Baicalein ameliorates osteoporosis via AKT/FOXO1 signaling

In this study, we used bioinformatics and an in vitro cellular model of glucocorticoid-induced osteoporosis to investigate mechanisms underlying the beneficial effects of baicalein (BN) against osteoporosis. STITCH database analysis revealed 30 BN-targeted genes, including AKT1, CCND1, MTOR, and PTEN. Functional enrichment analysis demonstrated that BN-targeted genes were enriched in 49 Kyoto Encyclopedia of Genes and Genomes (KEGG) pathways. MIRWALK2.0 database analysis identified 110 enriched KEGG pathways related to osteoporosis. A Venn diagram demonstrated that 26 KEGG pathways were common between osteoporosis and BN-targeted genes. The top 5 common KEGG pathways were prostate cancer, bladder cancer, glioma, pathways in cancer, and melanoma. BN-targeted genes in the top 5 shared KEGG pathways were involved in PI3K-AKT, MAPK, p53, ErbB, and mTOR signaling pathways. In addition, glucocorticoid-induced osteoporosis in MC3T3-E1 cells was partially reversed by BN through inhibition of AKT, which, by upregulating FOXO1, enhanced expression of bone turnover markers (ALP, OCN, Runx2, and Col 1) and extracellular matrix mineralization. These findings demonstrate that BN suppresses osteoporosis via an AKT/FOXO1 signaling pathway.


INTRODUCTION
Osteoporosis is a metabolic bone disease commonly diagnosed in the elderly [1]. Chronic osteoporosis is manifest by progressive brittleness of bones and greater incidence of non-stress fractures because of excessive bone resorption, low bone mineral density, and deterioration of bone micro-structure [2,3]. Several genetic and environmental factors, such as Vitamin D deficiency and low estrogen levels, contribute to the progression of osteoporosis [4]. However, molecular mechanisms underlying osteoporosis remain unclear.
Baicalein (BN) is one of the most abundant flavonoid in Scutellaria baicalensis, which is widely used in Chinese herbal medicine for various ailments [5]. For example, BN reduces cerebrovascular resistance, improves cerebral blood circulation, and prevents platelet agglutination [5]. BN is clinically used for treatment of paralysis in patients with cerebrovascular disease [6]. The beneficial effects of BN on osteoporosis have also been reported. For example, BN acts as a lipoxygenase inhibitor and increases bone formation in osteoporosis model mice [7]. Moreover, BN promotes osteoblastic differentiation of MC3T3-E1 cells via protein kinases and transcription factors such as P-4E/BP1 and P-S6K1 [8]. However, the molecular mechanisms underlying the therapeutic effects of BN on osteoporosis are not clear.
Bioinformatics analysis has been widely used to unravel molecular and cellular mechanisms of several human AGING diseases [9][10][11]. Therefore, in this study, we performed comprehensive bioinformatics analysis and in vitro experiments in the cellular model of glucocorticoidinduced osteoporosis (GIO) to identify molecular mechanisms underlying the therapeutic effects of BN in osteoporosis.

Identification of shared KEGG pathways between BN-targeted genes and osteoporosis
We identified 49 enriched KEGG pathways (p <0.05) by performing functional enrichment analysis of baicalein-targeted genes using DAVID. Furthermore, we identified 110 enriched KEGG pathways associated with human osteoporosis using the miRWalk2.0 database. We further identified 26 common KEGG pathways between osteoporosis and baicalein-targeted genes using the Venn diagram ( Figure 3). Among these, prostate cancer, bladder cancer, glioma, pathway in cancer, and melanoma were the top five common KEGG pathways ( Table 1).

Identification of baicalein-targeted hub genes
Functional enrichment analysis of BN-targeted genes is shown in Figure 4A. Among the 30 BN-targeted genes,   common KEGG pathways by intersecting those related to baicalein-target genes (n=49 KEGG pathways) and osteoporosis (n=110 KEGG pathways).  Figure 4B).

AGING
The FDR values, gene counts, and rich factors are shown in Figure 4C. The KEGG pathway, hsa05200: pathways in cancer, showed the highest FDR value, rich factor, and gene numbers ( Figure 4C).

Analysis of KEGG pathways related to BN-targeted genes
The top 5 baicalein-targeted KEGG pathways were associated with apoptosis inhibition, cell cycle progression, impaired G1 and G2 cell cycle arrest, genomic instability, tumor growth, cell growth and proliferation, angiogenesis, G1/S cell cycle progression, uncontrolled proliferation, and increased survival ( Figure 5). This suggested that baicalein regulated osteogenesis through PI3K-AKT, MAPK, p53, ErbB, and mTOR signaling pathways.

BN partially reverses GIO through suppression of AKT expression
We then analyzed AKT phosphorylation levels in MC3T3-E1 cells treated with different concentrations of baicalein (1 μM, 10 μM, and 100 μM) by ELISA. BN treatment significantly reduced p-AKT levels in MC3T3-E1 cells; p-AKT levels were significantly AGING lower in the 10 μM BN-treatment group compared to the 1 μM and 100 μM BN-treatment groups ( Figure  6A). QRT-PCR analysis results showed that the GIO treatment significantly decreased the expression levels of bone turnover markers such as ALP, OCN, Runx2, and Col 1 in MC3T3-E1 cells, but these effects were partially reversed by BN ( Figure 6B-6E). Furthermore, ALP staining assay results showed that BN partially rescued impaired extracellular matrix mineralization in GIO-treated MC3T3-E1 cells ( Figure 6F, 6G).  AGING BN regulates osteogenesis via AKT/FOXO1 signaling QRT-PCR analysis showed that expression of FOXO1 was significantly increased in BN-treated MC3T3-E1 cells; FOXO1 levels were highest in the 10 μM BN-treatment group ( Figure 7A). Furthermore, qRT-PCR analysis confirmed that AKT levels were significantly reduced in AKT-siRNA-transfected MC3T3-E1 cells compared to the corresponding controls ( Figure 7B). AKT silencing significantly increased FOXO1 levels in MC3T3-E1 cells ( Figure 7C). QRT-PCR analysis AGING showed that FOXO1 silencing significantly reduced expression levels of the bone turnover markers (ALP, OCN, Runx2, and Col 1) in GIO-treated MC3T3-E1 cells, but these effects were partially reversed by BN treatment ( Figure 7D-7G). ALP staining results showed that BN treatment partially rescued impaired extracellular matrix mineralization in GIO-treated FOXO1silenced MC3T3-E1 cells ( Figure 7H, 7I).

DISCUSSION
Osteoporosis is a chronic metabolic bone disorder observed commonly in aged individuals, and is characterized by low bone mineral density, increased brittleness of bone, damaged bone micro-structure, and significant increase in the number of non-stress fractures [12]. BN is one of the most abundant flavonoids in Scutellaria baicalensis and is widely used in Chinese herbal medicine to treat various human diseases for centuries. [13]. BN promotes osteogenic differentiation and bone formation [14][15][16], but the molecular mechanisms underlying the beneficial effects of BN on osteoporosis are not clear. Pathogenetic mechanisms underlying several human diseases have been discovered through bioinformatics analyses of microarray data and wholegenome sequencing studies [17].
In the present study, we identified 110 KEGG pathways that were associated with differentially expressed genes (DEGs) in human osteoporosis and 49 KEGG pathways associated with BN-targeted genes. PI3K-AKT, MAPK, p53, ErbB, and mTOR signaling pathways were the top five shared KEGG pathways among these two groups.
Previous studies have shown that BN-targeted genes in the PI3K-AKT signaling pathway were associated with osteogenesis [18,19]. Many studies have shown that PI3K-AKT signaling pathway plays a significant role in osteoporosis [20,21]. For an example, a recent study demonstrated that inhibition of PI3K-AKT signaling pathway delayed osteoporosis progression in postmenopausal women by suppressing inflammation and formation of osteoclasts [22]. Xiao et al demonstrated that inhibition of miR-148a ameliorated ovariectomyinduced osteoporosis via PI3K-AKT signaling pathway [23]. In the present study, we showed that BN significantly suppressed AKT expression and activation in MC3T3-E1 cells. BN treatment also stimulated the expression of bone turnover markers and enhanced extracellular matrix mineralization in GIO-induced MC3T3-E1 cells. These results suggested that the BN treatment suppresses osteoporosis via AKT.
FOXO family of transcription factors such as FOXO1, FOXO3, FOXO4, and FOXO6 play a crucial role in cellular defense mechanisms against oxidative stress [24]. Liao et al reported that FOXO1 knockdown decreased osteoblast differentiation by suppressing the expression of antioxidant enzymes, manganese superoxide dismutase and catalase, thereby increasing the levels of reactive oxygen species [25]. Feng et al demonstrated that inhibition of PI3K-AKT signaling pathway significantly increased FOXO1 transcriptional activity, which suppressed osteoporosis by enhancing antioxidant mechanisms [26]. In our study, we demonstrated that BN promoted expression of FOXO1 by suppressing AKT. We also observed that BN treatment and AKT silencing improved osteogenic differentiation of GIO-induced MC3T3-E1 cells by increasing FOXO1 expression. Therefore, our results demonstrate that BN treatment suppresses osteoporosis via AKT/FOXO1 signaling pathway. Future studies are required to confirm our findings in vivo and further investigate mechanisms underlying the beneficial effects of BN on osteogenesis.

ELISA
MC3T3-E1 cells were grown for 48 h in serum-free medium. The concentration of phospho-AKT in each group of cells was measured using the ELISA kit according to manufacturer's instructions. The colorimetric values were normalized to total cell numbers in each well. Then, the phospho-AKT levels in each sample were calculated using the standard curve.

ALP staining
MC3T3-E1 cells were washed twice with PBS and then fixed with 10% formalin for 15 minutes. Then, the cells were processed using the ALP color-development kit according to manufacturer's instructions and the color development was performed by incubating cells with BCIP/NBT liquid substrate for 24 hours. Absorbance was measured at 405 nm in a plate reader. The experiments were performed in triplicate.

Identification of BN-target genes and construction of PPI network
We identified baicalein-targeted genes and constructed the protein-protein interaction (PPI) network between them using the Search Tool for Interacting Chemicals (STITCH) database [27]. We analyzed the PPI network of baicalein-targeted genes using the Cytoscape 3.7.2 version software [28] and measured the degree, betweenness, and closeness of each gene in the network. The hub genes were identified based on the degree analysis using the cytoHubba plugin. Cytoscape plug-in ClueGO was used to identify enriched KEGG pathways representing the baicalein-targeted genes and the relationship between them. The KEGG pathway enrichment results were visualized using the ggplot2.7 package.

Identification of shared KEGG pathways between BN-targeted genes and osteoporosis
MiRWalk2.0 software [29] was used to retrieve human osteoporosis-related KEGG pathways using p<0.05 as the threshold parameter. VennDiagram (http://www. ehbio.com/ImageGP/index.php/Home/Index/index.html) was then used to identify shared KEGG pathways between human osteoporosis and baicalein-targeted genes.

Identification of hub genes
GOplot R package [30] was used to visualize functional enrichment analysis of KEGG pathways enriched with baicalein-targeted genes and identified hub genes from among the top five shared KEGG pathways between baicalein-targeted genes and osteoporosis.

Identification of KEGG pathways related to BNtargeted genes
We selected the top 5 shared KEGG pathways with the smallest p values and used ScienceSlides to schematically represent KEGG pathways related to baicalein-targeted genes, the hub genes and their related mechanisms of action.

Statistical analysis
The data are represented as means ± SD. Prism 8.0 software was used for all statistical analyses. One-way analysis of variance with Tukey's post hoc test was used to compare three or more groups of data. Twotailed Student's test was used to compare data between two groups. P < 0.05 was considered statistically significant.

AUTHOR CONTRIBUTIONS
PC conceived and designed the study; PC, YL and ZL supervised the study; PC and YL performed bioinformatics analysis and experiments; ZY, XW and XZ analyzed the data; XZ and ZL provided advice and technical assistance; PC, YL and ZL wrote the manuscript. All authors approved the final manuscript.

CONFLICTS OF INTEREST
The authors declare that they have no conflicts of interest.