Zanthoxylum schinifolium extracts enhance 3T3-L1 adipocyte differentiation via CHOP inhibition and PPARγ activation

ABSTRACT

Zanthoxylum schinifolium (ZS), which is widely used as a seasoning and traditional medicine in East Asia, has demonstrated pharmacological potential. This study investigated the effects of the leaf and twig extracts of ZS (LZSE and TZSE, respectively), which are native to the Honam region of Korea, on adipocyte differentiation and assessed the ligand-binding energy score of their components to bind peroxisome proliferator-activated receptor gamma (PPARγ), a critical regulator of adipogenesis and metabolic health. LZSE and TZSE were prepared using 70% ethanol, and their molecular effects on adipocyte differentiation were evaluated in 3T3-L1 preadipocytes. Single compounds from the extracts were identified using UPLC-ESI-Q-TOF-MS, and their ligand-binding energy scores were calculated via in silico molecular docking studies. PPARγ activity was further confirmed through reporter assays. LZSE and TZSE significantly promoted adipocyte differentiation, as demonstrated by morphological changes and the increased mRNA and protein levels of key adipogenic and lipogenic genes, such as PPARγ and CCAAT-enhancer-binding protein alpha (C/EBPα). LZSE specifically enhanced adipogenesis without inducing cytotoxicity, attributed to the inhibition of C/EBP homologous protein (CHOP) and stimulation of mitotic expansion. Additionally, UPLC-ESI-Q-TOF-MS identified several active compounds in LZSE and TZSE, and in silico docking revealed the high binding affinity of these compounds for the full-agonist ligand-binding domain of PPARγ. LZSE and TZSE could emerge as novel antidiabetic drug candidates based on their effects on adipocyte differentiation and PPARγ activation. Furthermore, the active compounds identified in these extracts hold promise as tentative PPARγ agonists, highlighting their therapeutic potential in the treatment of metabolic disorders.

1. Introduction

Obesity is widely recognized as one of the most serious global metabolic conditions globally that affecting millions of people and imposes a significant burden on healthcare systems. Its prevalence, together with its association with numerous comorbidities, underscores its critical status in society (Must and McKeown 2000; Cawley et al. 2021). According to research, obesity is influenced by the complex interplay of factors, including dietary habits, genetic predispositions, underlying health conditions and sociodemographic factors such as sex and ethnicity (Hüls et al. 2021; Górczyńska-Kosiorz et al. 2024). Fundamentally, obesity arises from an imbalance between energy intake and expenditure, leading to excessive fat accumulation (Ludwig et al. 2022). While various indigenous plants have been traditionally used as medicines, functional foods, and detoxifying agents, relatively few extensive studies have investigated their bioactive compounds at the molecular level. These compounds have potential therapeutic roles at the molecular level, targeting diseases such as diabetes, cancer, and metabolic disorders (Kim et al. 2022; Lee et al. 2024; Pak et al. 2024). Furthermore, natural extract-based supplements and cosmetics are commercially available, offering new avenues for healthcare applications. Many of these extracts have been refined into single-compound formulations for treating specific diseases, expanding their applicability (Wang et al. 2021; Dar et al. 2023; Hyun et al. 2024).

Adipocyte differentiation plays a pivotal role in obesity management, as it governs the body’s capacity for fat storage and metabolism. Studies on the regulation of adipocyte differentiation have insights into the potential of these cells as therapeutic targets (Jakab et al. 2021; Tang et al. 2021; Kan et al. 2022; Yang et al. 2024). Adipocytes, beyond serving as energy storage units, act as endocrine cells, regulating systemic energy homeostasis through hormone secretion. Given the fundamental functions of adipocytes, pharmacological interventions targeting healthy, newly formed adipocyte differentiation may alleviate obesity-related diabetes by reducing circulating free fatty acids (Dubois et al. 2006; Wang et al. 2008; Kim et al. 2015), improving insulin sensitivity (Haliakon et al. 1997; Nadler et al. 2000; Hu et al. 2022) and mitigating metabolic dysfunctions such as type 2 diabetes (Danforth 2000; Guilherme et al. 2008; Hammarstedt et al. 2018). Notably, agonists for peroxisome proliferator-activated receptor gamma (PPARγ), such as thiazolidinediones, exemplify this approach by enhancing adipogenesis while mitigating hyperglycemia and inflammation (Ahmadian et al. 2013). Adipogenesis in 3T3-L1 cells involves a multistep process characterized by mitotic clonal expansion (MCE) (Chen et al. 2017; Go et al. 2021; Tang et al. 2003; Kim et al. 2016), transcription factor activation and lipid accumulation. Key players in this process include PPARγ and CCAAT-enhancer-binding proteins (C/EBPs) (Wu et al. 1999; FreemanFJ and Spiegelman 2002; Guo et al. 2015) which are regulated by stress-response markers, including C/EBP homologous protein (CHOP), as previously reported (Tang and Lane 2000; Li et al. 2006; Yoon et al. 2022). In particular, PPARγ acts as a master regulator, orchestrating the expression of genes involved in adipogenesis and lipogenesis (Gerhold et al. 2002; Rizzatti et al. 2013; Chen et al. 2021).

Zanthoxylum schinifolium (SANCHO, ZS), a plant indigenous to East Asia, is widely recognized for its culinary and medicinal applications. Studies have identified various bioactive properties of ZS, including anti-bacterial (Choi et al. 2012), anti-cancer (Li et al. 2013), and anti-inflammatory effects (Lee et al. 2009), attributable to its rich composition of antioxidants and lipids such as oleic and palmitic acids (Kim et al. 2021). Specifically, 100% ethanol or 100% methanol ZS leaf extracts inhibit adipogenesis by modulating the PI3K/AKT signaling pathway and reactive oxygen species (ROS) generation (Choi et al. 2015; Choi et al. 2024). In in vivo models, n-hexane ZS seed oil has been shown to alleviate hyperlipidemia via the PKA pathway (Kim et al. 2016). Additionally, Yeon et al. and Kong et al. demonstrated that a 100% ethanol extract of ZS from Sichuan, China exerted anti-obesity effects in high-fat diet (HFD)-induced obese mice (Kong et al. 2013; Yeon et al. 2021). However, the pharmacological properties of natural plant extracts can vary considerably depending on several factors, including the extraction solvent – which determines the profile of bioactive constituents (Ngo et al. 2017; Abubakar and Haque 2020) – as well as geographic origin (Liu et al. 2018; Delporte et al. 2021), seasonal variations (Mwamatope et al. 2021; Ramasar et al. 2022), the plant parts used (Muanda et al. 2011; Pudziuvelyte et al. 2020; Kim et al. 2022), and other environmental factors (Muanda et al. 2011; Pudziuvelyte et al. 2020; Kim et al. 2022). Additionally, differences in adipogenesis induction protocols, such as variations in the composition and concentration of differentiation media, and the timing and duration of treatment, can influence the observed extent of adipocyte differentiation.

In the case of ZS, previous investigations have employed various plant parts (seeds, fruits, and whole plant), different extraction solvents (ethanol, methanol, and n-hexane), and most studies have employed ZS samples from various geographic origins, often commercially sourced or obtained from China, introducing variability that may underlie inconsistencies in reported pharmacological effects. Furthermore, studies have reported that the variation in ZS plant components based on their origin involves diverse environmental factors (Cho et al. 2003; Kim et al. 2022).

Notably, the regulatory effects of a 70% ethanol extract of ZS obtained from plants native to the Honam region of Korea have not been previously investigated. In this study, we examined the effects of ZS leaf (LZSE) and twig (TZSE) extracts on adipocyte differentiation. Both extracts were obtained using 70% ethanol from ZS plants collected in the Honam region. Furthermore, mechanistic insights were explored using UPLC-Q-TOF-MS and molecular docking analysis. Our findings suggest that both extracts promote adipogenesis via CHOP suppression and PPARγ activation, indicating their potential as novel therapeutic agents for metabolic disorders.

2. Materials and methods

2.1. Leaf or twig Zanthoxylum schinifolium (ZS) extracts

The ethanol extracts of Zanthoxylum schinifolium leaves (LZSE) or twigs (TZSE) were obtained from the Honam National Institute of Biological Resources (Mokpo, Republic of Korea). The deposit numbers are HNIBR-NP153 (LZSE) and HNIBR-NP174 (TZSE). The plant materials were collected from Pangeumdo, Shinan-gun, Jeollanam-do, Republic of Korea (Honam region). Leaves and twigs (50 g each) were incubated with 500 mL of 70% ethanol:water (7:3, v/v) 150 rpm, at 18°C for 24 h. The mixture was filtered and evaporated on rotary evaporator at 50°C to remove the ethanol and water and then lyophilized.

2.2. Cell culture and adipocyte differentiation

3T3-L1 cells were purchased from the American Type Culture Collection (ATCC, Manassas, VA). 3T3-L1 preadipocytes were maintained in Dulbecco’s modified Eagle’s medium (DMEM; Welgene, Daegu, Republic of Korea) supplemented with 10% bovine calf serum (BCS; Welgene Inc.), 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B in 0.85% NaCl at 37°C with 5% CO2. To induce adipocyte differentiation, 3T3-L1 preadipocytes were seeded at a density of 5 × 10⁴ cells/mL. The cells reached confluency after 24 h and they were incubated for an additional 48 h to achieve an overconfluent state (Day 0). Overconfluent 3T3-L1 preadipocytes were treated with a 0.2 × MDI induction cocktail containing 0.1 mM 3-isobutyl-1-methylxanthine (Sigma-Aldrich, St. Louis, MO), 0.2 μM dexamethasone (Sigma-Aldrich), and 0.2 μg/mL insulin (Sigma-Aldrich) in DMEM supplemented with 10% fetal bovine serum (FBS; Welgene). On Day 2, the culture medium was replaced with a 0.2 × FI induction medium consisting of 10% FBS, 0.2 μg/mL insulin. Starting on Day 4, cells were maintained in 10% FBS-containing medium. The medium was replaced with the respective induction medium on Day 6 (Shrestha et al. 2024). The HEK293a cell line (KCLB, Seoul, Republic of Korea) was cultured by DMEM supplemented 10% FBS, 100 U/mL penicillin, 100 μg/mL streptomycin, and 0.25 μg/mL amphotericin B at 37°C with 5% CO2. Each growth medium of cells was replaced every 2 days and split when confluence reached 70–80% (Le et al. 2024).

2.3. Cell viability

Cell viability was assessed using the WST-8 Cell Viability Assay Kit (BIOMAX, Guri, Republic of Korea) following the manufacturer’s instructions. 3T3-L1 cells were seeded at a density of 1 × 10⁴ cells/well in a 96-well plate. The cells reached approximately 70–80% confluency after 24 h. The medium was then replaced with fresh culture medium (10% BCS + DMEM) containing various concentrations of LZSE (50, 100, 200 and 300 μg/mL) or DMSO (control group), followed by incubation for 24 h. Then, each well was treated with 10% (v/v) WST-8 reagent for 1 h and the absorbance of the medium, including the formazan product, was measured at 450 nm using an iMark™ microplate reader (Bio-Rad Laboratories, Hercules, CA). Cell viability (%) was calculated relative to the absorbance of the control group, which was set at 100% (Eom et al. 2022).

2.4. Oil red O staining

Oil Red O staining was performed to evaluate lipid droplet formation during adipocyte differentiation. Cells were fixed with 4% formaldehyde, rinsed with phosphate-buffered saline (PBS), and treated with 100% propylene glycol at 30 rpm for 30 min twice. Subsequently, the cells were incubated with Oil Red O solution at room temperature for 15 min, followed by replacement with 85% propylene glycol and additional incubation for 1 min. The cells were rinsed with PBS until all Oil Red O debris was removed. The stained cells were observed under an NIB400 microscope (Nexcope, Ningbo, China) at ×100 magnification (Shrestha et al. 2024).

2.5. Quantitative real-time PCR

Total RNA was extracted from 3T3-L1 cells using Rio EX reagent (GeneAll, Seoul, Republic of Korea) following the manufacturer’s protocol and quantified using a Nano-Photometer N60 spectrophotometer (Implen, CA, USA). For qRT-PCR, reverse transcription of total RNA was performed using the ReverTra Ace qPCR RT Kit (Toyobo, Osaka, Japan) according to the manufacturer’s instructions. In total, 1 μg of RNA was used for cDNA synthesis with the T100 Thermal Cycler (Bio-Rad). Relative mRNA expression levels were quantified using the CFX Connect Real-Time PCR System (Bio-Rad) and real-time PCR Master Mix containing SFCgreen I (BIOFACT, Daejeon, Republic of Korea). The expression levels of the target genes were normalized to β-actin or 36B4 using the comparative CT method. Most of the primer sequences used in the experiment are listed in our previous studies (Yeon et al. 2021; Eom et al. 2022; Le et al. 2024; Yeon et al. 2024), and the remaining primer sequences are as follows: β-actin, (F) 5′-GGC TGT ATT CCC CTC CAT CG-3′ and (R) 5′-CCA GTT GGT AAC AAT GCC ATG-3′; Glut4, (F) 5′-CAT TGC TTC TGG CTA TCA C-3′ and (R) 5′-AAC ATG CTG GTT GAA TAG TAG-3′.

2.6. Western blotting

Proteins were extracted with Mammalian Protein Extraction Reagent (Thermo Fisher Scientific, Waltham, MA, USA) containing the Xpert Protease Inhibitor Cocktail (GenDEPOT, Baker, TX, USA) and purified by centrifugation at 14,000 rpm for 15 min at 4°C. Protein concentrations were quantified using the Pierce™ BCA Protein Assay Kit (Thermo Fisher Scientific) according to the manufacturer’s protocol. The protein samples were separated by 10% SDS-PAGE and transferred onto PVDF membranes (Merck, Darmstadt, Germany). The membranes were blocked with 5% BSA in 0.2% TBS-T solution for 1 h at room temperature on a shaker and rinsed three times with TBS-T. Subsequently, the membranes were incubated overnight at 4°C on a shaker with the following primary antibodies: anti-PPARγ antibody (Cat. No. 2443, Cell Signaling Technology, Danvers, USA), anti-FABP4 antibody (Cat. No. SC-271529, Santa Cruz Biotechnology, Dallas, USA), anti-GAPDH antibody (Cat. No. SC-365062, Santa Cruz Biotechnology), anti-HSP90 antibody (Cat. No. SC-13119, Santa Cruz Biotechnology), anti-PI3 Kinase p85 antibody (Cat. No. 4292, Cell Signaling Technology), anti-Phospho-PI3 Kinase p85(Tyr485)/p55 (Tyr199) antibody (Cat. No. 4228, Cell Signaling Technology), anti-ERK antibody (Cat. No. 91025S, Cell Signaling Technology), anti-phospho ERK antibody (Cat. No. 4370S, Cell Signaling Technology), anti-CHOP antibody (Cat. No. 2895, Cell Signaling Technology), anti-C/EBPα antibody (Cat. No. sc-365318, Santa Cruz Biotechnology), and anti-C/EBPβ antibody (Cat. No. sc-7962, Santa Cruz Biotechnology). After incubation, the membranes were washed three times with 0.2% TBS-T and subsequently incubated for 1 h at room temperature with horseradish peroxidase-conjugated secondary antibodies (mouse or rabbit IgG, 1:5000 dilution, Bio-Rad). Immunoreactive bands were visualized using ECL reagent (Bio-Rad) and captured with an iBright CL1500 imaging system (Thermo Fisher Scientific).

2.7. Luciferase and β-galactosidase assay

To evaluate the PPARγ activity of LZSE and TZSE, a luciferase assay was conducted alongside a β-galactosidase assay to normalize the transfection efficiency, as previously described. Following 24 h of transfection, the cells were treated with DMSO, rosiglitazone (10 μM), LZSE, or TZSE at concentrations of 30, 100, and 300 μg/mL for 24 h. Cell lysates were then prepared and subjected to both luciferase and β-galactosidase assays. The luciferase activity levels were normalized against the absorbance values obtained in the β-galactosidase assay corresponding to each sample (Le et al. 2024).

2.8. Flow cytometry assay

To evaluate changes in the cell cycle distribution, 3T3-L1 cells were dissociated using trypsin – EDTA solution (Welgene), washed with DPBS and fixed in 70% ethanol at −20 °C overnight. Fixed cells were stained with 2 μg/mL propidium iodide (PI; Cat. No. P3566, Thermo Fisher Scientific), washed twice with DPBS and treated with RNase A (Cat. No. PE290-25 h, BIOFACT). The fluorescence intensity of PI and cell counts were analyzed using a CytoFLEX flow cytometer (Beckman Coulter, Brea, CA, USA) with Kaluza analysis software (Beckman Coulter). The results are presented as histograms for each dataset, generated using the J.V. Watson cell cycle phase algorithm in Kaluza analysis software.

2.9. UPLC-ESI-Q-TOF-MS

The ZS samples were dissolved in LC – MS grade water and analyzed by UPLC-Q-TOF-MS on a Waters ACQUITY UPLC I-Class Plus System coupled with a Xevo G2-XS mass spectrometer (Waters Corporation, Milford, MA, USA). An ACQUITY UPLC BEH C18 column (2.1 × 100 mm; 1.7 μm; Waters, Milford, MA, USA) was used as the stationary phase and 0.1% formic acid in water (A) and 0.1% formic acid in acetonitrile (B) were used as the mobile phases (gradient mode: 0–7 min, B 7–15%; 7–10 min, B 15–15%; 10–10.5 min, B 15–26%; 10.5–14.5 min, B 26–28%; 14.5–15 min, B 28–36%; 15–18 min, B 36–44%) with a flow rate of 0.2 mL/min and an injection volume of 1 μL. Mass spectrometric analysis was conducted using an electrospray ionization source in the negative ionization mode. The optimized parameters for the analysis included a cone voltage of 40 V, desolvation temperature of 350°C, and mass range of m/z 50–1,500. Calibration was performed with 0.5 mM HCOONa and mass correction was applied using 0.2 ng/mL [Leu5]-enkephalin.

2.10. Molecular docking

To evaluate the ligand-binding energy scores of the constituents of LZSE and TZSE identified as potential PPARγ agonists, we followed a detailed protocol referenced from a previous dissertation (Le et al. 2024). The PPARγ structure (PDB ID: 4EMA) was obtained from the RCSB Protein Data Bank and the 3D conformer structures of the ligands, including rosiglitazone (CID: 77999), were downloaded from PubChem. The preparation of PPARγ and ligands using MGL Tools 1.5.7. The PPARγ protein structure was processed by removing water molecules, adding polar hydrogen atoms, and assigning Koll-man charges. Ligands, initially obtained as SDF files, were converted to the PDBQT format using Open Babel. These converted ligands were further prepared by adding Gasteiger charges and optimizing torsion for the AutoGrid setup. The grid box for docking was designed to encompass key amino acids within the full-agonist ligand-binding domain (LBD), including Ser289, His323, His449, and Tyr473. Molecular docking simulations were performed using AutoDock Vina 1.1.2. The interactions between ligands and PPARγ, including the chemical bonds formed, were visualized and analyzed using Discovery Studio 2021.

2.11. Statistical analysis

All results are presented as the mean ± standard error of the mean (SEM), while n indicates the number of culture wells analyzed. Comparisons between two treatment groups were performed using the unpaired two-tailed Student’s t-test with Prism 9 software (GraphPad Software, California, USA). For comparisons involving more than two groups, a two-way ANOVA was conducted using the same software. p < 0.05 was considered statistically significant for all tests.

3. Results

3.1. LZSE stimulates 3T3-L1 adipocyte differentiation without cytotoxicity

To evaluate the cytotoxicity of LZSE, 3T3-L1 preadipocytes were exposed to various concentrations (50, 100, 200, and 300 μg/mL) of LZSE or DMSO as a control. As presented in Figure 1A, LZSE did not induce cytotoxicity, even at the highest concentration (300 μg/mL). On the contrary, cell viability rates significantly increased in the LZSE-treated groups, reaching approximately 102.9 ± 3%, 106.9 ± 2.21%, 108.7 ± 1.96%, and 111.3 ± 1.7% in the presence of LZSE concentrations of 50, 100, 200, and 300 μg/mL, respectively. To investigate whether LZSE promotes adipocyte differentiation, 3T3-L1 preadipocytes were treated with various concentrations of LZSE (50–200 μg/mL) or DMSO in 0.2 × adipogenesis induction medium. The number of differentiated adipocytes, indicated by lipid droplet accumulation, increased in the presence of LZSE in a dose-dependent manner compared with the control group (Figure 1B). These results suggest that LZSE promotes adipocyte differentiation by enhancing lipid droplet formation and maintaining cell viability in a dose-dependent manner. Next, to investigate the molecular mechanisms underlying these effects, the expression of key adipogenic and lipogenic genes was analyzed using quantitative real-time PCR (qRT-PCR) and western blotting. Adipogenic genes, including PPARγ, C/EBPβ, C/EBPα, FABP4, and adiponectin, were significantly upregulated by LZSE (200 μg/mL), with fold increases of 2.4 ± 0.13, 2 ± 0.05, 2.5 ± 0.21, 2.3 ± 0.14, and 5.7 ± 0.29, respectively (Figure 1C). Similarly, the expression of lipogenic genes such as ACC, FAS, SCD1, and SCD2 was also elevated, with fold increases of 2.3 ± 0.14, 2.1 ± 0.21, 5.7 ± 0.29, and 3.4 ± 0.21, respectively (Figure 1D). Consistent with these findings, the level of PPARγ and FABP4 protein was increased by LZSE in a concentration-dependent manner (Figure 1E and 1F). Taken together, these results suggest that LZSE promotes adipocyte differentiation by upregulating the expression of adipogenic and lipogenic genes and enhancing lipid droplet formation, as evidenced by both molecular and morphological changes.

Figure 1.

Effects of LZSE on 3T3-L1 adipocyte differentiation. Each indicated variable dose of LZSE or DMSO (0 µg/mL) was administered according to the respective experimental scheme. (A) The viability of 3T3-L1 preadipocytes treated with various concentrations of LZSE (0, 50, 100, 200, and 300 µg/mL) was measured using a WST-8 assay kit. (B) Oil Red O staining was performed on 3T3-L1 cells differentiated in induction medium containing 0.17× MDI (3-isobutyl-1-methylxanthine, dexamethasone and insulin) in the absence (0 µg/mL) or presence of LZSE (50, 100, and 200 µg/mL). Scale bar, 200 µm; magnification, × 100. (C, D) Quantitative real-time PCR was performed to measure the mRNA expression of adipogenic and lipogenic genes after 6 days of differentiation in the absence (0 µg/mL) or presence of LZSE. (E) Western blot analysis was performed to measure PPARγ and FABP4 protein expression levels. HSP90 was used as a loading control. (E, F) Quantification of PPARγ and FABP4 protein levels from panel. All data are presented as the mean ± SEM from two independent experiments (n = 4). *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus the control group.

3.2. The promoting effects of LZSE are significant at an early stage of adipocyte differentiation

Adipocyte differentiation can be functionally divided into three stages: early, middle and late. To determine the specific stage at which LZSE exerts its promoting effect, treatments were conducted according to the experimental scheme shown in Figure 2A. A moderate dose of LZSE (100 μg/mL) was used to assess the extent of its effect. As in previous experiments, Oil Red O staining was performed on fixed cells from each differentiated group to evaluate lipid droplet formation. As shown in Figure 2B, adipocytes containing lipid droplets, an indicator of adipocyte differentiation, were more prevalent in the groups treated with LZSE during the earlier stages of differentiation. Compared with the fully treated positive control group (2), the group treated only during the earliest stage (3) exhibited a comparable level of differentiation. Conversely, as the treatment period shifted to later stages, the number of differentiated adipocytes gradually decreased. The group treated at the late stage (5) showed a similar level of differentiation as the untreated control group. Additionally, the group treated during the early – middle stage (6) displayed more differentiated adipocytes than the middle – late stage (7). To elucidate the molecular mechanisms underlying these observations, qRT-PCR and western blotting were conducted to measure the expression levels of adipogenic and lipogenic genes. At the mRNA level, the expression of adipogenic genes such as PPARγ, C/EBPβ, C/EBPα, and FABP4 was markedly higher when LZSE treatment occurred earlier. Similarly, the mRNA expression of the lipogenic genes followed the same trend (Figure 2C). At the protein level, the expression of FABP4, a key marker of adipocyte differentiation, was significantly elevated in cells treated during the early stages compared to those treated at later time points (Figure 2D and 2E). Consistent with this, PPARγ protein levels also increased markedly with early-stage treatment, further supporting the role of LZSE in promoting adipogenesis during the early phase of differentiation. Collectively, these findings suggest that the adipogenesis-promoting effects of LZSE are most pronounced during the early stages of differentiation. This highlights the importance of treatment timing in maximizing the pro-adipogenic activity of LZSE.

Figure 2.

Stage-specific effects of LZSE on the differentiation of 3T3-L1 preadipocytes. (A) Schematic representation of the experimental schedule for adipocyte differentiation. Cells were treated with DMSO (group 1) or 100 µg/mL LZSE for various durations, as indicated in green (treatment period). (B) Lipid accumulation was evaluated by Oil Red O staining after 6 days of differentiation. Scale bar, 200 µm; magnification, × 100. (C) mRNA expression levels of adipogenic and lipogenic genes were measured by qRT-PCR after 6 days of differentiation. Data are presented as the mean ± SEM from two independent experiments (n = 4). (D) Protein expression levels of PPARγ and FABP4 were analyzed by Western blotting. Both HSP90 and GAPDH were used as loading controls. (E) Quantification of relative PPARγ and FABP4 protein levels using ImageJ software. Data are presented as the mean ± SEM from two independent experiments (n = 3). *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus the control group (1) and #, p < 0.05; ##, p < 0.01; ###, p < 0.001 versus the LZSE-treated group for 6 days (2).

3.3. LZSE regulates adipocyte differentiation in the early stage by repressing CHOP expression and promoting mitotic clonal expansion

Adipocyte differentiation involves multiple factors that coordinate adipogenesis and subsequent lipogenesis. The initial steps in adipocyte differentiation are primarily driven by the activation of C/EBPα and C/EBPβ, which regulate the expression of PPARγ, a master transcription factor of adipogenesis (Darlington et al. 1998; Farmer 2006). However, C/EBP homologous protein (CHOP) plays a critical inhibitory role during early adipogenesis by suppressing C/EBPα and C/EBPβ activation and inducing cell stress responses, such as growth arrest and ER stress (Yang et al. 2017). Once CHOP expression is suppressed, C/EBPα and C/EBPβ activation facilitates the expression of PPARγ, which binds to peroxisome proliferator response elements (PPREs) on adipogenic and lipogenic gene promoters, driving differentiation (Madsen et al. 2014). To investigate the detailed mechanism by which LZSE promotes adipogenesis, LZSE (100 μg/mL) or DMSO was applied for 0, 24, and 48 h as illustrated in the experimental scheme (Figure 3A). As presented in Figure 3B, CHOP mRNA levels initially decreased in both groups 24 h after induction but increased by 48 h in the control group. However, in the LZSE-treated group, CHOP expression was consistently inhibited, displaying reductions of 0.696 ± 0.043-fold and 0.366 ± 0.020-fold at 24 and 48 h, respectively, compared with those in the control group. Meanwhile, the mRNA levels of C/EBPα and PPARγ were significantly increased in the LZSE group, whereas C/EBPβ levels remained unchanged compared with the control group. These trends were consistent with the protein expression patterns observed by western blotting (Figure 3C and 3D). These findings suggest that LZSE suppresses CHOP expression during early adipogenesis, enabling the upregulation of C/EBPα, which in turn promotes PPARγ expression.

Figure 3.

Regulatory mechanism of LZSE during the early stage of adipocyte differentiation. (A) Schematic representation of the experimental schedule for early-stage adipogenesis following treatment with DMSO (control) or 100 µg/mL LZSE for 0, 24, or 48 h. (B) Relative gene expression levels were measured by qRT-PCR under the indicated conditions during the early differentiation phase. Data are presented as the mean ± SEM from two independent experiments (n = 6). (C) Protein expression of key adipogenic markers was analyzed by western blotting. (D) Relative protein levels of PPARγ and FABP4 were quantified using ImageJ software. Data are presented as the mean ± SEM from two independent experiments (n = 3). (E) Expression of cell cycle-related genes was assessed by qRT-PCR. Data are presented as the mean ± SEM from two independent experiments (n = 6). (F) Cell cycle distribution was assessed by flow cytometry after propidium iodide (PI) staining. The percentage of cells in G1, S and G2/M phases was calculated using the J.V. Watson algorithm. Data are presented as the mean ± SEM from two independent experiments (n = 4). *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus the control group and #, p < 0.05; ##, p < 0.01; ###, p < 0.001 versus the DMSO-treated control group (24 or 48 h).

In addition to transcriptional regulation, mitotic clonal expansion (MCE) during the early stages of adipogenesis plays a critical role in adipocyte differentiation. To explore this, we performed qRT-PCR to analyze the mRNA levels of cell cycle-related genes in the same groups. As illustrated in Figure 3E, the mRNA levels of Cyclin D1, Cyclin E, CDK4, and CDK2 showed significant changes in the LZSE-treated group. Cyclin D1 expression in the LZSE group was slightly increased at 24 h but significantly decreased at 48 h compared with that in the control group. Conversely, Cyclin E, CDK4, and CDK2 expression was higher in the LZSE group than in the control group. Cell cycle analysis further confirmed these findings (Figure 3F). At 24 h, the proportion of G1 phase cells was lower in the LZSE group (66.19 ± 0.70%) compared with the control group (70.58 ± 0.38%). Conversely, the S and G2/M phase populations were significantly increased (10.29 ± 0.30% and 23.50 ± 0.84%, respectively) compared with that in the control group (8.64 ± 0.25% and 20.76 ± 0.59%, respectively). This trend in cell cycle regulation observed at 24 h was sustained at 48 h, although the difference in each cell phase population were statistically slightly lower than those observed at 24 h. Although the precise mechanism by which LZSE accelerates the cell cycle during early adipocyte differentiation remains unclear, these results suggest that LZSE promotes MCE, as evidenced by the altered mRNA expression of cell cycle-related genes and changes in the cell cycle distribution. In summary, LZSE facilitates adipocyte differentiation by suppressing the expression of CHOP, thereby inhibiting key transcription factors involved in adipogenesis, and by accelerating MCE during the early stages of adipocyte differentiation.

3.4. Twig extract also promotes adipocyte differentiation by enhancing PPARγ activity similarly as LZSE

To explore the effects of other parts of ZS, we compared the adipogenic effects of LZSE and TZSE. TZSE stimulated adipogenesis and lipogenesis, as observed morphologically (Figure 4A). At the mRNA level, LZSE and TZSE increased the expression of adipogenic and lipogenic genes compared with the effects of the control (Figure 4B and 4C). Notably, the mRNA expression of adipogenic genes such as PPARγ, C/EBPα was elevated in the TZSE group, surpassing that observed in the LZSE group (Figure 4B). A similar pattern was observed for lipogenic genes, including SCD1, SCD2, FAS, and ACC (Figure 4C). Furthermore, FABP4 protein expression was higher in the TZSE group than in the LZSE group. However, protein levels of PPARγ between the two groups (Figure 4D and 4E). PPARγ, known as the master regulator of adipogenesis, both influences gene expression and functions through its activity. PPARγ binds to PPREs (5′-AGGnCAXAGGnCA-3′) within the promoter regions of target genes, including Glut4, Plin2, FABP4, and PEPCK, to regulate their expression (Kawai et al. 2010; Moreno-Santos et al. 2019; Hernandez-Quiles et al. 2021). To assess the activity of PPARγ and its binding affinity for PPRE sequences, we conducted luciferase assays normalized by β-galactosidase activity. As presented in Figure 4F, the LZSE group exhibited significantly higher luciferase activity than the control group. At higher concentrations (300 μg/mL), both LZSE and TZSE significantly elevated luciferase activity by 5.41 ± 0.26-fold and 3.38 ± 0.09-fold, respectively, compared with the control level. These results indicate that both LZSE and TZSE enhance PPARγ activity and its binding to PPRE sequences, functioning as effective transcriptional regulators. Based on the observed increase in PPARγ activity, we analyzed the mRNA expression levels of PPARγ target genes, including Glut4, Plin2, FABP4, and PEPCK, in fully differentiated adipocytes with rosiglitazone. Additionally, matured adipocytes were further treated with LZSE or TZSE for 48 h. As illustrated in Figure 4G, the LZSE group displayed significantly elevated mRNA expression levels of Glut4 (1.547 ± 0.181-fold), Plin2 (1.498 ± 0.059-fold), FABP4 (2.017 ± 0.240-fold), and PEPCK (3.072 ± 0.498-fold). Similarly, Glut4 (1.397 ± 0.096-fold), Plin2 (1.375 ± 0.079-fold), FABP4 (1.943 ± 0.201-fold), and PEPCK expression (1.713 ± 0.179-fold) was significantly enhanced in the TZSE group. These findings suggest that both LZSE and TZSE promote PPARγ activity, serving as transcriptional regulators of key adipogenic and lipogenic genes during adipocyte differentiation.

Figure 4.

LZSE and TZSE promote PPARγ activity and target gene expression, enhancing adipocyte differentiation. (A) 3T3-L1 cells were differentiated in the presence of DMSO (control), 100 µg/mL LZSE, or TZSE for 6 days. Cells were then fixed with 4% formaldehyde and stained with Oil Red O to visualize lipid accumulation. Scale bar, 200 µm; magnification, × 100. (B and C) mRNA expression levels of adipogenic (B) and lipogenic (C) genes were assessed by qRT-PCR following differentiation in the presence of 100 µg/mL LZSE or TZSE. (D) Protein expression levels of PPARγ and FABP4 were analyzed by Western blotting; HSP90 was used as the loading control. (E) Quantification of PPARγ and FABP4 protein levels was performed using ImageJ software. (F) HEK cells were transfected with the indicated plasmid DNA. After 24 h, cells were treated with DMSO (control), 10 µM rosiglitazone, or varying concentrations of LZSE or TZSE for additional 24 h. Luciferase activity was measured and normalized to β-galactosidase activity. Data are presented as the mean ± SEM from two independent experiments (n = 8). (G) Fully differentiated 3T3-L1 adipocytes induced by 10 µM rosiglitazone were further treated with DMSO (control), rosiglitazone or 100 µg/mL LZSE or TZSE for 48 h. Expression levels of PPARγ target genes were assessed by qRT-PCR and normalized to 36B4. Data are presented as the mean ± SEM from two independent experiments (n = 4). *, p < 0.05; **, p < 0.01; ***, p < 0.001 versus the control group (DMSO) and #, p < 0.05; ##, p < 0.01; ###, p < 0.001 versus the LZSE-treated group.

3.5. Identification of metabolites in two parts of Zanthoxylum schinifolium extracts

The chemical constituents of LZSE and TZSE were analyzed using UPLC-ESI-Q-TOF-MSe. The analysis identified peaks corresponding to phenolic acids, flavanones, lignins, and other compounds (Figure 5 and Table 1). In total, 23 phytochemicals were identified, including six hydroxybenzoic acids (1, 5, 13–15, 16), two hydroxy acids (2, 3), two tyrosol derivatives (6, 7), seven caffeoylquinic acid derivatives (8–11, 18, 19, 22), one lignin (17), one terpenoid (20), two hesperidin derivatives (21, 23), and other compounds (4, 12). These identifications were based on the interpretation of the fragmentation patterns observed in the negative ionization mode. The MS/MS spectra and retention times were used to profile the fragmentation information of the compounds by comparing them with reference data and the existing literature. Both LZSE and TZSE contained various caffeoylquinic acid (CQA) derivatives. Among these, neochlorogenic acid (8), chlorogenic acid (9), and 4-CQA (11) were identified as the major components of both LZSE and TZSE, as observed in the base peak intensity (BPI) chromatogram from UPLC-ESI-Q-TOF-MSe. These compounds were identified at m/z 353.0898 [M − H]− in the MS spectrum, with fragment ions at m/z 191.0551, 179.0332 (caffeic acid) and 135.0448 (decarboxylated caffeic acid) in the MS/MS spectrum. The identities of these compounds were confirmed via comparison with standard compounds. Compound 21 exhibited a [M − H]− ion at m/z 609.1846 in the negative ion mode, with a major ion fragment at m/z 301.0735 [M-H-rutinoside], corresponding to its aglycone (compound 23). Notably, hesperidin (21) and hesperetin (23) were found exclusively in TZSE, highlighting the unique chemical profile of TZSE.

Figure 5.

Base peak intensity BPI chromatogram of the 70% ethanol extracts of ZS. BPI chromatogram of LZSE (A) and TZSE (B).

Table 1.

List of identified compounds of LZSE and TZSE.

	RT	Observed mass	Error	Molecular	Fragment	Tentative component
	(min)	[M – H]–	(ppm)	formula
1	1.15	191.0645		C7H6O5	191.0551, 149.0456, 96.9608, 850.87	Quinic acid*
2	1.31	133.0335	3.8	C4H6O5	115.0027, 96.9612	Malic acid
3	1.43	191.02	4.2	C6H8O7	155.9509, 111.0075, 87.0004	Citric acid
4	2.28	399.1527	1.7	C14H26O10	353.1435[M-H], 311.0774, 221.1011[M-H-apiosyl], 161.0435[glucoside], 101.0238	Isopropyl apiosylglucoside
		[M + HCOOH-H]
5	2.47	315.0728	3.8	C13H16O9	315.0707, 191.0551, 152.0100[M-H-glu], 108.0217	Protocatechuic acid 4-glucoside
6	2.57	315.1092	3.8	C14H20O8	315.1070, 153.0552 [M-H-glu], 123.0458	Hydroxytyrosolglucoside isomer 1
7	2.87	315.1092	3.8	C14H20O8	315.1070, 153.0552[M-H-glu], 123.0458	Hydroxytyrosolglucoside isomer 2
8	3.31	353.0898	7.1	C16H18O9	375.0687, 353.0800, 191.0551, 179.0332, 135.0448	Neochlorogenic acid*
9	4.78	353.0898	7.1	C16H18O9	375.0687, 353.0800, 191.0551	Chlorogenic acid*
10	4.99	353.0855	7.1	C16H18O9	375.0687, 353.0800, 191.0551, 179.0332, 135.0448	Caffeoylqunic acid
11	5.28	353.0898	7.1	C16H18O9	375.0687, 353.0800, 191.0551, 179.0332, 135.0448	4-Caffeoylquinic acid*
12	6.35	401.143	−4.5	C18H26O10	269.1042[M-H-pentose], 161.0435, 149.0435	Benzyl β-primeveroside
13	6.69	337.0942	5.6	C16H18O8	337.0922, 191.0551	4-Coumaroylquinic acid*
14	6.99	311.0774	2.3	C15H16O8	333.1068, 179.0332[M-H-xyloside], 149.0456, 135.0448	Caffeoyl-xyloside
15	7.79	367.1029	0	C17H19O9	389.0825, 367.1029, 191.0546	3-Feruloylquinic acid*
16	8.29	371.0979	0.3	C16H19O10	195.0660[M-H-glucuronide], 113.0229	Dihydroferulic acid 4-O-glucuronide
17	10.43	521.2026	0.6	C26H34O11	359.1476[M-H-glu], 341.1371, 329.1384	Lariciresinol-glucoside
18	12.22	515.1176	−2.7	C25H24O12	353.0880, 191.0551	1,3-Dicaffeoylquinic acid*
19	12.34	515.1169	6.2	C25H24O12	353.0880, 191.0551, 179.0332, 149.0456, 135.0448	1,5-Dicaffeoylquinic acid*
20	12.41	503.2485	−1.4	C24H40O11	371.2056	Leeaoside
21	12.56	609.1846	4.3	C28H34O15	301.0735 [M-H-rutinoside]	Hesperidin
22	12.63	515.1176	−2.7	C25H24O12	353.0880, 191.0551	4,5-Dicaffeoylqunic acid*
23	17.27	301.0735	7.6	C16H14O6		Hesperetin

* By comparing it with the standard compounds it was confirmed that it was the same compounds

Error (ppm): the difference between experimental mass and theoretical mass

3.6. In silico molecular docking

PPARγ is a transcription factor belonging to the PPAR family of nuclear receptors. It is regulated by ligands such as steroids, lipids, vitamins and cholesterol metabolites. PPARγ plays critical roles in adipocyte differentiation, lipid metabolism, inflammatory responses, carcinogenesis and metabolic homeostasis (Ahmadian et al. 2013). Its molecular functions are primarily mediated by conformational changes induced by ligand binding at the PPARγ ligands include endogenous ligands, full agonists, partial agonists, antagonists, inverse agonists and phosphorylation inhibitors. The relatively large LBD of PPARγ enables the binding of various ligands with low affinity. Upon ligand binding, PPARγ undergoes conformational changes, stabilizing in a lower-energy state and functioning as a transcription factor to regulate target gene expression (Ahmadian et al. 2013; Brust et al. 2018; Frkic et al. 2018; Miyamae 2021). Consequently, identifying PPARγ agonists is essential for modulating molecular pathways associated with various diseases. For example, rosiglitazone, a thiazolidinedione and a well-known PPARγ agonist, has been widely recognized for its anti-diabetic properties. However, its adverse effects, including increased risks of cardiovascular events and edema, have raised concerns regarding its clinical efficacy (Waksman 2008; Bortolini et al. 2013; He et al. 2014).

In this study, ZS extracts exhibited significant adipogenic potential and PPARγ activation, suggesting their potential roles as PPARγ agonists. Based on this observation, we hypothesized that specific compounds within the extracts are responsible for PPARγ activation. To test this hypothesis, 16 compounds identified in ZS by UPLC-Q-TOF-MS were selected according to their well-documented structures and the availability of 3D conformer data in PubChem. The selected 3D conformers were docked into the full-agonist LBD of PPARγ using molecular docking methods, with rosiglitazone (CID: 77999), which had a binding score of −7.4 kcal/mol, serving as the positive control. The 16 selected compounds had binding energy scores ranging from −7.4 to −5.03 kcal/mol, forming essential interactions with key amino acids in the LBD of PPARγ (Figure 6 and Fig. S1-S3). The docking results and binding outcomes of the selected compounds are summarized in Table 2, along with their respective CIDs. These findings suggest that LZSE and TZSE contain potential compounds capable of activating PPARγ based on their high affinity for its LBD, in line with known agonists. This provides a molecular basis for the observed adipogenic effects of the ZS extracts.

Figure 6.

Molecular docking of compounds with PPARγ. Binding energy scores (ΔG, kcal/mol) are shown along with visualizations of the interactions between each compound and the PPARγ ligand-binding domain. The following compounds are represented by their respective colors: navy, rosiglitazone; brown, 4-caffeoylquinic acid; blue, neochlorogenic acid; yellow, chlorogenic acid; magenta, 3-feruloylquinic acid; mint, lariciresinol glucoside; cyan, 4,5-dicaffeoylquinic acid; light green, hesperidin; pink, 1,5-dicaffeoylquinic acid; black, 4-coumaroylquinic acid; purple, 1,3-dicaffeoylquinic acid; deep green, hesperetin; light pink, dihydroferulic acid glucuronide; grey, caffeoylquinic acid; deep sky, leeaoside; deep purple, benzyl β-primeveroside; orange, protocatechuic acid glucoside.

Table 2.

Binding energy score of compounds

Compound	Score (kcal/mol)	Extract containing	Ligand color	Pubchem CID
Rosiglitazone	−7.94	–	Navy	77999
4-Caffeoylquinic acid	−7.94	LZSE, TZSE	Brown	58427569
Neochlorogenic acid	−7.74	LZSE, TZSE	Blue	5280633
Chlorogenic Acid	−7.73	LZSE, TZSE	Yellow	1794427
3-Feruloylquinic acid	−7.56	LZSE, TZSE	Magenta	10133609
Lariciresinol glucoside	−7.47	TZSE	Mint	11972395
4,5-dicaffeoylqunic acid	−7.38	LZSE	Cyan	5281780
Hesperidin	−7.07	TZSE	Light green	10621
1,5-dicaffeoylqunic acid	−7.04	LZSE	Pink	122685
4-Coumaroylquinic acid	−6.96	LZSE, TZSE	Black	5281766
1,3-dicaffeoylqunic acid	−6.95	LZSE	Purple	6474640
Hesperetin	−6.79	TZSE	Deep green	72281
Dihydroferulic acid glucuronide	−6.49	TZSE	Light pink	190069
Caffeoylquinic acid	−6.09	LZSE	Grey	10155076
Leeaoside	−5.62	TZSE	Deep sky	101432441
Benzyl β-primeveroside	−5.08	LZSE, TZSE	Deep purple	131248
Protocatechuic acid-glucoside	−5.03	LZSE	Orange	129722880

4. Discussion

The strong association between type 2 diabetes and obesity contributes to a wide spectrum of metabolic disorders, including carcinogenesis, chronic inflammation, reduced lifespan, and hyperlipidemia. Consequently, natural extracts have been extensively investigated for their therapeutic potential in alleviating these pathologies. In this study, we demonstrate that Zanthoxylum schinifolium (ZS) sourced from the Honam region of Korea serves as a potent promoter of adipogenesis.

Unlike previous studies, our work focused on 70% ethanol extracts of ZS specifically harvested from the Honam region, under the hypothesis that geographic origin may significantly influence the biological activity of ZS, particularly with respect to adipocyte differentiation. Notably, LZSE treatment promoted adipogenesis in a dose-dependent manner without inducing cytotoxicity (Figure 1), in contrast to previous reports describing ZS as anti-adipogenic. This discrepancy highlights the importance of regional and phytochemical variation in determining the pharmacological properties of plant-based therapeutics.

To validate the context-specific effects of LZSE, we examined it under the adipogenic induction conditions previously used in anti-adipogenic studies (1× induction medium). Under these conditions, LZSE failed to inhibit adipocyte differentiation, as evidenced by the absence of lipid droplet accumulation in Oil red O staining (Fig. S4A), and unaltered expression of adipogenic genes such as PPARγ and FABP4 (Fig. S4B). These results suggest that the observed discrepancy is unlikely attributable to differentiation protocol alone, and instead may stem from differences in the regional composition of ZS extracts.

The PI3K and MEK/ERK signaling pathways are well-established regulators of adipocyte differentiation, particularly during the early phase, where they posttranscriptionally enhance the expression of key transcription factors such as C/EBPα and PPARγ (Sakaue et al. 1998; Prusty et al. 2002). To further investigate the mechanistic divergence from prior findings, we examined the effect of LZSE on canonical adipogenic signaling pathways. Although phosphorylation of PI3K and ERK was elevated by standard adipogenic induction (1× MDI), LZSE treatment did not elicit any additional change in their activation (Fig. S5). This result contrasts with the findings of Choi et al., who reported methanolic ZS extracts suppressed adipogenesis via modulation of ERK and PI3K/AKT signaling (Choi et al. 2015). These findings suggest that LZSE derived from the Honam region promotes adipogenesis through mechanisms independent of PI3K/ERK signaling. While our data do not preclude a potential role for these pathways under different temporal or dosage conditions, further studies are warranted to clarify their involvement, if any, in LZSE-mediated effects.

To elucidate the molecular mechanisms underlying the pro-adipogenic activity of LZSE, we focused on CHOP, a negative regulator of adipogenesis. CHOP inhibits the expression of C/EBPα and C/EBPβ, both of which are essential for the transcriptional activation of PPARγ, the master regulator of adipocyte differentiation. Our data revealed that LZSE downregulated CHOP expression, allowing for increased C/EBPα levels and subsequent activation of PPARγ (Figure 3A-3D). In addition, we observed that LZSE enhanced mitotic clonal expansion (MCE), a prerequisite for early adipogenesis, by upregulating cell cycle-related genes and promoting cell cycle progression through the S and G2/M phases (Figure 3E and 3F). Collectively, these findings suggest that LZSE promotes adipocyte differentiation by simultaneously suppressing CHOP and accelerating MCE during early differentiation.

Given the central role of PPARγ in regulating adipogenesis and metabolic homeostasis, we further investigated whether ZS extracts act as PPARγ agonist. Luciferase reporter assays demonstrated that both LZSE and TZSE significantly enhanced PPARγ transcriptional activity (Figure 4F), supporting their role as functional agonists. Consistent with this, molecular docking simulations revealed that multiple phytochemicals identified in ZS, including caffeoylquinic acid derivatives, lignins, and flavanones, exhibited strong binding affinity to the ligand-binding domain (LBD) of PPARγ. These compounds may contribute to the overall adipogenic activity of ZS extracts. These results align with earlier studies that have identified natural PPARγ agonists through computational screening approaches.

Funding Statement

This work was supported by the National Research Foundation of Korea [grant numbers: NRF-2019R1A2C1005719 and RS-2022-NR070862], and the Glocal University Project of Mokpo National University in 2025.

Author’s contributions

Jiseok Lee: Conceptualization, Methodology, Software, Validation, Formal Analysis, Investigation, Data Curation, Writing – Original Draft Preparation, Writing – Review & Editing, Visualization. Bo-Ram Kim: Methodology, Software, Validation, Formal Analysis, Investigation, Resources, Data Curation, Writing – Original Draft Preparation, Writing – Review & Editing, Visualization. Kyeoungtae Park: Conceptualization, Methodology, Software, Validation, Formal Analysis, Investigation, Writing – Review & Editing, Visualization. Eunbin Kim: Methodology, Validation, Formal Analysis, Investigation, Writing – Review & Editing, Visualization. Jin-Woo Jeong: Methodology, Software, Validation, Resources, Writing – Review & Editing, Supervision. Jung Jin Kim: Conceptualization, Writing – Review & Editing, Supervision. Sung-Suk Suh: Conceptualization, Writing – Review & Editing, Supervision. Jong Bae Seo: Conceptualization, Writing – Original Draft Preparation, Writing – Review & Editing, Supervision, Project Administration, Funding Acquisition.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Supplemental Material

Supplemental data for this article can be accessed online at https://doi.org/10.1080/19768354.2025.2536022.