Atropisomerism by Design: Discovery of a Selective and Stable

Subscriber access provided by STEPHEN F AUSTIN STATE UNIV

Article

Atropisomerism by Design: Discovery of a Selective and Stable Phosphoinositide 3-Kinase (PI3K) # Inhibitor Jayaraman Chandrasekhar, Ryan Dick, Joshua Van Veldhuizen, David J Koditek, Eve-Irene Lepist, Mary E. McGrath, Leena Patel, Gary Phillips, Kassandra Sedillo, John R. Somoza, Joseph Therrien, Nicholas Alexander Till, Jennifer Treiberg, Armando Garcia Villaseñor, Yelena Zherebina, and Stéphane Perreault J. Med. Chem., Just Accepted Manuscript • DOI: 10.1021/acs.jmedchem.8b00797 • Publication Date (Web): 17 Jul 2018 Downloaded from http://pubs.acs.org on July 17, 2018

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 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 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.

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 30 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

Atropisomerism by Design: Discovery of a Selective and Stable Phosphoinositide 3-Kinase (PI3K) b Inhibitor Jayaraman Chandrasekhar,1 Ryan Dick,2 Joshua Van Veldhuizen,1 David Koditek,2 Eve-Irene Lepist,† Mary E. McGrath,2 Leena Patel,1 Gary Phillips,1 Kassandra Sedillo,1 John R. Somoza,2 Joseph Therrien,1 Nicholas A. Till,† Jennifer Treiberg,1 Armando Villaseñor,2 Yelena Zherebina,2 and Stephane Perreault*,1

1

Gilead Sciences, Inc., 199 E Blaine Street, Seattle, Washington 98102, United States

2

Gilead Sciences, Inc., 333 Lakeside Drive, Foster City, California 94404, United States

ABSTRACT Atropisomerism is a type of axial chirality in which enantiomers or diastereoisomers arise due to hindered rotation around a bond axis. In this manuscript, we report a case in which torsional scan studies guided the thoughtful creation of a restricted axis of rotation between two heteroaromatic systems of a phosphoinositide 3-kinase (PI3K) b inhibitor, generating a pair of atropisomeric compounds with significantly different pharmacological and pharmacokinetic profiles. Emblematic of these differences, the metabolism of inactive (M)-28 is primarily due to the cytosolic enzyme aldehyde oxidase, while active (P)-28 has lower affinity for aldehyde oxidase, resulting in substantially better metabolic stability. Additionally, we report torsional scan and experimental studies used to determine the barriers of rotation of this novel PI3Kb inhibitor.

ACS Paragon Plus Environment

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

INTRODUCTION Chirality exists in different forms. Although central or point chirality (a central atom bearing different substituents) is by far the most commonly encountered, other forms of chirality, such as axial chirality, helical chirality, and planar chirality, have been investigated over the years.1 Atropisomerism is a type of axial chirality in which enantiomers or diastereoisomers arise due to restricted or hindered rotation around a bond axis.2 Atropisomer interconversion occurs via an intramolecular dynamic process that only involves bond rotation. This time-dependent event is primarily affected by steric hindrance, stereoelectronic effects, solvent, and temperature. Similar to central stereoisomers, atropisomeric drug molecules have been shown to exhibit distinct biological activity on- and off-target.3,4 Differences in toxicological and pharmacokinetic properties are also encountered. The presence of an axis of atropisomerism can complicate several aspects of drug discovery, including synthesis, analytical characterization, drug metabolism, and pharmacokinetics. Strategically, these potential issues can be alleviated by introducing symmetry to remove the axis of chirality or reducing steric hindrance to allow free rotation. Alternatively, purposely increasing steric hindrance around atropisomeric axes to prevent interconversion can potentially result in improved potency, selectivity, safety, and chemical stability.5 The rate of interconversion of an atropisomeric drug molecule is an important parameter to consider.6 Many compounds with atropisomeric centers advance into development, but the majority of them have low rotational barriers (DErot) and are classified as class 1 atropisomers, meaning they are rapidly equilibrating mixtures with all DErot < 20 kcal/mol.6a It has been recommended that an atropisomeric compound containing one or more atropisomeric axes with DErot > 30 kcal/mol (class 3) should be developed as a single atropisomer. Compounds in class 2 (20 kcal/mol < DErot < 30 kcal/mol) can be challenging to develop given that stereochemical integrity can be compromised over the time course of drug production and administration to patients.


Page 2 of 30

Page 3 of 30 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


We are interested in the development of selective inhibitors of the Class I PI3Kb lipid kinase for the treatment of various malignancies.7 Class I PI3Ks comprise four different isoforms (PI3Ka, PI3Kb, PI3Kd, and PI3Kg) and catalyze the phosphorylation of phosphatidylinositol-4,5-bisphosphate (PIP2) to produce the second messenger phosphatidylinositol-3,4,5-trisphosphate (PIP3), which propagates intracellular signaling.8 The most common genetic alteration of the PI3Kb signaling pathway found in human cancer is the inactivation of the phosphatase and tensin homolog (PTEN) tumor suppressor gene. Inactivation of PTEN leads to loss of its lipid phosphatase activity, causing accumulation of PIP3 and aberrant downstream signaling.9 Preclinical studies have shown that down-regulation of PI3Kb in PTENdeficient cancer cells results in pathway inactivation and subsequent inhibition of growth in both cellbased and in vivo settings.10 In this manuscript, we describe the utility of atropisomerism in the discovery of a potent, selective, and metabolically stable PI3Kb inhibitor. Torsional scan studies guided the strategic design of a restricted axis of rotation around a carbon-nitrogen bond between two heteroaromatic systems, generating a pair of atropisomeric compounds with differential pharmacological and pharmacokinetic profiles. Additionally, we report torsional scan and experimental studies used to determine the rotational barriers of this novel PI3Kb inhibitor. RESULTS AND DISCUSSION A compound arising from our PI3K program is the ATP competitive PI3Kb inhibitor 1 (Figure 1). The binding mode of 1 has not been characterized crystallographically in PI3Kb. However, it can be modeled in a published co-crystal structure of mouse p110b (catalytic subunit of PI3Kb, residues 114-1064, pdb: 4BFR).7c Compound 1 is likely to bind to the ATP binding site of the kinase domain of p110β with the 2aminopyridine moiety acting as a two-point hinge binder with Val848. The quinoline adopts an orthogonal conformation relative to the benzimidazole in order to partially fill the specificity pocket



between Met773 and Trp781. The triazole is oriented towards the affinity pocket, and Lys799 potentially interacts with both the triazole and the benzimidazole nitrogen atoms.

Figure 1. Docked pose of 1 in the co-crystal structure of mouse p110b (catalytic subunit of PI3Kb, residues 114-1064, pdb: 4BFR, coordinates included in Supporting Information). The quinoline adopts an orthogonal conformation relative to the benzimidazole. For clarity, some residues have been removed. Blue dashed lines show hydrogen bond contacts between the inhibitor and the protein.

Based on this modeled orthogonal binding mode of compound 1 in PI3Kb, we envisioned restricting the free rotation around the C–N bond connecting the two bicyclic systems by generating an axis of atropisomerism. The general idea behind this strategy was to minimize the loss in entropy to reach the preferred binding conformation, ultimately resulting in a gain in binding free energy. Additionally, in the context of kinase inhibitor research, this approach can possibly lead to improved kinome selectivity resulting from a limitation of accessible conformations.5d


Page 4 of 30

Page 5 of 30 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


Consequently, we computationally explored the rotational barriers (DErot) of a large set of 4-(1Hbenzo[d]imidazol-1-yl)quinolines with different substitution patterns that could impact the rate of interconversion between atropisomers (Table 1). In the torsional scan studies using molecular mechanics (MMFFs force field), the rotation around the C–N bond connecting the quinoline and the benzimidazole was driven in both directions (clockwise and counterclockwise) for both possible transition states (TSin and TSout).11 As the bond torsion approaches coplanarity, the strain energy increases. The best estimates for the barriers via TSin and TSout were obtained from the lowest energy pathway in either direction. For compounds with extremely hindered rotation, the truly coplanar states could never be achieved, revealing significant hysteresis due to high energy penalties.12 The rotational barriers of those compounds are reported as >40 kcal/mol. For all monosubstituted compounds (2–9), bond rotation via TSout is consistently predicted to have the lowest energy pathway. The analogs are all indicated to be class 1 or 2 in terms of the magnitude of the barrier for atropisomer interconversion according to Laplante et al.6a The 2-methyl substituent on the benzimidazole (9) leads to the highest TSout rotational barrier of this group with a DErot = 28.3 kcal/mol. Disubstitution appears to be needed to shift the analogs to class 3 category. With a few exceptions, the lowest energy pathway for the disubstituted compounds 10–27 is also TSout. Among those compounds, ten have DErot > 30 kcal/mol, putting them in the class 3 atropisomer category. The combination of a 2substituted benzimidazole with a 3- or 5-substituted quinoline appears to be a consistently reliable way to create non-interconverting, class 3 atropisomerism.



Table 1. Rotational barriers of related 4-(1H-benzo[d]imidazol-1-yl)quinolinesa

Monosubstituted 4-(1H-benzo[d]imidazol-1-yl)quinolines Cpd

A

B

C

D

D Erot (TSin) (kcal/mol)

D Erot (TSout) (kcal/mol)

2

H

H

H

H

29.6

14.1

3

H

F

H

H

33.7

15.5

4

H

CH3

H

H

>40

22.5

5

H

H

F

H

30.1

24.3

6

H

H

CH3

H

31.7

26.6

7

H

H

H

F

35.0

20.2

8

H

H

H

CH3

>40

18.5

9

CH3

H

H

H

32.2

28.3

Disubstituted 4-(1H-benzo[d]imidazol-1-yl)quinolines

a

B

C

D

D Erot (TSin) (kcal/mol)

D Erot (TSout) (kcal/mol)

CH3

F

H

H

37.7

33.9

CH3

CH3

H

H

45.4

43.7

12

CH3

H

F

H

34.1

32.4

13

CH3

H

CH3

H

>40

>40

14

CH3

H

H

F

35.4

28.7

15

CH3

H

H

CH3

>40

>40

16

H

F

F

H

32.5

27.3

17

H

F

CH3

H

33.4

39.2

18

H

CH3

F

H

39.6

27.6

19

H

CH3

CH3

H

38.2

>40

20

H

F

H

F

>40

20.5

21

H

F

H

CH3

>40

27.2

22

H

CH3

H

F

>40

23.5

23

H

CH3

H

CH3

>40

>40

24

H

H

F

F

35.0

23.1

25

H

H

F

CH3

>40

25.1

26

H

H

CH3

F

36.1

30.6

27

H

H

CH3

CH3

>40

>40

Cpd

A

10 11

Using MMFFs force field, the best estimates for the rotational barriers DErot via TSin and TSout were obtained from the lowest energy pathway in either direction

(clockwise or counterclockwise). The lowest energy pathway is highlighted in blue. Because of significant distortion, all values above 40 kcal/mol are reported as >40 kcal/mol.


Page 6 of 30

Page 7 of 30 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


According to the previous estimated rotational barriers, compound 1, which is closely related to compound 3, is predicted to have a low rotational barrier (DErot < 20 kcal/mol). Applying the findings of compound 10, we envisaged restricting free rotation around the C–N bond of 1 by introducing a methyl group at the C2-position of the benzimidazole in combination with the 5,8-difluoroquinoline ((rac)-28, Table 2). This modification results in enhanced PI3Kb biochemical potency, presumably by minimizing entropic loss to reach the preferred binding conformation; it also affords improved selectivity over the other three Class I PI3K isoforms.13 Chiral supercritical fluid chromatography (SFC) analysis of (rac)-28 revealed the presence of atropisomers, which were not observed for compound 1. The C2-methyl group likely hinders rotation around the C–N bond due to a larger steric clash with the C5-fluorine of the quinoline moiety. This observation encouraged us to separate both atropisomers by chiral preparative SFC to provide enantiopure (M)-28 and (P)-28 (Table 2). Upon evaluation in our Class I PI3K biochemical assays, the single atropisomers (M)-28 and (P)-28 exhibit substantially different potency profiles in all four isoforms. (P)28 displays an IC50 of 2 nM against PI3Kb with good selectivity over the other three isoforms, while (M)28 is significantly less active.14



Table 2. Potency profiles against Class I PI3Ksa

a

1

(rac)-28

(M)-28

(P)-28

PI3Ka IC50 (nM)

233

264

>10,000

188

PI3Kb IC50 (nM)

11

3

1423

2

PI3Kd IC50 (nM)

67

43

>10,000

42

PI3Kg IC50 (nM)

440

4156

>10,000

4269

The activity against each Class I PI3K was evaluated in in vitro kinase assays containing 2*Km steady state concentrations of ATP (average of ≥ 2 determinations;

see Supporting Information for more details).

The absolute configurations of the single atropisomers were unambiguously determined by X-ray crystallography of (P)-28 bound in the PI3Kd ATP binding site (Figure 2, pdb: 6DGT).15 The axial configurations were determined by analysis of Newman projections and assignment of atom priority according to the CIP-system.2b,16 As predicted, the 2-aminopyridine moiety acts as a two-point hinge binder, making two hydrogen-bond interactions with Val828. The quinoline occupies the specificity pocket between Met752 and Trp760, a closed pocket in the apo structure of the enzyme.16,17 The dihedral angle of the quinoline and benzimidazole C–N bond is 85°. The triazole is directed towards the affinity pocket, making hydrogen bond contacts with Lys779, Asp787, and Tyr813.


Page 8 of 30

Page 9 of 30 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


Figure 2. 2.60 Å X-ray crystal structure of (P)-28 bound in the PI3Kd ATP binding site (pdb: 6DGT). For clarity, some residues have been removed. Blue dashed lines show hydrogen bond contacts between the inhibitor and the protein.

The limited number of accessible conformations of (P)-28 may impart improved selectivity over other protein and lipid kinases that show reduced conformational flexibility of their ATP pocket. Profiling of the freely rotating compound 1 and the locked single atropisomer (P)-28 against more than 400 human kinases using the DiscoveRx KINOMEscan® platform at a compound concentration of 10 µM revealed superior selectivity of the atropisomeric compound (P)-28 (Figure 3a for 1 and Figure 3b for (P)-28).18 In this active site-directed competition assay, (P)-28 has higher affinity binding for PI3Kb (lager blue circle) and still shows considerably less off-kinase activity (fewer and smaller red circles).



Figure 3. Protein and lipid kinase selectivity profiles of 1 (Figure 3a) and (P)-28 (Figure 3b) using the DiscoveRx KINOMEscan® platform at a compound concentration of 10 µM (see Supporting Information for data report). PI3Kb highlighted with a blue circle. Image generated using TREEspot™ Software Tool and reprinted with permission from KINOMEscan®, a division of DiscoveRx Corporation, © DISCOVERX CORPORATION 2010.

Further characterization of the single atropisomers revealed very low intrinsic clearance in human hepatocytes (hHep) for the potent atropisomer (P)-28 (0.08 L/h/kg), contrasted with much a higher value for (M)-28 (1.97 L/h/kg) (Table 3). The large difference in intrinsic hepatic clearances prompted us to determine their respective metabolic profiles in cryopreserved human hepatocytes (see Supporting Information for complete analyses of the metabolic profiles of (M)-28 and (P)-28)). In the case of metabolically less stable (M)-28, the major metabolites result from the oxidation of the quinoline moiety to the corresponding quinolinone M1 (mainly observed as glucuronide adduct) and the formation of three unique metabolites stemming from triazole oxidation (M2-4, Figure 4). Interestingly, the analogous oxidation of the quinoline could not be identified and the oxidative turnover of the triazole seems to be much slower for (P)-28 in the same experiment. The latter observations are consistent with the lower intrinsic hepatic clearance value of the single atropisomer (P)-28. Table 3. Intrinsic clearances and fraction metabolized by AO (M)-28

(P)-28

a

1.97

0.08

b

0.15

Atropisomerism by Design: Discovery of a Selective and Stable

Recommend Documents