Computational Approach for the Design of Flavone based CDK2/CyclinA Inhibitors: A Simulation Study Employing Pharmacophore based 3D QSAR

 

Aravinda Pai, B. S. Jayashree*

Department of Pharmaceutical Chemistry, Manipal College of Pharmaceutical Sciences,

Manipal Academy of Higher Education, Manipal-576104, Karnataka, India

 

ABSTRACT:

The present work aimed at designing selective and potent inhibitors of CDK2/CyclinA as anticancer agents. A five point pharmacophore (AAADR) model was developed for the reported molecules from literature, the pharmacophore model was used to build predictive 3D QSAR equation. The selected 3D QSAR models revealed the importance of hydrogen bond acceptors, hydrogen bond donors and aromatic rings for selectively towards the target enzyme. The developed models were statistically robust (CDK2/Cyclin A, Q²- 0.6380, R² value of 0.9857, SD-0.1667, F-320.9 and Pearson coefficient value of 0.7916). The built model could be useful for the design and development of novel and selective inhibitors of CDK2/Cyclin A.

 

 

 

INTRODUCTION:

The Cyclin dependent kinases (CDKs), belongs to a family of serine–threonine kinases that plays a pivotal role in the total process of cell division and proliferation. [1]. Cyclin dependent kinases are vigilantly controlled by the ATP dependent phosphorylation and specific interactions with their Cyclin partners [2]. Cancer is a well-known disease characterized by a berrant cell proliferation and majority of cancer cells witness the improper regulation of CDKs. As a result, CDK inhibitors have been designed and screened for their antiproliferative effect [3]. Each CDKs initiate phosphorylation of specific substrates in different stages of the cell cycle [4,5]. The mal functioning of Cyclin-dependent kinases (CDKs) highly deregulates normal cell cycle progression [6].

 

 

The enzyme CDK2/Cyclin E complex sustains hyper phosphorylation of the retinoblastoma tumor suppressor protein (pRb) during late G1 phase, which in turn controls the transcription factor levels (E2F). During S phase entry, CDK2/Cyclin A phosphorylates and subsequently disables the transcription factor E2F. Deregulated and increased levels of transcriptional activity by E2F leads to cell death by apoptosis. This clearly indicates the inhibition of CDK2 and its involvement in eliciting tumor cell apoptosis [7]. Inhibitors of cyclin dependent kinases are emerging anticancer therapeutics, which poses novel mechanistic action [8]. Importantly many heterocyclic derivatives are currently under development, which includes the first generation CDK inhibitor flavopiridol (alvocidib). Flavopiridol is one of the synthetic flavonoid and well developed as a CDK inhibitor because of its high affinity towards CDKs [9-12]. Apoptosis induction by flavopiridol is an extremely interesting area of investigation [13]. Flavopiridol has inhibitory effect on anti-apoptotic ligands, which involves bcl-2 [14], XIAP [15], cyclinD1 [16], and phospho-survivin [17]. The overall effect of flavopiridol on transcriptional antagonism is to promote apoptosis or to suppress cell proliferation. [18, 19].

 

 

Despite, the extensive well supportive clinical trial data, phase II effectiveness of flavopiridol as a mono therapy, failed to show significant therapeutic benefit in many multicenter trials [20–24]. Latest data by researchers at Ohio State University, indicates the dose and schedule dependent activity of flavopiridol [25]. In particular, flavopiridol poses high protein binding affinity (>90%) to human plasma proteins. Flavopiridol is a promising drug with nano molar activity in certain cancer cell lines. But high protein binding limits its usage as an individual therapeutically active agent. Newer analogues should be designed and to be explored for their anticancer activity and also to improve their pharmacokinetic profiles. With this background it was thought to design some flavopiridol analogues and to determine their binding affinity at the active site of CDK/Cyclin A. A literature data was taken, which has already reported few flavopiridol analogues binding to the same enzyme. The pharmacophore based 3D QSAR was performed in order to identify the inhibitors of CDK2/Cyclin A.

 

METHODOLOGY:

Work station and computer packages:

Pharmacophore generation and 3D QSAR model building were performed, utilizing PHASE v 3.0 version, which is embedded in the Maestro 11.8 platform from Schrodinger, LLC New York, and USA. The software was installed in a work station with Linux platform.

 

Data set used for experimental analysis:

In the current simulation study, the data set of 23 compounds were taken from the literature (26,27) with their invitro activity on CDK2/Cyclin A. Out of 23 compounds, 18 compounds were grouped under training set and 5 compounds were grouped under test set.

 

Pharmacophore hypothesis generation:

The pharmacophore models were generated using phase 3.0 module available with Schrodinger molecular simulation software [21]. In QSAR studies, the minimum energy structure is required to achieve accurate 3D descriptor values. The structures were drawn using 2D sketch option available with Maestro. The 2D structures were later converted to their corresponding 3D structures. Geometry optimization was performed using semiemperical OPLS force field.

 

PHASE uses a structure cleaning utility called ligprep, which adds hydrogen, generates stereoisomers and various conformers and predicts proper ionization states at a particular PH. The ligands were grouped into active and inactive sets by assigning a suitable activity cut-off range.

Generation of sites (pharmacophoric):

The chemical features of selected CDK2 inhibitors were defined by five point pharmacophoric features: one H bond acceptor (A) and three aromatic rings (R). The pharmacophore was developed based on the intersite distance and angles.

 

Scoring pharmacophores on the basis of their activitythreshold:

The resulting sets of pharmacophore were scored by specific algorithms and later ranked. The scoring will identify the best pharmacophore hypothesis. The scoring algorithm considered the effective contributions from the alignment of site points and magnitude of vectors, selectivity and activity with overall conformational energies.

 

Identifying common pharmacophore:

After the selection of best hypothesis, AAADR (Figure 1) it was further analyzed (Table 1). The pharmacophore AAADR (Figure 2) encompassed the following special features: Hydrogen bond acceptors (A (three) (two arrows with a pink sphere), one aromatic ring (R) (grey circle) and one hydrogen bond donor (D) (blue sphere).

 

Figure 1: Common pharmacophore hypothesis

 

 

Figure 2: QSAR model for the electron withdrawing group interactions

 

 

Table 1: Common pharmacophore hypothesis data

In	ID	Survival	Survival inactive	Post-hoc	Site	Vector	Volume	Selectivity	Matches	Energy	Activity	Inactive
1	AAADR.10	3.958	2.563	0	1	1	0.959	1.185	2	1.238	6.509	1.395

 

Table 2: QSAR statistics

	ID	# Factors	SD	R-squared	F	P	Stability	RMSE	Q-squared	Pearson-R
1	AAADR.10	1	0.4139	0.8986	141.8	2.30E-09	0.943	0.576	0.6757	0.8424
		2	0.2631	0.9616	187.7	2.42E-11	0.9102	0.626	0.6169	0.8078
		3	0.1664	0.9857	320.9	3.89E-13	0.8585	0.638	0.6021	0.7916

 

 

Figure 3: QSAR model for hydrogen bonding donor interactions

 

 

Figure 4: QSAR model for hydrophobic nonpolar interactions

 

Building QSAR models:

QSAR model was developed based on selected common hypothesis by segregating the molecules in to training set (70%) and test set (30%) based on their inhibitory activity, structural similarity and substitution pattern. PHASE provides dual options for the geometric alignment of 3dimensional structures. [21,23]. In the present study, an atom- based QSAR model was used, which is an effective tool for the study of structure–activity relationships. The atom-based QSAR, uses an algorithm which simulates the molecule as overlapping set of van der Waal’s spheres. The individual atoms or spheres are classified into one of six interactive categories. The categories are as follows- hydrogen bond donors (D), hydrogen bond acceptors (A), positive ionic (P) and negative ionic (N ) etc.

 

QSAR equation is generated by employing van der Waal’s models of the specifically aligned training set fragments in a dimension specific grid cubes. Each cube is tagged with a binary code in order to record the interaction of each type of atoms. This representation leads to specific binary patterns related to a particular sequence of atoms and useful for creating partial least-squares (PLS)models [19].

 

RESULTS AND DISCUSSIONS:

The present study aimed at elucidating the 3 dimensional structure for the substituted flavopiridol analogues as the selective inhibitors of the enzyme CDK2.The pharmacophore modeling and structure activity relationship studies were built using PHASE module of Schrödinger molecular modelling interface. The Phase generated hypothesis aids in predicting the relative binding of the test molecules at the active site of CDK2. Hence, pharmacophore-based alignment was most frequently used module for the development of 3D-QSAR model and further supports in identifying the key structural features that is required for the selective inhibition of CDK2. Further, to generate the useful pharmacophore hypothesis, a set of 23 compounds processing inhibitory activity against CDK2 were selected on the basis of the salient structural features that is crucial for binding at the active site. Four minimum sites and six maximum sites were chosen in order to arrive at obtaining a preferred combination of sites or features common to the highly active molecules. About two hundred common pharmacophore models were generated by continuous iteration method. The hypothesis AAADR10 (Figure 1) was selected based on the survival active and inactive scores (as shown in Table 1). The pharmacophore hypothesis was further used to build 3D QSAR models.

 

 

Figure 5: QSAR model for positive ionic interactions

Table 3: Predictive ability of the developed QSAR model

 

Table 4: Site score distances and angles

 

Graph 1: Actual vs Phase predicted activity of test set molecules      Graph 2: Actual vs Phase predicted activity of training  set molecules

 

The 3D QSAR model was then evaluated based on its predictive ability both in the training as well as the test sets (Tables 2, 3 and 5). The QSAR interactions like Hydrogen bond acceptor interaction is represented in Figure 6, hydrogen bond donor interactions were presented in Figure 3 and hydrophobic non polar interactions were presented in Figure 4.

 

Electron withdrawing group interactions and positive ionic interactions were presented respectively in the figures 2 and 5. Graph 1 and 2 represents the predictive ability of QSAR models in the training and test sets.

 

CONCLUSIONS:

A ligand-based pharmacophore model was built based on a series of novel flavone analogues inhibiting CDK2/Cyclin A. The pharmacophore and 3D QSAR based on atom connectivity algorithm was generated and validated using a set of 23 flavone analogues as reported in the literature. A five point pharmacophore (AAADR) model was developed for the dataset under consideration and the developed model was used to derive the predictive atom based 3D QSAR models. The selected 3D QSAR models revealed the importance of hydrogen bond acceptors, hydrogen bond donors and aromatic rings for selectively towards the target enzyme. The developed models were statistically robust (CDK2/Cyclin A, Q²- 0.6380, R² value of 0.9857, SD-0.1667, F-320.9 and Pearson coefficient value of 0.7916) The built model could be useful in predicting the structural requirements needed for the selective inhibition of CDK2/Cyclin A

 

CONFLICTS OF INTEREST:

None declared.

 

ACKNOWLEDGEMENTS:

Manipal Academy of Higher Education, Schrodinger Inc. USA. 

 

REFERENCES:

1.       J.W. Harper, P.D. Adams, Chem. Rev. 101 (2001) 2511.

2.       M. Mihara, S. Shintani, A. Kiyota, T. Matsumura, D.T.W. Wong, Int. J. Oncol. 21(2002) 95.

3.       Sielecki, T. M.; Boylan, J. F.; Benfiled, P. A.; Trainor, G. T. J. Med. Chem. 2000, 43, 1; (b) Toogood, P. I. Med. Res. Rev. 2001, 21, 487; (c) Hirai, H.; Kawanishi, N., Iwasawa, Y. Curr. Top. Med. Chem. 2005, 5, 167.

4.       Morgan, D. O. Cyclin-dependent kinases: engines, clocks, and microprocessors. Annu. ReV. Cell DeV. Biol. 1997, 13, 261-291.

5.       Sherr, C. J. Cancer cell cycles. Science 1996, 274, 1672-1677.

6.       Nurse, P.; Masui, Y.; Hartwell, L. Nat. Med. 1998, 4, 1103-1106.

7.       Fischer, P. M. Cell Cycle 2004, 3, 742-746.

8.       Senderowicz AM. Cyclin-dependent kinases as targets for cancer therapy. In Giaccone G, Schilsky R, Sondel P, editors. Cancer chemotherapy and biological response modifiers, annual 20. Elsevier Science; 2002. p. 169–88.

9.       Losiewicz MD, Carlson BA, Kaur G, Sausville EA, Worland PJ. Potent inhibition of cdc2 kinase activity by the flavonoid L86-8275. Biochem Biophys Res Commun1994;201:589–95.

10.     Carlson BA, Dubay MM, Sausville EA, Brizuela L, WorlandPJ. Flavopiridol induces G1 arrest with inhibition of cyclin-dependent kinase CDC2 and CDK4 in human breast carcinoma cells. Cancer Res 1996;56:2973–8.

11.     Carlson BA, Pearlstein RA, Naik RG, Sedlacek HH, Sausville EA, Worland PJ. Inhibition of CDK2, CDK4 and CDK7 by flavopiridol and structural anologs. In: Proceedings of the American Association of Cancer Research. 1996.

12.     Dumant, JA, HMR Report No. B/98/0189/R, B/98/0188/R,8/98/0187/R. HMR Oncology, 1996.

13.     Cartee L, Wang Z, Decker RH, Chellappan SP, Fusaro G, Hirsch KG, et al. The Cyclin-dependent kinase inhibitor (CDKI) flavopiridol disrupts phorbol 12-myristate13-acetate-induced differentiation and CDKI Expression while enhancing apoptosis in human myeloid leukemia cells. Cancer Res 2001;61(6):2583–91.

14.     Konig A, Schwartz GK, Mohammad RM, et al. The novel cyclin-dependent kinase inhibitor flavopiridol down regulates Bcl-2 and induces growth arrest and apoptosis in chronic B-cell leukemia cell lines. Blood1997;90:4307–12.

15.     Motwani M, Jung C, Sirotnak FM, She Y, Shah MA, Gonen M, et al. Augmentation of apoptosis and tumor regressions by flavopiridol in the presence of CPT-11 in HCT116 coloncancer monolayers and xenografts. Clin Cancer Res2001;7:4209–19.

16.     Carlson B, Lahusen T, Singh S, Loaiza-Perez A, Worland PJ, Pestell R, et al. Down-regulation of cyclin D1 bytranscriptional repression in MCF-7 human breast carcinoma cells induced by flavopiridol. Cancer Res1999;59(18):4634–41.

17.     Wall NR, O’Connor DS, Plescia J, Pommier Y, Altieri DC. Suppression of survivin phosphorylation on Thr34 by flavopiridol enhances tumor cell apoptosis. Cancer Res 2003;63:230–5

18.     Motwani MV, Sirotnak F, She Y, Commes T, Schwartz GK. Anovel target for modulating sensitivity to CPT-11 in colon cancer cells. Cancer Res 2002; 62:3950–5.

19.     Shah MA, Kortmansky J, Motwani M, Drobjnak M, Gonen M, Yi S, et al. A phase I clinical trial of the sequential combination of irinotecan followed by flavopiridol. Clin Cancer Res 2005; 11(10):3836–45.

20.     Shapiro GI, Patterson A, Lynch C, et al. A phase II trial of the cyclin-dependent kinase inhibitor flavopiridol in patients  with previously untreated stage IV non-small cell lung cancer. Clin Cancer Res 2001; 7(6):1590–9.

21.     Schwartz GK, Ilson D, Saltz L, O’Reilly E, Tong W, Maslak P, et al. Phase II study of the cyclin-dependent kinase inhibitor flavopiridol administered to patients with advanced gastric carcinoma. J ClinOncol2001; 19(7):1985–92.

22.     Stadler WM, Vogelzang NJ, Amato R, Sosman J, Taber D, Liebowitz D, et al. Flavopiridol, a novel cyclin-dependent kinase inhibitor, in metastatic renal cancer: a University of Chicago Phase II Consortium Study. J ClinOncol2000;18(2):371–5

23.     Burdette-Radoux S, Tozer RG, Lohmann R, et al. NCIC CT Gnphase II study of flavopiridol in patients with previously untreated metastatic malignant melanoma. In: Proc Am Soc Clin Oncol. 2002.

24.     Aklilu M, Kindler HL, Donehower RC, Mani S, Vokes EE. Phase II study of flavopiridol in patients with advanced colorectal cancer. Ann Oncol 2003;14(8):1270

25.     Lin T, Dalton JT, Wu B, et al. Flavopiridol given as a 30minute intravenous (IV) bolus followed by 4-hr continuous IV infusion results in clinical activity and tumor lysis in refractory chronic lymphocytic leukemia. Proc ASCO2004; 23:571 [abstract #6564].

26.     Schoepfer, Joseph, et al. "Structure-based design and synthesis of 2-benzylidene-benzofuran-3-ones as flavopiridol mimics." Journal of medicinal chemistry 45.9 (2002): 1741-1747.

27.     Ahn, Yu Mi, et al. "Design, synthesis, and antiproliferative and CDK2-cyclin a inhibitory activity of novel flavopiridol analogues." Bioorganic & medicinal chemistry 15. 2 (2007): 702-713.

 

 

 

 

Received on 09.01.2019          Modified on 10.02.2019

Accepted on 01.03.2019        © RJPT All right reserved