Regulation of GABAA and 5-HT Receptors Involved in Anxiolytic Mechanisms of Jujube Seed: A System Biology Study Assisted by UPLC-Q-TOF/MS and RT-qPCR Method

The increase of the prevalence of anxiety greatly impacts the quality of life in China and globally. As the most popular traditional Chinese medicinal ingredient for nourishing health and tranquilizing mind, Jujube seed (Ziziphus jujuba Mill., Rhamnaceae) (SZJ) has been proved to exert anxiolytic effects in previous reports. In this study, a system biology method assisted by UPLC-Q-TOF/MS and RT-qPCR was developed to systematically demonstrate the anxiolytic mechanisms of SZJ. A total of 35 phytochemicals were identified from SZJ extract (Ziziphus jujuba Mill. var. spinosa [Bunge] Hu ex H.F. Chow), which interact with 71 anxiolytic targets. Protein-protein interaction, genes cluster, Gene Ontology, and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways analysis were subsequently conducted, and results demonstrated that regulation of serotonergic and GABAergic synapse pathways were dominantly involved in the anxiolytic mechanisms of SZJ extract. The effects of SZJ extract on mRNA expressions of multiple GABAA (gamma-aminobutyric acid type A) and 5-HT (serotonin) receptors subtypes were further validated in human neuroblastoma SH-SY5Y cells using RT-qPCR. Results showed that SZJ extract (250 μg/mL) significantly up-regulated the mRNA level of GABRA1 and GABRA3 as well as HTR1A, HTR2A, and HTR2B in non-H2O2 treated SH-SY5Y cells. However, it exerted an inhibitive effect on the overexpressed mRNA of GABRA1, GABRA2, HTR1A, and HTR2A in H2O2 treated SH-SY5Y cells. Taken together, our findings suggest that anxiolytic mechanisms of SZJ mostly involve the regulation of GABAergic and serotonergic synapse pathways, especially a two-way modulation of GABRA1, HTR1A, and HTR2A. Our current results provide potential direction for future investigation of SZJ as an anxiolytic agent.


INTRODUCTION
Anxiety is characterized as excessive and persistent worry about the future, which in turn can impact one's ability to carry out activities of daily living. Anxiety can be divided into generalized anxiety disorder, panic disorder, obsessive-compulsive disorder, social anxiety disorder, and posttraumatic stress disorder (Möhler, 2012;Cohen et al., 2015). Physiological anxiety symptoms include pounding heart, difficulty breathing, upset stomach, muscle tension, sweating, and feeling faint or shaky (Teychenne et al., 2015). The global prevalence of anxiety is estimated at 16.6% across the life span, and it becomes a burden of healthcare and quality of life (Somers et al., 2006); therefore, it is important to develop effective and safe solutions for treatment (Starcevic, 2006). Tricyclic antidepressants, serotonin-specific reuptake inhibitors, and benzodiazepines have been developed to mitigate anxiety. While effective, these classes of drugs come with many side effects, such as insomnia, sexual dysfunction, suicidal ideation, and/or drug-dependency (Lakhan and Vieira, 2010). Therefore, the use of complementary and alternative medicines to improve anxiety has received increased attention. Traditional medicinal materials, such as Jujube seed (Ziziphus jujuba Mill., Rhamnaceae) (SZJ) , saffron (Crocus sativus L.) (Lopresti et al., 2018;Milajerdi et al., 2018), valerian root (Valeriana officinalis L.), and passion flower (Passiflora incarnata L.) (Müller et al., 2003;Cass, 2004;Benke et al., 2009) have shown anti-anxiety benefits.
Although previous studies have shown promising anxiolytic effects of SZJ, the underlying mechanism has not been systematically and comprehensively investigated. In our current work, we developed an integrated strategy of system biology assisted by ultra-performance liquid chromatography quadrupole-time of flight mass spectrometer (UPLC-Q-TOF/ MS) and real-time quantitative reverse transcription polymerase chain reaction (RT-qPCR) to uncover the active phytochemicals and anxiolytic mechanism of SZJ extract. This approach provides a modern and practical way to study complicated chemical systems with multiple pathways and connected targets, which is otherwise a difficult challenge in mechanistic research of traditional Chinese medicine ingredients.

MATERIAL AND METHOD
Phytochemical Analysis of SZJ Extract Using Ultra-Performance Liquid Chromatography Quadrupole-Time of Flight Mass Spectrometer (UPLC-Q-TOF/MS)

Characteristics of the SZJ Extract
Commercial SZJ extract (batch number HS-180651) was purchased from Honsea Sunshine (Guangzhou, China). Dried seed of Ziziphus jujuba Mill. var. spinosa (Bunge) Hu ex H.F. Chow was used for SZJ extract production in Guangzhou, China. SZJ was extracted by supercritical fluid CO 2 to remove lipid fraction, and the residue was further extracted by 50% ethanol, followed by vacuum concentration and vacuum drying. The final extraction ratio is 5:1, and total content of jujuboside A and B was quantitatively detected as 0.20% using HPLC method. A voucher of the batch used has been deposited at -16°refrigerator, sampler chamber of Amway (China) R&D Center (Shanghai, China).

Sample Preparation
Thirty mg of SZJ extract powder was precisely weighted and then transferred to a centrifugal tube, with 1.5 mL of methanol (Mass Pure Grade from MERCK), ultrasonicated for 30 min (KQ-300DB, 300W,40kHZ) at ambient temperature, followed by centrifuge (12000 rpm, 5 min, SIGMA 3K15, SIGMA). The obtained supernatant is filtered through 0.22 mm filter member prior to UPLC-Q-TOF/MS analysis.

Data Analysis
Data recording and processing was performed by Analyst software (Version 1.6, AB Sciex, USA). The compounds were tentatively characterized based on their retention time, mass accuracy of precursor ions, MS/MS spectra, and fragmentation pathways, referring to the SCIEX natural products HR-MS/MS Spectral Library, standard references, and previous literatures.

Acquisition of Potential Targets Regarding Anxiolytic Benefits for Identified Phytochemicals
The targets of identified phytochemicals and their potential metabolisms were acquired from multiple databases. In addition to retrieving candidate targets from TCMIP and CTD platforms, PharmMapper Server (http://www.lilab-ecust.cn/ pharmmapper/) (Wang et al., 2017) was employed to fish the potential targets for those phytochemicals of which no available candidate targets were found in TCMIP and CTD. Those targets were excluded if their reliable score was lower than 0.8 from TCMIP, interaction counts less than 5 from CTD, and/or their normalized fit score lower than 0.8 from PharmMapper. After removal of the duplicates, erroneous, and non-Homo sapiens targets, the rest were then mapped to an anxiety-related gene targets database to screen out the intersecting targets.

Protein-Protein-Interaction (PPI) and Clusters Analysis.
The selective target genes of SZJ extract were imported to STRING (Version 11.0, https://string-db.org/) (Szklarczyk et al., 2019) to obtain PPI results. The interaction score set as high confidence (>0.7). The STRING analysed results were then imported into Cytoscape (Version 3.6.1) (Su et al., 2014), and cluster analysis of target genes was conducted using a Molecular Complex Detection (MCODE) plug-in according to the method in the literature (Wang et al., 2020).

Enrichments Analysis Along With Network Construction.
The target genes contained in the MCODE enriched clusters were imported into the database for annotation, visualization, and integrated discovery (DAVID, version 6.8) (https://david. ncifcrf.gov/tools.jsp) (Huang et al., 2009) to conduct Gene Ontology (GO) terms enrichment including biological processes, cell component, and molecular function. ClueGo plug-in (Version 2.5.7) (Bindea et al., 2009) was further employed to analyze and demonstrate their participated biologic process and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways, respectively. The latest databases of GO-Biologic Process annotation EB1-Uniport and KEGG pathway were selected. Visual style was set as "groups," minimum gene number, and percentage contained in a term set as 5 and 5%. Statistical method of two-sided hypergeometric test and Bonferroni step down p-value correction was used. Cut-off value of kappa score of GO term/pathway network connectivity was set as 0.5, and only term/pathway with p-value < 0.05 was shown.

Samples Preparation
Gamma-amino butyric acid (GABA) was kindly provided by Toong Yeuan International Group (Shanghai, China). GABA was considered as positive control in this test. SZJ extract and GABA ingredients were dissolved in deionized water and diluted in culture solution before use.

Samples Concentration Determination
CellTiter-Glo assay was used to evaluate the available concentration of test samples according to the method described in literature (Fallahi-Sichani et al., 2013). The inhibition on SH-SY5Y cell viability of a series of concentrations of SZJ extract and GABA samples, from 3 mg/mL-0.46 mg/mL, were respectively evaluated. The nonlinear fitting curve of logarithm concentration response to cells viability was then simulated by GraphPad Prism 7.0 (GraphPad Software Inc., San Diego, CA, USA). The concentration of SZJ extract for further test referred to the correspondence value of 90% cells viability.

Samples Stimulation and Grouping
A total of 6 group samples were simultaneously tested: control group (deionized water), 250 mg/mL SZJ extract group, 100 mg/ mL GABA group, 100 mM H 2 O 2 group, 250 mg/mL SZJ extract with 100 mM H 2 O 2 group, and 100 mg/mL GABA with 100 mM H 2 O 2 group. Stimulation duration of all group samples was 48 h. Duplicates for each group were set as 3.

RNA Isolation and Reverse Transcription
RNA isolation and reverse transcription were conducted following the method reported in the literature (Fuchsova et al., 2016). Briefly, RNAprep Pure Cell/Bacteria kit (Tiangen, China) was used to extract total RNA according to the manufacturer's instructions. NanoDrop One C (ThermoFisher, USA) was used to determine RNA yield and purity by absorbance ratios A260/A280 and A260/A230. OD260/OD280 ratios of the RNA of all samples were in the range of 1.8-2.0. 2 mg of total RNA used to synthesize the first strand complementary DNA (cDNA) using high-capacity cDNA reverse transcription kit (ThermoFisher, USA) according to manufacturer's directions. All reverse transcription products were 10-fold dilution.

Oligonucleotide Primers
PrimerBank was applied to search primers for the amplification of human GABRA1, GABRA2, GABRA3, HTR1A, HTR1B, HTR2A, HTR2B, and internal reference genes (GAPDH and ACTB). Nucleotide sequence of primers are listed in Table 1.

Real-Time PCR (qPCR).
Levels of mRNA were quantified by conducting qPCR reactions with SsoAdvanced Universal SYBR Green Supermix (Bio-Rad, USA) according to the manufacturer's directions. CFX equipped with software (Bio-Rad, USA) was used to perform measurements. PCR was in the 20 mL reaction system containing 0.5 mM primer, 10 mL Mix, 5 mL cDNA, and 3 mL RNase-free ddH 2 O, and the amplification and dissolution curve condition was shown in Table 2. qPCR amplification of ACTB (b-actin) and GAPDH (glyceraldehyde-3-phosphate dehydrogenase) transcription were used as the internal control to verify that equal amounts of RNA were used in each reaction. Fold expression was defined as the fold change relative to control cells.

Statistical Analysis
Bio-Rad CFX Manager software (Bio-Rad, USA) was employed to analyze raw expression data (CT values). For further statistical analysis we used values normalized to the normalization factor calculated as a geometric mean of the expression of two reference genes. All data were expressed as the mean ± SD, and a two-way ANOVA followed by Tukey test was applied for statistical analysis using GraphPad Prism 7.0.

Phytochemicals Identification of SZJ Extract
Thirty-five phytochemicals were identified from SZJ by using the UPLC-Q-TOF/MS method. The MS chromatogram in negative ion model of SZJ extract is shown in Figure 1. The chemical information of identified phytochemicals is seen in Table 3. Among them, 11 phytochemicals were identified by comparison with standard references. The rest of the compounds were identified through comparison with literature data. These phytochemicals are mainly classified into four subcategories: (1) saponins including jujuboside A and jujuboside B; (2) flavones and their C-glycosides including catechin, epicatechin, vicenin-2, swertisin, nicotiflorin, isovitexin and its analogues, and spinosin and its analogues; (3) organic acids including triterpenic acid (e.g., alphitolic acid), fatty acid (e.g., linoleic acid), and glycosylated organic acids (pseudolaroside B and oleuropein); and (4) alkaloids including indoleacetic acid derivatives (e.g., N-glc-indoleacetic acid) and isoquinoline alkaloids (zizyphusine).

Anxiolytic Effect-Related Targets of Phytochemicals in SZJ Extract and Their Metabolites
In an integrated search of multiple databases, a total of 476 targets were found to be relevant with anxiety related disorders or diseases, of which 455 targets were acquired for 35 phytochemicals and 11 metabolites. All the interactions among the phytochemicals and targets are listed in Supplementary  Table 1. After the two clusters were compared and analyzed, 71 target intersects were further determined and are listed in Table 4. Among the interactions of phytochemicals and metabolites on those 71 targets (data is not shown), (epi) catechin had the most interactions (degree =22), followed by kaempferol (degree =19), palmitic acid (degree =17), oleic acid (degree =15), betulinic acid (degree =14), apigenin (degree =14), zizyphusine (degree =12), and naringenin (degree =11), etc.

GO and KEGG Pathway Enrichment and Analysis
GO enrichment analysis was conducted on the 35 core targets by using DAVID. All the enriched GO terms are seen in Supplementary Table 2. The top 10 significant terms in biological process, molecular function, and cell component categories are shown in Figure 3. The results demonstrated that GO terms were mainly concentrated in neurotransmitter receptors signaling, particularly GABA and serotonin receptor signaling. Functions of activating neurotransmitter receptors and ligandgated ion channel via receptor complex, and regulating synaptic transmission were mainly involved. Cytoscape ClueGo plug-in was further applied to visualize the interaction network of biological process, as shown in Figure 4. All statistically significant biological processes were listed in Supplementary Table 3.
In addition, 35 identified core targets were imported to ClueGo for KEGG pathway enrichment, resulting in 14 statistically significant pathways. The targets-pathway network is shown in Figure 5, demonstrating that neuroactive ligandreceptor interaction is the most significant pathway with involvement of 18 targets, followed by serotonergic synapse pathway (8 targets), taste transduction pathway (7 targets), etc.
Other nervous system related pathways including GABAergic synapse and retrograde endocannabinoid signaling, signaling transduction related pathways including TNF signaling  pathway, and longevity regulating pathways were also significantly enriched. These results are consistent with the results from GO enrichment analysis. Taken together, these findings suggest that SZJ extract mainly exerts an anxiolytic effect via modulation of serotonergic and GABAergic systems. GABA A receptors in the central nervous system are known as the major targets for the first line treatment of anxiety (Poisbeau et al., 2018). GABA A receptors subunits, mainly including GABRA1, GABRA2, and GABRA3, have been reported to mediates anxiolysis (Möhler, 2012). Similarly, preclinical studies have suggested that most 5-HT receptors subtypes participate in anxiety-like processes, and blocking or stimulating individual 5-HT receptor subtypes might cause the anxiolytic-like effect (Żmudzka et al., 2018). HTR1A, HTR1B, HTR2A, and HTR2B were commonly targeted in preclinical studies for the anxiety treatment (Graeff et al., 1996;Clinard et al., 2015;Spiacci et al., 2016;Colangeli et al., 2019) as well as the anxiolytic-like studies of SZJ (Wang et al., 2010;Liu et al., 2015). Hence, among the potential targets acted by SZJ, GABRA1, GABRA2, GABRA3, HTR1A, HTR1B, HTR2A, and HTR2B were preferentially selected for further validation.

Effects of SZJ Extract on mRNA Expression of GABA A and 5-HT Receptor Subtypes
Based on the findings from system biology analysis, RT-qPCR was employed to evaluate the effects of SZJ extract on mRNA expression of GABRA1, GABRA2, GABRA3, HTR1A, HTR1B, HTR2A, and HTR2B. As shown in Figure 6, in the non-H 2 O 2 treated cells, GABA 100 mg/mL significantly enhanced the expression of GABRA1 (p< 0.0001), GABRA2 (p< 0.0001), and HTR1A (p< 0.001) comparing with the control group, while it significantly decreased the expression of GABRA3 (p< 0.01) and HTR2B (p< 0.05). Similar as GABA, SZJ extract 250 mg/mL exhibited significant effect on enhancing expression of GABRA1 (p< 0.001) and HTR1A (p< 0.05). However, contrast to GABA group, obvious expression enhancement of GABRA3 (p< 0.01) and HTR2B (p< 0.0001) were observed at 250 mg/mL SZJ extract. These differences in regulating GABA A receptor subtypes following SZJ and GABA stimulations could help to explain the different mechanisms of anxiolytic effect of the two ingredients.  Furthermore, because oxidative stress mechanisms have been well explored in anxiety disorders (Bouayed et al., 2009;Salim, 2017), the effects of SZJ extract on GABA A and 5-HT receptors were also evaluated in hydrogen peroxide (H 2 O 2 ) induced oxidative stress condition. As a result, a remarkable increase (p< 0.0001) in GABRA1, GABRA2, HTR1A, HTR1B, and HTR2A expression following 100 mM H 2 O 2 stimulation was observed. Currently, though the alteration of GABA A and 5-HT receptors under strong oxidative stimulation is unclear, our result in some extent suggested that overexpression of GABRA1, GABRA2, HTR1A, HTR1B, and HTR2A may be involved in oxidative stress-induced anxiety. Intriguingly, the sharp increase induced by H 2 O 2 in GABRA1, GABRA2, HTR1A, and HTR2A were significantly antagonized by both SZJ extract 250 mg/mL and GABA 100 mg/mL. Hence, inhibition of GABRA1, GABRA2, HTR1A, and HTR2A overexpression is also possibly involved in anxiolytic-like mechanisms of SZJ.
Traditional system biology analysis would take ADME (absorption, distribution, metabolism, and excretion) screening strategy that may exclude these glycosides with low oral bioavailability and low drug-likeness (Yue et al., 2017a;Yue et al., 2017b). For instance, ginsenosides are the dominant phytochemicals in ginseng and are thought to be contributed to its multiple bioactivities (Ru et al., 2015;Kim et al., 2017). However, the above analysis strategy, i.e., ADME screening, would exclude the ginsenosides along with their contributions on efficacy when performing systematic research of ginseng. Therefore, such an analytical strategy is incomplete and not systematic. In fact, it has been well demonstrated that the metabolites of the ginsenosides are responsible for the specific bioactivities (Chen et al., 2018;Kim, 2018). Similarly, the chemoinformatics and pharmacoinformatics approach indicated that jujubogenin was the effective GABA A agonist, neither jujuboside A nor jujuboside B (Chen, 2009). Gut microbes play an important role in favoring phytochemicals transformation into metabolites endowed with biological activity (Dey, 2019). As a result, the strategy that involves the metabolites of glycosides in gastrointestinal environment (e.g., gut microbes) will be a more reasonable approach to understand the actual efficacy and mechanism of herbal materials in which the glycosides are considered as the main active components. The GABA A receptors are chloride channels and are composed of several subunit classes (a, b, g, d, and ϵ) (Olsen and Sieghart, 2008). GABAergic neurotransmission plays an important role in anxiety status. Previous studies have shown that deficit of GABA A receptors and reduction of GABA transmission were observed in people with anxiety-like symptoms (Horowski and Dorow, 2002;Nutt and Malizia, 2004;Hasler et al., 2008). In contrast, positive modulation of GABA A receptors and enhancement of GABA transmission have shown anxiolytic effects. Classic benzodiazepines reduce anxiety by interacting with the GABA A receptors via the benzodiazepine binding site, which is present at the interface of a1, a2, a3, or a5 subunits and g subunit of GABA A receptors (Möhler, 2012). Other classes of compounds, GABA, barbiturates, and alcohol also could act at different benzodiazepine binding sites to increase the opening of the chloride channel resulting in enhancement of inhibitory synaptic transmission (Harris, 1990;Schousboe and Redburn, 1995). The results of our system biology study suggested the GABA A receptors signaling is a significant pathway involving in anxiolytic effect of SZJ. In fact, pharmacologic study has found that spinosin, a major C-glycoside flavonoid in SZJ, exerted anxiolytic-like effects via modulation of GABA A and 5-HT receptors (Liu et al., 2015). Similarly, 6′′′-feruloylspinosin and FIGURE 3 | Top 10 significantly enriched GO terms in biologic process (red), cellular components (green), and molecular function (blue) categories. The bubble diagram was made using JMP software 14.2.0 (SAS Institute Inc., USA). The bubble size is proportional to its involved targets percentage in the term.
spinosin have been reported to significantly enhance the expression of GABRA1 and GABRA5 mRNA in rat hippocampal neurons . In addition, it has been found that stimulation of jujuboside A at 50 µg/mL could increase the mRNA transcription levels of GABRA1, GABRA5, GABRB1, and GABRB2 in hippocampal neurons (You et al., 2010;Wang et al., 2015); however, long time stimulation of jujuboside A at a high dose of 100 µg/mL result in the decrease of GABRA1 and GABRB2 mRNAs expression (You et al., 2010). These results suggested a two-way modulatory effect of SZJ on GABRA1 and other GABA A receptors. Such benefits were similar to what we found in this work, that is, SZJ extract enhanced mRNA level of GABRA1 in non-H 2 O 2 treated SH-SY5Y cells, but inhibited the H 2 O 2 -induced overexpression of GABRA1. Therefore, combining with the results from literatures and our results, it was suggested that SZJ exhibited anxiolytic effects through modulating GABA A receptors, in which a two-way modulation of GABRA1 may play an important role.
It was well established that the alteration of various behaviors in anxiety disorders including appetite, mood, sleep, and cognitive function have been linked to the serotonergic system (Liu Y. et al., 2018;Liu et al., 2019). Serotonin receptors are FIGURE 4 | The interaction networks of enriched biological processes. ClueGO was applied to analysis procedure, and multiple color circles indicate that they are involved in multiple biological processes. prevalent throughout the nervous system and the periphery, and they potentially control the serotonergic neurotransmission throughout the brain and neuronal activity to alleviate neuropsychiatric disorders (Okazawa et al., 1999). Generally, the activation of HTR1A, HTR2A receptors can produce anxiolytic effects, whereas inactivation of them increases anxiety-like behaviors (Clinard et al., 2015;Spiacci et al., 2016). Involvement of other 5-HT receptors including HTR1B, HTR1B, and HTR2C in the mechanisms of anxiety have also been recognized (Graeff et al., 1996;Griebel et al., 1997;McCorvy and Roth, 2015). Similarly, our system biology analysis found that the serotonergic synapse pathway was dominant in anxiolytic effects mechanism of SZJ extract, in which different subtypes of 5-HT receptors were involved. Notably, as the same effect on GABRA1, SZJ extract also showed a two-way modulation on HTR1A and HTR2A in our RT-qPCR test. Genetic studies in animal models have suggested that anxietylike behavior can increase when the HTR1A function is eliminated or overexpressed (Overstreet et al., 2003). Hence, these results suggested the involvement of modulating serotonergic synapse pathway, specifically two-way modulation of HTR1A and HTR2A in anxiolytic effects of SZJ. In addition, the cannabinoid receptors (CNR) are extensively expressed in areas of the nervous system and have been found closely associated with anxiety behavior (Akirav, 2011). It has been well illustrated that endocannabinoid (eCB) reduces the serotonin release in the central nervous system and increases the expression and function of HTR1A in the hippocampus via the activation of CNR1 (Haj-Dahmane and Shen, 2011;Patel et al., 2017). Beyond CNR1, eCB system could exert actions on other targets including CNR2, transient receptor potential vanilloid receptor type 1 (TRPV1), or cyclooxygenase-2 (COX2) to participate in improvement of anxiety (Patel et al., 2017). In addition, cyclic AMP-responsive element-binding protein (CREB) has been suggested to be crucial for the role of HTR1A in modulating anxiety-related behaviors via mediating hippoacampus structural plasticity . Intriguingly, this systematic analysis work showed phytochemicals in SZJ extracts potentially FIGURE 5 | Targets-pathway network associated with anxiolytic effects of SZJ extract. A cytoscape ClueGo plug-in was applied to enrich the pathways and construct the network. act on the above mentioned targets including CNR1, CNR2, TRPV1, COX2, and CREB. These results, to some extent, suggested that the mechanism of action of SZJ in anti-anxiety may also involve those pathways/targets that indirectly modulate eCB and serotoninergic systems. More attention needs to be paid to those targets/pathways in further experimental studies on anxiolytic effects of SZJ.
The phytochemicals in herbal medicines are the substantial basis for their pharmacologic actions. Those phytochemicals with good bioactivity and high content are considered to be the chemical markers in quality control of the herbal medicines. Jujuboside A and spinosin are used to quality markers of SZJ crude drug in Chinese Pharmacopoeia (Edition 2020). Combining the results from literatures reports (Han et al., 2009;Abdoul-Azize, 2016) and our results, the modulations of GABAergic and serotoninergic systems seem the major mechanisms of SZJ exerting anxiolytic effects, as well as the traditional efficacy of nourishing heart and calming mind. Based on that, we abstracted the phytochemicals-targets-pathway sub-networks of GABAergic and serotoninergic synapse pathways. As shown in Figure 7, it demonstrated that metabolites of C-glycosides (spinosin, etc.) and jujubosides (jujuboside A, etc.) including apigenin, kaempferol, naringenin, genkwanin, and jujubogenin were involved in modulation of GABAergic and serotoninergic synapse pathways. The result provides the extra evidence to support that C-glycosides and jujubosides are responsible for the anxiolytic effects of SZJ, and they support jujuboside A and spinosin as chemical markers for quality control of SZJ and its preparations. Beyond the C-glycosides and jujubosides, the involvements of triterpenic acid (e.g., betulinic acid) and alkaloid (zizyphusine) were also observed in modulation of GABAergic and serotoninergic synapse pathways. Specifically, betulinic acid is a function of the modulating GABAergic system via multiple subtypes of GABA A receptors, whereas zizyphusine is a function of the modulating serotoninergic system via multiple subtypes of 5-HT receptors. Notably, it has been reported that zizyphusine was identified as one of the principal components in SZJ by principal component FIGURE 6 | Effects of SZJ extract on mRNA gene expressions of different subtypes of GABA A and 5-HT receptors. GraphPad Prism 7.0 software was applied for statistical analysis and graphing. All data were expressed as mean ± SD, and a two-way ANOVA followed by Tukey's test was applied. Two-way ANOVA analysis results presented that, GABRA1, F (2, 12) =57.27: p < 0.0001; GABRA2, F (2, 12) =112.50: p < 0.0001; GABRA3, F (2, 12) =40.10: p < 0.0001; HTR1A, F (2, 12) =677.80: p < 0.0001; HTR2A, F (2, 12) =13.07: p = 0.0010; HTR1B, F (2, 12) =42.99: p < 0.0001; HTR2B, F (2, 12) =151.50: p < 0.0001. Post hoc Tukey's test results were represented by comparing with non-H 2 O 2 treated control group, *p < 0.05, **p < 0.01, ***p < 0.001 and ****p < 0.0001; compared with H 2 O 2 treated control group, ### p < 0.001 and #### p < 0.0001.
analysis (Sun et al., 2014). And according to records in TCMIP, zizyphusine is exclusively derived from Ziziphus jujube (fruit or seeds), and its bioavailability and druglikeness is much better than C-glycosides and jujubosides. These findings recommend that the involvement of zizyphusine in quality control of SZJ extract and pharmacologic actions in anxiolytic effect is worth investigating in the future. Because the pharmacologic study of zizyphusine is poor at present, more attention could be paid to research and development of zizyphusine as a potential natural anti-anxiety drug.
There are some limitations of our current research. First, we did not conduct the behavioral test to confirm the anxiolytic-like effect of SZJ. Such benefit of SZJ was concluded on the base of previous pharmacodynamic studies, as well as clinical practice experience of traditional medicine. Our in vitro mRNA expression evaluation of GABA A and 5-HT receptors only used a single concentration of SZJ extract, which corresponds to approximately 90% cell viability in CellTiter-Glo assay. Leveraging the integrated approach of system biology, UPLC-Q-TOF/MS and RT-qPCR, present work contributed to the illustration of potential mechanism of action involved in the anxiolytic-like effect of SZJ. However, further in-depth preclinical studies are warranted to verify the results obtained from the current analysis.

CONCLUSIONS
In conclusion, our results systematically demonstrated that anxiolytic mechanisms of SZJ mainly involved the regulation of serotonergic and GABAergic synapse pathways, in which the two-way modulation of GABRA1, HTR1A, and HTR2A may play an important role. In addition to C-glycosides and jujubosides, triterpenic acids and zizyphusine identified in SZJ also contributed to the regulation of serotonergic and GABAergic synapse pathways. This study provides directional predictions of anxiolytic mechanism of SZJ and insights to improve the quality control of standard extraction.

DATA AVAILABILITY STATEMENT
The raw data supporting the conclusions of this article will be made available by the authors, without undue reservation, to any qualified researcher.

AUTHOR CONTRIBUTIONS
LC, XZ, and JK designed the study. XZ, YZ, LZ, and BL performed the experiments. LC, CH, and JD prepared the manuscript. All authors contributed to the article and approved the submitted version.

FUNDING
The Nutrilite Health Institute fully funded this study.