Network pharmacology-guided mechanism study uncovers inhibitory effect of Mahuang Decoction on lung cancer growth by impeding Akt/ERK signaling pathways

Abstract

Lung cancer (LC) ranks the leading cause of cancer-related death worldwide, due partially to the unsatisfactory therapeutic effect of the mainstream treatment. Alternatively, Chinese herb medicine (CHM) offers a bright perspective for treating complex diseases. Mahuang Decoction (MHD), a classic CHM formula, has been widely used in treating respiratory diseases in China for centuries, but its action mechanism has yet to be fully investigated. In this study, we first systemically explore the action mechanism of MHD by using network pharmacology and bioinformatic analysis tools, which uncovered a potential “new use of old drug” for MHD in cancer treatment. The therapeutic effect of MHD on LC was then validated by oral administration of MHD in the immunodeficient mice bearing xenografted LC tumors. To better understand the pharmacological activity of MHD against LC, we next constructed a drug/disease-target PPI network composed of 252 putative core therapeutic targets of MHD using Cytoscape. The subsequent enrichment analysis for these targets suggested that MHD could affect the apoptosis and cell cycle of LC cells via impeding Akt/ERK signaling pathways. Notably, these in silico analysis results were further validated by a series of cellular functional and molecular biological assays. Thus, our results show that MHD holds a great potential in LC treatment.

Introduction

Lung cancer (LC) is the leading cause of cancer-related death worldwide. More than half of the LC patients are diagnosed at advanced stages, and thus miss the opportunity for curative surgery [1,2]. Currently, radiotherapy, chemotherapy and targeted therapy are mainstream treatments for advanced LC [3]. However, the overall therapeutic effect of these treatments is unsatisfactory [4], therefore novel approaches are needed to improve outcomes for LC patients. Chinese herb medicine (CHM), as an essential component of Traditional Chinese medicine (TCM), has been widely used to treat various diseases in China for thousands of years. Recent studies showed that CHM offers an attractive treatment option for treating complex diseases such as cancer, owing to its unique clinical effects [5-8].

Mahuang Decoction (MHD) is a classic CHM formula from Treatise on Febrile Diseases. With its distinct expectorant and cough relieving effects, MHD has been documented to treat various respiratory diseases since the 2nd century AD. It is composed of four herbs, including Ephedrae Herba (Ma-Huang, MH), Cinnamomi Ramulus (Gui-Zhi, GZ), Armeniacae Semen Amarum (Xing-Ren, XR), and Glycyrrhizae Radix Et Rhizoma (Gan-Cao, GC) (Figure 1A). Recent studies have demonstrated that MHD has desirable pharmacological activities in treating asthma through mitigating airway inflammation and febrile [9,10]. More recently, MHD was reported to have adjuvant therapeutic effects on chronic cough and malignant pleural effusion [11,12]. However, the action mechanism of MHD as well as its potential clinical application beyond respiratory diseases has yet to be fully evaluated.

Figure 1.

Construction of the MHD ingredient-target systematic network and enrichment analysis of the putative targets. A. The quality matching diagram of four important pharmaceutical ingredients from MHD (MH, GZ, XR, GC). B. The systematic network was constructed by linking the candidate active ingredients and their putative targets of the 4 herbs contained in MHD. C. The diagram of candidate drug target number of different 4 herbs in MHD formulated in accordance with the TCM principle of monarch, minister, assistant and envoy. D. Putative drug targets were enriched in the representative signalling pathways using DAVID v6.8 (P < 0.05).

In the present study, we first conducted a virtual study to explore the action mechanism of MHD using network pharmacology and bioinformatic analysis tools, which suggested a “new use of old drug” for MHD in cancer treatment. Due to the well-established therapeutic effect of MHD on respiratory diseases, we next determined the growth of LC cells on immunodeficient mice after orally administered with MHD. Having validated the therapeutic effect of MHD on LC cells in vivo, we next performed a series of in vitro assays guided by network pharmacology and bioinformatics, so as to provide some insights into the action mechanism of MHD against LC.

Material and methods

Cell culture and reagents

Human LC LTEP-A-2 and Glc-82 cell lines were obtained from the China Infrastructure of Cell Line Resources (School of Basic Medicine Peking Union Medical College, China) and were maintained in RPMI-1640 supplemented with 10% (v/v) FBS and 100 U/ml streptomycin/penicillin in 5% CO₂ at 37°C. Antibodies against Bcl-2, Bax, Caspase 9/p35/p10, Cyclin D1 and beta-actin were bought from Proteintech. Antibodies against Cyclin B1, Cyclin A2, p-Akt (S473), pan-Akt, p-ERK (T202/Y204) and pan-ERK were bought from CST. Antibody against CDK2 was obtained from Abcam. The MHD was provided by the Tianjin Medical University Cancer Institute & Hospital TCM Pharmacy. The quality matching (g) of the four ingredients from MHD was as follows: MH:GZ:XR:GC = 3:2:2:1. The final concentration of MHD was 30 mg/ml.

Candidate ingredients composition of MHD

The chemical composition of all the 4 herbs was mainly obtained from Traditional Chinese Medicine Systems Pharmacology (TCMSP) Database (http://lsp.nwu.edu.cn/tcmsp.php) and Traditional Chinese Medicine Integrated Database (TCMID) (http://www/megabionet.org/tcmid/) in 2018, and in TCMSP, the parameters for selection of active ingredients were set as oral bioavailability (OB) ≥ 30% and drug-likeness (DL) ≥ 0.18 as standard [13,14]. In addition, literature-mining method (www.CNKI.net) was used to search for the ingredients that failed to meet the above parameters but have been reported to contain in the herbs.

Identifying putative drug targets and known LC-related targets

The systematic drug targeting approach was utilized to identify potential targets for medicinal composition of MHD [15]. The potential drug targets were obtained from TCMSP and SwissTargetPrediction databases (http://www.swisstargetprediction.ch/) (Supplementary Table 1). The known LC-related targets were obtained from Gene Expression Omnibus (GEO) database (Supplementary Table 2). Four gene expression datasets (GSE22863, GSE27262, GSE43458 and GSE101929) derived from human LC and adjacent normal tissues were included. The protein-protein interaction (PPI) data were analyzed using Bisogenet, a key plugin of Cytoscape, and the final result was integrated into a single graph from six analyses of the obtained PPI datasets.

Systematic network construction and correlation enrichment analysis

The interaction networks for the putative drug targets of MHD and the known LC-related targets based on the data obtained from the Bisogenet plugin were constructed and visualized using Cytoscape (Version 3.2.1) [16]. After merging the above two PPI networks, the topology parameters of each node in the merged network was calculated using CytoNCA, another important plugin in Cytoscape. The node with a score twice higher than the median of “Degree centrality” (DC) was considered important and appeared in the new systematic network (See more details in Supplementary Table 3). Functional and pathway enrichment analyses of the obtained putative and core targets were performed as before [17-19].

Xenografted immunodeficient mouse work

Four-week-old male BALB/c nude mice were purchased from Beijing Vital River Laboratory Animal Technology Co., Ltd. (China). All procedures for the animal experiments were conducted according to the Animal Ethics Committee of Tianjin Medical University Cancer Institute & Hospital. All the 14 animals were randomly divided into two groups of 7 mice each. At 5 weeks of age, LC Glc-82 cells (2×10⁷/mL) were inoculated subcutaneously into the right flank of 14 mice using 1-ml needles. Two weeks later, when the tumor was visible by the naked eye, the mice were perorally (p.o.) gavaged with either 200 µL normal saline (control) or MHD (300 mg/kg weight), and the animals were gavaged once a day during the experiments. Meanwhile, the tumor volumes were also measured once daily using the following formula: long diameter × (short diameter)²/2. On day 12, mice were sacrificed and the tumor tissues were weighed. None of mice died during the experiments.

Immunohistochemistry assays

The slides of tumor tissue sections were disposed of deparaffinization and antigen unmasking, and were then incubated with the antibody against Ki-67 (Abcam, UK) at 4°C overnight. After washing with PBS, the slides were incubated with Polymer Helper and Poly peroxidase-anti-mouse/rabbit IgG (PV-9000, ORIGENE, China), followed by further incubation with diaminobenzidine (DAB).

Cellular functional and mechanism assays

The cytotoxicity was assessed by using the instrument of xCELLigence RTCA. Measurements were taken continuously for 72 hours at 37°C, and the RTCA software was used for subsequent data analysis. The cell viability and colony formation assays were carried out as before [19,20]. The accumulated distance of cells were acquired on the Operetta CLS High Content Analysis System equipped with Harmony software (PerkinElmer, Waltham, MA, USA) using a ×20 long wide distance objective in a digital phase contrast mode at a temperature of 37°C and 5% CO₂. Apoptosis detection, cell cycle assay and western blot (WB) analyses were performed according to the manufacturer’s instructions.

Statistical analysis

All data were analyzed using SPSS 17.0 software (USA). Results were represented as mean with standard deviations (mean ± SD). The differences expressed were using the Student’s t-test, and P < 0.05 was considered as statistically significant.

Results

Candidate active ingredients and putative drug targets screening for MHD

To explore the action mechanism of MHD, we first conducted a virtual screening with combined OB and DL, two important ADME parameters, to identify the active ingredients in MHD. Eighteen potential ingredients with OB ≥ 30% and DL ≥ 0.18 from the herb constituents of MHD were obtained. Meanwhile, another 35 ingredients that either exhibit good pharmacological activities (with OB < 30% or DL < 0.18) or have been reported to be typical ingredients of MHD by literature mining were also collected for subsequent analysis. As such, a total of 53 ingredients from the four herbs in MHD were considered as the “candidate ingredients”. As shown in Table 1, the four herbs, namely MH, GZ, XR and GC contributed 20, 12, 10 and 11 candidate ingredients, respectively.

Table 1.

The main potential active ingredients identified by in four herbs

Herbs	Number	Components
Ephedrae Herba (Mahuang)	20	Leucopelargonidin, Quercetin, Delphinidin, Resivit, Kaempferol, Herbacetin, L-ephedrine, D-pseudoephedrine, L-norephedrine, D-norpseudoephedrine, L-methylephedrine, D-methylpseudoephedrine, Apigenin, Mahuannin A, Ephedrannin A, (+)-Catechin, Benzoic acid, Ferulic acid, Vanillic acid, Ephedran A
Cinnamomi Ramulus (Guizhi)	12	(+)-Catechin, Sitosterol, Beta-sitosterol, ent-Epicatechin, (-)-Taxifolin, Cinnamaldehyde, Cinnamic alcohol, O-Methoxycinnamaldehyde, Cinnamic acid, D-Camphene, ()-Terpinen-4-ol, Benzaldehyde
Armeniacae Semen Amarum (Xingren)	10	Amygdalin, Arachidic acid, Citral, Hexadecanoic acid, Linolenic acid, Mandelonitrile, Myristic acid, Stearic acid, Prunasin, (e)-2-Nonenal
Glycyrrhizae Radix Et Rhizoma (Gancao)	11	Glycyrrhizin, 18Beta-glycyrrhetinic acid, Liquiritigenin, Isoliquiritigenin, Glabridin, Licochalcone A, Liquiritin, Isoliquiritin, Glycyrol, Isoglycyrol, Glabrene

In some cases, CHM formula shows advantages in treating obscure and complicated disease, such as cancer, due to the synergistic effects among its multiple ingredients and their corresponding targets [21]. Thus, we next explored the putative targets of above 53 candidate ingredients in MHD using TCMSP and SwissTarget Prediction databases, which resulted in a total of 189 putative targets (Figure 1B). The numbers of putative targets in MH, GZ, XR and GC were 128, 86, 44 and 66, respectively (Figure 1C). Among these herbs, MH had more corresponding targets than the others, indicating it plays an important role in delivering pharmacological activities of MHD. Detailed information of these drug-related targets were listed in Supplementary Table 1. Of note, there were many overlapped targets among different ingredients, suggesting these ingredients may play important role in manifesting synergistic effects of MHD. In addition, the individual drug-target network was constructed to visualize the systematic interactions among these ingredients and their putative targets using Cytoscape 3.2.1 (Supplementary Figure 1).

KEGG enrichment analysis of the putative targets for MHD

Having identified the putative targets of MHD, we performed a KEGG enrichment analysis for these 189 targets using DAVID v6.8. Intriguingly, among the affected signals by MHD, the most enriched signaling pathway was Pathways in cancer, which for the first time suggested a potential application of MHD on cancer treatment (Figure 1D). To further evaluate the new use of MHD, the putative drug targets were also enriched in the representative diseases using DAVID v6.8. The result showed that cancer was indeed among the top diseases that could be potentially treated by MHD (Supplementary Figure 2).

Construction of PPI systematic networks and enrichment analysis of the core targets of MHD against LC

Since the curative effect of MHD on respiratory diseases has been well recognized, we then set about to explore the action mechanism of MHD on LC. Four gene expression datasets (GSE22863, GSE27262, GSE43458 and GSE101929) derived from human LC and adjacent normal tissues were obtained from Gene Expression Omnibus (GEO) database. The overlapped 155 disease targets among these datasets were collected as the “LC-related targets” for further analysis (Figure 2A-C). Detailed information of these LC-related targets was listed in Supplementary Table 2.

Figure 2.

The known LC-related targets were screened from Gene Expression Omnibus (GEO) database. A. Four heat maps from GEO chips, including GSE22863, GSE27262, GSE43458 and GSE101929. B. The Venn diagram of 155 common LC-related targets from 4 GEO chips. C. Construction of the LC-related disease targets network.

To better understand the complex interactions among the targets, we constructed a PPI network of putative drug-related target for MHD, which contains 5056 nodes and 114586 edges, using the Bisogenet, a key plugin for Cytoscape 3.2.1. Also, a LC-related target PPI network, containing 1674 nodes and 25339 edges, was constructed using the same method. Next, to further investigate the pharmacological mechanisms of MHD against LC, we intersected above two networks and thus obtained 1063 nodes and 19877 edges. Referring to a previous method, the score of DC, a topology parameter, for each node in the overlapping network were calculated by using CytoNCA plugin. Based on the score of DC (> 52), a network of significant targets for MHD against LC, containing 252 nodes and 6809 edges, was thereby constructed [22] (Figure 3 and Supplementary Table 3).

Figure 3.

In silico identification and systematic network construction of candidate core targets for MHD against LC. (A) The interactive PPI network of MHD putative drug targets was made of 5056 nodes and 114586 edges. (B) The interactive PPI network of LC-related disease targets was composed of 1674 nodes and 25339 edges. (C) The interactive PPI network of MHD against GC-related targets made of 1063 nodes and 19877 edges was shown. (D) PPI network of core targets extracted from (C), in which 252 nodes and 6809 edges were included.

We subsequently performed an enrichment analysis for these identified 252 core targets (nodes in the PPI network) by dividing them into GO biological process and KEGG signaling pathways. Specifically, the enriched biological processes were mainly focused on apoptosis and transcription, while the affected signaling pathways mainly included pathways in cancer, PI3K-Akt signaling pathway, cell cycle, MAPK signaling pathway, Epstein-Barr virus infection and viral carcinogenesis (Figure 4A, 4B and Tables 2, 3). These results indicated that MHD is likely to inhibit the growth of LC cells by regulating the key signaling pathways involving cell proliferation, apoptosis and cell cycle.

Figure 4.

GO and KEGG enrichment analysis of candidate core targets for MHD against LC. A. Candidate core targets were enriched in the representative biological processes (GO-BP) by using DAVID v6.8 (p-value < 0.05). B. Candidate core targets were enriched in the representative signaling pathways (KEGG) by using DAVID v6.8 (p-value < 0.05).

Table 2.

GO enrichment analysis of potential core targets for MHD against LC

Term	Gene count	P-value
GO-BP:0045944~positive regulation of transcription from RNA polymerase II promoter	61	1.73E-21
GO-BP:0006351~transcription, DNA-templated	55	5.17E-06
GO-BP:0000122~negative regulation of transcription from RNA polymerase II promoter	50	1.96E-19
GO-BP:0043066~negative regulation of apoptotic process	42	1.06E-20
GO-BP:0016032~viral process	40	1.38E-25
GO-BP:0045892~negative regulation of transcription, DNA-templated	39	8.99E-17
GO-BP:0045893~positive regulation of transcription, DNA-templated	38	1.84E-15
GO-BP:0006355~regulation of transcription, DNA-templated	37	0.00321423
GO-BP:0098609~cell-cell adhesion	33	8.60E-20
GO-BP:0006915~apoptotic process	31	1.90E-09

Table 3.

KEGG enrichment analysis of potential core targets for MHD against LC

Term	Gene count	P-value
hsa05200:Pathways in cancer	49	7.22E-19
hsa05169:Epstein-Barr virus infection	40	1.42E-23
hsa04151:PI3K-Akt signaling pathway	32	7.35E-09
hsa05203:Viral carcinogenesis	31	5.50E-14
hsa04110:Cell cycle	29	3.03E-18
hsa05202:Transcriptional misregulation in cancer	29	1.42E-14
hsa05168:Herpes simplex infection	29	1.32E-13
hsa05161:Hepatitis B	28	2.43E-15
hsa05034:Alcoholism	26	1.87E-11
hsa05166:HTLV-I infection	25	2.00E-07
hsa05215:Prostate cancer	24	1.16E-16
hsa04010:MAPK signaling pathway	24	7.24E-07
hsa04114:Oocyte meiosis	23	2.00E-13
hsa04120:Ubiquitin mediated proteolysis	23	2.51E-11
hsa05220:Chronic myeloid leukemia	22	2.27E-16
hsa04919:Thyroid hormone signaling pathway	22	4.76E-12
hsa04722:Neurotrophin signaling pathway	22	1.34E-11
hsa05160:Hepatitis C	22	1.02E-10
hsa05205:Proteoglycans in cancer	22	1.81E-07
hsa05206:MicroRNAs in cancer	22	5.15E-05

MHD inhibited LC growth in vivo and decreased the viability and motility of LC cells in intro

To investigate the direct cancer suppressive role of MHD in vivo, we determined the effect of MHD on xenografted LC on immunodeficient mice. As shown in Figure 5A and 5B, compared with the control mice, the growth of xenografted LC was greatly suppressed by MHD treatment. By day 12, the tumor volume of MHD-treated group were approximately 2.8-fold smaller than that of control group (P < 0.05) (Figure 5C). In line with this, the tumor weights showed a striking difference between the two groups (P < 0.05) (Figure 5D). The subsequent immunohistochemistry analysis showed that the number of Ki-67 positive cells was significantly decreased in MHD-treated tumors, compared to those in controls, indicating an anti-proliferative effect of MHD on these tumors (Figure 5E, 5F). These results provided an convincing evidence showing that MHD possesses a direct anti-LC activity.

Figure 5.

MHD suppressed development of the xenografted LC tumors on mice. A, B. The tumor volumes were measured and calculated once daily for 12 consecutive days. C. The photo of tumor sizes was shown on the day 12. D. The tumors were resected and weighted on the day 12. E. Immunohistochemistry staining for Ki-67 was performed by using the tumor slides from control and MHD-treated groups. F. Statistical analysis of the positive ratio of Ki-67 staining. *P < 0.05 based on the Student t-test.

Next, we further evaluate the growth suppressive effect of MHD on LC cells using xCELLigence RTCA instrument. The results from dynamic monitor of the cytoactivity revealed that the growth ability of LTEP-A-2 and Glc-82 LC cells were dramatically suppressed in MHD treatment group in a dose-dependent manner, compared to that in the control group (Figure 6A, 6B). Meanwhile, the results of CCK-8 assay showed that the viability of LC cells was significantly inhibited by MHD in a dose-dependent manner (Figure 6C, 6D). After treating the cells for 24 hours, IC₅₀ analyses showed that MHD exerted its 50% inhibitory effect on LTEP-A-2 cells at 173.40±4.89 μg/mL and Glc-82 cells at 278.90±4.30 μg/mL, respectively.

Figure 6.

MHD inhibited LC cells growth and decreased its viability. A, B. xCELLigence RTCA was used to evaluate the cytotoxicity at ~72 h with MHD treatment at 0, 25, 50, 100, 200, 400 μg/mL in LTEP-A-2 and Glc-82. C, D. CCK-8 assay was performed to measure cell viability at 24 h after MHD treatment at different dosages (0, 25, 50, 100, 200, 400 μg/mL). E-H. Cell colony formation assay was performed to measure cell clonality after MHD treatment for 24 h. **P < 0.01 based on the Student t-test. I-N. The Operetta CLS High Content Analysis System was used to evaluate the accumulated distance of cells at 24 h after MHD treatment at 170 μg/mL in LTEP-A-2 and 270 μg/mL in Glc-82. *P < 0.05 based on the Student t-test.

Furthermore, the ability of cell colony formation of the LC cells was determined in the presence of MHD. The results showed that the cell clonality of the LC cells was decreased in a dose-dependent manner following MHD treatment for 24 hours (Figure 6E-H). In addition, the average of accumulated distance of the migrating population in MHD treatment cells was also smaller than that of control cells, indicating that the cell mobility was inhibited by MHD treatment (Figure 6I-N). Thus, the above results demonstrated a direct inhibitory effect of MHD on the viability and motility of LC cells in vitro.

MHD induced apoptosis and cell cycle arrest of LC cells

We next carried out a serial of cellular functional assays to validate the in silico enrichment analysis results above. Firstly, the morphology of LC cells was observed after MHD treatment for 24 hours. Compared to the control cells, the MHD-treated cells showed typical characteristics of cell apoptosis, such as shrinkage, roundness and disappearance of stereopsis. Meanwhile, the MHD-treated cell nucleus showed dense Hoechst 33342 staining by fluorescence observation (Figure 7A). Also, the flow cytometry analysis revealed that the apoptotic population stained with Annexin V-FITC was significantly increased upon MHD treatment in a dose-dependent manner (Figure 7B-D). The following WB results also demonstrated that MHD promoted the accumulation of pro-apoptosis proteins Bax and cleaved Caspase-9, whereas the level of anti-apoptosis protein Bcl-2 was down-regulated by MHD in a dose-dependent manner (Figure 7H and Supplementary Figure 4A).

Figure 7.

MHD treatment resulted in apoptosis and disturbance of cell cycle progression in LC cells. A. The cell morphology was observed in white light and fluorescence field using an inverted microscope. B. Induction of apoptosis of LTEP-A-2 and Glc-82 LC cells after MHD treatment. LTEP-A-2 cells were treated with MHD at different concentrations (0, 85 and 170 μg/mL) for 24 h, and Glc-82 cells were also treated with MHD at different concentrations (0, 135 and 270 μg/mL) for 24 h, when apoptotic events was assessed by flow cytometry. C, D. Statistical analysis of the percentages of the apoptotic cells in LTEP-A-2 and Glc-82 cells. **P < 0.01 based on the Student t-test. E. Cell cycle analysis of LC cells following MHD treatment in LTEP-A-2 (0 and 170 μg/mL) and Glc-82 (0 and 270 μg/mL) for 24 h by flow cytometry. F, G. Statistical analysis of the proportions of the cells at different phases in LTEP-A-2 and Glc-82 cells. **P < 0.01 based on the Student t-test. H. LTEP-A-2 (0, 85 and 170 μg/mL) and Glc-82 (0, 135 and 270 μg/mL) cells were treated with MHD at different concentrations for 24 h. After proteins were extracted, the protein expression levels of Bcl-2, Bax, pro-caspase-9, cleaved-caspase-9, Cyclin D1, CDK2, Cyclin A2 and Cyclin B1 were analyzed by WB assay.

Furthermore, we performed BrdU-incorporated cell profiling assay to evaluate the effect of MHD treatment on LC cell cycle. The results showed that MHD treatment significantly inhibited the proliferation rate of LC cells by arresting them at S phase (Figure 7E-G). Consistently, accumulation of the S phase-specific marker Cyclin A2 and down-regulated Cyclin D1, CDK 2 and Cyclin B1 protein levels were observed in the MHD-treated LC cells (Figure 7H and Supplementary Figure 4A). Together, these results showed that MHD inhibited cell growth via inducing cell cycle arrest and apoptosis of LC cells.

MHD inhibited growth of LC cells through impeding Akt/ERK signaling pathways

To further explore the underlying mechanism of the inhibitory effect of MHD on the growth of LC cells, we next evaluated the activities of the key signaling pathways involving cell proliferation and viability. Among those, PI3K-Akt and MAPK signaling pathways were selected for further investigation based on the previous KEGG pathway enrichment analysis. Indeed, these pathways were significantly impeded by MHD treatment, evidenced by dramatically reduced levels of the key factors in these pathways, such as p-Akt (S473) and p-ERK (T202/Y204). Furthermore, the levels of pan-Akt and pan-ERK were also down-regulated by MHD treatment (Figure 8 and Supplementary Figure 4B). Therefore, these results indicated that the affected apoptosis and proliferation of LC cells by MHD treatment were likely resulted from simultaneous inhibition of Akt/ERK signaling pathways, which shows a typical “multi-ingredient, multi-target and multi-function” pharmacological characteristics of CHM.

Figure 8.

MHD treatment resulted in down-regulating the phosphorylation protein expression level of Akt and ERK signaling pathways. LTEP-A-2 (0, 85 and 170 μg/mL) and Glc-82 (0, 135 and 270 μg/mL) cells were treated with MHD at different concentrations for 24 h. After proteins were extracted, the expression levels of p-Akt (S473), pan-Akt, p-ERK (T202/Y204) and pan-ERK were analyzed by WB assay.

Discussion

Syndrome differentiation is the core principle in TCM clinical practice, and the treatment protocol for the patient is guided by the TCM syndrome instead of the specific disease defined by modern medicine. However, the same TCM syndrome may manifest in different diseases, which means a given CHM formula might be effective on different diseases with the same syndrome. Therefore, to bring the ancient TCM practice into line with the modern medicine, it is an essential step to explore the “new use of old formula”. Undoubtedly, to decipher the action mechanism of CHM with a better understanding of the synergistic action among the multiple active ingredients in CHM, and to explore how this synergistic action results in a synergistic effect through their corresponding targets, may shed light on the new clinical applications of CHM. However, it is now still a bottleneck to unveil the scientific basis of the action mechanism for a given CHM in a holistic perspective by conventional approaches.

In fact, the holistic view of TCM shares much with the concepts of emerging system biology and network pharmacology, which define the complex and multi-level interactions through systematic analyses of various networks. Based on this point, a novel TCM network pharmacology approach has been recently launched, along with a series of powerful computational tools for TCM research. In this study, we started to explore the action mechanism of MHD by utilizing the TCM network pharmacology and bioinformatics analysis tools, which led to an interesting discovery of a potential pharmacological activity for MHD in cancer treatment. Indeed, the in vivo assay demonstrated the significant growth inhibition of xenografted LC cells after oral administration of MHD in immunodeficient mice. To further explore the underlying mechanism by untangling the complex interactions among the targets, we constructed a integrative PPI network based on the MHD-related and LC-related networks. According to the topo-parameter DC in the network, a total of 252 core targets were thereby identified to involve in the pharmacological effect of MHD against LC.

The subsequent GO and KEGG signaling pathway enrichment analyses for these core targets showed that apoptosis and transcription are the most affected biological processes by MHD treatment, while the disturbed signaling pathways mainly includes pathways in cancer, PI3K-Akt signaling pathway, and MAPK signaling pathway, suggesting that the inhibitory effect of MHD on LC cells is likely resulted from disturbance of the key signaling pathways involving in apoptosis and cell proliferation. Indeed, our western blot results showed that the levels of p-Akt (S473) and p-ERK (T202/Y204) were both dramatically reduced upon MHD treatment in LC cells. These results are in line with those of previous studies, which showed the anti-cancer effects of multiple ingredients from the herbs contained in MHD. For instance, cinnamic acid, a key ingredient from Cinnamomi Ramulus (GZ), could induce cell apoptosis and decrease the proliferation rate of melanoma cells, while its derivatives induced apoptotic cell death in colon and cervical cancer cells [23,24]. Amygdalin, an ingredient contained in Semen Amarum (XR), has been reported to exhibit its cytotoxic effect on multiple solid tumors by inhibiting cell proliferation, inducing cell apoptosis and impairing the immune functions in vivo [25]. Furthermore, several studies have previously determined the association between the ingredients identified in the present study and the key signaling pathways affected by MHD treatment. For example, cinnamaldehyde, another ingredient from GZ, exhibited desirable pharmacological activities in inhibiting angiogenesis and metastasis of tumor cells by targeting PI3K/Akt pathway [26]. Similarly, herbacetin, an ingredient of Ephedrae herba (MH), could suppress the motility of breast cancer cells by impeding the same pathway [27]. Also, glycyrrhizin from Glycyrrhizae Radix Et Rhizoma (GC) suppressed the growth and migration of leukemia cell via blocking Akt/mTOR signalings [28]. Meanwhile, 18β-glycyrrhetinic acid, an ingredient from Glycyrrhizae Radix Et Rhizoma (GC), was found to suppress the cell porliferation through inhibiting ERK signaling pathway in NSCLC cells [29]. Together, these results support a view that MHD holds a promising potential with its multi-component, multi-targets, multi-levels and coordinated intervention effects on LC treatment (Supplementary Figure 3).

Disclosure of conflict of interest

None.

Abbreviations

LC

Lung cancer

MHD

Mahuang Decoction

CHM

Chinese herb medicine

TCM

Traditional Chinese medicine

FBS

Fetal bovine serum

CCK-8

Cell counting kit-8

OB

Oral bioavailability

DL

Drug-likeness

PPI

Protein-protein interaction

DC

Degree centrality