What is iDEA pkEXPRESS ?

Historically, inappropriate pharmacokinetic properties have been a reason for the failure of compounds in the later stages of development. This was largely owing to an inability to rectify poor pharmacokinetic characteristics present in many lead series promoted to lead optimization. With the adoption of high throughput screening, combinatorial chemistry and parallel synthesis in drug discovery, the need for early information on the absorption, metabolism, distribution and elimination (ADME) characteristics of compounds became increasingly important in the lead selection and optimization process. This drove the development of in-silico models and medium to high throughput in vitro screens to characterize the ADME properties of compounds. Using this information, chemists can steer away from “problematic” chemotypes while trying to enhance the specificity and selectivity of lead compounds.

iDEA pkEXPRESS is a comprehensive predictive ADME software system developed and validated to predict relevant pharmacokinetic (PK) and ADME characteristics of potential drugs. The system has been previously deployed as a desktop application or as an ADME compute engine in a larger cheminformatics system within large Pharma.

iDEA pkEXPRESS™ includes the following:

  • Physiological Absorption Model

  • Physiological Metabolism Model

  • Physiological Distribution and Elimination model

  • One and two dimensional sensitivity analysis

The physiological models were trained using a diverse data set of internally generated in vitro data, human pharmacokinetic data from successful and unsuccessful clinical trials, and chemical structure. The models were internally and externally validated to benchmark their performance. The following is brief overview of the models and capabilities of pkEXPRESS.

 

Physiological Absorption Model

Developed in collaboration with five major pharmaceutical companies, the Absorption Model in pkEXPRESS predicts human intestinal absorption. The systems patented dispersed plug flow model of absorption, simulates human physiology and accounts for the regional solubility, regional permeability, intestinal surface area and fluid flow in the gastrointestinal tract. Optimized using internally generated in vitro data, intravenous and oral pharmacokinetic clinical data and chemical structures, the 62 drugs in the Absorption Model training set span 19 different therapeutic classes and include development successes and drugs that failed clinical trials. Compounds in the training set were selected so that a broad range of solubility, permeability, transport properties and extent of absorption was represented. Using dose, chemical structure, solubility, Caco-2 permeability, and efflux (optional input), the Absorption Model predicts the extent of absorption, absorption rate, mass absorbed, soluble mass, insoluble mass and intestinal drug concentration of potential drugs.

 

Physiological Metabolism Model

pkEXPRESS’ physiological Metabolism Model was designed and validated to predict the extent of metabolism (FH) or bioavailability of potential drug compounds. The Distributed Liver Flow model simulates first pass metabolism from a compound’s predicted absorption rate from the Absorption Model, protein binding and metabolic stability. The Metabolism Model was optimized using internally generated in vitro data, a training set of literature and collaborator pharmacokinetic clinical data, and chemical structures. The 82 drugs in the training set come from 29 different therapeutic groups and were selected to cover a broad range of Km, CLint, protein binding and extent of metabolism. Other pharmacokinetic models, such as the Well-Stirred and Parallel-Tube model or the application of CLint, assume linear kinetics. These models often over predict the extent of metabolism when the metabolizing enzymes are saturated, which occurs during the initial passage of drug through the liver. To overcome this issue, the Metabolism Model in pkEXPRESS incorporates the Michaelis- Menton equation to describe turnover kinetics. By doing so, the model successfully predicts FH for compounds whether they demonstrate linear or non-linear turnover kinetics without overestimating first pass metabolism.

 

Physiological Distribution and Elimination Model

Based on the Distribution and Elimination model published by Kawai, et. Al (J. Pharm and Biopharm Vol 22 No.5 1994), the Distribution and Elimination model in pkEXPRESS uses published human physiological blood flow rates and organ and tissue volumes to predict a compound’s plasma concentration level time curve (PLTC), Cmax, tmax and area under the curve (AUC). From the PLTC, other relevant pharmacokinetic parameters such as Vss, Cltot, t1/2, and micro- and macro-constants for a one or two compartment model can be calculated. The model was trained using intravenous pharmacokinetic data from 147 published and proprietary clinical studies. The model predicts the distribution and elimination of potential drugs, using protein binding, red blood cell partitioning ratio and either the predicted rate of absorption or metabolism, obtained from the Absorption or Metabolism Models respectively.

 

Sensitivity Analysis

Sensitivity analysis can be used to investigate the impact various input parameters have on pkEXPRESS’ predictions. For example, see how the absorption of a compound is affected by changes in its solubility or how varying protein binding affects the bioavailability of a compound. By using pkEXPRESS’ one or two dimensional sensitivity analysis, you can quickly show how significant a parameter is in determining the absorption, metabolism, distribution and elimination of a compound. This information can be used to assess which assays are the most important to perform, identify critical properties on which to focus lead optimization resources, or identify compounds with ADME liabilities that are unlikely to be overcome.