Developing multitarget coumarin based anti-breast cancer agents: synthesis and molecular modeling study

A new series of 7-substituted coumarin scaffolds containing a methyl ester moiety at the C4-position were synthesized and tested for their in vitro anti-proliferative activity against MCF-7 and MDA-MB-231 breast cancer cell lines using Doxorubicin (DOX) as reference. Compounds 2 and 8 showed noticeable selectivity against MCF-7 with IC50 = 6.0 and 5.8 µM, respectively compared to DOX with IC50 = 5.6 µM. Compounds 10, 12, and 14 exhibited considerable selectivity against Estrogen Negative cells with IC50 = 2.3, 3.5, and 1.9 µM, respectively) compared to DOX with (IC50 = 7.3 µM). The most promising compounds were tested as epidermal growth factor receptor and aromatase (ARO) enzymes inhibitors using erlotinib and exemestane (EXM) as standards, respectively. Results proved that compound 8 elicited the highest inhibitory activity (94.73% of the potency of EXM), while compounds 10 and 12 displayed 97.67% and 81.92% of the potency of Erlotinib, respectively. Further investigation showed that the promising candidates 8, 10, and 12 caused cell cycle arrest at G0–G1 and S phases and induced apoptosis. The mechanistic pathway was confirmed by elevating caspases-9 and Bax/Bcl-2 ratio. A set of in silico methods was also performed including docking, bioavailability ADMET screening and QSAR study

Based on the above findings, we worked on synthesizing 7-substituted coumarin derivatives with flexible C 4 -position ester functionality to examine their impact on two BC cell lines, MCF-7 and MDA-MB-231 cell lines, as well as normal breast cells MCF-10A.The synthetic compounds were tested in vitro for their ability to inhibit the EGFR and ARO enzymes.
The most potent compounds, such as cell cycle examination and apoptosis markers, were selected for additional studies to discover the molecular process behind their anticancer activity.
Moreover, molecular modeling studies were conducted to identify the two enzymes' interaction patterns and investigate the relationship between their physicochemical features and inhibitory action.Additionally, QSAR analysis was performed.

Results
Chemistry.The steps in Figs. 2 and 3 were used to synthesize the intermediates and target molecules.The requisite 4-chloromethyl-7-hydroxycoumarin (1) 31 was first synthesized through a Pechman cyclocondensation reaction involving acid-catalyzed condensation of resorcinol with ethyl 4-chloroacetoacetate in a readily scalable procedure 32,33 .
In vitro biological evaluation.Measurement of in vitro cytotoxic activity by MTT assay.For the MCF-7 cell line, results shown in Table 1 revealed that the most powerful compounds were methyl benzoate derivatives 2 and 8 with IC 50 = 6.0 µM and 5.8 µM, respectively, compared to IC 50 = 5.6 µM for the standard drug DOX.

Structure-activity relationship (SAR) study.
Starting compound 1 was devoid of activity against both types of cell lines.
The introduction of benzoate substituent at position-4 in the coumarin scaffold was observed to potentiate the activity.This is possibly explained by specific hydrophobic interactions of the phenyl group.
Incorporation of the electron-withdrawing group at position-4 on the aromatic moiety in compound 3 showed moderate selectivity towards MCF-7 than MDA-MB-231.Conversely, the substitution of an electron-donating Figure 3. Synthesis of new acetyl coumarin derivatives (13-15).methoxy group at position-2 as in compound 4 resulted in a beneficial selectivity against MDA-MB-231 than MCF-7 cell lines was IC 50 values = 6.4 and 17.2 µM, respectively.
The activity of 7-(2-oxo-2-phenylethoxy)-4-methyl benzoate coumarin series (7-12) varied according to the type of substituents compared to intermediates 2-4.Incorporation of 2-oxo-2-phenylethoxy at position-7 of coumarin moiety in compounds 7, 9, and 11 diminished the activity against the MCF-7 cell line while resulting in poor anti-proliferative activity against the MDA-MB-231 cell line.Compound 9 bearing halogen atom at p-position was identified as the most significant cytotoxic candidate among the series revealing potent selectivity against the MDA-MB-231 cell line with IC 50 value of 10.4 µM.
In general, mono substitution with an electron-withdrawing group on phenyl moiety such as p-chloro atom located at C 7 -of coumarin ring in compounds 8, 10, 12 and 14 dramatically improved the activity against both cell lines, in comparison to the unsubstituted derivatives.In this context, it could be stressed that the target compound 8 showed high selectivity against MCF-7 than MDA-MB-231 (IC 50 = 5.8 µM versus IC 50 = 27.5 µM), while compounds 10, 12 and 14 showed superior selectivity against MDA-MB-231 with IC 50 = 2.3, 3.5 and 1.9 µM, respectively.Fortunately, compound 12 was the most potent among this series against both MCF-7 and MDA-MB-231 with IC 50 = 13.2 µM and 3.5 µM, respectively.Surprisingly, compound 6 with an electron-withdrawing chlorine atom exhibited a lower activity level against both cell lines, which was assumed to involve H-bonds with amino acids in the ligand-binding domain.
In vitro cytotoxicity against nonmalignant human cells.The compounds were also examined against the human nonmalignant cells MCF10A (breast) to investigate their selectivity between nonmalignant cells and cancerous cells.Compared to DOX (IC 50 = 12.4 µM), the most active compounds 8, 10, 12 and 14 were less toxic to MCF10A cells (IC 50 = 79.9,76.8, 72.1 and 263.1 µM, respectively) with good SI values (more than 10) (Table 1).
In vitro enzyme inhibition assays.The most active compounds were subjected further to in vitro enzyme inhibition via ELISA enzyme assay.www.nature.com/scientificreports/ • EGFR enzyme inhibition assay MDA-MB-231 cell lines could express mainly EGFR enzyme, whereas MCF-7 cell lines only produced aromatase enzyme 23 .So, an enzyme assay study was performed for ARO and EGFR enzymes in MCF-7 and MDA-MB-231 cell lines, respectively.
It was observed that compounds 10 and 12 showed the highest inhibition effect towards EGFR, displaying 97.67% and 81.92% of the potency of erlotinib, respectively (Table 2, Supplementary Fig. S1).Therefore, their potent selectivity against MDA-MB-231 might be due to EFGR inhibition pathway.

• (ARO) enzyme inhibition assay
To convert androgens into estrogens, the MCF-7 cell line was shown to possess enough aromatase enzyme activity 37 .Results are presented in Table 3, Supplementary Fig. S2, and revealed that target compound 8 is considered the most potent AIs with (IC 50 value = 0.114 µM), compared to EXM (IC 50 value = 0.108 µM), and this described the elevated value of IC 50 in MCF-7 (5.8 µM).Furthermore, target compound 2 exhibited half ARO inhibitory activity as that of reference EXM. 4, Supplementary Fig. S3.Regarding the MCF-7 cell line, the percentage of apoptotic cells was increased significantly in the pre-G1 phase by 18.67 folds with significant arrest in G0-G1 and S phases by 1.508 and 1.33 folds, respectively compared to the untreated control cells for compound 8.For MDA-MB-231 cell lines, when treated with compounds 10, 12 and 14, the percentage of apoptotic cells in the pre-G1 phase increased by 16.22, 22.46 and 19.9 folds with consequential cell cycle arrest at the S phase by 1.59, 1.08 and 1.38 folds, respectively.Results indicated that compound 8 tested against the MCF-7 cell line and compounds 10, 12 and 14 tested against the MDA-MB-231 cell line showed pre-G1 apoptosis and cell cycle arrest at the S phase in addition to G0-G1 by compound 8. Effect on the levels of Bax/Bcl-2.Apoptosis is provoked in a cell through two major apoptotic pathways, the extrinsic and intrinsic pathways.By overexpressing anti-apoptotic proteins like Bcl-2 or downregulating pro-apoptotic proteins like Bax through the intrinsic apoptotic pathway, cancer cells can develop apoptotic resistance 39 .In this study, we measured the levels of Bax and Bcl2 to evaluate the effects of compounds 8, 10, 12 and 14, which showed promising apoptosis-inducing activity on the intrinsic apoptotic pathway.As shown in (Supplementary Table S6 and Supplementary Fig. S7), the tested compounds 8, 10, 12 and 14 significantly increased the expression of the pro-apoptotic protein Bax by 5.27, 5.27, 5.22 and 4.9 folds, respectively, with a consequent decline in the anti-apoptotic protein Bcl-2 levels by 0.44, 0.377, 0.271 and 0.329 folds, respectively compared to the control.As a result, a significant increase in the Bax/Bcl-2 ratio was observed, supporting the claim that the tested compounds can improve the therapeutic response in breast cancer.

Cell cycle arrest. Results are presented in Table
Caspase-9 activation.Caspase-9, activated in apoptotic cells by intrinsic pathways, can be used as a marker for the induction of apoptosis by anticancer agents. 40.In the present study, the levels of caspase-9 were then measured in the compounds that showed a noticeable increase in the Bax/Bcl-2 ratio to identify a possible pathway for the antiproliferative activity of the most promising compounds.Significantly elevated levels of caspase-9  S8 and Supplementary Fig. S9) by 6.71 folds for compound 8, while compounds 10, 12 and 14 resulted in 6.42-, 7.33-and 4.37-folds elevation, respectively, compared to control cells.These results suggested that all tested compounds could induce a significant overexpression in caspase-9 protein level higher than that elicited by control, assuming that apoptosis was probably due to the activation of caspase-9 protein.

Statistical analysis and data interpretation.
Data analysis was performed by SPSS software, version 25 (SPSS Inc., PASW Statistics for Windows version 25.Chicago: SPSS Inc.).Apoptosis was described using numbers and percentages.Quantitative data were described using mean ± Standard deviation for normally distributed data after testing normality using Shapiro Wilk test.The significance of the obtained results was judged at the (≤ 0.05) level.Chi-Square and Z-tests were used to compare qualitative data between groups as appropriate.One Way ANOVA test was used to compare more than 2 independent groups with Post Hoc Tukey test to detect pair-wise comparison.
In silico computational study.Docking into EGFR crystal structure.Re-docking results of the co-crystallized ligand (Erlotinib) and its binding mode 18,30,41 were described in (Table 6, Supplementary Figs.S10, and  S11).Docking results showed that compounds 10 and 12 exhibited good binding interactions and docking scores similar to Erlotinib (Table 6).It is suggested that the planarity of quinazoline moiety in erlotinib seems important for stabilizing the EGFR binding site.Similarly, the planar nature of the coumarin scaffold in compounds 10 and 12 allowed them to be embedded into the active site of EGFR, as shown in their 2D interactions (Supplementary Fig. S12) and 3D binding modes as in Fig. 5.It was found that the C=O group of the coumarin skeleton overlapped with the N 1 atom of the quinazoline moiety of erlotinib, which equivalently served to anchor with Met 769, which was considered the key interaction improving the stabilization.By investigating the binding site of the receptor, the presence of o-methoxy or p-chloro phenyl moiety at position 4-methyl carboxylate of coumarin nucleus in compounds 10 and 12, respectively, were observed to be positioned in the hydrophobic pocket, which included side chains Leu 764, Leu 753, Met 742, Ile 720 and Ile 765 as for the ethynyl-phenyl group in erlotinib 42 (Supplementary Fig. S13).
Moreover, additional halogen bond interactions were shown with the amino acid His 781 with the chloride atom on C 4 of the phenyloxy side chain at C 7 of the coumarin nucleus with bond length (2.22 Å for compound 10 and 2.32 Å for compound 12).This interaction appears to help their immobilization; in the same way, the ether side chain in erlotinib as illustrated in (Fig. 4).
In order to rationalize the biological outcomes, it was believed that the poor EGFR inhibitory action of compounds 11 and 14 was related to their failure to bind with the essential amino acid and their low docking score of -5.7394 and -5.4982 kcal/ mol, respectively (Supplementary Fig. S14).Hence, the acquired data revealed that the binding modes of the investigated compounds were consistent with their EGFR inhibitory action.
On the other hand, despite compound 8 showing the same binding features as the most potent EGFR inhibitors 10 and 12 into the active site with a docking score of -6.7717 kcal/ mol and RMSD = 1.9314Å, it unexpectedly possessed weak EGFR inhibitory activity.However, it could be stated that other factors that affect activity related to lipophilicity and solubility may be involved.Docking into ARO crystal structure.Results of re-docked ligands (ASD and EXM) are summarized in (Table 6) and their binding modes are described in (Supplementary Fig. S15).Aromatase enzymes comprise a heme prosthetic group containing iron at the active reaction center 43 .
Regarding compound 8, the most potent AI showed an acceptable binding mode (Table 6, Fig. 5).When compared to the binding mode of ligands, it was found that the coumarin scaffold was stabilized by arene-arene interaction with the 5-membered ring of the heme moiety, which is considered the binding driving interaction.
Unsubstituted benzoate at position-4 of coumarin is essential for activity, binding with Met 374 by H-bond (2.84 Å) and providing a convenient allocation in the hydrophobic pocket, constituted of Val 373, Met 374, Leu 372, Ile 398 residues.
Introducing a substituent on the phenyl moiety (p-chloro or o-methoxy) resulted in a dramatic decrease in activity due to either a steric effect or a change in lipophilicity.In particular, the distance between the benzene ring placed in the hydrophobic pocket and the coumarin moiety is likely the same as the entire length of ASD (6.97 Å, 6.69 Å, respectively) and plays a significant role in the control of affinity.The substitution at position-4 on the methylene bridge could be responsible for the activity.
In addition, the phenyl ethoxy side chain at position-7 slightly twisted to be embedded into the binding site and seemed to induce additional weak H-bond with Thr 310 (3.64 Å) to close the entrance cavity between Asp 309 and Ser 478 20 .The longer the substituent at position-7 could justify the better affinity to the binding site than the free OH moiety.These three binding site interactions were fulfilled when the coumarin ring carried benzyl carboxylate methyl group at position-4 and p-chlorophenyl ethoxy substituent at position-7.Therefore, the better binding mode of target compound 8 when compared to compound 2 can be explained, in which it showed the same binding manner with the backbone amide group of Met 374 (3.42 Å) and the heme moiety through arene-H interaction.Although, the free hydroxyl group at position-7 exhibited a decrease in activity due to the inability to close the entrance side of the pocket (Fig. 5).The 2D binding poses of both compounds are shown in (Supplementary Fig. S16).
On the other hand, because of the correlation between the bulkiness of the position-4 substituent and the activity, it may be presumed that the o-methoxy substituent in compound 11 (IC 50 = 5.98 µM) generated steric hindrance in the rear side of the pocket and unstable binding mode with score (S) = − 0.1435 kcal/mol.In addition, compound 6 (IC 50 = 3.93 µM) did not illustrate the three-site binding hypothesis since its molecular volume is smaller than the pocket site.They both shared the lack of essential interaction in either the rear side or the entrance side and unacceptable values of RMSD of more than 2 (Supplementary Fig. S17).
Bioavailability and ADMET properties screening.To understand the pharmacokinetic behavior of the most active compounds 8, 10, 12 and 14, their physicochemical characters were calculated with the aid of the online application ADMETSAR server http:// lmmd.ecust.edu.cn/ admet sar1 by applying various qualitative ADMET models as described in (Table 7).The results revealed that all compounds possessed good intestinal absorption and could penetrate the blood-brain barrier.Besides, the candidate compounds illustrated good oral bioavailability, making them promising candidates as anti-BC agents.www.nature.com/scientificreports/Vsurf descriptors explained the hydrophilic and hydrophobic interactions, while the hydrophobic bindings were described by the vsurf_D descriptors (vsurf_D5 and vsurf_D3) 44,45 .
The reliability of the built model was confirmed by excellent linearity [$PRED = 0.2094 (pIC 50 ) + 4.215], with R 2 = 0.326 (Fig. 6) and the relative values between the predicted and experimental activities (Table 8).It was found that the predicted values are close to those experimentally investigated, indicating that the QSAR model is reliable and can be safely applied to predict more effective compounds.

Discussion
In the current study, a new series of nonsteroidal 4,7-disubstituted coumarin (1,2-benzopyrone) derivatives bearing methyl ester moiety at the C 4 -position were designed as anti-BC agents.Easy and good-yielding synthetic pathways were adopted to prepare the designed compounds.MTT assay method was used to evaluate their in vitro anti-BC activity against MCF-7, MDA-MB-231, and MCF-10A cell lines using Doxorubicin (DOX) as a reference drug.Generally, most of the tested compounds revealed MDA-MB-231 selectivity more than MCF-7.Particularly, compounds 2 and 8 exhibited obvious selectivity against MCF-7, while compounds 4, 5, 10, 12, and 14 exhibited considerable selectivity against the MDA-MB-231 cell line.
In conclusion, it might be emphasized that "mono substitution with electron-withdrawing groups as chlorine on phenyl moiety is crucial for the anti-BC activity against both cell lines as it could be involved in electrostatic interactions.
The cytotoxic activity of compounds 10 and 12 against MDA-MB-231 could be due to its potent EGFR inhibitory activity, while compound 8 against MCF-7 might be attributed to its ARO inhibitory activity.Regarding compound 14, it possessed weak inhibitory activity toward EGFR despite its potent anticancer activity in MDA-MB-231 cell lines.Consequently, further investigations for the most active compounds 8, 10, 12 and 14 were  www.nature.com/scientificreports/done to discover more mechanistic pathways for their antiproliferative activity, such as cell cycle analysis and apoptosis assay using flow cytometry.It was found that they were able to induce cell cycle arrest at G0-G1 and S phases.Their apoptotic mechanism was proved by a significant increase in both Bax/Bcl-2 ratio and caspase-9.
The inhibitory activities of compounds 2 and 8 on ARO enzymes and 10 and 12 on EGFR were confirmed by docking studies, highlighting significant binding interactions.ADMET study was also carried out to further describe the synthesized drugs' pharmacokinetics.The findings of the QSAR model revealed that there should be a balance between the compounds' hydrophilic and hydrophobic substituents, which was confirmed by the docking results.Therefore, the observed results showed that these compounds were susceptible to being modified further to serve as promising multitargeted anti-BC agents (Supplementary Fig. S18).

Methods
Chemistry.Melting points were determined on Stuart melting point apparatus and are uncorrected.Microanalyses were performed at Cairo University and performed on a Perkin-Elmer 240 elemental analyzer for C, H, and N elements, and the results were within the acceptable range of the theoretical values. 1 H and 13 C NMR were performed at Mansoura University and recorded on Brucker 400 MHz spectrometer and 100 MHz spectrometer, respectively.Chemical shifts are expressed in δ ppm with reference to TMS.Using a Nicolet iS10 infrared spectrometer, IR spectra were recorded at Mansoura University.Mass spectral analyses were performed on Thermo SCIENTIFIC DCQII at Azhar University.All the used chemicals and reagents were purchased from Aldrich Chemicals Co, USA.
Pathway (C).To a solution of 1 (0.32 g, 1.5 mmol) in DMF (10 mL), the appropriate sodium salt of (un)substituted benzoic acid (1.5 mmol), anhydrous K 2 CO 3 (0.29 g, 3 mmol) and the appropriate phenacyl bromide (1.5 mmol) were added.The reaction mixture was kept under stirring at room temperature for about 4 h, heated at 90 °C overnight, then cooled and poured over ice-water.The separated solid products were found to be an impure mixture of different products with very low yields.

Biological tests.
In vitro cytotoxic activity using MTT assay.In this study, all synthesized compounds were initially screened for their in vitro antitumor activities against two human BC cell lines, namely, MCF-7 (estrogen positive) and MDA-MB-231(triple negative), to select the most promising ones for further screenings via MTT assay using Doxorubicin, as a reference drug as described in the reported method 46,47 .
In vitro (EGFR) enzyme inhibition assay.Fourteen compounds that showed good cytotoxic activity against the TNBC cell line (MDA-MB-231) were examined for in vitro EGFR inhibition assay relative to erlotinib as a reference drug according to the reported protocol 48 .
In vitro (ARO) enzyme inhibition assay.Eleven compounds that showed good cytotoxic activity against the ER + cell line (MCF-7) were examined for in vitro ARO inhibition assay relative to EXM as a reference drug according to the reported method 49 .
Cell cycle arrest.DNA flow cytometric cell cycle analysis was performed on treated MCF-7 cells with IC 50 concentration of compound 8 and treated MDA-MB-231 cells with IC 50 concentration of compounds 10, 12 and 14 according to the reported procedure 50 .
Apoptosis assay.Annexin V-FITC/PI dual staining assay was performed according to the reported method 51 to further study the ability of compounds 8, 10, 12 and 14 to induce apoptosis by introducing flow cytometrybased analysis.
Measurement of the level of Bax Bcl-2 and Caspase-9.MCF-7 cells were treated with compound 8, while MDA-MB-231 cells were exposed to compounds 10, 12 and 14 according to the manufacturer's instructions 52 .
Molecular docking study with both EGFR and human ARO enzymes.Crystal structures of EGFR (PDB ID: 1M17) 18 and ARO (PDB ID: 3EQM) 53 were obtained from the PDB.Regarding enzyme inhibition assay, the most potent compounds were selected for docking study using Molecular Operating Environment (MOE) software.

Figure 1 .
Figure 1.(a) Main structural features of EXM (Ia) and Erlotinib (Ib).(b) Design chart with reported coumarin derivatives as a potent anti-breast cancer agent.

Figure 4 .
Figure 4. (a) Binding mode of compound 10 (blue) and erlotinib (green), (b) Binding mode of compound 12 (blue) and erlotinib (green), (c) Binding mode of compound 10 (red), compound 12 (green) and erlotinib (blue) occupying the active site.The 2 compounds had the same binding mode as the native ligand.They bound with the key residue Met 769.The presence of p-chloro atom improved their stability.The hydrophobic side chain in both compounds changed their orientation to embed in hydrophobic pocket.The coumarin moiety was buried in the pocket in parallel to quinazoline nucleus in erlotinib.

Figure 6 .
Figure 6.Experimental versus predicted pIC 50 of the tested compounds against MCF-7 human cancer cell line.

Table 3 .
IC 50 and percentage above IC 50 of exemestane towards ARO.Significant values are in bold.

Table 4 .
Results of cell cycle analysis in MCF-7 and MDA-MB-231 cell lines expressed as % of cells in each phase for the selected compounds (8 in MCF-7 and 10, 12 and 14 in MDA-MB-231).Similar superscripted small letters in same column denote significant difference between groups by Z test.Similar superscripted capital letters in same row denote significant difference between different stages by Z test.

Table 5 .
Results of annexin V-FTIC/PI dual staining assay expressed as % Apoptosis and % Necrosis induced by compound 8 in MCF-7 cells and compounds 10, 12 and 14 in MDA-MB-231 cells.Similar superscripted small letters in same column denote significant difference between groups by Z test.Similar superscripted capital letters in same row denote significant difference between different stages by Z test.Vol:.(1234567890)Scientific Reports | (2023) 13:13370 | https://doi.org/10.1038/s41598-023-40232-3www.nature.com/scientificreports/protein level in treated samples were observed in (Supplementary Table

Table 6 .
RMSD values and docking binding scores of the docked compounds.

Table 7 .
Some ADMET properties of the target compounds using admet SAR server.

Table 8 .
The experimental, predicted pIC 50 of the synthesized compounds and the residual.