Acetaminophen was purchased from MedChem Express (Monmouth Junction, NJ 08852, United States); Thiazolyl Blue Tetrazolium Bromide, α-Keto-glutaric Acid, L-Alanine, Thiobarbituric Acid, Trichloroacetic Acid, 2′-7′-Dichlorodihydrofluorescein diacetate (H₂DCFDA), silymarin, JNK inhibitor SP600125, MITOISO2 – Mitochondria Isolation Kit were purchased from Sigma–Aldrich (St. Louis, MO, United States); M-PER Mammalian Protein Extraction Reagent, NE-PER^® Nuclear and Cytoplasmic Proteins Extraction Kit, Halt protease inhibitor cocktail EDTA-Free 100X, Pierce bicinchoninic acid (BCA) Proteins Assay Kit, SuperSignal West Pico Chemiluminescent Substrate were all purchased from Thermo Fisher Scientific (Rockford, IL, United States). Rabbit polyclonal anti-CYP2E1 antibody (1:1500 dilution) was purchased from Sino Biological Inc. (Beijing, China); Rabbit polychonal anti-phospho-JNK1/JNK2 and JNK2 antibodies (1:1000 dilution) were purchased from Signalway Antibody (Baltimore, MD, United States); Rabbit polyclonal anti-MKP-1, Nrf2, Keap-1, AIF, COX IV, Lamin B and mouse polyclonal anti-Bax antibodies (all 1:1000 dilution) were purchased from Beijing Biosynthesis Biotechnology CO., LTD. (Beijing, China); Mouse monoclonal anti-βactin primary antibody (1:5000 dilution), horseradish peroxidase-conjugated goat anti-rabbit and anti-mouse IgG AP-linked secondary antibodies (1:2000 dilution) were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, United States); TRIzol^® Reagent was purchased from Ambion Life Technologies (Carlsbad, CA, United States); First-Strand cDNA Synthesis Kit was purchased from Promega (Madison, WI, United States); iTaq Universal SYBR Green Supermix Kit was purchased from Bio-Rad Laboratories (Hercules, CA, United States); All primers of the genes of interest were synthetized by TSINGKE Biological Technology Company (Beijing, China). All others reagent used in this study were of analytical grade.

Stem barks of K. grandifoliola were collected in June 2015 in Foumban (West Cameroon). The botanical identification of the plant was done at the Cameroon National Herbarium, where voucher specimen is kept under the reference number 23434 YA. K. grandifoliola most active fraction, namely KgF25 (methylene chloride/methanol 75:25, v/v) was prepared as previously described (Njayou et al., 2015, 2016). Therefore, 8 g of KgF25 were subjected to silica gel 60 (particle size 40–63 μm) column chromatography (column: d × h = 3.5 cm × 65 cm) and gradient of hexane/acetyl acetate (100:0 v/v to 0:100 v/v), then acetyl acetate/methanol (10:90 v/v to 0:100 v/v) were used for elution. The 47 chromatographic fractions containing crystalline solid obtained were analyzed by thin layer chromatography (TLC) and pooled into 5 sub-fractions namely KgF25sf1 (41 mg); KgF25sf2 (67 mg); KgF25sf3 (19 mg); KgF25sf4 (105 mg); and KgF25sf5 (55 mg) based on the similarity of their TLC profiles.

Following the preliminary results, the most active sub-fractions were analyzed by HPLC-ACN (Acetonitrile)-Standard-Method and compounds were purified by HPLC-ACN-Standard-Purify-Method. We used a Series 1200 Liquid Chromatograph (Agilent Technologies, Folsom, CA, United States) equipped with a vacuum degasser, a quaternary pump, an autosampler and a Diode-Array-Detector (DAD) connected to Agilent ChemStation software. An Eclipse XDB-C8 column (9.4 mm × 250 mm, 5 μm particle size) was used. The mobile phase was (A) water; (B) acetonitrile. The flow rate was 1 mL/min. The elution conditions were: B, 0–15 min, increasing gradient from 0 to 30% B; 15–20 min, linear gradient 100% B; 20–25 min, linear gradient 30% B. The system operated at 28°C and the injection volume of each sample dilute in methanol (1 mg/mL) was 5 μL. The detection wavelength was kept at 254 nm.

High-Resolution Mass Spectra (HRMS) of each pure sample dissolved in methanol was measured on an Agilent 1200 HPLC/6520 High-resolution mass spectrometer. NMR of each pure sample dissolved in DMSO (5 mg/mL) was recorded on BRUKER AV 500 spectrometer.

The normal human liver cell line L-02, (Cell Bank, Type Culture Collection of Institute of Microbiology, Chinese Academy of Sciences, Beijing, China) derived from an adult human normal liver were used. Cells were cultured in 100 mm dish and maintained in high glucose Dulbecco’s Modified Eagle’s Medium supplemented with 10% fetal bovine serum, L-glutamine (2 mM), penicillin (100 IU/mL), streptomycin (100 μg/mL) and amphotericin B (0.25 μg/mL) in an atmosphere of 5% CO₂ at 37°C.

In this study, silymarin and JNK-inhibitor SP600125 were used as reference compounds. APAP, silymarin, sub-fractions of KgF25 and isolated compounds were diluted in 20% DMSO [20% DMSO in phosphate-buffer saline (PBS)]. JNK-inhibitor SP600125 was dissolved in 100% DMSO. L-02 cells (approximately 2.10⁵ cells/ml) in triplicate were seeded into 24-well plate labeled as control, APAP, standard and test (references or plant sample + APAP) and incubated for 24 h. Thereafter, the medium was replaced with fresh medium and cells were incubated in absence (control group) or in presence of APAP (APAP group), or in presence of APAP and references or plant sample (test groups) for different time points (6; 12; 24; and 36 h) depending on the downstream analysis.

Cells were incubated in presence of APAP at the final concentrations of 0; 5; 10; 15; 20; and 30 mM. After 36 h, cell viability was measured using 3-(4, 5-dimethylthiosol-2-yl)-2, 5-diphenyl-2H-tetrazolium bromide kit (MTT; Sigma–Aldrich) according to the manufacturer’s instructions and the cell membrane integrity was assessed by measuring ALT activity released into the incubation medium using a calibration curve according to Reitman and Frankel (1957). Cell viability was expressed as percentage of the control and the half lethal concentration (LC₅₀) of APAP was determined and used as toxic concentration for the following experiments.

Cells were treated simultaneously with APAP (LC₅₀) and plant sub-fractions or silymarin at the final concentration of 100 μg/mL for the hepatoprotective activity screening. After 36 h, cell viability and cell membrane integrity were assessed as abovementioned. Selected active sub-fractions and silymarin were tested at the final concentration of 25; 50; and 100 μg/mL for the dose-response study. Isolated compounds were then tested at 0; 10; 20; 30; and 40 μM and JNK inhibitor SP600125 at the final concentration of 0; 5; 10; 15; and 20 μM in order to minimize the effect of DMSO in APAP metabolism. After end time point, cell viability and membrane integrity were assessed and the half efficient concentrations (EC₅₀) determined.

Cells were simultaneously treated with APAP and isolated compounds or JNK inhibitor at the determined concentration. After 36 h, cells were washed with cold PBS and stained with 10 μg/mL of PI solution for 5 min. Cells were then stained with 1 μg/mL of 4′,6-diamino-2-phenylindole (DAPI) solution for 25 min and image were recorded with a fluorescence microscope (Advanced Microscopy Group).

Intracellular ROS were measured according to reported method (Ji et al., 2009). Briefly, cells were treated with 20 μM H₂DCFDA, APAP (10 mM) and isolated compounds or JNK inhibitor at the determined concentrations for 36 h. After treatment, supernatant was collected and cells were washed with PBS, immediately lysed in an appropriate lysis buffer, the whole cell lysates were centrifuged (10,000 g, 5 min, 4°C) and aliquot of lysate was used for fluorescence measurement at excitation 485 ± 20 nm, emission 525 ± 20 nm in a black wall with clear bottom 96-well plate using a spectrophotometer (SpectraMax M5, Molecular Devices) and presented as percentage of control. Lipid membrane peroxidation was evaluated through the malondialdehyde (MDA) level in the cellular supernatant by the thiobarbituric acid (TBA) method by using its molar extinction coefficient (ε_MDA = 1.56 × 10⁵ M^-1.Cm^-1) as described by Buege and Aust (1978).

After treating cells with APAP and isolated compounds or JNK inhibitor for the indicated time, cells were harvested and lysed in lysis buffer. The whole cell lysates were centrifuged (10,000 g, 5 min, 4°C). The supernatant was collected and the GSH concentration was determined by using its molar extinction coefficient (ε_GSH = 13,600 M^-1.Cm^-1) by the 5,5-dithio-bis (2-nitrobenzoic acid) (DTNB) assay as described by Ellman (1959).

Cells were simultaneously treated with APAP and plant sub-fractions, silymarin, isolated compounds or JNK inhibitor at the determined concentration and incubated for different time points. Afterward, total proteins were extracted using M-PER Mammalian Protein Extraction Reagent (Thermo Scientific) containing 0.2% Halt protease inhibitor cocktail EDTA-Free 100X (Thermo Scientific); cytosolic and nuclear protein were extracted using NE-PER^® Nuclear and Cytoplasmic Proteins Extraction Kit (Thermo Scientific); mitochondrial and cytosolic fractions were prepared with MITOISO2 - Mitochondria Isolation Kit (Sigma–Aldrich); protein concentration in each sample was quantified with Pierce BCA Proteins Assay Kit (Thermo Scientific). All procedures were performed according to each manufacturer’s instructions.

Equal amount of protein in Laemmli loading buffer (10 μl containing ≈ 50 μg of protein) was separated by 12 % Sodium-Dodecyl-Sulfate Poly-Acrylamide Gel Electrophoresis (SDS-PAGE) and electro-transferred into a Nitrocellulose Blotting Membrane (GE Healthcare, Life Science, Germany). The membrane were blocked with 5% w/v dehydrated skimmed milk in Tris-Buffered Saline Tween-20 (TBST: 10 mM Tris-HCl; 150 mM NaCl; 0.05% tween-20; pH 7.6); incubated overnight at 4°C with primary antibodies, rinsed, and then incubated 1 h at 25°C with horseradish peroxidase-conjugated secondary antibodies. Membrane was stained SuperSignal West Pico Chemiluminescent Substrate (Thermo Fisher Scientific, United States) and detection was performed by Enhanced Chemiluminescent Method which combines MicroChemi Unit and GelCapture Software. Densitometry analysis of the protein bands was performed using ImageJ Software.

L-02 cells were simultaneously treated with APAP and isolated compounds or JNK inhibitor and incubated for the indicated time. After treatment, Total RNA was extracted from cells using TRIzol^® Reagent (Ambion, Life Technologies) according to the manufacturer’s instruction. RNA concentration and purity were determined by reading the absorbance at 230; 260; and 280 nm by using a ND-2000 NanoDrop Spectrophotometer (Thermo Scientific). The 260/280 ratio of our RNA preparation ranged from 1.8 to 2.1 while the 260/230 ratio were greater than 2. These values indicate good RNA quality and therefore could be used for reverse transcription. First-strand synthesis cDNA was synthetized in the thermo cycler (Eastwin, Life Science) using Moloney Murine Leukemia Virus Reverse Transcriptase (M-MLV RT, Promega, Woods Hollow Road Madison, United States) as directed by the manufacturer.

Real-time qPCR was performed in Applied Biosystems 7500 System using iTaq^TM Universal SYBR^® Green Supermix (Bio-RAD Laboratories, Hercules, CA, United States) according to the manufacturer’s instruction. Relative expression of target genes was normalized to the endogenous gene (GAPDH) used as internal control, analyzed by the 2^-ΔΔC_T method using GenEX Software and given as ratio compared to the control group. All primers of interest genes were synthetized by TSINGKE Biological Technology Company (Beijing, China). Their sequences are presented in the Table 1.

Results are presented as mean ± standard deviation (SD) of three independent experiments in triplicate. Comparisons between the mean values of various treatments groups were analyzed by one-way analysis of variance (ANOVA) followed by the Bonferroni’s post hoc test whenever significant differences were observed between the variances. Comparisons were made between untreated group (DMSO control group) and intoxicated group (APAP-intoxicated group), and between APAP-intoxicated group and treated groups (APAP + tested or reference compounds). Differences between compared groups were considered significant for p < 0.05. Analyses were performed using Prism 5.03 statistical software (Graph Pad Inc.).

High Performance Liquid Chromatography chromatogram of KgF25sf1 (Supplementary Figure S1A) showed several peaks with different retention time. The higher peak (retention time: 11.22 min) was designated as compound A (C-A). HPLC chromatogram of KgF25sf2 (Supplementary Figure S1B) showed two main peaks with the retention time at 12.05 and 12.33 min and were designated as compound B (C-B) and (compound C) (C-C) respectively. C-A, C-B and C-C were purified and characterized. Comparison of theirs HRMS and NMR spectral data (Supplementary Figures S2–S10) with data literature (Atta-ur-Rahman et al., 2008; Tanaka et al., 2011, 2012) allowed for the unambiguously identification of three known limonoids (Figure 1) as 17-epi-methyl-6-hydroxyangolensate, 7-deacetoxy-7-oxogedunin and 7-deacetoxy-7R-hydroxygedunin, respectively, for C-A, C-B and C-C.

As a normal hepatic cell line, L-02 cells are commonly used in hepatotoxicity studies induced by various xenobiotic (Xiao et al., 2014; Pang et al., 2016). Extra cellular level of ALT activity, a cytosolic enzyme is routinely used to evaluate the membrane integrity during APAP toxicity study in vitro (Xie et al., 2014). We evaluated the toxicity of APAP in L-02 cells by performing a dose-dependent study with APAP concentration range from 5 to 30 mM. Exposure of L-02 hepatocytes to APAP for 36 h dose-dependently resulted in significant (p < 0.05) loss of cell viability (Figure 2A) and increase cellular ALT leakage into the incubation medium (Figure 2B). The LC₅₀ was found to be 9.80 ± 2.61 mM, and therefore, 10 mM of APAP was chosen as toxic concentration for the following experiment in this work.

The sub-fractions were tested against APAP-induced hepatotoxicity. As presented in Figures 3A,B, respectively, sub-fractions KgF25sf1, KgF25sf2, and KgF25sf3 significantly (p < 0.05) maintained cell viability and inhibited the release of ALT from the cell. These activities were not statistically different from that of silymarin, a well-known hepatoprotective flavonoids mixture. The above sub-fractions, being the most active, have been selected for the next part of the study.

Incubation of cells with APAP (10 mM) for 36 h resulted in a significant (p < 0.05) decrease of cell viability (Figure 4A) and increase of ALT (Figure 4B) activity leakage into the incubation medium. When hepatocytes were co-treated with active sub-fractions or silymarin at various concentrations (25; 50; and 100 μg/mL), a concentration-dependent protective and inhibitory effects on cell viability (Figure 4A) and ALT leakage (Figure 4B) was observed. Likewise, a similar effect on these parameters (Figures 4C,D) was observed when cells were co-treated with isolated limonoids (10; 20; 30; and 40 μM) or JNK inhibitor SP600125 (5; 10; 15; and 20 μM). The EC₅₀ were 19.04 ± 3.42 μM, 20.89 ± 3.68 μM, 17.40 ± 2.79 μM, and 13.76 ± 2.37 μM, respectively, for compound A (17-epi-methyl-6-hydroxyangolensate), compound B (7-deacetoxy-7-oxogedunin), compound C (7-deacetoxy-7R-hydroxygedunin) and JNK inhibitor SP600125. Although the EC₅₀ of JNK inhibitor was lower than those of isolated limonoids, JNK inhibitor offered only partial protection (cell viability = 85.66 ± 3.51%) while limonoids from K. grandifoliola showed almost total protection when added at 40 μM (cell viability ≥ 95%) as confirmed by PI cell death staining (Figure 5). When active sub-fractions or isolated compounds were added at 100 μg/mL or 40 μM, the viability of hepatocytes were not statistically different from those of control group and the levels of extracellular ALT activity were lower. Based on this observation, these concentrations were chosen for the following experiments.

Acetaminophen hepatotoxicity is initiated by its biotransformation by CYP2E1 to the electrophile NAPQI, and overexpression of CYP2E1 plays a critical role in this process (Jaeschke et al., 2002). To evaluate whether the protective effect of active sub-fractions and isolated compounds was related to the expression of CYP2E1, the protein expression of CYP2E1 was determined by western blot analysis after co-treatment of cells with APAP (10 mM) and active sub-fractions (100 μg/mL) or isolated compounds (40 μM). The constitutive up-regulation of CYP2E1 expression by APAP treatment was significantly (p < 0.05) down-regulated when cells were co-treated with active sub-fractions (Figures 6A,B) or isolated compounds (Figures 6C,D). This effect was comparable to that observed in silymarin (100 μg/mL) treated cells (Figure 6A). NAPQI, the reactive metabolite of APAP can react with the sulfhydryl group of cysteine in GSH, leading to its depletion and therefore enhance cellular oxidative stress (Jaeschke et al., 2012). As shown in Figure 6E, exposure to 10 mM APAP significantly (p < 0.05) reduced cellular GSH levels within 6 h, followed by further depletion during the next 30 h. Co-treatment of L-02 cells with isolated limonoids significantly (p < 0.05) attenuated GSH depletion after 6 h, followed by progressive restoration during the next 30 h. After 36 h, cellular GSH levels in co-treated cells were comparable to that of normal untreated cells.

Increase production of intracellular ROS occurring after GSH depletion plays an important role in APAP-induced injury (Jaeschke et al., 2002). Figure 7A depicts the effect of APAP-induced intracellular ROS generation in L-02 cells and its significant annihilation by co-treatment with isolated limonoids from K. grandifoliola. Hepatocytes stressed with APAP showed significant (p < 0.05) increase in ROS generation as compared to untreated cells. Co-treatment with isolated compounds (40 μM) or JNK inhibitor (20 μM) significantly (p < 0.05) decreased ROS level as compared to APAP-treated cells. Overproduction of ROS initiates the process of lipid peroxidation in cell membranes and causes the destruction of cellular integrity as well as cell death (Kumari and Kakkar, 2012b). Lipid peroxidation was measured by estimating the concentration of MDA, an end product of lipid peroxidation into the incubation medium. As presented in Figure 7B, oxidative stress induced by free radical due to APAP in the hepatocytes caused a significant (p < 0.05) increase of MDA production. However, the MDA formation was significantly attenuated when cells were co-treated with isolated limonoids (40 μM) or JNK inhibitor (20 μM).

Recently, JNK activation and the mitochondrial translocation of p-JNK have been well established to play a major role in APAP-induced hepatotoxicity (Gunawan et al., 2006; Hanawa et al., 2008; Xie et al., 2014). To explore the effects of isolated compounds on JNK mediating cell death, total proteins were initially extracted 6 h after treating L-02 cells with 10 mM APAP or co-treated with APAP and active sub-fractions (100 μg/mL) or limonoids (40 μM). The lysates were probed for JNK2 and p-JNK by western blotting. Administration of APAP alone resulted in massive p-JNK while co-treatment with active sub-fractions (Figure 8A) or isolated compounds (Figure 8B) significantly (p < 0.05) inhibited JNK activation.

Activated JNK translocate into mitochondria, leading to mitochondrial dysfunction and oxidant stress (Jaeschke et al., 2012). To determine whether isolated limonoids affect p-JNK translocation into mitochondria, cytosolic and mitochondrial fractions were prepared from L-02 cells and JNK activation was evaluated at 6 and 12 h after treatment. Similar to what was observed in total cell lysates, 10 mM APAP exposure caused JNK activation in the cytosol at 6 and 12 h (Figures 9A,B). We observed in the mitochondria of APAP treated cells a modest, but significant (p < 0.05) increase in p-JNK 6 h after APAP exposure (Figure 9A) and massive p-JNK at 12 h (Figure 9B). Co-treatment with isolated compounds prevented JNK activation in the cytosol at 6 and 12 h and its translocation into mitochondria (Figures 9A,B). These effects were comparable to those observed in JNK inhibitor treated cells.

Amplification of the mitochondrial oxidant stress by the translocation of p-JNK triggers the opening of the MPT pore in the mitochondria, followed by the mitochondrial Bax (Bcl₂-associated X protein) translocation. These effects are responsible for the translocation of mitochondrial intermembrane proteins such as AIF and endonuclease G to the nucleus as main cause of nuclear DNA damage and cell death after APAP overdose (Bajt et al., 2008, 2011; Saito et al., 2010). To investigate the effect of isolated limonoids in these events, the release of AIF into the cytosol and mitochondrial translocation of Bax were observed by immunoblotting in APAP-intoxicated cells 12 h after exposure (Figures 10A,B). Co-treatment of cells with isolated compounds (40 μM) or JNK inhibitor (20 μM) significantly (p < 0.05) prevented the mitochondrial translocation of Bax and the release of AIF into the cytosol (Figures 10A,B).

Mkp-1, a phosphatase which preferentially inactivates the stress induced by JNK activation in most tissues, has been found to protect against APAP-hepatotoxicity (Hammer et al., 2006; Wancket et al., 2012). However, overexpression of ROS decreases its expression (Kamata et al., 2005). In search of the protective mechanism of isolated limonoids from K. grandifoliola, the Mkp-1 protein level has been evaluated by immunoblotting. As shown in Figures 11A,B, exposure of L-02 hepatocytes for 6 h resulted in a marked decrease of Mkp-1 protein level. Co-treatment with silymarin (100 μg/ml) (Figure 11A) or JNK inhibitor SP600125 (20 μM) (Figure 11B) only attenuated this decreasing. However, co-treatment with active sub-fractions (100 μg/ml) (Figures 11A,C) or isolated compounds (40 μM) (Figures 11B,D) markedly increased the expression of Mkp-1 by up-to 1.8-fold as compared to the untreated cells.

Gene’s expression studies are useful supplements to protein examination, as the mRNA levels represent a snapshot of the cell activity at a given time. Thus, the effect of isolated limonoids was evaluated by qRT-PCR on the expression of the antioxidant enzymes, CAT, SOD1, and GST chosen as target genes of Nrf2. In parallel, the expression of MAT1A was also evaluated. As presented in Figure 13, exposure of cells to 10 mM APAP for 36 h significantly reduced the mRNA levels of CAT, SOD1, GST, and MAT1A. In limonoids co-treated cells, the mRNA levels of CAT (Figure 13A) and GST (Figure 13C) were increased by up-to 2.8-fold as compared to the untreated cells. Similarly, the mRNA levels of SOD (Figure 13B) and MAT1A (Figure 13D) were also increased by up-to 1.77-fold.