Development of potential inhibitors of cell division protein kinase 2 by ligand based drug design

Abstract

Cyclin-dependent kinases (CDKs) are commonly known by their role in cell cycle regulation which affects cancer mechanism. In many cancer types, CDKs show extreme activity or CDK inhibiting proteins are dysfunctional. Specifically, CDK2 plays an indispensable role in cell division especially in the G1/S phase and DNA damage repair. Therefore, it is important to find new potential CDK2 inhibitors. In this study, ligand-based drug design is used to design new potential CDK2 inhibitors. Y8 L ligand is obtained from the X-ray crystal structure of human CDK2 (PDB ID: 2XNB) (www.pdb.org) and used as a structure model. By adding hydrophilic and hydrophobic groups to the structure, a training set of 36 molecules is generated. Each molecule examined with Spartan’14 and optimized structures are used for docking to CDK2 structure by AutoDock and AutoDock Vina programs. Ligand-amino acid interactions are analysed with Discovery Studio Visualizer. Van der Waals, Pi-Pi T-shaped, alkyl, pi-alkyl, conventional hydrogen bond and carbon-hydrogen bond interactions are observed. By docking results and viewed interactions, some molecules are identified and discussed as potential CDK2 inhibitors. Additionally, 8 different QSAR descriptors obtained from Spartan’14, Preadmet and ALOGPS 2.1 programs are investigated with multiple linear regulation (MLR) analysis with SPSS program for their impact on affinity value.

Keywords

CDK2 docking ligand-based drug design MLR Y8 L ligand

1 Introduction

Cancer is one of the multifactorial diseases and lots of players such as microorganisms, viruses, radiation, genetic, dietary habits and environmental factors increase the risk of contracting cancer [1 –4]. In recent years, the number of cancer cases has reached up to 14.1 million and unfortunately 58.2% of the cases end in with death [1]. The structure of cyclin dependent kinases (CDKs) have been clarified in recent years and the relationship between CDKs and cancer is illuminated [5]. Also, their molecular mechanisms of regulation have been defined completely and therefore it got easier to develop drugs to inhibit their function [5]. CDKs are necessary for cell cycle progression and there are 20 CDKs and 29 cyclins in human cells which take part in cell division progress [6, 7]. It has been shown that seven of them (CDK1/2/3/4/6/10/11) are closely related to cell division cycle [6]. Also, their roles inmetabolism and other important biological processes like neuronal differentiation, ciliogenesis and gene transcription are discovered in recent years [7 –11]. Amplification, overexpression or mutation cause dysregulation of CDKs, which triggers uncontrolled cell proliferation and cancer. CDK inhibitors are considered as potential drugs for many diseases such as cancer, diabetes, renal, neurodegenerative and infectious diseases [6 , 12–15]. Therefore, it is crucial to find novel CDK inhibitors because inhibition of CDKs help trigger apoptosis and fight back with cancer [14, 16]. In this research we have been working with CDK2 (PDB format: 2XNB) to find potential inhibitors to help cells to end uncontrolled cell proliferation. Activation of CDK2 helps cells to end the G1 phase and trigger DNA replication. Cyclins E and A are required for CDK activity and help cells to progress through to the G2 phase. Comprehensively, CDK2 and the cyclin E complex controls the entrance into S phase and initiates synthesis of DNA. Cyclin A and CDK2 complexes are associated with replication and progression through S phase and they also play a role as activator for CDK1 and cyclin B complex during the G2 phase [5]. Overexpression of CDK2 is associated with laryngeal squamous cell cancer, advanced melanoma and breast cancer [5 , 17–21]. Amplification and/or overexpression of cyclin A and E cause hyperactivation of CDK2 in human cancers especially in breast cancer, ovarian and endometrial carcinomas, lung and thyroid carcinoma, melanoma and osteosarcoma [6]. There are several CDK-2 inhibitors known as Alvocidib, AT-7519, Milciclib, A-674563 and LDC000067 [6]. In this study we also examined Alvocidib, A-674563 and LDC000067 inhibitors. Alvocidib, also known as flavopiridol competes with ATP and inhibits CDKs such as CDK1,2,4,6. A-674563 is commonly known as Akt1 inhibitor but also inhibits CDK-2. LDC000067 is a potent inhibitor of CDK9 and also inhibits CDK2. The inhibitory selectivity between these 3 inhibitors for CDK2 is like Alvocidib > A-674563 > LDC000067 [22]. The key stone of this research is Y8 L ligand which was defined as transcriptional CDK inhibitor and cancer agent in a study by Wang et al. [23]. In vivo tests of their work shows that Y8 L ligand caused a delay in tumor growth for human colorectal Colo-205 cell line and also treatment of Y8 L ligand enhanced the life span of the treated animals for P388/0 murine leukaemia survival model [23].

2 Methods

In this research we have been focused on Y8 L, this ligand is a unique ligand of human CDK2X-ray crystal structure obtained from Protein Data Bank server (PDB format:2XNB) with resolution of 1.85Å. First of all, the structure of Y8 L ligand is built on Spartan’14 by looking at the literature structure that was obtained from PDB bank [24]. Then, conformer analysis is performed for Y8 L ligand and the best structure is optimized with Spartan’14. After obtaining electrostatic potential map, hydrophilic and hydrophobic functional groups are added to the structure. By adding hydrophilic and hydrophobic functional groups, we obtained 36 different molecules and after designing each molecule, docking calculations are performed and the results obtained are used for designing the new derivatives of the model structure. 16 different functional groups (–NH₂, –COOH, –COOF, –OSO₂OH, –CONH₂, –COCH₃, –CH₃, –SH, –C₂H₃, –C₆H₅, NO₂, –C(CH₃)₃, –COH, –C₇H₁₃, –CH₂, C(CH₃)₃, –C₆H₁₁ and their combinations are used to design our training set. All the designed molecules are examined in Spartan’14 with both conformer distribution and equilibrium geometry calculation. Conformer distribution calculations are performed with Molecular Mechanics/MMFF (Molecular Mechanics Force Field) and equilibrium geometry calculations are performed with Semi Empirical/PM6 in Spartan’14 for 40 molecules including the training set, Y8 L ligand, alvocidib, A-674563 and LDC000067 inhibitors. After conformer analysis and geometry optimization, optimized form of each molecule is used for docking in Autodock and Autodock Vina programs [25]. The grid centre coordinates were obtained from the X-ray crystal structure of CDK2 and Y8 L complex with Discovery Studio Visualizer program (Table 1) [26]. After dockings, affinity values (kcal/mol) are obtained and interactions between designed molecules and amino acids are observed with distance.

Table 1
Grid box coordinates, sizes, and spacing

Coordinates Grid Box Sizes Grid Spacing

x = 24.939 x = 40 0.375 Å

y = 27.894 y = 40

z = 27.362 z = 40

Coordinates	Grid Box Sizes	Grid Spacing
x = 24.939	x = 40	0.375 Å
y = 27.894	y = 40
z = 27.362	z = 40

In this study, eight QSAR descriptors obtained by using three different programs and the relationship between these descriptors and affinity values obtained from the AutoDock program is examined by using regression analysis. Spartan’14 is used to obtain QSAR descriptors such as log P, dipole moment (debye), polarizability and weight (amu). Mlog P, Xlog P3 and average log S values are obtained by using ALOGPS 2.1 and mol2 formats of the optimized structures used for the ALOGPS 2.1 calculations [27]. Pure water solubility (mg/L) values are calculated from online PreADMET program with ADME applet [28]. Also, all descriptors used in this work can be seen in Table 3. All of the QSAR descriptors for 40 molecules (including the training set, Y8 L ligand, alvocidib, A-674563 and LDC000067 inhibitors) are obtained by Spartan’14, ALOGPS 2.1 and PreADMET programs, examined with multiple linear regression analysis in SPSS program to observe the effect of QSAR descriptors on affinity value obtained by AutoDock program and a general equation of the model is generated to use in forthcoming researches [29]. The X-ray crystal structure of CDK2 (PDB format: 2XNB) and the binding site of Y8 L ligand are showed in Fig. 1.

Fig. 1

Structure of CDK2 (PDB format: 2XNB) (a), the binding site of CDK2 (b), CDK2 and Y8 L complex (c).

Also, the structures of Y8 L ligand, Alvocidib, A-674563 and LDC000067 inhibitors are given in Fig. 2.

Fig. 2

The structures of (a) Y8 L ligand, (b) Alvocidib, (c) A-674563, (d) LDC000067.

The numbered attachment places of functional groups to the structure of Y8 L ligand is given in Fig. 3.

Fig. 3

Attachment positions for Y8 L ligand.

For this research, most of the QSAR descriptors were chosen from a perspective study by Zineb Almi et al. [30] such as logP, polarizability, dipole moment and weight. As mentioned in that research, octanol/water partition coefficient (log P), is accepted as an essential descriptor used in drug design like other descriptors. Additionally, other descriptors such as polarizability, solubility, weight and dipole moment are considered as useful descriptors [30 –33]. All the calculated descriptors are summarized in Table 2 in this study.

Table 2

Symbols and definitions of calculated QSAR descriptors

Log P	Log of the octanol/water partition coefficient^a
μ	Dipole moment^a
Polarizability	Polarizability^a
W	Weight^a
MLog P	Moriguchi octanol/water partition coefficient^b
XLog P3	Log of the octanol/water partition coefficient^b
Avg. Log S	Average solubility^b
PWS	Pure water solubility^c

^aCalculated with Ref 24, ^bCalculated with Ref 27, ^cCalculated with Ref 28.

In this research octanol/water partition coefficient calculated for three times with different programs such as Spartan’14 and ALOGPS 2.1 (log P, Mlog P, Xlog P3) to see the difference caused by the selected program and the method. Detailed structures of designed molecules are given in Table 3 in order to the attachment places. All designed molecules examined the way explained before. Additionaly, it can be seen that addition of -COOH, -CONH₂, -COCH₃, -C₆H₅, NO₂, -C(CH₃)₃, -C₇H₁₃, -CH₂C(CH₃)₃ and -C₆H₁₁ functional groups cause an undeniable raise in affinity from other functional groups.

Table 3

Attachment substituent of functional groups

No	R₁	R₂	R₃	R₄	R₅	R₆
1	–NH₂
2	–COOH
3	–COOF
4	–OSO₂OH
5	–CONH₂
6		–CH₃
7		–SH
8			–C₂H₃
9			–C₆H₅
10				–CH₃
11						–CH₃
12	–COOH	–CH₃	–CH₃
13	–COCH₃			–CH₃
14	–COCH₃		–CH₃		–CH₃
15	–CONH₂		–CH₃
16	–CONH₂				–CH₃
17	–COCH₃					–CH₃
18	–NO₂
19	–COH					–CH₃
20	–COCH₃				–CH₃
21	–COOH				–CH₃
22	–COCH₃					–CH₃
23	–COCH₃		–C(CH₃)₃
24	–NO₂		–C(CH₃)₃
25	–NO₂					–CH₃
26					–C(CH₃)₃
27	–NO₂				–CH₃
28			–C₇H₁₃
29			–CH₂C(CH₃)₃
30	–COOH		–C₆H₁₁
31	–COCH₃		–C₆H₁₁
32	–CONH₂		–C₆H₁₁
33	–NO₂		–C₆H₁₁
34	–COOH		–C₇H₁₃
35	–CONH₂		–C₇H₁₃
36	–NO₂		–C₇H₁₃

3 Results and Discussion

Ligand-amino acid interactions are examined in Discovery Studio Visualizer programme and the 2D diagrams of Y8 L ligand, 32nd, 23rd and 35th molecules are given in Fig. 4. As can be seen in Fig. 4, a total number of interactions Y8 L ligand made with CDK2 is 25, for 32nd molecule the total number is 28. Despite the loss of interaction with LEU83, new interactions with ILE10 and LYS89 caused a raise in affinity with new van der Waals interactions such as GLU12, GLU8 and LYS88. Interaction with LEU83 was also lost in 23rd molecule but new interactions with ILE10 and LYS89 were caused the raise in affinity with two new van der Waals interactions such as GLU12 and LYS88. Six of the interactions seen in Y8 L ligand were lost in 35th molecule but new interactions with ALA31, ALA144, PHE80, LYS33, LYS89 and GLY13 were caused the raise in affinity with new Van der Waals interactions like LYS20, LEU298, GLU8 and LEU83. As can be seen in Fig. 3 and Fig. 4, obtained form of Y8 L ligand from X-ray crystal structure of CDK2 (PDB ID: 2XNB) is different from the one in the literature (N atom between pyrimidine and benzene rings makes double bond in the literature form of Y8 L ligand) and loss of LEU83 conventional hydrogen bond caused by this difference.

Fig. 4

2D diagrams of ligand interactions between 2XNB structure and Y8 L ligand (a), 32nd molecule (b), 23rd molecule (c), 35th molecule (d).

Docking results of Y8 L ligand, Alvocidib, A-674563 and LDC000067 inhibitors are given in Table 4 with QSAR descriptors such as log P, dipole moment(μ), polarizability, weight, Mlog P, Xlog P3, average log S and pure water solubility.

Table 4

Docking and QSAR descriptors values of CDK2 inhibitors

	log P	μ	Polarizability	W	Mlog P	Xlog P3	Avg. log S	PWS	Affinity
Y8L	1.33	5.36	70.59	382.49	1.92	1.24	–2.36	898.7	–10.40
Alvocidib	–2.16	8.09	70.00	401.85	2.23	3.25	–3.75	26.04	–10.60
A-674563	0.08	4.32	70.40	358.45	2.35	3.56	–5.08	8.75	–10.20
LDC000067	–1.74	1.44	68.20	370.43	2.27	2.21	–5.08	1.03	–10.10

From the literature [22], it is known that the inhibitory selectivity between these 3 inhibitors decrease from Alvocidib to LDC000067 for cell division protein kinase 2 and the obtained affinity results with Autodock program (given in Table 4) fit to this literature information which confirms the accuracy of the obtained affinity values for these 3 molecules. Also, only the affinity value of Alvocidib is higher than Y8 L ligand. From the docking calculations, obtained affinity values of the training set and the values of QSAR descriptors are also given in Table 5.

Table 5

Docking and QSAR descriptors values of the training set

_No .	log P	μ	Polarizability	W	Mlog P	Xlog P3	Avg. log S	PWS	Affinity
1	1.23	7.47	71.54	397.51	1.98	0.96	–2.03	515.74	–10.2
2	1.30	3.90	72.97	426.50	1.39	1.24	–2.32	625.91	–10.80
3	1.64	2.02	73.54	444.49	1.50	2.11	–2.86	100.07	–10.50
4	1.23	4.48	74.68	478.55	2.07	–1.45	–1.54	387.53	–9.70
5	0.65	7.12	73.22	425.52	1.39	0.59	–2.31	450.41	–11.00
6	1.67	4.94	72.09	396.52	2.14	1.60	–2.57	556.49	–10.20
7	1.92	4.71	72.05	414.56	1.92	1.62	–3.46	89.21	–9.90
8	1.66	5.91	73.11	408.53	2.29	1.56	–2.63	345.04	–10.40
9	2.70	5.41	77.31	458.59	3.00	2.41	–3.15	41.71	–10.40
10	1.65	5.88	72.07	396.52	2.14	1.67	–2.63	556.49	–10.30
11	1.65	5.06	72.10	396.52	2.14	1.67	–2.63	556.49	–10.10
12	1.82	6.33	75.85	454.56	1.84	1.52	–2.56	140.68	–11.10
13	1.30	3.51	75.32	438.56	2.17	1.60	–2.49	493.48	–10.80
14	1.64	6.23	76.79	452.58	2.38	1.51	–2.50	86.39	–11.30
15	0.83	5.64	74.69	439.54	1.62	0.51	–2.38	164.26	–11.20
16	1.14	5.60	74.74	439.54	1.62	1.02	–2.51	133.53	–11.00
17	1.30	6.25	75.35	438.56	2.17	1.60	–2.49	493.48	–11.30
18	–0.15	2.82	72.38	427.49	2.06	1.62	–2.45	78.28	–10.70
19	1.03	4.20	73.90	424.53	1.95	1.61	–2.55	455.29	–10.70
20	1.47	6.39	75.35	438.56	2.17	1.60	–2.43	237.11	–11.20
21	1.79	4.49	74.47	440.53	1.62	1.67	–2.52	185.54	–10.90
22	1.30	6.28	75.33	438.56	2.17	1.60	–2.49	493.48	–11.40
23	2.37	5.38	79.49	480.64	2.80	2.39	–2.84	32.23	–12.10
24	1.24	6.16	78.06	483.60	2.94	2.85	–3.06	3.15	–11.50
25	0.17	2.22	73.87	441.52	2.29	2.06	–2.71	48.29	–10.90
26	3.04	4.52	76.33	438.60	2.80	2.86	–3.03	29.82	–10.70
27	0.33	2.71	73.88	441.52	2.29	2.06	–2.65	23.20	–10.60
28	3.43	6.66	79.83	478.67	3.42	3.86	–3.49	3.62	–11.30
29	3.14	4.98	77.81	452.63	3.01	2.77	–3.14	15.51	–11.10
30	2.98	3.66	80.71	508.65	2.67	3.33	–3.34	5.88	–11.40
31	2.66	4.13	81.60	506.68	3.21	3.25	–3.26	7.52	–11.50
32	2.33	6.65	80.98	507.66	2.67	2.68	–3.36	4.23	–12.20
33	1.52	3.68	80.11	509.64	3.36	3.71	–3.48	0.73	–11.00
34	3.40	3.93	82.14	522.67	2.87	3.87	–3.50	2.47	–11.40
35	2.75	6.99	82.47	521.69	2.87	3.22	–3.56	1.78	–11.70
36	1.94	3.99	81.53	523.66	3.57	4.25	–3.66	0.31	–11.20

28 of the 36 molecules showed higher affinity than Y8 L ligand to CDK2 and the highest affinity, –12.2 kcal/mol, was achieved by adding –CONH₂ and –C₇H₁₃ to the Y8 L ligand’s structure (Table 5). Additionally, it can be seen that addition of –COOH, –CONH₂, –COCH₃, –C₆H₅, NO₂, –C(CH₃)₃, –C₇H₁₃, –CH₂C(CH₃)₃ and –C₆H₁₁ functional groups cause an undeniable raise in affinity from other functional groups [33, 34]. To see which of the QSAR descriptors are most relevant with the affinity value, all the data obtained from 40 ligands including the training set, Y8 L, Alvocidib, A-674563 and LDC000067 are examined with multiple linear regression analysis with SPSS program.

As seen in Table 6, R² and adjusted R² values are higher than 0.8 which means our model is acceptable and 86% of the change in affinity can be explained by our 8 descriptors. In Table 7, significance values also show that log P, polarizability, W and MLog P affect affinity more than other calculated descriptors. Pure water solubility has the lowest effect on affinity according to significance and coefficient values. The raise in log P, weight and Mlog P increase affinity. Additionally, the raise in dipole moment and polarizability decrease affinity.

Table 6

Regression Model of all descriptors

R	R Square	Adjusted R Square	Std. Error of	R² Change the Estimate	Significance
0.929	0.862	0.827	0.23886	0.862	0,000

Table 7

Obtained results from SPSS multiple linear regression analysis

	Coefficient	Std. Error	t	Significance
Constant	13.327	2.945	4.525	0.000
log P	0.367	0.064	5.718	0.000
μ	–0.026	0.029	–0.906	0.372
Polarizability	–0.471	0.066	–7.104	0.000
W	0.020	0.005	4.497	0.000
Mlog P	0.609	0.128	4.749	0.000
Xlog P3	–0.125	0.077	–1.623	0.115
Avg. Log S	–0.205	0.113	–1.812	0.080
PWS	0.000	0.000	0.825	0.416

Obtained equation for all descriptors except PWS is given in Equation 1 with standard error coefficients. PWS, pure water solubility, is eliminated according to the obtained results.

Equation 1. Obtained equation from regression analysis cycles.

Affinity = 13.327 (±2.945) + 0.367 (±0.064) log P –0.026 (±0.029) μ –0.471 (±0.066)

Polarizability + 0.020 (±0.005) W + 0.609 (±0.128) Mlog P –0.125 (±0.077) Xlog P3 –0.205 (±0.113)

4 Conclusion

A group of 36 molecules are designed by using Y8 L ligand as a model structure. According to docking results, 24 of them showed higher affinity to CDK2 than Y8 L ligand and ligand-amino acid interactions are observed for Y8 L ligand and three molecules showed the highest affinity values and new interactions are seen in detail. Docking results and amino acid interactions showed that our training set contain new potential CDK2 inhibitors and could be useful for in vitro testing. For all the molecules eight QSAR descriptors are also calculated and performed with regression analysis and four of them found as the most related to affinity value. Our newly developed model, Affinity = 13.327 (±2.945) + 0.367 (±0.064) log P –0.026 232 (±0.029) μ –0.471 (±0.066) Polarizability + 0.020 (±0.005) W + 0.609 (±0.128) Mlog P –0.125 (±0.077) Xlog P3 –0.205 (±0.113) can be useful for predicting the affinity value of new Y8 L ligand derivatives as anti-cancer agents.

Conflicts of interest

There are no competing interests to declare.

Footnotes

Acknowledgments

We would like to express our appreciation to Dr. Türker Tekin Ergüzel from Üsküdar University for his guidance in multiple linear regression analysis (MLR) with SPSS program and Prof. Dr. Safiye Sağ Erdem from Marmara University for the software supports. The authors declare that there is no conflict of interest regarding the publication of this paper.

References

Torre

L.A.

, Bray

, Siegel

R.L.

, Ferlay

, Lortet-Tieulent

and Jemal

, Global cancer statistics, 2012, CA: A Cancer J Clin 65 (2015), 87.

Yokuş

and Çakır

, Kanser Biyokimyası, Dicle Üni Vet Fak Der 1 (2012), 7.

Williams

G.M.

, Mechanisms of chemical carcinogenesis and application to human cancer risk assessment, Toxicology 14 (2001), 166.

Williams

G.M.

, DNA reactive and epigenetic carcinogens, Exp Toxicol Pathol 44 (1992), 45.

Peyressatre

, Prével

, Pellerano

and Morris

M.C.

, Targeting cyclin-dependent kinases in human cancers: From small molecules to peptide inhibitors, Cancers (Basel) 7 (2015), 179.

Sánchez-Martínez

, Gelbert

L.M.

, Lallena

M.J.

and De Dios

, Cyclin dependent kinase (CDK) inhibitors as anticancer drugs, Bioorg Med Chem Lett 25 (2015), 3420.

Roskoski

, Cyclin-dependent protein kinase inhibitors including palbociclib as anticancer drugs, Pharmacol Res 107 (2016), 249.

Cingöz

and Goff

P.S.

, Cyclin-dependent kinase activity is required for type I interferon production, Proc Natl Acad Sci 201720431 (2018).

Tang

, Wang

, Cheng

, Si

, Sun

, et al., Wilms’ tumor 1-associating protein promotes renal cell carcinoma proliferation by regulating CDK2 mRNA stability, J Exp Clin Cancer Res 37 (2018), 1.

10.

Perrot

, Millington

C.L.

, Gómez-Escoda

, Schausi-Tiffoche

and Wu

P-Y.J.

, CDK activity provides temporal and quantitative cues for organizing genome duplication, PloS Genetics 14 (2018), 2.

11.

Paculová

and Kohoutek

, The emerging roles of CDK12 in tumorigenesis, Cell Div 12 (2017), 7.

12.

Mohan

and Asakura

, CDK inhibitors for muscle stem cell differentiation and self-renewal, J Sports Med 6(2) (2017), 65.

13.

Allegri

, Baldan

, Mio

, Puppin

, Russo

, et al., Effects of BP-14, a novel cyclin-dependent kinase inhibitor, on anaplastic thyroid cancer cells, Oncol Rep 35 (2016), 2413.

14.

Chang

M.Y.

, Shieh

D.E.

, Chen

C.C.

, Yeh

C.S.

and Dong

H.P.

, Linalool Induces Cell Cycle Arrest and Apoptosis in Leukemia Cells and Cervical Cancer Cells through CDKIs. Haenen GRMM, Int J of Mol Sci 16(12) (2015), 28169.

15.

Robb

C.M.

, Kour

and Contreras

J.I.

, Characterization of CDK(5) Inhibitor, 20-223 (aka CP63) for Colorectal Cancer Therapy, Oncotarget 9(4) (2018), 5216.

16.

Bose

, Simmons

G.L.

and Grant

, Cyclin-dependent kinase inhibitor therapy for hematologic malignancies, Expert Opin Investig Drugs 22(6) (2013), 723.

17.

Abdullah

, Wang

and Becker

, Expression analysis and molecular targeting of cyclin-dependent kinases in advanced melanoma, Cell Cycle 10(6) (2011), 977.

18.

Georgieva

, Sinha

and Schadendorf

, Expression of cyclins and cyclin dependent kinases in human benign and malignant melanocytic lesions, J Clin Pathol 54(3) (2001), 229.

19.

Weroha

S.J.

, Lingle

W.L.

, Hong

, Li

S.A.

and Li

J.J.

, Specific Overexpression of Cyclin E·CDK2 in Early Preinvasive and Primary Breast Tumors in Female ACI Rats Induced by Estrogen, Horm Cancer 1(1) (2010), 34.

20.

Liu

, Ma

, Lu

, Sun

, Guo

, et al., NF-κB induces abnormal centrosome amplification by upregulation of CDK2 in laryngeal squamous cell cancer, Int J Oncol 39 (2011), 915.

21.

Wang

, Yoshimi

, Suzui

, Yamauchi

, Tarao

, et al., Different expression patterns of cyclins A, D1 and E in human colorectal cancer, J Cancer Res Clin Oncol 122 (1996), 122.

22.

http://www.selleckchem.com/products/Flavopiridol.html

23.

Wang

, Griffiths

, Midgley

C.A.

, Barnett

A.L.

, Cooper

, et al., Discovery and Characterization of 2-Anilino-4- (Thiazol-5-yl) Pyrimidine Transcriptional CDK Inhibitors as Anticancer Agents, Chem Bio 17(10) (2010), 1111.

24.

Hehre

W.J.

, A Guide to Molecular Mechanics and Quantum Chemical Calculations, Wavefunction, Inc., Irvine, CA. (2003)

25.

Trott

and Olson

A.J.

, AutoDock Vina: improving the speed and accuracy of docking with a new scoring function, efficient optimization, and multithreading, J Comput Chem 31 (2010), 455.

26.

Dassault Systèmes BIOVIA, Discovery Studio Modeling Environment, Release 2017, San Diego: Dassault Systèmes, 2016.

27.

VCCLAB, Virtual Computational Chemistry Laboratory, http://www.vcclab.org, 2005.

28.

Seul, South Corea: Bioinformatics and Molecular Design Research Center; 2004. PreADMET program. Available from: URL: http://preadmet.bmdrc.org.

29.

SPSS Inc. Released 2007. SPSS for Windows, Version 16.0. Chicago, SPSS Inc.

30.

Almi

, Belaidi

, Lanez

and Noureddine

, Structure Activity Relationships, QSAR Modeling and Drug-like Calculations of TP Inhibition of 1,3,4-oxadiazoline-2-thione Derivatives, Int Lett Chem Phys Astron 18 (2014), 113.

31.

Fotios

, Constantinos

and Anna

, Lipophilicity and biomimetic properties to support drug discovery, Expert Opin Drug Discov 12(9) (2017), 885.

32.

Liu

, Quan

and Fang

, Effect of drug physicochemical properties on drug release and their relationship with drug skin permeation behaviors in hydroxyl pressure sensitive adhesive, Eur J Pharm Sci 93 (2016), 437.

33.

Adhikari

, Amin

S.A.

, Saha

and Jha

, Combating breast cancer with non-steroidal aromatase inhibitors (NSAIs): Understanding the chemico-biological interactions through comparative SAR/QSAR study, Eur J Med Chem 137 (2017), 365.

34.

Basith

, Cui

, Macalino

S.J.Y.

, et al., Exploring G Protein-Coupled Receptors (GPCRs) Ligand Space via Cheminformatics Approaches: Impact on Rational Drug Design, Front Pharmacol 9 (2018), 128.