Both AM and DM were previously isolated from A. pilosa (Huang et al., 2015). The purities of AM and DM (95%) were further confirmed by using NMR and MS spectra. NMR spectra were recorded on a Bruker 400 Ultra Shield NMR spectrometer (¹H 400 MHz; ¹³C 100.61 MHz). Low resolution mass spectrometry was taken on a Triple Quadrupole mass spectrometer (Micromass Quattro Micro API, Waters). Column chromatography was carried out on SiO₂ (Merck 9385) or a Combiflash system (Teledyne-Isco; Lincoln, NE, USA) using SiO₂ or C₁₈ prepacked columns. TLC was performed on precoated SiO₂ F₂₅₄ plates (Merck 5554 and 5744) and visualized under UV light and by spraying with H₂SO₄-MeOH (1:1) followed by heating. HPLC was carried out on a Merck-Hitachi instrument, with UV detection (254 nm). Tested compounds were purified to ≥ 95% HPLC (Merck LiChrospher 100 RP-18; 5 μm, 125 × 4 mm column). The antibodies and horseradish peroxidase conjugated anti-goat secondary antibody were purchased from Santa Cruz Biotechnology (Santa Cruz, CA, USA).

1,1-Diphenyl-2-picrylhydrazyl (DPPH) radical inhibition assay was completed by following our previous study (Gleichenhagen and Schieber, 2016). A volume of 10 μL of sample was added to 190 μL solution of 0.4 mM DPPH dispersed in 95% ethanol. The sample was shaken vigorously and kept in the dark at room temperature for 30 min. Control was prepared using pure ethanol. The absorbance was measured at 517 nm.

2, 2′-Azino-bis-(3-ethylbenzothiazoline-6-sulphonic acid) (ABTS) scavenging activity assay was completed using the colorimetric method (Brenes et al., 2016). ABTS radical cation was prepared by mixing 5 mL of 7 mM ABTS solution with 45 mL of 2.45 mM K₂S₂O₄ solution. This mixture was stored in the dark at room temperature for 12–16 h before test. Prior to the assay, this mixture is diluted with 80% ethanol at an approximate ratio of 1:45 and adjusted to yield an absorbance of 0.70 ± 0.05 at 734 nm. A volume of 190 μL ABTS solution was added to 10 μL of sample and the mixture was vortexed for 1 min. After incubation at room temperature for 6 min the absorbance was measured.

This assay was performed according to our previous study (Chen and Kang, 2014b). Each sample or ascorbic acid (positive control) was mixed with Tris-hydrochloric acid buffers (50 mM Tris-aminomethane+10 mM EDTA, pH 8.5) and 7.2 mM pyrogallol. The reaction was stopped by the addition of HCl solution. Absorbance of the mixture was determined at 420 nm against a blank, and this assay was replicated for 3 times.

Human HepG2 cells were obtained from GibcoBRL (now Invitrogen Corporation, Carlsbad, CA, USA), and cultured in Dulbecco's modified Eagle's medium (DMEM) containing 10% FBS and 1% penicillin streptomycin under 5% CO₂ atmosphere at 37°C. Compounds' stock solutions were prepared in DMSO, and the final concentration in cell culture ranged from 25 to 150 μM. Hydrogen peroxide (H₂O₂) with a concentration of 1 mM was used for stimulation. 3-(4, 5-Dimethylthiazol-2-yl)-2, 5-diphenylt etrazolium bromide (MTT) assay was used to determine the cell viability (Chen et al., 2014a). Cells were plated at a density of 1 × 10⁵ cells/well in a 12-well plate overnight and incubated with tested compounds for 4 h and then treated with 1 mM hydrogen peroxide (H₂O₂) for 24 h. MTT working solution of 0.3 mg/mL final concentration (Promega, CA, USA) was added to the culture plates and after reacting for 2 h. MTT formazan products were solubilized with DMSO on a shaker platform. Finally, the absorbance was measured at 450 nm using Fluorescence spectrophotometer SpectraMax M5 (Molecular Devices, CA, USA). The results were expressed as percentage of the control.

For detection of cytotoxic drug-induced apoptosis, a FITC Annexin V apoptosis detection kit (BD Pharmingen™, BD Biosciences) was used. Briefly, 6 × 10⁴ cell/mL of HepG2 cells were seeded in six-well plates in complete medium and allowed to attach for 24 h. On the next day, the medium was discarded and new medium, with 30 μM of the tested compounds added. N-acetylcysteine (NAC) was used as positive control (1.0 mM). Cells were further incubated for 72 h, in 5% CO₂ at 37°C. After the incubation period, cells were trypsinized, washed in PBS and stained according to kit manufacturer's instructions. Stained cells were analyzed using BD Accuri C6 flow cytometer (BD Pharmingen™, BD Biosciences) and data was processed with BD Accuri C6 software. Each sample was assessed using a collection of 10,000 events. The mean values and standard deviations were calculated from three independent experiments.

Intracellular ROS production in RAW 264.7 treated with hydrogen peroxide (H₂O₂) was monitored by flow cytometry according to dichlorofluorescein assay (Kim et al., 2009). 2′, 7′-dichlorofluorescein-diacetate (DCF-DA) can be deacetylated in cells, where it can react quantitatively with intracellular radicals to convert into its fluorescent product. Hydrogen peroxide or low-molecular weight phydroeroxides produced by cells oxidize DCHF to the highly fluorescent compound, 20, 70-dichloro-fluorescein (DCF). Therefore, DCF-DA was used to evaluate the generation of ROS in oxidative stress. In the present study, cells were plated at a concentration of 4 × 10⁵ cells/well in 12-well plates overnight, and then treated with AM and DM for 24 h in the present of H₂O₂. Later, the compound-treated cells were washed twice with cold-PBS to remove the extracellular compounds, and the fluorescence intensities of the stained cells were determined under absorbances of 485 and 535 nm (excitation and emission wavelengths, respectively) using a fluorescence spectrophotometer (SpectraMax M5, Molecular Devices, CA, USA).

Hydrogen peroxide induced DNA oxidation was examined, following a protocol described elsewhere (Ballinger et al., 2000). Briefly, 100 μL of DNA reaction mixture was prepared by adding pre-determined concentrations of test sample (or same volume of distilled water as a control), 200 μM final concentration of FeSO₄, 1.0 mM final concentration of H₂O₂ and 50 μg/mL final concentration of genomic DNA in the same order. Then, the mixture was incubated at room temperature for 30 min and reaction was terminated by adding 10 mM final concentration of EDTA. 20.0 μL of reaction mixture containing about 1.0 μg of DNA were electrophoresed with 0.5% low-melting-point agarose (Promega Co., Madison, WI, USA). Gels were then stained with 1 mg/mL ethidium bromide and visualized by UV light using AlphaEase gel image analysis software (Alpha Innotech, CA, USA).

The cells were harvested, washed twice with PBS buffer, and lysised in RIPA buffer consisting of 20 mM Tris-HCl pH 7.5, 150 mM NaCl, 1 mM EDTA, 1 mM EGTA, 1% NP-40, 1% sodium deoxycholate, 2.5 mM sodium pyrophosphate, 1 mM β-glycerophosphate, 1 mM Na₃VO₄, and 1 μg/mL leupeptin (Cell Signaling Technology; Beverly, MA, USA). Eluted protein concentration was measured using the DC protein assay (Bio-Rad, Richmond, CA, USA) according to the manufacturer's instructions. For the western blot assay, equal amounts of extracted proteins were separated by SDS-polyacrylamide gel electrophoresis (PAGE) and transferred to nitrocellulose membranes by semi-dry electro-blotting (Bio-Rad). Proteins that were extracted from either whole cell lysate or nuclei were separated by SDS-PAGE, transferred to nitrocellulose membranes, and detected with an Nrf2 antibody. Membrane was washed three times for 15 min with washing solution of PBS containing 0.05% Tween-20. After washing, the blots were incubated with the secondary antibody for 2 h. Then, it was washed three times with PBS washing solution, and protein bands were detected using enhanced chemiluminesence (ECL) system and visualized by chemiluminescent light. Immunoblots for phosphorylation of ERKs, JNKs, and p38 kinase were carried out as described in the protocol of the manufacturer, using phosphorylated antibodies against phosphorylated sites of ERKs, JNK, and p38 kinase. Non-phospho specific antibodies against ERKs and p38 kinase proteins provided by the assay kits were used to normalize the phosphorylation assay with the same transferred membrane blot.

The data were expressed as the mean ± SD. The data were analyzed by one-way ANOVA or two-way ANOVA (Tukey's test) using the IBM SPSS Statistics 20.0 for Mac (IBM Corporation, Armonk, NY). The Kolmogorov–Smirnov and the Levene's tests were used to examine the normality and equality, respectively. The differences between the mean values were considered to be statistically significant when p ≤ 0.05.

Both AM and DM, from A. pilosa, tested in this study were preliminarily and systematically identified using RP-HPLC with UV detector by comparing with authentic compounds isolated previously (Huang et al., 2015). AM and DM were found as effective scavengers for DPPH and ABTS free radicals. At 200 μM concentration level, both of AM (Figure 1B) and DM (Figure 1E) showed that the ABTS radicals scavenging activity reached plateau of 80%. There was no significant variation of DPPH inhibitory effect between AM (58.2%, Figure 1C) and DM (44.8%, Figure 1F), at the same concentration. Scavenging of DPPH radicals allows the evaluation of hydrogen-donating potency of these two compounds. The SOD-like activity of AM and DM, at concentration of 200 μM, were 50.7 and 51.2% respectively. No significant change of antioxidant effect was observed when comparing concentration range of 200–250 μM for both of AM (Figure 1D) and DM (Figure 1G). Based on these findings, subsequent studies were conducted at the concentrations of AM and DM ranging from 25 to 200 μM.

The growth of HepG2 cells in the presence of various concentrations of AM and DM was examined by using MTT assay. As shown in Figure 2A, AM and DM treatment did not cause significant change in MTT based cell viability, compared to those of the untreated control cells. Typical microscopic images of AM and DM extracts on human tumor cells are shown in Figure 2B. Compared to the control, the size as well as the shape of all cells were not significantly changed after treatment with AM or DM. Similar results were found in Figure 2C, cells shrinking and falling off from the adherent state were not observed even at the highest concentrations tested.

It has been reported that the ROS can contribute to mitochondrial damage, which may facilitate further release of ROS into the cytoplasm (Simon et al., 2000). To investigate the effect of AM and DM (from 50 to 200 μM) on the generation of intracellular ROS, cells were pretreated with the samples and then incubated with 1 mM H₂O₂ (Figure 3A). As expected, pre-treatment of cells with AM and DM also blocked the increase of intracellular ROS generation. Both of them lead to reduction of ROS in H₂O₂-induced HepG2 cells: from 36 to 48% for AM (50–200 μM), and from 60 to 66% for DM, at the same concentration range. The level of intracellular ROS was monitored by flow cytometry using a peroxide-sensitive fluorescent probe, DCFDA (2′,7′-dichlorofluorescein diacetate). Fluorescence intensity was directly proportional to the level of dichlorofluorescein (DCF) formed intracellularly, which can be easily detected by flow cytometry. In the presence of intracellular ROS, DCF-DA was oxidized into fluorescent 2′, 7′-dichlorofluorescein (H₂DCF). In flow cytometry analysis, an increase of intracellular ROS level was noticed in H₂O₂-induced cells_,as demonstrated by the shift of histogram of fluorescence from left (Figure 3B) to the right panel (Figure 3C). To elucidate the role of ROS in H₂O₂ induced apoptosis, the antioxidant of N-acetylcysteine (NAC) was used, which has been shown previously to be a radical scavenger functioning as both redox buffer and reactive oxygen intermediate scavenger. Cells exposed to 1 mM NAC plus 1 mM H₂O₂ for 24 h did not exhibit the characteristic features of cell shrinking, rounding, and partial detachment as cells exposed to H₂O₂ alone (Figure 3D). However, when cells were pre-treated with NAC, the fluorescence were decreased to 3.0-fold, and the distance of peak movement from left to right was observed (Figure 3E) as compared with H₂O₂ stimulated cells. A same trend of in peak movement was found after the addition of 50 or 100 μM AM and DM (Figures 3E–H).

Hydrogen peroxide as a mediator of cytotoxicity could cause most of the oxidative damage in biological systems by the reaction between H₂O₂ and the presence of metal ions (Nathan et al., 1979). The comet assay was used to detect cellular DNA damage (Olive and Banáth, 2006), since DNA within the nucleoid from each cell typically appears as either an intact spherical mass (i.e., no DNA damage) or as a “comet” (i.e., DNA damage) upon staining with ethidium bromide, visualized by fluorescence microscopy. A basic assumption is that the overall span of tail region of the comets is a general indication for the extent of DNA single-strand breakage (Kelly et al., 2001). The photographs shown in Figure 4 are representative views of the slides for untreated and treated cells. The number and overall tail spans of the comets were found to be consistently similar in viewing different areas of each slide upon replicating the experiments three times and overall comet scores generated reflecting the relative extents of DNA damage. As shown in Figure 4A, cells treated with H₂O₂ were characterized by clearly discernible comets compared to control cells. Pre-incubated cells with AM concentration as low as 50 αM resulted in a slight reduction of the overall comet tail spans. Moreover, 100 and 200 αM concentration of AM had additional protective effects against DNA damage. In the case of DM, is also worth of expecting that DM would be able to inhibit DNA damage induced by H₂O₂ even at low concentration doses supplied.

In addition to detection of oxidative DNA damage, the extracted DNA was subjected to PCR. As shown in Figure 4B, H₂O₂ treatment could induce DNA damage. This damage can be reduced in the presence of AM and DM. The addition of DM at concentration of 200 μM, to the reaction mixture, induced the significant protection to the damage. Higher concentrations of AM were also checked for the dose dependent activity, however there was no difference between each other. Densitometric analysis (Figure 3C) revealed that both of AM and DM significantly decreased the H₂O₂-induced formation of DNA strand breaks. No differences could be observed between pre-incubation with the AM and DM. For AM, the protective effect was significantly increased by 57% when the dose increased from 0 to 100 μM. Higher concentrations of AM were also checked for the dose dependent activity but no higher protection at 200 μM concentration was observed. Similar trend of protective effect was found by adding DM into the reaction mixture of H₂O₂ induced HepG2 cells. These results related to the simultaneous incubation of the cells with AM and DM were clearly demonstrated by the comets representative mages for each condition.

Interestingly, AM and DM increased Nrf2 expression levels in a dose-dependent manner, with the maximal expression at 200 μM in HepG2 cells (Figures 5A,E).

The integrity of the cytosolic was confirmed by the analysis of the compartment-specific cytosolic α-tubulin. Besides the changes in nuclear levels of Nrf2, the nuclear content of the Nrf2 inhibitor Keap1 was also modified by AM and DM. The calculated nuclear Keap1 content ratio in AM and DM-treated HepG2 cells remained comparable to control values under all doses, whereas ratios of nuclear Nrf2/Keap1 were increased by 0.6 ± 0.1- and 3.0 ± 0.4-fold, respectively, at 25 and 200 μM AM (Figure 5B). Quantification of three independent western blots showed that in the presence of 25 and 200 μM DM, Nrf2/Keap1 increased 0.2 ± 0.1- and 3.6 ± 0.4-fold, respectively (Figure 5F). HO-1 expression was further measured to determine whether AM and DM exhibits potential antioxidant activity by up-regulating the intracellular phase II enzyme in HepG2 cells. Western blot analysis was performed to detect the expression of HO-1 induced by different concentrations of AM (Figure 5C) and DM (Figure 5G). Treatment of HepG2 also induced a concentration-dependent enhancement in HO-1 protein expression (Figures 5D,H). Quantification of three independent western blots showed that in the presence of 25, 50, 100, and 200 αM AM, HO-1 protein increased 0.12 ± 0.03-, 0.23 ± 0.01-, 0.38 ± 0.04-, and 0.5 ± 0.02-fold, respectively. Besides, DM also caused a significant increase of HO-1 protein expression (Figure 5H). These data indicated that both AM and DM are HO-1 inducers. Meanwhile, gene expression of Nrf2/Keap1analysed by qRT-PCR (Figures 5I,J) also confirmed that the observed effects are at transcriptional level.

To determine the possible role, if any, of MAPKs in H₂O₂-mediated HO-1 gene activation, the effect of AM and DM on MAPK activities were examined (Figure 6). Firstly, the effect of AM and DM on three MAPK cascades leading to activation of ERK, p38 and JNK on the survival pathway was analyzed. The activation of these pathways was analyzed with activation-specific antibodies that selectively recognized the active and phosphorylated forms of ERK, p38, and JNK. HepG2 cells treated with H₂O₂ alone, total JNK, ERK, and β-actin were not affected but phosphorylations of JNK and ERK were up-regulated by AM and DM treatment. In contrast, AM and DM decreased the phosphorylated levels of p38 in a dose-dependent activation in HepG2. Reduced phosphorylation of p38 (approximately two-fold) was detected after treatment with AM or DM.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.