The effect of luteolin 7-glucoside, apigenin 7-glucoside and Succisa pratensis extracts on NF-κB activation and α-amylase activity in HepG2 cells*

The chemical composition of Succisa pratensis is not well known. The existing data indicate a substantial content of flavonoids, which include luteolin and apigenin 7-glucosides. The aim of this study was to elaborate the isolation protocol of these flavonoids from flowers and leaves of S. pratensis, to carry out their characterization, as well as evaluate the effect of S. pratensis extracts on activation of transcription factor NF-κB and α-amylase activity. The extraction protocol applied in this study allowed isolation and characterization of flavonoid fraction of S. pratensis. Their identity was confirmed by NMR spectra analysis, UV spectroscopy and electrospray ionization-tandem MS evaluation. Treatment of pancreatic α-amylase with S. pratensis extract inhibited this enzyme’s activity to an extent comparable to that of isolated luteolin and apigenin 7-glucosides. Incubation of HepG2 cells for 24 h with S. pratensis extracts or isolated flavonoids resulted in moderate reduction in NF-κB transcription factor activation evaluated in terms of translocation of its active subunits from cytosol into nucleus and subsequently diminished expression of the COX-2 gene. Expression of NF-κB was also reduced. The most significantly diminished NF-κB activation and expression, as well as COX-2 expression, was found to result from treatment with isolated flavonoids and ethyl acetate extract of S. pratensis leaves. These results indicate that S. pratensis flavonoids may modulate the metabolic and signaling pathways whose deregulation is related to pathogenesis of liver cancer. Further studies are required to confirm these observations and assess the chemopreventive and/or therapeutic potential of the S. pratensis herb.


INTRODUCTION
Flavonoids are commonly found in plants and they are widely used in medicine, either as isolated compounds or in the form of extracts from plants in which they are abundantly present. Several studies had shown that flavonoids, such as apigenin and luteolin, have antioxidant, anticancer, and anti-inflammatory activities (Choi et al., 2007;Shukla & Gupta, 2010;Bouzaiene et al., 2016;Xue et al., 2017;Yan et al., 2017). Moreover, in vitro studies demonstrated that anti-inflammatory properties of luteolin may be mediated by inhibition of the nuclear factor-κB (NF-κB). This transcription factor is a regulator of cell survival, immunity and inflammation. NF-κB p50-p65 subunits represent the major active complex in most of the cells. Normally, NF-κB is sequestered in the cytosol by its inhibitor, the IκB protein.
Inhibition of α-amylase limits the availability of glucose, which might affect the cancer cell survival (Palorini et al., 2016).
The aim of this study was isolation and identification of flavonoids from S. pratensis flowers and leaves, and comparison of the effect of isolated luteolin and apigenin 7-glucosides with different plant extracts containing the whole flavonoids' fraction on the α-amylase activity in a cell-free system. Moreover, the influence of these preparations on activation and expression of NF-κB in human hepatocellular carcinoma cells (HepG2 cells) was evaluated.
Products, Poznan University of Medical Sciences, Mazowiecka 33, 60-623 Poznan. The leaves were collected in July 2016 and dried. The voucher number of the plant was 1104.
Preparation of extracts. The dried (10 g) leaves or flowers of S. pratensis were extracted with 200 mL of methanol and ethyl acetate for 30 min using an ultrasonic bath. The extraction was carried out at 50 o C, 1000 W ultrasonic power, frequency of 37 kHz (Elma S180 H). The methanolic and ethyl acetate extracts were concentrated to dry mass and used to prepare concentration suitable for the α-amylase inhibitory test and HepG2 cells' exposure.
Isolation and identification of flavonoids. The flavonoids were isolated from a methanol-water (1:1) S. pratensis flower extract using column chromatography.
The identification of the compounds was performed on the basis of UV, ESI-MS, 1 H and 13 C NMR analyses. The NMR spectra ( 1 H, 13 C NMR) were recorded using a Bruker NMR Avance II 400 MHz spectrometer, CD 3 OD with TMS as an internal standard. The ESI -MS mass spectra were measured on a Waters/Micromass (Manchester, UK) ZQ mass spectrometer. The obtained data were comparable to the published values (Giang & Son, 2004;Gohari et al., 2011;Peng et al., 2016). The UV spectra were recorded on a UV/VIS Perkin Elmer Lambda 35 spectrophotometer in MeOH, and also after addition of specific reagents (NaOAc/H 3 BO 3 , AlCl 3 , AlCl 3 /HCl, NaOMe, NaOAc), according to Mabry and co-workers (Mabry et al., 1970).
For the biological assay, the methanolic or ethyl acetate extracts from the flowers (SK1, SK2) and leaves (SL1, SL2) were applied.
The starch solution (1%) used as a substrate was prepared by boiling and stirring 0.5 g of starch in 50 mL of sodium phosphate buffer for 5 min. The pancreatic α-amylase solution was prepared by mixing 0.01 g of the enzyme in 10 mL of the sodium phosphate buffer. The color reagent was a solution containing 3,5-dinitrosalicylic acid (0.1 g), sodium potassium tartrate (2.99 g), sodium hydroxide (0.16 g) and the phosphate buffer (completed to 10 mL).
Twenty-five microliters of each plant extract and 25 µL of the enzyme were mixed in a 96-well plate and incubated for 10 min at 25 o C. Then, 25 µL of the starch solution were added and the mixture was incubated un-der the same conditions. The reaction was stopped by adding 50 µL of a dinitrosalicylic color reagent. The mixture was incubated for 20 min at 85 o C (Incubator CLN32STD POL-EKO Aparature, Poland). After cooling this mixture to room temperature, the absorbance was measured at 540 nm (Spectrophotometer Thermo Fisher SCIENTIFIC Multiskan Go, Finland). In the control, well dimethyl sulfoxide (DMSO, St. Louis, MO, USA) replaced the plant extract.
The blank sample in which the enzyme was replaced with the buffer solution was used to correct the absorption of the mixture. An acarbose solution at concentrations of 0.25 µg/mL to 1.0 µg/mL was used as a positive control. The inhibition percentage of α-amylase was assessed by the formula: Percent inhibition of the enzyme activity at the concentration range of 0.25-15 µg/mL (isolated flavoinods) and 40-140 µg/mL (extracts) were calculated and IC 50 values were estimated by linear regression.
Cell culture and treatment. HepG2 (ATCC HB 8065) was purchased from American Type Culture Collection (Rockville, MD, USA). Liver cancer cells were maintained in Dulbecco's Modified Eagle's Medium (St. Louis, MO, USA) containing 10% fetal bovine serum (St. Louis, MO, USA) and antibiotics solution (St. Louis, MO, USA). The cells were grown in a humidified incubator at 37°C in the atmosphere of 5% CO 2 .
Viability assay. The effect of studied compounds on cell viability was assessed by the MTT assay, according to standard protocols described previously (Krajka-Kuźniak et al., 2013). Briefly, 10 4 HepG2 cells were seeded per well in a 96-well plate. After 24 hours of preincubation in complete medium, flavonoids or extracts were added to the culture medium at various concentrations, and cells were incubated for subsequent 24 hours. DMSO concentration did not exceed 0.1%. After 24 hours, cells were washed twice with warm PBS buffer and fresh medium containing the MTT salt (0.5 mg/mL) was added. After 4 hours of incubation, formazan crystals were dissolved in acidic isopropanol and absorbance was measured at 570 nm and 690 nm. All of the experiments were repeated three times.
Nuclear and cytosolic preparation. The cytosol and nuclear extracts from HepG2 were prepared using the Nuclear/Cytosol Fractionation Kit (BioVision Research, CA USA).
RNA isolation. The extraction of total RNA from cells was performed using GeneMatrix Universal RNA Purification Kit (EurX, Gdańsk, Poland) according to the manufacturer's instructions.
Quantitative PCR. Total RNA was subjected to reverse transcription using the RevertAid Kit (Fermentas, Burlington, Canada), followed by quantitative real-time PCR. For real-time analyses, the Max-ima SYBR Green Kit (Fermentas, Burlington, Canada) and a BioRad Chromo4 thermal cycler were used. The protocol started with 5 min enzyme activation at 95°C, followed by 40 cycles of 95°C for 15 s, 56°C for 20 s and 72°C for 40 s, and final elongation at 72°C for 5 min. Melting curve analysis was used for product specificity verification. Estimation of expression of TBP (TATA box binding protein) and PBGD (porphobilinogen deaminase) was used for data normalization. Sequences of primers, obtained from oligo. pl (Warsaw, Poland), that were used for the analysis of NF-ĸB p50, NF-ĸB p65, COX-2, TBP and PBGD are listed in Table 1.
Western blot analysis. Nuclear extracts (NF-ĸB p50, NF-ĸB p65) or cytosolic proteins (NF-ĸBp 50, NF-ĸB p65, COX-2) (100 µg) were separated on 10% SDS-PAGE slab gels and proteins were transferred to nitrocellulose membranes. After blocking with 10% skimmed milk, proteins were probed with rabbit polyclonal NF-ĸB p50, rabbit polyclonal NF-ĸB p65, goat polyclonal COX-2, rabbit polyclonal β-actin, rabbit polyclonal lamin antibodies (Santa Cruz, CA, USA). Either β-actin or lamin was used as an internal control. The alkaline phosphatase-labeled anti-goat IgG, or anti-rabbit IgG were used as the secondary antibodies in the staining reaction. Bands were visualized with BioRad AP Conjugate Substrate Kit NBT/BCIP. The amount of immunoreactive product in each lane was determined using the Quantity One software (BioRad Laboratories, Hercules, CA, USA). Values were calculated as relative absorbance units (RQ) per mg protein.
Statistical analysis. The data obtained were expressed as a mean of three replicates and the significant difference was considered at p<0.05. Biological activity was analyzed using a simple linear regression, and the coefficient of determination (R 2 ) was calculated and the statistical analysis was performed using the STATISTI-CA10 software followed by the Student's t-test.

Chemical analysis
The applied protocol described above allowed for isolation of luteolin 7-glucoside (1) and apigenin 7-glucoside (2). The chemical structure of the compounds was determined on the basis of 1 H 13 C NMR, ESI-MS and UV-spectroscopy analysis (Table 2) and by comparison with the data from the literature (Giang et al., 2004;Mabry et al., 1970;Gohari et al., 2011, Asgharia et al., 2015Peng et al., 2016). For comparison, in further studies, extract containing the flavonoid fraction was applied.

7-glucoside (2) and methanolic and ethyl acetate extracts from S. pratensis leaves and flowers on the HepG2 cells' viability
The MTT test was used to evaluate the effect of luteolin 7-glucoside (1), apigenin 7-glucoside (2) and S. pratensis extracts on cell viability. Within the concentration range of 0.1-50 µg/mL the tested compounds and extracts reduced the viability of the HepG2 cells in a dose-dependent manner (Fig. 1). Luteolin and apigenin 7-glucosides were more cytotoxic than the extracts.
On the basis of the above results of the MTT test, in further studies, the tested flavonoids and S. pratensis extracts were used at the concentrations of 2.5 µg/mL, 5 µg/mL and 10 µg/mL.

The effect of luteolin 7-glucoside (1), apigenin 7-glucoside (2) and S. pratensis extracts on NF-κB activation, and NF-κB and COX-2 expression
Activation of NF-κB was evaluated in terms of translocation of its active subunits from cytosol into the nucleus. As is shown in Fig. 2, luteolin and apigenin 7-glucosides decreased the nuclear level of p50 and p65 subunits by about 21-23%, at concentration of 10 µg/ mL. Ethyl acetate extract from leaves (SL2), had significantly diminished the nuclear level of NF-κB p50 and the cytosolic level of COX-2, by about 29% and 21%, respectively.
The effect of isolated luteolin, apigenin 7-glucosides and S. pratensis extracts on NF-κB and COX-2 mRNA levels in HepG2 cells is shown in Table 4. At the concentration of 10 µg/mL, both glucosides and all studied extracts reduced the mRNA of p65 subunit level by about 23-36% in comparison to the result obtained in the control group. A similar effect was observed for expression of NF-κB p50. Moreover, SK1 and SK2 diminished the level of NF-κB p50 mRNA at a concentration of 5 µg/mL. At the same concentration, SL1 and apigenin 7-glucoside had decreased the transcript level of NF-κB p65. Expression of COX-2 was decreased after incubation with both glucosides and extracts, but the difference was not statistically significant. Table 3. α-Amylase inhibitory activities and IC 50 values of luteolin 7-glucoside (1), apigenin 7-glucoside (2) and S. pratensis extracts.

DISCUSSION
Chemical composition of the S. pratensis herb was not well described so far. Our earlier study showed a substantial amount of polyphenolic compounds in the leaves' extracts from this plant (Witkowska-Banaszczak & Długaszewska, 2017). Thus, the aim of this study was further characterization of these components and evaluation of their biological activity. Two flavonoid glucosides, luteolin and apigenin were isolated. Their effect on the α-amylase activity and NF-κB activation was compared with that exerted by the extracts containing the whole flavonoid fraction obtained from flowers or plant leaves.
The results of our study confirmed inhibition of the α-amylase activity by luteolin and apigenin 7-glucosides (Kim et al., 2000;Funke & Melzing, 2006;Li et al., 2018). Moreover, luteolin 7-glucoside isolated from S. pratensis showed a stronger inhibitory effect on α-amylase activity than apigenin 7-glucoside. Since the biological activity of flavonoids depends on the number of hydroxyl residues in the core molecule, it may be concluded that the presence of an additional OH group in the B ring of luteolin 7-glucoside makes this compound a more potent inhibitor of α-amylase. These studies have confirmed earlier reports about the influence of hydroxyl substituents and sugar molecules in the flavonoids on inhibition of α-amylase (Tadera et al., 2006;Sales et al., 2012;Asgharia et al., 2015).
Comparison of the effect of isolated luteolin and apigenin 7-glucosides with herb extracts indicated a stronger inhibitory potential of S. pratensis leaves' extract. A high content of flavonoids in this preparation (1.18%) in comparison with the flowers extract (0.23%) may be responsible for this effect (Witkowska-Banaszczak & Długaszewska, 2017).
The results of this study also indicated that the methanolic extracts are more efficient than the ethyl acetate in α-amylase activity inhibition. Thus, such preparation should be recommended for further studies. Additional research into the chemical profile of the extracts will probably allow determining the compounds or groups of compounds affecting amylase inhibition.
α-Amylases catalyze the specific cleavage of α-1,4 glycosidic bonds in polysaccharides, such as starch and glycogen. Human α-amylase is one of the major secretory products of the pancreas (P-amylase) and salivary glands (S-amylases). Human S-and P-amylases are encoded by AMY-1 and AMY-2A genes, respectively. The AMY-2B gene encodes α -amylases with a very similar amino acid sequence to the S-and P-amylases (Yokouchi et al., 1990), which are expressed in the lung carcinoid tissue (Doi et al., 1991;Tomita et al., 1989). It was shown that human liver amylase is encoded by the same gene. Thus, this α-amylase isoform may be related to hepatocellular carcinoma (HCC) (Koyama et al., 2001). Chronic inflammation plays an important role in pathogenesis of HCC.
The key element of this process is activation of NF-κB (Hwang et al., 2011).
Thus, in this study we also evaluated the effect of luteolin and apigenin 7-glucosides, and methanolic and ethyl acetate extracts from S. pratensis leaves and flowers on the NF-κB transcription factor activation and expression. Luteolin and apigenin 7-glucosides decreased expression of NF-κB and COX-2. The expression of the latter is controlled by NF-κB. A similar effect was observed after treatment of HepG2 cells with extracts from S. pratensis leaves and flowers. However, activation of NF-κB evaluated in terms of translocation of its active subunits from the cytosolic to the nuclear fraction was significantly affected only by isolated flavonoids and ethyl acetate extracts from S. pratensis leaves. Thus, the effect of S. pratensis preparation on NF-κB signaling was moderate.
Data are means ± S.E.M. from three separate experiments. The values were calculated as mRNA level in comparison to control cells (expression equals 1). The asterisk (*) denotes a statistically significant difference from the control group, p<0.05. sults confirmed the observation made by other authors on luteolin potential as an anti-inflammatory agent acting through downregulation of NF-κB and subsequently downregulation of COX-2 expression. In this regard, Xue and others (Xue et al., 2017) showed that luteolin and apigenin downregulate COX-2 expression in HepG2 cells. Similar observations in the case of luteolin appeared in the paper of Hwang and others (Hwang et al., 2011). They showed that luteolin inhibits NF-κB signaling pathways in HepG2 cells. Additionally, other studies using pancreatic carcinoma cells (PANC-1, CoLo-357 and BxPC-3 cells) confirmed the inhibitory effect of luteolin on NF-κB (Cai et al., 2012).
Overall, the results of this study indicate that S. pratensis flavonoids may modulate the metabolic and signaling pathways whose deregulation is related to pathogenesis of liver cancer. Further studies are required to confirm these observations and assess the chemopreventive and/ or therapeutic potential of the S. pratensis herb.

Conflicts of Interest
Authors declare that they have no conflict of interest to disclose.