Titanium Dioxide Induces Apoptosis under UVA Irradiation via the Generation of Lysosomal Membrane Permeabilization-Dependent Reactive Oxygen Species in HaCat Cells

Titanium dioxide nanoparticles (TiO2 NPs) have wide commercial applications, owing to their small size; however, the biosafety of TiO2 NPs should be evaluated further. In this study, we aimed to investigate the cytotoxicity of TiO2 NPs in the presence and absence of ultraviolet A (UVA) irradiation in human keratinocyte HaCaT cells. TiO2 NPs did not significantly affect cell viability in the absence of UVA irradiation. Nonetheless, UVA-irradiated TiO2 NPs induced caspase-dependent apoptosis of HaCaT cells. Exposure of HaCaT cells to TiO2 NPs and UVA resulted in reactive oxygen species (ROS) generation and lysosomal membrane permeabilization (LMP); both effects were not observed in the absence of UVA irradiation. An analysis of the relationship between LMP and ROS, using CA-074 as a cathepsin inhibitor or NAC as an antioxidant, showed that LMP stimulates ROS generation under these conditions. These results imply that LMP-dependent oxidative stress plays a critical role in the UVA phototoxicity of TiO2 NPs in HaCaT cells.


Introduction
The potential toxicity of nanomaterials (NMs) to humans and the environment has been recognized since the dawn of nanotechnology [1]. This is because their physicochemical properties are highly dependent on small particle size, resulting in a high surface-tovolume ratio and enhanced transport across biological barriers [2]. Consequently, some materials, while benign in bulk form, become toxic when prepared in nano sizes [3].
Titanium dioxide nanoparticles (TiO 2 NPs) are the highest volume-NM employed in water purification, paint, food additives, and personal care products, including sunscreens. There is a high probability that these nanoparticles (NPs) will come in contact with human skin and mucous membranes, subsequently entering the body [4]. Numerous studies have confirmed that the overall cytotoxicity of TiO 2 NPs is rather low, compared with that of several widely produced NMs [5][6][7], and they are "generally regarded as safe (GRAS)" by regulatory agencies [8]. However, some countries lately have removed TiO 2 from the GRAS designation owing to in vivo evidence of its potential toxicity [9]. In addition, in vitro studies with mammalian cells have shown that TiO 2 NPs induce oxidative stress and apoptosis [10][11][12][13], suggesting that the safety profile of TiO 2 NPs is still unclear.
A key determinant of NP toxicity is their ability to penetrate cellular membranes and their subcellular localization. NPs often accumulate in the lysosomes [14] via processes such as autophagy and endocytosis. Autophagy, a main catabolic pathway in mammalian are photoexcited by UVA. Consequently, the loss of lysosomal membrane integrity is essential for TiO2 NPs and UVA-induced ROS surge. Moreover, our data strongly suggest that caspase-dependent apoptosis is a key mechanism of HaCaT cell death induced by their exposure to internalized and photoexcited TiO2 NPs.

Cell Culture
Human keratinocyte HaCaT cells (Cell Line Services, Eppelheim, Germany) were maintained in Dulbecco's modified Eagle medium containing 10% fetal bovine serum and 1% penicillin-streptomycin (all obtained from GIBCO-BRL, Grand Island, NY, USA). The cells were incubated under 5% CO2 at 37 °C.

Preparation and Characterization of TiO2 NP Suspension
TiO2 NPs (AEROXIDE ® TiO2 P25) were obtained from Evonik Industries AG (Essen, Germany). They were dispersed in a complete medium according to the National Institute of Standards and Technology (USA) special publication 1200-4 [41]. Particle size distribution of TiO2 NPs was determined by dynamic light scattering (DLS; ZEN5600, Malvern Panalytical, Worcestershire, England). Zeta potential was measured using the same equipment mentioned above.

Treatment with TiO2 NPs and UVA
A graphic summary of the experimental procedure is shown in Figure 1. HaCaT cells were plated in two 96-well plates and then incubated for 24 h. On a subsequent day, the surface-adhered cells were exposed to 25-200 μg/mL TiO2 NPs for 24 h. The treated cells were washed twice with warm phosphate-buffered saline (PBS) (Sigma-Aldrich, St. Louis, MO, USA) to separate extracellular TiO2 NPs. While one plate was directly irradiated with UVA for 20 min (light dose P = 5 J/cm 2 , exposure group), the other plate was protected from UVA exposure by covering with foil (dark group). Following post-irradiation stabilization for 20 min, HBSS was replaced with a complete medium, and the cells were incubated for 24 h.

Cell Viability Assay
The cells were seeded in 96-well plates and treated as described above. The CellTiter-Glo Luminescent Cell Viability kit (Promega, Madison, WI, USA) was employed to assess cell viability, according to the manufacturer's instructions. The IC50, the concentration at

Cell Viability Assay
The cells were seeded in 96-well plates and treated as described above. The CellTiter-Glo Luminescent Cell Viability kit (Promega, Madison, WI, USA) was employed to assess cell viability, according to the manufacturer's instructions. The IC 50 , the concentration at which cell growth is inhibited by 50%, compared with the untreated control, was estimated using GraphPad Prism software 7.0 (GraphPad Software Inc., San Diego, CA, USA). Since treatment with 200 µg/mL TiO 2 in the dark did not cause more than 50% viability decrease in HaCaT cells, the IC 50 was estimated from the available viability data using GraphPad Prism software 7.0.

Measurement of Lactate Dehydrogenase Release
A lactate dehydrogenase (LDH) detection kit (CytoTox96 ® Non-Radioactive Cytotoxicity Assay, Promega) was used to evaluate the integrity of the cell membrane. The release of LDH into the culture medium indicates irreversible cell death due to cell membrane damage [42,43]. After treating the cells as mentioned above, 50 µL of the supernatant was collected and tested according to the manufacturer's instructions. LDH leakage was calculated using the following equation. LDH min is the LDH leakage value of untreated cells. LDH max is the LDH leakage value of cells treated with 2% Triton-X100 (positive control) for 10 min.

Cell Cycle Analysis
DNA fragmentation is a hallmark of apoptosis [44], and therefore, apoptotic cells eventually develop a deficit in DNA content. In DNA content histograms, apoptotic cells often form characteristic "sub-G1" or "hypodiploid" peaks [45][46][47]. HaCaT cells were seeded at a density of 2 × 10 5 cells/well in 6-well plates and incubated for 24 h. After treatment, HaCaT cells were harvested and fixed in 70% (v/v) ethanol. To evaluate the DNA content, cells were treated with RNase and stained with propidium iodide. The cells were analyzed using a FACSVerse flow cytometer (BD Biosciences, San Jose, CA, USA). The data were analyzed using FlowJo (version X, BD Biosciences).

Apoptosis Determination
Cells were seeded at a density of 2 × 10 5 cells/well in 6-well plates and treated with TiO 2 and UVA. The cells were incubated with 5 µL of Annexin V and 5 µL of propidium iodide (PI) for 15 min at room temperature in the dark, according to the manufacturer's instructions (BD), and then subjected to flow cytometry to measure the apoptosis rate (%).

Cell Lysates and Western Blotting
HaCaT cells were seeded in 60 mm dishes and treated as mentioned above. The treated cells were scraped and harvested using PBS. After discarding the supernatant, the cell pellets were suspended in 2× Laemmli sample buffer with 5% β-mercaptoethanol (Bio-Rad, Hercules, CA, USA), and boiled for 7 min. The lysates were separated by 4-15% SDS-PAGE and transferred onto an Immobilon membrane (Millipore, Burlington, MA, USA). In this study, the following primary antibodies were used: PARP1 (1:1000, ab32071, Abcam, Cambridge, UK) and β-actin (1:5000, #3700, Cell Signaling Technology, Beverly, MA, USA). The secondary antibodies used were horseradish peroxidase (HRP)-conjugated goat antirabbit immunoglobulin G (IgG) and goat anti-mouse IgG (Invitrogen, Waltham, MA, USA). Bands were developed using the WesternBright ECL HRP substrate (Advansta, CA, USA). Quantification of the immunoblots was performed using Image J (NIH, MD, USA).

Lysosomal Integrity Assay
As AO is a lysosomotropic dye that exhibits a red fluorescence under acidic conditions and green fluorescence in the non-acidic environment, therefore lysosomal acidification can be observed using the green/red fluorescence ratio in live cells [48]. We adopted the method described by Sun and Gan [49]. For fluorescence microscopy, HaCaT cells were seeded in 12-well plates and treated as mentioned above. After washing with PBS, the cells were stained with 5 µg/mL AO in DMEM for 15 min in the dark at 37 • C. Next, cells were washed with PBS again and further visualized using a fluorescence microscope (Leica Microsystems, Wetzlar, Germany). To quantitate fluorescence using a microplate reader, the cells were seeded in 96-well plates and allowed to attach to the plate for 24 h. The cells were stained with 5 µg/mL AO for 15 min. After washing with PBS, the cells were subsequently treated with TiO 2 NPs and subjected to UVA irradiation. The treated cells were incubated for 8 h and then washed with HBSS. Fluorescence was measured at an excitation wavelength of 485 nm and two emission wavelengths of 530 (green AO) and 620 nm (red AO). Normal lysosomal integrity = total red fluorescence intensity of normal lysosomes / total green fluorescence intensity of normal lysosomes. Lysosomal integrity = total red fluorescence intensity / (total green fluorescence intensity × normal lysosomal integrity).

Measurement of ROS
When CM-H 2 DCF-DA is oxidized by ROS such as H 2 O 2 and free radicals, it can be detected by monitoring the increase in fluorescence [50]. The TiO 2 treatment procedure is described above. The treated cells were loaded with 10 µM CM-H 2 DCF-DA for 30 min in the dark at 37 • C and washed with HBSS. Fluorescence of the cells was measured using a microplate reader (λ ex = 495 nm/λ em = 525 nm) and also observed by fluorescence microscopy.

Inhibitor Study
To investigate the relationship between LMP, ROS generation, and cell death, we used CA-074, a cathepsin B inhibitor, and NAC, a well-known antioxidant. When the medium was replaced after TiO 2 NP treatment and UVA irradiation, each inhibitor was added to the media prior to subsequent cell incubation.

Statistical Analysis
All data are presented as mean ± standard error of the mean (SEM) or one standard deviation (SD) of at least three separate experiments. GraphPad Prism 7.0 software was used to perform statistical analyses. The normality of data was assessed using Kolmogorov-Smirnov test, and equal variance was assessed using Bartlett's test. For normally distributed data, statistical differences were determined using the analysis of variance, followed by Bonferroni's multiple comparison test. If the data were not normally distributed, Kruskal-Wallis test was performed followed by Dunn's test.

Characterization of the TiO 2 NP Suspension Employed for Cellular Exposure
DLS was performed to determine the particle size distribution of the TiO 2 NP suspension ( Figure 2). After dispersion in media, the z-average size of TiO 2 NPs was 209.7 ± 2.9 nm (diameter, intensity weighted) with a polydispersity index of 0.246 ( Figure 2a). In addition, the stability of the TiO 2 NP suspension during the test duration was tested. The particle size distribution profile was almost constant for 2 days (Figure 2b), indicating that the stability of the TiO 2 NP suspension was acceptable.
Since NP size distribution can undergo significant changes when transferred to environments used for biological studies [51], we characterized the dispersions of TiO 2 NPs in deionized water (DIW) and media by measuring size distributions and zeta potential values ( Figure 2c). As expected, the z-average size of TiO 2 NPs was larger in the media (209.7 nm) than in DIW (157.1 nm) (Figure 2c). The zeta potential was 41.2 mV in DIW and −10.8 mV in media ( Figure 2c). These results suggest that the hydrodynamic size and zeta potential TiO 2 NPs both are affected due to the interaction with media components.
Since NP size distribution can undergo significant changes when transferred to envi-ronments used for biological studies [51], we characterized the dispersions of TiO2 NPs in deionized water (DIW) and media by measuring size distributions and zeta potential values ( Figure 2c). As expected, the z-average size of TiO2 NPs was larger in the media (209.7 nm) than in DIW (157.1 nm) (Figure 2c). The zeta potential was 41.2 mV in DIW and −10.8 mV in media ( Figure 2c). These results suggest that the hydrodynamic size and zeta potential TiO2 NPs both are affected due to the interaction with media components.

UVA Irradiation Induces Toxicity in HaCaT Cells Treated with TiO2 NPs
The 3T3 cell line is the most used cell line to study the effect of phototoxicity. These cells have several advantages, including availability, easy and inexpensive cultivation, as well as reproducibility. Recently, Svobodova et al. [52] reported that human keratinocyte HaCaT cells are also appropriate for in vitro phototoxicity testing. We have investigated how the viability of HaCaT cells is affected by a combined action of TiO2 NPs exposure and UVA irradiation. Cell viability measurement demonstrated that HaCaT cells were largely unaffected to both TiO2 exposure (up to 200 μg/mL) and UVA irradiation (up to 5 J/cm 2 ) treatment if applied separately (Figure 3a). However, the combination of TiO2 NPs and UVA irradiation markedly reduced the viability of HaCaT cells (Figure 3a). To quantify TiO2 NP phototoxicity, we estimated the NP concentration at which 50% of cell growth (IC50 value) is inhibited with and without UVA irradiation. The IC50 of the samples was reduced 10-fold with UVA irradiation, compared with that of nonirradiated samples (Figure 3a). To confirm cellular damage, we have performed the LDH assay. TiO2 NPs alone

UVA Irradiation Induces Toxicity in HaCaT Cells Treated with TiO 2 NPs
The 3T3 cell line is the most used cell line to study the effect of phototoxicity. These cells have several advantages, including availability, easy and inexpensive cultivation, as well as reproducibility. Recently, Svobodova et al. [52] reported that human keratinocyte HaCaT cells are also appropriate for in vitro phototoxicity testing. We have investigated how the viability of HaCaT cells is affected by a combined action of TiO 2 NPs exposure and UVA irradiation. Cell viability measurement demonstrated that HaCaT cells were largely unaffected to both TiO 2 exposure (up to 200 µg/mL) and UVA irradiation (up to 5 J/cm 2 ) treatment if applied separately ( Figure 3a). However, the combination of TiO 2 NPs and UVA irradiation markedly reduced the viability of HaCaT cells (Figure 3a). To quantify TiO 2 NP phototoxicity, we estimated the NP concentration at which 50% of cell growth (IC 50 value) is inhibited with and without UVA irradiation. The IC 50 of the samples was reduced 10-fold with UVA irradiation, compared with that of nonirradiated samples (Figure 3a). To confirm cellular damage, we have performed the LDH assay. TiO 2 NPs alone at lower concentrations (up to 100 µg/mL) did not induce a noticeable LDH release. However, the LDH activity was upregulated in cells that were subjected to the combined treatment with UVA compared to those treated with TiO 2 NPs alone (Figure 3b). Taken together, these results clearly show that a combination of UVA irradiation and TiO 2 NP exposure significantly enhances the cytotoxicity.
at lower concentrations (up to 100 μg/mL) did not induce a noticeable LDH release. However, the LDH activity was upregulated in cells that were subjected to the combined treatment with UVA compared to those treated with TiO2 NPs alone (Figure 3b). Taken together, these results clearly show that a combination of UVA irradiation and TiO2 NP exposure significantly enhances the cytotoxicity.

UVA Irradiation Facilitates the Apoptosis of HaCaT Cells Treated with TiO2 NPs
Next, we examined the morphological changes in cells that occurred during the combined TiO2 NP and UVA treatment. The cells treated with TiO2 NPs or UVA irradiation alone remained attached to the substrate and largely retained their shape. In contrast, cells treated with both TiO2 NPs and UVA irradiation detached from the plate surface and became smaller and spherical (Figure 4a). Given that cell shrinkage, or cell volume loss, is often an indication of apoptosis [41], we verified whether the cell death induced by the combined treatment with TiO2 NPs and UVA was apoptotic in nature. We examined the changes in DNA content following treatment with TiO2 NPs and UVA irradiation ( Figure  4b). Only 12.7% of HaCaT cells treated with TiO2 NPs alone for 24 h were in the sub-G1 phase, and there was no dramatic change in the DNA content, compared to the control group. However, treatment with both TiO2 NPs and UVA increased the cell population in sub-G1 to 32.8%. Simultaneously, G1, S, and G2/M populations were also smaller. Next, we detected the population of apoptotic cells using Annexin-V binding and PI staining. While the population of apoptotic cells (Q2 + Q4) did not change under treatment with TiO2 alone compared to the control group, it increased 6.6% after UVA irradiation alone. In addition, the combination treatment with TiO2 NPs and UVA sharply increased the proportion of apoptotic cells to 26.9%. These results provide further evidence that a combination of TiO2 and UVA leads to HaCaT cell death via the apoptotic pathway.
Since caspases are known as executioners in apoptosis [53], we also examined whether the TiO2 NP and UVA-induced cell death was mediated by caspase activation. We found no evidence that poly (ADP-ribose) polymerase (PARP), a caspase substrate, was processed following the separate treatments of HaCaT cells with TiO2 NPs or UVA (Figure 4c). However, when both TiO2 NPs and UVA were applied, PARP was progressively processed (Figure 4c). To ascertain whether caspases are important in apoptosis

UVA Irradiation Facilitates the Apoptosis of HaCaT Cells Treated with TiO 2 NPs
Next, we examined the morphological changes in cells that occurred during the combined TiO 2 NP and UVA treatment. The cells treated with TiO 2 NPs or UVA irradiation alone remained attached to the substrate and largely retained their shape. In contrast, cells treated with both TiO 2 NPs and UVA irradiation detached from the plate surface and became smaller and spherical (Figure 4a). Given that cell shrinkage, or cell volume loss, is often an indication of apoptosis [41], we verified whether the cell death induced by the combined treatment with TiO 2 NPs and UVA was apoptotic in nature. We examined the changes in DNA content following treatment with TiO 2 NPs and UVA irradiation (Figure 4b). Only 12.7% of HaCaT cells treated with TiO 2 NPs alone for 24 h were in the sub-G1 phase, and there was no dramatic change in the DNA content, compared to the control group. However, treatment with both TiO 2 NPs and UVA increased the cell population in sub-G1 to 32.8%. Simultaneously, G1, S, and G2/M populations were also smaller. Next, we detected the population of apoptotic cells using Annexin-V binding and PI staining. While the population of apoptotic cells (Q2 + Q4) did not change under treatment with TiO 2 alone compared to the control group, it increased 6.6% after UVA irradiation alone. In addition, the combination treatment with TiO 2 NPs and UVA sharply increased the proportion of apoptotic cells to 26.9%. These results provide further evidence that a combination of TiO 2 and UVA leads to HaCaT cell death via the apoptotic pathway.
Since caspases are known as executioners in apoptosis [53], we also examined whether the TiO 2 NP and UVA-induced cell death was mediated by caspase activation. We found no evidence that poly (ADP-ribose) polymerase (PARP), a caspase substrate, was processed following the separate treatments of HaCaT cells with TiO 2 NPs or UVA (Figure 4c). However, when both TiO 2 NPs and UVA were applied, PARP was progressively processed (Figure 4c). To ascertain whether caspases are important in apoptosis induced by the combination of TiO 2 NPs and UVA, we investigated the effect of the pan-caspase inhibitor z-VAD-fmk on cell viability. The treatment of HaCaT cells with z-VAD-fmk blocked cotreatment-induced cell death in a dose-dependent manner (Figure 4d). Taken together, these results confirm that TiO 2 NPs under UVA irradiation induces the caspase-dependent apoptosis in HaCaT cells. induced by the combination of TiO2 NPs and UVA, we investigated the effect of the pancaspase inhibitor z-VAD-fmk on cell viability. The treatment of HaCaT cells with z-VADfmk blocked co-treatment-induced cell death in a dose-dependent manner (Figure 4d). Taken together, these results confirm that TiO2 NPs under UVA irradiation induces the caspase-dependent apoptosis in HaCaT cells.

TiO 2 NPs Accumulate in Lysosomes
To observe the cellular uptake of TiO 2 NPs and structural changes in the cellular organelles in TiO 2 NP-and UVA-treated cells, we have employed electron microscopy ( Figure 5). At first we have confirmed the cellular uptake of TiO 2 NPs. TiO 2 NPs were not detected in untreated or only UVA-irradiated cells (Figure 5a,c). After treatment with TiO 2 NPs alone and in combination with UVA, TiO 2 NPs aggregates were observed in phagosome-like structures (black arrowhead) but not in the nuclei (white arrow) or mitochondria (black arrow) (Figure 5b,d). Next, we observed alterations in the cellular organelle structures. The organelles did not substantially change except when treated with both TiO 2 NPs and UVA (Figure 5a,c). Remarkably, the mitochondrial structure remained almost unaltered in all of the imaged samples ( Figure 5, black arrow). Moreover, we found that the combination of TiO 2 NPs and UVA increased the number of phagosomal and lysosomal structures containing TiO 2 NPs (Figure 5d, black arrowhead). These membranes were noticeably ruptured (Figure 5d, white arrowhead).

Combination of TiO2 NPs and UVA Induces LMP
As our results suggested that lysosomal membranes were altered by the combination of TiO2 NPs and UVA (e.g., ruptured membranes in Figure 5), we tested the lysosomal integrity of HaCaT cells by staining them with AO. While red fluorescence was strong, green fluorescence was rather weak in the untreated cells. However, the cells treated with TiO2 NPs and UVA mostly showed diminished red fluorescence and increased green fluorescence (Figure 6a), indicating that the lysosomes were severely damaged. Next, we quantified the lysosomal integrity using a microplate reader. The integrity of the lyso-

Combination of TiO 2 NPs and UVA Induces LMP
As our results suggested that lysosomal membranes were altered by the combination of TiO 2 NPs and UVA (e.g., ruptured membranes in Figure 5), we tested the lysosomal integrity of HaCaT cells by staining them with AO. While red fluorescence was strong, green fluorescence was rather weak in the untreated cells. However, the cells treated with TiO 2 NPs and UVA mostly showed diminished red fluorescence and increased green fluorescence (Figure 6a), indicating that the lysosomes were severely damaged. Next, we quantified the lysosomal integrity using a microplate reader. The integrity of the lysosomes in cells treated with the combination of TiO 2 NPs and UVA decreased to 89.04 ± 3.98% and 86.02 ± 4.01% following the incubation with 100 and 200 µg/mL TiO 2 NPs, respectively, relative to lysosomes in untreated cells (100%) (Figure 6b). These results suggest that lysosomal membranes are damaged during HaCaT cell death triggered by the combined treatment with TiO 2 NPs and UVA.

Combination of TiO2 NPs and UVA Induces ROS Generation
The oxidative damage of cellular molecules caused by excessive ROS production is known as a major factor driving the phototoxicity of TiO2 NPs [26][27][28][29][30]. In this study, we assayed intracellular ROS production using CM-H2DCF-DA dye. Independent treatment with TiO2-NPs and UVA generated rather low levels of intracellular ROS in a concentra-

Combination of TiO 2 NPs and UVA Induces ROS Generation
The oxidative damage of cellular molecules caused by excessive ROS production is known as a major factor driving the phototoxicity of TiO 2 NPs [26][27][28][29][30]. In this study, we assayed intracellular ROS production using CM-H 2 DCF-DA dye. Independent treatment with TiO 2 -NPs and UVA generated rather low levels of intracellular ROS in a concentrationdependent manner (Figure 6c), whereas the combination of TiO 2 NPs and UVA significantly increased intracellular ROS. This was particularly noticeable in the cells treated with 200 µg/mL TiO 2 NPs, wherein, the intracellular ROS level increased by approximately 2.3 fold following treatment with UVA irradiation, compared with level in the non-irradiated sample (Figure 6c). Again, fluorescence microscopy images confirmed that TiO 2 NPs and UVA noticeably boosted the ROS levels when combined (Figure 6d). These results confirm that intracellular ROS production accompanies HaCaT cell death caused by the combination of TiO 2 NPs and UVA.

Relationship among LMP, ROS Generation, and Cell Death in HaCaT Cells Treated with TiO 2 NPs and UVA
We examined the link between lysosomal membrane damage, ROS generation, and cell death induced by the combination of TiO 2 NPs and UVA. First, we assessed the role of the lysosomal membrane damage and ROS generation in cell death by means of CA-074, a cathepsin B inhibitor, and NAC, a well-known antioxidant. The treatment of HaCaT cells with CA-074 or NAC effectively inhibited the cell death, as caused by the combined treatment with TiO 2 NPs and UVA (Figure 7a). This result implies that both lysosomal membrane damage and ROS generation are critical contributors to cell death under these conditions. Next, we investigated the relationship between lysosomal membrane damage and ROS generation. ROS are known to induce LMP, especially when highly reactive hydroxyl radicals are produced in the lysosomes [54][55][56]. Therefore, we wanted to determine whether the loss of lysosomal membrane integrity occurs when ROS are produced in response to the combined TiO 2 NPs and UVA treatment. Unexpectedly, neither CA-074 nor NAC were effective in preserving the lysosomal membrane integrity in cells treated with TiO 2 NPs and UVA (Figure 7b). This result indicates that ROS increase was not the primary reason for the decline in lysosomal membrane integrity. Furthermore, as LMP may increase the generation of intracellular ROS via the release of free iron into the cytosol [57][58][59], we evaluated whether ROS can be generated due to LMP. Notably, both CA-074 and NAC attenuated ROS generation in response to the combined treatment with TiO 2 NPs and UVA (Figure 7c). This result suggests that the loss of lysosomal membrane integrity is vital for TiO 2 NP and UVA-induced ROS generation. Taken together, these results show that the combination of TiO 2 NPs and UVA induces cell death via LMP-dependent ROS generation in HaCaT cells.

Discussion
TiO2 NPs are among the most widely used nanomaterials in paints, plastics, cosmetics, and personal care products; they are also used as food additives and drug delivery agents [60,61]. They are widely used owing to their biocompatibility, high chemical stability, photocatalytic properties, and pronounced sensitivity to heat and magnetism [62-

Discussion
TiO 2 NPs are among the most widely used nanomaterials in paints, plastics, cosmetics, and personal care products; they are also used as food additives and drug delivery agents [60,61]. They are widely used owing to their biocompatibility, high chemical stability, photocatalytic properties, and pronounced sensitivity to heat and magnetism [62][63][64]. The TiO 2 NPs are being explored as potential candidates for degrading organic pollutants [65] and inactivating microorganisms owing to their photocatalytic activity [60,66]. However, these photocatalytic properties were reported to trigger oxidative damage, cellular structure destruction, key protein inactivation, and DNA damage, leading to cell apoptosis or necrosis [67][68][69][70]. To support their safe and efficient use, it is essential to identify mechanisms underlying the photocatalytic activity and phototoxicity of TiO 2 NPs. In this study, we evaluated the phototoxicity of TiO 2 NPs on human skin keratinocytes subjected to UVA irradiation and investigated the mechanistic aspects of the resulting cell death.
As shown in Figure 3, TiO 2 NPs exerted UVA-induced and dose-dependent toxicity in HaCaT cells. As cell viability decreases after TiO 2 NP and UVA treatment, we explored the mode of death of these cells. The photocatalytic activity of TiO 2 NPs can induce apoptosis and necrosis through various mechanisms [67,68,[70][71][72]. Apoptosis is a form of programmed cell death characterized by caspase activation, cell contraction, apoptotic body formation, exposure outside the cell membrane, chromatin condensation, and DNA fragmentation [73][74][75]. Our data show that the DNA sub-G1 population increases following combined treatment with TiO 2 NPs and UVA (Figure 4b). In addition, PARP, a substrate of caspase-3, is cleaved and z-VAD-fmk, a pan-caspase inhibitor, increases the viability in HaCaT cells co-treated with TiO 2 NPs and UVA (Figure 4c,d). These results suggest that the TiO 2 NPs and UVA-induced cell death is caused by the caspase-dependent apoptosis.
We detected TiO 2 NPs inside HaCaT cells using transmission electron microscopy; this confirmed that the phototoxicity observed in this study is induced by the intracellular accumulation of TiO 2 NPs. Consistent with the results of Tucci et al. [76], TiO 2 NPs accumulated in the phagosomal and lysosomal structures of the cells but not in the nuclei or mitochondria (Figure 5b,d). Remarkably, we observed lysosomal membrane rupture in HaCaT cells treated with TiO 2 NPs and UVA (Figure 5d). The maintenance of lysosomal membrane integrity is vital for cell viability because ruptured lysosomal membranes induce the leakage of various digestive enzymes into the cytosol, resulting in membrane trafficking defects, abnormal energy metabolism, and cell death [77]. Recently, lysosomal damage has been suggested as a key mechanistic element of nanoparticle toxicity [78][79][80] because most endocytosed nanoparticles were found to accumulate in lysosomal compartments. In our study, we evaluated the lysosomal membrane integrity by AO staining and determined that combined treatment with TiO 2 NPs and UVA damaged the lysosomal membranes in HaCaT cells (Figure 6a,b). In addition, CA-074, a cathepsin B inhibitor that protects against lysosomal rupture [81], prevented cell death induced by the combination of TiO 2 NPs and UVA (Figure 7a). These results demonstrate that LMP is imperative for TiO 2 NP and UVA-induced cell death.
As ROS-mediated LMP is the prevailing mechanism [82], we examined the link between LMP and ROS generation when HaCaT cells are exposed to TiO 2 NPs and UVA. First, we found that TiO 2 NPs subjected to UVA irradiation upregulated ROS generation in HaCaT cells (Figure 6c,d). In addition, the inhibition of ROS increased cell viability (Figure 7a), confirming the critical regulatory role of ROS in the progression of HaCaT cell apoptosis. Next, we tested lysosomal membrane integrity and ROS production in cells treated with the cathepsin B inhibitor CA-074 and antioxidant NAC. Interestingly, while CA-074 prevented ROS generation, NAC did not prevent the loss of lysosomal membrane integrity (Figure 7b,c). Although several factors and chemicals have been reported to induce LMP and the associated cell death [83][84][85], it is not clear whether lysosomal leakage leads to oxidative stress during apoptotic cell death driven by the lysosomal pathway. Our study, for the first time, proves that lysosomal destabilization by TiO 2 NPs and UVA induces cellular ROS generation and oxidative stress.
Recent reports showed that besides TiO 2 NPs other NPs, including ZnO, Au, and Ag NPs, exhibit photocatalytic activity [86]. In particular, the phototoxicity of ZnO NPs is studied because they are often included in paints, cosmetics, and medical materials, along with TiO 2 NPs. Wang et al. [87] have reported that the phototoxicity of ZnO NPs under UVA and visible light irradiation is related to oxidative DNA damage. As ZnO NPs and TiO 2 have comparable bandgap energy (3.2 eV) and photocatalytic activity [88], they can be expected to cause a similar cellular toxicity. However, the important difference between these two materials is their chemical stability, as ZnO slowly dissolves under UV illumination, releasing metal ions, while TiO 2 is chemically stable.

Conclusions
Previous studies have shown that ROS generation is the primary cause of the phototoxicity in cells treated with TiO 2 under UVA irradiation. In this study, we examined the mechanism underlying the phototoxicity induced by TiO 2 in HaCaT cells. We present evidence that TiO 2 induces lysosomal membrane damage and subsequently stimulates ROS production. We focused on the effects of TiO 2 NP uptake and UVA sensitization on lysosomal membrane integrity in HaCaT cells. We found that major lysosomal destabilization occurs when accumulated TiO 2 NPs are photoexcited by UVA. Consequently, the loss of lysosomal membrane integrity is essential for TiO 2 NPs and UVA-induced ROS surge. Taken together, our findings strongly suggest that caspase-dependent apoptosis is a key mechanism underlying HaCaT cell death, induced by their exposure to photoexcited TiO 2 NPs.
Certain instruments and materials are identified in this paper to adequately specify experimental details. In no case does it imply endorsement by NIST or that it is necessarily the best product for the experimental procedure.