Exelgen Discovery

Exelgen Discovery

Operational Overview

Bude-Stratton Business ParkBudeCornwall

Copyright 2009 Exelgen Discovery CONFIDENTIAL2

Exelgen Discovery

Exelgen Discovery background

• Building on the foundations of Exelgen originally established in 1997

• Based in UK with partners in US - 20 scientific personnel plus additional support staff

• Experienced drug discovery Contract Research Organisation

– Chemistry, Design, Data Analysis and Mining

Background


Exelgen Discovery

To work with Life Science Companies to help discover compound(s) which have

– the appropriate biological activity

– the appropriate target selectivity

– an acceptable ADME-Tox profile

– a patentable position

Lead Optimization

Hit Validationor

Lead FindingHit Finding

All in as short a time as possible

Exelgen Discovery’s Core Business

Background


Exelgen Discovery

Exelgen Discovery’s Offerings

• Collaborative Services– Hit Finding and File Enhancement– Hit Follow-up/Hit-to-Lead– Lead Optimisation and LeadHopping– Project Consultancy & Management

• Contract Research• Screening Libraries• Novel Intermediates and Building Blocks• Custom Synthesis• Custom Analysis & Purification

Background


Exelgen Discovery

Library Enhancement Strategy… analyze existing collection

Corporate Screening Collection

Desired Compounds

Acceptable Compounds

Unacceptable Compounds

Poor ChemistryUndesirable PropertiesDiscard

Good ChemistryDrug-like Properties

Screen

Good ChemistryLead-like Properties

Buy

Compounds can be further subdivided by target

Collaborative Services


Exelgen Discovery

Toward Lead-like or Targeted Subsets… what is desirable

Poor absorption or permeation of an orally administered drug is more likely to occur if any two of these criteria are violated:

– Molecular weight is greater than 500– Lipophilicity is high (ClogP is greater than 5)– Number of Hydrogen bond donors is greater than 5– Number of Hydrogen bond acceptors is greater than 10

Compounds in a screening set should have drug-like or lead-like properties

Properties of Oral Drugs Categorized by Gene Family

Hopkins, et al, Nature Biotechnology 2006, 7, 805-815

90% MW

90% ClogP

90% HBD

90% HBA

90% Rbond

s

Aminergic GPCRs 460 5.6 2 6 8

Ion Channels 430 4.7 3 6 7

Nuclear Hormone Receptors 495 7.3 2 6 10

Peptide GPCRs 752 6.5 8 10 17

Phospho-diesterases 465 5.2 2 8 9

Protein Kinases 505 5.7 4 7 9

Serine Proteases 572 4.8 4 8 12

Lipinski’s “Rule of 5” is the best known filtering criteria

There are MANY others

=> Rules need to be tailored to specific customers needs



Exelgen Discovery

Enhancing a Compound Collection… analyze vendors

• Process vendor collection in same manner as corporate collection• Produce a lead-like subset• Compare corporate collection to vendor collection

– Eliminate any vendor compounds that are within specified cut-off distance of corporate collection

• Cluster remaining lead-like, novel subset– Grid spacing for vendor collection often looser than for corporate

collection

– Can also fill-in clusters with low occupancy of corporate compounds

• Select compounds from clusters based on client preferences– Preferred vendors– Best properties– Best price– Purity



Exelgen Discovery

Enhancement Can Be Tailored… only buy what’s needed

• Select sequentially– Preferred Vendors– Preferred Targets

• Select based on target– Similarity to known actives– Privileged substructures– Meet pharmacophore model– Meet SAR model

• Select based on properties – Preferred vendors– Best properties– Best price– Purity

Fill-in holes in chemistry space

Include areas not covered by original collection

Covered by Corporate Collection



Exelgen Discovery

Synthesis of Enhancement libraries… Sources of Scaffolds

ClogP

0

500

1000

1500

2000

2500

3000

3500

4000

4500

-1.75 -1.25 -0.75 -0.25 0.25 0.75 1.25 1.75 2.25 2.75 3.25 3.75 4.25 4.75

From “Ideas” Database

(~2500 ideas)

From Chemists“chemical intuition”

From “de novo” Scaffold

Generation

Scaffolds should be vetted for applicability

Scaffold Idea

Representative subset of reagents

“mini” virtual library

Check that library has reasonable property profile and no overlap with known compounds

For targeted libraries, use docking or similarity scores to verify desirability of the potential library

Shape-based hierarchical clustering is employed to select diverse scaffolds

27,417 meet “rule of 5”14,322 meet “rule of 4.5” 3,918 meet “rule of 4”using Aldrich reagents

ClogP MW

0

1000

2000

3000

4000

5000

6000

7000

8000

212.5 237.5 262.5 287.5 312.5 337.5 362.5 387.5 412.5 437.5 462.5 487.5

MW



Exelgen Discovery

Careful Reagent Selection… enhance synthetic success

Reagent Pool

Vendor Catalogs

Library Reagents

Depending on the library design goals, custom synthesized andnovel in-house reagents may also be used

Reaction-CompatibleReagents must be compatible with reaction conditions in current and following reaction steps

Reagant-CompatibleReagents must not contain competing functional groups

Drug-likeCannot contain known toxicophores, etc.

ReasonableAvailable at reasonable cost



Exelgen Discovery

Type of Library Needed Depends on What is Known

Lead Optimization

Hit Validationor

Lead FindingHit Finding

Oprea, Ed. ChemoInformatics in Drug Discovery, Wiley, 2005, p. 45K

no

wle

dg

e o

f T

arg

et

ProteinX-ray

Pharma-cophore

ProteinClass

None

Structure Based Design

Pharmacophore Based Design

Focused Sets

Diverse Libraries

Need for diversity is inversely proportional to

knowledge about the target

Degree of Diversity Needed

and where you are in the drug-development process



Exelgen Discovery

Focused libraries are rapidly generated by combining high quality sources of library ideas with techniques for filtering these ideas using knowledge of the target.

The process is iterative. Each new library adds to the knowledge base. So good SAR will be built into the libraries.

Lead Lead MoleculesMolecules

SynthesisSynthesis

ScreeningScreening

LibraryLibraryDesignDesignComputer DesignComputer Design

FocusedFocusedLibraryLibrary

CandidateCandidatePoolPool

Virtual ScreeningVirtual Screening

Ideas Ideas DatabaseDatabase

Knowledge Knowledge BaseBase

ChemistChemistIdeasIdeas

de novo de novo DesignDesign

Rapid Development of Focused Libraries



Exelgen Discovery


Ideas can come directly from the chemist or other sources, based on knowledge or intuition.

The ideas database consists of reaction protocols for scaffolds, whose analogs have been made or synthetic route has good precedent, combined with sets of well vetted reagents.

Our de novo tools use target and synthetic knowledge to generate relevant suggestions.


SynthesisSynthesis

ScreeningScreening











Exelgen Discovery


The knowledge base is a continually evolving repository of knowledge about the desired target.

Information may come from public sources, such as crystal structures or patents.

Information may also come from the customer, for example, results of earlier screening programs against the target of interest.


SynthesisSynthesis

ScreeningScreening











Exelgen Discovery


The ideas are coded into “virtual” libraries, providing a pool of millions of compounds as potential candidates.

How the virtual screening is done depends on the knowledge of the target, methods include:

• Property profiling• ADMET screening• Virtual docking• 2D & 3D similarity comparisons


SynthesisSynthesis

ScreeningScreening











Exelgen Discovery


The compounds that survive the virtual screening are all potentially active in the target.

Generating the library design is the process of reducing this still potentially very large pool to a subset of compounds that can be synthesized, have reasonable reagent reuse, have the best potential to show activity and incorporate SAR to help understand the screening results.


SynthesisSynthesis

ScreeningScreening











Exelgen Discovery


Converting a library design into a physical set of compounds to screen is an interactive process involving design and medicinal chemistry.

As the synthesis proceeds and the chemistry is better understood, some reagents may not be viable.

This knowledge is fed back to design, and replacements chosen to insure the final compounds retain the same level of desirability as the initial set.


SynthesisSynthesis

ScreeningScreening











Exelgen Discovery

Initial screening results in a number of hits and neighborhood analysis can be used to differentiate true positives from false positives.

IC50 data from identified leads can be fed back into the knowledge base to identify additional libraries to synthesis.

The leads can also be optimized using more refined techniques such as QSAR/CoMFA to optimize activity or pharmacological properties.


SynthesisSynthesis

ScreeningScreening












Exelgen Discovery

Focused Compound Selection

Compounds can be extracted from Existing Collections=> Library Enhancement– Greater compound diversity to choose from– Limited IP

Compounds can be Synthesized=> Library Design/Synthesis– Typically 200-300 compounds per scaffold– Novel IP Position

Techniques for Library Design=> Focus on techniques applicable to large virtual libraries– Large pool of candidates increases probability of finding good leads– Methodology

• High-throughput virtual docking• High-throughput similarity searching• Property profiling• CoMFA prediction on virtual libraries



Exelgen Discovery

Extracting a Focused Set from a Vendor Collection

Grid chemical space with physiologically relevant descriptor and create clusters

Select compounds in same clusters as known actives

0

0.5

1

1.5

2

2.5

1.00 0.95 0.90 0.85Neighborhood principle: small change in biological activity correspond to small change in valid metric

Use neighborhood principle to find valid metrics

– 2D fingerprints– BCUTs– Topomer fields



Exelgen Discovery

Scaffold Selection

All sources of scaffolds can benefit from analysis

Process most important for scaffolds with the least prior medicinal chemistry input

Representativereagents

Pool ofscaffolds

VirtualLibraries

Pool of Scaffoldcandidates

Desiredscaffold subset Appropriate

physicalproperties

LeadlikeADME/tox

Appropriatematch to

target

Efficiencyand costof design

SimilarityDockingModels

Diversity analysisMatch to target

Novelty

OptDesign

Goal is to rank potential scaffolds on ability to produce a good pool of synthetic candidates. This can be very fast using our virtual screening technology



Exelgen Discovery

Generating a Focused Library Design

• Collect knowledge of the target – Literature data– Customer proprietary/screening data

• Collect a pool of synthetic routes to novel scaffolds – Chemist suggestions– Ideas database– De novo scaffold generation

• Assess scaffold ideas– Generate representative libraries– Using target knowledge to assess potential activity

• Design libraries around best scaffolds– In collaboration with chemists to ensure most relevant

products



Exelgen Discovery

Full Library Design

SyntheticSynthetic SchemeScheme Reagent PoolReagent Pool

Virtual LibraryVirtual Library Monomer PoolMonomer Pool

Drug/Lead-like PoolDrug/Lead-like Pool

Initial DesignInitial Design

Final DesignFinal Design

Extract/TransformFilter• Reagent Compatible• Drug-like• Chemist Inspection

Reaction CompatibleDrug/Lead-likeChemist Inspection

Initial Design is Reviewed• Synthetic success validating reagents• Customer feedback on desirability of productsSeveral iterations needed if chemistry difficult

Extract Subset• Similarity to target• Docking scores• Pharmacophore match

Enumerate Library

Property Filtering

Library Definition

Add TargetKnowledge

Extract Design

Review



Exelgen Discovery

Physical Property Profiling

Poor absorption or permeation of an orally administered drug is more likely to occur if any two of these criteria are violated:

– Molecular weight is greater than 500– Lipophilicity is high (ClogP is greater than 5)– Number of Hydrogen bond donors is greater than 5– Number of Hydrogen bond acceptors is greater than 10

Typically libraries are designed to meet Lipinski’s “Rule of 5”

Properties of Oral Drugs Categorized by Gene Family

Hopkins, et al, Nature Biotechnology 2006, 7, 805-815

90% MW

90% ClogP

90% HBD

90% HBA

90% Rbond

s

Aminergic GPCRs 460 5.6 2 6 8

Ion Channels 430 4.7 3 6 7

Nuclear Hormone Receptors 495 7.3 2 6 10

Peptide GPCRs 752 6.5 8 10 17

Phospho-diesterases 465 5.2 2 8 9

Protein Kinases 505 5.7 4 7 9

Serine Proteases 572 4.8 4 8 12

But in reality properties need to be tailored to target being addressed



Exelgen Discovery

ADMET Predictions

Delaney, J. S. J. Chem. Inf. Comput. Sci. 2004, 44, 1000 – 1005.

ESOL – Estimated Aqueous Solubility

0%

1%

2%

3%

4%

5%

6%

7%

8%

9%

10%

-3.0 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 1.5 2.0

Poor BBBPermeability

Good BBBPermeability

Moderate BBB Permeability

-1.5

-1.0

-0.5

0.0

0.5

1.0

1.5

2.0

-2 -1.5 -1 -0.5 0 0.5 1

High bioavailability(good solubility and

Permeability)

Low bioavailability(poor solubility and

permeability)

Problematic bioavailability(poor solubility andgood permeability)

Problematic bioavailability(good solubility andpoor permeability)



Exelgen Discovery

Identify Reasonable Binding Modes

• Cluster on RMS distance between docked structures

• Visually inspect examples from each mode

Virtual Docking

Dock a Diverse Subset of the Virtual Library

• Start with filtered virtual library• Select by diversity or similarity

Dock Complete Virtual Library• Use identified base poses• Select products to synthesize using

scores and similarity

Best Worst Avg

P38 -25.6 2.8 -11.6

CDK2 -26.7 2.7 -12.1

Abl -32.7 -0.8 -17.2

FGFr -21.9 -5.1 -14.4

Mode 1

Mode2Mode 3

Representative Structures in Abl



Exelgen Discovery

NF O

O

N+

H

Topomer Distances

Split molecule into 2-3 fragmentsRule-based alignment of fragments onto a gridUse probe atom to calculate steric potential at grid points

7.6 uM

Query Structure

Lead Hop

D4.4

Topomer Fields

Topomer distances are the sum of the pair-wise differences between the fields summed over the fragments plus alignment and steric penalties

An example result from a Tripos validation study is shown above

Similarity Searching

Topomer based searching is effective in searching large virtual libraries

Topomer fields can also be used for CoMFA predictions in virtual libraries



Exelgen Discovery

Example - Tagamet and Zantac: both do the same thing in the body

Sulfur

Ring

Nitrogen rich section

What do they have in common in 3D?

H-bond Donor

H-bond AcceptorHydrophobe

What do they have in common in 2D?

Pharmacophore Modeling

Goal of pharmacophore modeling is to find matches to key 3D features



Exelgen Discovery

• Design is an Iterative process– What pleases the chemist often doesn’t please the computer and vice versa– All designs are a balance of competing requirements

• Designs have a good pedigree– Developed from synthetic routes provided by experienced medicinal

chemists– Reagent and reaction filters insure that designs can be converted to

products– Modification of scaffold and/or reagents are done in collaboration with

chemists to insure high value products are actually made• High throughput techniques allow for thorough investigation of options

– Large numbers of ideas can be evaluated for appropriate properties, docking score and similarity to targets

• Goal is to obtain lead compounds with– Improved biological activity– Improved target selectivity– Acceptable ADME-Tox profile– Patentable position

Summary of Focused Library Design



Exelgen Discovery

Summary of the Complete Process

• Most compounds that enter the drug discovery process fail– Use target knowledge early to eliminate poor candidates

• Reduce the attrition rate through intelligent (informatics-based) application of appropriate tools– Efficient library design process

– Property filters and predictions

– Activity prediction tools (receptor and ligand based)

– Chemistry expertise and knowledge• Goal is to obtain lead compounds with

– Appropriate biological activity

– Appropriate target selectivity

– Acceptable ADME-Tox profile

– Patentable position• All accomplished in as short a time as possible



Exelgen Discovery

Design Environment

– Building & rapid searching of large databases

• Search criteria - fingerprint, topomer, feature matching

• Input from experimental results & availability of reagents

– Captures chemistry we have developed/conceived - integrated with synthetic validation

• Evolution of knowledge base

Strategic & Operational interface… optimal process engineering

Production Environment

– State of the Art facilities

• Parallel Medicinal Chemistry

• High-throughput robotics

• Automated analysis & purification

Both seamlessly integrated with in-house informatics system



Exelgen Discovery

We know which compounds to make

Design Environment

ProductionEnvironment

Difficult

Simple

Distant Close

Highest interest & earliest delivery

High interest & long delivery

Synthesis development & Reagent synthesis

Low interest & later delivery

Low interest & quick delivery

Every compound we make is designed and tracked via the informatics system



Exelgen Discovery

Project Consultancy & Management

• Consultancy– Work with client to establish project objectives– Analyze overall resourcing including any outsourcing– Identify major project milestones– Troubleshoot key issues– Define overall project plan and workflow

• Management– Help define scope & goals of project– Monitor progress against goals & budget– Maintain regular contact with client counterpart– Active member of project Steering Committee



Exelgen Discovery

Contract Research

• Flexible approach– FTE-based model gives client greatest flexibility in steering

project and optimising resource usage– “Mix and Match” chemistry, design, analysis and purification

as workflow dictates

• Dedicated project team – Project Manager assigned at earliest possible stage– Key named personnel assigned for duration of project– All documentation project/client specific– Regular project updates provided to client

• Dedicated laboratory space– Provision for sensitive projects to have dedicated lab space if

required

Contract Research


Exelgen Discovery

Screening Libraries

• Off the shelf– Cherry-pick from ~25000 compounds

– Catalogue continually being updated

• Custom request– Can be made to customer’s specifications

• All compounds designed

– Typical library size 200-2000 compounds

– Derived from database containing >2500 ideas

– Supplied on exclusive basis if required

– Typically >85% purity by LC-MS

Screening Libraries


Exelgen Discovery

Novel Intermediates, Building Blocks & Custom Synthesis

• Wide range of intermediates and building blocks – Many available from stock– Rapid turnaround from order on larger amounts– Analoging around specific series readily achieved

• Custom synthesis routinely undertaken– From mg to multi-kg scale– Scale-up trialling where no precedent exists

Novel Intermediates


Exelgen Discovery

Custom Analysis & Purification

• Custom NMR and LC-MS services– Can be tailored to customer’s specific requirements– Single run or batchwise processing

• Custom analysis & purification– Utilising customer defined protocols– Protocol development undertaken– Purified samples provided in customer’s desired

format

Custom A&P


Exelgen Discovery

Summary

Exelgen Discovery provides a broad spectrum of competitively priced Discovery Products and Services. For further information or enquiries please contact:

Exelgen DiscoveryBude-Stratton Business ParkBude EX23 8LYCornwall, UK

Dr Phil BillingtonBusiness Development

Email:Mob:

Dr Julian SmithPrincipal Scientist

Email:Mob:

Tel/Fax: +44 (0)1288 356500

[email protected]+44 07973 493403

[email protected]+44 07805 571662

Exelgen Discovery

Documents

Transcript of Exelgen Discovery