Exploring the Scaffold Universe of Kinase Inhibitors - ACS Publications

Department of Life Science Informatics, B-IT, LIMES Program Unit Chemical Biology ... Our detailed organization of the kinase inhibitor scaffold unive...
0 downloads 0 Views 3MB Size
Subscriber access provided by TULANE UNIVERSITY

Article

Exploring the Scaffold Universe of Kinase Inhibitors Ye Hu, and Jürgen Bajorath J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/jm501237k • Publication Date (Web): 05 Sep 2014 Downloaded from http://pubs.acs.org on September 8, 2014

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Medicinal Chemistry is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Exploring the Scaffold Universe of Kinase Inhibitors

Ye Hu and Jürgen Bajorath*

Department of Life Science Informatics, B-IT, LIMES Program Unit Chemical Biology and Medicinal Chemistry, Rheinische Friedrich-Wilhelms-Universität, Dahlmannstr. 2, D-53113 Bonn, Germany.

ACS Paragon Plus Environment

1

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 55

Abstract

The scaffold concept was applied to systematically determine, analyze, and compare core structures of kinase inhibitors. From publicly available inhibitors of the human kinome, scaffolds and cyclic skeletons were systematically extracted and organized taking activity data, structural relationships, and retrosynthetic criteria into account. Scaffold coverage varied greatly across the kinome, and many scaffolds representing compounds with different activity profiles were identified. The majority of kinase inhibitor scaffolds were involved in welldefined yet distinct structural relationships, which had different consequences on compound activity. Scaffolds exclusively representing highly potent compounds were identified as well as structurally analogous scaffolds with very different degrees of promiscuity. Scaffold relationships presented herein suggest a variety of hypotheses for inhibitor design. Our detailed organization of the kinase inhibitor scaffold universe with respect to different activity and structural criteria, all scaffolds, and the original compound data assembled for our analysis are made freely available.

ACS Paragon Plus Environment

2

Page 3 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Introduction

For nearly three decades, the search for ‘privileged substructures’1 or ‘masterkeys’2 has been a major topic in medicinal chemistry to aid in the design of compounds that preferentially interact with members of a given target family. Such privileged structural motifs, if they can be identified, are typically considered as core structure templates for compound generation.2 Core structures are represented using molecular scaffolds,3 another popular concept in medicinal chemistry. Molecular scaffolds can be derived in different ways,3 for example, by calculating the maximum common core structure of a series of compounds,3 by applying retrosynthetic reaction rules,4 or by removal of R-groups from compounds (for which different rules might be applied).5 From a medicinal chemistry perspective, each scaffold definition has pros and cons for core structure representation and no single scaffold definition has become a generally applied standard. In fact, the term scaffold is often rather loosely used in the literature, often without clear definitions.3 This might be acceptable when individual compounds or small series are investigated but for a systematic determination and analysis of scaffolds, the consistent application of clear scaffold definitions is essential. The probably most widely used scaffold definition follows a molecular hierarchy and extracts scaffolds from compounds by removal of all ‘side chain’ substituents while retaining all ring systems and linkers between them.5 From such hierarchically derived scaffolds, one can further abstract by converting all heteroatoms to carbon and setting all bond orders to one, thus generating so-called ‘cyclic skeletons’,6 which represent subsets of topologically equivalent scaffolds. The hierarchical scaffold definition has the advantage of representing a generally applicable and consistent methodological framework for the generation of molecular scaffolds from essentially all classes of ring-containing compounds, but also has some potential limitations.3 For example, the addition of a ring to a scaffold always defines a new scaffold, although ring additions are often carried out during analog design. Hence, ACS Paragon Plus Environment

3

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 55

hierarchical scaffolds might separate analog series into subseries characterized by differences in ring or linker content, which might increase the complexity of core structure analysis, at least for combinatorial chemistry efforts. Nonetheless, hierarchical scaffolds typically provide a reasonable and chemically interpretable representation of molecular cores. Protein kinases are among the most intensely investigated targets for a variety of therapeutic applications.7,8 Accordingly, a wealth of kinase inhibitors have been reported over the past one or two decades. Most of the currently available inhibitors target the ATP cofactor binding site in kinases.9,10 Given the large numbers of kinase inhibitors that have become available in recent years, we have carried out a comprehensive analysis of inhibitor scaffolds, taking kinase target classification, structural relationships among scaffolds, and kinase inhibitor promiscuity into account. The results of our analysis provide an unprecedentedly detailed classification of systematically derived kinase inhibitor scaffolds. Many sets of scaffolds with different medicinal chemistry relevant characteristics have been generated, providing a wealth of differentiated scaffold information. The entire classification and all structural data are made freely available to the scientific community to provide a basis for further analysis and the exploration of core structures and privileged structural motifs for the development of kinase inhibitors.

ACS Paragon Plus Environment

4

Page 5 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Materials and Methods

Compound activity data From ChEMBL11 release 18, compounds with reported direct interactions (i.e., target relationship type “D”) with human kinase targets at the highest confidence level (i.e., target confidence score 9) were extracted. ChEMBL represents the currently most comprehensive pubic repository of compound data from medicinal chemistry sources and includes information from other databases.11 Two different types of potency measurements were separately considered including (assay-independent) equilibrium constants (i.e., Ki values) and (assay-dependent) IC50 values. Only inhibitors with explicitly defined Ki or IC50 values were considered to ensure a high level of data integrity. Approximate measurements such as “>”, “1000 1000

500

100

50

20

10

5

3

1

5 10 # Kinases

20 1

2-4

≥5

1

388

135

9

2-10

133

60

10

> 10

6

17

3

30 50

# Compounds

>50 1

2 3 4 5 10 # Kinases (promiscuity)

>10

(b) IC50 subset 1 # BM scaffolds 2

# Compounds

4

>1000 1000

500

100

50

20

10

5

3

1

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 48 of 55

5 10 # Kinases

20 1

2-4

≥5

1

4015

1019

100

2-10

1264

558

107

> 10

136

96

48

30 50

# Compounds

>50 1

2 3 4 5 10 # Kinases (promiscuity)

>10

ACS Paragon Plus Environment

48

Page 49 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Figure 15

18 compounds | 30 kinases

3|1

4|1

4|1

2|2

24 | 3

5|3

5|4

36 | 14

ACS Paragon Plus Environment

49

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 50 of 55

Figure 16

ACS Paragon Plus Environment

50

Page 51 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Table 1. Kinase inhibitor data sets.

Number of

Ki

IC50

Inhibitors

1760

17,775

Kinases

94

264

Interactions

2654

25,043

BM scaffolds

761

7343

CSKs

442

3503

For the Ki- and IC50-based kinase inhibitor subsets, the number of kinase inhibitors, kinases these inhibitors were active against, and kinase-inhibitor interactions is reported. In addition, the number of Bemis-Murcko (BM) scaffolds and cyclic skeletons (CSKs) extracted from all kinase inhibitors is given.

ACS Paragon Plus Environment

51

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 52 of 55

Table 2. Promiscuity of structurally related scaffolds.

# Scaffold pairs (structural relationships) Difference in promiscuity (∆promiscuity)

At least one scaffold with ≥ 2 compounds

Total

Both scaffolds with ≥ 2 compounds

Ki

IC50

Ki

IC50

Ki

IC50

0

1920

25,285

1056

13,581

233

3020

1

347

6957

233

4902

59

1323

2

113

2632

101

2063

37

619

3

76

1849

57

1535

13

483

4

33

965

33

813

14

271

5

4

858

4

766

1

252

6

4

373

4

361

0

139

7

-

403

-

399

-

142

8

14

310

3

300

0

91

9

-

108

-

101

-

31

10

-

85

-

83

-

32

11 – 20

-

463

-

417

-

126

> 20

-

71

-

57

-

22

For the Ki- and IC50-based subsets, the number of scaffold pairs forming structural relationships with increasingly large differences in promiscuity rates is reported. In addition, the number of scaffold pairs is reported in which at least one scaffold or both scaffolds represented multiple compounds.

ACS Paragon Plus Environment

52

Page 53 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Table 3. Scaffold-based RECAP-MMPs.

Number of

Ki

IC50

RECAP-MMPs

770

8945

Scaffolds forming RECAP-MMPs

404 (53.1%)

4354 (59.3%)

Same or overlap

766

8348

Distinct

4

597

MMPs, CSK equivalences or substructure relationships

770

8942

None

0

3

Activity comparison

Structural relationships

For the Ki- and IC50-based subsets, the number of scaffold pairs forming transformation sizerestricted RECAP-MMPs is reported. The number (and ratio) of scaffolds involved in the formation of these retrosynthetic MMPs is given. Furthermore, the number of RECAP-MMPs formed by scaffolds with the same or overlapping kinase activity and with distinct activity is reported. In addition, the number of RECAP-MMPs that were also involved in other structural relationships is provided.

ACS Paragon Plus Environment

53

Journal of Medicinal Chemistry

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 54 of 55

Table 4. Promiscuity of synthetically related scaffolds.

# RECAP-MMPs Difference in promiscuity (∆promiscuity)

At least one scaffold with ≥ 2 compounds

Total

Both scaffolds with ≥ 2 compounds

Ki

IC50

Ki

IC50

Ki

IC50

0

640

6825

351

3426

89

776

1

91

1080

60

742

9

184

2

23

398

21

310

6

76

3

11

171

10

134

2

36

4

4

110

4

86

1

27

5

1

81

1

70

1

14

6

-

48

-

47

-

14

7

-

56

-

53

-

16

8

-

65

-

63

-

8

9

-

19

-

17

-

4

10

-

16

-

16

-

4

11 – 20

-

58

-

49

-

14

> 20

-

18

-

10

-

4

For the Ki- and IC50-based subsets, the number of RECAP-MMPs formed by scaffold pairs with increasingly large differences in promiscuity is reported. In addition, the number of RECAP-MMPs is given in which at least one scaffold or both scaffolds represented multiple compounds.

ACS Paragon Plus Environment

54

Page 55 of 55

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Medicinal Chemistry

Table of Contents Graphic

ACS Paragon Plus Environment

55