Prior to the commencement of this study, ethical approval was obtained from the Research Ethics Committee of Kampala International University, Western Campus (Approval No. KIU-2024-532). In addition, authorization to conduct the study was granted by the Uganda National Council for Science and Technology (UNCST), under approval number HS5372ES.

Bisphenol A (Cat. No: 239658) was used to induce endocrine disruption and systemic toxicity in experimental animals at a dose of 100 mg/kg body weight administered orally once daily. Ethyl acetate (Cat. No: 270989) was employed as a solvent during the extraction and fractionation of plant materials, with approximately 100 mL used per extraction batch. Methanol (Cat. No: 34860) was used both as a solvent in the DPPH radical scavenging assay and for plant extraction procedures, with an estimated volume of 100 mL per assay and extraction cycle. DPPH (Cat. No: D9132) powder (0.004% w/v in methanol) was used to evaluate the antioxidant activity of the plant extracts through a radical scavenging assay. Ketamine (Cat. No: PHR1663) and Xylazine (Cat. No: X1251) were used in combination as anesthetics prior to animal sacrifice, administered intraperitoneally at doses of 80 mg/kg and 10 mg/kg body weight, respectively. Eosin-nigrosin (Cat. No: 134-25-8/N3268) stain was used for assessing sperm viability and membrane integrity, with one drop of each stain applied to a semen smear for microscopy. They were purchased from Sigma-Aldrich (St. Louis, MO, USA). All other reagents used in the study were of analytical grade and obtained from Reagent World Ltd. (Kampala, Uganda), an authorized distributor of Sigma-Aldrich products.

Fresh leaves of Bidens pilosa were collected in September 2024 from the vicinity of Ishaka Mosque, located in the Ishaka-Bushenyi District of Uganda. Botanical authentication was conducted by Dr Olet Eunice of the Department of Botany at Mbarara University of Science and Technology, where a voucher specimen (No. AAI-2024-001) was deposited in the departmental herbarium. The collected leaves were thoroughly air-dried in a shaded area to prevent photodegradation for five (5) days and then ground into a fine powder using a blending machine (Model XYZ123, Philips, Amsterdam, Netherlands).

A portion of the powdered material (100 g) was first filtered to remove coarse plant residues and then suspended in 400 mL of distilled water for 48 hours with intermittent shaking. The aqueous suspension was subsequently partitioned successively with solvents of increasing polarity using a separatory funnel. The solvents used for fractionation included n-hexane, dichloromethane, ethyl-acetate, and n-butanol. Each solvent was added, mixed thoroughly with the aqueous layer, and the resulting organic phase was collected following phase separation. This process was repeated three times per solvent to ensure exhaustive extraction.

Each solvent fraction— n-hexane, dichloromethane, ethyl-acetate, and n-butanol—was individually concentrated under reduced pressure using a rotary evaporator to obtain the respective crude fractions. The concentrated extracts were stored at 4 °C until further use for phytochemical screening and biological assays.

Phytochemical constituents in the crude extracts were identified using Gas Chromatography-Mass Spectrometry (GC-MS) analysis, performed on a Shimadzu QP2010 GC-MS system equipped with an RTX-5MS capillary column (30 m × 0.25 mm ID × 0.25 μm film thickness). Helium served as the carrier gas at a constant flow rate of 1.0 mL/min. The injector temperature was maintained at 250 °C. The oven temperature program began at 60 °C (held for 2 minutes), ramped at 10°C/min to 280 °C, and was then held isothermally for 10 minutes.

The mass spectrometer operated in electron ionization (EI) mode at 70 eV with a scan range of 40–600 m/z. Compound identification was performed by comparing the acquired mass spectra against entries in the National Institute of Standards and Technology (NIST) library database, along with comparisons of retention indices and fragmentation patterns to those of authentic reference standards where available. Chromatographic peaks corresponding to individual compounds were quantified based on peak area and height. Identified compounds were characterized by their chemical names, molecular weights, and molecular formulas, as determined through spectral matching and library searches.

The bioactive compounds identified through GC-MS analysis were further evaluated in silico to predict their physicochemical properties, drug-likeness, lipophilicity, and solubility profiles. These predictions were based on established medicinal chemistry filters, including Lipinski’s Rule of Five, the Ghose filter, Veber rule, and Egan rule, as implemented in the SwissADME web tool. ¹⁰ Additionally, pharmacokinetic parameters such as gastrointestinal (GI) absorption, P-glycoprotein (P-gp) substrate potential, cytochrome P450 (CYP) enzyme inhibition (focusing on CYP1A2 and CYP2D6), and skin permeability (log Kp) were also predicted using the same platform.

To gain insights into the possible biological activities of the compounds, molecular docking simulations were carried out using PyRx software, employing the AutoDock Vina algorithm. ¹¹ Docking studies were focused on two key protein targets—Nuclear factor erythroid 2-related factor 2 (Nrf2) and the Androgen Receptor (AR)—due to their central roles in oxidative stress regulation and male reproductive function, respectively.

High-resolution (≤2.5 Å) three-dimensional structures of Nrf2 and AR were retrieved from the Protein Data Bank (PDB). Preference was given to structures co-crystallized with ligands to aid in accurate definition of the active site. Protein preparation was performed using AutoDock Tools (ADT) v1.5.7 and UCSF Chimera v1.15. This involved the removal of water molecules, co-crystallized ligands, and heteroatoms, addition of polar hydrogens and Kollman charges, atom type assignment, and conversion to the PDBQT format required for docking. Energy minimization of the protein structures was carried out in Chimera using the AMBER ff14SB force field to relieve structural strain and enhance docking accuracy.

Compounds identified in Bidens pilosa through GC-MS analysis—namely Methyl (Z)-5,11,14,17-eicosatetraenoate, Pentadecanoic acid, Phytol, 9,12,15-Octadecatrienoic acid, Squalene, and Hexatriacontane—were selected for further molecular docking based on a relative abundance threshold of ≥2% peak area. Ligand structures were retrieved from the PubChem database in SDF format and subsequently converted to 3D conformations using Open Babel, which also handled format conversion and protonation state adjustment. Geometry optimization was performed using Avogadro software with the MMFF94 force field.

Ligands were parameterized using the General AMBER Force Field (GAFF) through ACPYPE or Antechamber to generate MOL2 and PDBQT files compatible with docking software. Among the ligands, 9,12,15-Octadecatrienoic acid, which exhibited the highest binding affinity in preliminary docking, was selected for further interaction analysis.

Molecular docking was conducted using AutoDock Vina v1.2.3, chosen for its accuracy and computational efficiency. The docking grid was centered on the active site of the target proteins, defined by known ligand-binding residues, and sized to accommodate ligand flexibility. Docking results were ranked according to binding energy scores (kcal/mol), with the top-ranked poses retained for post-docking analysis.

Visualization and interaction analysis were performed using multiple tools: Discovery Studio Visualizer (BIOVIA) was employed to examine hydrogen bonding, hydrophobic interactions, and molecular surface characteristics; PyMOL was used for high-resolution 3D visualization and figure preparation; and LigPlot+ was utilized to generate 2D interaction diagrams highlighting key amino acid contacts involved in ligand binding.

In this study, DPPH and ascorbic acid (used as a standard antioxidant) were procured from a certified supplier and used without further modification. Methanol of analytical grade served as the solvent for all preparations. Serial dilutions of the ethyl acetate extract of Bidens pilosa were prepared at concentrations of 20, 30, 40, and 50 μg/μL, while ascorbic acid solutions were prepared at concentrations of 0.1, 0.2, 0.3, and 0.4 μg/μL. A 0.1 mM DPPH solution was freshly prepared in methanol. For each reaction, 1 mL of the test sample or standard solution was mixed with 1 mL of the DPPH solution. The control consisted of 1 mL methanol mixed with 1 mL DPPH solution.

All mixtures were incubated in the dark at room temperature for 30 minutes to prevent photodegradation. Following incubation, absorbance readings were taken at 517 nm using a UV-Visible spectrophotometer (UV-1800, Shimadzu Corporation, Kyoto, Japan). The percentage of radical scavenging activity was calculated relative to the control.

The percentage of DPPH radical scavenging activity was calculated using the formula: % Inhibition = A control − A sample A control × 100

Where A control is the absorbance of the control and A sample is the absorbance of the test sample.

A standard curve was plotted for ascorbic acid to determine the IC ₅₀ (the concentration required to inhibit 50% of DPPH radicals). Similarly, the IC ₅₀ for the plant extract was calculated from its linear regression equation.

IC ₅₀ stands for “Half Maximal Inhibitory Concentration.”

IC ₅₀ is the amount of a substance needed to reduce an activity (like radical activity) by half. It’s commonly used to measure the potency of an antioxidant, drug, or inhibitor—the lower the IC ₅₀, the more potent the substance.

The up-and-down technique for acute toxicity was used to determine the LD50 of the methanol extract of Bidens pilosa as outlined by. ¹² This technique provides accurate information on the extract’s toxicological profile while minimizing the usage of animals. At the end of this procedure 5 mice were used for the acute toxicity study. The oral LD50 of ethyl-acetate extract of Bidens pilosa was found to be greater than 5000 mg/kg body weight suggesting that the plant is safe.

Thirty inbred adult male albino mice ( Mus musculus), each weighing between 35–45 g, were used for this study. The animals were sourced from the Animal House of Mbarara University of Science and Technology, Uganda. Upon procurement, the mice were transported to the Animal House of Kampala International University, Western Campus, where they were housed in clean, well-ventilated plastic cages (five mice per cage) and acclimatized for a period of two weeks prior to the commencement of experimental procedures. During acclimatization and throughout the study, the animals were maintained under standard laboratory conditions: a controlled room temperature of 22 ± 2°C, relative humidity of 70 ± 4%, and an inverted 12-hour light/dark cycle. The mice were given unrestricted access to standard pelleted rodent feed and clean drinking water ad libitum. Bedding material (wood shavings) was changed every two days to maintain hygiene.

All procedures involving animals were conducted in accordance with the guidelines of the Institutional Animal Care and Use Committee (IACUC, 2010) and adhered to the UK Animals (Scientific Procedures) Act 1986, as amended in 2012. The ethical approval for this study was obtained from the Kampala International University Institutional Research and Ethics Committee (KIU-REC).

The sample size was determined using the power resource equation; Minimum n = 10 / k + 1 , Maximum n = 20 / k + 1 where n is the number of animals per group, k is the number of groups. ¹³

Therefore, the minimum number of animals needed per group for the studies = (10/6) +1 = 2.7.

While the maximum number of animals needed per group for the studies = (20/6) +1 = 4.3.

Hence, we 5 animals were assigned to the 6 groups.

After 2 weeks of acclimatization, the mice were randomly assigned into six groups with five mice in each group to reduce bias in accordance with ARRIVE guidelines. ¹⁴

Group 1 (control) received 2 ml/kg body weight of distilled water; Group 2 (100 mg/kg/day of BPA) were administered only bisphenol A; Group 3 (100 mg/kg/day of BPA + 250 mg/kg BW extract of B. pilosa) received bisphenol A and co-treated with 250 mg/kg body weight of the leaf extract; Group 4 (100 mg/kg/day of BPA + 500 mg/kg BW extract of B. pilosa) received BPA and co-treated with 500 mg/kg body weight of leaf extract; Group 5 (100 mg/kg/day of BPA + 1000 mg/kg BW extract of B. pilosa) received BPA and co-treated with 1,000 mg/kg of leaf extract, while Group 6 (100 mg/kg/day of BPA + Vit. C 60 mg/kg BW) was administered BPA and vitamin C. All treatments were administered once daily via oral gavage for a period of 35 consecutive days. BPA was freshly prepared and administered each day at a consistent time in the morning to minimize circadian variation. The dosage for BPA (100 mg/kg/day) was adopted from a previous study by. ¹⁵ The dosage for B. pilosa (250 mg/kg, 500 mg/kg, and 1000 mg/kg) was adopted from a previous study by, ¹⁶ while that of Vitamin C was also adopted from previous study by. ¹⁷ The body weights of all animals in each group were recorded weekly, beginning from the first week of the experiment till the end of the administration period at the fifth week. At the end of the treatment period, all animals were fasted overnight and weighed prior to euthanasia procedures. Anesthesia was induced via intraperitoneal injection using a combination of ketamine hydrochloride at a dose of 80 mg/kg and xylazine hydrochloride at 8 mg/kg body weight. Adequate anesthesia was confirmed by the absence of a pedal reflex and other responses to external stimuli. Following confirmation of a deep plane of anesthesia, euthanasia was performed via exsanguination by cardiac puncture, consistent with the guidelines provided in the American Veterinary Medical Association (AVMA) Guidelines for the Euthanasia of Animals (2020 Edition). These procedures were carried out to minimize pain and distress and complied fully with institutional and international standards for the humane treatment of laboratory animals. Following confirmation of death by cessation of heartbeat and respiratory movement, the animals were dissected, and relevant tissues were harvested for histological and biochemical analyses.

The testes were homogenized in Tris buffer using a mortar and pestle to prepare a 10% (w/v) tissue homogenate. The homogenate was then centrifuged at 3,000 × g for 10 minutes, and the resulting supernatant was collected for subsequent biochemical analyses. The supernatant was used to measure malondialdehyde (MDA) levels, superoxide dismutase (SOD) activity, and catalase activity.

MDA levels, an indicator of lipid peroxidation, were assessed using the thiobarbituric acid reactive substances (TBARS) assay, in which MDA reacts with thiobarbituric acid to form a pink chromogen detectable spectrophotometrically at 532 nm.

SOD activity was determined by its ability to inhibit the autoxidation of epinephrine at pH 10.2, as described by, ¹⁸ and was expressed as units per mg of protein. Catalase activity was measured according to the method of, ¹⁹ by monitoring the decomposition of hydrogen peroxide at 240 nm, with activity expressed as units per mg of protein.

The data from the in vivo study were analyzed using GraphPad Prism ^® version 5.01 (San Diego, CA, USA). One-way analysis of variance (ANOVA) was used to assess the difference mean among the different groups; this was followed by Tukey’s post-hoc test, where necessary. Differences among the groups were considered significant where the p-value was equal or less than 0.05. The data for the study can be accessed as an extended file on figshare. ¹⁴

Gas Chromatography-Mass Spectrometry (GC-MS) analysis of the ethyl-acetate extract of Bidens pilosa revealed several major constituents as shown in Figure 1.

A higher area percentage in GC-MS analysis reflects the relative abundance and potential biological relevance of the detected compounds. Based on a minimum threshold of 2% area, six major compounds were identified for further characterization and analysis. These included Methyl (Z)-5,11,14,17-eicosatetraenoate, Pentadecanoic acid, Phytol, 9,12,15-Octadecatrienoic acid, Squalene, and Hexatriacontane ( Table 1). Compound identification was achieved by comparing the obtained mass spectra with reference spectra from established databases, considering key parameters such as retention time, molecular weight, molecular formula, and area percentage. The identities and bioactivity profiles of the selected compounds ( Table 2) were further validated using online chemical databases including PubChem, SwissADME, and ADMETlab 3.0, which collectively provide structural, pharmacokinetic, and physicochemical data for over 60,000 compounds.

Of the six identified compounds, only two were predicted to have high gastrointestinal (GI) absorption, while the remaining four exhibited low predicted absorption. Among them, Phytol was uniquely identified as a substrate of P-glycoprotein (P-gp). None of the compounds were predicted to inhibit cytochrome P450 2D6 (CYP2D6); however, Pentadecanoic acid and 9,12,15-Octadecatrienoic acid were predicted to inhibit CYP1A2. The predicted skin permeability (log Kp) values for the compounds ranged from 0 to 4.18 cm/s, as summarized in Table 3.

P-glycoprotein (P-gp) is a key transmembrane efflux transporter involved in the active removal of a wide range of xenobiotics from cells, playing a critical role in drug absorption, distribution, and excretion. In parallel, the cytochrome P450 enzyme family, particularly CYP2D6 and CYP1A2, is integral to drug metabolism. CYP2D6 metabolizes a diverse array of drugs, including antidepressants, antipsychotics, antiarrhythmics, and opioids. ²⁰ Inhibition of CYP2D6 can impair drug metabolism, leading to elevated plasma concentrations and an increased risk of toxicity. ²¹ Similarly, CYP1A2 is responsible for metabolizing various drugs and endogenous substances. Its inhibition may reduce clearance rates, thereby altering drug pharmacokinetics and therapeutic efficacy. ²² Additionally, transdermal drug delivery is influenced by skin permeability, which is determined by factors such as molecular weight, lipophilicity, and skin condition. Log Kp values are commonly used to estimate a compound’s ability to penetrate the skin barrier.

All six compounds were subjected to molecular docking against Vitamin C and testosterone to determine which exhibited the strongest binding affinity. Using the AutoDock Vina algorithm within the PyRx platform, 9,12,15-Octadecatrienoic acid emerged as the compound with the highest binding affinity. Consequently, ligands of 9,12,15-Octadecatrienoic acid, Vitamin C, and testosterone were selected for further protein–ligand interaction analysis. Binding interactions with the target proteins—Nuclear factor erythroid 2–related factor 2 (NRF2) and the Androgen Receptor (AR)—were evaluated using Discovery Studio Visualizer. The analysis highlighted key interactions such as hydrogen bonding, van der Waals forces, and hydrophobic contacts that contribute to the stability and specificity of ligand–protein binding ( Figures 2- 4).

Analysis of the three-dimensional protein–ligand complexes revealed that both 9,12,15-octadecatrienoic acid and vitamin C bind to the same active site on NRF2. Two-dimensional interaction diagrams confirmed the presence of conventional hydrogen bonds between the ligands and NRF2. Specifically, 9,12,15-octadecatrienoic acid formed hydrogen bonds with Phe55 and Leu60, while vitamin C interacted with Phe55, Leu58, and Lys52 ( Figures 5- 6). Given that vitamin C is a well-established antioxidant, the similar binding pattern observed with 9,12,15-octadecatrienoic acid suggests that it may also exhibit antioxidant activity.

Analysis of the three-dimensional protein–ligand complexes showed that both 9,12,15-octadecatrienoic acid and testosterone bind to the same active site on the androgen receptor (AR) ( Figures 7- 8). Two-dimensional interaction diagrams revealed the presence of conventional hydrogen bonds and alkyl interactions between the ligands and AR ( Figures 9- 10). Since testosterone and related androgens exert their physiological effects by binding to AR and regulating genes critical for spermatogenesis, sperm maturation, and overall testicular function, the observed binding pattern suggests that 9,12,15-octadecatrienoic acid may mimic androgenic activity and influence reproductive function.

This analysis compares the antioxidant activity of the ethyl acetate fraction of B. pilosa to that of ascorbic acid (vitamin C) using the DPPH free radical scavenging assay, where a lower IC ₅₀ value indicates stronger antioxidant activity. The DPPH radical is decolorized upon reduction, and this change can be quantitatively measured by the decrease in absorbance at 515–517 nm. The IC ₅₀ values of vitamin C and the ethyl acetate extract of B. pilosa are presented in Figure 11. Vitamin C exhibited the highest scavenging activity with an IC ₅₀ of 1.25 μg/ml, indicating strong antioxidant potential. In comparison, the B. pilosa extract showed moderate antioxidant activity, with an IC ₅₀ of 6.11 μg/ml. Although less potent than vitamin C, the extract still demonstrated notable free radical scavenging ability ( Figure 11).

Following five weeks of treatment, body weight was reduced in the control group (2 mL/kg distilled water), as well as in the groups treated with 250 mg/kg (Group 3), 500 mg/kg (Group 4), and 1000 mg/kg (Group 5) of the ethyl-acetate extract of Bidens pilosa, and 60 mg/kg of vitamin C (Group 6), when compared to the BPA-only group (100 mg/kg/day, Group 2). Notably, the group treated with 60 mg/kg of vitamin C (Group 6) exhibited a statistically significant reduction in body weight (p ≤ 0.05) compared to the BPA-treated group, suggesting a possible modulation of BPA-induced metabolic effects ( Figure 12).

Superoxide dismutase (SOD) activity in the testes and epididymis of mice is shown in Figure 13. A significant reduction (p < 0.05) in SOD activity was observed in the BPA-induced group compared to the control, indicating oxidative stress. However, treatment with Bidens pilosa extract at doses of 250 mg/kg and 500 mg/kg, as well as with 60 mg/kg of vitamin C, significantly restored SOD activity relative to the BPA group (p < 0.05). While extract- and vitamin C-treated groups exhibited slightly higher SOD activity than the control, these increases were not statistically significant.

Figure 14 presents the catalase (CAT) activity in the testes. Although there was no statistically significant difference in CAT activity between the BPA-induced and control groups, a modest reduction was noted in the BPA group. Treatment with 500 mg/kg of B. pilosa extract resulted in a significant increase in CAT activity compared to the BPA group (p < 0.05), though this improvement did not reach significance relative to the control.

Malondialdehyde (MDA) levels, a marker of lipid peroxidation, are displayed in Figure 15. MDA levels were significantly elevated (p < 0.05) in the BPA group relative to the control, confirming oxidative damage. Administration of B. pilosa extract at 500 mg/kg and 1000 mg/kg, as well as vitamin C, led to a significant reduction in MDA levels compared to the BPA-induced group (p < 0.05), indicating protective antioxidant effects.