The synthesis of the NPs was performed by sonochemical method using acoustic cavitation (30) at the Laboratory of Acoustic Cavitation of the Department of Physics and Geology of the University of Perugia. The ultrasound generator used was a “Horn Type” working at 20 kHz and 750 W, using a 13-mm titanium probe (Sonics & Materials, Newtown, CT, USA). In particular, anatase phase of TiO₂ NPs was prepared dissolving 85 ml titanium (IV) butoxyde, (Sigma-Aldrich Co., St. Louis, MO, USA) in 200 ml of pure ethanol (Sigma-Aldrich Co., St. Louis, MO, USA), successively 200 ml of water (Sigma-Aldrich Co., St. Louis, MO, USA) were added dropwise under ultrasound irradiation for 90 min at 50% of amplitude. The obtained precipitate was centrifuged and calcinated at 400°C for 3 h. Chemical-structural characterization of the synthesized NPs was performed by X-ray diffraction (XRD) and electron scanning spectroscopy (SEM). XRD was evaluated at room temperature with the diffractometer Philips X’Pert PRO MPD (Malvern Panalytical Ltd., Royston, UK) operating at 40 kW/40 mA with step size of 0.017° and a step scan of 70 s, using a Cu Kα radiation and X-Celerator detector. SEM analysis was performed using a LEO 1525 Field Emission Scanning Electron Microscope, ZEISS (Jena, Germany). Size distribution of nanoparticle dispersion was performed by dynamic light scattering (DLS) using a Nicomp 380 ZLS spectrometer sorgent 35 mW He–Ne laser at 654 nm and Avalanche PhotoDiode (APD) detector (PSS NICOMP; Santa Barbara, CA, USA), Briefly, starting from a concentration of 1 mg/ml of TiO₂ NPs dissolved in endotoxin free water (Sigma-Aldrich, St. Louis, MO, USA) to obtain a stock dispersion and sonicated for 30 min to limit the formation of nanoaggregates, dilutions of 10 μg/ml were immediately prepared in different media in order to identify the best experimental conditions for NPs application and more precisely both in SCs culture medium, i.e., Hamster Ovary cells F-12 (HAMF-12), (Euroclone, Milan, Italy), supplemented with 0.166 nM retinoic acid (Sigma-Aldrich Co., St. Louis, MO, USA), 5 ml/500 ml of insulin-transferrin-selenium (ITS), (cat. no. 354352, BD Biosciences, Franklin Lakes, NJ, USA) and, based on literature data (31), Dulbecco’s modified Eagle’s medium (DMEM) (Euroclone, Milan, Italy) supplemented with 0.166 nM retinoic acid (Sigma-Aldrich, St. Louis, MO), 5 ml/500 ml ITS (cat. no. 354352, BD Biosciences, Franklin Lakes, NJ, USA), and 2 mg/ml bovine serum albumin (BSA) (Sigma-Aldrich Co., St. Louis, MO, USA).

Animal studies were conducted in agreement with the guidelines adopted by the Italian Approved Animal Welfare Assurance (A-3143-01) and the European Communities Council Directive of November 24, 1986 (86/609/EEC). The experimental protocols were approved by the University of Perugia. Danish Duroc prepupertal pigs (15 to 20 days old) underwent bilateral orchidectomy after general anesthesia with ketamine (Ketavet 100; Intervet, Milan, Italy), at a dose of 40 mg/kg, and dexmedetomidine (Dexdomitor, Orion Corporation, Finland), at a dose of 40 g/kg (32), and were used as SC donors. Specifically, pure porcine prepupertal SCs were isolated and characterized, according to previously established methods (33, 34), as reported in the Supplementary Material .

NPs working solutions, at a concentration of 5 and 100 µg/ml according to MTT analysis in the Supplementary Material , were prepared, administered to SCs for 24 h (acute exposure) and 1, 2, and 3 weeks (chronic exposure) where medium changes were performed every 72 h. The control group was the unexposed SCs (0 TiO₂ NPs µg/ml). Samples to perform all necessary analyses were collected at each experimental point.

Cellular uptake of TiO₂ NPs at 5 and 100 µg/ml was detected with different qualitative and quantitative techniques such as TEM and inductively coupled plasma-optical emission spectrometry (ICP-OES), respectively.

Intracellular ROS were measured by treating unexposed and exposed SCs with 50 μM dichlorofluorescein diacetate (DCFH-DA) (Sigma-Aldrich Co., St. Louis, MO, USA) solution in Dulbecco’s phosphate-buffered saline (D-PBS) (Sigma-Aldrich Co., St. Louis, MO, USA) at 37°C for 30 min. Fluorescence was read by using a plate reader (DTX 880 Multimode Detector, Beckman Coulter). Data were normalized for cell viability (MTT assay) and expressed as the percentage of unexposed SCs. The sensitivity of the test was confirmed by adding 30 µM hydrogen peroxide (H₂O₂) (30 min) on unexposed SCs as positive control.

To evaluate the oxidative DNA damage, unexposed and exposed SCs were processed in the comet assay under alkaline conditions (alkaline unwinding/alkaline electrophoresis, pH >13), basically following the original procedure (35). A positive control consisted in treating SCs with 1 μM 4-nitroquinoline N-oxide (4NQO) (Sigma-Aldrich, Milan, Italy) for 1 h at 37°C (36). At the end of treatments, the cells were washed twice with 5 ml ice-cold D-PBS (Sigma-Aldrich, St. Louis, MO, USA), pH 7.4, and detached with 300 μl of 0.05% trypsin (Invitrogen, Milan, Italy) in 0.02% Na₄EDTA (Invitrogen, Milan, Italy). After 3 min, trypsinization was stopped by adding 700 μl fetal bovine serum (FBS), (Sigma-Aldrich, St. Louis, MO, USA). Cells were then collected by centrifugation (70×g, 8 min, 4°C). Briefly, cell pellets were gently resuspended in 300 μl of 0.7% low-melting point agarose (Sigma-Aldrich, St. Louis, MO, USA) (in Ca^2+/Mg^2+-free D-PBS, w/v) at 37°C, layered onto a conventional microscope slide precoated with 1% normal-melting point agarose (in Ca^2+/Mg²⁺-free D-PBS, w/v), and covered with a coverslip (Knittel-Glaser, Braunschweig, Germany). After brief agarose solidification at 4°C, the coverslip was removed and the slides were immersed in cold, freshly prepared lysing solution (2.5 M sodium chloride (NaCl), 100 mM Na₄EDTA, (Invitrogen, Milan, Italy); 10 mM tris(hydroxymethyl)aminomethane-hydrochloric acid (Tris-HCl) and sodium hydroxide (NaOH), pH 10, and 1% Triton X-100, (Sigma-Aldrich, Milan, Italy), added just before use, for at least 60 min at 4°C. The slides were then placed in a horizontal electrophoresis box (HU20, Scie-Plas, Cambridge, UK) filled with a freshly prepared solution (10 mM Na₄EDTA, 300 mM NaOH; pH 13, (Sigma-Aldrich, Milan, Italy). Prior to electrophoresis, the slides were left in the alkaline buffer for 20 min to allow DNA unwinding and expression of alkali-labile damage. Electrophoresis runs were then performed in an ice bath for 20 min by applying an electric field of 34 V (1 V/cm) and adjusting the current to 300 mA (Power Supply PS250, Hybaid, Chesterfield, MO, USA). The microgels were then neutralized with 0.4 M Tris-HCl buffer (pH 7.5), fixed 10 min in ethanol (10 min, Carlo Erba Reagenti, Milan, Italy), allowed to air-dry and stored in slide boxes at room temperature until analysis.

All the steps of the comet assay were conducted in yellow light to prevent the occurrence of additional DNA damage. Immediately before scoring, the air-dried slides were stained with 65 μl of ethidium bromide (EB), (Sigma-Aldrich, St. Louis, MO, USA) 20 mg/ml and covered with a coverslip. The comets in each microgel were analyzed (blind), at ×500magnification with an epi-fluorescent microscope (BX41, Olympus, Tokyo, Japan), equipped with a high sensitivity black and white charge-coupled device (CCD) camera (PE2020, Pulnix, UK), under a 100-W high-pressure mercury lamp (HSH-1030-L, Ushio, Japan), using appropriate optical filters (excitation filter 510–550 nm and emission filter 590 nm). Images were elaborated by Comet Assay III software (Perceptive Instruments, UK). A total of 100 randomly selected comets (50 cells/replicate slides) were evaluated for each experimental point.

Total RNA was isolated with Tri-reagent (Sigma-Aldrich Co., St. Louis, MO, USA) and quantified by reading the optical density at 260 nm. Subsequently, 2.5 μg of total RNA was subjected to reverse transcription (RT Thermo Scientific, Waltham, MA, USA) in a final volume of 20 μl. Real-time reverse transcriptase-polymerase chain reaction (RT-PCR) was performed using 25 ng of the cDNA and SYBR Green master mix (Stratagene, Amsterdam, the Netherlands), as previously described (37).

Gene expression vs. β-actin was evaluated by RT-PCR on a MX3000P Real-Time PCR System (Agilent Technology, Milan, Italy). The sequences of the oligonucleotide primers are listed in Table 1 . The thermal cycling conditions were 1 cycle at 95°C for 5 min, followed by 45 cycles at 95°C for 20 s and 58°C for 30 s. The data required for carrying out a comparative analysis of gene expression were obtained by means of the 2-(DDCT) method.

Aliquots of the culture media from all the experimental groups were collected, stored at −20°C for the assessment of AMH (AMH Gen II ELISA, Beckman Coulter; intraassay CV = 3.89%; interassay CV = 5.77%) and inhibin B (inhibin B Gen II ELISA, Beckman Coulter, Webster, TX, USA; intraassay CV = 2.81%; interassay CV = 4.33%) secretion as previously described (34).

Total protein extracts were prepared by lysing the cells in 100 μl of radioimmunoprecipitation assay lysis buffer (RIPA buffer), (Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) as previously described (38).

After centrifuging the mixture at 1,000×g (Eppendorf, NY, USA) for 10 min, the supernatant was collected, and total protein content was determined by the Bradford method (39). Sample aliquots were stored at −20°C for WB analysis. The cell extracts were separated by 4%–12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE), and equal amounts of protein (70 μg protein/lane) were run and then blotted on nitrocellulose membranes (Bio-Rad, Hercules, CA, USA). The membranes were incubated overnight in a buffer containing 10 mM TRIS (Sigma-Aldrich Co., St. Louis, MO, USA), 0.5 M NaCl (Sigma-Aldrich Co., St. Louis, MO, USA), 1% (v/v) Tween 20 (Sigma-Aldrich Co., St. Louis, MO, USA), rabbit antiextracellular signal-regulated kinases 1–2 (ERK1–2) (1:2000; Millipore, Burlington, MA, USA), mouse antiphospho-ERK1-2 (1:100; Millipore, Burlington, MA, USA), rabbit anti-c-Jun N-terminal kinases (JNK) (1:1000; Millipore, Burlington, MA, USA), rabbit antiphospho-JNK (1:500; Millipore, Burlington, MA, USA), rabbit antiposphop38 (1:2000; Millipore, Burlington, MA, USA), mouse anti-p38 (1:2000; Millipore, Burlington, MA, USA), antiserine/threonine protein kinase (Akt) (1:100; Cell Signaling, Danvers, MA, USA), rabbit antiphospho-Akt (1:1000; Cell Signaling, Danvers, MA, USA), rabbit antiphospho-NF-kB p65 antibody (1:1000; AbCam, Cambridge, UK), rabbit anti-NF-kB p65 antibody (1:1000; AbCam, Cambridge, UK), rabbit anticaspase 3 antibody (1:1000; Cell Signaling, Danvers, MA, USA), and mouse anti-GAPDH (1:5000; Sigma-Aldrich Co., St. Louis, MO, USA) primary antibodies.

Primary antibody binding was then detected by incubating membranes for an additional 60 min in a buffer containing horseradish peroxidase conjugated antirabbit (1:5000; Sigma-Aldrich Co., St. Louis, MO, USA) and/or antimouse (1:5000; Santa Cruz Biotechnology Inc.) IgG secondary antibodies. The bands were detected by enhanced chemiluminescence and acquired by ChemiDoc imaging System (Bio-Rad, Hercules, CA, USA).

Normality analysis was performed by Shapiro-Wilk test, and statistical comparisons were analyzed using one-way ANOVA followed by Tukey’s HSD post-hoc test (SigmaStat 4.0 software, Systat Software Inc., CA, USA). Values were reported as the means ± SEM of three independent experiments, each performed in triplicate. Differences were considered statistically significant at ^* p < 0.05 and ^** p < 0.001 compared with unexposed SCs and ^# p < 005 with respect to 5 μg/ml of TiO₂ NPs.

XRD analysis, as a methodology used to determine the structure of inorganic compounds, confirmed the synthesis of TiO₂ NPs. In fact, the resulting diffractogram was characteristic of titanium dioxide crystals in anatase form (JCPDS 00-001-0562) ( Figure 1A ). https://doi.org/10.5061/dryad.7d7wm37wf.

SEM evaluation, through a nondestructive technique, allowed to perform morphological and measurement investigation of TiO₂ NP surfaces. A representative SEM image of TiO₂ NPs in dry form is showed in Figure 1B , and the mean size distribution reports values of 20 ± 5 nm diameter, calculated by measuring over 100 particles in random fields of view. Results showed that TiO₂ NPs tended to form aggregates of submicrometric dimensions.

DLS analysis, being able to measure the hydrodynamic diameter of NPs dispersed in a liquid, confirmed some aggregation of TiO₂ NPs in suspension. The mean hydrodynamic diameter of TiO₂ NPs mainly distributed in a range of 100–300 nm ( Figure 1C ). Additional information on NP suspension setups can be found in Supplementary Material .

Light microscopy analysis was used to monitor over the course of the experiment any morphological changes in the SCs treated with the two TiO₂ NP concentrations with respect to the unexposed monolayer ( Figures 2A, D, G, J ). Morphological analysis revealed that SCs treated with 5 µg/ml TiO₂ NPs did not undergo substantial changes compared with the untreated monolayer at all times of exposure. In fact, cells exposed to TiO₂ NPs maintained the typical squamous shape of epithelial cells with vacuoles containing lipid hormones, well evident and abundantly distributed in the cell surface ( Figures 2B, E, H, K ). Conversely, SCs treated with 100 µg/ml TiO₂ NPs showed an evident reduction either of the cell number that constituted the monolayer ( Figure 2C ; Supplementary Figure 2S ) and changes in the morphology of the remaining cells that showed a spindle-like shape, with very elongated cytoplasmic protrusions, typical of a cellular stress-associated condition by the end of the first week ( Figure 2F ) that progressively increased until the end of the third week ( Figures 2I, L ).

TEM has been used both to verify the uptake of NPs and to assess any ultrastructural changes in SCs as a result of this interaction. TEM analysis of unexposed SCs outlined a typical architecture characterized by oval/elongated spindle-shaped morphology containing a quite large nucleus surrounded by abundant mitochondria and numerous aggregates of stacks of parallel and flat cisternae of endoplasmic reticulum membranes ( Figures 3A–C ). Although we did not notice substantial and/or severe ultrastructural alterations in SCs treated with 5 µg/ml TiO₂ NPs for 1 and 2 weeks ( Figures 3D, E ), in all the other exposure conditions (5 µg/ml for 3 weeks and 100 µg/ml for 1, 2, and 3 weeks), ( Figures 3F, I ) SCs showed several ultrastructural changes. Specifically, we noted a high percentage of severely damaged SCs, presenting different features of cell stress such as deeply invaginated and shrunk nuclei ( Figure 3G ), disorder/marginalization of chromatin components ( Figures 3G, I ), swollen very fragmented, or almost completely missing endoplasmic reticulum membranes ( Figures 3E, H ). In addition, mitochondrial morphology was severely affected ( Figure 3F ) and mitochondria number dramatically decreased. Finally, we noted the presence of several large vacuoles ( Figures 3H, I ) probably as a result of apoptotic mitochondria and/or enlarged endoplasmic reticulum and/or increased frequency of lipid droplets, https://doi.org/10.5061/dryad.7d7wm37wf.

ICP-OES was used to quantify the uptake of NPs expressed as the percentage of internalized NPs and the amount of metal adsorbed per cell number (expressed as ng/10⁵), at each concentration, after 5 h of treatment ( Figure 4 ). At the lowest concentration of 5 μg/ml, a 3% higher uptake rate was detected than at the 100-μg/ml concentration (1.3%) ( Figure 4 , dotted line). In contrast, the amount absorbed in ng/10⁵ cells was greater at the highest concentration (28.6 ± 15.01 vs. 3.46 ± 1.3, Figure 4 , black bars). This difference is likely attributable to gradual saturation of the particles within the cells with a progressive slowing of uptake as the concentration of NPs increases.

We assessed, by Western blotting analysis, the involvement of caspase-3 pathway that is partially or totally responsible for the proteolytic cleavage of many key proteins, such as the nuclear enzyme poly (ADP-ribose) polymerase (PARP) (40). Activation of caspase-3 requires proteolytic processing of its inactive zymogen (p35) into activated p19 and p17 fragments (40).

We demonstrated that NPs exposure, at each concentration, induced the activation of caspase-3 at the first and second week, with the cleavage of p35 into the p19 kDa active fragment, to decrease at the third week, most likely due to the progress toward the final degradation phase of apoptosis (41).

Only at the toxic dose, we observed a statistical increase of active p19 with respect to the inactive p35 fragments, expression of a more prominent apoptotic process, as highlighted by TEM analysis ( Figures 5A, B , ^# p < 0.05, ^## p < 0.001 vs. p35).

We evaluated the intracellular production of ROS and oxidative DNA damage, as the most important mechanisms activated by nanomaterial.

As shown in Figure 6A , the dose of 5 µg/ml TiO₂ NPs did not affect ROS intracellular level at 24 h and 1 week postexposure. On the contrary, at 2 and 3 weeks posttreatment, ROS level significantly decreased with respect to unexposed SCs (set as 100 in the graphs and represented as a black dotted line, Figure 6A , ^* p < 0.05). Conversely, 100 µg/ml TiO₂ NPs induced a significant increase of intracellular ROS amounts at the second and third week posttreatment with respect to unexposed SCs (set as 100 in the graphs and represented as a black dotted line, Figure 6A , ^* p < 0.05 and ^** p < 0.001) accompanied by a significant increase at the third week compared with 5 μg/ml of TiO₂ NPs, expression of oxidative stress onset with increasing concentration ( Figure 6A , ^# p < 0.05 vs. 5 μg/ml of TiO₂ NPs). As expected, H₂O₂ (positive control) induced a significant increase in ROS intracellular levels (data not shown).

The levels of oxidative DNA damage induced by TiO₂ NPs were measured as the % of DNA in tail by the alkaline comet assay over time, after acute (24 h) and chronic exposures (from 1 to 3 weeks) to 5 and 100 µg/ml of TiO₂ NPs. At the dose of 5 μg/ml, the oxidative DNA damage was detected after 24 h of acute exposure ( Figure 6B , ^* p < 0.05). The follow-up study carried out after 1 and 3 weeks of chronic exposure exhibited an increase that however did not result in a significant difference compared with unexposed SCs ( Figure 6B ). Instead, the dose of 100 μg/ml induced a significant increase over time and until the end of treatment with respect to the unexposed SCs and from the second week compared with 5 μg/ml of TiO₂ NPs, as demonstrated by the increased oxidative DNA damage with increasing concentration ( Figure 6B , ^## p < 0.001 vs. 5 μg/ml of TiO₂ NPs).

The effects of TiO₂ NPs exposure on SCs functional competence were studied through the evaluation of AMH and inhibin B gene expression and protein secretion, specific and important markers of SCs functionality.

Exposure of SCs toTiO₂ NPs induced a significant increase in AMH and inhibin B gene expression at 24 h followed by a significant decrease after 1 week up to the third week, at both concentrations with respect to unexposed SCs ( Figures 7A, B , ^* p < 0.05 and ^** p < 0.001 vs. unexposed SCs, set as 1 in the graphs and represented as a black dotted line). AMH secretion was significantly and steadily decreased at 24 h throughout the duration of the experiment, whereas inhibin secretion showed a significant decrease starting from the first week onwards ( Figures 7C, D , ^* p < 0.05 and ^** p < 0.001 vs. unexposed SCs).

To investigate the role and effectiveness of phase II detoxification enzymes against ROS production, we evaluated SOD, glutathione peroxidase (GHSPx), and HO-1 gene expression. In addition, we studied lactate dehydrogenase-A (LDH-A) gene expression as a well-established and reliable indicator of cellular toxicity and plasma membrane damage.

The gene expression of SOD1 and HO-1 increased throughout the treatment at both TiO₂ NP concentrations ( Figures 8A, B , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 1 in the graphs and represented as a black dotted line). The gene expression of GHSPx increased at the toxic dose of 100 µg/ml after the first and second week of exposure to undergo a significant reduction at the third week of the treatment ( Figure 8C , ^* p < 0.05 and ^** p < 0.001 vs. unexposed SCs, set as 1 in the graphs and represented as a black dotted line). In particular, only at the second week, the dose of 100 µg/ml of TiO₂ NPs showed a significant increase of HO-1 and GHSPx gene expression with respect to the dose of 5 µg/ml of TiO₂ NPs ( Figures 8B, C , ^# p < 0.05 vs. 5 μg/ml of TiO₂ NPs).

A significant increase in LDH-A mRNA expression was observed at 100 µg/ml of TiO₂ NPs throughout the whole experiment with respect to the unexposed SCs. In addition, a significant increase was observed from the second week even when compared with the dose of 5 µg/ml of TiO₂ NPs ( Figure 8D , ^* p < 0.05 and ^** p < 0.001 vs. unexposed SCs, set as 1 in the graphs and represented as a black dotted line; ^# p < 0.05 vs. 5 μg/ml of TiO₂ NPs).

To assess if TiO₂ NPs treatment could activate an inflammatory response, we evaluated the gene expression either of proinflammatory cytokines, such as, IL-1α, IL-6, and typical immunomodulatory factors expressed by SCs as transforming growth factor-β (TGF-β) and indoleamine 2,3-dioxygenase (IDO) (42).

The gene expression of IL-1α and IL-6 showed a significant increase at the first and second week of treatment with the subtoxic dose meanwhile at the toxic dose increased starting from the first week postexposure up to the end of the experiment ( Figures 9A, B , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 1 in the graphs and represented as a black dotted line). Moreover, IL-1α gene expression showed a significant increase compared with the dose of 5 µg/ml of TiO₂ NPs only at the third week. IL-6 gene expression, at the highest dose, showed a statistical significant increase from the first week throughout the whole experiment with respect to the lowest dose as expression of an enhanced inflammatory state was probably linked to the increasing TiO₂ NP concentrations ( Figures 9A, B , ^# p < 0.05 vs. 5 µg/ml of TiO₂ NPs).

The gene expression of TGF-β1 and IDO, at the subtoxic dose, showed an upregulation only at the first and second week respect to the unexposed SCs. At the toxic dose, the gene expression of TGF-β1 was upregulated at the first and second week; meanwhile, IDO showed an upregulation from the first to the third week with respect to the unexposed SCs (set as 1 in the graphs and represented as a black dotted line) and exhibited a significant increase compared with the lowest dose only at the third week ( Figures 9C, D , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs; ^# p < 0.05 vs. 5 µg/ml of TiO₂ NPs).

We performed Western blotting analysis to investigate the involvement and timing of activation of different MAPK family members (ERK1/2, JNK, and p38) and NF-kB signaling pathway after TiO₂ NPs exposure ( Figure 10A ).

The phosphorylation ratio of ERK1/2 showed a significant increase ( Figure 10B , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 100 in the graphs and represented as a black dotted line) at 24 h postexposure that persisted until the second week posttreatment, to be downregulated at the third week postexposure, at both concentrations.

The phosphorylation ratio of JNK remained unaltered at both concentrations, at 24 h and 1 week postexposure, but showed a significant increase from the second and third week only at the highest concentration of 100 µg/ml ( Figure 10C , ^** p < 0.001 vs. unexposed SCs, set as 100 in the graphs and represented as a black dotted line). In addition, the highest dose exhibited a significant increase starting from the second week in comparison with the lowest dose of NPs ( Figure 10C , ^# p < 0.05 vs. 5 µg/ml of TiO₂ NPs).

The phosphorylation ratio of p38 showed a significant increase at 24 h, at the lowest dose of 5 µg/ml but was followed by a progressive decrease that became significant at the second and third week after exposure ( Figure 10D , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 100 in the graphs and represented as a black dotted line). At the highest concentration of 100 µg/ml, p38-dependent pathway showed a significant increase with respect to the lowest dose from the second week ( Figure 10D , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 100 in the graphs and represented as a black dotted line; ^# p < 0.05 vs. 5 µg/ml of TiO₂ NPs).

The phosphorylation ratio of AKT showed a significant increase throughout the whole experiment at the subtoxic dose but with an inverse time-dependent relationship ( Figure 10E , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs (set as 100 in the graphs and represented as a black dotted line). At the toxic dose, a significant increase at 24 h was followed by a significant decrease over time ( Figure 10E , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs (set as 100 in the graphs and represented as a black dotted line). In comparison with the lowest dose, the toxic dose showed a significant decrease from the first week ( Figure 10E , ^# p < 0.05 vs. 5 µg/ml of TiO₂ NPs).

Finally, the phosphorylation ratio of p-NF-kB remained unchanged following acute exposure and at 1 week at both 5 and 100 µg/ml TiO₂ NPs. Subsequently, it showed a significant increase after the second and third week posttreatment at both concentrations ( Figure 10F , ^* p < 0.05, ^** p < 0.001 vs. unexposed SCs, set as 100 in the graphs and represented as a black dotted line).