Evaluation of in vitro activity and SAR study of the novel hetarylamino-3-aryl-1 H -indazole derivatives as inhibitors of protein kinase CK2

Aim. To identify novel protein kinase CK2 inhibitors among the 5-hetarylamino-1H-indazoles. Methods. Biochemical testing was carried out with the aid of γ- 32 P ATP in vitro kinase assay. Molecular docking via the Autodock 4.2.6 program package was executed, rescoring of docking results was performed using DrugScore scoring function. Results. Among the 17 studied 5-amino-3-arylindazole derivatives 11 inhibitors of protein kinase CK2 with IC 50 in nanomolar range were identified. The most active compound has IC 50 = 2 nM. SAR study and addi-tional molecular modeling of these compounds allowed us to select prospective substituents for construction of novel compounds with improved activity and physicochemical properties. Conclusions. As a result of this work 11 nanomolar protein kinase CK2 inhibitors were de-veloped and the binding modes of these compounds with the ATP-acceptor site were proposed using molecular docking methods. The physicochemical properties and SAR of substituents of studied compounds were analyzed and 6 novel compounds were designed for further development as protein kinase CK2 inhibitors. Summarizing, 5-heterylamino-1H-indazoles are a good basis for further CK2 inhibitors development.


Introduction
Serine/threonine protein kinase CK2 was discovered among the first ones in 1954 [1]. Despite a long history of CK2 investigations, a number of its functions are not completely determined. At the same time CK2 has unique properties that differ it from the majority of protein kinases [2]. In contrast to the common protein kinases CK2 cannot be turned off by the mechanisms such as phosphorylation or dephosphorylation, second messenger binding and reversible association with regulatory subunits, but can be only ''modulated'' by means of the mechanisms, which are still not completely understood [3]. Unlike the most of other oncogenic protein kinases, the malignant behavior of CK2 is not caused by deregulation of the kinase activity at a genetic level. CK2 has the unique ability to use GTP as well as ATP as the phosphate donors [4]. The ATP binding pocket of CK2 is smaller than that of the other protein kinases, that is why the ATP competitive inhibitors with higher selectivity for CK2 could be developed. CK2 has high pleiotropicity (CK2 phosphorylates more than 500 proteins) [5] -most of the substrates have been found to be transcriptional factors (60), effectors of DNA/RNA structure (50), signaling proteins (more than 80), a limited number of metabolic enzymes and even over 40 viruses use CK2 to phosphorylate essential proteins in their life cycle [6]. Experimental results clearly showed that CK2 is involved in regulation of transcription, translation, cell proliferation, survival and apoptosis [7]. Overexpression and overactivity of CK2 are associated with the development of neurodegenerative, inflammatory, cardio-vascular dis-eases, virus infections, and all the cancers having been examined [8,9]. Therefore, smallmolecular inhibitors of CK2 would be important compounds for the development of clinical agents.
In the previous work we have described chemical synthesis of the novel 5-hetarylamino-1H-indazole derivatives as prominent inhibitors of human protein kinase CK2 [10]. Here we are representing measurements of their in vitro activity and SAR study as well as molecular docking simulation.

Molecular docking
Flexible docking was conducted using the program Autodock 4.2.6 [11]. The 3D structure of human protein kinase CK2 was obtained from the Brookhaven Protein Data Bank (PDB ID: 3NSZ and 4UBA) [12,13].
Water molecules, ions, and ligand were removed from the PDB file of the receptor. Ligands were prepared by Vega ZZ (command line) [14] and MGL Tools 1.5.6 [11]. To carry out a calculation with the aid of the Autodock program, the incoming formats of receptor and ligands data were converted into PDBQTformat with Vega ZZ in the AUTODOCK force field. This format contains the coordinates of the atoms and partial charges. Hydrogen atoms were removed from nonpolar adenine parameters: translation step -2 Å, quaternion step -50, torsion step -50. Torsional degrees of freedom and coefficient were 2 and 0.274, respectively. Cluster tolerance -2 Å. External grid energy -1,000, max initial energy -0, max number of retries -10,000. Number of individuals in the population -300, the maximum number of energy evaluations -850,000, the maximum number of generations -27,000, number of top individuals, which survived to the next generation -1, rate of gene mutation -0.02, rate of crossover -0.8, mode of crossover -arithmetic. Alpha parameter of Cauchy distribution was 0, Beta parameter Cauchy distribution -1. The number of iterations of the Lamarckian genetic algorithm was 50 for each ligand.
Visual analysis of the best-scored complexes was performed using Discovery Studio

Biological evaluation
Compounds were tested using in vitro kinase assay [15]. Each test was carried out in triplicate in 30 μL reaction volume, which contains 6 μg of peptide substrate RRRDDDSDDD (New England Biolabs); 10 units of recombinant human CK2 holoenzyme (New England Biolabs); 50 μM ATP and 0.05-0.1 μCi γ-labeled 32 P ATP; CK2 buffer (20 mM Tris-HCl, pH 7.5; 50 mM KCl; 10 mM MgCl 2 ) and inhibitor in varying concentrations. The incubation time was 20 min at 30 °C. The reaction was stopped by adding an equal volume of 10 % orthophosphoric acid and the reaction mixture was loaded onto 20-mm discs of phosphocellulose paper (Whatman). Disks were washed three times with 1 % orthophosphoric acid solution, air-dried at room temperature and counted by the Cherenkov method in a beta-counter (LKB). As a negative control, an equal volume of dimethyl sulfoxide (DMSO) was added to the reaction mixture. Inhibition percentage was calculated as the ratio of substrate-incorporated radioactivity in the presence of inhibitor to the radioactivity incorporated in control reactions, i.e., in the absence of inhibitor. Serial dilutions of inhibitor stock solution were used to determine its IC 50 concentration. The IC 50 values represent means of triplicate experiments with SEM never exceeding 15 %.

Results and Discussion
Earlier we have found new promising CK2 inhibitors with IC 50 values in a range 0.007-0.1 μM in in vitro assay with radiolabeled ATP [15] among the 4-substituted quinazoline derivatives of 5-amino-3-arylindazoles 1-3 [20]. We had left 3-(3,4-dichlorophenyl) substituent of the most active compound 1 unmodified and have focused on the variation of heterocyclic residue at the 5-position of indazole moiety [10].
A complex of compound 1 with a CK2 ATPbinding site is shown in Figure 1. An indazole heterocyclic moiety is involved in the hydrophobic interactions with amino acid residues Val66, Ile95, Phe113, Val116, Met163 and Ile174 in the adenine-binding region. Also, the indazole moiety forms a hydrogen bond with Val116 amino acid residue of the CK2 hinge region. The 3,4-dichlorophenyl substituent is directed towards the exit of ATP-binding pocket. It forms hydrophobic interaction with Leu45 and Met163. The (2-trifluoromethyl) quinazoline moiety lies inside the hydrophobic pocket I. It has hydrophobic interaction with Phe113, Val53, Lys68 and Ile174. Additionally, 2-trifluoromethylquinazoline forms hydrogen bonds with Lys68 and Asp175.
Chemical structures, IC 50 , calculated LogP and LipE (CLogP and CLipE) values of the 14 new heterocyclic derivatives of substituted 5-amino-3-(3,4-dichlorophenyl)indazole are represented in the Table 1. In general, 5-amino-3-aryl indazole derivatives show high inhibitory activity toward CK2 but mostly have had extremely high lipophilicity (LogP>5). In this case, to estimate the efficiency of studied compounds as CK2 inhibitors, it was used a special composite parameter that linked inhibitory activity and lipophilicity parameter. This parameter is lipophilic efficiency or LipE [17].
LipE is the most "correct" [18] composed parameter, and it is determined using the following formula: CLipE = pIC 50 -ClogP. The structure and IC 50 values of the most effective studied inhibitors are shown in Table 2.
According to in vitro tests, 11 compounds (including 1-3 having been synthesized earlier) have submicromolar IC 50 values ranging from 0.002 µM (4) to 0.8 µM (6). It turns out that negative steric influence of the bulky substituents at 2 nd position of quinazoline heterocycle on the compounds inhibitory activity is not as crucial as that of the substituents at the 6 th position. Small methyl group at the 6 th quinazoline's position of compound 8 causes decrease of IC 50 by two orders of magnitude   is more than one thousand times higher than the IC 50 of 4. However, the most surprisingly poor results in this testing were shown by quinoline derivatives 11 and 12 as well as thienopyrimidine thieno[2,3-d]pyrimidine derivative 13. 11 has almost the same geometry as 4 but shows no activity towards CK2. This might be a result of a combination of the lower basicity of quinoline heterocycle compared to quinazoline one and lower water solubility of 11. Difficulties with water solubility also may be a reason of significant variation in IC 50 of more polar and hydrophilic pyrazolo [3,4-d] pyrimidine derivatives 14-17 which, according to IC 50 compound 16 in 0.002 µM, can be active enough.
Regarging the lipophilicity -the analysis of FDA approved protein kinase inhibitors showed that average LipE value is 4.99 with a range from 2 (vandetanib) to 8.5 (tofacitinib) [19]. Our best compounds 2, 4 and 16 have values of CLipE 2.3, 2.7 and 3, respectively (Table 1). It is significantly lower than CLipE values of known promising CK2 inhibitors BFO13, FLC26, FNH79 and CX-4945-4.94, 3.85, 4.48 and 4.84, respectively. Examined 5-amino-3-arylindazole derivatives have strong inhibitory potential toward CK2, but their physicochemical properties should be thoroughly optimized. In this case, the new 5-amino-3-arylindazole derivatives with the same binding mode and activities but with better CLipE values should be constructed.
The CLogP values of the most effective substituents R 1 and R 2 were calculated to determine the direction of optimization (see Figure 2). R 1 substituent [3-(trifluoromethyl)-1-quinazolyl] and R 2 substituent 3,4-dichlorophenyl had the highest CLogP values -2.6 and 3.6, respectively. These substituents were not taken into account for the design of the new 5-amino-3-arylindazoles.
Based on the SAR study, the results of molecular docking and calculated LogP values of each substituent, six new 5-amino-3-arylindazole derivatives were designed as possible effective human protein kinase CK2 inhibitors.
All new six compounds comply with Lipinski's rules. The structure and CLogP values of these compounds are represented in Figure 3.
The protein part of a crystal structure of the protein kinase CK2 complex with inhibi-   [20] (see Table 2).
To avoid the influence of different molecular weights on the results of scoring function, the drug score values were approximated to molecular weights and numbers of heavy atom s of compounds. It is required for more detailed differentiation of its possible inhibitory acti vity. Molecular docking showed that binding modes of compounds 18-23 (see Figure 4) are similar to the binding mode of FLC26 proved by crystallography. R 1 substituents are oriented to the hydrophobic pocket I and R 2 -to the exit of the ATP-binding pocket. Indazole heterocycle is located in an adenine-binding pocket and forms hydrogen bonds with hinge amino acid residues.
Among known inhibitors the compounds 1, CX-4945, compound 4, and We anticipate that designed 5-amino-3-arylindazole derivatives will have lipophilic efficiency comparable with the best known CK2 inhibitors.