Comparative DNA damage induced by eight nitrosamines in primary human and macaque hepatocytes

Abstract

N-nitrosamines have been increasingly detected in human drugs, raising serious safety concerns due to their potential mutagenicity and carcinogenicity. In order to expand upon the human data available on these drug impurities, we previously used metabolically competent HepaRG human hepatoma cells to evaluate the genotoxicity of eight small-molecule nitrosamines [N-cyclopentyl-4-nitrosopiperazine (CPNP), N-nitrosodibutylamine (NDBA), N-nitrosodiethylamine (NDEA), N-nitrosodimethylamine (NDMA), N-nitrosodiisopropylamine (NDIPA), N-nitrosoethylisopropylamine (NEIPA), N-nitroso-N-methyl-4-aminobutyric acid (NMBA), and N-nitrosomethylphenylamine (NMPA)]. In this study, we used the comet assay to further investigate the DNA damage induced by the eight nitrosamines in primary human hepatocytes (PHHs) from three individual donors and primary macaque hepatocytes (PMHs) from freshly isolated livers of six rhesus macaques. In addition, expression of genes encoding Phase I and II metabolic enzymes and the activities of the enzymes were compared in PHHs and PMHs, and Western blot was used to analyze protein biomarkers of DNA damage and apoptosis in PMHs. All eight nitrosamines induced significant DNA damage in both PHHs and PMHs; with the exception of NDMA, higher fold increases in % tail DNA were detected in PMHs compared to PHHs. Greater interindividual variability in CYP gene expression, enzyme activities, and DNA damage responses was observed in PHHs compared to PMHs. Benchmark concentration (BMC) modeling analysis showed that PHHs had more conservative BMC₅₀ values than PMHs for most nitrosamines tested. Nonetheless, correlation analysis demonstrated that DNA damage data generated by PMHs and 3D HepaRG spheroids were comparable to those of PHHs. Western blot analysis suggested a potential role for the ethyl group in regulating protein expression in the DNA damage and apoptosis pathways for nitrosamines. Overall, this study provides human-relevant DNA damage responses for the eight nitrosamines and indicates that differences in genotoxic potency between PHHs and PMHs are likely related to CYP enzyme activity.

1. Introduction

N-nitroso compounds are carcinogenic in multiple organs of approximately 40 animal species [1]. N-nitrosamines (or nitrosamines) usually require metabolic activation by cytochromes P450 enzymes (CYPs) to alkylate DNA and produce a carcinogenic effect [2]. Because some nitrosamines are extremely potent carcinogens, ICH M7(R2) places nitrosamine drug impurities in a “cohort of concern”, and unlike many other drug impurities, excludes setting Acceptable Intake (AI) levels using the Threshold of Toxicological of Concern approach [3]. The recent identification of nitrosamines as impurities in several common medications has raised significant safety concerns within the pharmaceutical industry and regulatory authorities worldwide [4].

The genotoxicity and carcinogenicity of some nitrosamines have been well-documented in bacteria, in various in vitro mammalian cell models and in animals in vivo. The recently developed Enhanced Ames Test (EAT), using hamster and rat liver S9 fractions for metabolic activation, has been effective in detecting the mutagenicity of small-molecule nitrosamines and nitrosamine drug substance-related impurities (NDSRIs) [5,6]. The Ames test has demonstrated a high sensitivity for identifying nitrosamines that have positive responses in the rodent bioassay, but with a relatively low specificity [7]. Given the high carcinogenic potential of this class of drug impurities, it is important to develop follow-up mammalian cell assays that can confirm findings in the Ames test, especially negative findings, and that ideally have improved specificity without compromising sensitivity. Another consideration is to better establish the relevance of responses in bacterial and animal tests for defining human risk.

To address these concerns, we previously used HepaRG cells, a human derived, TP53-proficient, and metabolically competent cell model, to evaluate the genotoxicity of eight small-molecule nitrosamines. The results showed that all eight nitrosamines induced DNA damage in three-dimensional (3D) HepaRG spheroids, while only three of them were positive in 2D cells [8]. These eight nitrosamines are N-nitrosodibutylamine (NDBA), N-nitrosodiethylamine (NDEA), N-nitrosodimethylamine (NDMA), N-nitrosodiisopropylamine (NDIPA), N-ethyl-N-nitroso-2-propanamine (NEIPA), N-nitroso-N-methyl-4-aminobutyric acid (NMBA), N-nitrosomethylphenylamine (NMPA), and N-cyclopentyl-4-nitroso piperazine (CPNP). All eight have been found in human drug products as impurities at levels exceeding the recommended AI limit [9–11].

The metabolic capacity and hepatic functionality of the HepaRG cell line have been well-characterized [12–14]. When cultured in a 3D format, 3D HepaRG spheroids express much higher levels of most Phase I enzymes than their 2D counterpart [15]. This observation aligns with the higher sensitivity of 3D spheroids in comparison to 2D cells for detecting the genotoxicity of nitrosamines as well as other known genotoxicants and carcinogens [8,15,16]. We also demonstrated that 3D HepaRG spheroids had the same sensitivity as primary human hepatocytes (PHHs) cultured in 2D format in detecting the DNA damage induced by four indirect-acting (i.e., requiring metabolic activation) and five direct-acting (i.e., requiring no metabolic activation) genotoxicants/carcinogens [16].

However, HepaRG cells are derived from a single donor and have relatively low levels of CYP2D6 activity [12]. Questions remain as to whether HepaRG cells can be used as a reliable in vitro cell model for follow-up mammalian studies to provide human-relevant genotoxicity data on nitrosamines. PHHs have been considered to be the “gold standard” for liver metabolism and in vitro toxicity evaluation [17]. The limited availability of PHHs, however, is a significant hurdle that prevents their routine application in toxicology studies. To overcome this challenge, primary hepatocytes isolated from non-human primates (NHPs) have been evaluated as a surrogate for genotoxicity assessment due to the physiological and genetic similarity of NHPs to humans. Primary hepatocytes isolated from the livers of rhesus macaques (PMHs) showed the same sensitivity and specificity as PHHs in detecting five noncarcinogens, four indirect-acting and seven direct-acting genotoxicants/carcinogens [18]. Whether PMHs are useful for detecting the genotoxicity of nitrosamines remains unknown, as nitrosamines are considered difficult to evaluate for mutagenicity using standard Ames test protocols [6].

In the present study, we used the CometChip assay to evaluate the DNA damage produced by eight nitrosamine drug impurities in PHHs derived from three individual donors. The nitrosamines were the same eight nitrosamines tested previously in HepaRG cells [8]. PMHs freshly isolated from the livers of six rhesus macaques were also evaluated for their suitability for assessing the cytotoxicity and genotoxicity of nitrosamines. A quantitative comparison was conducted on the comet concentration-response data generated from PHHs and PMHs using benchmark concentration (BMC) modeling. We further used Western blot analysis to explore potential mechanisms responsible for the cytotoxicity and genotoxicity of these nitrosamines in PMHs.

2. Materials and methods

2.1. Chemicals and materials

NEIPA (CAS# 16339-04-1) and NMPA (CAS# 614-00-6) were obtained from Enamine (Monmouth Jct., NJ). CPNP (CAS# 61379-66-6), NDBA (CAS# 924-16-3), NDIPA (CAS# 601-77-4), and NMBA (CAS# 61445-55-4) were purchased from Toronto Chemical Research (Toronto, ON, Canada), TCI America (Portland, OR), ChemService (West Chester, PA), and Chemspace (Monmouth Junction, NJ), respectively. All other chemicals were purchased from Sigma-Aldrich (St. Louis, MO). This includes NDEA (CAS# 55-18-5), NDMA (CAS# 62-75-9), 0.4 % trypan blue solution, bovine serum albumin (BSA), calcium chloride (CaCl₂), collagenase type IV, ethylene glycol-bis(2-aminoethylether)-N,N,N′,N′-tetraacetic acid (EGTA), D-(–)-fructose, heparin, 4-(2-hydroxyethyl)-1-piperazineethanesulfonic acid (HEPES), hyaluronidase type I–S from bovine, hydrocortisone 21-hemisuccinate, insulin, L-15 medium (Leibovitz), potassium chloride (KCl), sodium bicarbonate, sodium chloride (NaCl), and trypsin inhibitor, seven CYP450 enzyme substrates (phenacetin, bupropion, chlorzoxazone, diclofenac, dextromethorphan, midazolam, and omeprazole), their metabolites (4-acetamidophenol, hydroxybupropion, 6-hydroxychlorzoxazone, 4-hydroxydiclofenac, dextrorphan, 1-hydroxymidazolam, and 5-hydroxyomeprazole, respectively), and the internal standard hydroxybupropion-d6 (Bup–OH–D6). Additional components were sourced from Thermo Fisher Scientific (Waltham, MA), including GlutaMax^™, nonessential amino acids, and penicillin-streptomycin. Fetal bovine serum (FBS) and comet supplies were obtained from R&D systems (Minneapolis, MN). All chemicals were stored as the manufacturer’s recommendations.

2.2. Primary human hepatocytes (PHHs)

Cryopreserved PHHs from three individual donors (Supplementary Table 1), along with Universal Cryopreservation Recovery Medium (UCRM^™) and Universal Primary Cell Plating Medium (UPCM^™), were obtained from Discovery Life Sciences (Huntsville, AL). PHHs were cultured following the supplier’s instructions. Briefly, the vials of PHHs were quickly thawed in a 37 °C water bath, transferred into pre-warmed UCRM^™, and centrifuged at 100×g for 10 min at room temperature. The supernatant was discarded, and the cell pellet was gently resuspended in pre-cooled UPCM^™. Viable cells were stained with 0.4 % trypan blue solution and counted using a hemocytometer under a phase contrast light microscope (Leica Microsystems Inc.; Deerfield, IL). Thereafter, cells were plated in 96-well plates coated with PureCol^® collagen (Advanced BioMatrix, Inc.; Carlsbad, CA) at a density of 5 × 10⁴ cells in 100 μL of UPCM^™ per well and incubated overnight at 37 °C in a humidified atmosphere with 5 % CO₂.

2.3. Primary macaque hepatocytes (PMHs)

Fresh PMHs were isolated from the livers of six rhesus macaques (15-year-old males and females with body weights ranging from 7.2 to 11 kg), as described previously [18,19]. Briefly, the liver was perfused consecutively with warm washout solution (9.2 mM HEPES-NaCl buffer with 0.5 mM EGTA, 16.65 mM fructose, and 10 U/mL heparin) and digestion solution (9.2 mM HEPES-NaCl buffer containing 6.7 mM KCl, 5 mM CaCl₂, 16.65 mM fructose, 0.3 % BSA, 0.06 % collagenase type IV, 100 mg/L hyaluronidase, and 100 mg/L trypsin inhibitor). The digested liver tissue was dissociated on ice in suspension medium (L-15 medium supplemented with 0.2 % BSA, 1 M HEPES, 5 μg/mL insulin, 5 % FBS, and 0.22 % sodium bicarbonate). The released cells were filtered through two sieves with decreasing mesh sizes (150 and 80 μm nylon mesh). Filtered cells were pelleted by centrifugation, washed gently using suspension medium, and purified using 90 % isotonic Percoll to remove dead cells. Purified hepatocytes were resuspended in plating medium (L-15 medium supplemented with 1 × nonessential amino acids, 100 U/mL penicillin, 100 μg/mL streptomycin, 1 × GlutaMax^™, 1.4 μM hydrocortisone, 100 nM insulin, 17 % FBS, and 0.23 % sodium bicarbonate). Viable PMHs were counted and plated in 96-well plates at a density of 5 × 10⁴ cells per well in 100 μL of the plating medium. The plates were incubated overnight at 37 °C in a humidified atmosphere with 5 % CO₂.

2.4. Phase I and Phase II gene expression of PHHs and PMHs

The basal gene expression levels of 13 Phase I and 5 Phase II enzymes were measured at the mRNA level using quantitative real-time PCR (qPCR). Total RNA from PHHs and PMHs was extracted using the RNeasy Mini kit (Qiagen; Valencia, CA) following the manufacturer’s instructions. The quality and concentration of RNA were determined by a NanoDrop 8000 spectrophotometer (Thermo Fisher Scientific). cDNA was synthesized from 1 μg of isolated RNA using the High-Capacity cDNA Reverse Transcription kit (Applied Biosystems; Foster City, CA). qPCR was performed using the FastStart Universal Probe Master Mix (Rox) (Roche Applied Science; Indianapolis, IN) and TaqMan^™ probes (Applied Biosystems). TaqMan^™ probes of Phase I and Phase II enzymes used for PHHs and PMHs are listed in Supplementary Tables 2 and 3, respectively. The amplification was conducted using a ViiA 7 Real-Time PCR system (Applied Biosystems) under the same conditions described in our previous studies [8,15]. Glyceraldehyde 3-phosphate dehydrogenase (GAPDH) was used as a reference gene. The expression value for each gene was calculated using cycle threshold (Ct) values by equation E = 2^{−(Ct of test gene–Ct of reference gene)} × 10,000. This represents the relative mRNA expression level of a gene, with the expression of GAPDH arbitrarily set at 10,000 copies.

2.5. CYP450 activities of PHHs and PMHs

The basal activities of seven major CYP enzymes (CYP1A2, 2B6, 2C9, 2C19, 2D6, 2E1, and 3A4) were measured using our previously established methods [16]. Briefly, PHHs or PMHs were incubated with 100 μL of the treatment medium containing 40 μM chlorzoxazone (CYP2E1) or a cocktail of six CYP substrates, i.e., 100 μM phenacetin (CYP1A2), 100 μM bupropion (CYP2B6), 20 μM diclofenac (CYP2C9), 20 μM omeprazole (CYP2C19), 20 μM dextromethorphan (CYP2D6), and 50 μM midazolam (CYP3A4), at 37 °C in a humidified atmosphere with 5 % CO₂. Following a 2-h exposure to the CYP substrates, the supernatants were collected, and the individual metabolites released into the medium were quantified using high-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) on a Shimadzu 20A ultra-fast liquid chromatography (UFLC) system coupled with an AB SCIEX 3200 QTRAP mass spectrometer (SCIEX LLC; Framingham, MA) as previously described [16]. In addition, PHHs and PMHs were immediately washed with cold Dulbecco’s Phosphate-Buffered Saline (DPBS) and lysed in 30 μL radioimmunoprecipitation assay (RIPA) lysis buffer (Thermo Fisher Scientific). The protein concentrations were determined using the Pierce BCA protein assay kit (Thermo Fisher Scientific). CYP450 activities were expressed as pmol metabolite/min/mg protein.

2.6. Cell treatments

Twenty-four hours after plating, PHHs and PMHs were exposed to various concentrations of the eight nitrosamines in total volume of 100 μL UPCM^™ and PMH treatment medium (L-15 medium supplemented with 1 × nonessential amino acids, 100 U/mL penicillin, 100 μg/mL streptomycin, 1 × GlutaMax, 17 μM hydrocortisone, 100 nM insulin, 10 % FBS and 0.23 % sodium bicarbonate), respectively, for 24 h at 37 °C in a humidified atmosphere with 5 % CO₂. Six nitrosamines, CPNP, NDBA, NDIPA, NEIPA, NMBA, and NMPA, were dissolved in DMSO, while two nitrosamines, NDEA and NDMA, were dissolved in deionized water. All nitrosamine stock solutions were freshly prepared prior to each treatment. Working solutions were prepared from 100 × of the highest treatment concentration of each nitrosamine, and serial dilutions were performed to achieve the final concentrations. All experiments for each chemical treatment were repeated three times as PHH_A, PHH_B, and PHH_C, while experiments with PMH1 through PMH6 were conducted independently in duplicate.

2.7. Cytotoxicity assay

Following the 24-h treatment, nitrosamine-induced cytotoxicity was evaluated using the CellTiter-Glo^® Luminescent Cell Viability Assay kit (Promega; Madison, WI). Briefly, 100 μL of ATP assay reagent was added to each well of a 96-well plate at a ratio of 1:10, and the plate was incubated at room temperature for 10 min. Luminescence was then measured using a Cytation 5 Cell Imaging Multi-Mode Reader (BioTek; Winooski, VT). Relative viability (%) was expressed as a percent intensity of treated cells relative to the vehicle controls.

2.8. CometChip assay

Following the 24-h treatment, DNA damage in PHHs and PMHs was evaluated using the CometChip assay as described previously with minor modifications [18,20,21]. The CometChips were either made in-house using a mold or purchased commercially from R&D systems. To fabricate the CometChip, molten 1 % normal melting point agarose (Thermo Fisher Scientific) in DPBS was spread on the hydrophilic side of GelBond^® film (Lonza; Walkersville, MD) and a polydimethylsiloxane (PDMS) stamp with micropillars was placed on top of the lid. The stamp was removed after the agarose gel solidified, leaving an array of 35-μm-sized microwells. The gel on the GelBond^® film was placed onto a 96-well glass plate and inserted into the 96-well CometChip^® System to form the macrowells. Treated cells were detached by incubating the cells in 30 μL of TrypLE^™ Express (Thermo Fisher Scientific) at 37 °C for 5 min, and the reaction was stopped with 120 μL of the treatment medium. Cell suspensions were transferred into each well of a 96-well CometChip^® System and gravity-loaded into the micropores for 50 min at 37 °C in a humidified atmosphere with 5 % CO₂. The CometChip was then incubated in CometAssay Lysis Solution at 4 °C for 1 h, followed by unwinding DNA in alkaline buffer (0.2 M NaOH, 1 mM EDTA, 0.1 % TritonX-100, pH > 13) for 40 min and electrophoresis in the same buffer at 22 V and 4 °C for 50 min. After neutralization and equilibration with Tris-HCl buffer (Sigma-Aldrich), the CometChip was stained with 0.2 × SYBR Gold (Invitrogen; Carlsbad, CA) at 4 °C. Images of the comets were captured using the Cytation 5 Cell Imaging Multi-Mode Reader and analyzed using CometAssay Analysis Software (R&D Systems) to determine the percentage of DNA fluorescence in the tail (% Tail DNA).

2.9. Quantification of DNA damage concentration-responses

Benchmark concentration (BMC) modeling was conducted to compare the DNA damage responses for the eight nitrosamines in PHHs and PMHs using PROAST web-based software (version 70.1), with ‘compound’ used as a covariate. The BMC₅₀ values, corresponding to 50 % increases in DNA damage relative to the vehicle control responses, and the upper and lower bounds (BMCU and BMCL, respectively) of their 90 % confidence intervals (CIs) were calculated using the exponential and Hill models. The values from the exponential model were used for quantitative comparisons of relative DNA damage responses of eight nitrosamines and between PHHs and PMHs. The BMC₅₀ values generated from the PHH and PMH comet data also were compared with those calculated using the 3D HepaRG comet data from our previous study [8]. Responses with non-overlapping BMCU and BMCL intervals were considered significantly different from one another.

2.10. Western blot

PMHs from three rhesus macaques were seeded at a density of 1.5 × 10⁶ cells/well in 6-well plates coated with PureCol^® collagen and incubated overnight at 37 °C in a humidified atmosphere with 5 % CO₂. The cells were treated for 24 h with two concentrations of the eight nitrosamines: the highest concentration tested in the CometChip assay and 1/10 of the highest concentration. After treatment, PMHs were harvested and lysed in 150 μL RIPA lysis buffer containing 1 % Halt^™ Protease Inhibitor Cocktail (Thermo Fisher Scientific). Supernatants were collected by centrifugation and the protein concentrations were determined using the Pierce BCA protein assay kit. Protein samples were dissolved in Laemmli sample buffer (Bio-Rad; Hercules, CA) with 2.8 % β-mercaptoethanol (Sigma-Aldrich), boiled at 95 °C for 5 min, and 12 μg of protein lysates were loaded onto NuPAGE^™ 4–12 % Bis-Tris Midi Protein Gels (26-well, Invitrogen). Electrophoresis was performed in a SureLock Tandem Midi-Gel-Tank (Thermo Fisher Scientific) at 80 V for 5 min followed by 200 V for 40 min, and proteins were transferred onto PVDF membranes (Millipore Corporation; Billerica, MA) at 30 V for 30 min. After blocking with 5 % BSA in Tris-buffered saline (Bio-Rad) containing 0.1 % Tween 20 (Sigma-Aldrich) (TBST) for 1 h, the membranes were incubated overnight at 4 °C with the following primary antibodies: phosphorylated-histone H2A.X (γ-H2A.X; #2577), growth arrest and DNA damage 45-α (GADD45-α; #4632), phosphorylated-p53 (p-p53; #9284), p21Waf1/Cip1 (p21; #2947), cleaved-poly (ADP-ribose) polymerase (c-PARP; #5625), cleaved caspase-3 (c-Cas3; #9664), Bcl-xL (#2764), BcL2L10 (#3869), Bad (#9239), Bax (#2772), and GAPDH (#5174) (Cell Signaling Technology; Danvers, MA) at 1:1000 dilutions. The membranes were then incubated with horseradish peroxidase (HRP)-conjugated secondary antibody (Cell Signaling Technology) at a dilution of 1:10,000, and visualized using the Immobilon Western Chemiluminescent HRP Substrate (EMD Millipore; Billerica, MA). Protein bands were imaged using the FluorChem E System (ProteinSimple; San Jose, CA) and quantified by AlphaView SA (ProteinSimple).

2.11. Statistical analysis

Data are expressed as the mean ± standard deviation (SD) from at least three independent experiments. Statistical analyses were performed using SigmaPlot 13.0 (Systat Software; San Jose, CA). The statistical significance of mRNA gene expression data was evaluated by the two-tailed Student’s t-test. One-way analysis of variance (ANOVA) followed by Dunnett’s post hoc test was employed to determine the lowest effective concentration (LEC) between treatment groups and the vehicle control. Pearson’s correlation coefficient was used to analyze the correlation of BMC₅₀ values between PHHs and PMHs as well as between PHHs and 3D HepaRG spheroids using GraphPad Prism version 8 (GraphPad Software; La Jolla, CA). A p-value of < 0.05 was considered statistically significant.

3. Results

3.1. Gene expression of Phase I and Phase II enzymes in PHHs and PMHs

Relative average mRNA expression levels for PHHs and PMHs are compared in Fig. 1. Although rhesus macaques and humans are 93.5 % identical in terms of nucleotide sequence [22], PHHs and PMHs had significantly different expression levels for most genes tested. Compared to PHHs, PMHs had significantly higher expression for seven out of 13 genes encoding Phase I enzymes, including CYP1A1 (24.3-fold), CYP2A6/2A13/2A7 (11.1-fold), CYP2B6 (6.5-fold), CYP2C19 (10.5-fold), CYP2D6 (13.8-fold), CYP3A4 (8.4-fold), and CYP3A7 (23.0-fold), as well as one out of five genes encoding Phase II enzymes, UGT1A1 (6.6-fold). In contrast, PHHs displayed significantly higher expression for CYP1B1 (159.3-fold) and CYP2E1 (2.1-fold) than PMHs. No significant differences were observed for the expression of three genes encoding Phase I enzymes (CYP2C8, CYP2C9, and CYP3A5) and two genes coding for Phase II enzymes (NAT1 and SULT2A1). Gene expression for CYP1A2, SULT1A1, and UGT1A6 was not analyzed in PMHs due to the unavailability of TaqMan^™ probes for rhesus monkeys. The relative mRNA expression levels for individual PHH and PMH cultures from the different donors are shown in Supplementary Fig. 1.

Fig. 1. Gene expression profiles of 13 Phase I and 5 Phase II enzymes in primary human hepatocytes (PHHs) and primary macaques hepatocytes (PMHs).

Relative mRNA levels were quantified using real-time quantitative PCR (qPCR) and normalized to internal control GAPDH. Data are averaged from three individual donors for PHHs and six individual rhesus macaques for PMHs and presented as mean ± SD (n ≥ 3). Statistical significance was determined by Student’s t-test between PHHs and PMHs (*p < 0.05, **p < 0.01, ***p < 0.001). CYP, cytochrome P450; NAT, N-acetyltransferase; SULT, sulfotransferase; UGT, UDP-glucuronosyltransferase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; NT, not tested.

3.2. CYP450 enzyme activities in PHHs and PMHs

The metabolic capacity of PHHs and PMHs was quantified by measuring metabolite formation using HPLC-MS/MS (Fig. 2). Similar to the gene expression levels, PHHs and PMHs had significantly different activities for the seven CYPs tested. On average, PMHs consistently showed significantly higher activities for five CYPs (CYP1A2, 2B6, 2C19, 2D6, and 3A4), while PHHs exhibited higher CYP2C9 and CYP2E1 activities than PMHs. Specifically, CYP1A2 and CYP2B6 activities were 13.3–34.7-fold higher (75.3–85.1 vs. 2.2–6.4 pmol/min/mg protein) in PMHs than in PHHs; CYP2D6 and CYP3A4 activities were 2.6–4.8-fold higher (55.3–69.6 vs. 11.2–27.2 pmol/min/mg protein) and CYP2C19 activity was 2.6-fold higher (5.3 vs. 2.1 pmol/min/mg protein) in PMHs. In contrast, PHHs had 16.7-fold and 5.9-fold higher CYP2C9 and CYP2E1 activities (6.4 vs. 0.4 and 2.3 vs. 0.4 pmol/min/mg protein, respectively) than PMHs.

Fig. 2. CYP450 enzyme activities in PHHs and PMHs.

The activities of seven major CYP450 (CYP1A2, 2B6, 2C9, 2C19, 2D6, 2E1, and 3A4) were determined by measuring metabolites of each enzyme substrate using HPLC-MS/MS, after PHHs and PMHs were incubated with substrate cocktails in the medium for 2 h. CYP450 activities for each of the three PHH donors and six rhesus macaques are expressed as pmol/min/mg protein. The data are presented as mean ± SD (n ≤ 3). Statistical significance was determined by Student’s t-test between PHHs and PMHs (*p < 0.05, **p < 0.01, ***p < 0.001). CYP, cytochrome P450; < LOD, below the limit of detection.

Interindividual variability in CYP activities also was observed. The three PHH donors (A–C) showed greater variation compared to six PMH samples (1–6) for most of CYPs tested. CYP2C19 and CYP2E1 activities in PHH_C were below the detection limit (Fig. 2).

3.3. Nitrosamine-induced cytotoxicity in PHHs and PMHs

The ATP assay was used to evaluate the cytotoxicity induced by nitrosamines. The highest concentration of each nitrosamine tested in comet assays was the concentration resulting in ≥70 % relative viability compared to the vehicle control, or 10 mM when no cytotoxicity was observed [23]. Nitrosamine-induced cytotoxicity in PHHs is presented individually by donor in Fig. 3A, while the results for PMHs are shown in Fig. 4 as the average across six PMH samples. Following 24-h of exposure, four nitrosamines, NDEA, NDIPA, NEIPA, and NMBA, induced less than 30 % cytotoxicity at concentrations up to 10 mM in both PHHs and PMHs (Table 1). Three other nitrosamines, CPNP, NDBA, and NMPA, produced comparable cytotoxic effects in PHHs and PMHs, with CPNP and NMPA having slightly lower maximum tested concentrations and NDBA exhibiting slightly higher maximum tested concentrations in PHHs than in PMHs. NDMA was significantly more cytotoxic in PHHs than in PMHs, with the maximum tested concentrations for comet testing being 1250 μM and 10 mM in PHHs and PMHs, respectively.

Fig. 3. Nitrosamine-induced cytotoxicity and DNA damage in PHHs.

PHHs were exposed to eight nitrosamines for 24 h. (A) Relative viability (% of control) was measured using the ATP assay. The concentration that resulted in ≥70 % cell viability was selected to evaluate DNA damage responses. (B) DNA damage (% Tail DNA) was assessed using the CometChip assay. Blue, red, and green lines represent PHH_A, PHH_B, and PHH_C, respectively. Data are presented as the mean ± SD for each donor (n = 3). Statistical significance was determined by one-way ANOVA followed by Dunnett’s test (*p < 0.05, **p < 0.01, ***p < 0.001 vs. vehicle control). NS, not significant (p ≥ 0.05). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Fig. 4. Nitrosamine-induced cytotoxicity and DNA damage in PMHs.

PMHs were exposed to eight nitrosamines for 24 h. Relative viability (% of control; right y-axis and red line) was measured using the ATP assay and the concentration that resulted in ≥70 % cell viability was used to evaluate DNA damage responses. DNA damage (% Tail DNA; left y-axis and black bar) was assessed using the CometChip assay. Data are presented as the mean ± SD (n ≥ 3). Statistical significance was determined by one-way ANOVA followed by Dunnett’s test (*p < 0.05, **p < 0.01, ***p < 0.001 vs. vehicle control). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Table 1.

Comparison of cytotoxicity and DNA damage responses between primary human hepatocytes (PHHs) and primary macaque hepatocytes (PMHs).

Nitrosamine	Max Conc. (μM)a		ATP (%)b		LECc		Fold increased		Outcomee
	PHHs	PMHs	PHHs	PMHs	PHHs	PMHs	PHHs	PMHs	PHHs	PMHs
CPNP	2,500	5,000	78.9	76.7	100	312.5	2.6	8.0	++	+++
NDBA	1,250	1,000	68.8	77.1	100	11.7	2.7	10.6	++	+++
NDEA	10,000	10,000	74.0	77.8	200	39.1	7.6	11.8	+++	+++
NDIPA	10,000	10,000	77.6	80.7	1,250	5,000	2.2	3.9	++	++
NDMA	1,250	10,000	80.3	89.5	20	19.5	15.4	11.0	+++	+++
NEIPA	10,000	10,000	78.6	75.5	200	156.3	4.6	13.2	++	+++
NMBA	10,000	10,000	73.1	76.3	2,500	1,250	2.9	6.9	++	+++
NMPA	2,500	3,000	78.6	73.4	200	250	2.0	3.4	+	++

CPNP, N-cyclopentyl-4-nitrosopiperazine; NDBA, N-nitrosodibutylamine; NDEA, N-nitrosodiethylamine; NDIPA, N-nitrosodiisopropylamine; NDMA, N-nitrosodimethylamine; NEIPA, N-nitrosoethylisopropylamine; NMBA, N-nitroso-N-methyl-4-aminobutyric acid; and NMPA, N-nitrosomethylphenylamine.

^aThe highest concentration tested in the CometChip assay.

^bATP (%) was used as an indicator for cytotoxicity induced by nitrosamines, relative to the vehicle control.

^cLEC, the lowest effective concentration, determined by one-way ANOVA followed by Dunnett’s test, is the lowest concentration that induces a significant response in the assay.

^dThe fold increase of nitrosamine-induced DNA damage over the vehicle control at the maximum concentration shown in the table.

^e+, 1.5 ≤ ratio <2; ++, 2 ≤ ratio <5; and +++, ratio ≥5 (p < 0.05).

3.4. Nitrosamine-induced DNA damage in PHHs and PMHs

When the comet data were averaged from the three human donors and the six PMH samples, all eight nitrosamines significantly increased % tail DNA (DNA damage) in both PHHs and PMHs (Table 1, Fig. 3B and 4). However, greater interindividual variability was seen in PHHs than in PMHs. Specifically, five nitrosamines, CPNP, NDBA, NDEA, NDMA, and NEIPA significantly increased % tail DNA in PHHs from all three individual donors. While three nitrosamines, NDIPA, NMBA, and NMPA, induced statistically significant increases in DNA damage only in PHH_A and PHH_B, but not in PHH_C. In contrast, all eight nitrosamines gave the same positive calls for DNA strand breaks in all six PMH samples (Fig. 4).

When the Lowest Effective Concentration (LEC) values were compared, CPNP, NDEA, and NMPA had lower LECs in PHHs than in PMHs, while PMHs had lower LECs for NDBA, NDEA, NEIPA, and NMBA than PHHs. NDMA exhibited similar LECs with both PHHs and PMHs (Table 1). At the maximum tested concentrations, NDMA produced a higher fold-increase for DNA damage in PHHs compared to PMHs (15.4- vs. 11.0-fold over the control) at much lower concentrations (1250 μM vs. 10 mM), while the seven other nitrosamines caused greater levels of DNA damage in PMHs than in PHHs. Specifically, CPNP, NDBA, NDIPA, NMBA, and NMPA induced 2.0–2.9-fold and 3.4–10.6-fold increases in % tail DNA over the control in PHHs and PMHs, respectively. NDEA and NEIPA caused 7.6- and 4.6-fold increases in PHHs and 11.8- and 13.2- fold increases in PMHs, respectively.

3.5. Comparison of DNA damage potency using benchmark concentrations (BMC) analysis

DNA damage induction by the eight nitrosamines was compared using BMC analysis. BMC₅₀ values, along with their lower and upper 90 % CIs, were calculated using the CometChip data derived from PHHs and PMHs. Data from 3D HepaRG models generated in our previous study [8] were included for comparison. All eight nitrosamines induced DNA damage in all three liver cell models. The BMC₅₀ values for the DNA damage induced by the eight nitrosamines in PHHs, PMHs, and 3D HepaRG cells spanned approximately 4, 3, and 3 orders of magnitude, respectively (Table 2). In addition, while many of the responses had overlapping CIs, most nitrosamines generated more conservative (lower) BMC₅₀ values in PHHs and PMHs than in 3D HepaRG spheroids.

Table 2.

Comparison of the benchmark concentrations (BMC₅₀) and potency ranking in PHHs, PMHs, and 3D HepaRG models.

Nitrosamine	BMC50 (BMCL50–BMCU50)a			BMC potency ranking
	PHHs	PMHs	3D HepaRGb	PHHs	PMHs	3D HepaRG
CPNP	26.7 (9.6–70.1)	31.8 (16.0–61.7)	289 (175–468)	5	5	5
NDBA	12.3 (3.9–36.4)	1.0 (0.6–1.6)	60.9 (38.2–92.5)	4	1	2
NDEA	0.3 (0.04–2.0)	6.7 (4.6–9.4)	238 (147–372)	2	3	4
NDIPA	417 (141–1,240)	900 (571–1,420)	4365 (2,530–8,530)	8	8	8
NDMA	0.01 (0.001–0.04)	2.0 (1.3–2.9)	1.9 (1.0–3.6)	1	2	1
NEIPA	6.2 (1.3–27)	10.9 (6.4–18.2)	979 (597–1,580)	3	4	7
NMBA	71.2 (18–277)	103 (59.3–173)	555 (326–933)	6	6	6
NMPA	181 (81–389)	842 (603–1,200)	88.6 (45.6–183)	7	7	3

PHHs, primary human hepatocytes; PMHs, primary macaque hepatocytes; 3D HepaRG, three-dimensional HepaRG spheroids.

^aThe BMC₅₀ values, corresponding to 50 % increases in DNA damage responses over the vehicle control, and their lower and upper bounds (BMCL₅₀ and BMCU₅₀) of the 90 % confidence intervals were calculated using exponential models of PROAST.

^b3D HepaRG CometChip data are from our previous study (Seo et al., 2023b).

The BMC-derived potency rankings were determined based on the order of BMC₅₀ CIs shown in Supplementary Fig. 2, with nitrosamines having lower BMC₅₀ CIs ranking higher in potency. Across the eight nitrosamines and three liver cell models, NDMA showed the highest DNA damage potency in PHHs, followed by NDEA (Fig. 5A and Table 2). NDBA was the most potent in PMHs, followed by NDMA and NDEA, while NDIPA was the least potent in all liver cell models. Three nitrosamines (NDEA, NDMA, and NMPA) had significantly lower BMC₅₀ CIs in PHHs than in PMHs, while the BMC₅₀ CIs for NDBA were significantly lower in PMHs than in PHHs. The BMC₅₀ CIs of the four other nitrosamines overlapped with each other in these two types of primary hepatocytes. A Pearson correlation-coefficient analysis of BMC₅₀ values derived from the DNA damage data revealed a strong correlation between PHHs and PMHs as well as between PHHs and 3D HepaRG spheroids for the eight nitrosamines, with correlation coefficients of 0.755 and 0.750 (p < 0.05), respectively (Fig. 5B and C).

Fig. 5. Comparison of nitrosamine-induced DNA damage potency.

(A) Benchmark concentration (BMC)₅₀ values and the 90 % confidence intervals (CIs) were calculated from CometChip data using PROAST. The exponential model was applied to compare BMC values between different liver cell models. The bars represent the uncertainty of BMC₅₀ estimates (the range between BMCLs and BMCUs) and are used to differentiate between responses based on non-overlapping CIs. The DNA damage concentration-response data for 3D HepaRG spheroids were obtained from our previous study (Seo et al., 2023b). Red, PHHs; Blue, PMHs; Black, 3D HepaRG. The correlation of BMC₅₀ values between PHHs and PMHs (B) or between PHHs and 3D HepaRG spheroids (C) was evaluated using the Pearson correlation coefficient test. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

3.6. Expression of protein biomarkers of DNA damage and apoptosis induced by eight nitrosamines in PMHs

Western blots were performed to assess whether the eight nitrosamines activate DNA damage and apoptosis pathways in PMHs at the protein level (Fig. 6A). At the tested concentrations, NDEA and NEIPA induced significant, concentration-dependent increases in γH2A.X and proteins involved in DNA damage regulation, including GADD45-α, p-p53, and p21 following a 24-h exposure (Fig. 6B). NDMA showed statistically significant increases only in γH2A.X and p-p53 protein levels in PMHs. In addition to DNA damage markers, NDEA significantly elevated the apoptosis markers, c-PARP and c-Cas3, while NEIPA only increased c-Cas3. Overall, the nitrosamines increased expression levels of DNA damage and apoptosis protein markers in the order of NDEA > NEIPA > NDMA. Other apoptosis markers such as the anti-apoptotic proteins BclxL and BcL2L10 as well as pro-apoptotic proteins Bad and Bax did not show significant changes after a 24-h exposure to the eight nitrosamines in PMHs (Supplementary Fig. 3).

Fig. 6. Expression of DNA damage and apoptosis markers by eight nitrosamines in PMHs.

(A) Total cellular proteins were extracted from PMHs after a 24-h treatment with eight nitrosamines. The highest concentration tested in the CometChip assay (#2) and a 10-fold lower concentration (#1) were used for the treatment. Protein expression levels of DNA damage and apoptosis markers were detected by Western blot analysis. GAPDH was used as a loading control. (B) The intensity of γH2A.X, GADD45-α, p-p53, p21, c-PARP, and c-Cas3 was normalized to the intensity of GAPDH. The bar graphs are the mean ± SD of fold changes compared to the vehicle control from at least three independent experiments. Statistical significance was determined by one-way ANOVA followed by Dunnett’s test (*p < 0.05, **p < 0.01, ***p < 0.001 vs. NC, negative control). c-Cas3, cleaved caspase 3; c-PARP, cleaved poly (ADP-ribose) polymerase; GADD45-α, growth arrest and DNA damage 45-α GAPDH, glyceraldehyde-3-phosphate dehydrogenase; p-p53, phosphorylated-p53; and γH2A.X, phosphorylated histone H2A.X.

4. Discussion

The development of the EAT represents a significant milestone in evaluating the mutagenicity of small-molecule nitrosamine drug impurities and NDSRIs [24]. In many instances, Ames results have been further studied or confirmed using the in vivo transgenic rodent (TGR) gene mutation assay [25]. Given the high-cost and ethical considerations of using animals, as well as the time-intensive nature of conducting in vivo assays, it is critical to develop reliable in vitro mammalian cell assays, especially assays using human-derived cell models that are metabolically competent, as follow-up assays to confirm Ames findings. This is especially true for negative Ames findings. For example, the FDA is currently not accepting negative EAT results alone to qualify a nitrosamine drug impurity as non-mutagenic [5].

Our previous studies evaluated the genotoxicity of several small-molecule nitrosamines using human lymphoblastoid TK6 cells transduced with human CYP2A6 and using HepaRG cells cultured in both 2D and 3D formats [8,26]. This present study further evaluated the DNA damage induced by these nitrosamines in primary hepatocytes, a well-recognized in vitro cell model that closely resembles in vivo liver function [27]. Two types of primary hepatocytes (PHHs and PMHs, derived from human donors and NHPs, respectively) were used to assess nitrosamine-induced DNA damage. All eight assayed nitrosamines induced positive responses in both types of primary hepatocytes, but differences in cytotoxicity and % tail DNA were seen between PHHs and PMHs (Figs. 3 and 4). Generally, PMHs were more resistant to nitrosamine-induced cytotoxicity and detected higher levels of DNA damage compared to PHHs. A possible explanation for this difference is that the PMHs used in the present study were freshly isolated from the livers of rhesus macaques in our facility, while PHHs were purchased as cryopreserved cell cultures from a commercial vendor.

Even though differences in cell preparation and cryopreservation could have affected the results, in most cases, the different responses produced by PMHs and PHHs also can be accounted for by significant differences in gene expression and metabolic activity of several major CYPs known to metabolize nitrosamines (Figs. 1 and 2). For example, NDMA, metabolized by CYP2E1 [28], induced higher fold-increases for DNA damage in PHHs at lower concentrations compared to PMHs (Table 1), which aligned with the higher CYP2E1 gene expression in PHHs than in PMHs (Fig. 1). In contrast, NDEA and NDBA induced higher % tail DNA levels in PMHs than in PHHs. CYP2A6 and CYP1A1 were reported to be the major CYPs activating NDEA and NDBA, respectively [29], and both CYPs had significantly higher expression levels in PMHs than in PHHs. Similarly, other nitrosamines that also are metabolized by CYP2A6, such as NEIPA, NMBA, and NMPA [26], induced higher fold-increases for DNA damage in PMHs compared to PHHs. These observations demonstrate the importance of metabolic capacity of the in vitro cell model for evaluating the genotoxicity of nitrosamines.

In addition to the different DNA damage responses observed between species, interindividual variability was also seen among different donors. This was more obvious with hepatocytes derived from humans compared to rhesus macaques, likely due to the wider variation in levels of CYP gene expression and activity in PHHs than in PMHs (Fig. 2 and Supplementary Fig. 1). However, not all nitrosamines showed different DNA damage responses among donors. Nitrosamines that have a relatively short alkyl chain and are mainly metabolized by CYP2E1, i.e., NDMA (molecular weight 74 g/mol), had the most significant differences in % tail DNA among donors and the responses were closely related with the CYP2E1 activity of the donor. Specifically, NDMA induced the highest levels of DNA strand breaks in PHH_B which had the highest CYP2E1 activity (126 pmol/min/million cells; Supplementary Table 1), while PHH_C, with the lowest CYP2E1 activity (10.5 pmol/min/million cells) had the lowest (but still positive) DNA damage response among the three donors. When the number of the carbon atoms in the alkyl chains bound to the N-nitroso group increases, CYP2A6, along with other CYP450 isoforms (CYP2C8/2C9/2C19 and 3A4) become the predominant metabolizing enzymes for activating nitrosamines [29,30]. This was the case for NDEA (molecular weight 102 g/mol), and likely for NEIPA (molecular weight 116 g/mol), where the interindividual variability seen with NDMA was not as obvious for the relatively larger-sized nitrosamines that were activated by CYP2A6, i.e., NMBA, and NMPA (molecular weight 146 g/mol and 136 g/mol, respectively) (Fig. 3B). We previously found that NDMA induced significantly higher cytotoxicity, DNA damage, and mutation frequency in 3D HepaRG spheroids since the spheroids had approximately 3-fold higher CYP2E1 gene expression than 2D HepaRG cells [31]. These data highlight the importance of CYP2E1 in converting small-molecule nitrosamines such as NDMA and N-nitrosomethylethylamine (NMEA) into genotoxic and/or carcinogenic metabolites, especially for individuals with higher CYP2E1 levels [29].

Given that PHHs are considered to be the “gold standard” of in vitro cell models, the current study conducted quantitative comparison between PHHs and the two other liver models using statistical approaches such as BMC modeling, a preferred point of departure metric for quantitative interpretation of genotoxicity responses [32,33]. It is worth noting that while NDEA exhibits greater carcinogenic potency than NDMA in rodents [34], NDMA was the most potent genotoxicant in human liver cell models, followed by NDEA and NDBA. This inconsistency warrants further investigation to determine whether it stems from species-specific metabolic differences or other underlying mechanisms. Despite variations in the potency ranking of the nitrosamines across the three liver cell models, the BMC₅₀ values for both PMHs and 3D HepaRG spheroids correlated strongly with those of PHHs, indicating that both cell models generated DNA damage data that were comparable to PHHs. A closer comparison of the BMC₅₀ and CI values indicates that both PMHs and 3D HepaRG spheroids have their respective advantages and limitations. In particular, PMHs can be isolated using our established protocol and used immediately for the treatment, since cryopreserved cells have an approximately 15 % reduction of cell viability compared to freshly isolated PMHs [19,35]. The remaining cells can be cryopreserved, generating in-house stocks to ensure consistency in future analyses. However, PMHs have metabolic profiles that differ from those of PHHs, resulting in different potency ranking for test compounds, such as NDMA and NDBA (Table 2). In addition, similar to PHHs, PMHs cannot divide and are not suitable for assays requiring cell proliferation. In contrast, 3D HepaRG spheroids are derived from a progenitor cell line and can be readily stimulated to divide so that the genotoxic effects of any DNA damage can be assessed [36]. In fact, several core genotoxicity assays have been conducted in the 3D HepaRG cell model, including the comet assay, micronucleus assay and mutagenesis using error corrected next generation sequencing [15,16,31,37]. Although BMC values for 3D HepaRG spheroids were less conservative (higher) than those of PHHs for these eight nitrosamines, the great plasticity and the long-term functional stability of HepaRG cells make them a promising surrogate in vitro cell model for genotoxicity testing [38].

Taking advantage of a large stock of PMHs in our lab, we further explored the molecular mechanisms of nitrosamine-induced genotoxicity using PMHs. Given that all nitrosamines induced significant DNA damage in PMHs, surprisingly, only three out of the eight nitrosamines significantly increased the protein expression of the DNA damage marker γH2A.X (Fig. 6). Treatment with 27 mM NDMA induced apoptosis in HepG2 cells via a caspase-dependent pathway [39]. However, 10 mM NDMA did not upregulate the c-Cas3 apoptosis marker or its cellular substrate, c-PARP, in PMHs. In contrast, NDEA and NEIPA upregulated the expression of both apoptosis makers (except c-PARP for NEIPA) as well as GADD45-α, p-p53, and p21, proteins that regulate multiple cellular processes such as DNA repair, the cell cycle, cell growth, and apoptosis [40,41]. Both NDEA and NEIPA contain an ethyl group, suggesting that DNA ethylation results in the upregulation of proteins involved in DNA damage and apoptosis in PMHs.

In conclusion, the current study compared DNA damage caused by eight nitrosamines in primary hepatocytes derived from humans and macaques. Although differences in the % tail DNA were observed between the two species of hepatocytes and the responses appeared to be related to differences in CYP enzyme activities, all eight nitrosamines produced positive responses in both PMHs and PHHs. Generally, PHHs generated more conservative BMC values than PMHs. In addition, greater interindividual variability in CYP gene expression and activities as well as DNA damage responses was seen among humans compared to macaques. As far as using PHHs, PMHs or HepaRG cell data as part of a weigh-of-evidence analysis to confirm negative EAT findings, one of the nitrosamines, NMBA, although a rat carcinogen [42], was negative in the EAT [6]. The positive responses detected in genotoxicity assays employing PHHs, PMHs and HepaRG cells support the use of such assays to confirm findings, especially negative findings, in the EAT.

Supplementary Material

Appendix A. Supplementary data

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cbi.2025.111538.

Data availability

Data will be made available on request.