Synthesis of 1-substituted epibatidine analogues and their in vitro and in vivo evaluation as α4β2 nicotinic acetylcholine receptor ligands

Article abstract of DOI:10.1039/c3ra44379e

The highly potent natural alkaloid epibatidine remains a source of inspiration in the search for new analgesic drugs. In this paper, we describe an expansion of our previously reported synthesis of epibatidine analogues, and five synthetic alkaloids chara

Full text of DOI:10.1039/c3ra44379e

RSC Advances
View Article Online
View Journal | View Issue
PAPER
Synthesis of 1-substituted epibatidine analogues
and their in vitro and in vivo evaluation as a4b2
nicotinic acetylcholine receptor ligands†
Cite this: RSC Adv., 2014, 4, 2226
Thomas S. A. Heugebaert,a Melissa Van Overtveldt,a Ann De Blieck,a Benjamin Wuyts,b
Patrick Augustijns,b Eugenia Ponce-Gamez,c Alicia Rivera,c Dominic De Groote,d
´
Romain A. Lefebvre,e Patrick Wouters,f Theo Meert,f Jacques Devulderg
a
*
and Christian V. Stevens
The highly potent natural alkaloid epibatidine remains a source of inspiration in the search for new analgesic
drugs. In this paper, we describe an expansion of our previously reported synthesis of epibatidine analogues,
and five synthetic alkaloids characterized by a symmetric, 1-substituted 7-azabicyclo[2.2.1]heptane
skeleton, were evaluated for their biological activity. Two of these are binding selectively to the a4b2
subtype of the nicotinic acetylcholine receptor. Their Ki values were determined to be 40 and 290 nM.
After a favourable evaluation of these compounds' cytotoxicity and metabolic stability, they were
submitted to a rat tail flick test. The compounds did not show antinociceptive effects, which may be
caused by a combination of insufficient potency and poor brain penetration.
Received 14th August 2013
Accepted 14th November 2013
DOI: 10.1039/c3ra44379e
www.rsc.org/advances
Medical Association reframed from publishing commercials
that aunted which brand of cigarettes doctors prefer to
smoke.1 Fortunately, this trend has now reversed and, through
Introduction
Upon scouting through The New World in the early 1500's, the
Spanish conquistadores who observed Cubans taking in
‘Tabacos’ smoke could not have estimated the impact this age-
old custom would have on the further course of social and
scientic developments. The detrimental effect of smoking on
human health was unknown and the addictive nature of the
nicotine alkaloid caused a rapid dispersal of tobacco products
around the world. By the mid 1900's, smoking had become so
widespread and socially accepted that not even the American
the in-depth study of the effects of nicotine, some good may
have come out of it. The availability of epidemiologic studies
discerning smokers and non-smokers, has led to a vast amount
of data indicating the involvement of nicotine and, more
importantly, the nicotinic acetylcholine receptors (nAChR's) in
a broad range of behavioral and disease patterns. One random
example is the stunningly high level of nicotine addiction
(72–90%) in schizophrenic patients.2 Self-medication leads
them to subconsciously modulate their nicotinic receptors by
using nicotine. Aside from Schizophrenia,3 nAChR's have been
associated with many neurological disorders such as depres-
sion,4 Alzheimer's disease,5 Parkinson6 and ADHD7 and other
phenomena such as neuropathic pain8 and inammation.9
For each of these disorders, specic subtypes of the nAChR
are involved. The nAChR is a membrane bound ion channel,
consisting of 5 equal (e.g. subtype a7) or differing (e.g. subtype
a4b2) subunits. In the human genome, as much as 16 genes
have been identied encoding for distinct nAChR subunits
(a1–7, 9 and 10, b1–4, g, d and 3).10 The subunit composition of
the nicotinic receptor, which denes its subtype, is controlled
for each single cell by differential expression of the nicotinic
subunits. These then assemble in different ratio's to fully
functional receptors. The expression ratio varies with organ,
anatomic location, tissue type, individual, and may even change
during the aging process.11 Furthermore, nAChR's are not
exclusively present in the central nervous system (CNS) and may
also be expressed in other cells and organs, including skin,
aResearch Group SynBioC, Department of Sustainable Organic Chemistry and
Technology, Faculty of Bioscience Engineering, Ghent University, Coupure links 653,
B-9000 Ghent, Belgium. E-mail: Chris.Stevens@UGent.be; Thomas.Heugebaert@
Ugent.be; Fax: +32 92 646221; Tel: +32 92 645957
bLaboratory of Pharmacotechnology and Biopharmacy, Gasthuisberg, KU Leuven, 3000
Leuven, Belgium
c
´
Department of Cell Biology, Faculty of Sciences, University of Malaga, 29071 Malaga,
Spain
dDepartment of Pharmaceutics, Faculty of Pharmaceutical Sciences, Ghent University,
Harelbekestraat 72, 9000 Ghent, Belgium
eHeymans Institute for Pharmacology, Ghent University, De Pintelaan 185, 9000
Ghent, Belgium
fDepartment of Anesthesia, University hospitals Ghent, De Pintelaan 185, 9000 Ghent,
Belgium
gMultidisciplinary Pain Centre 3B2, University Hospitals Ghent, De Pintelaan 152,
9000 Ghent, Belgium
† Electronic supplementary information (ESI) available: Full experimental details
of the used synthetic and biological methods, off-target receptor binding results
and copies of 1H- and 13C-NMR spectra of the new compounds. See DOI:
10.1039/c3ra44379e
2226 | RSC Adv., 2014, 4, 2226–2234
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
leukocytes, kidney, urinary bladder, etc.12 The result is a
complex checkerboard of nicotinic receptor subtypes across the
organism, each of which performs a specic function.
With regard to small molecule development, it is important
to note that nicotinic ligands bind at the interface of two
adjacent subunits.13 This augments the number of possible
variations of the binding pocket, but allows for the synthesis Fig. 2 Synthetic approach to symmetric epibatidine analogues.
nAChR subtype selective compounds, as opposed to subunit
selective compounds.
Despite the mere subtle variations between receptor subtypes,
dozens of subtype selective ligands have been reported, about
thirty of which are currently in phases I, II or III of clinical
development.11 These compounds typically display a Ki of 1 to
100 nM at their respective targets, with some highly potent
examples reaching a Ki of 0.05 nM. Four nicotinic modulators are
currently marketed: nicotine itself, bupropion, varenicline and
cytisine, all situated in the eld of smoking cessation.
Fig. 3 Tested epibatidine analogues.
Central to our study is the epibatidine alkaloid (1, Fig. 1). It
was rst isolated in 1974 by Daly and Myers from the skin of the
Ecuadorian frog Epipedobates tricolor and displayed a remark- developed can be used as an easy entry to this type of
able analgesic potency, about 200-fold higher than that of compounds.21 The pyridine substituent is attached to the bicy-
morphine.14 Unfortunately, the toxicity of epibatidine is too clic core by means of a linker derived from the nitrile.
high for any human therapeutic use. Epibatidine is a highly
In this paper, ve analogues were evaluated for their bioac-
potent a4b2 nAChR agonist,15 but displays poor subtype selec- tivity. The selected analogues are representative for the class,
tivity16 and furthermore binds to the muscarinic acetylcholine containing an sp2 one-atom linker (5a), an sp3 one-atom linker
receptors,17 thus causing severe side effects. Consequently, a (5b), a mixed sp2/sp3 two-atom aminomethyl linker (5c) and two
variety of synthetic, selective a4b2 agonists have entered clinical sp3 two-atom linkers (5d and 5e). Note that the presence of a
development for pain treatment (e.g. tebanicline 3, sonicline 4, free alcohol or amine introduces a chiral centre in compounds
TC-2696 2 and TC-6499).11 Most have been discontinued, due to 5b, 5d and 5e. They were used as racemates for their preliminary
gastrointestinal side-effects or lack of efficacy. Recent studies, bioevaluation (Fig. 3).
however, have shown that co-administration of allosteric
In the selection of these compounds, the focus was placed on
modulators may alleviate some of these concerns18 and a connection of 5 bonds between both nitrogen atoms. Previous
currently the search for selective, effective and well tolerated studies have shown that an important aspect upon changing the
epibatidine analogues, useful for pain treatment continues.8a,19 connection type between the azabicycle and the pyridine ring is
the distance between the bridgehead nitrogen and the pyridine
nitrogen, the so-called internitrogen distance. To maintain
biological activity, this distance has to remain within an
Chemistry
optimal range, which is, depending on the compounds repor-
Several modications of the epibatidine scaffold have been
22
˚
¨
ted, between 4.6 and 8 Angstrom (A). For the selected
reported. However, very few analogues bearing the pyridine
substituent at the bridgehead of the azabicyclo[2.2.1]heptane
are present in literature.20 This substitution pattern offers a
major advantage because, in this way, chirality is stripped from
the bicyclic system by a plane of symmetry (Fig. 2). Synthetically,
the reversible cyanide addition protocol we have previously
compounds, conformational sampling showed an internitrogen
˚
distance varying between 4.2 and 6.2 A, and the rotational
energy barriers were found to be no larger than 17 kcal molꢀ1
.
23
As such, these compounds are expected to easily adopt the
conformation needed for receptor binding.
Compounds 5a–5c were synthesized as previously published
(Scheme 1).21a The major problem for the construction of the
bicyclic ring by intramolecular ring closure, namely that the
nucleophilic amine and the leaving group need to be positioned
in a trans relationship, was circumvented by using the acetone
cyanohydrine24 induced reversible cyanide addition.
From the bicyclic precursors 8, analogues 5a and 5b were
obtained by the addition of a lithiated pyridine, hydrolysis of
the resulting ketimine and either reductive or oxidative depro-
tection. The yield of 5b was improved, using an alternative
purication procedure aer this last step. Column chromatog-
raphy over aluminium oxide, using highly polar solvent
mixtures (CH2Cl2–MeOH 95/5) provides better recovery and
Fig. 1 Epibatidine and its derived selective a4b2 nAChR agonists
developed for pain treatment.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 2226–2234 | 2227

View Article Online
RSC Advances
Paper
Scheme 1 Synthesis of 5a–5c according to ref. 21a.
improves the isolated yield by 22%. Analogue 5c was obtained delivering 11 cleanly in 94% yield, albeit with long reaction
from 8b in three steps. Reduction of nitrile 8b to the corre- times (45 h). Also the hydrolysis of the initially formed ketimine
sponding amine was performed quantitatively by means of is a sluggish process, and when the reaction mixture was
LiAlH4. Next, the pyridyl group was introduced using a Buch- quenched without acidication, the intermediate ketimine,
wald–Hartwig coupling.25 Removal of the protective benzyl which exists primarily in its enamine form 12, could be isolated
group was performed by reuxing in methanol, using ammo- in 65% yield aer crystallisation.
nium formate as the reducing agent.
As direct reduction protocols of the intermediates 11 and 12
Expanding the synthesis to two-carbon linkers, the addition (using H2–Raney Ni; H2–Pd/C; NH4HCO2–Pd/C or 1,3-c-hex-
of lithiated a-picoline 9 was attempted (Scheme 2). Treatment of adiene-Pd/C) did not result in the formation of analogues 5d
two equivalents picoline with equimolar amounts of BuLi in and 5e, a two-step protocol was used, rst reducing the ketone
diethyl ether for 1.5 h at ꢀ20 ꢁC led to complete deprotonation, and enamine functionalities (Scheme 3). Reduction of the
as shown by a quench with benzaldehyde. The addition of the ketone was straightforward, and treatment with NaBH4 in
resulting picolyl anion to the neopentylic nitrile, however, MeOH delivered the corresponding alcohol 13 in excellent
proved to be a slow process.26 Only aer 45 hours at room yields. Treatment of enamine 12 under the same conditions,
temperature conversion of the starting material was complete, however, did not result in the formation of amine 14. The
delivering 52% of the corresponding ketone 11 aer acidic enamine tautomer is highly favoured over the imine, resulting
hydrolysis and purication. The addition could be accelerated in very slow reaction rates. Moving to slightly acidic conditions,
using a less polar 5 : 8 hexane : Et2O solvent mixture. In this the imine/enamine tautomerisation could be accelerated,
case, conversion was complete in 12 hours, less side products resulting in improved reaction rates. Treatment with NaBH3CN
were formed, and the resulting yield was increased to 75%. and one equivalent of acetic acid in MeOH delivered amine 14
Further optimisation showed that THF is most efficient, in high yield aer 5 days.
Subsequent debenzylation of alcohol 13 could be performed
with full conversion using either 1,3-cyclohexadiene (4 equiva-
lents) or hydrogen gas as the reductant. Even though the reac-
tion using cyclohexadiene proceeded much faster (3 h, reux),
more side products were formed and treatment with hydrogen
gas (6 bar, 20 h, room temp) was preferred. Aer ltration of the
palladium catalyst, analogue 5d could be crystallised in 69%
yield from the crude mixture.
The preferred source of reductants for debenzylation of
amine 14 was ammonium formate, resulting in 91% conversion
aer 7 hours at reux temperature. The resulting diamine 5e is
Scheme 2 Addition of the a-picolinyl anion to neopentylic nitrile 8b.
highly polar and water soluble, which results in a tedious
purication using classic chromatographic methods. However,
dissolution of the crude product in distilled water, followed by
washing with ether, EtOAc and dichloromethane, delivered the
pure compound in the aqueous phase. Removal of the water
delivered analogue 5e in 42% yield.
Biology
In vitro: binding affinity and IC50 determination
To get a general idea of the nicotinic binding affinity of these
compounds, they were submitted to a series of competitive
binding assays at a 1 mM concentration. The inhibition
Scheme
and 5e.
3 Reduction and debenzylation towards analogues 5d
2228 | RSC Adv., 2014, 4, 2226–2234
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
phases I, II or III of clinical development.11 More importantly,
pharmacokinetics and selectivity proles are parameters to be
investigated.
Table 1 Determination of nAChR binding affinity at 1 mM concentra-
tion of compounds 5a–5e
% inhibition of nAChR control specic bindinga
When re-evaluating the concentration response curves in
this respect, the synthesized compounds clearly display a
higher subtype selectivity than epibatidine. Epibatidine shows
an IC50 of 13 nM (Ki ¼ 12 nM, nH ¼ 0.5) at the muscle type
nicotinic receptor, which is believed to be the cause of respi-
ratory paralysis.16,17 This concentration is a mere factor 100 from
the activity at a4b2 (Ki ratio ¼ 137; compare sazetidine-A:
24 000;33 rivanicline: 1380).35 The synthetic compounds on the
125I]a-bungarotoxin other hand show no binding to this receptor, or the a7 subtype
with that respect, at the highest concentration tested (Ki ratio:
5b $ 350; 5c $ 2500).
b
c
a4b2
a7
Muscle typec
5a
5b
5c
5d
5e
8
49
82
12
ꢀ17
0
22
ꢀ3
ꢀ4
ꢀ7
4
ꢀ11
ꢀ7
ꢀ7
ꢀ9
a
Average of 2 measurements, values below 30% are considered as not
signicant. [3H]cytisine (agonist radioligand).
[
b
c
(antagonist radioligand).
As such, these compounds and more specically 5c which
percentage was obtained using known radioligands, and the has a Ki ratio in the range of known selective compounds,
results are shown in Table 1. Values are expressed as the represent a clear improvement over the natural alkaloid.
percentage decrease of control specic binding in the presence
of the compounds and were determined by scintillation
counting. Screening of three nicotinic receptor subtypes
In vitro: cytotoxicity and metabolic stability
revealed a promising prole for compounds 5b and 5c. They are With these favourable in vitro enzyme binding results, we
binding to the a4b2 subtype of the nicotinic receptor family and moved on to the evaluation of toxicity and ADME characteris-
leave other receptors untouched.27
tics. Cytotoxicity was evaluated at 4 different cell lines; 2
As the nicotinic receptors will be used as target for pain neuronal cell lines representing the target cells and 2 hepatoma
treatment,28 an extra series of neurotransmitter recognition cell lines as a model of metabolism mediated toxicity.36 The
sites relevant to either the nicotinic acetylcholine receptors or compounds were tested at a concentration range of 0.1 to
pain treatment was included in the affinity tests (see ESI† for 100 mM and tamoxifen and terfenadine were included as posi-
full data). The panel contained representatives of the major tive controls. Cell viability was evaluated by both mitochondrial
classes of the ligand gated ion channels (GABA, glycine, sero- metabolism (formazan dye formation, MTS) and cell membrane
tonin [5-HT] and nicotinic acetylcholine), the muscarinic integrity (lactate dehydrogenase release, LDH)37 and the results
acetylcholine receptors, and other CNS related receptors rele- are presented in Table 2.
vant to pain mechanisms such as the adrenergic,29 dopami-
Both compounds appeared to be in solution at all concen-
nergic,30 cannabinoid31 and benzodiazepine32 receptors. No trations tested. No effects on cell viability were observed against
signicant (>30%, by historic reference) binding to any of these either neuronal cell line (IMR32 and B65) with either compound
receptors could be detected for compounds 5b and 5c. Based on at concentrations up to 100 mM in either the MTS or LDH assays.
these observations, analogues 5b and 5c were studied in more A small reduction in cell viability was observed at 100 mM in
detail, starting with the determination of a concentration both hepatoma cell lines (HepG2 and H4-II-E-C3) with the MTS
response curve at the nicotinic receptors in order to further assay. In comparison to controls, no gross changes in cell shape
elucidate their subtype selectivity. Compounds 5a, d and e were or distribution in the well were observed by microscopy. No
excluded from further study.
effect was observed at a concentration of 30 mM or below, which
As can be seen from Fig. 4, both compounds display selective sets the dose leading to (minor) cytotoxicity well above that of
binding to the a4b2 subtype of the nicotinic receptor family. The the observed affinity values.
IC50 values at this subtype could be determined at 0.86 mM (Ki ¼
The small effect at high concentration on the hepatoma cell
290 nM, nH ¼ 0.8) for 5b, and 0.12 mM (Ki ¼ 40 nM, nH ¼ 0.9) for lines may simply be tissue/cell specic but it is also possible
5c. When comparing these results to those of the naturally that the effect is caused by a metabolite generated during the
occurring alkaloid with an IC50 of 0.26 nM (Ki ¼ 0.087 nM, nH ¼ 72 h incubation period. These results indicated that further
1.9), or the most potent selective a4b2 modulators (e.g. sazeti- investigation into the metabolic stability using hepatic subcel-
dine-A, Ki ¼ 0.4–1 nM (ref. 33) or tebanicline, Ki ¼ 0.05 nM (ref. lular fractions or primary hepatocytes was needed.
11)), two main conclusions can be drawn. The potency of these
Results of the incubation of the compounds with rat liver
compounds is clearly signicantly lower: the Ki values are microsomes at a concentration near their IC50 (1 mM) are shown
separated by 3 logarithmic units. However, while potency is an in Fig. 5. The microsomes were harvested by the method
important parameter, it is not the foremost.34 In fact, several reported by Gorrod and coworkers,38 loperamide was included
nicotinic ligands displaying two- or three-digit nM affinity as a reference compound. As shown, compound 5b was iden-
(pozanicline Ki ¼ 16 nM; GTS-21 Ki ¼ 130 nM/250 nM; TC-6499 tied as a low turnover compound (t1/2 ¼ 175 min) with a pre-
Ki ¼ 27 nM; dianicline Ki ¼ 105 nM; rivanicline, Ki ¼ 26 nM; CP- dicted hepatic clearance of 6.5 ml minꢀ1 kgꢀ1
.
39 Compound 5c
601932 Ki ¼ 21 nM; PHA-568487 Ki ¼ 44 nM; CP-810123 Ki ¼ was identied as a high turnover compound (t1/2 < 2 min) with a
16 nM; SSR-180711 Ki ¼ 14–78 nM) have currently reached predicted hepatic clearance of more than 63 ml minꢀ1 kgꢀ1
.
39
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 2226–2234 | 2229

View Article Online
RSC Advances
Paper
Fig. 4 Concentration response curves of compounds 5b, 5c and (ꢂ)-epibatidine hydrochloride at the a4b2, a7 and muscle type nicotinic
receptors. Two measurements for each data point were performed. Data are shown as the mean ꢂ SEM. IC50 values and Hill coefficient were
obtained by non-linear regression analysis of the competition curves using Hill equation curve fitting to Y ¼ D + [(A – D)/(1 + (C/IC50
)
nH)] with Y ¼
specific binding, D ¼ minimum specific binding, A ¼ maximum specific binding, C ¼ compound concentration, and nH ¼ slope factor.
Also, plasma protein binding at 5 mM was determined to be as negative control, and an incubation of Verapamil was included
relatively low. For compound 5b 96% was determined as as positive control. As expected, a similar trend was observed. The
unbound and for compound 5c 88% was unbound. As such, results indicate that compound 5b is metabolically stable. The
these compounds are well available for metabolism.
metabolism of compound 5c on the other hand appears to be
A further evaluation of metabolic stability was made including concentration dependent and is faster than that of Verapamil. A
the full complement of hepatic metabolizing enzymes (phase I high rst pass extraction is to be expected.
and II) by using freshly isolated rat hepatocytes. Results of the
In conclusion, the cytotoxicity (at 72 h, 100 mM) of compound
incubation of compounds 5b and 5c with these hepatocytes are 5c may be attributed to its metabolism, whereas compound 5b
shown in Fig. 6. An incubation without hepatocytes was included remains stable and its cytotoxicity is more likely to be inherent to
Table 2 Cytotoxicity assay of compounds 5b and 5ca
MTS viability test (% viable cells)
LDH viability test (% viable cells)
Compound concentration (mM)
Compound concentration (mM)
100
30
10
3
1
0.3
0.1
100
30
10
3
1
0.3
0.1
B65 cell line
5b
5c
Tamoxifen
Terfenadine
106
100
11b
6
112
110
8
110
112
100
9
108
111
102
113
96
106
105
102
108
114
88
89
6b
ꢀ8
85
85
86
87
89
86
85
89
87
88
89
87
88
13
40
87
83
86
12
IMR32 cell line
5b
5c
Tamoxifen
Terfenadine
88
87
5b
87
91
21
24
91
92
35
66
90
92
94
89
94
105
89
99
101
98
93
88
11b
9
90
83
0
93
86
56
ꢀ6
85
82
74
ꢀ1
93
81
85
82
78
84
11
8
HepG2 cell line
5b
5c
Tamoxifen
Terfenadine
77
59
3b
4
92
88
2
90
94
90
2
111
98
100
71
102
87
98
99
101
82
99
100
14b
13
100
100
48
100
98
76
100
100
99
100
100
100
100
98
100
3
14
36
97
H4-II-E cell line
5b
5c
Tamoxifen
Terfenadine
74
70
8b
94
84
ꢀ3
9
99
105
55
101
118
89
100
109
103
108
95
112
99
98
22b
4
99
99
46
50
99
99
99
56
100
99
100
97
99
99
100
100
99
99
17
14
59
a
Duplicate wells were used per concentration, cells were seeded and allowed to adhere and equilibrate during 1 day. Aer removal of the medium,
compounds were added in 200 ml culture medium and incubation was continued for 72 h. Final DMSO concentration was 1%. b Above solubility in
media by eye and nephelometry.
2230 | RSC Adv., 2014, 4, 2226–2234
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
As such, we decided to evaluate these compounds in an in
vivo rat tail ick test as a measure for acute nociception effects.41
A high-intensity thermal stimulus was applied to the tail of
200–250 g male Sprague Dawley rats and the latency to reexive
withdrawal of the tail was measured (Fig. 7). Both compounds
were tested, considering their different pharmacological
proles. Compound 5c was tested at 0.01 to 1 mg kgꢀ1, while
compound 5b was tested at doses up to 10 mg kgꢀ1. Saline was
included as a negative control, morphine (10 mg kgꢀ1) and
epibatidine (4 mg kgꢀ1) were included as a positive control.
Unfortunately, the compounds 5b and 5c do not induce anal-
gesia using the tail ick test.
Compound 5b did not produce evident signs of toxicity.
Some aversive behaviour of the rats, which probably associate
the injection of the compound with the immediate immobili-
zation in the tube of the tail ick apparatus, was observed.
Compound 5c produced a small toxicity effect observed by
increased irritability and increase of defecation at the three
doses that have been tested. It induced a small increase of
sensitivity to pain at the highest dose of 1 mg kgꢀ1. Rats
survived the acute treatment with both compounds at all doses.
Fig.
5 Metabolic stability of compounds 5b and 5c in rat liver
microsomes. Compounds 5b and 5c were incubated at 1 mM in rat liver
microsomes, 1 mg mlꢀ1 microsomal protein after homogenisation, 37
ꢁC. Serial samples were removed up until 60 min. All experiments were
performed in triplicate, data are shown as the mean ꢂ SD.
the compound. These results are unexpected, because 5b contains
a metabolically sensitive benzylic alcohol, whereas 5c does not.
In vivo: tail ick antinociception
In vivo: blood/brain partitioning
Given the in vitro results, the two compounds in hands showed
selective binding to the a4b2 nicotinic acetylcholine receptor. Their
affinity constants are well below the cytotoxic concentrations and
both compounds differ largely as to their metabolic stability.40
The unexpected inactivity of these compounds in the in vivo
tests prompted us to evaluate the in vivo compound concen-
trations and the distribution between the test animals' brain
tissue and blood plasma. First, non-specic brain tissue
binding was determined at 5 mM. Due to limitations on the
dynamic range of the test measuring non-specic brain tissue
binding (0–30%), no exact values can be given.42 However, free
fractions were higher than 30% for both compounds. It appears
that the compounds are relatively unbound in brain tissue and
for further calculations a range of 30 to 70% free fraction will be
considered.
Next, in vivo plasma and brain concentrations of both
compounds were determined at 30, 60 and 120 minutes aer
intraperitoneal injection (IP) at the highest concentrations used
in the tail ick model (10 mg kgꢀ1 for 5b and 1 mg kgꢀ1 for 5c,
Table 3).
Compound 5b, dosed at 10 mg kgꢀ1 IP, is well absorbed from
the injection site into the plasma: the maximum mean plasma
concentration, observed at 30 minutes, amounted to 2038 ng
mlꢀ1. The corresponding compound concentration in the brain
is 152 ng gꢀ1. The compound is cleared from the brain and the
plasma at the same rate. Taking the non-specic brain tissue
binding into account, the free concentration in the brain is
calculated to be between 45 and 100 ng gꢀ1 (0.22 to 0.5 mM) aer
30 minutes (30–70% free). As such, levels in the brain are on the
lower side with respect to the administered dose and the brain/
plasma ratio is only 0.09. Presumably, the alcohol functionality
renders the compound too polar to pass the blood/brain barrier.
For compound 5c plasma levels are much lower, which
matches with the metabolic instability and clearance of this
compound. What gets into the plasma does make it to the brain,
and the brain/plasma ratios are signicantly higher than for 5b.
However, the overall levels are low relative to the dose. The
Fig. 6 Metabolic stability of compounds 5b and 5c in freshly isolated
rat hepatocytes. All analyses were performed in triplicate, calibration
curves showed an R2 > 0.96, data are shown as the mean ꢂ SD.
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 2226–2234 | 2231

View Article Online
RSC Advances
Paper
Fig. 7 Rat tail flick tests of compounds 5b and 5c. A radiant heat from a halogen lamp was applied directly to the tail and the latency to reflexive
withdrawal of the tail was measured. The heat intensity was adjusted such that the tail flick baseline latencies were 3.0–5.0 s and cut-off was
imposed at 15 s to prevent tissue damage. In the acute experiment, tail-flick latencies were recorded 30 min prior to IP drug injection and at 0, 30,
60, 90, 120 and 150 min after drug treatment. The percentage of maximal possible antinociceptive effect (%MPAE) was calculated by comparing
the test latency before (basal response) and after a drug administration using the equation: %MPAE ¼ [(test response time ꢀ basal response time)/
(cut off time ꢀ basal response time)] ꢃ 100. 14 to 16 rats were tested per dose for each compound. Data are shown as the mean ꢂ SEM.
Table 3 Plasma/brain concentrations of compounds 5b and 5ca
5b (10 mg kgꢀ1 IP)
5c (1 mg kgꢀ1 IP)
Time Plasma
(h)
Brain
Brain/plasma Plasma
ratio
Brain
Brain/plasma
ratio
concentration (ng mlꢀ1
)
concentration (ng gꢀ1
)
concentration (ng mlꢀ1
)
concentration (ng gꢀ1
)
0
0
0
—
0
0
—
0.5
1
2
2038 ꢂ 1606
667 ꢂ 609
386 ꢂ 208
2430 ꢂ 935
152 ꢂ 113
59 ꢂ 49
31 ꢂ 17
183 ꢂ 58
0.09 ꢂ 0.02
0.09 ꢂ 0.01
0.08 ꢂ 0.01
—
11 ꢂ 15
10 ꢂ 2
1.2 ꢂ 0.5
17 ꢂ 10
19 ꢂ 27
50 ꢂ 3
13 ꢂ 15
50 ꢂ 2
1.7 ꢂ 0.3
5.0 ꢂ 1.0
8.7 ꢂ 6.9
—
Cmax
a
Three rats were tested for each compound, data are shown as mean ꢂ SD. High SD on xed-time mean concentration can be attributed to different
response times between animals, an effect oen seen with IP dosing. Peak concentrations were approximately equal for all animals and show lower
variability (Cmax, see full data in ESI†).
maximum observed brain concentration is 50 ng gꢀ1 at 1 hour nicotinic acetylcholine receptor with IC50 values of 0.86 and
aer administration. Taking into account the free fraction, the 0.12 mM and were found to have favourable cytotoxicity proles
free concentration can be estimated as 15 to 35 ng gꢀ1 (0.07 to and a differing metabolic stability. The compounds did not
0.15 mM). Compound concentrations at the 30 minutes and 2 display antinociceptive effects in a rat tail ick test, presumably
hours time points are lower.
due to low concentrations of the free compound in the brain as
From these measurements, it is clear that the compound compared to their affinity constants (Ki). Compound 5b is
concentrations in the brain are about two- (5b) to fourfold (5c) absorbed readily into the blood plasma, but does not penetrate
higher that their respective Ki. It was shown on numerous the brain tissue. Compound 5c on the other hand has a
occasions that EC50 values of nicotinic agonists can be a factor favourable blood/brain distribution, but plasma concentrations
100 to 1000 above the actual Ki. Therefore, it is clear that the in (and consequently brain concentrations) remain too low due to
vivo compound concentrations are too low relative to the its fast metabolism. Despite the absence of in vivo anti-
observed receptor affinity to observe any antinociceptive effects. nociceptive effects, and the room for improvement with respect
Furthermore, antinociceptive effects are generally associated to affinity, metabolic stability and CNS penetration, we believe
with a4b2 agonism, an aspect which was not evaluated in this this compound class represents a new opportunity in the search
study. Further functional assays would be needed to determine for selective nicotinic acetylcholine receptor modulators.
the (partial) agonistic or antagonistic nature of these ligands.
Abbreviations
Conclusions
GABA g-Aminobutyric acid (receptor)
This study has delivered proof of concept on the use of
BZD
Benzodiazepine receptor
bridgehead substituted epibatidine analogues, thus removing CB1
Cannabinoid receptor type 1
Dopamine receptor D1
D1
chirality from the bicyclic core structure. Two of the evaluated
D2S
Dopamine receptor D2 short splice variant
compounds showed selective binding to the a4b2 subtype of the
2232 | RSC Adv., 2014, 4, 2226–2234
This journal is © The Royal Society of Chemistry 2014

View Article Online
Paper
RSC Advances
S. U. Kim and Y. B. Lee, Exp. Mol. Med., 2008, 40, 11–18;
(d) H. R. Parri and K. T. Dineley, Curr. Alzheimer Res., 2010,
7, 27–39; (e) D. K. Williams, J. Wang and R. L. Papke,
Biochem. Pharmacol., 2011, 82, 915–930.
M1
M2
M3
Muscarinic acetylcholine receptor type M1
Muscarinic acetylcholine receptor type M2
Muscarinic acetylcholine receptor type M3
nAChR Nicotinic acetylcholine receptor
6 (a) M. Fujita, M. Ichise, S. S. Zoghbi, J.-S. Liow, S. Ghose,
D. C. Vines, J. Sangare, J.-Q. Lu, V. L. Cropley, H. Iida,
K. M. Kim, R. M. Cohen, W. Bara-Jimenez, B. Ravina and
R. B. Innis, Ann. Neurol., 2006, 59, 174–177; (b) X. A. Perez
and M. Quik, Mol. Cell. Pharmacol., 2011, 3, 1–6; (c)
X. A. Perez, T. Bordia, J. M. McIntosh and M. Quik, Mol.
Pharmacol., 2010, 78, 971–980; (d) M. Quik, K. O'Leary and
C. M. Tanner, Mov. Disord., 2008, 23, 1641–1652.
7 (a) T. E. Wilens and M. W. Decker, Biochem. Pharmacol.,
2007, 74, 1212–1223; (b) Y. Ohmura, I. Tsutsui-Kimura and
M. Yoshioka, J. Pharmacol. Sci., 2012, 118, 413–422.
8 (a) M. D. Meyer, Drug Dev. Res., 2006, 67, 355–359; (b)
S. H. Sindrup, Pain, 2009, 146, 227–228.
5-HT
SEM
SD
Serotonin receptor
Standard error of the mean
Standard deviation
3-(4,5-Dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-
tetrazolium
MTS
LDH
%
MPAE
IP
Lactate dehydrogenase
Percentage of maximal possible antinociceptive effect
Intraperitoneal.
Acknowledgements
9 (a) W. J. de Jonge and L. Ulloa, Br. J. Pharmacol., 2007, 151,
915–929; (b) P. K. Chatterjee, M. M. Yeboah, O. Dowling,
X. Xue, S. R. Powell, Y. Al-Abed and C. N. Metz, PLoS One,
2012, 7, e35361.
Financial support for this research from the Fund for Scientic
Research Flanders (FWO Vlaanderen, T. Heugebaert), the
Research Fund Ghent University (BOF, Bijzonder Onderzoeks-
fonds Universiteit Gent, M. Van Overtveldt) and the IOF10/
StarTT/020 and IWT-SBO-100014 funds is gratefully acknowl-
edged. The authors gratefully acknowledge the CREATe ADME/
Tox and DSSc BA groups at Janssen Pharmaceuticals inc.
(Turnhoutseweg 30, B-2340 Beerse, Belgium) for their help
determining the blood/brain distribution of the investigated
compounds. Also the smooth collaboration with dr. C. Mackie,
Director Discovery ADME/Tox at J&JPRD was much appreciated.
The photograph of the Epipedobates tricolor poison dart frog was
gracefully provided by Mr B. De Waele and Mr M. Bellemans.
10 (a) R. J. Lukas, J.-P. Changeux, N. N. Le, E. X. Albuquerque,
D. J. K. Balfour, D. K. Berg, D. Bertrand, V. A. Chiappinelli,
P. B. S. Clarke, A. C. Collins, J. A. Dani, S. R. Grady,
K. J. Kellar, J. M. Lindstrom, M. J. Marks, M. Quik,
P. W. Taylor and S. Wonnacott, Pharmacol. Rev., 1999, 51,
397–401; (b) E. X. Albuquerque, E. F. R. Pereira,
M. Alkondon and S. W. Rogers, Physiol. Rev., 2009, 89, 73–
120; (c) J. A. Dani, Biol. Psychiatry, 2001, 49, 166–174; (d)
J. A. Dani and D. Bertrand, Annu. Rev. Pharmacol. Toxicol.,
2007, 47, 699–729.
11 R. Hurst, H. Rollema and D. Bertrand, Pharmacol. Ther.,
2013, 137, 22–54.
Notes and references
12 (a) E. Hagforsen, M. Edvinsson, K. Nordlind and
G. Michaelsson, Br. J. Dermatol., 2002, 146, 383–391; (b)
H. Kurzen, H. Berger, C. Jaeger, W. Hartschuh, H. Naeher,
A. Gratchev, S. Goerdt and M. Deichmann, J. Invest.
Dermatol., 2004, 123, 937–949; (c) J. M. Beckel, A. Kanai,
S.-J. Lee, G. W. C. de and L. A. Birder, Am. J. Physiol., 2006,
290, F103–F110; (d) C. R. Gwilt, L. E. Donnelly and
D. F. Rogers, Pharmacol. Ther., 2007, 115, 208–222; (e)
J. Arredondo, A. I. Chernyavsky, D. L. Jolkovsky,
K. E. Pinkerton and S. A. Grando, FASEB J., 2008, 22, 1356–
1368; (f) A. Wongsriraksa, M. E. Parsons and C. J. Whelan,
Inammation Res., 2009, 58, 38–44.
1 JAMA, J. Am. Med. Assoc., 1946, 132, 73.
2 J. R. Hughes, D. K. Hatsukami, J. E. Mitchell and
L. A. Dahlgren, Am. J. Psychiatry, 1986, 143, 993–997.
3 (a) R. Freedman, C. E. Adams and S. Leonard, J. Chem.
Neuroanat., 2000, 20, 299–306; (b) C. E. Adams and
K. E. Stevens, Front. Biosci., Landmark Ed., 2007, 12, 4755–
4772; (c) S. Kohnomi, K. Suemaru, H. Kawasaki, T. Choshi,
S. Hibino and H. Araki, Yakugaku Zasshi, 2009, 129, 197–
201; (d) M. Hajos and B. N. Rogers, Curr. Pharm. Des.,
2010, 16, 538–554; (e) S. C. Leiser, M. R. Bowlby,
T. A. Comery and J. Dunlop, Pharmacol. Ther., 2009, 122,
302–311.
13 A. Taly, P.-J. Corringer, D. Guedin, P. Lestage and
J.-P. Changeux, Nat. Rev. Drug Discovery, 2009, 8, 733–
750.
4 (a) N. S. Philip, L. L. Carpenter, A. R. Tyrka and L. H. Price,
Psychopharmacology, 2010, 212, 1–12; (b) N. S. Philip,
L. L. Carpenter, A. R. Tyrka and L. H. Price, Sci. World J., 14 (a) J. W. Daly, T. Tokuyama, T. Fujiwara, R. J. Highet and
2012, 104105–104111; (c) Y. S. Mineur, O. Somenzi and
M. R. Picciotto, Neuropharmacology, 2007, 52, 1256–1262.
5 (a) D. S. Woodruff-Pak, Drug Dev. Res., 2002, 56, 335–346; (b)
G. A. Bohme, S. R. Letchworth, O. Piot-Grosjean, G. J. Gatto,
I. L. Karle, J. Am. Chem. Soc., 1980, 102, 830–836; (b)
T. F. Spande, H. M. Garraffo, M. W. Edwards, H. J. C. Yeh,
L. Pannell and J. W. Daly, J. Am. Chem. Soc., 1992, 114,
3475–3478.
M.-C. Obinu, W. S. Caldwell, M. Laville, P. Brunel, R. Pellerin, 15 B. Badio and J. W. Daly, Mol. Pharmacol., 1994, 45, 563–569.
J.-P. Leconte, A. Genevois-Borella, P. Dubedat, M. Mazadier, 16 J. P. Sullivan, M. W. Decker, J. D. Brioni, D. Donnellyroberts,
L. Pradier, M. Bencherif and J. Benavides, Drug Dev. Res.,
2004, 62, 26–40; (c) J. H. Moon, S. Y. Kim, H. G. Lee,
D. J. Anderson, A. W. Bannon, C. H. Kang, P. Adams,
M. Piattonikaplan, M. J. Buckley, M. Gopalakrishnan,
This journal is © The Royal Society of Chemistry 2014
RSC Adv., 2014, 4, 2226–2234 | 2233

View Article Online
RSC Advances
Paper
M. Williams and S. P. Arneric, J. Pharmacol. Exp. Ther., 1994,
271, 624–631.
17 N. M. J. Rupniak, S. Patel, R. Marwood, J. Webb,
J. R. Traynor, J. Elliott, S. B. Freedman, S. R. Fletcher and
R. G. Hill, Br. J. Pharmacol., 1994, 113, 1487–1493.
W. M. Awni and M. D. Saltarelli, US20100144795A1, 2010.
Unfortunately, these assays were not available to us at the
time. As the compounds eventually proved to be inactive in
vivo, determination of a3b4 affinity was not persued.
28 J. R. Traynor, Br. J. Anaesth., 1998, 81, 69–76.
18 C. Z. Zhu, C.-l. Chin, N. R. Rustay, C. Zhong, J. Mikusa, 29 I. Carroll, S. Mackey and R. Gaeta, Semin. Anesth., Perioper.
P. Chandran, A. Salyers, E. Gomez, G. Simler, L. G. Lewis, Med. Pain, 2007, 26, 17–21.
D. Gauvin, S. Baker, M. Pai, A. Tovcimak, J. Brown, 30 B. K. Taylor, C. Joshi and H. Uppal, Brain Res., 2003, 987,
V. Komater, G. B. Fox, M. W. Decker, P. B. Jacobson, 135–143.
M. Gopalakrishnan, C.-H. Lee and P. Honore, Biochem. 31 R. G. Pertwee, Prog. Neurobiol., 2001, 63, 569–611.
Pharmacol., 2011, 82, 967–976. 32 P. L. I. Dellemijn and H. L. Fields, Pain, 1994, 57, 137–152.
19 D. Gundisch and C. Eibl, Expert Opin. Ther. Pat., 2011, 21, 33 Y. Xiao, H. Fan, J. L. Musachio, Z.-L. Wei, S. K. Chellappan,
1867–1896.
A. P. Kozikowski and K. J. Kellar, Mol. Pharmacol., 2006, 70,
1454–1460.
20 (a) J. Carreras, A. Avenoza, J. H. Busto and J. M. Peregrina,
J. Org. Chem., 2007, 72, 3112–3115; (b) A. Avenoza, J. H. Busto 34 M. M. Hann, MedChemComm, 2011, 2, 349–355.
and J. M. Peregrina, Tetrahedron, 2002, 58, 10167–10171; (c) 35 (a) M. Bencherif, M. E. Lovette, K. W. Fowler, S. Arrington,
Y. Z. Xu, J. Choi, M. I. Calaza, S. Turner and H. Rapoport,
J. Org. Chem., 1999, 64, 4069–4078; (d) A. B. Patel and
J. R. Malpass, J. Med. Chem., 2008, 51, 7005–7009.
L. Reeves, W. S. Caldwell and P. M. Lippiello, J. Pharmacol.
Exp. Ther., 1996, 279, 1413–1421; (b) M. I. Damaj,
W. Glassco, M. D. Aceto and B. R. Martin, J. Pharmacol.
Exp. Ther., 1999, 291, 390–398.
21 (a) T. Heugebaert, J. Van Hevele, W. Couck, V. Bruggeman,
S. Van der Jeught, K. Masschelein and C. V. Stevens, Eur. 36 N. J. Hewitt, M. J. G. Lechon, J. B. Houston, D. Hallifax,
J. Org. Chem., 2010, 1017–1020; (b) A. De Blieck and
C. V. Stevens, Synlett, 2011, 1748–1752; (c) C. V. Stevens,
T. Heugebaert and A. De Blieck, US Pat., US2010/0093807
A1, 2010; (d) C. V. Stevens, T. Heugebaert and A. De Blieck,
H. S. Brown, P. Maurel, J. G. Kenna, L. Gustavsson,
C. Lohmann, C. Skonberg, A. Guillouzo, G. Tuschl, A. P. Li,
E. LeCluyse, G. M. M. Groothuis and J. G. Hengstler, Drug
Metab. Rev., 2007, 39, 159–234.
WO 2011054885 A1, 2011; (e) C. V. Stevens and 37 T. L. Riss and R. A. Moravec, Assay Drug Dev. Technol., 2004,
T. Heugebaert, US Pat., US 7884125 B2, 2011. 2, 51–62.
22 M. J. Dart, J. T. Wasicak, K. B. Ryther, M. R. Schrimpf, 38 J. W. Gorrod, D. J. Temple and A. H. Beckett, Xenobiotica,
K. H. Kim, D. J. Anderson, J. P. Sullivan and M. D. Meyer,
Pharm. Acta Helv., 2000, 74, 115–123.
23 Computations were performed by means of the
Gaussian03W soware, using the DGT-B3LYP methodology
with the 6-31G basis set.
24 T. S. A. Heugebaert, B. I. Roman, A. De Blieck and
C. V. Stevens, Tetrahedron Lett., 2010, 51, 4189–4191.
25 S. Wagaw and S. L. Buchwald, J. Org. Chem., 1996, 61, 7240–
7241.
26 Neopentylic substitution of bromide at a highly similar
position has been reported to fail. See: J. R. Malpass,
1975, 5, 453–463.
39 Calculation of hepatic clearance rates was performed using
the Simplied Well-Stirred model. All calculations remain
uncorrected for binding to microsomal and plasma
proteins, full details on these calculations are available in
the ESI.†
40 The metabolic stability of 5c is quite low, which may be
considered as problematic for the subsequent in vivo
testing. However, for comparison to 5b, and to assess the
possibility of saturation of hepatic clearance at higher
doses, it was still taken into the tail ick test.
A. B. Patel, J. W. Davies and S. Y. Fulford, J. Org. Chem., 41 (a) D. Le Bars, M. Gozariu and S. W. Cadden, Pharmacol. Rev.,
2003, 68, 9348–9355.
27 To assess the full selectivity prole, a screening at the a3b4
2001, 53, 597–652; (b) K. Kawakita and M. Funakoshi,
Physiol. Behav., 1987, 39, 235–240.
would be obligatory, as a3 binding has been associated 42 Due to dilution factors, this test amplies measuring errors
with side effects such as nausea and emesis. See:
E. E. Bain, W. M. Abi-Saab, S. Dutta, T. S. Garimella,
in the range of 30–100%. See full protocol and calculations
in the ESI.†
2234 | RSC Adv., 2014, 4, 2226–2234
This journal is © The Royal Society of Chemistry 2014