Lycium barbarum polysaccharide (LBP) inhibits palmitic acid (PA)-induced MC3T3-E1 cell apoptosis by regulating miR-200b-3p/Chrdl1/PPARγ

Background Obesity is closely related to osteoporosis. Lycium barbarum polysaccharides (LBPs) have anti-osteoporosis activity. Objective This study aimed to explore the role of LBPs in palmitic acid (PA)-induced osteoblast apoptosis. Methods The microarray data set GSE37676 was downloaded from Gene Expression Ominibus (GEO) database. Top 300 differentially expressed genes (DEGs) were used to construct a protein–protein interaction (PPI) network based on STRING database, and significant modules were analyzed and their key genes were screened by using Cytoscape software. COEXPEDIA database showed that there was co-expression between Chrdl1 and peroxisome proliferator-activated receptor (PPARγ). MC3T3-E1 cells were treated with 100–500 μg/mL of PA. Reverse transcription polymerase chain reaction (RT-PCR) and western blot assays were used to detect mRNA and protein levels. Cell Counting Kit-8 (CCK-8) assay and flow cytometry were used to detect cell viability and cell apoptosis. Results Chrdl1 was the key gene from the most significant module and downregulation in MC3T3-E1 cells treated with PA. MicroRNA miR-200b-3p and PPARγ were significantly upregulated among PA-treated MC3T3-E1 cells. The results of luciferase reporter gene assay showed that miR-200b-3p targeted Chrdl1 3’-UTR. Over-expressing miR-200b-3p promoted PA-induced cell apoptosis and inhibited cell viability. After pre-treating cells with PA and LBP, MC3T3-E1 cell apoptosis rate was relatively lower than that of mimics+PA200 group. Chrdl1 inhibition partly reversed miR-200b-3p effect on inhibiting apoptosis among MC3T3-E1 cells pre-treated with LBP and PA. Decreased C CASP3, PPARγ and increased Chrdl1 by miR-200b-3p inhibition were partly reversed by Chrdl1 inhibition. Conclusions LBPs inhibit PA-induced MC3T3-E1 cell apoptosis by mainly decreasing miR-200b-3p to upregulate Chrdl1, but miR-200b-3p/Chrdl1/PPARγ is not the only mechanism for LBPs protecting osteoblasts from PA.

T wo metabolic diseases that have attracted global attention at present are obesity and osteoporosis; both have severely affected the health, qualify of life, and working competence of patients (1). Research suggests that obesity is closely related to osteopenia (2). Osteoporosis is a type of systemic bone disease featured by decline in bone mineral density (BMD) and destruction of bone tissue microstructure (3). Various bone growth factors, pro-inflammatory cytokines, and hormones bind with its receptor to activate various transcription factors that promote the survival, maturation, and differentiation of osteoblasts, thus exerting a variety of Popular scientific summary • LBP suppresses the PA-induced osteoblast apoptosis via the miR-200b-3p/Chrdl1/PPARγ axis to play a protective role for obesity-induced osteoporosis, but miR-200b-3p/Chrdl1/PPARγ is not the only mechanism for LBPs protecting osteoblasts from PA.
physiological effects (4). Therefore, osteoblasts have an important part during osteogenesis. Obesity induces increased adipocyte differentiation, lipid accumulation, and reduced osteoblasts differentiation, thereby resulting in decreased bone formation (5). Moreover, research has indicated that obesity can change the expression of multiple signaling pathways involving bone apoptosis, bone formation, bone differentiation, and bone adsorption (6). In our experiment, we simulated the in vivo hyperlipidemia environment of osteoblasts through palmitic acid (PA) treatment in vitro to further examine the relationship between obesity and osteoporosis. As a traditional valuable herbal medicine in China, fructus lycii is a mature fruit of Solanaceae Lycium barbarum, which is mainly produced in Ningxia, Gansu, and Sinkiang. Lycium barbarum polysaccharide (LBP), one of the main active ingredients in fructus lycii, is suggested in pharmacological experiments to have anti-aging, anti-hyperlipidemia, obesity-improving, anti-fatigue, anti-tumor, anti-oxidation, and BMD-increasing effects (7,8). In China, fructus lycii is frequently used to treat bone injuries. According to reports, LBP can upregulate osteoblasts activity to promote its growth; besides, it suppresses the activity of peroxisome proliferator-activated receptor (PPARγ) and reduces fat production to improve obesity (9,10). Results of our previous research have demonstrated that LBP could suppress endoplasmic reticulum stress (ERS) to reduce the PA-induced osteoblast apoptosis (11). As it is well known, LBP is an active ingredient of traditional Chinese medicine, possessing multiple pharmacological molecular mechanisms. Therefore, this study aimed to further explore other molecular mechanisms by which LBP reduces the PA-induced osteoblast apoptosis.
MicroRNAs (miRNAs) have been recognized as single-stranded non-coding RNA molecules with the length of approximately 22 nucleotides and coded via endogenous genes. They have been extensively distributed in the eukaryotic cells of a variety of organisms. miRNAs can regulate cell growth, differentiation, and apoptosis to participate in the growth and development of organism as well as disease genesis and development (12,13). It has been found in research that the abnormally expressed miRNAs can specifically regulate the expression of bone formation-related signaling pathways and transcription factors which are thereby closely correlated with osteoporosis genesis (14). miR-206 can suppress the osteoblasts differentiation and miR-26 can promote osteoblast differentiation (15,16). Besides, miRNAs also exert indispensable role in the field of obesity. In thin and obese mice, the excessive intake of fructose elevated cholesterol and mRNA miR-27a levels, downregulated mRNA miR-33a level in the adipose tissues of liver, and elevated mRNA miR-21 level in plasma (17)(18)(19). miR-200b-3p is reported to inhibit breast cell proliferation meanwhile improving chondrocyte viability in osteoarthritis patients (20,21); however, the miR-200b-3p role in obesity-induced osteoporosis has not been reported yet. Consequently, this study adopted bioinformatic analysis and cytological experimental techniques to explore miR-200b-3p function and its mechanism in PA-induced osteoblast apoptosis.

Construction of the protein-protein interaction (PPI) network as well as module analysis
The online annotation approach for biological function and localization of protein in cell, that is, Search Tool for the Retrieval of Interacting Genes (STRING), was utilized for predicting the PPI data. The top 300 DEGs were mapped into STRING to assess relationships of PPI; typically, the confidence value of >0.4 was regarded as significant. Cytoscape was utilized to visualize PPI-integrated networks. Then the MCODE plug-in of Cytoscape was employed to analyze modules within the as-constructed PPI network, and the following default parameters were determined, including 'node score cutoff = 0.2', 'degree cutoff = 2', 'max depth = 100', and 'K-core = 2'.

COEXPEDIA analyses
The Chrdl1 co-expression was examined from COEX-PEDIA (http://www.coexpedia.org/) database, the context-related co-expression network database constructed based on single mouse and human microarray sample series on the basis of GEO database.
Cell culture MC3T3-E1 cells were provided by Shanghai Cell Bank of the Chinese Academy of Sciences (Shanghai, China), which were cultured in α-MEM medium containing 10% fetal bovine serum (FBS; Gibco; Thermo Fisher Scientific Inc., Waltham, MA, USA) in an incubator under humidity, 5% CO 2 , and 37°C conditions.

Cell transfection
The cell transfection was performed as described. miR-200b-3p mimics/inhibitors and their negative control, and Chrdl1 small interfering RNA (siRNA) and a small interfering negative control (siNC) were provided by Gene Pharma Co. Ltd (China). MC3T3-E1 cells were subjected to transfection with miR-200b-3p mimics/inhibitors/NC (50 pmol/mL), and siChrdl1 or siNC was transfected by Lipofectamine 2000 (Invitrogen) according to manufacturer's protocols.

Cell viability assay
Viability of cells was evaluated through Cell Counting Kit-8 (CCK-8) (Tongren, Shanghai, China). In each well of a 96-well plate, 4 × 10 3 cells were grown, followed by 24-h of incubation. At 1 h prior to the completion of incubation, each well was added with CCK-8 reagent. Then optical density (OD) at 450 nm was detected for each well using the microplate reader.

Dual-luciferase reporter gene assay
Chrdl1 mRNA 3′UTR that contained wild-type (wt) or mutant (mut) sequence within candidate binding site of miR-200b-3p was amplified before cloning to miRNA-Report luciferase reporter vector (Shanghai Gene Chem Co. Ltd., China). Then, luciferase reporter gene assay was conducted according to a previous method. Briefly, MC3T3-E1 cells were plated to 24-well plates 24 h before mutant type (Chrdl1mut) or wild type (Chrdl1-wt) luciferase vector co-transfected with miR-200b-3p mimics/inhibitors/NC through Lipofectamine 2000 (Invitrogen; Thermo Fisher Scientific Inc.). The luciferase activities were determined 24 h following transfection by Promega Dual-Luciferase Reporter Assay System (Promega Corp., Madison, WI, USA).

Flow cytometry assay
The Annexin V-FITC apoptosis detection kit (Nanjing KeyGen Biotech Co., Ltd., Nanjing, China) was used to detect apoptotic rate using flow cytometer in accordance with manufacturer protocols. In brief, cells were subjected to 24-h incubation before they were harvested, followed by ice-cold phosphate buffer solution (PBS) washing, binding buffer (500 µL) re-suspending, and addition of Annexin V stock solution (5 µL) for 10 min of incubation at 4°C. Later, 5-µL propidium iodide (PI) was added to cells, followed by immediate analysis by FACSCalibur flow cytometer (BD Biosciences, San Jose, CA, USA).

RT-PCR assay
Western blotting Cells were collected in each group, rinsed twice with PBS, and lysed with cold radioimmunoprecipitation assay buffer (RIPA buffer; Beyotime, Shanghai, China) supplemented with freshly prepared protease inhibitor cocktail (0.01%; Sigma, Shanghai, China). The cells were subjected to 30 min of incubation in ice. Then cell lysate was subjected to 10 min of centrifugation at 13,000 × g under 4°C to collect supernatants, which were (20-30-µg protein) isolated through 10% SDS-PAGE gel, followed by electrophoretic transfer onto the polyvinylidene fluoride (PVDF) membrane (Merck Millipore, Shanghai, China). Later, 5% skimmed milk was used to block protein blots, which were incubated with anti-GAPDH, anti-PPARγ, anti-Chrdl1 antibodies (Beyotime), followed by incubation with goat anti-rabbit or anti-mouse secondary antibody (Beyotime). The blots were observed through enhanced chemiluminescence (ECL; Thermo Fisher Scientific Inc., Shanghai, China).
Statistical analysis SPSS 19.0 (IBM SPSS Statistics, USA) was used to assess each variable. Numerical data were expressed as mean ± standard deviation (SD) and differences were detected by Student's t-test. Statistical analyses were performed through one-way analysis of variance (ANOVA) and Dunnett's test in succession. A two-sided difference at P < 0.05 was deemed to be significant statistical.

Results
DEGs identification mRNA expression profile datasets of MC3T3-E1 cells treated with/without ascorbic acid contained expression profile matrix from 45,102 gene probes. A total of 4,072 DEGs were identified by GEO2R, including 458 downregulated and 347 upregulated DEGs (P < 0.05, logFC > 1.5 or < -1.5). The DEGs are displayed in Fig. 1a.

PPI network construction and module analysis
All top 300 DEGs were mapped with STRING database for constructing the PPI network. Protein pairs having an integrated score of > 0.4 were screened. PPI network containing 288 edges was displayed with Cytoscape (Fig.  1b). Thereafter, MCODE, the plug-in used to search optimized scoring parameters to generate optimal network results, was utilized to search for modules from the network. Module 1 consisted of 28 edges and 11 nodes and had a higher score over other modules (Fig. 1c). Chrdl1 is a key gene from the most significant module.

Network of Chrdl1 co-expression
Gene co-expression that showed functional association with Chrdl1 was estimated by COEXPEDIA database to examine the effect and molecular mechanism of Chrdl1 on disease. Chrdl1 had a co-expression with 274 genes (Fig. 2). Interestingly, Chrdl1 was co-expressed with PPARγ.

LBP inhibited PA-mediated apoptosis in MC3T3-E1 cells under miR-200b-3p mimics transfection
For further studying LBP action mechanism in affecting PA-induced MC3T3-E1 cell apoptosis, flow cytometry (FCM) and CCK-8 assays were carried out to determine the viability and apoptosis of MC3T3-E1 cells with different treatments. According to our findings, miR-200b-3p mimics could inhibit MC3T3-E1 cell viability under PA treatment and further aggravate PA-induced cell apoptosis (Figs 5a-5c). Next, the function of miR-200b-3p mimics to some extent lessened cell viability while promoted cell apoptosis treatment with PA by treatment with 400 μg/mL of LBP. In addition, miR-200b-3p further increased PA-medicated PPARγ expression and decreased PA-medicated Chrdl1 expression (Fig. 5d). Meanwhile, the function of miR-200b-3p mimics increasing PPARγ protein level and decreasing Chrdl1 protein level of MC3T3-E1 cells treated with PA was reversed to some extent by treatment with 400 μg/mL of LBP.

LBP further inhibited PA-mediated apoptosis of MC3T3-E1 cells under miR-200b-3p inhibitor transfection
Based on the above-mentioned results, miR-200b-3p inhibited the PA-mediated apoptosis of LBP-treated MC3T3-E1 cells. According to our results, miR-200b-3p inhibitors could upregulate MC3T3-E1 cell viability under PA treatment and inhibit PA-induced cell apoptosis (Figs 6a-6c). Next, the function of miR-200b-3p inhibitors to some extent increased MC3T3-E1 cell viability and decreased their apoptosis under PA treatment with 400 μg/mL of LBP. In addition, miR-200b-3p inhibition decreased PA-medicated PPARγ expression and increased PA-medicated Chrdl1 expression (Fig. 6d). Meanwhile, the function of miR-200b-3p inhibitors decreasing PPARγ protein level and increasing Chrdl1 protein level of MC3T3-E1 cells treated with PA was reversed to some extent by treatment with 400 μg/mL of LBP.
inhibitors/siRNA-Chrdl1 co-transfection in MC3T3-E1 cells, LBP effects on increasing the viability of MC3T3-E1 cells but reducing their apoptosis were abolished partly within PA-treated cells (Figs 7b-7d). Furthermore, LBP inhibited PPARγ protein level and increased Chrdl1 protein level in MC3T3-E1 cells treatment with PA by inhibiting miR-200b-3p expression. However, the function of LBP regulating PPARγ and Chrdl1 protein levels in MC3T3-E1 cells treatment with PA were abolished by co-transfection with miR-200b-3p inhibitors and siR-NA-Chrdl1 (Fig. 7e).

Discussion
The obesity-induced osteoporosis has always been our research focus. LBP can improve obesity and increase BMD, and its clinical effects have been approved by evidence-based medicine (22). However, similar to many Chinese patent medicines, its precise effect has not been recognized by modern medicine. This would provide new thinking and method for the future clinical treatment and new drug development if its major action pathway could be determined at molecular level. Our research group had previously used PA to treat osteoblasts so as to simulate the cellular environment of obesity-induced osteoporosis, and we discovered that LBP could suppress endoplasmic reticulum (ER) stress to reduce the PA-induced osteoblast apoptosis (11). Nonetheless, this was only the preliminary exploration. More intensive studies are needed to examine whether LBP could alleviate obesity to improve osteoporosis. We utilized bioinformatics to analyze GSE37676 chip data and confirmed that Chrdl1 was the key gene of the most significant module in (e) RT-PCR assay was used to detect miR-200b-3p mRNA level; (f) RT-PCR assay was used to detect Chrdl1 mRNA level; (g, h) western blot assay was used to detect Chrdl1 protein level; (i) Chrdl1 interacted with miR-200b-3p, and Chrdl1-mt and Chrdl1-wt binding sites were displayed by TargetScan database; and (j) luciferase report system was used to detect the targeting miR-200b-3p with Chrdl1. GAPDH was used as a load control. Data are presented as mean ± SD. * P < 0.05 versus blank group/or mimics negative control group, # P < 0.05 versus PA200 group or inhibitors negative control group, and ^P < 0.05 versus LBP100 + PA200 group. the PPI network, which was markedly upregulated after AA treatment. The bone morphogenetic protein (BMP) antagonist Chrdl1 is the target of miR-200a family and involved in osteoclast differentiation (23). Chrdl1 can hinder the BMP-mediated enhanced invasion and migration of breast cancer cells (24). Hypermethylation of Chrdl1 promoter induces gastric cell proliferation (25). Based on the above results, we predicted that Chrdl1 can have a vital role during the apoptosis of osteoblasts. Consequently, the Co database was used to predict the Chrdl1 co-expression network. Intriguingly, Chrdl1 was co-expressed with PPARγ. PPARγ is an important regulatory (d) western blot assay was used to detect PPARγ and Chrdl1 protein levels. GAPDH was used as a load control. Data are presented as mean ± SD. * P < 0.05 versus mimics negative control group, # P < 0.05 versus PA200 group, and ^P < 0.05 versus mimics+PA200 group.
factor during the differentiation process of pre-adipocytes, and its activation could accelerate adipocyte differentiation, thus aggravating obesity in patients. According to reports, LBP can reduce PPARγ activity to relieve obesity (9,10). We speculated based on the above research results that Chrdl1 was closely correlated with the obesity-induced osteoporosis. RT-PCR experimental results verified that PA could evidently downregulate Chrdl1 expression depending on the dose; meanwhile, PPARγ mRNA level was notably upregulated with increase in Fig. 6. LBP further inhibited PA-mediated apoptosis in MC3T3-E1 cells transfected with miR-200b-3p inhibitors. After cells transfected with miR-200b-3p, inhibitors were treated with LBP and PA: (a) CCK-8 assay was used to detect cell viability; (b,c) flow cytometry was used to detect cell apoptosis; (d) western blot assay was used to detect PPARγ and Chrdl1 protein levels. GAPDH was used as a load control. Data are presented as the mean ± SD. * P < 0.05 versus inhibitors negative control group, # P < 0.05 versus PA200 group, and ^P < 0.05 versus inhibitors+PA200 group.
PA concentration, revealing that the PA-induced osteoblast apoptosis could be achieved through regulating the expression of Chrdl1 and PPARγ.
As it is well known, miRNAs can specifically regulate one or more target genes involved in genesis and development of numerous diseases and participate in regulating cell functions. Different miRNAs have exerted vital roles in obesity and osteoporosis (12,14,16,20). Chrdl1 has been recognized to be the miR-200b-3p target gene using the TargetScan database. According to our findings, miR-200b-3p was upregulated among MC3T3-E1 cells with increasing PA concentration. Next, after MC3T3-E1 cells were incurred to pre-treatment with 100 or 400 μg/mL of LBP, cell viability increased significantly among PA-treated MC3T3-E1 cells. Based on our results, LBP significantly increased PA-mediated Fig. 7. LBP affected PA-mediated apoptosis by regulating miR-200b-3p/Chrdl1/PPARγ. (a) Western blot and RT-PCR assays were used to detect the protein and mRNA levels of Chrdl1 in cells transfected with siChrdl1 after cells were transfected with inhibitors and siChrdl1 was treated with LBP and PA; (b) CCK-8 assay was used to detect cell viability; (c,d) flow cytometry was used to detect cell apoptosis; (e) western blot assay was used to detect PPARγ and Chrdl1 protein levels. GAPDH was used as a load control. Data are presented as mean ± SD. * P < 0.05 versus inhibitors control+siNC group, # P < 0.05 versus PA200 group, ^P < 0.05 versus inhibitors+PA200 group, and & P < 0.05 versus LBP400+inhibitors+PA200 group. decrease in Chrdl1 mRNA levels and inhibited PA-mediated increase in miR-200b-3p and PPARγ mRNA levels. It suggested that LBP could regulate the miR-200b-3p/ Chrdl1/PPARγ axis to exert a series of functions in PAtreated cells. To further verify the above-mentioned results, CCK-8, flow cytometry, and western blotting were used to detect viability, apoptosis as well as PPARγ and Chrdl1 expression in miR-200b-3p-transfected cells.
Overexpression of miR-200b-3p aggravated PA-induced apoptosis and further increased PA-mediated increase in PPARγ protein level and decrease in PA-mediated inhibition in Chrdl1 protein level. Meanwhile, the function of PA affected cell proliferation, apoptosis, and the expression of PPARγ and Chrdl1 proteins by LBP treatment. miR-200b-3p inhibition had opposite effect relative to that of miR-200b-3p overexpression in MC3T3-E1 cells co-treatment of PA with LBP.
These results proved that LBP could downregulate miR-200b-3p expression and suppress the PA-induced cell apoptosis; meanwhile, it could promote Chrdl1 expression and indirectly suppress PPARγ expression. When siChrdl1 and inhibitors were co-transfected into cells, the suppression of LBP on PA-induced cell apoptosis was also suppressed partially. This was because the PA-induced cell apoptosis was not only realized through upregulating miR-200b-3p but also by other molecules to exert pro-apoptotic effect. Nonetheless, it is certain that suppressing Chrdl1 and miR-200b-3p expression at the same time could evidently weaken the anti-apoptotic effect of LBP on PA-treated cells. Meanwhile, we also discovered that PPARγ expression in cells co-transfected with si-Chrdl1 and inhibitors displayed no significant difference compared with that in PA-treated cells.

Conclusion
PA can partially upregulate miR-200b-3p and PPARγ expressions to specifically suppress Chrdl1 expression and promote osteoblast apoptosis. LBP can suppress the PA-induced osteoblast apoptosis through the miR-200b-3p/ Chrdl1/PPARγ axis, which has also verified that LBP shows certain protection on the obesity-induced osteoporosis.