Synthesis and biological evaluation of125I/123I- labelled analogues of citalopram and escitalopram as potential radioligands for imaging of the serotonin transporter

Article abstract of DOI:10.1002/jlcr.1841

Two novel radioligands for the serotonin transporter (SERT), [ ¹²⁵I]{3-[5-iodo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl] -propyl}-dimethylamine ([¹²⁵I]-2) and S-[¹²⁵I]{3-[5-iodo- 1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl]-propyl}-dimethylamine ([ ¹²⁵I]-(S)-2) were synthesized in a Br/¹²⁵I exchange reaction. Binding experiments in rats yielded K_d values of 0.7 ± 0.06 and 0.52± 0.02 nM for [¹²⁵I]-2 and [ ¹²⁵I]-(S)-2, respectively. One hour after intravenous injection of [¹²⁵I]-2, 0.34% of the injected dose had accumulated in the brain. The highest hypothalamus-to-cerebellum ratio was reached 2 h after injection of [¹²⁵I]-(S)-2 and amounted to 2.4. Pre-treatment experiments with paroxetine resulted in effective reduction of the target-to-cerebellum ratios. The corresponding iodine-123 labelled compound S-[¹²³I]{3-[5-Iodo-1- (4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl]-propyl}-dimethylamine [ ¹²³I]-S-2 was investigated in a pig single photon emission computed tomography (SPECT) study. Between 60 and 110 min after IV injection, the midbrain-to-cerebellum ratio was 1.2. However, the uptake did not differ between high-density and medium-density regions questioning the feasibility of the radioligand in imaging cortical SERT distribution in vivo. These data suggest that the iodine-labelled derivatives of citalopram and escitalopram are not superior to another SPECT tracer for the SERT, namely [¹²³I]ADAM. Copyright

Full text of DOI:10.1002/jlcr.1841

Research Article
Received 8 December 2009,
Revised 26 July 2010,
Accepted 12 September 2010
Published online 30 November 2010 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.1841
Synthesis and biological evaluation of
125I/123I-labelled analogues of citalopram and
escitalopram as potential radioligands for
imaging of the serotonin transporter
Ã
Jacob Madsen,a y Betina Elfving,by Vibe G. Frokjaer,b Birgitte R. Kornum,b
Gerda Thomsen,b Lars Martiny,c and Gitte M. Knudsenb
Two novel radioligands for the serotonin transporter (SERT), [125I]{3-[5-iodo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-
1-yl]-propyl}-dimethylamine ([125I]-2) and S-[125I]{3-[5-iodo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl]-propyl}-
dimethylamine ([125I]-(S)-2) were synthesized in a Br/125I exchange reaction. Binding experiments in rats yielded Kd values
of 0.770.06 and 0.5270.02 nM for [125I]-2 and [125I]-(S)-2, respectively. One hour after intravenous injection of [125I]-2,
0.34% of the injected dose had accumulated in the brain. The highest hypothalamus-to-cerebellum ratio was reached 2 h
after injection of [125I]-(S)-2 and amounted to 2.4. Pre-treatment experiments with paroxetine resulted in effective
reduction of the target-to-cerebellum ratios.
The corresponding iodine-123 labelled compound S-[123I]{3-[5-Iodo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl]-
propyl}-dimethylamine [123I]-S-2 was investigated in a pig single photon emission computed tomography (SPECT) study.
Between 60 and 110 min after IV injection, the midbrain-to-cerebellum ratio was 1.2. However, the uptake did not differ
between high-density and medium-density regions questioning the feasibility of the radioligand in imaging cortical SERT
distribution in vivo.
These data suggest that the iodine-labelled derivatives of citalopram and escitalopram are not superior to another SPECT
tracer for the SERT, namely [123I]ADAM.
Keywords: serotonin transporter; escitalopram; SPECT; iodine labelling
been evaluated in humans.12–14 So far, [123I]ADAM remains the
only available SPECT radioligand for human studies. However,
robust quantification of cerebral SERT requires a 120-min data
Introduction
Positron emission tomography (PET) and single photon emission
computed tomography (SPECT) are non-invasive techniques,
acquisition and another potential disadvantage with this
which can be applied for in vivo investigations of biological
processes in humans using appropriate radioligands.1 Radio-
radioligand is the presence of a lipophilic metabolite seen in
some subjects.14
ligands are labelled with radionuclides such as 11C (t1/2
=
In this paper, we report the radiolabelling and initial biological
studies of iodinated derivatives of the selective serotonin
reuptake inhibitors (SSRIs), citalopram and escitalopram. The
novel radioligands [125I]f3-[5-iodo-1-(4-fluorophenyl)-1,3-dihy-
droisobenzofuran-1-yl]-propylg-dimethylamine ([125I]-2) and
20.4 min) or 18F (t1/2 = 109 min) for PET and 99mTc (t1/2 = 6.0 h) or
123I (t1/2 = 13.1 h) for SPECT. The advantages of SPECT are first
of all lower costs and the more facile handling of often
longer lived SPECT radioligands, whereas PET is a quantitative
method and the use of radioligands labelled with short-lived
isotopes allow multiple investigations in the same subject on
the same day.
aPET and Cyclotron Unit 3982, Copenhagen University Hospital, Blegdamsvej 9,
2100 Copenhagen Ø, Denmark
The serotonin transporter (SERT) is believed to play a central
role in mental disorders arising from disturbances in the
serotonin system.2 There are now several PET-radioligands
available for imaging of the SERT in the living human brain,
especially the diarylsulfides [11C]DASB,3,4 [11C]MADAM,5,6 and
more recently [11C]HOMADAM.7,8 For SPECT imaging, iodinated
diarylsulfides such as [123I]ODAM, [123I]IDAM and [123I]ADAM
have all been suggested as candidates for visualization of the
SERT in vivo.9–11 Within this series of compounds, [123I]ADAM is
the most promising candidate as a SPECT radioligand and has
bNeurobiology Research Unit 9201, Copenhagen University Hospital, Blegdams-
vej 9, 2100 Copenhagen Ø, Denmark
cRadiation Research Division, Risø DTU, Frederiksborgvej 399, 4000 Roskilde,
Denmark
*Correspondence to: Jacob Madsen, PET and Cyclotron Unit 3982, Copenhagen
University Hospital, Blegdamsvej 9, 2100 Copenhagen Ø, Denmark.
E-mail: jacob.madsen@rh.regionh.dk
yB. E. and J. M. have contributed equally to this work.
J. Label Compd. Radiopharm 2011, 54 185–190
Copyright r 2010 John Wiley & Sons, Ltd.

J. Madsen et al.
S-[125I]f3-[5-iodo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-
The radioligand [125I]-2 was injected in the tail vein and the
1-yl]-propylg-dimethylamine ([125I]-(S)-2) were investigated in in rats were decapitated 1, 2, 3, 4 and 6 h later. Rats injected with
vitro and ex vivo assays. S-[123I]f3-[5-Iodo-1-(4-fluorophenyl)-1,3-
[
125I]-(S)-2 were decapitated after 2 h. In all cases, the brains
dihydroisobenzofuran-1-yl]-propylg-dimethylamine ([123I]-(S)-2) were quickly removed and dissected on an ice-cold tile. Eight
was investigated for its potential use as radioligand for in vivo brain regions (olfactory tubercles, hypothalamus, cerebellum,
imaging of the SERT in the human brain using SPECT.
hippocampus, corpus striatum, thalamus, pre-frontal cortex and
remaining cortex) were dissected, weighted and the accumu-
lated radioactivity was counted. Paroxetine (3 mg/kg) was used
in competitive binding studies. The region of interest (ROI)-to-
cerebellum ratios were determined as [(ROI(cpm)/ROI(g))/
(cerebellum(cpm)/cerebellum(g))]. The percent dose per gram
brain tissue (% dose/g) was calcutated as [(ROI(cpm)/g ROI
tissue)/total injected dose] for each brain region.
Materials and methods
No-carrier-added [125I]sodium iodide solution (2 Ci/mmol) was
purchased from Amersham Biosciences. Paroxetine hydrochlor-
ide was purchased from Smith-Kline Beecham, West Sussex UK.
Male Sprague–Dawley rats weighing 206–396 g were used for
the biological studies (M&B, Denmark). HPLC was performed on
a Merck Hitachi fitted with in-line radio detection. Log P values
were calculated with QikProp. f3-[5-bromo-1-(4-fluorophenyl)-
1,3-dihydroisobenzofuran-1-yl]-propylg-dimethylamine (1), S-f3-
[5-bromo-1-(4-fluorophenyl)-1,3-dihydroisobenzofuran-1-yl]-
propylg-dimethylamine (S-1) and f3-[5-iodo-1-(4-fluorophenyl)-
1,3-dihydroisobenzofuran-1-yl]-propylg-dimethylamine (2) were
gifts from H. Lundbeck A/S, Copenhagen, Denmark. [123I]-(S)-2
was prepared by MAP Medical, Finland, from the stannylated
precursor, (S)-f3-[(4-fluorophenyl)-5-tributylstannyl-1,3-dihydroi-
sobenzofuran-1-yl]-propylg-dimethylamine.15
SPECT study
A Landrace pig, 33.5 kg, was used for the SPECT scan. The pig
was sedated with an i.m. injection of midazolam (0.1 mg/kg).
After 10–15 min, full anaesthesia was induced with 0.15 ml/kg
i.m. of a mixture of Midazolam and Zoletil 50 vet. (Virbac Animal
Health, France: 125 mg tiletamin and 125 mg zolazepam in a
total volume of 8 ml Midazolam (1 mg/ml)) and maintained
during the whole procedure using propofol (10 mg/ml, 51 ml/h).
The pig was intubated and kept on a respiration rate of 12 ml/
min. Arterial and venous catheters were inserted in the femoral
artery and vein, and used for injection of [123I]-S-2, saline and for
Radiosynthesis of ½125Iꢀ-2 and ½125Iꢀ-(S)-2
taking blood samples.
A saturation monitor on the tail
To a 1-ml vial charged with 1 (1.6 mg, 4.2 mmol), citric acid
(5.3 mg, 27.6 mmol), 2,5-dihydroxybenzoic acid (3.9 mg,
25.3 mmol), isoascorbic acid (3.6 mg, 20.4 mol) and CuSO4
(0.2 ml, 1 mg/ml) was added acidic acid (0.2 ml, 1 M) and
monitored oxygen saturation and heart rate, and blood gases
and electrolytes in arterial blood were measured with an ABL
Blood Gas analyzer (625, Radiometer, Denmark). All parameters
were kept stable during the entire scan session by adjusting
respiration volume and giving saline drop.
[
125I]sodium iodide (5 mCi). The vial was sealed and heated to
1401C for 2 h. After cooling to room temperature, the mixture
was loaded on a Seppak Plus C18 cartridge (Waters). The
cartridge was flushed with H2O (2 ml) and the radioactive
material was eluted with 70% CH3CN (2 ml). The eluted material
was evaporated to approximately 300 ml and diluted with 200 ml
HPLC eluent. The mixture was split into two portions before
purification. HPLC purification was performed on an analytical
After injection of 147 MBq [123I]-S-2 dynamic SPECT scanning
was conducted over a period of 4 h (30 ꢁ 2 min in the first hour
followed by 18 ꢁ 10 min). SPECT imaging was performed on a
triple-head IRIX camera (Philips, Cleveland, USA) fitted with low-
energy, general-all-purpose parallel holed collimators (LEGAP-
PAR). The spatial resolution of the camera is approximately
8.5 mm at 10 cm distance to the camera head. To allow later
correction for septal penetration and backscatter from high-
energy photons, the signal was recorded in two separate energy
windows. Window 1 was centered at 159 KeV, width 32 KeV, and
window 2 was centered at 200 KeV, width 32 KeV. Each camera
covered 120 degrees of the circular orbit. Scans were performed
in a continuous mode with an angular step of 3 degrees and 40
steps. The acquisition matrix was 128 ꢁ 128 with a pixel size of
2.33 mm in both directions.
˚
column (RP, OdDMeSi, 5 mm, 120 A, 4.6 ꢁ 250 mm). HPLC eluent
A: 0.1% TFA/10% CH3CH and B: 0.1% TFA/90% CH3CN were
used. From 0–40 min, the eluent B was linearly increased from 25
to 40% with a 1-ml/min flow rate. The purified product, [125I]-2,
was analysed on a second HPLC system using an analytical
column as above. The solvents A: 0.1% TFA/10% CH3CH and B:
0.1% TFA/90% CH3CN were used. From 0–30 min, solvent B was
increased from 40 to 60%.
Applying the above procedure [125I]-S-2 was prepared using
S-1 as the starting material.
For reconstruction, the projection image files were converted
to DICOM format in the scanner software (Odyssey 9.2B; Philips,
Cleveland, USA) and exported to UNIX platforms for further
processing by matlab routines developed at the NRU (Neuro-
In vitro binding studies
Saturation binding studies were performed with rat brain biology Research Unit, Copenhagen, Denmark). In Matlab 6.5,
homogenates of cerebrum (n = 3) according to a previous the images were converted from DICOM to Analyse format.
published procedure16 and the Kd values were determined at Before reconstructing the projection images, the raw data of the
371C.17
two energy windows were subtracted in sinogram space to
perform correction for penetration and backscatter from high-
energy photons. The number of counts in the high-energy
window was multiplied by 1.1, the factor was determined by
recording the two energy windows of a 123I-source contained in
6 mm lead. Reconstruction was performed by using standard
filtered back projection with a low pass 4th order Butterworth 3D
post-filter with a cut-off frequency of 0.4 Nyquist and using
Ex vivo binding studies
Three rats per time point were used for ex vivo brain distribution
studies. All animal experiments were carried out in accordance
with the European Communities Council Resolves of 24th
November 1986 (86/609/ECC) and approved by the Danish
Animal Research Inspectorate (Journal number 2003/561–745).
www.jlcr.org
Copyright r 2010 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2011, 54 185–190

J. Madsen et al.
uniform attenuation correction with an attenuation coefficient purity of the final product was 498% as determined by
of 0.10 cmꢂ1. The complete subtraction and reconstruction analytical HPLC and no precursor could be detected in the UV-
algorithm was carried out in Matlab 6.5. The resulting series of trace of the HPLC chromatogram. The identity of the product
images were converted into a dynamic file and a summed image was confirmed by co-elution with the reference substance 2 in
using all frames.
each case. The purified product was evaporated and formulated
For calculating time-activity curves (TACs) from midbrain, in 50% EtOH. In this medium, [125I]-2 has proven to be stable for
frontal cortex and cerebellum, the summed image was co- at least 2 months (495% pure).
registered to a pig brain MRI-based atlas18 using the program
Register,
a tri-planer viewer developed at the Montreal Log P value
Neurological Institute, McGill University, Canada. Registration
was achieved by manually marking homologue anatomical
positions in the two scans and minimizing the root mean square
distance between the landmarks. Then, TACs from the ROIs were
calculated as previously described.19 To minimize statistical
noise in the signal, the regions thalamus, diencephalon and
mesencephalon were pooled to obtain a larger midbrain region.
The iodo derivative 2 has a calculated log P value of 5.2.
In vitro binding studies
From saturation experiments in rat membrane homogenates
16
previously described in detail,
Scatchard plot analysis was
performed. The Scatchard plot was linear indicating only one
binding site and the Kd values (mean7SEM) were 0.770.06 nM
and 0.5270.02 nM for the iodinated derivatives [125I]-2 and
Results
[
125I]-(S)-2, respectively.
Radiochemistry
The radiosyntheses of [125I]-2 and [125I]-(S)-2 were performed in
a one step reaction from the corresponding bromo precursors 1
and S-1, respectively (Scheme 1).
For each radioligand, the radioactive product was collected
after 29.5–30.5 min in two separate HPLC runs. The bromo
precursors 1 and S-1 were eluted after 20.5–26 min. The
iodinated products, [125I]-2 and [125I]-(S)-2, were obtained in
approximately 35% radiochemical yield. The radiochemical
Ex vivo brain distribution studies in rats
Between 1.25 and 1.95 MBq of [125I]-2 or [125I]-(S)-2 was
intravenously injected in the tail vein of the rats. Rats receiving
[
125I]-2 were decapitated between 1 and 6 h after injection as
indicated in Table 1. The radioligand penetrated the blood–
brain barrier resulting in a brain uptake of 0.34% injected dose
after 1 h. After decapitation, the brain regions of interest were
125I
Br
O
O
N
N
[
125I]NaI
[
125I]-2
1M CH3COOH,
140 °C, 2h, 36%
1
F
F
125I
Br
O
O
N
[
125I]NaI
N
1M CH3COOH,
140 °C, 2h, 34%
S-1
[
125I]-S-2
F
F
Scheme 1. Radiosynthesis of [125I]-2 and [125I]-(S)-2.
Table 1. Regional brain distribution (% dose/g) in rat brain after intravenous injection of [125I]-2
Brain region
1 h (n = 3)
2 h (n = 3)
3 h (n = 3)
4 h (n = 3)
6 h (n = 2)
Olfactory tuberli
Hypothalamus
Cerebellum
Hippocampus
Corpus striatum
Thalamus
0.1870.03
0.2370.03
0.1870.03
0.2370.04
0.2370.05
0.2870.04
0.2470.04
0.2370.04
0.0970.02
0.1270.02
0.0870.01
0.1270.02
0.1170.02
0.1770.04
0.1170.02
0.1070.02
0.0870.01
0.1170.02
0.0770.01
0.1170.20
0.0970.02
0.1270.02
0.0970.01
0.0970.01
0.0470.01
0.0570.02
0.0370.01
0.0470.01
0.0470.01
0.0570.01
0.0470.01
0.0470.01
0.0370.00
0.0370.00
0.0370.00
0.0370.01
0.0370.01
0.0470.01
0.0370.00
0.0370.00
Pre-frontal cortex
Remaining cortex
Values are mean7SD.
J. Label Compd. Radiopharm 2011, 54 185–190
Copyright r 2010 John Wiley & Sons, Ltd.
www.jlcr.org

J. Madsen et al.
dissected and weighted, and the radioactive content was
counted and regional brain distributions (% dose/g) of
compound [125I]-2 was determined (Table 1).
SPECT study
123I]-(S)-2 was studied in vivo in the pig brain. Summed images
[
and TACs from the midbrain, frontal cortex and cerebellum are
shown in Figure 2. In this set-up, [123I]-(S)-2 peaks at around
20–40 min after injection. Uptake is highest in midbrain, slightly
lower in frontal cortex and lowest in cerebellum. Target-to-
cerebellar area under the curve (AUC) ratio is 1.20 for midbrain
and 1.13 for frontal cortex.
The maximal target-to-cerebellum ratios were found in the
thalamus and hypothalamus regions after 2 h. Rats receiving
[
and thalamus-to-cerebellum ratios were 2.1 and 2.4, respec-
tively. In the blocking experiments, paroxetine (3 mg/kg) was
injected in the tail vein of the rats (n = 3) 10 min prior to injection
of [125I]-2. After 2 h, the rats were decapitated and the
thalamus- and hypothalamus-to-cerebellum ratios were signifi-
cantly reduced (Figure 1).
125I]-(S)-2 were decapitated after 2 h and the hypothalamus-
Discussion
In our search for new SPECT tracers to study the SERT in vivo, we
have synthesized the novel iodinated radioligands [125I]-2 and
[
125I]-(S)-2, which are structurally related to the SSRI’s citalo-
pram and escitalopram. The iodo derivative 2 has a calculated
log P of 5.2 and displays high affinity for the SERT and is highly
selective as compared with the dopamine and noradrenaline
transporter (Table 2).20
3
2.5
2
[
125I]-2 Hypothalamus
125I]-2 Thalamus
[
[
125I]-(S)-2 Hypothalamus
125I]-(S)-2 Thalamus
[
Table 2. Inhibition of [3H]5-HT, [3H]DA and [3H]NA uptake
for citalopram and iodo derivative 2
[
125I]-2 Hypothalamus (+ paroxetine (3 mg/kg))
[
125I]-2 Thalamus (+ paroxetine (3 mg/kg))
R
O
N
1.5
1
F
1
2
3
4
6
blocking
2 h after
i.v.
Inhibition of [3H]amine uptake, IC50 (nM)
Compound
R
5-HT
DA
NA
Time (hours)
Figure 1. Thalamus- and hypothalamus-to-cerebellum ratios of [125I]-2 and [125I]-
(S)-2 at different time points after intravenous injection. In the blocking
experiments, the rats were sacrificed 2 h after injection. The values are mean7SD.
Citalopram CN
2
1.8
3.5
40 000
2600
6100
1300
I
Figure 2. Summed SPECT images and TACs from the midbrain, frontal cortex and cerebellum of [123I]-(S)-(2).
www.jlcr.org
Copyright r 2010 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2011, 54 185–190

J. Madsen et al.
AUC ratio was 1.20 for the high-binding region midbrain. The
TAC from the frontal cortex and the midbrain are almost
overlapping, although a difference in SERT concentrations
between the two regions is approximately 3.5, as measured
with [11C]DASB PET.27 This can partly be ascribed to partial
volume effects that are caused by the limited spatial resolution
of SPECT. This leads to an underestimation of regional
radioactivity distribution in regions with high radioactivity
concentration relative to the surroundings, and also to the
converse situation with overestimation in regions with low
radioactivity concentration.28 With a co-registered magnetic
resonance image, it is possible to correct for these effects of
partial volume.29 Even when partial volume effects and the small
volume of the pig brain (180 g) are taken into account, the TAC
in the pig brain are considerably less promising than in our rat
data. However, without a partial volume correction, we still
detect a difference between high and low activity regions, and
this difference is expected to be even higher in partial volume
corrected data.
0.3
0.25
0.2
Thalamus
Cortex
Cerebellum
0.15
0.1
0.05
0
0
1
2
3
4
5
6
Time (hours)
Figure 3. Regional brain distribution in the thalamus, cortex and cerebellum
(% dose/g) in rat brain at various time points after intravenous injection of [125I]-2.
Values are mean, see also Table 1.
The low target-to-background ratio in the brain could either
The radioligands [125I]-2 and [125I]-(S)-2 were synthesized be due to too high non-specific binding or to too low specific
from the corresponding bromo precursors 1 and S-1, respec- binding. The estimated log P value of 5.2, which is considered to
tively, in a Br/125I exchange reaction21 (Scheme 1). The radio- be relatively high, could result in slow washout of non-specific
iodinated products, [125I]-2 or [125I]-(S)-2, were well separated binding. We previously found that the non-specific binding of
from the bromo precursors 1 or S-1, respectively, under the the carbon-11 labelled S-enantiomer of citalopram (log P 3.6) is
HPLC conditions applied.
also high in rats,30,31 as it is in humans30 and as well when a
The Kd values of [125I]-2 (0.7070.06 nM) and [125I]-(S)-2 carbon-11 labelled citalopram derivative where the cyano group
(0.5270.02 nM) are comparable with the values obtained was changed to a methyl group (log P 4.6) was studied in non-
for [3H]citalopram (Kd = 1 nM)22 and [3H]escitalopram human primates.32
(Kd = 0.71 nM).16 In ex vivo rat studies, [125I]-2 penetrated the
It is also possible that the binding affinity of [125I]-(S)-2
blood–brain barrier (0.34% injected dose 1 h after injection) and (0.5270.02 nM) for SERT is too low. We have previously
the highest target-to-cerebellum ratios at all time points were estimated that the imaging of SERT in a high-density region in
found in the thalamus and hypothalamus. This corresponds well the human brain requires radioligands with Kd values between
with the known high SERT densities in these regions of the 0.03 and 0.3 nM.17 We do not know whether the binding affinity
brain.23 Cerebellum, which has a low density of SERT sites, was of [125I]-(S)-2 displays species differences in the rat and human
used as reference region.23 A maximal thalamus-to-cerebellum brain. However, based on our previous experiments that
ratio of 1.9 was observed 2 h after intravenous injection of included radiolabelled (1)McN5652, DASB, ADAM, MADAM,
[
125I]-2. After pre-treatment with the SSRI paroxetine, the major escitalopram, fluoxetine and paroxetine, only paroxetine dis-
parts of the specific binding in the thalamus and hypothalamus played higher affinity for monkey SERT compared with rat
regions were reduced (Figure 1).
SERT.17 If [125I]-(S)-2 also has lower or equal affinity for the
Since the established TAC indicated that a maximal ROI-to- human SERT compared with the rat SERT, our finding supports
cerebellum ratio was reached after 2 h this time point was used the notion that successful SERT radioligands should have an
for further experiments with the biologically more active affinity in the range given above.
enantiomer, [125I]-(S)-2.
Metabolic degradation of citalopram invariably occurs at the
Compared with [125I]-2 injection of [125I]-(S)-2 resulted, as dimethyl amino group.33 Assuming that the minor differences
expected, in an increased target-to-background ratio in ex vivo between the cynao- and iodo substituents present in citalopram
rat studies. The hypothalamus- and thalamus-to-cerebellum and 2 do not play a role for the metabolic process, the two
ratios were increased from 1.6 and 1.9 to 2.1 and 2.4, major radiometabolites of [123I]-(S)-2 would be either glucor-
respectively (Figure 1).
onides or the desmethyl derivative. This is in accordance with
The maximal uptake was reached after 1 h. The maximal our experience with S-[N-methyl-11C]citalopram in non-human
target-to-cerebellum ratio reached after 2 h (Figure 3) was primates where we found indications of formation of polar
relatively fast compared with [125I]ADAM,24 which peaks after 4 h metabolites only (unpublished data). Therefore, we do not
and slightly slower than [125I]IDAM,25 which peaks already after expect the formation of lipophilic radiolabelled metabolites
60 min.
from [123I]-(S)-2. Although glucoronides cannot cross the
The target-to-cerebellum ratio was somewhat lower than blood–brain barrier, we cannot exclude that a radiolabelled
125I]ADAM and [125I]IDAM, where hypothalamus-to-cerebellum demethylated metabolite to some extent may also contribute to
[
ratios are 5.0 and 2.8, respectively.26
the insufficient target-to-background signal in pigs.
TACs show that in the pig brain the uptake of [123I]-(S)-2
peaks at around 20–40 min after injection, slightly faster in
transporter rich than in transporter poor regions. Uptake was
higher in transporter rich regions as midbrain compared with
cerebellum with very low densities of SERT. Target-to-cerebellar
Conclusions
Our results indicate that despite its superior selectivity towards
the SERT, [123I]-(S)-2 is not better SPECT radioligand than
J. Label Compd. Radiopharm 2011, 54 185–190
Copyright r 2010 John Wiley & Sons, Ltd.
www.jlcr.org

J. Madsen et al.
[
123I]ADAM for SERT. It is difficult to predict the in vivo
¨
[11] P. D. Acton, S.-R. Choi, C. Hou, K. Plossl, H. F. Kung, J. Nucl. Med.
2001, 42, 1556.
[12] T. A. Kauppinen, K. A. Bergstrom, P. Heikman, J. Hiltunen, A. K.
Ahonen, Eur. J. Nucl. Med. Mol. Imaging 2003, 30, 132.
[13] K. Erlandsson, T. Sivananthan, D. Lui, A. Spezzi, C. E. Townsend, S.
Mu, R. Lucas, S. Warrington, P. J. Ell, Eur. J. Nucl. Med. Mol. Imaging
2005, 32, 1329.
characteristics of a radioligand based on in vitro data alone.
Studies in small animals such as rats can give good indications of
the fate of the radioligand. However, for a series of radioligands
with promising properties a conclusion should be based on an in
vivo PET/SPECT study in larger animals such as non-human
primates or pigs. Based on our data, [123I]-(S)-2 does not show a
great potential as a SERT-radioligand for human SPECT studies.
¨
[14] V. G. Frokjaer, L. H. Pinborg, J. Madsen, R. de Nijs, C. Svarer, A.
Wagner, G. M. Knudsen, J. Nucl. Med. 2008, 49(2), 247.
¨
[15] J. Madsen, P. Merachtsaki, P. Davoodpour, B. Bergstrom, B.
Langstrom, K. Andersen, C. Thomsen, L. Martiny, G. Knudsen,
˚
¨
Acknowledgements
Bioorg. Med. Chem. 2003, 11, 3447.
¨
[16] B. Elfving, B. Bjornholm, B. Ebert, G. M. Knudsen, Synapse 2001,
41, 203.
[17] B. Elfving, J. Madsen, G. M. Knudsen, Synapse 2007, 61(11), 882.
[18] H. Watanabe, F. Andersen, F. Simonsen, S. M. Evans, A. Gjedde, P.
Cumming, Neuroimage 2001, 14(5), 1089.
[19] F. Andersen, H. Watanabe, C. Bjarkam, E. H. Danielsen, P.
Cumming, Brain. Res. Bull. 2005, 66(1), 17.
[20] H. Lundbeck A/S, unpublished data.
This work is supported by The Centre for Drug Design and
Transport through a grant from The Danish Medical Research
Council. Anne Mette Dall and Inge Møller are gratefully
acknowledged for excellent technical assistance.
´
[21] M. Ceusters, D. Tourwe, J. Callaerts, J. Mertens, A. Peter, J. Org.
Chem. 1995, 60, 8324.
[22] P. Plenge, E. T. Mellerup, J. Neurochem. 1991, 56, 248.
References
´
[1] R. E. Henkin, M. A. Boles, G. L. Dillehat, J. R. Halama, S. M. Karesh,
R. H. Wagner, A. M. Zimmer (Ed.), in Nuclear Medicine, Vol. 1,
Mosby Year Book Inc., 1996.
[2] M. B. Keller, J. Clin. Psychiatry 2001, 61, 896.
[3] A. A. Wilson, N. Ginovart, M. Schmidt, J. H. Meyer, P. G. Threlkeld,
S. Houle, J. Med. Chem. 2000, 43, 3103.
[23] R. Cortes, E. Soriano, A. Pazos, A. Probst, J. M. Palacios,
Neuroscience 1988, 27, 473.
[24] S. Oya, S.-R. Choi, C. Hou, M. Mu, M.-P. Kung, P. D. Acton, M.
Siciliano, H. F. Kung, Nucl. Med. Biol. 2000, 27, 249.
[25] M.-P. Kung, C. Hou, S. Oya, M. Mu, P. D. Acton, H. F. Kung, Eur. J.
Nucl. Med. 1999, 26, 844.
[4] M. Ichise, J. S. Liow, J. Q. Lu, A. Takano, K. Model, H. Toyama, T.
Suhara, K. Suzuki, R. B. Innis, R. E. Carson, J. Cereb. Blood Flow
Metab. 2003, 23(9), 1096.
[5] J. Tarkiainen, J. Vercouillie, P. Emond, J. Sandell, J. Hiltunen, Y.
Frangin, D. Guilloteau, C. Halldin, J. Label. Compd. Radiopharm.
2001, 44, 1013.
[6] J. Lundberg, I. Odano, H. Olsson, C. Halldin, L. Farde, J. Nucl. Med.
2005, 46(9), 1505.
[7] J. Jarkas, J. R. Votaw, R. J. Voll, L. Williams, V. M. Camp, M. J.
Owens, D. C. Purselle, J. D. Brenner, C. D. Kilts, C. B. Nemeroff, M.
M. Goodman, Nucl. Med. Biol. 2005, 32, 211.
[26] S. Oya, S.-R. Choi, C. Hou, M. Mu, M.-P. Kung, P. D. Acton, M.
Siciliano, H. F. Kung, Nucl. Med. Biol. 2000, 27, 249.
[27] S. B. Jensen, D. F. Smith, D. Bender, S. Jakobsen, D. Peters, E. Ø.
Nielsen, G. M. Olsen, J. Scheel-Kru¨ger, A Wilson, P. Cumming,
Synapse 2003, 49(3), 170.
[28] E. J. Hoffman, S. C. Huang, M. E. Phelps, J. Comput. Assist. Tomogr.
1979, 3, 299.
[29] H. W. Muller-Gartner, J. M. Links, J. L. Prince, R. N. Bryan, E.
McVeigh, J. P. Leal, C. Davatzikos, J. J. Frost, J. Cereb. Blood Flow
Metab. 1992, 12, 571–583.
[30] S. P. Hume, A. A. Lammertsma, C. J. Bench, V. W. Pike, C. Pascall, R.
J. Dolan, Int. J. Rad. Appl. Instrum. B. 1992, 19(8), 851.
[31] J. Madsen, B. Elfving, K. Andersen, L. Martiny, G. M. Knudsen, J.
Label. Compd. Radiopharm. 2004, 47, 335.
[8] J. A. Nye, J. R. Votaw, J. Jarkas, D. C. Purselle, V. Camp, J. D.
Brenner, C. D. Kilts, C. B. Nemeroff, M. M. Goodman, J. Nucl. Med.
2008, 49, 2018.
¨
¨
[32] J. Madsen, P. Merachtsaki, P. Davoodpour, M. Bergstrom, B.
[9] P. D. Acton, M. Mu, K. Plossl, C. Hou, M. Siciliano, Z-P. Zhuang, S.
Oya, S.-R. Choi, H. F. Kung, Eur. J. Nucl. Med. 1999, 26, 1359.
˚
Langstrom, K. Andersen, C. Thomsen, L. Martiny, G. M. Knudsen,
Bioorg. Med. Chem. 2003, 11, 3447.
¨
¨
[10] P. D. Acton, M.-P. Kung, M. Mu, K. Plossl, C. Hou, M. Siciliano, S.
Oya, H. F. Kung, Eur. J. Nucl. Med. 1999, 26, 854.
[33] L. Dalgaard, C. Larsen, Xenobiotica 1999, 29(10), 1033.
www.jlcr.org
Copyright r 2010 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2011, 54 185–190