Scaffold transforming and in-silico design of potential androgen receptor antagonists in prostate cancer therapy

Abstract

Androgens like testosterone and dihydrotestosterone are essential for the growth and development of the prostate gland. Androgenic receptors are overexpressed, which promotes the progression of prostate cancer; therefore, androgenic receptors are a key target in the therapy of prostate cancer. Enzalutamide is used to treat prostate cancer; however, it also causes toxicities such as cardiovascular toxicity, acute myocarditis, hypertension, and seizures. The objective of this research was to create novel and safer analogues of enzalutamide, followed by the prediction of the pharmacokinetic and toxicity characteristics of these enzalutamide analogues. Molecular docking studies of analogues were also done to guess how ligands will work biologically in treating prostate cancer. A total of 195 analogues were generated, and among them, 23 bioisosteres were selected for further pharmacokinetic, toxicological screening and docking studies. The predicted physical-chemical, medicinal, and ADMET characteristics of the designed bioisosteres were optimal to good compared to enzalutamide. Additionally, the drug likeness and drug score of analogues were superior to enzalutamide. According to docking studies of analogues, EZ12, EZ8, and EZ10 formed hydrogen bonds of SER778 with replaceable amide groups in enzalutamide molecules. SER778 residue may be responsible for antagonistic activity towards androgen receptors. Based on the results of the ADMET, drug likeness, drug score, and docking study of designed enzalutamide analogues, the ligands EZ12, EZ8, and EZ10 could be used to find more possible antiandrogen drugs that could be used to treat prostate cancer.

Introduction

Cancer is the third most common cause of death in India and a significant global public health issue. The most recent predictions put the number of cancer cases in India at 1.46 million in 2022, which is a 5% rise from the 1.39 million cases projected for 2020. In 2022, the estimated cases of prostate cancer (PC) in patients were found to be 43,691 in men, which increased to 47,068 in 2025. Furthermore, PC was shown to be the second most common malignancy in men over 65 years, with numbers and proportions of 33,695 and 12.3, respectively (Sathishkumar et al. 2022; Kulothungan et al. 2022). Androgens such as testosterone and dihydrotestosterone (a metabolite of testosterone) are necessary for the normal growth and development of the prostate gland; nevertheless, an elevated quantity of these androgens may lead to the proliferation of PC cells (Mohler et al. 2011; Zhang et al. 2016; Stabile and Dicks 2003). The majority of deaths from PC are due to metastatic disease, identified either at diagnosis or after relapse following local therapies (Helgstrand et al. 2017). The terminal stage of advanced PC known as castration-resistant prostate cancer (CRPC) is the outcome of the tumor’s adaptation to an environment with low levels of testosterone. After a response to androgen deprivation treatment (ADT) that takes between 3 and 8 years, patients often begin to show signs of metastatic CRPC (Harris et al. 2009).

Enzalutamide (ENZ) is chemically 4-(3-(4-cyano-3-(trifluoromethyl)phenyl)-5,5-dimethyl-4-oxo-2-thioxoimidazolidin-1-yl)-2-fluoro-N-methyl-benzamide which is second generation antiandrogen receptor antagonist that reduces intertumoral androgen synthesis shown in Fig. 1 (Tran et al. 2009; Ji et al. 2020). On the other hands, it inhibited the androgen receptor (AR) signaling that prevent the AR translocation and bind to DNA, result in cellular apoptosis (Koivisto et al. 2020). Prolonged ADT exposure in the aging population is characterised by cardiovascular toxicity including acute myocarditis with most several evidences (Kulkarni et al. 2016; Iacovelli et al. 2017; Reyes et al. 2021; Cone et al. 2021). The group of patients treated with ENZ cause seizures and other hazardous consequences (Davis et al. 2019). Because of this, the bioisosteric approach may be used to logically modify a lead molecule into a more desirable therapeutic drug with increased potency, selectivity, changed physicochemical, metabolic, and toxicological properties (Dick and Cocklin 2020). Structure-based drug design (SBDD) is a broadly used techniques in contemporary drug discovery. Due to the abundance of structural data available, variety of SBDD techniques may be used including virtual screening, molecular docking, and in-silico ADMET property prediction. Molecular docking is the most widely used SBDD technique for predicting binding poses and estimating binding affinities in small-molecule drug discovery (Tanchuk et al. 2016).

Fig. 1.

Structure of enzalutamide and modified amide group; Diagrammatic representation of the best three analogues of enzalutamide with specific bioisosteric replacement as androgen receptor antagonist for prostate cancer treatment

In this work, we emphasise the use of computational tools to bioisosterically modification of amide group in ENZ molecule in order to generate novel and less toxic ENZ analogues for the treatment of PC.

Materials and methods

Scaffold transformation through Bioisosteric approach of Enzalutamide

ENZ is used as second generation antiandrogen drug in the management of PC but during the tratment, patients experienced several toxicities such as hepatotoxicity, hypertension, seizures and acute myocarditis. Threfore, it is essential to modify the ENZ structure in order to reduce toxic effects caused by it. MolOpt, online software used for generation of bioisosteres which is based on deep generative models, data mining, and similarity comparisons as bioisosteric transformation rules. A useful feature of MolOpt is that it can navigate historical bioisosteric group space and identify new bioisosteric transformation ideas (Shan and Ji 2020). Total 195 replacible groups for amide group in ENZ molecule were generated by MolOpt. Newely generated ENZ bioisosteres were used for further screening such as ADMET, DL, DS prediction and their docking studies is displayed in Fig. 1.

Pharmacokinetic and toxicological (ADMET) studies

Prediction of ADMET properties of newly designed ENZ bioisosteres were computed using ADMETlab 2.0. It is an integrated online platform having eighty-four quantitative and four qualitative regression models with authentic and extensive predictions of ADMET properties for novel ligands that mimic mammalian ADMET properties (Xiong et al. 2021; Dong et al. 2018; Wang et al. 2016; Lei et al. 2016).

Molecular docking study

Molecular docking approach used extensively in modern drug designing and development. it is widely used for study of targets (macromolecules such as DNA, RNA and Proteins) and small molecule (ligand) interactions. Docking is then used to predict the bound conformation and binding free energy of small molecules to the target (Feher and Williams 2012; Forli et al. 2016). A molecular docking study of designed ENZ bioisosteres was carried out using the crystal structure of the androgen receptor and this study involved a number of steps like preparation of the protein, preparation of the ligands and protein-ligand interactions study using AutoDock Vina (ADV) 1.5.6 (https://vina.scripps.edu) (Trott and Olson 2010).

Protein preparation

A protein’s three-dimensional structure was taken from protein data bank (PDB) database (https://www.rcsb.org/). The human androgen receptor T877A mutant ligand-binding domain with cyproterone acetate crystal structure (PDB ID: 2OZ7, Resolution = 1.80 Å) served as the model for this work (Bohl et al. 2007). An automated protein-ligand docking package is called AutoDock software. AutoDock Vina (ADV) creates crystal-clear 3D protein pictures. To prepare the protein, ADV was utilised. The improved protein was first prepared for docking by removing water molecules, adding hydrogen atoms, adding Kollman charges followed by the repair of missing atoms and then saved in .PDBQT format.

Ligand preparation

The 2D chemical structure of ligands were drawn using ChemDraw. The 2D structure was converted into 3D with the help of chem3D for all ligands and saved in .PDB format. Ligands were dropped into ADV and saved in PDBQT format. In continuation, protein was dropped into ADV for grid box formation to keep the ligand in the centre.

Protein-ligand interactions study

The docking simulations between the above-mentioned ligands (Table 1) and protein were processed using the ADV program. Then, a grid box of 40 × 40 × 40 points was constructed targeting the entire active site of protein (LEU701, LEU704, ASN705, LEU707, GLN711, MET742, ARG752, PHE764, SER778, MET780, MET787, PHE876, ALA877, LEU880, PHE891), where the grid center was at X = 26, Y = 5, Z = 5, and the dimensions of the grid box were size x = 30, size_y = 30, size_z = 30, with grid spacing of 1.0 Å and exhaustiveness = 8. Other docking parameters, including ADV were set to default, including the rate of gene mutation and the rate of crossover. The PyMol programme (Yuan et al. 2017) and Discovery studio (Kemmish et al. 2017) examined the output files in order to produce 2D and 3D protein-ligand interactions.

Table 1.

Structure and molecular properties of the analogues

MW; Molecular weight, nHA; Number of hydrogen bond acceptor, nHD; Number of hydrogen bond donor, nRot; Number of rotatable bonds, TPSA; Topological polar surface area, logP; The logarithm of partition coefficient value, logs; The logarithm of aqueous solubility value, STD; Standard

Drug likeness and drug score studies

The prediction of a drug likeness (DL) can be done in a variety of ways, some of which make use of topological descriptors, fingerprints of MDL structure keys, or other elements like cLogP and molecular weights. Drug score (DS) can evaluate a compound’s general potential for medication approval by combining its drug similarity, cLogP, logS, molecular weight, and toxicity issues into a single, helpful value. Osiris property explorer (PEO) was used to calculate the DS and DL of newely designed analogues (Rajan et al. 2019).

Results and discussion

Scaffold transformation through Bioisosteric approach

In the drug discovery technique, bioisosteric approach is generally used to improve pharmacokinetic properties and decreases unwanted toxicities. MolOpt generates 195 replacible groups for amide group in ENZ and their ADMET properties were predicted. Based on this, some selected designed ENZ bioisosteres were shown in Table 1. In-silico pharmacokinetic and toxicological investigations were conducted on the 195 analogues of N-methyl formamide groups in enzalutamide. The overall workflow of the study is shown in Fig. 2.

Fig. 2.

The overall workflow of the present work

Study of molecular properties

The prediction of molecular properties of the designed ENZ bioisosteres were calculated using the ADMETlab 2.0 is shown in Table 1. Lipinski rule of five for all analogues exhibit acccepted indicates the proper absorption and bioiavailability of the drug candidates (Karami et al. 2022). The result indicates that all analogues met the acceptance criteria same as ENZ that means designed analogues may be considered as drug candidates.

Study of medicinal properties

The medicinal properties of the designed ENZ bioisosteres were calculated with the help of ADMETlab 2.0 and their results are shown in Table 2. QED, a measure of drug-likeness based on the concept of desirability based on the eight drug-like related properties. The predicted QED score of designed analogues has been found in the range between 0.298 and 0.578. All analogues will be easy to synthesize as per synthetic accessibility criteria (< 6). The Lipinski rule for all analogues has been found acceptable, indicating good bioavailability. Pfizer rules for all analogues found under criteria, with the exception of EZ2, EZ5, EZ7, EZ11, EZ13, EZ17, and EZ20. GSK rule violations were found for all bioisisteres because they consist of two properties, MW and LogP, which should be ≤ 400 and ≤ 4, respectively.

Table 2.

Prediction of medicinal properties of the analogues

Entry No.	QED	Synth	Fsp3	MCE-18	Lipinski	Pfizer	GSK	GT
EZ1	0.504	3.377	0.318	79	Accepted	Rejected	Rejected	Accepted
EZ2	0.455	3.383	0.348	79	Accepted	Rejected	Rejected	Accepted
EZ3	0.521	2.831	0.273	53	Accepted	Accepted	Rejected	Accepted
EZ4	0.569	2.851	0.200	54	Accepted	Accepted	Rejected	Accepted
EZ5	0.457	3.538	0.286	89	Accepted	Rejected	Rejected	Rejected
EZ6	0.384	3.496	0.208	92	Accepted	Accepted	Rejected	Rejected
EZ7	0.578	2.873	0.250	52	Accepted	Rejected	Rejected	Accepted
EZ8	0.569	2.851	0.200	54	Accepted	Accepted	Rejected	Accepted
EZ9	0.505	3.616	0.273	96	Accepted	Accepted	Rejected	Accepted
EZ10	0.527	3.42	0.286	79	Accepted	Accepted	Rejected	Accepted
EZ11	0.509	3.597	0.348	96	Accepted	Rejected	Rejected	Accepted
EZ12	0.552	2.817	0.200	54	Accepted	Accepted	Rejected	Accepted
EZ13	0.537	3.132	0.273	64	Accepted	Rejected	Rejected	Accepted
EZ14	0.298	3.187	0.200	56	Accepted	Accepted	Rejected	Accepted
EZ15	0.396	3.066	0.304	68	Accepted	Accepted	Rejected	Accepted
EZ16	0.420	3.002	0.250	54	Accepted	Accepted	Rejected	Accepted
EZ17	0.534	3.079	0.318	70	Accepted	Rejected	Rejected	Accepted
EZ18	0.528	3.422	0.286	79	Accepted	Accepted	Rejected	Accepted
EZ19	0.512	2.873	0.318	51	Accepted	Rejected	Rejected	Accepted
EZ20	0.512	3.157	0.304	65	Accepted	Rejected	Rejected	Accepted
EZ21	0.500	2.929	0.190	56	Accepted	Accepted	Rejected	Accepted
EZ22	0.507	3.452	0.238	83	Accepted	Accepted	Rejected	Accepted
EZ23	0.331	3.687	0.222	94	Accepted	Accepted	Rejected	Rejected
Std.	0.549	2.837	0.238	54	Accepted	Accepted	Rejected	Accepted

QED; A measure of drug-likeness based on the concept of desirability, Synth; Synthetic accessibility score, Fsp3; The number of sp3 hybridized carbons/total carbon count, MCE-18; Medicinal chemistry evolution in 2018, GT; Golden triangle.

Study of pharmacokinetic (ADME) properties

Pharmacokinetic properties such as absorption (caco-2, MDCK and HIA), distribution (BBB, PPB and VD), metabolism (CYP1A2), excretion (CL and T_1/2) have been calculated using ADMETlab 2.0, and their scores are tabulated in Table 3. The caco-2 score of analogues EZ7, EZ10, EZ14, and EZ16 was > -5.15, which indicates proper in-vivo drug permeability. HIA scores of all met in the range between 0 and 0.3 indicate excellent oral bioavailability. The MDCK score, which predicts the in-vitro permeability of all analogues found excellent. The BBB score analogues EZ12, EZ17, and EZ23, has been found safe for CNS side effects, whereas ENZ is 0.964.

Table 3.

Prediction of pharmacokinetic profile of the analogues

Entry No.	Caco-2	HIA	MDCK	BBB	PPB (%)	VD	Fu (%)	CYP1A2		CL	T1/2
								Inh	Sub
EZ1	-5.123	0.008	EX	0.906	95.27	1.066	1.69	-	+	6.573	0.08
EZ2	-5.167	0.007	EX	0.862	96.81	1.09	1.45	-	+	6.625	0.07
EZ3	-5.206	0.01	EX	0.853	96.19	1.123	1.62	-	+	5.397	0.11
EZ4	-5.166	0.006	EX	0.869	95.71	1.293	1.93	+	-	4.776	0.08
EZ5	-5.170	0.011	EX	0.71	97.42	1.341	1.28	-	+	5.552	0.05
EZ6	-5.188	0.008	EX	0.813	98.31	1.863	1.07	-	+	5.669	0.08
EZ7	-5.024	0.009	EX	0.874	94.73	1.153	2.14	-	+	6.279	0.24
EZ8	-5.166	0.006	EX	0.869	95.71	1.293	1.93	+	-	4.776	0.08
EZ9	-5.183	0.018	EX	0.926	94.25	0.785	2.64	-	+	4.761	0.07
EZ10	-4.971	0.006	EX	0.856	91.88	1.077	3.06	-	+	4.758	0.17
EZ11	-5.283	0.007	EX	0.818	83.78	1.278	7.51	-	+	4.582	0.06
EZ12	-5.267	0.012	EX	0.337	96.44	0.214	1.53	-	+	2.547	0.21
EZ13	-5.279	0.011	EX	0.492	84.49	2.858	6.29	-	+	6.048	0.05
EZ14	-5.122	0.01	EX	0.552	88.95	1.387	3.65	-	+	6.600	0.10
EZ15	-5.382	0.01	EX	0.846	79.90	2.232	6.19	-	+	4.354	0.07
EZ16	-4.968	0.008	EX	0.95	94.46	0.908	2.19	-	+	4.833	0.44
EZ17	-5.267	0.01	EX	0.694	93.47	3.466	3.68	-	+	5.609	0.09
EZ18	-5.292	0.008	EX	0.502	85.02	1.435	6.74	-	+	6.252	0.16
EZ19	-5.338	0.014	EX	0.745	78.18	2.95	9.24	-	+	7.061	0.10
EZ20	-5.277	0.009	EX	0.668	87.40	2.884	5.80	-	+	6.406	0.04
EZ21	-5.373	0.006	EX	0.952	92.78	0.954	3.41	-	+	3.404	0.13
EZ22	-5.312	0.011	EX	0.21	92.63	0.374	3.15	+	-	4.072	0.25
EZ23	-5.150	0.006	EX	0.741	98.22	0.629	1.37	-	+	4.301	0.05
STD	-5.184	0.014	EX	0.964	94.90	1.047	1.92	-	+	6.997	0.12

Caco-2; The human colon adenocarcinoma cell lines, MDCK; Madin − Darby canine kidney cells, HIA; Human intestinal absorption, PPB; Plasma protein binding, BBB; Blood-brain barrier, VD; Volume distribution, Fu; The fraction unbound in plasms, EX; Excellent. (-); Indicates inhibitor, (+); Indicates substrate of human cytochrome P450 (five isozymes-1A2, 3A4, 2C9, 2C19 and 2D6), CL; The clearance of a drug, T_1/2; The half-life of a drug.

Analogues EZ11, EZ13-15, and EZ18-20 have a proper PPB (< 90%), which indicates good distribution volume and decreases the half-life of elimination compared to ENZ (94.90%). The volume of distribution (VD) of all analogues has an excellent score (0.04–+20). Cytochrome P450 (CYT P450) plays a crucial role in the metabolism of drugs. Analogues may be substrates or inhibitors to CYT P450. The CYT P450 enzyme is a substrate for analogues, which causes molecules to undergo metabolism. If analogues inhibit the enzyme, however, metabolism will not occur. Analogues EZ24, EZ8-12, EZ15, EZ16 and EZ21-23 has been found excellent clearance score (≥ 5) indicating a low risk of toxicity. T_1/2 score of all analogues found in the range (0 to 0.3) with exception of EZ16 which indicates proper clearance from the body.

Study of the toxicity properties

Toxicological characteristics of analogues such as drug induced liver injury (DILI), mutagenicity (Ames), rat oral acute toxicity (ROA), androgen receptor-a nuclear hormone receptor (NR-AR), and others were calculated using ADMETlab 2.0 (Table 4). The DILI and H-HT scores for all analogues showed the same as ENZ with toxic effects. The mutagenic score of all analogues was predicted to be safer, indicating that the analogues could not cause mutagenesis. The ROA prediction score of the analogues EZ2, EZ5-7, and EZ23 was found in a safer range (< 0.3) which is an important safety profile for drug candidates, whereas ENZ is toxic. The carcinogenicity of analogues is a serious issue because of their powerful effects on wellness and because they can damage the genome or disrupt cellular metabolism. NR-AR plays a crucial role in AR-dependent PC as well as androgen-related diseases. The NR-AR score of the analogues EZ12, EZ16, and EZ22 predicted that they could bind with NR-AR and may alter the activity of the androgen receptor.

Table 4.

Prediction of toxicity profile of the analogues

Entry No.	H-HT	DILI	Ames	ROA	NR-AR	NR-AR-LBD
EZ1	0.974	0.979	0.017	0.343	0.474	0.046
EZ2	0.971	0.978	0.018	0.274	0.502	0.051
EZ3	0.977	0.983	0.039	0.421	0.154	0.027
EZ4	0.976	0.988	0.041	0.354	0.445	0.053
EZ5	0.99	0.985	0.015	0.206	0.468	0.141
EZ6	0.969	0.979	0.019	0.297	0.598	0.158
EZ7	0.976	0.985	0.035	0.264	0.474	0.079
EZ8	0.976	0.988	0.041	0.354	0.445	0.053
EZ9	0.971	0.982	0.334	0.764	0.064	0.183
EZ10	0.97	0.982	0.052	0.51	0.636	0.074
EZ11	0.99	0.979	0.036	0.902	0.023	0.047
EZ12	0.977	0.994	0.032	0.751	0.708	0.104
EZ13	0.984	0.983	0.021	0.814	0.038	0.167
EZ14	0.981	0.99	0.109	0.327	0.646	0.098
EZ15	0.984	0.98	0.037	0.626	0.041	0.013
EZ16	0.984	0.982	0.025	0.416	0.723	0.062
EZ17	0.959	0.966	0.027	0.619	0.338	0.148
EZ18	0.972	0.977	0.043	0.841	0.452	0.047
EZ19	0.986	0.971	0.017	0.902	0.262	0.013
EZ20	0.984	0.983	0.021	0.733	0.035	0.134
EZ21	0.966	0.991	0.114	0.382	0.725	0.021
EZ22	0.958	0.985	0.038	0.71	0.439	0.076
EZ23	0.963	0.985	0.046	0.203	0.612	0.099
Std.	0.975	0.986	0.062	0.668	0.155	0.036

H-HT; The human hepatotoxicity, DILI; Drug-induced liver injury, Ames; Test for mutagenicity, ROA; Rat oral acute toxicity, NR-AR; Androgen receptor - a nuclear hormone receptor, NR-AR-LBD; Molecule bind with LBD of androgen receptor.

Drug likeness and drug score studies

With the use of PEO, we performed DL and DS predictions of newer ENZ analogues in the current study, as shown in Table 5. The ligand EZ13 had a higher DL score than ENZ (-6.38), followed by EZ15 and EZ20, both of which had − 7.51. Table 5 makes it quite evident that EZ12, EZ8 and EZ10 may not only be non-toxic but also have good DS. The positive DS values for the ligands EZ8 (0.27), EZ10 (0.26), and EZ12 (0.23), respectively, suggest that these compounds may be considered as drug candidates (antiandrogen properties).

Table 5.

Prediction of drug likeness and drug score of the analogues

Entry No.	Toxicity				DL	DS
	M	T	I	R
EZ1	G	G	G	G	-11.4	0.20
EZ2	G	G	G	G	-11.1	0.19
EZ3	G	G	G	G	-7.80	0.20
EZ4	G	G	G	G	-8.69	0.23
EZ5	G	G	G	G	-13.53	0.19
EZ6	G	G	G	G	-9.27	0.17
EZ7	G	G	G	G	-9.04	0.24
EZ8	G	G	G	G	-9.85	0.27
EZ9	G	G	G	G	-22.18	0.20
EZ10	G	G	G	G	-9.84	0.26
EZ11	G	G	G	G	-8.33	0.17
EZ12	G	G	G	G	-11.41	0.23
EZ13	G	G	G	G	-6.38	0.22
EZ14	O	O	G	G	-9.57	0.14
EZ15	G	G	G	G	-7.51	0.21
EZ16	G	G	G	G	-10.02	0.25
EZ17	G	G	G	G	-8.47	0.20
EZ18	G	G	G	G	-12.19	0.26
EZ19	G	G	G	G	-8.15	0.21
EZ20	G	G	G	G	-7.51	0.20
EZ21	O	O	G	O	-7.82	0.11
EZ22	G	G	G	G	-14.62	0.26
EZ23	G	G	G	G	-8.59	0.24
STD	G	G	G	G	-8.17	0.22

M; Mutagenic, T; Tumorigenic, I; Irritant, R; Reproductive, G; No toxicity risk, O; Toxicity risk, DL; Drug likeness, DS; Drug score.

Molecular docking study

The purpose of this work was to look at newly designed ENZ analogues that might interact with the protein and be applied to the therapy of PC. Of all the docked molecules, the top three molecules with the maximum docking score or minimum binding energy were EZ8, EZ10, and EZ12 (Fig. 1). The protein utilised in this investigation was sourced in 3D from the RCSB-PDB, a research collaboration for structural bioinformatics database. Proteins’ 3D crystallographic structures can be found in the protein data bank (PDB ID: 2OZ7). The protein was docked with the ligands (Table 1) using ADV software, which produced grid box dimensions that were consistent and helped to clarify the inhibitors’ binding affinities. Table 6 displays the binding affinities of the chosen ligands, which range between − 3.6 Kcal/mol and − 7.4 Kcal/mol.

Table 6.

Summary of docking study of ligands against PDB ID: 2OZ7

Entry no.	Docking score (Kcal/mol)	Interactions	Distance (Å)
EZ6	-4.6	GLN711, ARG752, SER778, LEU700, LEU701, LEU704, LEU707, LUE711, MET742, MET780, PHE764	2.60
EZ7	-4.1	GLN711, ARG752, SER778, LEU704, LEU707, MET780, LEU880	2.21
EZ8	-7.3	GLN711, ARG752, SER778, LEU704, LEU707, MET742, MET780, PHE764, MET745	2.14
EZ10	-6.9	GLN711, ARG752, SER778, LEU704, LEU707, MET742, MET780, PHE764, MET745	2.01
EZ12	-7.4	GLN711, ARG752, SER778, LEU704, LEU707, MET742, MET780, PHE764, MET745	2.12
EZ14	-4.6	GLN711, PHE697, VAL746, SER778, LEU704, MET745, PHE746, MET742, LEU873, LEU701, MET780	3.41
EZ16	-5.0	GLN711, ARG752, SER778, LEU704, LEU707, MET780, PHE764	1.83
EZ22	-3.6	GLN711, SER778, LEU704, VAL746, MET749, LEU873, LEU707, MET742, MET780, PHE764, MET745	3.09
STD	-3.8	GLN711, ARG752, SER778, LEU704, LEU707, MET742, MET780, PHE764, MET745	3.15

EZ12 has the most efficient binding with the AR protein with interaction energy (-7.4 kcal/mol). The EZ12 analogue had a carboxylic acid ring substituent in place of a N-methyl-formamide ring at the fluorophenyl ring in enzalutamide; the rest of the structure was similar. The EZ12 analogue showed conventional hydrogen bond interaction with the SER778, GLN711, and ARG752 residues through the hydrogen of the carboxylic acid, the fluorine of the trifluoromethyl, and the nitrogen of the cyano group, respectively. The fluorine atoms in the structure show maximum interactions with the amino acid residues of the binding site. The two fluorine atoms of the trifluoromethyl substitution showed halogen fluorine bonding with LEU704 and LUE707. The phenyl groups showed pi-sulphur interaction with MET742 and MET780 and pi-pi-T-shaped interaction with PHE764. Also, both phenyl groups showed pi-alkyl interactions with MET745, LEU704, and PHE764. The methyl group in the trifluoromethyl substitution showed an alkyl interaction with LEU704 and LUE707.

EZ8 has the second-most efficient binding with the AR protein with interaction energy (-7.3 kcal/mol). The hydroxylamine ring substituted for the N-methyl-formamide ring at the fluorophenyl ring in enzalutamide was present in the EZ8 counterpart, but the remainder of the structure remained the same. The EZ8 analogue exhibited typical hydrogen bond interactions with the residues SER778, GLN711, and ARG752 through the hydrogen of the hydroxylamine, the fluorine of the trifluoromethyl and the nitrogen of the cyano group, respectively. The structure’s fluorine atoms engage with the binding site’s amino acid residues to the greatest extent possible. Halogen fluorine bonding was seen between the two fluorine atoms of the trifluoromethyl substitution and LEU704 and LUE707. The phenyl groups interacted pi-sulphur with MET742, MET780, and pi-pi-T-shaped with PHE764. Furthermore, pi-alkyl interactions were observed between both phenyl groups and MET745, LEU704, and PHE764. An alkyl interaction was observed between the methyl group in the trifluoromethyl substitution and LEU704 and LUE707.

The interaction energy of EZ10 (-6.9 kcal/mol) indicates the third most efficient binding with the AR protein. The EZ10 analogue’s fluorophenyl ring in enzalutamide had a 1-hydroxyethanol ring substituent in lieu of an N-methyl-formamide ring, while the remainder of the structure was the same. The EZ10 analogue exhibited conventional hydrogen bond interactions with the residues SER778, GLN711, and ARG752 through the hydrogen of the 1-hydroxyethanol, the fluorine of the trifluoromethyl and the nitrogen of the cyano group, respectively. Surprisingly, 1-hydroxyethanol form two conventional hydrogen bonds with SER778 residue. In contrast, ENZ have a binding affinity score of -3.8 Kcal/mol. Figures 3, 4, 5 and 6 depict the 3D and 2D interactions of EZ12, EZ8, EZ10 and ENZ, respectively. The literature claims that the amide group of ENZ and the amino acid residue SER778 formed a hydrogen bond (Wang et al. 2021). The binding posture of ligands EZ12, EZ8, and EZ10 is comparable to that of ENZ. The binding scores of EZ12, EZ8, and EZ10 were − 7.4, -7.3, and − 6.9 Kcal/mol, respectively, which were superior to ENZ’s binding score. SER778, the active residue of protein (PDB ID: 2OZ7), might be the reason behind the antagonistic activity on AR. The docking analysis suggests that ligands EZ12, EZ8, and EZ10 could be potent androgen receptor inhibitors. The docking study suggests that ligands EZ12, EZ8, and EZ10 could be useful antagonists against the androgen receptor, which would lead to a potential treatment for PC.

Fig. 3.

3D (a) and 2D (b) interactions diagram of EZ12 against the receptor (PDB ID: 2OZ7)

Fig. 4.

3D (a) and 2D (b) interactions diagram of EZ8 against the receptor (PDB ID: 2OZ7)

Fig. 5.

3D (a) and 2D (b) interactions diagram of EZ10 against the receptor (PDB ID: 2OZ7)

Fig. 6.

3D (a) and 2D (b) interactions diagram of ENZ against the receptor (PDB ID: 2OZ7)

Conclusion

ENZ is one of the non-steroidal antiandrogen drugs used in PC therapy. Patients receiving ENZ exhibit a variety of toxicities. Initially, the bioisosteric replacement approach was carried out on the N-methyl formamide group of enzalutamide to generate a library of its analogues. From a library of 195 analogues, 23 were selected based on the medicinal properties, ADMET, drug likeness and DS prediction to confer the drug-like properties of these analogues. Using a molecular docking approach, the 23 analogues were investigated for binding affinity with the target protein. Among them, some compounds showed good interactions with the target protein (PDB ID: 2OZ7); compounds EZ12, EZ8, and EZ10 were the most active molecules, suggesting a plausible binding mode with the enzyme (SER788), which is responsible for the AR antagonistic activity. These findings imply that the therapeutic profile of enzalutamide can be enhanced by a minor structural change. The study’s established approach may serve as a model for the future development of novel small molecules for the treatment of prostate cancer. It also suggests that the newly created compounds, EZ12, EZ8, and EZ10, could be further explored for their potential as antiandrogen agents in the treatment of prostate cancer.

Author contributions

AKG did the designing of the newer analogues using bioisosteric approach and ADMET study and wrote the manuscript. PG help in molecular docking study. SKJ is the Mentor and supervised the whole work. Analyse and Checked the data and manuscript.

Declarations

Conflict of interest

The authors declare no competing interests.