Immediately Preceding Cambridge Healthtech Institute's 1st Annual Predictive Toxicology Conference

With 50% of drug failures attributed to ADME-Tox issues, it is critical to accurately predict these qualities earlier in the investigation of a lead to assure appropriate attrition from the drug development process. Empirical assays have been the mainstay of evaluating ADME and toxicity, but they suffer from low throughput and difficulty to automate. New techniques and technologies for predicting these qualities must become more accurate and have the ability to be validated against empirical findings. It will then be possible to drop compounds that are most likely to exhibit ADME or toxicity problems sooner, and provide ways for researchers to arrive at correct conclusions, but in less time. The more attractive compounds can then be put on the development fast track, saving time and money. These two, 1.5 day, back-to-back meetings will provide a forum for discussion of new techniques, technologies, and strategies that hold promise for improving the selection of potent, on-target leads with fewer negative side-effects, thus saving time and money.

Keynote Presentation
The Unfortunate Truths behind in silico ADME Modeling

Bioavailability and Absorption
A Comparison of Commercially Available Software for the Prediction of Aqueous Solubility
Application of Physiologically Based Pharmacokinetic Models to Prioritize Potential Clinical Drug Candidates
Identification of Bioavailability Issues Caused By Active Transport: Novel Assays for ABC Transporters
The Evaluation of Permeability and Bioavailability During Drug Discovery: Caco-2 vs. PAMPA

Drug Metabolism and Pharmacokinetics
Predictive In Silico CYP450 Metabolism: Assessments of a Structure-Electronic Filter Method
Simulating Physiology to Predict Pharmacokinetics in Man and Rat
ADMET In Silico Modeling: Towards Prediction Paradise?
Measurement of Gene Expression Signatures: A Simple, Repeatable, High Quality In Vitro and Ex Vivo Predictive Metabolism and Tox Assay
Role of Drug Metabolism in Drug Discovery and Development

ADME/PK Models
QSAR Prediction of ADME Parameters Using a New Machine Learning Tool-Random Forest
Comparison of Empirical and Physiologically Based Approaches for the Prediction of Human Pharmacokinetic Parameters
ADMET Studies Using Rats with Chimeric Livers and Normal Immune Systems
Strategies in Lead Selection and Optimization: Application of a Graphical Model and Automated In Vitro ADME Screening
Application of an Up-Front Measure of Uncertainty for Predictive ADME Models

Monday, November 17, 2003

12:00 Registration

1:00 Welcome by Session Chairperson
Dr. John M. Maclennan, RETT Corporation

2:00 Pitfalls of Biopharmaceutical Property Predictions: Stretching Data and Models Too Far 
Dr. Pieter F. Stouten, Pharmacia Italia, Pfizer Group 
When modeling ADME or its underlying physical properties, several pit-falls must be avoided. Examples are: 1) use of data of insufficient accu-racy or inappropriate type; and 2) the generalization of observations beyond the training set. Pitfall 1 is elucidated by comparing solubility models developed on 24-hour equilibration data (spanning 16 log units solubility) and 2-hour data (spanning 3.5 log units), respectively, and by comparing protein binding models based on %bound (chromatography) and Kd (isothermal titration calorimetry) data, respectively. Pitfall 2 is exemplified by comparing the performance of several (in-house and com-mercial) solubility prediction tools on various data sets, and also by com-paring results we obtained with literature information on modeling-bioavailability with simple chemical descriptors. To demonstrate the impact solubility models can have on projects, a library design example is shown where solubility predictions led to a library with 50% of the com-pounds having solubility >250 micromolar.

2:30 Application of Physiologically Based Pharmacokinetic Models to Prioritize Potential Clinical Drug Candidates Prior to In Vivo Experiments
Mr. Neil Parrott, PRNS, F. Hoffman-La Roche, LLC (Switzerland)
During drug discovery considerable resources are required to assess the pharmacokinetic properties of potential clinical candidates in vivo in animals, and there is considerable interest in reducing the amount of such testing. Physiologically based simulation tools have the potential to do this by predicting the full plasma concentration versus time profiles based upon in vitro and in silico input data as the sole input parameters. Such an approach can help to rank order compounds based on their predicted in vivo PK profiles in order to select the optimal molecules for further in vivo experiments, and we will present data for over 100 Roche compounds where we have compared predicted and observed PK in the rat in a retrospective validation.

3:00 Refreshment Break

3:30 Identification of Bioavailability Issues Caused By Active Transport: Novel Assays for ABC Transporters
Dr. John M. Maclennan, Chief Operating Officer, RETT Corporation
P-glycoprotein (Pgp) is an efflux transporter expressed in the gastrointestinal epithelium, the kidney, capillary endothelial cells in the CNS and tumor cells. Transport by Pgp has been associated with reduced systemic bioavailability of orally administered drugs and chemotherapeutic resistance in tumors. PGP substrates include anticancer drugs such as the anthracycline antibiotics and vinca alkaloids, steroids, verapamil, peptides and quinolines. The rapid determination of a compound's affinity for Pgp and whether it is an inhibitor/substrate of the transporter are key elements in determining the efficacy of a drug in drug discovery and development programs. A chromatographic approach to the screening of compounds for Pgp activity that can be used to correctly sort compounds with high affinity, low affinity and no affinity to Pgp in less than 30 min will be described. Co-authors: R. Moaddel, I. W. Wainer

4:00 The Evaluation of Permeability and Bioavailability During Drug Discovery: Caco-2 vs. PAMPA
Dr. Yakov Rotshteyn, Group Leader, Discovery Support, Purdue Pharma
The Parallel Artificial Membrane Permeability Assay (PAMPA) continues to gain acceptance as a high-throughput surrogate for Caco-2 monolayers in determining membrane permeability during lead selection. The results of this assay, in combination with an evaluation of metabolic stability or intrinsic clearance, are used to estimate bioavailability and to select compounds for pharmacokinetic evaluation. However, the results of PAMPA may be somewhat misleading for compounds that have acceptable passive transcellular permeability but whose net absorption may be limited by intestinal efflux transporters.

4:30 Panel Discussion with all Afternoon Speakers

Tuesday, November 18, 2003

8:30 Poster and Exhibit Viewing, Light Continental Breakfast

Drug Metabolism and Pharmacokinetics

9:00 Comments by Session Chairperson
Dr. Danni Harris, President, Molecular Research Institute

9:05 Predictive In Silico CYP450 Metabolism: Assessments of a Structure-Electronic Filter Method
Dr. Danni Harris, President, Molecular Research Institute
Rapid, accurate, structure-based, in silico predictive toxicology related to CYP450 metabolism requires: 1) target (model) accuracy, 2) ligand CYP450 configuration sampling adequacy and 3) sufficiently accurate quantum chemical descriptors to discern the relative propensity of metabolism of multiple ligands sites (Park and Harris, J. Med. Chem. ASAP March, 29, 2003). The ability of a CYP450 structure (CYP2D6, CYP3A4, CYP1A2, CYP2E1, CYP2C9) and an electronic filter to predict: 1) ligand recognition, and 2) metabolism of large databases of 3D-stuctures from pharmacophore-directed searches will be described as a feasible approach to insert predictive toxicology early in the drug discovery effort. Critical assessments and directions for the future will be provided.

9:35 Simulating Physiology to Predict Pharmacokinetics in Man and Rat
Dr. Simon Thomas, Head of Scientific Computing, Cyprotex Discovery Ltd.
We discuss a new software system that uses basic ADME and physiological data to predict the pharmacokinetics of new compounds administered orally or intravenously to humans or rats. Data from 3rd party evaluations will be shown that demonstrate how predictions compare to in vivo measurements. We discuss how the capability could be effectively applied to the optimization and selection of lead compounds in drug discovery.

10:05 ADMET In Silico Modeling: Towards Prediction Paradise?
Dr. Pil Lee, Senior Research Scientist, Pfizer Global Research & Development

10:35 Poster and Exhibit Viewing, Refreshment Break

11:05 Measurement of Gene Expression Signatures: A Simple, Repeatable, High Quality In Vitro and Ex Vivo Predictive Metabolism and Tox Assay
Dr. Ralph Martel, Assay Technologies Director, High Throughput Genomics, Inc.
ArrayPlate™ technology provides a gene expression assay that has the sensitivity of PCR and provides quality, reproducible (whole assay average CV's of 10%), repeatable (day-to-day, lab-to-lab), and high sample throughput quantities of data. Sample preparation is simple, requiring only lysis; no extraction or purification of RNA, no reverse transcription, no gene amplification. Data presented will demonstrate performance using in vitro and in vivo samples, including measurement of indicator and predictor genes from animal tox and metabolism models. Measurement of gene expression is no longer limited to "fold" changes, but instead can be studied in terms of 10% to 20% changes (1.1 to 1.2-fold), and entire signatures or fingerprints of gene expression can be measured simultaneously from a single small sample of cells, tissues, or whole organisms.

11:35 Role of Drug Metabolism in Drug Discovery and Development
Dr. Gondi N. Kumar, Pharmacokinetics and Drug Metabolism, Amgen Inc.

12:05 Predicting Pharmacokinetics of a Drug using Machine Learning
Dr. Kalyanasundaram Subramanian, VP, Systems Biology, Strand Genomics
One of the causes of attrition in the pharmaceutical R&D pipeline is non-optimal ADME characteristics of a drug. ADME determines not only the availability of a drug for its therapeutic effect but also its toxicity, drug-drug interactions, etc. Current predictive methods have only partially addressed this need; in vitro and in vivo ADME models are not typically predictive of ADME or toxicity observed in the clinic. We have applied advanced machine learning techniques to mine information available on drug-like molecules and have developed methods to make accurate and quantitative predictions of compound properties in vivo. This has the potential of allowing the gen-eration and optimisation of leads with increased probability of success in the clinic. This approach also provides a methodology by which different forms of in vitro and in vivo data can be integrated within an appropriate mathe-matical framework to improve predictions. November 18, 2003

12:35 Lunch on your own

ADME/PK Models

1:45 Comments by Session Chairperson
Dr. Christopher Tong, Biometrician, Biometrics Research, Merck Research Laboratories

1:50 QSAR Prediction of ADME Parameters Using a New Machine Learning Tool--Random Forest
Dr. Christopher Tong
A new ensemble learning tool, Random Forest, is used for accurate in silico prediction of ADME parameters based on molecular structures and properties. Random Forest is capable of both classification and regression in various situations, while resisting overfitting and being robust to parameter tuning and noise. Examples used to illustrate Random Forest for QSAR with both low and high dimensional data include predicting blood-brain barrier permeability of test compounds by determining MDR1 P-glycoprotein-mediated vectorial transport in polarized LLC-PK1 cell lines. Authors: Vladimir Svetnik, Andy Liaw, Robert P. Sheridan, J. Christopher Culberson, Bradley P. Feuston, Raymond Evers, and Dylan Hartley

2:20 Comparison of Empirical and Physiologically Based Approaches for the Prediction of Human Pharmacokinetic Parameters
Dr. Hannah Jones, Non-Clinical Drug Safety, F. Hoffman-LaRoche Ltd. (Switzerland)
In order to reduce failures related to ADME issues in the drug development process, it is important to predict human pharmacokinetics as early as possible. Empirical methods have been traditionally used for this purpose, however, recently more powerful physiologically based mechanistic models have been developed. In the current work these physiological models have been validated against the empirical approaches in a retrospective analysis using a series of 18 Roche compounds; the ultimate aim is to develop the best practice for prediction of human PK.

2:50 ADMET Studies Using Rats with Chimeric Livers and Normal Immune Systems
Dr. Carolyn Kahn, President and CEO, Hepaticus, Inc. 
Summary: The Hepaticus rat has a chimeric liver composed of rat and human cells admixed in the rat parenchyma. These immunocompetent animals should serve as a unique technological solution for modeling human liver response in a easy to manipulate small animal model that have the potential to be substituted in the standard pharmaceutical pre-clinical tests.

3:20 Poster and Exhibit Viewing, Refreshment Break

3:50 Strategies in Lead Selection and Optimization: Application of a Graphical Model and Automated In Vitro ADME Screening
Dr. Arun Mandagere, Discovery Technologies, Pfizer Global Research and Development

4:20 Application of an Up-Front Measure of Uncertainty for Predictive ADME Models
Dr. Troy Bremer, Senior Computational Scientist, Chemical Informatics Research, Lion Bioscience
One aspect of application of predictive ADME is whether new compounds are outside the model training space, necessitating that the model extrapolate rather than interpolate and potentially increasing prediction uncertainty. We pose an up-front measure of uncertainty for ADME predictions developed through the identification of the model's interpolation limits by the application of a kernel support vector one-class classifier. This methodology is applied to a caco-2 training and testing set to illustrate its utility in identifying the compounds for which a predictive ADME model may have sub-optimal performance.

4:50 Panel Discussion with all Afternoon Speakers

5:20 Networking Reception, Poster and Exhibit Viewing

Wednesday, November 19, 2003

8:00 Registration, Light Continental Breakfast

 

Joint Session ADME/Tox Models: The Interface 8:30 Comments by Session Chairperson Dr. Julie Penzotti, Director, Computational Chemistry, Rational Discovery, LLC 8:45 Applying Artificial Intelligence to ADMET Predictions: A Collaborative Approach Towards Improved Models Dr. David E. Clark, Argenta Discovery Ltd. The next generation of predictive ADMET models will be driven by: Improvements in molecular descriptors and statistical modeling methods Improvements in the quality and quantity of data available for model building This talk will describe a collaborative research project between Amedis Pharmaceuticals and Argenta Discovery aiming to address both of these issues: Amedis Pharmaceuticals has developed novel, proprietary molecular descriptors and modelling technologies, based on genetic programming and advanced statistical techniques Argenta Discovery is applying its in vitro ADMET expertise to generate high quality data, e.g., for inhibition of various cytochrome P450s Details of this collaboration and its progress towards improved models will be presented. 9:15 Predictive Modeling Techniques Being Used by Companies Dr. Ravi Mallela, Recommind Inc. Predictive modeling is still in its infancy. A variety of methods has been developed in an attempt to answer the challenge. This talk will provide real-life case studies that focus on the currently used methods, the advantages and disadvantages, and the conclusions we can draw from them. 9:45 Integrating ADME/Tox Prediction and Informatics Dr. Gregory Banik, General Manager, Informatics Division, Bio-Rad Laboratories, Inc. ADME/Tox prediction is becoming an increasingly important tool in the evaluation and prioritization of compounds early in the drug discovery process. A system will be described that facilitates the deployment of multiple simultaneous prediction approaches to maximize prediction accuracy. The applicability of this to high-throughput evaluation will be discussed, as will the informatics issues relevant to in silico prediction, along with possible solutions. 10:15 Poster and Exhibit Viewing, Refreshment Break 10:45 Whole Body Autoradiography Answers Pivotal Questions for Early ADME and Toxicology Dr. Eric Solon, Director of Autoradiography and Animal Resources, Quest Pharmaceutical Services Whole Body Autoradiography provides tissue distribution information and can help to identify target organs and early information regarding ADME properties. Known drug interactions can be employed in the study design and answer more specific questions. Several examples of how QWBA has been used in drug discovery for ADME and Tox support will be presented. 11:15 An Ensemble Approach to Building Predictive ADME/Tox Models for Decision Support Dr. Julie Penzotti, Director, Computational Chemistry, Rational Discovery, LLC The development of predictive models for ADME/Tox properties provides decision-support tools that can be used in library design, lead optimization and the selection of drug candidates for development. Because multiple (possibly unknown) mechanisms can lead to the same toxicity endpoint or ADME property, algorithms for their prediction must be capable of handling multiple modes of action. Our modeling methods for ADME/Tox combine computational chemistry with an ensemble approach to machine learning, thereby selecting useful molecular descriptors and multiple predictive models to produce a single composite model that generally exhibits greater predictive power and generality. 11:45 Panel Discussion with all Morning Speakers 12:15 Conclusion of Predictive ADME Conference

Lead Sponsoring Publication:
Sponsoring Publication:
Web partners:

---

Call for Sponsors and Exhibitors There are many sponsorship opportunities for your company to maximize its exposure and influence. They include conference-specific sponsorships, technology workshops, networking receptions, delegate bags, etc. We are also ready to work with you in customizing a solution to meet your specific marketing objectives. Make a lasting impression by taking advantage of these marketing tools. For exhibit and sponsorship information, please contact Carol Dinerstein at 781-972-5471 or dinerstein@healthtech.com.

Hotel Information
Boston Park Plaza Hotel
64 Arlington St.
Boston, MA 02116
T: 617-426-2000 o F: 617-426-5545
RoomRate: $159/single o $179./double
Cut-off Date: October 22, 2003

Please call the hotel directly to make your room reservation. Identify yourself as a Cambridge Healthtech Institute conference attendee to receive the reduced room rate. Reservations made after the cut-off date or after the group room block has been filled (whichever comes first) will be accepted on a space-and-rate-availability basis. Rooms are limited, so please book early.

Travel Information
Special Airline Discounts Available
Special Zone and Discount Fares have been established for this conference with United Airlines. Please call United Airlines Meeting Reservation Desk at 800-521-4041 and reference ID#579YS.