A promising PET tracer for imaging of α7 nicotinic acetylcholine receptors in the brain: Design, synthesis, and in vivo evaluation of a dibenzothiophene-based radioligand

Article abstract of DOI:10.3390/molecules201018387

Changes in the expression of α7 nicotinic acetylcholine receptors (α7 nAChRs) in the human brain are widely assumed to be associated with neurological and neurooncological processes. Investigation of these receptors in vivo depends on the availability of imaging agents such as radioactively labelled ligands applicable in positron emission tomography (PET). We report on a series of new ligands for α7 nAChRs designed by the combination of dibenzothiophene dioxide as a novel hydrogen bond acceptor functionality with diazabicyclononane as an established cationic center. To assess the structure-activity relationship (SAR) of this new basic structure, we further modified the cationic center systematically by introduction of three different piperazine-based scaffolds. Based on in vitro binding affinity and selectivity, assessed by radioligand displacement studies at different rat and human nAChR subtypes and at the structurally related human 5-HT₃ receptor, we selected the compound 7-(1,4-diazabicyclo[3.2.2]nonan-4-yl)-2-fluorodibenzo-[b,d]thiophene 5,5-dioxide (10a) for radiolabeling and further evaluation in vivo. Radiosynthesis of [¹⁸F]10a was optimized and transferred to an automated module. Dynamic PET imaging studies with [¹⁸F]10a in piglets and a monkey demonstrated high uptake of radioactivity in the brain, followed by washout and target-region specific accumulation under baseline conditions. Kinetic analysis of [¹⁸F]10a in pig was performed using a two-tissue compartment model with arterial-derived input function. Our initial evaluation revealed that the dibenzothiophene-based PET radioligand [¹⁸F]10a ([¹⁸F]DBT-10) has high potential to provide clinically relevant information about the expression and availability of α7 nAChR in the brain.

Full text of DOI:10.3390/molecules201018387

Molecules 2015, 20, 18387-18421; doi:10.3390/molecules201018387
OPEN ACCESS
molecules
ISSN 1420-3049
www.mdpi.com/journal/molecules
Article
A Promising PET Tracer for Imaging of α
7
Nicotinic
Acetylcholine Receptors in the Brain: Design, Synthesis, and
in Vivo Evaluation of a Dibenzothiophene-Based Radioligand
Rodrigo Teodoro , Matthias Scheunemann , Winnie Deuther-Conrad 1,†,*, Barbara Wenzel 1,
1
,†
1,†
2
2
1
1
3
Francesca Maria Fasoli , Cecilia Gotti , Mathias Kranz , Cornelius K. Donat , Marianne Patt ,
4
4
5
1
3
Ansel Hillmer , Ming-Qiang Zheng , Dan Peters , Jörg Steinbach , Osama Sabri ,
4
1
Yiyun Huang and Peter Brust
1
Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiopharmaceutical Cancer Research,
Permoserstraße 15, Leipzig 04318, Germany; E-Mails: r.teodoro@hzdr.de (R.T.);
m.scheunemann@hzdr.de (M.S.); b.wenzel@hzdr.de (B.W.); m.kranz@hzdr.de (M.K.);
cdonat@nru.dk (C.K.D.); j.steinbach@hzdr.de (J.S.); p.brust@hzdr.de (P.B.)
Consiglio Nazionale delle Ricerche, Institute of Neuroscience, Biometra-Institute University of
Milan, Via Luigi Vanvitelli 32, Milano 20129, Italy; E-Mails: f.fasoli@in.cnr.it (F.M.F.);
c.gotti@in.cnr.it (C.G.)
2
3
4
Department of Nuclear Medicine, University Hospital Leipzig, Liebigstraße 18, Leipzig 04103,
Germany; E-Mails: marianne.patt@medizin.uni-leipzig.de (M.P.);
osama.sabri@medizin.uni-leipzig.de (O.S.)
PET Center, Yale University, P.O. Box 208048, 801 Howard Avenue, New Haven, CT 06520-8048,
USA; E-Mails: ansel.hillmer@yale.edu (A.H.); ming-qiang.zheng@yale.edu (M.-Q.Z.);
henry.huang@yale.edu (Y.H.)
5
Dan PET AB, Rosenstigen 7, Malmö SE-21619, Sweden; E-Mail: info@danpet.eu
†
These authors contributed equally to this work.
*
Author to whom correspondence should be addressed; E-Mail: w.deuther-conrad@hzdr.de;
Tel.: +49-341-234-179-4613; Fax: +49-341-234-179-4699.
Academic Editor: Svend Borup Jensen
Received: 15 July 2015 / Accepted: 28 September 2015 / Published: 9 October 2015
Abstract: Changes in the expression of α
7
nicotinic acetylcholine receptors (α nAChRs) in
7
the human brain are widely assumed to be associated with neurological and neurooncological
processes. Investigation of these receptors in vivo depends on the availability of imaging

Molecules 2015, 20
18388
agents such as radioactively labelled ligands applicable in positron emission tomography
PET). We report on a series of new ligands for α nAChRs designed by the combination of
(
7
dibenzothiophene dioxide as a novel hydrogen bond acceptor functionality with
diazabicyclononane as an established cationic center. To assess the structure-activity
relationship (SAR) of this new basic structure, we further modified the cationic center
systematically by introduction of three different piperazine-based scaffolds. Based on
in vitro binding affinity and selectivity, assessed by radioligand displacement studies at
different rat and human nAChR subtypes and at the structurally related human 5-HT
receptor, we selected the compound 7-(1,4-diazabicyclo[3.2.2]nonan-4-yl)-2-fluorodibenzo-
b,d]thiophene 5,5-dioxide (10a) for radiolabeling and further evaluation in vivo.
3
[
18
Radiosynthesis of [ F]10a was optimized and transferred to an automated module. Dynamic
18
PET imaging studies with [ F]10a in piglets and a monkey demonstrated high uptake of
radioactivity in the brain, followed by washout and target-region specific accumulation under
1
8
baseline conditions. Kinetic analysis of [ F]10a in pig was performed using a two-tissue
compartment model with arterial-derived input function. Our initial evaluation revealed that
1
8
18
the dibenzothiophene-based PET radioligand [ F]10a ([ F]DBT-10) has high potential to
provide clinically relevant information about the expression and availability of α nAChR in
7
the brain.
Keywords: α nAChR; pharmacophore; positron emission tomography; neuroimaging;
7
fluorine-18
1
. Introduction
The long-standing interest in molecular imaging of nicotinic acetylcholine receptors (nAChRs) is
driven by findings from preclinical and clinical studies, which have demonstrated that dysfunction of
neuronal nAChRs is involved in the pathophysiology of many disorders [1]. The nAChRs are a
superfamily of ligand-gated ion channels that consist of a pentamer of protein subunits. In the
mammalian brain, different subunits assemble with much diversity; however, the α
subtypes predominate [2]. The α subunit is highly expressed in the hippocampus and hypothalamus, as
well as in cell types where receptor-mediated ion currents have not been reported. Therefore,
nAChR-mediated effects such as cognitive enhancement [3] may depend on the ionotropic signaling
while other effects may be independent of ion-channel currents [4]. Altered functional availability of α
7
and α
4
β
2
nAChR
7
α
7
7
nAChR has been implicated in a number of diseases of the human central nervous system (CNS),
including Alzheimer’s and Parkinson’s disease, schizophrenia and autism, as well as in lung cancer and
heart disease [5]. With highly selective ligands, α nAChRs are approachable targets not only for
7
therapeutic interventions, but also for non-invasive imaging such as Positron Emission Tomography (PET),
which can be used to investigate disease pathophysiology in vivo and support drug development.
While the cationic pharmacophore sufficient for full activation of all neuronal nAChRs is the
tetramethylammonium cation, several α
7
nAChR selective motifs have been identified based on the
observation that an additional hydroxyl group present in choline or quinuclidinol activates largely the

Molecules 2015, 20
18389
homomeric α
high-affinity small-molecule ligands of the α
anabaseine-, pyrrolidine-, and diazabicyclononane-based PET radiotracers such as [ C]GTS-21 [8],
7
but not the heteromeric nAChRs [6]. So far, at least nine structurally related families of
7
nAChR have been characterized [6,7], including
1
1
[11
and [ F]NS14490 [13]. Recent progress in α
nAChR selective binding of the antiviral interferon inducer tilorone [14], containing a tricyclic
C]A-844606 and [ C]A-582941 [9], [ C]CHIBA-1001 [10], [ C]NS14492 [11], [ F]NS10743 [12]
11
11
11
18
1
8
7
nAChR PET imaging comes from the discovery of the
α
7
fluorenone nucleus. Further structural modification by replacing this fluorenone moiety with
dibenzothiophene sulfone as alternative hydrogen bond acceptor (HBA) functionality and introduction
of 1,4-diazabicyclo[3.2.2]nonane as cationic center resulted in a novel compound with markedly
increased binding affinity for α
7
nAChR (K = 56 nM for tilorone and 0.023 nM for compound 48,
i
Figure 1, A1) [15]. This discovery by Schrimpf et al., in 2012 prompted us to develop a series of novel
1
8
derivatives as references for F-labeled ligands of α nAChRs and to further investigate the SAR of this
7
new pharmacophore. At the time of working on the radiolabeling of the most promising ligand for PET
imaging studies [16], researchers from the Johns Hopkins University had successfully completed the
1
8
radiosynthetic work on respective ligands [17]. While the evaluation of the resulting [ F]ASEM in
baboons and humans was published [18–20], we felt encouraged by the inherent potential of the Abbott
lead structure to continue our research on 18F-labelled diazabicyclononane-containing α
nAChR
7
ligands [16,21] by modifying both the HBA functionality and the cationic center.
1
8
18
Besides the elaboration of a more effective approach towards compound [ F]10a ([ F]DBT-10),
1
8
which we have selected for further development [16] and is identical to compound [ F]7c reported by
Gao et al. [17], we have synthesized a series of derivatives of this lead structure to further analyze
structural elements that determine the affinity of potential α
ligand-gated ion channels such as heteromeric nicotinic or 5-HT
that the affinity for α nAChR of dibenzothiophene-diazabicyclononanes (e.g., [ F]ASEM,
= 562 nM [17]) is notably higher than that of oxadiazolyl-substituted diazabicyclononanes (e.g.,
F]NS10743 K > 10 μM [12]). Therefore, we replaced the cationic center with alternative tertiary
7
nAChR ligands to other pentameric
3
receptors. It has been shown previously
1
8
β
4 2
K
i
[18
i
amine motifs. Insertion of three different diazabicycloalkanes resulted in N-methyl-substituted
derivatives of the ethylene- and propylene-bridged piperazines diazabicyclo[3.2.1]octane and
diazabicyclo[3.3.1]nonane, respectively [22,23].
O
A
N
B
N
N
N
O
(
S)
NMe
O
O
A1
A3
N
H
Tropisetron
3
H
S
N
C
O
NMe
O
N N
(
R)
H
HN
N
A2
A4
NMe
Monosubstituted dibenzothiophene sulf ones
NS12857
Figure 1. Structures of recent α nAChR ligands. Four examples of bicyclic amines coupled
7
to dibenzothiophene sulfone (A) [15], a tropane derived compound (B) [24] and a
propylene-bridged N-methylpiperazine (C) [25] are shown.

Molecules 2015, 20
18390
Previous reports have also shown that structural changes on the amino side chain (cationic center) of
a given aromatic unit may affect the binding profile with respect to α /α nAChR affinity [26,27].
7
β
4 2
Furthermore, a small series of bicyclic amines coupled to dibenzothiophene sulfone, including chiral
quinuclidines (Figure 1, A3 and A4) and a fused azetidine (3,6-diazabicyclo[3.2.0]heptane) (Figure 1,A2)
have been investigated by Schrimpf et al. [15]. In addition to nitrogen bridgeheaded aza- or diazabicylic
molecules derived from quinuclidine or diazabicyclo[3.2.2]nonane, several alkylene bridged piperidines
(
e.g., tropane derivatives such as tropisetron [24]) or piperazines (e.g., N-methyl-3-azagranatanine
derivative such as NS12857 [25]) have been identified as potent α nAChR selective ligands (Figure 1B,C).
7
Considering this, we were inspired to prepare a small series of fluorinated dibenzothiophene sulfones
with a set of (three) differently arranged piperazine-based diazabicyclic scaffolds.
To test for the effect of the designed structural modifications and to evaluate the target selectivity of
potential α
7
nAChR imaging probes, the binding affinities of all fluoro-substituted derivatives for the
nAChR, as well as the three most important subtypes of heteromeric nAChRs (human α
, and rat α *) and the structurally similar 5-HT receptor, were determined. For the most
human α
7
4
2
β ,
human α
3
β
4
6
β
2
3
suitable ligand 10a, a radiolabeling procedure was developed and successfully implemented, then
subsequently optimized and transferred to a fully automated setup. Finally, the pharmacokinetics of
[18
F]10a was evaluated in both piglets and monkey by dynamic PET scans under baseline and
blocking conditions.
2
2
2
. Results and Discussion
.1. Chemistry
.1.1. Rationale
1,4-Diazabicyclo[3.2.2]nonane-substituted dibenzothiophene sulfone 10 was found to have exceptional
high affinity for α nAChR and effective brain uptake [15,17]. This prompted us to develop novel
7
derivatives as candidates for 18F-labeled ligands and to further investigate the SAR of this new
pharmacophore (Table 1).
Initially, the three fluoro-substituted 1,4-diazabicyclo[3.2.2]nonane-containing derivatives 10a–c
were designed. While 10a and 10b are identical to compounds 7c and 7a, respectively, recently
published by Gao et al. [17], the isomer 10c has not been described so far. To gain further insight into
the SAR of dibenzothiophene-based scaffolds we embarked on the synthesis of a second series of
compounds. Previous results on the systematic side chain rigidification have highlighted the importance
of conformational restriction of the aminocycle for high α selectivity [15]. Since the tropane core has
7
been identified as a structural motif that could provide this selectivity [4,6] it was of interest to include
a set of three tropane-like diamines into the synthetic plan. With the carbon atom replaced at position 3
both in tropane and in N-methylgranatanine [28] by an nitrogen atom, the resulting secondary amino
group of each N-methyl piperazine building block could serve as position of attachment to the tricyclic
arene system.

Molecules 2015, 20
Table 1. Structures of fluoro-substituted dibenzothiophene-based ligands for α
18391
7
nAChR.
Compound
R
X1 X2 X3
1
0
H
H
F
H
F
H
H
H
F
1
0a
0b
1
H
H
F
1
0c
2a
H
H
F
1
H
H
H
H
1
1
1
2b
3b
4b
H
H
H
F
F
Four novel fluorine-substituted derivatives were envisioned, in which the homopiperazine-based
,4-diazabicyclo[3.2.2]nonane was replaced by conformationally restricted piperazine-based
1
substituents, bridged by either ethylene (8-methyl-3,8-diazabicyclo[3.2.1]octane = azatropane (13b), 3-
methyl-3,8-diazabicyclo[3.2.1]octane (14b)) or propylene units (9-methyl-3,9-diazabicyclo[3.3.1]nonane =
N-methyl-3-azagranatanine (12a,b)) [22].
2
.1.2. Synthesis of α
7
nAChR Ligands
Dibenzo[b,d]thiophene (1a) and dibenzo[b,d]thiophene 4-boronic acid (1b) served as starting
materials for the synthesis of dibenzo[b,d]thiophene-5,5-dioxides functionalized in the para- (Scheme 1)
or ortho- position (Scheme 2) to the sulfone group. The final coupling to the N-atom of a diazabicyclic
moiety was performed at position 7 or 3 of the respective tricyclic scaffold to afford the potential α
7
nAChR ligands 10, 10a–c, 12a,b, 13b, and 14b, as shown in Table 1.

Molecules 2015, 20
18392
Scheme 1. Synthesis of 3-bromo dibenzothiophene intermediates 3, 6a, 8a and 8c.
Reagents and conditions: (a) H (50%), HOAc, 120 °C, 5 h, 97% for 2 and 96% for 5a
, r.t., 24 h, 44% for 3, 64% for 6a, 46% for 8c, and 18% for 8d;
(65%), HOAc, 20–40 °C, 24 h, 26% for 4a; (d) TMAF·4 H O, cyclohexane/DMSO,
↓, 6 h (azeotropic drying with separation of H O), compound 5a or compound 6a, 95 °C,
2
O
2
(
b) NBS, conc. H
2
SO
4
(
c) HNO
3
2
↑
2
5
h, 75% for 7a and 74% for 8a.
Scheme 2. Synthesis of 3-bromo dibenzothiophene intermediates 6b and 8b. Reagents and
conditions: (a) tert-butyl nitrite, MeCN, 55 °C, 24 h, mixture of 4/4b (~88:12); (b) H
2
O
2
(
50%), HOAc, 120 °C, 5 h, 72% (two steps); (c) NBS, conc. H
2
SO
4
, r.t., 24 h, 78%;
O),
(d) TMAF·4H O, cyclohexane/DMSO, 6 h (azeotropic drying with separation of H
2
2
compound 6b, 95 °C, 5 h, 73%.
Synthesis of Bromo Dibenzothiophene Intermediates 3, 6a and 8a–8c
The brominated intermediates were obtained from dibenzo[b,d]thiophene (1a) by three different routes.
The 3-bromo-derivative 3 was prepared via sulfide to sulfone oxidation of 1a with H in acetic acid
followed by bromination with an excess of 2 in the NBS/H SO system in 44% yield [29]. Isolation of
2
O
2
2
4
compound 3 from the 3,7-dibrominated by-product (not shown) along with unreacted starting material
was obtained by fractional crystallization of the dibromo compound, followed by a chromatographic

Molecules 2015, 20
18393
separation of pure 3 from the filtrate. The 2-nitrodibenzothiophene sulfone 5a was prepared by nitration
of 1a using HNO (65%) in acetic acid [30,31] affording the 2-nitro compound 4a in moderate yield
26%), followed by oxidation with H in acetic acid to give 5a [32]. The 2-nitro-7-bromo-
dibenzothiophene sulfone 6a was obtained by bromination of 5a according to the corresponding
conversion of 2 to 3. Because the one benzo ring in 5a is strongly deactivated by the NO and SO
3
(
O
2 2
2
2
R
groups, no over-brominated by-products were detected and the 2-nitro-7-bromo-dibenzothiophene
sulfone 6a was afforded in 64% yield. To obtain different mono-fluorinated building blocks in a short
synthesis, the fluorine substitution either at positions para or ortho to the sulfone were intended to be
carried out with a simple uniform synthetic pathway. In addition to the known three-step conversion of
6
a into the 2-fluoro-7-bromo derivative 8a based on the Balz-Schiemann reaction with 45% overall
yield [17], we deemed a one-step conversion from 6a to 8a via fluorodenitration as preferable [33,34].
Tetramethylammonium fluoride (TMAF) proved to be a good source for active fluoride after azeotropic
drying of commercial TMAF tetrahydrate in refluxing cyclohexane/DMSO for 6 h. A smooth conversion
of 6a occurred at 95 °C within 5 h to give 8a in 75%yield.Likewise, the nitro group in the
2
3
-nitrodibenzothiophene 5,5-dioxide 5a was displaced by fluorine to give 7a in 74% yield. The vicinal
-bromination of 7a was accomplished by applying NBS in H SO (98%) at room temperature for 24 h
2
4
to give compound 8b. In addition to 8c, the 3,7-dibromo derivative 8d was detected. Chromatographic
separation gave 8d in 18% yield along with 2-fluoro-3-bromo compound 8c in 46% yield.
Synthesis of the 3-Bromo Dibenzothiophene Intermediates 6b and 8b
As result of the two-step synthesis of 5a from 1a (Scheme 1), the ortho-derivative 4b was
concomitantly formed as a minor nitration product. However, because it was difficult to separate from
the excess of 4a, this approach was considered inappropriate to gain an adequate amount of the 4-nitro
dibenzothiophene sulfone 5b. The intermediate 5b was recently reported [17]. The described procedure
for 4b is a nitration of commercially available arylboronic acid 1b with two equivalents of Bi(NO
3
3
) as
described by Maiti et al. [35]. In contrast, we successfully applied a metal-free approach by performing
an ipso nitrosation of 1b with tert-butyl nitrite in acetonitrile [36,37]. As shown in Scheme 2, the nitroso
derivative 4 was formed as the main product along with a minor amount of 4b. Without further
purification, the 4/4b mixture was readily oxidized in a one-pot reaction with H
2
O
2
/acetic acid to give
b in 72% yield over two steps from 1b. After bromination, the resulting 6b was converted directly into
b in 73% yield via our fluorodenitration protocol, in contrast to the earlier described independent
5
8
four-step synthesis applying a Pschorr reaction [17].
Synthesis of 10, Fluorinated Reference Compounds 10a–c, and Nitro Precursor 11
The 3-bromodibenzo[b,d]thiophene sulfones 3, 6a and 8a–c were reacted with 1,4-diazabicyclo
[
3.2.2]nonane (9a) under Pd-catalyzed Buchwald-Hartwig conditions to provide the previously reported
compounds 10a and 10b [17] and the novel fluorinated isomer 10c in 52%, 48% and 44% yields,
respectively. The non-fluorinated compound 10 [15] and the nitro derivative 11a [17] were also obtained
under Buchwald-Hartwig conditions in 73% and 53% yields, respectively, as depicted in Scheme 3.

Molecules 2015, 20
18394
Scheme 3. Synthesis of 10, fluorinated reference compounds 10a–c, and nitro precursor 11.
Reagents and conditions: (a) Pd (dba) , BINAP, Cs CO , toluene, 90 °C, 24–36 h, 52% for
0a, 48% for 10b, 44% for 10c, 73% for 10, and 53% for 11a.
2
3
2
3
1
Synthesis of the Fluorinated Reference Compounds 12a,b, 13b, and 14b
In order to expand the structural diversity of compounds derived from the novel pharmacophore
dibenzothiophene for SAR characterization, in a second series of compounds the homopiperazine-based
1
,4-diazabicyclo[3.2.2]nonane moiety was replaced by piperazine-based substituents, bridged by either
ethylene or propylene units. The bromine substituent in the meta position to the sulfone group
was used to incorporate these more rigid diamines. The Buchwald-Hartwig coupling of
3
-bromodibenzo[b,d]thiophene sulfones 8a and 8b with 9-methyl-3,9-diazabicyclo[3.3.1]nonane (9b),
-methyl-3,8-diazabicyclo[3.2.1]octane (9c) and 3-methyl-3,8-diazabicyclo[3.2.1]octane (9d) gave
8
compounds 12a,b, 13b, and 14b in moderate yields (Scheme 4).
Scheme 4. Synthesis of the fluorinated reference compounds 12a,b, 13b, and 14b.
Reagents and conditions: (a) Pd
2
(dba)
3
, BINAP, Cs
2
CO , toluene, 90 °C, 24–36 h, 38% for
3
1
2a, 42% for 12b, 47% for 13b, and 67% for 14b.
All final compounds (10–14) are crystalline solids and were fully characterized by NMR spectroscopy
(1
H, C, F, COSY, HSQC) and high resolution mass spectrometry.
13
19
2
.2. In Vitro Affinity Assays
All compounds were evaluated in vitro to measure their affinity and selectivity for the target receptor
α
7
nAChR in relation to the heteromeric receptor subtypes α
4
β
2
, α
6
β
2
, and α
3
4
β . The results of the
respective binding assays are summarized in Table 2.

Molecules 2015, 20
18395
Table 2. In vitro binding affinities towards human homomeric α
7
, heteromeric α
4
β
2
and α
β
3 4
nAChR, rat α * nAChR subtypes, and human 5-HT receptor.
6
β
2
3
Affinity (K
i
in nM)
Selectivity (K
i
ratio)
d,e
Compound
nAChR Subtype
5-HT
3
hα a
hα
β
2
b
hα
β
4
b
rα
β
2
* c
α
7
/α
4
β
2
α
7
/α
3
β
4
7
4
3
6
1
0
0.51 ± 0.32 318 ± 43.3 49.6 ± 14.7
0.60 ± 0.44 517 ± 375 119 ± 29.0
0.84 ± 0.16 211 ± 108 42.3 ± 4.73
8.53 ± 1.74 507 ± 212
30.9 ± 8.72 141 ± 11.7 96.0 ± 1.48 1180 ± 360
105 ± 23.9 301 ± 148 94.8 ± 6.48 1190 ± 230
40.9 ± 7.77 426 ± 197 224 ± 51.7 2260 ± 540
9.26 ± 2.23 >4000 >5000 1450 ± 370
517 ± 186
589 ± 217
435 ± 152
(35%)
440 (2%)
(42%)
(26%)
(10%)
(11%)
(10%)
(4%)
623
862
251
59
5
97
1
0a
0b
198
50
33
3
1
1
0c
2a
279 ± 24.4 1390 ± 340
1
1
1
1
2b
3b
4b
3
1
10
5
>400
>500
a
3
7
Human α nAChR in stably transfected SH-SY5Y cells, with radiotracer [ H]methyllycaconitine (0.5–1 nM),
K
[3
D
= 2.0 nM. b Human α
H]epibatidine (0.5–1 nM), K
4
β
2
and α
3
β
4
nAChR in stably transfected HEK-293 cells, with radiotracer
c
D
= 0.025 nM for hα
4
β2 nAChR, K
D
= 0.117 nM for hα
3
β
4
nAChR. Rat α
6
β
2
*
obtained from rat striatum by immunoimmobilization using anti-rα
6
nAChR antibody, with radiotracer
receptor recombinant-HEK293 cells,
[3
H]epibatidine (0.1 nM), K
= 0.025 nM. K
d
value in nM; human 5-HT
D
i
3
3
e
with radiotracer [ H]GR65630 (working concentration n = 0.69 nM; K
D
= 0.2 nM). percentage of inhibition
at 0.1 μM concentration of test compound.
2
.2.1. Affinity for α
7
nAChR
The affinity for α
7
nAChR was assessed for the human receptor protein expressed in a stably
3
transfected cell line and labeled with the selective α
7
nAChR antagonist [ H]methyllycaconitine.
= 0.51 nM, is almost identical to the value
The affinity of the lead of the series (compound 10), K
i
1
25
reported by Gao et al., obtained on [ I]α-bungarotoxin labelled rat cortical membranes. The binding
affinities of the fluoro-substituted compounds 10a (K = 0.6 nM vs. 1.4 nM of compound 7c [17]) and
0b (K = 0.84 nM vs. 0.4 nM of compound 7a [17]) are also comparable with previously published
results. All novel fluoro-containing compounds 10c, 12a,b, 13b, and 14b bind with remarkably lower
affinity to α nAChR. Interestingly, the vicinal substitution of the dibenzothiophene with both the
fluorine and the diazabicycle is inappropriate in terms of binding as reflected by the 2-fluoro-3-amino
derivative 10c, which possesses an about 10-fold lower affinity at the α subtype relative to the
-fluoro-7-amino derivative 10a and the 6-fluoro-3-amino derivative 10b.
In particular, we assume that steric effects impair the interactions between the cationic center of
compound 10c and the binding site of α nAChR [38]. Furthermore, as noticed for the new series of
i
1
i
7
7
2
7
dibenzothiophene derivatives (12a,b, 13b and 14b) an increase in the flexibility of the tertiary amine in
the cationic center is also not tolerated. Replacement of the NC-bridged homopiperazine moiety,
containing a bicyclic tertiary amine as a basic structural element of the reference compound 10 by three
different CC-bridged piperazine ring systems and carrying a methyl-substituted tertiary amino group,
has negative effects. The 9-methyl-3,9-diazabicyclo[3.3.1]nonane substituted compounds 12a and 12b
have significantly diminished affinity for α
7
nAChR relative to the matched pairs of
1
,4-diazabicyclo[3.2.2]nonane-substituted 10a and 10b. A similar effect is observed by substitution with

Molecules 2015, 20
18396
an azatropane in compound 13b (8-methyl-3,8-diazabicyclo[3.2.1]octane) and its isomer 14b
3-methyl-3,8-diazabicyclo[3.2.1]octane).
Thus, in terms of affinity for the α
(
7
nAChR we considered compound 10a as the most suitable
compound reported herein. The structural modifications within the cationic center of the molecules,
which we performed to increase selectivity by reducing off-target binding, unexpectedly impaired the
binding to α
7
nAChR to such an extent that the resulting compounds were not included in the
development of radioligands within this study.
2
.2.2. Affinities for α
4
β
2
, α
3
β
4
, and α
6
β nAChR
2
To evaluate off-target binding of the compounds towards heteromeric nAChR subtypes, respective
3
binding assays were performed with [ H]epibatidine-labeled human α
4
β
2
or human α
3
β
4
nAChR, both
6 2
stably expressed on human HEK cells, as well as native rat α β * nAChR immobilized by a
subunit-specific antibody.
Besides the α
nAChRs [39]. While functionally similar to α
selective, primarily in dopamine neurons in the brain [40–43]. Overlapping expression of α
with α * nAChR and also α * nAChR within the mesolimbic axis [44,45] requires careful analysis
4
β
2
* nAChR, the α
6
β
2
* nAChR belongs to the high-affinity family of β
* nAChR, the expression of the α * subtype is rather
nAChR
2
-containing
4
β
2
β
6 2
7
4
β
2
β
6 2
of subtype affinity of radioligands for nAChR. In heteromeric nAChR subtypes, the orthosteric ligand
binding site is formed at the interface of an α subunit (principal component) and an adjacent non-α
subunit (complementary component), equivalent to the (+) surface and the (−) surface of an α subunit in
homomeric nAChRs [46,47].
Consistent with the observation that in heteromeric neuronal nAChRs the non-α subunit is of
particular importance [38], as it makes the affinity for nicotine dependent on the presence of a β
regardless of the α subunit [47], the herein investigated compounds bind with generally moderate
values > 100 nM) and almost equal affinities towards both human α and rat α * nAChRs. It is
worth noting that the affinities for the α subtype of the α nAChR ligands investigated herein, with
2
subunit
(
K
i
4
β
2
β
6 2
4
β
2
7
the exception of 14b, are significantly higher than those of the compounds in the NeuroSearch (NS)
series [12]. Because these compounds contain identical cationic centers, differences in the receptor
subtype binding affinities are probably related to differences in the HBA and hydrophobic
functionalities. In the NeuroSearch compound series, an oxadiazole moiety appears as HBA which is
spatially more separated from and not forced in plane with the hydrophobic fluorophenyl moiety, while
compounds 10a–10c possess fused functionalities with both the HBA and the hydrophobic moiety
represented by the fluorine-substituted dibenzothiophene dioxide ring system. We hypothesize that in
particular the sulfonyl moiety of these novel compounds promotes binding to the α
comparable to the carbonyl moiety acting as a hydrogen acceptor along with a cationic pharmacophore
element in the off-target binding of α nAChR ligand CHIBA-1001 [48,49]. However, for the most
selective fluorine-containing ligand of the current study, compound 10a (selectivity > 800), interfering
effects on α nAChR PET imaging due to its binding to β -containing receptor subtypes are highly
4
β
2
subtype,
7
7
2
unlikely. The conformational changes due to the shift of the bridgehead carbons in the
diazabicyclo[3.2.1]octane that we assume to contribute to the significantly improved selectivity of
compound 14b in comparison to 13b will be analyzed in future studies. It appears that the basic N-methyl

Molecules 2015, 20
18397
group is sterically more deshielded in contrast to the azatropane 13b, resulting in a detrimental effect
both on α and α nAChR binding with only weakly attenuated affinity to the α subtype (Table 2).
An overlap in the receptor expression between α and α * nAChRs, in particular within autonomic
neurons [50,51], necessitates the investigation of the selectivity of potential ligands. By comparing data
obtained for compounds 10a and 10b (K : 119 nM and 43 nM) with the affinity values of the
corresponding compounds 7c and 7a reported by Gao et al., (K : 5000 nM and 709 nM) [17],
species-specific differences become obvious. Although the most selective compound 10a (selectivity
200) binds with sufficiently low affinity to human α nAChRs to afford imaging of the α subtype in
4
β
2
3
β
4
7
7
β
3 4
i
i
~
3
β
4
7
human brain, more than 10-fold higher affinities at the human than the rat receptor subtype illustrate the
significance of species-specific targeting at the early stage of the development of PET ligands.
The in vitro binding affinity of all compounds from this series at the 5-HT
3
receptor, as determined by
nAChR over
3
percentage of inhibition of the binding of [ H]GR65630 at 10 μM, indicates high α
7
5
-HT
of 10a for 5-HT
ASEM). Based on this value, compound 10a is assumed to possess an at least fivefold higher selectivity
than all other compounds, except 14b, reported herein. The K of 10a was estimated to be 440 nM
for 5-HT
In addition, investigation of interaction of 10a at 1 μM with further off-targets (α
3
selectivity (Table 2). From this direct comparison a much higher affinity of ASEM (10b) than that
3
may be expected considering the inhibition data at 100 nM (2% for 10a, 42% for
i
3
.
1
nAChR, SERT, DAT,
NET, VMAT, and choline transporter) by radioligand binding assays revealed no significant binding.
Altogether, the target affinity and nAChR subtype specificity of 10a, the most suitable
fluorine-substituted α
nAChR in humans will not be confounded by binding of the respective PET radiotracer to relevant
off-target sites. Therefore, compound 10a was selected for radiolabeling and further evaluation in vivo.
7
nAChR ligand reported herein, is certainly sufficient to ensure that assessment of
α
7
1
8
2
.3. Optimization of Manual Radiosynthesis of [ F]10a for Translation to an Automated Module
1
8
Manual radiosynthesis of [ F]10a was performed starting from the nitro precursor 11a and
systematically optimized by varying the base, solvent, reaction time and heating system (Scheme 5).
The amount of nitro precursor 11a was kept constant at 1 mg according to our previous practice [52].
A nitro-to-fluoro substitution of activated aryl moieties in general requires moderately high temperatures
(e.g., 130–160 °C) [53], reflected by low labeling efficiencies of <1% at 90 °C using MeCN. With
increasing temperature (150 °C) and the use of DMSO, labeling efficiency of up to 30% were achieved.
1
8
Scheme 5. Radiosynthesis of [ F]10a. Solvents tested: MeCN, DMSO, and DMF at
different temperatures with microwave irradiation or conventional heating.
1
8
In contrast to decomposition of the corresponding [ F]7c reported by Gao et al. [17], we did not
1
8
observe decomposition of [ F]10a in the presence of the strong base potassium carbonate in DMSO at

Molecules 2015, 20
18398
this temperature. In an attempt to improve labelling efficiencies prior to high-scale production of
[18
F]10a in a commercially available automated device, we investigated the aromatic radiofluorination
of 11a using DMF as solvent based on our recent experience [52] under different conditions (thermal vs.
microwave heating). Influence of the basicity of the metal salt on the labelling yield (potassium
carbonate vs. potassium oxalate) was also investigated.
Under thermal heating and the use of K
increase in the nitro-to-fluoro conversion with maximum labelling efficiencies of about 90% after
0 min of reaction time at 120 °C. Under these conditions, K CO is preferred, since we observed a
significantly lower labelling efficiency of ~30% using the weaker base K . Our attempt to combine
2
CO , we determined a time- and temperature-dependent
3
1
2
3
C
2 2
O
4
microwave dielectric heating with DMF as solvent with a mean-to-high dielectric constant rendered
comparably high labeling efficiencies (≈94%) under microwave pulse mode (power cycling mode,
1
50 °C, 75 W, 5 min).
A semi-automated radiosynthesis coupled with microwave was recently reported for [ F]10b
1
8
1
8
(
[ F]ASEM) using the DMSO/K
2
CO
3
system at 160 °C, resulting in a labeling efficiency of
4
5%–50% [54]. In the attempt to gain insight into the influence of the solvent on labeling efficiencies,
1
8
we radiolabeled the nitro precursor 11b to obtain [ F]10b using the DMF/K
2
CO system under
3
conventional and microwave-assisted heating at 120 °C. With only 1 mg of the nitro precursor 11b,
labeling efficiencies of 60%–70% were achieved with DMF as solvent in both heating modes. These
findings reinforce the superiority of DMF as a solvent of choice for this compound class, and allows the
1
8
18
fully automated radiosynthesis of [ F]10a and [ F]10b for transfer to clinical radiopharmacies.
Overall, the radiolabelling reactions proceeded cleanly. No significant amount of 18F-labeled
by-products was detected according to radio-TLC analyses (data not shown). The crude reaction mixture
1
8
was applied directly onto a semi-preparative HPLC column (Figure 2a) and [ F]10a was successfully
isolated in high radiochemical purities (≥98%) with a retention time of about 17 min. Analytical
radio-HPLC analysis of the final product spiked with the reference compound confirmed the identity of
[18
F]10a (Figure 2b). [ F]10a proved to be stable in physiological solutions and organic solvents at
18
4
0 °C for up to 90 min.
Figure 2. Cont.

Molecules 2015, 20
18399
Figure 2. (a) Representative radio- and UV-chromatograms obtained for isolation of
[18
F]10a from crude reaction mixture by semi-preparative HPLC (Reprosil-Pur C18-AQ
−1
column, 35% MeCN/H O/0.05% TFA, Flow rate: 10 mL·min ); (b) Analytical
2
1
8
radio-chromatogram (top) and UV-chromatogram (bottom) of purified [ F]10a spiked with
the reference compound 10a.
1
8
We determined the distribution coefficient of [ F]10a in the n-octanol-PBS system experimentally
by the shake-flask method with a log D7.2 value of 1.3 ± 0.1 (n = 3), making sufficient blood-brain barrier
1
8
permeability of [ F]10a likely.
1
8
The translation to remotely controlled radiosynthesis of [ F]10a was performed using the
®
TRACERLAB™ FXFN module. The reaction was performed using the Kryptofix 222/K
2
CO system at
3
1
8
1
20 °C in 10 min with 1 mg of precursor 11a in DMF. Also in the automated process [ F]10a was
obtained with extremely high labeling efficiencies of 86% ± 3%. After isolation via semi-preparative
1
8
®
HPLC, [ F]10a was trapped on a pre-conditioned Sep-Pak (Waters, Milford, MA, USA) C18 light
cartridge, and formulated in isotonic saline containing 10% of EtOH (v/v). The average decay-corrected
radiochemical yield was 35% ± 9% (n = 6) calculated at the end of the synthesis (EOS). Radiochemical
−1
purity of >99% and high specific activities of 855 ± 302 GBq·μmol (n = 6) were obtained in a total
synthesis time of 70 min. This rapid and versatile automated radiosynthesis will enhance the accessibility
1
8
of [ F]10a for widespread production in future clinical studies.
1
8
2
.4. Dynamic PET Studies of [ F]10a in Piglets
2
.4.1. Baseline Studies
1
8
After intravenous injection of 425 ± 78 MBq [ F]10a (n = 3), we observed high uptake of
radioactivity in the brain with peak concentrations at 8–10 min p.i., followed by washout. The highest
accumulation of radioactivity (SUVmax > 2.2) occurred in the thalamus, colliculi, and midbrain.
Somewhat lower accumulation (SUVmax ~1.9–2.1) was observed in other brain regions (Figure 3).
The summed PET image (inset of Figure 3) shows the corresponding distribution pattern in the brain of
a female piglet from 0 to 20 min p.i.

Molecules 2015, 20
18400
Figure 3. Time-activity curves in different brain regions obtained by dynamic PET scans of
[18
F]10a in piglets under baseline and blocking conditions (mean values; n = 3).
Standard deviations are shown for two examples (hippocampus baseline and cerebellum
block). *** p < 0.001 vs. baseline. Inset: Brain image in sagittal plane acquired from 0 to
1
8
2
0 min p.i. after injection of [ F]10a in a female piglet. Data are expressed as standardized
uptake value (SUV). Regions of interest (ROI) were drawn based on an overlay with
T1-weighted MR images of a pig brain. Abbreviations: CB = Cerebellum, CC = Corpus
callosum, St = Striatum, Th = Thalamus.
2
.4.2. Metabolite Analysis
18
Radioactive metabolites of [ F]10a were assessed in plasma of pigs for up to 120 min p.i. (Figure 4a).
For preparation of RP-HPLC samples, the proteins were precipitated and extracted two times with MeCN
with a reproducible recovery of >90% of the starting radioactivity in the supernatant. The patterns of
radioactive metabolites obtained by radio-TLC and radio-HPLC correlated well with each other. Parent
fraction accounted for 88% ± 6%, 24% ± 4%, and 19% ± 4% of total radioactivity at 2, 60, and 120 min
p.i., respectively.
1
8
One major radioactive metabolite was detected, which was more hydrophilic than [ F]10a and is
therefore assumed not to pass the blood-brain barrier (Figure 4b). Based on previous experience from
1
8
our group with the diazabicyclononane-containing α nAChR ligand [ F]NS10743 [21], we believe that
7
this metabolite is formed by enzymatic oxidation of the nitrogen at position 1 in the identical
motif of 10a.

Molecules 2015, 20
18401
(
a)
(b)
1
8
Figure 4. (a) Tri-exponential function fitting of the percentage of [ F]10a under baseline
conditions (n = 3) which was used to obtain metabolite-corrected arterial plasma input
functions for modelling; (b) Representative radio-chromatogram of a pig plasma sample
1
8
obtained at 30 min after intravenous injection of [ F]10a under baseline conditions.
2
.4.3. Blocking Studies
1
8
To evaluate the binding specificity of [ F]10a in vivo, we performed blocking studies (n = 3).
−
1
Animals were pretreated with 3 mg·kg of the α
7
nAChR ligand NS6740 [3] at 10 min prior to injection
1
8
−1 −1
of 429 ± 45 MBq [ F]10a, followed by a continuous infusion of NS6740 at 1 mg·kg ·h during the
course of the PET scan [21].
Despite a significant increase of about 40%–50%, in the brain uptake, radioactivity uptake under
blocking conditions was about 15%–20% lower than that under baseline conditions in all brain regions
evaluated from about 90 min p.i. until the end of the study (Figure 3). The early increase of brain uptake
is accompanied by a significant increase of the influx rate constant K
1
(Table 4). The most likely
nAChRs are expressed on
explanation is a blood flow driven effect. There is strong evidence that α
7
vascular endothelial and smooth muscle cells [55–57], which we assume were activated by NS6740
despite its classification as a “weak agonist” [3].
2
.4.4. Modeling
One-tissue and two-tissue compartment models (1TCM and 2TCM) were evaluated for
1
8
mathematical description of [ F]10a time-activity curves (TACs) in pig brain regions (Figure 3) using
metabolite-corrected arterial plasma input functions. The 2TCM was identified as the more appropriate
model based on its smaller Akaike information criterion (AIC) value. The four rate constants K
and k were fitted and the non-displaceable volume of distribution (VND = K /k ) and the binding potential
/k ) as defined by Innis et al. [58] were calculated for each region of interest (Tables 3 and 4).
The values of the influx rate constant K and the clearance rate constant k were comparable between
the different brain regions with a mean of 0.362 mL·cm ·min and 0.131 mL·cm ·min , respectively, in
the whole brain. The total volume of distribution V calculated from K /k (1 + k /k
) [58] was 17 mL·cm−3
in thalamus and 13 mL·cm−3 in cerebellum, i.e., slightly lower than the comparable values for
[18
F]ASEM in baboons [18].
1
, k
2
, k
3
,
4
1
2
(
BPND = k
3
4
1
2
−3
−1
−3
−1
T
1
2
3
4

Molecules 2015, 20
18402
Table 3. Kinetic rate constants, non-displaceable volume of distribution (VND), and
1
8
non-displaceable binding potential (BPND) of [ F]10a in different regions of piglet brain
using 240 min of scan data under baseline conditions (n = 3).
Rate Constant
Brain Region
V
ND
BPND
−
3
−1
(min−1)
(min−1)
k
(min−1)
K
1
(mL·cm ·min )
0.362 ± 0.038
0.387 ± 0.059
0.399 ± 0.038
0.395 ± 0.053
0.400 ± 0.023
0.382 ± 0.028
0.448 ± 0.027
0.450 ± 0.072
0.436 ± 0.045
0.440 ± 0.045
0.421 ± 0.048
k
2
k
3
4
Whole
Frontal Cortex
Parietal Cortex
Occipital Cortex
Hippocampus
Striatum
0.131 ± 0.062 0.093 ± 0.074 0.027 ± 0.018 3.32 ± 1.74 3.31 ± 1.02
0.182 ± 0.153 0.142 ± 0.165 0.027 ± 0.015 3.32 ± 2.06 4.22 ± 2.93
0.140 ± 0.089 0.126 ± 0.135 0.031 ± 0.021 3.88 ± 2.49 3.47 ± 1.65
0.123 ± 0.063 0.099 ± 0.087 0.030 ± 0.023 4.07 ± 2.23 3.16 ± 1.11
0.188 ± 0.083 0.145 ± 0.101 0.025 ± 0.013 2.43 ± 0.96 5.50 ± 1.01
0.144 ± 0.079 0.114 ± 0.093 0.027 ± 0.014 3.35 ± 1.81 3.98 ± 1.40
0.150 ± 0.071 0.111 ± 0.085 0.030 ± 0.018 3.67 ± 2.10 3.51 ± 1.13
0.134 ± 0.074 0.070 ± 0.042 0.027 ± 0.018 4.19 ± 2.37 2.68 ± 1.42
0.162 ± 0.117 0.121 ± 0.132 0.029 ± 0.022 4.04 ± 3.06 3.40 ± 1.73
0.149 ± 0.059 0.068 ± 0.042 0.026 ± 0.019 3.32 ± 1.40 2.76 ± 0.81
0.131 ± 0.052 0.067 ± 0.051 0.026 ± 0.020 3.62 ± 1.47 2.53 ± 0.57
Thalamus
Colliculi
Midbrain
Pons
Cerebellum
Table 4. Kinetic rate constants, non-displaceable volume of distribution (VND), and
1
8
non-displaceable binding potential (BPND) of [ F]10a in different regions of piglet brain
using 240 min of scan data under blocking conditions (n = 3).
Rate Constant
Brain Region
V
ND
BPND
−
3
−1
(min−1)
(min−1)
k
(min−1)
K
1
(mL·cm ·min )
0.598 ± 0.108 *
0.647 ± 0.117 *
0.659 ± 0.120 *
0.668 ± 0.103 *
0.603 ± 0.118 *
0.624 ± 0.143 *
0.762 ± 0.129 *
0.735 ± 0.182 *
0.685 ± 0.131 *
0.670 ± 0.092 *
0.711 ± 0.148 *
k
2
k
3
4
Whole
Frontal Cortex
Parietal Cortex
Occipital Cortex
Hippocampus
Striatum
0.062 ± 0.015 0.018 ± 0.022 0.011 ± 0.018 10.27 ± 3.94 * 0.84 ± 1.07 *
0.055 ± 0.011 0.014 ± 0.019 0.010 ± 0.020 12.34 ± 4.27 * 0.86 ± 0.92
0.055 ± 0.013 0.015 ± 0.020 0.011 ± 0.021 12.58 ± 4.52 * 0.82 ± 0.88 *
0.064 ± 0.019 0.022 ± 0.032 0.013 ± 0.023 11.36 ± 4.41 * 0.95 ± 1.03 *
0.063 ± 0.023 * 0.024 ± 0.030 0.011 ± 0.015 10.81 ± 5.02 * 0.84 ± 1.70 *
0.057 ± 0.015 0.020 ± 0.025 0.011 ± 0.017 11.78 ± 4.91 * 0.75 ± 1.29 *
0.070 ± 0.025 0.023 ± 0.031 0.012 ± 0.021 12.09 ± 5.31 * 0.88 ± 1.00 *
0.066 ± 0.007 0.011 ± 0.013 0.007 ± 0.015 11.41 ± 3.90 * 0.73 ± 0.74 *
0.063 ± 0.010 0.013 ± 0.014 0.012 ± 0.018 11.17 ± 3.64 * 0.61 ± 0.77 *
0.084 ± 0.030 0.019 ± 0.023 0.012 ± 0.016 8.93 ± 3.96 * 0.76 ± 1.16 *
0.076 ± 0.014 0.016 ± 0.019 0.011 ± 0.016 9.75 ± 3.57 * 0.71 ± 1.10 *
Thalamus
Colliculi
Midbrain
Pons
Cerebellum
*
p < 0.05 vs. baseline (Table 3).
Under blocking conditions K
was decreased between 42% and 70% resulting in a significant (up to 4-fold) increase of VND
Table 4) and preventing the use of the occupancy plot [59] for calculation of BPND. Therefore BPND, as
1
was significantly increased by 50% to 70% in all brain regions while
k
2
(
the ratio (at equilibrium) of specifically bound radioligand to that of nondisplaceable radioligand in
tissue [58], was directly calculated form the rate constants k3, which is proportional to the association
rate constant, and k
compartment. The baseline study revealed that the rate constant k
lowest in the cerebellum and the values of k were low, rather uniform and ranged from 0.025 to
.031 min . BPND values between 5.5 (hippocampus) and 2.5 (cerebellum) were reliably calculated from
4
, which is proportional to the dissociation rate constant from the specific
3
was highest in the hippocampus and
4
−1
0

Molecules 2015, 20
18403
these data (SDwhole brain ~30%). A similarly high BPND has recently been reported in baboons for the
1
8
structurally related [ F]ASEM [18] but not for any previous α
blocking conditions k was decreased by 70% to 90% in various brain regions while k
between 60% and 70%, resulting in a significant decrease of BPND by 72% to 85% in all regions but the
7
nAChR PET radioligands [60]. Under
3
4
was decreased
1
8
frontal cortex (Table 4), suggesting specific binding of [ F]10a to α
7
nAChR in a similar range as the
1
8
recently reported [ F]ASEM [18].
18
18
18
18
2
.5. Comparative PET Study of [ F]10a ([ F]DBT-10) and [ F]10b ([ F]ASEM) in a Rhesus Monkey
1
8
A PET study in a single rhesus monkey was performed for direct comparison of [ F]DBT-10
1
8
(
injected dose: 167 MBq) and [ F]ASEM (injected dose: 185 MBq). Regional time-activity curves for
both radioligands in the monkey brain are presented in Figure 5. Initial uptake levels were lower, and
1
8
18
kinetics slower for [ F]DBT-10 than [ F]ASEM. For both radioligands the uptake levels follow the
order of thalamus > frontal cortex = putamen > caudate > hippocampus > occipital cortex > cerebellum.
The total volume of distribution V for both radioligands was similar with the highest values in the
T
1
8
−3 18
−3
thalamus ([ F]DBT-10: 48.1 mL·cm ; [ F]ASEM: 46.7 mL·cm ) and the lowest in the cerebellum
1
8
−3 18
−3
(
[ F]DBT-10: 28.5 mL·cm ; [ F]ASEM: 26.8 mL·cm ). These values are almost two-fold higher
1
8
than those reported for [ F]ASEM in baboons [18], supporting the species differences in α nAChR
7
binding and distribution.
1
8
18
Altogether these preliminary data support the equal potency of both [ F]DBT-10 and [ F]ASEM for
1
8
PET imaging of α nAChRs, with [ F]ASEM having slightly faster kinetics.
7
1
8
18
Figure 5. Time-activity curves of [ F]DBT-10 and [ F]ASEM in selected brain regions of
a rhesus monkey.
2
.6. Toxicity Studies in Rats
1
8
To prepare [ F]10a for use in humans, toxicity studies were performed in rats according to EU cGLP.
Compound 10a was administered by a single intravenous injection to rats followed by an observation
period of 2 or 15 day. The study was performed with 4 test groups, including 1 control and 3 dose groups
−
1
(
(
6.2, 62 and 620 μg
·kg ), with 60 male and 60 female Wistar rats divided into two experiments: Day 2
40 males and 40 females) and Day 15 (20 males and 20 females). During clinical observation the
animals displayed no notable clinical effects. No statistically significant differences in body weights

Molecules 2015, 20
18404
between control and treated groups in either gender were detected. Food consumption corresponded with
body weight development.
All haematology parameters on Day 2 and Day 15 were within physiological range for this species.
Individual divergences of some haematology parameters were small and not correlated with treatment.
No test item effect on haematology parameters was observed.
There were no findings in clinical chemistry parameters which could be definitively considered as
adverse. The average values of all test groups were within the historical control ranges. Occasional changes
had no dose relationship, and they were therefore considered as a result of intra- and inter-individual
variability for this species. The results of pathology examination indicated that 10a after single
intravenous administration did not result in changes in pathological and histopathological parameters on
−1
either Day 2 or Day 15. The no observed effect level (NOEL) of 10a was determined to be 620 μg·kg .
3. Experimental Section
3
.1. General Information
Analytical thin-layer chromatography (TLC) was performed with Macherey-Nagel precoated plastic
®
sheets with fluorescent indicator UV254 (Polygram SIL G/UV254, Düren, Germany). Visualization of
the spots was effected by illumination with an UV lamp (254 nm and 366 nm). Dry-column flash
chromatography (DCFC) [61] was performed with vacuum on silica gel 60 (particle size 15–40 μm,
Ref. 815650) from Macherey-Nagel (Macherey-Nagel GmbH & Co. KG, Düren, Germany). NMR
1
13
13
19
spectra ( H, C, C-APT, F, COSY, HSQC, HMBC) were recorded with Varian spectrometer (Varian
Mercury-300BB and Mercury-400BB; Agilent Technologies, Santa Clara, CA, USA). Chemical shifts
are reported as δ (δ
, δ
H C
, δ
F
) values. Coupling constants are reported in Hz. Multiplicity is defined by s
(singlet), d (doublet), t (triplet), and combinations thereof; br (broad) and m (multiplet). ESI/Ion trap
mass spectra (LRMS) were recorded with a Bruker Esquire 3000 plus instrument (Billerica, MA, USA).
High resolution mass spectra were recorded on an FT-ICR APEX II spectrometer (Bruker Daltonics;
Bruker Corporation, Billerica, MA, USA) using electrospray ionization (ESI) in positive ion mode.
Melting points were determined on a Linström capillary apparatus (Wagner & Munz GmbH, Vienna,
Austria) in open capillary tubes and are uncorrected. Reagents and solvents were purchased from
commercial sources and used without further purification unless otherwise noted. Compounds 2, 4a, and
5
a were prepared starting from 1a (dibenzo[b,d]thiophene; Acros Organics, Geel, Belgium) according
to literature methods [30–32]. The preparation of intermediates 3, 5b, 8a, 8b, and the final compound
0 have been described previously, but with different synthetic routes or from starting materials not used
1
in our protocols [17,31,62]. Intermediate 6a (two steps from 4a), 6b (one step from 5b) and final
compounds 10a, 10b, 11a and 11b (one step each, from amine 9a), which have been prepared according
to the literature [62], were also reported by Gao et al. [17]. To the best of our knowledge, intermediates
7
3
4
a, 8c, 8d and final compounds 10c, 12a,b, 13b, and 14b have not been reported so far.
.2. Chemistry
-Nitrosodibenzo[b,d]thiophene (4). tert-Butyl nitrite (2.14 mL, 1.86 g, 18 mmol) was added in one
portion to a stirred suspension of dibenzo[b,d]thiophen-4-yl boronic acid (1b, 1.37 g, 6 mmol; Frontier

Molecules 2015, 20
18405
Scientific, Logan, UT, USA) in MeCN (24 mL) under argon at 22 °C. The mixture was stirred at 50–55 °C
while its colour turned from yellow to dark brown in the course of 24 h. The solvent was evaporated and
the residual solid was dissolved in CH
evaporated to leave a dark greenish brown residue (1.38 g), which was used for the oxidation step. A
sample (40 mg) was dissolved in cyclohexane/CH Cl (2:3, v/v, 3 mL) and filtered through a short plug
of silica gel (15–40 μm, 2 g). Subsequent elution with cyclohexane/CH Cl (2:3, v/v, 50 mL) gave a
= 0.56, heptane/EtOAc,
): δ
2
Cl
2
(80 mL), washed with H
2
O (25 mL), dried (MgSO
4
) and
2
2
2
2
green eluate which was evaporated to yield 30 mg of the title compound 4 (R
f
1
3
7
9
:1) as a green powder, m.p. 113–115 °C (lit. m.p. 115–117 °C [36]). H-NMR (300 MHz, CDCl
3
.48–7.63 (m, 2H, 7-H, 8-H), 7.88–7.94 (m, 1H, 6-H), 7.94 (t, J = 7.7 Hz, 1H, 2-H), 8.13–8.22 (m, 1H,
1
3
-H), 8.49 (dd, J = 7.7, 1.1 Hz, 1H, 3-H), 9.58 (dd, J = 7.6, 1.1 Hz, 1H, 1-H). C-NMR (75 MHz,
): δ 120.10 (C), 121.68 (CH), 123.67 (CH), 125.44 (CH), 125.84 (CH), 127.82 (CH), 128.21 (CH),
32.38 (C), 137.94 (C), 138.11 (C), 142.04 (C), 161.91 (C), 165.95 (C). The product is 88% pure,
CDCl
3
1
1
containing 12% of 4-nitrodibenzo[b,d]thiophene (4b), as determined by comparison with H-NMR of a
pure sample.
2
-Nitrodibenzo[b,d]thiophene 5,5-dioxide (5a) [32]. A mixture of 2-nitrodibenzo[b,d]thiophene (4a, 2.3
g, 10 mmol) in acetic acid (64 mL) was stirred at 80 °C while H (50%, 6.8 mL, 0.12 mol) was added.
The temperature was raised to 120 °C and after 1 h a second portion H (50%, 4.5 mL, 0.08 mol) was
2
O
2
2 2
O
added. Stirring was continued at 120 °C for 1 h and at 80 °C for 3 h. After cooling, the mixture was poured
into water (180 mL) to form a precipitate which was filtered, washed and dried to yield the pure title
compound 5a (2.52 g, 9.64 mmol, 96% yield, R
f
= 0.22, cyclohexane/CHCl
3
, 1:3) as a pale yellow
) δ 7.74 (td,
1
powder, m.p. 256.5–257.5 °C (lit. m.p. 258 °C [32]); H-NMR (400 MHz, DMSO-d
6
H
J = 7.6, 0.9 Hz, 1HAr, 7-H), 7.87 (td, J = 7.6, 1.0 Hz, 1HAr, 8-H), 8.07 (d, J = 7.7 Hz, 1HAr, 6-H), 8.29
(
m, 1HAr, 9-H), 8.42 (dd, J = 8.4, 2.0 Hz, 1HAr, 3-H), 8.48 (d, J = 7.8 Hz, 1HAr, 4-H), 9.04 (d, J = 2.0 Hz,
1
3
1
1
HAr, 1-H). C-NMR (100 MHz, DMSO-d
6
): δ
C
118.26 (CH), 122.28 (CH), 123.60 (CH), 123.95 (CH),
26.10 (CH), 129.10 (C), 132.09 (CH), 132.98 (C), 135.04 (CH), 137.08 (C), 141.44 (C), 151.72 (C).
LRMS: m/z (ESI) = 284.0 (M + Na)+.
4
-Nitrodibenzo[b,d]thiophene 5,5-dioxide (5b). In a procedure similar to the preparation of 5a,
compound 4 (1.33 g, 5.8 mmol) was reacted with H (6.6 mL, 0.11 g, 116 mmol) in acetic acid to give
.33 g of a dark yellow solid. The product was further purified by column chromatogaphy (silica gel
2
O
2
1
1
7
5–40 μm, 24 g) with cyclohexane/CH
2
Cl
2
(1:1→1:2) to yield the pure title compound 5b (1.1 g, 4.2 mmol,
1
f
3
2% yield, R = 0.2, CHCl ) as pale yellow crystals, m.p. 245.5–251 °C. H-NMR (400 MHz, DMSO-d
6
):
δ
8
0
1
4
H
7.75 (td, J = 7.6, 0.8 Hz, 1HAr, 7-H), 7.86 (td, J = 7.6, 1.0 Hz, 1HAr, 8-H), 8.04 (d, J = 7.7 Hz, 1HAr),
.08 (t, J = 8.0 Hz, 1HAr), 8.29 (d, J = 7.7 Hz, 1HAr), 8.39 (dd, J = 8.2, 0.7 Hz, 1HAr), 8.65 (dd, J = 7.7,
13
.7 Hz, 1HAr). C-NMR (100 MHz, DMSO-d
6
): δ 122.19 (CH), 123.21 (CH), 125.91 (CH), 128.29 (C),
C
29.16 (CH), 130.58 (C), 132.18 (CH), 134.45 (C), 134.71 (CH), 136.21 (CH), 137.26 (C), 143.39 (C,
-C). LRMS: m/z (ESI) = 284.0 (M + Na)+.
7-Bromo-2-nitrodibenzo[b,d]thiophene 5,5-dioxide (6a). Compound 5a (1.75 g, 6.7 mmol) was dissolved
in H SO
2 4
(96%, 30 mL). N-Bromosuccinimide (1.28 g, 7.2 mmol) was added to this solution in several
portions and the mixture was stirred at room temperature for 24 h. The suspension was poured into

Molecules 2015, 20
18406
ice-water (90 mL) and stirred for 10 min. The precipitate was filtered off, washed with H
2
O until neutral,
followed by MeOH, and dried at 60 °C to obtain a yellowish solid (2.29 g). The raw material was
recrystallized twice from MeCN to afford the pure title compound 6a (1.45 g, 4.27 mmol, 63.8% yield,
1
R
f
= 0.30, cyclohexane/CHCl
DMSO-d ): δ
-H, 3-H, 6-H), 9.07 (d, J = 1.8 Hz, 1HAr, 1-H). C-NMR (100 MHz, DMSO-d
3
, 1:4) as a pale yellow powder, m.p. 281.0–282.5 °C. H-NMR (400 MHz,
6
H
8.08 (dd, J = 8.4, 1.8 Hz, 1HAr, 8-H), 8.31 (d, J = 8.4 Hz, 1HAr, 4-H), 8.41–8.47 (m, 3HAr
,
1
3
9
6
): δ
C
118.52 (CH), 123.72
(
(
CH), 125.11 (C), 125.36 (CH), 125.85 (CH), 126.35 (CH), 128.41 (C), 132.19 (C), 137.90 (CH), 138.66
C), 141.11 (C), 151.78 (C). LRMS: m/z (ESI) = 364.0, 362.0 (M + Na)+.
7
-Bromo-2-fluorodibenzo[b,d]thiophene 5,5-dioxide (8a). TMAF tetrahydrate (0.22 g, 1.3 mmol, Acros)
was dried under an atmosphere of argon by azeotropic distillation with a mixture of DMSO (6 mL) and
cyclohexane (12 mL) using a water separator for 6 h (bath temperature: 115–120 °C). The bath
temperature was allowed to cool to 80 °C and the nitro derivative 6a (0.34 g, 1 mmol) was added under
stirring to the suspension of the dried TMAF in one portion. The mixture immediately turned dark brown
and was stirred for additional 5 h at 95 °C. Reaction progress was monitored by TLC
(
cyclohexane/CHCl
water (60 mL) and extracted with CH
The extracts were combined and washed with water (10 mL), dried (MgSO
residual yellow solid (0.283 g) was purified by column chromatography (silica gel 15–40 μm, 6 g) with
3
, 1:4). The portion of cyclohexane was evaporated. The residual oil was poured into
2
Cl (4 × 15mL).
2
4
) and evaporated. The
CH
2
Cl
2
to afford the pure title compound 8a, (0.23 g, 0.74 mmol, 73.7% yield, R
f
= 0.39,
):
1
cyclohexane/CHCl
3
, 1:4) as a colorless powder, m.p. 266.5–268.5 °C. H-NMR (300 MHz, DMSO-d
6
δ
H
7.52 (dt-like, J = 8.8, 8.7, 2.4 Hz, 1HAr, 3-H), 8.05 (dd, J = 8.3, 1.8 Hz, 1HAr, 8-H), 8.11 (dd,
J = 8.6, 4.9 Hz, 1HAr, 4-H), 8.16 (d, J = 8.3 Hz, 1HAr, 9-H), 8.18 (dd, J = 9.2, 2.3 Hz, 1HAr, 1-H), 8.35
1
3
): δ
110.74 (d, 2JCF = 25.6 Hz, CH), 118.22
(
(
d, J = 1.6 Hz, 1HAr, 6-H); C-NMR (75 MHz, DMSO-d
d, 2JCF = 24.1 Hz, CH), 124.53 (s, C), 124.89 (d, 3JCF = 10.2 Hz, CH), 125.09 (s, CH), 125.16 (s, CH),
6
C
1
1
2
28.98 (d, 4JCF = 2.4 Hz, C), 132.82 (d, 4JCF = 3.0 Hz, C), 133.45 (d, 3JCF = 10.7 Hz, C), 137.49 (s, CH),
39.31 (s, C), 165.85 (d, 1JCF = 252.2 Hz, C); F-NMR (282.36 MHz, DMSO-d
19
): δ −103.8 (m, FAr,
6
F
-F). LRMS: m/z (ESI) = 334.9, 336.9 (M + Na)+.
2-Fluorodibenzo[b,d]thiophene 5,5-dioxide (7a). Similar to the preparation of 8a, TMAF tetrahydrate
(
1.35 g, 8.18 mmol) was azeotropically dried with cyclohexane/DMSO and subsequently reacted with
a (1.525 g, 5.84 mmol). After work-up, the yellow solid (1.25 g) was purified by column
chromatography (silica gel 15–40 μm, 18 g) with cyclohexane/CH Cl (1:1→1:3) as eluent to yield the
= 0.4, cyclohexane/CHCl , 1:4) as a colorless
7.21 (dt, J = 8.45, 8.45, 2.27 Hz, 1HAr, H-3),
5
2
2
pure title compound 7a (1.03 g, 4.4 mmol, 75.2% yield, R
f
3
1
powder, m.p. 227–228 °C. H-NMR (400 MHz, CDCl
3
): δ
H
7
7
.45 (dd, J = 8.37, 2.26 Hz, 1HAr, H-1), 7.57 (dt, J = 7.56, 7.53, 1.09 Hz, 1HAr, H-7), 7.66 (dt, J = 7.65,
.58, 1.17 Hz, 1HAr, H-8), 7.75 (ddd, J = 7.70, 1.01, 0.57 Hz, 1HAr, 9-H), 7.82 (dd, J = 8.42, 4.92 Hz, 1HAr
,
13
H-4), 7.83 (ddd, J = 7.60, 1.09, 0.69 Hz, 1HAr, 6-H); C-NMR (100 MHz, CDCl
3
): δ
C
109.34 (d,
2JCF = 24.6 Hz, CH), 117.61 (d, 2JCF = 23.9 Hz, CH), 121.98 (s, CH), 122.40 (s, CH), 124.58 (d,
3JCF = 9.9 Hz, CH), 130.47 (d, 4JCF = 2.5 Hz, C), 131.23 (s, CH), 133.72 (d, 4JCF = 3.2 Hz, C), 134.13 (s,
CH), 134.92 (d, 3JCF = 9.7 Hz, C), 138.68 (s, C), 166.37 (d, 1JCF = 255.0 Hz, C); F-NMR (282.36 MHz,
19
DMSO-d
6
): δ
ppm −103.63 (m. 1F, 2-F); LRMS: m/z (ESI) = 257.0, 258.0 (M + Na)+.
F

Molecules 2015, 20
18407
3
-Bromo-2-fluorodibenzo[b,d]thiophene 5,5-dioxide (8c) and 3,7-dibromo-2-fluoro-dibenzo[b,d]-
thiophene 5,5-dioxide (8d). Similar to the preparation of compound 6a, compound 7a (0.56 g, 2.4 mmol)
was reacted with NBS (0.43 g, 2.4 mmol) in H SO . The solid residue obtained upon work-up (0.75 g)
consisted of a mixture of starting material (R = 0.32), mono- (R = 0.43) and dibrominated derivative
= 0.56) as shown by TLC analysis (cyclohexane/CHCl , 1:4). The mixture was subjected to
chromatographic purification (silica gel 15–40 μm, 28 g) with cyclohexane/CHCl (1:3) to yield the
-fluoro-3,7-dibromo derivative 8d (0.17 g, 0.43 mmol, 18%) from the first fraction and the 2-fluoro-3-
2
4
f
f
(
R
f
3
3
2
bromo derivative 8c (0.35 g, 1.12 mmol, 46.5%) from the second fraction. Compound 8c: m.p.
1
2
8
8
1
69–274.5 °C. H-NMR (400 MHz, DMSO-d
6
): δ 7.72 (t, J = 7.6 Hz, 1HAr, 7-H), 7.84 (t, J = 7.6 Hz, 1HAr,
H
-H), 8.01 (d, J = 7.7 Hz, 1HAr, 6-H), 8.22 (d, J = 7.7 Hz, 1HAr, 9-H), 8.32 (d, J = 8.9 Hz, 1HAr, 1-H),
1
3
): δ
110.78 (d, 2JCF = 23.4 Hz, C),
.55 (d, J = 6.2 Hz, 1HAr, 4-H). C-NMR (100 MHz, DMSO-d
11.57 (d, 2JCF = 26.5 Hz, CH), 122.07 (s, CH), 123.33 (s, CH), 127.71 (d, 3JCF = 1.5 Hz, CH), 129.16 (d,
6
C
4JCF = 2.1 Hz, C), 131.80 (s, CH), 133.30 (d, 3JCF = 9.9 Hz, C), 134.06 (d, 4JCF = 3.5 Hz, C), 134.77 (s,
CH), 137.28 (s, C), 162.00 (d, 1JCF = 251.8 Hz, C). F-NMR (376.4 MHz, DMSO-d
19
): δ −98.68 (m, 1
6
F
+
F
Ar, 2-F). LRMS: m/z (ESI) = 336.9, 335.0 (M + Na) . Compound 8d: m.p. 282.5–285 °C.
1H-NMR (300 MHz, CDCl
J = 8.2, 1.8 Hz, 1HAr, 8-H), 7.94 (d, J = 1.7 Hz, 1HAr, 6-H), 8.01(d, J = 6.0 Hz, 1HAr, 4-H).
13C-NMR (75 MHz, CDCl 110.07 (d, 2JCF = 25.8 Hz, CH), 111.88 (d, 2JCF = 12.2 Hz, C), 123.28
): δ
s, CH), 125.53 (s, C), 125.84 (s, CH), 128.11 (d, 3JCF = 1.9 Hz, CH), 128.75 (s, C), 132.68 (d,
3JCF = 8.8 Hz, C), 134.32 (d, 4JCF = 3.8 Hz, C), 137.41 (s, CH), 139.67 (s, C), 162.85 (d, J = 256.4 Hz, C).
): δ 7.49 (d, J = 7.8 Hz, 1HAr, 1-H), 7.60 (d, J = 8.2 Hz, 1HAr, 9-H), 7.79 (dd,
3
H
3
C
(
19F-NMR (282.4 MHz, CDCl
−98.68 (m, 1 FAr, 2-F). LRMS: m/z (ESI) = 414.9, 416.9, 413.0 (M + Na)+.
) δ
3 F
7
-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)-2-fluorodibenzo[b,d]thiophene 5,5-dioxide (10a). A mixture of
Pd (dba) (11 mg, 0.012 mmol) and BINAP (15 mg, 0.024 mmol) in toluene (1.5 mL) was stirred for
0 min at 90 °C. The red-orange colored solution of the catalyst was allowed to cool (22 °C) and added
to a mixture of 1,4-diazabicyclo[3.2.2]nonane (9a, 53 mg, 0.42 mmol) and 8a (0.126 g, 0.40 mmol) in
toluene (2 mL). Cs CO (Alfa Aesar, Karlsruhe, Germany; previously dried for 2 h at 120 °C, 4 mbar;
.39 g, 1.2 mmol) was then added, and the reaction mixture was stirred under an atmosphere of argon
for 24 h at 90° C. After cooling to room temperature, the solid was filtered off and washed with CH Cl
2 × 4 mL). The filtrate was evaporated and chromatographically purified (silica gel 15–40 μm, 8 g) with
a gradient from CHCl (100%) to CHCl /MeOH/NH (aq) (100:8:0.8). The fractions containing the
product were combined, evaporated and the solid residue was recrystallized from EtOH to afford the
2
3
3
2
3
0
2
2
(
3
3
3
title compound 10a (0.075 g, 0.21 mmol, 52% yield, R
f
= 0.18, CHCl
): δ
, 8′-H ), 3.09 (A-part of AA′BB′, 2H,
), 3.61 (B-part of AA′BB′, 2H, 3′-H ), 4.08 (m, not resolved, 1H,
′-H),6.89 (dd, J = 8.8, 2.5 Hz, 1HAr, 8-H), 7.00 (td-like, J = 8.5, 8.5, 2.2 Hz, 1HAr, 3-H), 7.10 (d,
3
/MeOH/NH
3
(aq.), 100:10:1) as
1
yellow crystals, m.p. 317–319 °C (dec.). H-NMR (400 MHz, CDCl
3
H
1.78 (qd-like, J = 9.7, 4.6 Hz,
2
2
5
H, 6′-H
a
, 9′-H
a
), 2.11 (m, 2H, 6′-H
b
, 9′-H
b
), 2.99 (m, 2H, 7′-H
a
a
′-H ), 3.14 (m, 2H, 7′-H
2
b
, 8′-Hb
a
2
J = 2.5 Hz,1HAr,6-H), 7.24 (dd, J = 8.8, 2.2 Hz, 1HAr, 1-H), 7.49 (d, J = 8.7 Hz, 1HAr, 9-H), 7.70 (dd,
1
3
J = 8.4, 4.9 Hz, 1HAr, 4-H). C-NMR (100 MHz, CDCl
3
): δ
C
26.77 (s, 2 CH
2
), 44.69 (s, CH ), 46.51 (s,
2
2
CH ), 51.74 (s, CH), 57.05 (s, CH
2
2
), 105.19 (s, CH), 107.66 (d, 2JCF = 25.1 Hz, CH), 114.70 (d,
2JCF = 24.3 Hz, CH), 117.03 (s, CH), 117.28 (d, 4JCF = 2.2 Hz, C), 123.15 (s, CH), 124.27 (d, 3JCF = 10.3 Hz,
CH), 132.88 (d, 4JCF = 3.0 Hz, C), 136.19 (d, 3JCF = 9.6 Hz, C), 140.61 (s, C), 151.15 (s, C), 166.58 (d,

Molecules 2015, 20
18408
1JCF = 253.6 Hz, C). F-NMR (282.4 MHz, CDCl
19
): δ
−104.5 (m, 1FAr, 2-F). HRMS m/z (ESI): calcd
3
F
+
for C19
H20FN
2
O
2
S (M + H) 359.12240, found 359.12223.
3
-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)-6-fluorodibenzo[b,d]thiophene 5,5-dioxide (10b). In
procedure similar to the preparation of 10a, compound 8b (0.14 g, 0.45 mmol) and amine 9a (0.059 g,
.47 mmol) were reacted in the presence of Cs CO (0.44 g, 1.35 mmol) and a catalyst made from
Pd (dba) (10.3 mg, 0.011 mmol) and BINAP (14 mg, 0.022 mmol). For purification the same protocol
a
0
2
3
2
3
as for 10a was followed to afford the title compound 10b (0.077 g, 0.21 mmol, 48% yield,
1
R
f
= 0.19, CHCl
300 MHz, CDCl
Hz, 2H, 6′-H , 9′-H
), 3.61 (t-like, J = 5.7 Hz, B-part of AA′BB′, 2H, 3′-H
3
/MeOH/NH
): δ 1.77 (qd-like, J = 9.7, 4.6 Hz, 2H, 6′-H
), 3.05–2.92 (m, 2H, 7′-H , 8′-H ), 3.20–3.06 (m, A-part of AA′BB′, 4H, 7′-H
), 4.08 (m, not resolved, 1H, 5′-H),
3
(aq), 100:10:1) as yellow crystals, m.p. 272.5–274 °C (dec.). H-NMR
(
3
H
a
, 9′-H ), 2.11 (ddtd, J = 12.4, 9.7, 5.0, 2.6
a
b
b
a
a
b
,
8
6
7
′-Hb
a
, 3′-H
2
2
.89 (dd, J = 8.8, 2.5 Hz, 1HAr, 2-H), 6.97 (t, J = 8.4 Hz, 1HAr, 7-H), 7.09 (d, J = 2.5 Hz, 1HAr, 4-H),
.35 (d, J = 7.6 Hz, 1HAr, 9-H), 7.50 (dt, J = 8.0, 5.2 Hz, 1HAr, 8-H), 7.53 (d, J = 8.7 Hz, 1HAr, 1-H).
13C-NMR (75 MHz, CDCl
): δ
26.78 (s, 2 CH
), 44.70 (s, CH
), 46.52 (s, 2 CH
), 51.76 (s, CH),
3
C
2
2
2
5
1
1
7.06 (s, CH
2
), 105.09 (s, CArH), 115.17 (d, 2JCF = 19.4 Hz, CArH), 115.85 (d, 4JCF = 3.4 Hz, CArH),
17.09 (s, CArH), 117.65 (d, 4JCF = 2.8 Hz, CAr), 123.25 (s, CArH), 123.63 (d, 2JCF = 17.7 Hz, CAr),
35.83 (d, 3JCF = 2.7 Hz, CAr), 136.17 (d, 3JCF = 7.9 Hz, CArH), 140.42 (s, CAr), 151.12 (s, CAr), 158.05 (d,
1JCF = 257.4 Hz, CAr). F-NMR (282.4 MHz, CDCl
19
): δ −115.66 (m, FAr, 6-F). HRMS m/z (ESI): calcd
3
F
+
for C19
H20FN
2
O
2
S (M + H) 359.12240, found 359.12235.
3
-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)-2-fluorodibenzo[b,d]thiophene 5,5-dioxide (10c). In a procedure
similar to the preparation of 10a, compound 8c (0.126 g, 0.4 mmol) and 9a (0.053 g, 0.42 mmol) were
reacted in the presence of Cs CO (0.39 g, 1.2 mmol) and a catalyst mixture made from Pd (dba)
11 mg, 0.012 mmol) and BINAP (14.9 mg, 0.024 mmol). For purification the same protocol for 10a
2
3
2
3
(
was followed to afford the title compound 10c (0.063 g, 0.18 mmol, 44% yield, R
f
= 0.20,
1
CHCl
CDCl
3
/MeOH/NH
3
(aq), 100:10:1) as pale yellow crystals, m.p. 305–309 °C (dec.). H-NMR (300 MHz,
3
): δ 1.75 (m, 2H, 6′-H
H
a
, 9′-H
a
), 2.11 (m, 2H, 6′-H
), 3.44 (t, J = 5.6 Hz, B-part of AA′BB′, 2H, 3′-H
H, 5′-H), 7.32 (s, 1HAr, 4-H), 7.36 (d, J = 5.0 Hz, 1HAr, 1-H), 7.43 (ddd, J = 7.7, 5.1, 3.4 Hz, 1HAr
b
, 9′-H
b
), 3.04 (m, 4H, 7′-H
2
, 8′-H
2
), 3.16 (t,
J = 5.6 Hz, A-part of AA′BB′, 2H, 2′-H
2
2
), 3.80 (m,
1
7
2
,
1
3
-H), 7.55–7.63 (m, 2HAr, 8-H, 9-H), 7.75 (d, J = 7.6 Hz, 1HAr, 6-H). C-NMR (75 MHz, CDCl
7.64 (s, 2 CH ), 46.70 (s, 2 CH
), 47.50 (d, 4JCF = 1.2 Hz, CH
), 109.97 (d, 2JCF = 24.9 Hz, CH), 111.82 (d, 3JCF = 5.5 Hz, CH), 120.89 (s, CH), 122.19 (s, CH),
3
): δ
C
), 55.83 (d, 4JCF = 6.0 Hz, CH), 56.30 (s,
2
2
2
CH
2
1
1
23.72 (d, 3JCF = 9.4 Hz, C), 129.25 (s, CH), 131.40 (d, 4JCF = 2.3 Hz, C), 133.98 (d, 4JCF = 2.9 Hz, C),
34.05 (s, CH), 137.87 (s, C), 142.90 (d, 2JCF = 10.5 Hz, C), 157.62 (d, 1JCF = 252.1 Hz, C). F-NMR
19
(282.4 MHz, CDCl
3
): δ
F
−112.95 (m, 1 FAr, 2-F). HRMS m/z (ESI): calcd for C19
H20FN
2
O
S (M+H)+
2
3
59.12240, found 359.12235.
6-Fluoro-3-(9-methyl-3,9-diazabicyclo[3.3.1]nonan-3-yl)dibenzo[b,d]thiophene 5,5-dioxide (12b). In a
procedure similar to the preparation of 10a, compound 8b (0.13 g, 0.42 mmol) and amine 9b (0.065 g,
.46 mmol) were reacted in the presence of Cs CO (0.28 g, 0.84 mmol) and a catalyst made from
Pd (dba) (11.5 mg, 0.013 mmol) and BINAP (15.7 mg, 0.0252 mmol). For purification the same
0
2
3
2
3

Molecules 2015, 20
18409
protocol for 10a was followed to afford the title compound 12b (0.065 g, 0.175 mmol, 42% yield,
1
R
f
= 0.31, CHCl
3
/MeOH/NH
): δ 1.52–1.64 (m, 3H, 7′-H
.59 (s, 3H, 9′-NCH ), 3.01 (s, 2H, 1′-H, 5′-H), 3.39–3.51 (m, 4H, 2′-H
-H, 2-H), 7.22 (d, J = 2.4 Hz, 1HAr, 4-H), 7.39 (dd, J = 7.7, 0.5 Hz, 1HAr, 9-H), 7.52 (ddd, J = 8.2, 7.8,
3
(aq), 100:10:1) as yellow crystals, m.p. 264–272 °C (dec.). H-NMR
(300 MHz, CDCl
3
H
a
, 6′-H
a
, 8′-H
a
), 1.93–2.15 (m, 3H, 7′-H
b
, 6′-H
b
, 8′-H ),
b
2
7
5
2
3
2
, 4′-H ), 6.97–7.03 (m, 2HAr,
2
1
3
.2 Hz, 1HAr, 8-H), 7.59 (d, J = 8.7 Hz, 1HAr, 1-H). C-NMR (75 MHz, CDCl
7.93 (s, 2Csec), 40.62 (s, CH ), 47.40 (s, 2CH
), 53.03 (s, 2CH), 105.41 (s, CArH), 115.46 (d, 2JCF = 19.4 Hz,
ArH), 116.05 (d, 4JCF = 3.4 Hz, CArH), 117.22 (s, CArH), 118.83 (d, 4JCF = 2.6 Hz, CAr), 122.99 (s,
3
): δ
C
18.99 (s, 1 Csec),
3
2
C
C
ArH), 123.77 (d, 2JCF = 18.1 Hz, CAr), 135.76 (d, 3JCF = 2.5 Hz, CAr), 136.20 (d, 3JCF = 7.8 Hz, 1CArH),
40.10 (s, CAr), 152.27 (s, CAr), 158.06 (d, 1JCF = 257.4 Hz, CAr). F-NMR (282 MHz, CDCl
19
): δ
S (M + H)+
1
3
−
115.56 (dd-like, J = 8.4, 5.1 Hz, 1F, 6-F). HRMS m/z (ESI): calcd for C20
H21FN
2
O
2
3
73.13805, found 373.13773.
7
-(1,4-Diazabicyclo[3.2.2]nonan-4-yl)-2-nitrodibenzo[b,d]thiophene 5,5-dioxide (11a). In a procedure
similar to the preparation of 10a, compound 6a (0.17 g, 0.5 mmol) and amine 9a (0.064 g, 0.51 mmol)
were reacted in the presence of Cs
2
CO
3
(0.49 g, 1.5 mmol) and a catalyst mixture made from Pd
2
(dba)
3
(18.3 mg, 0.02 mmol) and BINAP (25 mg, 0.04 mmol). For purification the same protocol for 10a was
followed to afford the title compound 11a (0.10 g, 0.18 mmol, 53% yield, R
f
= 0.21, CHCl
3
/MeOH/NH
): δ
, 8′-H ), 3.10
), 3.65 (B-part of AA′BB′, 2H, 3′-H ), 4.11
3
1
(
(
(
(
(
aq), 100:10:1) as dark red crystals, m.p. 291–294 °C (dec.). H-NMR (300 MHz, CDCl
3
H
1.79
qd-like, J = 9.6, 4.6 Hz, 2H, 6′-H
a
, 9′-H
a
), 2.12 (m, 2H, 6′-H
b
, 9′-H
b
), 2.99 (m, 2H, 7′-H
a
a
A-part of AA′BB′, 2H, 2′-H ), 3.13 (m, 2H, 7′-H
2
b
, 8′-Hb
a
2
m, not resolved, 1H, 5′-H), 6.94 (dd, J = 8.8, 2.6 Hz, 1HAr, 8-H), 7.11 (d, J = 2.5 Hz, 1HAr, 6-H), 7.63
d, J = 8.8 Hz, 1HAr, 9-H), 7.85 (d, J = 8.3 Hz, 1HAr, 4-H), 8.15 (dd, J = 8.3, 2.0 Hz, 1HAr, 3-H), 8.35 (d,
1
3
J = 1.9 Hz, 1HAr, 1-H). C-NMR (75 MHz, CDCl
3
): δ
C
26.70 (2 CH
2
), 44.71 (CH
2
), 46.47 (2 CH ), 51.72
2
2
(CH), 57.02 (CH ), 105.18 (CH), 115.16 (CH), 116.04 (C), 117.33 (CH), 122.59 (CH), 123.18 (CH), 123.67
(
CH), 135.15 (C), 140.17 (C), 141.86 (C), 151.46 (C), 151.82 (C). HRMS m/z (ESI): calcd for
+
C
19
H
20
N
3
O
4
S (M + H) 386.11690, found 386.11704.
1
8
3
.3. Manual Radiosynthesis of [ F]10a
1
8
18
18
No-carrier-added (n.c.a.) [ F]fluoride was produced via the O(p,n) F nuclear reaction by
1
8
irradiation of a [ O]H
2
O target (Hyox 18 enriched water; Rotem Industries Ltd, Mishor Yamin, Israel)
®
on a Cyclone 18/9 (IBA RadioPharma Solutions, Louvain-la-Neuve, Belgium) with 18 MeV proton
®
18
18
beam using a Nirta [ F]fluoride XL target or [ O]H
2
O recycled by the established in-house method [63].
1
8
18
−
−1
Starting with 1–2 GBq of n.c.a. F-fluoride, [ F]F -containing anion resin was eluted with a 20 mg mL
aqueous solution of K
2
CO
3
(1.78 mg, 12.9 mmol) and added to a 5 mL V-vial in the Discover PET wave
®
®
microwave CEM (CEM Corporation, Matthews, NC, USA) cavity in the presence of Kryptofix 222
1
8
(
11.2 mg, 29.7 mmol) in 1 mL MeCN. The aqueous [ F]fluoride was dried under vacuum and argon
flow in the microwave cavity (75 W, 20 cycles) at 50–60 °C for 10–12 min. Additional aliquots of MeCN
2 × 1.0 mL) were added for azeotropic drying. Thereafter, the precursor (1 mg) was added to the reactive
(
1
8
®
anhydrous K[ F]F-Kryptofix 222/K
2
CO -complex. The reaction parameters were optimized varying
3
the reaction time, temperature, solvent (MeCN, DMF, DMSO), and heating condition (thermal vs.

Molecules 2015, 20
18410
microwave-assisted). Thereafter, the crude reaction mixture was diluted with H
2
O/MeCN (4:1, v/v) and
1
8
directly applied onto an isocratic semi-preparative HPLC for isolation of [ F]10a. Fractions were
collected, diluted with 30 mL of H
2
O and sodium hydroxide was added to neutralize the solution
®
(
(
~100–200 μL of 1 M NaOH). Final purification was performed using a Sep-Pak C18 light cartridge
Waters, Milford, MA, USA) followed by elution with 1 mL of ethanol. To obtain an injectable solution
1
8
the solvent was concentrated under a gentle argon stream and [ F]10a was formulated in a sterile
1
8
isotonic saline solution (5%–10% EtOH, v/v). The identity of [ F]10a was verified by radio-HPLC
analysis of a sample of the radiotracer solution spiked with the non-radioactive reference 10a.
Radiochemical and chemical purities were assessed by radio-TLC and analytical HPLC. The mass
determination for specific activity was determined on the base of a calibration curve carried out under
the same analytical HPLC conditions.
1
8
18
18
3
.4. Automated Radiosynthesis of [ F]10a and [ F]10b ([ F]ASEM)
Remote controlled radiosynthesis was performed using a TRACERLAB™ FXFN synthesizer (GE
Healthcare, Waukesha, WI, USA) equipped with a PU-980 pump (JASCO, Gross-Umstadt, Germany),
WellChrom K-2001 UV detector (KNAUER GmbH, Berlin, Germany), NaI(Tl)-counter and automated
data acquisition (NINA software version 4.8 rev. 4; Nuclear Interface GmbH, Dortmund, Germany). For
transfer to the automated module we started with activities in the range of 5–12 GBq of n.c.a.
18F-fluoride. Based upon the conditions optimized manually, the nitro-to-fluoro displacement was
achieved by adding the respective nitro precursor (1 mg) dissolved in anhydrous DMF (1 mL) to the
1
8
®
K[ F]F-Kryptofix 222/K
cooling, the reaction mixture was diluted with 3 mL of H
semi-preparative Reprosil-Pur 120 C18-AQ HPLC column (250 × 20 mm, 10 μm) using a solvent
2
CO
3
-complex. The reaction mixture was stirred at 120 °C for 10 min. After
2
O/MeCN (4:1) and directly applied onto the
−1
composition of 35% MeCN/H O/0.05%TFA as eluent and a flow rate of 10 mL·min . The desired
2
product was collected in 40 mL of H
2
O, neutralized with 1 M NaOH and directly transferred to a
®
18
pre-activated Sep-Pak C18 light cartridge. The cartridge was washed with 2 mL of water and [ F]10a
1
8
or [ F]10b was obtained after elution with 1 mL of EtOH. Injectable solutions were obtained by partial
evaporation of the solvent under a gentle argon stream at 70 °C and dilution in isotonic saline (5%–10%
of EtOH, v/v). Radiochemical purity and specific activity were assessed following the chromatographic
methods described in quality control.
3
.5. Quality Control
1
8
®
To control the quality of the K[ F]F-Kryptofix 222/K
2
CO
3
-complex, reactions using ethylene glycol
1
8
18
®
ditosylate were performed randomly. Radio-TLC of [ F]10a and [ F]10b was performed on Polygram
ALOX N/UV254 plates (Macherey-Nagel GmbH & Co. KG) with dichloromethane/methanol (9:1, v/v).
The spots of the references were visualized using UV light at 254 nm. Radiochemical purity and specific
activity were determined using analytical radio-HPLC with a Reprosil-Pur C18-AQ column (250 × 4.6 mm,
−1
5
μm) and 26% MeCN/H
profiles for the in vivo metabolism analysis in plasma was assessed by using a gradient mode (0–10 min:
0% MeCN/20 mM NH OAcaq., 10–40 min: 10%→90% MeCN/20 mM NH OAcaq., 40–50 min:
2
O/0.05% TFA as eluent at a flow rate of 1 mL·min . Analytical radio-HPLC
1
4
4

Molecules 2015, 20
18411
9
0%→10% MeCN/20 mM NH
4
OAcaq., 50–60 min: 10% MeCN/20 mM NH
4
OAcaq.) on a Reprosil-Pur
−1
C18-AQ (250 × 4.6 mm, 5 μm) column at a flow rate of 1 mL·min .
3
.6. Determination of in Vitro Stability and Lipophilicity (Log D7.2)
1
8
In vitro radiochemical stability of [ F]10a was investigated in 0.9% NaCl solution, PBS (pH 7.2),
and 0.01 M Tris-HCl (pH 7.4 at 21 °C) at 40 °C and in EtOH at room temperature for up to 90 min.
Samples were taken at 15, 30, 60, and 90 min of incubation time and analyzed by radio-TLC
1
8
and radio-HPLC. Log D7.2 of [ F]10a was experimentally determined in n-octanol/phosphate-buffered
saline (PBS; 0.01 M, pH 7.2) at room temperature by the shake-flask method in multiple distribution.
Measurement was performed twice in triplicate.
3
.7. Biological Evaluation
3.7.1. Animals
Male Sprague-Dawley rats (275–300 g) were obtained from Charles River Laboratories, and female
piglets (German Landrace pigs DL × Large White/Pietrain; mean weight 15.8 ± 0.8 kg, mean age 8 weeks)
as well as female CD-1 mice (10–12 weeks, 22–26 g) were obtained from Medizinisch-Experimentelles
Zentrum, Universität Leipzig, and all animals were provided with food and water ad libitum. All of the
animal experiments were performed in accordance with the European Communities Council Directive
of 24th November 1986 (86/609/EEC), and experiments using female piglets and female CD-1 mice
were approved by the Animal Care and Use Committee of Saxony (TVV 08/13). Imaging experiments
in rhesus monkeys were performed in accordance with a protocol approved by the Yale Institutional
Animal Care and Use Committee.
3.7.2. Binding Assays
Affinity for Human α
7
nAChR
The affinity of the test compounds for human α
experiments as previously described [12]. In brief, SH-SY5Y cells stably expressing human α
were grown in DMEM/Ham’sF-12, supplemented with 10% FCS, stable glutamine, geneticin
7
nAChR was determined by radioligand displacement
7
nAChR [64]
−1
−1
−1
(
100 μg·mL ), penicillin/streptomycin (100 U·mL , 100 μg·mL ) at 37 °C in an atmosphere containing
% CO . Cells were harvested by scraping, sedimented (800 rpm, 3 min), diluted with 50 mM
TRIS–HCl, pH 7.4, and stored at −25°C until use. For determination of α nAChR affinity of reference
5
2
7
compounds, frozen cell suspensions were thawed, homogenized by a 27-gauge needle, and diluted with
incubation buffer (50 mM TRIS–HCl, pH 7.4, 120 mM NaCl, 5 mM KCl). Membrane suspension was
3
3
incubated with [ H]methyllycaconitine ([ H]MLA; ~0.3 nM final concentration; specific activity of
−
1
2
.220 GBq·mmol ; NEN Life Sciences Products, Boston, MA, USA) and various concentrations of the
test compound. Nonspecific binding was determined by co-incubation with 300 μM (−)-nicotine tartrate.
The incubation was performed at room temperature for 120 min and terminated by rapid filtration using
Whatman GF/B glass-fibre filters, presoaked in 0.3% polyethyleneimine, and a 48-channel harvester
(
Biomedical Research and Development Laboratories, Gaithersburg, MD, USA), followed by 4×

Molecules 2015, 20
18412
washing with ice-cold 50 mM TRIS-HCl, pH 7.4. Filter-bound radioactivity was quantified by liquid
scintillation counting. The 50% inhibition concentrations (IC50) were estimated from the competition
curves by nonlinear regression using GraphPadPrism software and the K values calculated according to
i
the Cheng-Prusoff equation [65].
Affinity for Human α
4
β
2
and α
3
4
β nAChR
The affinity of the test compounds for human α
4
β
2
and α
3
β
4
nAChR was determined as described
or α nAChR [66] were cultured.
above. In brief, HEK293 cells stably expressing human α
4
β
2
β
3 4
The cells were harvested and processed as described above, and membrane suspension was incubated
3
−1
with (±)-[ H]epibatidine (0.4 to 0.6 nM final concentration; specific activity of 1250 GBq·mmol ;
PerkinElmer Human Health, Rodgau, Germany) and various concentrations of the test compound.
Nonspecific binding was determined in the presence of 300 μM (−)-nicotine tartrate. Incubation of the
samples, separation of free from receptor-bound radioligand, and data evaluation were performed as
described above.
Affinity for Subunit-Specific Antibodies Immobilized Native Rat α
6
2
β * nAChR
The specificity of the anti-α
immobilized as previously described [68] by the subunit-specific antibody. The inhibition of
[3
H]epibatidine binding to the immobilized subtypes by the test compounds was measured by
pre-incubating increasing concentrations of the compounds for 1 h at room temperature, followed by
6
antibody was tested as described [67]. The receptors were prepared and
3
−1
incubation overnight at 4 °C with 0.1 nM (±)-[ H]epibatidine (specific activity of 1250 GBq·mmol ;
PerkinElmer Human Health, Rodgau, Germany). Incubations were performed in a buffer containing
−1
50 mM Tris-HCl, pH 7, 150 mM NaCl, 5 mM KCl, 1 mM MgCl
2
, 2.5 mM CaCl
2
, 2 mg·mL BSA, and
0
.05% Tween 20. Specific ligand binding was defined as total binding minus the binding in the presence
of 100 nM (±)-epibatidine dihydrochloride hydrate. After incubation, the wells were washed seven times
3
with ice-cold PBS containing 0.05% Tween 20, the bound [ H]epibatidine released by adding 200 μL of
2
M NaOH to each well and after incubation for 2 h determined by means of liquid scintillation in a
beta counter.
Saturation binding analysis revealed an affinity of (±)-[ H]epibatidine of 25 pM (CV 20%) for the
nAChRs immobilized by anti-α antibody from rat striatum, which contains two major α * subtypes
and α ) in very similar proportions.
The experimental data obtained from the binding experiments were analyzed by means of a
non-linear least square procedure using the LIGAND program [69]. The K values of the test compounds
3
6
β
6 2
(
α
6
β
2
β
3
6
α
4
β β
2 3
i
were also determined by means of the LIGAND program by simultaneously fitting the data obtained
from three to five independent saturation and competition binding experiments.
Affinity for Human 5-HT
3
Receptor
The affinity of the test compounds for human 5-HT
3
receptor was evaluated by a contract research
organization (Eurofins Panlabs, Inc., Taipei, Taiwan). All compounds were tested at 1 μM in radioligand
binding assays using human recombinant HEK-293 cells for the percentage inhibition of the binding by

Molecules 2015, 20
18413
3
the 5-HT
3
specific radioligand [ H]GR-65630 (working concentration = 0.69 nM, K
D
= 0.20 nM). In
addition, for compound 10a a K
i
value was estimated based on the inhibition of the radioligand binding
at seven concentrations of the test compound (20 μM–5 nM).
Affinity for Other Target Proteins
The activity of compound 10a was evaluated by a contract research organization (Eurofins Panlabs,
Inc.). The test concentration was 1 μM and radioligand binding assays for dopamine transporter (DAT),
serotonin transporter (SERT), norepinephrine transporter (NET), monoamine transporter (VMAT),
choline transporter, and α
1
nAChR performed according to the manufacturer’s protocols. Responses
were judged as significant at ≥50% inhibition or stimulation of the binding of the specific radioligand.
3.7.3. PET Studies in Piglets
Six female animals were used in this study with anesthesia and surgery performed as described
−
1
previously [70]. In brief, the animals received initially a premedication with midazolam (1 mg·kg i.m.)
followed by induction of anesthesia with 3% isoflurane in 70% N O/30% O . All incision sites were
2
2
infiltrated with 1% lidocaine, and the anesthesia was maintained with 1.5% isoflurane throughout the
surgical procedure. A central venous catheter was introduced through the left external jugular vein and
used for the administration of the radiotracer and drugs, and for volume substitution with heparinized
−
1
−1 −1
(
50 IE·mL ) lactated Ringer’s solution (2 mL·kg ·h ). An endotracheal tube was inserted by
tracheotomy for artificial ventilation (Servo Ventilator 900C, Siemens-Elema, Solna, Sweden) after
−1
−1
immobilization with pancuronium bromide (0.2 mg·kg ·h ). The artificial ventilation was adjusted to
obtain normoxia and normocapnia (Radiometer ABL 500, Copenhagen, Denmark). Polyurethane
catheters (i.d., 0.5 mm) were advanced through the left and the right femoral arteries into the abdominal
aorta to withdraw arterial blood samples for regular monitoring of blood gases and for radiotracer input
function measurements and metabolite analyses. The body temperature was monitored by a rectal
temperature probe and maintained at ~38 °C by a heating pad. After the surgical procedure has been
completed, anesthesia was maintained with 0.5% isoflurane in 70% N
2
O/30% O , and the animals were
2
allowed to stabilize for 1 h before imaging procedure.
PET Scanning Protocol
PET imaging was performed according to the protocol described recently [21]. In brief, a clinical
tomograph (ECAT Exact HR+; Siemens Healthcare GmbH, Erlangen, Germany) was used with animals
lying prone with the head placed in a custom-made head holder. For attenuation and scatter correction,
68
transmission scans were acquired using three rotating Ge rod sources. The radiotracer was injected as
1
8
an intravenous (i.v.) bolus (10 mL saline with 340–494 MBq [ F]10a) by a syringe pump within 2 min
−1
followed by flushing with 10 mL heparinized saline (50 IE·mL ). PET scanning started at the time of
injection (0 min) and dynamic emission data were acquired for a total of 240 min. Three animals were
investigated under baseline conditions, and another three under blocking conditions with administration
−
1
of 3 mg·kg i.v. of the α
7
nAChR partial agonist NS6740 at 3 min before radiotracer injection followed
−1
−1
by a continuous infusion throughout the scan (1 mg·kg ·h ).

Molecules 2015, 20
18414
Arterial blood was sampled using a peristaltic pump during the first 20 min of the scan followed by
manual sampling at 25, 30, 40, 50, 60, 90, 120, 150, 180, 210, and 240 min after injection. Plasma was
obtained by centrifugation (13,000 rpm, 2 min at 21 °C) and total radioactivity concentration was
measured using a gamma-counter (1470 Wizard; PerkinElmer, Shelton, CT, USA) cross-calibrated to
the PET scanner. In addition, arterial whole blood was sampled manually at 4, 16, 30, 60, and 120 min
1
8
p.i. and plasma obtained for metabolite analyses to determine the fraction of non-metabolized [ F]10a.
Arterial Input Function
The plasma input function was corrected for the presence of radioactive metabolites by determination
of the parent fraction as described previously [12]. Plasma samples were de-proteinized by addition of
ice-cold acetonitrile (500 μL MeCN/100 μL plasma) followed by centrifugation (13,000 rpm, 10 min).
Aliquots of the original plasma sample and of the supernatant obtained after centrifugation were counted
for radioactivity, and the supernatant was analyzed by radio-HPLC following the chromatographic
methods described in quality control. The parent fraction as a function of time was fitted iteratively with
the Levenberg-Marquardt algorithm, including decay correction, to calculate metabolite corrected input
function [71].
Quantification of PET Data
After correction for attenuation, scatter, decay and scanner-specific dead time, images were
reconstructed by filtered back-projection using a Hanning-filter of 4.9 mm FWHM into 40 frames of
increasing length. A summed PET image of the 240-min scan was reconstructed for each animal and
used for alignment with a T1-weighed MR image of a 6-week-old farm-bred pig as described
previously [72]. The time-activity curves (TACs) were calculated for the following volumes of interest
(
VOIs): frontal cortex, temporal cortex, parietal cortex, occipital cortex, hippocampus, striatum (defined
as mean radioactivity in caudate and putamen), cerebellum, thalamus, middle cortex, ventral cortex,
midbrain, pons, and colliculi. Radioactivity in all VOIs was calculated as the average of radioactivity
−3
concentration (kBq·cm ) in the left and right sides. To generate standardized uptake values (SUVs) the
TACs of the individual VOIs were normalized to the injected dose and corrected for animal weight (in
−1
kBq·g ). Kinetic analysis of the time-activity curves (TACs) was performed using an “in house” data
analysis tool [73].
3.7.4. PET Study in Rhesus Monkey
One male rhesus monkey (Macaca mulatta, 6 years old, body weight = 10 kg) underwent two PET
1
8
18
imaging sessions with [ F]10a and [ F]10b, respectively. The scans were conducted 19 days apart.
−
1
The monkey was initially sedated by intramuscular ketamine at 10 mg·kg at ~2 h before the PET scan
and anesthetized by 1.5%–3% isoflurane throughout the imaging session. PET scan was performed
using a FOCUS-220 PET scanner (Siemens Preclinical Solutions, Knoxville, TN, USA). After a
1
8
−1
18
transmission scan, 167 MBq of [ F]10a (specific activity of 362 GBq·μmol ) or 185 MBq of [ F]10b
−1
(
specific activity of 159 GBq·μmol ) were administered intravenously. Emission data were acquired in
list mode for 240 min. Raw list-mode PET data was histogrammed (frames of 6 × 30 s; 3 × 1 min; 2 × 2 min;

Molecules 2015, 20
18415
4
6 × 5 min to scan termination) and reconstructed with 2D filtered back projection, using a 0.15 cm−1
Shepp filter and including corrections for scanner normalization, detector dead time, randoms, and
radiation scatter and attenuation. The PET images were then registered to MR images, acquired with a
Siemens 3 T Trio scanner, with a 6-parameter rigid body registration. The MR native space was then
normalized using nonlinear affine registration to a high-resolution rhesus monkey atlas using BioImage
Suite 3.01 (http://www.bioimagesuite.org/index.html). Time-activity curves were extracted by mapping
atlas-defined regions to PET native space using the optimal transformation matrices calculated in the
registration and normalization steps. Regions extracted included amygdala, caudate, cerebellum, cingulate,
cortical regions (frontal, occipital, and temporal cortex), hippocampus, putamen, thalamus, and pons.
3
.7.5. Toxicity Studies in Rats
The acute toxicity of 10a was assessed when administered by a single intravenous injection to rats
followed by an observation period of 2 or 15 day (Hameln rds, Modra, Slovak Republic). The study was
−1
performed with 4 test groups, including 1 control and 3 dose groups (6.2, 62 and 620 μg·kg ), with
0 males and 60 females Wistar rats divided into two experiment cohorts: Day 2 (40 males and 40 females)
6
and Day 15 (20 males and 20 females). The animals were weighed and allocated to the test groups based
on their body weights. The animals were sacrificed in two time periods (day 2 and day 15 after test item
administration) and then examined macroscopically and histopathologically. The following parameters
were evaluated: mortality, clinical observation, body weights, food consumption, haematology (white
blood cells, red blood cells, hemoglobin, hematocrit, mean corpuscular volume, mean corpuscular
hemoglobin, platelets, lympohcytes, neutrophils, eosinophils, basophils, monocytes), clinical chemistry
(alkaline phosphatase, aspartate aminotransferase, alanine aminotransferase, glucose, cholesterol,
triacylglycerols, creatinine, urea, bilirubin, albumin, calcium, phosphorus, sodium, potassium chloride),
pathology and histopathology.
4
. Conclusions
We have designed and synthesized a number of compounds based on a new pharmacophore and tested
their in vitro binding affinity and selectivity for α
7
nAChR over other receptor subtypes. Among the new
nAChR binding affinity and selectivity, and was
selected for radiolabeling with F and evaluation as PET imaging radioligand for α nAChR in vivo.
[18
F]10a was produced in high radiochemical yield and purity both manually and in an automated
synthesis. PET imaging experiments in piglets and monkey demonstrated that [ F]10a possessed
appropriate pharmacokinetic properties and displayed specific and displaceable binding to α nAChR in
the brain. Taken together, the new radioligand [ F]10a ([ F]DBT-10) appears to be a promising agent
ligands, compound 10a was found to have high α
7
1
8
7
1
8
7
1
8
18
for PET imaging and quantification of α nAChR availability in the primate brain.
7
Supplementary Materials
Supplementary materials including protocols for the preparation of intermediate compounds and
NMR spectra of compounds 10, 10a, 10b, 10c, 11a, 12a, 12b, 13b and 14b can be accessed at:
http://www.mdpi.com/1420-3049/20/10/18387/s1.

Molecules 2015, 20
18416
Acknowledgments
The Deutsche Forschungsgemeinschaft is acknowledged for financial support (Project DE 1165/2-3). We
thank the staff of the Institute of Analytical Chemistry, Department of Chemistry and Mineralogy of the
University of Leipzig, for the NMR spectra, Karsten Franke, Helmholtz-Zentrum Dresden-Rossendorf
(
HZDR), and the cyclotron staff of the Department of Nuclear Medicine of the University Hospital
18
Leipzig for providing [ F]fluoride, as well as Tina Spalholz, HZDR, and Juliane Schaller, HZDR, for
technical assistance. Support to DanPET AB from CCJobs (Creating Competitive Jobs), FBÖ
TransTechTrans and NRU, COGNITO, Copenhagen, is gratefully acknowledged.
Author Contributions
Rodrigo Teodoro, Barbara Wenzel, Jörg Steinbach, Marianne Patt, Ming-Qiang Zheng and
Yiyun Huang designed and conducted the radiochemical experiments. Matthias Scheunemann and
Dan Peters conceived and performed the chemical syntheses. Francesca Maria Fasoli, Cecilia Gotti and
Winnie Deuther-Conrad conducted the in vitro binding studies. Winnie Deuther-Conrad,
Cornelius K. Donat, Peter Brust, Osama Sabri, Ming-Qiang Zheng, Ansel Hillmer and Yiyun Huang
designed and conducted the PET imaging experiments. Rodrigo Teodoro, Barbara Wenzel, Peter Brust,
Mathias Kranz, Winnie Deuther-Conrad, Cecilia Gotti, Ansel Hillmer and Yiyun Huang analyzed the
data. Rodrigo Teodoro, Matthias Scheunemann, Winnie Deuther-Conrad, Peter Brust, Osama Sabri,
Ansel Hillmer, Ming-Qiang Zheng and Yiyun Huang wrote and revised the manuscript.
Conflicts of Interest
The authors declare no conflict of interest.
References
1
2
3
. Dineley, K.T.; Pandya, A.A.; Yakel, J.L. Nicotinic ACh receptors as therapeutic targets in CNS
disorders. Trends Pharmacol. Sci. 2015, 36, 96–108.
. Weiland, S.; Bertrand, D.; Leonard, S. Neuronal nicotinic acetylcholine receptors: From the gene to
the disease. Behav. Brain Res. 2000, 113, 43–56.
. Briggs, C.A.; Gronlien, J.H.; Curzon, P.; Timmermann, D.B.; Ween, H.; Thorin-Hagene, K.;
Kerr, P.; Anderson, D.J.; Malysz, J.; Dyhring, T.; et al. Role of channel activation in cognitive
enhancement mediated by α
486–1494.
7
nicotinic acetylcholine receptors. Br. J. Pharmacol. 2009, 158,
1
4
5
6
7
. Papke, R.L. Merging old and new perspectives on nicotinic acetylcholine receptors. Biochem.
Pharmacol. 2014, 89, 1–11.
. Palma, E.; Conti, L.; Roseti, C.; Limatola, C. Novel approaches to study the involvement of
α
7
-nAChR in human diseases. Curr. Drug Targets 2012, 13, 579–586.
. Horenstein, N.A.; Leonik, F.M.; Papke, R.L. Multiple pharmacophores for the selective activation
of nicotinic α -type acetylcholine receptors. Mol. Pharmacol. 2008, 74, 1496–1511.
7
. Mazurov, A.; Hauser, T.; Miller, C.H. Selective α
Med. Chem. 2006, 13, 1567–1584.
7
nicotinic acetylcholine receptor ligands. Curr.