Discovery of anthraquinones as potent Notum inhibitors for treating osteoporosis by integrating biochemical, phytochemical, computational, and experimental assays

Abstract

Osteoporosis, a severe systemic skeletal disorder characterized by decreased bone mineral density, leads to increased risks of bone fragility and fracture. Although some herbal medicines (HMs) are clinically used for treating osteoporosis, the crucial anti-osteoporotic constituents and their mechanisms have not been well-elucidated. Notum, a negative regulator of Wnt/β-catenin signaling, has been validated as a druggable target for enhancing cortical bone thickness and alleviating osteoporosis. Herein, we showcase an efficient strategy for uncovering the key anti-Notum constituents from HMs via integrating biochemical, phytochemical, computational, and cellular assays. Following screening the anti-Notum potentials of HMs, Polygonum multiflorum Thunb. (PM), a commonly used anti-osteoporosis herb, showed potent and competitive inhibition against Notum. Phytochemical profiling coupling with docking-based virtual screening suggested that three anthraquinones, including rhein, emodin, and chrysophanol, showed high binding-potency towards Notum. Biochemical assays validated that three anthraquinones were strong competitive inhibitors of Notum, while rhein was the most potent one (IC₅₀ = 9.98 nM). Cellular investigations demonstrated that rhein markedly promoted osteoblast differentiation in dexamethasone-challenged MC3T3-E1 osteoblasts, while RNA sequencing showed that rhein remarkably regulated Wnt signaling-related and osteogenic differentiation-related genes. In vivo tests showed that rhein displayed favorable safety profiles in healthy mice and this agent significantly elevated bone mineral density, and augmented trabecula and cortical bone thickness in dexamethasone-induced osteoporotic mice. Collectively, this study showcases an efficient strategy for uncovering the key anti-Notum constituents from HMs, while rhein was identified as a naturally occurring Notum inhibitor that shows favorable safety profiles and impressive anti-osteoporosis effects.

Graphical abstract

Highlights

• •A high-efficient platform for discovering anti-Notum agents from herbs was constructed.
• •Polygonum multiflorum Thunb. (PM) was found with potent anti-Notum effects.
• •Three anthraquinones were identified as potent anti-Notum constituents in PM.
• •Rhein significantly promoted osteoblast differentiation via activating Wnt signaling.
• •Rhein showed favorable safety profiles and impressive anti-osteoporosis effects in vivo.

1. Introduction

Osteoporosis, a severe systemic skeletal disorder characterized by decreased bone mineral density (BMD), leads to a progressively increased risk of bone fractures and fragility [1,2]. As a common disorder in both older women and men, osteoporosis causes approximately 8.9 million fractures annually. Among these, hip fractures have a mortality rate as high as 20% [3]. Patients with osteoporosis often require long-term administration with specific bone medications, including antiresorptive and osteoanabolic agents [4,5]. However, these agents could also cause some undesirable side effects. For instance, the commonly used anti-resorptive medications (such as denosumab) may lead to several disorders including osteonecrosis of the jaw, gastrointestinal irritation, and ocular inflammation [6,7]. Long-term use of anabolic agents (such as teriparatide) can cause issues like constipation, depression, nausea, headache, stomach pain and palpitations, and even increase the risk of osteosarcoma [8,9]. Thus, there is a growing and urgent need to find more efficacious anti-osteoporosis drugs with improved safety profiles.

In clinical settings, some herbal medicines (HMs), such as Polygonum multiflorum Thunb. (He Shou Wu), Salvia miltiorrhiza Bge. (Dan Shen), Epimedium brevicornu Maxim. (Yin Yang Huo), and Rehmannia glutinosa Libosch. (Di Huang), have been frequently used for treating osteoporosis for a long history and shown beneficial potentials to the patients with osteoporosis [[10], [11], [12], [13], [14], [15], [16], [17]]. It is well-known that discovering the bioactive compounds from HMs or other natural resources represents a feasible way for discovering lead compounds and drug candidates [18,19]. However, unlike Western drugs, HMs are complex mixtures that contain a variety of chemically diverse compounds, making it challenging to efficiently identify the key anti-osteoporosis components from HMs for further pharmacological and mechanistic studies. Thus, the high-efficient strategies for uncovering the key anti-osteoporosis agents from HMs are always desirable.

Target-based drug screening has been widely used for high-throughput screening and charactering the bioactive substances from HMs and natural sources [20]. Over the past few decades, several proteins, including nuclear factor erythroid 2-related factor 2 (Nrf2) [21], peroxisome proliferator-activated receptors alpha/beta (PPARα/β) [22,23], Runt-related transcription factor 2 (RUNX2) [24], peroxisome proliferator-activated receptor gamma (PPARγ) [25], and Notum [26], have been identified and validated as druggable targets for treating osteoporosis [27,28]. Among these, Notum has emerged as a promising therapeutic target for promoting osteoblast differentiation and treating osteoporosis. Notum, a secreted carboxylesterase, negatively regulates the Wnt signaling pathway by removing the essential palmitoleate moiety from Wnt proteins [29]. The Wnt/β-catenin signaling is critical for osteoblast differentiation, bone development, homeostasis, and remodeling [30]. Although Wnt proteins are undruggable targets [31], blocking Notum offers a feasible strategy to indirectly activate Wnt signaling, thereby promoting osteoblast development. Notum is predominantly found in osteocytes and osteoblasts in cortical bone, and its expression is upregulated by glucocorticoids, leading to reduced cortical bone mass and osteoporosis [32]. Recent studies have demonstrated that a range of small molecules including natural products could potently inhibit Notum [[33], [34], [35]], while blocking Notum has been validated as an emerging therapeutic approach for stimulating endocortical bone formation, increasing cortical bone thickness of intact and ovariectomized rodents [36]. These findings suggest that Notum inhibitor therapies are promising strategies for enhancing cortical bone and treating osteoporosis.

Herein, a highly efficient strategy was adapted to rapidly identify potent anti-Notum constituents from HMs, while their anti-osteoporosis effects were comprehensively investigated both in vitro and in vivo (Fig. 1). In brief, a fluorescence-based high-throughput screening (HTS) assay was firstly utilized to test the anti-Notum potentials of the frequently used HMs. Among all tested HMs, the crude extract of Polygonum multiflorum Thunb. (PM) showed the most potent anti-Notum effect. It was also found that PM dose-dependently inhibited Notum in a competitive inhibition manner, suggesting that PM contained the naturally occurring competitive inhibitors of Notum. After that, phytochemical profiling-based LC-MS coupling with docking-based virtual screening was used to seek the PM constituents with high binding affinities towards the catalytic cavity of Notum. The phytochemicals in PM with high binding-potency towards Notum were then selected for further experimental validation. Further biochemical assays showed that three anthraquinones in PM were potent competitive inhibitors of Notum, while rhein was the most potent one (IC₅₀ = 9.98 nM). These findings prompted us to further investigate the anti-osteoporosis effects of rhein both in vitro and in vivo, as well as to validate the regulatory effects of this newly identified naturally occurring Notum inhibitor on Wnt/β-catenin signaling.

Fig. 1.

Methodological framework for identifying Notum inhibitors from herbal medicines and validating their anti-osteoporotic effects. Dex: dexamethasone; PM: Polygonum multiflorum Thunb.; OPTS: 8-octanoyloxypyrene-1,3,6-trisulfonate; H&E: hematoxylin and eosin.

2. Experimental and methods

2.1. Chemicals and instruments

Notum was expressed and purified according to the previously described protocol [37]. Citric acid, K₂HPO₄ and 8-octanoyloxypyrene-1,3,6-trisulfonate (OPTS) were purchased from Sigma-Aldrich (St. Louis, MO, USA). All tested compounds including rhein, emodin, chrysophanol and isoalantolactone, with high purity (HPLC > 98%), were purchased from Shanghai Yuanye Bio-Technology Co., Ltd. (Shanghai, China). The extract of all tested HMs was provided by Shanghai Standard Technology Co., Ltd. (Shanghai, China). Tris was gained from Sigma-Aldrich (St. Louis, MO, USA). MS-grade optima® acetonitrile (ACN) was purchased from Fisher Scientific (Harrisburg, PA, USA). HPLC-grade formic acid (FA; 98%) was obtained from Aladdin (Shanghai, China). Fetal bovine serum, penicillin-streptomycin (10,000 U/mL), and Dulbecco's modified Eagle medium (DMEM) cell culture medium were acquired from Gibco (Grand Island, NY, USA). Culture-treated plates were purchased from Corning (Shanghai, China). Immun-Blot® polyvinylidene fluoride (PVDF) membrane and ChemiDoc MP Imaging System were obtained from Bio-Rad Laboratories (Hercules, CA, USA). HPLC-grade methanol was used to prepare the stock solution of all tested compounds and crude extract of HM. The fluorescent signals were detected by SpectraMax® iD3 (Vienna, Austria). Dexamethasone (Dex) were purchased from Shanghai Yien Chemical Technology Co., Ltd. (Shanghai, China).

2.2. Notum inhibition assay

The details of Notum inhibition assay have been previously reported [38]. In brief, 200 μL incubation system containing o-phthalaldehyde (OPTS; 10 μM), Tris-HCl (0.1 M, pH = 7.4) and inhibitor (2 μL) were preincubated at 37 °C for 3 min. Time-dependent inhibition assays for each Notum inhibitor was conducted by performing series of parallel incubations with different pre-incubation times (3, and 33 min), prior to adding the substrate OPTS [39]. After initiating the hydrolytic reaction, the fluorescence signals of the hydrolytic product of OPTS were recorded under excitation at 403 nm, while the emission wavelength was set at 510 nm. After removing the background signals (the negative incubation samples without enzyme), the nonlinear regression calculated the IC₅₀ values of PM and its constituents. The experimental details of determining the inhibition kinetic analyses have also been previously reported [40].

2.3. Phytochemical profiling of PM

Phytochemical analysis of PM was performed using a UHPLC-QExactive Orbitrap system (Thermo Fisher Scientific, Grand Island, NY, USA) equipped with a mass spectrometer and electrospray ionization (ESI) source. Chromatographic separation was achieved on an ACQUITY UPLC BEH C₁₈ column (2.1 mm × 100 mm, 1.7 μm; Waters, Milford, MA, USA) with gradient elution using 0.1% formic acid in water and methanol as mobile phases, at a flow rate of 0.3 mL/min. The gradient elution program was as follows: 0–4 min, 4% methanol; 4–10 min, 4%–12% methanol; 10–30 min, 12%–70% methanol; 30–35 min, 70% methanol; 35–38 min, 70%–95% methanol; 38–42 min, 95% methanol; followed by column re-equilibration at 4% methanol from 42 to 45 min. Mass spectrometry was conducted in both positive and negative ion modes, with compounds recorded over a mass-to-charge ratio (m/z) range of 100–1500. The spray voltages were set to 3.8 kV for positive ions and −2.5 kV for negative ions. Sheath gas flow rates (nitrogen) were set at 35 and 40 arb for positive and negative ions, respectively, and the capillary temperature was maintained at 320 °C for both modes. PM was diluted to 100 mg/mL with 50% methanol and sonicated for 30 min to ensure complete dissolution, with an injection volume of 2 μL. The acquired data were processed and analyzed using Thermo Xcalibur software (version 3.0).

2.4. Molecular docking simulations

The three dimensional (3D) structure of the Notum was obtained from the PDB database (PDB ID: 6zyf). Molecular docking was conducted using Autodock software. Visualization was carried out using Pymol and Discovery Studio software. Isoalantolactone was used as a positive inhibitor of Notum [41].

2.5. Cell culture

The MC3T3-E1 cell line was obtained from Cell Bank, Chinese Academy of Science (Shanghai, China). The cells were cultured in DMEM media supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin solution in a humidified culture chamber at 37 °C with 95% air and 5% CO₂.

2.6. Alkaline phosphatase (ALP) staining and alizarin red S (ARS) staining

MC3T3-E1 cells were incubated in growth media in 96-well plates at 2 × 10³ cells/well density for 24 h. The culture medium was then switched to an inducible medium (adding 10 mM β-glycerophosphate, and 50 mg/mL ascorbic acid), and the cells were cultured with simultaneous addition of Dex and the target compound for 7 days. The cells were fixed in 4% paraformaldehyde at room temperature for 15 min, washed with phosphate-buffered saline (PBS), and stained with the ALP staining kit (Beyotime, Shanghai, China). The cells were cultured for 21 consecutive days, and osteogenic differentiation was measured using ARS staining (Beyotime, Shanghai, China). After rinsing with PBS, the cells were observed via microscope.

2.7. RNA sequencing

MC3T3-E1 cells were treated with Dex, followed by the addition of rhein. After cultured in parallel for 7 consecutive days, the total RNA of each group, including control (dimethyl sulfoxide (DMSO), St. Louis, MI, USA), Dex and rhein, was extracted one by one. RNA sequencing was performed on an Illumina HiSeq 2500 sequencing plantform (Illumina, San Diego, CA,USA) by Shanghai OE Biotech Co., Ltd. (Shanghai, China). A volcano plot display was used to show differentially expressed genes. Subsequent Gene Ontology (GO) and Kyoto Encyclopedia of Genes and Genomes (KEGG) pathway analyses were performed to observe the molecular mechanisms and provide data to support further research.

2.8. Quantitative polymerase chain reaction (qPCR)

Following Dex intervention, MC3T3-E1 cells were treated with a compound for 5 days. Following two washes with PBS, RNA extraction was performed in accordance with the instructions provided in the kit (RK21203, ABclonal, Wuhan, China). Reverse transcription (5X ABScript III RT Mix, RM21478; 4XABScript Neo RT Master Mix, RM21486; 20X gDNA Remover Mix, RM21483; ABclonal, Wuhan, China) was used to synthesize cDNA, amplified using qPCR to detect mRNA (Table 1). The expression levels were normalized to glyceraldehyde-3-phosphate dehydrogenase (GAPDH) and quantified using the 2^−ΔΔC_T method.

Table 1.

The sequences of all primers.

Primers	Forward	Reverse
GSK3β	TGGCAGCAAGGTAACCACAG	CGGTTCTTAAATCGCTTGTCCTG
β-Catenin	ATGGAGCCGGACAGAAAAGC	CTTGCCACTCAGGGAAGGA
RUNX2	GCCGGGAATGA TGAGAACTA	GGTGAAACTCTTGCCTCGTC
OPN	CTGGCAGCTCAGAGGAGAAG	CAGCATTCTGTGGCGCAAG
GAPDH	GGAGATTGTTGCCATCAACGA	GAAGACACCAGTAGACTCCACGACA

GSK3β: glycogen synthase kinase 3 beta; RUNX2: Runt-related transcription factor 2; OPN: osteopontin; GAPDH: glyceraldehyde-3-phosphate dehydrogenase.

2.9. Western blotting

The radioimmunoprecipitation assay buffer (RIPA) buffer was used to extract protein lysates from MC3T3-E1 cells, which were then loaded equally onto a 10% polyacrylamide gel. The following antibodies were used for first antibody incubation, such as anti-glycogen synthase kinase 3 beta (GSK3β) antibody (22104-1-AP, 1:1000, Proteintech, Chicago, IL, USA), anti-β-catenin antibody (D10A8, 1:1000, Cell signaling, Danvers, MA, USA), anti-RUNX2 antibody (A11753, 1:500, ABclonal, Woburn, MA, USA), anti-osteopontin (OPN) antibody (A21084, 1:1000, ABclonal, Woburn, MA, USA), and anti-GAPDH antibody (GB15004-100, 1:4000, Servicebio, Wuhan, China). Subsequently, cultivation was conducted with an horseradish peroxidase (HRP)-conjugated secondary antibody (anti-rabbit, 7074S, 1:4000; anti-mouse, 7076P2, 1:4000, Cell signaling, Danvers, MA, USA).

2.10. Safety assessment of rhein in mice

Sixteen C57BL/6J mice (eight males and eight females), obtained from Shanghai Slake Experimental Animal Co., Ltd. (Shanghai, China), weighed approximately 18–20 g and of specific pathogen-free (SPF) grade, were used in this study. This study was conducted in accordance in compliance with the animal ethical guidelines and approved by the Shanghai Institute for Food and Drug Control (Approval number: SIFDC-IACUC24013). Two experimental groups were established: a negative control group comprising eight mice (four males and four females) receiving an equal volume of 0.5% CMC-Na solvent and a rhein treatment group comprising eight mice (four males and four females) receiving 200 mg/kg orally once daily for 14 days. Upon completion of the experiment, serum and tissues, including heart, liver, spleen, lung, and kidney, were collected from all mice for safety assessments.

2.11. In vivo anti-osteoporosis effects of rhein

The male C57BL/6J mice were obtained from Shanghai Slac Laboratory Animal Co. Ltd. (Shanghai, China). This study was conducted in accordance with the ethical standards set forth by the Animal Ethics Committee of Shanghai University of Traditional Chinese Medicine (Approval number: PZSHUTCM2404280002). The mice were maintained under specific pathogen-free conditions and standard laboratory conditions. The environmental conditions were maintained at 25 ± 2 °C, 50% ± 5% humidity, and a 12 h/12 h light/dark cycle. Male mice aged 6 weeks were randomly assigned to 5 groups (n = 6). Except for the control group, all other groups received Dex at a 2.5 mg/kg dose via intraperitoneal injection (i.p.), administered 5 days per week. The methods of Dex-induced osteoporosis are based on the findings of previous studies [42,43]. The positive control group received daily parathyroid hormone 1–34 (PTH) (HY-P1252, MCE, Monmouth Junction, NJ, USA) (i.p.). The selection of a safe dose of rhein for drug treatment was informed by previous research [44] and acute toxicity test results. The treatment groups were administered rhein at low (40 mg/kg) and high (80 mg/kg) doses via oral gavage (i.g.) once daily, 5 weeks. At the end of the animal experiment, serum, femurs, and tibias were obtained from the mice for subsequent analysis.

2.12. Determination of the levels of bone turnover serum markers

The levels of bone metabolism index in mouse serum, including alkaline phosphatase (ALP), osteocalcin/bone gla protein (OT/BGT), osteoprotegerin (OPG), and collagen type I alpha 1 chain (Col1A1), were measured using commercial enzyme-linked immunosorbent assay (ELISA) kits (ABclonal Biotechnology Co., Ltd., Wuhan, China). The calcium (Ca) (D799342-0100) and phosphorus (P) (D799344-0100) kits (Sangon Biotech, Shanghai, China) were utilized to assess the calcium and phosphorus levels in the blood samples of each group of mice.

2.13. Micro-CT analysis of the femur

All extracted right femurs of the mice were fixed in a 4% paraformaldehyde solution. Subsequently, the specimens were scanned using the Skyscan1172 (Bruker, Ettlingen, Baden-Württemberg, Germany) to observe the trabecular and cortical bone. The scanning parameters were listed as follows: a voltage of 80 kVp, a current of 124 μA, a slice thickness of 9.91 μm, and a threshold range of 65–225. The distal femur growth plate layers 100–200 were defined as the distal femur, while layers 300–400 were defined as the mid-femur. The measurement results were then analyzed accordingly. The results were analyzed using MicroView software.

2.14. Hematoxylin and eosin (H&E) staining

The samples were dipped in a 10% ethylenediaminetetraacetic acid solution for continuous decalcification using a multifunctional microwave processor. The paraffin-embedded samples were cut into 4 μm thick sections, followed by deparaffinization and rehydration. The sections were subjected to H&E staining for 10 and 2 min, respectively, after which they were washed with distilled water and dehydrated. Once the H&E staining was completed, the sections were mounted with neutral balsam and examined under a microscope.

2.15. Immunofluorescence

The sections were dewaxed with xylene and ethanol, and antigen retrieval was performed with an antigen retrieval solution. Subsequently, a 30 min incubation period was conducted with the blocking solution. 4‘,6-diamidino-2-phenylindole (DAPI; G1407) was purchased by Wuhan Servicebio Technology Co., Ltd. (Wuhan, China). A 1:100 dilution of the anti-Notum antibody (ab106448, Abcam, Cambridge, MA, UK) and the anti-OPN antibody (a21084, ABclonal, Wuhan, China) was applied and incubated at 4 °C overnight. Subsequently, the membrane was incubated with the secondary antibody, 488-conjugated goat anti-mouse IgG (H + L) (AS037, ABclonal, Wuhan, China) and 594-conjugated goat anti-rabbit IgG (H + L) (AS039, ABclonal, Wuhan, China) at room temperature for 1 h, following a washing step. Subsequently, the membrane was stained with a fluorescent dye, and an anti-fluorescence extraction and sealing medium were used for sealing. The observation was conducted using a fluorescence microscope.

2.16. Double labeling for measuring bone-formation activity

Calcein (5 mg/kg, two weeks before execution) was administered via intraperitoneal injection to the mice, while xylenol orange (5 mg/kg, 24 h before execution) was intraperitoneally injected. Following collection, tibial sections were embedded in light-sensitive resin (EXAKT Technologies, Norderstedt, Germany), cured for 13.5 h, and then sectioned into 40 μm thick slices using the EXAKT system for preparation. The bone-formation activity was assessed using double labeling techniques.

2.17. Statistical analysis

Data were expressed as the mean ± standard deviation. Student's t-test was employed for group comparisons, with statistical significance set at P < 0.05. All statistical analyses were carried out using Prism software.

3. Result

3.1. Discovery of the anti-Notum herbs

Firstly, the anti-Notum effects of 103 HMs were screened under identical conditions at a high dose (100 μg/mL). Fig. 2A shows that most HMs exhibited weak or no inhibitory effects on Notum, while a few of HMs showed strong anti-Notum potentials. Among all tested HMs, PM exhibited the most significant inhibitory effect, showing the residual activity of 10.31%. After then, the dose-response inhibition curve of PM against Notum was plotted by using increasing concentrations of PM. As shown in Fig. 2B, PM strongly inhibited Notum in a dose-dependent manner, with an IC₅₀ value of 34.97 μg/mL. Further investigations showed that the inhibition of PM against Notum was not time-dependent, suggesting that PM constituents act as non-covalent inhibitors of Notum (Fig. S1A). Further inhibition kinetic analyses showed that PM extract strongly inhibited Notum in a competitive inhibition manner, showing the K_i value of 14.79 μg/mL (Figs. 2C and S1B).

Fig. 2.

Screening of the anti-Notum potentials of herbal products (n = 3). (A) Preliminary screening of 103 herbal medicines extracts (100 μg/mL) for anti-Notum activity. (B) IC₅₀ inhibition curves of Polygonum multiflorum Thunb. anti-Notum activity under incubation. (C) The Lineweaver-Burk plots of Polygonum multiflorum Thunb. OPTS: 8-octanoyloxypyrene-1,3,6-trisulfonate.

3.2. Global phytochemical profiling of PM

Next, the phytochemicals in PM are identified using UHPLC-QExactive Orbitrap high-resolution mass spectrometry (HRMS). The total ion chromatogram (TIC) of PM is shown in Fig. 3, in which a total of 64 compounds were tentatively identified. The retention time (min), ion mode, measured mass (Da), calculated mass (Da), error (ppm), and formula identification of each phytochemical were recorded, as shown in Table S1. Almost all PM constituents could be detected under negative ion mode, except for hypaphorine and salsolinol. The phytochemicals identified from PM could be divided into several compound classes, including organic acids, flavonoids, anthraquinones, and others.

Fig. 3.

Total ion flow diagram of Polygonum multiflorum Thunb. ultra performance liquid chromatography coupled with mass spectrometry (UPLC-MS) in negative and positive ion modes.

3.3. Docking-based virtual screening of Notum inhibitors

It has been reported that the catalytic cavity is the major ligand-binding site of Notum, while almost all reported naturally occurring Notum inhibitors are identified as competitive inhibitors [38]. In these cases, docking-based virtual screening of Notum inhibitors from PM constituents were conducted, for seeking the natural products with high binding-potency towards Notum. Specifically, each PM constituent was docked into the catalytic cavity of the 3D structure of Notum (PDB ID: 6zyf) one by one, while a previously reported Notum inhibitor (isoalantolactone) was used as a reference drug. As shown in Fig. 4A, three compounds including rhein, emodin, and chrysophanol, were identified as potential Notum inhibitors (Figs. 4B and C, and Table S2), as these compounds show outstanding binding-affinities towards Notum (minimum binding energy <−10 kcal/mol). These findings encouraged us to further test their anti-Notum effects.

Fig. 4.

Molecular docking results of Polygonum multiflorum Thunb. compounds and inhibitory effects of anthraquinones and isoalantolactone on Notum. (A) Docking scores of 64 compounds of Polygonum multiflorum Thunb. and isoalantolactone with Notum molecules. The color change from orange to black is indicative of the trend of the lowest binding energy from large to small, and the number represents the specific value of the minimum binding energy (kcal/mol). (B) Docking of rhein with Notum molecules. (C) Docking of emodin with Notum molecules. (D) Docking of chrysophanol with Notum molecules. (E–H) Inhibitory effects on Notum. Rhein (E), emodin (F), chrysophanol (G), and isoalantolactone (H).

3.4. Anti-Notum effects of anthraquinones

Next, the dose-inhibition curves of rhein, emodin, and chrysophanol were plotted using increasing doses. As shown in Figs. 4E−G, these three anthraquinones could strongly inhibit Notum in a dose-dependent manner. Fig. 4H shows the results of the isoalantolactone positive control. The IC₅₀ values were then calculated through nonlinear fitting curves to be 9.98 ± 0.84 nM, 286.4 ± 23.46 nM, and 945.8 ± 55.15 nM, for rhein, emodin, and chrysophanol, respectively (Table 2). It was evident from these results that rhein showed the most potent anti-Notum effect, which was much more potent than the positive agent isoalantolactone (IC₅₀ = 85.96 ± 5.18 nM). Time-dependent inhibition assays showed that the anti-Notum activities of rhein, emodin, and chrysophanol were not time-dependent, suggesting that these three anthraquinones acted as the non-covalent inhibitors of Notum rather than covalent inhibitors (Fig. S3). Further inhibition kinetic assays suggested that two newly identified strong Notum inhibitors (rhein and emodin) were canonical competitive inhibitors against Notum, showing the K_i values of 13.61 nM and 275.6 nM, respectively (Fig. S4). These findings suggest that both rhein and emodin act as competitive inhibitors of Notum, while rhein exhibits more potent anti-Notum activity.

Table 2.

Anti-Notum effects of rhein, emodin, and chrysophanol.

Identification	Molecule formula	Minimum binding energy (kcal/mol)	IC50 (nM)	Ki (nM)
Rhein	C15H8O6	−11.7	9.98 ± 0.84	13.61
Emodin	C15H10O5	−10.5	286.4 ± 23.46	275.6
Chrysophanol	C15H10O4	−10.2	945.8 ± 55.15	–
Isoalantolactone	C15H20O2	−9.7	85.96 + 5.81	–

–: no data.

3.5. Rhein and emodin promote osteoblast differentiation in MC3T3-E1 cells

Next, the promoting effects of rhein and emodin on osteoblast differentiation were tested in MC3T3-E1 cells. As shown in Fig. S5, the growth rates of MC3T3-E1 cells were not enhanced by rhein or emodin, suggesting that these two agents did not directly regulate osteoblast proliferation. As shown in Fig. S6, upon addition of either rhein or emodin in MC3T3-E1 osteoblasts, the level of ALP could be significantly up-regulated. Compared with emodin, rhein showed more significant pro-differentiation effects on MC3T3-E1 cells even at low dose (0.5 μM), while emodin displayed a slight pro-differentiation effects even at a relative high dose (5 μM). The large difference in the pro-differentiation effect between rhein and emodin could be partially contributed to the large difference in anti-Notum activity (>20-fold difference in IC₅₀ value). Since rhein showed more potent Notum inhibitory activity and more efficacious osteogenic differentiation effects, ALP staining and ASR staining were subsequently conducted to verify the osteogenic differentiation effects of this agent. As shown in Fig. 5, rhein significantly increased both the ALP staining intensity and the matrix mineralisation level of osteoblasts in a dose-dependent manner. Upon addition of 0.5 μM rhein, the reduced ALP staining intensity and the decreased matrix mineralisation level could be totally restored in Dex-challenged MC3T3-E1 cells. These observations clearly suggest that rhein cannot regulate osteoblast proliferation but this agent strongly promote osteoblast differentiation.

Fig. 5.

Rhein promotes MC3T3-E1 cells differentiation. (A) Alkaline phosphatase staining (ALP) was performed with BCIP/NBT kit after the MC3T3-E1 cells were treated with dexamethasone (Dex) and rhein for 7 days. (B) Alizarin red S (ARS) staining was used on day 21. ^###P < 0.001 vs. Control. ^∗∗P < 0.01 vs. Dex.

3.6. RNA sequencing revealed the regulated genes by rhein in osteoblasts

To further decipher the anti-osteoporosis mechanisms of rhein, RNA sequencing was used to investigate the regulated genes of rhein in Dex-challenged MC3T3-E1 osteoblasts. A range of differential expression genes were identified between three different groups (Figs. 6A and B). Compared to the control group, a total of 150 up-regulated genes and 631 down-regulated genes were identified from the Dex-challenged osteoblasts. After treatment with rhein for 7 days, a total of 78 genes were found to be downregulated, while 67 genes were found to be upregulated, in comparison to the Dex-challenged group. Among all tested samples from three groups, a total of 504 differentially expressed genes were identified (Fig. 6C). Compared to the control group, Wnt signaling-related genes (WNT6 and up GSK3β) and osteogenic differentiation-related genes (RUNX2 and BMP4) were significantly regulated in Dex-challenged group, while rhein significantly restored the expression of these key genes (Fig. 6D). GO enrichment and KEGG enrichment analyses demonstrated that Wnt signaling-related genes and osteogenic differentiation-related genes were significantly regulated by Dex, while rhein could restore the expression of these key genes in MC3T3-E1 cells (Figs. 6E and F). These findings encouraged us to further validate the regulatory effects of rhein on the expression levels of Wnt signaling-related and osteogenic differentiation-related proteins in osteoblasts.

Fig. 6.

RNA sequencing of MC3T3-E1 cells. (A) Comparison of the dexamethasone (Dex) group and the control group. (B) Comparison of the rhein group and the Dex group. (C) Venn diagram illustrating the differential expression comparisons between three groups. (D) Heatmap of differentially expressed genes. (E) Gene Ontology analysis, including biological functions, cellular components, and molecular functions. (F) The results of Kyoto Encyclopedia of Genes and Genomes enrichment analysis.

3.7. Rhein activates Wnt/β-catenin pathway in MC3T3-E1 cells

On the basis of the results from transcriptome sequencing, we then validate the changes in mRNA and protein levels of Wnt signaling-related proteins. Two pivotal marker proteins of the Wnt signaling (GSK3β and β-catenin) and two osteogenic differentiation markers (RUNX2 and OPN) were selected for further verification. At the mRNA expression level, rhein strongly reversed the down-regulated Wnt signaling-related genes and osteogenic differentiation-related genes by Dex (Figs. 7A–D). Concomitantly, similar effects were also observed at the protein expression level (Figs. 7E−I). These findings strongly support that rhein could strongly activate Wnt/β-catenin pathway in MC3T3-E1 cells.

Fig. 7.

Quantitative polymerase chain reaction (qPCR) and Western blotting analysis of MC3T3-E1 cells. (A) Glycogen synthase kinase 3 beta (GSK3β) mRNA expression level. (B) β-catenin mRNA expression level. (C) Osteopontin (OPN) mRNA expression level. (D) Runt-related transcription factor 2 (RUNX2) mRNA expression level. (E) The Western blotting analysis of each protein. (F) β-catenin protein expression level. (G) OPN protein expression level. (H) RUNX2 protein expression level. (I) GSK3β protein expression level. ^###P < 0.001 vs. Control, ^##P < 0.01 vs. Control, ^∗∗∗P < 0.001, ^∗∗P < 0.01, ^∗P < 0.05 vs. Dex. Dex: dexamethasone group; GAPDH: glyceraldehyde-3-phosphate dehydrogenase.

3.8. Rhein shows favorable safety profiles

Next, the safety profiles of rhein (i.g.) were investigated in healthy mice (Fig. 8A). As shown in Fig. 8B, no noticeable discrepancy was observed in the weight of the subjects in both groups at the outset and at the end of animal tests. Meanwhile, there was no significant difference in the serum safety indicators of mice (Figs. 8C−M). HE staining indicated that rhein did not exert any notable detrimental effects on the hearts, livers, spleens, lungs, kidneys, stomach, intestine, colon and brain of the mice (Fig. 8N). These observations clearly suggested that rhein did not exhibit any discernible toxic effects, thereby providing a foundation for further in vivo pharmacological investigations on its potential efficacy in the treatment of osteoporosis.

Fig. 8.

The safety evaluation of rhein in vivo. (A) Experimental design drawing. (B) The change in the body weight of mice from the rhein groups. (C–M) The serum levels of albumin (ALB) (C), alkaline phosphatase (ALP) (D), alanine aminotransferase (ALT) (E), aspartate aminotransferase (AST) (F), creatine kinase (CK) (G), creatinine (CREA) (H), glucose (GLU) (I), total cholesterol (TC) (J), triglycerides (TG) (K), total protein (TP) (L) and blood urea nitrogen (BUN) (M). (N) The characteristic pictures of hematoxylin and eosin staining of the heart, liver, spleen, lung, kidney, stomach, intestine, colon, and brain were sampled from the rhein group. n = 8, ^∗∗P < 0.01, ^∗P < 0.05 vs. Control. ns: not significant vs. Control.

3.9. Rhein ameliorates Dex-induced osteoporosis in mice

Finally, the in vivo anti-osteoporosis effects of rhein were investigated in Dex-induced osteoporosis mice (Fig. 9A). Following injection of Dex (i.p.) for five weeks, a significant reduction in the body weight was observed in Dex-challenged mice (Figs. 9B and C). Dex also caused the reduced levels of all tested bone metabolism-related serum biomarkers. By contrast, the levels of all tested bone metabolism-related serum biomarkers were significantly restored in PTH- or rhein-treated groups (Figs. 9D−I). Micro-CT results showed that trabeculae were sparse and the cortical bone was thinner in the Dex group, which proved that the osteoporosis model was successfully constructed. Following treatment with PTH or rhein, the cortical and cancellous bones were thickened significantly (Figs. 10A and B). Quantitative parameters of bone are shown in Figs. 10D−K. Double labeling showed that the bone formation rate per bone surface and the mineral apposition rate were remarkably improved following treatment with PTH or rhein (Fig. 10C). In the distal and middle femurs, both trabecular and cortical bone thicknesses are significantly thicker than the Dex-challenged group (Fig. 10D, E, H and I). Notably, the bone volume to tissue volume (BV/TV) and BMD of the rhein group at the distal femur did not increase significantly, but the BMD increase at the mid femur was more prominent (Fig. 10F, G, J and K). Immunofluorescence results showed that Notum was up-regulated in cortical bone (mainly distributed in osteoblasts) from Dex-challenged mice (Fig. 10L), while the Notum expression levels could be significantly suppressed following treatment with rhein. These findings clearly demonstrate that rhein acts as an effective anti-osteoporosis agent, evidence by increasing the thickness of cortical and cancellous bone, and enhancing bone density in Dex-induced osteoporosis mice.

Fig. 9.

The protective effects of rhein from dexamethasone (Dex)-osteoporosis mice on serum bone metabolism index. (A) The detailed schedule of animal experiments. (B) Body weight of mice in the initial group. (C) The change in the body weight of mice. (D–I) The bone metabolism index of serum levels, including alkaline phosphatase (ALP) (D), osteocalcin/bone gla protein (OT/BGT) (E), osteoprotegerin (OPG) (F), collagen type I alpha 1 chain (COL1A1) (G), calcium (Ca) (H), and phosphorus (I). n = 6, ^###P < 0.001, ^##P < 0.01 vs. Control; ^∗∗∗P < 0.001, ^∗∗P < 0.01, ^∗P < 0.05 vs. Dex. ns: not significant vs. Control. PTH: parathyroid hormone 1-34; R(L): rhein low-dose group; R(H): rhein high-dose group.

Fig. 10.

The protective effects of rhein on dexamethasone (Dex)-induced pathological bone loss via inhibiting Notum. (A) Image of the femur was generated using micro-CT. (B) Hematoxylin and eosin (H&E) staining of distal femur. (C) The characteristic images of double labeling. (D) Distal femur cortical thickness. (E) Distal femur trabeculae thickness. (F) Distal femur bone volume/total volume (BV/TV). (G) Distal femur bone mineral density (BMD). (H) Middle femur cortical thickness. (I) Middle femur trabeculae thickness. (J) Middle femur BV/TV. (K) Middle femur BMD. (L) Immunofluorescence analysis of osteopontin (OPN)/Notum. n = 6, ^###P < 0.001 vs. Ctrol; ^∗∗∗P < 0.001, ^∗∗P < 0.01, ^∗P < 0.05 vs. Dex. ns: not significant vs. Control. PTH: parathyroid hormone 1-34; R(L): rhein low-dose group; R(H): rhein high-dose group.

4. Discussion

Increasing evidence has suggested that Notum, a negative regulator of Wnt signaling, is a validated therapeutic target for treating osteoporosis [38]. As a key protein primarily secreted from osteoblasts, Notum has been identified as a crucial biomarker and a key mediator for negative regulating Wnt signaling [45]. Recently, Notum inhibitor therapy has been proved capable of increasing the bone density and enhancing cortical bone thickness in osteoporosis mice [36]. It is well-known that some HMs (such as He Shou Wu) are effective for treating osteoporosis, but the key anti-osteoporosis constituents have not been well-investigated. In fact, high-efficient uncovering the key bioactive constituents from HMs is always a big challenge for the medicinal chemists. As a practical and efficient strategy for drug discovery, fluorescence-based high-throughput screening (HTS) are frequently used for screening the active chemicals from compound library or herbal constituents, especially for discovering of enzyme inhibitors. Previously studies have reported a panel of fluorescent substrates of Notum [37,39,41], which greatly facilitate high-efficient screening and characterization of Notum inhibitors.

It should be noted that Notum is a canonical monomeric serine hydrolase, which possess a single flexible catalytic pocket [26,29]. Almost all reported naturally occurring Notum inhibitors are identified as competitive inhibitors, which block the substrate catalysis via direct binding at the catalytic pocket of Notum [34]. The unique structural features of Notum inspired us to use computer-aided virtual screening for high-efficient discovering of Notum inhibitors from herbal constituents. In this work, an efficient strategy was adapted for uncovering the key naturally occurring anti-Notum constituents from HMs, via integrating biochemical, phytochemical, computational, and cellular assays. After screening more than one hundred clinically used HMs, PM, one of the most frequently used anti-osteoporosis herb in clinic, showed the most potent anti-Notum activity. Interestingly, it was also found that PM displayed competitive inhibition on Notum, suggesting that some phytochemicals in PM acted as competitive Notum inhibitors. These findings encouraged us to further investigate the key anti-Notum constituents in PM, as well as to evaluate their anti-osteoporosis effects.

Fortunately, phytochemical profiling coupling with docking-based virtual screening showed that PM constituents showed high binding-potency towards Notum, while inhibition assays showed that three anthraquinones including rhein, emodin, and chrysophanol were strong competitive inhibitors of Notum, while rhein was the most potent one (IC₅₀ = 9.98 nM). To the best of our knowledge, the anti-Notum potency of rhein is higher than the previously reported natural products as Notum inhibitors including caffeine (IC₅₀ = 200 μM) and isoalantolactone (IC₅₀ = 60.6 nM) [41,46]. These findings encouraged us to further investigate the anti-osteoporosis both in vitro and in vivo. Cellular investigations showed that rhein markedly promoted osteoblast differentiation in dexamethasone-challenged MC3T3-E1 osteoblasts, while this agent could significantly activate Wnt signaling and up-regulate osteogenic differentiation-related genes and marker proteins. Animal tests showed that rhein displayed favorable safety profiles in healthy mice, while this agent could remarkably elevate bone mineral density, augment trabecula and cortical bone thickness in Dex-induced osteoporotic mice. Interestingly, it was also found that rhein could down-regulate the protein levels of Notum in cortical bone, suggesting that rhein could block Notum function via both inhibition and down-regulation. In future, the molecular mechanism of rhein to affect Notum biosynthesis or degradation should be investigated in-depth.

As a newly identified naturally occurring Notum inhibitor, rhein showed impressive anti-osteoporosis effects, mainly via enhancing mid-femur bone density and cortical bone thickness. Meanwhile, it is also reported that rhein and its analogs displayed other beneficial effects (such as promoting healing) for the older adults, via targeting PPAR-alpha [47,48]. These findings suggested that rhein holds a great promise as a drug candidate to develop novel anti-osteoporosis agents. However, it was also reported that rhein displayed very lower oral bioavailability and poor metabolic stability [49]. Thus, it is necessary to design and develop more efficacious rhein derivatives as novel anti-Notum agents. In the past few decades, many rhein derivatives have been designed and synthesized for different purposes. In future, the medicinal chemists can use structure-based drug design and other strategies to design and synthesize more rhein derivatives with improved anti-Notum efficacy and favorable safety profiles, as novel Notum inhibitors. It can be seen from Fig. S2 that the carboxylic acid group of rhein created strong interaction with the key amino acid residues surrounding the catalytic cavity of Notum, such interactions should be conserved in future drug design and optimization. Additionally, the bone-targeted nanodrug delivery systems could also be designed for delivering of rhein or its derivatives to treat osteoporosis with high treatment outcome [50].

5. Conclusion

In summary, this study presents a practical and comprehensive strategy for efficiently identifying potent Notum inhibitors from HMs and also showcases a panel of in-depth pharmacological investigations on the newly identified Notum inhibitors. Facilitated by the fluorescence-based HTS, Polygonum multiflorum Thunb. (PM), a commonly used anti-osteoporosis herb, showed potent and competitive inhibition against Notum. After then, a high-efficient strategy was adapted to uncover the key anti-Notum constituents in PM, via integrating phytochemical, computational and biochemical assays. The results clearly demonstrated that three anthraquinones in PM, including rhein, emodin, and chrysophanol, showed high binding-potency towards Notum. Biochemical assays validated that these three anthraquinones were strong competitive inhibitors of Notum, while rhein was the most potent one (IC₅₀ = 9.98 nM). Further cellular assays demonstrated that rhein significantly prompted cell differentiation of dex-challenged MC3T3-E1 osteoblasts, while this agent could also regulate the Wnt signaling-related genes in Dex-challenged MC3T3-E1 osteoblasts. In vivo tests demonstrated that rhein showed favorable safety profiles, while this agent could markedly elevate bone mineral density, augment the thickness of both cancellous and cortical bone in dex-induced osteoporotic mice. Collectively, this study showcases an integrated strategy for discovering the naturally occurring anti-Notum constituents from HMs, while rhein, a newly identified potent Notum inhibitor, holds great promise as a drug candidate for developing novel anti-osteoporosis agents.

CRediT authorship contribution statement

Jia Guo: Writing – review & editing, Writing – original draft, Software, Project administration, Investigation, Formal analysis, Data curation, Conceptualization. Yuqing Song: Writing – review & editing, Writing – original draft, Project administration, Investigation, Formal analysis, Data curation, Conceptualization. Mengru Sun: Investigation, Formal analysis, Conceptualization. Jun Qian: Resources, Investigation. Dihang See: Writing – review & editing. Tian Tian: Methodology, Investigation. Yunqing Song: Supervision, Software, Investigation. Wei Liu: Methodology, Formal analysis. Hongping Deng: Investigation. Yao Sun: Investigation, Conceptualization. Guangbo Ge: Writing – review & editing, Supervision, Investigation, Funding acquisition, Conceptualization. Yongfang Zhao: Supervision, Investigation, Funding acquisition, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following is the Supplementary data to this article: