Comprehensive analysis of the potential and mechanisms of soy isoflavones against breast cancer: An integrated study of network pharmacology, molecular docking, molecular dynamics simulation, Mendelian randomization, and experimental validation

2.1. Identification of targets for each soy isoflavone component and BC

The six components of soy isoflavones were sourced from the literature, as previously detailed. Their structural formulas and SMILES notations were retrieved from the PubChem database (https://pubchem.ncbi.nlm.nih.gov/, accessed on 20 November 2024). Table 1 presents the chemical structures of these six components. Potential targets for these compounds were identified using the Similarity Ensemble Approach (SEA) Search Server (https://sea.bkslab.org/, accessed on 20 November 2024) and Swiss Target Prediction (STP) database (http://swisstargetprediction.ch/, accessed on 20 November 2024) based on the SMILES notations.

Table 1. Active components and structure of soy isoflavones.

Compound name	Molecular formula	Molecular weight (g/mol)	Structure
Daidzin	C21H20O9	416.4
Daidzein	C15H10O4	254.24
Genistin	C21H20O10	432.4
Genistein	C15H10O5	270.24
Glycitin	C22H22O10	446.4
Glycitein	C16H12O5	284.26

Potential targets related to BC were obtained from three sources: Genecards (https://www.genecards.org/, accessed on 21 November 2024), TTD (https://db.idrblab.net/ttd/, accessed on 21 November 2024), and OMIM (https://www.omim.org/, accessed on 21 November 2024). Targets from Genecards were filtered by relevance score ≥ 10.

2.2. Identification of targets for soy isoflavones in BC and construction of a protein-protein interaction (PPI) network

The potential targets of soy isoflavones and BC from multiple databases were merged, and duplicates were removed. The intersection of these targets was determined using the Venn Diagram package and visualized with ggplot2 in R software (version 4.2.1). The resulting intersecting targets were used to generate a PPI network via the STRING database (https://cn.string-db.org/, accessed on 21 November 2024), with a minimum interaction score of 0.9. Cytoscape software (version 3.9.1) was employed for network visualization and analysis.

2.3. Identification of hub targets

The network analysis tool in Cytoscape software was employed to analyze the topological parameters of the PPI network, including degree and betweenness centrality, which serve as indicators of node importance. Targets with values exceeding twice the mean for both degree and betweenness centrality were categorized as hub targets.

2.4. Enrichment analyses

To further explore the effect of soy isoflavones on BC, Gene Ontology (GO), and Kyoto Encyclopedia of Genes and Genomes (KEGG) enrichment analyses were conducted. GO analysis focused on the biological process (BP), cellular component (CC), and molecular function (MF) categories, while KEGG analysis identified key pathways involved in the impact of soy isoflavones on BC.

2.5. Molecular docking analysis

SDF format files for the six active components were obtained from PubChem (accessed on 26 November 2024). From the PDB database, 3D protein structures in PDB format for the ten hub targets were sourced: AKT1 (PDB ID:1H10), epidermal growth factor receptor (EGFR) (PDB ID: 1IV0), estrogen receptor 1 (ESR1) (PDB ID: 2BJ4), HRAS (PDB ID: 121P), HSP90AA1 (PDB ID: 1BYQ), ITGB1 (PDB ID: 3G9W), PIK3CA (PDB ID: 2ENQ), PRKACA (PDB ID: 2GU8), SRC (PDB ID: 1A07), TP53 (PDB ID: 1A1U) (https://www.rcsb.org/, accessed on 26 November 2024). Ligands and protein receptors were optimized using Chem3D and PyMOL, respectively. Molecular docking of the six active components with the ten hub targets was performed using AutoDockTools (version 1.5.7), with binding energies calculated and visualized using PyMOL. The parametric details of molecular docking are available in the Table S1.

2.6. Molecular dynamics simulation

Molecular dynamics simulations were conducted using Gromacs (version 2022.2) to further examine the binding affinity and stability of proteins with ligands, focusing on two complexes: PRKACA-DA and PRKACA-GE. The force field parameters were obtained through the pdb2gmx tool of Gromacs and the AutoFF web page. The CHARMM 36 force field was applied for protein modeling, and ligand topologies were derived from the CGenff force field parameters. A 1 nm TIP3P-type cubic water box was added around the system for solvation. Ions were added to the system using the gmx genion tool to achieve electroneutrality of the system. Long-range electrostatic interactions were processed by the Particle Mesh Ewald method with the cutoff distance set to 1 nm. The integration step during the molecular dynamics simulation was set to 2 fs using the Verlet Leapfrog Algorithm. Prior to the molecular dynamics simulations, the system underwent energy optimization. The energy minimization process consisted of 3000 steps of the steepest descent method optimization, followed by 2000 steps of the conjugate gradient method optimization. The optimization steps are as follows: first, constrain the solute and minimize the energy for the water molecules; then, constrain the counterions and minimize the energy; and finally, minimize the energy for the whole system without constraints. The simulated operating conditions were an isothermal-isobaric ensemble system at a temperature of 310 K and constant pressure with a time step of 2 fs and a simulation time of 100 ns.

2.7. MR analysis

Two-sample MR was employed to investigate the causal relationship between soy consumption and BC risk. Genetic instruments representing soy intake were based on GWAS data for soya dessert (GWAS ID: ukb-b-998) and tofu intake (GWAS ID: ukb-b-5522) sourced from the IEU database (https://gwas.mrcieu.ac.uk/) with a significance threshold of p < 5 × 10^-7. Linkage disequilibrium clumping was performed (clump window = 10,000 kb, r² = 0.001) to ensure independence among single-nucleotide polymorphisms (SNPs). GWAS summary data for BC were retrieved from the Breast Cancer Association Consortium (BCAC), based on the study by Michailidou et al. [17], which included 122,977 cases and 105,974 controls. MR analyses were conducted using the TwoSampleMR and MRPRESSO R packages (version 4.3.1), with three methods, inverse variance weighted (IVW), MR-Egger, and weighted median, applied to assess the relationship between soy product consumption and the risk of BC, including ER-positive subtypes.

2.8. Cell culture

MCF-7 and MDA-MB-231 BC cell lines were utilized for in vitro cellular experiments. Both cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) (4.5 g glucose, Gibco, USA) supplemented with 10% fetal bovine serum (FBS) and 1% penicillin/streptomycin (Gibco, USA).

2.9. Cell viability assay

Cells were seeded into 96-well plates at a density of 5,000 cells per well. After the cells adhered, the culture medium was replaced with fresh medium containing varying drug concentrations, and the cells were cultured for 48 h. Following this, the medium was discarded, and 100 μL of fresh medium containing 10% CCK-8 reagent was added to each well. After a 2 h incubation, optical density (OD) at 450 nm was measured to assess cell viability.

2.10. Cell transfection

Cell transfection was performed using Lipofectamine 3000 and plasmids, with transfection efficiency evaluated by Western blot analysis.

2.11. Wound healing assay

Wound healing assays were conducted in 6-well plates. Once the cells reached 90% confluence, scratches were made using a 1 mL pipette tip. The medium was changed to the drug-containing medium containing two different concentrations, and cells were cultured further. Wound closure was monitored and photographed at 0 and 48 h under a microscope at 4x magnification.

3.1. Screening of potential anti-BC targets for soy isoflavones

The STP and SEA databases predicted 621 and 501 potential targets, respectively, for the six components of soy isoflavones. After removing duplicates, a total of 362 unique potential targets for soy isoflavones were identified. For BC, potential targets were retrieved from Genecards (3398 targets), TTD (97 targets), and OMIM (95 targets), resulting in 3508 unique targets after duplicates were excluded. The intersection of the drug and disease targets yielded 190 common targets for soy isoflavones in BC (Figure 1). A complete list of soy isoflavone targets, BC targets, and their intersections has been provided in Table S2.

Close

Figure 1.

Common targets of soy isoflavones and BC.

Export to PPT

3.2. Constructing active ingredient-anti-BC networks

To investigate the relationship between the six active ingredients of soy isoflavones and BC, the 190 intersecting genes were used to construct an active ingredient-anti-BC target network (Figure 2). Among the six ingredients, GEN was the most significant, with 87 targets, followed by DAI (79), GE (74), GLY (73), GL (72), and DA (66). However, the differences in target numbers across the ingredients were minimal, indicating similar potency among the components.

Close

Figure 2.

Construction of the soy isoflavones-active components-targets network.

Export to PPT

3.3. GO and KEGG pathway enrichment analyses

For functional analysis, GO and KEGG analyses were performed on the 190 intersecting genes to explore the biological effects of soy isoflavones on BC and the signaling pathways potentially involved. A total of 178 KEGG pathways and 2945 GO terms were identified, with 2643 terms related to BP, 112 related to CC, and 190 related to MF (Table S3). Key BPs involved in the action of soy isoflavones on BC included responses to oxidative stress, such as responses to oxygen levels, hypoxia, and reactive oxygen species. Other BPs included epithelial cell proliferation, response to steroid hormones, and wound healing. The main MFs regulated by soy isoflavones were primarily associated with protein kinase activity. The intersecting genes were predominantly localized in intracellular and extracellular structures, including membrane rafts, membrane microdomains, vesicle lumens, and secretory granule lumens. Figures 3(a–c) show the top 20 BP, CC, and MF terms.

Close

Figure 3.

GO and KEGG analysis of 190 intersecting targets. (a-c) Main (a) BPs, (b) MFs, and (c) CCs involved in the effects of soy isoflavones on BC. (d) KEGG analysis results of 190 intersecting targets.

Export to PPT

KEGG pathway enrichment analysis revealed that the 190 intersecting genes were predominantly enriched in the EGFR tyrosine kinase inhibitor resistance, PI3K-Akt signaling pathway, MAPK signaling pathway, estrogen signaling pathway, endocrine resistance, and multiple cancer pathways (Figure 3d). These results suggest that the therapeutic effects of soy isoflavones on BC may be linked to BC drug sensitivity, particularly in relation to endocrine therapy and anti-HER-2 targeted therapies.

3.4. Constructing the PPI network and identifying the hub targets

The PPI network of the 190 intersecting genes was constructed using the STRING database with a minimum interaction score of 0.9 (Figure 4a). After removing isolated targets, the network was visualized using Cytoscape, where node color and size represented degree values (Figure 4b). To identify hub targets, both the degree value and betweenness centrality of the nodes were calculated, as these metrics reflect the importance of the targets in the network. The mean degree and betweenness centrality values for the PPI network were 11.9 and 0.019, respectively. Targets with degree and betweenness centrality values greater than twice the mean were considered hub targets. A total of 10 hub targets were identified, including TP53, SRC, ESR1, EGFR, PIK3CA, HSP90AA1, PRKACA, HRAS, AKT1, and ITGB1 (Table 2) (Figure 4c).

Close

Figure 4.

Screening of 10 hub targets. (a-b) PPI network of 190 intersecting targets obtained from (a) the STRING database and (b) Cytoscape software. (c) Identification of 10 hub targets through screening.

Export to PPT

3.5. Molecular docking analysis

Molecular docking was performed in triplicate for 10 hub targets and six active compounds, and the average binding energy was determined. Lower binding energies indicate stronger binding affinities, with values < -5.0 kcal/mol suggesting potential binding activity and values < -7.0 kcal/mol representing strong binding interactions [18]. The results of the three replicate experiments have been presented in Table S4, while Figure 5 displays the average binding energies across all 60 protein-ligand complexes. All complexes have binding energies below -5 kcal/mol, with 42 complexes having binding energies below -7 kcal/mol. Notably, PRKACA-DA and PRKACA-GE exhibited the strongest binding, with a binding energy of -10.5 kcal/mol. PyMOL was used to visualize the molecular docking of the six top protein-ligand complexes (PRKACA-DA, PRKACA-GE, PRKACA-GL, HRAS-DA, ESR1-GEN, HRAS-GE) with the highest binding affinities (Figure 6). These protein-ligand complexes were stabilized by multiple hydrogen bonds. Specifically, in PRKACA, DA formed hydrogen bonds with ARG519, LYS168, THR183, and VAL123 (Figure 6a). GE interacted with LYS168, THR183, and VAL123 (Figure 6b). GL formed hydrogen bonds with ARG519, LYS168, ALA521, THR51, and VAL123 (Figure 6c). In HRAS, DA formed hydrogen bonds with THR35, GLY60, LYS117, and ASP119 (Figure 6d), while GE interacted with GLY60 and LYS117 (Figure 6e). In ESR1, GEN formed hydrogen bonds with HIS524 and ARG394 (Figure 6f).

Close

Figure 5.

Molecular docking binding energy (kcal/mol) of 10 hub targets and 6 soy isoflavone active components.

Export to PPT

Close

Figure 6.

Molecular docking results of the active component with the target. (a-f) Molecular docking plots: (a) PRKACA-DA, (b) PRKACA-GE, (c) PRKACA-GL, (d) HRAS-DA, (e) HRAS-GE, (f) ESR1-GEN.

Export to PPT

3.6. Molecular dynamics simulations

Molecular dynamics simulations were performed on the two complexes (PRKACA-DA and PRKACA-GE) with the highest binding capacities to further assess their stability. The root mean square deviation (RMSD) was used to evaluate the conformational stability of the proteins and ligands, with an RMSD within 1 nm indicating stable protein-ligand interactions [19]. Both complexes, PRKACA-DA and PRKACA-GE, reached equilibrium at 65 ns and 10 ns, respectively, with final RMSD values of approximately 2.3 Å. This indicated that both complexes maintained high stability upon binding (Figure 7a). The root mean square fluctuation (RMSF), a metric used to assess the flexibility of amino acid residues in proteins, revealed that the RMSF values for both complexes were less than 3 Å (Figure 7b), further confirming the stability of the protein-ligand interactions. The radius of gyration (Rg) was employed to assess structural changes in the protein-ligand complex, with larger fluctuations in Rg indicating system swelling. As shown in Figure 7(c), the Rg values of both complexes remained stable throughout the simulations, with average values of 20.4 ± 0.099 Å and 20.5 ± 0.091 Å, respectively. This suggests that the complexes remained compact during the simulations. Additionally, the solvent-accessible surface area (SASA) was evaluated to provide insight into protein folding and stability, and the SASA values remained stable throughout the simulations (Figure 7d). Hydrogen bonds, which play a key role in ligand-protein binding, were also analyzed. The number of hydrogen bonds between the two active components and the PRKACA protein fluctuated between 0 and 7, with the number typically being 4 for PRKACA-DA and 5 for PRKACA-GE (Figure 7e). This indicates that both active components maintained strong hydrogen bonding interactions with the PRKACA protein, supporting the stability of the complexes.

Close

Figure 7.

Results of molecular dynamics simulations of PRKACA-DA and PRKACA-GE. (a-e) The following values of the two complexes: (a) RMSD, (b) RMSF, (c) Rg, (d) SASA, (e) Hbond numbers.

Export to PPT

3.7. MR analysis

To further explore the causal relationship between soy isoflavones and BC risk, a two-sample MR analysis was conducted. The results did not support a causal relationship between soy product consumption and reduced BC risk (Figure 8). Given the known relationship between soy isoflavones, estrogen, and ER, the effect of soy product consumption on BC risk was specifically examined in patients with ER-positive BC. Like the overall analysis, the difference was not statistically significant.

Close

Figure 8.

Two-sample MR analysis between soy food consumption and BC risk. (a) Odds ratio plot between soy food consumption and BC risk. (b-c) Scatter plots of SNPs associated with (b) tofu and (c) soya dessert intake and overall BC risk. (d-e) Scatter plots of SNPs associated with (d) tofu and (e) soya dessert intake and the risk of ER-positive BC.

Export to PPT

3.8. DA and GE inhibit BC progression by inhibiting PRKACA

Molecular docking and molecular dynamics simulations revealed that among the soy isoflavones, DA and GE exhibited the strongest binding affinity to PRKACA. To evaluate the anticancer potential of these compounds, MCF-7 and MDA-MB-231 BC cell lines were used in a series of in vitro experiments. CCK-8 assays demonstrated that both DA and GE inhibited BC cell growth in a dose-dependent manner (Figures 9a,b). The IC50 values of DA for MCF-7 and MDA-MB-231 cells were 147.2 μM and 224.0 μM, respectively, while those of GE were 116.4 μM and 241.9 μM, respectively. Both compounds demonstrated potent inhibitory effects on BC cells, as evidenced by their low IC50 values, further supporting their potential role in suppressing BC progression. Additionally, wound healing assays indicated that DA and GE significantly inhibited BC cell migration (Figures 9c-f). These results collectively suggest that DA and GE impede the progression of BC by inhibiting cell proliferation and migration.

Close

Figure 9.

DA and GE inhibit BC progression in vitro. (a) DA and (b) GE inhibit BC cell growth in a dose-dependent manner. (c-d) Wound healing assays illustrating that (c) DA and (d) GE inhibit migration of BC cells. (e-f) Statistics on the area of wound migration of BC cells inhibited by (e) DA and (f) GE. (*p < 0.05; **p < 0.01; ***p < 0.001; ****p < 0.0001).

Export to PPT

PRKACA exhibits a strong binding affinity for both DA and GE. Western blot analysis further validated that DA and GE effectively inhibit the expression of PRKACA (Figure 10a). The TCGA-GTEx database analysis revealed that PRKACA expression was notably elevated in a range of tumor tissues, including BC (Figure 10b). Furthermore, patients with BC exhibiting high PRKACA expression experienced poorer overall survival (OS) (Figure 10c), underscoring the potential role of PRKACA as a prognostic marker in BC. To further validate the role of PRKACA in BC progression, plasmids were used to inhibit PRKACA expression in BC cells (Figures 10d-e). CCK-8 assays revealed that the knockdown of PRKACA significantly suppressed BC cell proliferation (Figure 10f), while wound healing assays demonstrated that PRKACA inhibition hindered BC cell migration (Figures 10g, h). These findings suggest that PRKACA may play an oncogenic role in BC, and the inhibitory effects of DA and GE on BC progression may be mediated by the suppression of PRKACA expression.

Close

Figure 10.

DA and GE inhibit BC progression by suppressing PRKACA expression. (a) DA and GE inhibit PRKACA expression in BC. (DA: 100μM, GE: 100μM) (b) PRKACA expression is elevated in various tumor tissues, including BC. (c) High PRKACA expression correlates with poor OS in BC. (d) PRKACA knockdown efficiency assessed by Western blot analysis following plasmid transfection. (e) Statistical results of PRKACA knockdown efficiency. (f) PRKACA inhibition suppresses BC cell proliferation. (g-h) PRKACA inhibition reduces the migration of (g) MCF-7 and (h) MDA-MB-231 cells. (*p < 0.05; ** p < 0.01; *** p < 0.001; **** p <0.0001).

Export to PPT

3.9. Discussion

The relationship between soy consumption and BC risk has long been debated. Early studies, noting the structural similarity between soy isoflavones and estrogen, cautioned that individuals at high risk for BC, particularly those with estrogen-sensitive BC, should minimize soy intake [20]. However, further research revealed that soy isoflavones exhibit markedly lower binding affinity to the ER than estrogen itself [21]. Soy isoflavones do not directly interact with the ER but instead function as modulators of estrogenic activity [22]. Numerous studies have demonstrated that soy isoflavones exert anti-estrogenic effects in the body, suggesting a protective role against BC rather than a contributory one [23]. This hypothesis has been validated by various case-control and prospective studies, which consistently report an inverse association between soy isoflavones and BC risk [24].

As the link between soy isoflavones and a reduced BC risk becomes clearer, understanding the underlying mechanisms has garnered increasing attention. While six major soy isoflavones are recognized, previous studies have largely focused on individual components. To explore the comprehensive mechanism by which soy isoflavones impact BC, a network pharmacology approach was employed, leveraging its multicomponent, multitarget, and multipathway characteristics. The network analysis revealed that all six soy isoflavone components (DA, GE, GL, DAI, GEN, and GLY) demonstrate substantial anti-BC activity, with values exceeding 65° for each component. Among these, GEN, the most extensively studied soy isoflavone, displayed the highest degree value. Enrichment analysis indicated that the effects of soy isoflavones on BC are largely mediated through the regulation of cellular oxidative stress, a mechanism that aligns with previous findings linking soy isoflavones to oxidative stress modulation [25]. Additionally, enrichment analysis revealed that estrogen signaling, and endocrine resistance pathways are key mediators of soy isoflavones’ activity against BC. These findings provide further evidence for the phytoestrogenic anti-BC mode of action of soy isoflavones.

Through systematic screening, 190 anti-BC targets of soy isoflavones were identified, leading to the construction of a PPI network. PPI networks serve as tools to elucidate the interactions between proteins within biological systems. These 190 targets were found to interact significantly with one another, and the top ten targets with the highest degree of intersection within the network were designated as hub targets: TP53, SRC, ESR1, EGFR, PIK3CA, HSP90AA1, PRKACA, HRAS, AKT1, and ITGB1. These hub targets are known to play critical roles in BC pathogenesis. TP53, a well-known tumor suppressor gene, is often inactivated or mutated in various cancers, including BC [26]. SRC, the first identified proto-oncogene in mammals, regulates cell proliferation, differentiation, and migration, and is instrumental in the initiation and progression of BC [27]. ESR1, which encodes ER1, is implicated in acquired resistance to estrogen-based therapies in patients with BC [28]. EGFR, frequently overexpressed in BC, particularly in triple-negative subtypes, is essential for maintaining BC cell characteristics [29]. PIK3CA and AKT1 are key proto-oncogenes in the PI3K/AKT signaling pathway, which is often aberrantly activated in BC due to mutations in these genes [30]. HSP90AA1, encoding heat shock protein 90 (HSP90), has been identified as a significant marker for BC onset and metastasis prediction [31]. HRAS and ITGB1 are involved in the regulation of apoptosis, migration, and invasion in BC cells [32,33]. Additionally, PRKACA has been implicated in mediating resistance to anti-HER2 therapies and endocrine treatments in BC [34,35].

These ten hub targets were identified as potential mediators of BC risk reduction through soy isoflavones. To further validate the interaction between soy isoflavones and these targets, molecular docking was employed. The six soy isoflavone components were found to exhibit strong binding affinities with all ten hub targets, likely due to the distinctive molecular structure of flavonoids [36]. The unique scaffold of flavones contributes to their diverse biological activities. Specifically, their hydroxyl groups can form hydrogen bonds with protein carboxyl groups, while their aromatic rings interact with protein hydrophobic regions. These structural properties underscore the widespread application of flavones in natural drug development [37]. Notably, PRKACA demonstrated the most robust binding capacity, with binding energies consistently below -8.5 kcal/mol across all soy isoflavone components. Among them, DA and GE showed particularly strong binding to PRKACA, with binding energies of -10.5 kcal/mol, indicating a high-affinity interaction. Previous studies have demonstrated that both DA and GE exhibit higher chemical reactivity and bioavailability than other related compounds, highlighting their potential as effective bioactive therapeutic agents [38,39]. Molecular dynamics simulations further corroborated these findings, showing that the PRKACA-DA and PRKACA-GE complexes maintained structural stability during binding, with numerous hydrogen bonds stabilizing the complexes. This suggests that both protein-ligand complexes exhibit strong affinity, consistent with the docking results.

Given the strongest binding affinity between DA, GE, and PRKACA, bioinformatics analysis revealed that PRKACA is highly expressed in BC and associated with poor prognosis, identifying it as a potential target for soy isoflavones in BC risk reduction. Cytological experiments confirmed that DA and GE inhibit BC cell proliferation in a dose-dependent manner. Moreover, wound healing assays verified that DA and GE suppress BC cell migration. Western blot analysis further showed that DA and GE reduce PRKACA expression in BC cells, suggesting that PRKACA downregulation may be the mechanism through which DA and GE hinder BC progression.

Although cytologic experiments revealed that soy isoflavones may have therapeutic potential to slow the progression of BC, the two-sample MR analysis did not support a causal relationship between soy product consumption and reduced BC risk. This lack of statistical significance may stem from the use of GWAS data derived from a European population. While the average daily intake of soy isoflavones in European countries is less than 2 mg, Asian populations, particularly the Chinese, Japanese, and South Koreans, consume significantly higher amounts, typically ranging from 25-50 mg/day [40]. A dose-response meta-analysis of soy isoflavone intake and BC risk in an Asian population showed a 3% reduction in the risk of BC for every 10mg/day increase in soy isoflavone intake [24]. Although the findings of our study do not support the use of soy isoflavones as a preventive dietary component. We expect that future GWAS studies in Asian populations may yield new mechanistic insights into this protective association. In addition, we are looking forward to the development of more synthetic flavonoid derivatives with enhanced bioavailability [41].

This study represents the first comprehensive exploration of the mechanisms underlying the effects of soy isoflavones on BC, examining the collective action of multiple active components rather than focusing on a single ingredient. Additionally, two active components, DA and GE, were identified and subjected to cytological analysis, with DA also being the first soy isoflavone studied for its impact on BC. However, the study has limitations. First, although the reduced expression of PRKACA has been preliminarily linked to the potential mechanism by which soy isoflavones exert anti-BC effects, further research is required to validate this hypothesis. Second, this study only used two BC cell lines (MCF-7 and MDA-MB-231), which do not cover all BC subtypes. Further research with more cell models is needed to better understand the roles of DA and GE in BC.