8/29/2014

1

Jonathan S Mason

Senior Research Fellow CADD

Zing Fragment-Based Drug Discovery Conference

July 19-22, 2014 Punta Cana

© Heptares Therapeutics 2013

The HEPTARES name, the logo and STAR are trade marks of Heptares Therapeutics Ltd

High End GPCR FBDD: Driving Hit Discovery to Candidate Design with Protein Structure, Including

Water Network Energetics

Non-Confidential

Take Home Messages

Full FBDD/SBDD now possible for GPCRs and can be game-

changing

Waters are not optional – missing dimension in many

computational studies (protein + ligand + water)

Waters important for potency-binding (druggability)

- & may be key for selectivity & kinetics as well as potency

- & protein structure-function

- & possible now to generate & score networks rapidly

Each new target structure has “unexpected/predicted” elements - e.g. new Family B and C structures

A single ligand complex structure is generally not sufficient

- e.g. Insights from multiple orexin structures, including water-mediated

8/29/2014

2

Non-Confidential

Heptares Therapeutics

Exceptional pipeline of new medicines targeting G protein-coupled receptors

Multiple Phase 1 & Phase 2 clinical programs reporting data during 2014-16

Indications: Alzheimer’s, Schizophrenia, Diabetes, ADHD, Chronic Migraine

Leading structure- and fragment-based GPCR discovery platform

Proprietary StaR® technology enables small molecule & biologics discovery

Deals include: AstraZeneca, MedImmune, Cubist, MorphoSys, Takeda

Investors: Clarus Ventures, MVM, Novartis, Takeda, Stanley Foundation

Experienced team in UK (>70 staff) and USA / Boston

Unique SBDD Platform

StaR® technology: Small number of mutations that increase receptor thermal stability in the presence of a ligand in detergent receptor that preferentially exists in a single conformation (e.g. agonist or antagonist) with significantly increased levels of stability 1/3 known GPCR X-ray structures

Non-Confidential

Iterative process for

making StaRs

StaR approach delivers stabilised GPCRs

Pharmacology Correlates with

StaR Conformation

A2A inactive

StaR

→ folded protein → crystals

8/29/2014

3

Non-Confidential

Experimentally Enhanced Homology Models

Non-Confidential

Heptares Corporate Deals

Heptares Product Pipeline

Clinical portfolio of GPCR drugs advancing internally and through new partnerships

8/29/2014

4

Non-Confidential

Heptares Technology Platform

Adenosine A2A

Antagonist for

Neurology

GPCR Structure Determination

Landmark structure reveals new

drug binding site in Family B GPCRs

GPCR Fragment-Based Design

FBDD of mGlu modulator; sub-

nanomolar affinity with superior PK

Controlling Receptor Kinetics

Drug kinetics by StaR® using SPR

related to X-ray crystal structures

Functional or blocking mAbs

generated using StaR® antigens

Stabilised Receptor (StaR®)

Increased stability

Antigens for mAb Discovery

Validated

pharmacology

Non-Confidential

10-30 Mutations to the Binding

site region

StaR Mutants screened on Biacore chips

Site directed mutagenesis

Congreve, M. et. al, J. Med. Chem., 2012, 55, 1898

StaR Proteins Allow Biophysical MappingTM of GPCR Binding Sites

Zhukov, A. et al., J. Med. Chem. 2011, 54, 4312.

Biophysical Map of binding site

Correlating binding data from multiple ligands with multiple mutant StaR proteins

Ligand-refined homology model and detection of protein-ligand binding modes

Correlation with crystallographic binding mode

Structure Kinetic & Activity Relationships

Fragments

BPM & X-ray structural data

for caffeine

Best-scoring binding mode

BPM refined X-ray structure

XAC ligand in A2AR

N

S

O

O

O

O

OH

Virtual screening

hits

HTL79 HTL638

N253

H250

N181L85

I66

Y271

H278

N253

H250

N181L85

I66

Y271

H278

8/29/2014

5

Non-Confidential

Common features of ligand binding sites can be recognised

Family B CRF1-structure showed much deeper binding, new Family C structures bridge A – B

Bortolato et al. BJP 2014

Non-Confidential

Family A, B, C, F binding sites

A

B

C

F

8/29/2014

6

Non-Confidential

‘High-end’ structure based drug design

Computational chemistry as an effective link between

high-quality structural, biophysical and

pharmacological data

Water network energetics including network

perturbation

New perspectives on druggability, selectivity and

kinetics using water relative free energies, 3D

physicochemical properties (GRID) and MD

simulations

Non-Confidential

Agenda

Heptares approach to GPCR X-ray structures

Biophysical techniques & fragment screening approaches applied to

GPCRs

Using structures to enable the design of optimized drug candidates for

difficult GPCR targets

The importance of water molecules & their energies in GPCR drug design

– from druggability analyses through potency & selectivity to kinetics

Insights from Structural Biology - Some new X-ray structures at Heptares

8/29/2014

7

Non-Confidential

GPCR Fragment Based Drug Discovery

Stabilised GPCRs can be used for fragment biophysical

screening similarly to enzyme targets by:

SPR

NMR – TINS

HCS (StaR proteins stable in 10% DMSO unlike wild type)

CE

Success with Family A, B & C

Identify orthosteric & allosteric ligand fragment hits,

particularly with label-free methods (e.g. TINS)

Non-Confidential

SPR kinetics /

stoichiometry

Thermal shift

CE

Radioligand binding

Fragment Screening Cascade

NMR/TINS

HCS

SPR

SAR

Structural model

BPM X-ray

Primary screening validated with • SPR • NMR • HCS • CE

Hits triaged by • SPR kinetics • SPR stoichiometry • Binding assays • Thermal shift

Hits validated by • SAR / analogues • X-ray / modelling • BPM / SDM

Primary Screening

Hit Confirmation

Hit Validation

8/29/2014

8

Non-Confidential

Selective fragments can be identified for each GPCR

- different to those found for enzymes

Rational starting points for design of selective or multi-target ligands

A2A: 531 fragments assayed 94 hits

8 1

81

70

579 fragments assayed

1400 fragments

assayed

7

0

0

Overlap of TINS hits

0

Beta AR

A2A

PDE

BACE Protease

531 fragments assayed

2

52

65

0

0

20

0

Fragment Screening – NMR/TINS

Immobilized protein

– only small amounts needed (~1mg)

Very sensitive:higher mM hits identified

(not found by SPR)

TINS = Target Immobilized NMR Screening

NMR (TINS) Screening (with ZoBio) - b1AR, A2A, GLP1

Non-Confidential

SPR Fragment Screening

A2A StaR

DPCPX

riboflavin

caffeine

theophylline

1-methylxanthine

3-methylxanthinexanthine

7-methylxanthine

hypoxanthineallopurinol

FADDPCPX

riboflavin

caffeine

theophylline

1-methylxanthine

3-methylxanthinexanthine

7-methylxanthine

hypoxanthineallopurinol

FAD

Highly stable surface – DPCPX controls

Inactive fragments

Weak (mM) fragments easily detected

More potent (uM) fragments

• Weakly binding fragments hits easily discriminated from inactives • Xanthines added to library as likely binders • Chip stable for days (David Myszka)

8/29/2014

9

Non-Confidential

SPR fragment screening

Analysis of hits

Hits are further analysed by

– Detection of non-specific binding on control surface (eg denatured

protein)

– Full concentration response curves

– Assessment of kinetics – fragments should be fast off

– Stoichiometry – fragments should ideally bind 1:1 with target

Non-Confidential

Vs. plots A2A and b1AR SPR screen of Maybridge

Fragment Library (500 compounds)

Standard

8/29/2014

10

Non-Confidential 19

A2A and b1AR Full Fragment Screen of Heptares

Fragments (>600 compounds)

Selective Hit

Caffeine

b1 Control

A2A Control

b1 Hits

A2A Hits

Hits binding to both receptors

Non-Confidential

Comparison with binding assay

(all A2A positives from HTL library)

20

POC A2A known hits readily detectable by SPR

KD 0.5 mM Ki 0.03 mM

KD 5 mM Ki 0.1 mM

KD 16 mM Ki 5 mM

2.4% hit rate of A2A selective fragments

1.8% hit rate of b1AR selective fragments

A2A TINS screening identified allosteric compounds

8/29/2014

11

Non-Confidential

Capillary Electrophoresis (with Selcia)

21

• Clear demonstration that a range of known fragments including caffeine can be detected

• Non-binding fragments such as aspirin do not compete for the probe ligand (DPCPX)

• Medium throughput method, allowing screening of ~1000 compounds per mg of protein

Non-Confidential 22

ChromaDepth glasses

put on

your

now

Thank you

Please

8/29/2014

12

Non-Confidential 23

Non-Confidential 24

8/29/2014

13

Non-Confidential 25

Non-Confidential 26

8/29/2014

14

Non-Confidential 27

Non-Confidential 28

8/29/2014

15

Non-Confidential 29

Non-Confidential 30

8/29/2014

16

Non-Confidential 31

Non-Confidential

32

8/29/2014

17

Non-Confidential

33

Non-Confidential

34

8/29/2014

18

Non-Confidential

35

Non-Confidential

b1AR: Structure-guided hit to lead to

crystallography progression

KD = 16 mM

LE = 0.41

N

NH

F

FF

Ki = 500 nM

LE = 0.61

N

NH

b1AR & b2AR structures used to guide selections

1. Warne, T. et al. Nature 2011, 469, 241

2. Rasmussen, S. G. F. et al. Nature 2007, 450, 383.

Fragments bind in a consistent and efficiently way to receptor (crystal co-complex structures)

• High affinity, high solubility

fragments

• Two molecules of

complementary structure

selected for co-crystallisation

• Structures of both solved in

b1AR from single crystals

• – 3.25 Å and 2.7 Å

Ki = 68 nM

LE = 0.65

N

NH

NH

N

N

NH

Ki = 224 nM

LE = 0.53

Docking and design in a protein-ligand binding model led to suggestion to introduce donor atom

• Receptor fragment screening

by SPR of a StaR construct

8/29/2014

19

Non-Confidential

b1-AR

Chemical Structure b1 pKia LE cLogP

b LLE

c

N

N

NH

5.20 0.44 1.59 3.61

N

NH

5.87 0.67 1.11 4.76

N

NH

ClCl

7.07 0.69 3.03 4.04

N

N

NH

6.65 0.53 2.05 4.60

N

N

NH

5.80 0.47 1.44 4.36

N

N

N

NH

SMe

S

6.70 0.43 3.69 3.01

N

NH

NH

7.17 0.65 1.03 6.14

b1AR FBDD

Non-Confidential

8/29/2014

20

Non-Confidential

Non-Confidential

8/29/2014

21

Non-Confidential

Non-Confidential

N

NH

NH2

3 A3 pKi = 6.4

NN

NH

1 H4 pKi = 7.0

N N

N

N

NH2

2 H4 pKi = 8.2

N

N

NH

F3C

4 b2AR pKD = 7.8

ONH

ONH

5 H1 pKi = 8.2 6 H1 pKi = 6.1

Other GPCR Family A FBDD

8/29/2014

22

Non-Confidential

CXCR4 Chemokine Receptor Antagonist

Hit 2 Lead

39/808 hits from diversity

screening (4.8%)

Follow up gives clear SAR

Rapid identification of

potent hit series

Hit Series

IC50 = 10 nM (LE = 0.34)

Good solubility

MWT ~300, cLogP 1.3, PSA ~76

N

NH

NH

NH

N

NH

NH

NH

Example fragment hit

IC50 = 150 mM (LE = 0.47)

Good solubility

MWT 144, cLogP 1.3, PSA ~39

Non-Confidential

High Concentration Screening

Lipid Receptor Agonist StaR

Agonist StaR binds known agonists with

higher affinity compared to wild-type

receptor

10% DMSO has no effect on ligand-

binding to StaR in membranes, unlike

wild-type receptor

Enables screening of fragment library at

high concentration (100 mM) which would

be impossible with wild-type receptor

Approx 6% hit rate (84 / 1419 fragments

inhibit > 30% binding n = 2)

2 3 4 5 6 7 8 9

2

3

4

5

6

7

8

9

pKI, Wild Type

pK

I, S

taR

0 20 40 60 80 100

0

20

40

60

80

100

% inhibition of specific binding (n = 2)

% in

hib

itio

n o

f s

pe

cif

ic b

ind

ing

(n

= 1

)

StaR

-12 -11 -10 -9 -8 -7 -6 -5 -4 -3 -2

0

1000

2000

3000

4000

1%

5%

10%

[DMSO]

Log [Agonist] (M)

Bo

un

d lig

an

d (

cp

m)

Wild-type

-11 -10 -9 -8 -7 -6 -5 -4 -3 -2 -1

0

5000

10000

15000

20000

25000

Log [Agonist] (M)

Bo

un

d lig

an

d (

cp

m)

8/29/2014

23

Non-Confidential

High Concentration Screening: mGlu5 and mGlu2

First example of stabilisation of a family C receptor

mGlu5 stabilisation carried out with a negative

allosteric modulator (NAM) with binding site in the

transmembrane region of the receptor

Very dramatic increase in expression with the StaR

StaR has significantly higher DMSO tolerance

Bespoke Family C fragment set yielded tractable hits

for mGlu5 and mGlu2 (6-8% hit rate, >30% cut off)

mGlu5

Initial fragment hit

Rapid progress to advanced lead

compounds in less than 1 year

Fragment Based Drug Discovery

for mGlu5

Non-Confidential

HTL compound 1

Initial fragment hit

Jan 2012

HTL compound 2

SAR / SBDD

Mar 2012

HTL compound 36

Improved ADME profile

Jul 2012

HTL compound 43

Initial lead molecule

Aug 2012

PK

optimisation

HTL compound 98

Excellent rat PK

HTL compound 101

No P450 issues, different

PK profile

HTL compound 74

Excellent rat PK

Rapid progress to advanced lead compounds in less than 1 year

Fragment Based Drug Discovery for mGlu5

8/29/2014

24

Non-Confidential

Comparison of Heptares GPCR Fragment Hits with

Enzyme Targets

Target Method Hit LE Target Method Hit LE

A2A TINS NMR 0.56 Protein kinase B

X-ray soaking

0.47

A2A

SPR 0.53 DPPIV HCS 0.46

Family A aminergic

SPR 0.41 Thrombin X-ray

soaking 0.40

Family A peptidergic

SPR 0.31 BACE SPR ~0.3

Family A lipid HCS agonist 0.55* HSP90 NMR 0.53

Family A Chemokine

CXCR4

HCS antagonist

0.47 PDE4 HCS 0.46

GPCRs highly comparable to enzyme targets in terms of quality of hits, when StaR proteins used for screening

* StaR assay

Non-Confidential

Agenda

Heptares approach to GPCR X-ray structures

Biophysical techniques & fragment screening approaches applied to GPCRs

Using structures to enable the design of optimized drug candidates for

difficult GPCR targets

The importance of water molecules & their energies in GPCR drug design

– from druggability analyses through potency & selectivity to kinetics

Insights from Structural Biology - Some newly solved X-ray structures from

Heptares

8/29/2014

25

Non-Confidential

Techniques to Aid GCPR Structure Based Design

A1R pocket

boundary

A2AR pocket

boundary

lipophilic hotspots

in yellow (C1 probe)

3,5-disubstitution increases

affinity & selectivity

BPM reveals

novel binding mode: ligand sits

deep in ‘ribose pocket’

Biophysical Mapping

Water network energetics

GRID hotspots & surfaces

Non-Confidential

Does SBDD make a difference?

X-ray structures of a broad range of M1 agonist ligands

have been solved

Med Chem design heavily guided by protein-ligand

crystallography

Flexibility/induced fit in 1-2Å range at various sites in cavity

(main chain and side chain)

Quite similar ligands bind with the common substructure

atoms not overlapping

Enabling power of SBDD, caution for pharmacophore-

based approaches

8/29/2014

26

Non-Confidential

Agenda

Heptares approach to GPCR X-ray structures

Biophysical techniques & fragment screening approaches applied to GPCRs

Using structures to enable the design of optimized drug candidates for difficult

GPCR targets

The importance of water molecules & their energies in GPCR drug design

– from druggability analyses through potency & selectivity to kinetics

Insights from Structural Biology - Some new X-ray structures at Heptares

Non-Confidential

Why Water?

Water molecules play an essential role in the structure and function of

biological systems

-appear to play a key role in the GPCR structures available to date, perhaps due

to the deep pockets present in the GPCR transmembrane binding sites?

Displacement of waters from a binding site is a key component of ligand

binding, with significant binding energy, and thus potency, often from the

entropic gain of the displacement

But all waters are not equal…

- Burying an ”unhappy” water [i.e. entropically and/or enthalpically worse

than bulk sovent] may affect both potency and kinetics

- Pertubation of the remaining waters will also affect binding

Opportunity to provide new insights into druggability, to drive SAR,

selectivity and find solutions to many SAR issues (predicition of ”magic methyl”,

SAR not explainable by direct ligand-protein interactions...) + kinetics

8/29/2014

27

Non-Confidential

b-ionone binding pocket

L3.40, P5.50, W6.48

Hydrophobic hindering

mechanism L3.43, F6.44,

M6.40

Water pocket N1.50

L2.46, D2.50 S7.45, N7.49, P7.50

Ionic lock R3.50, E6.30

Rhodopsin Activation

W6.48

D2.50

F6.44

Y7.53

S7.45

N7.49 M6.40

L3.43

b-ionone movement L3.40, P5.50, W6.48

Occupy vacated hydrophobic pocket

Opening of the Hydrophobic

hindering mechanism

Unstable water channel N1.50

L2.46, D2.50 S7.45, N7.49, P7.50

Ionic lock broken

Interaction Y7.53 Y5.58

isomerisation to the all-trans retinal form

W6.48

D2.50

F6.44

Y5.58

Y7.53

S7.45

N7.49

M6.40

L3.43

Water Channel

Tehan et al. Pharmacol Ther. 2014 S0163-7258(14)00032-1

Waters in GPCR Protein Flexibility and Activation

Bortolato et al. BJP 2014

Y5.58

GPCR Hydrophobic Core and Activation

Hydrophobic Core common feature across subfamilies

Non-Confidential

Receptor Activation and Agonist Design

Cluster of GPCR Core Residues Control Activation

Conserved “Hydrophobic Hindering Mechanism”

I3.40

I6.40

F6.44

L3.43

TM3 TM6

Beta-2, rhodopsin, A2a et al

L3.43 gripped by F6.44 & I6.40 - Holds TM3 & TM6 in

inactive state

Mutations => constitutive activity

Inverse agonists impede release of 3.43 so freezing

receptor in inactive state

Agonists promote binding-site helical movements and

steric clashes that trigger breakage of cluster

Extracellular

Cytoplasmic

F6.44

I3.43

M6.40

H2O

Extracellular

Cytoplasmic

b2 AR

8/29/2014

28

Non-Confidential

The New Wave: WaterMap (Schrödinger) + GRID/WaterFLAP (Mol. Discovery)

GRID Map contoured at:

C1= (lipophilic) probe:

Yellow -2.5 kcal/mol

Water (H-bonding) probe:

Green -6.0 kcal/mol

Surface defined by CH3 probe:

Grey 1.0 kcal/mol

S1

S4

Drug-like properties from GRID

Factor Xa

The myth: Basic S1 for

serine proteases (factor Xa)

The structure

that broke the myth

The clinical

candidate

Potency, selectivity from waters

CRY probe = C1= + DRY

Waters in GPCR Protein Ligand Binding and Kinetics

Non-Confidential

Factor Xa

Update: New WaterFLAP/GRID based protocol

o GRID/WaterFLAP (Molecular Discovery) based protocol - adds waters iteratively taking explicity into account other waters already added + ligand if present - with short MD optimization; o Scored using GRID-based OH2 + CRY + ENTropy

8/29/2014

29

Non-Confidential

Water network computation reproduces

high-resolution X-ray water network

Technical details:

o Build a map of water

occupancy seeded with

GRID/WaterFLAP

1. - MD simulation with

GROMACS 200ps

- Energies (chemical potentials)

assigned using CRY/H2O/ENT

GRID potentials

- Approx 2 hours on a cpu.

.

or 2. WaterMAP (MD) to optimize

& evaluate energies

A2A at 1.8Å Liu, W. et. al, Science, 2012, 337:232

Excellent correlation + we also know the relative energies of the waters, and

can re-evaluate with different ligands - a significant advance

Non-Confidential

Binding Site Analysis for Ligandability / Druggability

Druggability – ligandability with properties of an oral drug

Use a combined analysis of ‘unhappy’ waters & binding site

preference (lipophilic/hydrophobic)

Analyse numbers, connectivity & distribution of ‘unhappy’ waters

Analyse environment (lipophilic etc) & entropy ‘Unhappy’ water = >2 kcal vs bulk solvent (vacuum usually worse) MD/WaterMap

or combined OH2/CRY/ENTropy GRID/WaterFLAP

Druggabllity linked to both number and proximity of “unhappy” waters and the

mixed concurrent hydrophobic-hydrophilic properties preferred for a ligand

X-ray structures show that potent efficient ligands displace the “unhappy”

waters – observed for all GPCR structures

8/29/2014

30

Non-Confidential

Druggable regions with clusters of ‘unhappy’ waters (WaterMap) with GRID showing adjacent/concurrent lipophilic & H-bond hotspots

Dopamine D3 Histamine H1 Muscarinic M2

OH

O

O

N+

H

Druggability : Dopamine, Histamine & Muscarinic GPCRs

= hydrophilic hotspot

= hydrophobic hotspot Predicted water energy

Low DG High DG

Bulk-like

Mason et al. TIPS 2012, 249-60

Non-Confidential

Druggability : Chemokine GPCRs

κ opioid receptor m opioid receptor δ opioid receptor

8/29/2014

31

Non-Confidential

Waters & GRID hotspots important for Adenosine A2A

antagonist triziazine & chromone series

Used GRID hotspots to enable first full GPCR

SBDD to give a clinical candidate for A2A

The highly ligand efficient designed structures

displace a cluster of unhappy waters deep in the

pocket (missed by previous ZM-like ligands from HTS etc)

Mason, J.S. et al., In Silico Pharm, 2014, 1:23

Non-Confidential

Alternative series from virtual screening utilizes other lipophilic hotspot region - Lipophilic hotspots key for GPCR design

NN

N

N

NH2

ON

S

O

O

O

Andrews, S.P. et. al, Med. Chem. Commun., 2014, epub

o High LE 0.5

Chromone VS series A1 selective from

lipophilic (vs H-bonding) groups

8/29/2014

32

Non-Confidential

Water can explain chromone ‘magic methyl’

33x more potent

Mason et al. In Silico Pharm. 2013 Andrews et al. Med Chem Comm 2014

MD shows ligand moves to avoid a

”dewetted” / high energy water situation

Non-Confidential

Agenda

Update: Heptares approach to GPCR X-ray structures

Update:Using structures to enable the design of optimized drug candidates

for difficult GPCR targets

The importance of water molecules & their energies in GPCR drug design

– from druggability analyses through potency & selectivity to kinetics

Insights from Structural Biology - Some new X-ray structures at

Heptares

8/29/2014

33

Non-Confidential 65

Courtesy of Arthur Doweyko

Structural information can be revolutionary – previously we were

biased by ligand + analogs data

Beware of biases in how we see and “force-fit” data

Beware of force-fitting ligands

(pharmacophores) without taking

water into account

Non-Confidential

o Suvorexant (Merck) DORA - EMPA - SB-334,867 (+ Heptares ligands)

o Sit high in orexin 1 / 2 binding site

o Suvorexant in hydrophobic collapse conformation as predicted

o Many polar interactions water mediated

o Ligand pharmacophore-based overlays do not correspond to experimental

structure-based overlay

Family A GPCR Structures

– Mutliple ligand structures key

NNO

N

O

N

N

N

Cl

O

N

NH NH

NN

O

S

O

ON

N

O

N

O

N

8/29/2014

34

Non-Confidential

Family B GPCR Structures

50-receptor Family B includes many important

drug targets

– GLP-1, PTH, Glucagon, CGRP

CRF-1 (anxiety/depression) 7TM domain solved

at high resolution with bound antagonist – Major differences to Family A

– Useful modelling of Family B enabled for the first time

Fundamental advance opening up new avenues

for drug design

– Driver for in-house GLP-1 agonist programme

Also Glucagon structure from Srevens/Scripps,

but no bound igand visible

Non-Confidential

CRF1 Antagonist StaR Structure – First Family B GPCR

Trans-membrane

helical bundle forms a chalice-like structure

Family B GPCR Structures

Antagonist binding site is located in

the intracellular half of the receptor

CP-376395

FamA

ligands

13-23Å

NH

N O

Binds in a predominantly

hydrophobic pocket:

8/29/2014

35

Non-Confidential

Radar plot of Heptares mGlu5

PCC properties. A perfect

compound hits the bull’s eye.

New mechanism for broad range of neurological and psychiatric diseases

Clinically validated in Autism, Dyskinesia, Depression/Anxiety, Migraine

Heptares leadership in SBDD for mGlu Family C (8 important GPCRs)

High resolution X-ray structures of mGlu5 in complex with multiple ligands

Creation of novel allosteric modulators using antagonist StaR® proteins

Highly differentiated next-generation drugs with best-in-class properties

Advantages: superior control over PK:PD; enhanced oral BA; cleaner chemistry

lacking toxicity alerts / sites of metabolism; excellent affinity & potency

mGlu5 Antagonists & Family C GPCRs: Neurological Diseases

Heptares have X-ray structures of

mGlu5 – defines drug binding

Heptares agent: QD/low dose with

tight PK:PD relationship to avoid

Cmax-driven AEs seen in competitors

Family C GPCR Structures

Non-Confidential

mGlu5 bound to mavoglurant at 2.6Å resolution

TM1

TM2

TM3

TM4

TM5

TM6

TM7

TM1

TM2 TM3

TM4

TM5

TM6

TM7

mavoglurant

ECL1 ECL2

ICL1

ICL3 C-ter

N-ter

N-ter

ECL1

ECL2

ECL3

ECL3

OMIT (before inclusion) FO-FC = 2.0s

• Outward kink in TM7 propagated by highly conserved (>90%) Pro/Lys motif • Ligand visible / unambiguous

• Conserved S-S – ECL2 – top of TM3 • ICL1 = short a-helical structure • Continuous density for ICL3 / ECL1 / ECL3

Family C GPCR Structures

8/29/2014

36

Non-Confidential

• Conserved S-S – ECL2 – TM3 anchors ECL2 shallow across top of receptor

• Additional network of interactions ECL2-TM1 / TM2 ECL1 / TM3 Gln6473.32 (92% conserved)

Closed entrance consistent with native ligands binding to orthosteric site in ECD

Restricted entrance to allosteric pocket

Non-Confidential

mavoglurant

8/29/2014

37

Non-Confidential

Overall very similar – 0.98Å rmsd across Cα in TMD ECL2 of mGlu1 does not cross top of receptor as in mGlu5 FITM binds higher (overlays with carazolol from beta), much more analogous to class A orthosteric positions

Comparison of mGlu1 (purple) vs mGlu5 (green)

Confidential

Summary – GPCR FBDD/SBDD

Multiple new X-ray structures have fully enabled structure based drug

discovery for GPCRs including FBDD

Virtual screening and fragment screening provide improved approaches to

lead generation (SPR, TINS, HCS on StaR proteins)

Compounds from FBDD/SBDD have optimised profiles compared to those

derived from high throughput screening

FBDD/SBDD can address many previous issues in GPCR drug discovery – Selectivity

– Lack of chemical starting points

– Limited chemistry which includes toxicophores

– Receptor kinetics

Demonstrated that SBDD can be used across multiple GPCR families to

generate optimised and novel drug candidates

8/29/2014

38

Confidential

Conclusions – Water & Lipophilic Hotspots

o Solvent is as important as the ligand and the protein!

o Water network modeling/energetics and network perturbation is

essential to understand ligand binding (& kinetics)

• Believe it is important to take in account in an explicit way the perturbation

of the remaining water network resulting from ligand binding as well as

displaced waters

• Residence time - unhappy waters (trapped, being formed...)

o New perspective on druggability possible by analysis of binding site

using water free energies & 3D physicochemical properties (GRID) in

particular lipophilic hotspots

Confidential

Class A

Class C

Class B

Acknowledgements

Computational Chemistry

Andrea Bortolato

Dahlia Weis

Ben Tehan

Francesca Deflorian

Crystallography & Biophysics

Rob Cooke

João Dias (M1)

Andy Doré (mGlu5, A2A)

Andrei Zhukov (Biacore)

Chemistry

Miles Congreve

Giles Brown (M1)

John Christopher (mGlu5)

Malcolm Weir (CEO) Fiona Marshall (CSO)

Protein Engineering

Ali Jazayeri

Nathan Robertson (A2A, M1)

Jay Patel (mGlu5)

Giselle Wiggin (mGlu5)

Protein Expression

James Errey (mGlu5)

Maria Serrano-Vega (mGlu5)

Markus Koglin (M1)

Kris Okrasa (M1, mGlu5)

Pharmacology

Alastair Brown

Ed Hurrell (M1)

Kirstie Bennett (mGlu5)

Molecular Discovery

(GRID/ WaterFLAP)

Massimo Baroni

Gabriel Cruciani

Simon Cross

Schrödinger

(WaterMap)

Woody Sherman

Thijs Beuming

Robert Abel