Acute Exposure of Upcyte® Hepatocytes to Sub-lethal Concentrations of Graphene Oxide: Impairment of Phase-I Xenobiotic Metabolism and Albumin Transcription

Graphene Oxide (GO) is a promising candidate for nanomedicine applications. Due to the central role of liver in biotransformation of xenobiotics and drugs, the impact of GO on hepatic functional cells represents a crucial evaluation step for its potential implementation as drug. Primary human hepatocytes (PHH) are the election model for studying drug toxicity and metabolism, however current technical limitations may slow down the large-scale diffusion of this cellular tool in in vitro investigations. To assess the potential hepatotoxicity of GO, in this study, we propose an alternative approach employing second-generation upcyte ® hepatocytes as cell model, which show metabolic and functional proles akin to PHH. Cells have been acutely exposed to increasing GO concentrations for 24 hours. Upon sub-lethal concentrations of GO, stress-related cell responses to GO (such as apoptosis, oxidative stress and inammatory response) have been evaluated, along with a broad investigation of GO impact on specialized hepatic functions. Results show an IC 50 equal to 102.2 μg/mL, which is in line with recent data obtained by hepatocellular carcinoma-derived cells. However, at sub-lethal doses ( ≤ 80 μg/mL), it is detected a mild activation of early apoptosis, but not oxidative stress or inammatory response. Importantly, we observed a clear impact of GO on phase-I drug metabolism enzymes (e.g., CYP3A4, CYP2C9) through the inhibition of gene expression and metabolic activity. Conversely, phase-II enzyme system and phase-III eux transporters were not affected by GO. Finally, GO strongly downregulated the gene expression of Albumin.


Introduction
Graphene is a carbon-based material consisting in a single layer of sp 2 -hybridized carbon atoms arranged in two-dimensional (2D) honeycomb-shaped lattice [1]. In the last decade, graphene and its derivatives have been raising huge interests in many industrial elds (e.g., electronics, energy storage or thermal isolation) owing to their peculiar physical-chemical characteristics. In particular, graphene oxide (GO) is a promising nanomaterial for nanomedicine applications [2], including drug and gene delivery, bioimaging and tissue engineering [3]. GO is commonly obtained via graphite oxidation and its chemical structure is characterized by the presence of highly reactive oxygenated chemical functional groups, such as epoxy, hydroxyl and carboxyl groups [4,5]. These functionalities make GO water-soluble and easily conjugable to bio-macromolecules or small ligands. Therefore, for such intrinsic properties, GO can be considered a suitable candidate for loading drugs or other bioactive molecules and/or for acting as vehicle system of drugs. However, the potential technological boost of GO raises concerns related to its safety for human use [6]. In the development of innovation technological processes of GO as novel drug, the early application of toxicology tools may be useful to predict toxicity at an early stage of the material design and development, thus offering a rapid safety screening for many combinations of GO-based materials. This approach requires the quick identi cation of dose-limiting toxicities (e.g., sub-lethal doses), along with the material impact on organ speci c functionalities (e.g., hepatotoxicity).
Furthermore, it possibly requires the application of in vitro models accounting for human body compartments (e.g., gastrointestinal tract, lung, liver, spleen, and so forth), due to their central role in drug pharmacokinetic (i.e., administration, distribution, metabolism and elimination [6]). In vivo studies on biodistribution reported that, after entering the blood stream (through primary or secondary routes), GO may accumulate in many organs including liver, where interactions with liver functional cells may be likely [7][8][9][10]. Indeed in mouse models, it is evidenced an early degeneration and necrosis of hepatocytes in proximity of central vein [8] and the activation of IL-6 gene expression in parenchymal hepatocytes [10], upon animal exposure to GO. Data also indicate that, despite GO can be eliminated from the body via faeces and/or urine [7,8,11,12], it behaves as a bio-persistent material in both human relevant simulated biological uids (e.g., gastro-intestinal uid [13]) and epithelial in vitro models (such as gut and lung [13,14]). Moreover, on the side of in vitro investigations, recent studies (which mostly employ hepatocellular carcinoma-derived cells) have indicated an IC 50 approximately within the range of 50 and 100 µg/mL GO [15,16]. Although the useful information gathered by these in vitro models, however, it emerges a knowledge gap about GO impact on specialized hepatic functions (such as phase-I and II drug metabolism enzymes, phase-III e ux transporters and Albumin), which, among many others, are useful descriptors for an early safety assessment of a nanomaterial in the innovation process. In fact, the above-mentioned models show a very low gene expression and a reduced metabolic activity of xenobiotic-metabolizing enzymes [17]. Based on their metabolic and functional features, primary human hepatocytes (PHH) represent the gold standard especially for studying drug metabolism and toxicity, but their limited availability and technical drawbacks greatly hamper their use at large scale [18,19]. To avoid these limitations, in the present study we propose an alternative approach using proliferating, differentiated human hepatocytes generated using upcyte® technology, since hepatic functionalities (phase-I and -II enzymes gene expression, cytochromes P450 metabolic activity, Albumin production) and toxicological responses are similar to those of PHH [17,19,20]. These cells were obtained from PHH, after transduction with HPV E6 and E7 genes followed by positive selection of slowly proliferating OSMresponsive cells [19]. As a result, they are not immortalised, present a normal karyotype and do not display a transformed phenotype. Therefore, upcyte® hepatocyte cultures may have the capacity to display GO in vitro toxicity and/or its potential impact on hepatocyte metabolism and functions. In our investigations, 2D cultures of upcyte® hepatocytes were acutely exposed to increasing concentrations of GO in order to evaluate their potential effects on cell viability, apoptosis activation, in ammatory response, oxidative stress induction and, nally, hepatocyte-speci c functions. To the best of our knowledge, no data are currently available about the assessment of GO cytotoxicity in human differentiated, functional hepatocytes. Consequently, our study provides a piece of information on GO hepatotoxicity in vitro, related not only to cell viability but also to functionalities of hepatocytes.

Results
Characterization of GO in Milli-Q® water and in complete HHPM GO used in our investigations showed the typical ake-like shape, as revealed by scanning electron microscopy ( Figure S1A, Supplementary Material). This data is in line with previously reported observations [13]. To measure the hydrodynamic size, GO was dispersed in either Milli-Q® water or complete HHPM using increasing concentrations (4, 20, 80 µg/mL) and different incubation times (0, 2 and 24 hours; Fig. 1A). Results show that after 2 hours of incubation in complete HHPM, 4 µg/mL GO suspension presents three main peaks with a relative mean size of 7.9 nm ± 0.1 nm (peak #1), 338.9 nm ± 29.0 nm (peak #2) and 5131.4 nm ± 274.8 nm (peak #3), respectively. As shown in Figure S1B (Supplementary Material), the smaller peak (peak #1) is identi ed and likely refers to the protein components supplemented in complete HHPM [22]. Peaks #2 and #3 are, therefore, relative to two GO populations, with peak #2 that, based on its relative intensity (73.1%), corresponds to the most representative size population. Instead, peak #3 is indicative of partial GO agglomeration/aggregation, which is quite similar in both the dispersants used. As the GO concentration increases (20-80 µg/mL), it is noticed a very slow shift of peak #2 toward larger sizes in both complete HHPM and Milli-Q® water along with a progressive rise of the relative peak intensity. Overall, these data suggest that the suspension is poly-dispersed, having a mild tendency to agglomerate/aggregate. However, such effect does not increase after a longer time of incubation as no changes in terms of hydrodynamic size and intensity are observed after 24 hours of incubation in complete HHPM. The tendency to agglomerate/aggregate of GO at higher concentrations (20-80 µg/mL) is also evidenced by the presence of a sediment, as reported in Fig. 1B. Peak position and relative intensity of all the peaks identi ed in each tested condition are reported in Table 2 (Supplementary Material). The zeta potential analysis shows a negative surface charge of GO in Milli-Q® water, which goes from − 35 mV to -41 mV as a function of the concentrations applied (Fig. 1C). After 24-hour incubation in complete HHPM, the surface charge of GO increased to about − 27.5 mV, as a consequence of the possible formation of a biomacromolecular corona around GO [22]. The particle-corona complexes (see technical details in Materials and Methods) show mean hydrodynamic diameters quite comparable to those of bare GO ( Figure S1C and

GO impacts on cell viability and exerts membrane damage in upcyte® hepatocytes
To understand the possible detrimental impact of GO on upcyte® hepatocytes, we rst evaluated its impact on cell viability and cell membrane integrity. Con uent cell cultures were treated for 24 hours with increasing GO concentrations, as reported in Fig. 2A. A wide concentration range (4-320 µg/mL) was applied to calculate the half maximal inhibitory concentration (IC 50 ). As evidenced by the reduction of the metabolic activity, GO determined a dose-dependent reduction of cell viability, with an IC 50 equal to 102.2 µg/mL ( Fig. 2A). Note that, the highest GO concentration tested, corresponding to 320 µg/mL, caused a cell viability reduction comparable to the lethal effect exerted by the positive control, which was 0.03% Triton X-100 ( Figure S2A, Supplementary Material). As revealed by the cytotoxicity assay, treated cells also showed a signi cant release in the culture medium of the cytosolic lactate dehydrogenase (LDH), indicating a clear cytotoxicity due to cell membrane damage (Fig. 2B). In particular, it is possible observing a signi cant increase of cytotoxicity with respect to the control condition to about 11.5% (p = 0.0015) at 80 µg/mL GO, a concentration value lower than the IC 50 , whereas at lethal doses (320 µg/mL), the damage weakly increased only up to 17% (p < 0.0001). In addition, with the increase of GO concentration, the decrease of cell viability alongside the increase of released LDH showed a signi cant negative linear correlation (p < 0.0001 and Pearson's r = 0.9521; Figure S2B, Supplementary Material). The morphology analysis con rmed cytotoxicity results, revealing that, with the increase of GO concentration (starting from 80 µg/mL GO), cells showed an evident morphological change along with a size reduction.
Conversely, control cells were characterized by normal spreading round shape (Fig. 2C). The progressive GO deposition on the top surface of treated cells, especially at concentration values higher than IC 50 (102.2 µg/mL), was observed. This phenomenon is indicative of a partial colloidal instability of GO suspensions (as evidenced also by DLS analysis), so mechanical cell damage due to particle deposition cannot be excluded.
GO induces early apoptosis but not oxidative stress and in ammatory response in upcyte® hepatocytes at sub-IC 50 doses To assess if GO induces apoptosis, oxidative stress or in ammation in upcyte® hepatocytes, cells were treated with increasing sub-IC 50 doses of GO (2-80 µg/mL). Regarding apoptosis induction, we screened the extent of phosphatidylserine (PtdSer) exposure over time by treated cells (Fig. 3A). At the rst hours (3 and 6 hours) of GO stimulation, no signi cant variation of PtdSer was observed at any of tested concentrations. However, after 24 hours of treatment, PtdSer exposure signi cantly increased about 1.29folds with respect to the control, upon treatment with 20 µg/mL GO (p = 0.0352). This increment remained almost unvaried at higher concentrations (40 µg/mL GO, p = 0.0043; and 80 µg/mL GO, p = 0.0429). It is important to underline that GO showed just mild induction of PtdSer exposure and a different kinetic in comparison to the positive control (0.5 µM Staurosporine). Indeed, Staurosporine induced an earlier signi cant increase of PtdSer exposure, reaching a maximum peak at 6 hours (p < 0.0001). PtdSer level reached again the basal value after 24 hours of Staurosporine treatment in correspondence to the appearance of necrotic cells ( Figure S3, Supplementary Material). To evidence signs of cellular apoptosis, the intracellular amount of cleaved-PARP in GO-treated cells was also measured. We found no alteration of cleaved-PARP levels, with respect to the control, at any of tested concentrations, further con rming the mild effect induced by GO (Fig. 3B). Since oxidative stress is involved in many mechanisms of cytotoxicity (such as apoptosis, DNA damage, lipid peroxidation), we investigated the possible role of GO as oxidative stress inducer [30]. We found that GO did not modulate the gene expression of the two main cellular antioxidants, HO-1 and SOD1, at any of tested concentrations (Fig. 3C). As opposite, the positive control (400 µM H 2 O 2 ) induced a signi cant upregulation of HO-1 and SOD1 gene expression (p = 0.0094 and p = 0.0269, respectively). Upon GO treatment, western blot analysis of HO-1 and SOD1 con rmed the absence of signi cant variations of the corresponding protein levels with respect to the control (Fig. 3D). As far as in ammatory response is concerned, GO did not induce modi cation in the gene expression of TNFα, IL-1β, IL-6 and IL-8, considered as mediators of in ammation (Fig. 3E). As opposite, upcyte® hepatocytes treated overnight with complete medium supplemented with 100 µg/mL LPS (positive control) showed a signi cant upregulation of gene expression of IL-1β, IL-6 and IL-8 (p = 0.0482, p < 0.0001 and p = 0.0001, respectively).

GO impairs cytochrome P450 system in upcyte® hepatocytes at sub-lethal concentrations
To evaluate the GO impact on cytochrome P450 system of upcyte® hepatocytes, cells were treated for 24 hours with increasing sub-IC 50 doses (2-80 µg/mL). We measured the gene expression and the corresponding metabolic activity of CYP3A4 and CYP2C9, which are the most representative enzymes of that system. In parallel, the responsiveness of cytochromes P450 was assessed making use of wellknown drugs. For this aim, cells were daily incubated up to 72 hours [28,29]  Cipro oxacin-mediated downregulation to about 0.7-folds with respect to un-treated cells (p = 0.0444; Fig. 4E, upper panel). Moreover, we evaluated the metabolic activity of CYP3A4 and CYP2C9 using as substrates two uorescent compounds, 7-benzyloxy-4-tri uoromethylcoumarin (BFC) and 7-methoxy-4tri uoromethylcoumarin (MFC), respectively. Results indicate that the activity of CYP3A4 was statistically increased of about 290% in Rifampicin-treated cells with respect to the control (p = 0.0247; Fig. 4A, lower panel), whereas it appeared to be reduced to about 64% in the case of Cipro oxacin-treated cells (p = 0.0006; Fig. 4B, lower panel). Comparably, CYP2C9 activity was modulated by the same treatments consistently with gene expression regulation ( Fig. 4D and E, lower panels). Afterwards, we evaluated the effect of sub-IC 50 concentrations of GO on cytochrome P450 system. We observed a dose-dependent downregulation of CYP3A4 gene expression (starting from 4 µg/mL, p = 0.0015), which was equal to about 0.37-folds with respect to the control level at the highest concentration (80 µg/mL GO, p < 0.0001; Fig. 4C, upper panel). Consistent with the gene expression, CYP3A4 metabolic activity showed a signi cant and progressive dose-dependent reduction, reaching a level of about 18.6% with respect to the control, at concentration value of 80 µg/mL GO (p = 0.0002; Fig. 4C, lower panel). In parallel, CYP2C9 gene expression signi cantly decreased starting from 20 µg/mL GO (p = 0.0012), down to about 1.9% with respect to the control level with 80 µg/mL GO (p < 0.0001; Fig. 4F, upper panel). CYP2C9 metabolic activity was also signi cantly reduced, reaching about 10.9% of the basal level with the highest GO concentration (80 µg/mL, p = 0.0006; Fig. 4F, lower panel). For both CYP3A4 and CYP2C9, GO-mediated gene expression downregulation had a signi cant positive linear correlation with the metabolic activity impairment (p = 0.0119 and Pearson's r = 0.9095 for CYP3A4 and p = 0.0015 and Pearson's r = 0.9682 for CYP2C9; Figure S4A and B, Supplementary Material). Our results indicate a strong inhibition activity of GO on CYP3A4 and CYP2C9 gene expression and relative metabolic activities. Such an inhibitory response appears to be dose-dependent and the negative effect of 4 µg/mL GO could be approximately compared to that induced by 100 µM Cipro oxacin (∼33 µg/mL). The inhibition of cytochrome P450 system by sub-lethal doses of GO was con rmed also on CYP2B6 and CYP1A2 ( Figure S4E and H, Supplementary Material). For CYP2B6, gene expression decreased starting from 20 µg/mL GO (p = 0.0069), while GO affected CYP1A2 gene expression at lower doses (2 µg/mL; p = 0.0052). Upon 80 µg/mL GO treatment, CYP2B6 and CYP1A2 reached a gene expression equal to about 4.3% and 11.3% of the control level (p < 0.0001 for both), respectively. In contrast to the effects exerted on CYP3A4 and CYP2C9 gene expression, Rifampicin and Cipro oxacin had no effects on CYP2B6 and CYP1A2, revealing their speci c mechanism of action only on CYP3A4 and CYP2C9 in upcyte® hepatocytes ( Figure S4C, D and S4F, G, Supplementary Material). Conversely, GO seems to act on a wide range of targets within cytochrome P450 system. GO does not modulate gene expression of phase-II GST and phase-III ABCG2, but downregulates expression of Albumin After having observed the effects of GO on phase-I drug metabolism enzymes, we investigated the effect of sub-lethal GO exposures on phase-II and phase-III drug metabolism/transport enzymes in upcyte® hepatocytes. In particular, the gene expression of Glutathione S-Transferase (GST) and ATP Binding Cassette Subfamily G Member 2 (ABCG2) was evaluated. Results show that GST and ABCG2 gene expression was not modulated by GO at any concentration tested ( Fig. 5A and B). After that, we investigated the effect of GO on the gene expression of two xenobiotic-sensing receptors, i.e., PXR and CAR ( Figure S5A and B, Supplementary Material), nding that GO signi cantly downregulated PXR gene expression to about 12% with respect to the control level only at 80 µg/mL (p = 0.0224), whereas it was ineffective on CAR gene expression at all tested concentrations. Finally, we analysed the gene expression of Albumin, nding that Albumin was statistically, dose-dependently downregulated to about 5.2% with respect to the control level with 80 µg/mL (p < 0.0001; Fig. 5C).

Discussion
This study aimed at investigating in vitro hepatotoxicity of GO after 24-hour acute exposure in a 2D-cell culture model of second-generation upcyte® hepatocytes. These cells represent an interesting model, since they are generated from primary human hepatocytes (PHH) via genetic engineering in order to induce the expression of HPV E6 and E7 genes [20]. Second-generation upcyte® cells are derived from E6/E7 low cell colonies, which are slowly proliferating cells, able to proliferate upon treatment with OSM (IL-6-like protein) and to differentiate terminally in 4 days after OSM removal [19]. These cells are not immortalised, have a normal karyotype and are metabolically and functionally similar to differentiated PHH [19,20,31]. In addition, cells applied in this study are derived from the speci c donor 653-03, which apart the good basal activity of cytochromes P450 (CYPs), shows also an appropriate responsiveness to known inducers and inhibitors of CYP enzymes [31]. Based on these features, upcyte® hepatocytes may resemble the functionality of primary hepatocytes and then can be considered a reliable in vitro cell culture model for studying the acute effects induced by toxicants on the hepatic tissue. In this study, we focused our attention on GO, which, together with other nanomaterials belonging to the wider family of graphene, has raised great interests due to its potential in nanomedicine. However, hazard assessment, focused on human health, is required for a safe development of these materials [6,32]. To this regard, in view of medical applications, it then appears clear how important is the understanding of the potential impact of GO on hepatic tissue, since liver has a central role in the biotransformation and detoxi cation of xenobiotics, after their entrance in the circulatory system. Before the in vitro cellular study, GO was characterized in terms of morphology, lateral dimensions and surface charge by SEM, DLS and Zeta Potential analysis. When dispersed in Milli-Q® water, GO showed the typical ake-like structure of graphene, size distribution (about 300 nm) and a surface charge (from − 35 mV to -41 mV) comparable to data previously published [13]. In complete HHPM, GO adsorbed medium proteins [22] and had an appreciable tendency to sediment, possibly related to the increased surface charge (-27.5 mV). We focused on acute toxicity of increasing concentrations of GO (4-320 µg/mL) using a 24-hour exposure of upcyte® hepatocytes. By applying such a wide concentration range, we found an IC 50 equal to 102.2 µg/mL along with relevant cytotoxic effects (such as morphological changes, cell membrane damage and cell viability reduction) at concentrations higher than 80 µg/mL. This data appears to be in line with those obtained with HepaRG™ cell line, which, although with some differences, presents hepatocyte-like differentiated phenotype [33,34]. These cells, upon treatment with 50 µg/mL GO, showed indeed absence of toxicity [35]. Noteworthy, to the best of our knowledge, this is the only work which investigated the potential toxicological effect of GO using a cell line presenting features comparable to human hepatocytes, even though it is an organotypic co-culture of hepatocyte-and cholangiocyte-like cells, as opposite to upcyte® hepatocytes. Overall, our results indicated a GO-mediated toxicity at a concentration range also comparable to immortalized cell lines. HepG2 cells, when treated with GO, presented an IC 50 approximately within the range of 50 and 100 µg/mL along with a clear cell viability reduction at concentration values higher than 80 µg/mL [15,16]. In L02 cells, GO provoked only a moderate cell viability reduction at 100 µg/mL or higher concentrations [36]. Taken together, these data indicate that the toxicological responses on hepatic cells induced by GO treatments seem not particularly affected by the cell type intrinsic features. Afterwards, we studied stress-related cell responses (apoptosis, oxidative stress and in ammation) in upcyte® hepatocytes upon acute treatment with increasing sub-IC 50 doses of GO (2-80 µg/mL). In the range of 20-80 µg/mL GO, we found an increased temporal exposure of PtdSer, but unvaried protein levels of cleaved-PARP. PtdSer is normally present in the inner layer of plasma membrane, but it is exposed to the outer layer as "eat me" signal for macrophages during apoptosis [37]. In normal conditions, PARP is involved in DNA repair, but, when it is proteolytically cleaved by the executioner Caspase-3, the apoptotic process is irreversible [38,39]. Based on these mechanisms, the obtained results indicate an early-stage induction of apoptosis by GO, which does not undergo to maturation (no cell death in upcyte® hepatocytes at the end of the treatment, according to no changes in cleaved-PARP levels). Similarly, it has been reported a dose-dependent increase of early apoptotic cells after 24 hours of GO treatment, both in L02 cells [36] and in HepG2 cells [15], even though higher GO doses were used with L02 cells (100 and 300 µg/mL) with respect to HepG2 cells (up to 50 µg/mL). Moreover, Chatterjee et al. reported that apoptosis executioners (Caspase-8, -9 and − 3) were un-affected or even downregulated after 24 hours of treatment with both 20 µg/mL and 81 µg/mL of GO, showing that late apoptotic cells increased in number only after longer treatments (48 hours) [15]. In relation to intracellular antioxidant defences in upcyte® hepatocytes, we found no modulation of both gene expression and protein levels of cellular antioxidants HO-1 and SOD1 after 24 hours of GO treatment, suggesting that GO should have induced no or weakly the generation of endogenous reactive oxygen species (ROS). However, some recent studies show ROS generation after different times of exposure in both HepG2 cells and L02 cells, with differences in treatments and collection times [15,16,36,40].
Notably, upcyte® hepatocytes could be less susceptible to GO-induced oxidative stress in comparison with immortalized hepatic cells or their antioxidant defenses could be su cient to neutralize a mild generation of ROS. Finally, we evaluated the in ammatory response of upcyte® hepatocytes upon GO treatment, showing no pro-in ammatory response (none of four tested mediators was altered, at the experimental conditions applied in the study). Conversely, using immortalized hepatic cells (HepG2 cells), Liu et al. reported a signi cant upregulation of IL-6 gene expression upon 24-hour exposure to 5 µg/mL GO [16], whereas, upon 24-hour stimulation with 20 µg/mL GO, Chatterjee et al. observed changes in in ammatory gene expression, possibly mediated by TGFβ1 signalling [15]. A similar pro-in ammatory response was observed both in vitro and in vivo by Zhang et al. [10]. Different size GO (small-GO, range 50-200 nm; medium-GO, range 200-500 nm; and large GO, range 500-2000 nm) enhanced IL-6 gene expression and secretion by different hepatic cell lines (Hepa1-6, HepG2 and Huh7 cells). In vivo, intravenous injection of different sized GO in the IL-6 reporter mouse model showed an increase of IL-6 expression, speci cally in liver, by activation in parenchymal cells. IL-6 induction was directly dependent both by GO concentration and size. Moreover, a weak pro-in ammatory response was detected in a 3D airway model (primary human bronchial epithelium), sub-chronically treated with increasing concentration of aerosolized GO [14]. By contrast, Lee et al. demonstrated in vivo that GO polarized invariant natural killer T-cells towards an anti-in ammatory phenotype, dampened pro-in ammatory cytokine production and protected mice against liver injuries due to induced in ammatory responses [41]. Similarly, Han et al. reported that GO attenuated in ammation by the modulation of the polarization of mouse macrophages [42]. Based on these experimental evidence, it seems that GO could act as a pro-or an anti-in ammatory stimulus, depending on the size, the actual concentration applied, the oxidation status of the surface and the targeted cell type. Overall, these data indicate that GO was not toxic for upcyte® hepatocytes, when acutely administered in a sub-lethal range. In the second part of this study, we focused our attention on two main hepatocyte-speci c functions, such as drug metabolism/transport and Albumin production, taking advantage by peculiar features of upcyte® hepatocytes. Regarding the metabolism of xenobiotics/drugs, cytochromes P450 represent the most important phase-I enzymes in human functional hepatocytes [43]. For the presented study, we selected as representative enzymes and potential GO targets some CYPs (CYP3A4, CYP2C9, CYP2B6 and CYP1A2), which are the most abundant enzymes involved in biotransformation reactions of liver. Results show that, at sub-lethal doses, GO dosedependently downregulated all tested CYPs. For CYP3A4 and CYP2C9, the effect of inhibition was demonstrated also for the enzyme metabolic activities, suggesting either a direct inhibitory effect of GO or a reduced amount of such enzymes caused by the reduced corresponding transcription. Such ndings reinforced the preliminary evidence of GO inhibition effects on cytochrome P450 system reported by Strojny et al. [35]. In that work, although using different in vitro approaches (such as the differentiated HepaRG™ cells for monitoring CYP gene expression and a microsomal-based model for assessing CYP metabolic activity), it has been reported a strong reduction of transcription of many CYP isoforms (CYP3A4, CYP2B6, CYP1A2 and CYP2E1) and the inhibition of the catalytic activity of CYP3A4, CYP1A2 and CYP2D6 [35]. It is worthwhile mentioning that the presented data on CYP gene expression and metabolic activity were obtained using the same cell model and, hence, provided for the rst time a complete description of CYP-inhibition effect of GO in functional, differentiated hepatocytes. Continuing our investigations, we found that GST (phase-II enzyme) and ABCG2 (phase-III e ux transporter) were not affected by acute, sub-lethal GO exposure, suggesting that GO could interfere with body's xenobiotic detoxi cation function, preferentially at the level of phase I system, in our cell model. An impairment of cytochrome P450 system caused by GO exposure could have severe consequences for human health, such as an impaired detoxi cation from xenobiotics and drugs, with an increased risk of adverse side effects. Interestingly, Strojny et al. showed in HepaRG™ cells that GO (50 µg/mL) signi cantly downregulated CAR and PXR [35], which are nuclear receptors acting as ligand (xenobiotic)-dependent, transcription regulators of a large part of the phase-I, -II and -III executioners [44]. In upcyte® hepatocytes, CAR resulted to be un-affected by GO treatment and PXR was downregulated only at the highest concentration (80 µg/mL GO), suggesting that presumably corresponding protein levels were unaltered.
Such evidence imply that GO could interfere with gene expression of CYPs via PXR-and CAR-independent molecular pathways, in our cell system. Finally, GO downregulated in a dose-dependent manner the gene expression of Albumin, which is the main plasma protein produced by liver and a typical marker of differentiated hepatocytes. It is known that Albumin production is predominantly regulated at the level of transcription and, hence, its downregulation could impair functions carried out by this protein, such as the regulation of plasma colloid osmotic pressure or the transport of endogenous molecules, ions and drugs through the blood circulation [45].

Conclusions
In conclusion, we investigated the effects of GO upon acute stimulation in upcyte® hepatocytes. Using this cell model, we were able to evaluate different aspects of the cell response to GO, as cytotoxicity, stress-related responses (apoptosis, oxidative stress and in ammatory response), along with hepatocytespeci c functions, and this large-scale approach represents the novelty of our study. In fact, some evidence have been already reported in literature but spread in different studies, using different cell models. Moreover, upcyte® hepatocytes allowed us to study reliably the response of representative cytochromes P450 (CYP3A4, CYP2C9, CYP2B6 and CYP1A2), in terms of gene expression and metabolic activity, along with other hepatocyte markers (phase-II GST, phase-III ABCG2 and Albumin). This organic description of GO impact on hepatocyte functionality is a further new aspect about GO hepatotoxicity in vitro. Overall, our data raised doubts about an effective nanosafety of GO for human health, since an acute exposure of GO could have a negative impact on hepatocyte-speci c functions. GO-mediated impairment of cytochrome P450 system could determine altered body's detoxi cation from xenobiotics and drugs, with an increased risk of adverse effects. Furthermore, Albumin downregulation could result in hypoalbuminemia, and severe consequences as edema and ascites [45]. In this framework, as long terms perspective, it would be of interest to understand if liver architecture of the hepatic lobules, when receiving similar doses, could effectively isolate hepatocytes from a direct contact with GO, preserving the hepatic functions. Moreover, one other important aspect to elucidate may be the interaction between GO and body's immune system cells (i.e., Kupffer cells), which are normally resident in liver sinusoids and able to remove xenobiotics, as for instance nanomaterials, which may be carried to liver through the blood circulation [46,47]. For these reasons, more complex 3D in vitro models (e.g., liver organoid or liveron-a-chip) and, ultimately, in vivo studies could be bene cial to clarify the effective hepatotoxicity of GO.

GO synthesis and characterization
Graphene oxide (GO) was kindly provided by Dr E. Vazquez (Universidad de Castilla-La Mancha, Spain). Characterization of the nanomaterial dispersed in Milli-Q® water was previously provided by Guarnieri et al. [13]. GO morphology was evaluated by scanning electron microscopy (SEM), carried out by JSM-7500FA Field Emission Scanning Electron Microscope equipped with a thermionic source. An acceleration voltage of 10.0 kV was used. GO colloidal stability was analysed as a function of concentration (4,20 and 80 µg/mL) and incubation time (0, 2 and 24 hours at 37 °C), when dispersed in Milli-Q® water or complete HHPM. GO size distribution pro les were determined via dynamic light scattering (DLS) analysis using a Zetasizer Nano-ZS at 25 °C, even though such technique is not entirely accurate for the analysis of non-spherical particles. Five consecutive measurements were performed for each GO suspension. The number of runs per measurement, the attenuator and the optimal measurement position were set automatically. GO-free complete HHPM was used as background control. For aqueous suspensions, GO surface charge was measured via Zeta Potential (ZP) analysis using a Zetasizer Nano-ZS at 25 °C. Five consecutive measurements, with an automatically set number of runs per measurements, were taken of each GO suspension. To study particle-corona complexes, at the end of 24hour incubation in complete HHPM, GO suspensions were centrifuged at 15'000 g for 15 minutes at 4 °C and corresponding pellets were washed three times adding a volume of Milli-Q® water equal to the initial volume [21,22]. Particle-corona complexes were characterized via DLS and Zeta Potential, as described above. GO was endotoxin-free, as previously reported by Di Cristo et al. [14], in accordance with US Food and Drug Administration guidelines (https://www.fda.gov/regulatory-information/search-fda-guidancedocuments/guidance-industry-pyrogen-and-endotoxins-testing-questions-and-answers).

Upcyte® hepatocyte culture
Second-generation human upcyte® hepatocytes (donor 653-03) were cultured following manufacturer's indications. Upcyte® hepatocytes were seeded at the concentration of 10'000 cells/cm 2 into cell culture asks coated with 0.1 mL/cm 2 of 50 µg/mL collagen-type I in 20 mM acetic acid, and they were cultured in complete HHPM, in incubation in a humidi ed atmosphere at 37 °C, with 5% CO 2 . No antibiotics were added to the culture medium in order to do not alter cytochromes P450 activity. Culture medium was changed 3 times per week. Cells were expanded for 1 or 2 passages before being treated, as described below.

Cell Viability Assay
Cell viability upon GO treatment was evaluated by colorimetric resazurin reduction test. Upcyte® hepatocytes were cultured into collagen-coated, at bottom 96-well plates (cell growth area equal to 0.3 cm 2 , approximately) and, at the con uence, they were treated with different GO concentrations ( nal volume equal to 75 µL per well) for 24 hours. At the end of GO treatment, stimulation media were replaced with 100 µL per well of serum-free phenol red-free high glucose DMEM, supplemented with 44 µM resazurin sodium salt, after having extensively washed with DPBS [23]. After 1-hour incubation at 37 °C in a humidi ed atmosphere with 5% CO 2 in the dark, resazurin solution was transferred into a clean 96-well plate and uorescence was measured at 535 nm by Tecan Spark® reader. For each culture condition, three independent experiments were performed, each one with a technical triplicate. In each experiment, a couple of cell-free, collagen-coated wells per culture condition was incubated with the stimulation medium and used as blank value during the data analysis. The reported results are expressed as percentage values (means ± SD) compared to the control condition (set as 100%). The half maximal inhibitory concentration (IC 50 ) of GO was calculated using log(inhibitor) vs. normalized response curves model on Prism software.

Cytotoxicity Assay
To evaluate cell membrane damage upon GO treatment, colorimetric CytoTox96® Non-Radioactive Cytotoxicity Assay was used. Upcyte® hepatocytes were cultured into collagen-coated, at bottom 96well plates (cell growth area equal to 0.3 cm 2 , approximately). At the con uence, cells were treated with different GO concentrations ( nal volume equal to 75 µL per well) for 24 hours. As a positive control, con uent cells were treated with complete medium supplemented with 0.03% Triton X-100 for 24 hours. At the end of stimulation, conditioned media were collected, centrifuged at 15'000 g for 15 minutes at 4 °C, and analysed following manufacturer's instructions. For each culture condition, three independent experiments were performed, each one with a technical triplicate. In each experiment, a couple of cell-free, collagen-coated wells per culture condition was incubated with the stimulation medium and used as blank value during the data analysis. The results are expressed as percentage net increase of the absorbance (means ± SD) compared to the positive control (set as 100%), after having subtracted the basal value of the control condition (set as 0%).

Apoptosis and Necrosis Assay
To assess the induction of apoptosis and necrosis, RealTime-Glo™ Annexin V Apoptosis and Necrosis Assay was used, according to manufacturer's instructions. Upcyte® hepatocytes were cultured into collagen-coated, at bottom 96-well plates (cell growth area equal to 0.3 cm 2 , approximately). At the con uence, cells were treated for 24 hours with complete medium supplemented with GO at different concentrations, in the presence of detection reagent ( nal volume equal to 75 µL per well). As a positive control, con uent cells were treated for 24 hours with 0.5 µM Staurosporine (Stau), diluted in the same incubation medium. After 3, 6 and 24 hours of incubation, luminescence and uorescence (at 530 nm) were measured by Tecan Spark® reader for the induction of apoptosis and necrosis, respectively. For each culture condition, three independent experiments were performed, each one with a technical triplicate. In each experiment, a couple of cell-free, collagen-coated wells per culture condition was incubated with the incubation medium and used as blank value during the data analysis. Apoptosis data are expressed as n-fold increase over the control conditions (means ± SD) for each incubation time.
Necrosis results are expressed as net increase of absorbance for each condition (means ± SD), after having subtracted the basal value of the control cells obtained after 3 hours of incubation (set as 0 a.u.).

Reverse Transcription and Quantitative Real-Time PCR
For gene expression analysis of GO-treated cells, upcyte® hepatocytes were cultured into collagencoated, at bottom 24-well plates (cell growth area equal to 2.0 cm 2 , approximately) and, at the con uence, they were treated with different GO concentrations ( nal volume equal to 500 µL per well) for 24 hours. At the end of stimulation, cells were extensively washed with DPBS and incubated with TRIzol™ Reagent (500 µL per well) at -80 °C for at least one night. Total RNA was isolated according to manufacturer's instructions. RNA yield was determined measuring absorbance at 260 nm with NanoDrop One C , while RNA purity was examined considering A260/A280 and A260/A230 ratios. Total RNA (2 µg/sample in 20 µL of total volume reaction) was reverse-transcribed to rst-strand cDNA by SuperScript™ VILO™ cDNA Synthesis Kit, following manufacturer's instructions. Transcript levels of target genes were measured by quantitative Real-Time PCR (qPCR) using iTaq™ Universal SYBR® Green Supermix, on Applied Biosystems ViiA 7 Real-Time PCR System. Custom-made primer sequences, temperatures of annealing, corresponding amplicon sizes and qPCR e ciencies are reported in Table 1 (Supplementary Material). The gene expression of GAPDH was used as endogenous control (reference gene). For each primer pair, melting curve analysis was carried out in order to verify the production of a single amplicon and, consequently, primer speci city. Transcript levels were calculated using Pfa 's model for relative quanti cation [24]. In each run, samples were assayed in technical triplicate. The results are expressed as the average of three independent experiments, with the relative SD values.

Western blot
To evaluate GO effects on apoptosis and oxidative stress, protein content from GO-treated upcyte® hepatocytes was analysed by western blot. In particular, proteins were isolated starting from the same homogenates used for RNA isolation as described by Chomczynski [25]. The protein content was quanti ed by Pierce™ BCA Protein Assay Kit, following manufacturer's instructions. Electrophoresis was performed loading 20 µg of total protein per sample on NuPAGE™ 4-12% Bis-Tris gel, under reducing conditions. Proteins were transferred to an Amersham™ Protran™ 0.2 µm NC nitrocellulose blotting membrane. The blot was blocked with 5% non-fat milk in TBS Buffer, supplemented with 0.1% Tween 20 (T-TBS), for at least 1 hour at room temperature and then probed with primary antibodies raised against cleaved-PARP (1:1'000), HO-1 (1:1'000), SOD1 (1:500) or β-Actin (1:1'000), overnight in a cold room. β-Actin was considered as internal control. The blot was extensively washed with T-TBS and incubated with secondary anti-rabbit (1:15'000) or anti-mouse (1:10'000) HRP-linked IgG antibodies, for 1 hour at room temperature. Then, blot was washed with T-TBS and incubated with Clarity™ Western ECL Substrate, following manufacturer's instructions. The blot image was acquired by ChemiDoc™ MP Imaging System.
Each considered marker was analysed by western blot in three independent experiments. The densitometric analysis was performed by quantifying band densities by Fiji software (http:// ji.sc [26]).
The reported results are expressed as n-fold increase over the control condition (means ± SD).
Cytochrome P450 activity assay To evaluate the enzymatic activity of CYP3A4 and CYP2C9 in GO-treated cells, the conversion of 7benzyloxy-4-tri uoromethylcoumarin (BFC) or 7-methoxy-4-tri uoromethylcoumarin (MFC) to 7-hydroxy-4tri uoromethylcoumarin (HFC) was monitored as described by Donato et al. [27], with some modi cations. Upcyte® hepatocytes were seeded into collagen-coated, at bottom 96-well plates (cell growth area equal to 0.3 cm 2 , approximately) and, at the con uence, they were treated with different GO concentrations for 24 hours ( nal volume equal to 75 µL per well). As positive and negative controls, con uent cells were treated for 72 hours with complete medium supplemented with 50 µM Rifampicin or 100 µM Cipro oxacin, respectively, changing the stimulation medium every day, after an early wash with DPBS [28,29]. At the end of treatments, cells were washed with DPBS and they were incubated with 100 µL per well of 100 µM BFC or 150 µM MFC in incubation medium (1 mM Na 2 HPO 4 , 137 mM NaCl, 5 mM KCl, 0.5 mM MgCl 2 , 2 mM CaCl 2 , 10 mM glucose, 10 mM Hepes; pH 7.4 buffered solution) for 2 hours at 37 °C, in a humidi ed atmosphere with 5% CO 2 . After the incubation with BFC or MFC, the supernatant was collected, diluted 1:1 (v/v) with β-glucuronidase/arylsulfatase (150 Fishman units/mL and 1200 Roy units/mL, respectively) and incubated for 2 hours at 37 °C. At the end of this step, the reaction mixture was diluted 1:1 (v/v) with the quenching solution (0.25 M Tris in 60% acetonitrile). Finally, the formation of the uorescent HFC metabolite was quanti ed at the wavelength of 410 nm (excitation) and 510 nm (emission) by Tecan Spark® reader. Each culture condition was assayed in technical triplicate, in three independent experiments. In each experiment, a couple of cell-free, collagen-coated wells per culture condition was incubated with the stimulation medium and used as blank value during the data analysis.
The reported results are expressed as percentage values (means ± SD) over the control condition (set as 100%).

Statistical analysis
Statistical analysis was run on Prism software. Ordinary one-way ANOVA was performed for cell viability assay, cytotoxicity assay, apoptosis assay, qPCR analysis, western blot analysis and cytochrome P450 activity assay. If ANOVA detected statistically signi cant differences within the data set, Dunnett's multiple comparisons test (cytotoxicity assay, apoptosis assay, qPCR analysis and cytochrome P450 activity assay) or Tukey's multiple comparisons test (cell viability assay) were used to calculate signi cant differences. Two-way ANOVA was used for necrosis assay, with Sidak's multiple comparisons test. Unpaired t-Test was used to calculate signi cant differences within data sets obtained by qPCR analysis or cytochrome P450 activity assay between the control condition and the drug-treated cells. GO IC 50 was calculated using log(inhibitor) vs. normalized response curves model. Correlation between cell viability and cytotoxicity data and between relative gene expression and metabolic activity data of CYP3A4 and CYP2C9 was calculated using a linear regression. All tests were run setting a con dence interval of 95%. When p < 0.05, differences were considered statistically signi cant. All data are presented as means ± standard deviations (SD) of three independent experiments.

Declarations
Availability of data and materials The datasets used and/or analyzed during the current study are available from the corresponding author on reasonable request. Triton X-100 for 24 hours, while 'CTRL' represents un-treated cells. Results are expressed as percentage values compared to Triton (set as 100%), after having removed the basal value of CTRL (set as 0%).
Results were obtained from three independent experiments (means ± SD). The symbols '**' and '****' refer to p ≤ 0.0038 and p < 0.0001 respectively, calculated versus CTRL (ordinary one-way ANOVA). (C) Representative morphology of GO-treated and CTRL cells at the end of stimulation by optical microscopy.

Supplementary Files
This is a list of supplementary les associated with this preprint. Click to download.