Enhancement of neuroprotective activity of Sagunja-tang by fermentation with lactobacillus strains

Ginsenoside Rg₁ and glycyrrhizin were purchased from the Korea Food and Drug Administration and Tokyo Chemical Industry Co., Ltd. (Tokyo, Japan), respectively. Liquiritin, liquiritigenin, atractylenolide I, atractylenolide II, atractylenolide III, and pachymic acid were purchased from Faces Biochemical Co., Ltd. (Wuhan, China). The purities of these seven reference standards were > 98%, and their structures are shown in Fig. 1. Trifluoroacetic acid (TFA) and formic acid were purchased from Sigma-Aldrich Co. (St. Louis, MO, USA). Acetonitrile of high-performance liquid chromatography (HPLC) grade was obtained from J.T. Baker Inc. (Philipsburg, NJ, USA), and deionized water was prepared using an ultrapure water production apparatus (Millipore, Billerica, MA, USA).

All herbal constituents of SGT, including ginseng radix (KMAC112), atractylodis rhizome (KMAC059), poria sclerotium (KMAC061), and glycyrrhizae radix (KMAC003), were purchased from the Korea Medicinal Herbs Association (Yeongcheon, Korea). The origins of the samples were confirmed taxonomically by Prof. Ki Hwan Bae of the College of Pharmacy, Chungnam National University. All voucher specimens were deposited in the herbal bank of the Korean Medicine Application Center, Korea Institute of Oriental Medicine. Four kinds of medicinal herbs (2 kg; Table 1) were extracted into 20 L of water for 3 h using a COSMOS-660 extractor (Kyungseo Machine Co., Incheon, Korea). The extract was filtered through standard testing sieves (150 μm) and freeze-dried to yield SGT extract powder (240 g). To ferment SGT, the following eight bacterial strains were used to generate eight batches: Lactobacillus rhamnosus KFRI127, L. zeae KFR129, L. rhamnosus KFRI144, L. acidophilus KFRI150, L. fermentum KFRI162, L. plantarum KFRI166, L. acidophilus KFRI217, and L. helveticus KFRI341. The strains were obtained from the Korea Food Research Institute (Seongnam, Korea) and were cultured in de Man, Rogsa, and Sharpe broth at 37 °C for 24 h and reinoculated to each broth under the same conditions. The solutions of SGT extract were adjusted to pH 8.0 with 1 M NaOH and autoclaved for 15 min at 121 °C for use as the culture media in the bacterial fermentations. The solutions were inoculated with 5 mL of the inoculum (1% v/v, 2 × 10⁹ CFU/mL). The inoculated samples were incubated at 37 °C for 48 h. The FSGT was lyophilized at 4 °C after filtering through a 60-μm nylon net filter (Millipore, Billerica, Mass, USA).

Individual stock solutions of liquiritin (400 μg/mL), ginsenoside Rg₁ (500 μg/mL), liquiritigenin (200 μg/mL), glycyrrhizin (900 μg/mL), atractylenolide I (200 μg/mL), atractylenolide II (200 μg/mL), and atractylenolide III (200 μg/mL) dissolved in methanol were mixed and diluted serially to five concentrations for use in generating calibration curves. Quality control samples for method validation were diluted serially to three concentrations. The SGT and FSGTs for quantitative analysis were dissolved at a concentration of 20 mg/mL in methanol. The samples were filtered through a 0.2 μm membrane filter (Whatman International Ltd., Maidstone, UK) and stored at 4 °C before use.

The analysis was performed on an HPLC system (Hitachi High-Technologies Co., Tokyo, Japan), consisting of a pump (L-2130), an auto sampler (L-2200), a column oven (L-2350), and a diode array UV/VIS detector (L-2455). The system control and data analyses were executed by an EZchrom Elite software (version 3.3.1a) system. The separation was performed on a RS tech C₁₈ column (5 μm, 250 × 4.60 mm) with a flow rate of 1.0 mL/min and a column oven temperature of 40 °C. The mobile phase consisted of 0.1% trifluoroacetic acid in water (A) and acetonitrile (B). To improve the chromatographic separation capacity, the gradient elution system was performed as follows: 90–65% A, 0–25 min; 65–55% A, 25–40 min; 55–30% A, 40–55 min; 30% A, 55–65 min. The sample injection volume was 20 μL.

Method validation for selectivity, linearity, limit of detection (LOD), limit of quantification (LOQ), precision and accuracy, and recovery in the present study was executed according to the guideline of International Conference on Harmonization (ICH).

The standard calibration curve for the linearity assay was prepared with five different concentrations of diluted standard solutions and performed in triplicate independently. The calibration curves were constructed by plotting the value of the peak versus the concentration of each analyte. The lower LOD and lower LOQ for each analyte were determined on the basis of signal-to-noise ratios (S/N) of 3.3 and 10, respectively.

The precision of the analytical method was evaluated by intra- and inter-day tests. Standard solutions at three different concentrations (low, medium, and high) were analyzed. The intra-day test was performed by analyzing a mixed standard solution in five replicates during one day. The inter-day test was performed by analyzing five replicates of the same standard solution on each of three consecutive days. Precision was expressed as relative standard deviation (RSD, %), which is generally acceptable within 3%. The related equation was as follows: RSD (%) = [standard deviation (SD)/mean measured amount] x100. The recovery tests were performed to evaluate the accuracy of the method. The recoveries of analytes were determined by adding three different concentrations of each standard solution into the SGT solution (10 mg/mL) in triplicate. Recovery (%) was calculated according to the following equation: Recovery (%) = [found amount – original amount]/spiked amount x 100.

SH-SY5Y cells, a human neuroblastoma-derived cell line with neuron-like characteristics, were provided by Prof. Jaewon Lee, Pusan National University, Korea. The cells were differentiated into neuron-like phenotype by induction of retinoic acid and cultured with RPMI 1640 media (Lonza, Walkersville, MD, USA) with heat-inactivated 10% fetal bovine serum (HyClone Laboratories, Utah, USA), 2 mM glutamine, and 1% penicillin/streptomycin antibiotic mixture (Corning Incorporated, NY, USA) in a humidified 5% CO₂ incubator at 37 °C.

Cells (1 x 10⁴ cells/mL) were seeded into 96-well plates. After 24 h, the cells were pretreated with different concentrations of samples (250 and 500 μg/mL) for 6 h and co-treated with 50-μM of the toxicants H₂O₂ and etoposide for 24 h. After treatment, cell viability was analyzed using a Cell Counting Kit-8 (CCK-8) solution (Dojindo Laboratories, Kumamoto, Japan). Color intensity was measured at 450 nm using ELISA microplate reader.

ROS products were measured using a fluorogenic dye 2′, 7′-dichlorodihydrofluorescein. diacetate (H₂-DCFDA), which is oxidized by intracellular ROS. Cells (5x10³ cells/well) were seeded into black 96-well plates and pretreated with vehicle or samples for 6 h, treated with 50 μM H₂-DCFDA for 30 min, and then incubated with 50 μM H₂O₂ for 30 min. The cells were washed with phosphate-buffered saline, and fluorescent compounds were detected using a fluorescence microplate reader (SpectraMax i3; Molecular Devices, CA, USA) with excitation and emission wavelengths of 495 nm and 529 nm, respectively.

Cells were seeded into a confocal dish (coverglass-bottom dish). After pretreatment with vehicle or samples for 6 h, the cells were co-treated with 100 μM H₂O₂ for 1 h. The cells were further incubated with JC-1 (chloride salt; Biotium, Hayward, CA, USA) staining solution (5 μg/mL) at 37 °C for 15 min and rinsed with culture media. MMPs were estimated by measuring the fluorescence of free JC-1 monomers (green) to JC-1 aggregates in mitochondria (red) as observed through a Nikon ECLIPSE TE2000-U microscope (Nikon Instruments Inc., Tokyo, Japan), and the quantification of the red and green fluorescence intensity ratio in individual cells was performed using Nikon NIS-Elements microscope imaging software.

Simultaneous quantitative and qualitative analyses of seven compounds in SGT were performed using HPLC–DAD and LC/MS, respectively. To improve the chromatographic separation capacity, 0.1% TFA (v/v) in water (A) and acetonitrile (B) were used as mobile phases in a gradient elution system. Furthermore, analyte resolution was found to be better with acetonitrile than with methanol. The present chromatographic conditions were used to establish the specific HPLC retention times (t_R) and UV detection wavelengths for the seven standard compounds, which were used to identify the seven compounds in SGT and in one of the Lactobacillus-fermented SGT batches (SGT166). As shown in Fig. 2, the retention times of liquiritin, ginsenoside Rg₁, liquiritigenin, glycyrrhizin, atractylenolide III, atractylenolide II, and atractylenolide Ι in the standard mixture were 15.25 (220 nm), 22.20 (200 nm), 23.91 (220 nm), 36.66 (254 nm), 44.76 (220 nm), 52.68 (220 nm), and 57.15 (275 nm) minutes, respectively. Under the same conditions, the retention times (min) of the observed components were 15.28 (liquiritin), 22.12 (ginsenoside Rg₁), 23.95 (liquiritigenin), 36.59 (glycyrrhizin), 44.74 (atractylenolide III), 52.67 (atractylenolide II), and 57.14 (atractylenolide Ι) in SGT and 15.22 (liquiritin), 22.10 (ginsenoside Rg₁), 23.95 (liquiritigenin), 36.55 (glycyrrhizin), 44.71 (atractylenolide III), 52.64 (atractylenolide II), and 57.11 (atractylenolide Ι) in SGT166. Pachymic acid as a standard compound of poria sclerotium was not detected in our optimized chromatographic method; therefore, pachymic acid was identified by LC/MS analysis (Fig. 3). In positive-ion mode, molecular ions for each compound in the standard mixture were observed at m/z 419.133 [M + H]⁺ for liquiritin, m/z 257.080 [M + H]⁺ for liquiritigenin, m/z 823.481 [M + Na]⁺ for ginsenoside Rg1, m/z 823.411 [M + H]⁺ for glycchirizine, m/z 231.137 [[M + H]⁺ for atractylenolide I, m/z 233.153 [M + H]⁺ for atractylenolide II, m/z 249.148 [M + H]⁺ for atractylenolide III, and m/z 529.388 [M + H]⁺ for pachymic acid. Under the same conditions, molecular ions for each components were detected at m/z 419.132 [M + H]⁺ for liquiritin, m/z 257.080 [M + H]⁺ for liquiritigenin, m/z 823.479 [M + Na]⁺ for ginsenoside Rg1, m/z 823.409 [M + H]⁺ for glycchirizine, m/z 231.137 [[M + H]⁺ for atractylenolide I, m/z 233.153 [M + H]⁺ for atractylenolide II, m/z 249.147 [M + H]⁺ for atractylenolide III, and m/z 529.387 [M + H]⁺ for pachymic acid in SGT; m/z 419.132 [M + H]⁺ for liquiritin, m/z 257.080 [M + H]⁺ for liquiritigenin, m/z 823.479 [M + Na]⁺ for ginsenoside Rg1, m/z 823.409 [M + H]⁺ for glycchirizine, m/z 231.137 [[M + H]⁺ for atractylenolide I, m/z 233.153 [M + H]⁺ for atractylenolide II, m/z 249.148 [M + H]⁺ for atractylenolide III, and m/z 529.388 [M + H]⁺ for pachymic acid in SGT166 (Additional file 1: Table S1).

To establish the calibration curve, mixed standard solutions of five different concentrations were analyzed in triplicate. Calibration curves were generated by plotting chromatographic peak area as a function of analyte concentration. Linear regressions of these data were expressed as Y = Ax + B, where A is the slope of the calibration curve, B is the y-intercept of the calibration curve, x is the concentration of marker components, and Y is the peak area. Correlation coefficients (R²) were calculated to assess linearity. The calibration data of each standard compound showed good linearity (R² > 0.9999). The LOD and LOQ were determined at signal-to-noise (S/N) ratios of 3 and 10, respectively. The LOD was between 0.002 μg/mL and 0.023 μg/mL, and the LOQ was between 0.006 μg/mL^, and 0.069 μg/mL (Additional file 1: Table S2). The RSD (%) was < 3%, which indicated acceptable precision. The RSDs of the intra- and inter-day assays were between 0.02 and 1.80% and between 0.02 and 1.90%, respectively, with accuracy between 97.72 and 102.20% for the intra-day assay and between 98.11 and 103.06% for the inter-day assay. Accuracy was determined by a recovery test with an appropriate amount of SGT (10 mg/mL) that had been spiked with three different (low, medium, and high) quantities of authentic standards. The amounts of each of the seven compounds in the spiked SGT were calculated from the corresponding calibration curve. The results showed good accuracy with an overall recovery between 96.22 and 104.09% for the compounds concerned. These results demonstrated that the developed HPLC method was sufficiently accurate and sensitive for simultaneous quantitative evaluation of seven components from SGT (Additional file 1: Tables S3 and S4).

The developed HPLC-based analytical method was applied to the simultaneous quantitation of seven components, including liquiritin, ginsenoside Rg₁, liquiritigenin, glycyrrhizin, atractylenolide III, atractylenolide II, and atractylenolide Ι, in SGT and its FSGTs. Among them, the amounts of ginsenoside Rg₁ in SGT and the FSGTs were lower than the LOQ even though Rg₁was observed in HPLC chromatograms of SGT and FSGTs (Fig. 2). Therefore, it was recorded as not detected (nd). Except for ginsenoside Rg₁, the six components in SGT and the FSGTs were identified by comparing their retention times and UV absorbances with those of the standard compounds. The contents of liquiritigenin, glycyrrhizin, and atractylenolide I–III were increased, whereas the content of liquiritin was decreased in the FSGTs relative to that of the SGT. These results indicate that the detected chemical components could be used as biomarkers for studying SGT and FSGTs.

In this study, following fermentation by all eight kinds of Lactobacillus strains, the conversion between liquiritin and liquiritigenin, an aglycone form of liquiritin, was clearly confirmed: the amount of liquiritigenin increased between 1275.84 and 1399.69%, and the amount of liquiritin decreased by 97.83 to 98.90%. Furthermore, glycyrrhizin contents in SGT was markedly changed by fermentation: the amount of glycyrrhizin increased between 199.44 and 279.28% (Table 2). In this study, fermentation of SGT by L. plantarum 166 (SGT166) showed distinct component changes among the eight kinds of Lactobacillus strains. As presented in Table 3, the conversion rates of liquiritigenin (1399.69%), glycyrrhizin (279.28%), atractylenolide III (29.96%), atractylenolide II (5.98%), and atractylenolide I (36.67%) in SGT166 were higher than those in SGT (100%).

Human neuroblastoma SH-SY5Y cells are widely used as an in vitro model in neuroscience research, including studies of neurobiology, neuronal differentiation, and neuroprotective events [21]. To determine the neuroprotective effects of SGT and SGT166, SH-SY5Y cells were treated with SGT or SGT166 at 250 μg/mL and 500 μg/mL for 24 h, respectively. The treatment with SGT or SGT166 had no significant cytotoxicity on SH-SY5Y cells; therefore, these concentrations were used in further studies (data not shown).

To access the neuroprotective effects of SGT or SGT166, H₂O₂ and etoposide were separately used to treat SH-SY5Y cells as a toxicant to induce ROS generation. The cells were pretreated with SGT or SGT166 for 6 h and then separately exposed to 50 μM H₂O₂ and 50 μM etoposide for 24 h. CCK analysis showed that SGT166 significantly protected toxicant-induced SH-SY5Y cell loss, whereas SGT had a much lesser effect (Fig. 4). To further investigate the effect on endogenous ROS in SH-SY5Y cells, SGT and SGT166 were treated separately with 50 μM H₂O₂ or 50 μM etoposide. After 24 h, the levels of intracellular ROS were measured using H₂-DCFDA, a fluorescent dye oxidized to fluorescent DCF by ROS. Toxicant treatment caused a marked increase in intracellular ROS generation, and SGT166 pretreatment significantly reduced ROS production, whereas SGT showed a negligible blocking of ROS production (Fig. 5). For this reason, the protective effects of SGT and SGT166 against toxicant-induced disruption of MMP in SH-SY5Y cells were evaluated using JC-1 staining. As shown in Fig. 6, treatment with 50 μM H₂O₂ or 50 μM etoposide decreased red fluorescence and increased green fluorescence in the cells, which indicated MMP loss. Particularly, 250 μg/mL SGT166 protected against MMP disruption induced by all toxicants in SH-SY5Y cells, which showed approximately three-times significantly greater activity than that of SGT.