The Performance of HepG2 and HepaRG Systems through the Glass of Acetaminophen-Induced Toxicity

Investigation of drug-induced liver injuries requires appropriate in vivo and in vitro toxicological model systems. In our study, an attempt was made to compare the hepatocarcinoma HepG2 and the stem cell-derived HepaRG cell lines both in two- and three-dimensional culture conditions to find the most suitable model. Comparison of the liver-specific characteristics of these models was performed via the extent and mechanism of acetaminophen (APAP)-induced hepatotoxicity. Investigating the detailed mechanism of APAP-induced hepatotoxicity, different specific cell death inhibitors were used: the pan-caspase inhibitor zVAD-fmk and dabrafenib significantly protected both cell lines from APAP-induced cell death. However, the known specific inhibitors of necroptosis (necrostatin-1 and MDIVI) were only effective in differentiated HepaRG, which suggest a differential execution of activated pathways in the two models. By applying 3D culture methods, CYP2E1 mRNA levels could be elevated, but we failed to achieve a significant increase in hepatocyte function; hence, the 3D cultivation especially in APAP toxicity studies is not necessarily worth the complicated maintenance. Based on our findings, the hepatocyte functions of HepaRG may stand between the properties of HepG2 cells and primary hepatocytes (PHHs). However, it should be noted that in contrast to PHHs having many limitations, HepaRG cells are relatively immortal, having a stable phenotype and CYP450 expression.


Introduction
The investigation of drug-induced hepatotoxicity at in the center of toxicological studies since drug-induced liver injury is a major cause of late-stage clinical drug attrition, market withdrawal, and acute liver failure [1]. The prediction of clinical drug-induced liver damage is of paramount importance at the earliest possible stage of development.
The most widely used experimental model related to human acute liver failure is the acetaminophen (APAP)-based model [2]. APAP is a commonly used antipyretic and analgesic drug with a large therapeutic window, but at a high dose or in combination with alcohol or other xenobiotics, it causes centrilobular hepatic necrosis, resulting in acute liver failure [3]. APAP overdose is one of the most frequent reasons for acute liver failure in humans, accounting for nearly 50% of all cases [2,4]. Approximately 5-10% of APAP is oxidized by CYP450s (CYP1A2, CYP2E1, and CYP3A4) into the highly reactive metabolite, N-acetyl-p-benzoquinone imine (NAPQI) [5], which is detoxified upon conjugation with Briefly, for the differentiation process, cells were seeded homogenously in 96, 24, or 6-well plates (at a seeding density of 9 × 10 3 , 5.5 × 10 4 , or 2 × 10 5 cells/well, respectively). For the first 14 days, the cells were maintained in William's E medium (Sigma-Aldrich ® ) containing ADD710C-HepaRG ® Growth Medium Supplement with antibiotics (Biopredic) and Glutamax (Gibco™), which was followed by an additional 14-day differentiation phase. The differentiation process was performed in William's E medium (Sigma-Aldrich ® ) containing ADD720C HepaRG ® Differentiation Medium Supplement with antibiotics (Biopredic) and Glutamax (Gibco™). A batch of HepaRG cells used in the experiments thath underwent 14 days of growth and 14 days of differentiation was termed "differentiated HepaRG" throughout this paper.

Three-Dimensional (3D) Cell Cultures
For spheroid cell culture, cells were maintained in Thermo Scientific™ Nunclon™ Sphera™ flasks. For the nanofiber scaffold, 3D cell culture Nanofiber multiwell plates with random oriented nanofibers (Merck, Darmstadt, Germany) were used. APAP treatments were performed on 14-day 3D HepG2 cultures. For HepaRG 3D cultures, the standard 14 + 14-day differentiation process was performed before the experiments.

APAP Treatment of HepG2 and HepaRG Cells
HepG2 cells were seeded homogenously in either 96, 24, or 6-well plates (at a seeding density of 1.5 × 10 4 , 1.5 × 10 5 , 8 × 10 5 cells/well, respectively). The cells were seeded in complete growth medium and were incubated for 24 h; then, they were replaced with the APAP supplemented complete growth medium for the treatment. The treatment of HepG2 cells with APAP was performed for 24 h.
Twenty-four h before treatment of HepaRG cells, the differentiation medium was replaced by induction medium (William's E (Sigma-Aldrich ® ) supplemented with ADD650C-HepaRG ® Serum-free Induction Medium Supplement with antibiotics (Biopredic) and Glutamax (Gibco™)). Then, treatment of HepaRG cells with APAP was performed for 24 h in induction medium.

Determination of LC 50 Values via MTT Assay
Cell viability for LC 50 curves determination was measured in 96-well plates. Cells were treated as described above. The treatment medium from the plate was discarded and replaced with DMEM (HepG2) or William's E medium (HepaRG) supplemented with 1/10 volume 5 mg/mL MTT dissolved in PBS. The plate was incubated with the medium supplemented with MTT for 40 min in cell culture incubator; then, it was replaced with dimethyl sulfoxide (DMSO) to dissolve the formazan crystals and further incubated for 10 min at 37 • C. The absorbance was determined by microplate spectrophotometer (Thermo Scientific™ Multiskan™ GO) at 570 nm.

Evaluation of Cell Viability via Aspartate Aminotransferase (AST) Enzyme Activity
The AST kit (Diagnosticum Zrt, Budapest, Hungary) was used to determine AST enzyme activity according to the manufacturer's instructions. Briefly, after treatment, supernatant samples were taken from the cells. After adding the reagent to the supernatant samples, the plate was incubated for 1 min at 37 • C; then, the absorbance was repeatedly determined by microplate spectrophotometer (Thermo Scientific™ Multiskan™ GO) at 340 nm for 3 min.

Reverse Transcription and Real-Time PCR Analysis
Total RNA was isolated from using innuPREP RNA Mini Kit (Analytik Jena, Jena, Germany). Reverse transcription was accomplished using a RevertAid First-Strand cDNA Synthesis Kit (Thermo Scientific™) following the manufacturer's guidelines and protocol. cDNA amplification has been done by a Real-Time PCR System (Thermo Scientific™ PikoReal™) and Sensifast™ SYBR ® No-ROX Kit (Bioline, London, UK).

Measurement of Viable Cell Number Using Flow Cytometry
After treatment, the culture medium was discarded, the cells were washed twice with PBS, trypsinized, and resuspended in HBSS (Hanks' Balanced Salt Solution, Sigma-Aldrich ® ). A suitable volume from the cell suspension supplemented with propidium iodide (PI) dye (with 10 µg/mL final concentration) was used for the determination of viable cell number using the CytoFLEX (Beckman Coulter, Brea, CA, USA) Flow Cytometer. The emission of PI was measured on the ECD channel (610/20 nm). Data were analyzed using FlowJo ® software.

Isolation and Quantitation of Protein Samples
Cells were treated as described above and were lysed in RIPA protein isolation buffer (150 mM NaCl, 1% NP-40, 50 mM Tris pH 8,0) supplemented with 1% protease inhibitor cocktail (Sigma-Aldrich ® ), 1% phosphatase inhibitor cocktail (Sigma-Aldrich ® ), and 1 mM PMSF. Samples were incubated on ice for 30 min and centrifuged at 14,000× g for 15 min at 4 • C. The supernatant was used for protein analysis and stored at −80 • C. Protein samples were quantified using the Pierce™ BCA Protein Assay Kit (Thermo Scientific™) according to the manufacturer's guidelines.

Determination of Caspase-3/7 Activation
Cells were treated and prepared as described above. First, 3 × 10 5 (HepG2) or 4 × 10 5 (HepaRG) cells were centrifuged at 300 g for 5 min. Cells were resuspended in 50 µL of assay buffer (20 mM HEPES, pH 7.4, with 1% CHAPS, 5 mM DTT, and 2 mM EDTA) and stored at −80 • C for 2-3 days. After thawing, the lysates were supplemented with 17 nM Ac-DEVD-AMC (a fluorogenic substrate of caspase-3/7 proteases). The mixture was incubated at 37 • C for 1 h, and the fluorescence was determined by a fluorescent plate reader (Varioskan LUX, excitation: 380 nm, emission: 445 nm). The results were normalized to the protein content of the sample that was determined by Thermo Scientific™ Pierce™ BCA Protein Assay Kit, according to the manufacturer's instructions.

GSH Measurement
For the determination of cellular GSH, monochlorbimane (mClB) derivatization followed by HPLC separation and fluorescent detection was used [37,38]. First, 10 5 trypsinized cells in HBSS (Hanks' Balanced Salt Solution, Sigma-Aldrich ® ) were diluted in Tris buffer (20 mM, pH 8.0) up to 100 ul, which was supplemented with 1 U/mL glutathione-Stransferase enzyme (GST) and mClB to reach 1 mM final concentration. After a 15 min incubation in the dark at RT, the derivatization was stopped with the addition of 100% trichloroacetic acid (TCA). The solution was centrifuged at 15,000× g for 10 min, and the supernatant was used for GSH determination. For separation, a Waters Acquity UPLC H-Class system was used, equipped with an Acquity UPLC BEH C18 2.1 × 50 mm column with an average particle diameter of 1.7 µm. Gradient elution was used as 0.25% sodium-acetate (pH 3.5) and methanol. The detector was a Waters Acquity FLR fluorescent detector with excitation and emission set to 395 and 477 nm, respectively. Quantitation was achieved by measuring GSH standards.
2.12. Visualization of Cell Viability, Caspase-3/7 Activity, Reduced Glutathione, and Hepatocytes (of HepaRG) by Fluorescent Microscopy Cells were examined during and after treatments with a Nikon™ Eclipse TS2R microscope using a 4x/10x/20x phase contrast objective and a Nikon™ DS-Ri2 camera.
For visualization of cell death/viability, Hoechst 33342 (5 µg/mL) and PI dye (10 µg/mL) were added to the medium, and the cells were incubated for 30 min and for 5 min (respectively) at 37 • C. The emission of PI was examined on the TRITC channel (579-640 nm), and of Hoechst 33342 on the DAPI channel (375/28) of a Nikon™ Eclipse TS2R microscope with a Nikon™ Intensilight Epi-fluorescence Illuminator light source and a Nikon™ DS-Ri2 camera.
For visualization of caspase-3/7 activity, medium was supplemented with CellEvent™ Caspase-3/7 Green Detection Reagent (Thermo Fisher Scientific, Invitrogen™) to reach 5 µM final concentration, and the cells were incubated for at least 30 min at 37 • C. The emission of the reagent was examined on the FITC channel (516-556 nm) of a Nikon™ Eclipse TS2R microscope.
Live imaging of intracellular reduced glutathione was measured by labeling the cells with ThiolTracker™ Violet (Invitrogen™) at a final concentration of 20 µM for 30 min at 37 • C. The emission of the reagent was examined on the DAPI channel (375/28 nm) of a Nikon™ Eclipse TS2R microscope.
For the HepaRG cell line, immunofluorescent staining was used to distinguish between epithelial-like and hepatocyte populations in differentiated cells. β-catenin and E-cadherin proteins appear in the HepaRG cell line only on the surface of mature hepatocyte cells [30,35]. Cells were first washed with PBS and then fixed in −10 • C methanol for 5 min. Then, it was blocked in PBS containing 2% BSA for 30 min at RT, after which the cells were washed with PBS and labeled for 1.5 h at RT using the Anti-E-cadherin Antibody (G-10) Alexa Fluor ® 488: sc-8426 (Santa Cruz Biotechnology) at a concentration of 1.33 ug/1 mL PBS and Anti-β-catenin Antibody (15B8) Alexa Fluor ® 488 sc-53483 (Santa Cruz Biotechnology) at a concentration of 1.33 ug/1 mL PBS. After washing with PBS, the emission of conjugated antibody was examined on the FITC channel (516-556 nm) of a Nikon™ Eclipse TS2R microscope.

Statistical Analyses
All statistical analyses (one-way ANOVA or nonparametric Kruskal-Wallis ANOVA and Median Test) were carried out using TIBCO ® (Palo Alto, CA, USA) Statistica™ program (version: 13.5.0.17). p values were calculated with Dunnett's test (after one-way ANOVA) or multiple comparisons (after Kruskal-Wallis test). LC 50 values were determined using Graph Pad Prism (version: 8.0.1). Data are presented as mean ± SD from at least 3 independent experiments.

Results and Discussion
The use of experimental animals in pharmacology and toxicology is time-consuming, costly, and raises animal welfare issues; in addition, the predictive accuracy of animal in vivo testing for human adverse health effects is often questionable [39,40]. Furthermore, there is a growing need to reduce the use of experimental animals. In vitro cell-based models are often used to investigate preclinical hepatotoxicity. Due to differences in the toxicity response of different species, the use of human cell lines is advisable [41]. In in vitro models of primary human hepatocytes, immortalized human hepatic cell lines have been used, but they are limited regarding their viability, hepatic gene expression, and function [42]. Of the many options, three-dimensional (3D) models [19][20][21][22][23][24][25][26][27] and stem cell-derived models [43] have also become areas of significant interest. Developing appropriate toxicological model systems is not an easy task, but it will help the effectiveness of toxicological studies.

Acetaminophen Sensitivity of HepG2 and Differentiated HepaRG
HepG2 and HepaRG cell lines were used in our experiments. Both of them are of hepatic origin; however, their retention of hepatic function is markedly different. Liverspecific enzymes metabolize APAP through sulfation, glucuronidation, and to a lesser extent, hydroxylation [44]. The latter reaction is catalyzed by various isoforms of CYP450s and results in the formation of the reactive metabolite NAPQI. At high APAP doses, NAPQI depletes glutathione and forms protein adducts, resulting in the diminished activity of specific enzymes, oxidative stress, and ultimately hepatocyte death [44].
We wanted to investigate the degree of liver-specific characteristics of HepG2 and differentiated HepaRG lines via the extent of APAP-induced hepatotoxicity. Thus, both cell lines were treated with increasing concentrations of the drug; then, the cell viability was determined by MTT assay (Figure 1, left panels) and by the release of an intracellular hepatocyte-specific enzyme, aspartate aminotransferase (AST) (Figure 1, right panels). Among the liver injury markers, aminotransferases (AST, ALT) are the most commonly used in both clinical diagnosis and research involving hepatocyte damage [45].
Although the MTT assay is widely used to assess the cytotoxic potential of different compounds, our results revealed that it underperformed in the case of HepaRG cells. The MTT assay in HepG2 resulted in a toxicity profile in accordance with our expectations and previous observations [46,47]. The LC50 was found to be 10 mM (Figure 1a  Although the MTT assay is widely used to assess the cytotoxic potential of different compounds, our results revealed that it underperformed in the case of HepaRG cells. The MTT assay in HepG2 resulted in a toxicity profile in accordance with our expectations and previous observations [46,47]. The LC50 was found to be 10 mM (Figure 1a, Appendix B, left panel).
On the other hand, in HepaRG cultures, the toxicity could be found biphasic: a first, more sensitive phase between 1 and 20 mM and a second phase between 20 and 80 mM of APAP ( Figure 1c, Appendix B, right panel). This phenomenon was also supported by fluorescence microscopy: lower APAP concentrations (first phase) resulted in marked cell death, which was limited exclusively to hepatocyte islets, whereas biliary epithelial-like cells are resistant to APAP in this concentration range (Figure 2a,b). Immunfluorescent staining was also used to distinguish between non-parenchymal biliary epithelial-like cells and hepatocytes ( Figure 2c). β-catenin and E-cadherin proteins appears in the Hep-aRG cell line only on the surface of mature hepatocytes [30,35]. Immunostaining also supported the reduction of hepatocyte islands at 20 mM APAP ( Figure 2c). Thus, the survival of non-parenchymal biliary epithelial-like cells at low APAP concentrations (up to 20 mM) masked hepatocyte-specific death assessed by MTT assay. However, the very high APAP On the other hand, in HepaRG cultures, the toxicity could be found biphasic: a first, more sensitive phase between 1 and 20 mM and a second phase between 20 and 80 mM of APAP (Figure 1c, Appendix B, right panel). This phenomenon was also supported by fluorescence microscopy: lower APAP concentrations (first phase) resulted in marked cell death, which was limited exclusively to hepatocyte islets, whereas biliary epithelial-like cells are resistant to APAP in this concentration range (Figure 2a,b). Immunfluorescent staining was also used to distinguish between non-parenchymal biliary epithelial-like cells and hepatocytes (Figure 2c). β-catenin and E-cadherin proteins appears in the HepaRG cell line only on the surface of mature hepatocytes [30,35]. Immunostaining also supported the reduction of hepatocyte islands at 20 mM APAP (Figure 2c). Thus, the survival of non-parenchymal biliary epithelial-like cells at low APAP concentrations (up to 20 mM) masked hepatocyte-specific death assessed by MTT assay. However, the very high APAP concentration (80 mM) is toxic for the non-parenchymal biliary epithelial-like cells, too (due to nonspecific reasons such as hyperosmolarity). concentration (80 mM) is toxic for the non-parenchymal biliary epithelial-like cells, too (due to nonspecific reasons such as hyperosmolarity). . PI-positivity (orange) as a characteristic cell death nuclear staining appears exclusively in hepatocyte islands in both 2D and 3D (a,b). Immunostaining of mature hepatocytes for E-cadherin and β-catenin (green) also confirms that characteristic hepatocyte/biliary-like phenotype in . PI-positivity (orange) as a characteristic cell death nuclear staining appears exclusively in hepatocyte islands in both 2D and 3D (a,b). Immunostaining of mature hepatocytes for E-cadherin and β-catenin (green) also confirms that characteristic hepatocyte/biliary-like phenotype in control cultures is progressively lost with increasing APAP concentration (c). Hoechst nuclear counter stain, blue. Scale bars: 100 µm.
By monitoring AST release (Figure 1b,d), APAP toxicity could be measured in both HepG2 and differentiated HepaRG cultures, and increased sensitivity toward hepatocytespecific cell death was achieved in the latter case. Although both cell lines show a compa-Life 2021, 11, 856 9 of 20 rable level of AST activity, in HepaRG, the enzyme activity originates from only a fraction of the population, as hepatocytes constitute approximately 50% of the cells. Thus, the application of the determination of AST activity in non-solely hepatocyte-containing cultures as a cell death (viability) marker is suggested.

Investigation of Cell Death Pathways by Acetaminophen Cytotoxicity in HepG2 and Differentiated HepaRG
The nature and exact mechanism of programmed cell death pathways involved in APAP toxicity in different in vitro and in vivo models are still under investigation. Apoptosis, necroptosis, and ferroptosis are all considered to play a role either distinctly or in parallel (apoptosis and necroptosis) [8,9,[48][49][50]. Furthermore, the cell death pathways involved in APAP-induced hepatocyte cell death are characteristic [8]. Through their characteristic combination, the investigated in vitro and in vivo models can be qualified.
In order to investigate the programmed cell death initiated by APAP in our experimental setup, we assessed the potential of known specific cell death inhibitors in alleviating cytotoxicity. HepG2 and differentiated HepaRG cells cultured in a monolayer were treated with 15 mM APAP in the presence and absence of various cell death inhibitors; then, cell death was measured by the AST assay ( Figure 3).
By monitoring AST release (Figure 1b,d), APAP toxicity could be measured in both HepG2 and differentiated HepaRG cultures, and increased sensitivity toward hepatocytespecific cell death was achieved in the latter case. Although both cell lines show a comparable level of AST activity, in HepaRG, the enzyme activity originates from only a fraction of the population, as hepatocytes constitute approximately 50% of the cells. Thus, the application of the determination of AST activity in non-solely hepatocyte-containing cultures as a cell death (viability) marker is suggested.

Investigation of Cell Death Pathways by Acetaminophen Cytotoxicity in HepG2 and Differentiated HepaRG
The nature and exact mechanism of programmed cell death pathways involved in APAP toxicity in different in vitro and in vivo models are still under investigation. Apoptosis, necroptosis, and ferroptosis are all considered to play a role either distinctly or in parallel (apoptosis and necroptosis) [8,9,[48][49][50]. Furthermore, the cell death pathways involved in APAP-induced hepatocyte cell death are characteristic [8]. Through their characteristic combination, the investigated in vitro and in vivo models can be qualified.
In order to investigate the programmed cell death initiated by APAP in our experimental setup, we assessed the potential of known specific cell death inhibitors in alleviating cytotoxicity. HepG2 and differentiated HepaRG cells cultured in a monolayer were treated with 15 mM APAP in the presence and absence of various cell death inhibitors; then, cell death was measured by the AST assay ( Figure 3).  Table A1. Data are normalized to untreated, and each data point represents the average ± SD of at least three independent experiments. * significantly different (p < 0.05) from untreated (0 mM acetaminophen); # significantly different (p < 0.05) from group control (15 mM acetaminophen + vehicle-treated).
Our results show that the pan-caspase inhibitor zVAD-fmk [28] and dabrafenib [51] significantly protected both cell lines from APAP-induced cell death (Figure 3). Since distinct molecular pathways govern caspase-dependent apoptosis and necroptosis, the activation of characteristic proteins was also investigated. The caspase-mediated cleavage of the 113 kDa nuclear enzyme Poly (ADP-ribose) polymerase 1 (PARP-1) to an 89 and a 24  Table A1. Data are normalized to untreated, and each data point represents the average ± SD of at least three independent experiments. * significantly different (p < 0.05) from untreated (0 mM acetaminophen); # significantly different (p < 0.05) from group control (15 mM acetaminophen + vehicle-treated).
Our results show that the pan-caspase inhibitor zVAD-fmk [28] and dabrafenib [51] significantly protected both cell lines from APAP-induced cell death (Figure 3). Since distinct molecular pathways govern caspase-dependent apoptosis and necroptosis, the activation of characteristic proteins was also investigated. The caspase-mediated cleavage of the 113 kDa nuclear enzyme Poly (ADP-ribose) polymerase 1 (PARP-1) to an 89 and a 24 kDa fragment is a known hallmark of apoptosis. PARP-1 cleavage was determined in APAP-treated HepG2 and HepaRG cells (Figure 4, left panels). The 89 kDa cleaved PARP1 fragment appeared in both cell lines upon APAP treatment but more markedly in HepG2 (Figure 4, left panels). kDa fragment is a known hallmark of apoptosis. PARP-1 cleavage was determined in APAP-treated HepG2 and HepaRG cells (Figure 4, left panels). The 89 kDa cleaved PARP1 fragment appeared in both cell lines upon APAP treatment but more markedly in HepG2 (Figure 4, left panels). What can be in the background of the effect of dabrafenib in the alleviation of the hepatotoxic effect of APAP?
Since RIPK3 was considered to play a key role in APAP-induced hepatotoxicity [49] and dabrafenib showed strong inhibition on RIPK3 [51], our first thought was that dabrafenib inhibited RIPK3 in our HepaRG cultures, too. However, we could not find any RIPK3 expression neither at mRNA nor at protein levels in our cultures (data not shown). Furthermore, the role of RIPK3 in APAP-induced hepatotoxicity has been the matter of intense debate [52]. Hence, the inhibitory role of dabrafenib on RIPK3 must be ruled out. At the same time, sterile-alpha motif and leucine zipper containing kinase (ZAK), a member of the MAP3K family, is known to be involved in apoptosis [53]. The overexpression of ZAK could induce apoptosis in human OS cells [53]. Furthermore, it was found that the bioflavonoid fisetin could upregulate the expression of ZAK that mediated the activation of the downstream JNK/ERK pathway, thus triggering cell apoptosis in an AP-1-dependent manner in human OS cells [53]. It was also described previously that dabrafenib inhibited ZAK kinase [54]. What can be in the background of the effect of dabrafenib in the alleviation of the hepatotoxic effect of APAP?
Since RIPK3 was considered to play a key role in APAP-induced hepatotoxicity [49] and dabrafenib showed strong inhibition on RIPK3 [51], our first thought was that dabrafenib inhibited RIPK3 in our HepaRG cultures, too. However, we could not find any RIPK3 expression neither at mRNA nor at protein levels in our cultures (data not shown). Furthermore, the role of RIPK3 in APAP-induced hepatotoxicity has been the matter of intense debate [52]. Hence, the inhibitory role of dabrafenib on RIPK3 must be ruled out. At the same time, sterile-alpha motif and leucine zipper containing kinase (ZAK), a member of the MAP3K family, is known to be involved in apoptosis [53]. The overexpression of ZAK could induce apoptosis in human OS cells [53]. Furthermore, it was found that the bioflavonoid fisetin could upregulate the expression of ZAK that mediated the activation of the downstream JNK/ERK pathway, thus triggering cell apoptosis in an AP-1-dependent manner in human OS cells [53]. It was also described previously that dabrafenib inhibited ZAK kinase [54].
Thus, the c-Jun N-terminal kinase (JNK) activated total c-Jun was determined in the presence/absence of dabrafenib and APAP by Western blot (Figure 4, right panels). Indeed, an increase in total c-Jun was found by increasing APAP concentrations, which could be diminished by dabrafenib treatment to levels below those of the untreated samples. In addition, ZAK is expressed most prominently in liver, it signals to JNK through MKK4 [54], and MKK4 is the major MAP2K, which activates JNK in acute liver injury [55]. Furthermore, SIRT2-mediated deacetylation favors the phosphorylation of JNK by MKK4. Hence, it was not surprising that SIRT2-KO mice exhibited increased acetylation of JNK, which was associated with significantly reduced JNK activity in the liver. Consequently, SIRT2-KO mice showed lower cell death, minimal degenerative changes, improved liver function, and survival following APAP treatment [56]. All these observations together with our results may support the finding that dabrafenib can exert its hepatoprotective effect through the inhibition of ZAK and the following JNK pathway. Although the hallmark of RIPKdependent necrosis, RIPK1 phosphorylation was shown by APAP treatment, and RIPK1 inhibition decreased reactive oxygen species (ROS) levels produced in APAP-injured hepatocytes in an animal model [57]; furthermore, the RIPK1 inhibitor Nec-1 also decreased the rate of hepatotoxicity in primary mice hepatocytes [9]. Both Nec-1 and Nec-2 showed no effect on our HepG2 cells and had only marginal beneficial effects in our HepaRG cultures (Figure 3). APAP treatment of both HepG2 and HepaRG resulted in PARP cleavage and the appearance of the 89 kDa fragment, further supporting the involvement of apoptosis (Figure 4, left panels).
Aside fromthe differences in hepatocyte function, the observations that (1) known specific inhibitors of necroptosis (necrostatin-1 and MDIVI) were only effective in differentiated HepaRG and (2) the degree of protection of zVAD-fmk was higher in HepG2 than in HepaRG suggest a differential execution of activated pathways. Specific inhibitors of ferroptosis (ferrostatin-1 and liproxstatin-1) were ineffective against cell viability loss in both cell lines, on the contrary of their effectiveness using in vitro and in vivo mouse models [9,10]. On the basis of all these observations, it seems HepaRG stands somewhere between HepG2 and primary hepatocytes from the hepatocyte functional point of view. However, it should be noted that the maintenance of HepaRG has some advantages over the maintenance of primary hepatocytes. A major drawback of primary hepatocytes is their limited lifespan. Isolated and in vitro cultured hepatocytes do not expand and gradually dedifferentiate, resulting in the loss of their liver-specific functions [58,59]. Hence, primary hepatocytes can only be used for a few days, making them unsuitable for long-term experiments. Primary human hepatocytes (PHHs) are considered to be the gold standard to study drug-induced liver injury, but because PHHs are usually isolated from whole livers or resected liver tissue [60], the availability of these cells is limited, and due to the interindividual variability between donors, they can differ significantly in drug response [32,[61][62][63]. On the other hand, HepaRG cells are relatively immortal, and they have a stable phenotype and CYP450 expression over time. These properties of HepaRG allow us to grow identical cells in virtually unlimited amounts [32,61,64].

Caspase Activity and GSH Level in APAP-Treated HepG2 and Differentiated HepaRG Cells
The pan-caspase inhibitor zVAD-fmk was able to protect both HepG2 and differentiated HepaRG from APAP-induced cell death (Figure 3). Thus, caspase activation was also investigated. Both cell lines were cultured in a monolayer and treated with increasing concentrations of APAP or 15 mM APAP in the presence or absence of an inhibitor. Caspase 3/7 activity was determined by fluorimetry ( Figure 5, top graphs) [65]. We also investigated real-time caspase activation by a fluorescently labeled caspase substrate using fluorescence microscopy ( Figure 5, bottom images).
APAP-induced caspase activation was concentration-dependent in both cell lines, further supporting the role of apoptotic mechanisms. As it could be expected, the presence of dabrafenib significantly decreased caspase activity. In parallel, an increase of the fluorogenic caspase 3/7 substrate CellEvent™ was observed in HepaRG, which could be inhibited by dabrafenib. This observation further reinforces our above detailed assumption on the possible role of dabrafenib in the inhibition of apoptosis via its inhibitory role on ZAK [54]. APAP-induced caspase activation was concentration-dependent in both cell lines, further supporting the role of apoptotic mechanisms. As it could be expected, the presence of dabrafenib significantly decreased caspase activity. In parallel, an increase of the fluorogenic caspase 3/7 substrate CellEvent™ was observed in HepaRG, which could be inhibited by dabrafenib. This observation further reinforces our above detailed assumption on the possible role of dabrafenib in the inhibition of apoptosis via its inhibitory role on ZAK [54].
From the methodological perspective, it should be emphasized that to assess the degree of caspase activation in the HepaRG culture correctly, incorporating both cells and cellular fragments/debris was crucial; otherwise, cellular structures found to be positive for caspase activity could be easily lost during washing steps. From the methodological perspective, it should be emphasized that to assess the degree of caspase activation in the HepaRG culture correctly, incorporating both cells and cellular fragments/debris was crucial; otherwise, cellular structures found to be positive for caspase activity could be easily lost during washing steps.
Conjugation with glutathione is an important moment of hepatic APAP metabolism [44]. At lower doses, APAP biotransformation proceeds without physiological disturbance; however, higher doses cause glutathione depletion, which leads to oxidative stress and oxidative damage, initiating signaling pathways that can drive the cell to programmed cell death [44]. Consequently, the level of reduced cellular glutathione is a suitable marker for monitoring APAP metabolism in hepatocytes. Therefore, the reduced form of cellular glutathione was determined in monolayer cultured HepG2 and differentiated HepaRG ( Figure 6). ance; however, higher doses cause glutathione depletion, which leads to oxidative stress and oxidative damage, initiating signaling pathways that can drive the cell to programmed cell death [44]. Consequently, the level of reduced cellular glutathione is a suitable marker for monitoring APAP metabolism in hepatocytes. Therefore, the reduced form of cellular glutathione was determined in monolayer cultured HepG2 and differentiated HepaRG (Figure 6). Glutathione decreased in both cell lines, with a more pronounced decrease seen in HepaRG since 15 mM APAP halved the cellular reduced glutathione pool. This observation highlights again that HepaRG has kept its hepatic function to a greater extent than HepG2, and it is more suitable for toxicological studies. It is also important to emphasize that normalization of the measured glutathione by cell count or protein concentration can bias the results toward surviving biliary epithelial-like cells. In order to visualize the differential depletion of glutathione among the cell types present in differentiated HepaRG culture, we labeled APAP-treated cells with a thiol-tracking probe ( Figure 6, right images). Glutathione decreased in both cell lines, with a more pronounced decrease seen in HepaRG since 15 mM APAP halved the cellular reduced glutathione pool. This observation highlights again that HepaRG has kept its hepatic function to a greater extent than HepG2, and it is more suitable for toxicological studies. It is also important to emphasize that normalization of the measured glutathione by cell count or protein concentration can bias the results toward surviving biliary epithelial-like cells. In order to visualize the differential depletion of glutathione among the cell types present in differentiated HepaRG culture, we labeled APAP-treated cells with a thiol-tracking probe ( Figure 6, right images).
Live cell fluorescent imaging revealed intensive labeling of hepatocyte islets in untreated cells (Figure 6, right images), which consistently with the hepatic phenotype contain the highest concentration of cellular glutathione among mammalian cells [66,67]. Glutathione within hepatocyte islets showed a proportional decrease with increasing APAP concentrations and approached that achieved by buthionine sulfoximine (BSO) depletion. These observations further confirm the hepatocyte-mediated metabolism of APAP and the accompanying reduction of cellular glutathione.

The Effect of 3D Culture Techniques (Spheroid and Nanofiber) on Acetaminophen Cytotoxicity in HepG2 and Differentiated HepaRG Cells
The efficient metabolism of APAP corresponds to the level of phase I enzymes in hepatocytes. Most frequently, the dominating role in the conversion of APAP to the highly reactive metabolite NAPQI is ascribed to the isoform CYP2E1 [28,68]. HepG2 and differentiated HepaRG are known to possess a different degree of hepatic functions; this difference also expands to the level of enzymes involved in the process of drug metabolism [69]. Developments in cell culture techniques aim at narrowing the gap between in vitro and in vivo models. Regarding hepatic in vitro models, 3D culture methods are extensively used to increase hepatic function [19][20][21][22][23][24][25][26][27].
We aimed at the investigation of the effect of two commonly used 3D culture methodsspheroid and nanofiber culture-on the hepatic function and APAP sensitivity of HepG2 and HepaRG. Cells were cultured in 3D by two methods as shown, and CYP2E1 mRNA was quantified; in the case of HepaRG, the level of CYP2E1 was also measured throughout the differentiation process (Figure 7, upper panels). tathione within hepatocyte islets showed a proportional decrease with increasing APAP concentrations and approached that achieved by buthionine sulfoximine (BSO) depletion. These observations further confirm the hepatocyte-mediated metabolism of APAP and the accompanying reduction of cellular glutathione.

The Effect of 3D Culture Techniques (Spheroid and Nanofiber) on Acetaminophen Cytotoxicity in HepG2 and Differentiated HepaRG Cells
The efficient metabolism of APAP corresponds to the level of phase I enzymes in hepatocytes. Most frequently, the dominating role in the conversion of APAP to the highly reactive metabolite NAPQI is ascribed to the isoform CYP2E1 [28,68]. HepG2 and differentiated HepaRG are known to possess a different degree of hepatic functions; this difference also expands to the level of enzymes involved in the process of drug metabolism [69]. Developments in cell culture techniques aim at narrowing the gap between in vitro and in vivo models. Regarding hepatic in vitro models, 3D culture methods are extensively used to increase hepatic function [19][20][21][22][23][24][25][26][27].
We aimed at the investigation of the effect of two commonly used 3D culture methods-spheroid and nanofiber culture-on the hepatic function and APAP sensitivity of HepG2 and HepaRG. Cells were cultured in 3D by two methods as shown, and CYP2E1 mRNA was quantified; in the case of HepaRG, the level of CYP2E1 was also measured throughout the differentiation process (Figure 7, upper panels). Figure 7. The effect of 3D culture techniques (spheroid and nanofiber) on CYP2E1 mRNA expression and APAP-induced cell death in HepG2 and differentiated HepaRG (a). CYP2E1 mRNA was determined by real-time RT-PCR from both cell lines cultured either in a 2D monolayer or 3D (spheroid or nanofiber). Each CYP2E1 expression is normalized to the expression of the 2D cultured HepG2 line. In the case of HepaRG, CYP2E1 was also monitored throughout the differentiation process (9 and 28 days). Cell death induced by different concentrations of acetaminophen (APAP, 0 mM-untreated, 10 mM, 15 mM, and 20 mM) or 15 mM acetaminophen and inhibitors (zVAD-fmk 40 μM, dabrafenib (Dabr) 10 μM, necrostatin-1 (Nec-1) 50 μM, or liproxstatin-1 (Lipr-1) 1 μM) was measured by the AST assay (b). Data are normalized to Figure 7. The effect of 3D culture techniques (spheroid and nanofiber) on CYP2E1 mRNA expression and APAP-induced cell death in HepG2 and differentiated HepaRG (a). CYP2E1 mRNA was determined by real-time RT-PCR from both cell lines cultured either in a 2D monolayer or 3D (spheroid or nanofiber). Each CYP2E1 expression is normalized to the expression of the 2D cultured HepG2 line. In the case of HepaRG, CYP2E1 was also monitored throughout the differentiation process (9 and 28 days). Cell death induced by different concentrations of acetaminophen (APAP, 0 mM-untreated, 10 mM, 15 mM, and 20 mM) or 15 mM acetaminophen and inhibitors (zVAD-fmk 40 µM, dabrafenib (Dabr) 10 µM, necrostatin-1 (Nec-1) 50 µM, or liproxstatin-1 (Lipr-1) 1 µM) was measured by the AST assay (b). Data are normalized to untreated (0 mM), and each data point represents the average ± SD of at least three independent experiments. * significantly different (p < 0.05) from untreated (0 mM APAP); # significantly different (p < 0.05) from 15 mM APAP.
Regarding 2D culture, undifferentiated HepaRG expressed 10 times more CYP2E1 mRNA than HepG2, and it increased further by 100-fold (Figure 7a) at the end of the differentiation.
By applying 3D culture methods, CYP2E1 mRNA levels could also be elevated. In HepG2, nanofiber culture resulted in the most robust increase, while in differentiated HepaRG, both 3D methods induced CYP2E1 expression to a similar degree (Figure 7a). Despite the effectiveness achieved by the nanofiber method in HepG2, the resulting mRNA levels were closest to those measured in undifferentiated HepaRG (9 days, Figure 7a). The elevated level of CYP2E1 mRNA in 3D cultured HepG2 could also be the result of prolonged cultivation, as HepG2 grown for 21 days in 2D monolayer showed a similar increase in hepatic function [70].
Based on the effectiveness in the induction of CYP2E1 by 3D culture-observed most prominently in HepaRG-we were interested in whether it also increases APAP sensitivity. Thus, HepaRG was differentiated on nanofiber plates and treated with different concentrations of APAP. To investigate for a deviation in the cytotoxic profile, cells were also incubated in the presence or absence of inhibitors of interest (Figure 7b).
APAP increased AST release in a dose-dependent manner; however, the extent was not significantly different compared to 2D cultured differentiated HepaRG. The applied inhibitors were also protective to a similar degree, supposing no difference in sensitivity nor the cytotoxic profile despite increased levels of CYP2E1.
Summarily, the hepatic functions of HepaRG stand closer to those of primary hepatocytes, but the 3D cultivation, especially in APAP toxicity studies, is not necessarily worth the more complicated and expensive maintenance.

Conclusions
The investigation of drug-induced hepatotoxicity and the prediction of clinical druginduced liver damage requires appropriate in vivo and in vitro toxicological model systems. In our study, an attempt was made to compare different three-dimensional and stem cellderived models to find the most appropriate one.
The degree of liver-specific characteristics of HepG2 and differentiated HepaRG lines via the extent of APAP-induced hepatotoxicity was followed.
The MTT assay in HepG2 resulted in a toxicity profile in accordance with our expectations and previous observations. However, our results revealed that it underperformed with HepaRG. In the case of the HepaRG line, the survival of non-parenchymal biliary epithelial-like cells at low APAP concentrations masked hepatocyte-specific death assessed by MTT assay. Thus, the application of the determination of AST activity in non-solely hepatocyte containing cultures as a cell death (viability) marker is suggested.
Since the pathways involved in APAP-induced hepatocyte cell death are characteristic, the investigated in vitro and in vivo models can be qualified by them. Thus, the effect of known specific cell death inhibitors and the activation of characteristic proteins was investigated. The pan-caspase inhibitor zVAD-fmk and dabrafenib significantly protected both cell lines from APAP-induced cell death. The differences in hepatocyte function and the observations that (1) known specific inhibitors of necroptosis (necrostatin-1 and MDIVI) were only effective in differentiated HepaRG and (2) the degree of protection of zVADfmk was higher in HepG2 than in HepaRG suggest a differential execution of activated pathways. On the basis of all these observations, it seems HepaRG stands somewhere between HepG2 and primary hepatocytes from the hepatocyte functional point of view. However, it should be noted that the maintenance of HepaRG has more advantages over the maintenance of primary hepatocytes such as relative immortality, stable phenotype, and CYP450 expression. These properties of HepaRG allow us to grow identical cells in virtually unlimited amounts.
By applying 3D culture methods, CYP2E1 mRNA levels could also be elevated. In HepG2, nanofiber culture resulted in the most robust increase, while in differentiated HepaRG, both 3D methods induced CYP2E1 expression to a similar degree. Albeit APAP treatment increased the AST release in a dose-dependent manner of 3D cultured HepaRG cells, it was not significantly different compared to 2D cultured differentiated HepaRG.
Summarily, the hepatic functions of HepaRG stand closer to those of primary hepatocytes, but the 3D cultivation, especially in APAP toxicity studies, is not necessarily worth the more complicated and expensive maintenance.

Data Availability Statement:
The data presented in this study are available on request from the corresponding author. The data are not publicly available due to the Material Transfer Agreement of the University of Technology and Economics and Inserm Transfert SA on the HepaRG cell line.
Acknowledgments: Anti-P-c-Jun (Cell Signaling, 9261S), Anti-c-Jun (Cell Signaling, 9165S) antibody was a generous gift from Orsolya Kapuy (Semmelweis Egyetem, Budapest, Hungary). The authors are grateful for the kind technical support for Viktor Podhragyai.

Conflicts of Interest:
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript, or in the decision to publish the results.