In silico strategies to predict drug metabolism and drug-drug interactions

Abstract

Early consideration of Absorption, Distribution, Metabolism, Excretion and Toxicity (ADMET) properties is highly necessitated for the reduction of high attrition rates associated with new chemical entities (NCEs) during the later phases of drug development. In addition to the experimtal testing a plethora of computational methods have been applied to complement the screening of NCEs with undesirable ADMET properties. The assessment of ADMET properties and the interactions of NCEs with Cytochrome P450 (CYP450) family of enzymes remains a central constituent of the current drug discovery process. The CYP450 superfamily of heme containing proteins are a group of clinically important enzymes that are involved in the metabolism of structurally diverse endogenous and exogenous compounds with CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 subtypes playing a major role in the overall metabolism of currently marketed drugs.

The CYP450 binding site is highly promiscuous allowing the metabolism of a broader range of structurally diverse chemical entities thus, leading to the modulation of enzymatic activity through the induction or inhibition of CYP450 subtype. The induction of CYP450 isoform might lead to an increased enzymatic activity, reduced efficacy and high clearance however, the inhibition of CYP medited metabolism might lead to the undesirable drug- drug interactions and toxicological outcomes. The adverse drug-drug interactions associated with CYP450s has led to the withdrawal of various marketed drugs (Seldane, Posicor, Hismanal, Propulsid, Lotronex, Baycol, and Seraone). Therefore, there is an earnest need for the early categorization of compounds as inhibitors, substrates and inducers of the CYP subtypes that would ultimately reduce the drug-drug interactions related adverse outcomes and result in the overall improvement of the therapeutic profile of a drug .

The major goal of this thesis was to elucidate the physicochemical properties and drug efficiency metrics for the inhibitors of the major metabolic enzymes across various activity levels. Herein, we identified a set of highly efficient inhibitors of the five CYP ( CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4) subtypes. We also investigated the binding hypothesis of the

Abstract

selected inhibitors against the specific CYP450 isozymes. Towards this goal, the refined databases of CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 inhibitors were collected for which physicochemical descriptors were calculated, the Lipophilic and Ligand Efficiency profiling was also performed. Initially, the concept of the drug efficiency was applied to binders of the true targets but here we apply these to a class of anti-targets the CYP450s.

The CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 inhibitors with best potency/lipophilicity and optimal fit within the binding site were identified as the highly efficient inhibitors that were further used for the exploration of the binding hypothesis using molecular docking, molecular dynamics simulations (MD) and binding energy predictions. Results show the affinities of the selected inhibitors against the CYP450 subtypes and the contribution of important binding site residues towards the overall stabilization of the inhibitor bound complexes through the WaterSwap method. The residue wise decompositions were also visualized which might provide an additional insight of the binding site residues essential for ligand binding that might assist in the optimization of leads for the development of new therapeutic options by increasing the specificity and reducing the adverse toxicological outcomes.

Additionally, the metabolic attributes of CYP450 substrates were also explored using the structurally diverse datasets of CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 substrates. Important parameters including metabolic stability, intrinsic clearance and lipophilicity were also monitored for the selected data sets to define meaningful relationships. The metabolic stability and intrinsic clearance of CYP450 substrates are linked with the respective lipophilicity (logP or logD) which necessitates the normalization of lipophilicity (logP or logD) of a given CYP450 substrate with respect to its metabolic stability (LipMetE) and intrinsic clearance (log;oCLiy yy). Therefore, the LipMetE values of already known substrates of CYP1A2, CYP2C9, CYP2C19, CYP2D6 and CYP3A4 including some marketed drugs have been calculated to elucidate the relationship between lipophilicity (logD74) and in

vitro clearance. Moreover, drug efficiency metrics including Lipophilic

Abstract

Efficiency (LipE) and Ligand Efficiency (LE) have also been evaluated and the thresholds of these parameters have been defined for the CYP450 substrates exhibiting normalized LipMetE. Our results indicate that for a given range of LipMetE, greater the logD value of the substrate the more avidly it will bind to a given CYP450 resulting in more intrinsic clearance (log10CLincu)- It is obereved that majority of the model substrates of CYP450 isoforms and the already marketed drugs in our data sets exhibit logD74 values of ~2.5 with LipMetE values in the range of 0-2.5 and LipE values of <3. The consideration of the studied parameters and optimal property thresholds of substrates in the ADME profiling could assist in reducing the drug failure rate in the later stages of clinical investigations. Overall, this work could help in increasing the therapeutic efficiency of drugs by reducing the adverse toxicological outcomes ultimately paving the way towards safer drug

development.