The use of cultured human hepatocytes as a test system to evaluate cell proliferation as a key event in nongenotoxic carcinogenesis Authors: David E Cowie, Samuel M Cohen, Manuela Goettel, Brian G Lake, Stephanie Nadzialek, Uku Rooni, Maaike E Schutte, Helen Tinwell, Josh Turiccki, Tomoya Yamada, Christian Strupp Published: 2026 ABSTRACT AbstractThe objective of this study was to evaluate the applicability of replicative DNA synthesis (RDS) in cultured human hepatocytes as an in vitro model to evaluate the carcinogenicity of nongenotoxic chemicals and species differences. Investigations were performed with 39 cryopreserved human hepatocyte preparations from predominantly Caucasian male and female donors aged 10 months to 80 years and were conducted by two separate laboratories, which employed different culture conditions and methodology for evaluating effects on hepatocyte RDS. For all male and female human hepatocyte preparations of all ages examined, treatment with either epidermal growth factor (EGF) and/or hepatocyte growth factor (HGF) resulted in a stimulation of RDS. In contrast, the treatment of human hepatocytes with the constitutive androstane receptor (CAR) activators phenobarbital and CITCO and the peroxisome proliferator-activated receptor alpha (PPARα) activator WY-14,643 did not result in any increases in RDS. These findings are in agreement with previous studies where, unlike EGF and HGF, nongenotoxic CAR and PPARα activators are mitogenic agents in rodent but not in human hepatocytes. While some donor to donor variability was observed, the qualitative inducibility of RDS by EGF and/or HGF in human hepatocytes was not sex-, age-, or, based on a limited number of samples examined, ethnicity-dependent. These studies demonstrate that cultured cryopreserved human hepatocytes are an established, reproducible and relevant in vitro test system for investigating the nongenotoxic carcinogenic potential of chemicals and species differences, and worth progressing to formal validation according to OECD principles.Supplementary InformationThe online version contains supplementary material available at 10.1007/s00204-026-04411-1.Keywords: Human hepatocytes, Replicative DNA synthesis, In vitro models, Species differences in liver tumour formation, Constitutive androstane receptor activators, Peroxisome proliferator-activated receptor alpha activators, Carcinogenicity, Cell proliferation, New approach method validation Introduction For many years the carcinogenic hazard of chemicals to humans has been assessed by long-term studies in experimental animals. Current regulations for plant protection products requires the assessment of carcinogenicity in two species (usually the rat and mouse) to derive data for human health decision making (European Commission 2009). The predictivity of the rodent cancer bioassay on chemicals for humans is not always a given and many publications have considered their potential usefulness and difficulties (Ames et al. 1990, 2017, 2024; Berry 2018; Cohen 2004, 2010; Cohen et al. 2019; Doe et al. 2019, 2022; Goodman 2018; Heusinkveld et al. 2020; Hilton et al. 2023; Luijten et al. 2020; Osimitz et al. 2013; Smith and Perfetti 2018; Terry et al. 2021). The relevance of data from chronic rodent bioassays for humans has been questioned on the basis of both the species extrapolation of cancer effects and in terms of dose extrapolation. Long-term rodent bioassays are normally conducted at up to maximum tolerated dose (MTD) levels, which may be orders of magnitude higher than potential human exposure and lead to non-specific mitogenicity at such doses. While tumours arising from a specific mode of action (MOA) are generally well reproducible in the rodent cancer bioassay, tumours at or above MTD dose levels are typically not reliable for human health assessment and are of limited scientific value (Rhomberg et al. 2007). Analysis of tumour incidence data from long-term rat and mouse bioassays demonstrates that the liver is the major site of chemically induced tumour formation in both species (Gold et al. 2001; Huff et al. 1991; Heusinkveld et al. 2020). Frameworks for analysing the modes of action (MOAs) by which chemicals produce tumours in experimental animals and the relevance of such animal tumours for human risk assessment have been developed by the International Life Sciences Institute and the International Programme on Chemical Safety (IPCS). Such MOA/human relevance frameworks are invaluable tools to permit human risk assessment of chemicals which produce tumours in experimental animals. The IPCS framework involves determining if it is possible to establish a MOA for tumour formation in experimental animals by identifying a series of key events (Boobis et al. 2006; Cohen et al. 2004; Meek et al. 2003, 2014a,b; Sonich-Mullin et al. 2001). To assess the human relevance of the proposed MOA, effects on key events are compared first qualitatively and then quantitatively between experimental animals and humans. Andersen et al. (2014) have proposed an extension to the basic IPCS framework where both key events and associative events are identified. For liver tumour formation mediated by xenobiotics a number of MOAs have been established, which can be divided into genotoxic (i.e. DNA damage as a molecular initiating event) and non-genotoxic MOAs (Holsapple et al. 2006; Cohen 2010; Cohen and Arnold 2011, 2016). Genotoxic agents, such as aflatoxin B1, have been shown to produce liver tumours in both experimental animals and humans (IARC 2012). MOAs for non-genotoxic agents can be divided into receptor mediated and non-receptor mediated MOAs, the latter including cytotoxicity, infection and metal overload. Two receptor mediated MOAs for rodent liver tumour formation comprise activation of the constitutive androstane receptor (CAR) and activation of the peroxisome proliferator-activated receptor alpha (PPARα). Robust MOAs have been established for both non-genotoxic CAR and PPARα activators which produce liver tumours in rats and/or mice (Corton et al. 2014, 2018; Elcombe et al. 2014; Klaunig et al. 2003; Lake 2018; Yamada et al. 2021, 2025). Examination of these two MOAs demonstrates a number of similarities (Yamada et al. 2025), the only major difference between these MOAs being the initial activation of the nuclear receptor (i.e. CAR or PPARα). As shown in Fig. 1, for both MOAs the initial key event is nuclear receptor activation, which results in altered gene expression specific to receptor activation. This leads to increased hepatocellular proliferation, which can be measured as increases in hepatocyte replicative DNA synthesis (RDS). Prolonged treatment then results in clonal expansion leading to altered hepatic foci and ultimately to the formation of hepatocellular adenomas and carcinomas. For both MOAs, associative events include altered expression of genes involved in hepatic metabolism, together with induction of hepatic cytochrome P450 (CYP) and other xenobiotic metabolising genes and liver hypertrophy. The European Partnership for Alternative Approaches to Animal testing (EPAA), conducted a retrospective evaluation of 411 agrochemicals which had been evaluated for carcinogenicity by the US Environmental Protection Agency (US EPA) and the European Chemicals Agency (ECHA); 170 were identified as non-genotoxic carcinogenic chemicals (Heusinkveld et al. 2020). Of the 107 chemicals producing liver tumours in the rat and/or mouse, 55 and 11 chemicals were demonstrated to have a CAR or a PPARα MOA, respectively. Schematic representation of the MOAs for rodent liver tumour formation by nongenotoxic CAR and PPARα activators. The initial key event is nuclear receptor (NR) activation (i.e. CAR or PPARα), with other key events shown below leading to the adverse outcome of liver tumour formation. Associative events shown to the right of the figure include increases in genes involved in hepatic metabolism and xenobiotic metabolism, together with liver hypertrophy and increased liver weight As demonstrated by the established MOAs for non-genotoxic rodent liver CAR and PPARα activators, increased cell proliferation has a key role in chemical carcinogenesis (Cohen 2024; Cohen and Arnold 2011; Peffer et al. 2018a; Strupp et al. 2023; Wolf et al. 2019; Wood et al. 2015; Yamada et al. 2025). Assessment of cell proliferation is considered an essential hallmark to be addressed in recent outputs from an OECD Expert Group dedicated to assessing Integrated Approaches for the Testing and Assessment (IATAs) of nongenotoxic chemical carcinogens (Jacobs et al. 2020). The effects of chemicals on cell proliferation in target organs can be assessed in both in vivo and in vitro studies employing various techniques (Strupp et al. 2023; Wood et al. 2015). Commonly used methods include administration of a nucleoside analogue, such as 5-bromo-2´-doxyuridine (BrdU), which is incorporated into newly synthesised DNA and detected by immunohistochemical staining, or immunohistochemical detection of Ki-67, a protein expressed during cell replication (interphase and specifically the mitotic phase of the cell cycle, but not when cells are not dividing [phase G0]). For the liver, experimental animals in in vivo studies and cultured hepatocytes in in vitro studies are treated with both the test chemical and BrdU or untreated in the case of Ki-67. Fixed tissue sections or hepatocytes are then processed for BrdU or Ki-67 immunostaining and the hepatocyte labelling index (i.e. percentage of hepatocyte nuclei undergoing RDS) is determined. For hepatocytes, BrdU incorporation can also be determined by an ELISA assay (Goll et al. 1999; Hirose et al. 2009). That chemicals may produce liver tumours in rat and mouse bioassays has led to additional studies in other species (e.g. primates) and in transgenic rats and mice (e.g. receptor knockout animals, mice with humanised livers), to attempt to address human safety. While such investigations may help establish a MOA, performing such studies for each individual chemical contradicts the principles of the 3Rs (Replacement, Reduction, and Refinement) and emphasises the need for the development of alternative approaches (Capelôa et al. 2024). Many studies have demonstrated that cultured hepatocytes are a useful in vitro model system to assess species differences in the effects of chemicals on RDS (Corton et al. 2018; Lake 2009; Yamada et al. 2021, 2025), and are considered relevant in regulatory decision making (Capelôa et al. 2024). The objective of the present study is to present robust evidence to further establish the usefulness of cultured human hepatocytes as an in vitro model system to evaluate the effects of chemicals on RDS in hepatocytes, in particular to address reproducibility and experimental variability in repeat experiments, as well as potential differences due to sex, age or ethnicity. We report data on the retrospective analysis of studies with a large number of individual human hepatocyte preparations from male and female donors of a wide range of ages. These studies have been performed with two CAR activators, namely phenobarbital (PB) and 6-(4-chlorophenyl)imidazo[2,1-b][1,3]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime) (CITCO). PB and its sodium salt (NaPB) are prototypical CAR activators which have been extensively studied in both experimental animals and humans and for which data from a number of human epidemiological studies are available (Elcombe et al. 2014; IARC 2001; Whysner et al. 1996; Yamada et al. 2021, 2025). CITCO is a potent and selective activator of human CAR and induces CYP2B6 mRNA levels in cultured human hepatocytes (Maglich et al. 2003; Omiecinski et al. 2011). Investigations were also performed with [4-chloro-6-(2,3-xylidino)-2-pyrimidinylthio]acetic acid (WY-14,643), which is a potent PPARα activator and non-genotoxic hepatocarcinogen in the rat and mouse (Corton et al. 2014, 2018; Yamada et al. 2025), and also with epidermal growth factor (EGF) and hepatocyte growth factor (HGF), which were employed as positive controls for increased RDS (Kang et al. 2012; Kimura et al. 2023). Non-receptor mediated MOAs (e.g. cytotoxicity triggering repair proliferation), and receptor-mediated effects different from CAR and PPARα were not evaluated and are not within the scope of this manuscript. Materials and methods Materials As described below, data were collected from studies performed in two laboratories, where the commercial sources of the cryopreserved human hepatocyte preparations, tissue culture reagents and other reagents, including dimethyl sulphoxide (DMSO), BrdU, PB and its sodium salt, CITCO, WY-14,643, EGF and HGF, are described. Data collection Data was collected from CropLife Europe (CLE) companies that conducted human hepatocyte studies as part of registration of active substances between 2011 and 2020. A total of 36 such studies were identified from the following CLE companies: ADAMA, BASF, Bayer, Gowan, Janssen PMP, Sumitomo Chemical Co Ltd and Syngenta. For all studies, details of the human hepatocyte preparations used, compound concentration ranges examined and other experimental details were collated, tabulated and analysed. Data visualisation was conducted and figures prepared using Spotfire (Gӧteborg, Sweden) version 14.0.5. Human hepatocytes Culture and treatment of human hepatocytes All the studies described in this paper were performed at either Concept Life Sciences (formerly CXR Biosciences Ltd; 2, James Lindsay Place, Dundee Technopole, Dundee, DD1 5JJ, UK) or Environmental Health Science Laboratory (Sumitomo Chemical Co Ltd, Konohana-ku, Osaka 554–8558, Japan), which are designated as Laboratory 1 and Laboratory 2, respectively. In a typical experiment, human hepatocytes were treated with a range of concentrations of the test compounds (including those which produced some cytotoxicity) and effects on hepatocyte RDS determined. Appropriate negative (e.g. medium containing the vehicle DMSO only) and positive (e.g. medium containing either EGF and/or HGF) controls were included in all studies. In some investigations, additional assays were performed to determine the effects of the test compounds on CYP enzyme activity and/or mRNA levels. While there were some differences in the design of the 36 individual human hepatocyte experiments (using a total of 39 individual human hepatocyte preparations as detailed in Table 1), some general conditions used by Laboratory 1 (Goettel et al. 2024; Haines et al. 2018; Lake et al. 2020; Wiemann et al. 2019) and Laboratory 2 (Kondo et al. 2020; Okuda et al. 2017; Yamada et al. 2015) are briefly detailed below. For Laboratory 1, cryopreserved human hepatocytes were thawed in hepatocyte thaw medium (Life Technologies, Warrington, Cheshire, UK) and viability determined by trypan blue exclusion. For RDS studies, human hepatocytes were cultured in collagen coated 6-well plates in cryopreserved hepatocyte recovery medium (Life Technologies) in a humidified incubator at 37 °C under an atmosphere of 95% air/5% CO2 for 6 h. At the end of the attachment period, treatment was commenced by changing the culture medium to Leibovitz L15 medium containing 50 μg/ml L-ascorbic acid and other additions as described by Plant et al. (1998) and cultured in a humidified incubator at 37 °C under air. If required, test compounds were added to the culture medium in DMSO at a final concentration of 0.1% (v/v), with DMSO being added to all media including control incubations and incubations containing PB, other test compounds and EGF. For Laboratory 1, due to compound solubility issues, some experiments were performed with a medium concentration of 0.5% (v/v) DMSO. Subsequently the medium was changed at 24h intervals to fresh medium containing the test compounds for a total of 96 h treatment. To study hepatocyte RDS 10 μM BrdU was added to the culture medium for the last 72 h of treatment. As a positive control for RDS studies human hepatocytes were also treated with EGF (normal concentration 25 ng/ml) for the last 72 h of treatment. At the end of the treatment period the human hepatocyte monolayers were fixed in methanol. For Laboratory 2, cryopreserved human hepatocytes were thawed according to the suppliers’ instructions and suspended in Williams’ medium E containing 2 mM L-glutamine, 0.1 μM bovine insulin, 1 μM dexamethasone, 10 mM nicotinamide, 0.2 mM L-ascorbic acid, 0.5 ng/ml EGF and 10% (v/v) fetal bovine serum. Viability was determined by trypan blue exclusion. Human hepatocytes were plated at a density of 3.5 × 104 viable cells/well in collagen coated 96-well plates and cultured in a humidified incubator at 37 °C under an atmosphere of 95% air/5% CO2 for either 18 or 24 h. Treatment was then commenced by changing the culture medium to serum free Williams´ medium E containing the above additions. If required, test compounds were added to the culture medium in DMSO at a final concentration of 0.1% (v/v), with DMSO being added to all media including control incubations and incubations containing PB, other test compounds, HGF and EGF. Subsequently the medium was changed after 24h to fresh medium containing the test compounds for a total of 48 h treatment. To study hepatocyte RDS 10 μM BrdU was added to the culture medium for the last 24 h of treatment. As a positive control for RDS studies human hepatocytes were also treated for 48 h with either EGF (concentration range 0.5–100 ng/ml) and/or HGF (concentration range 1–100 ng/ml). At the end of the treatment period the media was removed from the human hepatocyte monolayers, which were then air dried and processed as described below. Table 1Donor details and sources of the 39 cryopreserved human hepatocyte preparationsNo.Donor IDSexDonor ethnicityDonor age (years)Donor sourceLab 1Lab 21FCLFemaleHispanic0.8335 √ 2GHAFemaleCaucasian15 √ 3MMMFemaleCaucasian35 √ 4CEMMaleCaucasian114 √ 5SSRMaleCaucasian225 √ 6LLAMaleCaucasian265 √ 7BHLMaleCaucasian285 √ 8ZNNMaleCaucasian284 √ 9HFC498FemaleCaucasian351 √ 10HU1765FemaleCaucasian378 √ 11HU1552FemaleCaucasian388 √ 12LMPFemaleCaucasian385 √ 13HMC1034MaleCaucasian391 √ 14HMC3034MaleCaucasian391 √ 15HEP187076FemaleCaucasian413 √ 16HU4246MaleCaucasian466 √ 17SRTMaleCaucasian482 √ 18HU8287FemaleCaucasian498 √ 19IZBFemaleCaucasian494 √ 20HU8210MaleCaucasian518 √ √ 21HU8239FemaleCaucasian528 √ 22IPHFemaleCaucasian525 √ 23DSXFemaleHispanic564 √ 24HU1663MaleCaucasian568 √ 25HU1951FemaleCaucasian568 √ 26DOOMaleCaucasian575 √ 27HU4237FemaleCaucasian577 √ 28HU4238FemaleCaucasian576 √ 29CDPMaleCaucasian585 √ 30HU8219MaleCaucasian588 √ √ 31HU8196FemaleAfrican-American598 √ 32HEP187066FemaleCaucasian633 √ 33HFC429FemaleCaucasian641 √ 34HCF420FemaleCaucasian641 √ 35QOQMaleCaucasian665 √ 36HU1420FemaleCaucasian687 √ 37HU1474MaleCaucasian687 √ 38HU1258MaleCaucasian737 √ 39ETAFemaleCaucasian805 √ Male and female cryopreserved human hepatocytes were obtained from the commercial sources listed below, with individual preparations used by Laboratories 1 and 2 indicated1. BD Gentest, 6 Henshaw Street, Woburn, MA 01801, USA2. BIOIVT, 1450 South Rolling Road, Baltimore, MD 21227, USA3. Biopredic International, 8–18 Rue Jean Packer, 35,000 Rennes, France4. Bioreclamation IVT, 1450 South Rolling Road, Baltimore, MD 21227, USA5. Celsis IVT, 1450 South Rolling Road, Baltimore, MD 21227, USA6. GIBCO, 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK7. Invitrogen (GIBCO), 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK8. Life Technologies, 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK Donor details and sources of the 39 cryopreserved human hepatocyte preparations Male and female cryopreserved human hepatocytes were obtained from the commercial sources listed below, with individual preparations used by Laboratories 1 and 2 indicated 1. BD Gentest, 6 Henshaw Street, Woburn, MA 01801, USA 2. BIOIVT, 1450 South Rolling Road, Baltimore, MD 21227, USA 3. Biopredic International, 8–18 Rue Jean Packer, 35,000 Rennes, France 4. Bioreclamation IVT, 1450 South Rolling Road, Baltimore, MD 21227, USA 5. Celsis IVT, 1450 South Rolling Road, Baltimore, MD 21227, USA 6. GIBCO, 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK 7. Invitrogen (GIBCO), 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK 8. Life Technologies, 7 Kingsland Grange, Warrington, Cheshire, WA1 4SR, UK Determination of human hepatocyte RDS For Laboratory 1, the methanol fixed human hepatocyte monolayers were processed for BrdU immunocytochemistry, with 5 replicate wells being examined for all treatments (control, each concentration of the test compounds and EGF). Hepatocyte RDS was determined as the labelling index (i.e. the percentage of hepatocyte nuclei undergoing RDS). For Laboratory 2, the effect of the test compounds on human hepatocyte RDS was determined with a Cell Proliferation ELISA BrdU (chemiluminescent) kit (Roche, Mannheim, Germany). The air dried human hepatocyte monolayers were fixed and processed according to the kit manufacturer’s instructions, with 4, 8 or 12 replicate wells being examined for all treatments (i.e. control, each concentration of the test compounds, EGF and HGF). The luminescence of the samples was measured with a luminometer and hepatocyte RDS calculated as a percentage of control from the luminescent intensity of the treated samples compared to untreated controls. To account for the different methodologies used by Laboratory 1 and Laboratory 2, in the overall data set human hepatocyte RDS values were expressed as fold change compared to the concurrent control values. Additional assays For Laboratory 1, the potential cytotoxicity of test compounds to cultured human hepatocytes was determined by measurement of ATP levels using a CellTitre-Glo luminescent assay kit (Promega, Southampton, UK). In studies conducted by Laboratory 2, the potential cytotoxicity of test compounds to cultured human hepatocytes was determined by either the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyl tetrazolium bromide (MTT) assay or the (2-[2-methoxy-4-nitrophenyl]-3-[4nitrophenyl]-5-[2,4-disulfophenyl]-2H-tetrazolium, monosodium salt (WST-8) assay (Dojindo Laboratories, Kumamoto, Japan). In some studies performed by Laboratory 1, effects on some CYP enzymes, comprising 7-pentoxyresorufin O-depentylase (PROD), 7-benzyloxyresorufin O-debenzylase (BROD) and 7-benzyloxyquinolone O-debenzylase (BQ), activities were determined, together with effects on CYP2B6 and CYP3A4 mRNA levels. Additional studies conducted by Laboratory 2 included effects on CYP2B6 and CYP4A11 mRNA levels. Results Human donor demographics Details of the 39 cryopreserved human hepatocyte preparations used by either Laboratory 1 and/or Laboratory 2 in the studies conducted for the CLE companies are shown in Table 1. Of the 39 human hepatocyte preparations examined, 17 were from male and 22 were from female donors. The majority of the donors were of Caucasian origin (36/39), with two Hispanic donors and a single African-American donor. For all subjects, donor age ranges were 11–73 years for male donors and 10 months-80 years for female subjects. Donor age ranges were 11–73 years and 1–80 years for male and female Caucasian subjects, respectively, 10 months and 56 years for the two Hispanic female subjects and 59 years for the one female African-American subject. For all subjects, the donor numbers aged 0–20, 21–30, 31–40, 41–50, 51–60, 61–70 and > 70 years were 4, 4, 6, 5, 12, 6 and 2, respectively, with some 74% of the donors being aged 31–70 years (Table 1). Effect of EGF and HGF on RDS in cultured human hepatocytes Data for the response of the 39 human hepatocyte preparations used in the 36 CLE studies performed employing either EGF (concentration range 0.5–100 ng/ml) and/or HGF (concentration range 1–100 ng/ml) as positive controls for induction of hepatocyte RDS is shown in Fig. 2. Where required, test compounds were dissolved in DMSO (final concentration 0.1 or 0.5% v/v), which was added to all media including human hepatocytes treated with control medium and either EGF or HGF. As with the other investigations (Figs. 3–7), human hepatocyte labelling index values were very similar in controls treated with either 0.1 or 0.5% (v/v) DMSO. Due to differences in the design of the individual studies, the concentrations of EGF and/or HGF varied between individual experiments and hence there was some variation in the number of data points available for each concentration of EGF and HGF and also for the CAR and PPARα activators used (Figs. 2–7). As shown in Fig. 2, treatment with 1–25 ng/ml EGF and 1–100 ng/ml HGF produced concentration-dependent increases in hepatocyte RDS. Maximal median increases in hepatocyte RDS to 9.4 (range 3.5–28.2) and 3.3 (range 1.3–10.2) fold control were observed with 25 ng/ml EGF and 100 ng/ml HGF, respectively. In the individual CLE hepatocyte studies, statistically significant increases in hepatocyte RDS were normally observed after treatment with concentrations of ≥ 10 ng/ml of either EGF or HGF and in some studies with a concentration of 1 ng/ml EGF. Effect of treatment with control medium (containing either 0.1 or 0.5% DMSO), 0.5–100 ng/ml EGF and 1–100 ng/ml HGF on replicative DNA synthesis (RDS) in cultured male and female human hepatocytes from 39 donors. Results are presented as box and whisker plots expressed as median fold change in hepatocyte RDS plotted on a logarithmic scale compared to concurrent control values. Horizontal lines bisecting the rectangles represent median values, with the rectangles representing lower and upper quartiles. The whiskers represent the farthest data point lying within 1.5 × the inter-quartile range (IQR) from the box, the upper whisker being the upper quartile range + 1.5 * the inter quartile range, and the lower whisker being the lower quartile – 1.5 * the inter quartile range. Individual human hepatocyte experiments classified as outliers are indicated by dots Effect of treatment with control medium (containing either 0.1 or 0.5% DMSO), 0.5–100 ng/ml EGF and 1–100 ng/ml HGF on replicative DNA synthesis (RDS) in cultured male (red symbols) and female (blue symbols) human hepatocytes from 39 donors aged 10 months to 80 years. Results are presented for individual human hepatocyte preparations as mean values for each treatment expressed as fold change in hepatocyte RDS plotted on a logarithmic scale compared to concurrent control values Correlation between the effect of treatment of male and female human hepatocytes with control medium (containing either 0.1 or 0.5% DMSO) and 5–1000 μM PB on replicative DNA synthesis (RDS), expressed as median fold change compared to concurrent control, with effects on PROD (r2 0.01, A), BROD (r2 0.00, B) and BQ (r2 0.01, C) enzyme activities, expressed as pmol/min/mg protein. Data are presented as effects of DMSO or individual concentrations of PB in studies with 15 donors for PROD and BROD activities and 6 donors for BQ activity Figure 3 shows the effect of treatment with either 0.5–100 ng/ml EGF or 1–100 ng/ml HGF on hepatocyte RDS in 39 male and female subjects of various ages. As shown in Table 1, relatively few donors were < 20 or > 70 years old, with the majority of the liver donors being aged 31–70 years. However, while there was some variability between subjects, increases in hepatocyte RDS following treatment with either EGF or HGF were observed in hepatocytes from the 39 male and female subjects of all ages examined (Fig. 3). For laboratory 2 most studies were performed with either 1 or 10 ng/ml EGF, with analysis of the data (two-way analysis of variance) revealing no significant interactions between either male and female subjects or with donor age. Effect of CAR and PPARα activators on RDS in cultured human hepatocytes The effect of treatment of male and female human hepatocytes with the CAR activators PB (combined data for 5–1000 μM concentrations) and CITCO (100 μM); the PPARα activator WY-14,643 (50 and 100 μM) and with EGF (combined data for 0.5–100 ng/ml concentrations) and HGF (combined data for 1–100 ng/ml concentrations) as positive controls is shown in Fig. 4. Hepatocyte RDS median fold induction values after treatment with all concentrations of PB, 100 μM CITCO, 50 and 100 μM WY-14,643 were unchanged from control levels, being 1.0, 1.1, 1.0 and 0.9, respectively. In contrast, the treatment of human hepatocytes with 0.5–100 ng/ml EGF and 1–100 ng/ml HGF increased hepatocyte RDS median fold induction values to 3.5 and 2.1–fold control, respectively. Effect of treatment with control medium (containing either 0.1 or 0.5% DMSO), 0.5–100 ng/ml EGF (all concentrations combined), 1–100 ng/ml HGF (all concentrations combined), 5–1000 μM PB (all concentrations combined), 100 μM CITCO, 50 and 100 μM WY-14,643 (WY) on replicative DNA synthesis (RDS) in cultured male and female human hepatocytes from 39 donors. Results are presented as box and whisker plots expressed as median fold change in hepatocyte RDS plotted on a logarithmic scale compared to concurrent control values. Horizontal lines bisecting the rectangles represent median values, with the rectangles representing lower and upper quartiles. The whiskers represent the farthest data point lying within 1.5 × the inter-quartile range (IQR) from the box, the upper whisker being the upper quartile range + 1.5 * the inter quartile range, and the lower whisker being the lower quartile – 1.5 * the inter quartile range. Individual human hepatocyte experiments classified as outliers are indicated by dots While Fig. 4 presents combined data for all concentrations of PB examined, results for the effect of the individual concentrations of PB on RDS in male and female human hepatocytes is shown in Fig. 5. Unlike studies with EGF and HGF (Fig. 4), the treatment of male and female human hepatocytes with 5–1000 μM PB did not result in any indication of a concentration-dependent response, with hepatocyte RDS median fold induction values for 5, 10, 50, 100, 500, 750 and 1000 μM PB being 1.2, 0.9, 1.1, 0.9, 1.0, 0.9 and 0.9 fold control, respectively (Fig. 5). Effect of treatment with control medium (containing either 0.1 or 0.5% DMSO) and 5–1000 μM PB on replicative DNA synthesis (RDS) in cultured male and female human hepatocytes from 39 donors. Results are presented as box and whisker plots expressed as median fold change in hepatocyte RDS plotted on a logarithmic scale compared to concurrent control values. Horizontal lines bisecting the rectangles represent median values, with the rectangles representing lower and upper quartiles. The whiskers represent the farthest data point lying within 1.5 × the inter-quartile range (IQR) from the box, the upper whisker being the upper quartile range + 1.5 * the inter quartile range, and the lower whisker being the lower quartile—1.5 * the inter quartile range. Individual human hepatocyte experiments classified as outliers are indicated by dots Reproducibility studies with a single human hepatocyte preparation Examination of data from all of the CLE human hepatocytes studies performed revealed that hepatocytes from one donor, namely a male Caucasian subject aged 51 years (donor identification HU8210 in Table 1), had been examined in a total of fourteen studies (Fig. 6). A total of ten of these studies were performed by Laboratory 1 and four studies by Laboratory 2, employing different culture conditions and different procedures to evaluate the effect of treatment on human hepatocyte RDS levels. In these studies, EGF was used as a positive control for hepatocyte RDS at a concentration of 25 ng/ml by Laboratory 1 and at concentrations of 1 and 10 ng/ml by Laboratory 2. The effect of the CAR activator PB was studied at various concentrations comprising 10, 100, 500 and 1000 μM. While different experimental procedures and EGF concentrations were used, robustly reproducible statistically significant increases in human hepatocyte RDS following treatment with EGF were observed in all fourteen studies (Fig. 6). In contrast, treatment with PB at concentrations up to 1000 μM had no effect on hepatocyte RDS in all fourteen studies performed. This indicates that the methodology is robust and results can be successfully reproduced in different laboratories. Effect of treatment of human hepatocytes from a single male Caucasian donor aged 51 years (donor HU8210 in Table 1) with control medium (containing either 0.1 or 0.5% DMSO), 10, 100, 500 or 1000 μM PB and 1, 10 or 25 ng/ml EGF on replicative DNA synthesis (RDS). A total of 10 studies were performed by Laboratory 1 (employing 25 ng/ml EGF as a positive control), with 4 studies being performed by Laboratory 2 (employing 1 and 10 ng/ml EGF as a positive control). Results for each study are presented as mean values for each treatment expressed as median fold change in RDS plotted on a logarithmic scale compared to the concurrent control Correlation of effects of PB on hepatocyte RDS with effects on CYP enzyme activities In some Laboratory 1 studies, the effect of treatment of male and female human hepatocytes with PB on both hepatocyte RDS and some CYP enzyme activities was investigated. While treatment with 5–1000 μM PB produced some increases in PROD (Fig. 7A), BROD (Fig. 7B) and BQ (Fig. 7C) enzyme activities in cultured human hepatocytes, there was no correlation between the effects of 5–1000 μM PB on hepatocyte RDS and PROD (r2 0.01), BROD (r2 0.00) and BQ (r2 0.01) enzyme activities. However, unlike hepatocyte RDS, some positive correlations were observed between the medium concentrations of PB and increases in the PROD (r2 0.16), BROD (r2 0.33) and BQ (r2 0.55) enzyme activities measured (Supplementary Fig. 1). Discussion The results reported in this paper comprise a substantial database of studies with 39 individual cryopreserved human donor hepatocyte preparations conducted by CLE companies using two experimental laboratories, which employed both different culture conditions and methodology to assess the effects of treatment with test chemicals on hepatocyte RDS. These studies were performed for CLE companies to derive data on the effects of active ingredients on RDS in human hepatocytes and were included in regulatory submissions. Investigations were performed with EGF and HGF as positive control agents, together with studies with the CAR activators PB and CITCO and the PPARα activator WY-14,643. The data obtained for the individual human hepatocyte preparations clearly demonstrates a robust response to induction of RDS in human hepatocytes treated with either EGF or HGF, whereas human hepatocytes are refractory to the mitogenic effects of the CAR activators PB and CITCO and the PPARα activator WY-14,643. Human hepatocytes are an established in vitro test system for a range of applications including xenobiotic metabolism and kinetic studies, xenobiotic-induced toxicity studies and induction of CYP and other xenobiotic metabolising enzymes (Godoy et al. 2013; Hewitt et al. 2007a,b; Parkinson et al. 2004; Soldatow et al. 2013). While xenobiotic metabolism studies are normally performed using hepatocyte suspensions over short time periods, studies of induction of xenobiotic metabolising enzymes and effects on hepatocyte RDS require longer treatment times involving the use of plated hepatocytes. Cultured human hepatocytes have been employed to evaluate species differences in the effects of chemicals on hepatocyte RDS for over 30 years. In one of the first published studies, while PB induced RDS in cultured rat hepatocytes, neither PB nor the PPARα activator nafenopin stimulated RDS in cultured human hepatocytes, whereas human hepatocyte RDS was increased following treatment with EGF (Parzefall et al. 1991). These studies were conducted with human hepatocytes isolated by perfusion of liver samples freshly obtained from human donors undergoing liver surgery. Several other studies using cultured human hepatocytes isolated by perfusion of liver samples have demonstrated that CAR activators (e.g. PB) and PPARα activators (e.g. bezafibrate, ciprofibrate, clofibric acid, di-(2-ethylhexyl)phthalate, di-(isononyl)phthalate; 2-ethylhexanoic acid, fomesafen, methylclofenapate, mono-(2-ethylhexyl)phthalate and nafenopin), in contrast to effects in rodent hepatocytes, do not stimulate RDS in cultured human hepatocytes (Elcombe et al. 1996; Goll et al. 1999; Hasmall et al. 1999, 2000; Hirose et al. 2009; Perrone et al. 1998). However, as employed in the present studies, advances in liver perfusion techniques and the development of improved cryopreservation methods have resulted in the commercial availability of plateable human hepatocytes with a high viability from many individual donors (Hewitt and Li 2015). Using such preparations, MOA/human relevance studies have been performed for a number of nongenotoxic chemicals which exhibit either a CAR or a PPARα MOA for rodent liver tumour formation. Literature examples include benfluralin (Strupp et al. 2020), fluxapyroxad (Goettel et al. 2024), metazachlor (Wiemann et al. 2019), metofluthrin (Hirose et al. 2009; Yamada et al. 2015, 2019), momfluorothrin (Okuda et al. 2017), natural pyrethrins (Osimitz and Lake 2009), nitrapyrin (LaRocca et al. 2017), permethrin (Kondo et al. 2020), piperonyl butoxide (Lake et al. 2020) and sedaxane (Peffer et al. 2018a). The present studies were performed with cryopreserved male and female human hepatocyte preparations from 39 individual donors, comprising thirty-six Caucasian, two Hispanic and one African-American subjects. As clearly demonstrated in this study, with hepatocyte preparations from both male and female donors of a wide range of ages, treatment with EGF and/or HGF resulted in a consistent stimulation of RDS, whereas no increases in RDS were observed after treatment with either PB, CITCO or WY-14,643 (Figs. 2–6). The effects of PB and EGF were also very consistent when one human hepatocyte preparation was examined ten times by Laboratory 1 and four times by Laboratory 2 (Fig. 6). Despite interindividual differences in the magnitude of response, these studies clearly demonstrate a consistent species difference in that, unlike rodent hepatocytes, while EGF and HGF can stimulate RDS in cultured human hepatocytes, such effects are not observed with the CAR and PPARα activators examined. Based on this data, it can be recommended for laboratories conducting RDS studies in cultured human hepatocytes, that different lots of cryopreserved hepatocytes are screened to define minimum/maximum response data to set concentrations of either EGF or HGF and then to conduct studies only with preparations that fall within the expected normal range. Despite the large number of literature studies performed, the applicability of investigations employing cultured human hepatocytes to assess the human relevance of nongenotoxic rodent liver CAR activators has been questioned (van Kesteren et al. 2022). The issues raised included that the use of cultured hepatocytes to evaluate species differences in RDS is not sufficiently substantiated and while much experimental and epidemiology data are available for the “archetypical CAR activator” PB, few data are available for other CAR activators on effects on RDS in human hepatocytes and for human epidemiology studies. These authors also raised the issue of the number of human donors required, the age range of the donors and whether hepatocytes from both male and female subjects need to be evaluated. Another issue raised was the applicability of EGF and/or HGF as suitable positive controls for human hepatocyte RDS studies in the absence of suitable chemicals (van Kesteren et al. 2022; Shizu and Yoshinari 2020). All the issues raised for rodent liver CAR activators by van Kesteren et al. (2022) also apply to the use of cultured human hepatocytes for studies with rodent liver PPARα activators (Yamada et al. 2022, 2025). With respect to the point raised by van Kesteren et al. (2022) that the use of cultured hepatocytes to study species differences in RDS has not been sufficiently substantiated, there is now an extensive database based both on the present data and many literature studies (Corton et al. 2018; Yamada et al. 2021, 2025). The large number of individual human hepatocyte preparations studied clearly demonstrate a marked species difference in the effects of nongenotoxic CAR and PPARα activators on RDS in cultured human as compared to rodent hepatocytes. Moreover, such studies have been performed in a number of different laboratories utilising human hepatocyte preparations from both male and female donors of a wide range of ages. As noted by van Kesteren et al. (2022), PB is an “archetypical CAR activator” having been extensively studied in both experimental animals and in humans (Elcombe et al. 2014; IARC 2001; Lake 2018; Monro 1993; Whysner et al. 1996; Yamada et al. 2021). In deference to the comments of van Kesteren et al. (2022), much data are now available for the effects of a large number of other established rodent liver CAR activators and also PPARα activators in cultured human hepatocytes (Corton et al. 2018; Yamada et al. 2021, 2025) and as described below additional human epidemiology data are also available. Another issue raised was the number, age and sex of human donors required to provide evidence to establish the lack of human relevance of a MOA for a rodent liver CAR activator (van Kesteren et al. 2022). While the majority of the studies reported in this paper were conducted with hepatocyte preparations from Caucasian donors, it is clear that the responses obtained, comprising increases in human hepatocyte RDS after treatment with EGF and/or HGF, as against no increases after treatment with the CAR and PPARα activators studied, appear to be independent of donor age, sex and (based on the limited number of non-Caucasian samples examined) also of ethnicity. Although it has been suggested that only a single human hepatocyte preparation is required to exclude the human relevance of a CAR activator (Peffer et al. 2018b), several previous studies cited in this paper have employed more than just one human hepatocyte preparation in order to enhance the reliability of the experimental findings at that time. The data described in this paper demonstrate that it is not necessary to examine the effects of nongenotoxic rodent CAR and PPARα activators in a large number of human hepatocyte preparations. Clearly, as shown in Fig. 2, there will be interindividual differences between subjects in response to a chemical stimulus. Further work may be required to ascertain the number of donors required to allow for individual variability in human hepatocyte RDS studies. Studies have investigated how metabolic polymorphisms can affect hazard and risk assessment (Carnesecchi et al. 2019; Kasteel et al. 2020; Darney et al. 2021). Pharmacogenomic data on nuclear receptors relevant to non-genotoxic liver carcinogenesis (Neary and Owen 2017) could further inform donor selection for in vitro primary hepatocyte proliferation assays. Both EGF and HGF are known to play a significant role in liver biology and regeneration, contributing to the restoration of liver tissue following injury or partial hepatectomy (Kang et al. 2012; Kimura et al. 2023). With respect to the use of EGF and/or HGF as positive controls for the stimulation of RDS in cultured human hepatocytes in the absence of suitable chemicals, the investigations described in this paper, together with many other studies cited in this paper and elsewhere (Corton et al. 2018; Yamada et al. 2021, 2025), have clearly demonstrated that no non-genotoxic rodent liver CAR and PPARα activators induce RDS in cultured human hepatocytes. It should be noted that the present studies and a previous study included the effect of treatment with the potent human CAR activator CITCO (Maglich et al. 2003; Soldatow et al. 2016). Overall, this represents a fundamental qualitative species difference between the effects of chemicals on RDS in cultured rat and mouse hepatocytes, compared to cultured human hepatocytes. As such there are no chemicals which can be utilised as positive controls for in vitro RDS studies with cultured human hepatocytes, with EGF and/or HGF being the only available positive controls for such studies, One other potential issue is that studies with purified preparations of rat hepatocytes have demonstrated the requirement for the presence of Kupffer cells in the stimulation of hepatocyte RDS after treatment with PPARα activators (Parzefall et al. 2001). However, this is not an issue for human hepatocyte preparations, as like rodent hepatocyte preparations prepared by normal cell isolation procedures, the resulting preparations will contain both parenchymal and non-parenchymal cells (Corton et al. 2018). While the current investigations employed monolayer (i.e. two-dimensional) cultures, the results obtained are supported by data from a spheroid (three-dimensional) culture system containing both parenchymal and non-parenchymal cells where PB induced CYP mRNA levels in both rat and human hepatocytes (Plummer et al. 2019). However, while HGF induced RDS in both rat and human hepatocytes, treatment with PB only induced RDS and levels of the cell proliferation gene nucleophosmin 1 in the rat and not in the human spheroid culture system (Plummer et al. 2019; Elcombe et al. 2025). Hence, the two-dimensional culture system employed in these studies fulfils the requirement for an assay system to evaluate species differences in the effects of chemicals on hepatocyte RDS. Adding complexity will not improve the assay performance, which is important for potentially scaling up this study type for screening purposes. Emerging mechanistic data further support these species differences, as recent molecular analyses suggest that interactions of YAP (Yes‑associated protein) with CAR, and possibly PPARα, may contribute to species‑specific differences in hepatocyte proliferation, consistent with YAP’s role in the Hippo pathway (Shizu et al. 2024; Yamada et al. 2025). The results of the studies described in this paper together with the literature studies cited and reviewed elsewhere (Corton et al. 2018; Yamada et al. 2021, 2025) clearly demonstrate that human hepatocytes are refractory to the mitogenic effects of rodent liver CAR and PPARα activators. However, other effects of rodent CAR and PPARα activators can be observed in humans. For example, in both rodents and humans, CAR activators can produce liver hypertrophy and induction of hepatic xenobiotic metabolising enzymes, whereas PPARα activators can produce a hypolipidemic effect (Corton et al. 2018; Yamada et al. 2021, 2025). In the present studies, while PB had no effect on RDS in cultured human hepatocytes, increases in CYP enzyme activities were observed (Fig. 7 and Supplementary Fig. 1). The data for the lack of effect of rodent CAR and PPARα activators from in vitro studies with cultured human hepatocytes are supported by in vivo studies in chimeric mice with humanised livers and by available human epidemiology data. The chimeric mouse model, where transplanted human hepatocytes replace the host hepatocytes, has been used for many applications including studies of xenobiotic metabolism and toxicity (Tateno and Kojima 2020; Yamada 2021). The treatment of chimeric mice with human hepatocytes with the CAR activators PB, metofluthrin and momfluorothrin and with the PPARα activator fenofibrate was shown not to stimulate RDS in human hepatocytes (Okuda et al. 2017; Tateno et al. 2015; Yamada et al. 2014). With respect to available human epidemiology data, the rodent liver CAR activator PB has been employed as a sedative, hypnotic and antiepileptic drug in humans for several decades at therapeutic dose levels which result in similar blood levels to those which produce liver tumours in mice (Monro 1993). As such, an extensive epidemiology data set is available which confirms a lack of tumour formation in humans receiving PB for extended periods (La Vecchia and Negri 2014; Stritzelberger et al. 2021; Yamada et al. 2021, 2025). In addition, epidemiological data showing no increase in liver tumour formation are also available for two other rodent liver CAR activators, namely oxazepam and carbamazepine (Friedman et al. 2009; Stritzelberger et al. 2021; Yamada et al. 2021). Similarly, human epidemiological data are also available for a number of rodent liver PPARα activators, including bezafibrate, clofibrate, gemfibrozil and fenofibrate, where no excess cancer mortality was observed (Bonovas et al. 2012; Corton et al. 2018; Klaunig et al. 2003). Overall, the results described in this paper provide further evidence for the usefulness of cultured human hepatocytes in establishing species differences in response to the hepatic effects of mitogenic nongenotoxic chemicals. It can be concluded that the method can be conducted in different laboratories, employing different experimental techniques, and that the qualitative inducibility of RDS is not sex-, age- or (based on a limited number of samples examined) ethnicity-dependent. Donor-to-donor variability in amplitude of response exists, which may be addressed by screening preparations of cryopreserved hepatocytes for suitability to be employed in such studies and expressing effects relative to control. This review additionally provides detailed aspects important for the characterisation of primary hepatocytes as a test system for regulatory testing, including exposure conditions, dose selection procedures, endpoints assessed, measures of variability and appropriate performance chemicals. These pieces of information are required by OECD Guidance Document 34 (OECD 2005) and relevant to a retrospective assessment of the validation status of an assay. Systematic validation of cell proliferation, including other cell types than hepatocytes, appears highly recommendable. Indeed, the value of studies with human hepatocytes has been highlighted in a recent European Union publication on the usefulness of in vitro mechanistic information in evaluating MOAs for rodent liver CAR activators (Capelôa et al. 2024). The method is considered ready for a thorough validation according to OECD Guidance Document 34 to assess its robustness systematically, introduce more standardisation and increase regulatory uptake of this new approach method in regulatory decision making. Supplementary Information Below is the link to the electronic supplementary material. Supplementary file1 (DOCX 100 kb) Acknowledgements CropLife Europe’s Human Health Expert Group as the industry umbrella organisation of plant protection product companies facilitated the data collection across member companies and provided funding for preparation of the manuscript. ADAMA, BASF, Bayer, Gowan, Janssen PMP, Sumitomo Chemical Co Ltd and Syngenta kindly provided the data meta-analysed in this publication. DEC organised and statistically analysed the substantial amount of data provided by CropLife Europe members. SMC reviewed the developing script and provided guidance and input based on his substantial experience and expertise in this area. Data availability The data that support the findings of this study are available as positive control compounds in regulatory studyreports supporting registration of agrochemicals. They are either publically available in regulatory assessments or, ifnot, may be made available from the data owners via the corresponding author upon reasonable and justifi edrequest. Data would in that case be provided in a sanitized form. Declarations Competing interests CS, DEC, HT, JT, MG, MES and MG are employed by companies developing and producing plant protection products. TY was employed until the end of March 2025 by Sumitomo Chemical Company Ltd. CS is a member of the OECD working group on nongenotoxic carcinogenicity. SMC and BGL consult for the Sumitomo Chemical Co Ltd., provider of the Laboratory 2 data for this publication. The authors declare that they have no conflicts of interest associated with this work. Footnotes Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. References Associated Data Supplementary Materials Supplementary file1 (DOCX 100 kb) Data Availability Statement The data that support the findings of this study are available as positive control compounds in regulatory studyreports supporting registration of agrochemicals. They are either publically available in regulatory assessments or, ifnot, may be made available from the data owners via the corresponding author upon reasonable and justifi edrequest. Data would in that case be provided in a sanitized form.