“Bromodomains: A new class of epigenetic targets ripe for small molecule drug discovery”

Jason WitheringtonEpiNova DPU

ELRIG – Manchester 2012

Outline

Brief introduction to Epigenetics

Luck strikes!.....discovery of small molecule bromodomain inhibitors

Exploiting serendipity through SBDD/FBDD

Brief overview of preclinical iBET biology

Epigenetics - Chromatin

DNA is packaged around histones and other proteins to form chromatin

Chromatin is highly dynamic material which undergoes remodelling to allow suppression or activation of genes

A number of Epigenetic mechanisms control chromatin remodelling including post-translational modifications (PTMs) on histone tails

Dysregulation of histone PTMs implicated in human disease

Epigenetics : Histone post-translational modifications

· PTMs can have a direct impact on physical properties of individual nucleosomes e.g. neutralisation of charge

· PTM are recognised by specialised reader domains.

PTM of histone tails

• >70 sites are known mostly located in the unstructured N-terminal tails.

• > 8 types of modification have been reported.

• AA modified include : K, R, S, T, Y, H, E

• Most of these are reversible and dynamic.

• PTM rarely occur in isolation => complex pattern of modification = histone code.

• Reader domains rarely occur in isolation

Apo-A1 phenotypic assay• Apo-A1 target for dyslipidemia

• Upregulator reporter HTS identified several lead series including a BZD series.

• Medicinal chemistry successfully optimised molecules to candidate selection without knowledge of molecular target.

• Extensive profiling of compounds did not identify target for these molecules Chemoproteomics

NN

R2

N

R1

N

Benzodiazepine

5’-UTR ApoA1

3’-UTR ApoA1

-1.4kb

Human ApoA1 promoter

Firefly luciferase

How were new medicines discovered ?

Nat. Rev Drug Discovery 10, 507 (2011)

Between 1998-2008:

More first-in-class drugs were discovered by phenotypic screening

More follower drugs were discovered by target-based screening

1999-2008

Chemoproteomics – Overview of approach

Wash and Elution• Stringency• Compound / SDS

Matrix

A I

Separate on 1-D Gel • active compound

specific bands & low backgrounds

PMM

LC/MS/MS

PROTEINIDENTITY

Biologically relevantsystem

HepG2 & THP1

Derivatised CompoundsActive BZDInactive BZD

NN

R2

N

R1

N

NN

R2

N

R1

N

J Med Chem (2011) 54, 3827

200

40

50

150kDa

3020

10075

BZ

D a

ctiv

e m

atrix

BZ

D I

nact

ive

m

atrix

Mat

rix a

lone

Competition experiments suggest that actives from BZD and other series specifically interact with BET proteins

Compound Key

RED = Active

BLACK = Inactive

Chemoproteomics BET (BromoDomain & Extra Terminal) proteins identified

+S

erie

s X

inac

tive

200

40

50

150

kDa

30

20

100

75

+ B

ZD

act

ive

+ B

ZD

inac

tive

+ S

erie

s X

ac

tive

All bands identified as BET family proteins Brd2, Brd3, Brd4

Brd4 knockdown induces Apo-A1 upregulation

Apo-A1 activators are ligands for the BET proteins

Is this interaction responsible for Apo-A1 upregulation?

– Increase in ApoA1 mRNA on addition of BZD

– Increase in ApoA1 mRNA on BRD4 knockdown

Active BZD – 1mM BRD4 siRNA: 500nM

DC

T

24hr

48hr96hr

0hr

24hr 48hr

72hr

-1.6

-1.4

-1.2

-1

-0.8

-0.6

-0.4

-0.2

0

0.2

1

Incr

ease

in A

po-A

1 m

RN

A

BET binding correlates with Apo-A1 cellular activity

BZDs

BRD4 FP p IC50

Apo

-A1

pec1

70

NO

NN

N

N

O

Cl

I-BET762

Theoretical difficulties in tackling epigenetic PPI were not realised Many diverse and potent compound obtained using cellular activity to guide SAR.

Bromodomains bind to acetylated lysine residues

Challenges with targeting epigenetic readers

Reader domains often bind PTM weakly=> no hot spots?

Multi-valency of protein-protein interactions => Tethered ligands

MegaDalton protein-protein/DNA complexes => will inhibiting a single interaction be enough for biological efficacy?

If protein-protein inhibition is poorly tractable=> how tractable are targeting epigenetic readers?

PREVIOUS PHARMACEUTICAL FOCUS ON EPIENZYMES NOT EPIREADERS

BRDs control gene transcription

Transcriptional co-regulators involved in histone binding complexes Brd4 binds to cdk9/cyclinT (pTef-B) to positively regulate RNA pol II mediated transcription at

multiple promoters

ET domainbromodomain2bromodomain1

Ac Ac Ac

Acetylated lysines on Histones within euchromatin

Pol II

transcriptionBRD BRD

pTef-B

P

ApoA1 compounds bind to BET BUT where specifically do the compounds interact?

Chemoproteomics implicate Bromodomain of Brd2, 3, 4

Bromodomain 1 Bromodomain 2 ET domain

X

220

12010080

60

4050

30

20

kDa

1. GFP control

2. Flag Brd2 FL

3. Flag Brd2 N

4. Flag Brd2 C

Hek293 cells

1 2 3 4PD: BZD active

Western Blot: anti FLAG

1 801

473

FL

N

C

X

Biophysical data demonstrates specific binding

Some BIAcore Results - GW842819X BZD active

-20

0

20

40

60

80

100

120

1e-9 1e-8 1e-7 1e-6 1e-5 1e-4

Re

sp

on

se

RU

Conc M

Steady State: Fc=1 Spot=2-r corr

-10

0

10

20

30

40

50

60

1e-9 1e-8 1e-7 1e-6 1e-5

Re

sp

on

se

RU

Conc M

Steady State: Fc=1 Spot=1-r corr

-20

-10

0

10

20

30

40

50

60

-50 0 50 100 150 200 250 300

Fc=1 Spot=1-r corr

-40

-20

0

20

40

60

80

100

120

-50 0 50 100 150 200 250 300

Fc=1 Spot=2-r corr

Brd2 67-200 Brd2 338-473

KD = 2.25e-7MRmax = 41 => 40% surfaceis active.

KD < 7.0e-8MRmax = 74 => 70% surfaceis active.koff ~ 0.01s-1 ± 0.03 s-1

kon

N

NN

N

N O

O

U

GW841819X

KD 225nM@25°CBRD2_1

KD <70nM@25 °CBRD2_2

BZD tool binds both N and C-terminal domains but kinetics and affinity at 25°C are different for each

40 50 60

norm

alis

ed C

D

• 0

0.2

0.4

0.6

0.8

1Brd2(1-473)

Temp (oC)

INACTIVE X

INACTIVE Y

ACTIVE X

ACTIVE Y

Tool compounds stabilise all Brd2 bromodomain constructs

ACTIVE X INACTIVE X

ACTIVE Y INACTIVE Y

Isothermal Titration Calorimetry demonstrates specific binding to both BRDs

C-terminalbromodomain

N-terminalbromodomain

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

Brd2 67-2001

Brd2 67-200 Brd2 338-4733

Brd2 338-4732

C-terminalbromodomain

N-terminalbromodomain

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

1:1 46nM(16°C)

1:1 52nM(26°C)

2:1 30nM(26°C)

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

N

N

N

N

N O

O

U

GW841819X

ACTIVE BZD

I-BET762 is a highly selective inhibitor of BET bromodomains

Tm profiling

5-7oC1-3oC<1oC

I-BET762

iBET Broader Selectivity Profiling

Inactive against a wide range of proteins

Where do the compounds bind?

• N-terminal bromodomain of Brd2 is typical helical structure

• Their role is to recognise acetylated marks on histones and other proteins

• Compounds shown to displace the tetraAcH4 peptide

• Antagonise protein-protein interaction

[compound] (uM)

0.0001 0.001 0.01 0.1 1 10

rati

o

0.04

0.08

0.12

0.16 GW841819X

FRET assay for displacement of tetraacetylated H4

N

N

N

N

N O

O

N

NN

N

NO

O

U

GW841819X

First Small Molecule X-ray co-crystal confirms binding in the acetylated lysine pocket

H4 peptide

Recognition of carbonyl of

AcK preserved (N156,Y113)

F-(VP)-Y-(CAS)-N AcK binding

site

Common to 44 out of 58

bromodomains H2O structure in pocket

preserved. NH interactions of AcK not

preserved

Interactions of BZD outside the AcK pocket

N

N

N

N

N O

O

ACTIVE BZD

BrdT – Nature (2009)

Bromodomains can deliver both probes and drug like molecules

iBET 762

clogP, PSA, MWt ~2, ~80, ~400

BRD2/3/4 pIC50 6.8/6.7/6.7

hERG EC50

Ion Works (Dof) 100uM

Patch Express 61uM

Rat (Mouse) PK* Clb (mL/min/Kg);

Vss (L/kg); t½ (h), %Fpo

63 (24), 1.8 (1.7), 0.5 (0.8), 27 (22)

Dog PK* 5, 1.8, 5.9, 44

Unbound fraction in blood (R/D/Mou/H) 0.18 /0.24/0.21/0.19

CYP inhibition IC50s (uM) > 33

P450 TDI <2-fold

*3mg/kg p.o.; 1mg/kg i.v.

Optimisation of dimethyl isoxazole HTS lead to in vivo probe I-BET 151

N

O

O

N

NH

N

O

N

NO

N

NH

NH2

O

CLi microsomes (mL/min.Kg)

CLb ml/min.kg Vd L/kg T½ h F %

Rat <0.53 18 2.1 1.7 66

Dog 17 38 3.0 1.2 16

minipig <0.53 15 1.6 1.2 65

Human 1.1

BMCL, 2012, 2963BMCL, 2012, 2968

HTS Lead I-BET 151

GSK525762 and GSK1210151 bind BET proteins using similar “hot spots”

WPF ZA Channel

AcK pocket

I-BET 762

I-BET 151

N

N

NN

HN

O

O

Cl

N

O

ON

HN

N

ON

Bromodomain Family and Structural CoverageBRPF1

BRD1

BPRF3BRD9

BRD7KIAA1240

ATAD2

WDR9_2PHIP_2

BRWD3_2

BAZ2B

BAZ2AZMYND11

CREBBPEP300

BAZ1A

BRD8

BRDT_1

BRD4_1

BRD2_1 BRD3_1BRD3_2

BRDT_2

BRD2_2

BRD4_2

TAF1_1

TAF1L_1TAF1_2

TAF1L_2

PRKCBP1

CECR2FALZ

GCN5L2PCAF

WDR9_1

BRWD3_1

PHIP_1

ASH1L

PB1_1PB1_3

PB1_2PB1_4

PB1_5

SMARCA2SMARCA4

TRIM33

TIF1TRIM66

MLL

SP110

SP100LOC93349

SP140

TRIM28

BAZ1B

Structure knownAtypical AcK Binding Residue

T

T

T

T

Y

Y

Y

Y

>50 bromodomainsIn isolation or combination with other domains

Multiple opportunities for clinical utility

Across the family there is significant structural divergence outside of the AcK binding region

BC Loop ZA Loop

Exploiting Structural Knowledge :Fragments – Generation of a Hit-ID platform for Bromodomains

Knowledge of key ligand-protein interactions derived from the Bet programme lead-like compounds

N

N

N

N

NH

O

O

Generation of a pharmacophore model Selection of a focussed screening set

Creation of a fragment toolchest that binds in the AcK recognition pocket of the bromodomain

N

O

N

O

Confirmation of the binding mode using crystallography

>20% inhib at 200uM

NMe O

N

OMe

O N

N

OMe

Fragment based discovery

"Fragment-based discovery of bromodomain inhibitors part 1: Inhibitor Binding Modes and Implications for lead discovery

Author(s): Chung, Dean, Woolven and Bamborough

1400 Fragments screened >40 Fragments crystallised

Key Structural waters identified Pharmacophore refined

pIC50 BRD 2 pIC50 BRD 3 pIC50 BRD 4

< 4.0 < 4.0 < 4.0 (LE< 0.43)

pIC50 PBMC TNF < 4.7

O

N

O

N

S NH

O

O

pIC50 BRD 2 pIC50 BRD 3 pIC50 BRD 4

5.2 5.9 5.6 (LE 0.38)

pIC50 PBMC TNF 6.5

PharmacophoreWPF shelf

Application of FBBD for Bromodomains

"Fragment-based discovery of bromodomain inhibitors part 2: optimization of phenylisoxazole sulfonamides“

Author(s): Bamborough, Paul; Diallo, Hawa; Goodacre, Jonathan; Gordon, Laurie; Lewis, Antonia; Seal, Jon; Wilson, David; Woodrow, Michael; Chung, Chun-waACCEPTED

Application of Encoded Library Technology (ELT)

NO

LIBRARIES

ENRICHMENT

LIBRARIES

1 2 3 TARGETS

Structural knowledge Construction and screening of libraries

Identification of Features

ELT hits against target 2

Exploitation of Screening output

1. Hits

2. Screening tools

3. Probes

Preclinical Biology

Nodal AND gene specific intervention?

X X

pI:C LPS

TNFaIL6 IFNb

I-BET

unaffected blocked blocked

TNFa IL-6 IFNb

BET compound displaces BRD4 from IFNb and IL-6 promoters (ChIP)

LPS drives recruitment of Brd4 to selective promoters

Compounds prevent this recruitment and block transcriptional activation

BR

D4

/ H3

IL-6IFNb

Control LPS0.00

0.20

0.40

0.60

0.80

1.00

1.20

1.40DMSO

BET

Control LPS0.00

0.05

0.10

0.15

0.20

0.25DMSO

BET

Soren Beinke

Targeted intervention by I-BET

Secondary response genesPrimary response genes

CpG high

H3K4m3

H3K9Ac

Pol II

H3K9m1/2/3

Use of Chemoproteomics for target class expansion

Pharma industry mainly reliant on recombinant platforms

Large screening panels required for selectivity profiling

(human/rat etc)

Brds occur in isolation & combination with other domains

Protein complexes modify function

Different complexes may form under different activation

states &/or different tissues EpiNova-Cellzome alliance provides a complementary screening platform to address the above

proteins binding

directly or indirectly to

I-BET

proteins binding directly or

indirectly to histone marks

BET protein imuno-

complexes

= BET inhibitor

(I-BET)

AcAcetylated

H4 tail (K4,K8,K12)

H4

Antibody against BRD2/3/4

Triple purification strategy

BET interacting proteins: MS-proteomic analysis

Nature (2011) 478, 529

BET interacting proteins: MS-proteomic analysis

log10[GSK525762A] (M)

3H-t

hym

idin

e(%

cp

m o

f s

olv

en

t c

on

tro

l)

Cell V

iability

(% liv

e c

ells

co

mp

are

dto

so

lve

nt c

on

trol)

-9 -8 -7 -6 -5 -40

50

100

150

0

50

100

150

ProliferationViability

I-BET762 is effective in multiple models of Multiple Myeloma

**

[I-BET762] Survival of OPM-2 Xenografts

0 20 40 60 80 1000

20

40

60

80

100

G1 Vehicle qdx58

G2 3mg/kg qdx68

G3 10mg/kg qdx68

G4 30mpk(15-31)/20mpk(43-68)

G5 30mg/kg qodx35

days after inoculation

Pe

rce

nt

surv

iva

l

log10 [I-BET762]

BET interacting proteins: MS-proteomic analysis

I-BET151 is a novel & selective inhibitor of BET proteins with improved PK properties

I-BET151

I-BET151 has selectivity for MLL

leukaemias

I-BET151 mediates disease control in MLL leukaemia

models

NOD-SCID

Transplant human MV411 leukaemia cells

Transplantsyngeneic MLL-AF9 leukaemia cells

C57BL/6

Summary

Chemoproteomics has been employed to identify a chemical opportunities against a previously intractable target class

Chemoproteomics has been utilised to allow the efficient selectivity profiling across the “Bromonome” using endogenous cell lysates

Chemoproteomics has demonstrated utility in defining clinical opportunities through complex identification

Effect of BET inhibition on LPS induced shock

0h

LPS

therapeutic

I-BET

1.5hpreventative

I-BET

-1h

Nature, 468, p1119, 2010

Summary…..

Use of chemoproteomics can be a powerful way to identify output of phenotypic screening

Previously “undruggable” reader class of epigenetic proteins are ripe for drug discovery

The iBET bromodomain family of proteins have profound preclinical biology (more this afternoon)

Kevin Lee