Synthesis, radiofluorination and first evaluation of (±)-[ 18F]MDL 100907 as serotonin 5-HT2A receptor antagonist for PET

Article abstract of DOI:10.1002/jlcr.1563

In some psychiatric disorders 5-HT_2A receptors play an important role. In order to investigate those in vivo there is an increasing interest in obtaining a metabolically stable, subtype selective and high affinity radioligand for receptor binding studies using positron emission tomography (PET). Combining the excellent in vivo properties of [¹¹C]MDL 100907 for PET imaging of 5-HT_2A receptors and the more suitable half-life of fluorine-18, MDL 100907 was radiofluorinated in four steps using 1-(2-bromoethyl)-4-[¹⁸F]fluorobenzene as a secondary labelling precursor. The complex reaction required an overall reaction time of 140 min and (±)-[¹⁸F]MDL 100907 was obtained with a specific activity of at least 30 GBq/μmol (EOS) and an overall radiochemical yield of 1-2%. In order to verify its binding to 5-HT_2A receptors, in vitro rat brain autoradiography was conducted showing the typical distribution of 5-HT _2A receptors and a very low non-specific binding of about 6% in frontal cortex, using ketanserin or spiperone for blocking. Thus, [18F]MDL 100907 appears to be a promising new 5-HT_2A PET ligand. Copyright

Full text of DOI:10.1002/jlcr.1563

Research Article
Received 7 May 2005,
Revised 24 July 2008,
Accepted 7 October 2008
Published online 17 November 2008 in Wiley Interscience
(www.interscience.wiley.com) DOI: 10.1002/jlcr.1563
Synthesis, radiofluorination and first
evaluation of (7)-[18F]MDL 100907 as
serotonin 5-HT2A receptor antagonist for PET
a
Ã
a
b
¨
Ute Muhlhausen, Johannes Ermert, Matthias M. Herth,
and Heinz H. Coenena
In some psychiatric disorders 5-HT2A receptors play an important role. In order to investigate those in vivo there is an
increasing interest in obtaining a metabolically stable, subtype selective and high affinity radioligand for receptor binding
studies using positron emission tomography (PET). Combining the excellent in vivo properties of [11C]MDL 100907 for PET
imaging of 5-HT2A receptors and the more suitable half-life of fluorine-18, MDL 100907 was radiofluorinated in four steps
using 1-(2-bromoethyl)-4-[18F]fluorobenzene as a secondary labelling precursor. The complex reaction required an overall
reaction time of 140 min and (7)-[18F]MDL 100907 was obtained with a specific activity of at least 30 GBq/mmol (EOS) and
an overall radiochemical yield of 1–2%. In order to verify its binding to 5-HT2A receptors, in vitro rat brain autoradiography
was conducted showing the typical distribution of 5-HT2A receptors and a very low non-specific binding of about 6% in
frontal cortex, using ketanserin or spiperone for blocking. Thus, [18F]MDL 100907 appears to be a promising new 5-HT2A
PET ligand.
Keywords: radiofluorination; 1-(2-bromoethyl)-4-[18F]fluorobenzene; (7)-[18F]MDL 100907; 5-HT2A antagonist; PET
exhibits
a
high binding affinity to 5-HT2A receptors
Introduction
(Ki = 0.45–2.2 nM).10 However, it was shown that the BBB
permeability of MDL 100907 is more than four times that of
the metabolite MDL 105725, and that MDL 100907 does not
Serotonergic 5-HT2A receptors are of central interest in the
complex pathophysiology of human cerebral disorders such as
anxiety, depression, Alzheimer’s disease and schizophrenia.1–3 In undergo significant metabolism to MDL 105725 in the brain.10
Furthermore, 11C-labelled MDL 100907, with the radiolabel at
order to investigate the role of these receptors there is an
either the 20 or the 30 position, has been prepared and
increasing interest in obtaining a selective and high affinity
evaluated. Rat and baboon studies have been carried out with
the 20 labelled compound while the 30 labelled compound has
been evaluated in monkey and humans. Data from both animal
and human studies indicated that [11C]MDL 100907 labelled
either at the 20 or 30 position is a useful radioligand for in vivo
studies using PET.4,11,12 If 30-O-demethylation would play a
major role, than differences should have occurred with the 20-O-
radiomethylated analogue.
Nevertheless, a major draw back of [11C]MDL 100907 is the
short half-life of carbon-11 (20 min). This not only necessitates
the availability of a cyclotron close to the PET scanner but also
causes concern to reach a state of reversible binding during the
PET scan.7,11 These problems can be avoided by the use of
fluorine-18 with its half-life of 109.7 min. In summary [11C]MDL
radiolabelled ligand for direct in vivo receptor binding studies
using positron emission tomography (PET). For this purpose
several radiotracers have been developed, of which [11C]MDL
100907 and [18F]altanserin proved to be most useful.
Both MDL 100907 and altanserin bind to the 5-HT2A receptor
with high affinity. While for MDL 100907 a Ki value of 0.2 nM4
and a KD value of 0.56 nM (using [3H]MDL 100907)5 were
reported, for altanserin a Ki of 0.13 nM6 and a KD of 0.3 nM7 were
determined. Binding to other 5-HT receptor subtypes is very low
for MDL 100907,5 and for altanserin moderate to low.7 A further
difference between these two tracers is the binding to receptors
outside the serotonergic system. Altanserin shows a relatively
high affinity for D2 (62 nM) and especially for adrenergic-a1
receptors (4.55 nM),6,8 whereas the affinity of MDL 100907 to
these receptors is insignificant.9
Another disadvantage of altanserin is the formation of at least
four different metabolites in humans, which lower the bioavail-
ability and may cross the blood–brain barrier (BBB).8 On the
aInstitute of Neurosciences and Biophysics (INB-4): Nuclear Chemistry, Research
Center Ju¨lich GmbH, D-52425 Ju¨lich, Germany
other hand, pharmacokinetic studies with MDL 100907 in rats bInstitute of Nuclear Chemistry, University of Mainz, Fritz-Strassmann-Weg 2, D-
55128 Mainz, Germany
and dogs indicate that the drug undergoes extensive first-pass
metabolism that significantly reduces its bioavailability. The
*Correspondence to: Ute Mu¨hlhausen, Institute of Neurosciences and Biophysics
(INB-4): Nuclear Chemistry, Research Center Ju¨lich GmbH, D-52425 Ju¨lich,
Germany. E-mail: u.muehlhausen@fz-juelich.de.
major metabolite of MDL 100907 in animals has been found to
be 3‘-O-demethylated MDL 100907 (MDL 105725), which also
J. Label Compd. Radiopharm 2008, 52 6–12
Copyright r 2008 John Wiley & Sons, Ltd.

U. Mu¨hlhausen et al.
a) Huang et al.
R1
100907 is a very specific high affinity ligand for 5-HT2A receptors,
only hampered by the short half-life of C-11, while the binding
of [18F]altanserin lacks specificity.
COOEt
COOEt
For combining the superior pharmacological and pharmaco-
kinetic properties of MDL 100907 for PET imaging of 5-HT2A
receptors and the more suitable half-life of fluorine-18, the
isotopically labelled compound (7)-[18F]MDL 100907 was
prepared by a four step radiosynthesis. From the known
metabolism of MDL 100907 it might be expected that the
30-O-demethylated metabolite [18F]MDL 105725 will be formed
in vivo. The 20-labelled [11C]MDL 100907 proved to be useful
in rat and baboon even though [11C]MDL 105725 probably
occurs in vivo.12 Therefore, it appears worthwhile to find out if
the metabolism of 18F-labelled MDL 100907 will interfere in PET
measurements of 5-HT2A receptors in the brain. In vitro rat brain
autoradiography was performed in order to confirm the binding
of n.c.a. (7)-[18F]MDL 100907 in analogy to the cerebral pattern
of 5-HT2A receptors. In this study, aiming at establishing a
radiosynthesis of 18F-labelled MDL 100907, only the racemic
compound (7)-[18F]MDL 100907 was prepared but of course
the identical labelling procedure can also be performed with the
optically resolved precursor to yield (1)-[18F]MDL 100907.
N
N
H
b) Ullrich et al.
R2
R1,2
1, 2
OMe
O
N
OMe
O
MeO
N
N
R1,2
R1,2
5, 6
3, 4
a)
b)
R2: Boc
OMe
O
MeO
Results and discussion
NH
R1:
7
Chemistry
F
For the organic synthesis of MDL 100907 principally two
different procedures are described in the literature13–15 (cf.
Scheme 1). In the method by Huang et al.14 ethyl 1-(4-
fluorophenethyl)piperidine-4-carboxylate (1) is initially synthe-
sized by coupling of 2-(4-fluorophenyl)ethyl bromide to ethyl
isonipecotate. Then 1 is subsequently coupled to 1,2-dimethoxy-
benzene and reduced to MDL 100907.13,14 The method by
Ullrich et al.15 changes the reaction sequence using 1-tert butyl
4-ethyl piperidine-1,4-dicarboxylate (2) for coupling to 1,2-
dimethoxybenzene (veratrole). After reduction and deprotection
of the resulting piperidine derivative, reaction with 2-(4-
fluorophenyl)ethyl bromide was conducted yielding MDL
100907 (9). Even though the six step route of Ullrich et al.15
includes one more reaction step, it is better suited with respect
to enantiomeric purity of MDL 100907. The intermediate
(7)-(2,3-dimethoxyphenyl)-(piperidin-4-yl)methanol (8) can op-
tically be resolved15 and allows the introduction of a variety of
N-substituents. Compound 8 was used as the labelling precursor
in the work presented here. Both methods use an activated
Weinreb amide (3, 4) for coupling of the piperidine moiety to
1,2-dimethoxybenzene.
In order to improve and shorten the method by Ullrich et al.15
a five step synthesis was developed where instead of the
activated Weinreb amide 4 the aldehyde tert-butyl 4-formylpi-
peridine-1-carboxylate (11) was employed (cf. Scheme 2). For
this purpose commercially available piperidin-4-yl-methanol was
protected using di-tert-butyl dicarbonate. The resulting alcohol
10 was oxidized with oxalylchloride and DMSO in a Swern
oxidation to get 11. Subsequent reaction with 1,2-dimethoxy-
benzene, which was activated with sec-BuLi first, yielded
tert-butyl 4-((2,3-dimethoxyphenyl)(hydroxy)methyl)piperidine-
1-carboxylate (12). In contrast to the method by Ullrich et al.
no reduction of the coupling product is necessary, which saves
one reaction step. Deprotection of 12 with TFA yields (7)-(2,3-
dimethoxyphenyl)-(piperidin-4-yl)methanol (8), which can be
OMe OH
MeO
NH
8
OMe OH
MeO
N
F
9
MDL 100907
Scheme 1. Synthesis of MDL 100907 by Huang et al.14 (a) and Ullrich et al.15 (b).
alkylated with various phenylethyl bromides like 2-(4-fluorophe-
nyl)ethyl bromide. In order to get (7)-MDL 100907 (9) 8 was
deprotonated using NaHCO3 and reacted with 2-(4-fluorophe-
nyl)ethyl bromide as described by Ullrich et al. Using this
procedure the reaction route to 8 was abbreviated by one step
and the average overall yield was optimized (54 vs 43%15 or 48%
(own work)).
Radiochemistry
So far the synthesis of a suitable labelling precursor for the
radiofluorination to no-carrier-added (n.c.a.) [18F]MDL 100907
([18F]9) by direct nucleophilic substitution was not successful.
Therefore, a more complex labelling procedure was chosen
here. Using 1-(2-bromoethyl)-4-[18F]fluorobenzene ([18F]13) as a
J. Label Compd. Radiopharm 2008, 52 6–12
Copyright r 2008 John Wiley & Sons, Ltd.
www.jlcr.org

U. Mu¨hlhausen et al.
O
O
O
H
CH2OH
CH2OH
[K222][18F]F
Boc2O
Na2CO3
oxalylchloride
DMSO
O2N
18F
DMF, 135 °C
H2O/THF
100 °C
DCM
-50 °C
N
H
N
N
[
18F]14
Boc
10
Boc
11
O
OMe OH
veratrole
sec-BuLi
Br
Br2/HOAc/HCl
TES/TFA
TFA
MeO
chloroform/
ethyl acetate
100 °C
THF
- 50 °C
18F
90 °C
N
Boc
[
18F]15
12
OMe OH
MeO
Br
NH
18F
8
[
18F]13
Scheme 2. Improved synthesis of (7)-(2,3-dimethoxyphenyl)(piperidine-4-yl)-
methanol (8).
Scheme 4. Radiosynthesis of 1-(2-bromoethyl)-4-[18F]fluorobenzene ([18F]13).
OMe OH
4-nitroacetophenone as labelling precursor. The difference lies
in the bromination reaction of 4-[18F]fluoroacetophenone
([18F]14). While Hwang et al.16 use an acidic bromine solution,
Dence et al.17 conduct this reaction step with polymer-bound
perbromide. However, both reaction procedures could not be
reproduced as described. This was also observed in a study by
de Vries et al.18 who used the method by Dence et al.17 for the
synthesis of 2-bromo-40-[18F]fluoroacetophenone ([18F]15). The
labelling reaction was reproducible but despite considerable
effort the bromination step failed or gave poor yields. Therefore,
they could only produce an overall radiochemical yield (RCY) for
MeO
Br
+
NH
18F
[
18F]13
8
Cs2CO3/KI
DMF, 85 °C
[
18F]15 of 3.972.7% instead of 65%.17
After failure of the bromination reaction using the method by
Dence et al.17 the principal reaction sequence of Hwang
et al.16 was adopted here with some modifications of the
procedure (cf. Scheme 4).
OMe OH
MeO
As a first step commercially available 4-nitroacetophenone
was labelled with n.c.a. [18F]fluoride using Kryptofixs 222 and
potassium carbonate. The labelling was conducted in DMF as
solvent with conventional heating at 1301C. 4-[18F]Fluoroaceto-
phenone ([18F]14) was obtained with a RCY of 60–70%. To
remove the solvent and unreacted [18F]fluoride, solid phase
extraction (SPE) was used followed by passing through a drying
cartridge to remove water. Removal of water was essential for
the next reaction step in order to achieve radical bromination in
the side chain and to avoid side reactions.
After removal of the solvent, bromination was carried out in
chloroform/ethyl acetate, adding an acidic bromine solution
(hydrochloric and acetic acid) at a temperature of 1001C.
Analysis of the reaction mixture using reverse-phase high
performance liquid chromatography (HPLC) showed two radio-
active products. As described in the literature the desired 2-
bromo-40-[18F]fluoroacetophenone ([18F]15) as well as the
dibrominated 2,2-dibromo-40-[18F]fluoroacetophenone were
formed.16 The RCY of [18F]15 was 62–76%, that of the side
product 23–37%. An earlier report on a low fraction of o5% of
N
18F
( )-[18F]MDL 100907
18F]9
[
Scheme 3. Radiosynthesis of (7)-[18F]MDL 100907 ([18F]9).
secondary labelling precursor, the piperidine derivative (7)-(2,3-
dimethoxyphenyl)(piperidine-4-yl)methanol (8) was alkylated
yielding n.c.a. (7)-[18F]MDL 100907 ([18F]9) (cf. Scheme 3).
1-(2-Bromoethyl)-4-[18F]fluorobenzene ([18F]13)
The preparation of 1-(2-bromoethyl)-4-[18F]fluorobenzene
([18F]13) is described as a high-yield three-step synthesis by
Hwang et al.16 For the preparation of 2-bromo-40-[18F]fluoro-
acetophenone ([18F]15) basically two different methods
are described in the literature.16,17 Both of them use
www.jlcr.org
Copyright r 2008 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2008, 52 6–12

U. Mu¨hlhausen et al.
Figure 1. Autoradiography of [18F]MDL 100907 ([18F]9) distribution in horizontal rat brain sections: (a) total binding; (b) non-specific binding in the presence of ketanserin
(10 mM); (c) non-specific binding in the presence of spiperone (10 mM); cb, cerebellum; cx, cortex; st, striatum; t, thalamus.
dibrominated side product using the same reaction condi- In vitro autoradiography
tions16,19 could not be reproduced here. For the next reaction
step the solvent as well as the remaining acids and traces of
water had to be removed. This was accomplished by using an
For verifying the cerebral distribution of (7)-[18F]MDL 100907
([18F]9) according to the location of 5-HT2A receptors, in vitro
autoradiography with cryosections of rat brain was conducted.
Alumina N cartridge followed by a drying cartridge. Relative to
High binding of (7)-[18F]MDL 100907 ([18F]9) was observed in
the cortex with different intensities in various regions (cf.
Figure 1(a)). Especially in the frontal part 5-HT2A receptors have
been reported to be localized in the rat brain.20 Moderately high
binding was observed in striatum and low but still specific
binding was measured in thalamus and cerebellum. This binding
pattern is in good accordance with the distribution of 5-HT2A
[
18F]14 the RCY of [18F]15 after work up was 39–48%.
For reduction of [18F]15 to 1-(2-bromoethyl)-4-[18F]fluoro-
benzene ([18F]13), triethylsilane and TFA were used. The
reaction temperature was varied between 90 and 1001C. It
was observed that the RCY at lower temperatures was higher
and that formation of a side product, which supposedly is 2-
bromo-1-(40-fluorophenyl)ethanol, was less. Very problematic
with this reaction step was a loss of radioactivity, probably due receptors in rat brain20,21 and reflects the binding of [3H]MDL
100907 previously described.20 For blocking of (7)-[18F]MDL
to defluorination via formation of volatile triethylfluorosilane.
HPLC analysis of the reaction mixture showed a RCY for
1-(2-bromoethyl)-4-[18F]fluorobenzene ([18F]13) between 52
and 86%, depending on reaction temperature, while the side
100907 ([18F]9) binding ketanserin (10 mM) as well as spiperone
(10 mM) were used (cf. Figure 1 (b) and (c)). With both an
excellent displacement of (7)-[18F]MDL 100907 ([18F]9) was
product was formed in a RCY of 10–37%. Furthermore,
no [18F]15 could be detected anymore. For further reaction
[
order to remove water, which is critical to achieve coupling
of [18F]13 to (7)-(2,3-dimethoxyphenyl)(piperidine-4-yl)metha-
nol (8) (cf. Scheme 3), elution of purified [18F]13 after SPE was
conducted over a drying cartridge. [18F]13 was isolated with a
RCY of 13–26% relative to [18F]15. The overall RCY of [18F]13
relative to [18F]fluoride was 2.5–7% with a radiochemical purity
of 4 99% in a preparation time of 110 min.
achieved, resulting in a non-specific binding of only about 6% in
frontal cortex with both blockers. In relation to the total binding
(as average of the brain regions investigated) non-specific
binding was only about 15% with ketanserin and 10% with
spiperone. This reflects the reported pharmacological character-
istics of [3H]MDL 10090720 as expected, due to authentic
labelling of the compound with fluorine-18.
18F]13 had to be isolated by semi-preparative HPLC. In
Experimental
General
(7)-[18F]MDL 100907 ([18F]9)
The chemicals were analytical grade or better and were used
without further purification (Sigma-Aldrich, Steinheim, Ger-
many). Alumina N and Oasis HLB 1 cc cartridges were purchased
from Waters (Eschborn, Germany), LiChrolut RP-18e cartridges
from Merck (Darmstadt, Germany). Analytical TLC was per-
formed on aluminium-backed sheets (Silica gel 60 F254) and
normal-phase column chromatography was performed using
Silica gel 60, both from Merck (Darmstadt, Germany).
NMR spectra (1H-200 MHz) were obtained on a DPX Avance
200 spectrometer (Bruker, Rheinstetten, Germany). Chemical
shifts are reported in parts per million, solvent peaks were
referenced appropriately.
In order to get (7)-[18F]MDL 100907 ([18F]9), the piperidine
derivative 8 was deprotonated using caesium carbonate and
reacted with [18F]13 (Scheme 3). In contrast to observations by
Hwang et al.16 who tried to couple [18F]13 to piperidine under
various conditions, the formation of the elimation product of
[18F]13, 4-[18F]fluorostyrol, was moderate (o15%) under the
conditions applied here. (7)-[18F]MDL 100907 ([18F]9) was isolated
by semi-preparative HPLC with a RCY of 40–50%, relative to [18F]13.
The radiochemical purity was499% and the specific activity was at
least 30 GBq/mmol (end of synthesis). The preparation time of
(7)-[18F]MDL 100907 ([18F]9) was 140 min and the overall RCY
1–2% relative to [18F]fluoride. There is still room for optimization of
the reaction procedure concerning RCY as well as handling. This will
become even more attractive if ongoing work on a preferred direct
labelling approach is not successful.
HPLC was performed on the following system from Dionex
(Idstein, Germany): an Ultimate 3000 LPG-3400A HPLC pump, an
Ultimate 3000 VWD-3100 UV/VIS-detector (272 nm), a UCI-50
chromatography interface, an injection valve P/N 8215.
J. Label Compd. Radiopharm 2008, 52 6–12
Copyright r 2008 John Wiley & Sons, Ltd.
www.jlcr.org

U. Mu¨hlhausen et al.
Radioactivity was detected with a Gabi Star NaI(Tl) radio- tert-Butyl 4-((2,3-dimethoxyphenyl)(hydroxy)methyl)piperidine-1-
carboxylate (12)
activity-detector from Raytest (Straubenhardt, Germany) and
analysis of HPLC data accomplished with Chromeleon 6.80
software.
Under an atmosphere of argon, 1,2-dimethoxybenzene (1.30 mL,
10.0 mmol) dissolved in absolute THF (25 mL) was cooled to
ꢁ501C and sec-butyl lithium (1.3 M in n-hexane, 10 mL,
13.0 mmol) was added dropwise. The mixture was stirred at
room temperature for 75 min before cooling again to ꢁ501C.
Then compound 11 (2.0 g, 9.4 mmol) dissolved in absolute THF
(10 mL) was added and the mixture stirred at room temperature
for 3 h. The reaction was quenched with water (50 mL) and
extracted with diethyl ether (3 ꢀ 50 mL). The combined organic
phases were washed with brine (2 ꢀ 70 mL), dried over sodium
sulfate, and the solvent was evaporated under reduced pressure.
Purification by column chromatography (n-hexan:ethyl acetate,
Reversed-phase HPLC was carried out using a Gemini 5 mm
C18 110A column, for analytical separations with a dimension of
250 mm ꢀ 4.6 mm (flow 1 mL/min) and for semi-preparative
applications 250 mm ꢀ 10 mm (flow 5 mL/min) from Phenome-
nex (Aschaffenburg, Germany). Analysis or isolation of [18F]13,
[
18F]14, and [18F]15 was done by isocratic elution with
acetonitrile and water 65:35 (v/v). For [18F]MDL 100907 ([18F]9)
isocratic elution with acetonitrile, water and TEA 50:50:0.1 (v/v/v)
at a pH of 9.0 (phosphoric acid) was applied. For purpose of
identification of all radioactive products the respective non-
radioactive standard compounds were co-injected and co-
eluted with the radioactive products.
1
2:1) gave 12 (2.0 g, 5.7 mmol) in 61% yield. H NMR (DMSO-d6):
6.98 (m, 3H, ArH), 5.04 (d, 1H, OH), 4.63 (t, 1H, CHOH), 3.92 (s, 3H,
OMe), 3.85 (s, 3H, OMe), 2.64 (m, 3H, P2,6), 1.75 (m, 3H, P2,4,6),
1.39 (s, 9H, Boc), 1.26 (m, 4H, P3,5).
Phosphor imager plates were scanned with a laser phosphor
imager BAS 5000 (Fuji, Du¨sseldorf, Germany) utilizing
software from the vendor (Version 3.14, Raytest, Straubenhardt,
Germany). The resolution of a phosphor imager scan is
25 mm.
(7)-(2,3-Dimethoxyphenyl)-(piperidin-4-yl)methanol (8)
Radiochemical reactions were conducted in 5 mL conical vials
(Reactivial) from Wheaton Scientific (Millville, IL, USA), which
were closed with a silicon septum.
At 01C 12 (1.25 g, 3.65 mmol) was carefully dissolved in
trifluoroacetic acid (12.5 mL, 161 mmol). The resulting solution
was stirred at room temperature for 3 h. After addition of ethyl
acetate (50 mL) and water (30 mL) the mixture was adjusted at
pH 5 using sodium carbonate. The aqueous phase was extracted
with ethyl acetate (3 ꢀ 50 mL) and the combined organic phases
washed with brine (2 ꢀ 70 mL). After drying over sodium sulfate,
the solvent was evaporated under reduced pressure to give 8
Chemistry
tert-Butyl 4-(hydroxymethyl)piperidine-1-carboxylate (10)
1
(0.9 g, 3.60 mmol) in 99% yield. H NMR (DMSO-d6): 6.96 (m, 3H,
Piperidin-4-yl-methanol (5.02 g, 43.4 mmol) and sodium carbo-
nate (4.59 g, 43.2 mmol) were suspended in a mixture of water
(29 mL) and THF (11 mL). Di-tert-butyl dicarbonate (10.45 g,
51.6 mmol) was added and the resulting mixture stirred at 951C
for 2.5 h. After cooling to room temperature, water (100 mL)
was added and the product extracted with ethyl acetate
(3 ꢀ 50 mL). The combined organic phases were washed
with brine (2 ꢀ 70 mL) and after drying over sodium
sulfate the solvent was evaporated under reduced pressure.
This yielded the pure product 10 (9.34 g, 43.4 mmol) in 100%.
1H NMR (DMSO-d6): 3.98 (m, 2H, H2,6), 3.34 (s, 2H, CH2), 2.62
(m, 2H, H2,6), 1.52 (m, 3H, H3,4,5), 1.45 (s, 9H, Boc), 1.06
(m, 2H, H3,5).
ArH), 4.91 (d, 1H, OH), 4.58 (t, 1H, CHOH), 3.80 (s, 3H, OMe), 3.71
(s, 3H, OMe), 2.87 (m, 2H, P2,6), 2.34 (m, 2H, P2,6), 1.68 (m, 1H,
P4), 1.51 (s, 1H, NH), 1.14 (m, 4H, P3,5).
Radiochemistry
N.c.a. [18F]fluoride was produced via the 18O(p,n)18F nuclear
reaction by the bombardment of an isotopically enriched
[18O]water target with 17 MeV protons at the JSW cyclotron
BC 1710 (FZ Ju¨lich).22 The [18F]fluoride solution was azeotropi-
cally dried as described in the literature upon addition of
Kryptofixs 222 and potassium carbonate for anion activation.23
4-[18F]fluoroacetophenone ([18F]14)
tert-Butyl 4-formylpiperidine-1-carboxylate (11)
To the dry [K222][18F]F complex, 4-nitroacetophenone (2–3 mg,
E10 mmol), dissolved in absolute DMF (0.5 mL), was added and
the mixture stirred for 10 min at 1301C, which gave RCY of
60–70%. After dilution with water (20 mL), [18F]14 was fixed on a
conditioned Oasis HLB 1 cc cartridge (Waters), washed with
water (5 mL), dried with air and eluted with diethyl ether (4.5 mL)
through a glass column (LiChrolut 65 ꢀ 10 mm, Merck) filled
Under an atmosphere of argon a solution of oxalylchloride
(2 mL, 22.0 mmol) in absolute dichloromethane (50 mL) was
cooled to ꢁ501C. Absolute dimethylsulfoxide (3.5 mL,
44.0 mmol) in absolute dichloromethane (10 mL) was added
dropwise. Then 10 (4.30 g, 20.0 mmol) was added. After stirring
for 20 min at ꢁ501C triethylamine (14 mL) was added
˚
and the suspension was allowed to warm up to room with 4-A molecular sieves and sodium sulfate (170 mg). The
solvent was evaporated under a stream of argon at 700 mbar.
temperature. The reaction was stopped by the addition
of water (100 mL) after 2.5 h and the mixture extracted
with dichloromethane (4 ꢀ 50 mL). The combined organic
phases were washed with water (3 ꢀ 70 mL) and dried
over sodium sulfate. The solvent was evaporated under reduced
2-Bromo-40-[18F]fluoroacetophenone ([18F]15)
The residue of [18F]14 was taken up in choroform/ethyl acetate
(1:1, 1 mL) and an acidic solution of bromine (3.2 g Br2 in 25 mL
acetic acid and 1.25 mL concentrated hydrochloric acid; 0.1 mL)
was added. The mixture was stirred at 1001C for 6 min. In order
1
pressure to give 11 (4.27 g, 20.0 mmol) in 100% yield. H NMR
(DMSO-d6): 9.60 (s, 1H, CHO), 3.84 (m, 2H, H2,6), 2.96 (m, 2H,
H2,6), 2.51 (m, 1H, H4), 1.85 (m, 2H, H3,5), 1.44 (s, 9H, Boc), 1.42 to destroy the excess bromine at the end of the reaction, a 5%
aqueous solution of sodium bisulfite (1.5 mL) was added. After
(m, 2H, H3,5).
www.jlcr.org
Copyright r 2008 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2008, 52 6–12

U. Mu¨hlhausen et al.
separating the layers, the organic phase was diluted with diethyl 1-(2-bromoethyl)-4-[18F]fluorobenzene ([18F]13), synthesized as
ether (3.5 mL) and passed over a conditioned Alumina N
cartridge (Waters) followed by a glass column (LiChrolut
a secondary labelling precursor in three steps, could successfully
be coupled to the piperidine-derivative 8, to prepare
(7)-[18F]MDL 100907 ([18F]9) with specific activities of at least
30 GBq/mmol in an overall RCY of 1–2% and a preparation time
of 140 min. If on-going work on the principally more attractive
direct labelling approach to (7)-[18F]MDL 100907 ([18F]9) does
not succeed, the reaction procedure described here will serve as
a basis for optimization studies. Using optically resolved
(1)-(2,3-dimethoxyphenyl)(piperidine-4-yl)methanol as labelling
precursor in this approach will yield (1)-[18F]MDL 100907. In
vitro autoradiography using rat brain slices verified the excellent
distribution of (7)-[18F]MDL 100907 with high binding to 5-HT2A
receptors and very low non-specific binding of about 6%
in frontal cortex using ketanserin or spiperone for blocking.
In further studies the in vivo formation of the 30-O-demethylated
metabolite [18F]MDL 105725 has to be evaluated, especially
whether it interferes with PET measurements of 5-HT2A
receptors in the brain. Thus, it appears of high interest to
explore the potential of [18F]MDL 100907 ([18F]9) as a new
ligand for PET studies of 5-HT2A receptors.
˚
65 ꢀ 10 mm, Merck) filled with 4A molecular sieves and sodium
sulfate (170 mg). The RCY of [18F]15 after work up was 39–48%
relative to [18F]14. The solvent was evaporated under a stream
of argon at 700 mbar.
1-(2-Bromoethyl)-4-[18F]fluorobenzene ([18F]13)
For the reduction of [18F]15, triethylsilane (0.05 mL) and TFA
(0.5 mL) were added to the residue. The mixture was stirred at
90–951C for 15 min. After dilution with water (15 mL) it was
passed over a conditioned Oasis HLB 1 cc cartridge (Waters),
washed with water (5 mL), dried with air and eluted with
acetonitrile (1 mL). The solution was injected into a semi-
preparative reversed-phase HPLC. Elution occurred at k ’ = 3.71.
The RCY of isolated [18F]13 was 13–26% relative to [18F]15.
After dilution with water (20 mL) the collected fraction was fixed
on a conditioned Oasis HLB 1cc cartridge (Waters) again, washed
with water (5 mL) and eluted with dry DMF (1 mL) through a
˚
glass column (LiChrolut 65 ꢀ 10 mm, Merck) filled with 4A
molecular sieves and sodium sulfate (170 mg).
(7)-[18F]MDL 100907 ([18F]9)
Acknowledgement
The secondary labelling precursor [18F]13 (in 0.7 mL DMF) was
added to a suspension of (7)-(2,3-dimethoxyphenyl)(piperidine-
4-yl)methanol (8) (12.6 mg, 50 mmol), caesium carbonate
(16.3 mg, 50 mmol) and potassium iodide (8.3 mg, 50 mmol) in
dry DMF (50 mL) under an atmosphere of argon. The mixture was
stirred at 851C for 8 min and after dilution with water (15 mL)
The authors wish to thank Mrs Wiebke Sihver for radio-
pharmacological evaluation studies, Marcus H. Holschbach for
NMR measurements and Fabian Ku¨gler for support in organic
syntheses.
passed through
a conditioned LiChrolut RP18e cartridge
(Merck). The cartridge was washed with water (5 mL) and eluted
with acetonitrile (1 mL). The solution was injected into a semi-
preparative reversed-phase HPLC. Elution occurred at k ’ = 3.44.
The RCY of isolated [18F]9 was 40–50% relative to [18F]13. After
dilution with water (20 mL) the collected fraction was fixed on a
conditioned Oasis HLB 1 cc cartridge (Waters), washed with
water (5 mL) and eluted with ethanol (0.7 mL).
References
[1] M. Naughton, J. B. Mulrooney, B. E. Leonard, Hum. Psychopharma-
co. 2000, 15, 397–415.
[2] E. K. Perry, R. H. Perry, J. M. Candy, A. F. Fairbairn, G. Blessed,
D. J. Dick, B. E. Tomlinson, Neurosci. Lett. 1984, 51, 353–357.
[3] B. Dean, W. Hayes. Schizophr. Res. 1996, 21, 133–139.
[4] C. Lundkvist, C. Halldin, N. Ginovart, S. Nyberg, C.-G. Swahn,
A. A. Carr, F. Brunner, L. Farde, Life Sci. 1996, 58, 187–192.
[5] M. P. Johnson, B. W. Siegel, A. A. Carr, Naunyn-Schmiedeberg’s Arch.
Pharmacol. 1996, 354, 205–209.
[6] J. E. Leysen, in Neuromethods: Drugs as Tools in Neurotransmitter
Research (Eds.: A. A. Boulton, G. B. Baker, A. V. Juorio), Humana
Press, Clifton, NJ, 1989, pp. 299–350.
[7] H. Kristiansen, B. Elfving, P. Plenge, L. H. Pinborg, N. Gillings,
G. M. Knudsen, Synapse 2005, 58, 249–257.
[8] P. Z. Tan, R. M. Baldwin, C. H. Van Dyck, M. Al-Tikriti, B. Roth,
N. Khan, D. S Charney, R. B. Innis, Nucl. Med. Biol. 1999, 26,
601–608.
In vitro autoradiography
For autoradiography, frozen brains of Wistar rats were cut
horizontally at ꢁ181C into 20 mm thick slices (Leica AG
Microsystems, Germany), mounted onto gelatin-coated object
glasses and stored at ꢁ801C until use. For the binding assays,
the brain sections were thawed and dried at 231C and treated
after a literature protocol20 with some modifications. After
preincubation for 10 min at 231C in 50 mM Tris-HCl, pH 7.4, the
cryosections were incubated with 30 pM (7)-[18F]MDL 100907
([18F]3) in buffer for 45 min at 231C (n = 6 cryosections). Adjacent
sections were incubated in the presence of 10 mM ketanserin or
spiperone. After treatment the sections were washed twice for
5 min in Tris-HCl, dipped in deionized water, dried, and exposed
to a phosphor imager plate (Fuji).
[9] H. Hall, L. Farde, C. Halldin, C. Lundkvist, G. Sedvall, Synapse 2000,
38, 421–431.
[10] D. O. Scott, T. G. Heath, J. Pharm. Biomed. Anal. 1998, 17, 17–25.
[11] H. Ito, S. Nyberg, C. Halldin, C. Lundkvist, L. Farde, J. Nucl. Med.
1998, 39, 208–214.
[12] C. A. Mathis, K. Mahmood, Y. Huang, N. R. Simpson, J. M. Gerdes,
J. C. Price, Med. Chem. Res. 1996, 6, 1–10.
[13] A. A. Carr, J. M. Kane, D. A. Hay, PCT Int. Appl. 1991; WO 91/18602.
[14] Y. Huang, K. Mahmood, C. A. Mathis, J. Labelled Cpd. Radiopharm.
1999, 42, 949–957.
[15] T. Ullrich, K. C. Rice, Bioorg. Med. Chem. 2000, 8, 2427–2432.
[16] D.-R. Hwang, C. S. Dence, J. Gong, M. J. Welch, Appl. Radiat. Isot.
1991, 42, 1043–1047.
Conclusion
Combining the excellent in vivo properties of [11C]MDL 100907
with the more suitable half-life of fluorine-18, made the
radiosynthesis of (7)-[18F]MDL 100907 ([18F]9) attractive, even
by a rather complex four step radiosynthesis. For this purpose
[17] C. S. Dence, T. J. McCarthy, M. J. Welch, Appl. Radiat. Isot. 1993, 44,
981–983.
J. Label Compd. Radiopharm 2008, 52 6–12
Copyright r 2008 John Wiley & Sons, Ltd.
www.jlcr.org

U. Mu¨hlhausen et al.
[18] E. F. J. de Vries, J. Vroegh, P. H. Elsinga, W. Vaalburg, Appl. Radiat. [22] S. M. Qaim, J. C. Clark, C. Crouzel, M. Guillaume, H. J. Helmecke,
¨
B. Nebeling, V. W. Pike, G. Stocklin, PET radionuclide production, in
Isot. 2003, 58, 469–476.
[19] W. R. Banks, D. R. Hwang, Appl. Radiat. Isot. 1994, 45, 599–608.
¨
Radiopharmaceuticals for PET (Eds.: G. Stocklin, V. W. Pike), Kluwer
Academic Publishers, Dordrecht, 1993, pp. 15–26.
´
´
´
[20] J. F. Lopez-Gimenez, G. Mengod, J. M. Palacios, M. T. Vilaro,
¨
¨
¨
[23] H. H. Coenen, B. Klatte, A. Knochel, M. Schuller, G. J. Stocklin,
Naunyn-Schmiedeberg’s Arch. Pharmacol. 1997, 356, 446–454.
´
[21] A. Pazos, R. Cortes, J. M. Palacios, Brain Res. 1985, 346, 231–249.
J. Label. Compd. Radiopharm. 1986, 23, 455–466.
www.jlcr.org
Copyright r 2008 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2008, 52 6–12