Towards superior cancer drugs by modeling approaches: Modification and prediction of the mode of action

• Author / Creator

  Mahshad Moshari

• Cancer is a leading cause of death that imposes significant economic and social suffering on a global scale. Based on the projected growth and ageing of populations, the global burden of cancer is set to increase. Despite the vast investment in cancer research, the rate of translation of research developments into clinical practice and drug discovery is still low, mainly because drug development is a lengthy, expensive, and complicated process. Computer-aided, in silico, drug discovery is a powerful technology that offers a more effective, cheaper, and faster alternative for drug design. Based on the known structure of biomacromolecules and small targeting molecules, computational methods employ virtual screening techniques for a broad range of applications, from hit identification to lead optimization, and to drug target prediction. In terms of cancer research, microtubules are a key component of anti-cancer drug design approaches. Microtubule dynamics play a critical role during cell division making these biopolymers and their subunits superb targets for anti-cancer therapy. Diverse chemotherapeutic agents with the potential to target microtubules are currently among the most effective group of anticancer drugs available. Colchicine is an effective anti-mitotic drug that tightly binds between α, β tubulins and inhibits their assembly into microtubules; however, its application as an anti-cancer drug is limited due to its serious drug interactions and toxicity. Regardless, iii
  colchicine, due to its effective antiproliferation activity, has been used as a lead compound to generate potential chemotherapeutic drugs with desired pharmacological profiles.
  In this thesis, several computational techniques including ab initio quantum chemistry calculations, molecular docking, and virtual screening were employed to identify libraries of colchicine derivatives with the highest binding affinities towards β tubulin. Good R2 values of the linear regression between the binding affinities and IC50 values of compounds against different human cancer cell lines including breast cancer (MCF7), lung carcinoma (A549), and colon cancer (LoVo) were achieved. Computational studies on the novel derivatives were conducted on different groups of colchicine analogues including a group of double-modified 4-halothiocolchicines derivatives, a series of triple-modified 4-chloro-7-carbamatethiocolchicines, 4-bromothiocolchicine, and 4-chlorothiocolchicine analogues. A library of double-modified carbamate or thiocarbamate derivatives of 10-demethoxy-10-methylaminocolchicine, 7-deacetyl-10-thiocolchicine, and 4-iodo-7-deacetyl-10-thiocolchicine were also studied. Moreover, a 3D QSAR model was generated that has the ability to predict the IC50 values for 70 novel derivatives of colchicine based on their binding affinities. The IC50 values of these colchicine analogues were predicted against two commonly-used cell lines including MCF7 and A549. The chemical structure of a library of 50 new compounds and their corresponding in vitro activities (IC50 values) were used as input data to construct two models for the cell lines mentioned above. The input data was split into training and test sets using a Kohonen map. External independent validation was done using 15 independent compounds. Docking was used to estimate the binding and electrostatic energies between the colchicine derivatives library and βII tubulin. The identified estimates were used as two novel descriptors. The models were generated using a commonly used Artificial Neural Network. The generated QSAR models showed good performance on the test set for both A549 and MCF7 cancer cell lines (assessed through the high values of q2 and R2). Besides, it was shown that these models had predictive ability and desired generalization on the independent validation set of compounds. In an effort to discover new chemotherapeutic agents, the mode of action of a novel microtubule inhibitor scoulerine, was evaluated using a combination of computational approaches. To the best of our knowledge, this is the first time that the computational prediction and experimental validation of the molecular mode of action of scoulerine as a potential anticancer drug were investigated. Human tubulin structures at both free and microtubule states were modelled. Docking of the optimized structure of scoulerine was subsequently performed and the highest affinity binding sites located in both the free tubulin and in a microtubule were identified. Our findings show that binding in the vicinity of the colchicine binding site and near the laulimalide binding site are the most likely locations for scoulerine. These computational predictions were confirmed by thermophoresis assays using scoulerine and tubulin in both free and polymerized form. These results suggest a unique property of a dual mode of action for scoulerine with both microtubule stabilization and tubulin polymerization inhibition ability.

• Subjects / Keywords

  • Cancer
  • Drug design
  • Molecular dynamic simulations
  • Molecular docking
  • QSAR modelling

• Graduation date

  Fall 2021

• Type of Item

  Thesis

• Degree

  Doctor of Philosophy

• DOI

  https://doi.org/10.7939/r3-mrwq-ch55

• License

  This thesis is made available by the University of Alberta Libraries with permission of the copyright owner solely for non-commercial purposes. This thesis, or any portion thereof, may not otherwise be copied or reproduced without the written consent of the copyright owner, except to the extent permitted by Canadian copyright law.