Anti-allodynic effect of 2-(aminomethyl)adamantane-1-carboxylic acid in a rat model of neuropathic pain: A mechanism dependent on CaV2.2 channel inhibition

Article abstract of DOI:10.1016/j.bmc.2014.02.006

Neuropathic pain is a serious physical disabling condition resulting from lesion or dysfunction of the peripheral sensory nervous system. Despite the fact that the mechanisms underlying neuropathic pain are poorly understood, the involvement of voltage-gated calcium (Ca_V) channels in its pathophysiology has justified the use of drugs that bind the Ca_V channel α₂δ auxiliary subunit, such as gabapentin (GBP), to attain analgesic and anti-allodynic effects in models involving neuronal sensitization and nerve injury. GBP binding to α₂δ inhibits nerve injury-induced trafficking of the α₁ pore forming subunits of Ca_V channels, particularly of the N-type, from the cytoplasm to the plasma membrane of pre-synaptic terminals in dorsal root ganglion neurons and dorsal horn spinal neurons. In the search for alternative forms of treatment, in this study we describe the synthesis and pharmacological profile of a GABA derivative, 2-aminoadamantane-1-carboxylic acid (GZ4), which displays a close structure-activity relationship with GBP. Behavioral assessment using von Frey filament stimuli showed that GZ4 treatment reverted mechanical allodynia/hyperalgesia in an animal model of spinal nerve ligation-induced neuropathic pain. In addition, using the patch clamp technique we show that GZ4 treatment significantly decreased whole-cell currents through N-type Ca _V channels heterologously expressed in HEK-293 cells. Interestingly, the behavioral and electrophysiological time course of GZ4 actions reflects that its mechanism of action is similar but not identical to that of GBP. While GBP actions require at least 24 h and imply uptake of the drug, which suggests that the drug acts mainly intracellularly affecting channels trafficking to the plasma membrane, the faster time course (1-3 h) of GZ4 effects suggests also a direct inhibition of Ca²⁺ currents acting on cell surface channels.

Full text of DOI:10.1016/j.bmc.2014.02.006

Bioorganic & Medicinal Chemistry 22 (2014) 1797–1803
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Anti-allodynic effect of 2-(aminomethyl)adamantane-1-carboxylic
acid in a rat model of neuropathic pain: A mechanism dependent
on CaV2.2 channel inhibition
b
c
c
Grigoris Zoidis a, , Alejandro Sandoval , Jorge Baruch Pineda-Farias , Vinicio Granados-Soto ,
⇑
Ricardo Felix d,
⇑
a Faculty of Pharmacy, Department of Pharmaceutical Chemistry, University of Athens, Panepistimioupoli-Zografou, GR-15771 Athens, Greece
b School of Medicine FES Iztacala, National Autonomous University of Mexico (UNAM), Tlalnepantla, Mexico
c Departmento de Farmacobiología, Centro de Investigación y de Estudios Avanzados (Cinvestav), Sede Sur., México, D.F., Mexico
d Department of Cell Biology, Cinvestav, Mexico City, Mexico
a r t i c l e i n f o
a b s t r a c t
Article history:
Neuropathic pain is a serious physical disabling condition resulting from lesion or dysfunction of the
peripheral sensory nervous system. Despite the fact that the mechanisms underlying neuropathic pain
are poorly understood, the involvement of voltage-gated calcium (CaV) channels in its pathophysiology
Received 1 November 2013
Revised 29 January 2014
Accepted 7 February 2014
Available online 15 February 2014
has justified the use of drugs that bind the CaV channel
(GBP), to attain analgesic and anti-allodynic effects in models involving neuronal sensitization and nerve
injury. GBP binding to 2d inhibits nerve injury-induced trafficking of the 1 pore forming subunits of CaV
a2d auxiliary subunit, such as gabapentin
a
a
Keywords:
channels, particularly of the N-type, from the cytoplasm to the plasma membrane of pre-synaptic termi-
nals in dorsal root ganglion neurons and dorsal horn spinal neurons. In the search for alternative forms of
treatment, in this study we describe the synthesis and pharmacological profile of a GABA derivative,
2-aminoadamantane-1-carboxylic acid (GZ4), which displays a close structure–activity relationship with
GBP. Behavioral assessment using von Frey filament stimuli showed that GZ4 treatment reverted
mechanical allodynia/hyperalgesia in an animal model of spinal nerve ligation-induced neuropathic pain.
In addition, using the patch clamp technique we show that GZ4 treatment significantly decreased whole-
cell currents through N-type CaV channels heterologously expressed in HEK-293 cells. Interestingly, the
behavioral and electrophysiological time course of GZ4 actions reflects that its mechanism of action is
similar but not identical to that of GBP. While GBP actions require at least 24 h and imply uptake of
the drug, which suggests that the drug acts mainly intracellularly affecting channels trafficking to the
plasma membrane, the faster time course (1–3 h) of GZ4 effects suggests also a direct inhibition of
Ca2+ currents acting on cell surface channels.
Allodynia
Calcium channels
Neuropathic pain
Gabapentin
Ó 2014 Elsevier Ltd. All rights reserved.
1. Introduction
usually accompanied by a painful sensation that is only partially
relieved by analgesics.
Neuropathic pain is an important clinical problem that drasti-
cally impairs the quality of life of patients. This condition is man-
ifested as tactile allodynia (a painful sensation to a non-noxious
stimulus) and or hyperalgesia (an enhanced sensation to a painful
stimulus). Neuropathic pain may be caused by nerve injury and is
It is well acknowledged that certain groups of sensory neurons
are the main mediators of pain sensation. Sensory neurons trans-
mit information from the periphery to the spinal cord which then
transfers this information to the brainstem and forebrain. Though
the molecular basis underlying neuropathic pain are not well
understood, nerve damage may cause hyperexcitability by increas-
ing the firing frequency and/or generation of spontaneous action
potentials from dorsal root ganglion (DRG) neurons.1
⇑
Corresponding authors. Addresses: Faculty of Pharmacy, Department of Phar-
maceutical Chemistry, University of Athens, Panepistimioupoli-Zografou, GR-15771
Athens, Greece. Nel.:+30 210 7274809 (G.Z.), Departamento de Biología Celular,
Cinvestav-IPN, Avenida IPN #2508, Colonia Zacatenco, México, D.F. CP 07300,
Mexico. Tel.: +52 55 57 47 39 88 (R.F.).
High-threshold N-type (CaV2.2) voltage-gated Ca2+ channels,
which transduce electrical activity into other cellular functions,
regulate Ca2+ homeostasis and are thought to play a major role in
processing pain information.2 The function and distribution of
E-mail addresses: zoidis@pharm.uoa.gr (G. Zoidis), rfelix@cell.cinvestav.mx
(R. Felix).
http://dx.doi.org/10.1016/j.bmc.2014.02.006
0968-0896/Ó 2014 Elsevier Ltd. All rights reserved.

1798
G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
these channels may vary among different types of neurons, how-
ever they are predominantly expressed in nerve terminals, where
they control neurotransmitter release. Indeed, genetic and phar-
macological studies have revealed that N-type channels are impor-
tant for pain sensation in disease models.2 These data suggest that
N-type channel inhibitors or modulators could be developed into
useful drugs to treat neuropathic pain.3
In a recent report, we described the synthesis and pharmacolog-
ical profile of a GABA adamantane derivative (AdGABA), which dis-
played a close structure-activity relationship with gabapentin
(GBP).4 Since GBP has been shown to reduce the functional expres-
sion of neuronal N-type (CaV2.2) channels,5,6 we examined the con-
sequences of the AdGABA treatment on membrane neuronal
recombinant CaV2.2 channels. Our results showed an alteration
in CaV2.2 channel functional expression after chronic treatment
with AdGABA presumably caused by a direct interaction of the
solvent. Chemical shifts are reported in d (ppm) with the tetra-
methylsilane or solvent (DMSO-d6) as internal standard. Splitting
patterns are designated as s, singlet; d, doublet; dd, doublet of dou-
blets; t, triplet; td, triplet of doublets; q, quartet; m, multiplet; br,
broad; v br, very broad; sym, symmetrical. Coupling constants (J)
are expressed in units of hertz (Hz). The spectra were recorded at
293 K (20 °C) unless otherwise specified. Carbon multiplicities
were established by DEPT experiments. The 2D NMR experiments
(HMQC and COSY) were performed for the elucidation of the struc-
tures of the newly synthesized compounds.
Analytical thin-layer chromatography (TLC) was conducted on
precoated Merck silica gel 60 F254 plates (layer thickness 0.2 mm)
with the spots visualized by iodine vapors and/or UV light. Column
chromatography purification was carried out on silica gel 60 (70–
230 mesh). Elemental analyses (C, H, N) were performed by the
Service Central de Microanalyse at CNRS (France), and were within
0.4% of the theoretical values. Elemental analysis results for the
tested compound correspond to >95% purity. The commercial re-
agents were purchased from Alfa Aesar, Sigma–Aldrich, and Merck.
Organic solvents used were in the highest purity, and when neces-
sary, were dried by standard methods. Solvent abbreviations: DME,
dimethoxyethane; Et2O, ethyl ether; MeOH, methanol; EtOH, etha-
nol; AcOEt, ethyl acetate; DMSO, dimethylsulfoxide.
drug on the
ABA may not bind to the same site as GBP in
To obtain further insight into how structural features may affect
a
2d auxiliary subunit of the channels. However, AdG-
a
2d.7
the anti-nociceptive activity of compounds based on c-aminobu-
tyric acid that modulate N-type Ca2+ channels, here we explored
the biological effect of the 2-(aminomethyl)adamantane-1-carbox-
ylic acid (GZ4). In this compound part of the c-aminobutyric acid
pharmacophore was introduced to the adamantane carbon skele-
ton (Fig. 1A). As we reported previously, GBP and AdGABA show
similarities in the orientation in both the carboxymethylene and
aminomethylene moiety. However, in the case of GZ4 the carbox-
ylic acid group possessed a different orientation (Fig. 1B). As we
shall discuss later, these results might explain some differences
in the mode of action of these compounds. Last, but not least, data
from behavioral testing showed that intrathecal application of GZ4
significantly reversed nerve ligation-induced tactile allodynia in a
rat model of neuropathic pain.
2.1.1.1.
5.
Methyl(2-oxotricyclo[3.3.1.13,7]decane-1-carboxylate
A mixture of the acid 4 (1.08 g, 5.6 mmol) and thionyl chlo-
ride (4 mL) was heated at 65 °C for 15 min. The excess thionyl chlo-
ride was removed under vacuum, and the resulting chloride was
esterified in an methanolic solution (20 mL) to give, after 1 h at
room temperature (RT) and 0.5 h at 70 °C, 1.14 g (quantitative
yield) of methylester 5 as a white solid; mp = 86–88 °C; IR (mull)
v 1731 cmÀ1 (CO), 1711 cmÀ1 (CO); 1H NMR (400 MHz, CDCl3) d
1.57–1.96 (complex m, 7H, 4e, 6, 8a, 9e, 10-H), 2.00–2.06 (m, 4H,
4a, 5, 7, 8e-H), 2.12 (d, 1H, J = 12.0, 9a-H), 3.69 (s, 3H, CO2CH3)
ppm; 13C NMR (100 MHz, CDCl3) d 26.7 (5-C), 26.8 (7-C), 30.8 (3-
C), 32.1 (4-C), 34.0 (9-C), 35.7 (8, 10-C), 39.4 (6-C), 42.7 (1-C),
61.1 (COOCH3), 174.4 (C@O), 217.4 (2-C) ppm. Anal. Calcd for
2. Experimental
2.1. Chemistry
C12H16O3: C, 69.21; H, 7.74. Found: C, 69.34; H, 7.68.
2.1.1. General
2.1.1.2. Methyl 2-cyanotricyclo[3.3.1.13,7]decane-1-carboxylate
6. Solid t-BuOK (0.60 g, 4.3 mmol) was added portion wise to
Melting points were determined using a Büchi capillary appara-
tus and are uncorrected. IR spectra were recorded on a Perkin–El-
mer 833 photometer. The 1H and 13C NMR spectra were obtained
on either a Bruker MSL 400 (400 MHz 1H; 100 MHz 13C) or Bruker
600 (600 MHz 1H) spectrometer, using CDCl3 or DMSO-d6 as
a stirred (argon atmosphere) solution of ketoester 5 (0.48 g,
2.31 mmol) and TosMIC (0.55 g, 2.8 mol) in a mixture of 8 mL of
DME and 0.3 mL of absolute EtOH maintained at 5–10 °C. The cool-
ing was removed and stirring continued for 30 min and the mix-
ture was then heated (48 °C) for 60 min. The suspension obtained
was cooled to RT with stirring. The precipitate (TosK) was filtered
and washed with DME. The combined DME solutions were concen-
trated and purified by flashing the concentrate over silica gel using
as eluents n-pentane–Et2O (3:1) to afford cyanoester 6 (410 mg,
81%); mp 55–57 °C (n-hexane); IR (mull) v 2237 cmÀ1 (CN),
1728 cmÀ1 (CO); 1H NMR (400 MHz, CDCl3), d (ppm) 1.69–2.16
(m, 12H, 4, 5, 6, 7, 8, 9, 10-H), 2.28 (d, 1H, J = 2.7 Hz, 3-H), 3.19
(d, 1H, J = 2.7 Hz, 2-H), 3.72 (s, 3H, CH3) ppm; 13C NMR
(100 MHz, CDCl3) d 26.8 (5-C), 26.9 (7-C), 30.8 (3-C), 32.3 (4-C),
34.2 (9-C), 35.6 (8, 10-C), 38.1 (2-C), 39.4 (6-C), 42.8 (1-C), 120.6
(CN), 174.6 (C@O). Anal. Calcd for C12H15NO2: C, 70.22; H, 7.37.
Found: C, 70.39; H, 7.62.
2.1.1.3.
7.
2-Cyanotricyclo[3.3.1.13,7]decane-1-carboxylic
A solution of the cyanoester 6 (9.37 g, 42.8 mmol) in etha-
acid
nol (50 mL) was added to a solution of NaOH (3.03 g, 75 mmol) in
water (20 mL), the flask was stoppered and left standing at RT for
48 h. After evaporation of the solvent the precipitated salt was dis-
solved in warm water. The solution was chilled in an ice bath and
acidified with concentrated HCl. The precipitated solid was filtered
Figure 1. Gabapentin and gabapentinoid drugs. (A) Structural comparison of the
antiepileptic/antinociceptive gabapentin, AdGABA and compound
8 (GZ4). (B)
Superposition of gabapentin (blue), AdGABA (pink) and compound GZ4 (green).

G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
1799
off, washed with water and dried: yield 8.16 g (93% from the
cyanoester 6); mp 160–162 °C (Et2O–n-pentane); IR (mull) v
2234 cmÀ1 (CN), 1707 cmÀ1 (C@O); 1H NMR (400 MHz, CDCl3), d
1.67–1.72 (m, 4H, 4e, 6, 8a-H), 1.83–1.94 (m, 2H, 8a, 10a-H),
2.00–2.13 (m, 5H, 4a, 5, 7, 8e, 10e, 9e-H), 2.14–2.22 (br d, 1H, 9a-
H), 2.32 (d, 1H, J = 2.0 Hz, 3-H), 3.19 (s, 1H, 2-H), 9.88 (br s, COOH)
ppm; 13C NMR (CDCl3, 100 MHz), d 26.6 (7-C), 26.7 (5-C), 30.8 (3-
C), 32.0 (4-C), 33.8 (9-C), 35.6 (6, 8-C), 38.1 (2-C), 39.3 (10-C), 42.8
(1-C), 120.5 (C„N), 180.8 (C@O) ppm. Anal. Calcd for C12H15NO2: C,
70.22; H, 7.36. Found: C, 70.39; H, 7.62.
the degree of mechanical sensitivity based on the up–down method
described by Dixon (1980).11 Each filament was applied in a perpen-
dicular fashion to the glabrous plantar surface of each hind paw
with a force causing the filament to buckle. This degree of force
was maintained for 5 s or until a positive response in the form of
sharp withdrawal or paw licking was observed. No withdrawal
response prompted the use of next filament with a higher buckling
force. A positive response prompted the use of next filament
with a lower buckling force. This paradigm continued until 4
measurements were obtained, beginning with the one before the
first change in response, or until five consecutive positive (assigned
score 0.25 g) or negative (assigned score 15 g) measurements
occurred.
2.1.1.4. 2-Aminomethyl-1-tricyclo[3.3.1.11,7]decanecarboxylic
acid hydrochloride GZ4.
A solution of cyanoacid 7 (6.5 g,
31.7 mmol) in absolute ethanol (50 mL) and HCl (37%, 5 mL) was
hydrogenated in the presence of PtO2 (1 g) catalyst under a pres-
sure of 50 psi, at 25 °C, for 5 h. A white solid was formed, which
was dissolved in methanol. The suspension was filtered to remove
the catalyst and the filtrate was evaporated until a small volume
remained. Dry ether was then added and the mixture was chilled.
The amino acid hydrochloride 8 solid formed, was filtered and
washed with dry ether: yield 6.43 g (almost quantitative); mp
Animals were tested for possible side effects observed as a
reduction of righting, stepping, corneal and pinna reflexes as previ-
ously described. No apparent side effects of GZ4 were observed.
2.2.2. Transient channel transfection and electrophysiological
recording
Human embryonic kidney (HEK)-293 cells (American Type
Culture Collection, ATCC) were grown in Dulbecco’s modified
Eagle’s medium (DMEM)-high glucose supplemented with 10% horse
204–206 °C (dec., MeOH–Et2O), IR (Nujol) v (C@O) 1717 cmÀ1 1H
;
NMR (400 MHz, DMSO-d6), d 1.42 (d, 1H, J = 12.8 Hz, 4e-H), 1.55–
1.86 (complex m, 9H, 4a, 6, 8, 9, 10-H), 1.88 (br s, 1H, 7-H), 1.95
(br s, 1H, 3-H), 2.03 (br s, 1H, 5-H), 2.20 (br d, 1H, J = 10.8 Hz, 2-
H), 2.57 (dd, 1H, J = 12.4 and 2.8 Hz, CHAHMNH2ÁHCl), 2.95 (t, 1H,
J = 12.0 Hz, CHAHMNH2ÁHCl), 8.14 (br s, NH2), 12.51 (br s, COOH)
ppm; 13C NMR (CDCl3, 100 MHz), d 26.8 (7-C), 26.9 (5-C), 27.1
(3-C), 29.5 (4-C), 32.3 (9-C), 36.2 (6-C), 37.0 (8-C), 38.2 (CH2NH2Cl),
41.2 (10-C), 42.3 (1-C), 42.7 (2-C), 177.4 (C@O) ppm. Anal. Calcd for
serum, 2 mM
penicillin and 100
humidified atmosphere. After splitting the cells on the previous day
and seeding at 50–60% confluency, cells were transfected with the
cDNA clones encoding recombinant CaV channel subunits.
Cell expression constructs were made by standard techniques,
and their fidelity was verified by DNA sequencing. The a1B-pKCRH2
L
-glutamine, 110 mg/l sodium pyruvate, 100 U/ml
lg/ml streptomycin at 37 °C in a 5% CO2/95% air
construct containing the cDNA clone encoding the rabbit brain N-
type Ca2+ channel CaV2.2 pore-forming subunit (GenBank accession
number D14157)12 was provided by Dr. B. Adams (Utah State Uni-
C12H20ClNO2: C, 58.65; H, 8.20. Found: C, 58.89; H, 8.54.
2.2. Biology
versity, USA). The cDNA coding the rat brain CaVa
2d-1 (M86621)13
and the rat brain CaVb3 (M88751),14 provided by Dr. K. Campbell
(University of Iowa, USA), were subcloned into the pcDNA3 vector
(Invitrogen). The rat neuronal CaV1.3 (AF370009)15 was cloned into
the pcDNA6/His vector (Invitrogen). HEK-293 cells were transiently
transfected using Lipofectamine Plus transfection reagent (Invitro-
2.2.1. Spinal nerve ligation-induced neuropathic pain model
and measurement of tactile allodynia
Sprague Dawley rats (140–160 g) were housed in cages on a
standard 12/12 h light/dark cycle and had free access to food and
drinking water before experiments. All experiments were
performed according to the Guidelines on Ethical Standards for
Investigation of Experimental Pain in Animals8 and the Mexican
regulation (NOM-062-ZOO-1999). In addition, these experiments
were approved by our local Ethics Committee (Protocol 455,
Cinvestav).
Spinal nerve ligation was produced in animals as follows. Intra-
peritoneal injection of ketamine/xylazine (45:12 mg/kg) was used
to anesthetize the rats. The back skin of the animals was shaved
and prepared aseptically for surgery. A midline incision was made
at the dorsal region followed by a blunt dissection to expose the
corresponding spinal cord region. The left L5 and L6 spinal nerves
were exposed and ligated with suture distal to the dorsal root gan-
glion. In the sham group, the surgical procedure was identical, but
the spinal nerves were not ligated. Rats were allowed to recover for
9 d before a second surgery to implant an intrathecal catheter. For
this procedure, animals were anesthetized with a ketamine/xyla-
zine (45:12 mg/kg, i.p.) and placed in a stereotaxic head holder
to expose the atlantooccipital membrane.9 A PE-10 catheter was
then passed intrathecally to the level of the lumbar junction. Rats
were allowed to recover from surgery for 5 d. Animals showing
signs of motor impairment were excluded from the studies and
euthanized.
gen) as per the manufacturer’s instructions using 1.6
cDNA clones encoding CaV channel subunits, together with 0.4
lg of the
l
g
of the Green Lantern plasmid encoding GFP (Life Technologies).
Two days after transfection, cells were briefly split at ꢀ10%
confluence and patch-clamp recording was performed ꢀ4 h later
from fluorescent cells. Likewise, HEK-293 cells stably expressing
the CaV3.1a channel (AF190860)16 were grown as previously
reported.17 For electrophysiological recordings, cells were lift-off
plates, re-seeded on poly-L-lysine (0.05%)-precoated glass cover-
slips and used ꢀ4 h after plating.
Whole cell patch-clamp recordings were performed as de-
scribed previously.18,19 Currents were recorded at RT (22–24 °C)
using Clampex 10 software and an Axopatch 200B amplifier
(Molecular Devices), digitized at 5.71 kHz and filtered at 2 kHz.
The resistance of the patch electrodes was 2-4 M
with the internal solution. After establishing the whole-cell mode,
capacitive transients were cancelled with the amplifier. Series
X when filled
resistance values were typically 2–10 MX, and no records were
used in which the voltage error (as defined by Ver = Imax  Ra)
was >5 mV. Leak and residual capacitance currents were sub-
tracted online by a standard P/4 protocol. Membrane capacitance
(Cm) was determined as described previously and used to normal-
ize currents.20 Current recordings were obtained from individual
cells and performed within the initial 15 min after breaking into
the cells to minimize rundown. The recorded cells were preincu-
bated with the drug for 1–3 h before electrophysiological
recording. Differences in transfection efficiency/expression were
avoided by co-transfecting the cells (in 1:4 dilution with respect
Development of tactile allodynia was determined using the 50%
withdrawal threshold method (PWT) as described previously.10
Sham and spinal nerve ligated rats were placed in individual clear
chambers positioned on top of a mesh screen for 30–40 min accli-
mation. A series of von Frey filaments were then used to measure

1800
G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
to the channel subunits) with a GFP construct, and recordings were
made only in GFP expressing cells. Currents through CaV channels
were recorded in the following extracellular solution (in mM): 120
tetraethylammonium chloride (TEA-Cl), 10 BaCl2, 10 4-(2-hydroxy-
ethyl)-1-piperazine-ethanesulfonic acid (HEPES), and 10 glucose,
with pH adjusted to 7.4 with TEA hydroxide (osmolarity
ꢀ300 mOsm kgÀ1). The intracellular solution consisted of (in
mM): 135 CsCl, 2 MgCl2, 10 HEPES, 4 MgATP, and 10 ethylenegly-
col-bis-(b-aminoethylether)N,N-tetraacetic acid (EGTA), pH 7.1,
osmolarity ꢀ280 mOsm kgÀ1
.
GZ4 was dissolved in sterile distilled water (2.9 mM stock solu-
tion) and diluted to the desired concentration in the cell culture
medium. HEK-293 cells expressing recombinant CaV channels were
preincubated (1–3 h) with GZ4 (500 lM) and then subjected to
electrophysiological recordings.
3. Results
Figure 3. Effects of intrathecal GZ4 on spinal nerve ligation-induced allodynia.
Animals with tactile allodynia were treated intrathecally with increasing doses ( g/rat)
3.1. GZ4 synthesis
l
of GZ4 or DMSO followed by the test for paw withdrawal thresholds (PWT) to von Frey
filament stimulation as indicated. Data presented are the means SEM from 6 rats in
each group. ⁄P < 0.05, by two-way repeated measures ANOVA).
The synthesis of the novel GABA derivative (GZ4), which dis-
plays a close structure–activity relationship with gabapentinoid
drugs, was achieved as shown in Figure 2. For the synthesis of the
oxirane 2, protoadamantanone 1 was used as starting material.21,22
This was converted to oxirane 2, which was obtained as an insepa-
rable mixture of epimers (endo:exo 1:15), as was evident from the
1H NMR spectral data, upon treatment with dimethylsulfoxonium
methylide. Ring opening of compound 2 with a solution of H2SO4
(0.085 M) in acetone afforded diol 3.23 Oxidation of diol 3 with
Jones reagent under mild conditions afforded the desired ketoacid
4.24 Esterification of compound 4 gave ketoester 5.25 Reductive
cyanation26 of compound 5 was accomplished in high yield using
toluenesulphonylmethyl isocyanide (TOSMIC) and yielded in
cyanoester 10; saponification of the latter afforded the desired
cyanoacid 7. Finally, hydrogenation of cyanoacid 7 over platinum
oxide catalyst in ethanol at 55 p.s.i., in the presence of HCl (37%),
gave the desired aminoacid GZ4 in its hydrochloride form.
Since gabapentinoid drugs have been shown to reduce the func-
tional expression of neuronal N-type (CaV2.2) channels, we exam-
ined the potential effects of the GZ4 treatment initially on
neuropathic pain relief in a rat model of neuropathic pain and then
on recombinant CaV2.2 channels heterologously expressed in
HEK293 cells.
of nerve ligated and sham animals. Our data indicated that the
nerve ligation induced a gradual reduction in the hindpaw with-
drawal threshold to the mechanical stimulation to a level consid-
ered allodynic (not shown). Data from our previous studies have
shown that an adamantane derivative of GABA, AdGABA, produces
strong inhibitory effects on whole-cell currents through neuronal
N-type (CaV2.2) Ca2+ channels4,7 as well as on L-type channels of
the CaV1.3 class,19 and exhibited analgesic activity on mice in the
hot plate test.4 Hence, in order to investigate whether GZ4, a com-
pound structurally related to AdGABA, has a potential analgesic
activity, we injected the compound intrathecally in nerve ligated
rats with allodynia followed by behavioral sensitivity assessment.
Bolus GZ4 treatment (10–300 lg) reversed allodynia in these ani-
mals, and its effects started 1 h after injection, peaked in 2 h, and
diminished after 4 h of GZ4 application (Fig. 3). Although it is
difficult to estimate the concentration in the spinal cord, there is
evidence that the therapeutic intrathecal doses of gabapentin in
several studies are in the range of 10–300 lg. Thus, the concentra-
tions of GZ4 used in our study are similar to those reported to be
effective for gabapentin.
3.3. GZ4 treatment inhibits macroscopic currents through
recombinant Ca2+ channels in HEK-293 cells
3.2. GZ4 treatment in spinal nerve ligated rats blocks allodynia
We first tested whether spinal nerve ligation led to the develop-
ment of neuropathic pain, by determining the sensitivity to von
Frey filament stimulation at the plantar surface of the hindpaw
To study whether a possible dysregulation of Ca2+ channel func-
tional expression contributes to the development of allodynia at the
Figure 2. GZ4 synthesis. Reagents and conditions: (a) trimethylsulfoxonium iodide, NaH, DMSO, argon, 3 h 25 °C and then 8 h 55 °C; (b); H2SO4 (98%), H2O, 75 min, reflux,
(two steps yield: 60%) (c) Jones reagent (1 mM) (61%); (d) (i) SOCl2, 65 °C, 15 min, (ii) MeOH (quant); (e) TOSMIC, abs EtOH, DME, t-BuOK, 0 °C, argon, 30 min at 20 °C, and 1 h
at 48 °C (81%); (f) EtOH, NaOH, H2O, reflux, 2.5 h and then HCl (93%); (g) H2/PtO2, EtOH, HCl, 50 p.s.i., 5 h (almost quantitative).

G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
1801
Figure 4. CaV2.2 channel current inhibition by GZ4 in HEK-293 cells. (A) Representative IBa traces under control conditions and inhibition in the presence of 500
l
M GZ4 at
different time points up to 3 h. (B) Bar chart showing data for mean current density through recombinant N-type Ca2+ channels (CaV2.2
a1/CaVb3/CaVa2d-1) expressed in HEK-
293 cells measured at the peak of the inward current in response to a 140 ms voltage step command to 10 mV from a Vh of À80 mV. Data under control condition (Ctl) and in
the presence of 500 lM GZ4 are shown. The number of recorded cells is shown in parenthesis and the asterisk denotes significant difference from the respective control
(⁄P < 0.05, by Student’s t-test).
spinal level of the nerve ligated rats, we next tested the effects of
GZ4 on neuronal CaV2.2 channels heterologously expressed in
HEK-293 cells. To this end, cells expressing the recombinant
Figure 5A shows the effect of GZ4 on IBa elicited in HEK-293
cells by depolarization to a wide range of voltages. As can be seen,
treatment with GZ4 (3 h) inhibited inward currents at voltages
ranging from À20 to +30 mV. Likewise, the participation of the
CaV2.2 channels (a1B/a2d/b3) were kept in culture up to 3 h in the
presence of 500 M GZ4 prior to being subjected to electrophysio-
l
a2d subunit in the regulation of CaV channel voltage-dependent
logical recording. The results of this analysis showed a significant
inhibition of CaV2.2 current density after drug treatment. The traces
shown in Figure 4A illustrate representative whole-cell Ba2+ cur-
rents (IBa) through CaV2.2 channels elicited by 140 ms depolarizing
pulses to +10 mV from a holding potential (Vh) of À80 mV in HEK-
293 cells kept in a culture in the absence (control) and the presence
inactivation is well documented. This auxiliary subunit produces
a hyperpolarizing shift in half-inactivation potential of CaV chan-
nels.19,27 Consistent with this, GZ4 treatment resulted in a signifi-
cant change in the steady-state inactivation behavior of the CaV
channels. Indeed, IBa in cells exposed to the drug exhibited an
apparent ꢀ10 mV hyperpolarizing shift in half-inactivation poten-
tial (Fig. 5B). The rightward shift in the prepulse inactivation curve
could act to increase Ca2+ currents under physiological conditions.
Hence, at a resting potential of À80 mV, there are more channel
(about 5%) available to be open, although this is not sufficient to
counteract the significant reduction of ꢀ35% in maximum current
observed after drug application. This finding may suggest that the
inhibitory effect of the drug could be underestimated in our model
system.
of GZ4. Figure 4B shows that incubation of 1–2 h with GZ4 (500 lM)
had a small effect on current density through recombinant CaV2.2
channels. In contrast, a significant inhibition of channel activity
was observed after 3 h of drug treatment, supporting the notion
that GZ4 may inhibit the functional expression of CaV2.2 channels.
Current inhibition was incomplete (ꢀ30%) with a GZ4 concen-
tration of 500 lM, a value that is comparable to that obtained for
GBP6,18 but greater than that observed with AdGABA.4 It should
be noted, however, that AdGABA-induced inhibition of channel
activity was observed upon chronic treatment (>24 h) with the
drug, and that acute treatment (<8 h) had no significant effects
on current amplitude.
To corroborate role of the
ulation of N-type CaV2.2 channels, we next expressed recombinant
channels with the 1B/b3 subunit composition in HEK-293 cells and
tested the action of the drug. We observed that Ca2+ currents in
a2d subunit in the GZ4-mediated reg-
a
Figure 5. GZ4 treatment affects recombinant CaV2.2 current density and inactivation. (A) Current–voltage relationships for N-type CaV channel transfected cells (CaV2.2a1
/
CaVb3/CaV 2d-1) in control conditions or after 3 h of treatment with 500 M GZ4 (n = 3–4). (B) Voltage dependence of current inactivation. Currents were measured in
a
l
transfected cells during depolarizations to +20 mV, preceded by 2 s inactivating pulses (prepulses) of various amplitudes applied from a Vh of À80 mV, in the absence and
presence of the drug. Steady-state inactivation curves were fitted with a Boltzmann equation of the form: I = Imax {1 + exp[(V À Vm)/k]}, where Imax represents the maximal
½
current amplitude, V the potential for half-maximal inactivation of Imax, and k is a slope factor. The resulting V (mV) values for control and GZ4 treated cells were À73.1 and
½
½
À67.1, respectively (n = 4–6). (C) Current–voltage relationships for CaV2.2
a
1/CaVb3 channel transfected cells (without CaVa2d-1) in control conditions or after 3 h of treatment
with 500 lM GZ4 (n = 6).

1802
G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
Figure 6. GZ4 treatment affects recombinant CaV1.3 but not CaV3.1 channel activity. (A) Current-voltage relationships for L-type CaV channel transfected cells (CaV1.3a1
CaVb3/CaV 2d-1) in control conditions or after 3 h of treatment with 500 M GZ4 (n = 9). (B) Current–voltage relationships for CaV1.3 1/CaVb3 channel transfected cells
(without CaV 2d-1) in control conditions or after 3 h of treatment with 500 M GZ4 (n = 5–7). (C) Current–voltage relationships for T-type CaV3.1 channel transfected cells
(without auxiliary subunits) in control conditions or after 3 h of treatment with 500
/
a
l
a
a
l
lM GZ4 (n = 5–6).
cells expressing
(4–5-fold) than those in cells expressing
a
1B/b3 channels were substantially smaller
Conotoxins are small peptides in the venom of tropical marine
cone snails from the genus Conus that target ion channels includ-
ing Na+ and Ca2+ channels. Indeed, there is considerable interest
in finding clinically relevant drugs to block Na+ and Ca2+ channels
involved in pain signaling.3,28 Despite the fact that several cono-
toxins have been tested in clinical trials, peptide toxins as drug
leads have a few drawbacks, such as unstable disulfide bonds,
protease degradation and inefficient transport to its intended site
a2d, and that the voltage
dependence of activation was right shifted about 10 mV, what is
expected for macroscopic currents through channels lacking the
auxiliary subunit. Interestingly, the above mentioned effects of
GZ4 on N-type CaV2.2 channel currents did not occur in the ab-
sence of the
dition in HEK-293 cells transfected with CaV2.2 channels (
a
2d subunit. Ca2+ currents recorded in the control con-
a
1B/b3)
were of similar magnitude either in the absence or in the presence
of action.3 Only the
x-conotoxin MVIIA from Conus magus has
of GZ4 (Fig. 5C).
been approved as an analgesic drug (named Ziconotide) for severe
Previously, we showed that the co-expression of the
a2d sub-
chronic pain.28,30
unit renders the CaV1.3 L-type channels sensitive to gabapentinoid
drugs GBP and AdGABA.19 Therefore, we sought to determine
whether these channels were also sensitive to the actions of GZ4.
In the current report, we describe the synthesis of a GABA
analog, GZ4, with certain structure–activity relationship to gaba-
pentin. By using a combined approach of spinal nerve ligation-in-
duced neuropathic pain model as well as electrophysiological
recordings and molecular biology, we show relevant aspects
regarding the mechanism of action of this novel gabapentinoid
drug. Our results show that this compound mimics some of the ac-
tions of GBP, i.e. inhibits N-type Ca2+ currents most likely acting on
To this end, we co-expressed the
nels ( 1D/b3) in HEK-293 cells and tested the effects of GZ4
(500 M) after 3 h of incubation. Interestingly, GZ4 had a signifi-
cant effect on peak L-type current amplitudes in cells expressing
2d/b3 recombinant channels (Fig. 6A). Average peak L-type
a2d subunit and the CaV1.3 chan-
a
l
a1D/a
current density at À30 mV was À60.1 8.3 in absence and
the a2d subunit of the CaV channel complex heterologously ex-
À130.3 15.3 pA/pF in presence of the drug. Consistent with the
pressed in HEK-293 cells. However, differences in the time courses
of GZ4 and gabapentin action suggest divergences in their mode of
action. It is acknowledged that gabapentin in vitro actions require
at least 24 h and imply uptake of the drug, which suggests that
gabapentin acts mainly intracellularly affecting channels traffick-
ing to the plasma membrane.5,6 Indeed, recent studies have shown
contribution of the
a2d subunit to this regulation, the inhibitory
effects of GZ4 on L-type CaV1.3 channel currents were lost in the
absence of the Ca2+ channel auxiliary subunit (Fig. 6B). Last, incu-
bation (3 h) with the GZ4 (500 lM) had no effect on voltage-gated
Ca2+ though T-type CaV3.1 channels stably expressed in HEK-293
cells used also as a negative control (Fig. 6C).
that GBP inhibits post-Golgi forward trafficking of the a2d subunit
in a manner that is prevented by dominant-negative construct of
the small GTPase Rab11, which disrupts trafficking through the
recycling endosome compartment. This finding indicates that
4. Discussion
GBP may disrupt the interaction between a2d and sorting proteins
Chronic pain is one of the most prevalent and disabling condi-
tions in clinical practice. However, its therapeutic management is
challenging. Diverse pharmacological options are available to man-
age different pain mechanisms including antidepressants, anticon-
vulsants opioid and topical analgesics.
Given that neuropathic pain is associated with changes in volt-
age-gated Ca2+ activity and expression, gabapentin and pregabalin
are now been widely used in the treatment of pain. These two
gabapentinoid drugs act as neuromodulators by selectively bind-
in this compartment, and that this may represent a rate-limiting
step in calcium channel trafficking.31 On the other hand, the faster
time course (1–3 h) of GZ4 effect suggests a direct inhibition of
Ca2+ currents through the alteration of the biophysical properties
of cell surface channels. Our electrophysiological data support this
hypothesis.
Although the reasons for the apparent differences in the time
course of GZ4 and GBP action are currently unknown, they may
lie in the structural differences. Gabapentin and AdGABA show
similarities in the orientation in both the carboxymethylene and
aminomethylene moiety. However, in the case of GZ4 the carbox-
ylic acid group possessed a complete different orientation (Fig. 1B)
and at the same time the two methylenes of the GABA moiety were
introduced into the adamantane skeleton and adopted a very spe-
cific stereochemical orientation. This study demonstrates distinct
ing to the
a
2d subunit of voltage gated Ca2+ channels in various
regions of the brain and the dorsal horn of the spinal cord.28 They
also have a peripheral analgesic action.29 These actions result in
an inhibition of the release of excitatory neurotransmitters that
are important in the production of pain. Likewise, several
conotoxins are being investigated as antinociceptive drugs.3

G. Zoidis et al. / Bioorg. Med. Chem. 22 (2014) 1797–1803
1803
structure–activity relationships for the tested compounds, and
Acknowledgments
therefore it is reasonably to speculate that the changes to the size
and conformation of the GABA moiety had a significant effect on
biological activity. Most probably, GZ4 and gabapentin may be
affecting differentially the tertiary structure of the receptor target
This work was supported by funds from The National Council
for Science and Technology (Conacyt, Mexico; Grants 128707;
179294) to R.F. and V.G.S., respectively, and from PAPIIT-UNAM
(Grant IN221011) to A.S. A doctoral fellowship from Conacyt to
J.B.P.F. is gratefully acknowledged.
(the CaV channel
ture studies.
a2d subunit). These are interesting topics for fu-
It is also worth noting that high concentrations of GZ4 (500
lM;
Figs. 4 and 5), as well as of GBP,5,6 are required to observe significant
effects on IBa when recordings are made in single cells kept in
culture. This may be related to the over-expression in the heterolo-
gous system of the CaV channel b auxiliary subunit which may
change the binding properties of the channel complex. In addition,
the possibility exists that the amino acids in the culture medium,
may compete for the uptake of the gabapentinoids. However, the
conditions present in the in vitro system differ significantly from
those in vivo, in which the drug showed a more potent effect
(Fig. 3).
In line with this, there have been a number of studies docu-
menting the beneficial effects of gabapentinoid drugs in different
models of neuropathic pain as well as in different types of neuro-
pathic pain like neuropathy due to cancer, HIV infection, diabetic
neuropathy, trigeminal neuralgia and post-operative neuropathic
pain.32 In contrast to what has been described in vitro, the clinical
assays show that gabapentinoids produce anti-allodynic effects
ꢀ1 h following oral administration.33 This time course is consistent
with the pharmacokinetic distribution of the compound. The dis-
parity in results between the heterologous expression systems
and the clinical studies may be due to several factors including
channel subunit heterogeneity among native versus recombinant
channels and in the pathologic state of the tissue that is, hyperal-
gesic or normal. In addition, it is worth noting that gabapentin
has been shown to directly inhibit NMDA receptors and TRPA1
channels that might also be responsible for its anti-nociceptive
activity.32
Data from our behavioral testing showed that intrathecal GZ4
application significantly reversed nerve ligation-induced tactile
allodynia compared to that in allodynic rats treated with vehicle.
While the underlying mechanisms of GZ4 are unknown and still
under investigation, our findings on the electrophysiological
recording suggest that GZ4 may act inhibiting the functional
expression of N-type Ca2+ currents. Taken together, our results sug-
gest that GZ4 might serve as new therapeutic agent. It is possible
that some benefits of the GZ4 treatment may be of clinical signifi-
cance, although this requires further investigation. The introduc-
tion of a mechanism-based drugs with a specific pharmacological
action, may lead to more rational treatment for the patients with
neuropathic pain.
References and notes
1. Djouhri, L.; Fang, X.; Koutsikou, S.; Lawson, S. N. Pain 2012, 153, 1824.
2. Adams, D. J.; Berecki, G. Biochim. Biophys. Acta 2013, 1828, 1619.
3. Knapp, O.; McArthur, J. R.; Adams, D. J. Toxins (Basel) 2012, 4, 1236.
4. Zoidis, G.; Papanastasiou, I.; Dotsikas, I.; Sandoval, A.; Dos Santos, R. G.;
Papadopoulou-Daifoti, Z.; Vamvakides, A.; Kolocouris, N.; Felix, R. Bioorg. Med.
Chem. 2005, 13, 2791.
5. Vega-Hernandez, A.; Felix, R. Cell. Mol. Neurobiol. 2002, 22, 185.
6. Hendrich, J.; Van Minh, A. T.; Heblich, F.; Nieto-Rostro, M.; Watschinger, K.;
Striessnig, J.; Wratten, J.; Davies, A.; Dolphin, A. C. Proc. Natl. Acad. Sci. U.S.A.
2008, 105, 3628.
7. Martínez-Hernández, E.; Sandoval, A.; González-Ramírez, R.; Zoidis, G.; Felix, R.
Toxicol. Appl. Pharmacol. 2011, 250, 270.
8. Zimmermann, M. Pain 1983, 16, 109.
9. LoPachin, R. M.; Rudy, T. A.; Yaksh, T. L. Physiol. Behav. 1981, 27, 559.
10. Chaplan, S. R.; Bach, F. W.; Pogrel, J. W.; Chung, J. M.; Yaksh, T. L. J. Neurosci.
Methods 1994, 53, 55.
11. Dixon, W. J. Annu. Rev. Pharmacol. Toxicol. 1980, 20, 441.
12. Fujita, Y.; Mynlieff, M.; Dirksen, R. T.; Kim, M. S.; Niidome, T.; Nakai, J.;
Friedrich, T.; Iwabe, N.; Miyata, T.; Furuichi, T.; Furutama, D.; Mikoshiba, K.;
Mori, Y.; Beam, K. G. Neuron 1993, 10, 585.
13. Kim, H. L.; Kim, H.; Lee, P.; King, R. G.; Chin, H. Proc. Natl. Acad. Sci. U.S.A. 1992,
89, 3251.
14. Castellano, A.; Wei, X.; Birnbaumer, L.; Perez-Reyes, E. J. Biol. Chem. 1993, 268,
3450.
15. Helton, T. D.; Xu, W.; Lipscombe, D. J. Neurosci. 2005, 25, 10247.
16. Lee, J. H.; Daud, A. N.; Cribbs, L. L.; Lacerda, A. E.; Pereverzev, A.; Klöckner, U.;
Schneider, T.; Perez-Reyes, E. J. Neurosci. 1999, 19, 1912.
17. Avila, T.; Andrade, A.; Felix, R. J. Cell Physiol. 2006, 209, 448.
18. Sandoval, A.; Andrade, A.; Beedle, A. M.; Campbell, K. P.; Felix, R. J. Neurosci.
2007, 27, 3317.
19. Andrade, A.; Sandoval, A.; Gonzalez-Ramirez, R.; Lipscombe, D.; Campbell, K.
P.; Felix, R. Cell Calcium 2009, 46, 282.
20. Avila, G.; Sandoval, A.; Felix, R. Cell. Mol. Neurobiol. 2004, 24, 317.
21. Majersky, Z.; Hamesrak, Z. Org. Synth. 1988, 50, 958.
22. Kolocouris, N.; Zoidis, G.; Fytas, C. Synlett 2007, 1063.
23. Zoidis, G.; Benaki, D.; Myrianthopoulos, V.; Naesens, L.; Clercq, E. D.; Mikros, E.;
Kolocouris, N. Tetrahedron Lett. 2009, 50, 2671.
24. Farcasiu, D. J. Am. Chem. Soc. 1976, 98, 5301.
25. Peters, J. A.; Van Der Toorn, J. M.; Van Bekkum, H. Tetrahedron 1975, 31, 2273.
26. Oldenziel, O.; Van Leusen, D.; Van Leusen, A. J. Org. Chem. 1977, 42, 3114.
27. Felix, R.; Gurnett, C. A.; De Waard, M.; Campbell, K. P. J. Neurosci. 1997, 17, 6884.
28. Pexton, T.; Moeller-Bertram, T.; Schilling, J. M.; Wallace, M. S. Expert Opin.
Invest. Drugs 2011, 20, 1277.
29. Carlton, S. M.; Zhou, S. Pain 1998, 76, 201.
30. Miljanich, G. P. Curr. Med. Chem. 2004, 11, 3029.
31. Tran-Van-Minh, A.; Dolphin, A. C. J. Neurosci. 2010, 30, 12856.
32. Kukkar, A.; Bali, A.; Singh, N.; Jaggi, A. S. Arch. Pharm. Res. 2013, 36, 237.
33. Schmidt, P. C.; Ruchelli, G.; Mackey, S. C.; Carroll, I. R. Anesthesiology 2013, 119,
1215.