Osteogenic, antiresorptive, and toxicological evaluation of alpha-mangostin in MC3T3-E1 cells and medaka fish

Abstract

Osteoporosis, characterized by excessive bone resorption, requires new therapeutic agents with high efficacy and favorable safety profiles. In this study, we isolated alpha-mangostin (aMG) from the peel of Garcinia mangostana L. and investigated its osteogenic, antiresorptive, and toxicological activities. The osteogenic effects were evaluated in vitro using MC3T3-E1 pre-osteoblast cells, whereas acute toxicity and bone-protective activity were assessed in vivo using medaka fish (Oryzias latipes). In MC3T3-E1 cells, aMG promoted osteogenic differentiation by enhancing proliferation, alkaline phosphatase activity, collagen synthesis, and mineralization at 1 and 10 µM, whereas cytotoxicity was observed at 100 and 1,000 µM. In medaka embryos and larvae, aMG induced acute toxicity with LC₅₀ values (concentrations causing 50% mortality) of 91.27 and 3.44 µM, and EC₅₀ values (concentrations causing 50% of developmental abnormalities) of 103.66 and 3.3 µM, respectively, leading to tail deformities, hemorrhage, and reduced hatching and heart rates. At sub-toxic concentrations (0.25–1 µM), aMG significantly attenuated Rankl-induced bone loss in the neural arches of osteoporotic medaka larvae. Collectively, this study provides the first evidence of the dual osteogenic and antiresorptive actions of aMG, along with its toxicity profile, highlighting its potential as a plant-derived anti-osteoporosis agent.

INTRODUCTION

Osteoporosis is a long-term bone disorder characterized by progressive loss of bone mass and disruption of trabecular and cortical architecture, ultimately increasing the risk of fragility fractures [1]. It has become a significant public health burden, particularly in aging populations [1,2]. Although bisphosphonates and hormone replacement therapy are currently used with proven efficacy, safety issues and adverse effects limit their long-term application [3,4]. This highlights the pressing need for new and safer agents, particularly those from natural sources.

Xanthones are natural polyphenolic compounds found in plants, fungi, and lichens [5]. Their chemical framework typically consists of a tricyclic skeleton in which two aromatic rings are joined via a carbonyl bridge and an oxygen atom. These metabolites exhibit diverse pharmacological activities including antibacterial, antiviral, antioxidant, and anti-inflammatory effects [6-9], as well as anticancer, antidiabetic, and other bioactive properties [10-14]. Recent evidence suggests that xanthones may exert bone-protective actions and hold promise as candidates for anti-osteoporotic therapy [15-17].

A compound with anti-osteoporotic potential may act as an osteogenic and/or antiresorptive agent, enhancing bone formation and/or suppressing osteoclastogenesis, respectively [18]. Several xanthone derivatives have demonstrated antiresorptive activity: synthetic alkoxyl xanthones, including 3,6-diisopropoxyxanthone, inhibited bone resorption in vitro and in vivo [19]; natural xanthones from Halenia corniculata suppressed osteoclast differentiation [20]; and gambogic acid targeted CXCR4-related pathways to block osteoclastogenesis [16]. Extracts from mangosteen peel also reduced bone resorption and osteoclast activity in mouse and guinea pig models [15,17]. The osteogenic properties of xanthones remain largely unexplored.

Alpha-mangostin (aMG), a major xanthone isolated from mangosteen (Garcinia mangostana L.) pericarp, has attracted considerable attention due to its broad biological activities, including anti-inflammatory, anticancer, antibacterial, antifungal, antidiabetic, cardiovascular protective, and anti-obesity effects [21]. It has also been investigated in neurodegenerative diseases and infectious models [22,23] and exhibits cytotoxicity in different cancer cell lines [24-26]. Importantly, some studies suggest the antiresorptive activity of aMG [27,28]; however, its osteogenic properties and toxicity profile in bone-related models remain poorly understood.

In this study, we isolated aMG from mangosteen peel and investigated its osteogenic potential in MC3T3-E1 pre-osteoblasts by measuring their proliferation, alkaline phosphatase (ALP) activity, collagen synthesis, and mineralization levels. Because high concentrations of aMG showed cytotoxicity in vitro, we further evaluated its acute toxicity in vivo using wild-type medaka (Oryzias latipes) embryos and larvae, a well-established vertebrate model in toxicology and bone biology owing to its transparency, high fecundity, and conserved bone metabolism [29]. The toxicity endpoints included lethality (LC₅₀), morphological abnormalities (EC₅₀), and developmental parameters, such as heart rate and hatching success. Sub-toxic doses were then tested in a receptor activator of nuclear factor kappa B ligand (Rankl)-induced transgenic medaka model of osteoporosis [30,31], where heat-shock-induced Rankl expression stimulates osteoclastogenesis and bone loss in the neural arches. Our findings provide novel in vivo evidence of bone-protective activity and clarify the toxicity profile of aMG, supporting its further evaluation as a natural therapeutic candidate for osteoporosis and other bone-related disorders.

METHODS

Materials

All chemicals were obtained from commercial suppliers and used without further purification. Dried mangosteen peel was obtained from Thanh Binh Herbal Tea Co., Ltd. in Vietnam. The MC3T3-E1 murine preosteoblast cell line was purchased from the American Type Culture Collection (ATCC).

Isolation of aMG

The dried pericarp of G. mangostana L. (2 kg) was finely ground and extracted with dichloromethane (4 × 1 L) using an ultrasound-assisted extractor for 45 min at 100 W. The solid was filtered off, and the solvent was removed under reduced pressure to afford 64.15 g of crude product. The crude product (1 g) was purified by chromatography column (80 g of silica gel, gradient 0%–40% chloroform in petroleum ether) to afford the desired product aMG [32] (0.30 g, 30% yield). ¹H-NMR (400 MHz, CDCl₃) δH (ppm): 13.78 (s, 1H), 6.83 (s, 1H), 6.29 (s, 1H), 5.31–5.25 (m, 2H), 4.10–4.09 (d, J = 6.2 Hz, 2H), 3.81 (s, 3H), 3.47–3.45 (d, J = 7.1 Hz, 2H), 1.84 (s, 3H), 1.83 (s, 3H), 1.77 (s, 3H), 1.69 (s, 3H). ¹³C-NMR (101 MHz, CDCl₃) δC (ppm): 182.1, 161.6, 160.6, 155.8, 155.1, 154.6, 142.6, 137.1, 135.7, 132.2, 123.2, 121.5, 112.2, 108.5, 103.6, 101.6, 93.3, 62.1, 26.6, 25.9, 25.8, 21.5, 18.2, 17.9. ESI-HRMS (m/z): C₂₄H₂₇O₆ [M+H]⁺ m/z = 411.1808, found 411.1797.

Osteoblast differentiation assays

In vitro cell culture method: The MC3T3-E1 murine pre-osteoblast cell line (ATCC) was maintained as a monolayer culture in α-minimum essential medium supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin at 37°C in a humidified incubator with 5% CO₂. Cells were passaged every two days using 0.05% trypsin-EDTA. To evaluate the differentiation potential of MC3T3-E1 cells into osteoblasts, cells were cultured in differentiation medium supplemented with 10 mM β-glycerophosphate and 50 µg.ml^–1 ascorbic acid when they reached 80% confluence.

Evaluation of cell proliferation under the influence of test compounds: The 3-[4,5-dimethylthiazol-2-yl]-2,5 diphenyltetrazolium bromide (MTT) assay was used to assess the proliferation of MC3T3-E1 cells in response to test compounds. This assay measured cell viability based on the enzymatic conversion of MTT into a colored formazan product in metabolically active cells. MC3T3-E1 cells were seeded in 96-well plates at a density of 1 × 10⁴ cells/well and treated with test compounds dissolved in 10% dimethyl sulfoxide (DMSO) at final concentrations of 1,000, 100, 10, and 1 µM. Wells containing only the solvent served as negative controls. After incubation at 37°C in 5% CO₂ for three days, 20 µl of MTT solution (5 mg ml^–1) was added to each well and incubated for an additional 4 h. The supernatant was then removed, and 100 µl of DMSO was added to dissolve the formazan crystals. The absorbance was measured at 490 nm using an ELISA plate reader (BioTek ELx800). Cell viability was calculated using the following equation:

$$\text{\%Viability}=\frac{({\text{OD}}_{\text{sample}}-{\text{OD}}_{\text{blank}})}{{\text{OD}}_{\text{negative control}}-{\text{OD}}_{\text{blank}}}\times 100\%$$ (1)

ALP activity assay: MC3T3-E1 cells were seeded in 96-well plates and cultured in differentiation medium containing 10 mM β-glycerophosphate and 50 µg ml^–1 of ascorbic acid. The test compounds were added at various concentrations, and the culture medium was replaced every 2–3 days. ALP activity was measured on day 7 using an ALP Assay Kit (Abcam) following the manufacturer’s instructions. The results were expressed as a percentage of ALP activity relative to that of the negative control, using the following formula:

$$\text{\%ALP Activity}=\frac{({\text{OD}}_{\text{sample}}-{\text{OD}}_{\text{blank}})}{{\text{OD}}_{\text{negative control}}-{\text{OD}}_{\text{blank}}}\times 100\%$$ (2)

Collagen synthesis assay: To assess collagen synthesis, MC3T3-E1 cells were cultured in differentiation medium and treated with the test compounds for 10 days. The cells were then washed with phosphate-buffered saline (PBS) and fixed with Bouin’s solution for 1 h. After fixation, the plates were washed under running water for 15 min and air-dried. The cells were stained with 0.1% Sirius Red in picric acid for 1 h, followed by washing with 0.01 M HCl to remove excess dye. The bound dye was dissolved in 0.1 M NaOH, and the absorbance was measured at 550 nm using an ELISA plate reader (BioTek). Collagen production was quantified as follows:

$$\text{\%Collagen synthesis}=\frac{({\text{OD}}_{\text{sample}}-{\text{OD}}_{\text{blank}})}{{\text{OD}}_{\text{negative control}}-{\text{OD}}_{\text{blank}}}\times 100\%$$ (3)

Calcium mineralization assay: Calcium deposition was assessed using Alizarin Red S staining. MC3T3-E1 cells were cultured for 15 days in a differentiation medium containing the test compounds. The cells were washed twice with PBS and fixed with 70% ethanol for 1 h. After fixation, the cells were stained with 40 mM Alizarin Red S (pH 4.4) for 15 min and washed with deionized water. The bound stain was dissolved using 10% cetylpyridinium chloride and incubated for 15 min with gentle shaking. The absorbance was measured at 561 nm using a microplate reader (BioRad). The percentage of calcium mineralization was calculated as follows:

$$\text{\%Mineralization}=\frac{({\text{OD}}_{\text{sample}}-{\text{OD}}_{\text{blank}})}{{\text{OD}}_{\text{negative control}}-{\text{OD}}_{\text{blank}}}\times 100\%$$ (4)

Fish lines and fish maintenance

This study used wild-type medaka fish for acute toxicity experiments. To assess the bone protective activity of aMG, wild-type and transgenic medaka fish rankl:HSE:CFP model of osteoporosis were used [30,31]. The fish were maintained as described by Pham et al. [31] in recirculating systems (Techniplast) under a 14-h light/10-h dark cycle at 28°C–30°C. Embryos were collected from breeding pairs (one male and three females aged 6–12 months) and reared at 30°C in 50 ml E3 medium (5 mM NaCl, 0.17 mM KCl, 0.33 mM CaCl₂, 0.33 mM MgCl₂) at a density of 20 individuals per petri dish (15 × 90 mm), with the medium changed daily. Embryos hatched naturally within 7–14 days post-fertilization (dpf) [29], and for experimental synchronization, hatching was induced using natural enzymes on the day of the experiment. Animal studies were performed after receiving approval from the Institutional Animal Care and Use Committee (IACUC) at the Dinh Tien Hoang Institute of Medicine, Hanoi, Vietnam (IACUC approval No. IRB-A-2312).

Evaluation of acute toxicity

The acute toxicity of aMG in wild-type medaka embryos and larvae was assessed according to OECD guidelines [33]. The embryonic (egg) stage referred to the period when the fish were still covered by the chorion, and the early larval stage was when the fish have hatched and can swim freely [34]. First, limit tests were performed for aMG at the maximum solubility of 120 µM on 20 embryos/larvae. Dose range-finding tests then were conducted to determine the dose range used for the acute toxicity of the substance, which were 12, 30, 60, 90, and 120 µM for embryos and 1.5, 2.0, 2.5, 3.0, 4.0, 5.5, 7.0, 9.0, and 12.0 µM for larvae. For the embryo tests, fertilized embryos (6 hours post-fertilization [hpf]) were placed in 6-well plates with 10 embryos per well in 4 ml of E3 medium containing the test substance. The treatments lasted for four days (6–102 hpf), with the medium changed daily. For larvae, the experiment was conducted similarly, but the treatments lasted from 7 to 11 dpf. The solvent control group was treated with 0.1% DMSO. Mortality (absence of heartbeat, lack of movement, or coagulation) and morphological abnormalities (e.g., necrosis, reduced heart rate, developmental delay, and reduced eye pigmentation) were recorded every 24 h. Each experiment was performed in triplicate, with 10 embryos/larvae per group. Survival rates, abnormality rates, and values of LC₅₀ and EC₅₀, the dose that causes death and abnormality in 50% of the tested animals, respectively, were determined. The LC₅₀ and EC₅₀ values of aMG were determined using sigmoidal regression analysis in GraphPad Prism (Prism 9; GraphPad Software Inc.).

aMG treatment to Rankl larvae

The antiresorptive activity of aMG was assessed using the transgenic rankl:HSE:CFP larvae model of osteoporosis subline c7 (hereafter called Rankl fish/embryos). This fish expressed Rankl and CFP reporter signals upon heat shock at 39oC, which resulted in osteoclastogenesis and mineralized bone damage in the neural arches of the vertebrae [30,31]. Based on the acute toxicity test results in larvae, aMG concentrations of 0.1, 0.25, 0.5, 0.75, and 1.0 µM were selected for the experiments. This range represents approximately 1/30–1/3 of the previously determined LC₅₀ value. Alendronate (50 µg/ml, Sigma A4978), a known antiresorptive bisphosphonate, served as the positive control [31], and 0.1% DMSO was used as the solvent (negative) control. Additionally, a group of wild-type fish subjected to heat shock was included as a biological control to observe normal bone development under heat stress. aMG treatments were performed as previously described [31]. Briefly, 7 dpf Rankl larvae were randomly divided into experimental groups, with five larvae per well in 24-well plates. Each well contained 1 ml of E3 medium with either a tested dose of aMG (final DMSO concentration of 0.1%), 0.1% DMSO alone, or 0.1% DMSO with 50 µg/ml alendronate. At 9 dpf, larvae were subjected to a heat shock at 39°C for 90 min to induce Rankl expression and osteoporosis-like phenotypes. Two hours after heat shock, the medium of all fish groups was replaced with fresh, group-specific medium. At 11 dpf, larvae were fixed with 4% paraformaldehyde (PFA) and stained with Alizarin Red S to visualize the mineralized bone matrix. The experiments were repeated until a statistical sample size of at least 30 larvae per group (n ≥ 30) was achieved.

Staining of mineralized bone matrix

Eleven dpf fish larvae were fixed with 4% PFA and stained with Alizarin Red S (Sigma A5533) to visualize mineralized bone structures, as described previously [30,31]. Imaging of live embryos or larvae for toxicity experiments was performed using a steroscope (Carl Zeiss). Imaging of fixed larvae stained with Alizarin Red S was conducted following a previously described protocol [31]. The stained fish were mounted in 70% glycerol and imaged using an Axioplan 2 microscope (Carl Zeiss) with a 10X objective. Image stitching and measurement of mineralized neural arch lengths were performed using ImageJ software.

Quantification of bone mineralization and damage in medaka larvae

Given the observed bone damage in the neural arches of Rankl medaka larvae, the Index of Mineralization (I_M) was used to quantify bone mineralization and assess the extent of damage, as previously described [31]. The I_M was defined as the total length of the first 15 mineralized neural arches, expressed in pixels (px). ($I_M={\displaystyle {\sum }_{k=1}^{15}L}$, where k is the ordinal number of the neural arch and L is the length of each arch). aMG was considered to have an antiresorptive and bone-protective effect in this fish model if it significantly increased I_M values compared to the DMSO control group. To further quantify this protection, the Index of Mineralization Protection (I_P) for each dose was calculated using the following formula:

$$I_P=\frac{I_{M\left(+Rankl+aMG\right)}-I_{M\left(+Rankl-aMG\right)}}{I_{M\left(-Rankl-aMG\right)}}\times 100\%$$ (5)

Statistical analysis

Data are presented as mean ± standard deviation (SD) or standard error of the mean (SEM). GraphPad Prism software (Prism 9; GraphPad Software Inc.) was used for statistical analysis and graph construction. Differences between experimental groups were assessed using one-way ANOVA followed by Tukey's post-hoc test. Statistical significance was defined as follows: *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001.

RESULTS

Isolation of aMG

aMG was isolated from the pericarp of G. mangostana as a yellow solid with high performance liquid chromatography (HPLC) purity of 98.3% (Supplementary Fig. 1). The ¹H-NMR spectrum (400 MHz, CDCl₃) revealed a characteristic downfield signal at δ_H 13.78 (s, 1H), corresponding to a chelated hydroxyl proton (OH-1). Two isolated aromatic protons were observed at δ_H 6.83 (s, 1H, H-5) and δ_H 6.29 (s, 1H, H-4). The spectrum also displayed multiple signals corresponding to two isoprenyl groups, including δH 5.31–5.25 (m, 2H, H-12, H-17), δ_H 4.10–4.09 (d, J = 6.4 Hz, 2H, H-11), δ_H 3.47–3.45 (d, J = 7.2 Hz, 2H, H-16). Additionally, a singlet at δ_H 3.81 (s, 3H) indicated the presence of a methoxy group (7-OCH₃). Four methyl group resonances were observed at δ_H 1.84, 1.83, 1.77, and 1.69 (Supplementary Fig. 2). The ¹³C-NMR spectrum (101 MHz, CDCl₃) displayed 24 carbon signals. The presence of a conjugated carbonyl carbon was confirmed by the signal at δ_C 182.1 (C-9). Twelve aromatic carbon signals were observed at δ_C 161.6, 160.6, 155.8, 155.1, 154.6, 142.6, 137.1, 112.2, 108.7, 103.5, 101.6, and 93.3, corresponding to the xanthone skeleton. Signals at δ_C 135.7 and 132.2 were assigned to quaternary isoprenyl carbons (C-13 and C-18), while signals at δ_C 123.3 and 121.5 corresponded to tertiary isoprenyl carbons bearing a single proton (C-12 and C-17). The methoxy carbon appeared at δ_C 62.1 (7-OCH₃). Additionally, secondary carbons in the isoprenyl groups were observed at δ_C 26.6 and 21.5 (C-11, C-16), while four methyl carbons were identified at δ_C 25.8, 25.9, 18.2, and 17.9 (C-15, C-20, C-14, and C-19) (Supplementary Fig. 3). The ESI-HRMS spectrum exhibited a molecular ion peak at m/z 411.1797 [M+H]⁺, which corresponds to the molecular formula C₂₄H₂₇O₆. The calculated theoretical mass was m/z 411.1808, confirming the expected molecular composition (Supplementary Fig. 4). Based on these spectral data and a comparison with the reference literature [35,36], the structure of the isolated compound was confirmed to be aMG (Fig. 1).

Fig. 1. Chemical structure of α-mangostin.

Osteogenic properties of aMG

The MC3T3-E1 pre-osteoblast cell line possesses the capacity to differentiate into osteoblasts responsible for bone formation. During the differentiation process, MC3T3-E1 cells exhibit key osteoblastic characteristics, including the production of ALP, enhanced collagen synthesis from procollagen as a major component of the extracellular matrix, and subsequent mineralization of the matrix [37]. Owing to these osteogenic features, the MC3T3-E1 cell line is widely utilized as an in vitro model for evaluating the anti-osteoporotic potential of compounds [38].

Fig. 2 illustrates the experimental workflow (Fig. 2A) and the osteogenic effects of aMG at concentrations of 1, 10, 100, and 1,000 µM. At high concentrations (1,000 and 100 µM), aMG was cytotoxic and markedly reduced cell viability (Fig. 2B). In contrast, low concentrations promoted proliferation: 10 µM increased cell viability to 117.73% (p < 0.01), while 1 µM yielded 105.20%, suggesting a proliferative effect at sub-cytotoxic doses. Consistent with these findings, ALP activity was strongly suppressed at 100 µM, decreasing to 21.16% of the control level (Fig. 2C). However, both 10 µM and 1 µM significantly enhanced ALP activity, reaching 116.82% (p < 0.05) and 130.11% (p < 0.01), respectively (Fig. 2C). At 1 µM, aMG also increased collagen production to 114.13% (p < 0.05) (Fig. 2D, F). Moreover, mineral deposition was significantly elevated at both 1 µM and 10 µM, achieving 110.81% (p < 0.05) and 124.64% (p < 0.01), respectively (Fig. 2E, G).

Fig. 2. Effects of α-mangostin (aMG) on osteoblast differentiation and mineralization.

(A) Timeline representing the experimental design for assessing the osteogenic effects of aMG on MC3T3-E1 pre-osteoblasts. Differentiation and aMG treatment were initiated when cells reached 80% confluence (designated as 0 days post-exposure [dpe]). Cell proliferation, alkaline phosphatase (ALP) activity, collagen synthesis, and mineralization were assessed at 3, 7, 10, and 15 dpe, respectively. (B) Effects of aMG at 1, 10, 100, and 1000 µM on the proliferation of MC3T3-E1 cells. (C–E) Effects of aMG at 1, 10, and 100 µM on ALP activity (C), collagen synthesis (D), and mineralization (E) in differentiated MC3T3-E1 cells. Representative images of Sirius Red staining for collagen synthesis (F) and Alizarin Red staining for mineralization (G) in cells treated with 1 and 10 µM aMG. Data are presented as mean ± standard deviation (n = 3). ns, not significant; DMSO, dimethyl sulfoxide. *p < 0.05, **p < 0.01, ***p < 0.001 vs. DMSO 10% control; statistical differences were determined by one-way ANOVA followed by Tukey’s post-hoc test.

aMG shows high acute toxicity on medaka embryos and larvae

The limit tests revealed that for aMG, the OECD recommended dose of maximum solubility [33] of 120 µM caused 90% death of the tested embryos and larvae after 72 h of exposure. Based on dose-range finding tests and zebrafish data [39], the following doses were selected for acute toxicity assessment of aMG in medaka: 12, 30, 60, 90, and 120 µM for embryos, and 1.5, 2.0, 2.5, 3.0, 4.0, 5.5, 7.0, 9.0, and 12.0 µM for larvae. The procedure for toxicity assessment of aMG in medaka embryos and larvae after 96 h of exposure is described in Fig. 3A.

Fig. 3. Acute toxicity of α-mangostin (aMG) on medaka embryos and larvae.

(A) Schematic diagram of the experimental timeline. Embryos (starting at 6 hours post-fertilization [hpf]) and larvae (starting at 7 days post-fertilization [dpf]) were exposed to aMG, and toxicity was assessed every 24 hours up to 96 hours post-exposure (hpe). (B, H) Daily survival rates of embryos (B) and larvae (H) exposed to the indicated concentrations of aMG over 96 h (n = 30 per group). (C, I) Stacked bar charts showing the rates and types of abnormalities observed in embryos (C) and larvae (I) at 96 hpe (n = 30 per group). (D) Heart rates of embryos at 48 hpe (n = 20 per group). (E) Hatching rates recorded at 7, 9, and 11 dpf for embryos treated with 12 and 30 µM aMG (n = 30 per group). (F, J) Representative images showing the morphological development of embryos (F) and larvae (J) over time. (G) High-magnification images highlighting specific malformations as indicated by arrows (e.g., tail deformity, hemorrhage, and reduced eye pigmentation) in embryos at 72 and 96 hpe. Scale bar = 500 µm. The colored arrows in (F) correspond to the specific abnormalities defined in the figure keys. Statistical analysis: Data are presented as mean ± standard deviation or percentages where applicable. ns, not significant; DMSO, dimethyl sulfoxide. *p < 0.05, **p < 0.01 ***p < 0.001, ****p < 0.0001 compared to the control group (DMSO). Statistical differences were determined using one-way ANOVA followed by Tukey’s post-hoc test.

aMG induced dose-and time-dependent toxicity in both embryos and larvae (Fig. 3). Estimated LC₅₀ values decreased with increasing doses and longer exposure times (Table 1). After 96 h of exposure, LC₅₀ values were 91.27 µM for embryos and 3.44 µM for larvae, indicating that larvae were substantially more sensitive than embryos. This substance caused mortality in some embryos after three days of exposure (dpe) at 90 and 120 µM (Fig. 3B). In contrast, it induced substantial larval mortality as early as 1 dpe at much lower doses (> 5.5 µM; Fig. 3H), with 9.0 and 12.0 µM causing 86.7% and 90% mortality, respectively, at 1 dpe (Fig. 3H). None of the tested doses, including the highest (120 µM), caused 100% embryo mortality by 4 dpe (Fig. 3B). In contrast, 12.0 µM caused complete larval mortality by 2 dpe, and 5.5–9.0 µM killed all larvae by 3 dpe (Fig. 3H). Doses ≤ 30 µM did not cause embryo mortality, and doses ≤ 2 µM did not cause larval mortality.

Table 1.

LC₅₀ and EC₅₀ values of alpha-mangostin on medaka embryos and larvae

Hour post exposure		LC50 estimate (µM)	LC50 95% CI (µM)	EC50 estimate (µM)	EC50 95% CI (µM)	TI	TI range
Embryos	72	117.44	109.89–130.24	117.47	106.04–138.44	1	0.79–1.23
	96	91.27	87.43–95.04	103.66	97.45–110.3	0.88	0.79–0.98
Larvae	24	6.47	6.15–6.82	-	-	-	-
	48	5.23	5.05–5.41	-	-	-	-
	72	4.13	4.13–4.28	3.81	3.6–4.11	1.08	1.01–1.19
	96	3.44	3.31–3.57	3.3	3.17–3.45	1.04	0.96–1.13

TI (Teratogenic Index) = LC₅₀/EC₅₀. CI, confidence interval; - N/A, not applicable.

aMG also exhibited teratogenic effects on medaka development. Dose-dependent developmental deformities and functional impairments were observed in surviving embryos (Fig. 3C, F, G) and larvae (Fig. 3I, J) following aMG exposure. Abnormality observed in surviving embryos during four dpe included deformed tails, hemorrhage, necrosis, reduced eye pigmentation, and tail malformation (Fig. 3F, G). At 4 dpe, 80% of embryos surviving the 120 µM aMG treatment exhibited these defects (Fig. 3C, F). In larvae, immobilization was the only abnormality recorded at lower doses (2.0, 2.5, and 3.0 µM; Fig. 3I), whereas the highest dose (4.0 µM) additionally caused spinal curvature and delayed yolk sac absorption (Fig. 3I, J). In addition, aMG reduced embryo heart rate and hatching rate in a dose-dependent manner (Fig. 3D, E). At 2 dpe, doses of 60 µM and above decreased heart rate; in particular, 90 and 120 µM significantly reduced heart rate to 66.19% and 62.63%, respectively, compared to wild-type (Fig. 3D). Medaka raised in laboratory conditions, including ours, generally hatch between 7 and 14 dpf [29]. In our experiments, at 7 dpf, none of the embryos exposed to 12 or 30 µM aMG hatched, whereas approximately 20% of the wild-type embryos hatched. By 11 dpf, 93.33% of wild-type embryos had hatched, but only 33.33% and 23.33% of embryos exposed to 12 µM and 30 µM, respectively, successfully hatched (Fig. 3E).

aMG reduces bone damage in Rankl-induced osteoporotic medaka larvae

The antiresorptive activity of aMG was tested on Rankl induced osteoporosis larvae subline c7 (Rankl fish/Rankl larvae), which exhibit mineralized bone damage in the neural arches of vertebrae upon heat shock [31]. Based on the acute toxicity results, aMG concentrations of 0.1, 0.25, 0.5, 0.75, and 1.0 µM (corresponding to approximately 1/30–1/3 of the LC₅₀ value) were chosen for these experiments. Rankl larvae at 7 dpf were pretreated with aMG at the designated doses in a final concentration of 0.1% DMSO; heat shock was applied at 9 dpf, and bone staining was performed at 11 dpf (Fig. 4A). Positive and solvent control groups underwent the same procedure but received 50 µg/ml alendronate or 0.1% DMSO, respectively. A wild-type group subjected to heat shock without treatment was also included as a biological control.

Fig. 4. Protective effects of α-mangostin (aMG) against Rankl-induced bone loss in medaka larvae.

(A) Schematic timeline of the experimental protocol. Treatment with aMG or alendronate (Alen) was initiated at 7 days post-fertilization (dpf). The larvae were fixed and stained for bone mineralization analysis at 11 dpf. (B) Representative lateral view images of the first 15 mineralized vertebrae in larvae at 11 dpf. Experimental groups include: Wild-type non-induced control (WT, -Rankl), Rankl-induced vehicle control (DMSO 0.1%, +Rankl), positive control (Alen 50 µg/ml, +Rankl), and aMG-treated groups (0.1, 0.25, 0.5, 0.75, 1 µM; +Rankl). Arrowheads indicate neural arch integrity: black (intact), red (severely damaged/lost), and green (slightly damaged). Scale bar = 0.25 mm. (C) Quantification of the Mineralization Index (I_M, measured in pixels) across all groups. The sample sizes (n) are listed directly below the x-axis for each corresponding group. (D) Mineralization Protection Index (I_P, %) calculated for effective doses of aMG and Alen compared to the DMSO control. Statistical analysis: Data are presented as mean ± standard deviation. Significant differences compared to the DMSO 0.1% (+Rankl) group are denoted as *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001. Rankl, receptor activator of nuclear factor kappa B ligand; DMSO, dimethyl sulfoxide; ns, not significant. Statistical comparisons were determined using one-way ANOVA followed by Tukey’s post-hoc test.

Results showed that aMG reduced Rankl-induced mineralized bone damage in the neural arches of Rankl larvae (Fig. 4). While wild-type larvae displayed intact mineralized neural arches (black arrowheads, Fig. 4B), Rankl-induced larvae (+Rankl) exhibited varying degrees of damage (red and green arrowheads, Fig. 4B). In contrast, less damage was observed in Rankl larvae treated with either alendronate or aMG (green arrowheads in Fig. 4B). Statistical analysis of the I_M - defined as the total length of mineralized neural arches in the first 15 vertebrae per larva [31] revealed that aMG significantly reduced bone damage at doses of 0.25, 0.5, 0.75, and 1.0 µM (I_M values: 2,502.66, 2,475.21, 2,304.27, and 2,253.93, respectively), compared to the vehicle control group (0.1% DMSO, I_M = 1,720.56), with significance levels of p < 0.0001, p < 0.001, p < 0.01, and p < 0.05, respectively (Fig. 4C). Similarly, treatment with alendronate (50 µg/ml) resulted in a significantly higher I_M (2,523.89 vs. 1,720.56; p < 0.0001), confirming its protective effect (Fig. 4C). The index of protection (I_P ; see above for I_P formula), which reflects the percentage of mineralized bone preserved by treatment, was calculated at each effective dose. I_P analysis showed that aMG at 0.25 and 0.5 µM provided protective effects comparable to those of alendronate (50 µg/ml), with I_P values of 24.7%, 23.8%, and 25.3%, respectively (Fig. 4D).

DISCUSSION

Osteoporosis is a growing global health concern that necessitates the continuous exploration of novel therapeutic agents, particularly those derived from natural sources. G. mangostana, commonly known as mangosteen, has garnered significant attention for its rich array of xanthones, with aMG being the most abundant and extensively studied. aMG is renowned for its diverse pharmacological activities, including antioxidant, anti-inflammatory, antimicrobial, and anticancer properties, positioning it as a compound with considerable pharmaceutical potential [9]. However, despite extensive research into aMG's various bioactivities, its specific effects on bone health, particularly its osteogenic and antiresorptive capabilities, remain less comprehensively explored. Furthermore, while aMG-based products have been increasingly consumed and the therapeutic potential of aMG has gained growing interest, its toxicity profile has not been adequately studied [40].

In the present study, aMG was successfully isolated from G. mangostana peel collected in Vietnam using ultrasound-assisted extraction, yielding 64.15 g of crude extract from 2 kg of dried material. Subsequent silica gel chromatography afforded aMG with a 30% yield. With an HPLC purity of 98.3%, the isolated product fully met the analytical standards. The chemical structure of aMG was confirmed by ESI-HRMS and detailed ¹H- and ¹³C-NMR analyses. This extraction-based approach provides a cost-effective, reproducible, and scalable method, offering a reliable local natural source for further derivatization and biological evaluation.

Following isolation, the in vitro osteogenic effects of aMG were evaluated using MC3T3-E1 pre-osteoblasts, while its antiresorptive capacity was assessed in vivo using a medaka fish model of osteoporosis. Concurrently, the in vitro and in vivo acute toxicities of aMG were also investigated.

aMG’s potent osteogenic and antiresorptive effects

In healthy bones, a delicate equilibrium exists between bone formation (osteogenesis) by osteoblasts and bone resorption by osteoclasts. Imbalances, such as increased bone resorption or decreased bone formation, commonly lead to osteoporosis and osteopenia [1]. An effective agent for bone health would either promote bone formation (osteogenic action) or inhibit bone resorption (antiresorptive action), or ideally, both [18].

Bone formation begins with the proliferation and differentiation of osteoprogenitor cells into osteoblasts, which produce the organic matrix required for subsequent mineral deposition [41]. Functional indicators of osteogenesis, including cell proliferation, ALP activity, collagen synthesis, and calcium mineralization, represent the sequential stages of osteoblast development, from early expansion to matrix maturation and final mineral deposition [41]. Our in vitro investigations into the osteogenic effects of aMG using MC3T3-E1 pre-osteoblast cells provided valuable insights. aMG at the concentrations of 1 µM and 10.0 µM significantly enhanced all four functional osteogenic markers: cell proliferation (e.g., 117.73% at 10.0 µM, p < 0.01), ALP activity (e.g., 130.11% at 1.0 µM, p < 0.01), collagen synthesis (114.13% at 1.0 µM, p < 0.05), and calcium mineralization (e.g., 124.64% at 10 µM, p < 0.01) (Fig. 2). These coordinated improvements across all major stages of osteoblast maturation support the osteogenic potential of aMG and align with the expected downstream outcomes of osteogenic transcription factors, such as Runx2, Osterix, and bone-matrix-related markers, including osteocalcin (OCN) [42]. Although classical osteogenic biomarkers such as Runx2 and OCN were not assessed in this study, the observed functional improvements in ALP, collagen, and mineralization collectively suggest a pro-osteogenic effect. Future studies incorporating transcriptional or protein-level analyses of osteogenic markers are needed to substantiate this conclusion and clarify upstream regulatory events.

The RANKL/RANK pathway promotes osteoclast formation and bone resorption and is a common target of antiresorptive drugs [43]. This study evaluated the bone-protective effect of aMG in a Rankl-induced medaka model that mimics human osteoporosis condition through heat-shock-induced Rankl expression. This Rankl fish model has been successfully employed to identify several compounds with antiresorptive activity [31,44,45]. Our findings demonstrate that aMG possesses significant antiresorptive potential. It reduced bone damage in the neural arches of Rankl larvae at all five tested concentrations, with statistically significant reductions observed at four doses (0.25, 0.5, 0.75, and 1.0 µM). The bone protective I_P values (24.7% and 23.8%) of the most effective doses of 0.25 and 0.5 µM, respectively, were comparable to that of alendronate (25.3% at 50 µg/ml) (Fig. 4), an established positive control bisphosphonate used in this fish model [31,45,46]. Our results are aligned with previously published in vitro and in vivo data. In 2017, Hong et al. [27] showed that aMG inhibited RANKL-induced osteoclastogenesis in RAW264.7 cells by suppressing ERK and JNK signaling at 0.4 and 0.8 µM. In 2022, Zhang et al. [28] reported that aMG at 2.0 μM suppressed osteoclast formation in RANKL- and M-CSF-induced bone marrow-derived macrophages (BMMs) and reduced bone loss in an lipopolysaccharide-induced osteolysis mouse model by inhibiting NF-κB and MAPK pathways.

Although our study did not measure the transcriptional or protein-level changes in these signaling molecules in medaka, the phenotypic rescue strongly suggests that aMG modulates conserved osteoclastogenic pathways. The absence of mRNA or protein data represents a limitation of this study, and future work should include PCR or immunodetection of RANKL-related signaling components to clarify the underlying mechanism in this fish model.

Taken together, our findings demonstrate that aMG exerts both osteogenic and antiresorptive actions, suggesting a dual mode of activity beneficial for restoring bone homeostasis. These dual effects are advantageous for osteoporosis therapy [18]. Importantly, this is the first study to provide in vivo evidence from a fish model supporting the bone-protective effect of aMG, offering essential preclinical data to support its further evaluation as a potential therapeutic candidate for osteoporosis.

Toxicity profile of aMG

Owing to its widespread use in commercial products and promising therapeutic applications, toxicity studies of aMG are essential. However, current rodent-based data are limited and inconsistent, often influenced by variations in animal models, administration routes, and experimental design. For example, a study by Nelli et al. [47] investigating anti-diabetic effects of aMG reported no signs of toxicity or mortality in rats after oral administration of doses up to 2,000 mg/kg over 48 h. In stark contrast, Choi et al. [48] observed mortality in mice following intraperitoneal injection, reporting an LC₅₀ of 150 mg/kg after 72 h. These discrepancies highlight the urgent need for further research to establish a robust safety profile for aMG.

In our in vitro experiments using MC3T3-E1 pre-osteoblast cells, aMG exhibited cytotoxicity at high concentrations, most notably at 100 μM and 1,000 μM. This underscores the importance of identifying safe, non-toxic concentrations that still yield therapeutic effects. To this end, we conducted acute toxicity testing of aMG in wild-type medaka embryos and larvae to determine suitable concentrations for in vivo studies. Teleost fish models such as medaka and zebrafish are widely recognized for their utility in toxicity testing due to their genetic similarity to vertebrates, external development, and transparency [29,49].

Our medaka data demonstrated that aMG caused dose-dependent mortality and developmental abnormalities (Fig. 3). Importantly, aMG was significantly more toxic to larvae (LC₅₀: 3.44 μM) than to embryos (LC₅₀: 91.27 μM), likely due to the protective barrier of the chorion of the embryos and the increased uptake and metabolism in free-swimming, feeding larvae [29]. The teratogenic effects included tail malformation, hemorrhage, immobilization, and reduced eye pigmentation. At higher concentrations (> 60 μM), aMG also reduced heart rate, and at > 12 μM, impaired hatching success. The teratogenic index of aMG was approximately 1 in both embryos and larvae (Table 1), indicating that mortality was primarily associated with developmental defects. These findings are consistent with those of previous zebrafish studies. Kittipaspallop et al. [50] reported an LC₅₀ of 5.75 μM in embryos after 3 dpe, along with axis malformation, pericardial and yolk sac edema, and cardiac defects. Similarly, Pimtong et al. [39] observed an even lower LC₅₀ of 1.48 μM in zebrafish larvae after 5 days, with additional liver-specific toxicity. The similarity in the toxic effects across both medaka and zebrafish models supports a conserved mechanism of aMG toxicity during early vertebrate development. Interestingly, aMG appeared more toxic to medaka larvae (LC₅₀: 3.44 μM, Table 1) than to MC3T3-E1 cells, where 10 μM promoted proliferation and differentiation (Fig. 2). However, previous studies have also found that aMG induces cytotoxicity in other cell types, killing 64.94% of BMMs at 4 μM [28] and affecting RAW264.7 macrophages at concentrations ≥ 3.2 μM [27], highlighting variable cell sensitivity.

Taken together, despite its promising pharmacological effects, aMG exhibits significant toxicity in the medaka model at low micromolar concentrations. These findings are crucial for defining a safe therapeutic window and underscore the importance of using fish models for preclinical safety evaluations.

In this study, aMG was successfully isolated from Vietnamese mangosteen and demonstrated, for the first time, to exhibit dual bone-protective activities. In vitro, aMG promoted osteoblast differentiation and matrix mineralization in MC3T3-E1 cells, while in vivo, it exerted significant antiresorptive effects in a medaka osteoporosis model, comparable to a clinical reference compound and consistent with previous rodent findings. In addition, acute toxicity testing in medaka embryos and larvae provided new insights into developmental stage-dependent sensitivity and contributed to defining safe exposure ranges. Together, these results establish aMG as a promising natural compound for bone health, supporting its further evaluation in mechanistic and translational research.

SUPPLEMENTARY MATERIALS

Supplementary data including four figures can be found with this article online at https://doi.org/10.4196/kjpp.25.370