This one-day scientific symposium covered: technologies that guide you to successful compound discovery and development; intuitive molecular design and 3D-QSAR interpretation; and knowledge based toxicity prediction.
|09:00||Keynote presentation: Understanding Compound Quality||Dr Paul Leeson, Paul Leeson Consulting Ltd|
|10:00||Practical Application of Multi-Parameter Optimization to Guide Successful Drug Discovery||Dr Matt Segall, Optibrium|
|10:30||Improving the Plausibility of Success in Drug Discovery with the Use of Inefficient Metrics||Dr Mike Shultz, Novartis|
|11:30||Analyzing Selectivity Through Multi-dimensional Activity Cliff Analysis||Dr Tim Cheeseright, Cresset|
|12:00||The Open Innovation Drug Discovery (OIDD) Virtual Screening – Mining Blindly||Dr Maria Alvim-Gaston, Eli Lilly and Company|
|12:30||Predicting Metabolites – Enhancing an Expert System with Machine Learning||Dr Christopher Barber, Lhasa Limited|
|13:00||Lunch and software demonstrations|
|14:30||Predicting Adverse Drug Reactions: What Works and What Doesn’t||Dr Nigel Greene, Pfizer|
|15:00||COSMO Polarization Charge Densities as Key Information for Solubility and Partitioning Property Prediction and 3D-QSAR||Prof Andreas Klamt, COSMOlogic|
|15:30||Modeling of Chemical and Physical Stability of Pharmaceuticals||Dr Yuriy Abramov, Pfizer|
|16:30||Development of a Drug Discovery Simulation Laboratory Exercise in the Pharmaceutical Sciences Graduate Program Curriculum||Dr Chase Smith, Massachusetts College of Pharmacy and Health Sciences|
|17:00||Development of Kinome-wide and Isoform Selective Inhibitors of GSK3α and GSK3β for the Treatment of Psychiatric Disorders||Dr Florence Wagner, Broad Institute|
|18:00||Drinks reception and software demonstrations|
Dr Paul Leeson, Paul Leeson Consulting Ltd
A significant body of data suggests that controlling the molecular properties of hits, leads and drug candidates in the drug discovery phase should help to reduce risks of poor drug metabolism and toxicity. Concern for the health of pharmaceutical pipelines stems from analyses of small molecules patented by the industry, which are on average more lipophilic and larger than approved oral drugs. This is relevant too because only ~4% of candidate drugs reach the market and it has been acknowledged that compound-related risks, in dose, exposure and toxicity, can be carried from discovery into more costly clinical development.
Compound quality is controllable, being fixed at the point of design and varies substantially across organizations. Improving compound quality can be facilitated, inter alia, by selection of lead-like chemical starting points, use of predictive chemistry tools and application ligand efficiency measures to guide optimization.
Dr Matt Segall, Optibrium
A high quality drug must exhibit a balance of many properties, including potency, ADME and safety. Multi-parameter Optimization (MPO)  methods guide the selection and design of compounds to identify those with the highest chance of success, while minimizing opportunities missed by inappropriately rejecting compounds. In drug discovery this is particularly challenging due to the complex, often conflicting nature of the property requirements, combined with uncertain data because of experimental variability or predictive error.
But, how do we know what the profile of property criteria should be for a specific project? The criteria will depend on the ultimate goal of the project, e.g. therapeutic indication and route of administration, and are typically chosen based on the subjective opinion of the project team. However, analysis of historical data, using methods called rule induction , can guide the determination of the most appropriate profile, which can then be used prospectively to prioritize new compounds.
In this presentation we will discuss practical approaches for deriving and applying multi-parameter property profiles to guide compound optimization, illustrated with applications to drug discovery projects.
 M.D. Segall. Multi-Parameter Optimization: Identifying high quality compounds with a balance of properties. Curr. Pharm. Des. (2012) 18(9) pp. 1292-1310
 I. Yusof, F. Shah, N. Greene and M.S. Segall. Finding the Rules for Successful Drug Optimization. Drug. Discov. Today (2014) 19(5) pp. 680-687
Dr Tim Cheeseright, Cresset
During lead optimization the stepwise progression of compound activity is often disrupted by compounds that cause a disproportionately large (positive or negative) change in the biological response. These activity cliffs have long been recognized as an important source of information about the requirements of the protein for the series of interest. Activity cliff analysis has traditionally been done in 2D, but we have developed methods for expanding the dimensionality of activity cliff detection to include the 3D shape and electrostatic character of the ligands. In contrast to fingerprint similarity methods, accurate 3D similarity methods treat bioisosteres correctly which allows the identification of cliffs which the 2D methods fail to find.
The detection of activity cliffs for the primary activity end point is a valuable addition to the arsenal of drug discovery scientists. However, modern drug discovery rarely proceeds through the optimization of a single end point. More often project teams are tasked with optimizing the primary activity while minimizing the effect on a secondary, selectivity target or on a critical ADMET parameter. We have therefore studied the application of the 3D activity cliff analysis to multiple activity endpoints. These ‘selectivity cliffs’ highlight where molecular changes have a large effect on the activity against one target but not another. I will discuss the challenges of visualizing this data and present some novel techniques to deal with this.
Dr Maria Alvim-Gaston, Eli Lilly and Company
Maria Alvim-Gaston*, Gregory Durst£, and Marta Pineiro-Nunez*
*Open Innovation Drug Discovery, Discovery Chemistry; £ Computational Chemistry and Cheminformatics, Discovery Chemistry
The Open Innovation Drug Discovery (OIDD) program at Eli Lilly offers to examine submitter’s molecules for suitable drug properties and to direct molecules to various assays without directly reviewing chemical structures, thus maintaining structural confidentiality with our clients. This is done by calculating 2D and 3D molecular fingerprints for each submission and passing only those fingerprints on for making decisions. This is straight forward for 2D fingerprints, but presents a challenge for 3D fingerprints. Cresset’s standard 3D similarity score is computed as a weighted average of 3D shape similarity and field similarity. The standard weighting is 50-50 and the field points are usually represented around the 3D structure. This standard representation is not appropriate in the OIDD context. In the OIDD case we calculate and pass on only the field points and sizes with the explicit structure information removed from the file. This field point only file is a 3D abstraction of the submitted molecule, but it proves a suitable and useful construct to compare to other molecular field point files for making further decisions. With this metric, only the field points of the candidate molecule need to be transmitted across the OIDD firewall, meeting the confidentiality requirements. The application of this modified metric within the OIDD Virtual Screening initiative will be discussed.
Dr Christopher Barber, Lhasa Limited
The computer-based prediction of metabolites based upon structure has a wide number of applications – from a chemist’s desire to tune the metabolic profile of a lead, or a biologist’s requirement to predict likely toxic metabolites, to an analyst’s need to assign a peak in a bio-sample. Expert systems can provide transparent predictions with commentary and support based upon human knowledge whereas machine learning approaches are able to absorb new data more quickly but frequently show poor interpretability through the choice of descriptors and/or the model building methodology. However, these approaches are not mutually exclusive and the combination of both offers the potential for a new range of powerful predictive systems.
This talk will describe the science and results behind our work to apply machine learning approaches in order to enhance predictions made from the extensive biotransformation rule-base found within Meteor Nexus.
Dr Nigel Green, Pfizer
The ability to predict the adverse safety effects of small molecules has long been an aspiration in the pharmaceutical industry as it offers the advantage of lowering costs associated with the identification of candidate drugs whilst increasing the speed of development. Numerous methods and approaches to predicting adverse events have been developed but computational models that utilize physicochemical properties, structural alerts, polypharmacology assessments and mechanistic in vitro assays offer the most promise. One such approach is now being used to help guide early medicinal chemistry efforts and can be built into product development strategies to mitigate unwanted health effects in drug candidates that ultimately go to commercialization. By combining these diverse data types in an optimized, holistic model, a prediction of the exposure at which a compound may demonstrate a threshold level of toxicity in an in vivo study can be made. By combining this prediction with assessments of projected efficacious concentrations some success has been achieved in predicting the likely therapeutic index of a novel molecule. The use of such approaches allows medicinal chemists to steer early design efforts away from unproductive space, potentially reducing the use of in vivo experimentation for compounds with no hope of success.
Prof Andreas Klamt, COSMOlogic
The surface polarization charge density σ, which a virtual conductor would place on the surface of an embedded molecule, can nowadays be easily calculated for almost any molecule based on quantum chemical DFT/COSMO calculations. Within the COSMO-RS theory it has been demonstrated that the free energy of molecules in liquid phases and as a consequence properties as vapor pressures, solubilities, partition coefficients, etc. can be very well calculated by statistical thermodynamics based on σ-profiles of solutes and solvents. This method meanwhile is widely used in chemical engineering, but also in pharmaceutical drug development for solvent screening and other purposes. The success of COSMO-RS proves that the polarization charge density s essentially holds all information required for the quantification of molecular interactions in the liquid phase, especially about electrostatic interactions and hydrogen bonding.
If σ is such a good descriptor for the interactions in solution, it should also be powerful for the quantification of receptor-ligand interactions, because these are of the same nature. This idea led to the development of local, grid-based σ-profiles (LSPs), which can be used for accurate ‘field-based’ alignment and 3D-similarity of ligands (COSMOsim3D). In a further step, the LSPs have been demonstrated to be optimally suited descriptors for molecular field analysis. This COSMOsar3D method turns out to outperform traditional 3D-QSAR methods as COMFA or COMSIA with respect to accuracy of the trained models. Even more important may be the robustness of the COSMOsar3D models, which have are almost insensitive to the choice of the grid position and spacing and do not depend on any cut-off parameters.
In this presentation I will give an analysis of the origin of the superiority of the polarization charge density σ compared to the traditionally used electrostatic potential (ESP/MEP).
Dr Yuriy Abramov, Pfizer
One of the major concerns in modern drug discovery and development is chemical and physical stability of small molecule pharmaceuticals. Chemical stability is crucial for compounds at all stages of pharmaceutical R&D, from early drug discovery to formulation of liquid or solid dosage forms. Physical stability is typically related to stability of the pharmaceutical solid form.
QSPR models of oxidative chemical stability were built based on a large data set of electrochemical measurements. In addition a quantum chemical approach was proposed for oxidative chemical stability ranking of small organic molecules. Examples of the models application to pharmaceutical compounds will be discussed.
A typical physical instability issue of solid pharmaceuticals is related to a hydrate formation during formulation or product shell life. Transformation from anhydrate to hydrate solid form can have a significant impact on product performance and may also lead to a chemical instability. A model describing propensity of an API solid form to hydrate formation will be presented. In addition a rational coformer selection to enhance hydration stability of a co-crystalline form of API at a high relative humidity will be discussed.
Dr Chase Smith, Massachusetts College of Pharmacy and Health Sciences
The development of a 5-week long laboratory exercise that simulates an early stage drug discovery program and hit-to-lead optimization for use in an introductory course in the Pharmaceutical Sciences program at MCPHS University (Worcester/Manchester) will be discussed. Using the ADME QSAR module of the Stardrop™ software package, the students were introduced to triaging primary antimalarial screening data, evaluating calculated drug like properties and finally the selection of a hit series. The students then embarked on a hit-to-lead optimization through a decision making process involving improvement of the calculated drug like properties, improvement of metabolic stability using the web based SMARTCyp© algorithms and the calculation of analog activity using a QSAR model from the mobile application SAR Table©.
Dr Florence Wagner, Broad Institute
The serine/threonine kinase Glycogen Synthase Kinase-3 beta (GSK3β), part of the canonical WNT signaling pathway, has been implicated in multiple human disorders including non-insulin-dependent diabetes mellitus, cardiac hypertrophy and cancer as well as neurological and psychiatric disorders. Several converging lines of evidence make GSK3β an attractive target for the treatment of psychiatric disorders. First, a growing number of direct genetic associations for WNT signaling have been established in bipolar disorder, schizophrenia and autism. Secondly, the rapid anti-depressant effects of ketamine may depend on GSK3β signaling. From a therapeutic perspective, it has been demonstrated in pre-clinical studies that potent GSK3α/β inhibitors may be efficacious in models of lithium insensitivity. While numerous GSK3α/β inhibitors are reported, none possess isoform selectivity nor a desirable kinome-wide selectivity profile and suitable pharmacokinetic properties required of a CNS drug. Moreover, GSK3α/β inhibitors are known to increase β-catenin level and may lead to increased risk of neoplasms. Therefore, we set out to identify small-molecule GSK3α and GSK3β isoform selective inhibitors to delineate the biological function of each isoforms and for their potential use in a variety of disorders including psychiatric and neurological disorders.
Yuriy Abramov graduated with honors from Moscow Mendeleev University of Chemical Technology of Russia with specialization in Physical Chemistry. He received his PhD at Karpov Institute of Physical Chemistry, Moscow, Russia and started his carrier as an Assistant Professor at Moscow Mendeleev University of Chemical Technology of Russia. In 1995 Yuriy Abramov was invited as a Research Fellow at the National Institute for Research in Inorganic Materials, Japan. After that Yuriy Abramov was awarded a position of Research Assistant Professor at Chemistry and Medicinal Chemistry Department, State University of NY, Buffalo USA.
Yuriy Abramov started his career at Pfizer in 2001 as a computational chemist in Discovery, supporting multiple therapeutic areas including Inflammation, Immunology and Oncology. This led to a computational chemistry position in Pharmaceutical Sciences Department in 2008. Dr. Abramov’s computational support of Drug Discovery & Development projects at Pfizer was recognized by multiple awards.
Maria graduated from UFRJ, Federal University of Rio de Janeiro, RJ Brazil with a BS in Pharmacy in 1988 and continue with her education getting an Industrial pharmacy degree in 1989 at the same University.
She joined the Chemistry Department of UFRJ to obtain her master degree in science in the Organic Chemistry field, concluding her degree in 1992. After concluding her MS, she was selected by CNPq, a Brazilian government institution to work abroad as a visiting scholar for 2 years. Maria then joined the Research Institute of Pharmaceutical Science at University of Mississippi, in August of 1993 to carry on a research program on synthesis of antimalarial analogs of Artemisinin. At the end of this research program, Maria was invited by the Medicinal Chemistry department of Ole Miss to stay and join the PhD program.
Immediately upon receiving her PhD in August of 2000, she joined Eli Lilly & Company where she held various positions in both computational chemistry and discovery operations. Her current role is to develop computational models for Lilly’s Open Innovation Drug Discovery program (OIDD), a novel program for expanding open access for scientists external to the pharmaceutical industry.
Maria has published several scientific and computational chemistry papers in peer-reviewed journals. Maria was selected as one of the TOP 100 innovator of Eli Lilly & Company in 2013. She is also an active member of the Organization for Latin American group within Lilly.
Chris completed a PhD in synthetic organic chemistry under Professor Kocienski at Southampton University, UK in 1991.
For the following 20 years he worked at Pfizer as a medicinal chemist leading teams that delivered development candidates across a number of therapeutic areas including cardiovascular, pain and sexual health. He led Pfizer’s exploratory portfolio for anti-infectives which spanned from identifying leads for novel viral targets across to delivering potential clinical candidates for HIV. His work has targeted enzymes, receptors and ion channels using small molecules, peptides and antibody and PEG conjugates.
A recurring theme during his time at Pfizer was a focus upon applying in silico approaches to improve efficiency and to leverage early an understanding of ADME and toxicity in drug design for which he won a number of awards.
In 2011, a move from Pfizer to Lhasa turned this passion for in silico approaches to ADME and toxicity into a full time occupation. As the Director of Science, his focus has been on developing techniques to efficiently extract knowledge from data and applying this to the predictions of toxicity, metabolism and degradation.
After a DPhil in Chemistry at the University of Oxford, Tim gained experience as both a medicinal chemist and a molecular modeler at Peptide Therapeutics and Medivir. Tim joined Cresset in 2002. As Director of Products he is responsible for delivering easy to use applications that solve key problems in small molecule drug design and discovery. We use molecular fields to describe molecules in terms of binding rather than as 2D structures enabling the virtual screening, scaffold hopping, and smart molecule and library design.
Dr Greene received his BSc in Chemistry and Computational Science from the University of Leeds (UK) in 1991. He was awarded his PhD also from the University of Leeds (UK) in 1994 where his work focused on the synthesis of organophosphorus ligands and transition metal complexes for the catalysis of acetic acid production.
Following on from his PhD, Nigel worked at Lhasa Ltd. in Leeds promoting the use and development of the Derek for Windows computer program for toxicity prediction. After 5 years at Lhasa, Nigel then moved to Tripos Inc. where he promoted the use of molecular modeling software within the pharmaceutical industry.
In 2001, Dr Greene moved to the USA to work in Pfizer’s Drug Safety Research and Development department in Groton, CT where he lead the Computational Toxicology group aimed at developing new methods and tools for in silico toxicity assessment to be used as part of the drug discovery and development process. With the formation of the Compound Safety Prediction group in 2009, Dr Greene transitioned to this new function within Pfizer as part of Worldwide Medicinal Chemistry where he leads the computational toxicology group focusing on computational efforts to address early assessment of safety in the Pfizer portfolio.
In addition to his current role, Nigel’s other activities outside of Pfizer include the current Chair of the Board of Trustees for Lhasa Ltd. and has served on the National Research Council (NRC) committee that develop A Framework to Guide Selection of Chemical Alternatives sponsored by the US Environmental Protection Agency and has more recently asked to serve on the NRC committee tasked with developing a recommendation for the ‘Incorporation of 21st century science into Risk-Based Evaluations’.
Andreas Klamt studied physics in Göttingen and received his diploma degree in 1984 in theoretical metal physics. Then he moved on to the Max-Planck-Institute for Metal Research in Stuttgart, where he received his PhD in 1987.
After his PhD he directly started to work at Bayer AG in Leverkusen in the area of Computational Chemistry. Being involved in many different projects ranging from environmental distribution modelling, NMR-shift prediction, and chiral separation to drug design, he specialized on solvation and physical property calculation and developed the methods COSMO and COSMO-RS which meanwhile are widely used in the computational chemistry community. After being head of the central department for computational chemistry at Bayer for 3 years, he left Bayer in 1999 and founded COSMOlogic, a company for computational chemistry and fluid phase thermodynamics software and consulting, which since then he grew from a two people start-up to a company with 14 employees.
Since 2005 Andreas Klamt also is lecturer at the institute of Physical and Theoretical Chemistry at University of Regensburg and teaches block courses on solvation modelling and physical property predictions in solution. In 2012 he was appointed as honorary professor.
Paul Leeson is an independent medicinal chemistry consultant with more than 35 years’ experience in major pharmaceutical companies. He obtained BSc and PhD degrees in chemistry from the Universities of Liverpool and Cambridge, followed by postdoctoral work at Sussex University. His industrial career began at Smith Kline and French Research Laboratories, and has taken him to Merck Sharp and Dohme, then to Wyeth (USA) and from 1997-2011, AstraZeneca, where he was head of medicinal chemistry at the Charnwood site and leader of AstraZeneca’s Global Chemistry Forum. From 2011-14 he was a consultant for GlaxoSmithKline. Paul’s drug discovery contributions have been in the cardiovascular, neuroscience, respiratory and inflammation therapy areas and he has a special interest in compound quality. In 2014 he received the Nauta Award for Pharmacochemistry from the European Federation for Medicinal Chemistry (EFMC).
Matt is CEO of Optibrium. He has a Master of Science in computation from the University of Oxford and a PhD in theoretical physics from the University of Cambridge. As Associate Director at Camitro (UK), ArQule Inc. and then Inpharmatica, he led a team developing predictive ADME models and state-of-the-art intuitive decision-support and visualization tools for drug discovery. In January 2006, he became responsible for management of Inpharmatica’s ADME business, including experimental ADME services and the StarDrop software platform. Following acquisition of Inpharmatica, Matt became Senior Director responsible for BioFocus DPI’s ADMET division and in 2009 led a management buyout of the StarDrop business to found Optibrium, which develops software to intuitively guide decision-making through complex and uncertain data.
Dr. Chase Smith is currently an Associate Professor of Medicinal Chemistry at MCPHS University School of Pharmacy (Worcester, MA). Prior to this, he held various Senior Scientist positions over a 12 year span at Forma Therapeutics, Microbiotix, Inc. and GPC Biotech focusing mainly on early stage drug discovery programs. He has also held teaching positions at The College of the Holy Cross (Worcester, MA) and Ohio Northern University (Ada, OH) Additionally, Dr. Smith is a Reserve U.S. Navy Engineering Duty Officer and has been deployed to East Africa as part of the Combined Joint Task Force, Horn of Africa.
COSMOlogic is a small and dynamic company operating in the border region of computational chemistry, physical chemistry, chemical engineering and thermodynamics. Our vision is to use the powerful, modern techniques of quantum chemistry and molecular modelling to solve practical problems of chemists and chemical engineers in industry and academia.
Cresset provides chemistry software and consulting services to the pharmaceutical, agrochemical, flavor, fragrance and chemical industries. Cresset’s patented molecular comparison methods enable chemists to design and optimize the best molecules for their projects. Use Cresset’s software to visualize electrostatics to make better chemistry decisions, analyse SAR data to pinpoint and understand activity and selectivity cliffs, build more comprehensive chemical libraries, run virtual screens from the desktop and scaffold hop to new series and non-obvious new IP. Cresset’s highly experienced consulting team offers flexible working models and have expertise in all areas of ligand and structure-based drug design.
Lhasa Limited is a not-for-profit organization that facilitates collaborative data sharing projects in the pharmaceutical, cosmetics and chemistry-related industries. A pioneer in the production of knowledge-based systems for forward thinking scientists, Lhasa Limited continues to draw on over thirty years of experience to create user-friendly, state of the art in silico prediction and database systems. Lhasa Limited believes in ‘Shared Knowledge, Shared Progress’ and its not-for-profit, member driven status is designed to facilitate collaborative working and confidential data sharing between organizations. Lhasa’s software solutions include Derek Nexus and Sarah Nexus for predicting toxicity, Vitic Nexus for managing chemical information, Meteor Nexus for predicting metabolic fate and Zeneth for predicting forced degradation pathways.
Optibrium provides elegant software solutions for small molecule design, optimization and data analysis. Optibrium’s portfolio of products includes StarDrop, Sentira and Asteris. The company’s flagship platform, StarDrop, brings confidence to the selection and design of high quality candidate compounds. StarDrop creates an intuitive, highly visual and flexible environment to facilitate and speed up lead identification and optimisation, quickly targeting effective candidate compounds with a high probability of success downstream.
Founded in 2009, Optibrium continues to develop new products and research novel technologies to improve the efficiency and productivity of the drug discovery process. Optibrium works closely with its broad range of customers and collaborators that include leading global pharma, agrochemical and flavoring companies, biotech and academic groups.
Keynote presentation: Understanding Compound Quality, Dr Paul Leeson, Paul Leeson Consulting Ltd
Practical Application of Multi-Parameter Optimization to Guide Successful Drug Discovery, Dr Matt Segall, Optibrium Ltd
Improving the Plausibility of Success in Drug Discovery with the Use of Inefficient Metrics, Dr Mike Shultz, Novartis
Analyzing Selectivity Through Multi-dimensional Activity Cliff Analysis, Dr Tim Cheeseright, Cresset
Predicting Metabolites – Enhancing an Expert System with Machine Learning, Dr Christopher Barber, Lhasa Limited
Predicting Adverse Drug Reactions: What Works and What Doesn’t, Dr Nigel Greene, Pfizer
COSMO Polarization Charge Densities as Key Information for Solubility and Partitioning Property Prediction and 3D-QSAR, Prof Andreas Klamt, COSMOlogic
Modeling of Chemical and Physical Stability of Pharmaceuticals, Dr Yuriy Abramov, Pfizer
Development of a Drug Discovery Simulation Laboratory Exercise in the Pharmaceutical Sciences Graduate Program Curriculum, Dr Chase Smith, Massachusetts College of Pharmacy and Health Sciences