MRC-5 Human Lung Fibroblasts Alleviate the Genotoxic Effect of Fe-N Co-Doped Titanium Dioxide Nanoparticles through an OGG1/2-Dependent Reparatory Mechanism

The current study was focused on the potential of pure P25 TiO2 nanoparticles (NPs) and Fe(1%)-N co-doped P25 TiO2 NPs to induce cyto- and genotoxic effects in MRC-5 human pulmonary fibroblasts. The oxidative lesions of P25 NPs were reflected in the amount of 8-hydroxydeoxyguanosine accumulated in DNA and the lysosomal damage produced, but iron-doping partially suppressed these effects. However, neither P25 nor Fe(1%)-N co-doped P25 NPs had such a serious effect of inducing DNA fragmentation or activating apoptosis signaling. Moreover, oxo-guanine glycosylase 1/2, a key enzyme of the base excision repair mechanism, was overexpressed in response to the oxidative DNA deterioration induced by P25 and P25-Fe(1%)-N NPs.


Introduction
In recent years, NPs have become a major development opportunity for biomedical [1], agricultural [2], and different industrial applications such as electronics [3], aerospace and automotive coatings [4], active food packaging [5], or environmental remediation [6]. According to some estimates, the most produced metal-based nanomaterials worldwide are made of silver, titanium, zinc, or gold [7,8]. In particular, TiO 2 NPs are widely used in commercially available products or pilot applications, including cosmetics and sunscreens, paints, food products and active packaging, photoactive cement, or innovative textiles [9][10][11][12], due to their special optical and photocatalytic characteristics.
Different studies proved that TiO 2 NPs, especially anatase-rutile mixtures, exhibit cytoand genotoxicity [13][14][15], which probably derive from their ability to generate excessive ROS levels that could determine oxidative stress when the cellular antioxidant systems are overwhelmed. The toxicological response of living systems to TiO 2 NPs depends on the particles' physicochemical properties that can be modulated by various procedures [16,17]. TiO 2 is present in several stoichiometric (anatase, rutile, brookite) [18,19] and non-stoichiometric crystalline structures [20,21], the toxicity of the first ones being exhaustively investigated over time. Anatase and rutile are the main crystalline forms of TiO 2 . Evidence shows that their toxic effects might differ when nano-dimensioned particles under ultraviolet or visible irradiation are considered [22,23]. Concerning the non-stoichiometric forms, results on their toxicological effects are nearly inexistent. To the best of our knowledge, only one study investigated the biological activity of TiO 2−x , reporting eryptosis via ROS and Ca 2+ signaling [24].
Within the nano-size range, smaller NPs are generally considered more cytotoxic because of their large specific surface area, which probably leads to the formation of higher response of MRC-5 human pulmonary fibroblasts exposed to TiO 2 P25 NPs and Fe(1%)-N doped TiO 2 P25 ones, respectively. Also, lysosomal modifications and DNA integrity were investigated in relation to the oxidative lesions induced by the tested NPs.

Physicochemical Characteristics of TiO 2 NPs
The two types of TiO 2 NPs that were used in the present work were (i) commercially available P25 NPs and (ii) the same NPs co-doped with Fe and N atoms by direct impregnation in an aqueous solution of 1% FeCl 3 and in the presence of urea (P25-Fe(1%)-N) (see Section 4.1.). The concentration of FeCl 3 that we chose was mainly based on a previous paper from our group where P25 NPs impregnated by dispersion in 1% FeCl 3 had an enhanced photocatalytic effect in both long-wave UV (368 nm) and visible light (610 nm) [70]. Moreover, Kalantari et al. showed that co-doping with Fe and N considerably increased the TiO 2 NPs' photocatalytic activity compared to mono-doped TiO 2 NPs [71].
There is evidence that a higher amount of Fe atoms on TiO 2 NPs' surface would improve the photocatalytic effect of NPs [72][73][74]. However, it was already shown that the phototoxicity of TiO 2 NPs could be proportionally increased by 1% to 10% Fe-doping due to the generation of oxidizing agents via the Fenton reaction [74]. Therefore, we considered a low amount of dopant would be more appropriate for investigating the potentially toxic effects of pulmonary exposure to Fe-N doped TiO 2 NPs.
The chemical content of P25 NPs and P25-Fe(1%)-N NPs was revealed by X-ray photoelectron spectroscopy (XPS) measurements ( Figure 1a) and the corresponding binding energies ( Table 1). The results proved the presence of Ti and O atoms in both types of NPs (Figure 1b,c). Also, Fe and N atoms were identified in the P25-Fe(1%)-N sample (Figure 1d,e). Moreover, P25 NPs were made of stoichiometrically structured TiO 2, as revealed by the ratio of 2.02 between the main signals of Ti 2p 3/2 (458.65 eV) and O 1s (529.98 eV). The O 1s peaks near 532 eV might be assigned to hydroxyl groups or adsorbed water molecules on the surface, and the 530-531 eV peaks to Ti-O chemical bonds, respectively. The signal at 710.40 eV is characteristic of Fe 2p 3/2 , revealing that P25-Fe(1%)-N NPs contained Fe 3+ . The peak at 399.62 eV might be assigned to oxidized nitrogen, i.e., O-Ti-N bindings. Also, the N 1s peak at 401.19 eV usually reflects interstitial nitrogen. The ratios between intensities of the XPS peaks (Fe/TiO and N/TiO) showed that P25 NPs prepared in FeCl 3 had 2.1% Fe atoms and 0.5% N atoms on their surface. As revealed in the images obtained by transmission electron microscopy (TEM), both 'types of NPs generally had near polyhedral shapes with round corners (Figure 2a,b); some spheres could also be observed. The dimension of most P25 NPs was between ∼10-50 nm and had a mean particle size of 29 nm (Figure 2a). The size range of P25-Fe(1%)-N NPs was larger, with most being ∼15-60 nm. However, the mean particle size of P25-Fe(1%)-N NPs was similar to that of P25 NPs, i.e., ∼28 nm ( Figure 2b). More analyses regarding the characterization of TiO 2 NPs were provided in our previously published papers [70,75].  `types of NPs generally had near polyhedral shapes with round corners (Figure 2a,b); some spheres could also be observed. The dimension of most P25 NPs was between ∼10-50 nm and had a mean particle size of 29 nm (Figure 2a). The size range of P25-Fe(1%)-N NPs was larger, with most being ∼15-60 nm. However, the mean particle size of P25-Fe(1%)-N NPs was similar to that of P25 NPs, i.e., ∼28 nm ( Figure 2b). More analyses regarding the characterization of TiO2 NPs were provided in our previously published papers [70,75].

Oxidative DNA Damage Induced by TiO2 NPs in MRC-5 Cells
The concentrations of NPs, i.e., 10 µg/mL and 50 µg/mL, respectively, used by us were chosen based on our previous work [76] in which we proved that P25 NPs could cause a significant increase of oxidative stress in MRC-5 cells in a time-and dose-dependent manner while P25-Fe(1%)-N NPs had no influence on ROS level compared to the control group of cells.
ROS can damage the cell considerably by impairing the constitutive molecules of cellular structures. One of the damages induced by a high level of ROS is the oxidation of guanosine, a modification that might affect the integrity of DNA molecules. We investigated the impact of TiO2 NPs on the DNA molecules of MRC-5 cells by measuring the level of 8-hydroxydeoxyguanosine (8-OHdG), a commonly used marker for DNA oxidative lesions. Our results showed that exposure to P25-Fe(1%)-N NPs could increase the level of 8-OHdG in a time-dependent manner ( Figure 3) in MRC-5 cells. The levels of 8-OHdG induced by both doses of P25-Fe(1%)-N NPs and the dose of 10 µg/mL non-doped P25 NPs were generally similar and have not exceeded 130% compared to control after 72 Figure 2. Observation of TiO 2 NPs size and morphology. Representative transmission electron microscopy images (above) and particle size distribution histograms (below) of (a) P25 NPs and (b) Fe(1%)-N doped P25 NPs. Scale bar: 50 nm.

Oxidative DNA Damage Induced by TiO 2 NPs in MRC-5 Cells
The concentrations of NPs, i.e., 10 µg/mL and 50 µg/mL, respectively, used by us were chosen based on our previous work [76] in which we proved that P25 NPs could cause a significant increase of oxidative stress in MRC-5 cells in a time-and dose-dependent manner while P25-Fe(1%)-N NPs had no influence on ROS level compared to the control group of cells.
ROS can damage the cell considerably by impairing the constitutive molecules of cellular structures. One of the damages induced by a high level of ROS is the oxidation of guanosine, a modification that might affect the integrity of DNA molecules. We investigated the impact of TiO 2 NPs on the DNA molecules of MRC-5 cells by measuring the level of 8-hydroxydeoxyguanosine (8-OHdG), a commonly used marker for DNA oxidative lesions. Our results showed that exposure to P25-Fe(1%)-N NPs could increase the level of 8-OHdG in a time-dependent manner ( Figure 3) in MRC-5 cells. The levels of 8-OHdG induced by both doses of P25-Fe(1%)-N NPs and the dose of 10 µg/mL non-doped P25 NPs were generally similar and have not exceeded 130% compared to control after 72 h of exposure. However, the higher dose of P25 NPs caused an increase of the level of 8-OHdG up to 235% compared to the control after MRC-5 cells were exposed for 24 h. Also, 8-OHdG concentration was lowered in the cells as time passed, reaching 166% compared to the control at 72 h. Interestingly, the reduction of 8-OHdG recorded at 72 h at 50 µg/mL P25 NPs contrasts with the high level of ROS measured previously by us in MRC-5 cells [76]. h of exposure. However, the higher dose of P25 NPs caused an increase of the level of 8-OHdG up to 235% compared to the control after MRC-5 cells were exposed for 24 h. Also, 8-OHdG concentration was lowered in the cells as time passed, reaching 166% compared to the control at 72 h. Interestingly, the reduction of 8-OHdG recorded at 72 h at 50 µg/mL P25 NPs contrasts with the high level of ROS measured previously by us in MRC-5 cells [76].

Influence of TiO2 NPs on the Morphology of MRC-5 Cells
Actin cytoskeleton plays a key role in the mechanical support of cells, also defining their morphology. Fluorescent microscopy images displayed in Figure 4 showed that TiO2 NPs had no negative impact on the MRC-5 cells' actin cytoskeleton organization. The microscopic images suggested that MRC-5 cells maintained their fibroblast-like morphology regardless of the conditions applied in our study (type of TiO2 NPs, concentration of NPs, exposure time). Normally, these lung fibroblasts are elongated spindle-shaped bipolar cells. No disrupted filaments or cytoskeleton rearrangements were observed, while bundles of F-actin appeared very dense, indicating a high cellular density.

Influence of TiO 2 NPs on the Morphology of MRC-5 Cells
Actin cytoskeleton plays a key role in the mechanical support of cells, also defining their morphology. Fluorescent microscopy images displayed in Figure 4 showed that TiO 2 NPs had no negative impact on the MRC-5 cells' actin cytoskeleton organization. The microscopic images suggested that MRC-5 cells maintained their fibroblast-like morphology regardless of the conditions applied in our study (type of TiO 2 NPs, concentration of NPs, exposure time). Normally, these lung fibroblasts are elongated spindle-shaped bipolar cells. No disrupted filaments or cytoskeleton rearrangements were observed, while bundles of F-actin appeared very dense, indicating a high cellular density.

Influence of TiO 2 NPs on Lysosomes' Formation and Lysosomal Membrane Integrity in MRC-5 Cells
There are no statistically significant differences regarding the accumulation of lysosomes inside MRC-5 cells exposed to P25 and P25-Fe(1%)-N NPs (Figure 5a,b). However, we noted that the lysosome quantity increased by ∼14-18% compared to the control when the doses of 50 µg/mL at 72 h were applied. The distribution of cathepsin B suggested that the membrane of lysosomes was affected by the 72 h exposure to TiO 2 NPs. Cathepsin B is a key proteolytic enzyme localized in lysosomes under physiological conditions. Therefore, when labeled with Alexa Fluor 594, cathepsin B is present in fluorescent red vesicles in healthy cells, as can be observed in our control cells (Figure 5c).

Influence of TiO2 NPs on Lysosomes' Formation and Lysosomal Membrane Integrity in MRC-5 Cells
There are no statistically significant differences regarding the accumulation of lysosomes inside MRC-5 cells exposed to P25 and P25-Fe(1%)-N NPs (Figure 5a,b). However, we noted that the lysosome quantity increased by ∼14-18% compared to the control when the doses of 50 µg/mL at 72 h were applied. The distribution of cathepsin B suggested that the membrane of lysosomes was affected by the 72 h exposure to TiO2 NPs. Cathepsin B is a key proteolytic enzyme localized in lysosomes under physiological conditions. Therefore, when labeled with Alexa Fluor 594, cathepsin B is present in fluorescent red vesicles in healthy cells, as can be observed in our control cells (Figure 5c).

Influence of TiO2 NPs on Lysosomes' Formation and Lysosomal Membrane Integrity in MRC-5 Cells
There are no statistically significant differences regarding the accumulation of lysosomes inside MRC-5 cells exposed to P25 and P25-Fe(1%)-N NPs (Figure 5a,b). However, we noted that the lysosome quantity increased by ∼14-18% compared to the control when the doses of 50 µg/mL at 72 h were applied. The distribution of cathepsin B suggested that the membrane of lysosomes was affected by the 72 h exposure to TiO2 NPs. Cathepsin B is a key proteolytic enzyme localized in lysosomes under physiological conditions. Therefore, when labeled with Alexa Fluor 594, cathepsin B is present in fluorescent red vesicles in healthy cells, as can be observed in our control cells ( Figure 5c). When MRC-5 cells were treated with P25 and P25-Fe(1%)-N NPs, the red signal appeared in a diffused pattern, indicating that the lysosomal membrane was permeabilized and cathepsin B was released into the cytosol. Even though permeabilization occurred in all treated pulmonary fibroblasts, images showed that the red signal is more clustered in cells exposed at P25-Fe(1%)-N NPs, suggesting their effect on lysosome integrity is less pronounced than the one of P25 NPs.

Effect of TiO 2 NPs on the Integrity of DNA from MRC-5 Cells
Considering the generation of ROS and oxidative lesions induced by TiO 2 NPs in MRC-5 cells, we further decided to investigate whether they critically affect the integrity of DNA molecules. Fragmentation of DNA was investigated by Comet assay that indicated no significant changes between the different conditions tested, although raised levels of DNA oxidation might be considered a marker of double-strand breaks. It can be visually observed that no small fragments of DNA detached and migrated faster, as in the case of the positive control (Figure 6a). The damage of DNA molecules was expressed in percentages of DNA in the comet tail. Based on the quantified fluorescence (Figure 6b), the tail DNA% in samples varied in the 2.4-4.6% range, while in the negative control cells, it has not exceeded 4%. These results could strengthen the evidence that a molecular mechanism ameliorates the TiO 2 NPs-dependent oxidative damage of DNA within MRC-5 fibroblasts.
cells treated with P25 NPs and Fe(1%)-N doped P25 NPs (10 and 50 µg/mL) at 72 h of exposure. White arrows indicate the vesicular disposition of cathepsin B in control cells. Nuclei (blue) were stained with DAPI. Scale bar: 20 µm.
When MRC-5 cells were treated with P25 and P25-Fe(1%)-N NPs, the red signal appeared in a diffused pattern, indicating that the lysosomal membrane was permeabilized and cathepsin B was released into the cytosol. Even though permeabilization occurred in all treated pulmonary fibroblasts, images showed that the red signal is more clustered in cells exposed at P25-Fe(1%)-N NPs, suggesting their effect on lysosome integrity is less pronounced than the one of P25 NPs.

Effect of TiO2 NPs on the Integrity of DNA from MRC-5 Cells
Considering the generation of ROS and oxidative lesions induced by TiO2 NPs in MRC-5 cells, we further decided to investigate whether they critically affect the integrity of DNA molecules. Fragmentation of DNA was investigated by Comet assay that indicated no significant changes between the different conditions tested, although raised levels of DNA oxidation might be considered a marker of double-strand breaks. It can be visually observed that no small fragments of DNA detached and migrated faster, as in the case of the positive control (Figure 6a). The damage of DNA molecules was expressed in percentages of DNA in the comet tail. Based on the quantified fluorescence (Figure 6b), the tail DNA% in samples varied in the 2.4-4.6% range, while in the negative control cells, it has not exceeded 4%. These results could strengthen the evidence that a molecular mechanism ameliorates the TiO2 NPs-dependent oxidative damage of DNA within MRC-5 fibroblasts.

Cell Death Signaling in MRC-5 Cells Exposed to TiO 2 NPs
To investigate whether the oxidative lesions produced by TiO 2 NPs generated damages that trigger cell death signaling, protein expression of cathepsin B, p53, caspase-8, -9, and -3 were quantified using Western Blot analyses. Cathepsin B presented relatively constant levels in MRC-5 cells exposed to P25 and P25-Fe(1%)-N NPs regardless of the exposure time or dose applied (Figure 7a,b).

Cell Death Signaling in MRC-5 Cells Exposed to TiO2 NPs
To investigate whether the oxidative lesions produced by TiO2 NPs generated damages that trigger cell death signaling, protein expression of cathepsin B, p53, caspase-8, -9, and -3 were quantified using Western Blot analyses. Cathepsin B presented relatively constant levels in MRC-5 cells exposed to P25 and P25-Fe(1%)-N NPs regardless of the exposure time or dose applied (Figure 7a,b). On the contrary, TiO2 NPs significantly changed the expression of p53 protein in MRC-5 cells in a time-and dose-dependent manner. After 24 h of exposure to 10 µg/mL of P25 and P25-Fe(1%)-N NPs, the level of p53 decreased by 3%, respectively, 12% relative to the control. The expression of p53 started to drop considerably when the highest dose of TiO2 NPs was applied. At 24 h of exposure, P25 NPs led to the diminution of p53 expression by 64%. By comparison, the effect of iron-doped TiO2 NPs was slightly milder, leading to a decreased expression by nearly 38% relative to the control. However, the results indicated that the inhibitory effect of TiO2 NPs on p53 expression was more evident as time went on. Thus, p53 expression in NP-treated MRC-5 cells exhibited a massive reduction regardless of the dose or exposure time. As can be seen in Figure 7c,d, the level of expression of p53 dropped below 10% relative to the control and was totally suppressed by the treatment with 10 µg/mL and 50 µg/mL of P25 NPs, respectively.
Both p53 and cathepsin B are involved in the initiation of programmed cell death pathways [77,78]. However, in contrast with our results, p53 normally undergoes overexpression during apoptosis [79]. In the present study, we proved that neither initiator caspases-8 and -9 nor the effector caspase-3 were activated by the TiO2 NPs applied to pulmonary fibroblasts. Based on the molecular mass, protein bands displayed on the obtained blot profiles corresponded to the uncleaved, i.e., non-activated procaspases ( Figure  7e). The measured level of the apoptosis-inducing markers, i.e., cathepsin B, p53, caspase-8, -9, and -3, correlated well with the high DNA integrity revealed by the Comet assay. On the contrary, TiO 2 NPs significantly changed the expression of p53 protein in MRC-5 cells in a time-and dose-dependent manner. After 24 h of exposure to 10 µg/mL of P25 and P25-Fe(1%)-N NPs, the level of p53 decreased by 3%, respectively, 12% relative to the control. The expression of p53 started to drop considerably when the highest dose of TiO 2 NPs was applied. At 24 h of exposure, P25 NPs led to the diminution of p53 expression by 64%. By comparison, the effect of iron-doped TiO 2 NPs was slightly milder, leading to a decreased expression by nearly 38% relative to the control. However, the results indicated that the inhibitory effect of TiO 2 NPs on p53 expression was more evident as time went on. Thus, p53 expression in NP-treated MRC-5 cells exhibited a massive reduction regardless of the dose or exposure time. As can be seen in Figure 7c,d, the level of expression of p53 dropped below 10% relative to the control and was totally suppressed by the treatment with 10 µg/mL and 50 µg/mL of P25 NPs, respectively.
Both p53 and cathepsin B are involved in the initiation of programmed cell death pathways [77,78]. However, in contrast with our results, p53 normally undergoes overexpression during apoptosis [79]. In the present study, we proved that neither initiator caspases-8 and -9 nor the effector caspase-3 were activated by the TiO 2 NPs applied to pulmonary fibroblasts. Based on the molecular mass, protein bands displayed on the obtained blot profiles corresponded to the uncleaved, i.e., non-activated procaspases (Figure 7e). The measured level of the apoptosis-inducing markers, i.e., cathepsin B, p53, caspase-8, -9, and -3, correlated well with the high DNA integrity revealed by the Comet assay.

The Reparatory Role of 8-oxoguanine DNA Glycosylase in MRC-5 Cells Exposed to TiO 2 NPs
As the results suggested so far, a molecular reparatory mechanism might be the reason for the low genotoxic effect of TiO 2 NPs on MRC-5 cells. Therefore, we decided to investigate a key enzyme involved in the base excision repair, which can recognize oxidized guanine within the DNA, namely OGG1/2. The protein expression of OGG1/2 was assessed by Western Blot analysis, with the representative blot profiles displayed in Figure 8a. Interestingly, our results showed that the level of OGG1/2 slightly decreased when MRC-5 cells were exposed for 24 h at both doses of P25 or P25-Fe(1%)-N NPs. The expression exhibited an insignificant diminution of at most 6% relative to the control. The effect of TiO 2 NPs on the level of OGG1/2 became evident at 72 h of exposure when the reparatory protein exhibited a substantial overexpression (Figure 8b). In general, P25 and P25-Fe(1%)-N NPs caused the doubling of OGG1/2 expression level, confirming that the innate base excision repair mechanism coped with the oxidative damage induced by TiO 2 NPs and thus maintained the integrity of DNA molecules. The level of OGG1/2 was slightly higher in the pulmonary fibroblasts treated with P25-Fe(1%)-N NPs in comparison with the one measured in cells exposed to undoped TiO 2 NPs. Differences in OGG1/2 expression were comprised between 83.5 and 125% relative to the control.

NPs
As the results suggested so far, a molecular reparatory mechanism might be the reason for the low genotoxic effect of TiO2 NPs on MRC-5 cells. Therefore, we decided to investigate a key enzyme involved in the base excision repair, which can recognize oxidized guanine within the DNA, namely OGG1/2. The protein expression of OGG1/2 was assessed by Western Blot analysis, with the representative blot profiles displayed in Figure  8a. Interestingly, our results showed that the level of OGG1/2 slightly decreased when MRC-5 cells were exposed for 24 h at both doses of P25 or P25-Fe(1%)-N NPs. The expression exhibited an insignificant diminution of at most 6% relative to the control. The effect of TiO2 NPs on the level of OGG1/2 became evident at 72 h of exposure when the reparatory protein exhibited a substantial overexpression (Figure 8b). In general, P25 and P25-Fe(1%)-N NPs caused the doubling of OGG1/2 expression level, confirming that the innate base excision repair mechanism coped with the oxidative damage induced by TiO2 NPs and thus maintained the integrity of DNA molecules. The level of OGG1/2 was slightly higher in the pulmonary fibroblasts treated with P25-Fe(1%)-N NPs in comparison with the one measured in cells exposed to undoped TiO2 NPs. Differences in OGG1/2 expression were comprised between 83.5 and 125% relative to the control.

Discussion
Considering iron doping might intensify the use of TiO 2 NPs in consumer goods, we chose to investigate the possible associated toxicological risks due to the tuning of their photocatalytic properties toward visible illuminance. Therefore, we compared the toxicity of P25 TiO 2 NPs doped with iron and nitrogen with the same undoped NPs. The doses used by us were based on our previously published work [76] as well as on representative papers [80][81][82][83].
Moreover, inhalation is a major route by which TiO 2 NPs enter the human body; therefore, we have chosen MRC-5 cells, which are human pulmonary fibroblasts, as the experimental model.
Previous studies stated that TiO 2 NPs could also interact with the microtubules and other components of the cellular cytoskeleton [84,85]. We decided to investigate the influence of P25 and P25-Fe(1%)-N NPs on the actin cytoskeleton as they could provide valuable information regarding the morphology of treated cells. Even though some reports show that TiO 2 NPs can disrupt actin filaments [86,87], we did not observe any significant changes in the organization of the cytoskeleton between exposed samples and control cells. In addition, considering the role of the actin cytoskeleton in internalization mechanisms [88], endocytosis of the P25 and P25-Fe(1%)-N NPs with sizes about 50 nm might be produced to a lesser extent in the MRC-5 cells. This fact is supported by the dimension of the large aggregates of TiO 2 NPs formed [75], which would not be able to enter the cells through caveolae (with a diameter between 50nm and 80 nm) or clathrin-mediated endocytosis (with a diameter of ≈120 nm) [89]. However, Thurn et al. [90] stated that the uptake of aggregates could be possible through macropinosomes with a dimension of 500-2000 nm.
In our previous work [76], we already showed the significant difference between the ability of P25 and P25-Fe(1%)-N NPs to produce ROS. While P25 NPs could induce high levels of ROS in a time and dose-dependent manner, doping with iron ions totally suppressed the generation of oxidative stress [76]. Similarly, iron doping inhibited the production of TiO 2 -induced ROS in HaCaT keratinocytes [64]. In contrast, doping TiO 2 NPs with copper led to higher ROS production in A549 cells [69], and doping them with zinc enhanced the oxidative stress induced in MCF-7 cells [68].
Moreover, our previous paper [76] investigated the effect of TiO 2 NPs on the enzymatic antioxidant mechanism of MRC-5 cells. When the ROS level exceeded the neutralizing ability of antioxidant enzymes, some of the free radicals began to impair intracellular biomolecules. A part of the oxygen-derived free radicals produced lipid peroxidation that attacks organelles' membranes, while others damage DNA after entering the nucleus as well as proteins [91].
Some of the most commonly studied biomarkers indicating oxidative damage on DNA molecules, are 8-hydroxylated guanine species, mainly 8-oxoguanine (8-oxoG) and its isomer, 8-OHdG. In our study, P25 NPs increased the level of 8-OHdG in MRC-5 cells in a time-dependent manner, the results being in accordance with the level of ROS produced. Interestingly, we observed an attenuated but significant increase of 8-OHdG level in pulmonary fibroblasts treated with P25-Fe(1%)-N NPs. This might be explained by the fact that iron-doped TiO 2 NPs could generate some reactive species in the first hours of exposure that had probably produced their effects before initiating the antioxidant mechanisms [76]. We considered this might represent preliminary evidence that human pulmonary fibroblasts are able to counteract excessive oxidation caused by TiO 2 NPs.
Some studies showed that the increased level of guanine oxidation products within cells might be linked in certain circumstances with DNA fragmentation [92,93]. On the contrary, our results indicated that the integrity of DNA molecules from MRC-5 cells was not affected by the increased level of 8-OHdG caused by exposure to TiO 2 NPs. Similarly, Hackenberg et al. [94] showed that TiO 2 NPs did not induce DNA fragmentation in lymphocytes obtained from the peripheral blood of human donors. Bhattacharya et al. [95] obtained the same result when they applied TiO 2 NPs on both BEAS-2B (normal human bronchial epithelial cells) and IMR-90 (normal human pulmonary fibroblasts) cell cultures, showing that IMR-90 cells exhibited high levels of 8-OHdG after 24 h of exposure to TiO 2 NPs. Contrariwise, the potential of TiO 2 NPs to induce DNA double-stranded breaks was demonstrated in HUVEC cells. The genotoxic effect of TiO 2 NPs was more pronounced as their particle size diminished, producing more DNA damage [96].
Besides damages caused by TiO 2 NPs-induced oxidative stress on DNA molecules, we investigated the influence of this on the membrane of lysosomes from MRC-5 cells. So far, different studies have demonstrated that NPs could induce the permeabilization of lysosomal membranes. For example, Li et al. [97] found that the membrane of lysosomes from MRC-5 cells could be affected by Au NPs. In addition, membrane permeabilization was induced in THP-1 cells by ZnO NPs [98], in HepG2 NPs by Ag NPs [99], or in 3T3 cells by Si NPs [100]. The previously mentioned studies that investigated Au and Si NPs associated the damages caused on lysosomal membranes with an increase in the generation of ROS.
One of the roles of lysosomes is to enzymatically digest spent cellular organelles. As high levels of ROS in MRC-5 cells treated with TiO 2 NPs might damage different intracellular structures, an increase in the number of lysosomes was expected. Our results suggested that the number of lysosomes was not significantly influenced by P25 and P25-Fe(1%)-N NPs exposure. However, the lysosomal membrane was significantly impaired at 72 h of exposure. The damaged membrane of lysosomes probably allowed the release of lysosomal content, especially cathepsins, enzymes that can be involved in activating caspase-dependent cell death pathways [101]. Neither the expression of the p53 protein nor those of caspase-3, -8, and -9 indicated that apoptosis was activated in MRC-5 cells by TiO 2 NPs, although we observed that cathepsin B diffuses from lysosomes into the cytosol. The insignificant differences between the expression of cathepsin B validated that the diffuse red signal obtained through immunofluorescence resulted only from the lysosomal membrane permeabilization. As cathepsin B is a lysosome-resident protein, the result confirmed that TiO2 NPs did not significantly influence lysosomal formation in MRC-5 cells.
Besides the innate antioxidant defense system that acts directly on generated ROS, eukaryotic cells can cope with oxidative damage of DNA due to different reparatory mechanisms, including the base excision repair (BER) mechanism. OGG1/2 has a crucial role in the removal of oxidized guanine species, being the enzyme responsible for their recognition, hence the initiation of the BER process [102]. We found that MRC-5 cells overexpressed OGG1/2 when exposed to both P25 and P25-Fe(1%)-N TiO 2 NPs for 72 h, suggesting the BER mechanism was induced. The constant level of OGG1/2 noticed after 24 h of exposure might be explained by a delay between transcription and translation processes. The first result that suggested a reparatory mechanism had been activated was the decrease of 8-OHdG level in the case of 50 µg/mL P25 NPs exposure at 72 h and further the unaffected DNA integrity revealed by Comet assay.
Du et al. [103] revealed that OGG1 is overexpressed in a dose-dependent manner in human hepatocytes L02 by a combined treatment of TiO 2 NPs and lead, whereas Zijno et al. [104] showed that OGG1 level increased in human colon Caco-2 cells following treatment with TiO 2 NPs. Also, Xia et al. [105] found that human kidney HEK293T cells express OGG1 in response to the oxidative damage caused by TiO 2 NPs that act synergistically with CdCl 2 .
In contrast with our results, control of BER activity is managed by p53 through its ability to regulate the cell cycle [106]. We found that the expression of p53 was totally inhibited. Therefore, the point mutations caused by 8-OHdG in the sequence of DNA [43] might have been transmitted during cell division prior to the activation of the reparatory mechanism. However, the BER pathway can function in a p53-independent manner, as other proteins might arrest the cell cycle [107].

Physicochemical Characterization of TiO 2 NPs
Two types of TiO 2 NPs were used in this study: (i) Degussa P25 (Aeroxide ® P25) purchased from Sigma Aldrich (St. Louis, MO, USA) and (ii) Degussa P25 co-doped with Fe and N atoms that were obtained experimentally by direct impregnation in 1% FeCl 3 6H 2 O and urea. The method of impregnation of TiO 2 NPs with Fe and N, as well as the characteristics of the two types of TiO 2 NPs, were described in detail in the previous publications of our research group [70,75]. Briefly, powders of P25 and P25-Fe(1%)-N NPs consisted of approx. 83% anatase (with a crystallite size of around 30 nm) and approximately 17% rutile (with a crystallite size of around 50 nm) [70]. Moreover, our group showed that these types of TiO 2 NPs formed large aggregates when they were suspended in MEM supplemented with 10% FBS. Zeta potential values around -10 mV also confirmed the low stability of TiO 2 NPs [75].
XPS measurements provided in this work were obtained in an analysis chamber using a monochromatized Al K α1 X-ray source (1486.74 eV). The electrons were analyzed with a 150 mm hemispherical electron energy analyzer (Phoibos, Specs Gmbh, Berlin, Germany). TEM images and measurements were performed on a JEOL 200 CX transmission electron microscope (accelerating voltage: 200 kV). In this experiment, MRC-5 human lung fibroblasts were exposed to 10 and 50 µg/mL TiO 2 NPs for 24 and 72 h. Stock suspensions of 2 mg/mL TiO 2 NPs (P25 and P25-Fe(1%)-N) were prepared by adding 10 mg of each NP's type in 5 mL of phosphate-buffered saline (PBS), pH ≈ 7.4. For improving particles' dispersion, suspensions were sonicated for 5 min at room temperature using the ultrasonic processor UP50H (Hielscher Ultrasonics GmbH, Teltow, Germany). Then, stock suspensions were exposed for 30 min to UV light to be sterile when used. MRC-5 cells were detached as described above and seeded into 75 cm 2 culture flasks. P25 and P25-Fe(1%)-N NPs were added directly into the culture medium at the abovementioned final concentrations. Cells used as the control for each assay underwent the same procedures but were grown in an NP-free culture medium.

Measurement of 8-Hydroxy-2 -Deoxyguanosine Level
The level of 8-OHdG was measured using a commercially available enzyme-linked immunosorbent assay (ELISA) kit purchased from Abcam (ab201734; Cambridge, UK). Previously, DNA from MRC-5 cells exposed to TiO 2 NPs was isolated and quantified. Afterward, the DNA was digested with P1 nuclease and treated with alkaline phosphatase; thus, nucleotides were transformed into nucleosides. Further, DNA samples were processed using the 8-OHdG ELISA kit according to the manufacturer's instructions, and finally, their absorbance was measured at 450 nm using a microplate reader (TECAN GENios, Grödig, Austria).

Fluorescence Microscopy Analysis
Fluorescent staining was used to analyze the actin cytoskeleton morphology and dynamic, lysosomes' number and density, as well as cathepsin B localization. To observe actin filaments, MRC-5 cells cultured in flasks and exposed to TiO 2 NPs were fixed with 4% paraformaldehyde for 20 min at room temperature. Then, cell membranes were permeabilized with a mixture of 0.1% Triton X-100 and 2% bovine serum albumin (BSA) for 30 min. F-actin was labeled by incubating the cells for 1 h with 20 µg/mL phalloidin-fluorescein isothiocyanate (FITC; Sigma Aldrich, St. Louis, MO, USA). The staining of cell nuclei has been done by 4 ,6-diamidino-2-fenilindol (DAPI; Molecular Probes, Life Technologies, Carlsbad, CA, USA). Images of the actin cytoskeleton were acquired using the inverted fluorescence microscope Olympus IX71 (Tokyo, Japan).
The fluorescent staining of lysosomes was performed by incubating MRC-5 cells with 100 nM LysoTracker Green DND-26 (Molecular Probes, Invitrogen) for 30 min at 37 • C in a humidified atmosphere containing 5% CO 2 . Hoechst 33342 (Molecular Probes, Invitrogen) was used to counterstain cell nuclei. Images of stained lysosomes were taken with Olympus IX71 inverted fluorescence microscope (Tokyo, Japan). Green fluorescence intensity in different fields of view per each sample was quantified using the ImageJ 1.53u software available online at https://imagej.nih.gov/ij/ (National Institute of Health, Bethesda, MD, USA) and displayed as a mean relative to the control.
Immunofluorescent localization of cathepsin B was performed by seeding MRC-5 cells on coverslips at a density of 2 × 10 4 cells/cm 2 . After fibroblasts were allowed to adhere overnight, they were exposed to TiO 2 NPs, as described in Section 4.2. Further, MRC-5 cells underwent fixation and permeabilization as described above in the case of F-actin. Cathepsin B was labeled by incubating cell plates (overnight, 4 • C) with Alexa Fluor 594coupled anti-cathepsin B antibody (Santa Cruz Biotechnology Inc., Dallas, TX, USA). The staining of cell nuclei has been done by DAPI. Labeled cathepsin B was visualized at 60x objective of the fluorescence microscope Nikon Eclipse E200 (Tokyo, Japan).

Comet Assay
Comet assay was performed using a single-cell electrophoresis kit (Cell Biolabs, INC, San Diego, CA, USA). After exposure to the two types of TiO 2 NPs, MRC-5 cells were collected, resuspended in PBS, and diluted until the density of 1 × 10 5 cells/mL was reached. A volume of 10 µL of each cellular suspension was mixed with 100 µL low melting agarose maintained at 37 • C. Further, a volume of 75 µL from this mixture was stretched uniformly in thin films on a Comet glass slide. The agarose was allowed to jellify by incubating the slides on a horizontal surface in the dark at 4 • C for 15 min. Then, cells embedded in agarose were lysed (using the lysis solution within the kit at 4 • C, 60 min) and further treated with an alkaline solution (4 • C, 30 min). Afterward, the slides were washed with deionized water and subjected for 20 min to low voltage horizontal electrophoresis migration (20V). Subsequently, the slides were washed with 70% ethanol. Finally, DNA molecules from the agarose-embedded cells were stained with the Vista Green fluorescent dye. The negative control was represented by MRC-5 cells cultivated in an NP-free growth medium. The positive control underwent the same procedure, but NP-free cultured cells embedded in agarose were exposed at 70 µM H 2 O 2 (5 min, 4 • C). Images of the comets were acquired using the fluorescence microscope Olympus IX 71 (Tokyo, Japan). Fluorescence from representative images was quantified using the OpenComet plugin within the ImageJ 1.53u software (National Institute of Health, Bethesda, MD, USA) and displayed as a percentage of tail DNA expressed relative to the negative control.

Western Blot Analysis
Western Blot technique was used to determine the expression level of p53, cathepsin B, caspase-3, -8, -9, and OGG1/2 proteins. In advance, total protein extracts of samples were prepared, and their concentration was measured by the Bradford method. Harvested MRC-5 fibroblasts suspended in PBS were subjected to 3 cycles of 30 s ice-assisted sonication using the ultrasonic processor UP50H (Hielscher, Teltow, Germany) to disrupt the cell membranes. Obtained lysates were centrifuged at 3000× g, at 4 • C for 10 min, and then each supernatant containing the total protein extract was individually collected and stored at −80 • C until further use.

Protein Concentration
Protein concentration was measured using the Bradford method [108]. Briefly, the optical density of the reaction product between Bradford Reagent (Sigma Aldrich, St. Louis, MO, USA) and total protein extracts was measured at 595 nm using a FlexStation 3 Spectrophotometer. Protein concentrations of all samples were calculated based on a BSA standard curve between 0 and 1.25 mg/mL (0-18.8167 µM).

Statistical Analysis
The means of three independent experiments were expressed as percentages relative to the control ± standard deviation. Statistical differences between each treatment and the control were evaluated using the Student's two-tailed t-test. The statistical significance was displayed based on the p values as follows: * for p < 0.05; ** for p < 0.01; *** for p < 0.001. All the data were analyzed and visualized using GraphPad Prism software (version 8; GraphPad Software Inc., San Diego, CA, USA).

Conclusions
Our results could suggest that the oxidative lesions caused by TiO 2 NPs in human pulmonary fibroblasts could be partially neutralized by co-doping them with low amounts of nitrogen and iron ions. Moreover, the toxic effects of P25-Fe(1%)-N NPs can be considered attenuated compared to the undoped P25, albeit they were not totally suppressed. The main impairments probably produced by ROS in pulmonary fibroblasts were related to the oxidation of DNA components and lysosomal membrane permeabilization that led to the leakage of lysosomes' content into the cytoplasm. Additionally, overexpression of OGG1/2 in correlation with the integrity of DNA molecules indicated that probably the BER mechanism successfully managed the intranuclear damages induced by TiO 2 NPs. Therefore, we hypothesized that MRC-5 cells might be more resilient than other cell types to the effects induced by TiO 2 NPs. This conclusion could also be supported by the fact that pulmonary cells are usually more exposed to exogenous ROS-producing compounds that enter the lungs by inhalation, and their reparatory mechanisms are probably more active. However, other implications might be involved. The inhibited expression of p53 suggested that the cell cycle of pulmonary fibroblasts was not arrested during reparatory processes, as normally happens, indicating that the DNA errors, which probably occurred, might persist during cell division. In conclusion, our study showed that intracellular mechanisms of pulmonary fibroblasts could be stressed by TiO 2 NPs even though cell viability was not affected. Moreover, iron doping of TiO 2 NPs might be considered a suitable strategy to attenuate the effects of TiO 2 NPs on MRC-5 cells. We consider that this research contributes to the knowledge regarding the interaction of doped P25 NPs with molecular mechanisms of in vitro cultured cells and might be a support for the design of safer and more efficient TiO 2 NPs.