MicroRNA regulation of CYP 1A2, CYP3A4 and CYP2E1 expression in acetaminophen toxicity

MicroRNAs (miRNAs) that regulate the cytochrome P-450 isoforms involved in acetaminophen (APAP) toxicity were examined in HepaRG cells treated with APAP (20 mM). In-vitro studies found that APAP protein adducts were increased at 1 h, followed by ALT increases at 12 and 24 h. CYP1A2, CYP3A4 and CYP2E1 mRNA levels were decreased, while miRNAs were increased for miR-122-5p, miR-378a-5p, miR-27b-3p at 6 h and miR-125b-5p at 12 h and miR-27b-3p at 24 h. Putative miRNA binding sites on the 3′UTRs of the CYPs were identified in-silico. Overexpression of miR-122-5p and miR-378a-5p in cells suppressed protein expression of CYP1A2, CYP3A4 and CYP2E1. Luciferase reporter assays confirmed the interaction between miR-122 and the 3′UTR of the CYP1A2 and CYP3A4. Thus, the in-vitro experiments showed that miR-122-5p and miR-378a-5p upregulation were associated with translational repression of CYPs. Serum samples of children with APAP overdose had significant elevation of miR-122-5p, miR-378a-5p, miR-125b-5p and miR-27b-3p, compared to healthy controls and receiver operator curves of the miRNAs had AUCs of 91 to 100%. Collectively, the data suggest that miRNA elevations in APAP toxicity represent a regulatory response to modify CYP1A2, CYP3A4 and CYP2E1 translation due to cellular stress and injury.

IU/L + 0.44). Consistent with previous data 29 , a dose response pattern was observed for APAP protein adducts in the APAP 5 and 20 mM exposed cells (Fig. 1B). Figure 1A,B demonstrates that ALT levels increased over time in both dose groups at 24 h (*p < 0.05).
Cells treated with 5 or 20 mM APAP released miR-122-5p, miR-378a-5p, miR-27b-3p, and miR-125b-5p into media as a function of both APAP dose and time. Consistent with the ALT and adduct profiles described above, the most marked changes in miRNA profiles were observed in the APAP 20 mM cells. Figure 1C shows down-regulation of all miRNAs at 1 h in the APAP 5 mM cells, followed by elevation of miR-122-5p at 6 h, miR-378a-5p at 6 h, and miR-27b-3p at 24 h. In contrast, the APAP 20 mM cells (Fig. 1D) had a significant elevation of miR-122-5p at 6 h, miR-378a-5p at 6 h, miR-125b-5p at 12 h, and miR-27b-3p at 24 h. Thus, the data demonstrate Time and dose-dependent changes in HepaRG cells with APAP treatment for ALT, APAP protein adducts, miRNA (miR-122-5p, miR-378a-5p, miR-27b-3p, and miR-125b-5p and mRNA levels (CYP2E1, CYP3A4, and CYP1A2) levels. Effect of 5 and 20 mM APAP treatment of HepaRG cells at 1, 6, 12 and 24 h time points. (A and B) Show levels of ALT and APAP protein adducts in cell medium; *p < 0.05 compared to controls for APAP 5 mM and 10 mM. Parenthesis around the asterisk denotes significance of ALT for 5 mM APAP treatment at 12 and 24 h. (C and D) Depict miRNA expression in cell medium. Data were normalized for HepaRG cell culture medium with Let-7d and spiked-in C. elegans miR-39; *p < 0.05 compared to controls. (E and F) Describe CYP1E2, CYP1A2 and CYP3A4 mRNA levels determined by qRT-PCR. TaqMan assays were duplexed with GAPDH to normalize the mRNA expression. Parenthesis around the asterisk differentiate 24 h CYP3A4 from CYP1A2; *p < 0.05 compared to controls. "h" Denotes hour. Error bars represent Standard error of the mean.
that HepaRG cells generate dose-response and temporal data for toxicity and oxidative drug metabolism endpoints known to be important in models of APAP toxicity. Furthermore, miRNA profiles were distinct between the 5 ("low dose") and 20 mM ("high dose") exposed cells.
Effect of APAP treatment on CYP2E1, CYP1A2 and CYP3A4 mRNA expression. To understand temporal relationships between miRNA profiles and gene expression for DME important in APAP toxicity, gene expression of CYP2E1, CYP1A2 and CYP3A4 was examined in cells exposed to APAP 5 mM or 20 mM. As shown in Fig. 1E, CYP2E1 mRNA expression in cells treated with 5 mM APAP was down regulated by 2.2 fold (p < 0.05) at 24 h. Transient down-regulation of CYP1A2 mRNA expression was observed with 5 mM APAP at 1 h (p < 0.05; Fig. 1E), while increased expression was apparent at 24 h. CYP3A4 mRNA expression was increased at all-time points for the APAP 5 mM (Fig. 1E),while the APAP 20 mM cells had down regulation of CYP2E1 at 6, 12 and 24 h (*p < 0.05, Fig. 1F). In contrast, the 20 mM APAP dose produced marked down regulation of CYP1A2 at all-time points, with a 26 fold decrease by 12 h (*p < 0.05, Fig. 1F). Moreover, cells exposed to 20 mM APAP had decreased CYP3A4 mRNA expression at 12 and 24 h (*p < 0.05). miRNA target sites located on promoters, 5′UTR, CDS and 3′UTR of CYP genes. The miR-Walk2.0 in-silico meta-analysis showed that hsa-miR-122-5p, hsa-miR-378a-5p, hsa-miR-27b-3p and hsa-miR-125b-5p have target sites within the 3′UTR region of CYP1A1, CYP1A2, CYP3A4 and CYP2E1 genes. Moreover, a total of 47 interactions were identified between the CYP1A1, CYP1A2, CYP3A4 and CYP2E1 genes and the 4 miRNAs (Table S1). Among these interactions, 15, 5, 16 and 11 were detected on the promoter, 5′ UTR, CDS and 3′ UTR regions, respectively. The mRNA-miRNA interactions suggest preferential control at transcriptional, as well as translational levels. The mRNA 5′ and 3′UTR regions were predicted to harbor several miRNA binding sites. Collectively, the in-silico analysis suggests cooperativity and multiplicity of miRNA binding sites on gene regulation during APAP toxicity. miR-122 targets 3′ UTR of CYP1A2 and CYP3A4. To further examine the predicted relationships between miR-122 and CYP1A2 and CYP3A4 (Table S1), HEK293T cells were transfected with 3′UTR luciferase plasmids with miR-122 mimic and inhibitor. Luciferase activity in HEK293T cells expressing the CYP1A2-Luc-3′UTR was inhibited by overexpressing the miR-122 mimic ( Fig. 2A) and reporter activity returned to control level (determined by transfection of pEZX-MT06) following transfection with the miR-122 inhibitor. However, overexpression of the miR-122 mimic marginally decreased CYP3A4 3′UTR luciferase activity, while addition of the miRNA-122 inhibitor partially reversed luciferase activity to the level of control plasmid (Fig. 2B). 3′UTR reporter assays demonstrated that the miR-122 mimic reduced luciferase activity, confirming the presence of miR-122-binding sites within the 3′UTR of CYP1A2 and CYP3A4 genes. Mohri and colleagues 32 previously confirmed the human CYP2E1 3′UTR target for miR-378a.
Gain-of-function and loss-of function studies: impact of miRNAs on protein expression and toxicity. In further studies, the effect of the miR-122-5p and miR-378 mimics and inhibitors on protein expression was examined. CYP1A2 protein expression was reduced by 90% at 12 h in APAP 20 mM cells, compared to controls (Fig. 3A). Transfection with the miR-122-5p mimic attenuated the effect of APAP on CYP1A2 protein expression; CYP1A2 protein expression was reduced by 40% compared to controls. In contrast, transfection of the miR-122-5p inhibitor restored CYP1A2 protein expression to that of controls (Fig. 3A).
APAP reduced CYP3A4 protein levels by 83% (Fig. 3B), whereas transfection of cells with the miR-122 mimic reduced CYP3A4 protein levels by only 28%. In addition, transfection of cells with the miR-122 inhibitor restored CYP3A4 expression to that of controls (Fig. 3B). Figure 3C demonstrates that APAP reduced CYP2E1 protein expression by only 26%, whereas cells transfected with the miR-378 mimic had reduced CYP2E1 protein expression by 33%. Transfection with the miR-378 inhibitor restored CYP2E1 protein to that of controls.
Non-APAP exposed cells transfected with mimics and inhibitors of miR-122 and miR-378a had no changes in protein expression for CYP1A2, CYP3A4 and CYP2E1 [Supplementary Figure S1A-F and Figure S2)]. Thus, collectively the data show that gain of function of miR-122 repressed CYP1A2 and CYP3A4, while gain of function of miR-378a repressed CYP2E1 protein levels. Figure 3D illustrates the effect of mimic and inhibitor transfections on trends in ALT levels. APAP treated cells had increased ALT levels compared to controls. Both the APAP/miR-122 mimic treated cells and the APAP/miR-378 mimic treated cells had higher ALT levels than the APAP only cells. Cells treated with the miR-122 inhibitor or the miR-378 inhibitor had ALT levels comparable to the APAP only cells.
Elevated levels of APAP protein adducts, ALT, and CYP-binding miRNA in APAP overdose subjects. To examine the clinical relevance of the in-vitro data, serum samples from children with APAP overdose were analyzed for APAP protein adducts, ALT, and miRNA. Demographic, toxicity, and APAP protein adduct data are provided in Table 1. Consistent with previous data 25,33,34 , ALT levels and APAP protein adducts were markedly elevated in the serum samples of APAP overdose subjects compared to healthy controls.

Discussion
The present study used HepaRG cells and clinical samples to examine the expression of miRNAs identified to regulate CYPs known to be involved in the metabolism and subsequent development of APAP toxicity [35][36][37] . Initial in-silico analysis determined that miR-122-5p, miR-378a-5p, miR-27b-3p and miR-125b-5p target CYP1A2, CYP3A4 and CYP2E1 genes. Direct suppression of CYP1A2 and CYP3A4 through 3′UTR binding of miR-122 was confirmed by luciferase assays. Gain-of-function and loss-of-function studies demonstrated that miR-122 and miR-378a downregulated CYP 1A2, CYP3A4, and CYP2E1 expression in cells at the translational level. Treatment of cells with two concentrations of APAP ("low" and "high" dose) generated time course toxicity and metabolism data consistent with previous data 29,38 , and indicated differential down regulation of CYP2E1 and CYP1A2 mRNA as a function of APAP dose. In contrast, upregulation of CYP3A4 expression occurred with APAP 5 mM, replicating findings from a recent report 31 , while APAP 20 mM led to down regulation of CYP3A4 expression. APAP treatment resulted in decreased expression of miRNA (miR-122-5p, miR-125b-5p, miR-378a-5p and miR-27b-3p) in cells and increased expression of miRNA's in media (Fig. 1C,D). These data are consistent with a report by Wang and colleagues 9 who noted the down-regulation of hepatic miRNAs and subsequent elevation in miRNAs in the plasma of mice treated with a toxic dose of APAP 9 .
Elevations of CYP-regulating miRNAs correlated with elevations of APAP protein adducts in APAP treated cells and in patients with APAP toxicity. Initial studies of miRNA expression in APAP toxicity reported elevations of the liver-specific miR-122-5p and its correlation with ALT elevation 9,10 ; miR-122 may also represent a predictive biomarker of APAP liver injury 11 and resolving liver injury 13,14 . Yamaura et al. used a rat model of APAP toxicity and described the elevation of miR-122, as well as miR-192, miR-685, miR-193 and miR-29c 12 . High throughput omic analysis using human APAP overdose samples suggested that miR-122-5p 15-18 , miR-27b-3p 15,16 and miR-125b-5p 15,17 were sensitive and noninvasive biomarkers for APAP toxicity. In the present study, we observed a strong association between miR-122-5p, miR-378a-5p, miR-125b-5p, and miR-27b-3p and APAP liver injury (defined by ALT elevation; Fig. 4B).
Prediction algorithms have identified a number of miRNAs that may target CYPs 3′UTR regions [39][40][41] in drug metabolism. To our knowledge, the present study is the first to use this approach in APAP toxicity. Of the potential miRNA binding sites identified within the gene regions of the CYPs (CYP2E1, CYP3A4 and CYP1A2) (Table S1) the greatest number of interactions were found for the promoter, CDS and 3′UTR. Taken together in APAP exposed cells down regulation of CYP1A2, CYP3A4 and CYP2E1 (Fig. 1E,F) are mediated by interactions between miRNA binding sites and the CYPs gene regions (promoter, CDS and 3′ UTR). Our findings are in accordance with the previous studies reporting cooperative miRNA-target interactions for 3′UTR and the CDS or 5′UTR region, resulting in strong gene downregulation [42][43][44][45] . The analysis also identified region-specific miRNA binding sites (promoter and CDS) for CYP members, consistent with a previous report that suggested that some miRNAs preferentially base-pair to the CDS region, suggesting a key role in rapid inhibition of translation 44 . These interactions, or co-targets, have previously been demonstrated in a number of studies 42-45 , but not in APAP toxicity. As each miRNA can target roughly 200 transcripts 46 , the number of putative interactions may be very large to have an effect on gene regulation in APAP toxicity.
Target prediction and in-vitro functional studies using 3′UTR reporter assays showed that CYP1A2 and CYP3A4 were a direct target of miR-122. The present study also corroborates the work of Mohri and colleagues 32 , who showed that 3′ UTR of CYP2E1 was regulated by miR-378 in HEK293 cells, mainly via translational repression 32 . The data presented herein are consistent with the hypothesis that miR-122-5p and miR-378a-5p upregulation in APAP exposed cells may represent a protective mechanism to lower CYPs expression (Fig. 6) and the formation of the reactive metabolite, N-acetyl-para-quinone imine (NAPQI), and resulting generation of APAP protein adducts. The data also suggest that the predominant effect of miR-122 and miR-378a on CYP1A2, CYP3A4 and CYP2E1 is exerted at the translational level ( Fig. 3A-C) and is consistent with recent data from primary cells and tissues from miRNA mutant mice 5 showing that 48% of target genes are regulated by translational repression. Thus, the data provide new insights into the function of miRNA in the regulation of CYP expression in APAP toxicity and suggests that miRNAs may be a potential target for novel therapeutic approaches in APAP toxicity.
One limitation of the present study its reliance on HepaRG cells which do not fully capture the inherent biologic variability and microenvironment of the human liver. It is generally regarded that APAP toxicity is a direct necrotic event, although signaling cascades invoked by cytokine release from inflammatory cells and other pathways may modulate the toxicity. Since the liver comprises a variety of cell types, further well characterized in-vivo studies are necessary to fully understand the miRNA regulated signaling pathways in APAP toxicity.
To summarize, serum miR-122-5p was highly increased in children with APAP overdose, similar to data published in adults. Three other miRNAs (miR-378a-5p, miR-125b-5p, and miR-27b-3p) were also increased, but the relative increase was lower than that of miR-122-5p. Strong correlations were noted for ALT and miR-122-5p, miR-378a-5p and miR-125b-5p (R = 0.76-89, p < 0.05), while moderate correlations were noted for APAP protein adducts and the miRNAs (R = 0.58-0.67, p < 0.05). These elevations may represent cell-to-cell communication in an attempt to minimize toxic events in cells or stimulate hepatocyte repair responses. A better understanding of the function of miRNAs in APAP-induced hepatotoxicity may have relevance for the development of new treatments for APAP toxicity and/or understanding individual susceptibility to drug toxicity. The results provide new insights into the protective role of miR-122 and miR-378a in suppressing the expression of DMEs.

Methods
Chemicals and materials. APAP, nuclease free water and ethanol (Molecular grade) were purchased from Sigma-Aldrich (St. Louis, MO). miRNA and RNA isolation, cDNA synthesis and pre-amplification reagents were purchased from Qiagen (Valencia, CA). Collagen coated 24 well plates and phosphate buffered saline were obtained from ThermoFisher Scientific (Carlsbad, CA). The miR-122 mimic, miR-122 inhibitor, miR-378a mimic, miR-378a inhibitor, AllStars Negative control siRNA and Hs Cell Death Control siRNA were purchased from Qiagen (Valencia, CA). HepaRG cells. HepaRG cells (ThermoFisher Scientific,Carlsbad, CA) were seeded in base medium per the manufacturer's protocol. Cell viability was determined as per Trypan blue method and 330,000 cells were seeded per well in collagen-coated 24 well plates. As per protocol, from day 2-9 base medium was supplemented with Tox Medium Supplement (ThermoFisher Scientific, Carlsbad, CA). APAP treatment, mRNA/miRNA isolation, cDNA synthesis and quantitative real-time PCR (qRT-PCR). Cells were incubated with 5 or 20 mM APAP 26,29-31 on day 7 for 1, 6, 12, or 24 h time points.
Human samples. Peripheral blood samples were collected as part of a multicenter investigation of APAP toxicity in children of ages 2-18 years, approved by University of Arkansas for Medical Sciences (UAMS) institutional review board (IRB) 33 . Subjects were categorized as controls (n = 15 healthy children with no use of APAP in the preceding 2 weeks) or APAP overdose (n = 12; children that required hospitalization for treatment of APAP overdose) and informed consent was obtained from all participants per UAMS IRB policy guidelines and all experiments were performed in accordance with the relevant guidelines and regulations. A single blood sample was collected from control subjects, whereas daily samples were collected from the APAP overdose subjects. Blood samples were centrifuged within 30 minutes of collection and the serum was stored at −80 °C. Demographic information available for study subjects included subject age, gender, weight, height, body surface area, past medical history, reason for hospitalization, relevant history concerning recent APAP dosing and concomitant medications. Clinical laboratory results included APAP concentration and ALT determinations.
Quantitation of APAP protein adducts. APAP protein adducts were analyzed using a method previously published by our laboratory 24,47,48 . Briefly, 100 µl of serum or cell supernatants was gel filtered, hydrolyzed with protease, precipitated, and injected onto an HPLC system. APAP cysteine was resolved on a 150 mm C18 column and detected using an ESA CoulArray electrochemical detector. Concentrations of adducts were determined relative to a standard curve of authentic APAP cysteine and reported as nmoles a-cys/mg protein (in-vitro data) and nmol/L a-cys (human serum samples).

Transfection of microRNA mimics and inhibitors. Mimics and inhibitors of miR-122 and miR-378
were used to overexpress or knockdown miRNA in HepaRG cells. Cells were grown in 24 well plate (see cell culture) and transfected with 5 nM mimic, 50 nM or 100 nM inhibitor using RNAiMAX Transfection Reagent (ThermoFisher Scientific, Carlsbad, CA). At 24 h, the transfection medium was replaced with HepaRG cells Tox Working Medium. At 72 h, cells were cultured in glutathione free media. APAP (20 mM) was added for 12 h and cells were harvested at 96 h post transfection for protein analysis. Transfection efficiency of 65-89% was observed with 5-50 nM concentration of AllStars Hs Cell Death Control siRNA. Preliminary studies found no interference in protein expression with negative control AllStars siRNA.
Protein isolation and Western blot analysis. After transfections, cells were lysed with ice-cold RIPA lysis buffer (ThermoFisher Scientific, Carlsbad, CA) containing protease inhibitors (Pierce Protease Inhibitor Mini Tablets, Rockford, IL and Mammalian Protease Arrest, GBiosciences, St. Louis, MO) and centrifuged at 12,500 rpm for 10 min at 4 °C. Cell supernatants were used for immunoblots. Proteins were quantified using Pierce BCA Protein Assay Kit (ThermoFisher Scientific, Carlsbad, CA) and 20 ug of protein was loaded on to precast polyacrylamide Bis-tris 10% gels (GenScript, Piscataway, NJ). The membranes were blocked with 5% skim milk and incubated with primary antibodies against CYP2E1, CYP1A2 and CYP3A4 (Abcam, Cambridge, MA) and β-actin (Sigma, St. Louis, MO). Next membranes were incubated with HRP-conjugated secondary antibodies to goat anti-rabbit IgG H&L (Abcam, Cambridge, MA). The protein bands were visualized using the SuperSignal West Pico Chemiluminescent Substrate (ThermoFisher Scientific, Carlsbad, CA) and densitometry was performed with the ImageJ software program (NIH, Bethesda, MD).

Statistical analysis.
Statistical analyses of cellular data used one-way ANOVA with post hoc Dunnett's test to determine differences between treatment groups and the control. The unpaired Student's t-test was used for comparison between two cell groups. The non-parametric Mann-Whitney U test was used for pairwise comparisons between clinical groups for toxicity (ALT), metabolism (APAP protein adducts) and miRNA data. Spearman rank correlation analysis was used to assess the pairwise relationships among miRNAs, ALT and APAP protein adducts; data are reported as medians, minimums, and maximums. A p value < 0.05 was considered significant for all analyses. Receiver operating curve (ROC) analysis was used to assess the performance of each miRNA as a binary classifier of toxicity outcomes.

Disclosure.
Dr. James has a patent application pending for the measurement of APAP protein adducts in human blood samples.