To Elucidate the Inhibition of Excessive Autophagy of Rhodiola crenulata on Exhaustive Exercise-Induced Skeletal Muscle Injury by Combined Network Pharmacology and Molecular Docking

Abstract

Autophagy can remodel skeletal muscle in response to exercise. However, excessive autophagy can have adverse effects on skeletal muscle. Although Rhodiola crenulata (R. crenulata) is thought to regulate autophagy, its active ingredients and mechanisms of action remain unclear. In this study, molecular docking and network pharmacology were used to screen for autophagy-related targets of R. crenulata. Subsequently, protein–protein interaction (PPI) analysis was used to find the relationships between the inverse docking targets and autophagy-related targets and therefore highlight the key targets. And then the Database for Annotation, Visualization, and Integrated Discovery (DAVID) database was recruited to explain the functions and enrichment pathways of the target proteins. Finally, the potential targets were validated by immunohistochemistry of a mouse model of exhaustive exercise-induced skeletal muscle injury. We found a network of 15 major constituents of R. crenulata with 30 autophagy-related and 105 inverse-docking targets by molecular docking and network pharmacology. The results of PPI analysis indicated that 16 inverse-docking targets interacted 8 autophagy-related proteins. Further pathway analysis showed that R. crenulata could regulate exercise-induced skeletal muscle autophagy through mammalian target of rapamycin (mTOR), AMP activated protein kinase (AMPK) and Forkhead box protein O (FoxO). The results of our animal experiments indicated that R. crenulata could suppress the expression of Ubiquitin-like protein ATG12 (ATG12), Beclin-1 (BECN1), and Serine/threonine-protein kinase ULK1 (ULK1), while increasing the expression of MTOR, NAD-dependent protein deacetylase sirtuin-1 (SIRT1), and Microtubule-associated protein tau (MAPT). In conclusion, this study demonstrated that R. crenulata may protect skeletal muscle injury induced by exhaustive exercise via regulating the mTOR, AMPK, and FoxO singling pathway.

INTRODUCTION

Autophagy, a highly conserved eukaryotic cellular recycling process, has key effects on cell survival and maintenance by the degradation of organelles, proteins and macromolecules, and by the recycling of their breakdown products. However, abnormal autophagy will result in various human diseases including heart disease, neurodegeneration, myopathies, cancer, aging and metabolic diseases.¹⁾ Under normal conditions, autophagy plays a key role in maintaining homeostasis and cell function, such as the remodeling of skeletal muscle. However, under stress conditions including hunger and exercise fatigue, autophagy is up-regulated.^2,3) The excessive autophagy is thought to have adverse effects on skeletal muscle, leading to the enhancement of skeletal muscle catabolism and even atrophy.^4,5)

Traditional Chinese Medicine (TCM) remedies are increasingly considered to regulate autophagy.⁶⁾ Rhodiola crenulata (R. crenulata), Tibetan ginseng is one of the most popular traditional Tibetan medicines. It has primarily been used to treat cardiovascular disease, hypobaric hypoxia, microbial infection, tumor and muscular weakness.⁷⁾ In recent years, multiple compositions from R. crenulata have been extracted and analyzed and its regulatory effect on autophagy has been extensively studied.⁸⁾ R. crenulata extract could inhibit the fatigue index of weight-loaded swimming test mice.⁹⁾ In addition, salidroside, extracted from R. crenulata, could suppress inflammatory reactions, apoptosis and autophagy of serum and liver tissues.¹⁰⁾

Our research group has long been committed to the research of Tibetan medicine R. crenulata. Previous studies have showed that Salidroside, tyrosol, gallic acid and ethyl gallic acid may be the basis of the effective components of R. crenulata. Duoxuekang (containing R. crenulata), R. crenulata or its active ingredients have a protective effect on cerebral ischemia (oxygen) and myocardial ischemia by up-regulating the expression of hypoxia-inducible factor (HIF)-1α mRNA.¹¹⁾ R. crenulata and salidroside can inhibit the opening of mitochondrial permeability transition pore (MPTP) in neurons, maintain the morphology and function of mitochondria, and reduce the intracellular Ca²⁺ concentration. R. crenulata can improve brain mitochondrial energy metabolism, inhibit neuronal apoptosis, increased antioxidant potential of rats by regulating HIF-1α/microRNA 210/ISCU1/2 (COX10) signaling pathway to protect plateau hypoxia brain injury.¹²⁾ Clinically, R. crenulata can treat altitude sickness and has significant effects on cardiovascular and cerebrovascular system, nervous system, respiratory system and other diseases, with fewer adverse reactions.

However, whether the active ingredients from R. crenulata regulate the exercise-induced autophagy in skeletal muscles remains poorly known. At present, the network pharmacology is proved to be widely used to identify the active compositions of some TCMs and their action mechanisms.¹³⁾ R. crenulata plays multiple roles through its multiple ingredients, and is characterized by network pharmacology, such as multi-target and multi-pathway process. Therefore, in this work, network pharmacology and molecular docking were combined to explore the active components, specific targets and action mechanisms of R. crenulata. Immunohistochemistry was used to verify the results of network pharmacology and molecular docking. The roadmap for our experiment is presented in Fig. 1.

Fig. 1. Roadmap of This Study

(Color figure can be accessed in the online version.)

MATERIALS AND METHODS

Constituent Compounds of R. crenulata

In our previous studies, evidence has shown that R. crenulata aqueous extract can inhibit the expression level of mitochondrial autophagy protein p62. Then, we determined the main compounds in R. crenulata oral liquid. R. crenulata oral liquid contains 1.869 mg/mL gallic acid, 1.271 mg/mL salidroside, 0.536 mg/mL tyrosol, 0.011 mg/mL catechin, 0.099 mg/mL caffeic acid, and 0.409 mg/mL p-cournaric acid. These chemical constituents were confirmed by the literature about R. crenulata.^{8,12,14–21)} The other constituent compounds of R. crenulata were selected according the above references. The compounds were to be mentioned at least 3 times in all the previous articles. As results, 15 compounds of R. crenulata were selected for further study. The constituent compounds of R. crenulata are listed in Table 1. In order to acquire the 3D structure for each compound, the obtained 15 compounds were further searched at Pubchem database (https://pubchem.ncbi.nlm.nih.gov), and then their structures were saved in “SDF” file.

Table 1. Details of Compounds in R. crenulata

Name	Pubchem ID	Compound name	CAS ID
M1	689043	Caffeic acid	331-39-5
M2	9064	Catechin	154-23-4
M3	5316128	Crenulatin	63026-02-8
M4	72276	Epicatechin	490-46-0
M5	65064	Epigallocatechin gallate	989-51-5
M6	13250	Ethyl gallate	831-61-8
M7	370	Gallic acid	149-91-7
M8	6549	Linalool	78-70-6
M9	637542	p-Coumaric acid	501-98-4
M10	5280343	Quercetin	117-39-5
M11	10314695	Rosiridin	100462-37-1
M12	159278	Salidroside	10338-51-9
M13	3483754	Salipurposide	529-41-9
M14	222284	Sitosterol	83-46-5
M15	10393	Tyrosol	501-94-0

Identifying Autophagy-Related Targets

The autophagy-related targets were obtained by using 2 kinds of methods. One method was to obtain the differently expressed genes (DEGs) between normal cells and autophagy specimens using the microarray data GSE81719 downloaded from the Gene Expression Omnibus database. This dataset consisted of 4 normal cell samples and 4 loss of autophagy cell samples. DEGs were identified using Bioconductor/R limma package, and p-Value less than 0.01 and Fold Change more than 2 were used as the Cut-off value. The other method was to seek the known autophagy-related targets from two existing databases at TTD database (http://database.idrb.cqu.edu.cn/TTD/) and Drugbank database (http://www.drugbank.ca/).²²⁾

Molecular Docking of Autophagy-Related Targets

The autophagy-related proteins were defined as the proteins collected from the above databases. The human protein structures related to autophagy were searched at the UniProt database (https://www.uniprot.org/), and then the proteins with the highest resolution were selected. Meanwhile, the X-ray crystal structures of these proteins were searched at the RCSB PDB database (https://www.rcsb.org/). All processes of molecular docking are completed by the Maestro (version 11.1) software. For the amendment of protein structure, the addition of hydrogen atoms, and assigning bond orders, etc. were conducted by Maestro protein preparation wizard. The scaling factor and partial charge cutoff of van der Waals radius scaling 0.25 and 1 Å were used to generate the grids on active sites, respectively. LigPrep wizard was used for ligand optimization and energy minimization. Subsequently, a conformation for each ligand was produced. No parameters such as ionization were changed. There was no tautomer of proteins, and the specified chirality was retained. The docking procedure was performed after preparing the ligands and defining the grids on active sites of proteins. The output Glide score was presented as kcal/mol. A visualizer was used to examine the specific interaction between ligands and proteins. The proteins with the absolute value of Glide score greater than 3 were chosen as the proteins with good effect.

The Prediction and Screening for Inverse-Docking Targets

Since the components from R. crenulata may indirectly play roles in the autophagy-related proteins, the structures of obtained 15 compounds were searched at SwissTargetPrediction database (http://www.SwissTargetPrediction.ch/) to predict the proteins for each small molecule.²³⁾ The species of the obtained proteins were limited to “Homo sapiens,” and then the top 15 inverse-docking proteins were chosen. Due to the non-standard naming, the UniProtKB was used to acquire official symbols, thereby more accurately and consistently collecting the central hub proteins with functional information.

Protein–Protein Interaction (PPI) Analysis

To further investigate the indirect role R. crenulata plays in the autophagy-related proteins, and to research the relationship between autophagy proteins in molecular docking and inverse-docking proteins in R. crenulata, the PPI between the above mentioned inverse-docking proteins and the autophagy-related proteins were established at String database (http://string-db.org/).²⁴⁾ Accordingly, the species were also limited to “Homo sapiens” and a confidence score more than 0.4 was used. The protein derived from a protein-coding gene locus was represented by a node.

Building the Network

For more convenient research the relationships between the 15 compounds from R. crenulata with autophagy-related proteins and inverse-docking proteins. The Cytoscape software (version 3.7.1) was used to established networks.

Gene Ontology (GO) and Signaling Pathway Analysis

GO analysis and signaling pathway analysis for the targets were analyzed through the Database for Annotation, Visualization, and Integrated Discovery (DAVID).²⁵⁾ The results were assigned into 3 categories: biological processes, molecular function, and Kyoto encyclopedia of genes and genomes (KEGG) signaling pathway analysis. p ≤ 0.05 was regarded as significance. The ggplot2 was used to plot a bubble chart, which is a contributed visualization package in the R programming language.

Selection of Verification Targets

As previously described, a node was defined as the hub target if its degree was 2 times more than the median degree of all the nodes in the network.²⁶⁾ In the PPI network, the primary hub targets were selected by calculating the topological feature for each node via Cytoscape plugin (CytoNCA).

Animals

Male Institute of Cancer Research (ICR) mice (6 weeks) weighing 25 ± 2 g were purchased from Chengdu Dashuo Experimental Animal Co., Ltd. (Chengdu, China). All mice were grown under standard laboratory conditions with 50–60% humidity and 12-h light/12-h dark cycle at 23 ± 2°C. All mice were randomly assigned into 5 groups with 10 mice in each group. They had free access to standard diets and disinfected water. All animal experimental procedures were approved by the Animal Research Ethics Committee of Chengdu University of Traditional Chinese Medicine (Chengdu, China).

Loaded Swimming Test

Oral liquid of R. crenulata was purchased from Tibet Tibetan medicine group Co., Ltd. (authorized approval number: B20070002). Fifty mice were randomly divided into 5 groups with10 mice in each group for Oral Liquid of R. crenulata administration: (1) control group (CON), (2) exhaustive exercise group (EE), (3) low-dose group (L-R, 1.02 mL/kg/d), (4) medium-dose group (M-R, 3.03 mL/kg/d), and (5) high-dose group (H-R, 6.06 mL/kg/d). The three R groups were given intragastric administration once a day for two weeks. The control group and exhaustive exercise group were given the same dosage of disinfected water as the individual body weight. Three days before the load swimming test, all mice were familiar with 20 min of swimming training without load. Then, based on our previous studies, load swimming was used to challenge mice to establish quadriceps femoris injury caused by fatigue exercise. One hour after the last dose, all the mice were undergoing a lead-loaded swimming test (about 5% of each mouse’s weight) which was attached to their tail roots. All mice were trained separately under the same conditions (25 ± 1°C, 30 cm depth). After exhausting swimming, the quadriceps of mice were collected and fixed using 4% paraformaldehyde (pH = 7.2) overnight at 25°C.

Immunohistochemistry

Five micrometer thick sections were generated from the paraffin-embedded sections for mouse quadriceps using a cryotome (RM2016; Leica Microsystems GmbH, Germany). Subsequent to deparaffinization, the sections were blocked with bovine serum albumin (cat no.A8020; Beijing Solarbio Science & Technology Co., Ltd., China) at 25°C for 30 min, and subsequently incubated with the primary anti-ATG12 (1 : 400, cat no. ab155589; Abcam, U.K.), anti-BECLIN-1 (1 : 200, cat no. ab62557; Abcam), anti-MTOR (1 : 100, cat no. #2983; Cell Signaling Technology, U.S.A.), anti-SIRT1 (1 : 500, cat no. ab189494; Abcam), anti-ULK1 (1 : 100, cat no. #8054; Cell Signaling Technology) and anti-MAPT (1 : 600, cat no.#46687; Cell Signaling Technology) antibodies overnight at 4°C. After incubation with secondary antibodies (1 : 3000, cat no.70-GAR007; Hangzhou MultiSciences Biotech Co., Ltd., China) for 50 min at room temperature, the sections were stained with diaminobenzidine (DAB; Dako, Denmark) according to the manufacturer’s instructions. Photographs were acquired using a Nikon Eclipse TI-SR microscope (Nikon Corporation, Japan). Semiquantitative analysis for the specimens was performed using Image-Pro Plus (v. 6.0, Media Cybernetics, Inc., Rockville, MD, U.S.A.).

RESULTS

The Direct Relationship between R. crenulata and Autophagy-Related-Proteins

Thirty known autophagy-related targets were finally collected. Among them, MYO9A, DHRS3, GSTA4, CD180, and PRMT6 were found from DEGs between normal cells and autophagy specimens (Fig. 2). In these proteins, 3D structures of CD180, GSTA4, and PRMT6 were searched from UniProt database. The other 27 targets were found from the TDD and Drugbank databases (Table 2). Thirty X-ray crystal structures of the autophagy-related proteins from the RCSB database were acquired as targets for docking analysis for 15 primary compounds from R. crenulata. For reducing the false positive rate in the molecular docking process, we therefore selected small molecules with the absolute value of Glide score >3 for further network construction, with the aim of finding enough molecules potentially bound to autophagy related proteins. Thus, compounds and targets with the absolute value of Glide score >3 are presented in Fig. 3. Table 2 shows the professional database of 30 protein sources for molecular docking, official protein and corresponding gene names. Supplementary Table 1 shows 412 docking results with the absolute value of Glide score more than 3 for 15 compounds of R. crenulata and 30 proteins, and the highest absolute value of Glide score is 8.43.

Fig. 2. Heatmap of DEGs from GSE81719

Rows represented genes, and columns represented samples. The heatmap was color-coded according to the Z-score; red color represented a value of the high expression and green color represented a value of the low expression. (Color figure can be accessed in the online version.)

Table 2. Information of Known Autography-Related Targets for Molecular Docking

Gene name	Uniprot ID	Protein name	RCSB ID	Database
DAPK2	Q9UIK4	Death-associated protein kinase 2	2A2A	Drugbank
PARK7	Q99497	Protein/nucleic acid deglycase DJ-1	2R1U	Drugbank
PRKAG1	P54619	5'-AMP-activated protein kinase subunit gamma-1	2UV4	Drugbank
MAPK8	P45983	Mitogen-activated protein kinase 8	2XRW	Drugbank
CD180	Q99467	CD180 antigen	3B2D	DEGs
PRMT6	Q96LA8	Protein arginine N-methyltransferase 6	4QQK	DEGs
DAPK3	O43293	Death-associated protein kinase 3	3BHY	Drugbank
GSTA4	O15217	Glutathione S-transferase A4	3IK7	DEGs
MAPK14	Q16539	Mitogen-activated protein kinase 14	3LFF	Drugbank
CDK5R1	Q15078	Cyclin-dependent kinase 5 activator 1	3O0G	Drugbank
PIK3R2	O00459	Phosphatidylinositol 3-kinase regulatory subunit beta	3O5Z	Drugbank
MTOR	P42345	Serine/threonine-protein kinase mTOR	4DRI	Drugbank
S100A8	P05109	Protein S100-A8	4GGF	Drugbank
SIRT1	Q96EB6	NAD-dependent protein deacetylase sirtuin-1	4KXQ	Drugbank
ATG12	O94817	Ubiquitin-like protein ATG12	4NAW	TTD
CISD2	Q8N5K1	CDGSH iron-sulfur domain-containing protein 2	4OOA	Drugbank
HLA-DPA1	P04440	HLA class II histocompatibility antigen	4P5M	Drugbank
DAPK1	P53355	Death-associated protein kinase 1	4PF4	Drugbank
ATG5	Q9H1Y0	Autophagy protein 5	4TQ1	TTD
ULK1	O75385	Serine/threonine-protein kinase ULK1	4WNO	Drugbank
ULK3	Q6PHR2	Serine/threonine-protein kinase ULK3	4WZX	Drugbank
HLA-DRB1	P04229	HLA class II histocompatibility antigen	4X5W	Drugbank
ATG101	Q9BSB4	Autophagy-related protein 101	5C50	TTD
TBK1	Q9UHD2	Serine/threonine-protein kinase TBK1	5EP6	Drugbank
BECN1	Q14457	Beclin-1	5HHE	Drugbank
ATG4B	Q9Y4P1	Cysteine protease ATG4B	5LXI	TTD
LRRK2	Q5S007	Leucine-rich repeat serine/threonine-protein kinase 2	5MY9	Drugbank
ATG16L1	Q676U5	Autophagy-related protein 16-1	5NUV	TTD
TLK2	Q86UE8	Serine/threonine-protein kinase tousled-like 2	5O0Y	Drugbank
CDO1	Q16878	Cysteine dioxygenase type 1	6BPU	Drugbank
HMGB1	P09429	High mobility group protein B1	6CIK	Drugbank

Fig. 3. The Network Consisted of 15 Compounds with 30 Known Autophagy Targets

Compounds and targets with the absolute value of Glide score >3 are linked in straight lines. Proteins and components with the absolute value of Glide score >7 were highlighted in red straight lines. The green ellipses represented the compounds, and the blue rectangles represented the proteins. (Color figure can be accessed in the online version.)

The Network between the Compounds and Inverse Docking Targets

R. crenulata could also indirectly affect the autophagy-related targets via other proteins. More other proteins were acquired by performing the inverse docking in SwissTargetPrediction. According to the candidate compounds from R. crenulata and their potential targets, a compound target network was established. Fifteen nodes and 103 targets were shown in Supplementary Table 2. The interactions were presented in this network. As shown in Fig. 4, a network between the compounds and inverse docking targets was established. The network suggested that these central targets including carbonic anhydrase 2 (CA2) could be regulated by multiple compounds. However, peripheral nodes like epidermal growth factor receptor (EGFR), could only be modulated by one compound.

Fig. 4. The Network Consisted of 15 Compounds with 103 Inverse-Docking Targets

The green ellipses represented the compounds, and the blue rectangles represented the proteins. (Color figure can be accessed in the online version.)

PPI Network Analysis

The top 20 autophagy-related proteins and the inverse-docking proteins were analyzed using the PPI network (Fig. 5). The nodes of inverse-docking proteins not related to the autophagy proteins were deleted in the established PPI network (Fig. 6), in which 8 inverse-docking targets interacted with 16 autophagy-related proteins.

Fig. 5. The Complex Network between Compounds with Inverse Docking Proteins and Autophagy-Related Proteins

The green ellipses represented the compounds, the yellow rectangles represented the inverse-docking proteins, and the blue rectangles represented the autophagy-related proteins. (Color figure can be accessed in the online version.)

Fig. 6. PPI Network of Inverse-Docking Proteins and Autophagy-Related Proteins

The yellow rectangles represented the inverse-docking proteins, and the blue rectangles represented the autophagy-related proteins. (Color figure can be accessed in the online version.)

Pathway Enrichment Analysis for Candidate R. crenulata Targets

To know further the possible roles the 24 candidate proteins played, DAVID database was used as described in Materials and Methods. The molecular functions, biological processes and signaling pathways results were shown in Fig. 7, the molecular functions and biological processes related to the regulations of autophagy and protein serine/threonine kinase activity were obtained. Candidate targets could mainly be assigned to participate in regulation of autophagy and Alzheimer’s disease. In these pathways, exercise could suppress the starvation-induced autophagy via reactivation of mTOR signal in the skeletal muscles of hungry mice.²⁷⁾ AMP activated protein kinase (AMPK) could stimulate the autophagy in skeletal muscle cells via its effect on the transcriptional function of FoxO and participate in autophagosome formation.²⁸⁾ Therefore, the mTOR, AMPK, FoxO singling pathway was related to exhaustive exercise-induced skeletal muscle injury.

Fig. 7. A: The Bubble Chart of GO Enrichment Analysis of the Genes in PPI Network

B: The bubble chart of KEGG enrichment analysis of the genes in PPI network. “Gene Ratio” represented the proportion of the target genes in GO entries. A higher ratio represented a better enrichment. The size of the dots indicated the number of genes and the color of the dots reflected the −log₁₀ (p-value). (Color figure can be accessed in the online version.)

Selection of Verification Targets

The median value of “Degree” of nodes in the PPI network was 2.5. Thus, the nodes with “Degree” > 5 were set as the main hub targets. We selected ULK1, MTOR, SIRT1, BECN1, MAPT, and ATG12 as verification targets. The detail of topological characteristics of the targets in PPI network is listed in Table 3.

Table 3. Topological Characteristics of Targets in the PPI Network

Target name	Degree	Betweenness	Closeness	Eigenvector	LAC	Network
ULK3	4	0	0.191667	0.246981695	3	4
ULK1	6	8.233333333	0.207207	0.335510105	3.666667	5.15
SIRT1	8	121.2666667	0.221154	0.330686539	2.25	4.319048
PRKAG1	3	0	0.190083	0.196297362	2	3
PARK7	2	0	0.191667	0.055602767	1	2
MTOR	9	67.76666667	0.214953	0.408782691	3.333333	6.553571
MBNL2	2	0	0.179688	0.037289105	1	2
MBNL1	3	9.633333333	0.198276	0.08506611	1.333333	2.5
MAPT	8	109.2	0.209091	0.143340841	1.25	4.452381
LRRK2	5	62.73333333	0.214953	0.191463277	1.2	2.333333
HMGB1	2	0	0.198276	0.128111675	1	2
HLA-DRB1	1	0	0.045365	1.47E-09	0	0
HLA-DPA1	2	2	0.045455	8.93E-09	0	0
DAPK2	1	0	0.185484	0.072626017	0	0
CISD2	1	0	0.185484	0.072625987	0	0
CDK5R1	1	0	0.178295	0.023373043	0	0
CA9	1	0	0.18254	0.067750305	0	0
CA4	1	0	0.178295	0.023373043	0	0
CA2	1	0	0.045365	5.47E-08	0	0
BECN1	11	116.7666667	0.219048	0.438295484	2.909091	7.696429
BACE2	1	0	0.172932	0.018628359	0	0
BACE1	4	38.5	0.201754	0.114226364	1	1.333333
ATG4B	5	0.5	0.193277	0.294888109	3.6	4.75
ATG12	6	7.4	0.207207	0.342665106	4	5.4

The binding sites of salipurposide, salidroside, ethyl gallate, gallic acid, quercetin, and epigallocatechin gallate with the amino acids of corresponding proteins, with the highest absolute value of Glide score, are shown in Fig. 8, which suggests that the components of R. crenulata may interact with the autophagy-related proteins. The compounds bound well to the structures of autophagy-associated protein receptors and combined with lipophilicity through the hydrogen bonding, π–π interaction and van der Waals interaction between the ligand and active site residues of the proteins.

Fig. 8. Schematic Representation of the Molecular Docking between Autophagy Target Protein and Ligand Atoms of the Main Components from R. crenulata

(A–F) A: ATG12 and salipurposide B: BECN1 and salidroside C: MTOR and ethyl gallate D: SIRT1 and gallic acid E: MAPT and quercetin F: ULK1 and epigallocatechin gallate. Droplet shapes represent amino acid groups that interact with compounds in 4 Å. Capital letters represent the abbreviation for amino acids. Purple arrows and red lines represent hydrogen bonds and π–π bonds between ligands and proteins, respectively. (Color figure can be accessed in the online version.)

Immunohistochemistry

The levels of ATG12, BECN1, MTOR, SIRT1, MAPT, and ULK1 protein expression in quadriceps tissues were detected by immunohistochemistry (Fig. 9). The loaded swimming test in the EE group shared significantly enhanced expression of ATG12, BECN1, ULK1 and decreased expression of MTOR, SIRT1, MAPT. However, R. crenulata could suppress the expression of ATG12, BECN1 and ULK1, while increasing the expression of MTOR, SIRT1, and MAPT. Statistically significant differences were found in the positive average area expression rate of all the proteins in the five groups, suggesting that R. crenulata may repress excessive autophagy of EE-induced mouse skeletal muscle cells by regulating the above six autophagy-related proteins.

Fig. 9. Immunohistochemistry Staining Showing ATG12, BECN1, MTOR, SIRT1, MAPK and ULK1 Protein Expression in Quadriceps Tissues of Mice

A: Section 200-fold field photograph. B: Average Optical Density Statistics Table. Data were presented as average number ± standard deviation (S.D.). Different letters represented the significant differences between species, and the same letter indicated no significant differences between species, according to Tukey’s multiple comparison tests (p < 0.05). (Color figure can be accessed in the online version.)

DISCUSSION

Recently, with a growing amount of research on Traditional Chinese Medicine, the “multi-targets, multi-components” model has been considered as a more effective strategy for understanding drug actions to treat complex diseases.^29–31) Amazingly, network pharmacology broke the previous theory of single component and target, and was consistent with the multiple-components and multiple-targets TCM theory for the treatment of complicated diseases.³²⁾ Therefore, we used combined network pharmacology and reversed molecular docking to clarify the mechanisms of R. crenulata on the reduction of exercise-induced skeletal muscle autophagy.

Autophagy is the basics of the homeostasis and stress response of skeletal muscles under normal conditions. But excessive autophagy is harmful to muscle health and plays a pathogenic role in some muscle diseases such as muscle atrophy, weakness and fiber degeneration.³³⁾ It was reported that R. crenulata extract could alleviate fatigue by inhibition of autophagy.^10,34) Excessive ULK1 would lead to the reduction of the synthesis of skeletal muscle protein and the quality of muscle.^35,36) While decrease in the autophagy marker AGT12 would inhibit autophagy, which induces endoplasmic reticulum stress.³⁷⁾ Additionally, the Bcl-2 protein family can inhibit excessive autophagy by reducing the content of BECN1, which plays a vital role in maintaining the normal physiological function of skeletal muscle.³⁸⁾ Evidence has shown that exhaustive exercise-evoke hypoxia can down-regulate the activity of the MTOR protein through the mTOR pathway, thereby leading to dysfunctional or impaired mitochondrial function.³⁹⁾ What's more, SIRT1 is an important autophagy-related gene and energy metabolic receptor in the body.⁴⁰⁾ And high levels of SIRT1 can delay senescence and improve skeletal muscle function.⁴¹⁾ Meanwhile, evidence also demonstrated that MAPT may be associated with fatigue-related symptoms caused by some diseases, such as Alzheimer’s disease.⁴²⁾

Previous studies have suggested that R. crenulata could significantly increase the swimming endurance of mice.⁴³⁾ Our results of network pharmacology and molecular docking showed that R. crenulata could regulate multiple autophagy-related proteins, which provides potential active ingredients and targets of R. crenulata to regulate autophagy. The results of further pathway enrichment analysis also showed that R. crenulata could regulate EE-induced autophagy through the mTOR, AMPK, and FoxO pathways, and that ULK1, MTOR, SIRT1, BECN1, MAPT, and ATG12 may be the key targets. Finally, we validated the effect of R. crenulata on the expression of the above six autophagic proteins by immunohistochemical method. As shown in Fig. 9, the load swimming test enhanced the expression of ATG12, BECN1, and ULK1 and decreased the expression of MTOR, SIRT1, and MAPT. After being given oral liquid of R. crenulata, the expression of these proteins was gradually reversed, which is partly consistent with previous literature.⁴⁴⁾

Considering all the above-mentioned data, autophagy of skeletal muscle was enhanced in mice after exercise, but reduced by administration of oral liquid of R. crenulata. We therefore conclude that R. crenulata can regulate EE-induced autophagy of skeletal muscle through the ULK1, MTOR, SIRT1, BECN1, MAPT, and ATG12 protein, which are involved in the mTOR, AMPK, and FoxO singling pathway.

Acknowledgments

This work was supported by the National Key R&D Program of China (2017YFC1703904), Sichuan Science and Technology Program (2019YJ0480), and the National Natural Science Foundation of China (81973569). State Administration of Traditional Chinese Medicine of the People’s Republic of China (201507002).

Conflict of Interest

The authors declare no conflict of interest.

Supplementary Materials

The online version of this article contains supplementary materials.

REFERENCES

• 1) Parzych KR, Klionsky DJ. An overview of autophagy: morphology, mechanism, and regulation. Antioxid. Redox Signal., 20, 460–473 (2014).
• 2) Rogov V, Dotsch V, Johansen T, Kirkin V. Interactions between autophagy receptors and ubiquitin-like proteins form the molecular basis for selective autophagy. Mol. Cell, 53, 167–178 (2014).
• 3) Drake JC, Wilson RJ, Yan Z. Molecular mechanisms for mitochondrial adaptation to exercise training in skeletal muscle. FASEB J., 30, 13–22 (2016).
• 4) Munteanu I, Kalimo H, Saraste A, Nishino I, Minassian BA. Cardiac autophagic vacuolation in severe X-linked myopathy with excessive autophagy. Neuromuscul. Disord., 27, 185–187 (2017).
• 5) Nair U, Klionsky DJ. Activation of autophagy is required for muscle homeostasis during physical exercise. Autophagy, 7, 1405–1406 (2011).
• 6) Yang X, Xie J, Liu X, Li Z, Fang K, Zhang L, Han M, Zhang Z, Gong Z, Lin X, Shi X, Gao H, Lu K. Autophagy induction by xanthoangelol exhibits anti-metastatic activities in hepatocellular carcinoma. Cell Biochem. Funct., 37, 128–138 (2019).
• 7) Fu YY, Li LW, Hao SJ, Guan R, Fan GY, Shi CC, Wan HB, Chen WB, Zhang H, Liu GC, Wang JH, Ma LL, You JL, Ni XM, Yue Z, Xu X, Sun X, Liu X, Lee SMY. Draft genome sequence of the Tibetan medicinal herb Rhodiola crenulata. GigaScience, 6, 1–5 (2017).
• 8) Wen SY, Zhou JT, Chen YY, Ding LQ, Jiang MM. Study on ¹H-NMR fingerprinting of Rhodiolae Crenulatae Radix et Rhizoma. Chung Kuo Chung Yao Tsa Chih, 40, 2629–2633 (2015).
• 9) Wei W, Li ZP, Zhu T, Fung HY, Wong TL, Wen X, Ma DL, Leung CH, Han QB. Anti-fatigue effects of the unique polysaccharide marker of Dendrobium officinale on BALB/c Mice. Molecules, 22, 155 (2017).
• 10) Feng J, Niu PQ, Chen K, Wu LW, Liu T, Xu SZ, Li JJ, Li SN, Wang WW, Lu XY, Yu Q, Liu N, Xu L, Wang F, Dai WQ, Xia YJ, Fan XM, Guo CY. Salidroside mediates apoptosis and autophagy inhibition in concanavalin A-induced liver injury. Experimental and Therapeutic Medicine, 15, 4599–4614 (2018).
• 11) Zhang Y, Meng XL, Wu W, Lai X, Wang Y, Zhang J. Duoxuekang, a traditional Tibetan medicine, reduces hypoxia-induced high-altitude polycythemia in rats. A Sponsored Supplement to Science, 63–64 (2012).
• 12) Wang X, Hou Y, Li Q, Li X, Wang W, Ai X, Kuang T, Chen X, Zhang Y, Zhang J, Hu Y, Meng XL. Rhodiola crenulata attenuates apoptosis and mitochondrial energy metabolism disorder in rats with hypobaric hypoxia-induced brain injury by regulating the HIF-1α/microRNA 210/ISCU1/2 (COX10) signaling pathway. J. Ethnopharmacol., 241, 111801 (2019).
• 13) Hou J, Chen W, Lu H, Zhao H, Gao S, Liu W, Dong X, Guo Z. Exploring the therapeutic mechanism of Desmodium styracifolium on oxalate crystal-induced kidney injuries using comprehensive approaches based on proteomics and network pharmacology. Frontiers in Pharmacology, 9, 620 (2018).
• 14) Li X, Wang X, Hong D, Zeng S, Su J, Fan G, Zhang Y, Liu Y, Chen C, Qiu J, Fang Z, Wu H, Zhang X, Wang S. Metabolic discrimination of different Rhodiola species using ¹H-NMR and GEP combinational Chemometrics. Chem. Pharm. Bull., 67, 81–87 (2019).
• 15) Han F, Li Y, Ma L, Liu T, Wu Y, Xu R, Song A, Yin R. A rapid and sensitive UHPLC-FT-ICR MS/MS method for identification of chemical constituents in Rhodiola crenulata extract, rat plasma and rat brain after oral administration. Talanta, 160, 183–193 (2016).
• 16) Chen D, Fan J, Wang P, Zhu L, Jin Y, Peng Y, Du S. Isolation, identification and antioxidative capacity of water-soluble phenylpropanoid compounds from Rhodiola crenulata. Food Chem., 134, 2126–2133 (2012).
• 17) Li T, Ge Z, Zhang H. Study on the chemical constituents of Rhodiola crenulata. West China Journal of Pharmaceutical Sciences, 27, 367–370 (2012).
• 18) Yang Y, Feng Z, Jiang J, Zhang P. Chemical constituents of roots Rhodiola crenulata. Chung Kuo Chung Yao Tsa Chih, 48, 410–413 (2013).
• 19) Ni F, Xie X, Liu L, Zhao Y, Huang W, Wang Z, Xiao W. Flavonoids from roots and rhizomes of Rhodiola crenulata. Chin. Tradit. Herbal Drugs, 47, 214–218 (2016).
• 20) Huo J, Wang J, Wu Z, Zhang G. Analysis of chemical constituents of Rhodiola crenulata using high performance liquid chromatography-mass spectrometry. Natural Product Research and Development, 24, 1405–1407 (2012).
• 21) Lyu X, Fan F, Wen J, Fan G, Zhang J, Zhang Y. Simultaneous determination of 5 chemical components in 3 kinds of Rhodiola rosea by HPLC. China Pharmacy, 29, 2515–2519 (2018).
• 22) Li YH, Yu CY, Li XX, Zhang P, Tang J, Yang Q, Fu T, Zhang X, Cui X, Tu G, Zhang Y, Li S, Yang F, Sun Q, Qin C, Zeng X, Chen Z, Chen YZ, Zhu F. Therapeutic target database update 2018: enriched resource for facilitating bench-to-clinic research of targeted therapeutics. Nucleic Acids Res., 46 (D1), D1121–D1127 (2018).
• 23) Gfeller D, Michielin O, Zoete V. Shaping the interaction landscape of bioactive molecules. Bioinformatics, 29, 3073–3079 (2013).
• 24) Szklarczyk D, Gable AL, Lyon D, Junge A, Wyder S, Huerta-Cepas J, Simonovic M, Doncheva NT, Morris JH, Bork P, Jensen LJ, Mering CV. STRING v11: protein–protein association networks with increased coverage, supporting functional discovery in genome-wide experimental datasets. Nucleic Acids Res., 47 (D1), D607–D613 (2019).
• 25) Jiao X, Sherman BT, Huang W, Stephens R, Baseler MW, Lane HC, Lempicki RA. DAVID-WS: a stateful web service to facilitate gene/protein list analysis. Bioinformatics, 28, 1805–1806 (2012).
• 26) Zhang Y, Li Z, Yang M, Wang D, Yu L, Guo C, Guo X, Lin N. ONE NLJP. Identification of GRB2 and GAB1 coexpression as an unfavorable prognostic factor for hepatocellular carcinoma by a combination of expression profile and network analysis. PLOS ONE, 8, e85170 (2013).
• 27) Zheng DM, Bian Z, Furuya N, Oliva Trejo JA, Takeda-Ezaki M, Takahashi K, Hiraoka Y, Mineki R, Taka H, Ikeda S, Komatsu M, Fujimura T, Ueno T, Ezaki J. A treadmill exercise reactivates the signaling of the mammalian target of rapamycin (mTor) in the skeletal muscles of starved mice. Biochem. Biophys. Res. Commun., 456, 519–526 (2015).
• 28) Sanchez AMJ, Csibi A, Raibon A, Cornille K, Gay S, Bernardi H, Candau R. AMPK promotes skeletal muscle autophagy through activation of forkhead FoxO3a and interaction with Ulk1. J. Cell. Biochem., 113, 695–710 (2012).
• 29) Law BY, Mok SW, Wu AG, Lam CW, Yu MX, Wong VKJM. New potential pharmacological functions of Chinese herbal medicines via regulation of autophagy. Molecules, 21, 359 (2016).
• 30) Li W, Yuan G, Pan Y, Wang C, Chen H. Network pharmacology studies on the bioactive compounds and action mechanisms of natural products for the treatment of diabetes mellitus: a review. Frontiers in Pharmacology, 8, 74 (2017).
• 31) Ai X, Hou Y, Wang XB, Wang XY, Liang Y, Zhu Z, Wang P, Zeng Y, Li X, Lai X, Meng XL, Li Q. Amelioration of dry eye syndrome in db/db mice with diabetes mellitus by treatment with Tibetan Medicine Formula Jikan Mingmu Drops. J. Ethnopharmacol., 241, 111992 (2019).
• 32) Li S, Zhang B. Traditional Chinese medicine network pharmacology: theory, methodology and application. Chin. J. Nat. Med., 11, 110–120 (2013).
• 33) Grumati P, Coletto L, Schiavinato A, Castagnaro S, Bertaggia E, Sandri M, Bonaldo P. Physical exercise stimulates autophagy in normal skeletal muscles but is detrimental for collagen VI-deficient muscles. Autophagy, 7, 1415–1423 (2011).
• 34) Chen T, Yao L, Ke DZ, Cao WG, Zuo GW, Zhou L, Jiang J, Yamahara J, Li YH, Wang JW. Treatment with Rhodiola crenulata root extract ameliorates insulin resistance in fructose-fed rats by modulating sarcolemmal and intracellular fatty acid translocase/CD36 redistribution in skeletal muscle. BMC Complement. Altern. Med., 16, 209 (2016).
• 35) Green DR, Levine B. To be or not to be? How selective autophagy and cell death govern cell fate. Cell, 157, 65–75 (2014).
• 36) You JS, Anderson GB, Dooley MS, Hornberger TA. The role of mTOR signaling in the regulation of protein synthesis and muscle mass during immobilization in mice. Dis. Model. Mech., 8, 1059–1069 (2015).
• 37) Huang S, Lin Y, Liang Z, Wu Z, Chen Y, Chen C. Erythropoietin helix B surface peptide modulates miR-21/Atg12 axis to alleviates cardiomyocyte hypoxia-reoxygenation injury. American Journal of Translational Research, 11, 2422–2430 (2019).
• 38) He C, Bassik MC, Moresi V, Sun K, Wei Y, Zou Z, An Z, Loh J, Fisher J, Sun Q, Korsmeyer S, Packer M, May HI, Hill JA, Virgin HW, Gilpin C, Xiao G, Bassel-Duby R, Scherer PE, Levine B. Exercise-induced BCL2-regulated autophagy is required for muscle glucose homeostasis. Nature, 481, 511–515 (2012).
• 39) Zhang Y, Jing Y, Qiao J, Luan B, Wang X, Wang L, Song Z. Activation of the mTOR signaling pathway is required for asthma onset. Scientific Reports, 7, 4532 (2017).
• 40) Lin N, Li XY, Zhang HM, Yang Z, Su Q. MicroRNA-199a-5p mediates high glucose-induced reactive oxygen species production and apoptosis in INS-1 pancreatic beta-cells by targeting SIRT1. Eur. Rev. Med. Pharmacol. Sci., 21, 1091–1098 (2017).
• 41) Kume S, Uzu T, Horiike K, Chin-Kanasaki M, Isshiki K, Araki S, Sugimoto T, Haneda M, Kashiwagi A, Koya D. Calorie restriction enhances cell adaptation to hypoxia through Sirt1-dependent mitochondrial autophagy in mouse aged kidney. J. Clin. Invest., 120, 1043–1055 (2010).
• 42) Hu J, Geng R, Li K. Advance in relationship between Tau protein and Alzheimer’s disease. Chin. J. Rehabil. Theory Pract., 11, 1031–1034 (2014).
• 43) Xie L, Li Y, Liu X, Chen S, Wang K, Chen Y, Wang Q. Establishment of the swimming endurance system in mice and the verification of the anti-fatigue effect in hongjingtian. Chinese Journal of Comparative Medicine, 26, 71–76 (2016).
• 44) Chen XZ, Wu YP, Yang TH, Wei MT, Wang YX, Deng XB, Shen CC, Li WT, Zhang H, Xu WY, Gou LT, Zeng Y, Zhang YH, Wang ZQ, Yang JL. Salidroside alleviates cachexia symptoms in mouse models of cancer cachexia via activating mTOR signalling. J. Cachexia Sarcopenia Muscle, 7, 224–232 (2016).