Analytical procedures utilized a variety of laboratory instruments and materials. Weighing was conducted using an analytical scale (Ohaus ^®), while absorbance readings were obtained via a UV–Vis spectrophotometer (Beckman Coulter ^®). Chromatographic separation was performed using radial chromatography with silica gel 60 PF254 containing gypsum (Merck, Cat. No. 1.07749.1000). For TLC profiling, aluminum-backed silica gel 60 F254 plates (Merck, Cat. No. 1.05554.0001; 0.25 mm thickness) were used and visualized under 254 nm UV light or after spraying with 10% H ₂SO ₄ (v/v) in methanol.

Advanced compound identification was conducted using a Vanquish™ UHPLC Binary Pump (Thermo Scientific™, Germany) coupled to a Q Exactive™ Hybrid Quadrupole-Orbitrap™ Mass Spectrometer. Separation was achieved using a Thermo Scientific™ Accucore™ Phenyl-Hexyl column (100 mm × 2.1 mm ID × 2.6 μm). Solvents used for extraction, chromatography, and sample preparation included methanol (MeOH; Merck, Cat. No. 106009), ethanol (EtOH; Merck, Cat. No. 100983), ethyl acetate (EtOAc; Merck, Cat. No. 109623), and Milli-Q water (Merck Millipore). Additional laboratory tools comprised Pyrex ^® beakers, droppers, probes (Terumo ^®), microscope (Olympus ^®), ELISA microplate reader (Thermo Scientific™ Multiskan), and standard laboratory glassware.

Reagents used in animal treatments included: Sodium acetate 1 M (Merck, Cat. No. 106268), Sodium carboxymethyl cellulose (NaCMC) 0.5% (Sigma-Aldrich, Cat. No. C4888) as a suspension vehicle, Aqua pro injection (Andeska Lab, Indonesia, Cat. No. AQPRO-INJ-500), Aquadest (Andeska Lab, Cat. No. AQ-DES-1000), COVID-19 vaccine (Indovac ^®, Bio Farma, Indonesia).

Biological materials included male white mice ( Mus musculus, 6–8 weeks old, 25–30 g) from PT Indoanilab (Indonesia), standard mouse pellet diet (Vita Feed ^®), and apigenin isolated from Peronema canescens leaves.

Leaves of Paronema canescens Jack. (sungkai) were collected from Aia Pacah, Padang City, West Sumatera, Indonesia. The plant was identified and authenticated by Dr. Nurainas at the Herbarium ANDA, Department of Biology, Universitas Andalas, with voucher specimen number 717/K-ID/ANDA/X/2024. Samples were air-dried and ground to a fine powder before extraction.

Seventeen kilograms of dried sungkai leaf powder were macerated in 96% ethanol (2:5 w/v) at room temperature for 72 hours. The extraction process was repeated every 24 hours by decanting the extract and re-macerating the residue. The combined filtrates were concentrated using a rotary evaporator to yield approximately 1.2 kg of crude extract.

Isolation and characterization

Five hundred grams of the concentrated extract were fractionated via liquid-liquid extraction using increasing polarity solvents: n-hexane, ethyl acetate, and ethanol. The ethanol-soluble fraction was further subjected to vacuum liquid chromatography using gradients of n-hexane: EtOAc (100:0 to 0:100) and methanol. This yielded 14 initial fractions, which were combined based on TLC spot profiles into five main fractions.

Fraction X (FX), displaying prominent UV-active spots, was purified using radial chromatography with silica gel-coated plates (0.5 mm thickness) and solvent gradients of n-hexane: EtOAc (6:4, 5:5, 3:7). Approximately 10 g of a yellow solid (Isolate 1) was obtained.

Spectroscopic analyses identified Isolate 1 as apigenin ( Table 1), based on the following data: •

UV-Vis (MeOH): λ_max 210, 267, and 336 nm

Ligand and protein preparation

The 3D structure of apigenin (CID: 5280443) and Levamisole (CID: 26879) was retrieved from the PubChem website ( https://pubchem.ncbi.nlm.nih.gov/). Crystal structures of granzyme B (PDB: 1FQ3), perforin (PDB: 3NSJ), and interferon-γ (PDB: 1FG9) were retrieved from the Protein Data Bank ( https://www.rcsb.org/). Protein and ligand were optimized with MOE 2022.02 (licensed proprietary software, license was obtained) to a gradient convergence of 0.001 kcal/mol/Å ².

Molecular docking

The active sites of each target protein were identified using the Site Finder tool in MOE. The Molecular docking simulations were carried out using the default settings of the MOE applications. Ligand–protein interactions were evaluated based on binding affinity (kcal/mol); more negative binding affinities were considered to exhibit favorable binding stability and orientation. Levamisole, a known immunomodulatory agent, was used as a reference compound for comparison. Following docking, ligand–protein interactions were analyzed through 2D and 3D visualizations using the Biovia Discovery 2022 application.

Experimental design

This study was a randomized controlled in vivo study conducted to evaluate the immunostimulatory effect of apigenin, isolated from Paronema canescens (sungkai) leaves, in male mice immunized with the COVID-19 vaccine. This study utilized 25 healthy mice ( Mus musculus), each weighing between 25 and 40 grams. Only healthy male mice aged 6–8 weeks and weighing between 25 and 40 grams were included. Mice were obtained from a certified vendor and allowed a 7-day acclimatization period before treatment. Mice showing signs of illness, injury, or abnormal behavior during the acclimatization period were excluded. No animals were excluded after allocation, and all animals that began treatment completed the protocol and were included in the final analysis.

The animals were randomly assigned to five groups (n = 5 per group) as follows: •

Group 1 (Negative control): Received only the drug carrier, 0.5% sodium carboxymethyl cellulose (NaCMC).

All mice underwent a 7-day acclimatization period to allow adaptation to the laboratory environment and reduce stress-related variability. On Day 1, all groups except the negative control (Group 1) were immunized with the COVID-19 vaccine. Apigenin was administered intramuscularly once daily for 7 consecutive days (Days 1–7) to Groups 3, 4, and 5 according to their respective dosages. On Day 8, a booster dose of the COVID-19 vaccine was administered to all mice in Groups 2 through 5. On Day 9, before blood collection, animals were sedated using diethyl ether as an inhalation anesthetic. The anesthetic was administered by placing a cotton pad soaked in diethyl ether into a sealed treatment chamber. Each mouse was then gently placed into the chamber and monitored until sedation was achieved. Standard equipment used during animal procedures included feeding sonde, sterile disposable syringes, and surgical blades (scalpels). Blood samples were collected following euthanasia by guillotine decapitation targeting the carotid artery. This method was selected to ensure rapid and humane termination under deep sedation, minimizing animal distress and ensuring optimal blood volume recovery for downstream analyses. All procedures were performed by trained personnel under the supervision of a laboratory veterinarian to ensure animal welfare and procedural consistency. After blood collection, the samples were allowed to clot at room temperature and subsequently centrifuged to separate the serum. The isolated serum was aliquoted and stored under appropriate conditions until further analysis.

Granzyme B expression assay

The level of granzyme B was quantified using an ELISA kit (BT Lab ^®, Cat. No. E0846Mo) according to the manufacturer’s instructions. Standards were prepared at concentrations of 1600, 800, 400, 200, 100, 50, and 0 pg/mL. Each well received 50 μL of the granzyme B standard or sample, followed by 50 μL of detection antibody. The plate was incubated at room temperature (25°C) for 1 hour, then washed three times with 350 μL of wash buffer (BT Lab ^®, provided in the kit). After washing, 100 μL of TMB substrate (BT Lab ^®, supplied with the kit) was added per well and incubated in the dark for 10 minutes. The reaction was stopped with 100 μL of stop solution (BT Lab ^®), and absorbance was read at 450 nm using a microplate reader. Granzyme B concentrations were calculated based on the standard curve generated from known concentrations.

Interferon gamma level examination

Serum IFN-γ levels were measured using a commercial ELISA kit (BT Lab ^®, Cat. No. E0002Mo) according to the manufacturer’s instructions. Each well was loaded with 50 μL of serum or standard, followed by antibody incubation and washing steps as recommended. Subsequently, 100 μL of TMB substrate was added to initiate the enzymatic reaction, followed by 10 minutes of incubation. The reaction was stopped with 100 μL of stop solution, and absorbance was measured at 450 nm. The concentrations were derived from a standard curve generated using IFN-γ standards provided in the kit.

Statistical analysis

All data were presented as mean ± standard deviation (SD). Differences between groups were analyzed using one-way analysis of variance (ANOVA). Post hoc comparisons were performed using Duncan’s multiple range test to determine statistically significant differences among groups. Statistical significance was defined as p < 0.05. All analyses were carried out using SPSS Statistics version 15.0 (SPSS Inc., Chicago, IL, USA).

All animal procedures were conducted following the ethical standards for animal experimentation and complied with institutional and national guidelines. Ethical approval for this study was obtained from the Health Research Ethics Committee of the Faculty of Pharmacy, Andalas University, under approval number: 93/UN16.10.D.KEPK-FF/2024. All efforts were made to minimize animal suffering, and only the minimum number of animals required to achieve statistical significance was used. Sedation and euthanasia procedures were performed to ensure humane treatment throughout the experimental protocol.

Isolate 1 was successfully obtained from fraction FX of the sungkai leaf extract, yielding 10 grams of yellow powder solid that was soluble in ethanol. The TLC profile showed a single spot under UV light at 254 and 365 nm, indicating the presence of a single compound.

The UV spectrum of Isolate 1 exhibited three maximum absorption peaks at λmax 210, 267, and 336 nm. These peaks suggested the presence of a conjugated chromophore system ( Figure 1).

FTIR spectrum showed characteristic peaks at Vmax 3286 cm ⁻¹ (O–H stretch), 1653 cm ⁻¹ (C=O stretch), 1587–1446 cm ⁻¹ (aromatic C=C), and 1298–1031 cm ⁻¹ (C–O stretch), indicating the presence of hydroxyl, carbonyl, and aromatic functional groups ( Figure 2).

The ¹H-NMR spectrum revealed five proton signals in the δH 6–8 ppm region. Signals at δH 7.92 and 6.92 ppm appeared as doublets with J = 8.75 Hz, suggesting ortho-coupled aromatic protons. Signals at δH 6.47 and 6.18 ppm showed meta coupling (J ≈ 2 Hz), while a singlet at δH 6.78 ppm was attributed to an isolated proton between two quaternary carbons ( Figure 3).

The ¹³C-NMR spectrum exhibited 13 carbon signals, all in the sp ² region. Six signals between δC 157.3–181.8 ppm corresponded to =C–OH and –C=O carbons, while the remaining seven signals (δC 94.0–128.5 ppm) represented sp ² methine and quaternary carbons ( Figure 4).

COSY analysis confirmed proton-proton correlations, notably between δH 7.92 and 6.92 ppm, supporting ortho substitution patterns. No direct coupling was observed among the remaining three protons ( Figure 5).

Long-range correlations observed in the HMBC spectrum confirmed the positioning of key functional groups. δH 6.78 ppm correlated with C-2, C-4, C-10, and C-1’, indicating carbonyl at C-4 and phenolic group at C-2. δH 6.18 and 6.47 ppm showed meta correlations to C-5, C-6, C-7, C-8, C-9, and C-10, confirming OH substitution at C-5 and C-7 ( Figure 6).

Using UPLC-MS/MS in negative ion mode, the molecular ion peak [M–H] ^- was detected at m/z 268.992, corresponding to the molecular formula C ₁₅H ₁₀O ₅. This confirmed the identity of the compound as apigenin ( Figure 7).

The IHD value was calculated to be 11, which supports a structure containing three rings and eight additional degrees of unsaturation, consistent with a flavonoid backbone ( Figure 8).

The docking results demonstrated that apigenin exhibited binding affinities of −5.56 kcal/mol with both granzyme B and perforin, and −4.95 kcal/mol with IFN-γ. These values were comparable to those of levamisole, which showed binding affinities of −5.08 kcal/mol (granzyme B), −5.36 kcal/mol (perforin), and −4.82 kcal/mol (IFN-γ) ( Table 2). In terms of molecular interactions, apigenin formed hydrogen bonding with His151 in granzyme B, and additional interactions including Pi-alkyl (Leu39, Lys40), sulfur-X (Lys149), and van der Waals (Phe57). For perforin, interactions included Pi-alkyl (Ala8, Leu11, Ile73), Pi-Pi stacking (Tyr53), and Pi-sulfur (Met77). For IFN-γ, apigenin exhibited multiple hydrogen bonds (Arg143, Arg372, Gln374, Asp544), Pi-alkyl interaction (Pro546), and ionic interactions including Pi-cation/anion (Arg378, Gly551). Levamisole, in comparison, exhibited hydrogen bonding with Ser195 (granzyme B), Arg372 (IFN-γ), and various hydrophobic and Pi-based interactions with all three proteins ( Table 3).

The administration of apigenin isolated from Peronema canescens (sungkai) leaves was evaluated for its immunostimulant effects in male white mice induced with the COVID-19 vaccine. The study involved five experimental groups: a negative control group receiving 0.5% NaCMC, a positive control group receiving the COVID-19 vaccine, and three treatment groups receiving apigenin at doses of 1 mg/kgBW, 25 mg/kgBW, and 50 mg/kgBW, respectively. The ELISA results indicated a dose-dependent increase in granzyme B levels across the groups. The negative and positive control groups showed relatively lower granzyme B levels, with mean values of 150.83 and 151.98, respectively. In contrast, the treatment groups demonstrated higher values, reaching 160.75 at 1 mg/kgBW, 161.38 at 25 mg/kgBW, and 162.09 at 50 mg/kgBW ( Figure 9a). Statistical analysis using one-way ANOVA confirmed significant differences among the groups (p < 0.05), and Duncan’s post hoc test revealed that all treatment groups formed a separate subset from the controls, indicating a significant stimulatory effect of apigenin on granzyme B production.

Similarly, perforin levels showed a notable increase with escalating doses of apigenin. The negative and positive control groups had lower mean values of 1,602 and 1,789, respectively, whereas the treatment groups exhibited higher values of 1,846 at 1 mg/kgBW, 2,917 at 25 mg/kgBW, and 3,086 at 50 mg/kgBW ( Figure 9b). Statistical validation through one-way ANOVA revealed a significant increase in perforin levels (p < 0.05), and Duncan’s test indicated a consistent trend with the granzyme B data, supporting the conclusion that apigenin administration enhances perforin expression.

For interferon-γ, the ELISA data also demonstrated a dose-dependent increase. The negative and positive control groups exhibited mean levels of 107.18 and 109.10, respectively, while the treatment groups showed elevated levels of 114.20, 130.32, and 130.57 for doses of 1, 25, and 50 mg/kgBW, respectively ( Figure 9c). Although body weight was not significantly different among the groups, interferon-γ levels increased significantly with higher doses of apigenin. Normality testing with the Shapiro-Wilk test (p > 0.05) and homogeneity testing (p > 0.05) confirmed the appropriateness of one-way ANOVA, followed by Duncan’s test, which indicated the highest stimulation occurred at the 50 mg/kgBW dose.