Subscriber access provided by UNIV OF NEW ENGLAND ARMIDALE
Discovery of Novel KRAS-PDE# Inhibitors by Fragment-based Drug Design Long Chen, Chunlin Zhuang, Junjie Lu, Yan Jiang, and Chunquan Sheng J. Med. Chem., Just Accepted Manuscript • Publication Date (Web): 06 Mar 2018 Downloaded from http://pubs.acs.org on March 6, 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 7 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
Discovery of Novel KRAS-PDEδ Inhibitors by Fragment-based Drug Design Long Chen§, Chunlin Zhuang§, Junjie Lu§, Yan Jiang, Chunquan Sheng* School of Pharmacy, Second Military Medical University, 325 Guohe Road, Shanghai 200433, China Supporting Information ABSTRACT: Targeting KRAS-PDEδ protein-protein interactions with small molecules represents a promising opportunity for developing novel antitumor agents. However, current KRAS-PDEδ inhibitors are limited by poor cellular antitumor potency, and the druggability of the target remains to be validated by new inhibitors. To tackle these challenges, herein, novel, highly potent KRAS-PDEδ inhibitors were identified by fragment-based drug design, providing promising lead compounds or chemical probes for investigating the biological functions and druggability of KRAS-PDEδ interaction.
INTRODUCTION 1, 2
Oncogenic KRAS signaling is an important antitumor pathway. The KRAS protein is often mutated in different kinds of cancers and especially in a large proportion (90%) of pancreatic cancers.3, 4 Targeting KRAS signaling is becoming an important field in anticancer drug discovery and has achieved great success.5-9 Recently, inhibition of mammalian KRAS-PDEδ protein-protein interaction (PPI) by small molecules has been seen as a promising opportunity for the discovery of novel antitumor agents.10-16 PDEδ, also named PDE6D, determines the KRAS dynamic distribution in the cell.17-19 The farnesylated KRAS protein is solubilized after binding with PDEδ, which enhances KRAS diffusion throughout the cell.7, 18, 20, 21 After discharged from the PDEδ binding pocket, the farnesylated KRAS is trapped by the recycling endosome and then re-localized to the plasma membrane by vesicular transport. Aberrant oncogenic signaling ultimately results from the high concentration of KRAS at the plasma membrane.18,
ing lead compounds or chemical probes for investigating the biological functions and druggability of KRAS-PDEδ interactions.
22
Recently, Waldmann’s group reported pioneering works in the discovery of small-molecule KRAS-PDEδ inhibitors (Figure 1), including benzimidazole inhibitor deltarasin (1),10 pyrazolopyridazinone inhibitor deltazinone (2),11, 14 and bis(sulfonamide) inhibitor deltasonamide (3).16 Compound 1 was the first PDEδ inhibitor with nanomolar binding affinity; however, it was a non-selective inhibitor and showed apparent cytotoxicity.10 Inhibitors 2 and 3 showed better selectivity toward PDEδ than inhibitor 1, but they were limited by poor cellular antitumor potency.15 Previously, our group reported novel bis-quinazolinone 4 and quinazolinone-pyrazolopyridazinone 5 as inhibitors based on a virtual screening and structural biology-guided drug design.13 Compound 5 exhibited excellent binding affinity (KD = 2 ± 0.5 nM) and moderate antitumor activity (IC50 = 18.6 ± 1.3 µM) against Capan-1 pancreatic cancer cells.13 Nevertheless, its cellular potency needs improvement. Additionally, a new class of KRAS-PDEδ inhibitors is highly desirable to validate the druggability of the target. Fragment-based drug design (FBDD) is emerging as a powerful tool in drug discovery.23-27 In this study, novel KRAS-PDEδ inhibitors were identified using a FBDD strategy based on the co-crystal complexes of fragment-like inhibitors with PDEδ.10, 11, 13, 16 Interestingly, several new inhibitors showed improved cellular antitumor activities, which makes them promis-
Figure 1. Chemical structures of representative KRAS-PDEδ inhibitors.
RESULTS AND DISCUSSION
FBDD of KRAS-PDEδ Inhibitors. Recently, our group identified quinazolinone fragment hit 6 (Figure 2) as a novel KRASPDEδ inhibitor (KD = 467 ± 65 nM) using a structure-based virtual screening (SBVS).13 The crystal structure of the complex of PDEδ with fragment 6 (PDB entry: 5X73) indicated that the two molecules of inhibitor 6 bound to Arg61 and Tyr149 in the active site mainly through hydrogen-bonding interactions (Figure 2A). Similarly, Waldmann’s work revealed that two molecules of benzimidazole fragment 7 (KD = 165 ± 23 nM, Figure 2) could also bind in the two active sites (Figure 2C, PDB entry: 4JV6).10 Inspired by these binding modes, computational FBDD28 was used to rationally design novel PDEδ inhibitors by linking the two fragment-like inhibitors. Based on molecular docking and intermolecular distance measurements, fragment-linking was proposed at different sites on the two inhibitors. First, the distance from the benzene ring of benzimidazole inhibitor 7 to the nitrogen atom of
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
the amide in quinazolinone inhibitor 6 is 5.3 Å (Figure 2B). Second, the distance between the benzene ring of inhibitor 6 to the imidazole nitrogen atom of inhibitor 7 is 5.0 Å (Figure 2D). These distances were both suitable for using an ether linker between the two methylenes. Thus, two series of compounds were designed and docked into the PDEδ protein. As shown in Figure 2E and 2F, two representative compounds (8a and 9a) could insert well into the binding pocket and form hydrogen bonds with Arg61 and Tyr149, mimicking the binding modes of the unlinked fragment-like inhibitors 6 and 7 (Figure S1A and S1B in Supporting Information). Although the predicted orientation of the benzimidazole group in compounds 9a and 7 was flipped (Figure S1B and S1C in Supporting Information), the hydrogen-bonding interaction with Tyr149 was retained.
Page 2 of 7
7-substituted derivatives (8e-8g) were slightly lower than that of 8a. Among the prepared compounds, compound 8b, with a 6fluoro substituent, had the best PDEδ inhibitory activity (KD = 4.5 ± 1.1 nM). However, compounds 8a-8g showed no cellular antitumor activity toward Capan-1 pancreatic cancer cells. The ligand efficiency (LE)29 of the target compounds was also calculated (Table S1 in Supporting Information). Moderate LE of the fragment-like inhibitors 6 (LE= 0.08) and 7 (LE= 0.20) was observed. After fragment linking, compounds 8a-g showed improved LE (LE range: 0.23 ~ 0.31). Table 1. Chemical structures, PDEδ binding affinities, and cellular level activities of compounds 8a-g. O
N 5.3 Å O
Link
N
er
O
N
R
N
N H
N
N
F
N H F
8a-8g N
Compound
R
PDEδ[a] (KD, nM)
Capan-1[b] (IC50, µM)
8a
H
9.0 ± 2.3
> 100
8b
6-F
4.5 ± 1.1
> 100
8c
6-Cl
11.6 ± 2.6
> 100
8d
6-CH3
9.3 ± 2.5
> 100
8e
7-F
12.1 ± 3.6
> 100
8f
7-Cl
16.0 ± 3.4
> 100
8g
7-OCH3
9.8 ± 2.2
> 100
2
/
8±3
48 ± 6
5
/
2 ± 0.5
18.6 ± 1.3
[a]
Determined by fluorescence anisotropy assay. od.
Figure 2. Design of novel KRAS-PDEδ inhibitors by a computational FBDD strategy. (A) Co-crystal structure of quinazolinone fragment 6 with PDEδ protein (PDB entry: 5X73); (B) the first fragment-linking strategy. (C) Co-crystal structure of benzimidazole fragment 7 with PDEδ protein (PDB entry: 4JV6); (D) the second fragment-linking strategy. (E) Proposed binding mode of compound 8a with PDEδ protein; (F) proposed binding mode of compound 9a with PDEδ protein. Yellow dashed lines represent hydrogen-bonding interactions. Red dashed lines represent intermolecular distance measurements.
Biological Evaluations, Structure-activity Relationship and Binding Modes. Initially, fragment-like PDEδ inhibitors 6 and 7 were linked by an ethyl ether (–(CH2)2O-) linker, and pyridine/benzene replacement was used to improve the water solubility. Based on these structural features, compound 8a was synthesized and assayed. Compared to inhibitors 6 and 7, compound 8a (KD = 9.0 ± 2.3 nM) showed improved PDEδ inhibitory activity, and its activity was comparable to that of compound 2 (KD = 8 ± 3 nM). When substituents were introduced at the 6 position of the quinazolinone scaffold, the resulting compounds (8b-8d, KD range: 4.5 ~ 11.6 nM) showed superior or comparable binding affinities to that of compound 8a. In contrast, the activities of the
[b]
Determined by the CCK8 meth-
When the position of the fragment linker was changed to the quinazolinone N3-phenyl group and the imidazole nitrogen atom, the resulting compounds (9a-9c) generally showed decreased binding affinity (KD range: 146 ~ 192 nM), low LE (0.10 ~ 0.13, Table S1) and no in vitro antitumor activity (Table 2). The removal of the benzene ring from the benzimidazole scaffold increased the binding affinity for PDEδ (compounds 9d-9i, KD range: 22 ~ 170 nM). Also, compounds 9d, 9e and 9g showed improved LE (0.22 ~ 0.30, Table S1). The phenyl substitutions on the imidazole ring were important for the inhibitory activity. For example, methyl derivative 9e and unsubstituted derivative 9f were less potent than the corresponding compounds with the phenyl substituents. To improve the water solubility, further pyridine/benzene replacement provided compounds 9g-9i, which showed similar activities. Among 9g-9i, compound 9i possessed moderate inhibitory activity against Capan-1 cells (IC50 = 52 ± 2.5 µM), which is comparable to that of compound 2 (IC50 = 48 ± 6 µM). Table 2. Structures, PDEδ binding affinities, and cellular level activities of compounds 9a-i designed by the fragment-linking strategy.
ACS Paragon Plus Environment
Page 3 of 7 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 O O
N N
O N
5.0 Å Linker
N
N H
N
N
R Ar 9a-9c
HN
O O
F
N N
N H
R
N
Ar 9f-9i
Compound
Ar
R
PDEδ[a] (KD, nM)
Capan-1[b] (IC50, µM)
9a
2-F-Ph
Ph
168 ± 25
> 100
9b
2-F-Ph
CH3
146 ± 22
> 100
9c
pyridin-2-yl
Ph
192 ± 35
> 100
9d
2-F-Ph
Ph
29 ± 11
> 100
9e
2-F-Ph
CH3
53 ± 19
> 100
9f
2-F-Ph
H
80 ± 15
> 100
9g
pyridin-2-yl
Ph
22 ± 8
52 ± 2.5
spectively. In particular, compound 11b showed good antitumor activity (Capan-1, IC50 = 8.8 ± 2.4 µM), and it was more potent than compound 2 (Capan-1, IC50 = 48 ± 6 µM) and quinazolinone inhibitor 5 (Capan-1, IC50 = 18.6 ± 1.3 µM). When bulkier groups, namely, cyclobutyl (11c), cyclopentyl (11d) and cyclohexyl (11e), were introduced, their PDEδ inhibitory activities and cytotoxicities were lower probably due to steric hindrance. Quantitative activity-activity relationships (QAAR)30 analysis indicated that there was good positive correlation (R2 = 0.62, Figure S2 in Supporting Information) between the inhibitory activity against PDEδ and Capan-1 cells. Furthermore, the binding mode of compounds 10g and 11b was investigated by molecular docking studies. As shown in Figure 3B, the binding mode of compound 11b was similar to that of compound 10g. The cyclopropyl group of 11b formed additional hydrophobic interactions with Ile129, Val145 and Leu147. Finally, compound 11b had a good combination of PDEδ binding affinity (KD = 38 ± 17 nM) and cytotoxicity (IC50 = 8.8 ± 2.4 µM) and was therefore subjected to further biological evaluations. Table 3. Structures, PDEδ binding affinities, and cellular level activities of compounds 10a-g and 11a-e. N
O
[a]
pyridin-2-yl
CH3
166 ± 33
> 100
9i
pyridin-2-yl
H
170 ± 28
> 100
2
/
/
8±3
48 ± 6
5
/
/
2 ± 0.5
18.6 ± 1.3
Determined by fluorescence anisotropy assay. od.
[b]
F
N H
10a-10g
11a-11e
R
R’
PDEδ[a] (KD, nM)
Capan-1[b] (IC50, µM)
10a
H
H
67 ± 24
34 ± 3.4
10b
6-CH3
H
56 ± 26
42 ± 3.9
10c
6-OCH3
H
146 ± 51
74 ± 6.5
10d
7-CH3
H
105 ± 33
78 ± 7.2
10e
7-OCH3
H
167 ± 45
82 ± 6.4
10f
7-F
H
101 ± 34
35 ± 3.2
10g
6-F
H
85 ± 27
33 ± 2.9
11a
H
-CH3
37 ± 22
47 ± 8.4
11b
H
38 ± 17
8.8 ± 2.4
11c
H
50 ± 25
11.6 ± 3.8
11d
H
60 ± 31
12.4 ± 4.3
11e
H
81 ± 30
54.6 ± 7.6
2
/
/
8±3
48 ± 6
5
/
/
2 ± 0.5
18.6 ± 1.3
Determined by fluorescence anisotropy assay. od.
Notably, there was a small hydrophobic pocket around the quinazolinone NH and no hydrogen-bonding interaction was observed, suggesting that the incorporation a suitable hydrophobic group improved the binding affinity. Thus, as expected, when methyl and cycloalkyl groups were introduced, the resulting compounds (11a and 11b, respectively) showed improved PDEδ inhibitory activities (KD = 37 ± 22 nM and KD = 38 ± 17 nM), re-
F
Compound
[a]
Figure 3. Proposed binding mode of compounds 10g (A) and 11b (B) with PDEδ.
N
N R'
F
Inspired by our previously identified quinazolinone inhibitors (4 and 5),13 the phenyl group in compound 9f was replaced by a piperidinyl group. Interestingly, resulting compound 10a showed improved cellular activity (Capan-1, IC50 = 34 ± 3.4 µM). However, further introduction of substituents on the quinazolinone scaffold did not have positive effects on the cytotoxicity. Subsequently, molecular docking studies indicated that compound 10g could insert well into the two binding pockets of PDEδ and form hydrogen bonds with both Arg61 and Glu88 (Figure 3A).
N
N
R
Determined by the CCK8 meth-
N
O N
N
9h
N
[b]
Determined by the CCK8 meth-
Compound 11b Inhibited the Phosphorylation of Akt and Erk in Capan-1 Cells. The RAS family regulates the MAPK and PI3K-Akt-mTOR pathways4 and influences cell growth, proliferation, and differentiation.6 Thus, the effects of the most potent compound (11b) on the phosphorylation of extracellular signalregulated kinase (Erk) and protein kinase B/Akt were evaluated. The epidermal growth factor (EGF) was used to induce the expression of MAPK and Akt.13 Phosphorylation levels of Erk and Akt were determined using Capan-1 cells unstimulated and stimulated with EGF and then treated with compounds 11b or 2 for 1 h (Figure 4). When unstimulated cells were treated with compound
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
2 (10 µM), no differences were observed in Akt and Erk phosphorylation. When compound 11b (10 µM) was used, Akt phosphorylation was downregulated, but there was no effect on Erk phosphorylation. At 20 µM, compound 2 significantly decreased Akt phosphorylation and slightly decreased Erk phosphorylation. The phosphorylation of both Akt and Erk was dramatically reduced by treatment with compound 11b at the same dose. When EGF-stimulated cells were treated with compound 2, there was no change in the phosphorylation of Akt, but the phosphorylation of Erk decreased in a dose-dependent manner. In contrast, treatment with compound 11b significantly downregulated the phosphorylation of both Akt and Erk.
Page 4 of 7
membrane, and compound 2 was less potent at the same concentration (20 µM).
Figure 5. (A) Apoptosis in the Capan-1 cell line induced by 48 h of treatment with compound 2 (50 µM) and compound 11b (10 µM and 25 µM). (B) Immunostaining of Capan-1 with anti-panRAS (red) and anti-PDEδ (green) after treatment with various compounds for 2 h (2, 20 µM; 11b, 20 µM). The vehicle control for the above experiments was 0.1% DMSO. (∗∗∗) p < 0.001.
Figure 4. (A) Phosphorylation levels of Erk and Akt using oncogenic KRAS-dependent Capan-1 cells that were unstimulated or stimulated with EGF (125 ng/mL, 5 min). From top to bottom: phosphorylated Akt on S473 (p-Akt), total level of Akt1 (t-Akt1), phosphorylated Erk on Thr202 and Tyr204 (p-Erk), total level of Erk (t-Erk), and loading control (GAPDH). (B) Gray intensity analysis of the Western blots in which the quantification of pErk/t-Erk ± SEM (left) and p-Akt/t-Akt ± SEM (right) was normalized to the EGF-stimulated control (0.1% DMSO). (∗) p
