An in silico evaluation of CDK Inhibitors targeting BCL2, TS and mTOR

The cyclin dependent kinase (CDK) inhibitors have recently been found to be of potential use as anticancer drugs. The present research work focuses on screening of compounds targeting multiple pathways involved in human cancers along with CDK-regulated cell cycle for prospective anticancer potential. Molecular docking study of selected compounds were performed to determine the binding affinity of selected compounds towards respective targets of cancer cells to verify if there is any physical interaction of these inhibitors with their reported target proteins as claimed in the existing literatures. Prior to docking, molecular pathway prediction and gene set enrichment analyses were performed to identify the target molecules by using appropriate bioinformatics tools. Interestingly, the results of in silico molecular docking have been found to be in line with the laboratory findings that are obtained from the literatures. Specifically, few of our selected CDK inhibitors, namely Abemaciclib, Palbociclib, AMG 925 and RGB 286638 showed good binding scores against BCL2, TS, mTOR in addition to CDKs (4, 6 and 9). On the basis of scientific evidence based on published scholarly articles and according to molecular docking results, it can be inferred that these CDK inhibitors as anticancer agents may play a very promising role in cancer treatment and can be used as potential lead compounds for the development of target therapy against human cancers. However, more intensive research is needed to confirm the feasibility of these compounds to be used in treating cancer and it is expected that this work will provide a stimulating impetus for the development of chemotherapeutic agents in future.


Introduction
Cyclin dependent kinase (CDK) enzymes belongs to a group of related proteins that control the cell cycle progression. CDKs are heteromeric serine/threonine protein kinases and among the 12 different isoforms of CDKs, only 5 (CDK1-CDK7) have been shown to drive the cell cycle directly (Ding et al., 2020). CDK deregulation, either through direct or indirect means, is found in most cancer cells. Pharmacologically, CDK inhibition has become an attractive strategy towards mechanism-based and non-genotoxic therapies in oncology. Over the last decade, many potential drug candidate molecules are discovered which are capable of inhibiting CDKs, blocking cell-cycle progression, modulating transcription and inducing apoptosis selectively in cancer cells. A CDK (cyclin-dependent kinase) inhibitor potentially inhibits the cellular functions of CDKs. They are employed to treat cancers and hey act by preventing overproliferation of cancer cells (Law et al., 2015). CDK inhibitors are considered as target therapy. Majority of the currently available anticancer drugs are to some extent have some serious limitations including toxicity and importantly therapy failure. Anticancer drugs fail to Asian J. Med. Biol. Res. 2021, 7 (2) 127 provide efficacy in most of the cases when they target only single, selective pathway (Wang et al., 2019). Scientists all over the world are therefore in quest of some super molecules that will be able to produce effective anticancer activity by targeting more than one pathways. Herein, we present the evidence of physical interaction in silico between few selected investigational molecules along with approved CDKIs and some putative targets as a baseline study for development of anticancer agents.

Materials and Methods
Docking methodology is used to predict the new binding modes and affinities of small molecules within the binding site of particular receptor targets and is currently used as a standard computational tool in drug design for lead compound optimization and in virtual screening studies to find novel biologically active molecules (Meng et al., 2011).

Selection of compounds for docking
By a systematic literature search, we found 49 compounds with CDK inhibitory property. Among these compounds include molecules already approved by the USFDA and those in various phases of clinical trial. These compounds target different type of CDK such as CDK1, CDK2, CDK4, CDK6, CDK7 and CDK9 etc. From these compounds, we screened only seven compounds (Table1) as these have been found in the literature to target multiple pathway components along with CDK especially CDK4, CDK6, CD7 and CDK9.  (Thomas et al., 2012), AMG-925 (Li et al., 2014), and Alvocidib (Zocchi et al., 2018) (Table 1). These compounds were chosen as these were reported to target mainly cyclin dependent kinases (CDKs) along with other target proteins (Table 2). The 3D structures of the target proteins in PDB format were obtained from RCSB Protein Data Bank (RCSB-PDB) (Bank, 2021, Berman et al., 2000, a reservoir for three dimensional structures of proteins identified by Xray crystallography, NMR etc. The 3D structures of the selected target proteins were retrieved from RCSB-PDB in PDB format. The 3D structures of the selected drug molecules were retrieved from PubChem in SDF (Structure data file) format.

Active pocket site identification
Active sites of the target proteins were predicted using the tools CASTp (Dundas et al., 2006). This provides a resource for locating, delineating and measuring concave surface regions on 3D structures of proteins. After analyzing the pockets suggested by the above tools, a consensus active site was chosen for docking analyses. The choice of active site was also done keeping the conserved region of protein 3D structure under consideration.

Preparation of Target Proteins and Ligands
For docking, the structure of the target proteins (Table 2) were converted into PDBQT format after performing necessary modifications in Auto Doc Tools (ADT) (version 1.5.6) (Morris et al., 1998). The modification involves stripping all heteroatoms (including waters), re-protonating all atoms, and re-assigning Gasteiger charges to all atoms (Morris et al., 1998). The ligands were converted to PDBQT with Open Bable tool (O'Boyle et al., 2011). Open Bable basically purifies and modifies the ligand and convert them into format of interest.

Docking and post-docking visualization for binding actions
For this computational study, we used various available software and online databases as outlined in Table 3. The docking of the ligands with the target proteins was done by using YASARA (Krieger and Vriend, 2014). Then, the prepared PDBQT format of the proteins and the ligands were imported for docking. Next, a Grid box was generated into the predicted active sites of the target proteins. The docking parameters were set using the software Autodock Tools (Morris et al., 1998). Finally, the output files were saved in PDBQT format. The output files were opened in Autodock Tools (Morris et al., 1998) for visualizing the interaction of the specific active sites of target proteins with the ligands of interest.

Results
Almost all of our selected molecules showed remarkable binding affinities for most of the target proteins. RGB286638 showed excellent binding affinity with BCL2 (8.124kcal/mol), CDK4 (8.413kcal/mol), CDK6 (8.528kcal/mol), CDK9 (8.007kcal/mol), mTOR (9.674kcal/mol) and TS (10.063kcal/mol). Evidently, this molecule holds a good potential as a multi-target anticancer drug candidates. Along with the classical pathway inhibition like thymidylate synthase (TS) and mTOR, inhibition of cell cycle and anti-apoptotic protein might be a good strategy to combat therapy-resistant cancers. Likewise, abemaciclib, a launched drug which have activity against CDK also showed excellent binding affinity with BCL2 (7.982kcal/mol), CDK4 (8.525kcal/mol), CDK6 (7.717kcal/mol), CDK9 (8.472kcal/mol), mTOR (8.748kcal/mol) and TS (8.551kcal/mol) thereby showing similar anticancer activity like RGB286638. These two molecules thus demonstrated strong binding against the selected targets (Table 4 and Figure 1). Other molecules showing strong binding interactions against the selected targets are presented in Table 4. Among the other putative targets include DHFR, one of the classical anticancer targets as well as MAPK8 and MAPK9, which are often implicated in various types of human cancers (Gkouveris and Nikitakis, 2017).

Discussion
After analyzing the docking results, we observed that, all the compounds showed excellent binding interactions with our selected target molecules. Our computational findings indicate that these compounds have significant potential to become good anticancer leads to target multiple pathways including cell cycle regulatory network.
In particular, our findings clearly demonstrate that both abemaciclib (launched) and RGB286638 (in clinical trial) can potentially target the anti-apoptotic protein BCL2, TS as well as mTOR. Among these, the antiapoptotic BCL2 has been one of the potential drug able targets that led the scientists to the development of target therapy as anti-BCL2 anticancer agents, more popularly known as BCL2 inhibitors (Thomas et al., 2013;Huang, 2000). Since anti-apoptotic BCL2 play a major role in tumor cell survival, BCL2 inhibitors have been developed as direct inducers of apoptosis. One such molecule,  received breakthrough TS TS therapy designation from the FDA due to its remarkable efficacy in chronic lymphocytic leukemia (CLL) and acute myeloid leukemia (AML) (Soderquist RS and Eastman A, 2016). TS and mTOR are two of the most popular targets against conventional anticancer drugs that have been in the focal points of the scientists for many decades. Besides, mitogen-activated protein kinases (MAPKs) including MAPK8 and MAPK9 have found to be overexpressed and hence implicated in various human cancers. Also, they mediate resistance against chemotherapy and apoptosis (Gkouveris and Nikitakis, 2017). Since our docking study identifies all these targets along with the CDKs (CDK4, 6 and 9) for abemaciclib and RGB286638, it can be suggested that these two compounds can produce promising anticancer activities by inhibiting these target proteins (Figure 2).

Conclusions
In this computational study, we summarized the binding interaction of seven inhibitors against putative targets that provide us an insight for rational drug development that require extensive structural modifications and medicinal investigations. After molecular modification of structures, these compounds need to be tested on cell lines overexpressing their respective targets to confirm their better interaction and improved potency to develop them as potential anticancer drugs or drug leads. In short, more detailed and precisely designed pharmacological, biochemical and toxicological investigations in relevant in vivo and in vitro models are required to confirm the effectiveness, usefulness and toxicity to validate the feasibility of these promising molecules as lead compounds.