Band Alignment-Driven Oxidative Injury to the Skin by Anatase/Rutile Mixed-Phase Titanium Dioxide Nanoparticles Under Sunlight Exposure

Band Alignment-Driven Oxidative Injury to the Skin by Anatase/Rutile Mixed-Phase Titanium Dioxide... Abstract Anatase/rutile mixed-phase titanium dioxide (TiO2) nanoparticles (NPs) have been found in cosmetics and cotton textiles. Once exposed to sunlight, mixed-phase TiO2 NPs are even more toxic to cells than pure phase NPs; however, the underlying mechanism remains unclear. Considering the unique anatase/rutile heterojunction structure existing in mixed-phase NPs, the potent toxicity of mixed-phase TiO2 NPs probably originates from the high reactive oxygen species (ROS) production because the anatase/rutile heterojunction is constituted by the staggered energy bands that facilitate the electron-hole separation at the interface due to the band alignment. In the present study, a library of mixed-phase TiO2 NPs with different anatase/rutile ratios was established to investigate the potential property-activity relationship and further clarify the underlying molecular mechanism. Under sunlight exposure, these mixed-phase TiO2 NPs could produce significant abiotic ROS and induce hierarchical oxidative stress to HaCaT skin cells and mice skin. The ROS magnitude and toxicity potential of these NPs were found to be proportional to their energy band bending (BB) levels. This means that the toxicity of mixed-phase TiO2 NPs can be correlated to their heterojunction density, and the toxicity potential of mixed-phase TiO2 NPs can be weighed by their BB levels. titanium oxide, nanoparticles, property-activity relationship, oxidative injury, sunlight exposure The rapidly expanding use of nanomaterials in various aspects of human life has dramatically increased human exposure to nanomaterials. Concerns on the potential adverse human health effects of nanomaterials are escalating (Gunter et al., 2007; Seaton et al., 2010). Titanium dioxide (TiO2) nanoparticles (NPs), such as anatase and rutile, have been widely used as ingredients in cosmetics and cotton textiles (Barker and Branch, 2008; Contado and Pagnoni, 2008; Doganli et al., 2016; Meilert et al., 2005). These applications potentially increase the risk of human skin exposure to TiO2 NPs, especially under outdoor sunlight irradiation. Ultraviolet (UV) or sunlight exposure on one hand can remarkably deepen the penetration of NPs into skin (Bennett et al., 2012) and on the other hand can facilitate reactive oxygen species (ROS) generation through exciting electrons into conduction band to form superoxide (O2•−) and generating holes in valence band to form hydroxyl radicals (HO•), which has led to serious concerns on the risk of TiO2 NPs to human skin health (Brezova et al., 2005; Serpone et al., 2007). Unfortunately, more severe toxicity is caused by mixed-phase TiO2 NPs compared with pristine NPs (Gerloff et al., 2012; Wu et al., 2009; Yin et al., 2012). In vitro study indicates mixed-phase TiO2 NPs can induce the enhanced Lactate Dehydrogenase (LDH) leakage and oxidative stress responses in HaCaT skin cells (Yin et al., 2012), whereas in vivo study reveals a large reduction in the collagen content of mouse skin when exposed to mixed-phase TiO2 NPs (Wu et al., 2009). Obviously, mixed-phase TiO2 NPs are more dangerous than pristine NPs, however, the underlying mechanism remains unclear. Mixed-phase TiO2 NPs usually contain high density of anatase/rutile heterojunctions that are constructed by staggered band edges arising from the different band gaps of anatase and rutile (Tiwari et al., 2016). Previous studies have demonstrated that the band alignment at the interface of heterojunction can drive electrons flow from rutile to anatase with holes moving in the opposite directions (Deák et al., 2016; Kullgren et al., 2015; Scanlon et al., 2013; Sun et al., 2015), leading to an efficient electron-hole separation. The free electron and hole can react with oxygen and water molecules to form O2•− and HO• radicals, respectively, potentially causing potent oxidative stress-mediated cellular injury (Zhang et al., 2014). This means there probably existing a property-activity relationship between heterojunction and toxicity of mixed-phase TiO2 NPs, which can be used to deeply understand the toxicity mechanism of mixed-phase TiO2 NPs and further predict their toxicity potential. However, it is currently hard to establish this relationship because there is a lack of a feasible way to determine the density of heterojunction. So, it is necessary to develop a method to reflect the density of heterojunction. Since the heterojunction-assisted electron-hole separation can lead to charge accumulation at the interface of heterojunction, a band bending (BB) effect usually occurs (Zhang et al., 2014). The degree of BB is proportional to the amount of accumulated charges and further proportional to the density of heterojunction (Edmonds et al., 2011). Thus, determination of the BB level becomes a feasible approach to reflect the density of heterojunction and further predict the toxicity potential of mixed-phase TiO2 NPs. In this study, a series of anatase/rutile mixed-phase TiO2 NPs with different anatase/rutile ratios were prepared to display different densities of heterojunction, which could be used to investigate the potential property-activity relationship of mixed-phase NP sin keratinocyte HaCaT skin cells and mouse skin under exposure to simulated sunlight. It was hypothesized that the staggered energy bands at the interface of anatase/rutile heterojuctions could trigger the electron-hole separation and the BB until Fermi energies of anatase and rutile were aligned (Figure 1A), and the formed free charges could produce ROS and cause oxidative injury in cells and mice skin, where the density of heterojunction and the extent of injury could be weighed by the BB level (Figure 1B). Energy structures and the BB levels of these NPs were systemically examined to correlate with ROS production and induced oxidative stress toxicological responses in cells and mice. Actually, it was found that under sunlight irradiation, TiO2 NPs with various anatase/rutile ratios could generate different magnitudes of O2•− and HO• radicals, and cause different levels of heme oxygenase-1 (HO-1) expression, interleukin 8 (IL-8) release, and mitochondrial dysfunction in HaCaT skin cells, and ultimately leading to cell death. Further in vivo evaluation of the activity of superoxide dismutase (SOD) and the content of malondialdehyde (MDA) in mouse skin also supported this relationship between the injury and the BB. Thus, the biological effects of mixed-phase TiO2 NPs can be well elucidated by their semiconductor properties. Figure 1. View largeDownload slide Potential heterojunction-toxicity relationship of mixed-phase TiO2 NPs. (A) Band alignment-driven electron-hole separation, BB induction, and ROS production at the interface of heterojunction. (B) Density variation of heterojunction during the transformation of anatase to rutile phase. Figure 1. View largeDownload slide Potential heterojunction-toxicity relationship of mixed-phase TiO2 NPs. (A) Band alignment-driven electron-hole separation, BB induction, and ROS production at the interface of heterojunction. (B) Density variation of heterojunction during the transformation of anatase to rutile phase. MATERIALS AND METHODS Chemicals Titanium (IV) isopropoxide (≥97%), 1-butanol (anhydrous, 99.8%), and poly-(ethyleneglycol)-block-poly(propyleneglycol)-block-poly(ethyleneglycol)PEO20-PPO70-PEO20 (Pluronic P123) were purchased from Sigma-Aldrich. All chemicals were reagent grade and used without further purification or modification unless otherwise indicated. Water throughout all experimental procedures was obtained by using a Milli-Q water purification system. Synthesis of pristine and mixed-phase (anatase/rutile) TiO2 NPs Mixed-phase TiO2 NPs were synthesized through inverse micelle template sol-gel method under solvothermal conditions and using titanium isopropoxide as a precursor (Luo et al., 2015). In a typical synthesis, 0.01 mol (2.84 g) titanium isopropoxide was dissolved in a solution containing 0.094 mol (7 g) of 1-butanol, 0.016 mol (1 g) of concentrated HNO3 and 1.72 × 10−4 mol of P123 surfactant in a 150-ml beaker at room temperature (RT) under magnetic stirring. The obtained clear gel was dark yellow in appearance and placed in an oven running at 120°C for 4 h. The pristine and mixed-phase TiO2 NPs were obtained by calcining the dark orange and transparent (rigid) films under air conditions for 4 h at various temperatures (600°C–1000°C). After the temperature dropped to RT, the products were taken out and washed with water and ethanol alternately for several times, and dried in an oven at 80°C overnight. Physicochemical characterization of pristine and mixed-phase TiO2 NPs X-ray diffraction (XRD) patterns were collected using D8 ADVANCE (BRUKER company, Germany) (Cu Kα radiation) over a range of 15°−80° 2θ. Transmission electron microscopy (TEM) and high resolution TEM (HRTEM) were performed using a JEOL 2100 F microscope operated at 200 keV. Specific surface area was determined by the BET method from N2 sorption isotherms, and acquired using a Micrometrics ASAP 2010 sorption instrument. Zeta potential and dynamic light scattering (DLS) data were collected on a Malvern Nanosizer ZS. Micro-Raman spectra were recorded in backscattering geometry using LabRam HR 800 (HORIBA Jobin Yvon) coupled with an Olympus BX 41microscope; the confocal hole and the silt width were both fixed at 200 μm, and the excitation wavelength of 633 nm was used. Ultraviolet photoelectron spectroscopy (UPS) measurements were performed on a PREVAC XPS/UPS system spectrometer equipped with monochromatic Al Kα radiation (hν = 1486.7 eV) as X-ray source and a He I radiator (hν = 21.2 eV) as UV source. All spectra were collected at RT in an ultra-high-vacuum environment (base pressure of analysis chamber < 5× 10−8 Torr). All UPS-binding energies are referenced to the Fermi edge of a clean Au foil. For UPS measurement, indium tin oxide (ITO) slide was first washed by water, acetone, and ethanol, respectively for several times in order to remove the contamination. The TiO2 NPs dispersed with deionized water at a concentration of 2.5 mg/ml were dropped onto the ITO surface and left to dry. The valence band maximum (VBM) was determined by linear extrapolation of the leading edge of the UPS spectrum, and the work function was determined from one-half the height of the secondary electron onset. Band-gap energies were obtained from diffuse reflectance (DR) UV-vis spectroscopic analysis (PerkinElmer UV-VIS Spectrometer Lambda 35) by using BaSO4 as a reference material. All measurements were conducted in ambient air using a bandwidth of 1.0 nm. Collected DR UV-vis spectra were converted into Kubelka−Munk function (F[R∞]) spectra. Sunlight exposure experiments were conducted on Sun 2000 Solar Simulator (Abet Technologies). Confirmation of the percentage of rutile phase in the mixed-phase TiO2 NPs by XRD and Raman spectra The percentage of rutile phase in the mixed-phase of TiO2 NPs was determined based on XRD analysis through the equation: WR = IR/(0.884IA+ IR), where WR is the weight percentage of rutile phase, IR and IA are the integrated intensity of rutile (110) and anatase (101) peaks. The percentage of rutile phase can be further confirmed by Raman spectra analysis. The Raman spectra ranging from 300 to 700 cm−1 was deconvoluted into anatase phase (containing peaks at 395, 518, and 639 cm−1) and rutile phase (containing peaks at 448 and 609 cm−1). The percentage of rutile phase was calculated by A448+609 cm−1/(A395+518+639 cm−1+A448+609 cm−1), where A448+609 cm−1is the integrated peak area of rutile phase at 448 and 609 cm−1 and A395+518+639 cm−1 is the integrated peak area of anatase phase at 395, 518, and 639 cm−1 (Frank et al., 2012). Abiotic total ROS measurement of pristine and mixed-phase TiO2 NPs The total ROS generation of TiO2 NPs was determined by 2,7-dichlorodihydrofluorescein diacetate (H2DCFDA) fluorescence (Sigma). According to the previous method (Zhang et al., 2014), the 2′,7′-dichlorofluorescein (DCF) final working concentration was 29 μmol/l. This could be obtained by the following process: 50 μg of H2DCFDA, 17.3 μl of ethanol and 692 μl of a 0.01 mol/l sodium hydroxide solution were mixed well. The resulting solution was incubated at 4°C for 30 min under dark condition, and 3500 μl of a sodium phosphate buffer (25 mmol/l, pH =7.4) was added to obtain 29 μmol/l DCF working solution. To each well of a 96-multiwell black plate (Costar, Corning, New York), we added 80 μl of 29 μmol/l DCF working solution. In total 20 μl of 1 mg/ml NP aqueous suspension was subsequently added to each well (final concentration is 200 μg/ml), followed by 15 min of exposure to simulated sunlight (Sun 2000 Solar Simulator, Abet Technologies) and another 2 h of incubation at RT under dark condition. DCF fluorescence emission spectra in the range of 500–600 nm were collected using a SpectraMax M5 microplate reader with an excitation wavelength of 490 nm. In comparison, the DCF fluorescence enhancement by NPs without sunlight exposure was also carried out according to the similar procedure except exposure to the simulated sunlight. Superoxide and hydroxide radical measurements by electron spin resonance Superoxide radicals (O2•−) and hydroxide radicals (HO•) were detected by electron spin resonance (ESR) technology. ESR signals of spin-trapped paramagnetic species with 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) were recorded with a Brucker A300 spectrometer. For O2•− detection, 15 μl of 1.0 mol/l DMPO was mixed well with 50 μl of 50 μg/ml TiO2 suspension in methanol, while for HO• detection, 10 μl of 1.0 mol/l DMPO was mixed with 50 μl of 50 μg/ml TiO2 NPs aqueous suspension. The resulting mixture was irradiated by simulated sunlight for 15 min and then immediately subjected to the ESR measurement of O2•− and HO•, respectively. Controlled experiments without simulated sunlight were also carried out according to the similar procedure except the irradiation with simulated sunlight. Cell culture Human Keratinocytes cell line HaCaT (ATCC Number: PCS-200-011) were cultured in vented T-75 cm2 flasks (Corning, Fisher Scientific) in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal calf serum (GiBCO), 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine at 37°C in a humidified atmosphere of 5% CO2. The culture medium was replaced at 70%–80% confluency every 2 days. Cellular uptake of pristine and mixed-phase TiO2 NPs determined by ICP-OES There were 8 × 104 HaCaT cells in 800 μl of DMEM were incubated in 12-well plates (Costar, Corning) for overnight growth. The medium was removed, and cells were incubated with 800 μl of 200 μg/ml TiO2 NP DMEM suspension for 6 h. After removal of culture medium, cells were washed with PBS 3 times and harvested through trypsinization, and digested by concentrated nitric acid for ICP-OES assessment. Cytotoxicity assessment by MTS assays Viability of HaCaT cells exposed to different TiO2 NPs with or without simulated sunlight exposure was determined by an MTS assay. 1 × 105 HaCaT cells in 100 μl of DMEM were seeded in each well of 96-well plates and incubated for 24 h. The medium was removed, and cells were treated with 100 μl of TiO2 NP suspensions at concentrations of 0.4–200 μg/ml for 6 h, followed by exposure to the simulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for 18 h. Then, the culture medium was removed and cells were washed with PBS 3 times. Each well received 100 μl of culture medium containing 16.7% of MTS stock solution, followed by 1 h of incubation at 37°C in a humidified 5% CO2 incubator. The plate was centrifuged at 2000 × g for 10 min in Xiangyi L535R with a microplate rotor to spin down the NPs and cell debris. In toal 80 μl of the supernatant was transferred into a new 96-multiwell plate. The absorbance of formazan was read at 490 nm on a SpectraMax M5 microplate reader (Molecular Devices Corp., Sunnyvale, California). In comparison, the viability of cells treated with TiO2 NPs but without sunlight exposure was also tested following the similar procedure. Cellular ROS measurement by flow cytometer There were 8 × 104 HaCaT cells in 800 μl of DMEM were plated in each well of 12-well plates (Costar, Corning) for overnight growth. After removal of the medium, and cells were incubated with 800 μl of 50 μg/ml different TiO2 NPs suspension in DMEM medium for 6 h, and then cells were exposed to simulated sunlight for 15 min at a power density of 0.1 W/cm2 and further cultured for 6 h. Cells only exposed to the sunlight were used as a control. Next, the culture medium was removed and cells were stained with 100 μl of 10 μmol/ml DCFH-DA for 30 min. After washing with PBS for 3 times, the harvested cells were transferred to polystyrene tubes for acquisition and analysis by flow cytometer (BD Accuri C6). In comparison, ROS generation of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. Cellular Glutathione (GSH) measurements based on a 5,5′dithio 2-nitrobenzoic acid (DNTB) method In total 8 × 104 HaCaT cells were plated in each well of 12-well plates (Costar, Corning) for overnight growth. After removal of the medium, and cells were incubated with 800 μl of 50 μg/ml different TiO2 NP suspension in DMEM medium for 6 h, and then cells were exposed to simulated sunlight for 15 min at a power density of 0.1 W/cm2 and further cultured for 6 h. Cells were washed with PBS 3 times and resuspended in cell lysis buffer (Beyotime). After centrifugation, 10 μl of lysate was mixed with 150 μl of 30 μg/ml DNTB in 1 of 96-well plates. The absorption of the resulting mixture at 412 nm was recorded on a SpectraMax M5 microplate spectrophotometer. In comparison, GSH levels of the cells treated with TiO2 NPs but without exposure to sunlight were also determined following the similar procedure. Western blot analysis for HO-1 expression There were 1.6 × 105 HaCaT cells in 1.6 ml of culture medium seeded into each well of 6-well plates (Costar, Corning). After overnight growth, cells were treated with 1.6 ml of 50 μg/ml NP suspension for 6 h, and then exposed to the stimulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for 6 h. Cells only exposed to the sunlight were used as a control. Cells were washed with PBS 3 times and harvested through scraping. Cell pellets were resuspended in cell lysis buffer containing protease inhibitors. The lysate was sonicated briefly and centrifuged, and the protein content in the supernatant was measured by the Bradford method. A 30 μg amount of total protein from each sample was electrophoresed by 10% SDS-PAGE and transferred to a PVDF membrane. After blocking, membranes were incubated with antihuman HO-1 or β-actin monoclonal antibody (1:1000) (Beyotime) for 12 h. After being washed with TBS/T buffer 3 times, membranes were incubated for 12 h with horseradish peroxidase conjugated secondary antibody (dilution 1:1000) (Beyotime, Nanjing). After being washed, membranes were developed using an enhanced Super Signal West Pico chemiluminescent substrate (Thermo Scientific, USA). The intensity of HO-1 was normalized to that of β-actin using Image J software. In comparison, HO-1 expression of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. IL-8 cytokine quantification by ELISA There were 1 × 105 HaCaT cells in 100 μl of DMEM plated in each well of a 96-multiwell black plate (Costar, Corning) for overnight growth. The medium was removed, and cells were treated with NP suspensions (50 μg/ml) in DMEM for 6 h, and exposed to the stimulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were cultured for another 6 h. Plates were centrifuged at 2000 × g for 10 min in Xiangyi L535R with a microplate rotor to spin down the cell debris and NPs. In total 50 μl of the supernatant from each well was used for measurement of IL-8 activity in HaCaT cells using an OptEIA (BD Biosciences, California) ELISA kit according to the manufacturer’s instructions. Briefly, a 96-well plate was coated with 100 μl of monoclonal antiIL-8 antibody overnight. After removal of the unbound antibody, a standard cytokine dilution series or 50 μl of each supernatant were pipetted into the precoated wells for antigen capture. After 2 h of incubation, the unbound growth factor was removed and each well was washed with a buffer 5 times and an enzyme-linked secondary polyclonal antibody added. Following washing, a substrate solution (1:250) was added into each well to allow color development. After termination of the reaction, the colorimetric intensity was measured at 450 nm on a SpectraMax M5 microplate reader. In comparison, IL-8 release of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. Fluorescence microscopy to investigate the mitochondrial dysfunction Mitochondrial membrane depolarization and superoxide generation were investigated by 5, 5’, 6, 6’-tetrachloro-1, 1’, 3, 3’-tetraethylbenzimidazolocarbocyanine iodide (JC-1) and MitoSox Red fluorescent indicators, respectively. There were 2 × 105 HaCaT cells in 800 μl of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at power intensity of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 6 h at 37°C in a humidified 5% CO2 incubator. Then, the culture medium was removed, and cells were washed with PBS 3 times. Each well received 800 μl of fresh culture medium containing Hoechst 33342 (1 μmol/l) and JC-1 (5 μmol/l) or Hoechst 33342 (1 μmol/l) and MitoSox Red (5 μmol/l). After 30-min incubation, fluorescent images were taken on Olympus BX-51 Optical System Microscope (Tokyo, Japan) with 20× objective. In comparison, the mitochondrial membrane depolarization and superoxide generation of cells treated with TiO2 NPs but without sunlight exposure were also investigated following the similar procedure. Flow cytometric analysis of cell cycle Cell cycle distribution was examined by evaluating the relative cellular DNA content with a flow cytometric technique. In total 2 × 105 HaCaT cells in 800 μl of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at light power intensity of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 18 h at 37°C in a humidified 5% CO2 incubator. Then the cells were detached by trypsin and washed 3 times with PBS. The harvested cells were fixed with ice-cold 70% (v/v) ethanol at 4°C for at least 6 h. Afterwards, cell pellets were washed twice with ice-cold PBS and resuspended in 1 ml of PBS containing (1 mg/ml) RNase and (50 μg/ml) propidium iodide (PI), and then incubated for 30 min in the dark at RT. Finally, samples were examined by a flow cytometer (BD Accuri C6). Cell cycle was analyzed according to the distribution of DNA content and divided into G1, S, and G2/M phases. Flow cytometric analysis of apoptotic cells There were 2 × 105 HaCaT cells in 800 μL of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at a power densityof 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 18 h at 37°C in a humidified 5% CO2 incubator. Then the cells were detached by trypsin (without ethylenediaminetetraacetic acid) and washed 3 times with PBS. Cells were re-suspended in 500 μl of binding buffer (10 × 10−3 M 4-[2-hydroxyethyl]-1-pipera zineethanesulfonic acid, 140 × 10−3 M NaCl, 2.5 × 10−3 M CaCl2, pH = 7.4), and mixed with 5 μl of 100 μg/ml FITC-conjugated Annexin V and 5 μl of 100 μg/ml PI, and incubated in the dark for 10 min. Then, the cell apoptosis was analyzed by a flow cytometer. Senescence-associated β-galactosidase staining There were 2 × 105 HaCaT cells in 800 μL of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 6 h at 37°C in a humidified 5% CO2 incubator. Then, the cells were fixed in 4% glutaraldehyde solution for 15 min at RT. The fixed cells were washed with PBS 3 times and then incubated with 200 μl of β-galactosidase staining solution (contain 20 mg/ml X-gal) (KeyGEN Bio TECH) over night at 37°C without CO2. Cells treated with 100 μM H2O2 were performed as positive control. Animals and treatment All animal studies were performed in Center for Experimental Animals, Jilin University, and the procedures involving experimental animals were in accordance with protocols approved by the Committee for Animal Research of Jilin University (SYXK [JL]2008-0011), China. The Balb/c hairless mice (4-weeks old, female, around 18 g) were randomly divided into 6 groups with 6 mice per group. Animals were acclimatized to laboratory conditions for 1 week prior to experiments and were on standard animal chow and water ad libitum with 12-h light/dark cycle. The temperature of the room was maintained at 18°C–22°C and the relative humidity about 50%–60%. One group was for the control group without any NP treatment, while the other 5 experimental groups were treated by 5 different types of TiO2 NPs including A, A7:R3, A4:R6, A3:R7, and R. Suspensions of different types of TiO2 NPs of 8 wt% were made with Pentalan-408 (pentaerythritoltetraethylhexanoate), which is similar to the sunscreen formulation (Pal et al., 2016), followed by 5 min of sonication to provide a uniform formulation. The dorsal surface of mice in experimental groups was exposed to 100 μl of above TiO2 NP formulations with a 4 cm2 (2 × 2 cm2) area for 3 h under simulated sunlight exposure, and the remaining residual NPs were removed from skin with lukewarm water. The mice in control group were only exposed to simulated sunlight for 3 h. After 14 consecutive day treatment, the mice were sacrificed and the skin was removed for further assay. SOD activity and MDA content analysis of the mouse skin The dorsal skin that was exposed to TiO2 NPs was exercised and washed with ice-cold PBS 3 times, and then homogenized in ice-cold PBS by ultrasonic cell sonicator at 200 W for 10 min (0°C). Homogenates were centrifuged at 8000× g for 15 min (4°C) and the supernatants were collected and stored at −20°C. The activities of SOD and the contents of MDA in dorsal skin homogenates were examined by standard reagent kit (Beyotime, Nanjing). Histopathologic analysis of the mouse skin Balb/c hairless mice were sacrificed at the termination of the experiment under anesthesia. The dorsal skin about 8 µm thick containing the epidermis that was exposed to TiO2 NPs were taken and washed with cold PBS for 3 times. After that, the skin tissues were fixed in 4% paraformaldehyde in 0.4 M phosphate buffer (pH 7.6). The skin tissues were dehydrated and embedded in paraffin, and then cut into 4 μm sections. The sections were stained with hematoxilin and eosin (H&E) and subsequently processed for histopathological examination under a light microscope. The thickness of the epidermis was measured afterwards by means of the light microscope (Olympus BX-51 Optical System Microscope [Tokyo, Japan]) with Image J system. Three to five views under the microscope were choice for measurement. The average thickness was recorded for compare. Statistical analysis All data were expressed as mean ± SD. All values were obtained from at least 3 independent experiments. Statistical significance was evaluated using 2-tailed heteroscedastic Student’s t tests according to the TTEST function in Microsoft Excel. The significant difference between groups was considered statistically significant when the p-value was < .05. RESULTS Physicochemical Characterization of Mixed-Phase TiO2 NPs With Different Anatase/Rutile Ratios The ratios of anatase/rutile were precisely adjusted by controlling the process of phase transformation from anatase (A) to rutile (R) through 600°C–1000°C, while other reaction conditions were kept consistent for achieving similar physicochemical properties. The phase ratio of A to R in mixed-phase TiO2 was determined by XRD (Luo et al., 2015). Figure 2A indicates that with the temperature increasing, the diffraction peak intensity of rutile phase gradually increases, while the corresponding intensity of anatase phase declines. Based on the XRD analysis, the anatase/rutile ratios of mixed-phase TiO2 NPs were determined as 7:3, 4:6, and 3:7, and the corresponding particles were named as A7:R3, A4:R6, and A3:R7, respectively. Raman spectrum is another widely used method to determine the phase transfer from anatase to rutile (Fang et al., 2008; Frank et al., 2012). Figure 2B shows that with the temperature increasing the Raman peak at 638 cm−1 attributed to anatase phase is gradually shifted until to 612 cm−1 that is attributed to rutile phase, corroborating the formation of mixed-phase NPs with different phase ratios. TEM images revealed the spherical morphology of these TiO2 NPs (Figure 2C), and HRTEM images (Figure 2C) further displayed the formed anatase/rutile heterojunctions in A7:R3, A4:R6, and A3:R7 regarding the coexisting anatase/rutile lattice fringes. The lattice fringe of anatase was about 0.35 nm, orientated in the (101) direction, while it was 0.32 nm for rutile corresponding to (110) direction (Ruan et al., 2013). The primary sizes of these NPs ranged from 80 ± 8 to 100 ± 9 nm (Figure 3A), and the corresponding specific surface area were determined by Brunauer–Emmett–Teller (BET) ranging from 39.9 ± 2.2 to 89.9 ± 4.7 m2/g (Figure 3B). Assessments of hydrodynamic sizes using DLS method demonstrated that all the TiO2 NPs could be well dispersed in water and their hydrodynamic sizes ranged from 315.0 ± 27.8 to 375.9 ± 11.1 nm (Figure 3C). ζ-potential measurements indicated all these TiO2 NPs had positive surface charges ranging from +22.58 ±2.47 to +26.99 ±1.83 mV in water (Figure 3D). These TiO2 NPs could be also well dispersed in cell culture medium (DMEM), showing similar hydrodynamic sizes of 354.0 ± 20.0 to 394.6 ± 4.2 nm (Supplementary Figure 1). In summary, the characterization data reveal that mixed-phase TiO2 NPs with different anatase/rutile ratios have been successfully obtained, showing homologous physicochemical properties comparable to pristine TiO2 NPs. Figure 2. View largeDownload slide Determination of anatase/rutile phase ratios in mixed-phase TiO2 NPs. (A) XRD patterns (The red and black vertical line represent the standard XRD pattern for anatase (JCPDS card No. 04-0477) and rutile TiO2 (JCPDS card No.04-0551), respectively, (B) Raman spectra and (C) TEM and HRTEM images. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Figure 2. View largeDownload slide Determination of anatase/rutile phase ratios in mixed-phase TiO2 NPs. (A) XRD patterns (The red and black vertical line represent the standard XRD pattern for anatase (JCPDS card No. 04-0477) and rutile TiO2 (JCPDS card No.04-0551), respectively, (B) Raman spectra and (C) TEM and HRTEM images. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Figure 3. View largeDownload slide Physicochemical properties of pristine and mixed-phase TiO2 NPs. (A) Primary size; (B) Specific surface area; (C) Hydrodynamic sizes; and (D) Zeta potentials. A 50 μg/ml NPs dispersed in water for measurement of hydrodynamic sizes and zeta potentials. Figure 3. View largeDownload slide Physicochemical properties of pristine and mixed-phase TiO2 NPs. (A) Primary size; (B) Specific surface area; (C) Hydrodynamic sizes; and (D) Zeta potentials. A 50 μg/ml NPs dispersed in water for measurement of hydrodynamic sizes and zeta potentials. BB Level for Evaluating ROS Production The formation of anatase/rutile heterojunctions provides the opportunity for the electron transfer between anatase and rutile phase if the staggered band edges get matched. The energy band structures of pristine and mixed-phase TiO2 NPs were constructed based on their band-gap energies (Eg), Fermi energies (Ef), and VBM energies that were determined by UV-visible diffuse reflection spectroscopy and UPS, respectively. Figure 4A and Supplementary Table 1 summarize the band energy information. The conduction- and valence-band edges of rutile TiO2 NPs were higher than those of anatase TiO2 NPs, respectively, meaning the excited electrons preferentially move from rutile to anatase while holes from anatase to rutile, which is consistent with previous results (Kullgren et al., 2015; Scanlon et al., 2013). This electron-hole separation at the interface of heterojunction can result in free electron and hole accumulation in anatase and rutile sides, respectively, which can induce the BB (Supplementary Figure 2) (Zhang and Yates, 2012). Since the degree of BB is proportional to the amount of accumulated electrons or holes that depend on the density of heterojunction, the degree of BB can reflect the density of heterojunction in mixed-phase TiO2 NPs. Figure 4B presents the BB value variation of these NPs, where A4:R6 shows the largest one, followed by A7:R3 and A3:R7. Pristine TiO2 (A and R) did not show any BB effect because there is no heterojunction interface and charge accumulation. The BB analysis suggested that the notable but varied charge accumulation existing in the mixed-phase TiO2 NPs with the phase ratio changing. The free electrons and holes at the interface of heterojunction can react with oxygen and water to form O2•− and HO•, respectively (Brezová et al., 2014; Dvoranova et al., 2014). The fluorescent dye, DCF, is usually used to generally detect the abiotic ROS production on nanomaterials (Zhang et al., 2014). Supplementary Figure 3A shows the significantly enhanced DCF fluorescence intensity induced by mixed-phase TiO2 NPs compared with pristine NPs after exposure to simulated sunlight, where A4:R6 exhibits the strongest fluorescence intensity, followed by A7:R3 and A3:R7. This trend is consistent with that of the degree of BB. However, only weak fluorescence enhancement was observed in these NPs without exposure to sunlight (Supplementary Figure 3B). Further ESR spectra distinguished the ROS types and revealed the more remarkable O2•− and HO• generation in mixed-phased TiO2 NPs than pristine particles (Figs. 4C and 4D) (He et al., 2014), supporting the ROS generation trend in the DCF assay. In comparison, without sunlight exposure, O2•− and HO• radicals were rarely induced by both pristine and mixed-phase TiO2 NPs (Supplementary Figure 4). Figure 4. View largeDownload slide Electronic properties of pristine and mixed-phase TiO2 NPs. (A) Energy structures, (B) BB degree. (C) ESR spectra for O2•−. (D) ESR spectra for HO•. Figure 4. View largeDownload slide Electronic properties of pristine and mixed-phase TiO2 NPs. (A) Energy structures, (B) BB degree. (C) ESR spectra for O2•−. (D) ESR spectra for HO•. All above results demonstrate that free electrons and holes can be derived from heterojunctions in mixed-phase TiO2 NPs and give rise to the induction of abiotic ROS generation upon sunlight exposure, and the degree of BB can well reflect the magnitude of ROS in mixed-phase TiO2 NPs. Toxic Potential Reflected by BB Level Both O2•− and HO• radicals are highly reactive, capable of oxidizing a vast of biomacromolecules and cell organelles, and causing cell oxidative injury (Li et al., 2012). After incubated with these different TiO2 NPs (200 µg/ml) for 6 h, HaCaT cells showed similar cellular uptake of TiO2 NPs based on the cellular Ti contents as determined by inductively coupled plasma optical emission spectrometry (ICP-OES) (Supplementary Figure 5). To benchmark the relative toxicity of the mixed-phase TiO2 NPs compared with pristine particles, the cellular viability of HaCaT cells was assessed by MTS assays upon exposure to 0.4-200 µg/ml TiO2 NPs for 6 h, followed by 15 min of exposure to simulated sunlight and another 18 h of incubation. Figure 5A shows that mixed-phase TiO2 NPs can induce more robust decline in cell viability than anatase and rutile TiO2, and A4:R6 still exhibits the most potent toxicity, followed by A7:R3 and A3:R7, which is consistent with the tends of ROS production and BB levels of these mixed-phase TiO2 NPs. To reveal the critical ROS role in toxicity induced by these TiO2 NPs, N-acetyl-cysteine (NAC) as ROS scavenger was used to pretreat the HaCaT cells (Xue et al., 2011), and the cell viability was assessed again by MTS assay following treatment with TiO2 NPs under sunlight exposure. The result shows the cytotoxicity of mixed-phase TiO2 NPs can be significantly reduced in HaCaT cells after NAC pretreatment (Supplementary Figure 6), demonstrating ROS scavenger can protect skin cells from ROS attacking of mixed-phase TiO2 NPs under sunlight exposure. In comparison, without sunlight exposure, the cell viability of HaCaT cells was weakly reduced by these TiO2 NPs (Supplementary Figure 7A). This toxicity study demonstrates that mixed-phase TiO2 NPs can cause more severe toxicity than pristine TiO2 NPs in HaCaT cells upon exposure to sunlight, with the larger BB level corresponding to the stronger toxicity. Sunlight-mediated ROS generation plays a critical role in induction of toxicity of mixed-phase TiO2 NPs. Figure 5. View largeDownload slide Cytotoxicity, cellular ROS generation and GSH depletion in HaCat cells exposed to TiO2 NPs under sunlight condition. (A) MTS assay, (B) Flow cytometric analysis for cellular DCF fluorescence (inset showing the fold increase in integral area), and (C) Cellular GSH assessment based on DNTB methods. Cells were treated with 0.4-200 µg/ml (for MTS assay) or 50 µg/ml (for ROS and GSH assessments) of TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h (for MTS assay) or 6 h (ROS and GSH assessments) of incubation, respectively. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 9, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 5. View largeDownload slide Cytotoxicity, cellular ROS generation and GSH depletion in HaCat cells exposed to TiO2 NPs under sunlight condition. (A) MTS assay, (B) Flow cytometric analysis for cellular DCF fluorescence (inset showing the fold increase in integral area), and (C) Cellular GSH assessment based on DNTB methods. Cells were treated with 0.4-200 µg/ml (for MTS assay) or 50 µg/ml (for ROS and GSH assessments) of TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h (for MTS assay) or 6 h (ROS and GSH assessments) of incubation, respectively. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 9, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Intracellular ROS Production and GSH Depletion Since abundant O2•− and HO• radicals are generated in mixed-phase TiO2 NPs upon exposure to simulated sunlight, the molecular mechanism underlying the toxicity of these TiO2 NPs is probably ascribed to activation of oxidative stress (Hamzeh and Sunahara, 2013). Intracellular ROS production and GSH depletion usually are indicative of oxidative stress injury. Flow cytometry analysis was carried out to detect the intracellular ROS level using H2-DCF-DA, while the classic DNTB method was employed to determine the intracellular GSH level. Figure 5B shows the mixed-phase TiO2 NPs can induce more potent intracellular DCF fluorescence than pristine TiO2 NPs in HaCaT cells after exposure to sunlight. The fluorescence enhancement trend based on the fold increase in integral area (inset of Figure 5B) is still in accordance with that of the degree of BB, showing the largest fluorescence enhancement induced by A4:R6, followed by A7:R3 and A3:R7. Figure 5C further displays the corresponding cellular GSH levels, where mixed-phase TiO2 NPs can more remarkably reduce the cellular GSH level than pristine particles, and A4:R6 still is the most potent one. In comparison, without sunlight exposure, ROS and GSH levels in HaCaT cells were rarely affected by TiO2 NPs (Supplementary Figs. 7B and 7C). Cellular ROS production and GSH depletion strongly imply the toxicity mechanism of mixed-phase TiO2 NPs under sunlight exposure is correlated with oxidative stress. Hierarchical Oxidative Stress Responses According to oxidative stress paradigm, the hierarchical oxidative stress responses are characterized by an antioxidant defense response (tier 1), the initiation proinflammatory (tier 2), and mitochondrial-mediated cytotoxicity (tier 3) (Nel et al., 2006). The lowest levels of oxidative stress (tier 1) are associated with the activation of cytoprotective enzymes such as HO-1. Western blotting analysis (Figure 6A) revealed that, with simulated sunlight exposure, A4:R6 triggered the most abundant HO-1 expression, followed by A7:R3 and A3:R7, which is proportional to the degree of BB. Failure to restore redox equilibrium in tier 1 is capable of activating proinflammatory signaling pathways such as the Jun kinase and NF-κB cascades, which are involved in the transcriptional activation of cytokine, chemokine, and adhesion gene promoters. IL-8 as typical tier 2 proinflammatory response has been investigated to demonstrate the inflammatory effects in HaCaT cells (Comfort et al., 2014). An ELISA assay to detect IL-8 release in the cellular supernatant of HaCaT cells indicated the incremental increase in chemokine production with the increase of BB in mixed-phase TiO2 NPs (Figure 6B). Escalation of oxidative stress response to tier 3 can trigger mitochondrial dysfunction, including mitochondrial superoxide generation and mitochondrial membrane depolarization, ultimately leading to cell death. The fluorescent dyes, MitoSox Red and JC-1, were applied to detect the mitochondrial superoxide generation and mitochondrial membrane depolarization, respectively, through fluorescence microscopy. Figure 6C shows both mitochondrial-mediated toxicological responses are found the most significant in the cells exposed to A4:R6, and these mitochondrial dysfunction effects still followed the trend of BB degrees. Also, in comparison, without sunlight exposure, above tier 1 to tier 3 hierarchical oxidative stress responses including HO-1 expression, IL-8 release and mitochondrial dysfunction were not induced by mixed-phase and pristine TiO2 NPs (Supplementary Figure 8). All these in vitro hierarchical oxidative stress responses corroborate the activation of oxidative stress by mixed-phase TiO2 NPs under sunlight exposure. Figure 6. View largeDownload slide Hierarchical oxidative stress responses in HaCaT cells exposed to TiO2 NPs under sunlight condition. (A) Western blot analysis for HO-1 expression; (B) ELISA assessment for IL-8; and (C) Fluorescence images of cells stained by Mitosox and JC-1 to detect mitochondrial superoxide generation and membrane depolarization. Cells were treated with 50 µg/ml of various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 6 h of incubation. Flow cytometric analysis for cell cycle phase arrest (D) and apoptotic cells (E) of the HaCaT cells. Cells were treated with 50 µg/ml various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h of incubation, and stained with PI for cell cycle detection or Annexin V-FITC/PI for apoptosis detection. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 6. View largeDownload slide Hierarchical oxidative stress responses in HaCaT cells exposed to TiO2 NPs under sunlight condition. (A) Western blot analysis for HO-1 expression; (B) ELISA assessment for IL-8; and (C) Fluorescence images of cells stained by Mitosox and JC-1 to detect mitochondrial superoxide generation and membrane depolarization. Cells were treated with 50 µg/ml of various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 6 h of incubation. Flow cytometric analysis for cell cycle phase arrest (D) and apoptotic cells (E) of the HaCaT cells. Cells were treated with 50 µg/ml various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h of incubation, and stained with PI for cell cycle detection or Annexin V-FITC/PI for apoptosis detection. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Cell Cycle Distribution, Apoptosis, and Senescence Oxidative stress has been known capable of inducing cell death through apoptosis and senescence (Fulda et al., 2010; Haines et al., 2013; Kim et al., 2017; Ozben, 2007). To identify the type of cell death induced by mixed-phase TiO2 NPs, additional investigation on cell cycle distribution, apoptosis and senescence were performed. The effect of these NPs on the cell cycle progression was examined by flow cytometry using PI staining. Figure 6D shows that all these TiO2 NPs can lead to a significantly increased accumulation of the G2/M-phase cells and reduced number of G1-phase and S-phase cells, where A4:R6 induces the most significant G2/M arrest (10.8%), followed by A7:R3 (9.0%), and A3:R7 (8.5%). Then, induction of apoptosis by these NPs was evaluated by flow cytometry using Annexin V-FITC/PI staining. Figure 6E shows that A, A7:R3, A4:R6, A3:R7, and R TiO2 NPs can induce 24%, 37.5%, 58.2%, 34.8%, and 27.8% of the early apoptotic cells, and 1.1%, 0.1%, 0.5%, 0.4%, and 0.1% of the late apoptotic cells, respectively, where A4:R6 still induces the largest magnitude of apoptotic cells. Moreover, cellular senescence was detected by bright field microscope based on senescence-associated beta-galactosidase (SA-β-gal) activity (Marazita et al., 2016). As shown in Supplementary Figure 9, after treatment with mixed-phase or pure phase TiO2 NPs, HaCaT cells did not display positive (SA-β-gal) staining, while 100 μM H2O2 as a positive control could induce noticeable staining (Choo et al., 2014). In Vivo Mouse Skin Damage In vivo SOD activity and MDA content are biomarkers of antioxidant defense, and their increased levels are indicative of oxidative stress injury. To further confirm the significant in vitro toxicity of mixed-phase TiO2 NPs in skin cells, the dorsal skin of Balb/c hairless mice were exposed to various pristine and mixed-phase TiO2 NPs for 14 days with 3 h of daily simulated sunlight exposure, and in vivo SOD activity and MDA content were analyzed. As shown in Figure 7A, compared with the control group, the topical treatment of TiO2 NPs followed by sunlight exposure causes significantly increased SOD activities in the mice skin, where mixed-phase TiO2 NPs induce higher SOD activities than pristine particles, and A4:R6 showed the highest activity, followed by A7:R3 and A3:R7. Figure 7B indicates the MDA contents of mice skin exposed to TiO2 NPs, can display the similar tendency as shown in SOD activity. Both in vivo SOD activity and MDA content assessments confirm the in vitro results that mixed-phase TiO2 NPs can cause more severe oxidative stress than pristine particles under sunlight exposure. Moreover, the oxidative stress-mediated inflammation usually can result in pathological changes in skin ultrastructures (Pal et al., 2016). The epidermal thickness of mouse skin was evaluated in skin tissues after H&E staining. Figure 7C shows the mixed-phase TiO2 NPs can more significantly increase the thickness of mouse skin than anatase (the epidermal thickness was about 24.3 μm) and rutile (32.6 μm) NPs, and the most pronounced effect is found on A4:R6. The epidermal thickness for A7:R3, A4:R6, and A3:R7 were about 44.6, 53.8, and 35.7 μm. Obviously, the more severe in vivo skin damage is induced by mixed-phase TiO2 NPs compared with pristine particles, which intensely correlates to the abiotic O2•− and HO• radicals generation and can be predicted by BB degrees. Figure 7. View largeDownload slide In vivo SOD and MDA level in the BALB/C hairless mice skin tissue and the histopathological evaluation after dermal exposure to TiO2 NPs. (A) SOD activity, (B) MDA level, and (C) TiO2 exposure induced epidermal thickness in mouse skin. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 7. View largeDownload slide In vivo SOD and MDA level in the BALB/C hairless mice skin tissue and the histopathological evaluation after dermal exposure to TiO2 NPs. (A) SOD activity, (B) MDA level, and (C) TiO2 exposure induced epidermal thickness in mouse skin. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. DISCUSSION Pristine TiO2 NPs (anatase and rutile) are usually considered as safe nanomaterials and have been widely used as ingredients in cosmetics and cotton textiles (Barker and Branch, 2008; Contado and Pagnoni, 2008; Doganli et al., 2016; Meilert et al., 2005). Although TiO2 NPs have also been well known capable of causing cell injury under UV light exposure (George et al., 2014; Shakeel et al., 2016; Uchino et al., 2002), the low skin penetration ability of TiO2 NPs is considered beneficial for reducing their adverse effects to human skin (Sadrieh et al., 2010; Wu et al., 2009). However, most these skin penetration studies do not consider the influence of sunlight exposure. Actually, the sunlight exposure can significantly deepen the penetration of NPs into skin (Bennett et al., 2012), meaning the NPs probably are more dangerous under sunlight exposure. Thus, it is necessary to investigate the toxicity of TiO2 NPs on skin under sunlight exposure not only because of the photocatalytic properties of TiO2 NPs but also because of the sunlight-assisted skin penetration. Anatase/rutile mixed-phase TiO2 NPs recently were found in cosmetics and cotton textiles (Doganli et al., 2016; Meilert et al., 2005), and exhibited more severe toxicity than pristine TiO2 NPs (Wu et al., 2009; Yin et al., 2012). Although the stronger toxicity of mixed-phase TiO2 NPs can be easily ascribed to their active electronic property arising from the heterogeneous interface, there is a lack of a clear property-activity relationship to understand the toxicity mechanism and predict the toxicity potential based on the semiconductor property of mixed-phase TiO2 NPs. Moreover, since anatase/rutile mixed-phase TiO2 NPs can have different anatase/rutile ratios based on distinctive preparation process, there also lacks a feasible strategy to evaluate the toxicity difference based on their physicochemical property. In the present study, a series of mixed-phase TiO2 NPs with different anatase/rutile ratios were prepared to evaluate their oxidative injury on HaCaT cells and mouse skin under sunlight exposure, unveiling the potential property-toxicity relationship. Since the different anatase/rutile ratios could lead to different densities of anatase/rutile heterojunction, the BB concept was introduced herein to reflect the density of heterojunction, which is beneficial for quantitatively assessing the biological effect based on the semiconductor property. Actually, mixed-phase NPs were proved to be able to promote electron transfer from rutile to anatase based on their energy structures and produced more significant O2•− and HO• radicals than pristine NPs (Figs. 4C and 4D) under sunlight exposure, leading to more severe oxidative injury on HaCaT cells (Figure 5A) and mouse skin (Figure 7). The BB levels of A7:R3, A4:R6, and A3:R7 mixed-phase TiO2 NPs were found to be proportional to their abiotic radical amounts as well as in vitro and in vivo toxicological response levels and the extent of damage. Previous studies have successfully made attempts to correlate the electronic property of semiconductor nanomaterials with their toxicity (Burello and Worth, 2011; Zhang et al., 2012). This study further corroborates that the electronic property of semiconductor nanomaterials can be used to evaluate their toxicity potential. Under sunlight exposure, mixed-phase NPs was proved to be able to generate a large amount of O2•− and HO• radicals. As typical ROS, O2•− and HO• radicals can react with a range of biomolecules and exert potent adverse effects in damage of cell organelles and disruption of cellular redox homeostasis, leading to direct and indirect cellular ROS production (Chang et al., 2017; Liu et al., 2016; Meng et al., 2009). To overcome the ROS production, cells usually trigger either a defensive or an antioxidative response eliciting a chain of adverse biological responses (Manke et al., 2013). Cells exposed to mixed-phase NPs probably can induce similar biological responses, which is helpful for clarification of the molecular mechanism. Typical intracellular ROS production (Figure 5B and Supplementary Figure 7B) and GSH depletion (Figure 5C and Supplementary Figure 7C) of HaCaT cells were induced by mixed-phase TiO2 NPs, suggesting the high possibility of activation of oxidative stress signaling pathway. Further hierarchical oxidative stress response assessments corroborated mixed-phase TiO2 NPs could elicit HO-1 phase II enzyme expression (tier 1), IL-8 proinflammatory CXC chemokine release (tier 2), and mitochondrial dysfunction (tier 3) including mitochondrial membrane depolarization and superoxide generation (Figure 6C and Supplementary Figure 8). More importantly, all mixed-phase TiO2 NPs-induced in vitro toxic response levels as well as in vivo mouse skin damage extents (Figure 7) were proportional to O2•− and HO• amounts generated in mixed-phase TiO2 NPs (A4:R6, A7:R3, and A3:R7). Obviously, the in vitro and in vivo toxicity induced by mixed-phase TiO2 NPs under sunlight exposure are based on activation of oxidative stress. Oxidative stress can induce multiple types of cell death, such as apoptosis and senescence. Significant G2/M arrest associated with generation of apoptotic cells induced by mixed-phase TiO2 NPs confirms the cell death is possibly ascribed to apoptosis. G2/M phase can ensure the accuracy of allocation during mitosis. When some damaged DNA enters into the next phase of the cell cycle from 1 phase before being repaired, this damage can be “fixed down” to result in genetic instability of genome and potential likelihood of death. These results are consistent with the report that severe oxidative stress of TiO2 NPs is capable of inducing cell apoptosis (Saira et al., 2016; Wang et al., 2015). It has been reported that cellular senescence is usually associated with G1 arrest (Tao et al., 2017; Yi et al., 2013). Integrated with the absence of senescence–associated SA-β-gal activity, it can be concluded the type of cell death induced by mixed-phase TiO2 NPs is not correlated with cellular senescence. In summary, under sunlight exposure, the energy band alignment in mixed-phased TiO2 NPs can drive the electron transfer between anatase and rutile, resulting in the electron-hole separation at the interface of heterojunction. The formed free electron and hole can react with oxygen and water to produce O2•− and HO• radicals, respectively, resulting in HaCaT cell death and mouse skin damage. The underlying mechanism is involved in activation of oxidative stress as manifested by hierarchical oxidative stress toxicological responses. Without sunlight exposure, mixed-phased and pristine TiO2 NPs cannot induce notable toxicity in HaCaT cells and mouse skin. The density of heterojunction of mixed-phased TiO2 NPs is theoretically proportional to the amount of produced free electrons and holes at the interface of heterojunction, which can be experimentally reflected by the degree of band banding. Thus, the degree of BB in mixed-phased TiO2 NPs can be used to assess their ROS generation, ultimately predicting their induced toxicity potential in HaCaT cells and mouse skin. SUPPLEMENTARY DATA Supplementary data are available at Toxicological Sciences online. FUNDING This work was supported by the National Natural Science Foundation of China (Grant Nos. 21573216, 21501170, 21777152 and 21703232) and Hundred Talent Program of Chinese Academy of Sciences, Science and Technology Development Project Foundation of Jilin Province (20160101304JC, 20160520134JH, and 20180520145JH). REFERENCES Barker P. J. , Branch A. ( 2008 ). The interaction of modern sunscreen formulations with surface coatings . Prog. Org. Coat . 62 , 313 – 320 . Google Scholar CrossRef Search ADS Bennett S. W. , Zhou D. , Mielke R. , Keller A. A. ( 2012 ). Photoinduced disaggregation of TiO2 nanoparticles enables transdermal penetration . PLoS One 7 , e48719 – e48725 . Google Scholar CrossRef Search ADS PubMed Brezová V. , Barbieriková Z. , Zukalová M. , Dvoranová D. , Kavan L. ( 2014 ). EPR study of 17O-enriched titania nanopowders under UV irradiation . Catal. Today 230 , 112 – 118 . Google Scholar CrossRef Search ADS Brezova V. , Gabcova S. , Dvoranova D. , Stasko A. ( 2005 ). Reactive oxygen species produced upon photoexcitation of sunscreens containing titanium dioxide (an EPR study) . J. Photoch. Photobio. B 79 , 121 – 134 . Google Scholar CrossRef Search ADS Burello E. , Worth A. P. ( 2011 ). A theoretical framework for predicting the oxidative stress potential of oxide nanoparticles . Nanotoxicology 5 , 228 – 235 . Google Scholar CrossRef Search ADS PubMed Contado C. , Pagnoni A. ( 2008 ). TiO2 in commercial sunscreen lotion: Flow field-flow fractionation and ICP-AES together for size analysis . Anal. Chem . 80 , 7594 – 7608 . Google Scholar CrossRef Search ADS PubMed Chang Y. , Li K. , Feng Y. , Cheng Y. , Zhang M. , Wang Z. , Wu Z. , Zhang H. ( 2017 ). Achievement of safer palladium nanocrystals by enlargement of {100} crystallographic facets . Nanotoxicology 11 , 907 – 922 . Google Scholar CrossRef Search ADS PubMed Choo K. B. , Tai L. , Hymavathee K. S. , Wong C. Y. , Nguyen P. N. N. , Huang C.-J. , Cheong S. K. , Kamarul T. ( 2014 ). Oxidative stress-induced premature senescence in Wharton’s jelly-derived mesenchymal stem cells . Int. J. Med. Sci . 11 , 1201 – 1207 . Google Scholar CrossRef Search ADS PubMed Comfort K. K. , Braydich-Stolle L. K. , Maurer E. I. , Hussain S. M. ( 2014 ). Less is more: Long-term in vitro exposure to low levels of silver nanoparticles provides new insights for nanomaterial evaluation . ACS Nano 8 , 3260 – 3271 . Google Scholar CrossRef Search ADS PubMed Deák P. , Kullgren J. , Aradi B. , Frauenheim T. , Kavan L. ( 2016 ). Water splitting and the band edge positions of TiO2 . Electrochim. Acta 199 , 27 – 34 . Google Scholar CrossRef Search ADS Doganli G. , Yuzer B. , Aydin I. , Gultekin T. , Con A. H. , Selcuk H. , Palamutcu S. ( 2016 ). Functionalization of cotton fabric with nanosized TiO2 coating for self-cleaning and antibacterial property enhancement . J. Coat. Technol. Res . 13 , 257 – 265 . Google Scholar CrossRef Search ADS Dvoranova D. , Barbierikova Z. , Brezova V. ( 2014 ). Radical intermediates in photoinduced reactions on TiO2 (an EPR spin trapping study) . Molecules 19 , 17279 – 17304 . Google Scholar CrossRef Search ADS PubMed Edmonds M. T. , Pakes C. I. , Mammadov S. , Zhang W. , Tadich A. , Ristein J. , Ley L. ( 2011 ). Surface band bending and electron affinity as a function of hole accumulation density in surface conducting diamond . Appl. Phys. Lett . 98 , 102101 – 102103 . Google Scholar CrossRef Search ADS Fang D. , Huang K. , Liu S. , Huang J. ( 2008 ). Fabrication and photoluminiscent properties of titanium oxide nanotube arrays . J. Braz. Chem. Soc . 19 , 1059 – 1064 . Google Scholar CrossRef Search ADS Frank O. , Zukalova M. , Laskova B. , Kürti J. , Koltai J. , Kavan L. ( 2012 ). Raman spectra of titanium dioxide (anatase, rutile) with identified oxygen isotopes (16, 17, 18) . Phy. Chem. Chem. Phys . 14 , 14567 – 14572 . Google Scholar CrossRef Search ADS Fulda S. , Gorman A. M. , Hori O. , Samali A. ( 2010 ). Cellular stress responses: Cell Survival and Cell Death . Int. J. Cell Biol . 2010 , 1. George S. , Gardner H. , Seng E. K. , Chang H. , Wang C. , Yu Fang C. H. , Richards M. , Valiyaveettil S. , Chan W. K. ( 2014 ). Differential effect of solar light in increasing the toxicity of silver and titanium dioxide nanoparticles to a fish cell line and zebrafish embryos . Environ. Sci. Technol . 48 , 6374 – 6382 . Google Scholar CrossRef Search ADS PubMed Gerloff K. , Fenoglio I. , Carella E. , Kolling J. , Albrecht C. , Boots A. W. , Förster I. , Schins R. P. F. ( 2012 ). Distinctive toxicity of TiO2 rutile/anatase mixed phase nanoparticles on Caco-2 Cells . Chem. Res. Toxicol . 25 , 646 – 655 . Google Scholar CrossRef Search ADS PubMed Gunter O. , Vicki S. , Ken D. ( 2007 ). Toxicology of nanoparticles: A historical perspective . Nanotoxicology 1 , 2 – 25 . Google Scholar CrossRef Search ADS Haines D. D. , Juhasz B. , Tosaki A. ( 2013 ). Management of multicellular senescence and oxidative stress . J. Cell. Mol. Med . 17 , 936 – 957 . Google Scholar CrossRef Search ADS PubMed Hamzeh M. , Sunahara G. I. ( 2013 ). In vitro cytotoxicity and genotoxicity studies of titanium tioxide (TiO2) nanoparticles in chinese hamster lung fibroblast cells . Toxicol. In Vitro 27 , 864 – 873 . Google Scholar CrossRef Search ADS PubMed He W. , Liu Y. , Wamer W. G. , Yin J. J. ( 2014 ). Electron spin resonance spectroscopy for the study of nanomaterial-mediated generation of reactive oxygen species . J. Food Drug Anal . 22 , 49 – 63 . Google Scholar CrossRef Search ADS PubMed Kim C. , Kang D. , Lee E. K. , Lee J.-S. ( 2017 ). Long noncoding RNAs and RNA-binding proteins in oxidative stress, cellular senescence, and age-related diseases . Oxid. Med. Cell. Longev . 2017 , 1. Kullgren J. , Aradi B. , Frauenheim T. , Kavan L. , Deák P. ( 2015 ). Resolving the controversy about the band alignment between rutile and anatase: The role of OH–/H+adsorption . J. Phys. Chem. C 119 , 21952 – 21958 . Google Scholar CrossRef Search ADS Li Y. , Zhang W. , Niu J. , Chen Y. ( 2012 ). Mechanism of photogenerated reactive oxygen species and correlation with the antibacterial properties of engineered metal-oxide nanoparticles . ACS Nano 6 , 5164 – 5173 . Google Scholar CrossRef Search ADS PubMed Liu N. , Li K. , Li X. , Chang Y. , Feng Y. , Sun X. , Cheng Y. , Wu Z. , Zhang H. ( 2016 ). Crystallographic facet-induced toxicological responses by faceted titanium dioxide nanocrystals . ACS Nano 10 , 6062 – 6073 . Google Scholar CrossRef Search ADS PubMed Luo Z. , Poyraz A. S. , Kuo C.-H. , Miao R. , Meng Y. , Chen S.-Y. , Jiang T. , Wenos C. , Suib S. L. ( 2015 ). Crystalline mixed phase (anatase/rutile) mesoporous titanium dioxides for visible light photocatalytic activity . Chem. Mater . 27 , 6 – 17 . Google Scholar CrossRef Search ADS Manke A. , Wang L. , Rojanasakul Y. ( 2013 ). Mechanisms of nanoparticle-induced oxidative stress and toxicity . Biomed. Res. Int . 2013 , 942916 – 942931 . Google Scholar CrossRef Search ADS PubMed Marazita M. C. , Dugour A. , Marquioni-Ramella M. D. , Figueroa J. M. , Suburo A. M. ( 2016 ). Oxidative stress-induced premature senescence dysregulates VEGF and CFH expression in retinal pigment epithelial cells: Implications for Age-related Macular Degeneration . Redox Biol. 7 , 78 – 87 . Google Scholar CrossRef Search ADS PubMed Meilert K. T. , Laub D. , Kiwi J. ( 2005 ). Photocatalytic self-cleaning of modified cotton textiles by TiO2 clusters attached by chemical spacers . J. Mol. Catal. A Chem . 237 , 101 – 108 . Google Scholar CrossRef Search ADS Meng H. , Xia T. , George S. , Nel A. E. ( 2009 ). A predictive toxicological paradigm for the safety assessment of nanomaterials . ACS Nano 3 , 1620 – 1627 . Google Scholar CrossRef Search ADS PubMed Nel A. , Xia T. , Mädler L. , Li N. ( 2006 ). Toxic potential of materials at the nanolevel . Science 311 , 622 – 627 . Google Scholar CrossRef Search ADS PubMed Ozben T. ( 2007 ). Oxidative stress and apoptosis: Impact on cancer therapy . J. Pharm. Sci . 96 , 2181 – 2196 . Google Scholar CrossRef Search ADS PubMed Pal A. , Alam S. , Chauhan L. K. S. , Saxena P. N. , Kumar M. , Ansari G. N. , Singh D. , Ansari K. M. ( 2016 ). UVB exposure enhanced the dermal penetration of zinc oxide nanoparticles and induced inflammatory responses through oxidative stress mediated by MAPKs and NF-κB signaling in SKH-1 hairless mouse skin . Toxicol. Res . 5 , 1066 – 1077 . Google Scholar CrossRef Search ADS Ruan P. , Qian J. , Xu Y. , Xie H. , Shao C. , Zhou X. ( 2013 ). Mixed-phase TiO2 nanorods assembled microsphere: Crystal phase control and photovoltaic application . CrystEngComm. 15 , 5093 – 5099 . Google Scholar CrossRef Search ADS Sadrieh N. , Wokovich A. M. , Gopee N. V. , Zheng J. , Haines D. , Parmiter D. , Siitonen P. H. , Cozart C. R. , Patri A. K. , McNeil S. E. et al. , . ( 2010 ). Lack of significant dermal penetration of titanium dioxide from sunscreen formulations containing nano- and submicron-size TiO2 particles . Toxicol. Sci . 115 , 156 – 166 . Google Scholar CrossRef Search ADS PubMed Saira F. , Mackey M. A. , Qureshi R. , Mahmoud M. A. ( 2016 ). In vitro investigations of gold nanocages: Toxicological profile in human keratinocyte cell line . Food Chem. Toxicol . 97 , 89 – 95 . Google Scholar CrossRef Search ADS PubMed Scanlon D. O. , Dunnill C. W. , Buckeridge J. , Shevlin S. A. , Logsdail A. J. , Woodley S. M. , Catlow C. R. A. , Powell M. J. , Palgrave R. G. , Parkin I. P. et al. , . ( 2013 ). Band alignment of rutile and anatase TiO2 . Nat. Mater . 12 , 7 , 798 – 801 . Google Scholar CrossRef Search ADS PubMed Seaton A. , Tran L. , Aitken R. , Donaldson K. ( 2010 ). Nanoparticles, human health hazard and regulation . J. R. Soc. Inter . 7 , S119 – S129 . Google Scholar CrossRef Search ADS Serpone N. , Dondi D. , Albini A. ( 2007 ). Inorganic and organic UV filters: Their role and efficacy in sunscreens and suncare product . Inorg. Chim. Acta 360 , 794 – 802 . Google Scholar CrossRef Search ADS Shakeel M. , Jabeen F. , Shabbir S. , Asghar M. S. , Khan M. S. , Chaudhry A. S. ( 2016 ). Toxicity of nano-titanium dioxide (TiO2-NP) through various routes of exposure: A review . Biol. Trace Elem. Res . 172 , 1 – 36 . Google Scholar CrossRef Search ADS PubMed Sun X. , Dai W. , Wu G. , Li L. , Guan N. , Hunger M. ( 2015 ). Evidence of rutile-to-anatase photo-induced electron transfer in mixed-phase TiO2 by solid-state NMR spectroscopy . Chem. Commun . 51 , 13779 – 13782 . Google Scholar CrossRef Search ADS Tao Y.-F. , Wang N.-N. , Xu L.-X. , Li Z.-H. , Li X.-L. , Xu Y.-Y. , Fang F. , Li M. , Qian G.-H. , Li Y.-H. et al. , . ( 2017 ). Molecular mechanism of G1 arrest and cellular senescence induced by LEE011, a novel CDK4/CDK6 inhibitor, in leukemia cells . Cancer Cell Int . 17 , 35. Google Scholar CrossRef Search ADS PubMed Tiwari A. , Mondal I. , Ghosh S. , Chattopadhyay N. , Pal U. ( 2016 ). Fabrication of mixed phase TiO2 heterojunction nanorods and their enhanced photoactivities . Phy. Chem. Chem. Phys . 18 , 15260 – 15268 . Google Scholar CrossRef Search ADS Uchino T. , Tokunaga H. , Ando M. , Utsumi H. ( 2002 ). Quantitative determination of OH radical generation and its cytotoxicity induced by TiO2-UVA treatment . Toxicol. In Vitro 16 , 629 – 635 . Google Scholar CrossRef Search ADS PubMed Wang Y. , Cui H. , Zhou J. , Li F. , Wang J. , Chen M. , Liu Q. ( 2015 ). Cytotoxicity, DNA damage, and apoptosis induced by titanium dioxide nanoparticles in human non-small cell lung cancer A549 cells . Environ. Sci. Pollut. Res. 22 , 5519 – 5530 . Google Scholar CrossRef Search ADS Wu J. , Liu W. , Xue C. , Zhou S. , Lan F. , Bi L. , Xu H. , Yang X. , Zeng F.-D. ( 2009 ). Toxicity and penetration of TiO2 nanoparticles in hairless mice and porcine skin after subchronic dermal exposure . Toxicol. Lett . 191 , 1 – 8 . Google Scholar CrossRef Search ADS PubMed Xue C. , Liu W. , Wu J. , Yang X. , Xu H. ( 2011 ). Chemoprotective effect of N-acetylcysteine (NAC) on cellular oxidative damages and apoptosis induced by nano titanium dioxide under UVA irradiation . Toxicol. In Vitro 25 , 110 – 116 . Google Scholar CrossRef Search ADS PubMed Yi G. , He Z. M. , Zhou X. K. , Xian L.W. , Yuan T. , Jia X. , Hong J. , He L , Liu J. ( 2013 ). Low concentration of metformin induces a p53-dependent senescence in hepatoma cells via activation of the AMPK pathway . Int. J. Oncol . 43 , 1503 – 1510 . Google Scholar CrossRef Search ADS PubMed Yin J.-J. , Liu J. , Ehrenshaft M. , Roberts J. E. , Fu P. P. , Mason R. P. , Zhao B. ( 2012 ). Phototoxicity of nano titanium dioxides in HaCaT keratinocytes-generation of reactive oxygen species and cell damage . Toxicol. Appl. Pharmacol . 263 , 81 – 88 . Google Scholar CrossRef Search ADS PubMed Zhang H. , Ji Z. , Xia T. , Meng H. , Low-Kam C. , Liu R. , Pokhrel S. , Lin S. , Wang X. , Liao Y.-P. et al. , . ( 2012 ). Use of metal oxide nanoparticle band gap to develop a predictive paradigm for oxidative stress and acute pulmonary inflammation . ACS Nano 6 , 4349 – 4368 . Google Scholar CrossRef Search ADS PubMed Zhang H. , Pokhrel S. , Ji Z. , Meng H. , Wang X. , Lin S. , Chang C. H. , Li L. , Li R. , Sun B. et al. , . ( 2014 ). PdO doping tunes band-gap energy levels as well as oxidative stress responses to a Co(3)O(4) P-type semiconductor in cells and the lung . J. Am. Chem. Soc . 136 , 6406 – 6420 . Google Scholar CrossRef Search ADS PubMed Zhang Z. , Yates J. T. Jr. ( 2012 ). Band bending in semiconductors: Chemical and physical consequences at surfaces and interfaces . Chem. Rev . 112 , 5520 – 5551 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices) http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Toxicological Sciences Oxford University Press

Band Alignment-Driven Oxidative Injury to the Skin by Anatase/Rutile Mixed-Phase Titanium Dioxide Nanoparticles Under Sunlight Exposure

Loading next page...
 
/lp/ou_press/band-alignment-driven-oxidative-injury-to-the-skin-by-anatase-rutile-QvahZi3MwK
Publisher
Oxford University Press
Copyright
© The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com
ISSN
1096-6080
eISSN
1096-0929
D.O.I.
10.1093/toxsci/kfy088
Publisher site
See Article on Publisher Site

Abstract

Abstract Anatase/rutile mixed-phase titanium dioxide (TiO2) nanoparticles (NPs) have been found in cosmetics and cotton textiles. Once exposed to sunlight, mixed-phase TiO2 NPs are even more toxic to cells than pure phase NPs; however, the underlying mechanism remains unclear. Considering the unique anatase/rutile heterojunction structure existing in mixed-phase NPs, the potent toxicity of mixed-phase TiO2 NPs probably originates from the high reactive oxygen species (ROS) production because the anatase/rutile heterojunction is constituted by the staggered energy bands that facilitate the electron-hole separation at the interface due to the band alignment. In the present study, a library of mixed-phase TiO2 NPs with different anatase/rutile ratios was established to investigate the potential property-activity relationship and further clarify the underlying molecular mechanism. Under sunlight exposure, these mixed-phase TiO2 NPs could produce significant abiotic ROS and induce hierarchical oxidative stress to HaCaT skin cells and mice skin. The ROS magnitude and toxicity potential of these NPs were found to be proportional to their energy band bending (BB) levels. This means that the toxicity of mixed-phase TiO2 NPs can be correlated to their heterojunction density, and the toxicity potential of mixed-phase TiO2 NPs can be weighed by their BB levels. titanium oxide, nanoparticles, property-activity relationship, oxidative injury, sunlight exposure The rapidly expanding use of nanomaterials in various aspects of human life has dramatically increased human exposure to nanomaterials. Concerns on the potential adverse human health effects of nanomaterials are escalating (Gunter et al., 2007; Seaton et al., 2010). Titanium dioxide (TiO2) nanoparticles (NPs), such as anatase and rutile, have been widely used as ingredients in cosmetics and cotton textiles (Barker and Branch, 2008; Contado and Pagnoni, 2008; Doganli et al., 2016; Meilert et al., 2005). These applications potentially increase the risk of human skin exposure to TiO2 NPs, especially under outdoor sunlight irradiation. Ultraviolet (UV) or sunlight exposure on one hand can remarkably deepen the penetration of NPs into skin (Bennett et al., 2012) and on the other hand can facilitate reactive oxygen species (ROS) generation through exciting electrons into conduction band to form superoxide (O2•−) and generating holes in valence band to form hydroxyl radicals (HO•), which has led to serious concerns on the risk of TiO2 NPs to human skin health (Brezova et al., 2005; Serpone et al., 2007). Unfortunately, more severe toxicity is caused by mixed-phase TiO2 NPs compared with pristine NPs (Gerloff et al., 2012; Wu et al., 2009; Yin et al., 2012). In vitro study indicates mixed-phase TiO2 NPs can induce the enhanced Lactate Dehydrogenase (LDH) leakage and oxidative stress responses in HaCaT skin cells (Yin et al., 2012), whereas in vivo study reveals a large reduction in the collagen content of mouse skin when exposed to mixed-phase TiO2 NPs (Wu et al., 2009). Obviously, mixed-phase TiO2 NPs are more dangerous than pristine NPs, however, the underlying mechanism remains unclear. Mixed-phase TiO2 NPs usually contain high density of anatase/rutile heterojunctions that are constructed by staggered band edges arising from the different band gaps of anatase and rutile (Tiwari et al., 2016). Previous studies have demonstrated that the band alignment at the interface of heterojunction can drive electrons flow from rutile to anatase with holes moving in the opposite directions (Deák et al., 2016; Kullgren et al., 2015; Scanlon et al., 2013; Sun et al., 2015), leading to an efficient electron-hole separation. The free electron and hole can react with oxygen and water molecules to form O2•− and HO• radicals, respectively, potentially causing potent oxidative stress-mediated cellular injury (Zhang et al., 2014). This means there probably existing a property-activity relationship between heterojunction and toxicity of mixed-phase TiO2 NPs, which can be used to deeply understand the toxicity mechanism of mixed-phase TiO2 NPs and further predict their toxicity potential. However, it is currently hard to establish this relationship because there is a lack of a feasible way to determine the density of heterojunction. So, it is necessary to develop a method to reflect the density of heterojunction. Since the heterojunction-assisted electron-hole separation can lead to charge accumulation at the interface of heterojunction, a band bending (BB) effect usually occurs (Zhang et al., 2014). The degree of BB is proportional to the amount of accumulated charges and further proportional to the density of heterojunction (Edmonds et al., 2011). Thus, determination of the BB level becomes a feasible approach to reflect the density of heterojunction and further predict the toxicity potential of mixed-phase TiO2 NPs. In this study, a series of anatase/rutile mixed-phase TiO2 NPs with different anatase/rutile ratios were prepared to display different densities of heterojunction, which could be used to investigate the potential property-activity relationship of mixed-phase NP sin keratinocyte HaCaT skin cells and mouse skin under exposure to simulated sunlight. It was hypothesized that the staggered energy bands at the interface of anatase/rutile heterojuctions could trigger the electron-hole separation and the BB until Fermi energies of anatase and rutile were aligned (Figure 1A), and the formed free charges could produce ROS and cause oxidative injury in cells and mice skin, where the density of heterojunction and the extent of injury could be weighed by the BB level (Figure 1B). Energy structures and the BB levels of these NPs were systemically examined to correlate with ROS production and induced oxidative stress toxicological responses in cells and mice. Actually, it was found that under sunlight irradiation, TiO2 NPs with various anatase/rutile ratios could generate different magnitudes of O2•− and HO• radicals, and cause different levels of heme oxygenase-1 (HO-1) expression, interleukin 8 (IL-8) release, and mitochondrial dysfunction in HaCaT skin cells, and ultimately leading to cell death. Further in vivo evaluation of the activity of superoxide dismutase (SOD) and the content of malondialdehyde (MDA) in mouse skin also supported this relationship between the injury and the BB. Thus, the biological effects of mixed-phase TiO2 NPs can be well elucidated by their semiconductor properties. Figure 1. View largeDownload slide Potential heterojunction-toxicity relationship of mixed-phase TiO2 NPs. (A) Band alignment-driven electron-hole separation, BB induction, and ROS production at the interface of heterojunction. (B) Density variation of heterojunction during the transformation of anatase to rutile phase. Figure 1. View largeDownload slide Potential heterojunction-toxicity relationship of mixed-phase TiO2 NPs. (A) Band alignment-driven electron-hole separation, BB induction, and ROS production at the interface of heterojunction. (B) Density variation of heterojunction during the transformation of anatase to rutile phase. MATERIALS AND METHODS Chemicals Titanium (IV) isopropoxide (≥97%), 1-butanol (anhydrous, 99.8%), and poly-(ethyleneglycol)-block-poly(propyleneglycol)-block-poly(ethyleneglycol)PEO20-PPO70-PEO20 (Pluronic P123) were purchased from Sigma-Aldrich. All chemicals were reagent grade and used without further purification or modification unless otherwise indicated. Water throughout all experimental procedures was obtained by using a Milli-Q water purification system. Synthesis of pristine and mixed-phase (anatase/rutile) TiO2 NPs Mixed-phase TiO2 NPs were synthesized through inverse micelle template sol-gel method under solvothermal conditions and using titanium isopropoxide as a precursor (Luo et al., 2015). In a typical synthesis, 0.01 mol (2.84 g) titanium isopropoxide was dissolved in a solution containing 0.094 mol (7 g) of 1-butanol, 0.016 mol (1 g) of concentrated HNO3 and 1.72 × 10−4 mol of P123 surfactant in a 150-ml beaker at room temperature (RT) under magnetic stirring. The obtained clear gel was dark yellow in appearance and placed in an oven running at 120°C for 4 h. The pristine and mixed-phase TiO2 NPs were obtained by calcining the dark orange and transparent (rigid) films under air conditions for 4 h at various temperatures (600°C–1000°C). After the temperature dropped to RT, the products were taken out and washed with water and ethanol alternately for several times, and dried in an oven at 80°C overnight. Physicochemical characterization of pristine and mixed-phase TiO2 NPs X-ray diffraction (XRD) patterns were collected using D8 ADVANCE (BRUKER company, Germany) (Cu Kα radiation) over a range of 15°−80° 2θ. Transmission electron microscopy (TEM) and high resolution TEM (HRTEM) were performed using a JEOL 2100 F microscope operated at 200 keV. Specific surface area was determined by the BET method from N2 sorption isotherms, and acquired using a Micrometrics ASAP 2010 sorption instrument. Zeta potential and dynamic light scattering (DLS) data were collected on a Malvern Nanosizer ZS. Micro-Raman spectra were recorded in backscattering geometry using LabRam HR 800 (HORIBA Jobin Yvon) coupled with an Olympus BX 41microscope; the confocal hole and the silt width were both fixed at 200 μm, and the excitation wavelength of 633 nm was used. Ultraviolet photoelectron spectroscopy (UPS) measurements were performed on a PREVAC XPS/UPS system spectrometer equipped with monochromatic Al Kα radiation (hν = 1486.7 eV) as X-ray source and a He I radiator (hν = 21.2 eV) as UV source. All spectra were collected at RT in an ultra-high-vacuum environment (base pressure of analysis chamber < 5× 10−8 Torr). All UPS-binding energies are referenced to the Fermi edge of a clean Au foil. For UPS measurement, indium tin oxide (ITO) slide was first washed by water, acetone, and ethanol, respectively for several times in order to remove the contamination. The TiO2 NPs dispersed with deionized water at a concentration of 2.5 mg/ml were dropped onto the ITO surface and left to dry. The valence band maximum (VBM) was determined by linear extrapolation of the leading edge of the UPS spectrum, and the work function was determined from one-half the height of the secondary electron onset. Band-gap energies were obtained from diffuse reflectance (DR) UV-vis spectroscopic analysis (PerkinElmer UV-VIS Spectrometer Lambda 35) by using BaSO4 as a reference material. All measurements were conducted in ambient air using a bandwidth of 1.0 nm. Collected DR UV-vis spectra were converted into Kubelka−Munk function (F[R∞]) spectra. Sunlight exposure experiments were conducted on Sun 2000 Solar Simulator (Abet Technologies). Confirmation of the percentage of rutile phase in the mixed-phase TiO2 NPs by XRD and Raman spectra The percentage of rutile phase in the mixed-phase of TiO2 NPs was determined based on XRD analysis through the equation: WR = IR/(0.884IA+ IR), where WR is the weight percentage of rutile phase, IR and IA are the integrated intensity of rutile (110) and anatase (101) peaks. The percentage of rutile phase can be further confirmed by Raman spectra analysis. The Raman spectra ranging from 300 to 700 cm−1 was deconvoluted into anatase phase (containing peaks at 395, 518, and 639 cm−1) and rutile phase (containing peaks at 448 and 609 cm−1). The percentage of rutile phase was calculated by A448+609 cm−1/(A395+518+639 cm−1+A448+609 cm−1), where A448+609 cm−1is the integrated peak area of rutile phase at 448 and 609 cm−1 and A395+518+639 cm−1 is the integrated peak area of anatase phase at 395, 518, and 639 cm−1 (Frank et al., 2012). Abiotic total ROS measurement of pristine and mixed-phase TiO2 NPs The total ROS generation of TiO2 NPs was determined by 2,7-dichlorodihydrofluorescein diacetate (H2DCFDA) fluorescence (Sigma). According to the previous method (Zhang et al., 2014), the 2′,7′-dichlorofluorescein (DCF) final working concentration was 29 μmol/l. This could be obtained by the following process: 50 μg of H2DCFDA, 17.3 μl of ethanol and 692 μl of a 0.01 mol/l sodium hydroxide solution were mixed well. The resulting solution was incubated at 4°C for 30 min under dark condition, and 3500 μl of a sodium phosphate buffer (25 mmol/l, pH =7.4) was added to obtain 29 μmol/l DCF working solution. To each well of a 96-multiwell black plate (Costar, Corning, New York), we added 80 μl of 29 μmol/l DCF working solution. In total 20 μl of 1 mg/ml NP aqueous suspension was subsequently added to each well (final concentration is 200 μg/ml), followed by 15 min of exposure to simulated sunlight (Sun 2000 Solar Simulator, Abet Technologies) and another 2 h of incubation at RT under dark condition. DCF fluorescence emission spectra in the range of 500–600 nm were collected using a SpectraMax M5 microplate reader with an excitation wavelength of 490 nm. In comparison, the DCF fluorescence enhancement by NPs without sunlight exposure was also carried out according to the similar procedure except exposure to the simulated sunlight. Superoxide and hydroxide radical measurements by electron spin resonance Superoxide radicals (O2•−) and hydroxide radicals (HO•) were detected by electron spin resonance (ESR) technology. ESR signals of spin-trapped paramagnetic species with 5, 5-dimethyl-1-pyrroline N-oxide (DMPO) were recorded with a Brucker A300 spectrometer. For O2•− detection, 15 μl of 1.0 mol/l DMPO was mixed well with 50 μl of 50 μg/ml TiO2 suspension in methanol, while for HO• detection, 10 μl of 1.0 mol/l DMPO was mixed with 50 μl of 50 μg/ml TiO2 NPs aqueous suspension. The resulting mixture was irradiated by simulated sunlight for 15 min and then immediately subjected to the ESR measurement of O2•− and HO•, respectively. Controlled experiments without simulated sunlight were also carried out according to the similar procedure except the irradiation with simulated sunlight. Cell culture Human Keratinocytes cell line HaCaT (ATCC Number: PCS-200-011) were cultured in vented T-75 cm2 flasks (Corning, Fisher Scientific) in Dulbecco’s Modified Eagle’s Medium (DMEM) containing 10% fetal calf serum (GiBCO), 100 U/ml penicillin, 100 μg/ml streptomycin, and 2 mM L-glutamine at 37°C in a humidified atmosphere of 5% CO2. The culture medium was replaced at 70%–80% confluency every 2 days. Cellular uptake of pristine and mixed-phase TiO2 NPs determined by ICP-OES There were 8 × 104 HaCaT cells in 800 μl of DMEM were incubated in 12-well plates (Costar, Corning) for overnight growth. The medium was removed, and cells were incubated with 800 μl of 200 μg/ml TiO2 NP DMEM suspension for 6 h. After removal of culture medium, cells were washed with PBS 3 times and harvested through trypsinization, and digested by concentrated nitric acid for ICP-OES assessment. Cytotoxicity assessment by MTS assays Viability of HaCaT cells exposed to different TiO2 NPs with or without simulated sunlight exposure was determined by an MTS assay. 1 × 105 HaCaT cells in 100 μl of DMEM were seeded in each well of 96-well plates and incubated for 24 h. The medium was removed, and cells were treated with 100 μl of TiO2 NP suspensions at concentrations of 0.4–200 μg/ml for 6 h, followed by exposure to the simulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for 18 h. Then, the culture medium was removed and cells were washed with PBS 3 times. Each well received 100 μl of culture medium containing 16.7% of MTS stock solution, followed by 1 h of incubation at 37°C in a humidified 5% CO2 incubator. The plate was centrifuged at 2000 × g for 10 min in Xiangyi L535R with a microplate rotor to spin down the NPs and cell debris. In toal 80 μl of the supernatant was transferred into a new 96-multiwell plate. The absorbance of formazan was read at 490 nm on a SpectraMax M5 microplate reader (Molecular Devices Corp., Sunnyvale, California). In comparison, the viability of cells treated with TiO2 NPs but without sunlight exposure was also tested following the similar procedure. Cellular ROS measurement by flow cytometer There were 8 × 104 HaCaT cells in 800 μl of DMEM were plated in each well of 12-well plates (Costar, Corning) for overnight growth. After removal of the medium, and cells were incubated with 800 μl of 50 μg/ml different TiO2 NPs suspension in DMEM medium for 6 h, and then cells were exposed to simulated sunlight for 15 min at a power density of 0.1 W/cm2 and further cultured for 6 h. Cells only exposed to the sunlight were used as a control. Next, the culture medium was removed and cells were stained with 100 μl of 10 μmol/ml DCFH-DA for 30 min. After washing with PBS for 3 times, the harvested cells were transferred to polystyrene tubes for acquisition and analysis by flow cytometer (BD Accuri C6). In comparison, ROS generation of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. Cellular Glutathione (GSH) measurements based on a 5,5′dithio 2-nitrobenzoic acid (DNTB) method In total 8 × 104 HaCaT cells were plated in each well of 12-well plates (Costar, Corning) for overnight growth. After removal of the medium, and cells were incubated with 800 μl of 50 μg/ml different TiO2 NP suspension in DMEM medium for 6 h, and then cells were exposed to simulated sunlight for 15 min at a power density of 0.1 W/cm2 and further cultured for 6 h. Cells were washed with PBS 3 times and resuspended in cell lysis buffer (Beyotime). After centrifugation, 10 μl of lysate was mixed with 150 μl of 30 μg/ml DNTB in 1 of 96-well plates. The absorption of the resulting mixture at 412 nm was recorded on a SpectraMax M5 microplate spectrophotometer. In comparison, GSH levels of the cells treated with TiO2 NPs but without exposure to sunlight were also determined following the similar procedure. Western blot analysis for HO-1 expression There were 1.6 × 105 HaCaT cells in 1.6 ml of culture medium seeded into each well of 6-well plates (Costar, Corning). After overnight growth, cells were treated with 1.6 ml of 50 μg/ml NP suspension for 6 h, and then exposed to the stimulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for 6 h. Cells only exposed to the sunlight were used as a control. Cells were washed with PBS 3 times and harvested through scraping. Cell pellets were resuspended in cell lysis buffer containing protease inhibitors. The lysate was sonicated briefly and centrifuged, and the protein content in the supernatant was measured by the Bradford method. A 30 μg amount of total protein from each sample was electrophoresed by 10% SDS-PAGE and transferred to a PVDF membrane. After blocking, membranes were incubated with antihuman HO-1 or β-actin monoclonal antibody (1:1000) (Beyotime) for 12 h. After being washed with TBS/T buffer 3 times, membranes were incubated for 12 h with horseradish peroxidase conjugated secondary antibody (dilution 1:1000) (Beyotime, Nanjing). After being washed, membranes were developed using an enhanced Super Signal West Pico chemiluminescent substrate (Thermo Scientific, USA). The intensity of HO-1 was normalized to that of β-actin using Image J software. In comparison, HO-1 expression of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. IL-8 cytokine quantification by ELISA There were 1 × 105 HaCaT cells in 100 μl of DMEM plated in each well of a 96-multiwell black plate (Costar, Corning) for overnight growth. The medium was removed, and cells were treated with NP suspensions (50 μg/ml) in DMEM for 6 h, and exposed to the stimulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were cultured for another 6 h. Plates were centrifuged at 2000 × g for 10 min in Xiangyi L535R with a microplate rotor to spin down the cell debris and NPs. In total 50 μl of the supernatant from each well was used for measurement of IL-8 activity in HaCaT cells using an OptEIA (BD Biosciences, California) ELISA kit according to the manufacturer’s instructions. Briefly, a 96-well plate was coated with 100 μl of monoclonal antiIL-8 antibody overnight. After removal of the unbound antibody, a standard cytokine dilution series or 50 μl of each supernatant were pipetted into the precoated wells for antigen capture. After 2 h of incubation, the unbound growth factor was removed and each well was washed with a buffer 5 times and an enzyme-linked secondary polyclonal antibody added. Following washing, a substrate solution (1:250) was added into each well to allow color development. After termination of the reaction, the colorimetric intensity was measured at 450 nm on a SpectraMax M5 microplate reader. In comparison, IL-8 release of the cells treated with TiO2 NPs but without exposure to sunlight was also examined following the similar procedure. Fluorescence microscopy to investigate the mitochondrial dysfunction Mitochondrial membrane depolarization and superoxide generation were investigated by 5, 5’, 6, 6’-tetrachloro-1, 1’, 3, 3’-tetraethylbenzimidazolocarbocyanine iodide (JC-1) and MitoSox Red fluorescent indicators, respectively. There were 2 × 105 HaCaT cells in 800 μl of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at power intensity of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 6 h at 37°C in a humidified 5% CO2 incubator. Then, the culture medium was removed, and cells were washed with PBS 3 times. Each well received 800 μl of fresh culture medium containing Hoechst 33342 (1 μmol/l) and JC-1 (5 μmol/l) or Hoechst 33342 (1 μmol/l) and MitoSox Red (5 μmol/l). After 30-min incubation, fluorescent images were taken on Olympus BX-51 Optical System Microscope (Tokyo, Japan) with 20× objective. In comparison, the mitochondrial membrane depolarization and superoxide generation of cells treated with TiO2 NPs but without sunlight exposure were also investigated following the similar procedure. Flow cytometric analysis of cell cycle Cell cycle distribution was examined by evaluating the relative cellular DNA content with a flow cytometric technique. In total 2 × 105 HaCaT cells in 800 μl of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at light power intensity of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 18 h at 37°C in a humidified 5% CO2 incubator. Then the cells were detached by trypsin and washed 3 times with PBS. The harvested cells were fixed with ice-cold 70% (v/v) ethanol at 4°C for at least 6 h. Afterwards, cell pellets were washed twice with ice-cold PBS and resuspended in 1 ml of PBS containing (1 mg/ml) RNase and (50 μg/ml) propidium iodide (PI), and then incubated for 30 min in the dark at RT. Finally, samples were examined by a flow cytometer (BD Accuri C6). Cell cycle was analyzed according to the distribution of DNA content and divided into G1, S, and G2/M phases. Flow cytometric analysis of apoptotic cells There were 2 × 105 HaCaT cells in 800 μL of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at a power densityof 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 18 h at 37°C in a humidified 5% CO2 incubator. Then the cells were detached by trypsin (without ethylenediaminetetraacetic acid) and washed 3 times with PBS. Cells were re-suspended in 500 μl of binding buffer (10 × 10−3 M 4-[2-hydroxyethyl]-1-pipera zineethanesulfonic acid, 140 × 10−3 M NaCl, 2.5 × 10−3 M CaCl2, pH = 7.4), and mixed with 5 μl of 100 μg/ml FITC-conjugated Annexin V and 5 μl of 100 μg/ml PI, and incubated in the dark for 10 min. Then, the cell apoptosis was analyzed by a flow cytometer. Senescence-associated β-galactosidase staining There were 2 × 105 HaCaT cells in 800 μL of DMEM were plated in each well of a 12-multiwell plate (Costar, Corning) for overnight growth. After removal of the medium, and cells were treated with 800 μl of 50 μg/ml TiO2 NPs for 6 h, and then exposed to the simulated sunlight for 15 min at a power density of 0.1 W/cm2. After the sunlight exposure, the cells were further cultured for another 6 h at 37°C in a humidified 5% CO2 incubator. Then, the cells were fixed in 4% glutaraldehyde solution for 15 min at RT. The fixed cells were washed with PBS 3 times and then incubated with 200 μl of β-galactosidase staining solution (contain 20 mg/ml X-gal) (KeyGEN Bio TECH) over night at 37°C without CO2. Cells treated with 100 μM H2O2 were performed as positive control. Animals and treatment All animal studies were performed in Center for Experimental Animals, Jilin University, and the procedures involving experimental animals were in accordance with protocols approved by the Committee for Animal Research of Jilin University (SYXK [JL]2008-0011), China. The Balb/c hairless mice (4-weeks old, female, around 18 g) were randomly divided into 6 groups with 6 mice per group. Animals were acclimatized to laboratory conditions for 1 week prior to experiments and were on standard animal chow and water ad libitum with 12-h light/dark cycle. The temperature of the room was maintained at 18°C–22°C and the relative humidity about 50%–60%. One group was for the control group without any NP treatment, while the other 5 experimental groups were treated by 5 different types of TiO2 NPs including A, A7:R3, A4:R6, A3:R7, and R. Suspensions of different types of TiO2 NPs of 8 wt% were made with Pentalan-408 (pentaerythritoltetraethylhexanoate), which is similar to the sunscreen formulation (Pal et al., 2016), followed by 5 min of sonication to provide a uniform formulation. The dorsal surface of mice in experimental groups was exposed to 100 μl of above TiO2 NP formulations with a 4 cm2 (2 × 2 cm2) area for 3 h under simulated sunlight exposure, and the remaining residual NPs were removed from skin with lukewarm water. The mice in control group were only exposed to simulated sunlight for 3 h. After 14 consecutive day treatment, the mice were sacrificed and the skin was removed for further assay. SOD activity and MDA content analysis of the mouse skin The dorsal skin that was exposed to TiO2 NPs was exercised and washed with ice-cold PBS 3 times, and then homogenized in ice-cold PBS by ultrasonic cell sonicator at 200 W for 10 min (0°C). Homogenates were centrifuged at 8000× g for 15 min (4°C) and the supernatants were collected and stored at −20°C. The activities of SOD and the contents of MDA in dorsal skin homogenates were examined by standard reagent kit (Beyotime, Nanjing). Histopathologic analysis of the mouse skin Balb/c hairless mice were sacrificed at the termination of the experiment under anesthesia. The dorsal skin about 8 µm thick containing the epidermis that was exposed to TiO2 NPs were taken and washed with cold PBS for 3 times. After that, the skin tissues were fixed in 4% paraformaldehyde in 0.4 M phosphate buffer (pH 7.6). The skin tissues were dehydrated and embedded in paraffin, and then cut into 4 μm sections. The sections were stained with hematoxilin and eosin (H&E) and subsequently processed for histopathological examination under a light microscope. The thickness of the epidermis was measured afterwards by means of the light microscope (Olympus BX-51 Optical System Microscope [Tokyo, Japan]) with Image J system. Three to five views under the microscope were choice for measurement. The average thickness was recorded for compare. Statistical analysis All data were expressed as mean ± SD. All values were obtained from at least 3 independent experiments. Statistical significance was evaluated using 2-tailed heteroscedastic Student’s t tests according to the TTEST function in Microsoft Excel. The significant difference between groups was considered statistically significant when the p-value was < .05. RESULTS Physicochemical Characterization of Mixed-Phase TiO2 NPs With Different Anatase/Rutile Ratios The ratios of anatase/rutile were precisely adjusted by controlling the process of phase transformation from anatase (A) to rutile (R) through 600°C–1000°C, while other reaction conditions were kept consistent for achieving similar physicochemical properties. The phase ratio of A to R in mixed-phase TiO2 was determined by XRD (Luo et al., 2015). Figure 2A indicates that with the temperature increasing, the diffraction peak intensity of rutile phase gradually increases, while the corresponding intensity of anatase phase declines. Based on the XRD analysis, the anatase/rutile ratios of mixed-phase TiO2 NPs were determined as 7:3, 4:6, and 3:7, and the corresponding particles were named as A7:R3, A4:R6, and A3:R7, respectively. Raman spectrum is another widely used method to determine the phase transfer from anatase to rutile (Fang et al., 2008; Frank et al., 2012). Figure 2B shows that with the temperature increasing the Raman peak at 638 cm−1 attributed to anatase phase is gradually shifted until to 612 cm−1 that is attributed to rutile phase, corroborating the formation of mixed-phase NPs with different phase ratios. TEM images revealed the spherical morphology of these TiO2 NPs (Figure 2C), and HRTEM images (Figure 2C) further displayed the formed anatase/rutile heterojunctions in A7:R3, A4:R6, and A3:R7 regarding the coexisting anatase/rutile lattice fringes. The lattice fringe of anatase was about 0.35 nm, orientated in the (101) direction, while it was 0.32 nm for rutile corresponding to (110) direction (Ruan et al., 2013). The primary sizes of these NPs ranged from 80 ± 8 to 100 ± 9 nm (Figure 3A), and the corresponding specific surface area were determined by Brunauer–Emmett–Teller (BET) ranging from 39.9 ± 2.2 to 89.9 ± 4.7 m2/g (Figure 3B). Assessments of hydrodynamic sizes using DLS method demonstrated that all the TiO2 NPs could be well dispersed in water and their hydrodynamic sizes ranged from 315.0 ± 27.8 to 375.9 ± 11.1 nm (Figure 3C). ζ-potential measurements indicated all these TiO2 NPs had positive surface charges ranging from +22.58 ±2.47 to +26.99 ±1.83 mV in water (Figure 3D). These TiO2 NPs could be also well dispersed in cell culture medium (DMEM), showing similar hydrodynamic sizes of 354.0 ± 20.0 to 394.6 ± 4.2 nm (Supplementary Figure 1). In summary, the characterization data reveal that mixed-phase TiO2 NPs with different anatase/rutile ratios have been successfully obtained, showing homologous physicochemical properties comparable to pristine TiO2 NPs. Figure 2. View largeDownload slide Determination of anatase/rutile phase ratios in mixed-phase TiO2 NPs. (A) XRD patterns (The red and black vertical line represent the standard XRD pattern for anatase (JCPDS card No. 04-0477) and rutile TiO2 (JCPDS card No.04-0551), respectively, (B) Raman spectra and (C) TEM and HRTEM images. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Figure 2. View largeDownload slide Determination of anatase/rutile phase ratios in mixed-phase TiO2 NPs. (A) XRD patterns (The red and black vertical line represent the standard XRD pattern for anatase (JCPDS card No. 04-0477) and rutile TiO2 (JCPDS card No.04-0551), respectively, (B) Raman spectra and (C) TEM and HRTEM images. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.) Figure 3. View largeDownload slide Physicochemical properties of pristine and mixed-phase TiO2 NPs. (A) Primary size; (B) Specific surface area; (C) Hydrodynamic sizes; and (D) Zeta potentials. A 50 μg/ml NPs dispersed in water for measurement of hydrodynamic sizes and zeta potentials. Figure 3. View largeDownload slide Physicochemical properties of pristine and mixed-phase TiO2 NPs. (A) Primary size; (B) Specific surface area; (C) Hydrodynamic sizes; and (D) Zeta potentials. A 50 μg/ml NPs dispersed in water for measurement of hydrodynamic sizes and zeta potentials. BB Level for Evaluating ROS Production The formation of anatase/rutile heterojunctions provides the opportunity for the electron transfer between anatase and rutile phase if the staggered band edges get matched. The energy band structures of pristine and mixed-phase TiO2 NPs were constructed based on their band-gap energies (Eg), Fermi energies (Ef), and VBM energies that were determined by UV-visible diffuse reflection spectroscopy and UPS, respectively. Figure 4A and Supplementary Table 1 summarize the band energy information. The conduction- and valence-band edges of rutile TiO2 NPs were higher than those of anatase TiO2 NPs, respectively, meaning the excited electrons preferentially move from rutile to anatase while holes from anatase to rutile, which is consistent with previous results (Kullgren et al., 2015; Scanlon et al., 2013). This electron-hole separation at the interface of heterojunction can result in free electron and hole accumulation in anatase and rutile sides, respectively, which can induce the BB (Supplementary Figure 2) (Zhang and Yates, 2012). Since the degree of BB is proportional to the amount of accumulated electrons or holes that depend on the density of heterojunction, the degree of BB can reflect the density of heterojunction in mixed-phase TiO2 NPs. Figure 4B presents the BB value variation of these NPs, where A4:R6 shows the largest one, followed by A7:R3 and A3:R7. Pristine TiO2 (A and R) did not show any BB effect because there is no heterojunction interface and charge accumulation. The BB analysis suggested that the notable but varied charge accumulation existing in the mixed-phase TiO2 NPs with the phase ratio changing. The free electrons and holes at the interface of heterojunction can react with oxygen and water to form O2•− and HO•, respectively (Brezová et al., 2014; Dvoranova et al., 2014). The fluorescent dye, DCF, is usually used to generally detect the abiotic ROS production on nanomaterials (Zhang et al., 2014). Supplementary Figure 3A shows the significantly enhanced DCF fluorescence intensity induced by mixed-phase TiO2 NPs compared with pristine NPs after exposure to simulated sunlight, where A4:R6 exhibits the strongest fluorescence intensity, followed by A7:R3 and A3:R7. This trend is consistent with that of the degree of BB. However, only weak fluorescence enhancement was observed in these NPs without exposure to sunlight (Supplementary Figure 3B). Further ESR spectra distinguished the ROS types and revealed the more remarkable O2•− and HO• generation in mixed-phased TiO2 NPs than pristine particles (Figs. 4C and 4D) (He et al., 2014), supporting the ROS generation trend in the DCF assay. In comparison, without sunlight exposure, O2•− and HO• radicals were rarely induced by both pristine and mixed-phase TiO2 NPs (Supplementary Figure 4). Figure 4. View largeDownload slide Electronic properties of pristine and mixed-phase TiO2 NPs. (A) Energy structures, (B) BB degree. (C) ESR spectra for O2•−. (D) ESR spectra for HO•. Figure 4. View largeDownload slide Electronic properties of pristine and mixed-phase TiO2 NPs. (A) Energy structures, (B) BB degree. (C) ESR spectra for O2•−. (D) ESR spectra for HO•. All above results demonstrate that free electrons and holes can be derived from heterojunctions in mixed-phase TiO2 NPs and give rise to the induction of abiotic ROS generation upon sunlight exposure, and the degree of BB can well reflect the magnitude of ROS in mixed-phase TiO2 NPs. Toxic Potential Reflected by BB Level Both O2•− and HO• radicals are highly reactive, capable of oxidizing a vast of biomacromolecules and cell organelles, and causing cell oxidative injury (Li et al., 2012). After incubated with these different TiO2 NPs (200 µg/ml) for 6 h, HaCaT cells showed similar cellular uptake of TiO2 NPs based on the cellular Ti contents as determined by inductively coupled plasma optical emission spectrometry (ICP-OES) (Supplementary Figure 5). To benchmark the relative toxicity of the mixed-phase TiO2 NPs compared with pristine particles, the cellular viability of HaCaT cells was assessed by MTS assays upon exposure to 0.4-200 µg/ml TiO2 NPs for 6 h, followed by 15 min of exposure to simulated sunlight and another 18 h of incubation. Figure 5A shows that mixed-phase TiO2 NPs can induce more robust decline in cell viability than anatase and rutile TiO2, and A4:R6 still exhibits the most potent toxicity, followed by A7:R3 and A3:R7, which is consistent with the tends of ROS production and BB levels of these mixed-phase TiO2 NPs. To reveal the critical ROS role in toxicity induced by these TiO2 NPs, N-acetyl-cysteine (NAC) as ROS scavenger was used to pretreat the HaCaT cells (Xue et al., 2011), and the cell viability was assessed again by MTS assay following treatment with TiO2 NPs under sunlight exposure. The result shows the cytotoxicity of mixed-phase TiO2 NPs can be significantly reduced in HaCaT cells after NAC pretreatment (Supplementary Figure 6), demonstrating ROS scavenger can protect skin cells from ROS attacking of mixed-phase TiO2 NPs under sunlight exposure. In comparison, without sunlight exposure, the cell viability of HaCaT cells was weakly reduced by these TiO2 NPs (Supplementary Figure 7A). This toxicity study demonstrates that mixed-phase TiO2 NPs can cause more severe toxicity than pristine TiO2 NPs in HaCaT cells upon exposure to sunlight, with the larger BB level corresponding to the stronger toxicity. Sunlight-mediated ROS generation plays a critical role in induction of toxicity of mixed-phase TiO2 NPs. Figure 5. View largeDownload slide Cytotoxicity, cellular ROS generation and GSH depletion in HaCat cells exposed to TiO2 NPs under sunlight condition. (A) MTS assay, (B) Flow cytometric analysis for cellular DCF fluorescence (inset showing the fold increase in integral area), and (C) Cellular GSH assessment based on DNTB methods. Cells were treated with 0.4-200 µg/ml (for MTS assay) or 50 µg/ml (for ROS and GSH assessments) of TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h (for MTS assay) or 6 h (ROS and GSH assessments) of incubation, respectively. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 9, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 5. View largeDownload slide Cytotoxicity, cellular ROS generation and GSH depletion in HaCat cells exposed to TiO2 NPs under sunlight condition. (A) MTS assay, (B) Flow cytometric analysis for cellular DCF fluorescence (inset showing the fold increase in integral area), and (C) Cellular GSH assessment based on DNTB methods. Cells were treated with 0.4-200 µg/ml (for MTS assay) or 50 µg/ml (for ROS and GSH assessments) of TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h (for MTS assay) or 6 h (ROS and GSH assessments) of incubation, respectively. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 9, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Intracellular ROS Production and GSH Depletion Since abundant O2•− and HO• radicals are generated in mixed-phase TiO2 NPs upon exposure to simulated sunlight, the molecular mechanism underlying the toxicity of these TiO2 NPs is probably ascribed to activation of oxidative stress (Hamzeh and Sunahara, 2013). Intracellular ROS production and GSH depletion usually are indicative of oxidative stress injury. Flow cytometry analysis was carried out to detect the intracellular ROS level using H2-DCF-DA, while the classic DNTB method was employed to determine the intracellular GSH level. Figure 5B shows the mixed-phase TiO2 NPs can induce more potent intracellular DCF fluorescence than pristine TiO2 NPs in HaCaT cells after exposure to sunlight. The fluorescence enhancement trend based on the fold increase in integral area (inset of Figure 5B) is still in accordance with that of the degree of BB, showing the largest fluorescence enhancement induced by A4:R6, followed by A7:R3 and A3:R7. Figure 5C further displays the corresponding cellular GSH levels, where mixed-phase TiO2 NPs can more remarkably reduce the cellular GSH level than pristine particles, and A4:R6 still is the most potent one. In comparison, without sunlight exposure, ROS and GSH levels in HaCaT cells were rarely affected by TiO2 NPs (Supplementary Figs. 7B and 7C). Cellular ROS production and GSH depletion strongly imply the toxicity mechanism of mixed-phase TiO2 NPs under sunlight exposure is correlated with oxidative stress. Hierarchical Oxidative Stress Responses According to oxidative stress paradigm, the hierarchical oxidative stress responses are characterized by an antioxidant defense response (tier 1), the initiation proinflammatory (tier 2), and mitochondrial-mediated cytotoxicity (tier 3) (Nel et al., 2006). The lowest levels of oxidative stress (tier 1) are associated with the activation of cytoprotective enzymes such as HO-1. Western blotting analysis (Figure 6A) revealed that, with simulated sunlight exposure, A4:R6 triggered the most abundant HO-1 expression, followed by A7:R3 and A3:R7, which is proportional to the degree of BB. Failure to restore redox equilibrium in tier 1 is capable of activating proinflammatory signaling pathways such as the Jun kinase and NF-κB cascades, which are involved in the transcriptional activation of cytokine, chemokine, and adhesion gene promoters. IL-8 as typical tier 2 proinflammatory response has been investigated to demonstrate the inflammatory effects in HaCaT cells (Comfort et al., 2014). An ELISA assay to detect IL-8 release in the cellular supernatant of HaCaT cells indicated the incremental increase in chemokine production with the increase of BB in mixed-phase TiO2 NPs (Figure 6B). Escalation of oxidative stress response to tier 3 can trigger mitochondrial dysfunction, including mitochondrial superoxide generation and mitochondrial membrane depolarization, ultimately leading to cell death. The fluorescent dyes, MitoSox Red and JC-1, were applied to detect the mitochondrial superoxide generation and mitochondrial membrane depolarization, respectively, through fluorescence microscopy. Figure 6C shows both mitochondrial-mediated toxicological responses are found the most significant in the cells exposed to A4:R6, and these mitochondrial dysfunction effects still followed the trend of BB degrees. Also, in comparison, without sunlight exposure, above tier 1 to tier 3 hierarchical oxidative stress responses including HO-1 expression, IL-8 release and mitochondrial dysfunction were not induced by mixed-phase and pristine TiO2 NPs (Supplementary Figure 8). All these in vitro hierarchical oxidative stress responses corroborate the activation of oxidative stress by mixed-phase TiO2 NPs under sunlight exposure. Figure 6. View largeDownload slide Hierarchical oxidative stress responses in HaCaT cells exposed to TiO2 NPs under sunlight condition. (A) Western blot analysis for HO-1 expression; (B) ELISA assessment for IL-8; and (C) Fluorescence images of cells stained by Mitosox and JC-1 to detect mitochondrial superoxide generation and membrane depolarization. Cells were treated with 50 µg/ml of various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 6 h of incubation. Flow cytometric analysis for cell cycle phase arrest (D) and apoptotic cells (E) of the HaCaT cells. Cells were treated with 50 µg/ml various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h of incubation, and stained with PI for cell cycle detection or Annexin V-FITC/PI for apoptosis detection. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 6. View largeDownload slide Hierarchical oxidative stress responses in HaCaT cells exposed to TiO2 NPs under sunlight condition. (A) Western blot analysis for HO-1 expression; (B) ELISA assessment for IL-8; and (C) Fluorescence images of cells stained by Mitosox and JC-1 to detect mitochondrial superoxide generation and membrane depolarization. Cells were treated with 50 µg/ml of various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 6 h of incubation. Flow cytometric analysis for cell cycle phase arrest (D) and apoptotic cells (E) of the HaCaT cells. Cells were treated with 50 µg/ml various TiO2 NPs for 6 h, followed by 15 min of sunlight exposure and another 18 h of incubation, and stained with PI for cell cycle detection or Annexin V-FITC/PI for apoptosis detection. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Cell Cycle Distribution, Apoptosis, and Senescence Oxidative stress has been known capable of inducing cell death through apoptosis and senescence (Fulda et al., 2010; Haines et al., 2013; Kim et al., 2017; Ozben, 2007). To identify the type of cell death induced by mixed-phase TiO2 NPs, additional investigation on cell cycle distribution, apoptosis and senescence were performed. The effect of these NPs on the cell cycle progression was examined by flow cytometry using PI staining. Figure 6D shows that all these TiO2 NPs can lead to a significantly increased accumulation of the G2/M-phase cells and reduced number of G1-phase and S-phase cells, where A4:R6 induces the most significant G2/M arrest (10.8%), followed by A7:R3 (9.0%), and A3:R7 (8.5%). Then, induction of apoptosis by these NPs was evaluated by flow cytometry using Annexin V-FITC/PI staining. Figure 6E shows that A, A7:R3, A4:R6, A3:R7, and R TiO2 NPs can induce 24%, 37.5%, 58.2%, 34.8%, and 27.8% of the early apoptotic cells, and 1.1%, 0.1%, 0.5%, 0.4%, and 0.1% of the late apoptotic cells, respectively, where A4:R6 still induces the largest magnitude of apoptotic cells. Moreover, cellular senescence was detected by bright field microscope based on senescence-associated beta-galactosidase (SA-β-gal) activity (Marazita et al., 2016). As shown in Supplementary Figure 9, after treatment with mixed-phase or pure phase TiO2 NPs, HaCaT cells did not display positive (SA-β-gal) staining, while 100 μM H2O2 as a positive control could induce noticeable staining (Choo et al., 2014). In Vivo Mouse Skin Damage In vivo SOD activity and MDA content are biomarkers of antioxidant defense, and their increased levels are indicative of oxidative stress injury. To further confirm the significant in vitro toxicity of mixed-phase TiO2 NPs in skin cells, the dorsal skin of Balb/c hairless mice were exposed to various pristine and mixed-phase TiO2 NPs for 14 days with 3 h of daily simulated sunlight exposure, and in vivo SOD activity and MDA content were analyzed. As shown in Figure 7A, compared with the control group, the topical treatment of TiO2 NPs followed by sunlight exposure causes significantly increased SOD activities in the mice skin, where mixed-phase TiO2 NPs induce higher SOD activities than pristine particles, and A4:R6 showed the highest activity, followed by A7:R3 and A3:R7. Figure 7B indicates the MDA contents of mice skin exposed to TiO2 NPs, can display the similar tendency as shown in SOD activity. Both in vivo SOD activity and MDA content assessments confirm the in vitro results that mixed-phase TiO2 NPs can cause more severe oxidative stress than pristine particles under sunlight exposure. Moreover, the oxidative stress-mediated inflammation usually can result in pathological changes in skin ultrastructures (Pal et al., 2016). The epidermal thickness of mouse skin was evaluated in skin tissues after H&E staining. Figure 7C shows the mixed-phase TiO2 NPs can more significantly increase the thickness of mouse skin than anatase (the epidermal thickness was about 24.3 μm) and rutile (32.6 μm) NPs, and the most pronounced effect is found on A4:R6. The epidermal thickness for A7:R3, A4:R6, and A3:R7 were about 44.6, 53.8, and 35.7 μm. Obviously, the more severe in vivo skin damage is induced by mixed-phase TiO2 NPs compared with pristine particles, which intensely correlates to the abiotic O2•− and HO• radicals generation and can be predicted by BB degrees. Figure 7. View largeDownload slide In vivo SOD and MDA level in the BALB/C hairless mice skin tissue and the histopathological evaluation after dermal exposure to TiO2 NPs. (A) SOD activity, (B) MDA level, and (C) TiO2 exposure induced epidermal thickness in mouse skin. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. Figure 7. View largeDownload slide In vivo SOD and MDA level in the BALB/C hairless mice skin tissue and the histopathological evaluation after dermal exposure to TiO2 NPs. (A) SOD activity, (B) MDA level, and (C) TiO2 exposure induced epidermal thickness in mouse skin. Data are shown as mean ± SD, Student’s t test, *p < .05 as compared with untreated cells, n = 3, and #p < .05 for pairwise comparison between A7:R3, A4:R6, and A3:R7. DISCUSSION Pristine TiO2 NPs (anatase and rutile) are usually considered as safe nanomaterials and have been widely used as ingredients in cosmetics and cotton textiles (Barker and Branch, 2008; Contado and Pagnoni, 2008; Doganli et al., 2016; Meilert et al., 2005). Although TiO2 NPs have also been well known capable of causing cell injury under UV light exposure (George et al., 2014; Shakeel et al., 2016; Uchino et al., 2002), the low skin penetration ability of TiO2 NPs is considered beneficial for reducing their adverse effects to human skin (Sadrieh et al., 2010; Wu et al., 2009). However, most these skin penetration studies do not consider the influence of sunlight exposure. Actually, the sunlight exposure can significantly deepen the penetration of NPs into skin (Bennett et al., 2012), meaning the NPs probably are more dangerous under sunlight exposure. Thus, it is necessary to investigate the toxicity of TiO2 NPs on skin under sunlight exposure not only because of the photocatalytic properties of TiO2 NPs but also because of the sunlight-assisted skin penetration. Anatase/rutile mixed-phase TiO2 NPs recently were found in cosmetics and cotton textiles (Doganli et al., 2016; Meilert et al., 2005), and exhibited more severe toxicity than pristine TiO2 NPs (Wu et al., 2009; Yin et al., 2012). Although the stronger toxicity of mixed-phase TiO2 NPs can be easily ascribed to their active electronic property arising from the heterogeneous interface, there is a lack of a clear property-activity relationship to understand the toxicity mechanism and predict the toxicity potential based on the semiconductor property of mixed-phase TiO2 NPs. Moreover, since anatase/rutile mixed-phase TiO2 NPs can have different anatase/rutile ratios based on distinctive preparation process, there also lacks a feasible strategy to evaluate the toxicity difference based on their physicochemical property. In the present study, a series of mixed-phase TiO2 NPs with different anatase/rutile ratios were prepared to evaluate their oxidative injury on HaCaT cells and mouse skin under sunlight exposure, unveiling the potential property-toxicity relationship. Since the different anatase/rutile ratios could lead to different densities of anatase/rutile heterojunction, the BB concept was introduced herein to reflect the density of heterojunction, which is beneficial for quantitatively assessing the biological effect based on the semiconductor property. Actually, mixed-phase NPs were proved to be able to promote electron transfer from rutile to anatase based on their energy structures and produced more significant O2•− and HO• radicals than pristine NPs (Figs. 4C and 4D) under sunlight exposure, leading to more severe oxidative injury on HaCaT cells (Figure 5A) and mouse skin (Figure 7). The BB levels of A7:R3, A4:R6, and A3:R7 mixed-phase TiO2 NPs were found to be proportional to their abiotic radical amounts as well as in vitro and in vivo toxicological response levels and the extent of damage. Previous studies have successfully made attempts to correlate the electronic property of semiconductor nanomaterials with their toxicity (Burello and Worth, 2011; Zhang et al., 2012). This study further corroborates that the electronic property of semiconductor nanomaterials can be used to evaluate their toxicity potential. Under sunlight exposure, mixed-phase NPs was proved to be able to generate a large amount of O2•− and HO• radicals. As typical ROS, O2•− and HO• radicals can react with a range of biomolecules and exert potent adverse effects in damage of cell organelles and disruption of cellular redox homeostasis, leading to direct and indirect cellular ROS production (Chang et al., 2017; Liu et al., 2016; Meng et al., 2009). To overcome the ROS production, cells usually trigger either a defensive or an antioxidative response eliciting a chain of adverse biological responses (Manke et al., 2013). Cells exposed to mixed-phase NPs probably can induce similar biological responses, which is helpful for clarification of the molecular mechanism. Typical intracellular ROS production (Figure 5B and Supplementary Figure 7B) and GSH depletion (Figure 5C and Supplementary Figure 7C) of HaCaT cells were induced by mixed-phase TiO2 NPs, suggesting the high possibility of activation of oxidative stress signaling pathway. Further hierarchical oxidative stress response assessments corroborated mixed-phase TiO2 NPs could elicit HO-1 phase II enzyme expression (tier 1), IL-8 proinflammatory CXC chemokine release (tier 2), and mitochondrial dysfunction (tier 3) including mitochondrial membrane depolarization and superoxide generation (Figure 6C and Supplementary Figure 8). More importantly, all mixed-phase TiO2 NPs-induced in vitro toxic response levels as well as in vivo mouse skin damage extents (Figure 7) were proportional to O2•− and HO• amounts generated in mixed-phase TiO2 NPs (A4:R6, A7:R3, and A3:R7). Obviously, the in vitro and in vivo toxicity induced by mixed-phase TiO2 NPs under sunlight exposure are based on activation of oxidative stress. Oxidative stress can induce multiple types of cell death, such as apoptosis and senescence. Significant G2/M arrest associated with generation of apoptotic cells induced by mixed-phase TiO2 NPs confirms the cell death is possibly ascribed to apoptosis. G2/M phase can ensure the accuracy of allocation during mitosis. When some damaged DNA enters into the next phase of the cell cycle from 1 phase before being repaired, this damage can be “fixed down” to result in genetic instability of genome and potential likelihood of death. These results are consistent with the report that severe oxidative stress of TiO2 NPs is capable of inducing cell apoptosis (Saira et al., 2016; Wang et al., 2015). It has been reported that cellular senescence is usually associated with G1 arrest (Tao et al., 2017; Yi et al., 2013). Integrated with the absence of senescence–associated SA-β-gal activity, it can be concluded the type of cell death induced by mixed-phase TiO2 NPs is not correlated with cellular senescence. In summary, under sunlight exposure, the energy band alignment in mixed-phased TiO2 NPs can drive the electron transfer between anatase and rutile, resulting in the electron-hole separation at the interface of heterojunction. The formed free electron and hole can react with oxygen and water to produce O2•− and HO• radicals, respectively, resulting in HaCaT cell death and mouse skin damage. The underlying mechanism is involved in activation of oxidative stress as manifested by hierarchical oxidative stress toxicological responses. Without sunlight exposure, mixed-phased and pristine TiO2 NPs cannot induce notable toxicity in HaCaT cells and mouse skin. The density of heterojunction of mixed-phased TiO2 NPs is theoretically proportional to the amount of produced free electrons and holes at the interface of heterojunction, which can be experimentally reflected by the degree of band banding. Thus, the degree of BB in mixed-phased TiO2 NPs can be used to assess their ROS generation, ultimately predicting their induced toxicity potential in HaCaT cells and mouse skin. SUPPLEMENTARY DATA Supplementary data are available at Toxicological Sciences online. FUNDING This work was supported by the National Natural Science Foundation of China (Grant Nos. 21573216, 21501170, 21777152 and 21703232) and Hundred Talent Program of Chinese Academy of Sciences, Science and Technology Development Project Foundation of Jilin Province (20160101304JC, 20160520134JH, and 20180520145JH). REFERENCES Barker P. J. , Branch A. ( 2008 ). The interaction of modern sunscreen formulations with surface coatings . Prog. Org. Coat . 62 , 313 – 320 . Google Scholar CrossRef Search ADS Bennett S. W. , Zhou D. , Mielke R. , Keller A. A. ( 2012 ). Photoinduced disaggregation of TiO2 nanoparticles enables transdermal penetration . PLoS One 7 , e48719 – e48725 . Google Scholar CrossRef Search ADS PubMed Brezová V. , Barbieriková Z. , Zukalová M. , Dvoranová D. , Kavan L. ( 2014 ). EPR study of 17O-enriched titania nanopowders under UV irradiation . Catal. Today 230 , 112 – 118 . Google Scholar CrossRef Search ADS Brezova V. , Gabcova S. , Dvoranova D. , Stasko A. ( 2005 ). Reactive oxygen species produced upon photoexcitation of sunscreens containing titanium dioxide (an EPR study) . J. Photoch. Photobio. B 79 , 121 – 134 . Google Scholar CrossRef Search ADS Burello E. , Worth A. P. ( 2011 ). A theoretical framework for predicting the oxidative stress potential of oxide nanoparticles . Nanotoxicology 5 , 228 – 235 . Google Scholar CrossRef Search ADS PubMed Contado C. , Pagnoni A. ( 2008 ). TiO2 in commercial sunscreen lotion: Flow field-flow fractionation and ICP-AES together for size analysis . Anal. Chem . 80 , 7594 – 7608 . Google Scholar CrossRef Search ADS PubMed Chang Y. , Li K. , Feng Y. , Cheng Y. , Zhang M. , Wang Z. , Wu Z. , Zhang H. ( 2017 ). Achievement of safer palladium nanocrystals by enlargement of {100} crystallographic facets . Nanotoxicology 11 , 907 – 922 . Google Scholar CrossRef Search ADS PubMed Choo K. B. , Tai L. , Hymavathee K. S. , Wong C. Y. , Nguyen P. N. N. , Huang C.-J. , Cheong S. K. , Kamarul T. ( 2014 ). Oxidative stress-induced premature senescence in Wharton’s jelly-derived mesenchymal stem cells . Int. J. Med. Sci . 11 , 1201 – 1207 . Google Scholar CrossRef Search ADS PubMed Comfort K. K. , Braydich-Stolle L. K. , Maurer E. I. , Hussain S. M. ( 2014 ). Less is more: Long-term in vitro exposure to low levels of silver nanoparticles provides new insights for nanomaterial evaluation . ACS Nano 8 , 3260 – 3271 . Google Scholar CrossRef Search ADS PubMed Deák P. , Kullgren J. , Aradi B. , Frauenheim T. , Kavan L. ( 2016 ). Water splitting and the band edge positions of TiO2 . Electrochim. Acta 199 , 27 – 34 . Google Scholar CrossRef Search ADS Doganli G. , Yuzer B. , Aydin I. , Gultekin T. , Con A. H. , Selcuk H. , Palamutcu S. ( 2016 ). Functionalization of cotton fabric with nanosized TiO2 coating for self-cleaning and antibacterial property enhancement . J. Coat. Technol. Res . 13 , 257 – 265 . Google Scholar CrossRef Search ADS Dvoranova D. , Barbierikova Z. , Brezova V. ( 2014 ). Radical intermediates in photoinduced reactions on TiO2 (an EPR spin trapping study) . Molecules 19 , 17279 – 17304 . Google Scholar CrossRef Search ADS PubMed Edmonds M. T. , Pakes C. I. , Mammadov S. , Zhang W. , Tadich A. , Ristein J. , Ley L. ( 2011 ). Surface band bending and electron affinity as a function of hole accumulation density in surface conducting diamond . Appl. Phys. Lett . 98 , 102101 – 102103 . Google Scholar CrossRef Search ADS Fang D. , Huang K. , Liu S. , Huang J. ( 2008 ). Fabrication and photoluminiscent properties of titanium oxide nanotube arrays . J. Braz. Chem. Soc . 19 , 1059 – 1064 . Google Scholar CrossRef Search ADS Frank O. , Zukalova M. , Laskova B. , Kürti J. , Koltai J. , Kavan L. ( 2012 ). Raman spectra of titanium dioxide (anatase, rutile) with identified oxygen isotopes (16, 17, 18) . Phy. Chem. Chem. Phys . 14 , 14567 – 14572 . Google Scholar CrossRef Search ADS Fulda S. , Gorman A. M. , Hori O. , Samali A. ( 2010 ). Cellular stress responses: Cell Survival and Cell Death . Int. J. Cell Biol . 2010 , 1. George S. , Gardner H. , Seng E. K. , Chang H. , Wang C. , Yu Fang C. H. , Richards M. , Valiyaveettil S. , Chan W. K. ( 2014 ). Differential effect of solar light in increasing the toxicity of silver and titanium dioxide nanoparticles to a fish cell line and zebrafish embryos . Environ. Sci. Technol . 48 , 6374 – 6382 . Google Scholar CrossRef Search ADS PubMed Gerloff K. , Fenoglio I. , Carella E. , Kolling J. , Albrecht C. , Boots A. W. , Förster I. , Schins R. P. F. ( 2012 ). Distinctive toxicity of TiO2 rutile/anatase mixed phase nanoparticles on Caco-2 Cells . Chem. Res. Toxicol . 25 , 646 – 655 . Google Scholar CrossRef Search ADS PubMed Gunter O. , Vicki S. , Ken D. ( 2007 ). Toxicology of nanoparticles: A historical perspective . Nanotoxicology 1 , 2 – 25 . Google Scholar CrossRef Search ADS Haines D. D. , Juhasz B. , Tosaki A. ( 2013 ). Management of multicellular senescence and oxidative stress . J. Cell. Mol. Med . 17 , 936 – 957 . Google Scholar CrossRef Search ADS PubMed Hamzeh M. , Sunahara G. I. ( 2013 ). In vitro cytotoxicity and genotoxicity studies of titanium tioxide (TiO2) nanoparticles in chinese hamster lung fibroblast cells . Toxicol. In Vitro 27 , 864 – 873 . Google Scholar CrossRef Search ADS PubMed He W. , Liu Y. , Wamer W. G. , Yin J. J. ( 2014 ). Electron spin resonance spectroscopy for the study of nanomaterial-mediated generation of reactive oxygen species . J. Food Drug Anal . 22 , 49 – 63 . Google Scholar CrossRef Search ADS PubMed Kim C. , Kang D. , Lee E. K. , Lee J.-S. ( 2017 ). Long noncoding RNAs and RNA-binding proteins in oxidative stress, cellular senescence, and age-related diseases . Oxid. Med. Cell. Longev . 2017 , 1. Kullgren J. , Aradi B. , Frauenheim T. , Kavan L. , Deák P. ( 2015 ). Resolving the controversy about the band alignment between rutile and anatase: The role of OH–/H+adsorption . J. Phys. Chem. C 119 , 21952 – 21958 . Google Scholar CrossRef Search ADS Li Y. , Zhang W. , Niu J. , Chen Y. ( 2012 ). Mechanism of photogenerated reactive oxygen species and correlation with the antibacterial properties of engineered metal-oxide nanoparticles . ACS Nano 6 , 5164 – 5173 . Google Scholar CrossRef Search ADS PubMed Liu N. , Li K. , Li X. , Chang Y. , Feng Y. , Sun X. , Cheng Y. , Wu Z. , Zhang H. ( 2016 ). Crystallographic facet-induced toxicological responses by faceted titanium dioxide nanocrystals . ACS Nano 10 , 6062 – 6073 . Google Scholar CrossRef Search ADS PubMed Luo Z. , Poyraz A. S. , Kuo C.-H. , Miao R. , Meng Y. , Chen S.-Y. , Jiang T. , Wenos C. , Suib S. L. ( 2015 ). Crystalline mixed phase (anatase/rutile) mesoporous titanium dioxides for visible light photocatalytic activity . Chem. Mater . 27 , 6 – 17 . Google Scholar CrossRef Search ADS Manke A. , Wang L. , Rojanasakul Y. ( 2013 ). Mechanisms of nanoparticle-induced oxidative stress and toxicity . Biomed. Res. Int . 2013 , 942916 – 942931 . Google Scholar CrossRef Search ADS PubMed Marazita M. C. , Dugour A. , Marquioni-Ramella M. D. , Figueroa J. M. , Suburo A. M. ( 2016 ). Oxidative stress-induced premature senescence dysregulates VEGF and CFH expression in retinal pigment epithelial cells: Implications for Age-related Macular Degeneration . Redox Biol. 7 , 78 – 87 . Google Scholar CrossRef Search ADS PubMed Meilert K. T. , Laub D. , Kiwi J. ( 2005 ). Photocatalytic self-cleaning of modified cotton textiles by TiO2 clusters attached by chemical spacers . J. Mol. Catal. A Chem . 237 , 101 – 108 . Google Scholar CrossRef Search ADS Meng H. , Xia T. , George S. , Nel A. E. ( 2009 ). A predictive toxicological paradigm for the safety assessment of nanomaterials . ACS Nano 3 , 1620 – 1627 . Google Scholar CrossRef Search ADS PubMed Nel A. , Xia T. , Mädler L. , Li N. ( 2006 ). Toxic potential of materials at the nanolevel . Science 311 , 622 – 627 . Google Scholar CrossRef Search ADS PubMed Ozben T. ( 2007 ). Oxidative stress and apoptosis: Impact on cancer therapy . J. Pharm. Sci . 96 , 2181 – 2196 . Google Scholar CrossRef Search ADS PubMed Pal A. , Alam S. , Chauhan L. K. S. , Saxena P. N. , Kumar M. , Ansari G. N. , Singh D. , Ansari K. M. ( 2016 ). UVB exposure enhanced the dermal penetration of zinc oxide nanoparticles and induced inflammatory responses through oxidative stress mediated by MAPKs and NF-κB signaling in SKH-1 hairless mouse skin . Toxicol. Res . 5 , 1066 – 1077 . Google Scholar CrossRef Search ADS Ruan P. , Qian J. , Xu Y. , Xie H. , Shao C. , Zhou X. ( 2013 ). Mixed-phase TiO2 nanorods assembled microsphere: Crystal phase control and photovoltaic application . CrystEngComm. 15 , 5093 – 5099 . Google Scholar CrossRef Search ADS Sadrieh N. , Wokovich A. M. , Gopee N. V. , Zheng J. , Haines D. , Parmiter D. , Siitonen P. H. , Cozart C. R. , Patri A. K. , McNeil S. E. et al. , . ( 2010 ). Lack of significant dermal penetration of titanium dioxide from sunscreen formulations containing nano- and submicron-size TiO2 particles . Toxicol. Sci . 115 , 156 – 166 . Google Scholar CrossRef Search ADS PubMed Saira F. , Mackey M. A. , Qureshi R. , Mahmoud M. A. ( 2016 ). In vitro investigations of gold nanocages: Toxicological profile in human keratinocyte cell line . Food Chem. Toxicol . 97 , 89 – 95 . Google Scholar CrossRef Search ADS PubMed Scanlon D. O. , Dunnill C. W. , Buckeridge J. , Shevlin S. A. , Logsdail A. J. , Woodley S. M. , Catlow C. R. A. , Powell M. J. , Palgrave R. G. , Parkin I. P. et al. , . ( 2013 ). Band alignment of rutile and anatase TiO2 . Nat. Mater . 12 , 7 , 798 – 801 . Google Scholar CrossRef Search ADS PubMed Seaton A. , Tran L. , Aitken R. , Donaldson K. ( 2010 ). Nanoparticles, human health hazard and regulation . J. R. Soc. Inter . 7 , S119 – S129 . Google Scholar CrossRef Search ADS Serpone N. , Dondi D. , Albini A. ( 2007 ). Inorganic and organic UV filters: Their role and efficacy in sunscreens and suncare product . Inorg. Chim. Acta 360 , 794 – 802 . Google Scholar CrossRef Search ADS Shakeel M. , Jabeen F. , Shabbir S. , Asghar M. S. , Khan M. S. , Chaudhry A. S. ( 2016 ). Toxicity of nano-titanium dioxide (TiO2-NP) through various routes of exposure: A review . Biol. Trace Elem. Res . 172 , 1 – 36 . Google Scholar CrossRef Search ADS PubMed Sun X. , Dai W. , Wu G. , Li L. , Guan N. , Hunger M. ( 2015 ). Evidence of rutile-to-anatase photo-induced electron transfer in mixed-phase TiO2 by solid-state NMR spectroscopy . Chem. Commun . 51 , 13779 – 13782 . Google Scholar CrossRef Search ADS Tao Y.-F. , Wang N.-N. , Xu L.-X. , Li Z.-H. , Li X.-L. , Xu Y.-Y. , Fang F. , Li M. , Qian G.-H. , Li Y.-H. et al. , . ( 2017 ). Molecular mechanism of G1 arrest and cellular senescence induced by LEE011, a novel CDK4/CDK6 inhibitor, in leukemia cells . Cancer Cell Int . 17 , 35. Google Scholar CrossRef Search ADS PubMed Tiwari A. , Mondal I. , Ghosh S. , Chattopadhyay N. , Pal U. ( 2016 ). Fabrication of mixed phase TiO2 heterojunction nanorods and their enhanced photoactivities . Phy. Chem. Chem. Phys . 18 , 15260 – 15268 . Google Scholar CrossRef Search ADS Uchino T. , Tokunaga H. , Ando M. , Utsumi H. ( 2002 ). Quantitative determination of OH radical generation and its cytotoxicity induced by TiO2-UVA treatment . Toxicol. In Vitro 16 , 629 – 635 . Google Scholar CrossRef Search ADS PubMed Wang Y. , Cui H. , Zhou J. , Li F. , Wang J. , Chen M. , Liu Q. ( 2015 ). Cytotoxicity, DNA damage, and apoptosis induced by titanium dioxide nanoparticles in human non-small cell lung cancer A549 cells . Environ. Sci. Pollut. Res. 22 , 5519 – 5530 . Google Scholar CrossRef Search ADS Wu J. , Liu W. , Xue C. , Zhou S. , Lan F. , Bi L. , Xu H. , Yang X. , Zeng F.-D. ( 2009 ). Toxicity and penetration of TiO2 nanoparticles in hairless mice and porcine skin after subchronic dermal exposure . Toxicol. Lett . 191 , 1 – 8 . Google Scholar CrossRef Search ADS PubMed Xue C. , Liu W. , Wu J. , Yang X. , Xu H. ( 2011 ). Chemoprotective effect of N-acetylcysteine (NAC) on cellular oxidative damages and apoptosis induced by nano titanium dioxide under UVA irradiation . Toxicol. In Vitro 25 , 110 – 116 . Google Scholar CrossRef Search ADS PubMed Yi G. , He Z. M. , Zhou X. K. , Xian L.W. , Yuan T. , Jia X. , Hong J. , He L , Liu J. ( 2013 ). Low concentration of metformin induces a p53-dependent senescence in hepatoma cells via activation of the AMPK pathway . Int. J. Oncol . 43 , 1503 – 1510 . Google Scholar CrossRef Search ADS PubMed Yin J.-J. , Liu J. , Ehrenshaft M. , Roberts J. E. , Fu P. P. , Mason R. P. , Zhao B. ( 2012 ). Phototoxicity of nano titanium dioxides in HaCaT keratinocytes-generation of reactive oxygen species and cell damage . Toxicol. Appl. Pharmacol . 263 , 81 – 88 . Google Scholar CrossRef Search ADS PubMed Zhang H. , Ji Z. , Xia T. , Meng H. , Low-Kam C. , Liu R. , Pokhrel S. , Lin S. , Wang X. , Liao Y.-P. et al. , . ( 2012 ). Use of metal oxide nanoparticle band gap to develop a predictive paradigm for oxidative stress and acute pulmonary inflammation . ACS Nano 6 , 4349 – 4368 . Google Scholar CrossRef Search ADS PubMed Zhang H. , Pokhrel S. , Ji Z. , Meng H. , Wang X. , Lin S. , Chang C. H. , Li L. , Li R. , Sun B. et al. , . ( 2014 ). PdO doping tunes band-gap energy levels as well as oxidative stress responses to a Co(3)O(4) P-type semiconductor in cells and the lung . J. Am. Chem. Soc . 136 , 6406 – 6420 . Google Scholar CrossRef Search ADS PubMed Zhang Z. , Yates J. T. Jr. ( 2012 ). Band bending in semiconductors: Chemical and physical consequences at surfaces and interfaces . Chem. Rev . 112 , 5520 – 5551 . Google Scholar CrossRef Search ADS PubMed © The Author(s) 2018. Published by Oxford University Press on behalf of the Society of Toxicology. All rights reserved. For permissions, please e-mail: journals.permissions@oup.com This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/journals/pages/about_us/legal/notices)

Journal

Toxicological SciencesOxford University Press

Published: Apr 12, 2018

There are no references for this article.

You’re reading a free preview. Subscribe to read the entire article.


DeepDyve is your
personal research library

It’s your single place to instantly
discover and read the research
that matters to you.

Enjoy affordable access to
over 18 million articles from more than
15,000 peer-reviewed journals.

All for just $49/month

Explore the DeepDyve Library

Search

Query the DeepDyve database, plus search all of PubMed and Google Scholar seamlessly

Organize

Save any article or search result from DeepDyve, PubMed, and Google Scholar... all in one place.

Access

Get unlimited, online access to over 18 million full-text articles from more than 15,000 scientific journals.

Your journals are on DeepDyve

Read from thousands of the leading scholarly journals from SpringerNature, Elsevier, Wiley-Blackwell, Oxford University Press and more.

All the latest content is available, no embargo periods.

See the journals in your area

DeepDyve

Freelancer

DeepDyve

Pro

Price

FREE

$49/month
$360/year

Save searches from
Google Scholar,
PubMed

Create lists to
organize your research

Export lists, citations

Read DeepDyve articles

Abstract access only

Unlimited access to over
18 million full-text articles

Print

20 pages / month

PDF Discount

20% off