Multiparametric in vitro genotoxicity assessment of different variants of amorphous silica nanomaterials in rat alveolar epithelial cells

Abstract The hazard posed to human health by inhaled amorphous silica nanomaterials (aSiO2 NM) remains uncertain. Herein, we assessed the cyto- and genotoxicity of aSiO2 NM variants covering different sizes (7, 15, and 40 nm) and surface modifications (unmodified, phosphonate-, amino- and trimethylsilyl-modified) on rat alveolar epithelial (RLE-6TN) cells. Cytotoxicity was evaluated at 24 h after exposure to the aSiO2 NM variants by the lactate dehydrogenase (LDH) release and WST-1 reduction assays, while genotoxicity was assessed using different endpoints: DNA damage (single- and double-strand breaks [SSB and DSB]) by the comet assay for all aSiO2 NM variants; cell cycle progression and γ-H2AX levels (DSB) by flow cytometry for those variants that presented higher cytotoxic and DNA damaging potential. The variants with higher surface area demonstrated a higher cytotoxic potential (SiO2_7, SiO2_15_Unmod, SiO2_15_Amino, and SiO2_15_Phospho). SiO2_40 was the only variant that induced significant DNA damage on RLE-6TN cells. On the other hand, all tested variants (SiO2_7, SiO2_15_Unmod, SiO2_15_Amino, and SiO2_40) significantly increased total γ-H2AX levels. At high concentrations (28 µg/cm2), a decrease in G0/G1 subpopulation was accompanied by a significant increase in S and G2/M sub-populations after exposure to all tested materials except for SiO2_40 which did not affect cell cycle progression. Based on the obtained data, the tested variants can be ranked for its genotoxic DNA damage potential as follows: SiO2_7 = SiO2_40 = SiO2_15_Unmod > SiO2_15_Amino. Our study supports the usefulness of multiparametric approaches to improve the understanding on NM mechanisms of action and hazard prediction.


Introduction
Nanotechnology is a fast-growing field associated with the development and use of nanomaterials (NM) that exhibit unique and/or improved physicochemical properties given their nanoscale size and high surface area per unit mass (Shatkin 2008).Owing to their attractive properties, NM have a wide range of applications (Clift, Jenkins, andDoak 2020, Vance et al. 2015) that led to an increase in their production in the last decades.Consequently, human exposure to these materials has also increased, particularly in occupational settings where inhalation is an important route of exposure (Clift, Jenkins, and Doak 2020, Pietroiusti et al. 2018, Stone et al. 2017, Vance et al. 2015).
Despite several efforts, uncertainties with respect to hazards and risks of NM to human health and the environment remain.In this regard, the genotoxicity of NM is an aspect of special concern, since DNA or chromosomal damage may lead to mutations, which in turn can increase the risk of cancer development and/or other diseases (El Yamani et al. 2017, Kryston et al. 2011).Available studies on the genotoxic potential of NM often lack a correlation between in vivo and in vitro findings, making data inconsistent and contradictory (Chen et al. 2014, Gurcan et al. 2020, Yazdimamaghani et al. 2019).Several aspects make NM toxicity profile evaluation complex, namely the large number of available NM variants, whose intrinsic properties (as synthesized), as well as the method of NM synthesis (e.g.pyrogenic, sol-gel, microemulsion) that have been shown to influence their toxicity profile (Croissant et al. 2020, Karkossa et al. 2019, Tsuji et al. 2006, Zhang et al. 2012).Consequently, their safety assessment has been carried out on a caseby-case basis (Kwon, Koedrith, and Seo 2014).At the same time, NM intrinsic properties might be modified by the environment, which can influence NM interactions and effects on the biological systems (Kroll et al. 2011, Kwon, Koedrith, andSeo 2014).In addition, factors related to the study experimental design including the test systems used that can exhibit different sensitivity, the exposure conditions (e.g.medium composition, dose range, and exposure time), as well as NM ability to interfere with assay components make toxicity data hard to compare (Ajdary et al. 2018, Guadagnini et al. 2015, Kroll et al. 2011, Kroll et al. 2012, Pauluhn 2009).
To overcome these issues, NM safety assessment needs to be done in a comprehensive manner following standardized protocols to reduce experimental variation.Several EU and national projects have meanwhile undertaken huge efforts to better define and harmonize NM genotoxicity testing (summarized in Dusinska et al. (2019)).
So far, the available literature supports the ability of NM to induce genotoxicity through direct interactions of the NM with DNA, however, the genotoxic damage seems to arise mostly from indirect mechanisms, such as NM interactions with other molecules important to DNA replication and repair or from secondary mechanisms as oxidative stress and inflammation (Chen et al. 2014, Gurcan et al. 2020, Swedish Chemicals Agency 2016, Yazdimamaghani et al. 2019).Accordingly, a battery of assays covering different mechanisms is recommended for assessing NM genotoxicity, though no strict protocol exists.The OECD guidelines for genotoxicity testing, with the exception of the Ames test, seem to be suitable for NM.However, these guidelines have not yet been formally validated for NM, and special attention should be paid to some experimental details (OECD 2014).Thus, for in vitro NM genotoxicity testing, a gene mutation test, preferably in mammalian cells, such as the mouse lymphoma assay (MLA) (OECD 2016d) or the hypoxanthine-guanine phosphoribosyl transferase gene (HPRT) test (OECD 2016b), in combination with a chromosome mutation test such as the cytokinesisblock micronucleus (CBMN) test (OECD 2016c) or the chromosomal aberration (CA) test (OECD 2016a) is recommended (Doak andDusinska 2017, Landsiedel et al. 2022).Regardless, other assays can provide useful information on NM mechanisms of action.The comet assay has been proved to be a simple and sensitive method to assess early DNA damage such as single-and double-strand breaks (SSB and DSB), alkali-labile sites, being one the most used assays for genotoxicity testing of NM (Azqueta andDusinska 2015, Doak andDusinska 2017).To maintain genome integrity, cells have developed complex mechanisms of DNA damage response (DDR) that act to repair damage and minimize the odds of lethal or permanent genetic damage.DDR encompasses multiple repair mechanisms and signal transduction pathways, including cell cycle checkpoint arrest and/or apoptosis (Smith et al. 2010).Among the different forms of DNA damage, DSB are the most severe lesions that can lead to chromosome alterations, genomic instability, and even to tumorigenesis if not efficiently repaired (Wan et al. 2019).In response to DSB induction, histone H2AX molecules are rapidly phosphorylated at serine 139 (Ser 139) of DSB sites, referred to as gamma-H2AX (c-H2AX) foci (Rogakou et al. 1998, Valdiglesias et al. 2013).This early DDR event causes chromatin remodeling and may provide a platform for the recruitment of other enzymes responsible for DSB signaling and DNA repair (Podhorecka, Skladanowski, and Bozko 2010), hence cH2AX levels are considered a biomarker of DNA repair and genomic instability (Huang and Darzynkiewicz 2006, Nikolova, Marini, and Kaina 2017, Valdiglesias et al. 2013).
Amorphous silica (aSiO 2 ) NM are extensively used in many fields (Nayl et al. 2022).They can be synthesized by different methods (physical, chemical, biological) in varying sizes and surface functionalities.All these factors (size, surface modification, and synthesis method) can contribute for their toxicological potential though their exact contribution is not yet well understood (Croissant et al. 2020).Herein, we have investigated the in vitro cyto-and genotoxic potential of a set of seven aSiO 2 NM variants with different size (7, 15, and 30 nm) and surface modification (unmodified, phosphonate-[negatively charged], aminopropyltrimethoxysilane-[positively charged] and trimethylsilyl-modified [2 and 3%; different degrees of hydrophobicity]).The rat RLE-6TN cell line was chosen as in vitro model for assessing pulmonary toxicity of the NM due to its similarity to alveolar type II epithelial cells and its established usefulness in studying human alveolar cell function (Jiao et al. 2017, Oda et al. 2011).On a first step, all aSiO 2 NM variants were tested for cytotoxicity and DNA damage (comet assay).On a second step, the most toxic aSiO 2 NM variants were selected to further explore the mechanisms involved in genotoxicity and/or associated DNA repair responses.For that purpose, we assessed cell cycle progression and c-H2AX content by flow cytometry in alveolar epithelial cells exposed to the most toxic aSiO 2 NM variants.

Cell culture
Rat alveolar type II epithelial RLE-6TN cells (passage 39) were obtained from the American Type Culture Collection (ATCC; CRL-2300 TM ; Manassas, USA).Cells were cultured in Ham's F12 medium supplemented with 2 mM L-glutamine, 0.01 mg/mL bovine pituitary extract, 0.005 mg/mL insulin, 2.5 ng/mL IGF, 0.00125 mg/mL apo-transferrin, and 2.5 ng/mL EGF, 10% heat-inactivated FBS, 100 U/mL penicillin, and 100 mg/mL streptomycin, and grown at 37 � C, 5% CO 2 in a humidified atmosphere.Cells of an early passage were subcultured twice a week at a 1:2 split ratio, using Accutase V R solution to a maximum of 20 passages.For toxicity assessment, cells (circa 80% confluence) were detached and seeded in 96or 12 well-plates, depending on the assay.The cell media was changed 24 h after seeding and the cells were allowed to grow for further 24 h until exposure.

Assessment of the cytotoxic effects induced by the aSiO2 NM variants
Cells were seeded in 96-well plates (2 � 10 4 cells/ well) and incubated with increasing concentrations (0-56 mg/cm 2 ; 200 mL) of the tested NM, serumfree cell medium (NC), and vehicle (serum-free cell culture media containing 100 lg/mL of Pluronic V R F-108 as dispersant) during 24 h at 37 � C, 5% CO 2 .Two cytotoxicity endpoints were determined: lactate dehydrogenase (LDH) release and WST-1 reduction, indicators of plasma membrane integrity and metabolic (mitochondrial) activity, respectively.LDH release was determined using the Cytotoxicity Detection Kit (Roche, Mannheim, Germany), according to manufacturer's instructions.Following the exposure period, incubation media were gently pipetted to a 96-well round bottom microplate and centrifuged for 5 min at 2210 � g to remove the cell debris and residual NM.Cells lysed with 0.2% Triton X-100 (30 min) for total LDH release were used as the positive control (PC).Briefly, 100 mL of freshly prepared reaction mixture was added to 100 mL of each sample and incubated at room temperature (RT), protected from light.Absorbance was measured at 490 and 655 nm (reference wavelength) in a Cambrex ELx808 microplate reader (Biotek, Winooski, USA).LDH release values were normalized considering the total LDH release (PC) using the following equation: ([optical density [OD] sample -OD blank]/[mean OD PC] � 100), where the blank corresponds to cell medium containing the same concentration of the respective aSiO 2 NM variant.To test for possible NM interferences, total LDH release was determined in the absence and in the presence of the highest tested concentration of each aSiO 2 NM variant.Cell metabolic activity was evaluated using WST-1 Cell Proliferation Reagent Kit (Roche, Mannheim, Germany), according to the manufacturer's instructions.Cells exposed to ethanol 70% (30 min) were used as PC.After exposure, the incubation medium was removed and cells washed with PBS pH 7.4 prior incubation for 2 h at 37 � C, 5% CO 2 with 100 mL/well of WST-1 reagent diluted 1:10 in serum-free cell culture medium.Sample's absorbance was measured at 450 and 630 nm (reference wavelength) in a Cambrex ELx808 microplate reader.To test for potential interferences of the aSiO 2 NM in the assay, PC was determined in the absence and in the presence of the highest tested concentration of each aSiO 2 NM variant.WST-1 reduction values were normalized considering the NC mean value according to the following formula ([OD sample]/[mean NC] � 100).Moreover, for both assays a cell-free system test with the NM-tested concentrations was conducted in order evaluate a potential activation of the assay enzymatic reactions by NM.

Assessment of the genotoxic effects induced by the aSiO 2 NM
The genotoxic potential and DDR were investigated by evaluating DNA damage, phosphorylation of histone H2AX (c-H2AX) and changes in the cell cycle of RLE-6TN cells after 24 h of exposure to the aSiO 2 NM.

DNA damage
DNA damage was assessed by the alkaline comet assay as previously described by Bessa et al. (2017), with slight modifications.Minimum Information for Reporting Comet Assay procedures and results (MIRCA) recommendations were followed in this manuscript (Moller et al. 2020).Briefly, RLE-6TN cells were seeded into 12-well plate (5 � 10 5 cells/well) and incubated for 24 h at 37 � C, 5% CO 2 in a humidified atmosphere with non-cytotoxic concentrations of the NM (7-28 mg/cm 2 ) and the alkylating agent MMS (250 lM) used as PC.After NM incubation, cells were rinsed 3 times with PBS pH 7.4 (w/o Ca and Mg) and gently harvested using a scraper.The cell suspensions were transferred to a microcentrifuge tube and centrifuged for 5 min at 500 � g.The supernatant removed, the pellet gently resuspended in 200 mL of PBS 7.4 and kept on ice.Cell density was determined using a Neubauer cell chamber and 8 x 10 3 cells/200 mL in PBS pH 7.4 were transferred to a microtube and centrifuged for 5 min at 700 � g.Subsequently, the supernatants were discarded, and cells embedded in 100 mL of 1% LMP agarose.Then, 5 mL of each sample (400 cells per gel) were loaded in duplicate onto dry microscope slides precoated with 1% NMP using a medium throughput 12-gel comet assay unit (Severn Biotech Ltd. V R , Kidderminster, UK).Slides were maintained for 5 min at 4 � C for gel polymerization and after that immersed in ice-cold lysis solution (2.5 M NaCl, 100 mM Na 2 EDTA, 10 mM Trisbase, 10 M NaOH pH 10, 1% Triton-X 100) during 1 h at 4 � C, protected from light.After lysis, slides were washed 3 times with PBS 7.4 for 5 min and immersed in electrophoresis solution (1 mM Na 2 EDTA, 0.3 M NaOH pH 13) in the electrophoresis platform for DNA unwinding for 40 min, followed by electrophoresis in the same solution for 30 min at constant 25 V (0.9 V/cm) and 400 mA.The slides were then washed with ice-cold PBS pH 7.4 and ice-cold dH 2 O for 10 min each, followed by fixation with ethanol (70%) and ethanol (96%) for 15 min each, and the slides dried overnight.Prior to scoring, slides were stained with a 1:10,000 dilution of SYBR V R Gold in Tris-EDTA buffer (10 mM Tris-HCl, 1 mM EDTA, pH 7.5-8).The slides were scored using a fluorescence microscope (Motic BA410 ELITE) attached to an epifluorescence illuminator (Filter Set: Exciter D480/30x, Emitter D535/40 nm, Dichroic 505DCLP) with a 100� magnification, using the image analysis software Comet Assay IV (Perceptive Instruments, Suffolk, UK).Three independent experiments with two biological replicates/experiment were performed.At least 100 cells/replicate (50 in each replicate gel) were analyzed in each replicate slide.Percentage of DNA in the comet tail (% DNA in tail) was used as DNA damage descriptor.

Cell cycle progression and histone H2AX phosphorylation (c-H2AX) analysis
Potential changes in the cell cycle, as well as phosphorylation of histone gamma-H2AX (c-H2AX) at Ser 139 residue were assessed in parallel in the same samples by flow cytometry, as previously described by Valdiglesias et al. (2011) for the c-H2AX histone and by Ros� ario et al. (2020) for the cell cycle, with minor modifications.Briefly, RLE-6TN cells were seeded into 12-well plates (2 � 10 5 cells/ well) and incubated for 24 h with sub-cytotoxic concentrations (7-28 mg/cm 2 ) of the aSiO 2 NM (SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _15_Amino, and SiO 2 _40).Briefly, after the exposure period, the incubation medium was aspirated, and the cells rinsed three times with 500 mL of PBS 7.4.Then, cells were gently harvested by incubation with 250 mL of Accutase V R solution for 5 min at 37 � C and inactivated with 500 mL of complete cell culture medium.The cell suspensions were transferred to centrifuge tubes, centrifuged for 5 min at 200 � g and the supernatants discarded.The pellets were suspended in 1 mL of pre-chilled ethanol 70% for fixation and membrane permeabilization and stored at −20 � C until further analysis.At the time of analysis, cells were centrifuged at 300 � g for 5 min, the supernatants were removed, and the pellets suspended in 1 mL of flow cytometry grade PBS (PBScit) supplemented with 1% BSA.Then, cell suspensions were centrifuged again for 5 min at 300 � g, the supernatants discarded, and cells were labeled with 100 mL of 5 mg/mL phospho-histone H2AX (Ser139) Alexa Fluor V R 488 conjugated monoclonal antibody for 15 min at RT, protected from light.Thereafter, 1 mL of 1% BSA in PBScit was added to each sample, followed by sample centrifugation for 5 min at 300 � g.The supernatants were discarded, and the pellets suspended in staining solution (50 mg/mL of PI, 50 mg/mL of RNAse A prepared in ultrapure water) for 30 min at RT, protected from light.Samples were then filtered through a 70 mm cell strainer to remove residual NM and reduce cell clumping.Acquisitions were made using Guava V R easyCyte TM flow cytometer (Merck KGaA, Darmstadt, Germany).For each sample, 1 � 10 4 events were recorded at a low flow rate (0.24 lL/s).Debris and doublets were excluded by plotting width against area for side scatter (SSC).For c-H2AX analysis, gating was based on dot plot (Red-B height linear scale vs. Green-B height log scale) with quadrants, defined based on the PC samples (i.e. cells exposed to 10 mg/mL of BLM for 4 h).For cell cycle analysis, % of nuclei in each phase of the cell cycle (G 0 /G 1 , S and G 2 /M-phases) was based on the histogram outputs for Red-B peak height linear scale.The expression of c-H2AX in each phase of the cell cycle was also determined using the dot plot (Red-B height linear scale vs. Green-B height log scale) gated for each cell cycle phase.The % subG 1 population was also calculated and used as an indicator of apoptosis.Figure S1 presents histograms obtained for NC and variants of SiO 2 NM in cell cycle analysis, and the corresponding dot plots used for c-H2AX analysis.

Statistical analysis
All data were expressed as mean ± standard deviation (SD) of three independent experiments.Statistical analyses of data were performed using GraphPad Prism version 6 for Windows (La Jolla, USA).Data were tested for normal distribution and homogeneity of variances by D'Agostino-Pearson omnibus and Levene's tests, respectively.Cytotoxicity, c-H2AX, and cell cycle data were analyzed by one-way ANOVA followed by Dunnett's test for multiple comparisons.Comet assay data were analyzed by the Kruskal-Wallis test followed by the Dunn's multiple comparisons test.Concentration-response curves were obtained using the method of least squares and half-maximal and 25% inhibitory concentrations (IC 50 and IC 25 ) values were calculated.Significance was considered at a p value < 0.05.

In vitro cytotoxicity assessment
In vitro cytotoxicity testing of the selected aSiO 2 NM variants was carried out to select of sub-cytotoxic concentrations for in vitro genotoxicity testing, thereby ensuring reliability of data.Overall, both cytotoxicity assays demonstrated to be suitable for assessing the cytotoxic potential of the selected panel of NM under study.As depicted in Figure 1, under serum-free conditions, SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _ 15_Amino, and SiO 2 _15_Phospho were the most cytotoxic among the aSiO 2 NM variants tested, while exposure to the trimethylsilyl-modified hydrophobic aSiO 2 NM, either SiO 2 _7_TMS2, or SiO 2 _7_TMS3, did not induce significant cytotoxicity on RLE-6TN cells.SiO 2 _7 markedly decreased both cellular metabolic activity and plasma membrane integrity.Amino-and phosphonate surface modification significantly decreased SiO 2 _15 cytotoxicity as evidenced by the higher half-maximal inhibitory concentration (IC 50 ) values obtained from WST-1 concentration-response curves, 55.1 and 59.3 mg/cm 2 , respectively, compared to SiO 2 _15_Unmod (34.56 mg/cm 2 ) (Table 1).Although significant changes in plasma membrane integrity of cells exposed to high concentrations of SiO 2 _40 have been detected, no significant alterations in the metabolic activity were observed.NM interferences did not seem to have occurred since no significant differences in the PC values in the absence vs. in the presence of the highest tested concentration of any tested NM were detected in both assays (data not shown).

DNA damage
The genotoxic potential of the aSiO 2 NM variants was evaluated by assessing their ability to induce DNA damage in the rat alveolar epithelial (RLE-6TN) cell line.For that purpose, cells were exposed for 24 h to sub-cytotoxic concentrations (0-28 lg/cm 2 ) of the seven aSiO 2 NM variants and DNA damage was assessed by the alkaline comet assay that detects SSB, DSB, and alkali-labile sites in individual cells (Azqueta and Collins 2013).As shown in Figure 2, negative control (NC) cells exhibited a very low level of DNA damage close to 5% of DNA in tail.On the other hand, cells exposed to the PC, the alkylating agent MMS (250 lM), were significantly damaged, as expected, with a mean % DNA in tail value of 37.9%.SiO 2 _7, either unmodified or trimethylsilyl-modified did not cause any increase in DNA damage levels comparing with NC cells.Similar findings were obtained for SiO 2 _15, irrespective of its surface modification.Only SiO 2 _40 caused a significant increase in the DNA damage comparing with NC cells, with % DNA in tail values of 11.0 ± 3.8 in cells exposed to Figure 1.Cytotoxicity of the seven aSiO 2 variants under study in rat alveolar epithelial RLE-6TN cells after 24 h of exposure, measured by lactate dehydrogenase (LDH) release (A) and WST-1 reduction (B).Data are expressed as mean ± standard deviation of three independent experiments, each performed in triplicate.LDH release values were normalized considering the positive control (PC; cells lysed with 0.2% Triton X-100), while WST-1 reduction values were normalized considering the negative control (NC; serum-free cell culture medium).Vehicle (Veh; serum-free cell culture media containing 100 lg/mL of Pluronic V R F-108 as dispersant).Data were analyzed by one-way ANOVA followed by Dunnett's post hoc test for multiple comparisons.� p < 0.05 vs. NC.the highest tested concentration.To test for potential interferences of the NM in the comet assay, NC cells were mixed with each NM at the highest tested concentration (28 mg/cm 2 ), combined with LMA (final concentration 1%) and then subjected to the comet assay procedures.For all tested NM, no significant differences in % DNA in tail values compared to the NC were observed, supporting that these NM did not interfere in this analysis (data not shown).

Hazard ranking
According to the magnitude of the cytotoxic and genotoxic (DNA damage) effects observed in RLE-6TN cells, aSiO 2 NM variants were ranked for hazard based on the criteria listed in Table S3.The cytotoxicity ranking was based on WST-1 data (IC 50 / IC 25 estimates).Hence, SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _15_Amino, and SiO 2 _15_Phospho were ranked as highly cytotoxic (Table 2).In turn, SiO 2 _7_TMS2, SiO 2 _7_TMS3, SiO 2 _40 were classified as noncytotoxic.Regarding their genotoxic potential, while most of the aSiO 2 NM variants did not cause any change in DNA integrity, being classified as negative, the SiO 2 _40 variant induced a significant DNA damage, though not concentration-dependent, being considered as equivocal.

DNA damage response (DDR)
In order to further explore aSiO 2 NM DNA damage potential and the putative DDR pathways in RLE-  6TN cells, we have analyzed cell cycle progression and c-H2AX content in cells incubated for 24 h with sub-cytotoxic concentrations (0-28 mg/cm 2 ) of the aSiO 2 NM variants ranked with higher toxic potential (SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _15_Amino, and SiO 2 _ 40) based on cytotoxicity and DNA damage data (Table 2).Under our experimental conditions, i.e. in serum-free incubation medium, G 0 /G 1 phase (81.1 ± 2.8%) was the dominant subpopulation compared to S (5.8 ± 1.3%) and G 2 /M (13.1 ± 1.5%) phases in NC cells (Figure 3; representative histogram in Figure S1).While exposure to SiO 2 _40 did not affect cell cycle progression, some changes in cell cycle subpopulations have been observed after 24 h of exposure to SiO 2 _7, SiO 2 _15_Unmod, and SiO 2 _15_Amino.Exposure to 7, 14 and 28 mg/cm 2 of SiO 2 _7 induced a significant decrease in G 0 /G 1 subpopulation to 76.5%, 71.7%, and 63.4%, respectively.Moreover, in cells exposed to 14 and 28 mg/cm 2 of SiO 2 _7, the decrease in G 0 /G 1 subpopulation was accompanied by a significant increase in S and G 2 /M subpopulations (14 mg/cm 2 : 10.7% in S and 17.6% G 2 /M phases; 28 mg/cm 2 : 13.8% in S and 22.8% G 2 /M phases).A similar trend has been observed in RLE-6TN cells exposed to SiO 2 _15_Unmod that at lower concentrations (7 mg/cm 2) caused a slight but significant reduction in G 0 /G 1 cell subpopulation (78.8%) without significantly affecting the other two phases; and at highest tested concentration, SiO 2 _15_Unmod led to a significant decrease in G 0 /G 1 (74.6%)cell subpopulation at the expense of S (8.2%) and G 2 /M (17.2%) phases (Figure 3).In turn, exposure to SiO 2 _15_Amino led to a significant decrease in G 0 /G 1 subpopulation (64.7%) at the expense of the other two phases (12.0% in S and 23.3% G 2 /M phases) only at the highest tested concentration.Considering the number of apoptotic cells, assessed by the sub-G1 cell population, a significant increase was observed after exposure to the highest tested concentration (28 mg/cm 2 ) of all aSiO 2 NM variant tested (Figure S2), except for SiO 2 _40, where a significant increase in the number of apoptotic cells was detected after exposure to 14 mg/cm 2 .
Regarding the total c-H2AX levels, as shown in Figure 4, RLE-6TN cells exhibited a low basal level of c-H2AX (1.8 ± 0.4%), with 0.9 ± 0.2%, 0.3 ± 0.1%, and 0.6 ± 0.2% of cells in the G 0 /G 1 , S and G 2 /M phases, respectively.As expected, cells exposed to bleomycin sulfate (BLM; PC) exhibited a marked increase of total c-H2AX (88.6 ± 3.6%), mostly in non-S-phase populations (data not shown), which is in accordance with previous studies (Scarpato et al. 2013, Tomilin et al. 2001).Exposure to any of the tested aSiO 2 NM variants affected c-H2AX levels of RLE cells.Incubation with SiO 2 _7 or SiO 2 _40 caused a significant increase in total c-H2AX levels in cells exposed to all tested concentrations, with SiO 2 _7 inducing the highest increase in c-H2AX (30.8 ± 8.2%, for the highest tested concentration) compared to NC.In addition, a significant increase of c-H2AX levels was also detected in cells exposed to the two highest tested concentrations of SiO 2 _15_Unmod (14.0 ± 0.9%, for the highest tested concentration) and in cells exposed to the highest tested of SiO 2 _15_Amino (24.0 ± 6.8%) in relation to the NC.Moreover, at the highest tested concentration (28 mg/cm 2 ) of any tested variant, the rise in c-H2AX levels was visible in all phases of the cell cycle, while in cells exposed to other concentrations the increase in c-H2AX levels was only detected in G 0 /G 1 and G 2 /M subpopulations, except for SiO 2 _7, where exposure to 14 mg/cm 2 only induced an increase of c-H2AX levels in cells at G 0 /G 1 phase but cells incubated with 7 mg/cm 2 also showed a significant increase of c-H2AX in all cell cycle phases (Figure 4).

Hazard ranking
As shown in Table 3 and according to the criteria depicted in Table S3, SiO 2 _7, SiO 2 _15_Unmod, and SiO 2 _40 were classified as positive, equivocal and negative, for one of the three genotoxicity endpoints assessed.While SiO 2 _7 and SiO 2 _15_Unmod were both negative for DNA damage as assessed by the comet assay, the former were positive for cell cycle progression and equivocal for c-H2AX, while the latter were positive for c-H2AX and equivocal for cell cycle progression.Interestingly, amino surface modification of SiO 2 _15 seems to attenuate its genotoxicity potential since SiO 2 _15_Amino was assigned as negative for DNA damage and equivocal for both cell cycle progression and total c-H2AX.On the other hand, SiO 2 _40_were positive for c-H2AX but assigned as equivocal for its DNA damaging potential and negative for cell cycle progression.Therefore, considering all genotoxicity endpoints data sets, aSiO 2 NM can be ranked for its hazard to rat alveolar epithelial cells as follows:

Discussion
Alveolar epithelial cells play a major role in lung physiology and are expected targets for inhaled NM (Hiemstra et al. 2018).Hence, we have carried out genotoxicity testing of different aSiO 2 NM variants in the rat alveolar epithelial cell line RLE-6TN using a multiparametric approach.Our data showed that SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _15_Amino, and SiO 2 _ 15_Phospho impacted both membrane integrity and metabolic activity of RLE-6TN cells, being these effects more pronounced in the cells exposed to SiO 2 _7 and SiO 2 _15_Unmod.In turn, the surfacemodified variants SiO 2 _7_TMS2, SiO 2 _7_TMS3, SiO 2 _ 15_Amino, and SiO 2 _15_Phospho demonstrated a lower or no cytotoxic potential, suggesting that surface modification can prevent/mitigate the cytotoxic effects induced by their pristine counterparts.In fact, previous studies demonstrated that surface modification is an effective strategy to prevent the cytotoxic effects of aSiO 2 NM (Großgarten et al. 2018, Kasper et al. 2015, Lankoff et al. 2013, Marzaioli et al. 2014, Yoshida et al. 2012).In this regard, Karkossa et al. (2019) showed significant alterations in the global proteomics, targeted metabolomics and SH2 profiling of RLE-6TN exposed to SiO 2 _7 but not to SiO 2 _7_TMS2 and SiO 2 _7_TMS3.The hydrophobicity of the organosilane surface-modified variants seems to diminish their bioactivity, given their increased tendency to agglomerate that may impair their uptake, and consequently their toxicity (Halamoda-Kenzaoui et al. 2015, Wiemann et al. 2022, Wohlleben et al. 2016).Furthermore, stronger cellular responses to SiO 2 _ 15_Unmod compared to their surface-modified variants have also been observed in rat alveolar macrophage NR8383 (Wiemann et al. 2016) and epithelial RLE-6TN cells (Karkossa et al. 2019).In turn, SiO 2 _40 caused a lower LDH release and did not alter metabolic activity compared to control RLE-6TN cells, most likely related to its largest size and lowest surface area, which is in agreement with the general consensus that particles with smaller size induce stronger biological effects than larger particles likely due to their specific surface area, which greatly increases their surface reactivity and cellular uptake (Oberdorster et al. 2005, Shang, Nienhaus, and Nienhaus 2014, Skuland et al. 2014).In this respect, and supporting our findings, Karkossa et al. (2019) reported that aSiO 2 NM surface area showed a positive correlation with mitochondrial dysfunction since an increased abundance for analytes connected to mitochondrial dysfunction was observed for the smaller building NM such as SiO 2 _7, SiO 2 _ 15_Unmod, SiO 2 _15_Amino, and SiO 2 _15_Phospho.
The genotoxic potential of aSiO 2 NM is far from being established.While some in vitro studies reported positive effects (Decan et al. 2016, Gonzalez et al. 2014, Maser et al. 2015, Mu et al.

Equivocal
No changes up to 28 mg/cm 2 Negative Changes at all tested concentrations (7-28 mg/cm 2 ); concentration dependent trend Positive 2012), others showed no genotoxic effect from exposure to these NM (Guidi et al. 2015, Jin et al. 2007, Uboldi et al. 2012).DNA damage can be induced by direct interaction of the aSiO 2 NM with the DNA, or through indirect mechanisms such as the induction of oxidative stress (Kwon, Koedrith, and Seo 2014).Multi-omics analysis of RLE-6TN cells exposed to the same aSiO 2 NM variants tested herein revealed a significant enrichment of oxidative stress related pathways, mainly in cells exposed to SiO 2 _40 (Karkossa et al. 2019(Karkossa et al. , 2021)), supporting the view that the DNA damaging ability of SiO 2 _40 might be associated to the induction of oxidative stress in RLE-6TN cells.In agreement with this, previous in vitro studies have reported positive results for the genotoxic potential of aSiO 2 NM through indirect mechanisms in lung (Guichard et al. 2016, Maser et al. 2015, Mu et al. 2012).Nevertheless, differences in the genotoxic potential of aSiO 2 NM may also arise from their ability to enter the cell, distribute in the intracellular milieu and directly interact with the nucleus.Under serum-free conditions, as the ones adopted in this study, the physicochemical characteristic that most influences aSiO 2 NM ability to affect both plasma membrane and DNA integrity is the agglomerate size (Mu et al. 2012, Yazdimamaghani et al. 2019).In line with this view, Gonzalez et al. (2014) observed a hydrodynamic size-dependent increase of micronucleus (MN) frequency in A549 human lung carcinoma cells exposed to differently sized aSiO 2 NM, which was more pronounced in the absence of serum in the incubation medium.Moreover, Wiemann et al. (2021) also observed that the presence of serum mitigates the effects of precipitated and fumed aSiO 2 NM on alveolar macrophages by lowering their bioactivity and uptake in a particle specific manner.Under our experimental conditions (in serum-free incubation medium), SiO 2 _7_TMS2, SiO 2 _7_TMS3, and SiO 2 _40 did not exert a significant effect in the metabolic activity of rat alveolar epithelial cells.However, the remaining variants significantly decreased the metabolic activity of RLE-6TN cells and can be ranked for their cytotoxic potential as follows: SiO 2 _15_Unmod > SiO 2 _7 > SiO 2 _15_ Amino � SiO 2 _15_Phospho.Interestingly, aSiO 2 NM variants with high cytotoxic potential such as SiO 2 _ 7, SiO 2 _15_Unmod and SiO 2 _15_Amino were negative for DNA damage, while SiO 2 _40, the only aSiO 2 NM variant causing DNA damage did not induce cytotoxicity.
To deal with DNA damage, cells trigger DDR signaling cascades.To further explore the mechanisms involved in DNA damage and repair, we analyzed changes in cell cycle progression and c-H2AX levels in response to exposure to selected aSiO 2 NM variants with high cyto-and genotoxic potential, i.e., SiO 2 _7, SiO 2 _15_Unmod, SiO 2 _15_Amino, and SiO 2 _ 40.Although no DNA damage was observed in cells exposed to SiO 2 _7, SiO 2 _15_Unmod and SiO 2 _15_ Amino, cell cycle arrest in S and G 2 /M phases accompanied by a significant increase in histone H2AX phosphorylation was detected.Cell cycle arrest is an important mechanism during nanoparticle-induced DNA damage (Gonzalez et al. 2010) that provides time for repair mechanisms to occur prior progression to subsequent phases of the cell cycle (Uboldi et al. 2012), thus avoiding potential gene mutation (Schonn, Hennesen, andDartsch 2010, Smith et al. 2010).At the S-phase checkpoint, the fidelity of DNA replication and the presence of unrepaired DNA damage are checked (Huang et al. 2008).In turn, G 2 /M DNA damage checkpoint serves to prevent the cell from entering mitosis (M-phase) with genomic DNA damage (Gao et al. 2016).Cell cycle arrest at G 2 /M and proliferation inhibition by SiO 2 NM exposure have previously been reported in human umbilical vein endothelial cells (HUVEC) exposed to SiO 2 NM, however DNA damage, as assessed by alkaline comet assay, was observed (Duan et al. 2013).The phosphorylation of H2AX plays a crucial role in the DDR by facilitating the recruitment of DNA repair proteins to sites of damaged chromatin (DSB) and activating checkpoint proteins that arrest cell cycle progression (Podhorecka, Skladanowski, and Bozko 2010).Furthermore, histone H2AX phosphorylation also occurs in response to replication arrest (Ward and Chen 2001) and initiation of DNA fragmentation during apoptosis (Rogakou et al. 2000).Thus, increased levels of c-H2AX in RLE-6TN cells after exposure to the tested aSiO 2 NM (SiO 2 _7, SiO 2 _15_ Unmod, and SiO 2 _15_Amino), with no induction of DNA damage detected by the alkaline comet assay may be associated to the observed S-and G 2 /M cell cycle arrest.Moreover, the possibility of other types of DNA damage, such as DNA interstrand crosslinks (ICL), which can induce H2AX phosphorylation but may not be detected by the alkaline comet assay (Huang et al. 2004, Mogi andOh 2006), or early apoptosis induction cannot be ruled out.
Whenever DNA injuries in the cells exceed cellular repair capacity, apoptosis may occur (Lu et al. 2006, Tanaka et al. 2006).At the same time, apoptosis is also associated with the occurrence of DNA breakage, however the relationship between apoptosis and the comet assay is complex since apoptotic cells may not appear as damaged cells in the comet assay, as the small oligonucleotides of latestage apoptotic cells may disappear from the cellular area during electrophoresis (Azqueta et al. 2022).Herein, an increased number of cells in the sub-G1 population after exposure to the highest tested concentration of any of the three aSiO 2 variants, mostly in cells exposed to SiO 2 _15_Amino, was observed.This finding is in agreement with Karkossa et al. (2021) that observed increased levels of metabolites such as proapoptotic Bax protein, sphingomyelins, and citric acid cycle enzymes [glutamate dehydrogenase 1 (Glud1) and fumarate hydratase (Fh)], which are associated to induction of apoptosis.
Genetic damage of another kind such as oxidative DNA damage, not detected in the standard alkaline comet assay, may occur, therefore, more studies are needed to further explore the underlying mechanisms through which these aSiO 2 NM induce genotoxicity and modulate DDR.Although SiO 2 _40 induced DNA damage and increased cH2AX levels of rat alveolar epithelial cells, mostly at G 0 /G 1 -phase, no alterations in the cell cycle progression were observed, suggesting the DNA damage detected in the alkaline comet assay correspond to DSB.DSB are one the most dangerous types of DNA damage which if left unrepaired lead to cell death (Podhorecka, Skladanowski, and Bozko 2010).In line with this, our results showed an increase of sub-G1 apoptotic cells in RLE-6TN cells exposed to 14 mg/cm 2 of SiO 2 _40, which was the concentration that caused the highest level of DNA damage in the rat alveolar cells.
With respect to the genotoxicity of aSiO 2 NM, there is a lack of correlation between in vivo and in vitro findings described in the literature, where most of the in vitro studies revealed positive genotoxic effects, contrasting with the absence of significant genotoxic risk in vivo (Yazdimamaghani et al. 2019).However, evidence exists that in vitro studies can be predictive of in vivo genotoxicity.In this regard, Haase et al. (2017) showed that ZrO 2 .acrylateand SiO 2 .aminoNM were not genotoxic to human 3D bronchial EpiAirway TM 3D cultures nor to rats following short-term inhalation (STIS).In this study, we showed a positive genotoxic effect for SiO 2 _7 and SiO 2 _40 in rat alveolar epithelial RLE-6TN cells, which is in agreement with a previous in vivo study from our group where oxidative DNA damage in the lung cells of rats intratracheally instilled with the same variants was detected, though the damage seemed repairable (Brandão et al. 2021).In addition, the low DNA damaging potential of SiO 2 _15_Amino also agrees with previous observations in vitro in EpiAirway TM cultures and in vivo in rats after shortterm exposure (Haase et al. 2017).In vitro concentrations tested in this study matched in vivo calculated deposited doses, except for the highest tested dose, which may explain the absence of genotoxic potential of SiO 2 _15_Unmod in vivo (Brandão et al. 2021) but a positive genotoxic potential in vitro.In addition, pyrogenic aSiO 2 NM (SiO 2 _7 and SiO 2 _40) caused more pronounced effects in vivo (alterations in lung DNA, protein integrity, and gene expression) compared to colloidal aSiO 2 NM (SiO 2 _15_Unmod and SiO 2 _15_Amino) as reported herein, demonstrating that the route of synthesis seems to play a role on aSiO 2 NM toxicity, which is in accordance with previous reports describing more pronounced effects caused by pyrogenic than precipitated aSiO 2 NM (Arts et al. 2007, Di Cristo et al. 2016, Karkossa et al. 2021).
Although the implemented ranking approach has some limitations, namely the low number of endpoints considered, we were able to successfully identify aSiO 2 NM variants that might require closer evaluation.In future studies, adding other genotoxicity endpoints, such as micronuclei formation, will be crucial to enhance the accuracy and the comprehensiveness of the ranking process.

Conclusion
Our study highlights the suitability and usefulness of a multiparametric approach to define aSiO 2 NM mechanism of action, and rank their hazard to pulmonary cells, with a focus on genotoxicity.Overall, no major changes in the DNA integrity were detected in response to the aSiO 2 NM variants tested, though increases in H2AX phosphorylation were observed.Increasing the number of endpoints to assess would increase the robustness and reliability of the hazard ranking and provide a better understanding of their mode of action.The rat alveolar epithelial RLE-6TN cells seem to be a predictive model of aSiO 2 NM pulmonary toxicity in vivo.Surface modification with hydrophobic organosilanes, and in a lesser extent with phosphonate groups, seem to be an effective strategy to decrease the toxicity of aSiO 2 NM.

Figure 2 .
Figure 2. DNA damage (% DNA in tail) of RLE-6TN cells exposed to the aSiO 2 NM variants for 24 h, as assessed using the alkaline comet assay.Negative control (NC; serum-free cell culture medium); Vehicle: (Veh; serum-free cell culture media containing 100 lg/mL of Pluronic V R F-108 as dispersant; positive control (PC; cells exposed to 250 lM of methyl methane sulfonate [MMS]).Data are expressed as mean ± standard deviation of three independent experiments, each performed in duplicate.Data were analyzed by Kruskal-Wallis test followed by Dunn's multiple comparisons.� p < 0.05 vs. NC.

Figure 3 .
Figure 3. Cell cycle distribution of RLE-6TN cells.Percentage of cells in different phases after exposure for 24 h to the four aSiO 2 NM variants tested (SiO 2 _7, SiO 2 _15_Unmod; SiO 2 _15_Amino, and SiO 2 _40).Values are expressed as mean ± standard deviation of three independent experiments, each performed in triplicate.Data were analyzed by one-way ANOVA followed by Dunnett's post hoc test.� p < 0.05 vs. negative control (NC).

Figure 4 .
Figure 4. Flow cytometry analysis of histone H2AX phosphorylation in RLE-6TN cells.Percentage of phosphorylated histone H2AX (% c-H2AX), total and in each cell cycle phase, after exposure to the aSiO 2 NM variants (SiO 2 _7, SiO 2 _15_Unmod; SiO 2 _15_Amino and SiO 2 _40).Data are expressed as mean ± standard deviation of three independent experiments, each performed in duplicate.Data were analyzed by one-way ANOVA followed by Dunnett's post hoc test.Total c-H2AX: � p < 0.05 vs. NC; c-H2AX in each phase of cell cycle: # p < 0.05 vs. NC.

Table 2 .
Hazard ranking of the aSiO 2 NM variants based on in vitro cyto-and genotoxicity (DNA damage) data sets of rat alveolar epithelial RLE-6TN cells.

Table 3 .
Hazard ranking of the tested aSiO 2 NM based on the genotoxicity endpoints data sets of rat alveolar epithelial RLE-6TN cells.