The plant materials, purchased from Qinping Market in Guangzhou China, were morphologically authenticated by one of the authors, Dr. Tina T. Dong. The corresponding voucher specimens were deposited in Center for Chinese Medicine of The Hong Kong University of Science and Technology. The herbs were tested to be qualified according to the requirements of Chinese Pharmacopeia (2015 Edition) and Hong Kong Materia Medica Standards.

According to Yi Xin Fang (Yasunori, 2011), GR-PR and ATR-PO were prepared according to the weight ratio of 1:1 and 1:2, respectively. In preparing the herbal extracts, 20 g of herb mixture was boiled in 160 mL of water for 2 h and extracted twice. The extracts were combined, dried under vacuum and stored at −80°C. KXS₂₀₁₂ was obtained by mixing the water extracts of GR-PR and ATR-PO together in 1:5 weight ratio. The quality control of herbal extracts was described in previous study (Yan et al., 2015). The extracts were solubilized in dimethylsulfoxide (DMSO) to give stock solution at a series of concentration from 15 to 50 mg/mL and stored at −20°C.

PC12 cells, originated from rat adrenal medulla, were purchased from American Type Culture Collection (ATCC, Manassas, VA). Cells were cultured in Dulbecco's modified Eagle's medium (DMEM), supplemented with 6% fetal bovine serum (FBS) and horse serum (HS), 100 units/mL penicillin and 100 μg/mL streptomycin in a humidified CO₂ (7.5%) incubator at 37°C. Fresh medium was applied every other day. All culture reagents were purchased from Life Technologies (Grand Island, NY).

Cultured PC12 cells, after serum starvation for 3 h in DMEM having 1% FBS, HS, penicillin, and streptomycin, were treated with KXS₂₀₁₂ for 48 h for western blotting, or 24 h for real-time quantitative PCR. The cell viability assay was performed to determine a safe concentration range (0–100 μg/mL) of each extract, at which all extracts did not induce cell proliferation or death (Yan et al., 2015). In the co-treatment, cultured PC12 cells were treated with NGF (1.5 ng/mL) and KXS₂₀₁₂ (15 μg/mL) for 24 h for western blotting assay. In the pre-treatment, PC12 cells were treated with herbal extract for 48 h and washed by 1x PBS twice before NGF (1.5 ng/mL) application for 24 h. In phosphorylation study, the cultures were starved for 5 h in DMEM before NGF (1.5 ng/mL) or herbal application.

Total RNA was isolated from cell cultures by RNAzol RT reagent (Molecular Research Center, Cincinnati, OH) according to the manufacture's instruction. The amounts of RNAs were detected by UV absorbance at 260 nm. The total RNA was used to do the reverse transcription with moloney murine leukemia virus (MMLV) reverse transcriptase (Life Technologies), according to the protocol provided by the manufacturer. Real-time PCR was performed by using FastStart SYBR Green Master (Roche, Indianapolis, IN), according to the manufacturer's instruction. The SYBR green signal was detected by Mx3000P^TM muitiplex quantitative PCR machine (BD Biosciences Clontech, San Jose, CA).

After the treatment, the cells were solubilized in lysis buffer containing 0.125 M Tris-HCl, PH6.8, 4% SDS, 20% glycerol, 2% 2-mercaptoethanol, and analyzed immediately or stored frozen at −20°C. Proteins were separated on 8% SDS-polyacrylamide gels and transferred to a nitrocellulose. Transfer and equal loading of the samples was confirmed by staining the ponceau-S. The nitrocellulose was blocked with 5% fat-free milk in Tris-buffer saline/0.1% Tween 20 (TBS-T), and then incubated in the primary antibodies diluted in 2.5% fat-free milk in TBS-T over night at 4°C. The primary antibodies were: anti-NF200 (Sigma-Aldrich, St. Louis, MO, N4142, 1:1,000), anti-NF160 (Sigma-Aldrich, N2787, 1:5,000), anti-NF68 (Sigma-Aldrich, N5139, 1:2,000), anti-GAPDH (Abcam Ltd, Cambridge, UK, ab9485, 1:1,000,000), anti-phospho-Trk A (Cell Signaling, Danvers, MA, 9141, 1:1,000), anti-Trk A (Cell Signaling, 2505, 1:1,000), anti-phospho-Erk1/2 (Cell Signaling, 9101, 1:1,000), anti-Erk1/2 (Abcam, ab17942, 1:5,000), anti-phospho-CREB (Cell Signaling, 9198, 1:1,000), and anti-CREB (Cell Signaling, 9197, 1:5,000). After that, the nitrocellulose was rinsed with TBS-T and incubated for 2 h at room temperature in peroxidase (HRP)-conjugated anti-mouse secondary antibody, or peroxidase (HRP)-conjugated anti-rabbit secondary antibody (Life Technologies, 31430 or 31460, 1:5,000), diluted with 2.5% fat-free milk in TBS-T. After intensive washing with TBS-T, the immune complexes were visualized using the enhanced chemiluminescence (ECL) method (GE Healthcare, Piscataway, NJ). The intensities of bands in control and samples, run on the same gel and under strictly standardized ECL conditions, were compared on an image analyzer, using a calibration plot constructed from a parallel gel with serial dilutions of one of the sample.

A light microscope (Diagnostic Instruments, Sterling Heights, MI) equipped with a phase-contrast condenser, 10x objective lens and a digital camera (Diagnostic Instruments) was used to capture the images with the manual setting. For analyzing the number and length of neurite, ~100 cells were counted from at least 10 randomly chosen visual fields for each culture. Using the photoshop software, the cells were then analyzed for the number and length of neurite. The cells were scored as differentiated if one or more neuritis was longer than diameter of the cell body, and they were classified to different groups according to the length of neurite that it possessed, i.e., < 15, 15–30, and >30 μm (Xu et al., 2012). The number of neurite per cell was counting only those neurite that was longer than cell body.

All data were analyzed using one-way ANOVA followed by the Students t-test. Statistical significance were classed as ^*P < 0.05; ^**P < 0.01; ^***P < 0.001.

According to ancient preparation method of herbal mixture, GR-PR and ATR-PO were prepared in a weight ratio of 1:1 and 1:2, respectively. The extraction efficiency of GR-PR and ATR-PO were about 25.79 ± 3.25 and 10.32 ± 2.78%, respectively (Mean ± SD, n = 3). Two approaches were selected for the quality control of paired-herb extracts: (i) chemical fingerprinting; and (ii) minimal-marker requirement. The identified chemicals were shown in fingerprints (Figures 1A,B), according to our published reports (Zhu et al., 2010; Yan et al., 2015). Eight chemical markers, i.e., GR-derived ginsenosides (Rb₁, Rd, Re, Rg₁), PR-derived 3,6′-disinapoyl sucrose (330 nm), ATR-derived α-asarone (258 nm), and β-asarone (258 nm) and PO-derived pachymic acid, were selected for chemical quantification (Figure 1). In GR-PR extract, the amounts of Rb₁, Rd, Re, Rg₁ and 3,6′-disinapoyl sucrose were about 4.287 ± 0.065, 0.397 ± 0.046, 3.648 ± 0.079, 3.304 ± 0.084, and 9.678 ± 0.067, respectively, in mg/g of dried herbal extract. In ATR-PO extract, the amounts of α-asarone, β-asarone, and pachymic acid were 6.292 ± 0.017, 0.736 ± 0.014, and 0.021 ± 0.003 in mg/g of dried herbal extract (Mean ± SD, n = 3) (Figure 1C). The established chemical parameters served as the control for repeatability of below biochemical analyses. Thus, KXS₂₀₁₂ was referring to a mixture of GR-PR extract plus ATR-PO extract at 1:5 weight ratio.

The cellular study of KXS₂₀₁₂ on PC12 cells was employed to investigate the function of herbal extracts in neuronal differentiation. The neuronal differentiation of PC12 cells could be determined morphologically in measuring the length of neurite: NGF induced the neurite outgrowth in a dose-dependent manner (Supplementary Figure 1). Three mammalian neurofilament subunits, NF68 (~68 kDa), NF160 (~160 kDa), and NF200 (~200 kDa), form hetero-dimers in making the structural domain of neurite (Schimmelpfeng et al., 2004; Xiong et al., 2016). Application of NGF in cultured PC12 cells also induced the expressions of neurofilaments in dose-dependent manners (Supplementary Figure 2). In cultured PC 12 cells, application of KXS₂₀₁₂ caused up-regulation of neurofilaments, NF68, NF160, and NF200, in dose-dependent manners (Figure 2A). The treatment of KXS₂₀₁₂ at 25 μg/mL slightly induced neurite outgrowth in PC12 cells. As compared to the control group, the effect possessed no significant difference neither on the length of neurite nor the number of neurite (Figure 2B). Thus, the herbal extract was not able to promote neuronal differentiation as single treatment.

In depressed mammalian brain, the level of NGF is reduced, and NGF-induced neuronal differentiation is down-regulated. Thus, the effect of KXS₂₀₁₂ in the presence of low level of NGF in neuronal differentiation was determined here. KXS₂₀₁₂ at 15 μg/mL, a concentration just barely showed induction of neurofilament expressions, was applied together with a small amount of NGF onto cultured PC12 cells by the methods of pre- and co-treatments. A suitable concentration of NGF was selected: this concentration should have no or little effect on induction of neurite outgrowth or neurofilament expressions. NGF at 1.5 ng/mL was selected in pre-treatment and co-treatment studies (see Supplementary Figures 1, 2): this concentration was to ensure the responses from a low amount of KXS₂₀₁₂ could be revealed. In both pre- and co-treatments, the inductions of NF68, NF160, and NF200, induced by KXS₂₀₁₂ plus NGF, were significantly higher at 3- to 4-folds (Figure 3). However, the effect triggered by pre-treatment was more robust than that of co-treatment. In neurite outgrowth assay, the number of differentiated cells by pre-treatment of KXS₂₀₁₂ and NGF was reached to ~30% (Figure 4A), and ~20% increase was revealed in the co-treatment (Figure 4B). By counting the number of neurite per cell, the pre- and co-treatment of KXS₂₀₁₂ plus low dose of NGF caused an increase, significantly, by about 10% of cells having neurite (Figure 4). In addition, the differentiated cells, induced by KXS₂₀₁₂ plus low amount of NGF, contained longer neurite, e.g., >15 and >30 μm (Figure 4). Compared to pre-treatment study, the inductive role of KXS₂₀₁₂ showed less robust effect in promoting NGF-induced neurite outgrowth by co-treatment (Figure 4).

NGF achieves its function by binding to and activating Trk A in PC12 cells. The NGF-activated Trk A stimulates downstream signaling pathways leading to neuronal differentiation and promoting cell survival (Zhang et al., 2000; Vaudry et al., 2002). The role of KXS₂₀₁₂ in phosphorylating Trk A was tested here. NGF induced Trk A phosphorylation in a dose-dependent manner, and NGF at 1.5 ng/mL was selected in phosphorylation study, which barely showed phosphorylation capability (Supplementary Figure 3). KXS₂₀₁₂ or co-treatment procedure did not effectively phosphorylate Trk A; however, the pre-treatment of KXS₂₀₁₂ together with low level of NGF showed much better responsiveness (Figure 5A).

In addition to Trk A, the downstream effectors of CREB and Erk1/2 were also determined here by pre- and co-treatment of KXS₂₀₁₂ with low amount of NGF. In cultured PC12 cells, NGF induced CREB phosphorylation (~40 kDa) at ~5-folds; however, co-treatment with KXS₂₀₁₂ did not show significant difference, as compared to NGF single treatment (Figure 5B). In contrast, the phosphorylation of CREB was significantly induced by ~7-folds in the scenario of pre-treatment of KXS₂₀₁₂ plus NGF (1.5 ng/mL; Figure 5B). Similarly, the Erk1/2 phosphorylation (~42/44 kDa) was also being markedly up-regulated with pre-treatment of KXS₂₀₁₂, while no significant difference, compared to NGF group, was observed in the co-treatment (Figure 5C).

To explore the role of KXS₂₀₁₂ in NGF signaling, the expression of Trk A in cultured PC12 cells were analyzed. The expression of Trk A at ~140 kDa, recognized by its specific antibody, was increased by 160% in KXS₂₀₁₂ (15 μg/mL)-treated cultures (Figure 6A). NF68 served as a positive control. The mRNA encoding Trk A was analyzed by real-time PCR. Application of KXS₂₀₁₂ in PC12 cultures induced Trk A mRNA in a dose-dependent manner: the maximal induction of over 150% increase was revealed at ~25 μg/mL (Figure 6B). This was in consistent with the phosphorylation study of KXS₂₀₁₂ that the up-regulation of Trk A expression could be an outcome of increasing sensitivity to low level of NGF during differentiation in PC12 cells.

Moreover, K252a, an inhibitor of tyrosine phosphorylation of Trk A, was applied onto cultured PC12 cells for 3 h before the pre- and co-treatment procedures. In the presence of K252a (0.1 μM), the protein expression of NF68 was totally decreased to that of control in KXS₂₀₁₂-treated cultures (Figure 7), which suggested that KXS₂₀₁₂-promoted neuronal differentiation was mediated by Trk A. These findings suggested that KXS₂₀₁₂ promoted neuronal differentiation by up-regulating Trk A expression, which could increase the cellular response to low level of NGF.

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.