Dopant-Dependent Toxicity of CeO2 Nanoparticles Is Associated with Dynamic Changes in H3K4me3 and H3K27me3 and Transcriptional Activation of NRF2 Gene in HaCaT Human Keratinocytes

Despite advances in the preparation of metal oxide (MO) nanoparticles (NPs) as catalysts for various applications, concerns about the biosafety of these particles remain. In this study, we prepared transition metal-doped cerium oxide (TM@CeO2; TM = Cr, Mn, Fe, Co, or Ni) nanoparticles and investigated the mechanism underlying dopant-dependent toxicity in HaCaT human keratinocytes. We show that doping with Cr or Co but not Fe, Mn, or Ni increased the toxicity of CeO2 NPs in dose- and time-dependent manners and led to apoptotic cell death. Interestingly, while both undoped and transition metal-doped NPs increased intracellular reactive oxygen species (ROS), toxic Cr@CeO2 and Co@CeO2 NPs failed to induce the expression of NRF2 (nuclear factor erythroid 2-related factor 2) as well as its downstream target genes involved in the antioxidant defense system. Moreover, activation of NRF2 transcription was correlated with dynamic changes in H3K4me3 and H3K27me3 at the promoter of NRF2, which was not observed in cells exposed to Cr@CeO2 NPs. Furthermore, exposure to relatively non-toxic Fe@CeO2 NPs, but not the toxic Cr@CeO2 NPs, resulted in increased binding of MLL1 complex, a major histone lysine methylase mediating trimethylation of histone H3 lysine 4, at the NRF2 promoter. Taken together, our findings strongly suggest that failure of cells to respond to oxidative stress is critical for dopant-dependent toxicity of CeO2 NPs and emphasize that careful evaluation of newly developed NPs should be preceded before industrial or biomedical applications.


Introduction
Metal oxide nanoparticles (MONPs) have been used for various chemical and biological applications, for example, as chemical sensors, biosensors, drug delivery agents, and for cancer therapy and in electrochemical reactions, due to their unique physicochemical properties [1][2][3][4]. MONPs are produced and consumed in large quantities, and the breadths of their applications are rapidly expanding. However, concerns have been expressed regarding their adverse effects on human health and the environment, as MONPs could enter the human body through ingestion, infection, inhalation, or skin contact [5][6][7][8][9]. The toxicities of MONPs depend on particle size and surface area, dosage, exposure time, pH, and extent of agglomeration [7,[10][11][12][13][14]. In vitro and in vivo studies have suggested that induction of reactive oxygen species (ROS) by MONPs predominantly underlies their toxicities by causing oxidative stress and inflammation, leading to intracellular component damage and aberrant expressions of genes associated with cellular homeostasis [7,15]. In addition, changes in epigenetic modification, such as DNA methylation and histone modification, have recently been suggested as alternative mechanisms of MONPs-mediated toxicity [16]. However, the effects of MONPs on histone modification, especially at the ROS-related genes, and the effects of histone modifications on MONPs-mediated toxicity are not fully understood.
NRF2 (nuclear factor erythroid 2-related factor 2) is a transcription factor that controls the cellular antioxidant defense system [40]. Its function is mainly regulated at the posttranscriptional level. Upon the oxidative stimuli, NRF2 is freed from KEAP1 (Kelch like ECH associated protein 1), a negative regulator of NRF2, and enters the nucleus, where it activates an array of antioxidative metabolizing/detoxifying genes by binding to ATE (antioxidant response element) [41,42]. NRF2 is also regulated at the transcriptional level. Studies have shown that transcription factors, such as AhR, NF-kB, and even NRF2 itself, regulate the expression of NRF2 [43][44][45]. In addition, epigenetic modifications, such as DNA methylation and histone methylation, have recently been reported to be key regulators of NRF2 [46].
The effects of CeO 2 NPs on NRF2-KEAP1 signaling have been reported in several studies, but results are not conclusive [38,[47][48][49][50]. It has been shown that exposure to CeO 2 NPs induces oxidative stresses, increases nuclear NRF2 level, and eventually causes cell death [38]. However, it has also been reported CeO 2 NPs have protective effects due to the transcriptional and posttranscriptional activation of NRF2 signaling [47,48], and yet others have reported CeO 2 NPs exposure resulted in no significant change or even a reduction in NRF2 level [49,50]. Moreover, the effect of CeO 2 NPs on the epigenetic modification of the NRF2 gene has not been studied in detail. In this study, we synthesized five different TM@CeO 2 NPs (where TM = Cr, Mn, Fe, Co, or Ni) and investigated their effects on HaCaT human keratinocytes and the mechanism responsible for dopant-dependent toxicity. Our comparative analysis provides evidence that transcriptional activation of the NRF2 gene and dynamic changes in H3K4me3 and H3K27me3 histone modifications play a critical role in dopant-dependent toxicity of TM@CeO 2 NPs.

Effects of Transition Metal Doping on Cell Viability
To investigate the effects of transition metal doping on the toxicity of CeO 2 NPs, we first analyzed the crystal structure of TM@CeO 2 NPs by X-ray diffraction (XRD) and transmission electron microscopy (TEM). The XRD pattern of CeO 2 NPs was typical of fluorite structured CeO 2 without any obvious structural changes. All tested TM@CeO 2 NPs generated XRD spectra with peaks at 2θ = 28.7 • , 33.2 • , 47.7 • , 56.5 • , 59.2 • , 69.5 • , 76.9 • , and 79.3 • (Figure 1a), which corresponded to the reflections from the (111), (200), (220), (311), (222), (400), (331), and (420) planes of undoped CeO 2 NPs (JCPDS card No. . TEM images demonstrated undoped CeO 2 and TM@CeO 2 NPs had similar sizes (~20 nm) and shapes (Figure 1a, inset). In addition, the c axis lattice constants of TM@CeO 2 NPs were almost the same as that of undoped CeO 2 NPs (Figure 1b). These observations suggest that transition metal doping is unlikely to cause significant changes in the surface structures of CeO 2 NPs. We next conducted MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) and NRU (neutral red uptake) assays to assess the effects of transition metal doping on cell viabilities using three different cell lines, that is, HaCaT human keratinocytes, HEK293T cells (a human embryonic kidney cell line), and C3H10T1/2 mouse mesenchymal stem cells, respectively ( Figure 2 and Figure S1). Consistent with previous studies, which showed CeO 2 NPs were relatively non-toxic [51,52], the viability of HaCaT cells fed with undoped CeO 2 NPs was comparable with that of untreated control cells even at a concentration of 625 µg/mL for up to 72 h ( Figure 2). Furthermore, no significant decrease in viability was observed in cells treated with Mn-, Fe-, or Ni-doped CeO 2 NPs for 24 and 72 h ( Figure 2). In contrast, Co@CeO 2 and Cr@CeO 2 NPs exhibited significant toxicities ( Figure 2). While exposure to Co@CeO 2 NPs for 24 h had no significant effect on cell viabilities even at the highest concentration used (625 µg/mL) (Figure 2a,c), exposure to Co@CeO 2 NPs at 625 µg/mL for 72 h reduced cell viability by~30% ( Figure  2b,d). Notably, exposure to Cr@CeO 2 NPs caused a dose-and time-dependent decrease in viability ( Figure 2). In HaCaT cells, exposure for 24 h resulted in modest but meaningful reductions (~7% at 125 µg/mL and~15% at 625 µg/mL) and exposure for 72 h caused a further decreased the viability of HaCaT cells (~30% at 125 µg/mL and by >80% at 625 µg/mL) (Figure 2b,d). The viabilities of HEK293T cells were not significantly affected by exposure to relatively non-toxic NPs, but similar reductions were observed after exposure to Cr-or Co-doped NPs ( Figure S1a-d). Interestingly, Co@CeO 2 NPs, which showed modest but significant toxicity in both HaCaT and HEK293T cells, had no significant effect on the viability of C3H10T1/2 mouse mesenchymal stem cells, and only cells exposed to 625 µg/mL of Cr@CeO 2 NPs for 72 h showed a reduction in viability of~20%. These results indicated responsiveness to TM@CeO 2 NPs is cell-type dependent (Figure S1e-h). We next investigated whether differences in intracellular localization and cellular uptake efficiency predominantly determined dopant-dependent toxicity ( Figure S2). Both relatively non-toxic Fe@CeO 2 and toxic Cr@CeO 2 NPs were readily internalized and localized in the perinuclear region of HaCaT cells ( Figure S2a). Moreover, fluorescencebased cellular uptake assays revealed that uptake efficiencies of toxic Cr@CeO 2 NPs were no higher than those of Fe@CeO 2 NPs at 5~625 µg/mL after exposure up to 24 h ( Figure  S2b). Taken together, these data suggest that transition metal doping can affect the intrinsic toxicity of CeO 2 NPs, and that doping with Cr or Co, dose-and time-dependently increases CeO 2 nanoparticle toxicity.

Dopant-Dependent Toxicities of TM@CeO 2 NPs Were Associated with Apoptotic Cell Death in HaCaT Cells
We next investigated whether the observed decreases in cell viability were associated with apoptotic cell death ( Figure 3). HaCaT cells were used for the in vitro analysis because they are derived from normal adult skin cells, and skin is one of the primary tissues affected by NPs. In addition, we used NPs at 125 µg/mL as both toxic and non-toxic NPs resulted in comparable viabilities at this concentration after 24 h but differences in viability after incubation for 72 h ( Figure 2 and Figure S1). Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays revealed extensive and prolonged DNA fragmentation in cells treated with toxic Cr@CeO 2 NPs but lesser effects in cells exposed to Co@CeO 2 NPs, no significant fragmentation in cells fed with undoped CeO 2 or relatively non-toxic TM@CeO 2 NPs (TM = Mn, Fe, and Ni) (Figure 3a). RT-qPCR analysis confirmed increased expressions of pro-apoptotic sensor genes BID (BH3 interacting domain death agonist) and BAD (BCL2 associated agonist of cell death) and the pro-apoptotic effector gene BAX (BCL2 associated X) in cells exposed to 125 µg/mL of Cr@CeO 2 or Co@CeO 2 NPs after 72 h ( Figure 3b). Conversely, mRNA levels of the anti-apoptotic genes BCL-2 (B-cell CLL/lymphoma 2), BCL-XL (BCL2 like 1), and MCL-1 (Myeloid cell leukemia sequence 1) were markedly decreased in cells treated with Cr-or Co-doped CeO 2 NPs (Figure 3b). Notably, while no significant changes in pro-and anti-apoptotic gene expressions were observed in HaCaT cells exposed to undoped CeO 2 or relatively non-toxic TM@CeO 2 NPs after 24 h, prolonged exposure (72 h) resulted in modest but meaningful increases in pro-apoptotic gene expressions in cells (Figure 3b). These data indicate that decreases in cell viability by toxic TM@CeO 2 NPs (TM = Cr or Co) are at least in part due to apoptotic cell death. The qPCR data shown are representative of at least three independent experiments and are presented as mean ± SD (n = 3~5). * p < 0.05, ** p < 0.01.

Effect of Transition Metal Doping on Intracellular ROS Generation
Since exposure to MONPs often causes oxidative stress, such as intracellular ROS generation, and these stresses are believed to be major factors of NP toxicity, we next examined the effect of transition metal doping on intracellular ROS generation ( Figure 4). Surprisingly, we found that HaCaT cells exposed to NPs generated more ROS than un-treated cells regardless of toxicity ( Figure 4). However, levels of ROS measured in cells exposed to toxic TM@CeO 2 NPs (TM = Cr, Co) were significantly greater than levels in undoped CeO 2 NPs, whereas exposure to relative non-toxic TM@CeO 2 NPs (TM = Fe, Mn, Ni) resulted in the ROS level similar to those observed in undoped NPs (Figure 4a, b). Levels of intracellular ROS appeared to decrease after 72 h, but HaCaT cells treated with toxic TM@CeO 2 NPs (TM = Cr or Co) maintained higher ROS levels than those treated with relatively non-toxic NPs (Figure 4b). Considering that all tested NPs increased intracellular ROS generation but only Cr-and Cr-doped CeO 2 NPs showed discernible cytotoxicity, these results suggest that either ROS level or the ability of cells to respond to ROS more critically determine NPs-mediated toxicity than oxidative stress itself.

Dopant-Dependent Toxicity Was Associated with a Failure of Cells to Activate NRF2 Expression
Because oxidative stresses induced by reactive oxidants are mainly countered by the NRF2-KEAP1 signaling pathway (a major antioxidant defense system), we investigated whether ROS increases by CeO 2 or TM@CeO 2 NPs led to the activation of this pathway ( Figure 5). RT-qPCR analysis revealed increased expression of NRF2 and decreased expression of KEAP1 (a negative regulator of NRF2) in HaCaT cells treated with undoped CeO 2 NPs and similar results in cells treated with relatively non-toxic TM@CeO 2 NPs (TM = Mn, Fe, or Ni) (Figure 5a). Surprisingly, no significant increase in NRF2 mRNA level and decrease in KEAP1 mRNA level was observed in cells exposed to toxic TM@CeO 2 NPs (TM = Cr or Co) despite elevated intracellular ROS levels (Figures 4 and 5a). Immunoblot analysis confirmed increased NRF2 levels in nuclear and cytosolic fractions and decreased KEAP1 levels after exposing cells to relatively non-toxic CeO 2 or Fe@CeO 2 NPs, but not in cells exposed to toxic Cr@CeO 2 NPs (Figure 5b and Figure S3). Next, we examined the expression of downstream target genes of NRF2, which include CAT (catalase), SOD1 (superoxide dismutase 1, cytosol), SOD2 (superoxide dismutase 2, mitochondria), HO-1 (heme oxygenase 1), and NQO1 (NAD(P)H quinone dehydrogenase 1) (Figure 5c). As was expected, the expression of NRF2 target genes was markedly increased in HaCaT cells exposed to relatively non-toxic NPs but not in cells exposed to toxic TM@CeO 2 NPs (Figure 5c). These observations suggest that intracellular ROS increases induced by relatively non-toxic NPs can be countered in cells, at least in part, by activation of the antioxidant defense system mediated by NRF2, and that the failure of cells to cope with elevated ROS levels underlies the dopant-dependent toxicity of CeO 2 NPs. (a) RT-qPCR analysis of NRF2(nuclear factor erythroid 2-related factor 2) and KEAP1 (kelch like ECH associated protein 1) genes in non-treated controls and NPs-treated cells. The total RNAs were isolated from untreated control, CeO 2 NPs-treated, and TM@CeO 2 NPs-treated cells after the indicated treatment time and relative mRNA levels were measured by RT-qPCR. (b) Immunoblot analysis of NRF2 and KEAP1 before and after NPs exposure. Nuclear and cytosolic extracts were prepared from cells treated or not with NPs for 24 h and subjected to immunoblot analysis to detect NRF2, KEAP1, Lamin A/C, and Tubulin. Lamin A/C and Tubulin were used as controls for nuclear and cytosolic fractions, respectively. (c) RT-qPCR analysis of target genes of NRF2 in control and NPs-treated cells. Relative mRNA levels of CAT (catalase), SOD1 (superoxide dismutase 1, cytosol), SOD2 (superoxide dismutase 2, mitochondria), HO-1 (heme oxygenase 1), and NQO1 (NAD(P)H quinone dehydrogenase 1) were measured using cDNA prepared from the same cells used in (a). The mRNA levels of indicated genes (a,c) were normalized to mRNA level of GAPDH, and data are presented as ratios of mRNA levels in NP-treated cells to those in untreated cells at each time point (24 and 72 h). The qPCR results are representative of at least three independent experiments and presented as mean ± SD (n = 3~5). * p < 0.05, ** p < 0.01.

Dopant Dependent Toxicity Was Associated with H3K4me3 and H3K27me3 Modification at NRF2 Promoter
Since lysine methylation of core histones is known to be involved in both activation and repression of genes depending on the site and status of modification [53], we next investigated whether the failure of NRF2 expression following exposure to toxic TM@CeO 2 NPs was associated with changes in histone lysine methylation (Figure 6a, b). Chromatin immunoprecipitation (ChIP) assays revealed that the exposure of HaCaT cells to undoped CeO 2 or Fe@CeO 2 NPs resulted in significant increases in the trimethylation of histone H3 lysine 4 (H3K4me3) and a discernible decrease in the trimethylation of histone H3 lysine 27 (H3K27me3) at NRF2 promoter (Figure 6a, upper right and lower left panel). However, exposure to toxic Cr@CeO 2 NPs had little effect on H3K4me3 or H3K27me3 at the promoter (Figure 6a). Interestingly, trimethylation of histone H3 lysine9, which also marks repressed gene expression, was not affected by NPs exposure (Figure 6a, lower right panel). Because levels of histone methylation are determined by methylation and demethylation, we conducted a time course chromatin immunoprecipitation analysis to confirm that exposure to Cr@CeO 2 NPs did not promote H3K4me3 demethylation. As shown in Figure 6b, H3K4me3 level at the promoter of NRF2 gradually increased after exposure to undoped CeO 2 or Fe@CeO 2 NPs for up to 24 h, but no discernible change in H3K4me3 level was detected after treatment with Cr@CeO 2 NPs for the same time. Finally, we examined the binding of the MLL1 (mixed-lineage leukemia 1) complex (a major histone lysine methylase for H3K4 trimethylation) at NRF2 gene. As was expected, exposure to relatively non-toxic NPs but not to toxic Cr@CeO 2 NPs increased bindings of MLL1 and ASH2L (a key component of MLL1 complex) at NRF2 promoter (Figure 6c). Taken together, these data strongly suggest that oxidative stresses induced by CeO 2 and relatively non-toxic TM@CeO 2 NPs can be countered by transcriptional activation of NRF2 via dynamic changes in H3K4me3and H3K27me3, and that failure of NRF2 activation is an underlying cause of the dopant-dependent toxicity of TM@CeO 2 NPs.
In conclusion, our current study shows that TM@CeO 2 NPs could exhibit dopantdependent toxicity. Cr was the most toxic dopant among the transition metal tested, and Fe, Mn, or Ni appeared to have no significant effect on the intrinsic toxicity of CeO 2 NPs. In particular, our data support the idea that activation of NRF2 signaling pathway rather than oxidative stress per se critically determines NPs-mediated toxicity, as all tested CeO 2 NPs elevated intracellular ROS levels but only the relatively non-toxic NPs induced intracellular antioxidant defense mechanism at least in part by activating NRF2 expression. In addition, our observations of dynamic changes in H3K4me3 and H3K27me3 histone modifications and increased binding of MLL1 complex at the NRF2 promoter following NPs exposure suggest MLL1 complex participates in the regulation of NRF2 expression, which we hope provides new insights into the molecular mechanism responsible for activating NRF2 dependent antioxidant defense system. Lastly, it should be noted that despite the observed relatively non-toxic natures of undoped CeO 2 and Fe-, Mn-, and Ni-doped CeO 2 NPs, the safety of these NPs with respect to long-term exposure remains undetermined, and thus, the study emphasizes the importance of carefully evaluating engineered NPs for biological safety before they are adopted for industrial and biomedical purposes. Chromatins prepared from the cells before (0 h) and after (24 h) NPs exposure were precipitated with α-H3K4me3, α-H3K9me3, or α-H3K27me3 antibodies. qPCR analysis was performed to assess the enrichment of modified histones at the promoter and distal regions of the NRF2 gene. (b) Time course ChIP analysis for H3K4 trimethylation induced by NPs. Chromatins were prepared from HaCaT cells exposed to NPs for the indicated times and precipitated with α-H3K4me3 antibodies. (c) Binding of MLL (mixed-lineage leukemia) complex at NRF2 promoter increased after non-toxic NPs treatment but not after treatment with toxic Cr@CeO 2 NPs. Chromatins were prepared as described in (a) and precipitated with α-MLL1 (left) or α-ASH2L (right) antibodies. For each chromatin, ChIP using IgG was performed to check chromatin quality. qPCR analyses shown in (b,c) were performed as in (a). For the relative ChIP signal, the % input (indicated antibody) was calculated for all samples, and data are presented as ratios of % input (indicated antibody) in NP-treated cells to those in untreated control cells. qPCR data are representative of at least three independent experiments and are presented as mean ± SD (n = 3~5). * p < 0.05, ** p < 0.01.

Preparation of Transition Metal-Doped CeO 2 NPs
CeO 2 NPs were synthesized using a modified thermal method [25,26]. For transition metal doping, precursor solutions were prepared by one-pot synthesis. The desired amount (1 mol%) of each TM dopant (Cr, Mn, Fe, Co, and Ni) in the form of TM(NO 3 ) 3 ·9H 2 O (99% purity) was added to each synthetic gel solution with stirring until the solution became homogeneous and transparent. The solution was then transferred to a Teflonlined autoclave and heated at 220 • C for 10 h in a convection oven. The resulting CeO 2 and TM@CeO 2 NPs were filtered and washed with deuterium-depleted water (DDW) to remove residues. All substances used for doping were purchased from Sigma-Aldrich (Sigma, St Louis, MO, USA).

Characterization of TM@CeO 2 NPs
The structures of fabricated CeO 2 NPs and the five TM@CeO 2 NPs were analyzed by using a JEM-3010 high-resolution transmission electron microscopy (HR-TEM, JEOL, Tokyo, Japan) at 300 kV and X-ray diffraction (XRD) patterns were obtained using Ni-filtered Cu-Kα radiation from a D8 Advance diffractometer (Bruker, Karlsruhe, Germany).

Cell Culture and NPs Exposure
HaCaT human keratinocytes were kindly provided by Dr. S. Kwon (Inha University, Korea). Cells were maintained in Dulbecco's modified Eagle's medium (DMEM, Wel-GENE, Gyeongsan, Korea) supplemented with 10% fetal bovine serum (FBS, WelGENE, Gyeongsan, Korea) and 1% penicillin-streptomycin (GE Healthcare, Madison, WI, USA) in a humidified atmosphere with 5% CO 2 at 37 • C. For NPs exposure, 10 mg/mL of TM@CeO 2 NPs in DMEM supplemented with 10% FBS were prepared using a vortex mixer to prevent aggregation and then added to culture plates at the final concentrations of 5, 25, 125, or 625 µg/mL.

Cell Viability Assays
The effects of TM@CeO 2 NPs on cell viability were assessed by MTT and NRU assays, as previously described [54,55]. Briefly, HaCaT, HEK293T, and C3H10T1/2 cells were seeded at 2 × 10 4 cells per well in 96-well cell culture plates and cultured for 24 h. Cells were then exposed to undoped CeO 2 or TM@CeO 2 NPs for 24 or 72 h. For MTT assays, cells were washed twice with phosphate-buffered saline (PBS, GIBCO, Grand Island, NY, USA), and then MTT solution (Sigma, St Louis, MO, USA) was added to each well to a final concentration of 0.5 mg/mL. One hour later, formazan crystals that formed were dissolved in 50% dimethyl sulfoxide (

TUNEL Assay
Terminal deoxynucleotidyl transferase dUTP nick end labeling (TUNEL) assays were performed using the In Situ Cell Death Detection Kit, Fluorescein (Roche, Mannheim, Germany). HaCaT cells were seeded at 2 × 10 4 cells per well in a 96-well cell culture plate and 24 h later, exposed to 125 µg/mL of undoped CeO 2 or TM@CeO 2 NPs for 24 or 72 h. Following fixation with 4% paraformaldehyde (EMS, Hatfield, PA, USA) for 30 min and permeabilization with 0.1% Triton X-100 (Merck, Darmstadt, Germany) in PBS for 10 min, cells were incubated with TUNEL reaction solution for 1 h at 37 • C in the dark and stained with Hoechst 33342 (Invitrogen, Carlsbad, CA, USA) for 5 min. Images were taken at 20× magnification using an Olympus IX71 inverted microscope equipped with a U-RFL-T mercury lamp (Olympus, Tokyo, Japan). Cells treated with 1,000 units/mL of DNase I (Promega, Madison, WI, USA) for 10 min were used as positive controls.

Measurement of Intracellular Reactive Oxygen Species (ROS) Levels
Intracellular ROS levels were measured using dichlorofluorescein diacetate oxidation, as previously described [56]. Cells were seeded at 2 × 10 5 cells per well in 6-well plates, incubated for 24 h, and exposed to 125 µg/mL of undoped CeO 2 or TM@CeO 2 NPs for 24 or 72 h. Following exposure, cells were washed with PBS and incubated with 50 µM of 2 ,7 -dichlorofluorescin diacetate (DCFDA, Invitrogen, Carlsbad, CA, USA) for 30 min. Images were taken using an Olympus IX71 inverted microscope equipped with a U-RFL-T mercury lamp at excitation wavelengths of 488 nm and processed using Adobe Photoshop CC2018 software (Adobe Systems, San Jose, CA, USA). To quantify ROS levels, fluorescence intensities were measured using a Synergy HTX multi-mode microplate reader (Bio-Tek

Cell Fractionation and Immunoblot Analysis
HaCaT cells were washed with PBS and lysed with hypotonic buffer (10 mM HEPES pH 7.9, 10 mM KCl, 0.1 mM ethylenediaminetetraacetic acid (EDTA) pH 8.0, and 0.3% NP-40) in the presence of a protease inhibitor cocktail (Roche, Mannheim, Germany). Lysates obtained were passed through a 26G1/2 needle 10 times, incubated on ice for 10 min, and then centrifuged at 5,000× g for 10 min. Supernatants were used as a cytosolic fraction, and nuclear fractions were prepared by suspending pellets in radioimmunoprecipitation assay (RIPA) buffer (10 mM Tris-HCl, pH 7.5, 1 mM EDTA, pH 8.0, 150 mM NaCl, 1% Triton X-100, 1% sodium deoxycholate, and 0.1% sodium dodecyl sulfate (SDS)), sonicating at 40% amplitude for 5 × 30 s using a VCX130 sonicator (Sonics, Newtown, CT, USA), and then centrifugation at 13,000× g for 20 min. Immunoblot analysis was performed using a standard protocol. Detailed information regarding antibodies and working concentrations is provided in Supplementary Material (Table S1).

Quantitative RT-PCR Analysis (RT-qPCR)
For RT-qPCR analysis, total RNA was isolated using an RNeasy plus mini kit (QIA-GEN, Hilden, Germany), and cDNA was synthesized using a GoScript reverse transcription system (Promega, Madison, WI, USA), according to the manufacturer's instructions. Quantitative PCR was conducted using a QuantStudio 1 Real-Time PCR system (ThermoFisher Scientific, Waltham, MA, USA) using SYBR Green I (Invitrogen, Carlsbad, CA, USA) and i-StarTaq DNA polymerase (Intron, Sungnam, Korea). mRNA levels were normalized to GAPDH mRNA, and data are presented as indicated in Figures 3 and 5. Primer sets used are detailed in Supplementary Materials (Table S2).

Chromatin Immunoprecipitation (ChIP)-qPCR Analysis
ChIP assays were performed as previously described [57]. Briefly, 100~300 µg of sonicated chromatins were precleared for 2 h using protein A/G sepharose 4 Fast Flow (GE Healthcare, Madison, WI, USA) in the presence of 4 mg/mL salmon sperm DNA (Invitrogen, Carlsbad, CA, USA) and 0.5 mg/mL bovine serum albumin (Sigma, St Louis, MO, USA) and then subjected to immunoprecipitation using appropriate antibodies. Purified DNA obtained was analyzed by quantitative PCR (qPCR) using a QuantStudio 1 Real-Time PCR system. For quantification, the % input value per sample was calculated, and the data are presented as relative ChIP signals as indicated in Figure 6. The antibodies and primers used for ChIP-qPCR analysis are listed in Tables S1 and S3.

Statistical Analysis
Results of cell viability assays and all qPCR-based experiments are representative of at least three independent experiments (as indicated in the figure legends) and are presented as the means ± SDs. Statistical significance and p-values were determined by two-tailed t-tests of the indicated paired groups using Microsoft Excel (version 2102, Microsoft, Redmond, WA, USA). Differences were considered significant when p-values were < 0.05.

Supplementary Materials:
The following are available online at https://www.mdpi.com/1422-0 067/22/6/3087/s1, Figure S1: Effect of transition metal doping on the viabilities of HEK293T and C3H10T1/2 cells. Figure S2: Intracellular localization and cellular uptake efficiency of toxic Cr@CeO 2 NPs and relatively non-toxic Fe@CeO 2 NPs. Figure S3: Exposure to relatively non-toxic undoped or Fe-doped CeO 2 NPs led to increased NRF2 and decreased KEAP1 in HaCaT cells. Table S1: Information on the antibodies used in this study, Table S2: Information on the primers used for RT-qPCR, Table S3: Information on the primers used for ChIP-qPCR, References.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author.