QSAR Model of Indeno[1,2-b]indole Derivatives and Identification of N-isopentyl-2-methyl-4,9-dioxo-4,9-Dihydronaphtho[2,3-b]furan-3-carboxamide as a Potent CK2 Inhibitor

Casein kinase II (CK2) is an intensively studied enzyme, involved in different diseases, cancer in particular. Different scaffolds were used to develop inhibitors of this enzyme. Here, we report on the synthesis and biological evaluation of twenty phenolic, ketonic, and para-quinonic indeno[1,2-b]indole derivatives as CK2 inhibitors. The most active compounds were 5-isopropyl-1-methyl-5,6,7,8-tetrahydroindeno[1,2-b]indole-9,10-dione 4h and 1,3-dibromo-5-isopropyl-5,6,7,8-tetrahydroindeno[1,2-b]indole-9,10-dione 4w with identical IC50 values of 0.11 µM. Furthermore, the development of a QSAR model based on the structure of indeno[1,2-b]indoles was performed. This model was used to predict the activity of 25 compounds with naphtho[2,3-b]furan-4,9-dione derivatives, which were previously predicted as CK2 inhibitors via a molecular modeling approach. The activities of four naphtho[2,3-b]furan-4,9-dione derivatives were determined in vitro and one of them (N-isopentyl-2-methyl-4,9-dioxo-4,9-dihydronaphtho[2,3-b]furan-3-carboxamide) turned out to inhibit CK2 with an IC50 value of 2.33 µM. All four candidates were able to reduce the cell viability by more than 60% after 24 h of incubation using 10 µM.


Introduction
Human CK2 (casein kinase II) is an S/T protein kinase, able to phosphorylate the bovine casein in vitro. By this means, it was one of the first phosphorylating enzymes discovered (in 1954) [1]. CK2 is a heterotetrameric enzyme, composed of two catalytic subunits (α and/or α ) and two regulatory subunits (β). It is involved in different intracellular signaling pathways and is also known to be an                 In the next step, we used our validated QSAR model to predict the pIC50 values of the 25 naphtho[2,3-b]furan-4,9-dione-based compounds that were selected earlier using in silico screening as possible hits for the target enzyme [24]. As it is clear from Table 3, many of the compounds were predicted to be highly active inhibitors. Most of them fit well in the active site of the enzyme, as confirmed by the docking study and the S scores obtained for all hits ( Table 3). The above QSAR prediction and the docking results encouraged us to test the activity of these compounds in vitro. Unfortunately only four compounds were available to us and were tested for their in vitro CK2 activity, namely: 01893208: and 01236034: 2-methyl-3(morpholine-4carbonyl)naphto[2,3-b] furan-4,9-dione. The compounds were tested for inhibition with purified human protein kinase CK2 expressed in Escherichia coli, as described before [36], using a capillary electrophoresis-based activity assay. The in vitro results are depicted in Table 4. Among the tested compounds, only compound 00082235 showed a good CK2 inhibitory activity (75% at a concentration of 10 µM) with an IC50 value of 2.33 ± 0.25 µM, while the other three compounds were able to inhibit the target enzyme by around 35-45% at a concentration of 10 µM. The IC50 value was determined in In the next step, we used our validated QSAR model to predict the pIC 50 values of the 25 naphtho[2,3-b]furan-4,9-dione-based compounds that were selected earlier using in silico screening as possible hits for the target enzyme [24]. As it is clear from Table 3, many of the compounds were predicted to be highly active inhibitors. Most of them fit well in the active site of the enzyme, as confirmed by the docking study and the S scores obtained for all hits ( Table 3). The above QSAR prediction and the docking results encouraged us to test the activity of these compounds in vitro. Unfortunately only four compounds were available to us and were tested for their in vitro CK2 activity, namely: 01893208: furan-3-carboxamide, and 01236034: 2-methyl-3(morpholine-4-carbonyl)naphto[2,3-b] furan-4,9-dione. The compounds were tested for inhibition with purified human protein kinase CK2 expressed in Escherichia coli, as described before [36], using a capillary electrophoresis-based activity assay. The in vitro results are depicted in Table 4. Among the tested compounds, only compound 00082235 showed a good CK2 inhibitory activity (75% at a concentration of 10 µM) with an IC 50 value of 2.33 ± 0.25 µM, while the other three compounds were able to inhibit the target enzyme by around 35-45% at a concentration of 10 µM. The IC 50 value was determined in three independent biological replicates with similar results (2.046, 2.527, and 2.423 µM). The mean value was determined. It is important to mention that compounds 01893208 and 00082235 were included recently in a patent claimed by Yang et al. and were evaluated for the treatment of proliferative and infectious diseases [37].                   10.0672 a The percentage of inhibition of CK2 activity was determined for each compound at a fixed concentration of 10 µM. b The IC50 value was determined only for compounds with an initial inhibitory activity above 50% at a concentration of 10 µM. Here, the concentration was varied to precisely determine the IC50 value. c A rough estimation was obtained for less potent compounds from the experimental inhibition produced at 10 µM.
The effect of the four naphtho[2,3-b]furan-4,9-dione derivatives on cell viability was evaluated using an MTT assay with MCF-7 cells (human breast adenocarcinoma cell line). As it is shown in Figure 2, compounds 01893208, 37867960, and 00082235 were able to reduce the cell viability by more than 90% after 24 h and 48 h of incubation using 10 µM, while compounds 01893208 and 37867960 were able to reduce the cell viability by around 90% after 24 h and 48 h of incubation using 1 µM. In contrast, compound 01236034 was less cytotoxic and was able to reduce the cell viability to 60% after 48 h of incubation using 10 µM. These results indicate that the tested compounds showed a good 45 (11000) c (0.08) 10.0672 a The percentage of inhibition of CK2 activity was determined for each compound at a fixed concentration of 10 µM. b The IC 50 value was determined only for compounds with an initial inhibitory activity above 50% at a concentration of 10 µM. Here, the concentration was varied to precisely determine the IC 50 value. c A rough estimation was obtained for less potent compounds from the experimental inhibition produced at 10 µM. The effect of the four naphtho [2,3-b]furan-4,9-dione derivatives on cell viability was evaluated using an MTT assay with MCF-7 cells (human breast adenocarcinoma cell line). As it is shown in Figure 2, compounds 01893208, 37867960, and 00082235 were able to reduce the cell viability by more than 90% after 24 h and 48 h of incubation using 10 µM, while compounds 01893208 and 37867960 were able to reduce the cell viability by around 90% after 24 h and 48 h of incubation using 1 µM. In contrast, compound 01236034 was less cytotoxic and was able to reduce the cell viability to 60% after 48 h of incubation using 10 µM. These results indicate that the tested compounds showed a good effect on cell viability. It is important to note that some compounds with a naphtho [2,3-b]furan-4,9-dione backbone are known to have anticancer effects and are highly cytotoxic [38,39], and other natural naphtho [2,3-b]furan-4,9-dione derivatives, which are found in the lapacho (Tabebuia) tree, are known to have different biological activities, such as antibacterial, antifungal, anti-inflammatory, and antitumor activities [40]. a The percentage of inhibition of CK2 activity was determined for each compound at a fixed concentration of 10 µM. b The IC50 value was determined only for compounds with an initial inhibitory activity above 50% at a concentration of 10 µM. Here, the concentration was varied to precisely determine the IC50 value. c A rough estimation was obtained for less potent compounds from the experimental inhibition produced at 10 µM.
The effect of the four naphtho [2,3-b]furan-4,9-dione derivatives on cell viability was evaluated using an MTT assay with MCF-7 cells (human breast adenocarcinoma cell line). As it is shown in Figure 2, compounds 01893208, 37867960, and 00082235 were able to reduce the cell viability by more than 90% after 24 h and 48 h of incubation using 10 µM, while compounds 01893208 and 37867960 were able to reduce the cell viability by around 90% after 24 h and 48 h of incubation using 1 µM. In contrast, compound 01236034 was less cytotoxic and was able to reduce the cell viability to 60% after 48 h of incubation using 10 µM. These results indicate that the tested compounds showed a good effect on cell viability. It is important to note that some compounds with a naphtho [2,3-b]furan-4,9dione backbone are known to have anticancer effects and are highly cytotoxic [38,39], and other natural naphtho [2,3-b]furan-4,9-dione derivatives, which are found in the lapacho (Tabebuia) tree, are known to have different biological activities, such as antibacterial, antifungal, anti-inflammatory, and antitumor activities [40]. All naphtho [2,3-b]furan-4,9-dione compounds in this study were docked in the ATP binding site of the crystal structure of CK2, as mentioned above using PDB ID: 3C13 from the Protein Data Bank (PDB) [41], having a resolution of 1.95 Å. A conformational search for the 25 selected compounds was carried out using MOE, and all the resulting 219 conformations were used for the docking study. The conformations of all structures were docked in human CK2 enzyme using MOE and sorted according to their S score (energy-based scoring method implemented in MOE) to rank the best ligand in terms of the orientation and binding to the active site. The four tested compounds were selected and visual 2D and 3D inspections were carried out to exclude "false positives" due to assumptions and shortcomings in the docking methods and scoring function [42]. The active site was analyzed to determine the protein-ligand interaction to show exactly which atoms could interact. Figure 3 shows the 3D and 2D interactions of the four tested compounds and it is clear that they fit well in the ATP active site of CK2. The four docked compounds were inserted into the narrow ATP binding site of CK2 ( Figure 3) and all of them were able to create a π-π interaction between the aromatic ring of the naphtho [2,3-b]furan-4,9-dione and the side-chain residues of Ile174. Compound 01893208 was also able to create direct bonding with Asp120 and Glu123. Compound 37867960 and compound 00082235 showed direct bonding with Lys68 and via a water molecule with Glu81. Compound 01236034 had direct bonding with Val53 and via a water molecule with Glu81 as well. Emodin in the crystal structure created a π-π interaction with Phe113 and direct bonding with Lys68.
to their S score (energy-based scoring method implemented in MOE) to rank the best ligand in terms of the orientation and binding to the active site. The four tested compounds were selected and visual 2D and 3D inspections were carried out to exclude "false positives" due to assumptions and shortcomings in the docking methods and scoring function [42]. The active site was analyzed to determine the protein-ligand interaction to show exactly which atoms could interact. Figure 3 shows the 3D and 2D interactions of the four tested compounds and it is clear that they fit well in the ATP active site of CK2. The four docked compounds were inserted into the narrow ATP binding site of CK2 ( Figure 3) and all of them were able to create a π-π interaction between the aromatic ring of the naphtho [2,3-b]furan-4,9-dione and the side-chain residues of Ile174. Compound 01893208 was also able to create direct bonding with Asp120 and Glu123. Compound 37867960 and compound 00082235 showed direct bonding with Lys68 and via a water molecule with Glu81. Compound 01236034 had direct bonding with Val53 and via a water molecule with Glu81 as well. Emodin in the crystal structure created a π-π interaction with Phe113 and direct bonding with Lys68.
(1)  Several studies have been performed to develop an active CK2 inhibitor using different in silico techniques [23]; however, few "ligand-based drug design" (LBDD) techniques were able to put forward a successful candidate [23]. Our database of indeno [1,2-b]indoles was not only used to find a list of candidates via a pharmacophore approach, but also exploited to develop a new reliable QSAR model, which we used to predict the activity of these compounds. We demonstrated that some of the tested naphtho[2,3-b]furan-4,9-dione compounds were active using both in vitro evaluation and cellular assays. Nevertheless, we were not able to correlate our predicted IC50 values to the tested IC50 values of naphtho[2,3-b]furan-4,9-dione compounds since only four compounds were available to us.
As a matter of fact, the moderate inhibitory activity of some tested naphtho[2,3-b]furan-4,9-dione derivatives was not a surprise for us, as a variation between predicted and tested inhibition is expected in such QSAR studies. This was also the case in some compounds in the test set, where the r 2 for the test set was 0.77. Unfortunately, it was not possible to correlate the predicted and tested values for all naphtho[2,3-b]furan-4,9-dione derivatives presented in this study, since only four of them were tested. It is important to emphasize that the aim of this work was to describe the inhibitory activity of 20 novel indeno [1,2-b]indoles and to show that compounds with naphtho[2,3-b]furan-4,9dione backbone were appropriate for developing CK2 inhibitors, which was proved by compound 00082235. The future step will be optimizing this compound by synthesizing a series of compounds with different substituents on a naphtho[2,3-b]furan-4,9-dione backbone and performing a SAR study with the aim of finding a highly active CK2 inhibitor.

Computational Study
As a matter of fact, the moderate inhibitory activity of some tested naphtho[2,3-b]furan-4,9-dione derivatives was not a surprise for us, as a variation between predicted and tested inhibition is expected in such QSAR studies. This was also the case in some compounds in the test set, where the r 2 for the test set was 0.77. Unfortunately, it was not possible to correlate the predicted and tested values for all naphtho[2,3-b]furan-4,9-dione derivatives presented in this study, since only four of them were tested. It is important to emphasize that the aim of this work was to describe the inhibitory activity of 20 novel indeno [1,2-b]indoles and to show that compounds with naphtho[2,3-b]furan-4,9-dione backbone were appropriate for developing CK2 inhibitors, which was proved by compound 00082235. The future step will be optimizing this compound by synthesizing a series of compounds with different substituents on a naphtho[2,3-b]furan-4,9-dione backbone and performing a SAR study with the aim of finding a highly active CK2 inhibitor.

Computational Study
Molecular Operating Environment software package (MOE, 2016.01, Chemical Computing Group, Montreal, QC, Canada) was used to perform this study [29], running on an Intel Core, i5-6500CPU, 3.20 GHz processor.

Data Set for QSAR
The data for the QSAR study (IC 50 values) was created by using the above-mentioned compounds. The compounds were divided into a training set with 20 compounds having IC 50 values between 25 and 4100 nM, and a test set with 10 compounds having IC 50 values between 140 and 4160 nM. The compounds were divided into training and test sets randomly but maintained the same range of biological activity. All compounds were sketched using a building function integrated in MOE, converted to 3D, and their energy was minimized using an MMFF94 force field. Both the training and test sets were stored as a dataset in mdb files, and the pIC 50 value for each compound was calculated and added manually.

Molecular Descriptors
MOE offers a wide range of 2D and 3D molecular descriptors in order to calculate the molecular properties of compounds. In this work, all 2D descriptors (total 206) that MOE offers were calculated for all compounds in the training set. Then, the contingency (which is a statistical application) function integrated in the software was used to select the appropriate descriptors for this set of compounds. In the second step, we used the function "relative importance of descriptors," and all descriptors with less than 0.1 were removed.

QSAR Study
The QSAR model was developed by selecting the activity as "dependent variable" and the descriptors as model fields. After performing the regression analysis for the training set, the root mean square error (RMSE) and r 2 values of the fit were calculated and the model was saved as a QSAR fit model, which was then used for the prediction of activities of the test data set. The QSAR fit was then used for the validation and cross-validation. Plotting the tested pIC 50 values (X-axis) versus the predicted (PRED) values (Y-axis) was performed to assess the predictive ability of the model. The correlation coefficient (r 2 ) was determined for the test set using the QSAR fit model and was used to predict the activity of the naphtho[2,3-b]furan-4,9-dione derivatives as well. The correlation coefficient r 2 lies between 0 and 1, where 1 corresponds to an ideal fit.

Crystal Structure from PDB
Three dimensional structure of the CK2 complex with emodin was obtained from the Protein Data Bank (PDB) using PDB ID: (3C13) with a resolution of 1.95 Å [41]. The structure was optimized by using the QuickPrep function implemented in the MOE software. Then, water molecules were removed from the structure and 3D protonation was done to change the state into an ionization level. In the second step, energy minimization was performed using default parameters, where the force field was Amber 10.

Database Generation
The selected compounds from the ZINC database [23] (naphtho [2,3-b]furan-4,9-dione derivatives) were rebuilt with the MOE building option implemented in the software. The compounds were optimized by adding hydrogen atoms using the option found in the MOE software. The energy of the compounds was minimized using the following parameters gradient: 0.05, Force Field: MMFF94X, Chiral constraint, and Current Geometry. The conformation methodology was used to develop low energy conformations for each compound, applying the LowModMD method with an RMS gradient of 0.05, and all other parameters were used as their default. All the compounds and their conformations were saved in an mdb database and later employed for docking studies.

Docking Study
The docking of the naphtho[2,3-b]furan-4,9-dione derivatives into the active site of the CK2 enzyme (3C13) was achieved using MOE-Dock implemented in MOE. The docking parameters were set as Rescoring 1: London dG, Placement: triangle matcher, Retain 30, Refinement Force field, and Rescoring 2: GBVI/WSA dG, Retain 30. The docking part of MOE can give the correct conformation of the ligand to obtain a minimum energy structure. The top conformation for each compound was selected based on the S score, and visual inspection in 2D and 3D was carried out using MOE. Prior to docking, the initial ligand from the complex structure was extracted. For the scoring function, lower scores indicated more favorable poses. The unit for the scoring function is kcal/mol, and the S score refers to the final score, which was the score of the last stage that was not set to None. The Lig X function in MOE was used for conducting interactive ligand modification and energy minimization in the active site of the receptor.

Capillary Electrophoresis-Based Assay for Testing the Inhibitors of the Human CK2
The available naphtho[2,3-b]furan-4,9-diones were tested for their inhibitory activity toward the human CK2 holoenzyme following the procedure described earlier [36]. Briefly, the synthetic peptide RRRDDDSDDD was used as the substrate, which is reported to be most efficiently phosphorylated by CK2. The purity of the CK2 holoenzyme was greater than 99%. For initial testing, inhibition was determined relative to the controls at inhibitor concentrations of 10 µM in DMSO as a solvent. The reaction with the pure solvent without an inhibitor was used as a positive control and set to 100% inhibition. Reactions without CK2 were used as a negative control and were taken as 0% inhibition. IC 50 values were determined by measuring the CK2 inhibition at eight different concentrations ranging from 0.001 to 100 µM at appropriate intervals and calculated from the resulting dose-response curve [43].

Cell Viability Assay
The effect of CK2 inhibitors on the viability of MCF-7 cells was evaluated using an MTT assay [44]. MCF-7 breast cancer cells (kindly provided by the Department of Clinical Radiology of the University Hospital Muenster, Germany) were cultured in RPMI 1640 medium containing GlutaMax (Life Technologies, Darmstadt, Germany) and 10% fetal calf serum. The MTT assay was performed in 96-well plates. Cells were seeded at a density of 1 × 10 5 cells per well, then incubated for 24 or 48 h at 37 • C in a humidified atmosphere (5% CO 2 ). After overnight incubation, the seeding medium was removed and replaced with fresh medium containing the inhibitor at 0.01, 0.1, 1, and 10 µM. DMSO at a final concentration of 1% served as a control. Afterward, the MTT reagent (Sigma Aldrich, Darmstadt, Germany) was added at a final concentration of 0.5 mg/mL. After incubation for 2 h at 37 • C, the medium was discarded and 200 µL DMSO was added for the solubilization of the formazan.
After mixing, the absorption was determined at 570 nm with a reference wavelength of 630 nm using a microplate reader. CK2 inhibitors were tested in triplicate, and the experiments were repeated three times.

Conclusions
In this work, we described a set of indeno [1,2-b]indoles as CK2 inhibitors, together with their in vitro activities, and further presented a detailed description of a QSAR study built from these indeno [1,2-b]indoles. The obtained QSAR model was exploited to predict the activity of a group of naphtho [2,3-b]furan-4,9-diones that have never been tested as CK2 inhibitors before. Through our current results, we can confirm that a naphtho [2,3-b]furan-4,9-dione backbone is an appropriate skeleton for the development of novel inhibitors for this kinase target. More compounds having this backbone with different functional groups should be designed with the aim of finding highly active CK2 inhibitors.
Supplementary Materials: File S1. Chemical synthesis of all new indeno [1,2-b]indoles. File S2. The NMR and MS data for all new indeno [1,2-b]indoles. Table S1: Chemical structures of the indeno [1,2-b]indoles used for the training set and the test set in the SMILES format. Tables S2 and S3: calculated physicochemical parameters of the training set and the test set.