Probing the Inhibition of Microtubule Affinity Regulating Kinase 4 by N-Substituted Acridones

Microtubule affinity regulating kinase 4 (MARK4) becomes a unique anti-cancer drug target as its overexpression is responsible for different types of cancers. In quest of novel, effective MARK4 inhibitors, some acridone derivatives were synthesized, characterized and evaluated for inhibitory activity against human MARK4. Among all the synthesized compounds, three (7b, 7d and 7f) were found to have better binding affinity and enzyme inhibition activity in µM range as shown by fluorescence binding, ITC and kinase assays. Here we used functional assays of selected potential lead molecules with commercially available panel of 26 kinases of same family. A distinctive kinase selectivity profile was observed for each compound. The selective compounds were identified with submicromolar cellular activity against MARK4. Furthermore, in vitro antitumor evaluation against cancerous cells (MCF-7 and HepG2) revealed that compounds 7b, 7d and 7f inhibit cell proliferation and predominantly induce apoptosis in MCF-7 cells, with IC50 values of 5.2 ± 1.2 μM, 6.3 ± 1.2 μM, and 5.8 ± 1.4 μM respectively. In addition, these compounds significantly upsurge the oxidative stress in cancerous cells. Our observations support our approach for the synthesis of effective inhibitors against MARK4 that can be taken forward for the development of novel anticancer molecules targeting MARK4.

Selected compounds show significant binding with MARK4. Molecular docking was carried out to predict existing molecular interactions between the synthesized acridones and amino acid residues of MARK4 utilizing Autodock, Autodock Vina TM combined with PyRx TM for workflow management 34,35 . The crystal structure of MARK4 with PDB code 5ES1 in the absence of 5RC module was selected and acridones were docked into the catalytic domain of the kinase. Data from docking experiments are presented in Table S1 and Figs 2, S1-S3. The MARK4-acridone complexes were stabilized by various non-covalent interactions offered by the residues present in the active site cavity of MARK4 (Figs 2, S1-S3). Further analysis of the docking results revealed that all the selected compounds showed binding energy ranging between -9.2 kcal/mol to -9.8 kcal/mol (Table S1). Our first aim was to verify that the 5 occupies the catalytic center and second to design derivatives with increased interactions with the activation loop of MARK4 (e.g. DFG motif). All compounds 5 and 7a-j form close interactions to active site residues of MARK4 including one hydrogen bond and several van der Waals interactions (Figs 2, S1-S3). The 2-methylacridone moiety off almost all tested compounds interacts by two different binding modes with the following amino acid residues: Ile62, Val70, Ala83, Val116, Met132, Glu133, Tyr134, Gly138, Glu139 and Leu185. In Fig. S4 two representative binding modes of acridones are illustrated: one pose with the methyl group to extend towards Ala83, Val116 and Met132 (compounds 5, 7a, 7b, 7d, 7e and 7h) and one pose with the same group to display interactions with Ile62 and Tyr134 (compounds 7c, 7f, 7g, 7i and 7j).
Notably, the carbonyl group of acridone forms one hydrogen bond with the amino group of Ala135 in the hinge region of kinase. Molecular docking studies revealed that compounds 7b, 7f and 7h have the most promising binding energies (−9.5-−9.8 kcal/mol). Docking analysis illustrated that from these five analogues, all but 7f get the same binding pose with the acridone's methyl group interacting with Ala83, Val116 and Met132. The reliability of the applied docking protocol was assessed by re-docking the pyrazolopyrimidine inhibitor 5RC into the active site of the MARK4 (Fig. S5).

Fluorescence binding studies.
To determine the binding affinities of the synthesized acridones with MARK4, we used fluorescence emission spectra measurements. Protein sample was excited at 280 nm and emission was measured in 300-400 nm range with increasing concentrations of each acridone derivative (0-100 µM). A significant decrease in fluorescence intensity of MARK4 with increasing concentrations of each compound was fitted in the modified Stern-Volmer equation to calculate the binding constant, K a and the number of binding sites per protein molecule (n). All the synthesized acridones have been screened ( Supplementary Fig. S6, S7 and Table S3) and those with the highest affinity have been selected (compounds 5, 7b, 7d, 7f and 7h) for further evaluation (Fig. 3). The binding constants of compound 5, 7b, 7d, 7f and 7h were estimated as 6.3 × 10 4 M −1 , 1.0 × 10 6 M −1 , 1.0 × 10 6 M −1 , 1.9 × 10 6 M −1 , and 3.1 × 10 5 M −1 , respectively (Fig. 3). Comparison of the measured binding affinities shows that 7b, 7d and 7 f exhibit the highest affinity towards MARK4. Additionally, it was found that each compound binds to a single binding site on MARK4.
Enzyme inhibition assay and structure activity relationships. To evaluate the inhibition activity of synthesized acridone derivatives against recombinant MARK4, ATPase enzyme inhibition assay was performed (Table S3). Ester derivative 5 was active against MARK4 with an IC 50 value of 14.12 ± 1.02 μM (Fig. 4). Our studies were initiated by replacing the ester group of 5 with sterically and electronically diverse amines affording the corresponding amides 7a-j, in order to establish favorable interactions with residues Ala135, Val70, Asp196 and/or Glu182. The most potent compounds 7b and 7d in this screening carry a benzylamine (IC 50 value of 1.80 ± 0.04 μM) and p-MeO-benzylamine (IC 50 value of 2.20 ± 0.05 μM), respectively, whereas the methyl substituted benzylamide 7c was much weaker inhibitor. These results are in accordance with docking studies, which show that the benzylamino moiety of 7b and 7d forms favorable interactions with residues Ala135, Val70, Ala68 and Gly63. Replacement of the benzylamine with morpholine resulted in a significant loss of potency, while compounds containing substituted phenylamines displayed variable activities. Compound 7e was less potent than 7b, while, introducing -OCOMe-, -COMe or -OMe as the C4-substituent of the phenylamino ring yielded compounds 7g, 7i and 7j with lower activities and IC 50 values higher than 20 μM. It should be noted that compound 7h bearing an ortho-MeO-phenylamine was more active (IC 50 = 12.15 ± 1.10 μM) than the para-MeO-phenylamino derivative 7g (IC 50 > 20.0 μM). In this case, the docking studies indicate a better fit of 7h compared with 7g in the active site of MARK4 with a binding energy of −9.5 kcal/mol and a stronger predicted hydrogen bond with Ala135 (3.06 Å for 7h, 3.19 Å for 7g). A p-methyl-phenylamine was favorable for the activity and 7f inhibited MARK4 with an IC 50 value of 4.5 ± 0.52 μM. It is interesting to note that the results of fluorescence binding are consistent with that of enzyme activity results suggesting that the high binding affinity compounds significantly inhibit the enzyme activity of MARK4.
Isothermal titration calorimetry measurements. Molecular docking, enzyme inhibition and fluorescence binding studies showed that compounds 5, 7b, 7d, 7f and 7h are interacting to MARK4 and inhibit its activity. In order to measure the actual binding affinity and stoichiometry of selected compounds with the recombinant MARK4, isothermal titration calorimetry (ITC) has been carried out. A typical isotherm of ITC obtained after the titration of MARK4 with compound 7b, 7d and 7f was shown in Fig. 5. Negative heat impulses in the upper panel of each isotherm indicate exothermic nature of binding. The amount of heat produced as a result of each injection helps to deliver the molar ratio of studied compound to that of MARK4. Thermodynamic parameters associated with binding of MARK4-compounds (Ka, binding constant and ΔH, enthalpy change) are shown in Table 1. These results were obtained from single-site fitting model. We also tried ITC with compound 5 and 7h, but at lower stoichiometry these compounds do not show significant binding and at high concentration they precipitate with MARK4. Overall inference from docking, fluorescence, ITC and enzyme inhibition suggested that the compound 5, 7b, 7d, 7f and 7h bind with MARK4 and this binding is responsible for the inhibition of enzyme activity.
Cell proliferation assay. Initially, all synthesized acridone derivatives were evaluated for their cytotoxicity potential on MCF-7, HepG2 and HEK293 cell lines by MTT assay. These synthesized acridone derivatives were screened in the concentration range of 0-200 µM, for 24 and 48 h. The results show that at higher concentrations all compounds inhibit the proliferation of MCF-7 as well as HepG2 cells (Table S1). Consistent with earlier studies, cell proliferation studies also showed that compounds 5, 7b, 7d, 7f and 7h provoked superior toxicity in a concentration dependent manner on MCF-7 cells. It was also interesting that in the studied submicromolar concentration range these compounds don't show considerable cytotoxicity towards HEK293 cell lines. The IC 50 values for compound 5, 7b, 7d, 7f and 7h were found to be 9.4 ± 1.0, 5. also studied on HEK293 cells, it was observed that more than 85% of embryonic kidney cells were viable even after 72 h of treatment. Cytotoxicity results clearly suggested that the tested compounds are non-toxic to normal cells and specifically bear toxicity for cancerous cells. Thus, compounds 5, 7b, 7d, 7f and 7h were taken for further cell based studies such as apoptosis and reactive oxygen species (ROS) production. Apoptosis assay. Evasion of apoptosis is a striking hallmark of cancerous cells, it is an essential process that regulates abnormal growth of cells, but compromised signaling helps the cancerous cells to escape apoptosis 36 . MARK4 overexpression also supports the growth and evasion of cancerous cells 4 . Thus, the probability of apoptosis induction by inhibiting MARK4 was studied. The MCF-7 cells were serum starved and treated with IC 50 concentration of each acridone derivative for 24 h and subsequently annexin-V staining was used to assess the apoptotic potential of these compounds. Stained cells were analyzed by flow cytometry and found that the treatment with compounds 5, 7b, 7d, 7f and 7h considerably induces apoptosis in the MCF-7 cells (Fig. 6A). Analysis of the results suggested that treatment with compounds 5, 7b, 7d, 7f and 7h induces apoptosis in 17.91%, 69.23%, 43.10%, 41.35%, and 21.53% of MCF-7 cells, respectively as compared to the control cells (Fig. 6B). Consistent to the results from binding, enzyme inhibition and cell proliferation tests, 7b is active to a significantly higher extent (Fig. 6A,B). These results are consistent with previous observations which demonstrate that MARK4 as an inhibitor of hippo signalling in MCF-7 cells 3 and negative regulator of mTORC1 12 is a regulator of cell proliferation and migration of cancer cells.
Selected acridones increase the levels of reactive oxygen species. The respiratory cycle of mitochondria is the foremost source of reactive oxygen species (ROS), and ROS has the potential to induce cell apoptosis 37 . Thus, it was speculated that the treatment with selected acridone derivatives might lead to the production of ROS. The MCF-7 cells were treated with the IC 50 dose of compounds 5, 7b, 7d, 7f and 7h for 5-6 h and ROS levels were quantified by flow cytometry using 2-dichlorofluorescein diacetate (DCFDA) staining. Interestingly, it was found that treatment with compound 5, 7b, 7d, 7f and 7h increases the production of ROS. Representative histogram is shown in Fig. 6C, which suggests that incubation of MCF-7 cells with compounds 5, 7b, 7d, 7f and 7h shifts the position of respective histogram towards right (higher value), that shows an increase in the levels of ROS.
Besides flow cytometry, levels of ROS were also measured by fluorescence spectroscopy, after treatment with compounds 5, 7b, 7d, 7f and 7h, DCFDA staining and imaging on a fluorescence microscope (Fig. 6D). It is easily observed in the Fig. 6D that the intensifications of green fluorescence in case of treated cells denote the higher levels of ROS. Results of ROS measurement suggested that compound 5, 7b, 7d, 7f and 7h considerably increase the levels of ROS that might be also a reason for cellular death of MCF-7.
Other important observation suggested by ROS experiments is that compounds 7b, 7d and 7f induce ROS to a higher extent than compound 5 and 7h. Generation and accumulation of ROS results in oxidative stress and plays a crucial role in the governing of cancer cell behavior. Higher levels of ROS trigger an array of pro-apoptotic signal pathways, such as mitochondrial dysfunction and ER stress and, which in due course leads to worsening of cell function and apoptosis 37 .
Kinase inhibitor selectivity. The selectivity of kinase inhibitors is commonly evaluating them against a panel of closely related kinases based on the hypothesis that off-target interfaces are more likely to be associated  with same/different families of kinases with similar amino acid sequence. So, we evaluated the fraction of kinase targets that are within the same kinase family (CAMK family). The results of within-family selectivity of MARK-4 targeting compounds suggested that at a single dose of 10 µM, compound 7b, 7d and 7f more strongly inhibits MARK4 as compared to other kinases of same family ( Fig. 7A-C). It was clearly observed that, though selected compounds inhibit other kinases also, but non-significantly. The percent inhibition shows that for other kinases the IC 50 value will be higher than 10 µM, as at this dose none of the kinase inhibited more/equal than 50 percent ( Fig. 7A-C). These kinase selectivity results signify the use of these compounds as potential and selective inhibitors towards MARK4.

Tau phosphorylation studies.
To see the effect of most potential lead molecules (compound 7b, 7d and 7f) on the substrate protein (tau protein) of MARK4 enzyme inhibition activity studies are extended to a cell-based tau-phosphorylation assay. Cells were allowed to grow in the presence of IC 50 dose of compound 7b, 7d and 7f for 24 hrs. Following the treatment, the phosphorylation of tau has been assessed with the help of flow cytometry. Results of tau-phosphorylation assay suggested that the compound 7b, 7d and 7f decreases the phosphorylation level of tau protein (Fig. 7D). The outcomes of tau-phosphorylation assay as shown in Fig. 7D revealed that treatment of compounds 7b, 7d and 7f moves the histogram towards the lower side of untreated/control cells (shown by dark green color). These results suggested that the selected compounds inhibit the phosphorylation of tau and also clues the inhibition of MARK4 as tau is a substrate for MARK4.

Conclusions.
Conclusively, in this study it was perceived that the studied acridone derivatives significantly bind to the active site cavity of MARK4 that might be responsible for enzyme inhibition. The selected lead molecules also show the specificity towards MARK4. Inhibition of MARK4 hinders the proliferation of cancerous cells, enhance ROS generation and induce apoptosis. Results obtained from this study clearly support our assumption and previous studies that MARK4 will be an effective anticancer drug target. Furthermore, our findings indicate that acridone scaffold may be further employed in the discovery of potential MARK4 inhibitors, with an anticancer pharmacological profile, that will help to counter the progression of cancer and other MARK4 allied disorders.

Materials and Methods
Materials. Luria      Expression and purification of MARK4. The human MARK4 was cloned, expressed and purified by following our previously reported protocols 38,39 . In brief, the recombinant cells harbouring the expression construct of MARK4 were grown in Luria-Bertani broth and culture were induced by IPTG (1 mM) followed by overnight culture of cells at 16 °C with continuous vigorous shaking. The pellet was obtained by centrifuging the culture, dissolved in lysis buffer (50 mM Tris, 20 mM EDTA, 0.1 mM PMSF and 1% Triton-100) and inclusion bodies were prepared. Definite amount of inclusion bodies were taken and dissolved in sarcosine buffer (50 mM CAPS, 1.5% N-laurosyl sarcosine, pH 11.0) and were centrifuged for 25 min at 12,000 rpm and the supernatant was allowed to bind on preequlibrated Ni-NTA column GE healthcare (GE Healthcare Life Sciences, Uppsala, Sweden). After washing (with 5 mM imidazole in sarcosine buffer), protein was eluted with increasing concentration of imidazole (10 mM to 400 mM). Fractions containing MARK4 protein were pooled down and further purity was accessed using sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE) and confirmed with the help of Western blot using peptide specific primary antibodies 40 .  42 . Prior to docking analysis, the structure was emended by removing water and co-crystallized ligand , followed by addition of polar hydrogens and Gasteiger charges using Auto Dock Tool (ADT). The 2D and 3D structures of all the synthesized compounds 5, 7a-7j were generated and energy minimized by ChemBio3D Ultra 12.0. Both ligands and receptor were transformed to the proper format for docking through PyRx. Following the standard docking procedure, ligands were docked by defining a grid box with spacing 1 Å and size of 20 × 20 × 20 (Å) pointing in x, y and z directions around the protein active site. After preparing the coordinate files of MARK4 and respective compounds, they were subjected to molecular docking in order to see the bound conformations, binding affinity and possible protein-ligand interactions. The "exhaustiveness" was set to the value of 100 instead of the default 8 for all docking analysis. PyMOL viewer (Schrödinger, LLC) and "Receptor-Ligand Interactions" modules of BIOVIA/Discovery Studio 2017R2 were used for the visualization and structure analysis of the docked complexes of MARK4 and to generate two dimensional docking for the analysis of hydrogen bonds and hydrophobic interactions 43 .
Fluorescence measurements. Binding affinities of synthesized acridone derivatives with recombinant MARK4 was carried out by observing the fluorescence intensity change of emission spectrum of MARK4 by following our previously published protocol 44,45 . Each titration of protein was performed in triplicates and the average was taken for analysis. A significant decreased in fluorescence intensity of protein with increase in the concentration of acridone derivatives were used as the criteria for deducing the binding constant (K a ) as well as number of binding sites (n) present on the protein molecule using the modified Stern-Volmer equation 46 : Isothermal titration calorimetry. ITC experiments were performed on a VP-ITC microcalorimeter (MicroCal, Inc. GE, MicroCal, USA) by following our previously published protocols 41,47 . For sample preparation, recombinant MARK4 was extensively dialyzed in 20 mM Tris buffer, pH 8.5 and the working solutions of acridone derivatives were prepared in last dialyzing buffer. The titration data so obtained was analyzed with the help of MicroCal Origin 7.0 software provided with the instrument. The thermodynamic parameters of binding such as association constant (K a ), stoichiometry of binding (n), and enthalpy change (ΔH) were determined by fitting the binding isotherm into the 'one-set of sites' binding model.
Cell culture. HEK-293 and MCF-7 human cell lines were grown and maintained in a DMEM supplemented with 10% heat-inactivated fetal bovine serum (Gibco) and 1% penicillin, streptomycin solution (Gibco), in a 5% CO 2 humidified incubator at 37 °C. Routinely cells were cultured, maintained and trypsinized not more than 30 passages.

Cell viability assay.
To study the effect of synthesized acridone derivatives on cell viability and proliferation, a standard MTT method was carried out. Cells were plated (8000-9000/well) in a 96-well plate and incubated overnight. On the next day, cells were treated with increasing concentrations (0.1-80 μM) of synthesized acridone derivatives in a final volume of 200 µl for 48 h at 37 °C in a humidified chamber. At the end of treatment time point, 20 µl of MTT solution (from 5 mg/ml stock solution in phosphate buffer saline, pH 7.4) was added to each well and incubated further for 4-5 h at 37 °C in the CO 2 incubator. After stipulated time the supernatant was aspirated and the colored formazan crystal produced as a result of MTT reduction was dissolved in 100 µl of DMSO. The absorbance (A) of colored product was then measured at 570 nm on a multiplate ELISA reader (BioRad). The percentage of viable cells was calculated and used to estimate the IC 50 (50% inhibitory concentration) values for each acridone derivative. For cell proliferation studies paclitaxel has been taken as positive control.
Cell apoptosis assay. Annexin-V staining was used to analyse the apoptotic potential of synthesized acridone derivatives as described previously 47,48 . Briefly, MCF-7 cells were dosed with IC 50 concentration of selected acridone derivative for 24 h at 37 °C. The control cells were treated with media only. After 24 h treatment, nearly 2.0-2.5 × 10 6 cells were trypsinized and collected by centrifuging the cell suspension at 1800 rpm for 4 min. Collected cells were washed two times with 5 ml of PBS. Finally, cells were stained with FITC labeled Annexin-V antibodies using FITC-Annexin-V kit according to the manufacturer's guidelines (BD-Biosciences,