Discovery and optimization of a novel series of pyrazolyltetrahydropyran N-type calcium channel (Cav 2.2) blockers for the treatment of pain

Article abstract of DOI:10.1016/j.bmcl.2018.10.007

A novel series of pyrazolyltetrahydropyran N-type calcium channel blockers are described. Structural modifications of the series led to potent compounds in both a cell-based fluorescent calcium influx assay and a patch clamp electrophysiology assay. Representative compounds from the series were bioavailable and showed efficacy in the rat CFA and CCI models of inflammatory and neuropathic pain.

Full text of DOI:10.1016/j.bmcl.2018.10.007

Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Discovery and optimization of a novel series of pyrazolyltetrahydropyran N-
type calcium channel (Cav 2.2) blockers for the treatment of pain
Mark J. Wall⁎, Nalin L. Subasinghe, Michael P. Winters, Mary Lou Lubin, Michael F.A. Finley,
Ning Qin, Michael R. Brandt, Michael P. Neeper, Craig R. Schneider, Raymond W. Colburn,
Christopher M. Flores, Zhihua Sui
Janssen Research & Development, L.L.C., Welsh & McKean Roads, Spring House, PA 19477, USA
A R T I C L E I N F O
A B S T R A C T
Keywords:
A novel series of pyrazolyltetrahydropyran N-type calcium channel blockers are described. Structural mod-
ifications of the series led to potent compounds in both a cell-based fluorescent calcium influx assay and a patch
clamp electrophysiology assay. Representative compounds from the series were bioavailable and showed effi-
cacy in the rat CFA and CCI models of inflammatory and neuropathic pain.
Pain
N-type calcium channel
Cav 2.2
Pyrazoles
CFA model
CCI model
Chronic neuropathic pain is estimated to affect approximately 1% of
the United States population, and many patients are insufficiently
treated by currently available therapies.1,2 This unmet medical need has
led to an intensive effort to identify new treatment options, particularly
ones that act on critical parts of the pain neurocircuitry. The voltage-
gated N-type calcium channel (Cav2.2) is one such molecular target.3 It
is expressed on synaptic terminals of pain fibers in the spinal dorsal
horn where it regulates excitatory neurotransmission; and knock-out
mice lacking the N-type calcium channel exhibit reduced pain pheno-
types.4–6 ω-Conotoxin MVIIA, also known as ziconotide, is a 25-amino
acid cone snail toxin that potently and selectively antagonizes N-type
calcium channels7 and is currently approved for the treatment of severe
chronic pain as Prialt®. However, its narrow therapeutic window and
requirement for intrathecal delivery have limited its usefulness. An
orally active, small-molecule N-type calcium channel blocker, particu-
larly one exhibiting a frequency-dependent mechanism,8 might miti-
gate these problems, providing an improved therapeutic for certain
chronic pain indications. Consequently, there has been considerable
effort in the pharmaceutical industry to discover such a compound.
A high-throughput, cell-based, fluorescent calcium-influx screening
assay (FDSS)9,10 identified the diarylpyrazole 1 (Fig. 1) as a potent
inhibitor of the N-type calcium channel. In a whole-cell automated
patch clamp electrophysiology assay (QPatch),11 the compound showed
modest inhibition of 50% at 5 µM. The compound was also rapidly
metabolized in rat liver microsomes (RLM), with only 4% remaining
after 10-min incubation, although its human liver microsome (HLM)
stability was modestly better, with 58% remaining after 10 min.
Analysis of the metabolites of compound 1 identified the 3-meth-
oxypropyl group as a major liability for metabolic stability. Therefore, a
series of pyrazoles was prepared, wherein the pharmacologic properties
of various substitutions at the 3-position of the pyrazole scaffold were
evaluated. The synthesis of these compounds is outlined in Scheme 1.
Condensation of β-ketoester 2 and acylhydrazine 3 in dichloroethane
with phosphorus trichloride afforded 3-hydroxypyrazole 4 in a re-
gioselective manner. The hydroxypyrazole was converted to the triflate
using bis-N-phenyl triflimide, followed by palladium-catalyzed boryla-
tion with pinacol borane to give compound 5. A Suzuki-Miyaura reac-
tion of pyrazole boronate 5 with vinyl triflates followed by alkene re-
duction gave 6.
The synthesized compounds were evaluated for their ability to block
the N-type calcium channel (FDSS and QPatch) and for their metabolic
liability in both HLM and RLM. The results in Table 1 show that 4-
tetrahydropyran (7) can replace the 3-methoxypropyl chain of the
screening hit while maintaining its potency. However, there was still a
significant rat metabolic liability associated with this compound. It was
found that this could be mitigated by adding methyl groups to the 2-
and 6-positions of the tetrahydropyran ring. Addition of two methyls to
the ring, in either a geminal (8), cis (10) or trans (11) orientation
modestly improved RLM stability, whereas addition of four methyl
groups (9) produced a compound that was substantially more stable
⁎
Corresponding author.
E-mail address: mwall2@its.jnj.com (M.J. Wall).
https://doi.org/10.1016/j.bmcl.2018.10.007
Received 15 May 2018; Received in revised form 27 August 2018; Accepted 6 October 2018
0960-894X/©2018PublishedbyElsevierLtd.
Pleasecitethisarticleas:Wall,M.J.,Bioorganic&MedicinalChemistryLetters,https://doi.org/10.1016/j.bmcl.2018.10.007

M.J. Wall et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 1
SAR at the 3-position of pyrazole.
Fig. 1. High-throughput screening hit.
Entry R1
7
FDSS IC50
(µM)
QPatch IC50 (µM)
High/Low Freq
HLM/RLM %
remaining at 10 min
0.030
0.002
0.001
0.003
0.002
0.003
0.020
30/25a
75/8
8
0.055/0.150
0.016/0.024
0.012/0.091
61/30a
98/21
97/96
64/18
40/14
96/37
26/14
9
10
11
12
13
Scheme 1. Reagents and conditions: (a) PCl3, DCE, 60 °C; (b) PhNTf2, Hunig’s
base, DCE, 60 °C; (c) pinacol borane, Pd(dppf)Cl2; (d) R1-X, Pd(PPh3)4, Na2CO3,
Tol/EtOH, 80 °C; (e) 10% Pd/C, 30 psi H2, MeOH, RT.
26/14a
(i.e., 96–97% remaining after a 10-min incubation). Partial saturation
of the pyran ring (12) resulted in reduced RLM stability, and a bulkier
isopropyl group, in either a cis (14) or trans (13) orientration, also was
ineffective in stabilizing the tetrahydropyran ring.
NT
These compounds all displayed relatively high potency in the FDSS
assay, with the unsubstituted tetrahydropyran 7 being the least potent
(IC50 = 0.030 µM) and the tetramethyl-substituted analog 9 being the
most potent (IC50 = 0.001 µM). The compounds were further evaluated
14
a
0.007
55/18a
66/21
in
a high-resolution whole-cell automated patch clamp electro-
physiology QPatch assay. This assay also allowed for the examination of
the frequency dependence of inhibition, wherein high-frequency acti-
vation of the channel might better represent its functional state during
certain chronic pain conditions and low-frequency activation more
normal physiologic conditions. In this assay as well, addition of methyl
groups increased the potency of the series, with compound 9 having an
IC50 value of 0.016 µM at the higher frequency, and 0.024 µM at the
lower frequency. The trans-dimethyl compound 11 was also relatively
potent in blocking the channel in this assay, demonstrating appreciable
separation between the two activation states of the channel (i.e., ap-
proximately two-fold greater inhibition at high vs. low frequency), al-
though its RLM stability was insufficient for in vivo testing in rat pain
models.
Percent inhibition at 0.1 µM.
Having identified the tetramethylpyran 9 as an optimal substituent
for the 3-position of the pyrazole scaffold, the SAR of the 1- and 5-
positions of the pyrazole scaffold was investigated. These compounds
were prepared as shown in Scheme 2. The 4-ketopyran 16 was treated
with TosMIC to give the nitrile that was then hydrolyzed to the car-
boxylic acid, followed by conversion to the benzotriazole amide 17.
The benzotriazole was reacted with acylthiophenol under soft-enoliza-
tion conditions to give the β-ketothioester 18.12 The R2 group was then
introduced regioselectively via a condensation with R2-substituted hy-
drazines to give the 5-ketopyrazole 19.13 The ketopyrazole was con-
verted to the triflate 20 and the R3-group was introduced via a Suzuki-
Miyaura reaction.
Scheme 2. Reagents and conditions: (a) (i) TosMIC, (ii) KOH (iii) SOCl2, ben-
zotriazole; (b) Acylthiophenol, MgBr2, DIEA; DCM (c) R2-hydrazine hydro-
chloride, MeOH; (d) PhNTf2, DIEA, DCE (e) R3-B(OH)2, Pd(PPh3)4, Na2CO3,
Tol/EtOH.
The FDSS and QPatch data for the 1,5-substituted pyrazoles are
2

M.J. Wall et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
Table 2
Table 4
SAR at the 1- and 5-positions of the pyrazole.
Selectivity profile of compounds 9 and 22.
Compound
L-Type; Percent inhibition at 1 uM
hERG IC50 (uM)
9
0
20
26
22
12
two compounds exhibited long half-lives of 15.4 hr and 10.2 hr and oral
bioavailabilities of 36% and 49%, respectively. The volume of dis-
tribution of both compounds was relatively high, presumably due to the
lipophilic nature of the compounds, suggesting high tissue distribution.
The selectivity of compounds 9 and 22 over the L-type calcium
channel and the hERG potassium channel are shown in Table 4. Com-
pound 9 did not inhibit the L-type at 1 uM and compound 22 showed
modest inhibition of 12% at 1 uM. Both compounds had a relatively
weak hERG inhibition profile.
Entry R2
22
R3
FDSS IC50
(µM)
QPatch IC50 (µM)
High/low Freq
0.006
0.028/0.062
23
24
0.002
0.029/0.061
0.001
0.085/0.130
An ex vivo assay measuring the release of the neurotransmitter
calcitonin gene-related peptide (CGRP) from rat spinal cord slices was
used to determine the activity of compounds 9 and 22 in a physiolo-
gically relevant setting.9 In this assay, ω-conotoxin-GVIA (1 µM) was
used as a positive control. At a concentration of 1 µM, compound 9 and
compound 22 exhibited, respectively, 100% and 81% block of CGRP
release compared with the same concentration of ω-conotoxin-GVIA.
Compounds 9 and 22 were evaluated in the rat complete Freunds
adjuvant (CFA) model of inflammatory pain.14 This test predicts the
analgesic effect of numerous efficacious clinical agents, including the N-
type calcium channel blocker Prialt®.15 Compounds 9 and 22 showed
reversal of thermal hyperalgesia at 30 mg/kg (Fig. 2). Maximum re-
versal was observed 30 min post-dose with both compounds. After
180 min, latencies were similar to those in vehicle-treated animals. In
separate studies, tissue concentrations of the compounds were de-
termined. Compound 22 exhibited Cplasma = 1.1 µM, Cbrain = 2.2 µM
and Cspinal cord = 1.9 µM at 1 hr following a 30 mg/kg p.o. dose; and
compound 9 exhibited Cplasma = 0.7 µM, and Cbrain = 1.2 µM at 2 hr
following a 30 mg/kg p.o. dose.
25
26
0.005
0.006
0.044/0.200
13/9a
27
28
29
30
0.011
0.001
0.012
0.007
0.250/0.740
0.100/0.590
1/0a
22/0a
Compound 9 was also evaluated in the rat chronic constriction in-
jury (CCI) model of neuropathic pain16 (Fig. 3). At 1–3 h following a
dose of 30 mg/kg p.o. of compound 9, there was a 51% inhibition of
cold allodynia, as assessed by nocifensive responses to acetone appli-
cation to the affected hindpaw.
a
Percent inhibition at 0.1 µM.
shown in Table 2. Regarding the R2 group, it was observed that the
ortho-methoxy was important for maintaining potency in the QPatch
assay. The -NEt (28) was equally potent in the FDSS assay, but was 6-
fold and 25-fold less potent than -OMe (9) in the QPatch assay at high
and low frequency, respectively. For the R3 group, it was found that the
4-chlorophenyl could be replaced with the 2-chlorothiophene (22), 4-
ethoxyphenyl (23), 4-cyanophenyl (24) or 2-ethoxypyridinyl (25).
Compounds 22 and 25 showed, respectively, an approximately 2- and
4.5-fold difference in potency between high- and low-frequency acti-
vation states. Compound 25 was one of the few compounds wherein a
heterocycle could be incorporated to help lower the logP of the series.
Other attempts to incorporate polar functionality led to loss of potency.
Compounds 9 and 22 were selected for pharmacokinetic profiling in
rats, based on in vitro potency and metabolic stability (Table 3). The
In summary, we have identified a novel and potent series of pyr-
azolyltetrahydropyran N-type calcium channel blockers. Addition of
methyl groups to the tetrahydropyran ring stabilized the series to HLM
and RLM, leading to pharmacologically relevant exposures in plasma
and spinal cord. In particular, compound 22 exhibited favorable in vitro
and in vivo profiles. Further investigations will be aimed at identifying
2 0
**
C om pound 22
V e h ic le fo r 2 2
1 9
1 8
1 7
1 6
1 5
1 4
1 3
1 2
1 1
1 0
9
C o m p o u n d
V e h ic le fo r
9
9
**
**
**
**
*
Table 3
8
7
6
5
4
Pharmacokinetic parameters for compounds 9 and 22 in rat. AUClast and Cmax
calculated from PO administration.
Entry AUClast (h * ng/ Cl (ml/
t1/2 iv
(h)
Cmax (ng/
ml)
Vss (L/
kg)
F (%)
(-2 4 h )
0
30 60
1 2 0
1 8 0
3 0 0
ml)
min/kg)
T im e (m in ) a ꢀe r p .o . a d m in istra ꢁo n
9
1498
2660
37
15.4
10.2
127
369
33.3
10.2
36
49
22
15.9
Fig. 2. Rat CFA (radiant heat) time course. Compounds were administered at
30 mg/kg p.o. in 10% N-methyl pyrrolidone (NMP)/ 20% Solutol. Vehicle 10%
Compound 9 was administered at 10.0/2.0 mg/kg po/iv; compound 22 was
administered 6.0/1.2 mg/kg po/iv. Compounds were formulated in 20% hy-
droxypropyl-β-cyclodextran.
**
NMP/20% Solutol. One-way analysis of variance, Dunnett’s post test
–
*
p < 0.01, – p < 0.05.
3

M.J. Wall et al.
Bioorganic&MedicinalChemistryLettersxxx(xxxx)xxx–xxx
100
currents by whole-cell patch clamp. HEK293 cells stably expressing Cav2.2 subunits are
routinely grown as monolayer in Dulbecco’s Modification of Eagle’s Medium (DMEM; low
glucose Cat # 12320-032) supplemented with 10% FBS, 2 mM L-glutamine, 100 I.U./ml
penicillin, 100 lg/ml streptomycin, 400 lg/ml G418, and 200 lg/ml Zeocin (Split ratio:
1:5). Cells are maintained in 5% CO2 at 37 °C. JNJ compounds are prepared as 10 mM
stock in DMSO from neat, if available. Otherwise, the 5 or 10 mM DMSO stock solutions
provided inhouse are used. Calcium mobilization responses to KCl depolarization are
evaluated by measuring the intensity of Ca2+ fluorescent signal in the presence of BD
Calcium Assay Dye (BD Biosciences), utilizing a Functional Drug Screening System (FDSS)
by Hamamatsu. 24 h prior to assay, cells are seeded in clear-base poly-D-lysine coated
384-well plates (BD Biosciences, NJ, USA) at a density of 5000 cells per well in culture
medium and grown overnight in 5% CO2 at 37 °C. On assay day, growth media is removed
and cells are loaded with BD calcium assay dye (BD Bioscience) for 35 min at 37 C, under
5% CO2 and then for 25 min at room temp. Utilizing the FDSS, cells are challenged with
test compounds (at varying concentrations) and intracellular Ca2+ is measured for 5 min
prior to the addition of 50 mM KCl for an additional 3 min of measurement. IC50 values
for compounds ran in this assay are determined from six-point dose–response studies and
represent the concentration of compound required to inhibit 50% of the maximal re-
sponse. Maximal fluorescence intensity (FI) achieved upon addition of 50 mM KCl was
exported from the FDSS software and further analyzed using GraphPad Prism 3.02 (Graph
Pad Software Inc., CA, U.S.A.). Data is normalized to the maximum average counts of
quadruplicate wells for each data points in the presence of 50 mM KCl and to the
minimum average counts in the presence of buffer. Theoretical curves are generated using
nonlinear regression curve-fitting analysis of either sigmoidal dose response or sigmoidal
dose response (variable slope) and the IC50 values with the best-fit dose curve determined
by GraphPad Prism are reported. ω-Conotoxin MVIIA was run as a positive control in this
assay and had IC50 of 22 nM.
Compound 9
Vehicle
80
60
40
20
0
*
*
*
0
1
2
3
Hours after p.o. dose
Fig. 3. Effects of Compound 9 in CCI-induced cold allodynia – acetone test.
Compound dosed 30 mg/kg p.o. in 10% Solutol/20% HP-β-CD; vehicle was
10% Solutol/20% HP-β-CD. ∗p < 0.05 2-way ANOVA for repeated measures
with Bonferroni post-hoc test.
11.. This assay employs the whole-cell patch clamp technique to measure ion flux through
the N-type channel using an automated patch clamp device called the QPatch-HT, a 48
channel version of the QPatch16 from Sophion Biosciences. The N-type calcium channel
is voltage-dependent. Cells (HEK293) expressing the channel subunits are held in voltage-
clamp, providing maximal control over the state of the ion channel complexes within the
cell. The cell is maintained at a membrane potential (−80 mV) that keeps the channels in
a closed state except for brief depolarizing pulses (to +20 mV) to open the channel under
very controlled conditions, allowing current to flow into the cell. Barium is used as the
charge carrier because it provides enhanced current levels relative to calcium without
altering the function of the channel. In order to examine the frequency dependence of
inhibition, a train of depolarizing pulses (15 pulses at 5 Hz) is delivered to the cell once
every 30 s for 8 trains, and the resulting currents are measured during a control period (no
compound). This protocol is repeated with 4 trains for each addition of compound (ty-
pically 2 or 3 additions are performed). The current generated in the first (0.033 Hz) and
fifteenth (5 Hz) pulses of the 4th train in the presence of each drug concentration, re-
presenting low- and high-frequency stimulation, respectively, is normalized to the current
generated during the control period at the respective pulses. A final period of 2 trains in
the presence of 60 μM cadmium chloride is used to block all current and set the 0% level.
The average ‘% control low’ and ‘% control high’ values for several cells is determined. In
every experiment, a number of cells are treated in parallel with vehicle alone and used as
controls to monitor any rundown of current. All reported results for % inhibition by
compound are normalized against these control cells. ω-Conotoxin MVIIA was run as a
positive control in this assay and inhibited both low- and high-frequency currents with
estimated IC50 values of 48 and 49 nM, respectively. (Ref.9).
orally available N-type calcium channel blockers that may be useful in
treating severe chronic pain.
Acknowledgements
The authors would like to thank Ronghui Zhou for 2D NMR ex-
periments and Janssen Discovery Sciences ADME and Analytical teams
for their contributions to this work.
Appendix A. Supplementary data
Supplementary data (Full 1H and 2D NMR characterization of
compounds 9 and 22) to this article can be found online at https://doi.
org/10.1016/j.bmcl.2018.10.007.
References
12. Zhou G, Lim D, Coltart DM. Org Lett. 2008;10:3809–3812.
1. Jarvis MF. Drug Dev Res. 2006;68:287–288.
13. Dagli DJ, Gorski RA, Wemple J. J Org Chem. 1975;40:1741–1745.
2. Snutch TP. NeuroRx. 2005;2:662–670.
14.. Rats were injected with Complete Freund’s Adjuvant (CFA, 100 µL, 1:1 in sterile
saline) subcutaneously into the sub-plantar tissue of the left hind paw 24 h before the
study, after baseline radiant heat measurements were recorded. On the day of the study,
radiant heat latencies were assessed to confirm development of thermal hypersensitivity
(75% cut-off). Rats were dosed p.o. at 2.5 mL/kg with vehicle (10% NMP/20% solutol) or
30 mg/kg of compound. Radiant heat latencies were assessed at 30, 60, 120, 180, and
300 min after dosing.
3. Bell TJ, Thaler C, Castiglioni AJ, Helton TD, Lipscombe D. Neuron. 2004;41:127–138.
4. Kerr LM, Filloux F, Olivera BM, Jackson H, Wamsley JK. Eur J Pharmacol.
1988;146:181–183.
5. Saegusa H, Kurihara T, Zong S, et al. EMBO J. 2001;20:2349–2356.
6. Smith MT, Cabot PJ, Ross FB, Robertson AD, Lewis R. J Pain. 2002;96:119–127.
7. Olivera BM, Cruz LJ, de Santos V, et al. Biochemistry. 1987;26:2086–2090.
8. Winquist RJ, Pan JQ, Gribkoff VK. Biochem Pharmacol. 2005;70:489–499.
9. Finley MF, Lubin ML, Neeper MP, et al. Assay Drug Dev Technol. 2010;8:685–694.
10.. In the FDSS assay, functional activity of test compounds are determined by their
inhibitory effects on Ca2+ influx via N-type calcium channel after depolarizing plasma
membrane with 50 mM KCl. Calcium mobilization responses to KCl depolarization are
evaluated by measuring the intensity of Ca2+ fluorescent signal in the presence of BD
Calcium Assay Dye (BD Biosciences), utilizing a Functional Drug Screening System (FDSS)
by Hamamatsu. For this screen, a stable cell line (HEK parent) expressing Cav2.2 (N-type
calcium channel, Genbank accession number AAO53230) subunits is used. Cav2.2 sub-
units were expressed in pcDNA3.1 and pBudCE4.1 vectors under selection by 400 lg/ml of
G418 and 200 lg/ml of Zeocin. Cells were clonally isolated, expanded and screened by
Western blot analyses, and further tested for expression of characteristic N-type calcium
15. McGivern JG. Neuropsychiatr Dis Treat. 2007;3:69–85.
16.. In male Sprague–Dawley rats, 4 loose ligatures (4–0 chromic gut) were surgically
placed around the left sciatic nerve under inhalational anesthesia. Cold sensitivity was
assessed 2 wks after surgery by acetone evaporative cooling. In elevated observations
chambers containing wire mesh floors, 5 applications of acetone (100 μL) were applied to
the bottom of the hindpaw using a multidose syringe device. Individual responses were
quantified by measuring a brisk foot withdrawal in response to acetone application. Cold
hypersensitivity was assessed before (‘0’ time point) and hourly for 3 h after compound
administration. Individual scores were averaged over the five applications and percent
inhibition was calculated as a percentage of responses from the ‘0’ timepoint.
4