Quantitative Determination of Stilbenoids and Dihydroisocoumarins in Shorea roxburghii and Evaluation of Their Hepatoprotective Activity

A simultaneous quantitative analytical method for 13 stilbenoids including (−)-hopeaphenol (1), (+)-isohopeaphenol (2), hemsleyanol D (3), (−)-ampelopsin H (4), vaticanols A (5), E (6), and G (7), (+)-α-viniferin (8), pauciflorol A (9), hopeafuran (10), (−)-balanocarpol (11), (−)-ampelopsin A (12), and trans-resveratrol 10-C-β-d-glucopyranoside (13), and two dihydroisocoumarins, phayomphenols A1 (14) and A2 (15) in the extract of Shorea roxburghii (dipterocarpaceae) was developed. According to the established protocol, distributions of these 15 polyphenols (1–15) in the bark and wood parts of S. roxburghii and a related plant Cotylelobium melanoxylon were evaluated. In addition, the principal polyphenols (1, 2, 8, 13–15) exhibited hepatoprotective effects against d-galactosamine (d-galN)/lipopolysaccharide (LPS)-induced liver injury in mice at a dose of 100 or 200 mg/kg, p.o. To characterize the mechanisms of action, the isolates were examined in in vitro studies assessing their effects on (i) d-GalN-induced cytotoxicity in primary cultured mouse hepatocytes; (ii) LPS-induced nitric oxide (NO) production in mouse peritoneal macrophages; and (iii) tumor necrosis factor-α (TNF-α)-induced cytotoxicity in L929 cells. The mechanisms of action of these polyphenols (1, 2, and 8) were suggested to be dependent on the inhibition of LPS-induced macrophage activation and reduction of sensitivity of hepatocytes to TNF-α. However, none of the isolates reduced the cytotoxicity caused by d-GalN.


Simultaneous Quantitative Analysis
As shown in Figure 2  In order to optimize the extraction condition, the quality of the extracts in association with the contents of the stilbenoids (1-13) and dihydroisocoumarins (14 and 15) were examined. The extraction efficacies were compared for three solvent systems (methanol, 50% aqueous methanol, and water) under two different conditions (reflux for 120 min or sonication for 30 min, each twice). As shown in Table 1, "reflux in methanol" afforded the highest total contents of these polyphenols (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15). Therefore, all of the analytical samples were prepared using the method "reflux in methanol for 120 min, twice". As shown in Table 2, analytical parameters such as linearity, limits of detection and quantitation, and precision of the developed method were evaluated. The calibration curves were linear in the range studied (2.5-50 µg/mL), showing the correlation coefficients (R 2 ) of more than 0.9994 for each analyte. Linear regression equations of their calibration curves for each analyte were described in Table 2. The detection and quantitation limits were estimated to be 0.05-0.26 and 0.14-0.80 ng, respectively, indicating sufficient sensitivity of this method. The relative standard deviation (RSD) values were 0.13%-1.54% for intra-day assays and 0.13%-1.58% for inter-day assays. Accuracy was determined in recovery experiments using the methanol under reflux extract from the bark of S. roxburghii. As shown in Table 3, overall recovery rate was observed in the range of 95.1%-104.9% with RSD values lower than 1.3%. According to the established protocol, contents of the stilbenoids (1-13) and dihydroisocoumarins (14 and 15) in both bark and wood of S. roxburghii, as well as a related plant classified the same Dipterocarpaceae family Cotylelobium melanoxylon, which have all been reported to possess the common oligostilbenoids, were evaluated [16] ( Figure S2). The assays proved to be reproducible, precise, and readily applicable to the quality evaluation of these extracts. As shown in Table 4, total polyphenol content in the bark of S. roxburghii (72.60 mg/g in dry material) was found to be three-fold higher than the wood part (21.20 mg/g). Among them, two resveratrol tetramers, (−)hopeaphenol (1, 13.31 mg/g in dry material) and (+)-isohopeaphenol (2, 10.21 mg/g), a resveratrol trimer, vaticanol E (6, 11.57 mg/g), and a dihydroisocoumarin, phayomphenol A1 (13, 13.81 mg/g) were present relatively in abundance in the bark of S. roxburghii. As for C. melanoxylon, both of the total oligostilbenoid contents in the bark (286.73 mg/g) and wood (197.50 mg/g) were higher than those of S. roxburghii, and their distributions were biased towards vaticanols A (5, 76.45 mg/g in the bark), E (6, 120.75 mg/g in the bark), and G (7, 63.81 mg/g in the bark; 181.69 mg/g in the wood), which is supported by our previous report [16]. On the basis of this evidence, these Dipterocarpaceaeous plants, S. roxburghii and C. melanoxylon, have been shown to be useful as abundant resources for obtaining the bioactive oligostilbenoids. In order to optimize the extraction condition, the quality of the extracts in association with the contents of the stilbenoids (1-13) and dihydroisocoumarins (14 and 15) were examined. The extraction efficacies were compared for three solvent systems (methanol, 50% aqueous methanol, and water) under two different conditions (reflux for 120 min or sonication for 30 min, each twice). As shown in Table 1, "reflux in methanol" afforded the highest total contents of these polyphenols (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15). Therefore, all of the analytical samples were prepared using the method "reflux in methanol for 120 min, twice". As shown in Table 2, analytical parameters such as linearity, limits of detection and quantitation, and precision of the developed method were evaluated. The calibration curves were linear in the range studied (2.5-50 µg/mL), showing the correlation coefficients (R 2 ) of more than 0.9994 for each analyte. Linear regression equations of their calibration curves for each analyte were described in Table 2. The detection and quantitation limits were estimated to be 0.05-0.26 and 0.14-0.80 ng, respectively, indicating sufficient sensitivity of this method. The relative standard deviation (RSD) values were 0.13%-1.54% for intra-day assays and 0.13%-1.58% for inter-day assays. Accuracy was determined in recovery experiments using the methanol under reflux extract from the bark of S. roxburghii. As shown in Table 3, overall recovery rate was observed in the range of 95.1%-104.9% with RSD values lower than 1.3%. According to the established protocol, contents of the stilbenoids (1-13) and dihydroisocoumarins (14 and 15) in both bark and wood of S. roxburghii, as well as a related plant classified the same Dipterocarpaceae family Cotylelobium melanoxylon, which have all been reported to possess the common oligostilbenoids, were evaluated [16] ( Figure S2). The assays proved to be reproducible, precise, and readily applicable to the quality evaluation of these extracts. As shown in Table 4, total polyphenol content in the bark of S. roxburghii (72.60 mg/g in dry material) was found to be three-fold higher than the wood part (21.20 mg/g). Among them, two resveratrol tetramers, (−)-hopeaphenol (1, 13.31 mg/g in dry material) and (+)-isohopeaphenol (2, 10.21 mg/g), a resveratrol trimer, vaticanol E (6, 11.57 mg/g), and a dihydroisocoumarin, phayomphenol A 1 (13,13.81 mg/g) were present relatively in abundance in the bark of S. roxburghii. As for C. melanoxylon, both of the total oligostilbenoid contents in the bark (286.73 mg/g) and wood (197.50 mg/g) were higher than those of S. roxburghii, and their distributions were biased towards vaticanols A (5, 76.45 mg/g in the bark), E (6, 120.75 mg/g in the bark), and G (7, 63.81 mg/g in the bark; 181.69 mg/g in the wood), which is supported by our previous report [16]. On the basis of this evidence, these Dipterocarpaceaeous plants, S. roxburghii and C. melanoxylon, have been shown to be useful as abundant resources for obtaining the bioactive oligostilbenoids. Extraction efficiency was tested using the bark of Shorea roxburghii (loss of drying 7.44%); a value (%) relative to the content obtained by methanol under reflux is given in parenthesis; and b less than the quantitation limit. Table 2. Linearities, detection and quantitation limits, and precisions for stilbenoids (1)(2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13) and dihydroisocoumarins (14 and 15) from the bark of Shorea roxburghii. x is the concentration of the analyte solution (µg/mL), and y is the peak area of the analyte; b values are the amount of the analyte injected on-column; and c precision of the analytical method were tested using the methanol under reflux extract of the bark of Shorea roxburghii (n = 5).

Effects on D-GalN-Induced Cytotoxicity in Primary Cultured Mouse Hepatocytes
To characterize the mechanisms responsible for the hepatoprotective activity, the inhibitory effect of 23 polyphenol constituents, including principal polyphenols (1-15) isolated from the bark of S. roxburghii [13,14], on D-GalN-induced cytotoxicity in primary cultured mouse hepatocytes were examined. As shown in Table 6, none of the isolates led to a reduction in the cytotoxicity at concentrations of up to 100 µM. These results are similar to what was seen with curcumin [22,23,25,27], a naturally occurring hepatoprotective product obtained from turmeric [40]. Thus, the principal polyphenols (1-15) did not affect D-GalN-induced cytotoxicity. In contrast, trans-resveratrol inhibited the cytotoxicity (IC 50 = 40.8 µM), which was equivalent to that of silybin (IC 50 = 38.8 µM) [26,27,30,32], a naturally occurring hepatoprotective product obtained from milk thistle [41,42]. This evidence led us to confirm that trans-resveratrol and its oligomers do not have similar effects on D-GalN-induced cytotoxicity in hepatocytes. Table 6. Inhibitory effects of constituents from the bark of Shorea roxburghii on D-GalN-induced cytotoxicity in mouse primary hepatocytes.

Chemicals and Reagents
Methanol, acetic acid, and distilled water for HPLC were purchased from Nacalai Tesque Inc., Kyoto, Japan. All other chemicals were reagent grade, and were purchased from Wako Pure Chemical Industries, Ltd., Tokyo, Japan or Nacalai Tesque Inc., Kyoto, Japan.

Plant Materials
The bark and wood parts of Shorea roxburghii and Cotylelobium melanoxylon were collected from Phatthalung Province, Thailand, in September 2006 or September 2007 as described previously [13,16].

Sample Solution Preparation
An accurately weighed pulverized sample powder (ca. 2 g, conversion with loss on drying) was extracted with 20 mL of three solvent systems (methanol, 50% aqueous methanol, or water) under two conditions (reflux for 120 min or sonication for 30 min, each twice). The extracts were combined and centrifuged at 3000 rpm for 5 min, then the supernatants were diluted to 100 mL with the extraction solvent. An aliquot (1 mL) of the extract solution was transferred into a 10 mL volumetric flask and 50% aqueous methanol was added to make up the volume. The solution was filtered through a syringe filter (0.45 µm), and an aliquot of 2 µL was subjected to the HPLC analysis. To calculate the extraction yields, the remaining extraction solution (90 mL) was evaporated in vacuo (Table 1).

HPLC Instruments and Conditions
All analytical experiments in this study were performed with an LC-20A series Prominence HPLC system (Shimadzu Co., Kyoto, Japan), which consists of a UV-VIS detector, a binary pump, a degasser, an autosampler, a thermostated column compartment, and a control module. The chromatographic separation was performed on an Inertsil ODS-3 column (3 µm particle size, 2.1 mm i.d. × 100 mm, GL Sciences Inc., Tokyo, Japan) operated at 40 • C with mobile phase A (acetonitrile) and B (H 2 O containing 0.1% formic acid). The gradient program was as follows: 0 min (A:B = 10:90, v/v) → 20 min (30:70, v/v) → 30-40 min (50:50, v/v, hold). The flow rate was 0.2 mL/min with UV detection at 284 nm and the injection volume was 2 µL.

Calibration and Validation
The standard curves were prepared with four concentration levels in the range of 0.5-50 µg/mL. Standard curves were made on each day of analysis. Linearity for each compound was plotted using linear regression of the peak area versus concentration. The coefficient of correlation (R 2 ) was used to judge the linearity. The detection and quantitation limits for each analyte were determined by the signal-to-noise (S/N) ratio for each analyte by analyzing a series of diluted standard solutions until the S/N ratios were about 3 and 10, respectively, based on a 2 µL injection. Precision and accuracy of the analytical method were tested using a methanol under reflux extract of the bark of S. roxburghii. The intra-and inter-day precisions were determined by estimating the corresponding responses five times on the same day and on five different days, respectively ( Table 2). The recovery rates were determined by adding analytes of two different concentrations (50 and 125 µg/mL) to the sample solution of the homogeneous extract (Table 3).

Animals
Male ddY mice were purchased from Kiwa Laboratory Animal Co., Ltd. (Wakayama, Japan). The animals were housed at a constant temperature of 23 ± 2 • C and fed a standard laboratory chow (MF, Oriental Yeast Co., Ltd., Tokyo, Japan). All experiments were performed with conscious mice unless otherwise noted. The experimental protocol was approved by Kindai University's Committee for the Care and Use of Laboratory Animals (KAPR-26-001, 1 April 2014).

Effects on D-GalN/LPS-Induced Liver Injury in Mice
Protective effects on D-GalN/LPS-induced liver injury in mice were determined according to the previously described protocol [27]. Trans-resveratrol, curcumin, and silybin were used as reference compounds.

Effects on Cytotoxicity Induced by D-GalN in Primary Cultured Mouse Hepatocytes
Hepatoprotective effects on D-GalN-induced cytotoxicity in primary cultured mouse hepatocytes were determined according to the protocol described previously [27]. Trans-resveratrol, curcumin, and silybin were used as reference compounds.

Effects on Production of NO in LPS-Induced Mouse Peritoneal Macrophages
Assays for NO production in TGC-induced mouse peritoneal macrophages were performed as described previously [27]. Trans-resveratrol, N G -Monomethyl-L-arginine (L-NMMA), and caffeic acid phenethyl ester (CAPE) were used as reference compounds.

Effects on Cytotoxicity Induced by TNF-α in L929 Cells
Inhibitory effects on TNF-α-induced cytotoxicity in L929 cells were determined according to the protocol described previously [27]. Trans-resveratrol and silybin were used as reference compounds.

Statistics
Values are expressed as means ± SEM. One-way analysis of variance (ANOVA) followed by Dunnett's test was used for statistical analysis. Probability (p) values less than 0.05 were considered significant.