Article pubs.acs.org/jmc
Discovery of Potent and Selective Sirtuin 2 (SIRT2) Inhibitors Using a Fragment-Based Approach Huaqing Cui, Zeeshan Kamal, Teng Ai, Yanli Xu, Swati S. More, Daniel J. Wilson, and Liqiang Chen* Center for Drug Design, Academic Health Center, University of Minnesota, 516 Delaware Street SE, Minneapolis, Minnesota 55455, United States S Supporting Information *
ABSTRACT: Sirtuin 2 (SIRT2) is one of the sirtuins, a family of NAD+-dependent deacetylases that act on a variety of histone and non-histone substrates. Accumulating biological functions and potential therapeutic applications have drawn interest in the discovery and development of SIRT2 inhibitors. Herein we report our discovery of novel SIRT2 inhibitors using a fragment-based approach. Inspired by the purported close binding proximity of suramin and nicotinamide, we prepared two sets of fragments, namely, the naphthylamide sulfonic acids and the naphthalene−benzamides and −nicotinamides. Biochemical evaluation of these two series provided structure−activity relationship (SAR) information, which led to the design of (5-benzamidonaphthalen-1/2-yloxy)nicotinamide derivatives. Among these inhibitors, one compound exhibited high anti-SIRT2 activity (48 nM) and excellent selectivity for SIRT2 over SIRT1 and SIRT3. In vitro, it also increased the acetylation level of αtubulin, a well-established SIRT2 substrate, in both concentration- and time-dependent manners. Further kinetic studies revealed that this compound behaves as a competitive inhibitor against the peptide substrate and most likely as a noncompetitive inhibitor against NAD+. Taken together, these results indicate that we have discovered a potent and selective SIRT2 inhibitor whose novel structure merits further exploration.
■
INTRODUCTION Sirtuins are commonly classified as class-III histone deacetylases (HDACs) that use NAD+ as a cofactor to remove an acetyl group from the lysine residue, whereas classical HDACs are zinc-dependent. There are seven human sirtuins (SIRT1− SIRT7) with distinctive subcellular localizations and enzymatic functions.1 SIRT1 is mainly a nuclear protein, while SIRT2 predominantly resides in the cytoplasm. SIRT3−5 are mitochondrial, but SIRT6 and SIRT7 are localized to the nucleus. As depicted in Figure 1, SIRT1−3 remove the acetyl group from the acetylated lysine residue of substrate 1 in the presence of NAD+, releasing substrate 2 bearing a lysine, 2′-Oacetyl-ADP-ribose (3), and its regioisomer 3′-O-acetyl-ADPribose.2 Nicotinamide (4) (Figure 1), another product of sirtuin-catalyzed reactions, is a physiological, albeit weak, inhibitor of sirtuins.3 In contrast to SIRT1−3, SIRT4 lacks deacetylase activity but catalyzes ADP-ribosylation. Even though SIRT5 was identified as a deacetylase, it has recently been reported to primarily possess demalonylation and desuccinylation activities.4 SIRT6 catalyzes both deacetylation and ADP-ribosylation. SIRT7, the least-studied sirtuin, has recently been shown to deacetylate histone H3K18.5 Sirtuins deacetylate histone as well as numerous cellular substrates, and they have been implicated in a wide range of biological functions, including chromatin modification, gene transcription, metabolism, mitosis, stress response, DNA damage repair, and cell survival.1b,6 Consequently, modulation of sirtuins has emerged as a promising therapeutic strategy to © XXXX American Chemical Society
treat diverse diseases. Activation of SIRT1 has been proposed as a potential therapy for aging-related diseases such as metabolic syndromes and neurodegenerative diseases, including Alzheimer’s and Huntington’s diseases.7 In contrast, inhibition of SIRT2 may hold promise as a treatment for Parkinson’s disease and other disorders.7 SIRT2 inhibitors have been shown to rescue α-synuclein-induced toxicity and reduce cell death in cellular and Drosophila models of Parkinson’s disease.8 More recently, SIRT2 has been shown to be a key player in transcription regulation signaling initiated by bacterial infection, suggesting that blocking host SIRT2 may be a viable approach to fight bacterial infection.9 Intriguingly, both tumor-suppressing and tumor-prompting roles have been proposed for SIRT1−3.10 Even though the precise roles of sirtuins in cancer pathogenesis remain to be revealed, it is clear that they depend on the affected tissues and the stage of cancer progression. Further studies are needed to unravel the therapeutic applications of sirtuins in not only cancers but also other diseases, a challenging task that can be facilitated by the availability of potent and selective sirtuin inhibitors. In addition to nicotinamide, a wide range of chemical structures have been reported as sirtuin inhibitors,11 some of which are depicted in Figure 2. As examples of initial inhibitors, sirtinol (5),12 cambinol (6),13 and splitomicin (7)14 all feature a β-naphthol moiety and have been subjected to structural Received: May 19, 2014
A
dx.doi.org/10.1021/jm500777s | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Figure 1. Deacetylation catalyzed by sirtuins.
Figure 2. Examples of sirtuin inhibitors.
Figure 3. Selected SIRT2 inhibitors.
optimization. EX-527 (8),15 one of the most potent and selective SIRT1 inhibitors, has recently been reported to alleviate pathology of Huntington’s disease.16 Suramin (9)17 is slightly selective against SIRT1 over SIRT2 and SIRT3, but it suffers from high molecular weight and polarity. Because of SIRT2’s potential therapeutic applications, there has been a growing interest in the discovery of potent and selective SIRT2 inhibitors (Figure 3). AGK2 (10) with its modest activity and selectivity has been shown to mitigate αsynuclein toxicity and prevent against dopaminergic cell death in multiple models of Parkinson’s disease.8 Similarly, AK-1 (11) and its analogues have been evaluated for their therapeutic application in Huntington’s disease and have been subjected to extensive structural modifications.18 Ro31-8220 shows sub-
micromolar inhibition against SIRT2, but its antikinase activity excludes its use as a selective SIRT2 inhibitor.19 Tenovins, which are represented by tenovin-6 (12) and stem from a highthroughput screening for activators of p53, exhibit modest antiSIRT2 activity.20 In contrast, excellent anti-SIRT2 activity has been accomplished with macrocyclic peptides equipped with a mechanism-based warhead,21 but the peptidic nature limits their applications. More recently, substituted chroman-4-one 13 was shown to be a low-micromolar inhibitor of SIRT2 with high selectivity for SIRT2 over SIRT1 and SIRT3. 22 Compound 14, based on an anilinobenzamide chemotype, displays submicromolar inhibitory activity and an excellent selectivity profile.23 A new series of thieno[3,2-d]pyrimidine-6carboxamides, as exemplified by compound 15, were reported B
dx.doi.org/10.1021/jm500777s | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
Figure 4. Fragment-based approach toward the design of sirtuin inhibitors.
to possess remarkable low-nanomolar activity against SIRT1−3 but only modest isozyme selectivity.24 Aiming to identify SIRT2 inhibitors with improved potency and selectivity, we have used fragment-based ligand design to discover novel chemotypes, as discussed in this report.
■
RESULTS AND DISCUSSION Design and Biochemical Evaluations. Fragment-based drug design has emerged as a powerful strategy to explore chemical space and identify novel leads for further optimization.25 The fragment-based approach generally involves screening of a relatively small library of fragments and subsequent generation of new hit/lead compounds through structural manipulations, including linking, merging, and growth. The identification of novel scaffolds leads to further iterative optimization of activity, selectivity, and drug-like properties. Our design of fragments was inspired by close examination of a crystal structure of SIRT5 in complex with suramin, which mainly occupies the peptide substrate-binding site with one sulfonate group protruding into the nicotinamide-binding site.17a It has been shown that the naphthalene portion (circled in Figure 4) is close to the binding region proposed for nicotinamide. On the basis of the fact that suramin displays inhibitory activities against SIRT1, SIRT2, and SIRT5,17 it is reasonable to speculate that it binds to sirtuins in a similar fashion, which may allow for a fragment-based approach to the discovery of novel sirtuin inhibitors. To this end, we planned to assemble a focused library of fragments containing structural features from either suramin or nicotinamide. More importantly, because of the close binding proximity of these two known inhibitors, we envisioned that we could generate fragments by linking nicotinamide or its analogues to the structural elements extracted from suramin. Because of suramin’s high molecular weight and polarity, we decided to first focus on a significantly simplified naphthylamine moiety to which a sulfonate group is attached at various positions (16, Figure 4). To enhance structural diversity, the naphthylamine moiety was further acylated with various benzoyl chlorides to give substituted benzamides, a functionality that is also present in suramin. As a result, in this naphthylamide sulfonic acid series, a sulfonic acid is attached at position 2 (19−24, Figure 5), 4 (25−30), or 8 (31−36). Since suramin contains 4-methyl and 3-benzamido substituents, we decided to explore 4-methyl and 3-nitro substituents, which were installed individually or in combination. Furthermore, the 3-nitro group was reduced to
Figure 5. Fragments based on naphthylamide sulfonic acids.
give a 3-amino group, which if necessary could be further derivatized for structure−activity relationship (SAR) studies. In the other set of fragments, we decided to investigate compounds in which nicotinamide is connected to naphthalene at position 1 or 2 (17, Figure 4). More specifically, nicotinamide was attached at the para and meta positions (in relation to the primary amide) to give fragments 37−48 and 49−58, respectively (Figure 6). Simple linkers such as a nitrogen, oxygen, or sulfur atom were used. To investigate the importance of the nicotinamide nitrogen atom, analogous fragments based on benzamide were also included. We anticipated that screening of these two sets of fragments would provide necessary SAR information, allowing us to link selected fragments to generate inhibitors represented by the general structure 18 (Figure 4). These two series of fragments were tested against human recombinant SIRT1−3 using AMC-tagged deacetylation substrates. The naphthylamide sulfonic acids generally exhibited marginal inhibition when tested at 100 μM (Table S1 in the Supporting Information). However, there are two noticeable observations. First, the position of sulfonic acid does not affect the activity against SIRT1−3. Second, compounds 23, 29, and 35, all of which share 4-Me and 3-NO2 groups, appeared to show modest inhibition at 100 μM, suggesting that these substituents could be explored in constructing advanced inhibitors on the basis of the information gathered on the fragments. The naphthalene−benzamides and −nicotinamides were also tested against SIRT1−3 at 100 μM. Fragments 37−42, in C
dx.doi.org/10.1021/jm500777s | J. Med. Chem. XXXX, XXX, XXX−XXX
Journal of Medicinal Chemistry
Article
against SIRT1−3. Nevertheless, fragments 40 and 42 displayed more than 50% inhibition of SIRT2, suggesting that linking nicotinamide to naphthalene through an oxygen or sulfur atom might be productive. In comparison, fragments 49−53, in which benzamide or nicotinamide is attached to naphthalene via the meta position, displayed improved activity (Table 2). Examination of this Table 2. Inhibitory Activities of Meta-Substituted Fragments Based on Naphthalene−Benzamide or Naphthalene− Nicotinamidea
% inhibition at 100 μM or IC50 (μM) compound
X
Y
SIRT1
SIRT2
SIRT3
49 50 51 52 53 54 55 56 57 58
NH NH O O S NH NH O O S
CH N CH N N CH N CH N N
NA 33.6b 47 ± 5 13.5b 26.9b NA 40.9b 35 ± 7 4.89b 30.3b
65 ± 5 4.30b 84 ± 1 6.29b 12.1b 46 ± 1 9.70b 66 ± 2 7.57b 9.80b
NA 12.3b 49 ± 3 2.14b 27.9b NA 9.00b 35 ± 5 16.6b 30.1b
Figure 6. Fragments based on naphthalene−benzamide or naphthalene−nicotinamide.
which the benzamide or nicotinamide is attached at position 2 of naphthalene via the position para to the primary amide functionality, generally showed very weak inhibition (Table 1). Similarly, fragments 43−48 in which the linkage happens at position 1 also generally exhibited weak inhibitory activity
a
Inhibitory activities were determined as described in SIRT1−3 Biochemical Assays in the Experimental Section. NA indicates
