.05, ### p < 0.001 (comparison between TiO 2 -NP treatment times, 24 and 48 h). IP3R, inositol triphosphate receptor; VDAC1, voltage-dependent anion-selective channel protein 1; Grp75, 75 kDa glucose regulated protein; ND6, NADH dehydrogenase; SDHA, succinate dehydrogenase complex subunit A; Cyto C, cytochrome C; COXIV, cyto C oxidase IV.
doi:10.1371/journal.pone.0131208.g003  Grp75 [8,44,45]. Results of the fluorescence microscope imaging and western blot assay showed alterations in the MAM-related proteins following exposure to TiO 2 -NP (Fig 2A and  2B). ER stress decreases ER Ca 2+ levels, which consequently reduces the expression of IP3R. IP3R functions as a transporter of Ca 2+ from the ER to the mitochondria, and therefore, its decreased expression eventually diminishes the mitochondrial bioenergetics, leading to reduced cellular ATP production [46]. Furthermore, IP3R-mediated Ca 2+ release from the ER is associated closely with increased levels of free cytoplasmic Ca 2+ , which activates autophagy [47]. Our data show that Ca 2+ imbalance and abnormal autophagy increased following TiO 2 -NP exposure. In particular, decreased mitochondrial Ca 2+ levels were detected with Rhod-2 AM staining in TiO 2 -NP-treated cells. The changes in the Ca 2+ level affected the mitochondrial biogenesis disruption. In addition, we determined mitochondrial disorder by measuring the levels of proteins related to the respiration chain in the mitochondria and the ATP levels. To further support the correlation between ER stress and disruption of the MAMs, we treated cells with TUDCA, a bile acid that inhibits ER stress [29]. TUDCA pretreatment decreased ER stress and restored mitochondrial biogenesis and autophagy (Fig 5).
In conclusion, our data provide new evidence for the mechanism of TiO 2 -NP cytotoxicity, as demonstrated by increased ER stress effects, MAM disruption, dysfunction of mitochondria biogenesis, and autophagic cell death (Fig 6). The potential harmful effects of TiO 2 -NP nanoparticles have remained elusive. There are minimal regulatory guidelines from the government controlling their use for the industry. The safe widespread use of nanomaterials, including TiO 2 -NP, requires intensive study of their toxicities and mechanisms of action in relevant organisms. In future research, however, further works should be performed in order to discriminate whether observed results are due to the nanosized form or to the titanium component itself. To monitor the state of the endoplasmic reticulum (ER) after TUDCA treatment, calnexin, an ER state marker, was detected. Immunofluorescence assay of calnexin after TUDCA (ER stress inhibitor) treatment and with TiO 2 -NP; green, calnexin; blue, DAPI-nucleus. TUDCA, tauroursodeoxycholic acid. (TIF)