Gas phase production of [11C]methyl iodide-d3. Synthesis and biological evaluation of S-[N-methyl-11C]citalopram and deuterated analogues

Article abstract of DOI:10.1002/jlcr.820

Three ¹¹C-labelled tracers for the serotonin reuptake site, S-[N-methyl-¹¹C]citalopram ([¹¹C]-4), S-[N-methyl-d ₃-¹¹C]citalopram ([¹¹C]-12), and S-[N-methyl-¹¹C]citalopram-α,α-d<

Full text of DOI:10.1002/jlcr.820

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2004; 47: 335–348.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.820
Research Article
Gas phase production of [11C]methyl iodide-d3.
Synthesis and biological evaluation
of S-[N-methyl-11C]citalopram and
deuterated analogues
Jacob Madsen1,*, Betina Elfving2, Kim Andersen3, Lars Martiny4
and Gitte M. Knudsen2
1 PET and Cyclotron Unit 3982, Copenhagen University Hospital, Blegdamsvej 9,
Copenhagen 2100, Denmark
2 Neurobiology Research Unit 9201, Copenhagen University Hospital, Blegdamsvej 9,
Copenhagen 2100, Denmark
3 Medicinal Chemistry Research, H. Lundbeck A/S, Ottiliavej 9, Valby 2500, Denmark
4
(
Isotope Chemistry, Novo Nordisk, Malꢀv 2760, Denmark
Summary
Three 11C-labelled tracers for the serotonin reuptake site, S-[N-methyl-11C]citalopram
([11C]-4), S-[N-methyl-d3-11C]citalopram ([11C]-12), and S-[N-methyl-11C]citalopram-
a,a-d2 ([11C]-13) were synthesized and the distribution of radioactivity after injection
of radioligand was examined ex vivo in rats. The deuterated analogue of (S)-
desmethylcitalopram, (S)-1-(4-fluorophenyl)-1-(3-methylamino-[3-d2]-propyl)-1,3-di-
hydro-isobenzofuran-5-carbonitrile (11), was synthesized in a multi-step synthesis
from escitalopram (4) and used as precursor in the synthesis of [11C]-13. In analogy
with the reported gas phase synthesis of [11C]methyl iodide the first gas phase synthesis
1
of [11C]Methyl iodide-d3 is reported. The H/2H kinetic isotope effect related to the
synthesized compounds were investigated in ex vivo rat studies, where the brain
regions of interest to cerebellum ratios of the tracers [11C]-4, [11C]-12 and [11C]-13 were
compared. The ex vivo data indicated no significant differences in binding in any of the
investigated brain regions after injection of the three tracers. Copyright # 2004 John
Wiley & Sons, Ltd.
Key Words: deuterium; specific radioactivity; serotonin reuptake sites
*Correspondence to: J. Madsen, PET and Cyclotron Unit 3982, Copenhagen University Hospital,
Blegdamsvej 9, Copenhagen 2100, Denmark. E-mail: j.madsen@rh.dk
Contract/grant sponsor: The Centre for Drug Design and Transport
Copyright # 2004 John Wiley & Sons, Ltd.
Received 14 October 2003
Revised 31 December 2003
Accepted 1 February 2004

336
J. MADSEN ET AL.
Introduction
In the last decades a range of selective serotonin reuptake inhibitors (SSRIs),
which have proven to be highly efficient in the treatment of major depression
and other psychiatric disorders, have been developed.1 Several of the marketed
SSRIs have been radiolabelled with positron-emitting isotopes such as carbon-
11 (t1/2=20.3 min) or fluorine-18 (t1/2=109.7 min) and investigated as
potential positron emission tomography (PET) tracers for measurements of
the serotonin transporters (SERTs) in the primate brain. Fluvoxamine (1),2,3
sertraline (2),4 citalopram (3)5–7 and escitalopram (4)8 have been labelled with
carbon-11, whereas fluoxetine (5) has been labelled with carbon-119 as well as
fluorine-18,10 Figure 1.
None of these marketed SSRIs do, however, exhibit appropriate in vivo
properties as PET ligands, mostly due to insufficient target to background
ratios.11–13 Recently, diarylsulfides such as [11C]DASB14,15 and [11C]MA-
DAM16 have shown promising results as PET tracers for the SERT in
humans.17,18
Escitalopram (4), the therapeutically active enantiomer of citalopram (3),19
displays high affinity for the SERT and very high selectivity over the dopamine
transporter (DAT) and the noradrenaline transporter (NAT).19,20 The log P
value of escitalopram (4) is 3.6.21 Recent studies of SERT tracers such as
[11C]DASB and other diaryl sulfides indicate that a log P value in the range of
2–3 is optimal in order to obtain a sufficient BBB penetration as well as low
non-specific binding.22 Based on these observations the log P value of
escitalopram (4) could be argued to be in the upper end of the optimal range
for a PET ligand. However, despite a thalamus to cerebellum ratio in rats of
1.8 60 min after injection of S-[N-methyl-11C]citalopram ([11C]-4) only a very
*
NHCH
F C
3
3
NC
O
*
O
N
NH
2
N
F
Cl
O
*
Cl
Fluvoxamine (1)
Sertraline (3)
Citalopram (3)
NC
*
*
O
O
NHCH
3
N
*
F C
3
F
Escitalopram (4)
Fluoxetine (5)
Figure 1. SSRIs currently available on the marked as antidepressant drugs. The
* indicates the position of the label
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
337
low target to background ratio was found in human volunteers studied with
PET.8
The major metabolite found in patients treated with citalopram (3) is
desmethylcitalopram,23 which has a 10-fold lower affinity for the SERT as
compared to citalopram (3).19 In one study where a single dose of 20 mg
citalopram infusion was given to healthy males, the only detectable metabolite
in plasma after 140 min was the propanoic acid derivative (the racemic form of
compound 9, Scheme 1).24 The enzymes cytochrome P450 and monoamine
oxidase (MAO) have been identified as responsible for the in vivo metabolism
of citalopram (3).25–28 Metabolism of citalopram will therefore also lead to
formation of secondary hydrophilic metabolites such as formaldehyde and
dimethylamine. Accordingly, in a PET study with S-[N-methyl-11C]citalopram
([11C]-4), radiolabelled metabolites like S-[N-methyl-11C]desmethylcitalopram,
[11C]formaldehyde and/or [11C]dimethylamine could be formed. It can be
speculated that [11C]formaldehyde and [11C]dimethylamine due to their limited
lipophilicity do not really leave the brain within the time window of a PET
study and might therefore result in non-specific binding.29–31 There are, to the
best of our knowledge, no reports on the in vivo metabolism of citalopram (3)
applied in tracer doses, which could be different from the pharmacokinetic
data obtained after treatment with pharmacological doses.
NC
NC
NC
O
O
O
H
N
N
Cl
c,d
a,b
7
4
6
F
F
F
e,f,g
NC
NC
NC
O
O
O
O
OH
OH
h
D
D
O
9
8
10
F
F
F
NC
O
H
N
i,j
D
D
11
F
Reaction conditions: a) ClCOOCH2CCl3, toluene, reflux, 3h, b) Zn, KH2PO4, THF, reflux, 2h, 70% (11% of
7 was formed). c) BrCH2COOEt, K2CO3, reflux, 3h, d) ClCOOEt, toluene, reflux, 3h, 52% (from 4 the
overall yield of 7 amounts to 47%). e) NaI, acetone, reflux, 18h, f) NaHCO3, DMSO, 120oC, 20 min.,
g) NaOH, AgNO3, 0oC, 30 min., 36%. h) NaBD4, BF3.Et2O, THF, 0oC-rt, 18h, 76%. i) MsCl, pyridine, 0oC-rt,
1h, j) H2NMe in THF (2M), 80oC, 18h, 78%.
Scheme 1. Synthesis of the deuterated S-desmethyl citalopram derivative 11
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

338
J. MADSEN ET AL.
We hypothesized that deuterated analogues of escitalopram labelled with
carbon-11 could delay the in vivo metabolism in the brain and thereby reduce
formation of radiolabelled metabolites. This would potentially lead to
increased target to background ratios if the deuterium labels were placed in
a position where the proton abstraction by MAO or cytochrome P450 takes
place in a rate limiting step. 1H/2H Kinetic isotope effects (KIE) have
previously been observed in ex vivo rat studies and in PET studies of MAO
activity in the brain when for example [11C]dimethylphenetylamine and [11C]L-
deprenyl were compared with the corresponding deuterated counterparts (a-
hydrogens substituted with deuterium).32–34
In this paper, we report the synthesis of S-desmethylcitalopram-a,a-d2 (11)
and the radiosynthesis of S-[N-methyl -11C]citalopram ([11C]-4), S-[N-methyl-
d3-11C]citalopram ([11C]-12) and S-[N-methyl-11C]citalopram-a,a-d2 ([11C]-
13), respectively. Comparative ex vivo rat studies of [11C]-4, [11C]-12 and
[11C]-13 were performed in order to evaluate the binding properties of the
deuterated radioligands for the SERT.
Results and discussion
Chemistry
The synthesis of the precursors, S-desmethylcitalopram (6) and deuterium
labelled S-desmethylcitalopram 11, were initiated from readily available
escitalopram (4) in order to obtain the target molecules as pure enantiomers,
Scheme 1.
Treatment of escitalopram (4) with 2,2,2-trichloroethyl chloroformate
followed by reduction with zinc dust gave S-desmethylcitalopram (6) and
the chloro derivative 7 in 70 and 11% yield, respectively. Treatment of S-
desmethylcitalopram (6) with ethyl bromoacetate followed by ethyl chloro-
formate yielded 7 in 52% yield. Thus, the combined overall yield of 7 was
47% calculated from the amount of escitalopram (4). The chloro derivative
7 was converted to the carboxylic acid derivative 9 in a three-step procedure.
Compound 7 was first converted to the corresponding iodide in a Finkelstein
reaction35 followed by oxidation with dimethyl sulfoxide36,37 to yield aldehyde
8. The aldehyde 8 was further oxidized with a mixture of silver nitrate and
sodium hydroxide38 to the carboxylic acid derivative 9 in 36% overall yield.
A key step in the synthesis of 11 was the selective reduction of 9 with a
deuterated reagent in order to introduce two deuterium atoms in the a-
position relative to the hydroxy group in the phthalane side chain. For this
purpose, deuterated diborane was freshly prepared by reaction of boron
trifluoride diethyl etherate and sodium borodeuteride.39–41 The suspension was
slowly added to the carboxylic acid derivative 9 in dry tetrahydrofuran. The
reduction preceded smoothly giving the deuterium labelled alcohol 10 in 76%
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
339
yield with complete incorporation of deuterium. Interestingly, it was noted
that the order of addition of the deuterated diborane reagent effected the
extent of deuterium incorporated in 10. Thus, adding sodium borodeuteride to
the carboxylic acid derivative 9 followed by addition of the boron trifluoride
diethyl etherate resulted in formation of the alcohol 10 with 4–5%
incorporation of hydrogen in the side chain.
Treatment of the deuterium labelled alcohol 10 with methylsulfonyl chloride
in pyridine followed by amination with methylamine in THF gave the desired
product 11 in 78% yield.
The optical purity of 11 was determined by chiral HPLC and was >99% as
in the starting material. 1H NMR confirmed the incorporation of deuterium in
11 due to the absence of signals at d 2.85 ppm corresponding to the deuterated
positions. LC–MS analysis further confirmed the introduction of two
deuterium atoms, since S-desmethylcitalopram (6) and its deuterated counter-
part 11 showed identical retention times on HPLC and the molecular ion of 11
was increased by 2 mass units compared to compound 6.
Radiochemistry
Synthesis of [11C]methyl iodide-d3 in a gas phase reaction has not previously
been described. [11C]Methyl iodide-d3 was obtained in analogy with the
reported procedure for gas phase synthesis of [11C]methyl iodide.42 Bombard-
ing a gas target containing a mixture of N2 and D2 (95/5) with 16 MeV protons
produced [11C]methane-d4. [11C]Methane-d4 was in several cycles passed
through a quartz tube at 7208C containing I2 vapor and [11C]methyl iodide-d3
was synthesized. By changing the target gas from N2/D2 to N2/H2 [11C]methyl
iodide was produced as described above.42 The radiosynthesis of S-[N-
methyl-11C]citalopram ([11C]-4), S-[N-methyl-d3-11C]citalopram ([11C]-12) and
S-[N-methyl-11C]citalopram-a,a-d2 ([11C]-13), Scheme 2, were performed in
ethanol and purified on a normal phase semi-preparative HPLC column.7
By applying normal phase HPLC purification the 11C-labelled product
eluted before the S-desmethyl precursor and contamination of the product
NC
NC
11
CR
N
3
O
O
11CR3I
H
N
X X
PMP, EtOH, ∆
X X
6
X = H
F
F
11 X = D
[
[
11C]-4 R = H, X = H
11C]-12 R = D, X = H
[
11C]-13 R = H, X = D
Scheme 2. Radiosynthesis of S-[N-methyl-11C]citalopram ([11C]-4), S-[N-
methyl-d3-11C]citalopram ([11C]-12), and S-[N-methyl-11C]citalopram-a,a-d2
([11C]-13)
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

340
J. MADSEN ET AL.
with the precursor was thereby avoided. In addition, the synthesis time was
minimized due to quick evaporation of the low boiling organic solvents. After
evaporation of the solvents the residue was formulated in 2% ethanol. The
radiochemical yield of the final product was 30–40% (decay corrected
radiochemical yield based on the amount of trapped [11C]methylation agent).
By reverse phase analytical HPLC, the identity of the product was confirmed
by co-elution with an authentic sample of citalopram (3). The radiochemical
purity was determined to be >98%.
The specific radioactivity of the purified products (synthesis time was
45 min) was found to be in the range of 37–130 GBq/mmol at the end of
synthesis. Increasing the beam current and the irradiation time might increase
the specific radioactivity. The incorporation of the deuterium label in S-[N-
methyl-d3-11C]citalopram ([11C]-12) was confirmed by GC–MS and LC–MS
analysis. After decay mass spectra with molecular ions at m/z: 327[M+] and
328[M++1] corresponding to the carbon-12 analog of [11C]-12 was obtained.
The difference of 3 mass units (the molecular weight of citalopram is 324)
corresponds to the incorporation of three deuterium molecules in the CD3
unit. No compound with a molecular ion at m/z: 325[M++1] corresponding
to citalopram (3) could be identified.
Ex vivo studies
S-[N-Methyl-11C]citalopram ([11C]-4), S-[N-methyl-d3-11C]citalopram ([11C]-
12) or S-[N-methyl-11C]citalopram-a,a-d2 ([11C]-13) were injected into the tail
vein of the rats. Each rat received 5–25 MBq of the ligand in question with a
specific radioactivity ranging from 10 to 18 GBq/mmol, at the time of injection.
With the given specific radioactivity and assuming an SERT density in rat
thalamus of about 11 fmol/mg protein (B. Elfving, unpublished data) the
occupancy in the rat thalamus was calculated to constitute ꢀ0.2% and can
therefore be considered to be negligible. The optimal thalamus to cerebellum
ratio for carbon-11 labelled escitalopram has previously been shown to be
reached 60 min after tracer injection.7,8 At this time point the rats were
decapitated and the brains were dissected, weighed and the radioactive content
of the dissected regions was measured. Ratios between the radioactive content
of the regions of interest (ROI) and the reference region, cerebellum, was
determined (Figure 2). Cerebellum was used as reference region due to the low
density of SERTs.43 Pretreatment with paroxetine resulted in thalamus to
cerebellum ratios of approximately 1.2, indicating that the majority of the
specific binding could be blocked (data not shown).
As seen in Figure 2, there were no significant differences between the ROI to
cerebellum ratios for [11C]-4, [11C]-12 and [11C]-13 in any of the investigated
regions.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
341
S-[N-methyl-11C]citalopram ([11C]-4)
S-[N-methyl-d3-11C]citalopram ([11C]-12)
S-[N-methyl-11C]citalopram-α,α-d2 ([11C]-13)
5
4
3
2
1
0
Brain region
Figure 2. Ratios between six ROIs and cerebellum (mean Æ S.D., n=3) 60 min
after tracer injection
1
The absence of a H/2H KIE could indicate that the tracers were not
metabolized in the brain. It can, however, not be ruled out that brain
metabolism of carbon-11 labelled escitalopram by cytochrome P450 or MAO
actually takes place leading to formation of non-specific radiolabelled
metabolites. This would be the case if the in vivo metabolism of [11C]-12 and
[11C]-13 only to a minor extent was affected by the deuterium substitutions.
In conclusion, we have developed an efficient procedure to produce
[11C]methyl iodide-d3 with high specific radioactivity (up to 130 GBq/mmol)
in a gas phase reaction. The gas phase methylation system can produce either
[11C]methyl iodide-d3 or [11C]methyl iodide by changing the target gas. S-[N-
methyl-11C]citalopram ([11C]-4), S-[N-methyl-d3-11C]citalopram ([11C]-12) and
S-[N-methyl-11C]citalopram-a,a-d2 ([11C]-13) were synthesized in high specific
radioactivity and evaluated in ex vivo rat studies. No significant differences
between the ROI to cerebellum ratios for [11C]-4, [11C]-12 and [11C]-13 were
found in any of the investigated regions. The substitution of deuterium did
thus not have any significant effect on the ROI to cerebellum ratios. Whether
the rate of brain metabolism was largely unaffected by the deuterium
substitutions or brain metabolism by cytochrome P450 and MAO is minimal
is, however, not clear. A more direct way of addressing this question would be
by conducting direct measurements of the labelled metabolite content in the
rat brain.
Experimental
General
Chemicals were purchased from commercial suppliers and used without
further purification. THF was distilled under N2 from sodium-benzophenone.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

342
J. MADSEN ET AL.
Purifications were performed by flash chromatography (silica gel, 230–400
mesh). NMR spectra were obtained on a Bruker 500 instrument operating for
1H at 500 MHz and for 13C at 125 MHz. Chemical shifts are reported in ppm
using tetramethylsilane (TMS) as internal standard. Coupling constants are
given in Hz and coupling patterns are abbreviated as s=singlet, d=doublet,
t=triplet, q=quintet and m=multiplet. Melting points were determined with
.
a Buchi B-540 apparatus and are uncorrected.
Gas chromatograms and mass spectra were obtained on a Varian Saturn
2000 GC/MS spectrometer equipped with a phenomenex ‘Zebran 2B-5’ (L:
15 m, i.d. 0.25 mm) column. Column temperature started at 608C and was
linearly increased to 3008C in 12 min and maintained there for 2 min. LC–MS
was performed using an API 150EX LC/MS system from Applied Biosystems.
Chiral HPLC was performed on a Gilson Series SF3 System for packed-
column supercritical fluid chromatography (Column: Diacel OD,
4.6 Â 250 mm, 10 mm, 10% modifier: methanol with 0.5% trifluoroacetic acid
and 0.5% diethylamine, flow: 2 ml/min).
All irradiations were carried out with the Scanditronix MC32 cyclotron at
the PET and Cyclotron Unit at Copenhagen University Hospital, Rigshos-
pitalet. [11C]Methane-d4 and [11C]methane were produced in the 14N(p,a)11C
nuclear reaction by bombarding an aluminum gas target containing a mixture
of nitrogen (AGA 6.0 nitrogen) and 5% deuterium (AGA 2.8 deuterium) or
nitrogen (AGA 6.0 nitrogen) and 5% hydrogen (AGA 2.8 hydrogen) gas,
respectively. Radiolabelled compounds were purified on a semi-preparative
normal phase HPLC column and analyzed on a reverse phase analytical
HPLC column at room temperature using Gilson 306 pumps and a Gilson 117
UV detector (254 nm) followed by a radioactivity detector.
Citalopram (3) and escitalopram (4) were gifts from H. Lundbeck A/S,
Copenhagen, Denmark.
Chemistry
(S)-1-(3-Methylamino-propyl)-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-5-
carbonitrile (6). A solution of escitalopram (4) (23.0 g, 71 mmol) dissolved in
toluene (100 ml) was boiled under reflux when 2,2,2-trichloroethyl chlorofor-
mate (16.0 ml, 0.12 mol) was added drop wise during 30 min. After addition,
the mixture was boiled under refluxed for further 2 h. After cooling to room
temperature, Et2O (100 ml) and 4 M aqueous HCl (50 ml) was added and the
phases were separated. The aqueous phase was extracted with Et2O
(2 Â 100 ml). The combined organic phases were dried (MgSO4) and the
volatile solvents evaporated in vacuo yielding 43.4 g of yellow oil. The crude
mixture was dissolved in THF (200 ml) and 1.5 M aqueous KH2PO4 (200 ml)
and heated to 508C. To the mechanically stirred solution zinc dust (50 g,
0.76 mol) was added portion wise over 30 min. The resulting mixture was
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
343
boiled under reflux for further 2 h. After cooling to room temperature the
suspension was filtered through celite. The phases were separated and the
organic layer was evaporated in vacuo to give a mixture of 6 and 7. Et2O
(200 ml) and water (200 ml) were added to the residue and the aqueous phase
was acidified with 4 M aqueous HCl. The phases were separated and the
aqueous phase was extracted with Et2O (3 Â 100 ml). The combined organic
phases were dried (MgSO4) and the solvents evaporated in vacuo. The residue
was purified by column chromatography (EtOAc/heptane 1:3) yielding 2.4 g
(11%) of 7 as a colorless oil. The remaining aqueous phase was made alkaline
with 25% aqueous NH3 and extracted with Et2O (3 Â 200 ml). The combined
organic phases were dried (MgSO4) and the solvent was evaporated in vacuo
yielding 15.3 g (70%) of 6 as an yellow oil.
LC/MS (m/z): 311 (M++1), purity >97%.
1H NMR (CDCl3) d 1.35 (m, 1 H), 1.48 (m, 1 H), 2.15 (m, 1 H), 2.23
(m, 1 H), 2.36 (s, 3 H), 2.85 (t, J=7.1 Hz, 2 H), 5.15 (d, JAB=13.0 Hz, 1 H),
5.19 (d, JAB=13.0 Hz, 1 H), 7.00 (t, J=8.5 Hz, 2 H), 7.40 (m, 3 H), 7.50
(s, 1 H), 7.58 (d, J=8.0 Hz, 1 H).
13C NMR (CDCl3) d 24.3, 36.4, 39.0, 51.8, 71.2, 91.1, 111.6, 115.1,
115.3, 118.5, 122.7, 125.1, 126.6, 126.7, 131.7, 139.5, 140.3, 149.3, 162.0
(d, JC–F=239 Hz).
(S)-1-(3-Chloropropyl)-1-(4-fluorophenyl)-1,3-dihyroisobenzofuran-5-carboni-
trile (7). A mixture of S-desmethylcitalopram (6) (15.0 g, 48 mmol) in EtOH
(100 ml), K2CO3 (13.8 g, 100 mmol) and ethyl bromoacetate (6.0 ml, 54 mmol)
was boiled under reflux for 4 h. After cooling to room temperature the mixture
was filtered and the volatile solvents were evaporated in vacuo. The remaining
residue was dissolved in Et2O (200 ml) and washed with brine. The aqueous
phase was extracted with additional Et2O (2 Â 200 ml). The combined organic
phases were dried (MgSO4) and the solvent was evaporated in vacuo yielding
16.3 g of a yellow oil. The residue was dissolved in toluene (100 ml) and
the resulting mixture was added dropwise to a refluxing mixture of
ethyl chloroformate (42.1 ml, 0.44 mol) in toluene (100 ml) and boiled under
refluxed for 2 h. After cooling the low boiling components were removed
in vacuo. Purification by column chromatography (EtOAc/heptane 1:3) gave
8.1 g (52%) of 7 as a colorless oil. The overall yield of 7 from 4 amounts to
10.5 g (47%) taking into account the amount of 7 formed in the dealkylation
of 4.
1H NMR (CDCl3) d 1.66 (m, 1 H), 1.79 (m, 1 H), 2.26 (m, 1 H), 2.34
(m, 1 H), 3.56 (m, 2 H), 5.14 (d, JAB=13.2 Hz, 1 H), 5.19 (d, JAB=13.2 Hz,
1 H), 7.02 (t, J=8.5Hz, 2 H), 7.42 (m, 3 H), 7.52 (s, 1 H), 7.61 (d, J=8.0 Hz, 1 H).
13C NMR (CDCl3) d 27.3, 38.5, 45.0, 71.3, 90.8, 111.9, 115.4, 115.5, 118.6,
122.7, 125.3, 126.7, 126.8, 132.0, 139.1, 140.3, 149.1, 162.1 (d, JC–F=252 Hz).
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

344
J. MADSEN ET AL.
(S)-3-[5-Cyano-1-(4-fluorophenyl)-1,3-dihydro-isobenzofuran-1-yl]-propionic
acid (9). A solution of 7 (2.4 g, 7.6 mmol) and sodium iodide (20 g, 133 mmol)
in acetone (20 ml) was boiled under reflux overnight. After cooling to room
temperature, brine (100 ml) was added. The aqueous phase was extracted with
Et2O (3 Â 100 ml) and the combined organic phases were dried (MgSO4) and
the solvents were evaporated in vacuo. The remaining oil (3.2 g) was dissolved
in DMSO (50 ml) and heated to 1208C. Sodium hydrogencarbonate (10 g) was
added and the mixture was stirred for further 20 min at 1208C. The hot
mixture was poured into ice-water and extracted with Et2O (3 Â 100 ml). The
combined organic phases were dried (MgSO4) and evaporated in vacuo giving
2.4 g of 8 as a crude product. A small sample was purified by column
chromatography (EtOAc/heptane 1:3) yielding 45% of 8.
1H NMR (CDCl3) d 2.35 (m, 2 H), 2.45 (m, 1 H), 2.59 (m, 1 H), 5.13 (s, 2 H),
7.00 (t, J=8.7 Hz, 2 H), 7.42 (m, 3 H), 7.52 (s, 1 H), 7.61 (d, J=8.0 Hz, 1 H),
9.70 (s, 1 H).
13C NMR (CDCl3) d 33.6, 38.9, 71.0, 90.3, 112.0, 115.4, 115.5, 118.4, 122.6,
125.3, 126.6, 126.6, 132.0, 138.7, 140.2, 148.3, 162.0 (d, JC–F=252 Hz), 201.0.
A solution of crude 8 (2.4 g) in ethanol (25 ml) was added to a mixture of
2.5 M aqueous AgNO3 (8 ml) and 5 M aqueous NaOH (8 ml) at 08C and
stirred for 30 min. Water was added and the brown precipitate was filtered
off. The filtrate was extracted with Et2O (100 ml). The phases were
separated and the organic phase was discarded. The aqueous phase was
acidified with 4 M aqueous HCl and extracted with Et2O (3 Â 100 ml).
The combined organic phases were dried (MgSO4) and the solvents were
evaporated in vacuo. Purification by column chromatography (EtOAc/heptane
1:3+1% CH3COOH) gave 860 mg (36%) of 9 as a colorless oil.
1H NMR (CDCl3) d 2.30 (m, 2 H), 2.45 (m, 1 H), 2.56 (m, 1 H), 5.14
(d, JAB=13.0 Hz, 1 H), 5.19 (d, JAB=13.0 Hz, 1 H), 7.02 (t, J=8.7 Hz, 2 H),
7.42 (m, 3 H), 7.51 (s, 1 H), 7.60 (d, J=8.0 Hz, 1 H), 10.0–11.0 (broad s, 1 H).
13C NMR (CDCl3) d 29.1, 35.8, 71.3, 90.2, 112.0, 115.4, 115.5, 118.4, 122.7,
125.3, 126.6, 126.7, 131.9, 138.6, 140.2, 148.3, 162.0 (d, JC–F=252 Hz), 179.1.
(S)-1-(4-Fluorophenyl)-1-(3-hydroxy-[3-d2]-propyl)-1,3-dihydro-isobenzofur-
an-5-carbonitrile (10). Boron trifluoride diethyl etherate (300 ml, 2.44 mmol)
was added under N2 to a suspension of sodium borodeuteride (77 mg,
1.83 mmol) in dry THF (5 ml) cooled to 08C. The mixture was stirred at 08C
for 1 h. The suspension was slowly (1 h) added to a solution of 9 (475 mg,
1.53 mmol) in dry THF (5 ml) cooled to 08C. The reaction mixture was
allowed to warm to room temperature and stirred over night. Ice and
saturated aqueous NaHCO3 was added and the resulting mixture was
extracted with Et2O (3 Â 75 ml). The combined organic phases were dried
(MgSO4) and the solvent evaporated in vacuo. Purification by column chro-
matography (EtOAc/heptane 1:3) gave 345 mg (76%) of 10 as a colorless oil.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
345
1H NMR (CDCl3) d 1.44 (m, 1 H), 1.55 (m, 1 H), 1.88 (broad s, 1 H), 2.21
(m, 1 H), 2.31 (m, 1 H), 5.15 (d, JAB=13.0 Hz, 1 H), 5.20 (d, JAB=13.0 Hz,
1 H), 7.01 (t, J=8.7Hz, 2 H), 7.41 (m, 3 H), 7.50 (s, 1 H), 7.59 (d, J=7.1 Hz, 1 H).
1
13C NMR (CDCl3) d 27.0, 37.6, 61.8 (q, JC–D=21.6 Hz, CD2), 71.2, 91.0,
111.7, 115.2, 115.4, 118.5, 122.7, 125.2, 126.6, 126.7, 131.8, 139.2, 140.1, 149.2,
162.0 (d, JC–F=238 Hz).
(S)-1-(4-Fluorophenyl)-1-(3-methylamino-[3-d2]-propyl)-1,3-dihydro-isobenzo-
furan-5-carbonitrile (11). A solution of 10 (266 mg, 0.89 mmol) in pyridine
(1.3 ml) was cooled to 08C. Methanesulfonyl chloride (84 ml, 1.07 mmol) was
added and the resulting mixture was stirred for 30 min at 08C followed by
30 min at room temperature. Water (50 ml) was added and the resulting mixture
was extracted with toluene (3 Â 75 ml). The combined organic phases were
dried (MgSO4) and the solvents were evaporated in vacuo and re-evaporated in
vacuo 3 times from toluene. The crude product was dissolved in a 2 M solution
of methylamine in THF (6ml) and heated at 808C for 15 h in a sealed tube.
After cooling to room temperature, the solvents were evaporated in vacuo to
give 217 mg (78%) of 11. The oxalate was crystallized from acetone yielding
130 mg (36%).
Anal. (C19H17FN2OD2.C2H2O4) calcd. C: 62.68, H: 5.26, N: 6.96. Found C:
62.25, H: 5.45, N: 6.82. Mp.=170–1728C. LC–MS (m/z): 313 [M++1].
1H NMR (DMSO-d6) d 1.39 (m, 1 H), 1.48 (m, 1 H), 2.25 (m, 2 H), 2.46
(s, 3 H), 5.16 (d, JAB=13.4 Hz, 1 H), 5.21 (d, JAB=13.4 Hz, 1 H), 7.20
(t, J=8.9 Hz, 2 H), 7.60 (m, 3 H), 7.74 (d, J=7.5 Hz, 1 H), 7.80 (d, J=8.0 Hz,
1 H).
13C NMR (DMSO-d6) d 20.4, 32.2, 36.9, 47.4 (q, 1JCD=18.8 Hz, CD2), 71.0,
90.2, 110.6, 115.0, 115.2, 118.7, 123.0, 125.7, 126.8, 126.9, 132.0, 139.8, 140.0,
148.8, 161.3 (d, JC–F=238 Hz), 164.6.
The final product was analyzed on a chiral HPLC column applying
supercritical fluid flow conditions. No indications of racemisation could be
determined, ee>98% as in the starting material.
Radiochemistry
Synthesis of [11C]methyl iodide-d3. Bombarding a gas target (20 mA, 10 min)
containing a mixture of N2 and D2 (95/5) with 16 MeV protons produced
[11C]methane-d4. [11C]Methane-d4 was transferred from the target to the hot
cell and trapped on a Porapak N column cooled to À1908C with liquid
nitrogen. After trapping the Porapak N trap was heated to room temperature
and [11C]methane-d4 was liberated. In several cycles [11C]methane-d4 was
passed through a quartz tube at 7208C containing I2 vapor. At the end of each
cycle [11C]methyl iodide-d3 was trapped on the Porapak N column at room
temperature. Unreacted [11C]methane-d4 was re-circulated through the quartz
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

346
J. MADSEN ET AL.
tube until the radioactivity at the Porapak N column reached a maximum. At
this point [11C]methyl iodide-d3 was liberated by purging the Porapak N trap,
now at 1908C, with helium. From a typical bombardment (ꢀ3.4 mAh) 5–7
GBq of [11C]methyl iodide-d3 was obtained. By changing the target gas to
N2/H2 (95/5) [11C]methyl iodide was obtained similarly.42
Radiosynthesis of S-[N-methyl-d3-11C]citalopram ([11C]-12), S-[N-methyl-11C]-
citalopram ([11C]-4) and S-[N-methyl-11C]citalopram-a,a-d2 ([11C]-13). In
separate runs [11C]methyl iodide or [11C]methyl iodide-d3 was trapped in a
reaction vessel cooled to À458C containing EtOH (740 ml), the S-desmethyl-
precursor (6 or 11) (4–8 mg) and 1,2,2,6,6 pentamethyl piperidine (6 ml). After
trapping the sealed reaction vessel was heated to reflux for 5 min followed by
evaporation of the solvent leaving 50–100 ml in the vessel. The residue was
dissolved in chloroform (1 ml), which was automatically injected to the semi-
preparative HPLC column. The product was collected after 7.5 min. The
solvent was evaporated with a stream of nitrogen at atmospheric pressure at
708C and the product was formulated in 2% EtOH in water (2 ml).
Preparative HPLC: The column used was a Knauer nucleosil-100 Si (5 mm,
250 Â 8 mm). The products were eluted isocratic with as mixture of
CHCl3:CH3OH:28% NH4OH (90:10:0.1) and a flow rate of 3 ml/min. The
analytical column used was a Merck LiChrospher1 60 RP-select B (5 mm,
250 Â 4 mm). The mobile phases were A) CH3CN:0.05 M ammonium acetate
(60:40) and B) CH3CN. The product was eluted with a flow rate of 1 ml/min.
starting with A:B (90:10) for 5 min and then linearly increased to A:B (50:50)
over a period of 13 min.
The radiochemical purity was >98% determined with analytical HPLC.
The identity of the product was verified by co-elution of an authentic sample
of citalopram, which showed an identical retention time. [11C]-4, [11C]-12 and
[11C]-13 eluted after approximately 16 min and the precursors after approxi-
mately 13 min. The specific radioactivity was calculated by comparing the UV-
response of the product with a standard curve of UV-responses.
Ex vivo rat studies
Male Sprague-Dawley rats (254–311 g) were used for ex vivo rat studies. Sixty
minutes after tracer injection (300 ml, 5–25 MBq) the rats (n=3 in each group)
were decapitated and seven brain regions (olfactory tuberli, hypothalamus,
cerebellum, hippocampus, corpus striatum, thalamus, frontal cortex, cortex)
were rapidly dissected out, weighed and the radioactive content was measured.
For blocking studies paroxetine (3 mg/kg) dissolved in saline was injected
15 min prior to tracer injection. ROI to cerebellum ratios were determined as
[ROI(cpm)/ROI(mg)]/[cerebellum(cpm)/cerebellum(mg)].
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

GAS PHASE PRODUCTION OF
347
Acknowledgements
Peter Brꢀsen and Henrik Pedersen at H. Lundbeck A/S are gratefully
acknowledged for performing the supercritical fluid HPLC analysis and
helpful discussions. Inge Mꢀller at the Neurobiology Research unit is
gratefully acknowledged for excellent technical assistance with the ex vivo
rat studies. This work is supported by The Centre for Drug Design and
Transport through a grant from The Danish Medical Research Council.
References
1. Wong DT, Bymaster FP. Development of antidepressant drugs. In Neurochem-
istry Clinical Application. Tang L, Tang S (ed.). Plenum Press: New York, 1995;
77–95.
2. Matarrese M, Soloviev D, Todde S, Mangni F, Colombo D, Galli Kienle M,
Fazio F. Appl Radiat Isot 1997; 48: 749–754.
3. Wollmer P, Moresco RM, Simonelli P, Rigamonti M, Ortu G, LaGaetana R,
Soloviev D, Galli M, Fazio F. Nucl Med Biol 2000; 27: 177–181.
4. Lasne MC, Pike, VW; Turton, DR. Appl Radiat Isot, Int J Radiat Appl Instrum
Part A 1989; 40: 147–151.
5. Dannals RF, Ravert HT, Wilson AA, Wagner Jr HN. Appl Radiat Isot 1990; 41:
541–543.
6. Ram S. Appl Radiat Isot 1990; 41: 645–648.
7. Hume SP, Pascali C, Pike VW, Turton DR, Ahier RG, Myers R, Bateman DM,
Cremer JE, Manjil LG, Dolan R. Nucl Med Biol, Int J Radiat Appl Instrum Part B
1991; 18: 339–351.
8. Hume SP, Lammertsma AA, Bench CJ, Pike VW, Pascali C, Cremer JE,
Dolan RJ. Nucl Med Biol, Int J Radiat Appl Instrum Part B 1992; 19: 851–855.
9. Shiue CY, Shiue GG, Cornish KG, O’rourke MF. Nucl Med Biol 1995; 22:
613–616.
10. Hammadi A, Crouzel C. J Label Compd Radiopharm 1993; 33: 703–710.
11. Pike VW. Nucl Med Biol 1995; 22: 1011–1028.
12. Brust P, Scheffel U, Szabo Z. IDrugs 1999; 2: 129–145.
13. Dannals RF, Scheffel U, Suehiro M, Ricaurte G. Med Chem Res 1994; 5: 228–224.
14. Wilson AA, Houle S. J Label Compd Radiopharm 1999; 42: 1277–1288.
15. Wilson AA, Ginovart N, Schmidt M, Meyer JH, Threlkeld PG, Houle S. J Med
Chem 2000; 43: 3101–3110.
16. Tarkiainen J, Vercouillie J, Emond P, Sandell J, Hiltunen J, Frangin Y,
Guilloteau D, Halldin C. J Label Compd Radiopharm 2001; 44: 1013–1023.
17. Houle S, Ginovart N, Hussey D, Meyer JH, Wilson AA. Eur J Nucl Med 2000; 27:
1719–1722.
18. Lundberg J, Olsson H, Halldin C, Tarkainen J, Sovago J, Vercouillie J, Gulyas B,
Guilloteau D, Emond P, Chalon S, Hiltunen J, Farde L. Neuroimage 2002; 16:
3:S17.
!
19. Hyttel J, Bꢀgesꢀ KP, Perregaard J, Sanchez C. J Neural Transm [Gen Sect]
1992; 88: 157–160.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348

348
J. MADSEN ET AL.
20. Owens MJ, Knight DL, Nemeroff CB. Biol Psychiat 2001; 50: 345–350.
21. Elfving B, Bjꢀrnholm B, Ebert B, Knudsen GM. Synapse 2001; 41: 203–211.
22. Huang Y, Hwang D-R, Narendran R, Sudo, Y, Chatterjee R, Bae S, Mawlawi O,
Kegeles LS, Wilson AA, Kung HF, Laruelle M. J Cereb Blood Flow Metab 2002;
22: 1377–1398.
23. Øyehaug E, Østensen ET, Salvesen B. J Chromatogr Biomed Appl 1984; 308:
199–208.
.
24. Seifritz E, Baumann P, Muller MJ, Annen O, Amey M, Hemmeter U,
Hatzinger M, Chardon F, Holsboer-Trachsler E. Neuropsychopharmacology
1996; 14: 253–263.
25. Rochat B, Kosel M, Boss G, Testa B, Gillet M, Baumann P. Biochem Pharmacol
1998; 56: 15–23.
26. Olesen O, linnet K. Pharmacology 1999; 59: 298–309.
27. Dalgaard L, Larsen C. Xenobiotica 1999; 29: 1033–1041.
28. Kosel M, Amey M, Aubert AC, Baumann P. Eur Neuropsychopharmacol 2001;
11: 75–78.
(
(
.
29. Halldin C, Bjurling P, Stalnacke C-G, Jossan SS, Oreland L, Langstrom B.
Appl Radiat Isot, Int J Appl Instrum Part A 1989; 40: 557–560.
30. Shinotoh H, Inoue O, Suzuki K, Yamasaki T, Iyo M, Hashimoto K, Tominaga T,
Itoh T, Tateno Y, Ikehira H. J Nucl Med 1987; 28: 1006–1011.
31. Inoue I, Tominaga T, Yamasaki T, Kinemuchi H. J Neurochem 1985; 44:
210–216.
32. Hashimoto K, Inoue O, Suzuki K, Yamasaki T, Kojima M. Nucl Med Biol, Int
J Radiat Instrum Part B 1986; 13: 79–80.
33. Fowler JS, Wolf AP, MacGregor RR, Dewey SL, Logan J, Schlyer DJ,
(
Langstrom B. J Neurochem 1988; 51: 1524–1534.
34. Fowler JS, Wang G-J, Logan J, Xie S, Volkow ND, Macgregor RR, Schlyer DJ,
Pappas N, Alexoff DL, Patlak C, Wolf AP. J Nucl Med 1995; 36: 1255–1262.
35. Roedig, A. Metod Org Chem, Houben-Weyl 1960; 5(4): 595–605.
36. Kornblum N, Jones WJ, Anderson GJ. J Am Chem Soc 1959; 81: 4113–4114.
37. Zimmerman HE, Albrecht FX, Haire MJ. J Am Chem Soc 1975; 97: 3726–3740.
38. Campaigne E, LeSuer WM. Org Synth 1963; 4: 919–921.
39. Lane, CF. Chem Rev 1976; 76: 773–799.
40. Kotake Y, Lijima A, Yoshimatsu K, Tamai N, Ozawa Y, Koyanagi N, Kitoh K,
Nomure H. J Med Chem 1994; 37: 1616–1624.
41. Tollefson MB, Li JJ, Beak P. J Am Chem Soc 1996; 118: 9052–9061.
42. Larsen P, Ulin J, Dahlstrꢀm K, Jensen M. Appl Radiat Isot 1997; 48: 153–157.
43. Cortes R, Soriano E, Pazos A, Probst A, Palacios JM. Neuroscience 1988; 27:
473–496.
Copyright # 2004 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2004; 47: 335–348