Calycosin prevents bone loss induced by hindlimb unloading.

Bone loss induced by microgravity exposure seriously endangers the astronauts' health, but its countermeasures still have certain limitations. The study aims to find potential protective drugs for the prevention of the microgravity-induced bone loss. Here, we utilized the network pharmacology approach to discover a natural compound calycosin by constructing the compound-target interaction network and analyzing the topological characteristics of the network. Furthermore, the hind limb unloading (HLU) rats' model was conducted to investigate the potential effects of calycosin in the prevention of bone loss induced by microgravity. The results indicated that calycosin treatment group significantly increased the bone mineral density (BMD), ameliorated the microstructure of femoral trabecular bone, the thickness of cortical bone and the biomechanical properties of the bone in rats, compared that in the HLU group. The analysis of bone turnover markers in serum showed that both the bone formation markers and bone resorption markers decreased after calycosin treatment. Moreover, we found that bone remodeling-related cytokines in serum including IFN-γ, IL-6, IL-8, IL-12, IL-4, IL-10 and TNF-α were partly recovered after calycosin treatment compared with HLU group. In conclusion, calycosin partly recovered hind limb unloading-induced bone loss through the regulation of bone remodeling. These results provided the evidence that calycosin might play an important role in maintaining bone mass in HLU rats, indicating its promising application in the treatment of bone loss induced by microgravity.

Introduction

Bone loss induced by spaceflight has become one of the most important risk factors for astronauts, which threatens astronauts’ health and limits space exploration[1,2]. Among the 60 American and Russian astronauts who participated in the Mir Space Station (Mir) and the International Space Station (ISS) long-duration space missions (average 176 ± 45 days), about 92% of the astronauts suffered from more than 5% bone loss in at least one skeletal part, 40% of the astronauts suffered from more than 10% bone loss in at least one skeletal part[3]. When the Soviet astronauts flew in space for 75 to 184 days, the bone mineral density (BMD) of the calcaneus decreased 0.9–19.8%, and the spine BMD decreased about 0.3–10.8% for the astronauts who stayed at the Salute Space Station for 5 to 7 months[4]. At present, several preventative and therapeutic strategies have shown good efficacy for bone loss induced by spaceflight, such as physical exercise, mechanical stimulation and drug therapy[5,6], but still have certain limitations and significant deficiencies. Among them, drug treatments such as bisphosphonates have significant protective effects on bone loss induced by microgravity, but there were the inevitable side effects, such as osteonecrosis of the jaw, headache and nausea[7-9]. Therefore, new safer compounds could be found to prevent and treat bone loss induced by spaceflight.

Recent studies demonstrated that natural small molecule drugs derived from Traditional Chinese Medicine (TCM) for homology of medicine and food, have good prospects in the treatment of osteoporosis because of their safety, low side effects and low cost, such as curcumin, lycopene, and resveratrol, etc[10-12]. Radix Astragali, as a kind of TCM for homology of medicine and food, has been reported to have a positive therapeutic effect on bone disorder after spinal cord injury[13]. However, due to the complexity of its ingredients and the limitations of traditional drug research methods, it remains difficult to find the effective compounds for the treatment of bone loss. At present, the network pharmacology has been applied to comprehensively determine the potential active compounds and targets of complex drugs[14,15]. Network pharmacology integrates network biology analysis, gene connectivity and redundancy, and gene pleiotropy on the basis of systems biology and multi-directional pharmacology[16], which provides a new idea for drug discovery and mechanism research[17,18].

In this study, we initially used the network pharmacology to identify the effective natural compounds of Radix Astragali for the prevention of bone loss. Furthermore, the HLU rat model was used to evaluate the function of the compounds screened by network pharmacology analysis in bone loss. As a result, we found that calycosin partly recovered HLU-induced bone loss through the regulation of bone remodeling. This research will provide a promising candidate for the treatment of bone loss induced by spaceflight.

Methods

Active compounds and targets of Radix Astragali

The chemical components of the herb in Radix Astragali were determined through the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP, https://tcmspw.com/tcmsp.php): a common database for the study of TCM and the components of Chinese herbal medicine. To further obtain the effective compounds from Radix Astragali, we performed an in silico ADME approach, which integrated predict drug-likeness (PreDL), predict oral bioavailability (PreOB) and predict Caco-2 permeability (PreCaco-2). Oral bioavailability (OB), as one of the most important pharmacokinetic parameters, can determine whether a small molecule compound has drug activity. Herein, we employed an in-house system OBioavail1.1 to perform the OB screening[19]. Additionally, Caco-2 permeability prediction model PreCaco-2 was developed to predict absorption of the drug through a robust in silico. And the PreDL was developed to classify drug-like and nondrug-like chemicals by the molecular descriptors and Tanimoto coefficient (as shown in Eq. (1))[20].where A denotes the molecular properties of ingredients in herb, and B represents the average drug-likeness index of all compounds in DrugBank database (http://www.drugbank.ca/). In this work, the compounds matching OB ≥ 30%, DL ≥ 0.18 and Caco-2 ≥ -0.4 were screened as bioactive compounds.

Further, we conduct the prediction and discovery of the targets of these active compounds based on a new computational model termed SysDT through two powerful algorithms of Random Forest (RF) and Support Vector Machine (SVM), which integrates large-scale information of chemistry, genomics and pharmacology[21]. Subsequently, we individually convert these predicted target names into the standard gene symbols through the Uniprot database (https://www.uniprot.org).

Disease target

The targets associated to bone loss, were identified from the GeneCards database (https://www.genecards.org/), using “osteoporosis” as the search term, and all the genes found in the database were considered as the target set of osteoporosis.

Network construction and topological analysis

To analyze and find potential drugs in the treatment of bone loss, we constructed and analyzed the visualized compound-target (C-T) network in this study. Using all bioactive compounds of Radix Astragali and the common targets of these active compounds with bone loss disease, a network diagram of compound-target interactions was generated, in which a compound and a target were linked with each other. By analyzing the topological properties of the C-T network, we identified the key compounds according to the degree values and neighborhood connectivity of nodes. The visualization of the network was achieved by the Cytoscape software.

Animal experiments

Twenty-four healthy Sprague-Dawley male rats weighing 200 ± 20 g (SPF standard) were kept at a suitable temperature of 25 ± 3 °C with a 12 h light-dark cycle. After 7 days of adaptation, all rats were randomly divided into four groups with six rats in each group, including baseline group, ground control group, HLU group, HLU treatment group with calycosin, 30 mg· kg−1 ·day−1. Among them, the rats in the baseline group were sacrificed before hind limb unloading and preserved their tissues for further examination. The rats in the control group were free to move without HLU, but the rats tails in the HLU and HLU+Calycosin groups were suspended to unload their hindlimbs as suggested by MoreyHolton and Globus[22]. Calycosin was intragastrically administered once a day in the HLU+Calycosin group for 4 weeks; and an equal volume of 0.5% CMC-Na was received in the control and HLU groups, daily, for 4 weeks. Besides, the rats of each group were injected with 5 mg/kg calcein subcutaneously on days 13 and 3 before necropsy. All procedures strictly followed the Guidelines for the Care and Use of Laboratory Animals, and were approved by the Institutional Ethics Committee of Northwestern Polytechnical University.

BMD analysis

After the rats were injected intraperitoneally with anesthesia and deeply anesthetized, they were neatly arranged on test bench of the bone densitometer. We scanned the whole body of the rats by the small animal scan mode using dual-energy X-ray absorptiometry (DEXA) assay (Lunar Prodigy Advance DXA, GE healthcare, Madison, WI, USA)[23]. After all scans were completed, then we selected the femurs as the region of interest (ROI) to analyze the BMD value using GE’s purpose-designed software (enCORE 2006, GE Healthcare, Madison, WI, USA).

Micro-CT scanning

The right femur fixed with 4% paraformaldehyde (PFA) was fixed on the template and scanned along the long axis with a micro CT scanner (Skyscan 1276, Bruker microCT, Kontich, Belgium)[24]. The specific parameters during the scanning: voltage was 70 kVp, current was 114 μA, the image pixel was 10 μm, and layer spacing was 10 μm. Later, the scanning files were analyzed to reconstruct and analyze the three-dimensional image. And the trabecular bone with a 2 mm thickness and 1 mm in the growth plate was selected as the area of interest for further data analysis. The 3D parameters for qualitative analysis of the trabecular bone were as follows: the bone volume per tissue volume (BV/TV; %), trabecular number (Tb.N; 1/mm), trabecular spacing (Tb.Sp; mm), trabecular thickness (Tb.Th; mm), and trabecular pattern factor (Tb.Pf; 1/mm). The cortical bone with a 0.5 mm thickness and 4 mm below the femoral growth plate was selected as the region of interest for further data analysis. The 3D parameters for qualitative analysis of the cortical bone were the cortical thickness (Cr.Th; mm). Among them, the 3D parameters of the trabecular bone and cortical bone were qualitatively analyzed by CTan software, and the 3D images of the trabecular bone and cortical bone were constructed by CTvol software and CTan software (Skyscan 1276, Bruker microCT, Kontich, Belgium).

Three-point bending mechanical test

The mechanical properties of the left femurs were tested using a universal material testing machine (Intron company, USA). Before mechanical testing, the left femurs stored at −80 °C were soaked in saline solution to thaw for about 4 h, and then measured them on the material testing machine. Each sample was placed on two supports separated by a distance of 20 mm. The load was applied to the middle of the diaphysis and pressed down at a speed of 2 mm/min until fracture occurred[25]. The load and displacement data were then recorded by acquisition computer. According to the load-displacement curve and the internal and external diameters of the fractured bone, the mechanical properties were calculated by Matlab software, including: the maximum load (N), toughness (J/mm2), stiffness (N/mm), maximum stress (MPa) and Young’s modulus (GPa).

Dynamic bone histomorphometry

The femurs were fixed in 4% PFA and dehydrated in 60–70–80–90–95% alcohol, with each step taking for 24 to 72 h. It was then dehydrated twice with for 12 h to 24 h each time, as was xylene. Next, the femurs were put into the embedding bottle, and adding an appropriate amount of embedding agent. Routine sectioning about 10 µm by the hard tissue slicer (Leica Microtome, HistoCore AUTOCUT) and place sections in an oven at 60 °C overnight. The dynamic parameters of mineral deposition rate (MAR) (micrometers per day) were quantified using CaseViewer software.

Serum analysis

The rats’ blood was collected from the abdominal aorta. Serum was obtained by centrifuging at 1200 rpm for 10 min at 4 °C. The concentrations of bone turnover markers and cytokines of bone remodeling in serum were measured by using rat ELISA kits, including NTX, TRACP5b, PINP, RANKL, OPG, BGP, ALP, IL-4, IL-6, IL-8, IL-10, IL-12, TNF-α and IFN-γ. All testing procedures were carried out according to the manufacturer’s instructions.

Statistical analysis

All the results are presented as the mean ± SD. Statistical analyses were carried out using GraphPad Prism 8.0 (GraphPad Software Inc., San Diego, CA, USA). Statistical significant differences were performed by two-tailed Student’s t test analysis between two groups. P < 0.05 was considered statistically significant.

Table 1

The information of 23 interacting targets of calycosin related to osteoporosis.

	Target name	Name	Uniprot ID	Organism
1	ADRA1D	Alpha-1D adrenergic receptor	P25100	Homo sapiens
2	ADRA2C	Alpha-2C adrenergic receptor	P18825	Homo sapiens
3	ADRB1	Beta-1 adrenergic receptor	P08588	Homo sapiens
4	HTR3A	5-hydroxytryptamine receptor 3A	P46098	Homo sapiens
5	ACHE	Acetylcholinesterase	P22303	Homo sapiens
6	ADRA1B	Alpha-1B adrenergic receptor	P35368	Homo sapiens
7	ADRB2	Beta-2 adrenergic receptor	P07550	Homo sapiens
8	CAMKK2	Calmodulin	Q96RR4	Homo sapiens
9	PDE3A	CGMP-inhibited 3’,5’-cyclic phosphodiesterase A	Q14432	Homo sapiens
10	ESR1	Estrogen receptor	P03372	Homo sapiens
11	GABRA1	Gamma-aminobutyric acid receptor subunit alpha-1	P14867	Homo sapiens
12	CHRM1	Muscarinic acetylcholine receptor M1	P11229	Homo sapiens
13	CHRM3	Muscarinic acetylcholine receptor M3	P20309	Homo sapiens
14	NOS2	Nitric oxide synthase, inducible	P35228	Homo sapiens
15	NOS3	Nitric oxide synthase, endothelial	P29474	Homo sapiens
16	NCOA2	Nuclear receptor coactivator 2	Q15596	Homo sapiens
17	PTGS1	Prostaglandin G/H synthase 1	P23219	Homo sapiens
18	PTGS2	Prostaglandin G/H synthase 2	P35354	Homo sapiens
19	RXRA	Retinoic acid receptor RXR-alpha	P19793	Homo sapiens
20	SCN5A	Sodium channel protein type 5 subunit alpha	Q14524	Homo sapiens
21	F2	Prothrombin	P00734	Homo sapiens
22	PRSS1	Trypsin-1	P07477	Homo sapiens
23	OPRM1	Mu-type opioid receptor	P35372	Homo sapiens