Future Directions in Drug Future Directions in Drug Design & Discovery Design & Discovery [email protected][email protected][email protected][email protected][email protected][email protected]Jonathan S. Mason Jonathan S. Mason Head of Computational Chemistry Head of Computational Chemistry Heptares Heptares Therapeutics, Welwyn Garden City, UK & Therapeutics, Welwyn Garden City, UK & Chief Scientist Chief Scientist (Predictive Technologies & Drug Design) Medicinal Chemistry (Predictive Technologies & Drug Design) Medicinal Chemistry Lundbeck Research, Lundbeck Research, Valby Valby, Copenhagen, Denmark , Copenhagen, Denmark The menu The menu … … The drug discovery problems The drug discovery problems The drug design problems The drug design problems The opportunities The opportunities
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Future Directions in Drug Future Directions in Drug Design & DiscoveryDesign & Discovery
The drug discovery problemsThe drug discovery problemsThe drug design problemsThe drug design problemsThe opportunitiesThe opportunities
Attrition / ProductivityAttrition / Productivity
The Drug Discovery ProblemThe Drug Discovery Problem
Need to Need to -- choose the right targetschoose the right targets
-- identify suitable hitsidentify suitable hits-- progress rapidly to a molecule with the progress rapidly to a molecule with the
desired activity and propertiesdesired activity and propertiesthat does not fail (that does not fail (attritattrit) during ) during development development
The Drug Discovery ProblemThe Drug Discovery Problem
Toxicity Toxicity –– Safety IssuesSafety Issues-- large 5 year EU funded project large 5 year EU funded project
(IMI (IMI eToxeTox) to develop models based on ) to develop models based on pooled data from all major pooled data from all major pharmapharmacompanies companies (4 week rat)(4 week rat)
Drug DiscoveryDrug Discovery
NCELeadOptimization
Hit LeadIdentificationGenomics Target
IdentificationTarget
Validation
Phase I & IIPhase I & IIPhase I & II Phase IIIPhase IIIPhase III
IdeaIdeaIdea DrugDrugDrug
Drug/Biological Target Space
CompoundIdentification
TargetIdentification
TargetValidation
Target equityTarget equity
Chemical equityChemical equity
LeadSelectivity
Lead Differentiation
Lead equityLead equity
Risk managementRisk management
Design, make & screen compounds that are Design, make & screen compounds that are ““drugdrug--likelike”” & & ““leadlead--likelike””–– and have biological activity !and have biological activity !
Tackle Tackle attrition attrition && productivity at an early stage productivity at an early stage 100 Ideas 100 Ideas --------------------------1010------------------------ 11 NCENCE
Number 2:Number 2:Continue to think as much about Continue to think as much about properties as properties as
potency (physicochemical, LE, LLE etc)potency (physicochemical, LE, LLE etc)MultiMulti--parametric optimisationparametric optimisation
But nowBut now with kinetics with kinetics (off(off--rates etc)rates etc) from an from an early stage early stage –– from biophysical screening from biophysical screening (Surface Plasmon (Surface Plasmon Resonance(SPR)/BiacoreResonance(SPR)/Biacore etc)etc)
Future trends: The opportunitiesFuture trends: The opportunities……
Number 3:Number 3:Apply Apply ““rational designrational design””, SBDD to the major , SBDD to the major drug target class drug target class GPCRsGPCRsStabilise the GPCR to the desired conformation Stabilise the GPCR to the desired conformation (agonist, antagonist etc)(agonist, antagonist etc)Finally gain understanding at the atomic level of Finally gain understanding at the atomic level of antagonismantagonism--agonismagonism & use that in design& use that in designSubSub--type selectivity type selectivity SubSub--state selectivitystate selectivity
Fragment Fragment screeningscreening
for mono or polyfor mono or poly--pharmacologypharmacology
XX--ray ray structures structures with with ligandsligands
Future trends: The opportunitiesFuture trends: The opportunities……
Number 4:Number 4:Drive dataDrive data--driven design driven design in in cerebrocerebroAid intuition, experience, knowledge with Aid intuition, experience, knowledge with target derived information target derived information ((egeg hotspots for hotspots for functional groups, unfavourable waters etc)..functional groups, unfavourable waters etc)..
GRID GRID energetic energetic survey of survey of
binding sitebinding site
BindingBinding--site site water water
energeticsenergetics(GRID, (GRID,
WatermapWatermap……))
Using 3D stereo!Using 3D stereo!Now possible with NO Now possible with NO
loss of quality or loss of quality or interferenceinterference
(PLANAR)(PLANAR)
Future trends: The opportunitiesFuture trends: The opportunities……
Number 5:Number 5:New energy and datasets to model toxicityNew energy and datasets to model toxicityIMI IMI eToxeTox project with input from many key project with input from many key players and data pooled from all major players and data pooled from all major PharmaPharmawith European site.with European site.
Hopefully models based on more than chemical Hopefully models based on more than chemical substructures substructures ……
Future trends: The opportunitiesFuture trends: The opportunities……
Number 6:Number 6:Think about structures based on how proteins Think about structures based on how proteins see them, not how we synthesise them !see them, not how we synthesise them !2D 2D 3D 3D Biological profile/fingerprintBiological profile/fingerprint
Future trends: The opportunitiesFuture trends: The opportunities……
OffOff--target target screeningscreening
++In In silicosilico
Prediction ofPrediction ofoffoff--targettarget
profilesprofiles
Number 7:Number 7:
SimulationSimulationGet back to using molecular dynamics etc,Get back to using molecular dynamics etc,
-- bothboth the lock and the key are flexible & solvated!the lock and the key are flexible & solvated!
ModelModel RefineRefine Predict freePredict free--energy energy of bindingof binding
Future trends: The opportunitiesFuture trends: The opportunities……
Number 8:Number 8:Integration of data Integration of data –– now plenty available!now plenty available!Integration of methods, use to solve what the Integration of methods, use to solve what the real problem/bottleneck is, not restricted to real problem/bottleneck is, not restricted to ““traditionaltraditional”” software & methodssoftware & methods
Bioassay data for activity, Bioassay data for activity, ADME, ADME, ToxicicityToxicicity
External databasesExternal databases
Future trends: The opportunitiesFuture trends: The opportunities……
““Computational chemistryComputational chemistry””””Computational Medicinal Chemistry Computational Medicinal Chemistry ComputerComputer--Aided Drug Design/Discovery (CADD)Aided Drug Design/Discovery (CADD)
Number 9:Number 9:
PharmacophorePharmacophore--based concepts:based concepts:-- now for both proteins & now for both proteins & ligandsligands-- GRID energetic analysis of binding sites with GRID energetic analysis of binding sites with
functional groups present in drug moleculesfunctional groups present in drug moleculesVirtual ScreeningVirtual Screening-- more structuremore structure--based but based but ligandligand--based with based with relevant descriptors still useful for hit/tool relevant descriptors still useful for hit/tool findingfinding
Future trends:Future trends: Still GoodStill Good……
GRID mapsGRID maps--The real onesThe real ones……
PharmacophoresPharmacophores from from LigandsLigands & Proteins& ProteinsFLAPFLAP(new(new method from Mol. Discovery / U. Perugia)method from Mol. Discovery / U. Perugia)a)a) Molecular Interaction Fields (MIF) calculated in the active siteMolecular Interaction Fields (MIF) calculated in the active site of a of a
protein structure using GRIDprotein structure using GRIDb)b) MIF is condensed into a few TargetMIF is condensed into a few Target--based based pharmacophoricpharmacophoric points. points. c)c) All possible arrangements of 4 All possible arrangements of 4 pharmacophoricpharmacophoric points are generatedpoints are generated
StericSteric hindrance as a hindrance as a filter for the many filter for the many solutions found for solutions found for each each ligandligand when when ““dockeddocked”” in the in the protein active site.protein active site.
A A moleculemolecule/fragment /fragment willwill bebe posedposed//fittedfittedusingusing meaningfulmeaningfulcontactscontacts ……((pharmacophoricpharmacophoric) b) b
An early example of fragment An early example of fragment screening screening better leadsbetter leads
Factor Factor XaXa serine protease serine protease –– the transformation binding mode the transformation binding mode ““flipflip””
NH2 N H
NN
O
S
O
O
S C lS
O
O
S
C l
N H2
N H
N
N
O
π-cationinteraction
Displaces an energetically “unhappy” water in a lipophilic hotspot
Hydrophobic / lipophilicinteractions
Cation-anion intertaction withCatalytic aspartate
--
A great fragment, that could have been A great fragment, that could have been foundfound……..
An early example of fragment An early example of fragment screening screening better leadsbetter leads
GRID maps show GRID maps show lipophiliclipophilic hotspot in S1hotspot in S1-- needed to ignore dogma, forceneeded to ignore dogma, force--fittingfitting
-- Through dogma of S1 needing a basic group did not screen widely Through dogma of S1 needing a basic group did not screen widely or follow up on or follow up on in in silicosilico reversed binding and hotspotsreversed binding and hotspots
PolypharmacologyPolypharmacology / promiscuous behaviour is one of the / promiscuous behaviour is one of the many issues that is more probable with many issues that is more probable with clogPclogP > 3> 3
The Protein PerspectiveThe Protein Perspective: : PolypharmacologyPolypharmacology && ““Off TargetOff Target”” EffectsEffects
Future trends: The opportunitiesFuture trends: The opportunities……
Weak binding
Strong binding
Fragments: Why small is beautifulFragments: Why small is beautiful
““FragmentsFragments”” or or ““ComponentsComponents”” of a drug molecule are reduced of a drug molecule are reduced complexity starting points complexity starting points --Enable a much larger proportion / diversity of relevant drug spaEnable a much larger proportion / diversity of relevant drug space ce to be exploredto be explored::
-- new targets, new indications (new targets, new indications (egeg CNS) without screening file biasCNS) without screening file bias
-- Good highGood high--LE starting points, even though existing within compoundsLE starting points, even though existing within compoundsin the screening file are often missed in the screening file are often missed ““dressed updressed up”” wrongly wrongly (( greatergreater chances for chances for HH--bondbond mismatchmismatch oror stericsteric clashclash etc.etc.))
LowerLower hithit--raterate for for HTSHTS--sizedsized cpdscpds, , highhigh--LELE fragments fragments rarelyrarely show show upup in hits.in hits.
ReducedReduced ligandligand complexitycomplexity increasesincreases thethechances chances ofof a match but binding a match but binding willwill bebe lowlowFragment screening delivers Fragment screening delivers highhigh effectiveeffectivechemicalchemical diversitydiversity from small from small librarieslibraries
Some of the questions to considerSome of the questions to consider……
Can we find Can we find ““betterbetter”” leads ? leads ? –– OptimisableOptimisable (non(non--flat SAR) flat SAR) -- Better properties, high Better properties, high ligandligand efficiency (LE), efficiency (LE),
& & ligandligand lipophiliclipophilic efficiencefficienc (LLE: (LLE: --logKilogKi –– clogPclogP or or clogDclogD))-- AllostericAllosteric binding sitesbinding sites……
Can we finding a hit/lead when HTS fails ? Can we finding a hit/lead when HTS fails ? –– Sampling of drug relevant chemical spaceSampling of drug relevant chemical space-- SensitiveSensitive screening method (screening method (llabelabel--free: biophysical)free: biophysical)
What do we need to be successful ?What do we need to be successful ?–– LearningsLearnings from the pastfrom the past--New methods for membrane proteins New methods for membrane proteins –– stabilization of stabilization of conformation(sconformation(s) of interest ) of interest biophysical screening, kinetics & biophysical screening, kinetics & XX--ray structuresray structures
Can we use on the major drug class of Can we use on the major drug class of GPCRsGPCRs ??–– Better leads Better leads (as above)(as above)-- Control Control polypharmacologypolypharmacology
A recent example of fragment screening A recent example of fragment screening better leads: BACE better leads: BACE aspartateaspartate proteaseprotease
Poor/no HTS hits Poor/no HTS hits (many companies...)(many companies...)HydroxyethylamineHydroxyethylamine (HEA) compounds difficult to (HEA) compounds difficult to
optimise to clean brain optimise to clean brain penetrantpenetrant compounds compounds (many companies)(many companies)
Fragment screening by many different biophysical Fragment screening by many different biophysical methods able to find diverse new different / reduced methods able to find diverse new different / reduced basicitybasicity ““warheadswarheads””
BiophysicalBiophysical screening platforms (screening platforms (e.ge.g. . XX--rayray, SPR, TINS) , SPR, TINS) usedused to to identifyidentify & & facilitatefacilitate evolution evolution ofof weakweak mMmM fragments to fragments to nMnM leadsleads
Fragments: Are all the screening Fragments: Are all the screening libraries similar ? libraries similar ?
Analysis of a partially built inAnalysis of a partially built in--house house ““handhand--pickedpicked””fragment library and the fragments from 3 companies fragment library and the fragments from 3 companies offering fragment screening gave surprising results:offering fragment screening gave surprising results:
-- Practically no overlap at the compound level between any Practically no overlap at the compound level between any of the commercial vendor librariesof the commercial vendor libraries
Different hits will be obtained by fragment screening at multipDifferent hits will be obtained by fragment screening at multiple le vendors (as well as any differences from different methods)vendors (as well as any differences from different methods)
-- Overlap at the compound level between inOverlap at the compound level between in--house library house library and all vendor librariesand all vendor libraries
-- All 3 vendor libraries contained significant numbers of All 3 vendor libraries contained significant numbers of medchemmedchem desirable/acceptable hitsdesirable/acceptable hits
Number 2:Number 2:Continue to think as much about Continue to think as much about properties as properties as
potency (physicochemical, LE, LLE etc)potency (physicochemical, LE, LLE etc)
But nowBut now with kinetics with kinetics (off(off--rates etc)rates etc) from an from an early stage early stage –– from biophysical screening from biophysical screening (Surface Plasmon (Surface Plasmon Resonance(SPR)/BiacoreResonance(SPR)/Biacore etc)etc)
Future trends: The opportunitiesFuture trends: The opportunities……
Kinetics of binding is an important factor for the Kinetics of binding is an important factor for the in in vivovivo efficacy of many drugsefficacy of many drugs-- Methods already set up for fragment screening & optimization, Methods already set up for fragment screening & optimization,
such as immobilised protein with SPR screening (such as immobilised protein with SPR screening (BiaCoreBiaCore) can be ) can be directly used to measure on and off rates during optimizationdirectly used to measure on and off rates during optimization
Better Leads from Fragments ? Better Leads from Fragments ?
Build up a molecule that has multiple properties optimised Build up a molecule that has multiple properties optimised for the desired for the desired target(starget(s):):
-- LigandLigand efficiency, properties efficiency, properties [[egeg for CNS for for CNS for ““difficultdifficult”” targets]targets]
To identify, characterize and optimize the fragment binding To identify, characterize and optimize the fragment binding a purified protein sufficiently stable for immobilization / a purified protein sufficiently stable for immobilization / screening is neededscreening is needed-- Ok in general for enzymes Ok in general for enzymes -- Previously very difficult for membrane bound proteins Previously very difficult for membrane bound proteins
such as such as GPCRsGPCRs: now possible (: now possible (StaRStaR™™: : Stabilized Receptors)Stabilized Receptors)
Number 3:Number 3:Apply Apply ““rational designrational design””, SBDD to the major , SBDD to the major drug target class drug target class GPCRsGPCRsStabilise the GPCR to the desired Stabilise the GPCR to the desired conformation (agonist, antagonist etc)conformation (agonist, antagonist etc)Finally gain understanding at the atomic level Finally gain understanding at the atomic level of antagonismof antagonism--agonismagonism & use that in design& use that in design
Fragment Fragment screeningscreeningfor mono or for mono or
polypoly--pharmacologypharmacology
Biophysical Biophysical mapping mapping to to determine determine
Future trends: The opportunitiesFuture trends: The opportunities……
Use of new stabilized receptors (Use of new stabilized receptors (StaRsStaRs) ) technology with technology with GPCRsGPCRs for biophysical mapping for biophysical mapping & X& X--ray structures with ray structures with ligandsligands
full SBDD full SBDD (Structure(Structure--Based Drug Design)Based Drug Design)
Future trends: The opportunitiesFuture trends: The opportunities……
Heptares Therapeutics
GPCR Drug Discovery CompanyGPCR Drug Discovery CompanyIntegrated GPCR drug discovery capabilityIntegrated GPCR drug discovery capabilityUnique stabilised receptor (Unique stabilised receptor (StaRStaR™™) technology and strong IP ) technology and strong IP Cross TherapeuticCross Therapeutic
Breakthrough stabilised receptor Breakthrough stabilised receptor StaRStaRTMTM technologytechnologyAddressing deficiencies in GPCR drug pipelines caused by poorly Addressing deficiencies in GPCR drug pipelines caused by poorly tractable GPCR tractable GPCR targetstargetsEnables FULL suite of contemporary drug discovery technologies tEnables FULL suite of contemporary drug discovery technologies to be applied to o be applied to GPCRsGPCRs
StaRStaRTMTM technology is being used to progress a pipeline of best or firstechnology is being used to progress a pipeline of best or first t in class molecules across the GPCR in class molecules across the GPCR superfamilysuperfamily
Small molecule and antibodies Small molecule and antibodies –– twintwin--track approachtrack approachAim to progress promising drug leads internally and through straAim to progress promising drug leads internally and through strategic collaborationstegic collaborations
WellWell--funded through funded through ££21M Series A fund raise completed Feb 200921M Series A fund raise completed Feb 200940 employees located in 40 employees located in BioparkBiopark, Hertfordshire UK, Hertfordshire UK
Receptors embedded in cell membrane exist in multiple Receptors embedded in cell membrane exist in multiple conformations conformations
Highly unstable when removedHighly unstable when removedNot suitable for structure based drug discovery methodsNot suitable for structure based drug discovery methods
HeptaresHeptares’’ technology is used to make a stabilized technology is used to make a stabilized versions of target versions of target GPCRsGPCRs ('('StaRStaRTMTM') held in a specific ') held in a specific chosen conformationchosen conformation
Stable in functionallyStable in functionally--relevant, purified formrelevant, purified form
Bindingassay
FunctionalAssay
AR AR* AR*G
R R* R*G
Heptares StaRTM Technology
Heptares StaRTM Technology
●● Receptor embedded in cell membrane, highly unstable when Receptor embedded in cell membrane, highly unstable when removedremoved•• Aggregates and loses function when purified in detergentAggregates and loses function when purified in detergent•• Current drug discovery approaches rely on cellCurrent drug discovery approaches rely on cell--based assaysbased assays
●● General method required to engineer stabilized, less flexible General method required to engineer stabilized, less flexible receptor held in specified functional conformationreceptor held in specified functional conformation
●● HeptaresHeptares’’ proprietary technology is used to make a stabilized proprietary technology is used to make a stabilized version of target GPCR ('version of target GPCR ('StaRStaRTMTM') held in a chosen conformation') held in a chosen conformation•• Stable in functionallyStable in functionally--relevant, purified formrelevant, purified form•• Dependent on Dependent on ligandligand usedused
RR RRnn R*R*InactiveInactive Partially Active Partially Active Fully ActiveFully Active
•• A GPCR containing a small number of point mutations A GPCR containing a small number of point mutations that greatly improve its that greatly improve its thermostabilitythermostability
-- Stable in purified, detergent solubilised formStable in purified, detergent solubilised form-- Functional and drugFunctional and drug--binding characteristics preservedbinding characteristics preserved-- Completely novel and proprietaryCompletely novel and proprietary-- Suitable for Suitable for uHTSuHTS, , BiacoreBiacore (kinetics), crystallisation etc.(kinetics), crystallisation etc.-- Transferrable across GPCR familiesTransferrable across GPCR families
0 10 20 30 40 50 60 70
0
20
40
60
80
100
120
A2A-WT
A2A-Rant22
A2A-SS3
Temperature oC
% s
pec
ific
[3H
] Z
M24
1385
bo
un
d
What is a StaRTM
Types of Types of StaRsStaRs
Antagonist Antagonist StaRStaR
Agonist state 1Agonist state 1StaRStaR
NAM NAM StaRStaRPAM PAM StaRStaR
Agonist state 2Agonist state 2StaRStaR
Inverse Agonist Inverse Agonist StaRStaR
StaRStaR proteins are locked in the conformation derived from the proteins are locked in the conformation derived from the pharmacology of the pharmacology of the ligandligand used in their creationused in their creation
Proprietary Process for Creating StaRs to stabilise GPCRs Fragment screening Kinetics,
Biophysical mapping, X-ray structures etc
UnstableNative receptor
SelectConformation
SelectLigand
Purifyprotein
CrystalliseNewGPCR structure
SPR/BiacoreTINS/Zobio...
Fragment screeningKinetics
Thermostabilityand Selection
AATCAGCTTGTCGAA
Mutagenesis
Iteration
StaRRecombination
of mutants
StaRsAg/Antag
Conformations
5 10 15 20 25 30 35 40 45 500
25
50
75
100
Temperature
Enabling technology for GPCR Drug Discovery
Cell Based AssaysHigh Throughput Screening
RecombinantReceptorMembraneAssays
In vitroPharmacologySelectivity
In vivoPharmacology Efficacy andReceptor Occupancy
SPR KineticsandThermodynamics
Modelling bySDM withRadioligand
Modelling byBiophysicalMappingTM
Fragment ScreeningSPR / NMR direct binding
In vitroPharmacologyIsolated ReceptorConformations
Entrance to ligand binding site well defined – high resolution
Drug binding pocket
Activation and G-protein binding region retainedMultiple loop conformations resolved cf biased agonism
9 drug co-crystal structures now solved in detergentAgonists and AntagonistsLow and High Affinity
R68S
M90V
F338M
F327A
A282L
Y227A
Biophysical Screening on GPCRs
StaRsStaRs broadly applicable to biophysical approachesbroadly applicable to biophysical approaches•• StabilityStability•• Protein productionProtein production
●● GPCR targeted fragment libraryGPCR targeted fragment library•• In In silicosilico for virtual screeningfor virtual screening•• In vitro In vitro for diversity screeningfor diversity screening
Common Drug Binding site in all Common Drug Binding site in all GPCRsGPCRs
Family AFamily Aamine, nucleotide,amine, nucleotide,Lipid, Lipid, eicosanoideicosanoid
Family A/BFamily A/BPeptidePeptide
ReceptorsReceptors
Family CFamily CMetabotropicMetabotropic glutamateglutamate
Drug Binding Site located in Drug Binding Site located in transmembranetransmembrane domaindomain
Considered Considered ‘‘allostericallosteric sitesite’’ for Family B and Cfor Family B and C
Family A Family A allostericallosteric site overlaps with site overlaps with ligandligand binding pocketbinding pocket
LigandLigand
Number 1:Number 1:Drive design with experimental data!Drive design with experimental data!Synergise, not compete..Synergise, not compete..
Future trends: The opportunitiesFuture trends: The opportunities……
Biophysical Mapping using Biophysical Mapping using BiacoreBiacore
Site directed mutagenesisSite directed mutagenesis 1010--30 Mutations to the30 Mutations to theBinding site regionBinding site region
Mutant Mutant StaRsStaRs displayed displayed on a on a BiacoreBiacore chipchip
Detection of binding Detection of binding of different of different ligandsligands to to each mutant each mutant StaRStaR
Biophysical MapBiophysical Map
of binding siteof binding site
LigandLigand refined refined Homology Model and Homology Model and prediction of proteinprediction of protein--ligandligand binding modesbinding modes
Correlating binding data from Correlating binding data from Multiple Multiple ligandsligands with multiplewith multipleMutant Mutant StaRStaR proteinsproteins Label freeLabel free
Direct binding methodDirect binding method
Biophysical Biophysical MappingMappingTM: Definition : Definition of binding modes of binding modes
Matrix of Matrix of ligandligand--construct binding construct binding
datadata
DataDatainterpretation interpretation and modellingand modelling
Biophysical maps for each Biophysical maps for each chemical serieschemical series
Decide between Decide between alternative docked alternative docked
posesposes
FragmentFragment--based drug discovery/design (FBDD) based drug discovery/design (FBDD) for for membrane proteins: membrane proteins: GPCRsGPCRs
Fragment screening has been performed using Fragment screening has been performed using stabilized stabilized GPCRsGPCRs, giving results comparable to that , giving results comparable to that obtained using soluble (enzyme) targetsobtained using soluble (enzyme) targets
The stabilized GPCR can be immobilized and used The stabilized GPCR can be immobilized and used for fragment screening using:for fragment screening using:-- BiacoreBiacore /SPR: /SPR: Surface Plasma Resonance BiosensorSurface Plasma Resonance Biosensor-- TINS (TINS (ZobioZobio): ): Target Immobilized NMR ScreeningTarget Immobilized NMR Screening
Fragment Screening: Fragment Screening: The new possibilities for enzymes & The new possibilities for enzymes & GPCRsGPCRs
GPCRsGPCRs
Select Select conformationconformation
StabilizeStabilize
StaRStaR™™
StaRStaR proteins are locked in the proteins are locked in the conformation derived from the conformation derived from the pharmacology of the pharmacology of the ligandligand used used
Identify binding Identify binding position / modeposition / mode
EnzymesEnzymes
Immobilized protein Immobilized protein fragment screeningfragment screening
SPRSPRTINS (TINS (ZobioZobio))
TINS method for finding hits: TINS method for finding hits: fragment screening fragment screening
Immobilized protein Immobilized protein –– only small amounts needed (~1mg)only small amounts needed (~1mg)Very sensitive: higher Very sensitive: higher nMnM hits identified hits identified (not found by SPR)(not found by SPR)
TINS = Target Immobilized NMR Screening: Zobio
TINS method for finding hitsTINS method for finding hits
Fragment Screening: Fragment Screening: NNew possibilities for ew possibilities for GPCRsGPCRs (+(+ enzymes) enzymes)
TINS screening of TINS screening of ZobioZobiofragment library for 2 fragment library for 2 enzyme targets + enzyme targets + ββ11--AR AR GPCRGPCR ::
Sensible looking hits for Sensible looking hits for all targets all targets
8080--100 from ~1400100 from ~1400for enzyme targetsfor enzyme targets
~80 from 580 ~80 from 580 for for ββ11--ARAR
Specific for each targetSpecific for each target-- Only one fragment hit in Only one fragment hit in common between GPCR & common between GPCR & enzyme target hitsenzyme target hits
StaR™ technology used to generate thermostabile mutant of β1AR.‐ Stable in purified, detergent solubilised form
‐ Functional and drug‐binding characteristics preserved
‐ Completely novel and proprietary
Approach: TINS - Screen immobilized, DDM solubilized β1AR vs OmpA
TINS screening using stabilized TINS screening using stabilized ββ1AR was able to 1AR was able to identify ~80 hits from 580 fragments screened:identify ~80 hits from 580 fragments screened:-- reasonable looking structuresreasonable looking structures-- including compounds where activity could be confirmed including compounds where activity could be confirmed by biochemical screeningby biochemical screening
Are the GPCR Fragment Screening Hits from Are the GPCR Fragment Screening Hits from TINS (TINS (ZobioZobio) generally reasonable / valid ?) generally reasonable / valid ?
Yes !Yes !
Many reasonable / interesting hitsMany reasonable / interesting hits
Activity confirmed for many by biochemical assayActivity confirmed for many by biochemical assay
Not the same as found for the other targetsNot the same as found for the other targets
Further work ongoingFurther work ongoing
GPCR Fragment Screening ExamplesGPCR Fragment Screening Examples
●● ββ11-- AdrenoceptorAdrenoceptor by TINS NMRby TINS NMR•• ZoBioZoBio library (Gregg library (Gregg SiegalSiegal))•• TINS screening using stabilized TINS screening using stabilized ��1AR was able to identify ~80 hits 1AR was able to identify ~80 hits
from 580 fragments screenedfrom 580 fragments screened•• Promising chemical structuresPromising chemical structures•• Including compounds where activity could be confirmed by biochemIncluding compounds where activity could be confirmed by biochemical ical
screeningscreening●● Adenosine AAdenosine A2A2A by SPR/by SPR/BiacoreBiacore
•• University of Utah (David University of Utah (David MyszkaMyszka))•• XanthineXanthine derivatives identified as hitsderivatives identified as hits•• 200 200 μμMM screening concentrationscreening concentration
A2A Rant224 mutations
StaRStaR immobilised on chipimmobilised on chip Challenged with FragmentsChallenged with Fragments(~150(~150--250 250 DaDa))
Fragments progressed Fragments progressed to structure modelto structure model
FragmentFragment--basedbased
Drug DiscoveryDrug Discovery
Fragments Fragments ‘‘growngrown’’ to to fill active site and fill active site and give potent novel leadsgive potent novel leads
O
N
NH
N
N NH2
N
N
N
NH
O
O
Fragment Screening using Biacore
SPR fragment screeningAnalysis of hits
Hits are further analysed byHits are further analysed byDetection of nonDetection of non--specific binding on control specific binding on control surface (surface (egeg denatured protein)denatured protein)Full concentration response curvesFull concentration response curvesBiophysical Biophysical mappingmappingTMTM
Analysis Analysis vsvs receptor with binding site mutationsreceptor with binding site mutationsCoCo--crystal structurescrystal structures
3D Structural Information on Industrial Timescale 3D Structural Information on Industrial Timescale HeptaresHeptares GPCR CrystallographyGPCR Crystallography
State of the art facility for GPCR crystallisation set upState of the art facility for GPCR crystallisation set upVapour diffusion crystallographyVapour diffusion crystallographyCoCo--crystallisation with multiple crystallisation with multiple ligandsligandsNanolitreNanolitre lipidiclipidic cupid phase roboticscupid phase roboticsSynchrotron sources at Diamond, ESRF, SLSSynchrotron sources at Diamond, ESRF, SLS
StaR Drug Discovery Applications
StaRsStaRsAg/Ag/AntagAntag
ConformationsConformations
SPR/SPR/BiacoreBiacore
Fragment screeningFragment screeningKineticsKinetics
LigandLigand coco--structuresstructuresBinding site definitionBinding site definition
XX--ray structureray structure
Mutagenesis dataMutagenesis dataBinding site mutantsBinding site mutants
Refined homologyRefined homologyModelModel
Virtual screeningVirtual screening
DockingDockingVirtual screeningVirtual screening
NMRNMR
Fragment screeningFragment screeningBinding site definitionBinding site definition
Residues involved in Residues involved in LigandLigand bindingbinding
Number 6:Number 6:Think about structures based on how proteins Think about structures based on how proteins see them, not how we synthesise them !see them, not how we synthesise them !2D 2D 3D 3D Biological profile/fingerprintBiological profile/fingerprint
Future trends: The opportunitiesFuture trends: The opportunities……
OffOff--target target screeningscreening
++In In silicosilico
Prediction ofPrediction ofoffoff--targettarget
profilesprofiles
Describing molecules:Describing molecules:The Protein View The Protein View
2D structure ?2D structure ?
3D structure ?3D structure ?XX
Better!Better!
Fingerprint of binding to 150 diverse targets?Fingerprint of binding to 150 diverse targets?
Experiences from Experiences from BioPrintBioPrint Broad Broad Pharmacological ProfilingPharmacological Profiling
TheseThese fragment fragment corescores areare in in potent hit potent hit compoundscompounds withwithundesirableundesirable offoff--targettargetactivitiesactivities thatthat couldcould not not beberemovedremoved
Cluster compounds in biological spaceCluster compounds in biological space XX
O
NH
Ar
Ar
O
NH
Ar
Ar
NH
O
Ar
Ar
N
O
Ar
Ar N
N HAr
Ar
N
NAr
Ar
2D2DRelatively Relatively selectiveselective
N
NH
Ar
ArN
NH
Ar
Ar
ThisThis fragment fragment corecore waswas in a in a ””cleanclean”” potent hit potent hit compoundcompoundwithoutwithout thethe undesirableundesirable offoff--targettarget activitiesactivities candidatecandidate
2D2D
CanCan bebe criticalcritical to to getget corecore (fragment) (fragment) startingstarting point rightpoint right
ExperiencesExperiences from from BioPrintBioPrintBroad Pharmacological ProfilingBroad Pharmacological Profiling
VeryVery similarsimilar structuresstructures cancan have have veryvery differentdifferent polypharmacologypolypharmacology::
ExperiencesExperiences from from BioPrintBioPrintBroad Pharmacological ProfilingBroad Pharmacological Profiling
VeryVery differentdifferent polypharmacologypolypharmacology, , bothboth for for relatedrelated targetstargets and and broadlybroadly for for candidatescandidates for a 5HT for a 5HT targettarget::
(TOXIC)(TOXIC)
(TOXIC)(TOXIC)
(TOXIC)(TOXIC)
Correlation of toxic effect with BioPrintprofile was found for a CNS target
PDE target inhibitorsPDE target inhibitors
The prediction challengeThe prediction challenge
Potent hitsPotent hitsN
N H
Ar
Ar
N
N HAr
Ar
O
N H
Ar
Ar
NH
O
Ar
Ar
O
N
Ar
R
RefRef::
ClusterCluster usingusinga a circularcircularstructurestructure--basedbasedfingerprint ? fingerprint ? e.g. e.g. ScitegicScitegic FCFP6FCFP6
(not general)(not general)
Relatively Relatively selectiveselective
ClusterCluster usingusingin in vitrovitro biologicalbiologicalactivitesactivites, , diverse diverse targetstargetse.g. e.g. BioPrintBioPrint
cancan getget individualindividual models models for ~50%for ~50%
ClusterCluster usingusingstructuralstructuralfingerprints ? fingerprints ? e.g. e.g. DaylightDaylight
2D2D
O NH
S
In In SilicoSilico Prediction of Broad Profile for Prediction of Broad Profile for PolypharmacologyPolypharmacology ??
Predictions from ChEMBL dataset (7.2K targets, 0.6 million compounds, 2.4 million activities) + IUPAR +
For project compounds not yet viable to replace For project compounds not yet viable to replace experimental screening experimental screening –– too many false positives / negativestoo many false positives / negatives
Warning: Warning: ““SimilarSimilar”” compounds often do not have compounds often do not have similar offsimilar off--target activities!target activities!
<100nM <1μM <10μM
In In SilicoSilico Prediction of Broad Profile for Prediction of Broad Profile for PolypharmacologyPolypharmacology
Experimentally validated examples of predicting new Experimentally validated examples of predicting new targets for drugs targets for drugs (richer chemical space than for many new (richer chemical space than for many new project compounds)project compounds) but but ““bigger bigger picturepicture””maymay be missed:be missed:
Paxil (Paxil (paroxetineparoxetine):):••SEA Predicted beta adrenergic blocker SEA Predicted beta adrenergic blocker ––confirmed at 10 confirmed at 10 μμM on M on ββ11••Experiment (Experiment (BioPrintBioPrint): ):
–– 4 Active at 100nM4 Active at 100nM--11μμM M –– 9 Active at 19 Active at 1--5 5 μμM M
Good prediction of BioPrint assays, + some interesting predictions of other activities but false positivesProzac (Prozac (fluoxetinefluoxetine):):
••SEA Predicted beta adrenergic blocker SEA Predicted beta adrenergic blocker ––confirmed at 4.4confirmed at 4.4μμM on M on ββ11••Experiment (Experiment (BioPrintBioPrint): ):
––5 Active at 100nM5 Active at 100nM--11μμMM––4 Active at 14 Active at 1--5 5 μμM M –– 6 Active at 56 Active at 5--1010μμMM
Difficult to predict BioPrint assays, false positives + some interesting
predictions of other activitiesGood prediction
In In SilicoSilico Prediction of Broad Profile for Prediction of Broad Profile for PolypharmacologyPolypharmacology
Use of docking into XUse of docking into X--ray structures an exciting ray structures an exciting future future ““de novode novo”” approachapproach
--BUT from results seen so far a current NOBUT from results seen so far a current NO--GO!GO!
SummarySummarySummary
Fragments & SBDD Fragments & SBDD cancan provideprovide ””betterbetter”” startingstartingpoints for points for bothboth enzymeenzyme & GPCR & GPCR targetstargets((wherewhere binding is relatively optimal & properties (LE etc) good binding is relatively optimal & properties (LE etc) good vsvssubsub--optimal but more potent hits from HTS etc)optimal but more potent hits from HTS etc)
Analyse & Analyse & useuse all all thethe data to drive data to drive betterbetter drugdrugdiscoverydiscovery decisions:decisions:
-- PhysicochemicalPhysicochemical propertiesproperties etcetc-- Protein Protein structurestructure (SBDD) & profiling ((SBDD) & profiling (BioPrintBioPrint etcetc))-- But But dondon’’tt forgetforget in in cerebrocerebro ! ! AidAid creativitycreativity, intuition, , intuition, knowledgeknowledge withwith 3D. 3D. ThinkThink aboutabout relevancerelevance & & bewarebeware ofofbias/bias/overfittingoverfitting
FullFull ””rational designrational design”” (SBDD) (SBDD) nownow possiblepossible for major for major drug drug classclass ofof GPCRsGPCRs
FragmentsFragmentsProperties: physicochemical, LE, LLE etc Properties: physicochemical, LE, LLE etc -- kineticskineticsBiophysical screening / mappingBiophysical screening / mappingGPCR knowledge antagonists GPCR knowledge antagonists –– agonists etcagonists etcIn In cerebrocerebro -- with 3Dwith 3DToxicity predictionToxicity predictionDescribe structures in a biologically relevant wayDescribe structures in a biologically relevant way
Simulation (MD etc) Simulation (MD etc) FEBFEBIntegration of data and approachesIntegration of data and approaches –– no boundsno bounds
Gregg Gregg SiegalSiegal, Francis , Francis FigaroaFigaroa, Johan , Johan HollanderHollander, , EisoEiso AB (U. AB (U. LeidenLeiden & & ZobioZobio))TINSTINS
David David MyszkaMyszka, Rebecca Rich , Rebecca Rich ((UniversityUniversity of Utah)of Utah)SPRSPR
FBDD / Enzymes FBDD / Enzymes Lena Lena TagmoseTagmose (Lundbeck)(Lundbeck)
BioPrintBioPrintBioPrintBioPrint teams teams atat CerepCerep and Pfizerand Pfizer
FBDD/SBDD FBDD/SBDD -- GPCRsGPCRsAll the team All the team atat HeptaresHeptares (Platform(Platform--DiscoveryDiscovery--Management)Management)LMB: Richard Henderson, LMB: Richard Henderson, GebhardGebhard SchertlerSchertler, Chris Tate, Tony Warne, Chris Tate, Tony Warne