Sulfonamides as Selective NaV1.7 Inhibitors: Optimizing Potency

Article pubs.acs.org/jmc

Sulfonamides as Selective NaV1.7 Inhibitors: Optimizing Potency, Pharmacokinetics, and Metabolic Properties to Obtain Atropisomeric Quinolinone (AM-0466) that Affords Robust in Vivo Activity Russell F. Graceffa,*,† Alessandro A. Boezio,*,† Jessica Able,∇ Steven Altmann,∥,¶ Loren M. Berry,§ Christiane Boezio,† John R. Butler,† Margaret Chu-Moyer,† Melanie Cooke,# Erin F. DiMauro,†,+ Thomas A. Dineen,†,○ Elma Feric Bojic,∥,□ Robert S. Foti,§ Robert T. Fremeau, Jr.,∥,▲ Angel Guzman-Perez,† Hua Gao,‡ Hakan Gunaydin,‡,● Hongbing Huang,†,◇ Liyue Huang,§ Christopher Ilch,∥ Michael Jarosh,∥ Thomas Kornecook,∇ Charles R. Kreiman,† Daniel S. La,† Joseph Ligutti,∇ Benjamin C. Milgram,† Min-Hwa Jasmine Lin,§ Isaac E. Marx,†,◆ Hanh N. Nguyen,†,■ Emily A. Peterson,†,@ Gwen Rescourio,† John Roberts,§ Laurie Schenkel,† Roman Shimanovich,# Brian A. Sparling,† John Stellwagen,† Kristin Taborn,∥ Karina R. Vaida,† Jean Wang,∥ John Yeoman,† Violeta Yu,∥ Dawn Zhu,∥ Bryan D. Moyer,∇ and Matthew M. Weiss† †

Departments of Therapeutic Discovery, ‡Molecular Engineering, §Pharmacokinetics and Drug Metabolism, ∥Neuroscience, and Pharmaceutics, Amgen, Inc., 360 Binney Street, Cambridge, Massachusetts 02142, United States ∇ Department of Neuroscience, Amgen Inc., One Amgen Center Drive, Thousand Oaks, California 91320, United States #

S Supporting Information *

ABSTRACT: Because of its strong genetic validation, NaV1.7 has attracted significant interest as a target for the treatment of pain. We have previously reported on a number of structurally distinct bicyclic heteroarylsulfonamides as NaV1.7 inhibitors that demonstrate high levels of selectivity over other NaV isoforms. Herein, we report the discovery and optimization of a series of atropisomeric quinolinone sulfonamide inhibitors [Bicyclic sulfonamide compounds as sodium channel inhibitors and their preparation. WO 2014201206, 2014] of NaV1.7, which demonstrate nanomolar inhibition of NaV1.7 and exhibit high levels of selectivity over other sodium channel isoforms. After optimization of metabolic and pharmacokinetic properties, including PXR activation, CYP2C9 inhibition, and CYP3A4 TDI, several compounds were advanced into in vivo target engagement and efficacy models. When tested in mice, compound 39 (AM-0466) demonstrated robust pharmacodynamic activity in a NaV1.7-dependent model of histamine-induced pruritus (itch) and additionally in a capsaicininduced nociception model of pain without any confounding effect in open-field activity.

■

INTRODUCTION

NaV1.7 is a voltage-gated sodium channel that is primarily expressed in the peripheral nervous system (PNS), specifically in sympathetic and sensory neurons. It is involved in controlling action potentials and neuronal excitability and appears to play a prominent role in a wide range of inherited neuropathic pain syndromes.5 Mutations in SCN9A, the gene encoding NaV1.7, have been linked to hereditary pain disorders such as erythromelalgia (IEM),6 paroxysmal extreme pain disorder (PEPD),7 and small fiber neuralgia (SFN).8 These are gain-of-function mutations that modulate NaV1.7 channel

Approximately 2% of the world’s population is affected by neuropathic pain.2 In the United States alone, it was estimated that the costs attributed to pain were approximately $600 billion annually.3 Currently, treatments for pain include opioids, antidepressants, anticonvulsants, and analgesics, which are often used in combination. These therapeutic approaches have several shortcomings such as limited efficacy, unwanted side effects including sedation and coordination impairment, and possible addiction. Additionally, it is estimated that only 50% of the patients can obtain satisfactory relief with current therapies.4 For these reasons, it is not only desirable but necessary to develop new and effective treatments for pain. © 2017 American Chemical Society

Received: December 20, 2016 Published: March 21, 2017 5990

DOI: 10.1021/acs.jmedchem.6b01850 J. Med. Chem. 2017, 60, 5990−6017

Journal of Medicinal Chemistry

Article

the latter being an issue that plagued this series of compounds. Unfortunately, one metabolic liability that remained was inhibition of CYP2C9. Although we were able to identify compounds that were slightly less potent on CYP2C9, they routinely suffered from other liabilities, as epitomized by 2. It was generally found within these series, mitigating CYP2C9 inhibition required the incorporation of polarity. However, this then led to higher in vivo clearance in both rodent and dog due to transporter-mediated clearance.16,17 These opposing trends highlight some of the challenges that were associated with the identification of NaV1.7 inhibitors that could ultimately be pharmacologically relevant agents in humans. Concurrent with these efforts, work was ongoing to advance a separate series of benzoxazine sulfonamide inhibitors represented by 3.1,18 This series was appealing as it merged moderately potent inhibition of NaV1.7 with an acceptable rat in vivo pharmacokinetic profile. Unfortunately, akin to other early sulfonamide leads,16 analogues in this series demonstrated significant activation of PXR, were potent inhibitors of CYP2C9, and lacked selectivity over NaV1.5. In the course of investigating different bicyclic sulfonamide cores,16−18 it was found that compounds containing fully aromatic B-rings (e.g., 1) exhibited higher levels of selectivity over NaV1.5 than related compounds which contained sp3hybridized carbons within the B-ring (e.g., 3). This observation led us to postulate that high levels of selectivity could be achieved when the molecule adopted a conformation wherein the central B-ring and C-ring were orthogonal in nature with a narrow well-defined global minimum conformation. When comparing selective lead 1 with benzoxazine 3 in Figure 1, the nonselective benzoxazine showed a wider energy minimum, allowing multiple lower energy binding conformations that could disrupt selectivity over NaV1.5. It was envisioned that this hypothesis could be tested by the introduction of a carbonyl within the benzoxazine scaffold (e.g., 3), a modification that would favor an orthogonal relationship between the B- and C-rings with a sharper minimum (see Figures 1 and 2). Consequently, benzoxazinone 4 was prepared and we were pleased to find it maintained potency on NaV1.7 and demonstrated >143-fold selectivity over NaV1.5. However, chemical instability due to amide bond hydrolysis under basic conditions (1N NaOH @ RT) precluded advancement of this compound, athough it served as a promising starting point for further investigation. With the aim of addressing the chemical instability of benzoxazinones such as 4, quinolinone 5 was designed and prepared. It was envisioned that this compound could enforce the requisite conformation to engender high levels of selectivity over NaV1.5 and would not succumb to base-mediated hydrolysis of the amide as obsereved with 4. One feature of the quinolinone core that was intriguing was the possibility that hindered rotation around the junction of the B-ring and the Cring would manifest in an element of atropisomerism.20 The predicted barrier of rotation around this bond in compound 5 was calculated at the B3LYP/6-31G* level of theory and found to be 29.3 and 30.0 kcal/mol in the gas and aqueous phase, respectively. These values suggested that this compound would indeed exist as a mixture of stable atropisomers at room temperature. Subjection of 5 to chiral separation led to the isolation of configurationally stable atropisomers 6 and 7 (Table 2). Upon evaluation, it was found that 6 and 7 exhibited a large disparity in their potency on NaV1.7, with the former being approximately 100 times more potent than the latter.

activation or inactivation properties that result in hyperexcitability of peripheral nociceptors.9 In contrast, individuals with congenital insensitivity to pain (CIP)10 exhibit loss-offunction mutations which manifest in the inability to sense painful stimuli. In light of the human genetics implicating NaV1.7 in the pain processing pathway, this target has received considerable attention across the pharmaceutical industry.11 One of the long-standing challenges associated with NaV1.7 has been the identification of inhibitors demonstrating selectivity over the other sodium channel subtypes. Acceptable selectivity over the cardiac NaV1.5 channel subtype is particularly important, as inhibition of NaV1.5 is associated with a number of arrhythmic conditions, such as long QT syndrome,12 Brugada syndrome,13 conduction dysfunction,14 and atrial fibrillation.15

■

RESULTS AND DISCUSSION We have previously reported on a number of structurally distinct bicyclic heteroarylsulfonamide-containing NaV1.7 inhibitors, several of which are illustrated in Table 1.16−18 The Table 1. Representative Example of Potent NaV1.7 Inhibitors

NaV1.7 IC50 (μM)a NaV1.5 IC50 (μM)b PXR activation (POC)c CYP2C9 IC50 (μM)d rat iv CLe [CLu] (L/h/kg) dog iv CLf [CLu] (L/h/kg) cLogD PSA

1

2

3

0.03 >10 1 0.43 0.23 [26] 0.023 [3.6] 2.9 94

0.12 >10 54 1.2 2.0 [611] 0.11 [31] 1.1 116

0.12 2 116 0.25 0.24 [40] 3.6 81

a hNaV1.7 data were collected using a PatchXpress automated electrophysiology platform using a protocol where cells were held at a voltage yielding 20−50% channel inactivation. bhNaV1.5 data were collected on the same platform using a protocol where cells were held at −50 mV. cExpressed as a percentage of the activation response seen with rifampicin when both compounds are used at a concentration of 10 μM. dCYP inhibition, IC50 after coincubation with diclofenac. e0.5 mg/kg in DMSO to male rats. CLu = unbound clearance = total clearance/f u. f0.25 mg/kg in DMSO to male dogs. CLu = unbound clearance = total clearance/f u.

series represented by compounds 1 and 2 demonstrated potent inhibition of NaV1.7 and high levels of selectivity over the other NaV isoforms, including NaV1.5. One of the primary challenges associated with these series was the identification of compounds which not only demonstrated potent inhibition of NaV1.7 but were also isoform-selective and, equally importantly, demonstrated good pharmacokinetics. Eventually, compounds which met these criteria were identified, with compound 1 being a representative example. This compound was potent, selective, exhibited good pharmacokinetics in both rat and dog, and did not activate pregnane X receptor (PXR), 5991



Article

Figure 1. Torsional scan analysis of selected compounds.19

Figure 2. Evolution of NaV1.7 heteroarylsulfonamides.

findings, both atropisomers were inactive on NaV1.5 (NaV1.5 IC50 > 30 μM). Efforts next focused on replacing the 2-aminothiazole A-ring, a structural alert that is often associated with bioactivation.21 A number of different heterocyclic sulfonamide A-rings were investigated within the context of the quinolinone core. In analogy to findings with our earlier series of heteroaryl sulfonamides,16−18 the SAR associated with the A-ring was very steep and replacement of the 2-aminothiazole with a number of other heterocycles (pyridazin-3-amine, pyrimidin-2amine, pyrimidin-4-amine, and isoxazol-3-amine) led to a significant loss in potency (data not shown). Unable to find a suitable replacement for the 2-aminothiazole A-ring, a strategy similar to the one employed in the identification of 1 was employed within this series of quinolinone compounds. Specifically, replacement of the 2-aminothiazole with a 3aminoisoxazole in conjunction with replacement of the CF3 with a substituted aryl D-ring led to quinolinone sulfonamide 8, a potent inhibitor of NaV1.7. This observation was in line with our previous findings wherein the incorporation of an aryl Dring enabled the preparation of potent inhibitors with a broad range of heterocyclic A-rings. Akin to 5, this compound existed as a mixture of stable and separable atropisomers. Upon chiral separation, it was found that enantiomer 9 was significantly more potent than 10, with both compounds retaining a high

Table 2. Example of NaV1.7 Selective Atropisomers

NaV1.7 IC50 (μM)a NaV1.5 IC50 (μM)b

5

6

7

0.50 >30

0.24 >30

22 >30

a

hNaV1.7 data were collected using a PatchXpress automated electrophysiology platform using a protocol where cells were held at a voltage yielding 20−50% channel inactivation. bhNaV1.5 data were collected on the same platform using a protocol where cells were held at −50 mV.

This was a trend that persisted throughout this series of compounds, where one atropisomer was significantly more potent than the other. Additionally, in line with our previous 5992



Article

level of selectivity over NaV1.5 (IC50 > 30 μM). As noted previously, the identification of potent heteroarylsulfonamides which demonstrated a reduced level of CYP2C9 inhibition and favorable pharmacokinetics had previously been extremely difficult. Not only was 9 very selective over CYP2C9, but it also demonstrated low total and unbound clearance in both rat and dog and was well absorbed in both species (Tables 3 and 4). Table 3. Representative NaV1.7 Atropisomers Containing 2Aminoisoxazole A-Ring

Figure 3. Crystal structure of representative (P) atropisomer 11.

NaV1.7 IC50 (μM) CYP2C9 IC50b

a

8

9

10

0.027 0.56

0.059 12

18 0.76

a hNaV1.7 data were collected using a PatchXpress automated electrophysiology platform using a protocol where cells were held at a voltage yielding 20−50% channel inactivation. bCYP inhibition, IC50 after coincubation with diclofenac.

Table 4. Pharmacokinetic Profiles and Plasma Protein Binding of Atropisomer 9 iva

pob

species

CL [CLu] (L/h/kg)

Vdss (L)

t1/2 (h)

AUC (μM· h)

tmax (h)

%F

plasma protein binding ( f u)

rat dog

0.19 [55] 0.0045 [2.0]

1.8 0.14

2.6 4.0

65 600

2.7 4.0

66 62

0.0034 0.0022

Figure 4. (M)-Atropisomer 10 docked into the active site of CYP2C9.

that interconversion would not be problematic moving forward. A detailed study evaluating the racemization kinetics found that 9 was stable in a number of different media. Under neutral (e.g., toluene, water) and basic conditions (e.g., NaOH), racemization occurred only when heated above 100 °C, while no racemization was observed under acidic conditions, (e.g., 10% HCl at 37 °C for 24 h and 10% HCl at 100 °C for 1 h). The observation that racemization only occurred at high temperatures was subsequently exploited as it enabled us to recycle the “undesired” atropisomer (vide infra). Following this work, it was determined (using the Arrhenius equation) that the experimental barrier of rotation for 11 (31.6 kcal/mol, t1/2 = 32 years at 37 °C) correlated well with the calculated value (B3LYP/6-31G* level of theory) of 29.3 kcal/mol. To evaluate the potential for interconversion to occur in vivo, a related analogue (compound 12, Table 10) was advanced into a number of preclinical species. It was found that when dosed orally and intravenously, no interconversion was observed in vivo in monkey (iv, 1 h), dog (po, 24 h), rat (iv, 1 h), or mouse (po, 12 h). A series of minor modifications to the B-ring, primarily focusing on small increases in polarity and increased saturation, were explored. These changes were investigated to discern the impact that increased polarity within the core would have not only on potency and pharmacokinetics but also on the rate of racemization. Table 5 details the calculated rotational barriers for a subset of the analogues that were prepared in this endeavor. Adding additional polarity in the form of

a

0.5 mg/kg as a solution in DMSO (rat), 0.25 mg/kg in DMSO (dog); CLu = unbound clearance = total clearance/f u. bRat, 10 mg/kg; dog, 2 mg/kg; dosed as a solution in 1% Tween 80/2.0% HPMC, 97% water (pH = 10 with KOH).

In an effort to determine the absolute stereochemistry of the more potent atropisomer, compounds 9 and 10 were evaluated by vibrational circular dichroism (VCD) and found to be of the (P)- and (M)-configurations, respectively. This finding was subsequently supported with a single crystal X-ray of compound 11, which confirmed the more potent atropisomer as being of the (P) configuration (Figure 3). In an attempt to understand the difference in CYP2C9 inhibition between 9 and 10, the global minimum conformation of 10 was determined and docked into the active site of CYP2C9 (PDB 19RO) using the “Induced Fit” algorithm in MOE. As illustrated in Figure 4, CYP2C9 is able to accommodate the methoxy substituent on the C-ring of (M)atropisomer 10, whereas this substituent on (P)-atropisomer 9 would generate an unfavorable steric interaction with the protein. This finding provides a plausible explanation for the selectivity 9 demonstrates over CYP2C9. Subsequent to demonstrating the ability to attenuate CYP2C9 inhibition by incorporating an element of atropisomerism into this series of inhibitors, efforts focused on reinforcing the integrity of this stereocenter and confirming 5993



Article

Table 5. B-Ring Core Modifications SAR

a hNaV1.7 data were collected using a PatchXpress automated electrophysiology platform using a protocol where cells were held at a voltage yielding 20−50% channel inactivation. bCYP inhibition, IC50 after coincubation with diclofenac. cActivation barriers were computed in the gas phase for the neutral species at the B3LYP/6-31G* level of theory. dDetermined from Arrhenius equation. eExperimental activation energy.

Figure 5. Graphical illustration of key data for (P)-quinolinone sulfonamides.

naphthyridinone 13 and quinazolinone 14 resulted in a considerable loss in potency and hence were not explored further. Interestingly, the isomeric quinoxalinone 15 demon-

strated single-digit nanamolar inhibition of NaV1.7, however, this series of compounds proved to have chemical stability issues with regard to racemization upon sitting in methanol at 5994



Article

Table 6. NaV1.7 Potency SAR within the Quinolone Sulfonamide Scaffold

a

hNaV1.7 data were collected using a PatchXpress automated electrophysiology platform using a protocol where cells were held at a voltage yielding 20−50% channel inactivation. bExpressed as a percentage of the activation response seen with rifampicin when both compounds are used at a concentration of 10 μM. cCYP inhibition, IC50 after coincubation with diclofenac.

At the outset, several guidelines were established with respect to what criteria would be acceptable in a lead molecule for further progression. These included a NaV1.7 IC50 of

Sulfonamides as Selective NaV1.7 Inhibitors: Optimizing Potency

Recommend Documents