Brasil
Journal of the Brazilian Chemical Society
Sumário Anterior Atual Seguinte
Sumário
Full Paper • J. Braz. Chem. Soc. 36 (11) • 2025 • https://doi.org/10.21577/0103-5053.20250143 linkcopy

Open-access In silico Analysis of Soy Isoflavones and Their Metabolites in Prostate Cancer

AuthorshipSCIMAGO INSTITUTIONS RANKINGS

Abstract

This study investigated the effects of soy isoflavonoids and their metabolites on androgen receptor (AR) signaling in prostate cancer using in silico methods. Molecular docking analyses revealed that several compounds, notably 8-hydroxydaidzein (8HD) and dihydrogenistein (DHG), bind to AR with binding energies comparable to those of testosterone. This suggests that these isoflavonoids inhibit prostate cancer cell growth by competitively blocking androgen binding to the AR. In silico analysis of absorption, distribution, metabolism, excretion, and toxicity predicted favorable drug-like properties and bioavailability, while also identifying potential nephrotoxicity as a safety issue. Our findings highlight the potential of specific soy isoflavonoids, particularly 8HD and DHG, to modulate AR signaling pathways owing to their strong predicted binding to AR. These in silico results provide a mechanistic basis for further in vitro and in vivo investigations of soy isoflavonoids as potential chemopreventive or therapeutic agents against prostate cancer.

Keywords:
soy isoflavones; molecular docking; prostate cancer; anticancer; metabolites

Introduction

Prostate cancer is a major global health concern and ranks fourth among the most diagnosed cancers worldwide. The World Health Organization reported over 1.4 million new cases globally in 2022.1 Incidence rates vary considerably across populations.2 Asian populations exhibit lower rates than those in Western countries, a difference potentially linked to dietary habits such as the high consumption of soy-based foods in Asian diets.2,3 Key risk factors include age, ethnicity, family history, and environmental exposures, with age being a strong determinant.4-6 African-American men are twice as likely to die from prostate cancer and have a 60% higher risk of developing the disease than males of Caucasian descent.7

Developing successful prevention and treatment plans for prostate cancer requires an understanding of the factors contributing to disease. Hormonal influences, particularly those of androgens, play a crucial role in this process. Prostate cancer progression is closely related to androgen hormones, primarily testosterone (TES) and its more potent derivative, dihydrotestosterone (DHT).8 These hormones exert their effects by binding to the androgen receptor (AR). Consequently, AR signaling is a critical pathway that drives the progression of prostate cancer. Androgen deprivation therapy (ADT), which aims to suppress TES production and block AR signaling, is the standard treatment approach. Most cases eventually progress to castration-resistant prostate cancer, which is characterized by disease advancement despite significantly reduced TES levels.8,9 This progression underscores the complex and multifaceted role of AR signaling in prostate cancer and highlights the need for alternative therapeutic strategies targeting this pathway.

The central role of AR signaling and the challenges associated with current therapies, such as ADT, have driven current research to explore the treatment of prostate cancer with natural compounds, particularly soy isoflavonoids.10,11 These plant-derived compounds, also known as phytoestrogens, exhibit estrogen-like activity and are considered chemopreventive agents that may help prevent cancer development.12,13 According to epidemiological research, prostate cancer risk is inversely correlated with soy consumption, especially in Asian populations.14,15 Soy isoflavones, particularly genistein and daidzein, have shown notable anticancer effects in prostate cancer cells. Studies have shown that these compounds inhibit cell growth, proliferation, and metastasis, both in vitro and in vivo. Apoptosis induction, cell cycle modulation, effects on androgen- and estrogen-mediated pathways, and the suppression of angiogenesis and metastasis are some of the mechanisms of action of genistein and daidzein.16,17 These isoflavones are metabolized by the gut microbiota into various compounds, including equol, which is derived from daidzein.18,19 Equol has shown promise in lowering the risk of hormone-related malignancies.20

This study focused on a range of soy isoflavones (Figure 1) and their key metabolites (Figure 2). These include daidzein, genistein, glycitein, formononetin, equol, dihydrodaidzein (DHD), O-desmethylangolensin (ODMA), 8-hydroxydaidzein (8HD), 6-hydroxydaidzein (6HD), 3’-hydroxydaidzein (3HD), cis-4-hydroxyequol (C4HE), dihydrogenistein (DHG), 6’-hydroxy-O-desmethylangolensin (6H-ODMA), and tetrahydrodaidzein (TDZ). These metabolites are produced by both hepatic and intestinal microbial enzymes and represent forms that interact within the body.21-23

thumbnail of Figure 1
Figure 1
Molecular structures of daidzein, genistein, glycitein and formononetin.

thumbnail of Figure 2
Figure 2
Molecular structures of metabolites of soy isoflavones.

Given the established importance of AR signaling in prostate cancer pathogenesis and empirical evidence supporting the ability of soy isoflavonoids (including their metabolites) to influence cancer processes and modulate hormone pathways, direct investigation of their interaction with AR is crucial. Understanding whether these compounds can bind to AR interact with endogenous androgens, such as TES and DHT, could provide a specific mechanistic explanation for their observed protective effects, as highlighted in epidemiological and preclinical studies. Therefore, the primary objective of this study was to use in silico molecular docking, a computational method used to estimate the interactions between proteins and small molecules, to assess the ability of soy isoflavonoids to modulate AR signaling. The predicted binding energies of the soy isoflavonoids in AR were quantified. We identified the key structural features of these isoflavonoids that contribute to their interactions with the target proteins. Through this analysis, we evaluated the potential of soy isoflavonoids to modulate AR signaling and inhibit prostate cancer cell growth and progression. This in silico approach provides valuable insights into the use of these agents to prevent prostate cancer and offers a foundation for future experimental validation, guiding further research.

Methodology

Materials

The two-dimensional (2D) and three-dimensional (3D) structures of the soy isoflavones were obtained from the PubChem database.24,25 These included daidzein (PubChem CID: 5281708), genistein (PubChem CID: 5280961), glycitein (PubChem CID: 5317750), formononetin (PubChem CID: 5280378), equol (PubChem CID: 91469), DHD (PubChem CID: 176907), ODMA (PubChem CID: 89472), 8HD (PubChem CID: 5466139), 6HD (PubChem CID: 5284649), 3HD (PubChem CID: 5284648), C4HE (PubChem CID: 40424401), DHG (PubChem CID: 9838356), 6H-ODMA (PubChem CID: 20601635), and TDZ (PubChem CID: 129651510) (Figures 1 and 2). Two reference steroid compounds, TES (PubChem CID: 6013) and DHT (PubChem CID: 10635), and one non-steroidal antagonist, R-bicalutamide (PubChem CID: 56069), were sourced from PubChem (Figure 3). All of these were downloaded from the structure data file (.sdf) format. The UCSF Chimera v1.18 program was used for visual examination and manipulation of 3D ligand structures.26,27

thumbnail of Figure 3
Figure 3
Molecular structures of testosterone, dihydrotestosterone and R-bicalutamide.

The crystal structure of human AR (Protein Data Bank (PDB): 2AM9, resolution: 1.64 Å) was retrieved from the RCSB Protein Data Bank.28 The AR structure was examined using the UCSF Chimera program. The structure, bound to its native ligand TES, had a complete backbone but contained four residues with incomplete side chains and 14 residues with alternate locations.

Predicting biological activity

The online prediction of activity spectra for substances (PASS) service is a powerful tool for predicting the biological activities of chemical compounds, including their antiviral, antioxidant, antibacterial, antifungal, and antineoplastic properties. The 2D structures of the isoflavones were subjected to PASS service using the integrated Marvin JS tool.29,30

Prediction of pharmacokinetic properties and druglikeness

The physicochemical properties and druglikeness of all compounds were evaluated using the SwissADME web tool.31,32 Key molecular descriptors, including molecular weight, hydrogen bond donor/acceptor counts, topological polar surface area, and partition coefficients. A detailed in silico profiles for absorption, distribution, metabolism, excretion, and toxicity (ADMET) was generated using the pkCSM web tool.33,34 This analysis predicted crucial parameters such as human intestinal absorption (HIA), P-glycoprotein substrate status, blood-brain barrier (BBB) permeability, and AMES mutagenicity. Finally, a comprehensive toxicity profile was predicted using the ProTox 3.0 server.35,36 This prediction included median lethal dose for acute oral toxicity (LD50) values, toxicity class, and specific toxicological endpoints such as hepatotoxicity, carcinogenicity, and nephrotoxicity.

Molecular docking

Preparation of receptor

The AR structure was prepared for docking using the Dock Prep utility of the UCSF Chimera. This involved the removal of all solvents and hetero molecules, deletion of alternate conformations, and the completion of incomplete side chains. Subsequently, hydrogen atoms were added to protonate the structure at a physiological pH. Finally, Gasteiger charges were assigned to all receptor atoms.

Preparation of ligand

Ligand molecules were prepared using UCSF Chimera. Gasteiger charges were added, and energy minimization was performed using a the 100-step steepest descent method followed by a 10-step conjugate gradient method. The optimized structures were saved in the MOL2 format.

Molecular docking simulation

Molecular docking simulations were performed using AutoDock Vina.37,38 A docking grid of 20 × 20 × 20 Å dimensions was centered on the binding site of the AR at coordinates (26.78, 2.36, 4.61). The default settings were used for all other parameters.

To validate the docking protocol, the co-crystallized ligand was re-docked using the same parameters. The resulting docking pose for AR was nearly identical to the co-crystallized pose, with a root-mean-square deviation (RMSD) value of 0.241 Å. This RMSD value, being below the 2.0 Å threshold, confirms the validity of the docking protocol for predicting ligand-binding orientation.39 For each compound, the highest-scoring pose, representing the most stable conformation at the active site of the protein, was selected as the optimal docking pose (Figure 4). Protein ligand interactions were analyzed using the Protein-Ligand Interaction Profiler (PLIP) web tool.40 Visualizations of these interactions were generated using the PyMOL software.41

thumbnail of Figure 4
Figure 4
Validation of docking protocol. The co-crystallized ligand is shown in black, and the re-docked pose is shown in gray with an RMSD value of 0.241 Å.

The language of the manuscript was refined for clarity and readability using the artificial intelligence tools Grammarly and Gemini. The author evaluated all suggestions provided by these tools, incorporating only those changes that were deemed appropriate. The author is fully responsible for the final content of the article.

Results

PASS

The PASS program utilizes a vast database of known structure-activity relationships to make predictions with high accuracy.42 The program outputs two probability scores: Pa, representing the likelihood of a molecule being active, and Pi, representing the likelihood of it being inactive. These chances ranged from 0.00 to 1.00, and for a compound to be considered potentially active in a particular role, Pa should exceed Pi.29 The results indicated that all compounds in our study were likely to be active (Table 1).

Thumbnail
Table 1
Predicted biological activity spectra for the studied compounds using the PASS program

Physico-chemical properties, druglikeness and bioavailability

Druglikeness, which is determined by structural or physicochemical analyses of the generated compounds, refers to the likelihood of a molecule becoming an oral drug based on bioavailability. The SwissADME in silico program uses five filters (Lipinski, Veber, Egan, Ghose, Muegge) to estimate drug bioavailability, and all of them satisfied these five filters. The physicochemical properties of the compounds are detailed in Table 2.

Thumbnail
Table 2
Physicochemical properties and predicted druglikeness of isoflavones and reference compounds

ADMET analysis

Pharmacokinetic and toxicity predictions were performed using both tools. The pkCSM server predicted that all the analyzed soy isoflavonoids were non-inhibitors of hERG I and, except for DHD and C4HE, they lacked mutagenic potential in the AMES test (Table 3).

Thumbnail
Table 3
Predicted pharmacokinetic properties by the ADMET model using pkCSM servers

The toxicity profiles were generated using the ProTox 3.0 server (Table 4). The predicted LD50 values for acute oral toxicity ranged from 500 to 5000 mg kg-1, corresponding to toxicity classes 4 and 5. These classes represent a lower acute oral toxicity profile, which is a favorable characteristic for further drug development studies.43-46 Most soy isoflavonoids, including 8HD and DHD, were predicted to be inactive against hepatotoxicity and neurotoxicity, in contrast to the active controls (TES, DHT, and R-bicalutamide). However, all soy isoflavonoids and their metabolites are predicted to be nephrotoxic. Regarding carcinogenicity and immunotoxicity, 8HD and DHD were predicted to be inactive, whereas the other metabolites showed activity against these effects. All compounds were predicted to be inactive with respect to mutagenicity and cytotoxicity. Collectively, these findings suggest a favorable profile for 8HD and DHD but highlight a potential nephrotoxicity concern that warrants further investigation.

Thumbnail
Table 4
Predicted toxicity profiles using ProTox 3.0 servers

Molecular docking simulation

Molecular docking simulations were conducted to investigate the interactions between isoflavones and the AR binding pocket. Simulations revealed that isoflavones bind to AR in a manner similar to that of TES, an endogenous ligand. The isoflavones exhibited favorable binding energies, ranging from −8.161 to −9.287 kcal mol-1, approaching that of TES (−10.927 kcal mol-1) (Table 5). The binding energies of DHT and R-bicalutamide −10.928 and −7.538 kcal mol-1, respectively. These findings suggest that isoflavones can effectively occupy the AR binding pocket, indicating their potential to act as competitive ligands for receptor.

Thumbnail
Table 5
Binding energy and H-bonding interaction of the compounds with the amino acid residues of the androgen receptor

The binding poses and corresponding binding energies were used to assess the interactions between the isoflavones and AR. Binding energy is a commonly used parameter in the literature to evaluate the stability of protein-ligand complexes, with more negative values suggesting a higher binding affinity and potential for a strong interaction.47 The strong binding energies observed for the isoflavones are a direct result of specific interactions within the binding site of the protein, primarily hydrogen bonds with crucial residues such as Arg752 and Thr877 and, in some cases, π-stacking with Phe764, as detailed in Table 5.

The binding energies of all the complexes are listed in Table 5 and shown in Figures 5-7. All complexes exhibited favorable binding energies within the range established in the literature. The 8HD-AR complex exhibited the lowest energy, followed by the DHG-AR, 6HD-AR, 3HD-AR, and equol-AR complexes. All the binding energies presented were lower (indicating stronger binding) than those of the antagonist, R-bicalutamide, but higher (indicating weaker binding) than those of TES and DHT. These results suggest that these isoflavones interact effectively with the AR binding site.

thumbnail of Figure 5
Figure 5
Three-dimensional view of testosterone binding with the androgen receptor. Testosterone is shown in orange, and key interacting amino acid residues are shown in blue. Hydrogen bonds are shown as solid blue lines and hydrophobic interactions as gray dashed lines. Distances are shown in Å.

thumbnail of Figure 6
Figure 6
Three-dimensional view of isoflavonoids binding with the androgen receptor. The ligand for each panel: (a) daidzein; (b) genistein; (c) glycitein; (d) formononetin; (e) equol; (f) DHD; (g) ODMA; (h) 8HD. The color and interaction representations are the same as described in Figure 5.

thumbnail of Figure 7
Figure 7
Three-dimensional view of isoflavonoids binding with the androgen receptor. The ligand for each panel: (a) 6HD; (b) 3HD; (c) C4HE; (d) DHG; (e) 6H-ODMA; (f) TDZ. The color and interaction representations are the same as described in Figure 5.

In total, 168 PDB files containing AR crystal structures with TES and DHT natural ligands were examined, and all files were analyzed using the PLIP web tool. The results are summarized in Table S1 (Supplementary Information section). Hydrogen bonds were observed in these crystal structures with amino acids Asn705, Asn706, Gln711, Gln712, Arg752, Arg753, Thr877, Thr878, and Phe764. In the AR crystal structures containing DHT, hydrogen bonds were found with Asn705, Asn706, Gln711, Gln712, Arg752, Arg753, Thr877, and Thr878. The most common amino acids involved in hydrogen bonding across the examined PDB files were Asn705, Arg752, and Thr877. In the AR crystal structures containing TES, hydrogen bonds were observed with Asn705, Gln711, Arg752, Phe764, and Thr877, with Asn705 and Arg752 being the most frequently involved. In the different AR crystal structures containing both TES and DHT, Asn705 and Arg752 consistently formed hydrogen bonds.

In our study, when the interactions of amino acid residues that occur when soy isoflavones bind to AR were analyzed, all isoflavones formed at least one hydrogen bond with Arg752, and all isoflavones except for ODMA formed at least one hydrogen bond with Thr877. Met745 formed hydrogen bonds with daidzein, genistein, equol, ODMA, 6HD, 3HD, and C4HE. Analysis of other amino acids revealed that Leu704 was hydrogen-bonded with genistein and DHG, Asn705 with ODMA and 6H-ODMA, and Phe764 with glycitein, whereas π-stacking interactions were observed with genistein and 6H-ODMA (Table 5). Numerous hydrophobic interactions with amino acid residues occurred after these interactions. These interactions may contribute to biological response observed in the literature.

Structure-activity relationship analysis

To elucidate the structural determinants responsible for the observed binding energies and interaction patterns, a comprehensive structure-activity relationship analysis was conducted. This analysis correlated the chemical structures of the parent soy isoflavones (Figure 1), their diverse metabolites (Figure 2), and their specific atomic positions as defined by the standard isoflavonoid numbering scheme presented in Figure 8, with their respective docking scores and hydrogen bonding profiles, as presented in Table 5.

thumbnail of Figure 8
Figure 8
Structure and numbering of the isoflavonoid skeleton.

Importance of the intact isoflavone core and C-ring modifications

The fundamental tricyclic isoflavonoid scaffold, comprising rings A, C, and B (Figure 8), appears to be crucial for AR binding.

C-ring saturation

Modifications to the C-ring significantly influenced binding energy. Metabolites with a saturated C2-C3 double bond, forming an isoflavone (dihydroisoflavone) structure, such as DHD (−8.821 kcal mol-1) and DHG (−9.042 kcal mol-1), exhibited lower binding energies than their parent isoflavones, daidzein (−8.695 kcal mol-1) and genistein (−8.613 kcal mol-1), respectively. This suggests that the increased flexibility or altered conformation conferred by the saturation of the C2-C3 bond may allow for a more favorable fit within the AR binding pocket. DHG showed one of the lowest binding energies, retaining interactions with Leu704 (similar to genistein) and forming robust hydrogen bonds with Arg752 and Thr877 (Table 5).

C4-keto group reduction and isoflavan formation

Further reduction of the C4-keto group (C4=O in Figure 8) leading to isoflavan structures such as equol (−8.826 kcal mol-1), C4HE (−8.664 kcal mol-1), and TDZ (−8.536 kcal mol-1) resulted favorable binding energies, with equol and C4HE showing values comparable or superior to daidzein (−8.695 kcal mol-1). For instance, equol maintained key H-bonds with Arg752, Met745, and Thr877 (Table 5).

C-ring opening

In stark contrast, the metabolites in which the C-ring was opened, namely ODMA (−8.161 kcal mol-1) and 6H-ODMA (−8.219 kcal mol-1), which lacked the intact C-ring shown in Figure 8, displayed the highest (least favorable) binding energies among the tested compounds (Table 5). This strongly suggests that the integrity of the C-ring is crucial for effective AR interactions. Despite their overall weaker predicted binding, ODMA and 6H-ODMA uniquely form hydrogen bonds with Asn705 (Table 5), a residue known to interact with natural androgens. This may indicate an alternative binding mode for these C-ring opened structures, although they are less potent.

Influence of hydroxyl (–OH) substituents on A and B rings

The number and position of hydroxyl groups on the A and B rings (Figure 8) played a significant role in modulating AR binding.

Hydroxylation of daidzein scaffold

The introduction of an additional hydroxyl group onto the daidzein scaffold (Figure 1) generally improved the binding. For instance, 8HD with a hydroxyl group at the C8 position of ring A, and a docking score of −9.287 kcal mol-1 was the most potent AR binder among all tested compounds. Similarly, 6HD, hydroxylated at the C6 position of ring A (−9.039 kcal mol-1), and 3HD, hydroxylated at the C3’ position of ring B (−9.037 kcal mol-1) (Figure 2), exhibited lower (more favorable) binding energies than daidzein. This enhancement was likely due to the ability of these additional hydroxyls groups to form more extensive and strategically positioned hydrogen bonds. For example, 8HD, 6HD, and 3HD formed two hydrogen bonds with Thr877 compared to daidzein and consistently interacted with Arg752 (Table 5).

Effect of the C5-hydroxyl group in genistein

Genistein (Figure 1), which possesses an additional hydroxyl group at the C5 position of ring A compared to daidzein, showed a comparable binding energy (−8.613 kcal mol-1). However, it established unique interactions with Leu704 and Phe764 (including π-stacking with Phe764; Table 5), suggesting that C5-OH alters the binding orientation or allows access to different sub-pockets.

Influence of methoxyl (–OCH3) substituents

The replacement of hydroxyl groups with methoxy groups had a position-dependent effect on the binding energy. Formononetin (Figure 1), where the hydroxyl at the C4’ position of ring B (Figure 8) of daidzein is methoxylated, had a slightly stronger binding energy (−8.761 kcal mol-1 vs. −8.695 kcal mol-1 for daidzein). This stronger binding occurred despite the loss of a hydrogen-bond donor at the 4’-position. Glycitein (Figure 1), featuring a methoxy group at the C6 position and a hydroxyl at the C7 position of ring A (Figure 8), exhibited a weaker binding energy compared to daidzein (−8.278 kcal mol-1 vs. −8.695 kcal mol-1). This suggests that methoxylation at C6 may be less favorable than hydroxylation for AR binding, potentially due to steric hindrance or loss of H-bonding capability at this position.

Key amino acid interactions

Hydrogen bonds with Arg752 and Thr877 were consistently observed across most active isoflavonoids and their metabolites (Table 5), highlighting these residues as critical anchor points within the AR ligand-binding domain. The number and strength of these interactions are often correlated with lower binding energies. Interactions with Met745 were common among several potent binders. The π-stacking interaction with Phe764, observed for genistein and 6H-ODMA, likely contributes to the stabilization of these ligand-receptor complexes.

In summary, the structure-activity relationship analysis revealed that an intact tricyclic isoflavonoid core (rings A, B, and C, as shown in Figure 8), particularly a closed C ring, was crucial for strong AR binding. Saturation of the C-ring often leads to more favorable binding energies than those of the parent isoflavones. Increased hydroxylation, particularly at positions C8 and C6 (ring A) and C3’ (ring B), leads to stronger predicted binding, likely through enhanced hydrogen bonding capabilities. Methoxylation is less favorable than hydroxylation. Arg752 and Thr877 are the key interacting residues for this class of compounds.

Discussion

Molecular docking simulations were performed to investigated the potential of soy isoflavonoids as AR inhibitors. These simulations revealed interactions between soy isoflavonoids and AR at the amino acid level, suggesting that these compounds are promising candidates for prostate cancer therapy.

These results support the hypothesis that soy isoflavonoids and their metabolites competitively inhibit androgen binding to AR, reinforcing the hypothesis that these isoflavonoids are protective against prostate cancer.

In our study, the isoflavonoids, DHG, 3HD, 6HD, and 8HD exhibited the lowest AR binding energies. This suggests that DHG and 8HD may effectively block androgen action by binding to AR, potentially inhibiting the growth of prostate cancer cells. While our findings point specifically to this direct competitive inhibition, it is understood that soy isoflavonoids may influence the AR signaling pathway through multiple mechanisms. Broader literature suggests these can also include the inhibition of AR translocation to the nucleus, interference with TES synthesis, and conversion to DHT.48,49

Studies on the effects of soy isoflavonoids on prostate cancer have yielded mixed results. Several studies have indicated that soy isoflavonoids can reduce the risk of prostate cancer. For instance, soy isoflavonoids have been reported to control prostate tumor growth and metastasis and may enhance the therapeutic effects of radiotherapy.49 Another study proposed that soy isoflavonoids could inhibit prostate carcinogenesis by influencing genes that regulate cell cycle and apoptosis.48 In addition, some studies have shown that isoflavonoids such as genistein may exert anticancer effects by inhibiting the inflammatory response and cellular interactions.50

However, other studies have reported no clear associations or even contradictory results. For example, one study reported that pure genistein increased lymph node metastasis; however, a soy isoflavonoid mixture did not show this effect.49 Other studies have shown that soy isoflavonoids suppress the growth of prostate cancer cells. This effect is mediated by isoflavonoids that activate endogenous copper and trigger the formation of reactive oxygen species. When the copper chelator neocuproine is used, isoflavonoids can suppress the growth of prostate cancer cells and induce apoptosis by mobilizing endogenous copper.51 Depending on the species, dose, and individual metabolism, isoflavonoids have different effects. The effects of soy isoflavonoids may depend on factors such as the amount, form, and timing of consumption and the hormonal profile of individuals.50 Larger-scale, long-term, and multi-ethnic epidemiological studies are needed to reach more definitive conclusions.

Molecular docking studies have demonstrated the potential of isoflavonoids to bind to AR,52 and in vitro studies have confirmed their ability to inhibit prostate cancer cell growth.53 Several studies suggest that phytoestrogens may act as AR antagonists, potentially playing a preventive role in prostate cancer.54 For instance, isoflavonoids such as genistein have been shown to interact with AR and exhibit AR antagonist properties.54 Other studies observed that genistein showed a strong binding interaction with the agonist AR conformation and exhibited partial agonist activity when tested in the presence of a single androgen type.52,54 This suggests that the effects of isoflavonoids on AR may be complex and vary depending on the binding.

Our molecular docking studies support the potential of isoflavonoids to bind to AR and inhibit prostate cancer cell growth. However, its precise mechanisms of action are complex and require further investigation. Isoflavonoids are promising candidates for developing novel therapeutic strategies for prostate cancer.53,55 These findings are consistent with earlier investigations and epidemiological findings that show a negative relationship between the incidence of prostate cancer in Asian populations and the consumption of soy isoflavones.14,15 Although experimental studies have shown the ability of these compounds to inhibit prostate cancer cell growth and proliferation, it is essential to recognize that molecular docking simulations provide a theoretical framework that does not directly translate into biological activity. Further experimental studies are required to confirm the inhibitory effects of isoflavonoids on AR in cellular and animal models.

The predicted nephrotoxicity of all evaluated soy isoflavonoids poses a significant safety concern. Although computational, this finding of a consistent structural alert is a substantial hurdle for development, particularly for chronic therapies where long-term safety is paramount. Therefore, rigorous experimental validation is required to assess this risk and accurately define the therapeutic window of these compounds before they can be considered for clinical use.

Although this study offers promising results, it has certain limitations. Molecular docking simulations are performed using static protein structures, which may not fully capture the dynamic nature of protein-ligand interactions. Additionally, the in silico nature of this study necessitates further in vitro and in vivo validation to confirm the findings.

Conclusions

In this study, the potential of soy isoflavonoids and their metabolites to modulate AR signaling, a key pathway in prostate cancer, was systematically investigated using in silico molecular docking. Among the tested compounds, 8HD and DHG exhibited the lowest binding energies for AR, comparable to that of the endogenous ligand TES, suggesting their capacity to competitively inhibit androgen binding and thereby interfere with AR-mediated transcriptional activity.

These findings are consistent with epidemiological data indicating a reduced incidence of prostate cancer in populations with a high dietary intake of soy-based products. Furthermore, ADMET and drug-likeness analyses revealed favorable pharmacokinetic and safety profiles for most of the evaluated compounds, with the notable exception of predicted nephrotoxicity, underscoring their potential as lead candidates for further pharmacological development.

Nonetheless, the inherent limitations of molecular docking, including the use of rigid receptor conformations and the absence of biological system dynamics, necessitate cautious interpretation of these results. Therefore, subsequent studies should focus on experimental validation using in vitro assays and in vivo models to elucidate the precise biological effects and therapeutic efficacies of these isoflavonoids.

This study provides a theoretical framework supporting the role of specific soy isoflavonoids as promising modulators of AR signaling. These findings contribute to the growing body of literature advocating the exploration of dietary polyphenols for the chemoprevention and treatment of prostate cancer.

Supplementary Information

Supplementary data are available free of charge at http://jbcs.sbq.org.br as PDF file.

Data Availability Statement

The data that supporting the findings of this study are available in the article.

Acknowledgments

I would like to thank Ali Ceylan for all his devoted support.

References

  • 1 Bray, F.; Laversanne, M.; Sung, H.; Ferlay, J.; Siegel, R. L.; Soerjomataram, I.; Jemal, A.; Ca-Cancer J. Clin. 2024, 74, 229. [Crossref]
    » Crossref
  • 2 Delongchamps, N. B.; Singh, A.; Haas, G. P.; Cancer Control 2006, 13, 158. [Crossref]
    » Crossref
  • 3 Quinn, M.; Babb, P.; BJU Int. 2002, 90, 162. [Crossref]
    » Crossref
  • 4 Damaschke, N. A.; Yang, B.; Bhusari, S.; Svaren, J. P.; Jarrard, D. F.; Prostate 2013, 73, 1721. [Crossref]
    » Crossref
  • 5 Hsing, A. W.; Epidemiol. Rev. 2001, 23, 42. [Crossref]
    » Crossref
  • 6 Wigle, D. T.; Turner, M. C.; Gomes, J.; Parent, M. E.; J. Toxicol. Environ. Health, Part B 2008, 11, 242. [Crossref]
    » Crossref
  • 7 Powell, I. J.; J. Urol. 2007, 177, 444. [Crossref]
    » Crossref
  • 8 Nacusi, L. P.; Tindall, D. J.; Nat. Rev. Urol. 2011, 8, 378. [Crossref]
    » Crossref
  • 9 Chandrasekar, T.; Yang, J. C.; Gao, A. C.; Evans, C. P.; Transl. Androl. Urol. 2015, 4, 365. [Crossref]
    » Crossref
  • 10 Jacobsen, B. K.; Knutsen, S. F.; Fraser, G. E.; Cancer Causes Control 1998, 9, 553. [Crossref]
    » Crossref
  • 11 Applegate, C. C.; Rowles, J. L.; Ranard, K. M.; Jeon, S.; Erdman, J. W.; Nutrients 2018, 10, 40. [Crossref]
    » Crossref
  • 12 Hwang, K. A.; Choi, K. C.; Nutr. Cancer 2015, 67, 796. [Crossref]
    » Crossref
  • 13 Chavda, V. P.; Chaudhari, A. Z.; Balar, P. C.; Gholap, A.; Vora, L. K.; Phytother. Res. 2024, 38, 3060. [Crossref]
    » Crossref
  • 14 Yan, L.; Spitznagel, E. L.; Am. J. Clin. Nutr. 2009, 89, 1155. [Crossref]
    » Crossref
  • 15 Hwang, Y. W.; Kim, S. Y.; Jee, S. H.; Kim, Y. N.; Nam, C. M.; Nutr. Cancer 2009, 61, 598. [Crossref]
    » Crossref
  • 16 Hedlund, T. E.; Johannes, W. U.; Miller, G. J.; Prostate 2003, 54, 68. [Crossref]
    » Crossref
  • 17 Van der Eecken, H.; Joniau, S.; Berghen, C.; Rans, K.; De Meerleer, G.; Nutrients 2023, 15, 4856. [Crossref]
    » Crossref
  • 18 Rowland, I. R.; Wiseman, H.; Sanders, T. A.; Adlercreutz, H.; Bowey, E. A.; Nutr. Cancer 2000, 36, 27. [Crossref]
    » Crossref
  • 19 Setchell, K. D.; Clerici, C.; Lephart, E. D.; Cole, S. J.; Heenan, C.; Castellani, D.; Wolfe, B. E.; Nechemias-Zimmer, L.; Brown, N. M.; Lund, T. D.; Handa, R. J.; Heubi, J. E.; Am. J. Clin. Nutr. 2005, 81, 1072. [Crossref]
    » Crossref
  • 20 Lv, J.; Jin, S.; Zhang, Y.; Zhou, Y.; Li, M.; Feng, N.; Gut Pathog. 2024, 16, 35. [Crossref]
    » Crossref
  • 21 Chiou, Y.-S.; Wu, J.-C.; Huang, Q.; Shahidi, F.; Wang, Y.-J.; Ho, C.-T.; Pan, M.-H.; J. Funct. Foods 2014, 7, 3. [Crossref]
    » Crossref
  • 22 Heinonen, S.; Wahala, K.; Adlercreutz, H.; Anal. Biochem. 1999, 274, 211. [Crossref]
    » Crossref
  • 23 Yuan, J. P.; Wang, J. H.; Liu, X.; Mol. Nutr. Food Res. 2007, 51, 765. [Crossref]
    » Crossref
  • 24 Kim, S.; Thiessen, P. A.; Bolton, E. E.; Chen, J.; Fu, G.; Gindulyte, A.; Han, L.; He, J.; He, S.; Shoemaker, B. A.; Wang, J.; Yu, B.; Zhang, J.; Bryant, S. H.; Nucleic Acids Res. 2016, 44, D1202. [Crossref]
    » Crossref
  • 25 PubChem; https://pubchem.ncbi.nlm.nih.gov, accessed in August 2025.
    » https://pubchem.ncbi.nlm.nih.gov
  • 26 UCSF Chimera; Resource for Biocomputing, Visualization, and Informatics (RBVI); University of California: USA, 2024.
  • 27 Pettersen, E. F.; Goddard, T. D.; Huang, C. C.; Couch, G. S.; Greenblatt, D. M.; Meng, E. C.; Ferrin, T. E.; J. Comput. Chem. 2004, 25, 1605. [Crossref]
    » Crossref
  • 28 RCSB; https://www.rcsb.org/structure/2AM9, accessed in August 2025.
    » https://www.rcsb.org/structure/2AM9
  • 29 Filimonov, D. A.; Lagunin, A. A.; Gloriozova, T. A.; Rudik, A. V.; Druzhilovskii, D. S.; Pogodin, P. V.; Poroikov, V. V.; Chem. Heterocycl. Compd. 2014, 50, 444. [Crossref]
    » Crossref
  • 30 Lagunin, A.; Stepanchikova, A.; Filimonov, D.; Poroikov, V.; Bioinformatics 2000, 16, 747. [Crossref]
    » Crossref
  • 31 Daina, A.; Michielin, O.; Zoete, V.; Sci. Rep. 2017, 7, 42717. [Crossref]
    » Crossref
  • 32 SwissADME; http://www.swissadme.ch, accessed in August 2025.
    » http://www.swissadme.ch
  • 33 Pires, D. E.; Blundell, T. L.; Ascher, D. B.; J. Med. Chem. 2015, 58, 4066. [Crossref]
    » Crossref
  • 34 pkCSM; https://biosig.lab.uq.edu.au/pkcsm/, accessed in August 2025.
    » https://biosig.lab.uq.edu.au/pkcsm/
  • 35 Drwal, M. N.; Banerjee, P.; Dunkel, M.; Wettig, M. R.; Preissner, R.; Nucleic Acids Res. 2014, 42, W53. [Crossref]
    » Crossref
  • 36 ProTox-3.0; https://tox.charite.de/protox3/, accessed in August 2025.
    » https://tox.charite.de/protox3/
  • 37 AutoDock Vina; Forli Lab, The Scripps Research Institute, USA, 2023.
  • 38 Eberhardt, J.; Santos-Martins, D.; Tillack, A. F.; Forli, S.; J. Chem. Inf. Model. 2021, 61, 3891. [Crossref]
    » Crossref
  • 39 Cole, J. C.; Murray, C. W.; Nissink, J. W.; Taylor, R. D.; Taylor, R.; Proteins:Struct., Funct., Genet. 2005, 60, 325. [Crossref]
    » Crossref
  • 40 Adasme, M. F.; Linnemann, K. L.; Bolz, S. N.; Kaiser, F.; Salentin, S.; Haupt, V. J.; Schroeder, M.; Nucleic Acids Res. 2021, 49, W530. [Crossref]
    » Crossref
  • 41 The PyMOL Molecular Graphics System; Schrödinger: USA, 2024.
  • 42 Poroikov, V. V.; Filimonov, D. A.; Ihlenfeldt, W. D.; Gloriozova, T. A.; Lagunin, A. A.; Borodina, Y. V.; Stepanchikova, A. V.; Nicklaus, M. C.; J. Chem. Inf. Comput. Sci. 2003, 43, 228. [Crossref]
    » Crossref
  • 43 Guengerich, F. P.; Drug Metab. Pharmacokinet. 2011, 26, 3. [Crossref]
    » Crossref
  • 44 Hughes, J. P.; Rees, S.; Kalindjian, S. B.; Philpott, K. L.; Br. J. Pharmacol. 2011, 162, 1239. [Crossref]
    » Crossref
  • 45 Lee, H.; Kim, J.; Kim, J. W.; Lee, Y.; Front. Chem. 2025, 13, 1632046. [Crossref]
    » Crossref
  • 46 Uesawa, Y.; Toxicol. Res. 2024, 40, 1. [Crossref]
    » Crossref
  • 47 Guedes, I. A.; de Magalhaes, C. S.; Dardenne, L. E.; Biophys. Rev. 2014, 6, 75. [Crossref]
    » Crossref
  • 48 Mahmoud, A. M.; Yang, W.; Bosland, M. C.; J. Steroid Biochem. Mol. Biol. 2014, 140, 116. [Crossref]
    » Crossref
  • 49 Raffoul, J. J.; Banerjee, S.; Che, M.; Knoll, Z. E.; Doerge, D. R.; Abrams, J.; Kucuk, O.; Sarkar, F. H.; Hillman, G. G.; Int. J. Cancer 2007, 120, 2491. [Crossref]
    » Crossref
  • 50 Naponelli, V.; Piscazzi, A.; Mangieri, D.; Int. J. Mol. Sci. 2025, 26, 1114. [Crossref]
    » Crossref
  • 51 Farhan, M.; El Oirdi, M.; Aatif, M.; Nahvi, I.; Muteeb, G.; Alam, M. W.; Molecules 2023, 28, 2925. [Crossref]
    » Crossref
  • 52 Wang, H.; Li, J.; Gao, Y.; Xu, Y.; Pan, Y.; Tsuji, I.; Sun, Z. J.; Li, X. M.; Asian J. Androl. 2010, 12, 535. [Crossref]
    » Crossref
  • 53 Shenouda, N. S.; Zhou, C.; Browning, J. D.; Ansell, P. J.; Sakla, M. S.; Lubahn, D. B.; Macdonald, R. S.; Nutr. Cancer 2004, 49, 200. [Crossref]
    » Crossref
  • 54 Sivonová, M. K.; Kaplán, P.; Tatarková, Z.; Lichardusová, L.; Dusenka, R.; Jureceková, J.; Mol. Clin. Oncol. 2019, 10, 191. [Crossref]
    » Crossref
  • 55 Singh, A. N.; Baruah, M. M.; Sharma, N.; Sci. Rep. 2017, 7, 1955. [Crossref]
    » Crossref

Edited by

  • Editor handled this article:
    Paula Homem-de-Mello (Executive)

Publication Dates

  • Publication in this collection
    03 Nov 2025
  • Date of issue
    2025

History

  • Received
    10 June 2025
  • Accepted
    04 Sept 2025
Creative Common - by 4.0
This is an Open Access article distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.
location_on
Sociedade Brasileira de Química Instituto de Química - UNICAMP, Caixa Postal 6154, 13083-970 Campinas SP - Brazil, Tel./FAX.: +55 19 3521-3151 - São Paulo - SP - Brazil
E-mail: office@jbcs.sbq.org.br
rss_feed Acompanhe os números deste periódico no seu leitor de RSS
Reportar erro