36
Hit Identification and Lead Confirmation in the Discovery of Drugs Targeting Bromodomain Proteins Liu Liu, Shaomeng Wang Department of Internal Medicine and Comprehensive Cancer Center University of Michigan, Ann Arbor, MI 48109

Hit Identification and Lead Confirmation in Discovery of ......Hit Identification and Lead Confirmation in the Discovery of Drugs Targeting Bromodomain Proteins Liu Liu, Shaomeng Wang

  • Upload
    others

  • View
    5

  • Download
    0

Embed Size (px)

Citation preview

Hit Identification and Lead Confirmation in the Discovery of Drugs Targeting

Bromodomain Proteins

Liu Liu, Shaomeng Wang

Department of Internal Medicine and Comprehensive Cancer CenterUniversity of Michigan, Ann Arbor, MI 48109

Bromodomain (BRD) proteins

Family of Chromatin modification “Reader” proteins Recognizing acetylated histone lysine and recruiting

“Effectors” (Chromatin modifiers and remodeling enzymes)

Critical to maintain proper epigenetic regulationDysfunction linked to a large number of human diseasesTotal 46 human proteins containing 61 BRDs identified

8

9

36

21

4

1

3

1

4

12 2

TRIM24

TRIM28

p300

MLL1BRG1BRM

BRDT

BRD2

BRD3

BRD4

TRIM33

CBP

PB1ATAD2

Number of diseases indications

BRDs involved in various diseases

BET

BET

BET: bromodomain and extra-terminal

BET Bromodomain (BRD) proteins

BRD4 is an activator of c-Myc, a master regulator of diverse cellular functions and a long-time compelling therapeutic target

BRD3 and 4 fuse with NUT in NUT midline carcinomas BET Bromodomain proteins cooperate with MLL (MLL1)

fusion proteins in gene activation in acute leukemia (BRD4) Beyond oncogenesis (BRD3-4), these proteins have potential

roles in other therapeutic areas, including viral infections (BRD2/4), cholesterol regulation (BRD2), inflammation(BRD2/4) and spermatogenesis (BRDT), among others.

Small molecule inhibitors of BET BRDs also selectively suppresses other genes important in cancers, such as Bcl-2and some NF-κB-dependent genes.

Therapeutic potentials of BET-BRD inhibitors in cancers

BRD inhibitors have been shown to be effective in the following tumor types

– NUT midline carcinomas with BRD-NUT fusion protein– Acute leukemia with MLL1 fusion protein– Multiple myeloma (Delmore, et al, Cell, 2001, 146 (6)

904-917)– Lung cancer (Lockwood, et al, PNAS, 2012, 109 (47),

19408–19413)– Breast, prostate and head-neck cancers (our own data)

Phase I/II(NMC)

BET-BRD inhibitors available

Phase IIb(atherosclerosis)

BET-BRD inhibitors available

MWAffinities to BET-BRDs

BRD2BD1

BRD2BD2

BRD3BD1

BRD3BD2

BRD4BD1

BRD4BD2

(+)JQ-1 457 128(1) 59(1) 82(1) 49(1) 90(1)

IBET762 424 81(1, 2) 70(1, 2) 53(1, 2)

IBET151 415 20(2, 3) 100(2, 3)

Bromosporine 404 410(4) 290(4)

PFI-1 347 123(5) 98(4) 136(1) 303(1)

PFI-1(1) 108 144 80.0 76.3 47.4 195Abbot CMPD 4 409 1.4(6) 2.0(6)

(1) ITC(2) BRD proteins with both BD1 and BD2(3) SPR(4) ALPHA screen IC50s(5) Octet RED96(6) TR-FRET (Ki)

Develop new classes of BET-BRD inhibitors with improved affinities, cellular activities, in vivo efficacy and drug properties

Identify probe molecules for other BRDs with promising therapeutic potentials, which can be used as scaffolds for developing tool molecules and clinical candidates

Rationally design and synthesize a small probe compound library (MW between 133-300) based on existing inhibitors and crystal structures (29 high resolution structures available)

Screen this library to identify new scaffolds for inhibitors targeting both BET and other BRD subfamilies

Obtain high-resolution co-crystal structures of our lead compounds with BRD proteins followed by optimization based on SAR

Determine affinities to all BETs and selectivities to other BRDs Evaluate cellular activities, Mechanism of Action, and in vivo

efficacy

New BRD inhibitors discovery and optimization

Screening small probe molecules against BRD proteins

Natural ligands (e.g. the acetylated histone peptides) of most BRDs are largely unknown or too weak to be utilized

http://www.discoverx.com/services/drug-discovery-development-services/bromoscan-epigenetic-profiling

Bromoscan from DiscoveRx

Label free technologies: ThermoFluor, NMR: semi-quantitative, requiring relatively high

concentrations for both compounds and proteins ITC: too low throughput, large amount of proteins Optical: SPR, BLI (Octet RED96)

Label-free Octet RED96:BioLayer Interferometry Technology

Higher throughput, faster workflow Reads 8 channels simultaneously so wide concentration ranges can be

run in parallel Multiple assay conditions can be used simultaneously Analyzes more samples in a shorter timeframe than Biacore Assays can be developed in hours and days

The Octet is fluidics-free so there is no leaking, or clogging to deal with The Octet can analyze samples that are not feasible on a SPR: phage

samples, cell lysates, culture media, 100% serum, etc. Octet assays are not affected by changes in the refractive index of the buffer

used leading to significantly higher tolerance to DMSO The Octet replaces ELISA assays with a single-step, no wash assay The Octet uses a non-destructive, dip-and-read measurement method that

allows for full protein/sample recovery after analysis Operational advantages (lower maintenance and cost)

The advantages of Octet RED96 system over SPR

systems

Influence of BSA and Tween-20

BRD4 BD2/PFI-1 Super Streptavidin (SSA) sensors Inactive blocked Streptavidin and buffer as double references

0.1% BSA 0.01% Tween-20

DMSO influence the interaction between BET-BRDs and inhibitors

DMSO controlled lower than 0.5%

O N

N

Cl

NN

NHO

NH

OO

SO3H

NH2

NH2

SO3HO

O

FAM-762

KD values of reference compounds to BRD4 BD1-2 determined by BLI

KD = 418nM

KD = 42nM

CMPDs IBET-762 PFI-1

MW 424 374

BRD4BD1

Octet RED96 89 131

Published 53* 136

BRD4BD2

Octet RED96 42 418

Published 53* 303* BRD4 BD1-2, both reported data and our own data suggested no significant difference between BD1-2 and BD1 and BD2.

Super Streptavidin (SSA) sensors Inactive blocked Streptavidin (SAB4) and buffer as double references 0.1% BSA, 0.01% Tween-20, 0.5% DMSO

Screening assay design

Screening to BRD4BD2

250µM, 0.5% DMSO, finished in a few hours Positive hits: > average of buffer + 3 SD Hit 1, 2, 3 are derivatives from the same scaffold.

Hit 1 Hit 2

Hit 3

Hit confirmation by dose dependence titration

CMPD MW KD (µM) kon (1/Ms) kdis (1/s)

Hit 1 133 89.2 2.21E+02 1.97E-02

Hit 2 210 42.5 5.57E+02 2.37E-02

Hit 3 211 14.9 1.52E+03 2.27E-02

Binding affinities of Hit 3 to other BET-BRDs

BETs KD (µM)

BRD3BD1 58.7

BRD4BD1 14.9

BRD4BD2 18.3

Hit 1 (89µM)

Hit 1

Hit 3 (15µM)

Hit 3

N433

L387I385

W374

P375 F376

BRD4-BD2

N433

L387I385

W374

P375F376

BRD4-BD2

BRD4-BD1 BRD4-BD1

N140

L94L92

W81

P82F83 N140

L94L92

W81

P82F83

Docking of hit compounds to BRD4 BD1 and BD2

Co-crystal structures of BRD4 BD2 complexed with lead compound

Co-Crystal structure of Lead 1complexed with BRD4 BD2 (1.4 Å)

Modifications of hit 3 leading to the discovery of lead compound 1 with sub-µM affinities to BETs

KD values of optimized BRD inhibitors to BRD2-4 BD1 and BD2

Inhibitors MW

KD SD (nM)*

BRD2 BRD3 BRD4

BD1 BD2 BD1 BD2 BD1 BD2

IBET-762 424 12911 538 8517 528 8932 426

CMPD 1 415 342 246 189 185 406 287

CMPD 2 413 264 204 183 214 299 132

CMPD 3 450 412 369 218 4218 6018 3611

CMPD 4 423 348 114 177 111 2715 104

* SD obtained from three independent experiments

More than 50 compounds of this class were synthesized with significantly improved affinities.

Binding of CMPD 2 to 6 BET BRD proteins

No clear selectivity among BETs

Ki values to BRD2/4 BD2 determined by competitive FP assay

IDBRD2 BD2 BRD4 BD2

IC50 (nM) Ki (nM) IC50 (nM) Ki (nM)

IBET151 29839 6614 451111 8820

IBET762 14624 2611 14922 218

CMPD 1 28234 386 15915 N/A

CMPD 2 21232 269 25630 327

CMPD 3 33579 6620 379123 6925

CMPD 4 9831 N/A 13130 N/A

O N

N

Cl

NN

NHO

NH

OO

SO3H

NH2

NH2

SO3HO

O

FAM-762

FAM labeled IBET762 was used as tracer.

Only assays for BD2 proteins are available.

Assays for BD1 are not available due to too low dynamic range.

Ki of CMPD 4 can’t be accurately determined since they are lower than assay limits.

Comparison between KD from BLI (Octet RED96) and Ki from FP assays

BRD2 BD2 BRD3 BD2 BRD4 BD2

Ki (nM)(FP)

KD (nM)Octet

RED96

Ki (nM)(FP)

KD (nM)Octet

RED96

Ki (nM)(FP)

KD (nM)Octet

RED96

IBET-762 26±11 538 16±4 528 20±8 426

CMPD 1 38±6 246 21±5 185 8±3 287

CMPD 2 26±9 204 12±3 214 32±7 132

CMPD 3 66±20 369 41±26 4418 69±25 3611

CMPD 4 N/A 114 N/A 111 N/A 104

Compounds with low nM affinities determined by Octet RED96

BRD2 BD2 BRD3 BD2 BRD4 BD2

CMPD 5 12.74.2 13.92.0

CMPD 6 4.10.7 4.50.9 14.41.1

CMPD 7 3.50.7 1.50.3 3.60.1

CMPD 7/BRD3BD2

Very low dissociation rates (koff < 10-3 s-1) Column format required

Compounds with sub-nM affinities determined by Octet RED96

Cao, et al. Mol Cell 2014, 53, 247.

Binding affinities to other BRD Proteins

KD (nM) SelectivityBRD4

BD2/CREBBPBRD4 BD2 ADAT2A ADAT2B ASH1L CREBBP*

IBET-762 426 > 5000** > 5000 > 5000 > 5000 > 100

CMPD 1 287 > 5000 > 5000 > 5000 4758 170

CMPD 2 132 > 5000 > 5000 > 5000 1510400 117

CMPD 3 4014 > 5000 > 5000 > 5000 39591766 100

CMPD 4 104 > 5000 > 5000 > 5000 47982219 470

*CREBBP has the closest structural and sequence homology to BET BRD proteins** Compounds precipitate when concentration higher than 10 µM with 0.5% DMSO.

Cellular activities of selected BET-BRD inhibitors

ID

Cell growth assay IC50 (nM)

MV4;11 MOLM-13 AML-2 K562

MLL-AF4 MLL-AF9 MLL-AF6 BCR-ABL

IBET151 162112 22852 8131 >2000

IBET762 9345 24158 12754 >2000

CMPD 1 2310 7812 7421 >2000

CMPD 2 209 6614 5030 >2000

CMPD 3 3415 14431 10062 >2000

CMPD 4 10 74 72 460

CellTiter-Glo® Luminescent Cell Viability AssayData are meanSD (n=3-4)

Cellular Mechanism of Action (MOA)

AML-2 AML Cell Line with MLL1 Fusion Protein

µM 0 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1

I-BET-762 CMPD 1 CMPD 2 CMPD 3

0 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1

I-BET-762 CMPD 1 CMPD 2 CMPD 3

PARP

Cas-3

c-Myc

Actin

Bcl-2

40 h 60 h

BRD inhibitors induce down-regulation of c-Myc and cleavage of PARP in AML-2 cell line

Similar effects seen in Molm-13 AML cell line with MLL1 Fusion Protein

60 h0 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1 0.1 0.3 1

I-BET-762 CMPD 1 CMPD 2 CMPD 3

PARP

Cas-3

p21

c-Myc

Actin

K562 without MLL1 Fusion Protein

BRD inhibitors do not induce down-regulation of c-Mycand cleavage of PARP in K562 cell line

Cellular Mechanism of Action

BRD inhibitors potently and effectively induce apoptosis

BRD inhibitors induce cell cycle arrest

BRD inhibitors induce G1 arrest within 24 hr.

Antitumor activity of BRD inhibitors in MV4;11 model

CMPD6 has good microsomal stability (T1/2 >50 min) in human (58 min), mouse (57 min) and rat (>60 min) microsomes.

Summary

Novel chemical scaffolds for BET subfamily BRD protein inhibitors were successfully identified by screening low MW (133-300) probe compound library utilizing label free Octet RED96 assay

New classes of BET subfamily BRD protein inhibitors were developed based on rational design and SAR based optimization

3-10 times more potent than GSK clinical trial IBET-762 in both binding and cell viability assays

Good microsomal stability Comparable tumor growth inhibition than GSK IBET

compounds

Medicinal chemistryDr. Yujun ZhaoXu RanDr. Yongqiang ZhuHacer Karata

Biology and in vivo modelDr. Longchuan BaiDonna McEachern

Computational simulationDr. Chao-Yie Yang

X-ray crystallography core facility of the Life Science Institute at U of MichiganDr. Jennifer MeagherDr. Jeanne Stuckey

Pharmocokinetics (PK) core facility at U. of MichiganDr. Duxin SunDr. Xiaoqin Li

Acknowledgement

Welcome to Ann Arbor and University of Michigan

UMHS NCRC (former Pfizer site)