Novel Chalcone Derivatives as Potent Lyn Tyrosine Kinase Inhibitors: A Promising In-Silico Approach for Targeted Therapy in Triple-Negative Breast Cancer ^†

Abstract

Triple-negative breast cancer (TNBC) accounts for approximately 10–15% of breast cancer cases and poses a significant clinical challenge due to its aggressive nature and poorer survival outcomes compared to other subtypes. This is primarily attributed to the lack of estrogen, progesterone, and HER2 receptors, which render conventional hormone-based therapies ineffective. In this study, we employed in silico approaches to design and evaluate novel chalcone derivatives as potential inhibitors of Lyn tyrosine kinase, a critical enzyme implicated in TNBC progression. The designed compounds were screened for drug-likeness and toxicity, all meeting Lipinski’s rule of five and demonstrating favorable toxicity profiles. Molecular docking analyses identified five promising ligands, CHCN1, CHCN19, CHCN48, CHCN333, and CHCN94, that exhibited comparable binding affinities to key active site residues of Lyn kinase, including Asp385, Phe386, Gly387, Lys275, and Glu290. Among these, CHCN1 showed the highest binding affinity at –8.4 kcal/mol, likely due to interactions with Asp385 and Lys275. These results suggest that the chalcone derivatives may effectively disrupt Lyn-mediated signaling pathways essential for cancer cell survival, potentially inhibiting proliferation, metastasis, and invasion. Overall, this study provides valuable insights into the therapeutic potential of chalcone derivatives for TNBC, offering promising avenues for targeted intervention.

1. Introduction

Breast cancer remains the most frequently diagnosed cancer among women worldwide, with a steady increase in incidence since 2020. An estimated 13.1% of women in the U.S. are projected to be diagnosed with the disease in their lifetime [1]. With over 2 million new cases identified in 2020, breast cancer remains the most prevalent cancer diagnosed in women worldwide, highlighting a critical global health challenge [2]. Breast cancer is classified into three primary subtypes based on molecular markers for estrogen (ER), progesterone (PR), and human epidermal growth factor receptor 2 (HER2/ERBB2). Triple-negative breast cancer (TNBC), characterized by the absence of all three receptors, accounts for approximately 15% of all cases [3]. TNBC is particularly aggressive, with a higher likelihood of invasion, metastasis, and poor prognosis compared to other subtypes [4], making it a focal point for cancer research. Among the members of the Src kinase family, Lyn Tyrosine Kinase is uniquely implicated in TNBC progression, promoting increased cell migration and metastatic potential [5]. A study by [6] demonstrated that silencing Lyn, but not Src, significantly reduced invasion in mesenchymal breast cancer cell lines. Furthermore, the ratio of Lyn splice isoforms (LYNA to LYNB) has been correlated with TNBC patient outcomes, with a higher ratio associated with reduced overall survival [7]. Chalcones are a class of naturally occurring compounds predominantly found in plants. Their characteristic scaffold structure demonstrated anticancer activity, making them promising candidates for cancer treatment, including TNBC [8]. Recent studies, such as [9], have shown that chalcone derivatives, including novel acetamide forms, possess cytotoxic activity against TNBC cell lines. Although chemotherapy, particularly anthracycline- and taxane-based regimens, remains the frontline treatment for TNBC, the limited availability of effective targeted therapies underscores the urgent need for new therapeutic compounds. In this study, we evaluate the potential of newly designed chalcone derivatives to inhibit Lyn Tyrosine Kinase activity using a comprehensive in-silico approach, aiming to identify viable candidates for TNBC therapy.

2. Materials and Methods

2.1. Software, Hardware, and Databases

AutoDock Vina version 1.5.6 [10], UCSF Chimera [11], ChemDraw ultra.12 (https://chemistrydocs.com/chemdraw-ultra-12-0/ (accessed on 15 June 2025)), Discovery Studio (https://discover.3ds.com/discovery-studio-visualizer-download (accessed on 26 June 2025)), Spartan 04 (https://www.wavefun.com/products (accessed on 16 June 2025)), SwissAdme (https://www.swissadme.ch/ (accessed on 15 June 2025)), ProTox 3.0 (https://tox.charite.de/protox3/ (accessed on 16 June 2025)) and Windows (Intel processor, Corei7), Protein Data Bank (https://www.rcsb.org/ (accessed on 18 June 2025)).

2.2. Protein Crystal Structure Retrieval

The protein structure was retrieved from the Protein Data Bank (http://www.rcsb.org/ (accessed on 18 June 2025)) under the PDB ID; 2ZVA. It was determined using X-ray diffraction at a resolution of 2.60 Å, originating from the species Mus musculus. The co-crystallized ligand, Dasatinib, and the associated published data is accessible

2.3. Creation of Library

A library of five chalcone derivatives was designed using ChemDraw ultra. 12, where the 2D chemical structures and IUPAC names were generated. The corresponding SMILE’s were then exported into a separate file for further analysis.

2.4. Evaluation of Theoretical Oral Bioavailability and Toxicity

Using ProTox 3.0, the oral toxicity of the compounds were analyzed (CHCN1, CHCN19, CHCN48, CHCN333, and CHCN94), where the LD50 in mg/kg was assessed for all five compounds. To ensure that the designed compounds were druglike, further pharmacokinetic profiling was conducted using SwissADME where Lipinski’s rule of five was used as a metric to identify drug-like compounds. These included the number of hydrogen bond acceptors, the number of hydrogen bond donors, the MLogP value (partition coefficint), and the molecular weight.

2.5. Protein Structure Preparation

After the retrieval of high-resolution protein (2zva) from the Protein Data Bank (http://www.rcsb.org (accessed on 18 June 2025)), it was then exported into UCSF Chimera in the pdb file format, where protein preparation was conducted. All nonstandard amino acids, unbound water molecules, and ions were eliminated. Finally, Gasteiger charges were added. The final output file was transferred to Autodock Vs 1.5.6 for conversion to the corresponding pdbqt file format for molecular docking.

2.6. Ligand Structure Preparation

The newly designed compounds, having passed oral bioavailability and toxicity evaluation, were exported into Spartan 04, where energy minimization was performed and saved in the MOL2 format. This was further processed using UCSF Chimera for Gasteiger charges to be included. All other parameters were maintained at default. The corresponding files in pdb format were then converted in a similar way to pdbqt file formats using Autodock Vs 1.5.6.

2.7. Molecular Docking Analysis

To ensure the reliability of the docking protocol, the co-crystallized ligand (Dasatinib) was first removed from the enzyme’s crystal structure and then re-docked using the established docking parameters [12]. Successful validation was confirmed by the close alignment of the re-docked ligand with its original position in the active site, as observed in the crystal structure. The active site was defined using the coordinates from the crystallographic data, and an appropriate grid box was set around this region. The most favorable binding pose was selected for further analysis. Post-docking evaluation and visualization of ligand–protein interactions were conducted using Discovery Studio.

3. Results

3.1. Summary of the Designed Chalcone Library

Table 1. Represents a library of chalcone derivatives.

Compounds

represents a library of chalcone derivatives. A total of ten novel nitrogen-based chalcone analogs (CHCN1-CHCN94) were successfully designed in silico using ChemDraw for structural construction and visualization. The chalcone core was maintained while systematically modified to incorporate diverse nitrogen-containing groups.

3.2. Predicted Oral Bioavailability and Toxicity

The theoretical oral bioavailability of all compounds was assessed and presented in

Table 2. Drug-likeness and oral bioavailability analysis of test compounds.

Compound Name	CHCN1	CHCN19	CHCN48	CHCN333	CHCN94
Formula	C26H24Cl3NO2	C20H21Cl2NO3	C24H22ClNO2	C26H25Cl2NO3	C24H22ClNO3
Molecular weight	488.83 g/mol	394.29 g/mol	391.89 g/mol	470.39 g/mol	407.89 g/mol
Num. heavy atoms	32	26	28	32	29
Num. arom. heavy atoms	18	12	18	18	18
Fraction Csp3	0.19	0.255	0.12	0.19	0.12
Num. rotatable bonds	10	9	8	11	8
Num. H-bond acceptors	2	3	2	3	3
Num. H-bond donors	0	1	0	0	1
Molar Refractivity	136.60	108.18	116.23	132.67	118.25
TPSA	29.54 Å2	49.77 Å2	29.54 Å2	38.77 Å2	49.77 Å2
Log Po/w (MLOGP)	5.28	3.22	4.21	4.50	3.61
Inference	Yes	Yes	Yes	Yes	Yes
Lipinski’s Violation	1	0	0	0	0
Veber Violation	0	0	0	1	0
Ghose violations	3	0	0	2	0
Egan Violation	1	0	0	1	0
Muegge Violation	1	1	1	1	1
Bioavailability Score	0.55	0.55	0.55	0.55	0.55
Synthetic accessibility	3.38	2.94	2.90	3.45	3.01

. This includes the molecular weight, number of hydrogen bond acceptors, number of hydrogen bond donors, MlogP values, topological surface area, molar refractivity amongst others. These parameters are in congruence with Lipinski’s requirements for characterizing the drug-likeness property of drug molecules [13], where all compounds had their violations within an acceptable range. The Veber violation, Ghose violation, Egan violation, and Muegge violations, which also use the physicochemical properties of compounds to assess drug likeness, were included in this study.

ADMET Analysis

Further ADMET profiling results in

Table 3. Pharmacokinetics prediction output of the test compounds.

Compound Name	CHCN1	CHCN19	CHCN48	CHCN333	CHCN94
Silicos-IT LogSw	−10.23	−6.62	−8.29	−6.54	−7.70
Silicos-class	Insoluble	Poorly Soluble	Poorly soluble	Poorly soluble	Poorly soluble
Log Kp (cm/s)	−4.16 cm/s	−5.04 cm/s	−4.45 cm/s	−4.56 cm/s	−4.81 cm/s
GI Absorption	Low	High	High	High	High
BBB Permeant	No	Yes	Yes	No	Yes
Pgp substrate	Yes	No	Yes	Yes	Yes
CYP1A2 inhibitor	Yes	Yes	Yes	Yes	Yes
CYP2C19 inhibitor	No	Yes	Yes	No	Yes
CYP2C9 inhibitor	No	Yes	Yes	Yes	Yes
CYP2D6 inhibitor	No	Yes	Yes	Yes	Yes
CYP3A4 inhibitor	Yes	Yes	Yes	Yes	Yes

represent the pharmacokinetic profiling of the drug molecules. These include the Log Sw, implying oral solubility at a range of (−6.54 to −10.23), blood–brain permeation ability, substrate specificity, and possible metabolic pathway. Toxicity profiling in

Table 4. Toxicity profile of test compounds.

Properties	CHCN1	CHCN19	CHCN48	CHCN333	CHCN94
Oral acute toxicity	V	V	IV	IV	IV
Carcinogenicity	+	−	−	−	−
Hepatotoxicity	−	+	+	+	−
Androgen receptor binding	−	−	−	−	−
Thyroid receptor Binding	−	−	−	−	−
Estrogen receptor Binding	−	+	+	+	+
Aromatase binding	−	+	+	+	+

− = Inactive; + = active; Class IV = LD50 ≤ 2000 mg/kg; Class V = LD50 ≤ 5000 mg/kg.

, highlights the LD₅₀ values of the test compounds. It also highlights their mutagenic, carcinogenic, and hepatoxic tendencies as well.

3.3. Molecular Docking Results

The molecular docking results of the designed chalcone compounds with the target protein the target protein (2ZVA) are presented in Table 5, Table 6 and Table 7, where Table 5 presents the grid box coordinates, Table 6 shows the docking validation, and Table 7 provides the predicted affinities (docking scores) of the compounds.

3.3.1. Grid Point Generation

To define the docking search space, a grid box was centered on the active site of the enzyme, as identified by the position of the native ligand. The specific parameters for this grid box are provided in Table 5. All molecular docking was performed using AutoDock Vina, which generated the resulting protein-ligand complexes for further visualisation.

Table 5. Enzyme grid-box size center.

Enzyme	Grid-Box Size			Center
	X	Y	Z	X	Y	Z
Lyn-kinase	48	44	60	19.669	−9.388	23.998

Table 5. Enzyme grid-box size center.

Enzyme	Grid-Box Size			Center
	X	Y	Z	X	Y	Z
Lyn-kinase	48	44	60	19.669	−9.388	23.998

3.3.2. Validation of Docking Procedures

The docking protocol was validated by re-docking the native co-crystallized ligand into the protein’s active site. The resulting pose showed excellent alignment with the original crystal structure (Table 6), confirming the high accuracy of the method.

Table 6. Dock Validation of ligand against enzyme structure.

Enzyme	Crystal Structure Complex (Enzyme and Ligand)	Crystal Structure Complex with Re-Docked Ligand (Validation)
2zva

Table 6. Dock Validation of ligand against enzyme structure.

Enzyme	Crystal Structure Complex (Enzyme and Ligand)	Crystal Structure Complex with Re-Docked Ligand (Validation)
2zva

3.3.3. Binding Affinity of Ligands to Protease Enzymes

Table 7 showcases the binding energies of both the co-crystallized ligands and the five isolated compounds in their interactions with the enzymes.

Table 7. Docking scores of the designed chalcone derivatives against Lyn Tyrosine Kinase enzyme.

Compound Name	SMILE	Dock Scores (kcal/mol)
Lig 0	Dasatinib	−9.8
CHCN1	c1ccc(Cl)cc1c(c(cc2)OC)cc2C(=O)C=Cc3ccc(cc3)N(CCCl)CCCl	−8.6
CHCN19	COc(cc1)cc(O)c1C(=O)C=Cc2ccc(cc2)N(CCCl)CCCl	−7.1
CHCN48	COc(cc1)ccc1C(=O)C=Cc2ccc(cc2)N(CCl)Cc3ccccc3	−8.3
CHCN333	ClCCN(CCCl)c(cc1)ccc1C=CC(=O)c2cc(c(cc2)OC)Oc3ccccc3	−8.1
CHCN94	COc(cc1)c(O)cc1C(=O)C=Cc2ccc(cc2)N(CCl)Cc3ccccc3	−8.0

Lig 0 is the co-crystallized ligand (Dasatinib).

Table 7. Docking scores of the designed chalcone derivatives against Lyn Tyrosine Kinase enzyme.

Compound Name	SMILE	Dock Scores (kcal/mol)
Lig 0	Dasatinib	−9.8
CHCN1	c1ccc(Cl)cc1c(c(cc2)OC)cc2C(=O)C=Cc3ccc(cc3)N(CCCl)CCCl	−8.6
CHCN19	COc(cc1)cc(O)c1C(=O)C=Cc2ccc(cc2)N(CCCl)CCCl	−7.1
CHCN48	COc(cc1)ccc1C(=O)C=Cc2ccc(cc2)N(CCl)Cc3ccccc3	−8.3
CHCN333	ClCCN(CCCl)c(cc1)ccc1C=CC(=O)c2cc(c(cc2)OC)Oc3ccccc3	−8.1
CHCN94	COc(cc1)c(O)cc1C(=O)C=Cc2ccc(cc2)N(CCl)Cc3ccccc3	−8.0

Lig 0 is the co-crystallized ligand (Dasatinib).

3.3.4. Binding Poses and Binding Interaction Analysis of Designed Compounds Against Lyn Tyrosine Kinase Enzyme

The binding conformation and interaction of the isolated compounds (CHCN1-CHCN94) with residues on the active site of Lyn Tyrosine Kinase were elucidated and presented in Figure 1, Figure 2, Figure 3, Figure 4 and Figure 5. The binding interaction mode and amino acid interaction were also presented in

Table 8. Molecular interactions of amino acid residues of compounds from chalcone derivatives with protein Lyn Tyrosine Kinase.

Compounds	Hydrogen Bond Interaction	Hydrophobic Interaction
CHCN1	THR319 MET322 GLU320 LEU253	ALA384 LYS251 LEU253 VAL261 ALA273 LEU374 LEU253
CHCN19	THR319 LYS275 LEU253 LEU253	LEU253 MET294 ILE317 ALA273 LYS275 LEU253
CHCN48	THR319 LYS275 LEU253	LEU253 MET294 ILE317 ALA273 LYS275 LEU253
CHCN333	MET322	LEU253 LEU253 VAL261 ALA273 LEU374 VAL303 ALA384
CHCN94		MET282 VAL261 VAL261 ALA273

.

4. Discussion

According to Lipinski’s rules for oral absorptivity [13], the analyzed compounds had molecular weights ranging from 391.89 to 488.83 g/mol. Across all compounds, the maximum number of hydrogen bond acceptors was 3, and the maximum number of hydrogen bond donors was 1. Although the partition coefficient (MlogP) of CHCN1 was 5.28, slightly exceeding the ideal threshold for oral bioavailability, all compounds generally remained within acceptable limits, indicating favorable oral absorption potential. Veber’s rule, which sets strict criteria for topological polar surface area (TPSA) and the number of rotatable bonds, was met by all compounds except CHCN 333, which exceeded the permissible number of rotatable bonds with a score of 11. TPSA values for the compounds ranged from 29.54 to 49.77 Ų, further buttressing their drug likeness. According to Ghose’s criteria, which consider molar refractivity and the total number of atoms, all compounds except CHCN1 and CHCN333 fell within the acceptable range for molar refractivity, and all compounds had a total atom count between 47 and 57. Furthermore, the compounds exhibited an oral bioavailability score of 0.55 and a synthetic accessibility score ranging from 3.45 to 3.94, supporting their possible suitability for oral formulation and ease of synthesis (Table 2). CHCN333 was identified as a non-P-glycoprotein substrate, suggesting good bioavailability and consequent interaction in the active site. All compounds were substrates of either CYP1A2 or CYP3A4, suggesting a possible excretion pathway and drug interaction. The Silicos-IT LogSw values, reflecting molar solubility, ranged from −6.54 to −10.23 mol/L. Solubility results showed that only CHCN1 is insoluble. This points out that the other compounds have some relative potential for oral solubility or some other suitable formulation strategy that can deliver the drug adequately. Toxicity profiling revealed a good oral toxicity profile [14], with all compounds classified between class IV and V with LD₅₀ values from ≤ 2000 mg/kg to LD50 ≤ 5000 mg/kg (Table 3). Only CHCN1 showed potential carcinogenicity. Compounds CHCN1 and CHCN94 were not hepatotoxic, and none of the compounds exhibited androgen or thyroid receptor binding. Additionally, only CHCN1 lacked estrogen and aromatase receptor binding activity. This reflects the relative safety of the drug molecules. Molecular docking results were comparable to those of the native ligand dasatinib (−9.8 kcal/mol). Dock scores of the compound library were −9.8, −8. 6, −7.1, −8.3, −8.1, and −8.0 kcal/mol, respectively. CHCN1 had the best dock score of (−8. 6 kcal/mol), suggesting it can be taken further for its potential in chemotherapy. The DFG motif (Asp385, Phe386, and Gly387) acts like a switch that helps turn the kinase on or off.

When the kinase is active, Lys275 and Glu290 form a bond (a salt bridge) that holds the protein in the right shape so it can grab and use ATP for its activity [15]. CHCN1 reflected the Vanderwal interaction with Asp385, Phe386, Glu290, and Pi-alkyl interaction with Lys275. CHCN19 exhibited the Vanderwal interaction with Lys275 and Asp385. CHCN48 interacted via pi-alkyl with Lys275, and the Vanderwal interaction was observed with Glu290. In CHCN333, the Vanderwal interaction was seen with Lys275, Asp385, Glu290, and Phe386 (Table 8). Lastly, the Vanderwal interaction was recorded between CHCN94 and Asp385. CHCN48 had a docking score of −8.3 kcal/mol, which is comparably better than the others. This could be a result of the hydrogen and hydrophobic interactions formed with Lys275, which are absent in others. These results highlight the potential of these compounds in inhibiting Lyn Tyrosine Kinase to modulate cancer proliferation.

Although mustard-type chalcones contain electrophilic centers that could, in principle, undergo covalent interactions with nucleophilic residues or nucleic acids, the present docking study evaluates only non-covalent interactions with Lyn kinase. This approach was chosen to predict initial binding conformations and affinity trends. The reactive centers of the ligands were not positioned close to any nucleophilic residues within the binding pocket, suggesting that reversible binding may still occur prior to any potential covalent modification. Nonetheless, further covalent docking and experimental validation are required to confirm the nature of the binding. Although mustard-type chalcones contain electrophilic centers that could, in principle, undergo covalent interactions with nucleophilic residues or nucleic acids, the present docking study evaluates only non-covalent interactions with Lyn kinase. This approach was chosen to predict initial binding conformations and affinity trends. The reactive centers of the ligands were not positioned close to any nucleophilic residues within the binding pocket, suggesting that reversible binding may still occur prior to any potential covalent modification. Nonetheless, further covalent docking and experimental validation are required to confirm the nature of the binding.

5. Conclusions

The chalcone derivatives showed favorable ADME profiles (MW 391.89–488.83 g/mol, TPSA 29.54–49.77 Ų, bioavailability score 0.55) with low predicted toxicity (LD50 ≤ 2000–5000 mg/kg). Docking scores ranged from −7.1 to −9.8 kcal/mol, with CHCN1 as the most promising candidate. Key interactions were observed with Lys275, Glu290, Asp385, and Phe386, presenting better insight into their potential as Lyn Tyrosine Inhibitors in TNBC.