The development of potential new fluorine-18 labelled radiotracers for imaging the GABAA receptor

Article abstract of DOI:10.1016/j.bmcl.2012.11.066

Positron emission tomography (PET) using the tracer [¹¹C] Flumazenil has shown changes in the distribution and expression of the GABA _A receptor in a range of neurological conditions and injury states. We aim to develop a fluorine-18 labelled PET agent with comparable properties to [¹¹C]Flumazenil. In this study we make a direct comparison between the currently known fluorine-18 labelled GABA_A radiotracers and novel imidazobenzodiazepine ligands. A focussed library of novel compound was designed and synthesised where the fluorine containing moiety and the position of attachment is varied. The in vitro affinity of twenty-two compounds for the GABA_A receptor was measured. Compounds containing a fluoroalkyl amide or a longer chain ester group were eliminated due to low potency. The fluorine-18 radiochemistry of one compound from each structural type was assessed to confirm that an automated radiosynthesis in good yield was feasible. Eleven of the novel compounds assessed appeared suitable for in vivo assessment as PET tracers.

Full text of DOI:10.1016/j.bmcl.2012.11.066

Bioorganic & Medicinal Chemistry Letters xxx (2012) xxx–xxx
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
The development of potential new fluorine-18 labelled radiotracers for
imaging the GABAA receptor
⇑
Alexander Jackson , Benedicte B. Guilbert, Stuart D. Plant, Julian Goggi, Mark R. Battle, John L. Woodcraft,
Alessandra Gaeta, Clare L. Jones, Denis R. Bouvet, Paul A. Jones, Dennis M. O’Shea, Penny Hao Zheng,
Samantha L. Brown, Amanda L. Ewan, William Trigg
GE Healthcare, Medical Diagnostics, The Grove Centre, White Lion Road, Amersham, Buckinghamshire HP7 9LL, UK
a r t i c l e i n f o
a b s t r a c t
Positron emission tomography (PET) using the tracer [11C]Flumazenil has shown changes in the
distribution and expression of the GABAA receptor in a range of neurological conditions and injury states.
We aim to develop a fluorine-18 labelled PET agent with comparable properties to [11C]Flumazenil. In
this study we make a direct comparison between the currently known fluorine-18 labelled GABAA radio-
tracers and novel imidazobenzodiazepine ligands. A focussed library of novel compound was designed
and synthesised where the fluorine containing moiety and the position of attachment is varied. The
in vitro affinity of twenty-two compounds for the GABAA receptor was measured. Compounds containing
a fluoroalkyl amide or a longer chain ester group were eliminated due to low potency. The fluorine-18
radiochemistry of one compound from each structural type was assessed to confirm that an automated
radiosynthesis in good yield was feasible. Eleven of the novel compounds assessed appeared suitable for
in vivo assessment as PET tracers.
Article history:
Received 18 October 2012
Revised 16 November 2012
Accepted 18 November 2012
Available online xxxx
Keywords:
GABAA
PET
Flumazenil
Benzodiazepine
GE-194
FASTlab
Ó 2012 Elsevier Ltd. All rights reserved.
Single photon emission computed tomography (SPECT) and PET
are nuclear imaging techniques that allow the assessment of bio-
logical molecular processes in vivo. Both techniques require the
injection of a radio-labelled imaging agent into the subject prior
to the imaging procedure. PET imaging of GABAA receptor distribu-
tion in the human brain was first reported in 1985.1 Subsequently,
changes in the distribution and expression of the GABAA receptor
have been reported in a range of neurological conditions and injury
states including epilepsy, stroke, Alzheimer’s disease, traumatic
brain injury, amyotrophic lateral sclerosis and alcohol depen-
dence.2–9 The GABAA receptor is a ligand gated chloride ion channel
which has a pentameric structure, the subunits of which are fur-
[
11C]FMZ has been used in a range of clinical imaging studies and
is evidently a valuable PET tracer in clinical research.2–9 The struc-
turally related compound Ro15-4513 has also been labelled with
carbon-11 and used in clinical PET imaging studies.20,21 SPECT with
[
123I]Iomazenil (Ro16-0154) (half-life = 13.2 h) and has been used
extensively for localization of epileptogenic foci as a pre-operative
diagnostic tool.22,23
Due to the short half-life of carbon-11, [11C]FMZ is usually pro-
duced and used at the same site. This logistical issue prevents the
broader use of [11C]FMZ. The longer half-life of fluorine-18 makes
central production and distribution a possibility for [18F]FMZ: a
factor which could enable broader adoption of this tracer. The
radiosynthesis of [18F]FMZ has been reported by several groups
with varying degrees of success. Radiochemical yields have been
reported as medium-low and variable which is in agreement with
our experience.15–19 This has led some researchers to develop clo-
sely related derivatives of FMZ that can be radio-labelled by alter-
native strategies and aim to maintain or improve on the properties
of FMZ as a PET agent. These include N- and O-alkylation with
radio-labelled intermediates such as 2-[18F]fluoroethyl bromide
and tosylate to produce [18F]FEFMZ and [18F]FFMZ (Fig. 1).24–30
ther divided into various subtypes (a1–6, b1–4, c1–3, d, e, p, q
1–3).10,11
The most studied radiotracers for the GABAA receptor are imi-
dazobenzodiazepines which bind to an allosteric site on the recep-
tor located between the
a- and c-subunits, known as the
benzodiazepine (BZ) binding site. Ligand binding at the BZ site
modulates ion channel activity. For example, the drug Flumazenil
(FMZ) is a GABAA antagonist and has been used as an antidote in
the treatment of BZ drug overdose.12 FMZ binds to subtypes of
the GABAA receptor containing a1,2,3 and a5 subunits.13,14 FMZ
has been radio-labelled with carbon-11 ([11C]FMZ) (half-life = 20
min) and fluorine-18 ([18F]FMZ) (half-life = 110 min) (Fig. 1).15–19
[
18F]FEFMZ has been used to image human benzodiazepine recep-
tors. However, it is reported that this tracer has disadvantages
compared to FMZ such as lower brain uptake, lower affinity, faster
metabolism and faster kinetics.26 The peak brain uptake of
⇑
Corresponding author.
E-mail address: Alex.Jackson@GE.com (A. Jackson).
[
18F]FFMZ in mice is reported as 2.5% at 2 min with very slow
0960-894X/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.bmcl.2012.11.066
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066

2
A. Jackson et al. / Bioorg. Med. Chem. Lett. xxx (2012) xxx–xxx
O
O
O
R
O
O
N
N
N
N
N
R1H
R1
N
N
N
N
O
18F
F
N
F
F
N
N
F
F
11CH3
O
O
O
O
FFMZ, R1 = O(CH2)2F
[
11C]FMZ
[
18F]FFMZ
R = OEt R = Cl
, R1 = O(CH2)3F
, R1 = NH(CH2)2F
1
2
O
O
O
O
N
N
3, R1 = NH(CH2)3F
1) NaOH
2) SOCl2
N
N
18F
N
N
Scheme 1. Synthesis of FFMZ, 1, 2 and 3.
18F
O
O
[
18F]FMZ
[
18F]FEFMZ
should also have properties not inferior to [18F]FMZ as a neurology
PET agent when directly compared in pre-clinical in vitro and
in vivo models. Preferably, the agent should also have greater
in vivo stability than FMZ, which is known to rapidly metabolise
in the periphery. This study includes previously reported ana-
logues of FMZ and novel compounds in a direct comparison and
Figure 1. Analogues of FMZ labelled with carbon-11 and fluorine-18.
washout which is in contrast to
[ [
18F]FMZ.29 18F]FFMZ and
[
18F]FMZ have been directly compared in an imaging study in
rats.30 Despite the two tracers having comparable peak brain up-
take, the authors concluded that [18F]FMZ is the superior PET tracer
since [18F]FFMZ was rapidly metabolised. At 30 min, 10% of the sig-
nal was from parent compound in the brain. Consequently, there
was as large nondisplaceable fraction.
provides
a selection of compounds that appear suitable for
in vivo screening.
A library of compounds based on FMZ was synthesised and the
in vitro binding affinity for the GABAA receptor measured. Table 1
shows a summary of the data for compounds screened.
Scheme 1 shows the synthesis of analogues of FMZ where the
ester group R1 is modified (FFMZ and compounds 1–3). The aim:
improved radiochemistry and minimal changes to other properties
The aim of our programme is to find a GABAA PET tracer, which
can be labelled with fluorine-18, reliably and in a high yield using a
method that is amenable to automated production. The tracer
Table 1
In vitro affinity of literature compounds and compounds 1–22
General structure of compounds
General structure of
compounds 14-19
General structure of
FMZ, FFMZ, FEFMZ, 1-13
compounds 20-22
O
O
O
N
N
N
10
X
R1
R1
R1
N
N
N
Y
8
R3
8
R3
8
R3
N
N
N
Z
R2
R2
7
7
7
O
O
O
40
Compound
R1
R2
R3
Ki (nM)
1.3
4
>10,000
>10,000
>10,000
LogD7.4
cLogP41
FMZ
FFMZ
1
2
3
4
5
6
FEFMZ
7
8
9
10
11
12
13
OCH2CH3
OCH2CH2F
OCH2CH2CH2F
NHCH2CH2F
CH3
CH3
CH3
CH3
CH3
CH3
8-F
8-F
8-F
8-F
8-F
8-F
8-F
8-F
8-F
8-F
7-F
7-F
7-F
8-OCH2CH2F
8-OCH2CH2F
7-OCH2CH2F
7-Br
7-Br
7-Br
1.12
0.68
—
—
—
1.65
1.44
1.54
1.25
—
0.69
—
1.62
1.17
1.85
—
—
—
—
—
—
1.58
—
—
—
1.29
1.01
1.24
1.16
1.39
2.00
1.82
3.06
1.54
2.25
1.29
1.60
2.00
1.57
1.28
2.28
3.07
2.09
2.32
2.23
2.46
1.65
0.15
0.86
0.86
NHCH2CH2CH2F
OtBu
1.9
10.7
2.9
13.5
8.9
5.5
0.42
0.76
2.3
0.52
1.1
0.3
OCH2CH3
OCH2CH3
OCH2CH3
OtBu
OCH2CH3
OCH(CH3)2
OtBu
CH2CH3
Bn
CH2CH2F
CH2CH2F
CH3
CH3
CH3
CH3
CH3
CH3
—
—
—
—
—
OCH2CH3
OtBu
OtBu
14 (Bretazenil)
OtBu
15
16
17
18
19
20
21
22
OCH2CH2F
OCH2CH2CH2F
NHCH2CH2F
NHCH2CH2CH2F
OCH2CH3
OCH2CH3
OtBu
9.4
3.2
7-Br
7-Br
8-F
>10,000
>10,000
2.3
14
3.9
1.3
—
CH3
CH3
CH3
8-F, X = CH, Y = C, Z = N
7-F, X = CH, Y = N, Z = C
7-F, X = N, Y = CH, Z = C
OtBu
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066

A. Jackson et al. / Bioorg. Med. Chem. Lett. xxx (2012) xxx–xxx
3
H
H
N
L-Proline
DMSO,150 C
O
O
NH2
N
O
Triphosgene
O
R3
CO2H
N
F
R1
O
NaH, DMF, THF
Diethyl phosphoro-
chloridate,
R2NHCH2CO2H
ethyl isocyanoacetate
H
O
R3 = OMe
3 = OH
R3 = 2-Fluoro-
ethoxy
N
N
N
O
(a)
(b)
R
N
OEt
N
R3
R2
O
F
4, R1= OtBu, R2 = Me, R3 = 8-F
5, R1= OEt, R2 = Et, R3 = 8-F
O
19
NaH, DMF, THF
Diethyl phosphoro-
chloridate,
6, R1 = OEt, R2 = Bn, R3 = 8-F
FEFMZ, R1= OEt, R2 = CH CH2F, R3 = 8-F
ethyl ort -butyliso cyanoacetate
, R = O Bu, R2 = C2H2CH2F, R3 = 8-F
1
t
7
8
9
10
11
12
13
, R1 = OEt, R2 = Me, R3 = 7-F
O
N
8
(c)
, R1 = OiPr, R2 = Me, R3 = 7-F
9
R1
N
, R1 = OtBu, R2 = Me, R3 = 7-F
R
2 = DMB
, R1 = OEt, R2 = Me, R3 = 8-OCH2CH2F
, R1 = OtBu, R2 = Me, R3 = 8-OCH2CH2F
, R1 = OtBu, R2 = Me, R3 = 7-OCH2CH2F
(d), 69%
(e), 41%
R2 = H
N
R3
R2
FEFMZ
O
Scheme 2. Synthesis of compounds 4–13, 19 and FEFMZ. Reagents and conditions (a) BBr3, DCM, ꢀ78° C; (b) Cs2CO3, DMF, 2-fluoroethyl tosylate; (c) Ti(OiPr)4, 85 °C
i
(microwave, 60 W), 1–1.5 h, PrOH; (d) DCM, trifluoroacetic acid, trifluorosulfonic acid, 5 °C to rt; (e) NaH, DMSO, 2-fluoroethyl tosylate, 90 °C, 30 min.
O
F
O
R
N
N
N
OH
N
N
R1
F
HATU, R1H
(a)
(b)
HN
N
F
O
N
N
N
N
37%
F
O
Br
Br
O
O
O
O
O
15, R1 = O(CH2)2F
O
t
14
R = O Bu,
R = OH
N
H
N
, R1 = O(CH2)3F
, R1 = NH(CH2)2F
, R1 = NH(CH2)3F
16
17
18
(c)
N
H
(a)
O
O
N
N
Scheme 3. Synthesis of compounds 15–18. Reagents and conditions (a) 70 °C,
formic acid (96% in water), 99%.
N
O
O
N
O
N
F
N
21
80%
N
22
10%
N
F
O
O
CDI
19%
F
HN
O
N
Scheme 5. Synthesis of compounds 21 and 22. Reagents and conditions (a) (COCl)2,
rt, 0.5 h, DCM, DMF (1 drop); (b) tert-butyl 5-((methylamino)methyl)-1H-imidaz-
ole-4-carboxylate, rt, 24 h, DMF; (c) Cs2CO3, 60 °C, 1 h, CH3CN.
N
H
N
Cl
N
N
N
H
O
O
O
F
Compound 8 has been synthesised previously and is also known
as Ro15-5623.31 Taking the structure of FMZ and changing the sub-
stituent R2 from methyl- to 2-fluoroethyl, gives the literature com-
pound FEFMZ. The same approach can be taken and combined with
further substitutions, for example, R1 = OtBu in compound 7. R1 is
either an ethyl ester or tert-butyl ester in the compounds FMZ/4,
FEFMZ/7, 8/10 and 11/12. These pairs of compound were included
to examine the effect of this substitution on potency, in vivo stabil-
ity and physicochemical properties. In Scheme 3, four compounds
(15–18) were synthesized based on the structure of Bretazenil: a
potent GABAA partial agonist. These included variations at the ester
moiety R1 and phenyl ring R3 according the rationale described for
compounds 1–13. Finally, in Schemes 4 and 5, a further three com-
pounds (20–22) were synthesised where the phenyl ring is re-
placed with a pyridyl in combination with changes to R3 and the
ester group R1. This is expected to have an impact on the radio-
chemistry, physicochemical properties and in vivo stability.
NaH
DMF
23%
OH
Cl
N
O
O
N
N
O
23
X = Cl,
CsF
130 C (µw)
X = F, 20
N
X
N
O
Scheme 4. Synthesis of compounds 20 and 23.
compared to FMZ. Scheme 2 shows the synthesis of a range of com-
pounds with variations at R1, R2 and R3. Varying the position and/or
substituent R3 in compounds 8–13 (Table 1 and Scheme 2) is de-
signed to allow improvements to the radiochemistry whilst main-
taining favourable potency and physicochemical properties.
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066

4
A. Jackson et al. / Bioorg. Med. Chem. Lett. xxx (2012) xxx–xxx
FFMZ, 1, 2 and 3 were synthesised from FMZ in a three step pro-
than for [18F]FMZ. In order to make a meaningful assessment, the
radiosynthesis of [18F]FMZ was assessed using a GE TRACERlab
FX F-N automated radiosynthesis module with HPLC purifica-
tion.34,37 The procedure used similar conditions as reported by
Halldin et al.17 Typically, the non-decay-corrected end of synthesis
yield (EOS) was less than 5%. The highest EOS yield obtained in a
single experiment was 11.5%.35 Efforts were made to optimize
the synthesis further to improve the radiochemical yield and re-
duce variability, without success. Furthermore, the reaction tem-
perature of 160 °C used in these examples cannot currently be
used on all commercially available automated synthesis modules.
Efforts to synthesise [18F]FMZ using a lower reaction temperature
on a gave maximum EOS yields of 5% (n = 2).36 Other groups have
reported radiosynthesis of [18F]FMZ with EOS yields in the range
6–18%.17–19 However, there is variation in reported yields between
these reports despite significant optimization.
The difficulties associated with the fluorine-18 radiochemistry
of FMZ are mainly attributable to the labeling reaction, which pro-
ceeds by the SNAr mechanism at a position that is not activated by
an electron withdrawing group at the ortho- or para-positions. In
compounds 8, 9 and 10, the position of the fluorine is moved from
the 8- to the 7-position such that the carbonyl of the N-methyl
amide is at the ortho-position and is able to activate the aromatic
ring towards SNAr. To test this approach, the corresponding ni-
tro-radiolabelling precursor for compound 10, compound 24, was
prepared. The synthetic route outlined in Scheme 2 gave very
low yields (1%) when forming the imidazole ring. Therefore, a
new route was developed (Scheme 6).
cess (Scheme 1). Hydrolysis of the ethyl ester gave the carboxylic
acid in quantitative yield. This was treated with thionyl chloride
followed by the appropriate alcohol or amine to give the desired
compound FFMZ or 1–3.
The synthesis of 4–13, 19 and FEFMZ is shown in Scheme 2. The
general method for these compounds involves the synthesis of a
substituted isatoic anhydride from the corresponding substituted
2-aminobenzoic acid using triphosgene. Reacting the substituted
isatoic anhydride with an
a-amino acid gives a substituted 4-al-
kyl-3,4-dihydro-1H-benzo[e][1,4]-diazepine-2,5-dione. Here the
R3 group may be modified (e.g., methoxy to hydroxy and subse-
quently by alkylation to 2-fluoroethoxy). The next step for each
analogue involved the formation of the imidazole ring using
diethyl chlorophosphate and ethyl or tert-butyl isocyanoacetate.
Finally, where R2 is the protecting group 2,4-dimethoxybenzyl
(DMB), deprotection by treatment with acid or ceric ammonium
nitrate (where R1 = OtBu) followed by N-alkylation gives, for exam-
ple, 5, 6 or FEFMZ.
Compounds 15–18 were synthesised from the commercially
available compound Bretazenil (14) in two-steps (Scheme 3).
The synthesis of compound 20 is shown in Scheme 4.
2-Chloro-5-fluoropyridine-6-carboxylic acid was coupled to
ethyl 5-((methylamino)methyl)-1H-imidazole-4-carboxylate using
N,N0-carbonyldiimidazole. Cyclisation using sodium hydride in
dimethyl-formamide (DMF) at 100 °C gave compound 23 in 23%
yield. Compound 20 was formed using microwave irradiation
with cesium fluoride in DMF.
Compounds 21 and 22 were synthesised in three steps
(Scheme 5). Carboxylation of 2,4-difluoropyridine using standard
literature conditions gave 2,4-difluoronicotinic acid. The acid chlo-
ride was generated using oxalyl chloride, then reacted with tert-
butyl 5-((methylamino)methyl)-1H-imidazole-4-carboxylate in
37% yield. Cyclisation afforded 21 as the major product in 80%
yield, and 22 in 10% yield.
The affinity of compounds 1–22 for the GABAA receptor in vitro
was measured (Table 1) by a competitive radioligand binding assay
utilizing [3H]FMZ as the competitive agent.32 The data presented in
Table 1 show that substitution of the ester moiety for an amide is
not tolerated since compounds 2, 3, 17 and 18 did not displace
[3H]FMZ (up to 10 mM concentration). Also it appears that having
a longer alkyl chain at this position (as in compound 1) is not tol-
erated. Changing the position of the fluorine from the 8- to 7- and
changing ethyl ester to tert-butyl or isopropyl ester has minimal
impact on potency. Our results correlate with those reported by
Wong et al. and Cook and co-workers, where compound 8 had
comparable affinity to FMZ and related compounds with ester
groups having longer linear chains than propyl- gave diminished
affinity.31,33 Changing either a 7- or 8-fluoro to 7- or 8-fluoroeth-
oxy also appeared to be well tolerated.
A similar strategy was applied in compounds 20–22 where the
fluorine is ortho- to the amide, and/or ortho- or para- to the pyri-
dyl-nitrogen. The pyridyl-nitrogen is also able to promote SNAr,
so the precursors for 21 and 22 would be further activated. The
chloro-radiolabelling precursor 25 was prepared in Scheme 7.
The radio-labelled compounds with a fluoroethoxy group at the
7- or 8-position may be prepared indirectly by the alkylation of a
phenol with, for example, [18F]fluoroethyl tosylate. Therefore, the
phenolic precursor 26 was prepared by the general route outlined
in Scheme 2 (see Supplementary data for more detail on the syn-
thesis of 26).
The radiochemistry of selected compounds was assessed to
highlight possible advantages over [18F]FMZ (Scheme 8 and Ta-
ble 2). [18F]10 was prepared using a GE FASTlab™ module and
[
18F]21 was prepared using a method analogous to that used for
F
O
O
DMSO
Cs2CO3
RT, 12h
EtO
OEt
N
NO2
N
N
O
Changing the R2 substituent (as in compounds 5, 6, 7 and
FEFMZ) gave a small decrease in potency. For the compounds based
on Bretazenil (14), it was again shown that R1 amide analogues re-
sult in a decrease in affinity. However, interestingly, the larger es-
ter group in 16 did not cause a decrease in affinity. Literature
suggests that in a similar series of compounds, larger ester groups
such as cyclohexyl and cyclopropylmethyl cause a decrease in
affinity.31 Substitution of the tert-butyl ester in 14 for a fluoroethyl
ester as in 15 and substitution of the 7-bromo for an 8-fluoro sub-
stituent in 19 are tolerated. Compounds 20–23 all had low nano-
molar affinity. In summary, the binding assay showed that several
of the compounds have high affinity for the GABAA receptor and
may be suitable for further assessment as PET tracers.
72%
OEt
NO2
NH
EtO
O
(a)
(b)
61%
O
N
R1
N
R1 = OEt
(c), 88%
(d), 58%
R1 = OH
R1 = OtBu, 24
N
NO
2O
An important aspect of this study was to demonstrate that any
novel compounds with good affinity for the GABAA receptor, could
also be radio-labelled with fluorine-18 in significantly greater yield
Scheme 6. Synthesis of 24. Reagents and conditions (a) NBS, benzene, benzoyl
peroxide (cat), hV (60 W); (b) MeOH, DIPEA, MeNH2, reflux; (c) NaOH, H2O/EtOH,
reflux; (d) pyridine, tert-butanol, POCl3.
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066

A. Jackson et al. / Bioorg. Med. Chem. Lett. xxx (2012) xxx–xxx
5
Cl
O
Cl
O
pared with the EOS yields from our [18F]FMZ radiosynthesis, the
representative compounds 10, 12 and 21 can be prepared in signif-
icantly improved EOS yield. The measured specific activities ran-
HN
(b)
N
N
OH
N
N
28%
Cl
Cl
ged from 73 to 7 GBq/lmol, some of which are lower than ideal
(a)
O
O
O
and should be improved upon in future work. Furthermore, the
conditions used present no major obstacles for further develop-
ment of automated radiochemical synthesis.
O
(c)
44%
N
H
The two main criteria to determine whether a compound would
be selected for further assessment as a PET tracer were: affinity
comparable to FMZ, and improved radiochemistry. We have re-
ported an affinity screen of a number of known and novel GABAA
ligands having small structural variations. We have also reported
radiochemistry studies that demonstrate the feasibility of produc-
ing fluorine-18 labelled analogues in yields that would not hinder
the progression of these compounds as PET tracer candidates.
In conclusion, compounds 7, 8, 9, 10, 11, 12, 13, 15, 16, 20, 21
and 22 met the criteria to be assessed in an in vivo model to deter-
mine their pharmacokinetics and if appropriate, metabolism pro-
file and other properties for comparison to [18F]FMZ. This study
will be the subject of a future publication.
N
N
H
O
N
N
O
N
N
25
Cl
O
Scheme 7. Synthesis of 25. Reagents and conditions (a) (COCl)2, rt, 0.5 h, DCM, DMF
(1 drop); (b) rt, 24 h, DMF; (c) Cs2CO3, 60 °C, 2 h, CH3CN.
O
O
N
18F-, K2CO3
K2.2.2
DMF
N
N
OtBu
OtBu
N
Supplementary data
N
N
10-15%
EOS
18F
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmcl.2012.11.
066.
NO
O
2O
24
18
10
[
F]
18F
O
O
N
N
References and notes
OTs
N
OtBu
OtBu
18F
N
O
1. Persson, A.; Ehrin, E.; Eriksson, L.; Farde, L.; Hedstrom, C. C.; Litton, J. E.;
Mindus, P.; Sedvall, G. J. Psychiatr. Res. 1985, 19, 609.
2. Savic, I.; Persson, A.; Roland, P.; Pauli, S.; Sedvall, G.; Widen, L. Lancet 1988, 2,
863.
3. Niimura, K.; Muzik, O.; Chugani, D. C.; Shen, C.; Chugani, H. T. J. Nucl. Med. 1999,
40, 1985.
4. Savic, I.; Ingvar, M.; Stone-Elander, S. J. Neurol. Neurosurg. Psychiatry 1993, 56,
615.
5. Meyer, M.; Koeppe, R. A.; Frey, K. A.; Foster, N. L.; Kuhl, D. E. Arch. Neurol. 1995,
52, 314.
6. Yamauchi, H.; Kudoh, T.; Kishibe, Y.; Iwasaki, J.; Kagawa, S. J. Nucl. Med. 2005,
46, 1973.
7. Heiss, W. D.; Kracht, L.; Grond, M.; Rudolf, J.; Bauer, B.; Wienhard, K. Stroke
2000, 31, 366.
8. Heiss, W. D.; Sobesky, J.; Smekal, U.; Kracht, L. W.; Lehnhardt, F. G.; Thiel, A.
Stroke 2004, 35, 1892.
18-20%
EOS
N
N
O
HO
O
N
18
26
12
F]
[
O
O
18F-, K2CO3
K2.2.2
DMF
N
N
OtBu
OtBu
N
N
N
N
N
31%
EOS
18F
Cl
25
O
O
18
21
[
F]
9. Wicks, P.; Turner, M. R.; Abrahams, S.; Hammers, A.; Brooks, D. J.; Leigh, P. N.
Amyotrophic Lateral Scler. 2008, 9, 43.
Scheme 8. Radiosynthesis of [18F]10, [18F]12 and [18F]21.
10. Mehta, A. K.; Ticku, M. K. Brain Res. Rev. 1999, 29, 196.
11. Jakobsen, S.; Kodahl, G. M.; Olsen, A. K.; Cumming, P. Nucl. Med. Biol. 2006, 33,
593.
12. Goldfrank, L. R. Goldfrank’s Toxicologic Emergencies; McGraw-Hill Medical Publ.
Division: New York, 2002. ISBN 0-07-136001-8.
Table 2
Radiochemistry results
13. Sieghart, W. Pharmacol. Rev. 1995, 47, 181.
Compound Yield (EOS) (%)
21
RCP (%)
Specific activity (GBq/lmol)
14. Hammers, A. Neuroimaging Clin. N. Am. 2004, 14, 537.
15. Maziere, M.; Hantraye, P.; Prenant, C.; Sastre, J.; Comar, D. Int. J. Appl. Radiat.
Isot. 1984, 35, 973.
16. Halldin, C.; Stone-Elander, S.; Thorell, J.-O.; Persson, A.; Sedvall, G. J. Radiat.
Appl. Instrum., Part A. 1988, 39, 993.
[
[
[
18F]10
18F]12
18F]21
2
>99
>99
>99
32–73
12–29
7
18 1.5
31
17. Ryzhikov, N. N.; Seneca, N.; Krasikova, R. N.; Gomzina, N. A.; Shchukin, E.;
Fedorova, O. S. Nucl. Med. Biol. 2005, 32, 109.
18. Odano, I.; Halldin, C.; Karlsson, P.; Varrone, A.; Airaksinen, A. J.; Krasikova, R.
N.; Farde, L. NeuroImage 2009, 45, 891.
19. Massaweha, G.; Schirrmachera, E.; la Fougerea, C.; Kovacevica, M.; Wänglera,
C.; Jollya, D.; Gravela, P.; Readera, A. J.; Thielc, A.; Schirrmacher, R. Nucl. Med.
Biol. 2009, 45, 721.
20. Halldin, C.; Farde, L.; Litton, J.-E.; Hall, H.; Sedvall, G. Psychopharmacology 1992,
108, 16.
21. Lingford-Hughes, A.; Hume, S. P.; Feeney, A.; Hirani, E.; Osman, S.;
Cunningham, V. J.; Pike, V. W.; Brooks, D. J.; Nutt, D. J. J. Cereb. Blood Flow
Metab. 2002, 22, 878.
22. Kaneko, K.; Sasaki, M.; Morioka, T.; Koga, H.; Abe, K.; Sawamoto, H.; Ohya, H.;
Yoshiura, T.; Mihara, F.; Honda, H. Nucl. Med. Commun. 2006, 27, 893.
23. Torizuka, K.; Uemura, K.; Toru, M.; Shinohara, Y.; Nishimura, T.; Yonekura, Y.;
Nakagawara, J.; Matsuda, H.; Sakai, F.; Matsuda, K.; Fukuyama, H.; Morimoto,
K. Kakuigaku 1996, 33, 319.
[
18F]FMZ.38 For [18F]10, the EOS yield measured over three experi-
ments was 21 2%. For [18F]21, the single highest EOS yield over
two experiments was 31%. During the radiosynthesis, [18F]21 was
separated from unreacted precursor 25 by semi-preparative HPLC
(retention time of 15 vs 18 min for 25). [18F]12 was prepared using
a two-step process using [18F]fluoroethyl tosylate.39 The EOS yield
measured over three experiments was 18 1.5%. This two-step
radiosynthesis would be sufficient to enable further study of this
compound and could be implemented on an automated radio-
chemistry module. However, if 12 appeared to be particularly
attractive as a GABAA PET tracer, an alternative precursor for single
step radiolabelling could be prepared. Radiochemical purity for
24. Moerlein, S. M.; Perlmutter, J. S. Eur. J. Pharmacol. 1992, 218, 109.
25. Leveque, P.; Labar, D.; Gallez, B. Nucl. Med. Biol. 2001, 28, 809.
[
18F]10, [18F]12 and [18F]21 were all greater than 99%. When com-
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066

6
A. Jackson et al. / Bioorg. Med. Chem. Lett. xxx (2012) xxx–xxx
26. Grunder, G.; Siessmeier, T.; Lange-Asschenfeldt, C.; Vernaleken, I.; Buchholz,
HG.; Stoeter, P. Eur. J. Nucl. Med. 2001, 28, 1463.
250 ꢁ 10 mm column, isocratic elution with 75% 10 mM phosphoric acid and
25% acetonitrile, 4 mL/minute flow rate) gave the product
18F]FMZ with
[
27. Yoon, Y. H.; Jeong, J. M.; Kim, H. W.; Hong, S. H.; Lee, Y. S.; Kil, H. S.; Chi, D. Y.;
Lee, D. S.; Chung, J.-K.; Lee, M. C. Nucl. Med. Biol. 2003, 30, 521.
28. Mitterhauser, M.; Wadsak, W.; Wabnegger, L.; Mien, L.-K.; Togel, S.; Langer, O.
Nucl. Med. Biol. 2004, 31, 291.
29. Chang, Y. S.; Jeong, J. M.; Yoon, Y. H.; Kang, W. J.; Lee, S. J.; Lee, D. S. Nucl. Med.
Biol. 2005, 32, 263.
30. Dedeurwaerdere, S.; Gregoire, M.-C.; Vivash, L.; Roselt, P.; Binns, D.; Fookes, C.;
Greguric, I.; Pham, T.; Loc’h, C.; Katsifis, A.; Hicks, R. J.; O’Brien, T. J.; Myers, D. E.
Eur. J. Nucl. Med. Mol. Imaging 2009, 36, 958.
retention time 17.4 min which was collected using an electronic switching
valve on the TRACERlab into a flask containing water (25 mL). [18F]FMZ was
trapped on a C18 cartridge (Waters tC18 light), washed with water (3 mL) and
eluted with ethanol (0.5 mL) and phosphate buffered saline (4.5 mL) giving the
formulated product. EOS (non decay corrected) yield: 4.3 1.8%, n = 6.
35. As above, but: precursor solution is nitromazenil (sourced from Pharmasynth,
10 mg) in DMF (1 mL). Example yield: starting activity: 303 MBq @ time 14:43.
Product activity: 34.9 MBq @ time 16:10. See Supplementary data for details.
36. General radiosynthesis procedure is as for Ref. 34, but: reaction temperature is
31. Wong, G.; Koehler, K. F.; Skolnick, P.; Gu, Z. G.; Ananthan, S.; Schönholzer, P.;
Hunkeler, W.; Zhang, W.; Cook, J. M. J. Med. Chem. 1993, 36, 1820.
32. Briefly, a serial dilution of the test compound is incubated with a crude
homogenate of rat cerebellum in the presence of tritiated FMZ. After 90 min,
130 °C. Starting activity: 272 MBq @ time 11:10. Example yield: product
activity: 14 MBq @ time 12:58. See Supplementary data for details.
37. Three sources of nitromazenil precursor were used: (1) synthesised in-house,
(2) supplied by ABX, (3) supplied by Pharmasynth.
the assay is filtered through
homogenate and the bound radioligand. The amount of activity on the filter
mat is then measured using liquid scintillation.
a
glass fibre mat which filters out the rat
38. This was carred out on a GE FASTlab™ module. The general radiosynthesis
procedure is as for Ref. 34, but reaction temperature is 130 °C. Example yield:
starting activity: 637 MBq @ time 9:39. Product activity: 131 MBq @ time
11:15. See Supplementary data for details.
33. Huang, Q.; He, X.; Ma, C.; Liu, R.; Yu, S.; Dayer, C. A.; Wenger, G. R.; McKernan,
R.; Cook, J. M. J. Med. Chem. 2000, 43, 71.
39.
[
18F]Fluoroethyl tosylate was prepared using a FASTlab™ module and was
34. Radiosynthesis carred out on a GE TRACERlab FX F-N radiosynthesis module.
No carrier added [18F]fluoride was produced via 18O(p,n)18F reaction by proton
irradiation of an 18O-enriched water target of the GEMS PETtrace cyclotron.
The [18F]fluoride was trapped on an ion exchange cartridge (Waters QMA
SepPak Light) then eluted using a mixture of potassium carbonate (1.7 mg),
coupled to the phenolic precursor using manual non-automated methods. The
general procedure for [18F]fluoroethyl tosylate is as described in Ref. 33, but
the reaction conditions were 100 °C for 10 min. Following preparative HPLC,
the synthon was trapped on a cartridge (Waters tC18 light) and dried with
nitrogen gas. It was then eluted into
a second reaction vial containing
Kryptofix 2.2.2 (12 mg), water (80
l
l) and acetonitrile (1920
l
l). The solution
precursor and Cs2CO3 in DMF, heated (120 °C, 15 min), cooled, diluted with
water, purified by preparative HPLC followed by manual reformulation by
reversed phase SPE. See Supplementary data for detail.
was dried at 100 °C for 10 min then 120 °C for 20 min using nitrogen plus
vacuum flow and then cooled to 50 °C. A solution of nitromazenil (5–10 mg) in
N,N-dimethylformamide (0.6–1 mL) was added and the mixture heated at
160 °C for 30 min. The crude product was diluted with phosphoric acid
(10 mM, 2.5 mL). The diluted mixture was loaded into a 5 mL HPLC loop under
manual remote control. Semi-preparative HPLC (Phenomenex Prodigy ODS
40. Measured using the shake flask method with n-octanol and pH 7.4 buffer.
41. cLogP values were calculated using CambridgeSoft ChemDraw Ultra Version
12. The algorithm is licenced from BioByte Corp., 201 W. 4th St., #204,
Claremont, CA 91711-4707.
Please cite this article in press as: Jackson, A.; et al. Bioorg. Med. Chem. Lett. (2012), http://dx.doi.org/10.1016/j.bmcl.2012.11.066