UHPLC-MS-based metabolomics and chemoinformatics study reveals the neuroprotective effect and chemical characteristic in Parkinson’s disease mice after oral administration of Wen-Shen-Yang-Gan decoction

Chemicals and reagents

Acetonitrile and methanol (HPLC grade, >=99.9%) were supplied by Merck ltd. Formic acid was obtained from Sigma-Aldrich. Ultra pure water was purified using the Milli-Q system (Bedford, USA). The reference compounds of rhynchophylline (lot number: wkq18051803), verbascoside (lot number: 120524), linderane (lot number: wkq18031303), geniposidic acid (lot number: wkq18040312), coclaurine (lot number: wkq18036198) were supplied by Sichuan Weikeqi- Biotech Co., Ltd., nootkatone were obtained from Chengdu Chroma- Biotechnology Co., Ltd., rutin (lot number: 100080-201811), echinacoside (lot number: 111670-201907), quercetin (lot number:081-9003), and paeoniflorin (lot number:110736-201943) were obtained from National Institutes for Food and Drug Control.

Plant materials

Cistanches Herba, Alpiniae Oxyphyllae Fructus, Dioscoreae Rhizoma, Paeoniae Radix Alba, Linderae Radix and Uncariae Ramulus Cum Uncis were obtained from Zhengzhou Ruilong pharmaceutical CO., Ltd. and authenticated by Professor Shihui Qian (Jiansu Province Academy of Traditional Chinese Medicine, Nanjing, China).

Water extract of WSYGD preparation

The weighed powders of RCR (30g), BS (30g), WY (20g), YZ (30g), SY (20g), GT (20g) were mixed and immersed in 1500 mL of deionized water for 30 min then decocted using boiling for 1 h. The procedure was repeated twice. The extract was decanted and evaporated under reduced pressure. The mixed decoction was concentrated to high dosage (4g/mL), medium dosage (2g/mL), low dosage (1g/mL). The total ion chromatogram of WSYGD extract of medium dosage were shown in Supplementary Figure 1.

Animal treatment and sample preparation

Animals and treatments

Male C57BL/6J mice (ten months old, weighing 30-35g) were provided by Shanghai SLAC Laboratory Animal Co., Ltd (License No. SCXK (Hu), 2007-0005). Animals were housed in a controlled condition (22±2° C, 55±5% humidity) with a 12h circadian rhythm and had free access to food and water for two weeks to acclimatize them to the environment. Then, they were divided into control group, model group, Wen-Shen-Yang-Gan decoction groups (WSYGD-H, WSYGD-M and WSYGD-L) and drug group (Sinemet, carbidopa/levodopa, 25/100mg. 50mg/kg) randomly, 10 mice in each group. The mice in the constituent research divided into control group and model group, 6 mice in each group. All protocols and care for the mice were performed in accordance with the Guidelines for the Care and Use of Laboratory Animals and approved by the Animal Ethics Committee of Nanjing University of Chinese Medicine.

Efficacy of WSYGD to Parkinson’s disease

All of the mice except for control group were intragastrically administered with rotenone (30 mg/kg, lot number: R8875, sigma, USA). For the control group, an equal volume of normal saline was treated once daily for four weeks. In the 5th to 8th week, the administrated groups mice were received WSYGD solution once a day with different dose (4g/mL, 2g/mL, 1g/mL) and the positive group were received Sinemet solution (0.1ml/10g, MSD pharmaceutical Co. Ltd.). The mice of other two groups were given an equal volume of normal saline. The Behavioral of mice were evaluated by pole test and rotarod test at 24 h after experiment. All the animals were sacrificed 24h after the behavior test. The concrete drug regimen shown in Figure 1.

Figure 1. Schematic representation of experiment.

Constituents analysis of WSYGD

The male mice of C57BL/6J were divided into Control group and control-WSYGD group. The control group intragastrically administrated with normal saline each day. The mice of control-WSYGD group received WSYGD extract (dissolved in normal saline) at 14.70 g/kg. All animals were anesthetized with 10% urethane. The blood was obtained from the fosse orbital vein of animals at 0.25, 0.5, 1, 2, 3, 4, 6, 8 h after WSYGD treatment.

The blood samples (with heparin and without anticoagulant) were centrifuged (12000 rpm, 10 min, 4° C) to acquire plasma. The sample of various points from one animal were mixed and stored at -80° C until analysis.

Pole test

The pole test was performed on all mice to assess movement retardation and coordination. A steel pipe (50 cm long and 1 cm in diameter) which is tightly wrapped by white antiskid belt and has a spherical protrusion (2 cm in diameter) on the top as attachment points of mice. Place the mouse’s head upward on the spherical protrusion, record the time from its placement on the pole top to the point of turning its head downwards and the total time from its placement on the pole top to its climbing to the pole bottom with its rear limbs touching the ground. Three trials were performed with each mouse with an interval of two minutes and median data were taken across the trials.

Rotarod test

The rotarod test was performed with the help of ZH-600 mouse rotarod fatigue meter (Anhui Zhenghua biologic apparatus facilities Co., Ltd). The mice were placed on a rotation drum and measurement the time of standing on the drum as the speed accelerated from five to forty rpm in five-minute. Place the mice on a shifting drum, increase the rotation speed from 5 rpm to 40 rpm in 5 minutes, and record the time of standing on the drum. Final trials were performed three times for each animal in every group, and the values were averaged for every animal. The experimental environment, light and temperature for example, and timing were compatible in all the tests.

Immunofluorescence staining

After perfusion, the tissue samples of mice were rapidly removed and fixed with 4% paraformaldehyde overnight. Sucrose solution was utilized to store isolated brains to dehydrate gradiently and embedded in OCT for frozen tissue sections. The substantia nigra brain tissue sections in line with the brain atlas were prepared. Goat anti-mouse IgG (H&L, ab150113, Abcam, Cambridge, UK) was appended and hatched for one hour after being flushed with PBS. The sample was washed by PBS and sealed with anti- fluorescence. The flake was observed by Olympus BX63 fluorescent microscope (Olympus Corporation, Tokyo, Japan).

Western blotting (WB) analysis

Mid-brains of mice were rinsed with PBS and lysed with radioimmunoprecipitation assay buffer and then centrifuged (12000 rpm, 10 minutes, 4° C) Then, Lysates were segregated by 12% SDS-PAGE and transferred to polyvinylidene fluoride membrane (Millipore, Danvers, MA, USA), blocked in 5% BSA in PBS and pre-incubated with primary antibodies as the α-syn (ab59264, Abcam, Cambridge, UK), TH (SAB4200697, Sigma-Aldrich, USA) and GAPDH (60004-1-Ig, ProteinTech Group, Inc., USA). The GAPDH antibody was utilized as a loading control. On the second day, the PVDF membrane was incubated after flushed by TBST solution for one hour at 4° C. The membranes were exposed to anti-rabbit IgG antibodies (HRP, ZB-2301, Beijing Zhongshan Jinqiao Biotechnology Co., Ltd., China) for one hour at 4° C and scanned with a chemiluminescence image analysis system (Tanon-5200, Tanon Science and Technology Co., Ltd, China).

Quantization of cytokine levels by enzyme-linked immunosorbent assay (ELISA)

The blood was collected and clarified, then centrifugated at 12000 rpm for 10 min at 4° C, and the supernatant was utilized for analysis. The levels of supernatant cytokines (all of obtained from Shanghai ZCi Biotechnology Co., Ltd., China), such as IL-10, TNF-α, and IL-1β, were determined by ELISA kits. Standards and samples were incubated by diluent in a plate coated with primary antibody for one hour at room temperature. Plates were then washed three times and 100 μL of substrate solution (tetramethylbenzidine) was added to each well for chromogenic reaction for 15 minutes in complete darkness. Diluted standards and samples were appended to each well and incubation with biotin-labelled secondary antibodies. Then, the plate was incubated by horseradish peroxidase-conjugated streptavidin (SA-HRP). The reaction halted by adding 50 μL of stop solution into each well. The standard curve was constructed using the absorbance of each standard (450nm, vertical axis) against its concentration (horizontal axis). The concentration of inflammatory cytokines was calculated using the relative absorbance of the samples and standards.

Reference and plant sample preparation

A standard mixture containing rhynchophylline, verbascoside, linderane, geniposidic acid, coclaurine, nootkatone, rutin, echinacoside, quercetin and paeoniflorin was prepared in 100% methanol.

An aliquot of 10 milliliter of the WSYGD solution was decanted and evaporated under reduced pressure. The residue was dissolved in 10 milliliter of methanol and vortex-mixed for 5 min. The sample was centrifugated with 12000 rpm for 10 min at room temperature and then filtered through a 0.22 mm membrane.

Instrument and analytical conditions

Metabolite analysis was achieved using a Waters ACQUITY UPLCTM I-Class chromatography system (Waters Corp, Milford, MA, USA) equipped with ion mobility mass spectrometry detector. The separation was performed on a Thermo Scientific™ Syncronis™ C18 (100 mm×2.1 mm, 1.7 mum particle size, Thermo Fisher Scientific) and the injection volume was 5 μL at 25° C. A gradient elution profile was employed using mobile phase A (0.1% formic acid in acetonitrile) and mobile phase B (0.1% formic acid in water). The following gradient was utilized: 0 min, 5% A; 0.5 min, 5% A; 3 min, 40% A; 10.5 min, 95% A. The flow rate was set to 0.40 mL/min.

The mass spectrometry was carried out using a SYNAPT G2-Si Q TOF (Waters Corp, Wilmslow, UK) connected to the ACQUITY UPLC I-Class System via an ESI interface. The mass spectrometry analysis was performed in the positive mode and negative mode, with the following parameters: Capillary 1.5 kV, sample voltage 30 V, cone voltages 5 V, source temperature 120° C, desolvation temperature 500° C. desolvation gas flow rate 800 L/h, cone gas flow rate 50 L/h. The MS was set with 0.1 second of scan time to acquire in sensitivity mode. The accurate mass precursor and fragment ion data was obtained using the selected mass range of m/z 100-1200.

Data processing

The mechanism of WSYGD in PD

All data were expressed as mean ±SD and statistical significance by using one-way ANOVA followed by post-hoc analysis using the GraphPad Prism software, (version 8.0, San Diego, CA, USA). Ultimately, the data that P < 0.05 was deemed to be significant.

Composition identification

The resultant data matrices from the analytical technique were introduced into the Masslynx™ NT 4.1 software (Waters Corporation) for multi-variate statistical analysis including PCA and OPLS-DA. Pareto scaling was performed through Principal component analysis to catch sight of the separation between groups, and OPLS-DA to recognition important characteristics using Masslynx v.4.1 (Waters, Milford, MA, USA) [16]. We utilize the database of HMDB, Massbank and KEGG to screen and discriminate the potential markers.

Effects of WSYGD on behavioral test of rotenone-intoxicated mice

In the behavioral test of pole, the turning around time of mice head in the rotenone-intoxicated group was obviously longer than control group mice, while the turning around time of mice head in the WSYGD-H, WSYGD-M and WSYGD-L groups were significantly shortened, indicating statistically significant in comparison with model. With regard to the total time of climbing, the mice in rotenone-intoxicated group obviously longer than those in control group, and the time spent by WSYGD-H and WSYGD-M was significantly shortened, indicating statistically significant in comparison with model. In behavioral test of rotarod, the time of rod-standing time of rotenone-intoxicated group mice were obviously shorter than that of control, moreover, the WSYGD-H, WSYGD-M and WSYGD-L group mice were obviously longer than rotenone-intoxicated group, the results shown in Figure 2.

Figure 2. Effects of WSYGD on behavioral test of PD chronic model mice. In the pole test, (A) refers to the turning time and (B) refers to the total time; (C) refers to the retention time in the rotarod test. Control: blank group; Model: rotenone-intoxicated group; WSYGD-H: high dosage group; WSYGD-M: medium dosage group; WSYGD-L: Low dosage group; Sinemet: positive control. ***p<0.001 vs. Control; #p<0.05,##p<0.01, ###p<0.001 vs. Model.

Effects of WSYGD on midbrain dopamine (DA) neurons and pathological changes of α-syn in rotenone-intoxicated mice

The TH-positive cells number could be utilized to reflect the dopamine neurons numbers. The WB results shown that the TH expression in rotenone-intoxicated group was evidently down-regulated than that in the control group, and that in WSYGD-H, WSYGD-M and WSYGD-L groups were significantly increased in comparison with that in model group (Figure 3B). The same results were also indicated among the immunofluorescence test (Figure 3A). Simultaneously, the α-syn expression among the middle brain of mice in the rotenone-intoxicated group was evidently upward-regulated than that in the control group, and WSYGD-H, WSYGD-M and WSYGD-L could obviously decrease the expression of α-syn (Figure 3B).

Figure 3. Immunofluorescence staining of TH protein in the substantia nigra of mice (A); Western blotting analysis (B) and quantification (C) of relative α-syn and TH protein abundance. Control: blank group; Model: rotenone-intoxicated group; WSYGD-H: high dosage group; WSYGD-M: medium dosage group; WSYGD-L: Low dosage group; Sinemet: positive control. **p<0.01,***p<0.001 vs control; #p<0.01, ##p<0.01, ###p<0.001 vs Model.

Effects of WSYGD on peripheral immunity of rotenone-intoxicated mice

The result of ELISA shown that the inflammatory cytokines contents such as TNF-α, IFN-γ, IL-1β, IL-22, and IL-17 in the mice serum of rotenone-intoxicated group were obviously increased than those in normal group. However, IGF-1 and TGF-β1 in rotenone-intoxicated group were evidently lower than that in the normal group. Simultaneously, WSYGD could reverse such phenomenon (Figure 4).

Figure 4. Changes of serum inflammatory factors including IL-10, IL-17, IL-22, IL-10, TGF-β1, TNF-α, IGF-1, IL-1β and NF-κB in Mice. Control: blank group; Model: rotenone-intoxicated group; WSYGD-H: high dosage group; WSYGD-M: medium dosage group; WSYGD-L: Low dosage group; Sinemet: positive control. **p<0.001, ***p<0.001 vs control; #p<0.05, ##p<0.01, ###p<0.001 vs Model.

UPLC-MS characterization of chemical components from WSYGD

All mass spectrum data acquired in the LC-MS/MS analysis were utilized to recognize the metabolites structures which was essential for structural confirmation of the biomarkers by the procedures as described in the before-mentioned. The UPLC-QTOF/MS analysis platform provided the precise molecular mass within measurement errors (<5 ppm), allowing the potential element composition, unsaturation degree and fractional isotope abundance of the compounds to be determined. Total ion chromatograms were obtained by LC-MS/MS (Supplementary Figure 1). A total of 97 peaks were obtained and confirmed or tentatively preliminarily peculiarity (Figure 5 and Supplementary Table 1). For instance, the molecular weight of 37 peak is 623.1969, and the main fragment ions analyzed by MS/MS screening were surveyed at m/z 605.1870 [M-H₃O]-, 487.1452 [M-C₈H₉O₂]-,477.1397 [M-C₆H₁₁O₄]-, 179.0344 [M-C₂₀H₂₉O₁₁]-, 161.0239 [M-C₂₃H₂₅O₁₀]-. Therefore, the elemental composition of ion 37 was calculated to be C₂₉H₃₆O₁₅ and identified as acteoside.

Figure 5. Chemical structures of WSYGD identified using UPLC-ESI-Q-TOF MS.

Multivariate statistical analysis for in vivo discrimination of WSYGD

The raw data were collected by the SYNAPT G2-Si Q-TOF before analysis. The representative total ion chromatograms (TICs) of WSYGD in mice were shown in Supplementary Figure 2. To compare the difference between control group and WSYGD group, unsupervised PCA and supervised OPLS-DA were produced and the results data were displayed as PCA scores plot (Figure 6A, 6B). The samples from WSYGD and normal group clusters were obviously diverse indicating that the clear discrimination between two groups.

Figure 6. PCA score plot of all analyzed samples in positive-ion (A) mode and negative-ion (B) mode with the statistical parameters; S-plot of OPLS-DA in positive-ion (C) mode and negative-ion (D) mode with the statistical parameters; VIP value plot in positive-ion (E) mode and negative-ion (F) mode.

To find the potential quality markers, OPLS-DA analysis was performed to emerge S-plot (Figure 6C, 6D). As it could be observed in Figure 6C, 6D every point symbolizes an ion t_R-m/z pair; the X-axis symbolizes the ingredient contribution and the Y-axis symbolizes the ingredient confidence. The farther away the distance is from the origin the greater the contribution to classification. Thus, the variables located at the two ends of “S” sharp were the compounds which are more important to discriminating the two groups, namely potential characteristic chemical markers. Then, to minimize complex data, further filtering procedures based on VIP value (VIP>1) analysis were performed (Figure 6E, 6F). It turns out that 62 variables were screened out, including 23 and 39 ions in positive mode and negative mode, respectively. Simultaneously, the result of the variables being summarized in ESI, Table 1.

Prototypes constituents and metabolites screening of WSYGD in mouse plasma

Analysis of prototype constituents

To elucidate the MS/MS information of the exogenous using UPLC-ESI-Q-TOF-MS technique, a comprehensive comprehending of retention regularities and changes in mass spectra of the prototype constituents is essential. Combined the exact mass < 5.0 ppm, the composition of elemental can be explicit acknowledged. By analyzing their accurate mass and MS fragment pathways along with the available literature, 97 prototype ingredients were identified. Taken the compound of P13 as an example, P13 had a molecular ion [M-H] ^– at m/z 314.1393 and produce fragmentation with the peak at m/z 298.1083 [M-H-CH₄]^-, 286.1445 [M-H-CO]^-, 282.1133 [M-H-CH₃-OH]^-, 270.1135 [M-H-CH₄-CO]^-, 256.1348 [M-H-C₂H₃O₂]^-, 192.1022 [M-H-C₇H₆O₂]^-. The chemical formula of P13 was conjectured to be C₁₈H₂₁NO₄ based on the elemental composition, unsaturation degree and relative abundance. Therefore, ion P13 was deduced preliminarily as Coclaurine (Figure 7).

Figure 7. Chromatograms (A) and fragmentations and mode assignments (B) of coclaurine.

Characterization of mouse metabolites of WSYGD

The total ion chromatogram of the endogenous and exogenous compounds in plasma after oral administration of WSYGD was analyzed by the software of Metabolynx 4.1 (Supplementary Figure 2). The possible metabolic pathway of major compounds is indicated in Figure 8. Most of the metabolites remained the skeleton structural characteristics which absorbed in blood [17], the study of fragmentation pathways accelerated the authentication of metabolites from WSYGD in plasma.

Figure 8. The proposed metabolic profiles of paeoniflorin-related metabolites.

The retention time of M9 chromatographic peak was 5.050 min with the [M-H]^- ion at m/z 199.0970 and the characterized molecular ion at m/z 151.1002, 123.0988, 108.0978, and the fragmentation behavior was confirmed in accordance with previous literature [18]. Simultaneously, M1 was found on RT 0.402 minute with the ion of [M+H] ⁺ at m/z 201.1120. Therefore, M9 and M1 were identified as Paeonimetabolin I and Paeonimetabolin II.

Metabolite M2 exhibited the ion of [M-H]^- at m/z 297.0616 and was obtained at 0.720 minute. The chemical formula of M2 was C₁₃H₁₄O₈, which was 176 Da higher than that of benzoic acid and exhibited the fragmentation of 121.0132 [M-GluA] ⁺. M2 was tentatively assigned as glucuronidated product of Benzoic acid-O-glucuronide.

Metabolites M14 (t_R=1.35 min) showed the ion of [M-H]^- at m/z 375.1291, which the molecular formula was calculated to be C₁₆H₂₄O₁₀. The fragment ion which produced by [M-H]^- was exhibited at 375.1291 [M-C₇H₄O]^-. Simultaneously, it was 104 Da lower than paeoniflorin, and the characterized fragment ion at m/z 345.1186, 327.1124, 283.1115. Ultimately, M14 [19] was tentatively identified as desbenzoylpaeoniflorin. Metabolite M6 was detected on the Rt 2.51 min of chromatogram and produced an ion of [M-H]^- at m/z 507.1758 (C₂₁H₃₂O₁₄), which was 132 Da (C₆H₈O₆-CO₂) higher than desbenzoylpaeoniflorin, indicating that the ingredients were conjugated with glucuronic acid, happening phase II in metabolic processes. Simultaneously, the main products ion of 327.1046, 283.1035, 165.1022, 121.1049 was present based on the fragmentation pattern. Therefore, M6 was designated as desbenzoyl product of paeoniflorin-O-glucuronide. Metabolite M11(t_R=5.53 min) presented an ion of [M-H]^- at m/z 449.1440 (C₁₇H₂₄O₁₀). Therefore, M5 and M11 were designated as demethylene hydroxyl oxypaeoniflorin, demethylene hydroxyl paeoniflorin. M26 exhibited quasi-molecular [M-H]^- ion at m/z 521.1659 and the chemical formula calculated to be C₂₅H₃₀O₁₂, which was 42 Da lower than paeoniflorin. M26 was identified as acetylation paeoniflorin.

Paeoniflorin was easy isomerize into albiflorin, and the intensity of the peaks is different. M20 had an [M-H]^- ion at m/z 493.1714 (C₂₄H₃₀O₁₁), which was 15 Da higher than albiflorin. M22 emerged an ion of [M-H]^- at m/z 317.1025(C₁₇H₁₈O₆) and 30 Da more than albiflorin. Therefore, M20 and M22 were identified as methylalbiflorin and methyloxyalbiflorin. Metabolite M21 exhibited the ion of [M-H]^- at m/z 317.1025 (C₁₇H₁₈O₆), which was 162 Da higher than that of albiflorin. It was indicating that the constituent was obtained through the loss of glycoside group. M21 was designated as albiflorinaglycone.

M11(t_R 5.62 min) and M16 (t_R 6.56 min) emerged the ions of [M+H]⁺ at m/z 399.1950 and 399.1911, respectively. Simultaneously, the chemical of M11 and M16 have the same molecular formula, C₂₂H₂₆N₂O₅, and were 16 Da more than that of isocorynoxeine. M11 and M16 were acknowledged as isocorynoxeine-N-oxide and corynoxeine-N-oxide based on the reserve time, exact mass and diagnostic fragment ions [20]. The other metabolites, such as M2, M3, M4, M7, M8, M10, M12, M13, M15, M17, M18, M19, M23, M24, M25, were identified using the same methods.

Quantification of ten main constituents in WSYGD

The correlation coefficient (r²) of the calibration curve > 0.999, indicating good linearity. The LOD, LOQ, intra- and inter day precision, repeatability, and recoveries were tested and the obtained results were summarized in Tables 2, 3, demonstrating good repeatability for the simultaneous quantification of the ten main constituents in WSYGD. Their typical chromatograms were shown in Figure 9.

Figure 9. UHPLC-MS/MS spectra and their structures of ten mixture standard compounds in positive mode (A) (1 Geniposidic acid; 2 Coclaurine; 3 Rhynchophylline; 4 Nootkatone) and negative mode (B) (5 Rutin; 6 Echinacoside; 7 Acteoside; 8 Paeoniflorin; 9 Linderane; 10 Quercetin).