Synthesis and evaluation of N-(5-fluoro-2-phenoxyphenyl)-N-(2-[ 18F]fluoromethoxy-d2-5-methoxybenzyl)acetamide: A deuterium-substituted radioligand for peripheral benzodiazepine receptor

Article abstract of DOI:10.1016/j.bmc.2004.11.058

N-(5-Fluoro-2-phenoxyphenyl)-N-(2-[¹⁸F]fluoromethoxy-d ₂-5-methoxybenzyl)acetamide ([¹⁸F]2) is a potent ligand (IC₅₀: 1.71 nM) for peripheral benzodiazepine receptor (PBR). However, in vivo evaluation on rodents

Full text of DOI:10.1016/j.bmc.2004.11.058

Bioorganic & MedicinalChemistry 13 (2005) 1811–1818
Synthesis and evaluation of N-(5-fluoro-2-phenoxyphenyl)-N-
(2-[18F]fluoromethoxy-d2-5-methoxybenzyl)acetamide: a deuterium-
substituted radioligand for peripheral benzodiazepine receptor
c
a,b
Ming-Rong Zhang,a,b, Jun Maeda, Takehito Ito, Takashi Okauchi,c
Masanao Ogawa,a,b Junko Noguchi,a,b Tetsuya Suhara,c Christer Halldind
and Kazutoshi Suzukia
*
aDepartment of Medical Imaging, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan
bSHI Accelerator Service Co. Ltd, 5-9-11 Kitashinagawa, Shinagawa-ku, Tokyo 141-8686, Japan
cBrain Imaging Project, National Institute of Radiological Sciences, 4-9-1 Anagawa, Inage-ku, Chiba 263-8555, Japan
dKarolinska Institute, Department of Clinical Neuroscience, Psychiatry Section, Karolinska Hospital, S-17176 Stockholm, Sweden
Received 6 October 2004; revised 24 November 2004; accepted 24 November 2004
Available online 22 December 2004
Abstract—N-(5-Fluoro-2-phenoxyphenyl)-N-(2-[18F]fluoromethoxy-d2-5-methoxybenzyl)acetamide ([18F]2) is a potent ligand (IC50:
1.71 nM) for peripheralbenzodiazepine receptor (PBR). However, in vivo evaulation on rodents and primates showed that
this ligand was unstable and rapidly metabolized to [18F]FÀ by defluorination of the [18F]fluoromethylmoiety. In this study, we
designed a deuterium-substituted analogue, N-(5-fluoro-2-phenoxyphenyl)-N-(2-[18F]fluoromethoxy-d2-5-methoxybenzyl)acetamide
([18F]5) as a radioligand for PBR to reduce the in vivo metabolic rate of the non-deuterated [18F]2. The design principle was based
on the hypothesis that the deuterium substitution may reduce the rate of defluorination initiated by cleavage of the C–H bond with-
out altering the binding affinity for PBR. The non-radioactive 5 was prepared by reacting diiodomethane-d2 (CD2I2, 6) with a
phenolprecursor 7, followed by treatment with tetrabutylammonium fluoride. The ligand [18F]5 was synthesized by the alkyl-
ation of 7 with [18F]fluoromethyliodide- d2 ([18F]FCD2I, [18F]9). Compound 5 displayed a similar in vitro affinity to PBR (IC50:
1.90 nM) with 2. In vivo evaluation demonstrated that [18F]5 was metabolized by defluorination to [18F]FÀ as a main radio-
active component, but its metabolic rate was slower than that of [18F]2 in the brain of mice. The deuterium substitution decreased
the radioactivity level of [18F]5 in the bone of mouse, augmented by the percentage of specific binding to PBR in the rat brain deter-
mined by ex vivo autoradiography. However, the PET image of [18F]5 for monkey brain showed high radioactivity in the brain
and skull, suggesting a possible species difference between rodents and primates.
ꢀ 2004 Elsevier Ltd. All rights reserved.
1. Introduction
relationship between PBR and brain diseases (Alzhei-
merÕs disease, ParkinsonÕs disease, stroke, encephalitis
etc.) or tumors (lung cancer, liver cancer, gastric cancer
etc.) has promoted the development of a PET ligand,
which could be used for visualizing the distribution
and examining the change of PBR in these regions.9–12
Peripheralbenzodiazepine receptor (PBR) is ol cated on
the mitochondrialouter membrane in peripheralorgans
including the kidney, nasal epithelium, lung, heart, and
endocrine organs such as the adrenal, testis, and pitui-
tary gland, and in the central nervous system.1,2 Since
the PBR density was found to increase in injured brains
and tumor regions, PBR imaging of PBR in these areas
has become an attractive study target.3–8 The putative
We recently developed a PET ligand, [11C]N-(2,5-
dimethoxybenzyl)-N-(5-fluoro-2-phenoxyphenyl)acet-
amide ([11C]DAA1106, [11C]1, Scheme 1), for imaging
PBR in the primate brain.13,14 The in vitro and in vivo
evaluation of mice, rats, and monkeys demonstrated
that [11C]1 had high specific binding with PBR in the
brains and had about a 4-fold higher uptake into
the monkey brain than [11C]PK 11195,15,16 a commonly
Keywords: Peripheralbenzodiazepine receptor; Deuterium substitu-
tion; Defluorination; PET.
*
Corresponding author. Tel.: +81 43 206 4041; fax: +81 43 206
3261; e-mail: zhang@nirs.go.jp
0968-0896/$ - see front matter ꢀ 2004 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2004.11.058

1812
M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
It was demonstrated that defluorination of the fluoro-
CH3
CH2
OCH3
O
methylmoiety was a main metaboilc route of [ 18F]2 as
shown in Scheme 2.18 The carbon–hydrogen (C–H)
bond of the [18F]fluoromethylmoiety may firstyl be at-
tacked by enzyme and cleaved, followed by the elimina-
tion of hydrofluoride from the same a carbon atom to
generate a carbene structure (4). Carbene 4 was stabi-
lized by resonating it with the adjacent oxygen atom,
which further promoted the decomposition of [18F]2.
Thus, cleavage of the C–H bond is considered as a
rate-limiting step, which contributes to the metabolism
of [18F]2.
C
N
F
O
OR
[
[
[
11C]DAA1106 ( [11C]1]: R = 11CH3
18F]FMDAA1106 ( [18F]2): R = 18FCH2
18F]FEDAA1106 ( [18F]3): R = 18FCH2CH
Scheme 1. Chemicalstructures of [ 11C]DAA1106 ([11C]1) analogues.
In this study, we designed a deuterium-substituted ana-
logue of [18F]2, N-(5-fluoro-2-phenoxyphenyl)-N-(2-
[18F]fluoromethoxy-d2-5-methoxybenzyl)acetamide ([18F]5),
as a novelradioilgand for PBR ( Scheme 2). The design
principle was as follows: since the deuterium atom has
a similar and bioisoteric property to the hydrogen atom,
deuterium substitution may only have minimal influence
on the potentialaffinity of 5 for PBR. On the other
hand, the substitution can be expected to reduce the rate
of defluorination of the [18F]fluoromethylmoiety since
carbon–deuterium (C–D) bond is generally stronger to
break than the C–H bond. We firstly synthesized the
non-radioactive 5 and its 18-fluorine labeled analogue
[18F]5 and measured the in vitro binding affinity of
5 for PBR. We then compared the metabolic rate
of [18F]5 with that of [18F]2 in the plasma and brain
of mouse, and examined the regionaldistribution of
[18F]5 in the mice, rats, and monkey.
used PET ligand for imaging PBR in the primate brain.
Now, [11C]1 is being used to investigate PBR in the hu-
man brain in our facility.
Using [11C]1 as a lead compound, we further synthesized
and evaluated two novel [18F]alkyl analogues
[18F]FMDAA1106 ([18F]2) and [18F]FEDAA1106
([18F]3) by replacing the positron emitter 11C for 18F
(Scheme 1).17,18 The fluoromethylanaolgue
2 (IC50:
1.71 nM) and fluoroethylanalogue 3 (0.77 nM) had sim-
ilar or higher potency for PBR than 1 (1.62 nM). Ligand
[18F]3 had a similar property to [11C]1 as a potent and
specific PET ligand for PBR, and showed 1.5 times
higher uptake into the monkey brain than [11C]1. This
promising results led to the clinical investigation of
PBR using [18F]3 to follow [11C]1. In contrast, although
[18F]2 also passed through BBB and entered into the
brain, this ligand exhibited a much higher uptake in
monkey skulls, and it had at least 50 times higher radio-
activity in the bones of mice than [18F]3 at 120 min after
injection. Metabolite analysis demonstrated that [18F]2
was rapidly decomposed to [18F]FÀ in the plasma and
brain of mice. The high accumulation of [18F]FÀ into
the bone with a long residence time should decrease
the effective signaland give a olw sensitivity of PET
image of the brain. Moreover, the presence of [18F]FÀ
in the brain could interfere with the determination of
ÔrealÕ specific binding of [18F]2 to PBR, and augment
non-specific binding in the examined regions.
2. Results and discussion
2.1. Chemistry
The non-radioactive fluoromethyl-d2 analogue 5 was
synthesized as shown in Scheme 3. The reaction of
17
excess diiodomethane-d2 (CD2I2, 6) with N-(5-fluoro-
2-phenoxyphenyl)-N-(2-hydroxy-5-methoxybenzyl)acet-
amide19 (7) in the presence of NaH at 25 ꢁC gave
iodomethyl-d2 analogue 8 and a dimmer of 7 as a main
CH3
CH2
OCH3
O
C
N
F
H
In vivo metabolism
CH3
OCH3
O
18F-
O
+
O-C:
C
N
F
4
CH2
H
O
O-C 18F
CH3
CH2
OCH3
D
O
H
C
N
18F]2
deuterium substitution
[
F
O
O-C 18
F
D
[
18F]FDDAA1106 ([18F]5)
Scheme 2. In vivo metabolism of [18F]2 and design of [18F]5.

M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
1813
thesis time of 50 3 min. The identity of the desired
product was confirmed by co-injection with the non-
radioactive 5 using analytic HPLC. In the final product
solution, the radiochemical purity of [18F]5 was higher
than 98% with a specific activity of 40–70 GBq/lmol
as determined from the mass measured by the HPLC/
UV analysis. No significant peak corresponding to 7
or other impurities relative to 7 was observed in the
HPLC chart for the finalproduct. Moreover, the radio-
chemicalpurity of [ 18F]5 remained >95% after it was left
at 25 ꢁC for 180 min, and this product was stable as a
PET ligand while performing animal experiments.
CH
OCH
OR
3
3
O
C
N
F
CH
7: R = H
8: R = ICD
2
O
2
5: R = FCD
2
18
18
[
F]5: R = FCD
2
Bu NF
4
7
CD I
8
5
2 2
b
a
6
18
[
F]KF
c
7
18
18
6
Route 1
Route 2
[
F]FCD I
[ F]5
2.2. Evaluation
2
d
18
[
F]9
In a previous study, the rapid metabolism of fluoro-
methylilgand [ 18F]2 by defluorination in the brain and
plasma precluded its usefulness for the clinical inÀvestiga-
tion of PBR.18 The high accumulation of [18F]F in the
bone including the skull with a long residence time could
decrease the effective signalof PET imaging of [ 18F]2 in
the brain and create difficulties on the data analysis.
Furthermore, the presence of [18F]FÀ might augment
the non-specific binding in the examined regions, and
then interfere with the determination of the specific
binding for PBR. It was found that [18F]FÀ was gener-
ated by in vivo defluorination, which might be initiated
by cleavage of the C–H bond as a rate-limiting step.
Therefore, to reduce the rate of defluorination starting
from break of the C–H bond, we considered the useful-
ness of the isotopic effect of deuterium on hydrogen.
Due to the lower zero-point energy (difference: 6–9 kJ/
mol) and the lower vibration frequency (difference:
788 cmÀ1) of the C–D bond than the C–H bond, the
C–D bond is generally more difficult to break than the
C–H bond.22 It was reported that the cleaving rate of
the C–H bond was about 6.7 times faster at 25 ꢁC,
and 3.3 times faster at 200 ꢁC than that of C–D.22 There
were severalexampel s of the deuterium-substituted PET
ligands for MAO B23 and 5-HT2A24 receptors improving
the properties of the corresponding non-deuterated li-
gands by reducing their in vivo metabolic rates. In this
study, we compared the in vitro binding affinity, meta-
bolic rate and regional distribution of [18F]5 with
those of the non-deuterated [18F]2.
18
[
F]KF
e
18
8
[
F]5
Scheme 3. Synthesis of 5 and [18F]5: (a) NaH, DMF, 25 ꢁC, 3 h, 15%;
(b) THF, reflux, 8 h, 64%; (c) o-dichlorobenzene, 110 ꢁC, 2 min, 23%;
(d) NaH, DMF, À15 ꢁC, 2 min, 92%; (d) DMF, 30–120 ꢁC, 1–15 min,
0–35%.
by-product. The treatment of 8 with tetrabutylammo-
nium fluoride afforded 5 at a totalyiedl of 10% starting
from 7.
The radiosynthesis of [18F]5 was performed by two
routes, respectively (Scheme 3). One route was a two-
step reaction sequence, which involved the synthesis
of the radioactive intermediate [18F]fluoromethyl-d2
iodide20 ([18F]FCD2I, [18F]9), followed by the alkylation
of phenolprecursor 7 with [18F]9. Another route was
the direct nucleophilic replacement of 8 with [18F]FÀ.
In the two-step method, the intermediate [18F]9 was
prepared by the fluorination of 6 with [18F]FÀ in the
presence of K2CO3 and 4,7,13,16,21,24-hexaoxa-1,10-
diazabicyclo[8,8,8]hexacosane (Kryptofix 222) as
a
phase-transfer agent by using an automated synthesis
device.21 The formed [18F]9 was distilled from the
reaction mixture, passed through short columns filled
with Ascarite and phosphorus pentoxide, and trapped
in a DMF solution containing 7 and NaH at À15 ꢁC.
This procedure gave a pure and highly reactive [18F]9,
since the distillation could leave behind all non-volatile
impurities such as metalions from the cycol tron target,
the unreacted [18F]FÀ and Kryptofix 222/K2CO3. After
[18F]9 trapping was completed, the reaction of 7 with
[18F]9 proceeded perfectly to form [18F]5. In the direct
method, heating 8 with [18F]FÀ at 30–120 ꢁC could
give [18F]5, but the radiochemicalyiedl was not repro-
ducible (0–35%). Moreover, purifying the product from
the reaction mixture was often complicated because
excess Kryptofix/K2CO3 and the impurities resulting
from the target significantly decreased the purification
efficiency. Thus, [18F]5 was synthesized by the two-
step reaction method via [18F]9 for animalexperiments.
2.2.1. In vitro binding affinity for PBR and CBR. The in
vitro binding affinities (IC50) of 5 and 8 for PBR were
measured from competition for the [11C]1 binding using
quantitative autoradiography according to the proce-
dures as described previously.18,25 As shown in Table
1, 5 displayed a similar affinity than 2 for PBR, and this
value was also little lower than that of 1 and about 4-
fold higher than PK 11195. The iodomethyl-d2 analogue
8 had about 27 times weaker affinity than 5, suggesting
that the relative bulk group contained in this series of
derivatives was not favorable for expressing the binding
potency of PBR. On the other hand, 5 and 8 displayed
negligible affinities for central benzodiazepine receptor
(CBR) measured by using CBR-selective [11C]flu-
mazenil. The high selectivity of PBR/CBR of the two
Semi-preparative HPLC purification on the reverse
phase for the reaction mixture gave [18F]5 in 13 6%
(n = 4) radiochemicalyidel based on [ 18F]FÀ, cor-
rected for the decay at the end of bombardment in a syn-
analogues may be due to their structural difference from
20
the typicalbenzodiazepine structure.
Although the

1814
M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
Table 1. In vitro binding affinity (IC50) for PER and CBR, and
octanol/phosphate buffer distribution coefficient (log P)
100
90
80
70
60
50
40
30
20
10
0
CH
OCH
3
3
O
C
N
F
CH
2
O
OR
18
18
[
[
F]5 in brain
F]2 in brain
18
[
F]5 in plasma
F]2 in plasma
18
IC50(nM)a
log Pd
[
Ligand
R
PBRb
CBRc
5
FCD2
ICD2
FCH2
CHS
1.90
52.30
1.71
>10,000
>10,000
>10,000
>10,000
>10,000
3.76
4.26
3.70
3.65
8
2
1
PK11195
1.62
8.26
0
10
20
30
time (min)
40
50
60
a Values represent the mean obtained from nine concentrations of
compound using at least eight slices of rat brain (n = 3).
Figure 1. Percentage (mean s.d., n = 3) conversion of [18F]5 to
metabolite in the mouse plasma and brain at several time points after
i.v. injection of [18F]5 (5–10 MBq) into the mice. The unchanged [18F]5
and metabolite were analyzed by HPLC for CH3CN extracts from the
plasma and brain homogenate prepared as described in the
experimental.
b [1C]DAA1106 was incubated in the presence of the compounds
examined.
c
[
11C]Flumazenil was incubated in the presence of the compounds
examined.
d The log P values were determined in the phosphate buffer (pH = 7.4)/
octanol system using the shaking flask method. All results were
presented as mean values (n = 3) with a maximum range of 5%.
HPLC. We conducted two-way repeated ANOVA to
compare the time courses of [18F]5 with those of [18F]2
in the brain and plasma, respectively. In the plasma,
no significant interaction was determined between the
two ligands. The half life (T1/2) in the plasma
was 2.575 min for [18F]5 and 2.367 min for [18F]2. In
deuterium atom is larger than hydrogen, the deuterium
substitution does not alter the molecular similarity and
bioisoteric property between the two functionalgroups
of O-CD2F and O-CH2F. The difference between the
two molecules is too small to have significant influence
on their affinities for PBR.
the brain,
a significant interaction (F1/16 = 83,93,
P < 0.01) was determined between them. The half
life (T1/2) of [18F]5 in the brain was >60 min, whereas
that of [18F]2 was only 2.227 min. The mechanism of
the different metabolic rates between the two ligands
in the brain was that the C–D bond of [18F]5 was more
difficult to break by enzyme than the C–H bond of
[18F]2, which was the rate-limiting step contributory to
their metabolism.
2.2.2. Metabolism in the plasma and brain of mouse. The
metabolite analysis of [18F]5 was performed in the
brain and plasma of mice. After i.v. injection of [18F]5
(8 MBq) into mice (n = 3) at designated time points,
the plasma and brain homogenate were obtained and
immediately deproteinized with CH3CN. Extraction of
the radioactivity in the plasma and brain homogenate
into CH3CN was efficient, and the recovered yield was
>70%. The radioactive metabolite of [18F]5 was identi-
fied using HPLC with a highly sensitive detector26 for
radioactivity. In all samples, a radioactive component
was determined in addition to the unchanged [18F]5.
As estimated by its retention time (tR = 2.0 min), this
metabolite was much more polar than [18F]5
(tR = 10.2 min). Moreover, the tR of the peak remained
unchanged when treated with a base or acid. Using ion
exchange chromatography and TLC, this product was
assigned to [18F]FÀ, which was identified by co-injecting
and developing with non-radioactive standard KF solu-
tion. As it was improbable brain permeable, [18F]FÀ was
produced in the plasma and brain, respectively. This re-
sult suggested that defluorination of the [18F]fluorom-
ethylmoiety was the main metaboilc route of [ 18F]5,
and the deuterium substitution did not change its main
metabolism route.
The difference of the metabolic rates of [18F]2 and [18F]5
in mouse brain was remarkable, compared with that in
the mouse plasma. We considered that a same enzyme
is related to their metabolism in the plasma and brain.
In general, the level of enzyme metabolizing a parent li-
gand in plasma is much higher that in brain. Therefore,
the different level of the same enzyme metabolizing
[18F]2 and [18F]5 in the brain and plasma was a contrib-
uting factor for the difference of their metabolic rates.
The level of enzyme attacking the C–H and C–D bonds
in the plasma may be too high to distinguish difference
of the two bonds, whereas that in the brain may be
apt to realize the difference.
2.2.3. Regional distribution. The radioactivity distri-
bution of [18F]5 in the mice was examined and comp-
ared to that of [18F]2 at 5–120 min after injection. Table
2 shows the decay-corrected percent dose per gram (%
ID/g) data of [18F]2 and [18F]5 in the brain, blood and
bone. The radioactivity level of [18F]5 in the bone was
1.6% ID/g at 120 min after injection, and was 30% of
[18F]2 level in the bone. On consideration of the ten-
Figure 1 shows the percentages of unchanged [18F]5 and
[18F]2 in the plasma and brain of mice measured by

M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
1815
Table 2. Distribution (% injected dose/g tissue: mean s.d., n = 3) of [18F]ligand in Mice at 5, 15, 30, 60, and 120 min after injection
Ligand
Tissue
5 min
15 min
30 min
60 min
120 min
[
18F]2
Blood
Brain
Bone
0.95 0.09
2.65 0.85
0.58 0.23
1.34 0.54
3.26 1.07
1.98 1.99
0.76 0.32
4.39 1.38
2.51 1.35
0.54 0.18
4.04 1.95
3.91 0.86
0.49 0.18
3.47 0.94
5.64 0.72
[
18F]5
Blood
Brain
Bone
0.78 0.23
2.31 0.92
0.58 0.29
0.92 0.15
2.93 1.01
0.74 0.05
0.83 0.20
3.68 1.43
1.23 0.08
0.86 0.33
3.62 0.85
1.59 0.01
0.52 0.15
3.24 0.70
1.63 0.04
dency of [18F]FÀ to readily accumulate in bone, the
radioactivity examined in the bone was mostly due to
[18F]FÀ. In contrast, [18F]2 and [18F]5 displayed a simi-
lar level of radioactivity in the brain and blood. The
cause may be that the two ligands had similar affinity
for PBR, lipophilicity and other physiochemical proper-
ties. Although they were metabolized in the brain in dif-
ferent rates after they entered into the brain, the main
radioactive metabolite [18F]FÀ may be retained in the
brain without a rapid clearance, which did not give sig-
nificant difference for their total radioactivity levels in
the brain.
from the brain rapidly, which led to a different distribu-
tion of [18F]FÀ in the whole brain. In fact, the ratio of
olfactory bulb to frontal cortex was 2.1 0.3 for
[18F]5, whereas the ratio for [18F]2 was only 1.6 0.5
in the corresponding regions. This finding supported
that the [18F]FÀ level was lower for [18F]5 than for
[18F]2 in the whole brain including the frontal cortex.
Figure 3 shows a typicalPET summation image of mon-
key brain acquired from 30 to 180 min after [18F]5 injec-
tion (80 MBq/1.5 mL). Unexpectedly, although a high
radioactivity was observed in the monkey brain, the
PET image of [18F]5 displayed much higher accumula-
tion of radioactivity in skull. Moreover, the level of
radioactivity of [18F]5 in the monkey skull was similar
to that of [18F]218 by comparing their PET images at a
normalized injection dose. This image gave a visual evi-
dence that [18F]5 was metabolized by in vivo defluorina-
tion to [18F]FÀ, which was accumulated into the bone.
Figure 2 shows ex vivo autoradiograms of rat brains at
30 min after injection of [18F]5. As can be seen in the
sagittalsection of the controlbrain ( Fig. 2a), [18F]5
showed a high uptake, consistent with the distribution
data measured above (Table 2). High radioactivity levels
were observed in the olfactory bulb and cerebellum,
whereas low uptakes were seen in the other brain regions
such as the frontalcortex and striatum. The uptake pat-
tern of radioactivity was in accordance with the regional
distribution of PBR in the brain and was similar to those
of [11C]1 and [18F]3.14,17,27 Co-injection of [18F]5 with
the non-radioactive 1 (1 mg/kg) produced a reduction
of radioactivity in the brain regions having PBR (Fig.
2b). The radioactivity level in the olfactory bulb and cer-
ebellum was decreased to 30–40% of the control, while
the reduction ratio of [18F]2 by 1 was about 50%.18 In
spite of remarkable difference of the metabolic rate of
the two ligands in the brain, the difference of their non-
specific bindings determined by ex vivo autoradiography
was not so significant. From the metabolic route of
[18F]2 and [18F]5, we considered that the nonspecific
binding of the two ligands was mainly due to [18F]FÀ
in the brain. Although [18F]FÀ does not bind to the
receptor, the [18F]FÀ might be partly retained in the
olfactory bulb and cerebellum, and could not be released
by the co-injection of 1, which kept the nonspecific bind-
ing of [18F]2 and [18F]5 at a similar level. The left [18F]FÀ
might spread around the brain and did not clear out
In this study, the isotopic effect induced a distinct reduc-
tion in the metabolic rate of [18F]5 in the brain of mice
Figure 3. PET summation image of the monkey brain acquired
between 30 and 180 min after [18F]5 injection (80 MBq).
Figure 2. Ex vivo autoradiogram of [18F]5 in the sagittalsections of rat brains at 30 min after injection (20–25 MBq). (a) [ 18F]5; (b) [18F]5+1
(1 mg/kg).

1816
M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
although the deuterium substitution did not change the
main metabolism profile of [18F]5. The decreased rate of
[18F]5 reduced the uptake of [18F]FÀ in the bone of mice,
and increased slightly the percentage of specific binding
to PBR accounting for the totalbinding as reflected in
the autoradiogram of the rat brain. However, the PET
image of [18F]5 for the monkey brain was not improved
in comparison with [18F]2, which was disturbed by
[18F]FÀ accumulating into the skull. The significant dif-
ference of radioactivity distribution in the bones of mice
or rats and monkeys may reflect the species difference
between rodents and primates. Therefore, unlike [11C]1
and [18F]3, [18F]5 may be not a usefulPET radioilgand,
even though it entered the brain and had some specific
binding with PBR in the brain. The high accumulation
of [18F]FÀ in the bone should preclude the in vivo use-
fulness of [18F]5 for determining PBR binding in the pri-
mate brain. Moreover, the uptake of [18F]FÀ should
decrease the effective signalof PET image and the accu-
racy of data if this ligand was employed to quantita-
tively measure the PBR density in the brain.
chased from Aldrich Chem. (Milwaukee) and Wako
Pure Chem. Ind. (Osaka). The animalexperiments were
performed according to the recommendations of the
committee for the care and use of laboratory animals,
NationalInstitute of Radiool gicalSciences (NIRS).
4.2. Chemical synthesis
4.2.1. N-(5-Fluoro-2-phenoxyphenyl)-N-(2-iodomethoxy-
d2-5-methoxybenzyl)acetamide (8). A mixture of N-(5-
fluoro-2-phenoxyphenyl)-N-(2-hydroxy-5-methoxyben-
zyl)acetamide19 (7, 38 mg, 0.10 mmol), CD2I2 (6, 32 lL,
0.40 mmol), and NaH (8 mg, 0.34 mmol) in DMF
(2 mL) was stirred at 25 ꢁC for 3 h. The reaction was
stopped with AcOEt, and rinsed with water and satu-
rated NaCl solution. After the organic layer was dried
over Na2SO4, the solvent was removed to give a residue.
Column chromatograph of the residue on silica gel with
CHCl3/hexane (1/20) gave 8 (8.0 mg, 15%) as a colorless
1
oil; H NMR (300 MHz, CDCl3) d: 7.18–7.40 (2H, m),
6.15–7.33 (9H, m), 4.71 (dd, J = 7, 46 Hz), 3.78 (3H,
s), 2.20 (3H, s); FABMS (m/z): 524.1 (M++1); Anal.
(C23H19D2FINO4) C, H, N.
3. Conclusion
4.2.2.
N-(5-Fluoro-2-phenoxyphenyl)-N-(2-fluorometh-
In this study, the deuterium-substituted [18F]5 for PBR
was designed, synthesized, and evaluated. Ligand
[18F]5 was synthesized by the alkylation of the desmethyl
precursor 7 with [18F]9 in reproducible radiochemical
yields. Compared with the non-deuterated [18F]2,
[18F]5 displayed a slower metabolic rate in the brain of
mice, a lower uptake into the mouse bone, and higher
specific binding to PBR in the rat brain with a similar
binding affinity for PBR. However, the deuterium sub-
stitution did not decrease the radioactivity level of
[18F]5 in the monkey skull due to its remarkable in vivo
defluorination, suggesting species difference between ro-
dents and primates. Although [18F]5 was not be used for
the in vivo investigation of PBR in the primates, our re-
sults indicated that isotope effect is an effective tool for
improving in vivo behaviors of a parent radioligand.
oxy-d2-5-methoxybenzyl)acetamide (5). A mixture of 8
(16 mg, 0.03 mmol) and tetrabutylammonium fluoride
(1.0 M solution in THF, 2.0 mL) was heated at reflux
for 8 h. After THF was removed, the reaction mixture
was re-dissolved in AcOEt, and rinsed with water and
saturated NaCl solution. After the organic layer was
dried over Na2SO4, the solvent was removed to give a
residue. Column chromatograph of the residue on silica
gelwith CHCl /hexane (1/20) gave 5 (10 mg, 70%) as a
3
colorless oil; H NMR (300 MHz, CDCl3) d: 7.38–7.49
(2H, m), 6.71–7.13 (8H, m), 6.28–6.44 (1H, m), 4.75
(2H, dd, J = 7, 46 Hz), 3.71 (3H, s), 2.12 (3H, s);
FABMS (m/z): 416.3 (M++1); Anal. (C23H19D2F2NO4)
C, H, N.
1
4.3. Radiosynthesis
4.3.1. [18F]Fluoride ([18F]FÀ). [18F]FÀ was produced by
the 18O(p,n) 18F reaction on 10–20 atom% H218O using
18 MeV protons (14.2 MeV on target) from the cyclo-
tron and separated from [18O]H2O using Dowex 1-X8
anion exchange resin in an irradiating room. The pro-
duced [18F]FÀ was released from the resin with aqueous
4. Experimental
4.1. General
1H NMR spectra were recorded on a JNM-GX-270
spectrometer (JEOL, Tokyo) with tetramethylsilane as
K2CO3 (3.3 mg/0.3 mL) into a vialcontaining CH CN
3
an internalstandard. Allchemicalshifts (
d) were re-
(1.5 mL)/Kryptofix 222 (25 mg) and transferred into a
reaction vessel set in a hot cell.
ported in parts per million (ppm) downfield from the
standard. FAB-MS were obtained on a JEOL NMS-
SX102 spectrometer (JEOL, Tokyo). Column chroma-
tography was performed on Merck Kieselgel gel 60
F254 (70–230 mesh). 18-Fluorine (18F) was produced
by 18O(p,n)18F nuclear reaction using a CYPRIS HM-
18 cyclotron (Sumitomo Heavy Industry, Tokyo). If
not otherwise stated, radioactivity was measured with
an IGC-3R Curiemeter (Aloka, Tokyo). HPLC was per-
formed using a JASCO HPLC system (JASCO, Tokyo):
effluent radioactivity was monitored using a NaI (Tl)
scintillation detector system. All chemical reagents with
the highest grade commercially available were pur-
4.3.2.
N-(5-Fluoro-2-phenoxyphenyl)-N-(2-[18F]fluoro-
methoxy-d2-5-methoxybenzyl)acetamide ([18F]5). After
[18F]FÀ in the reaction vesselwas dried to remove
H2O and CH3CN at 120 ꢁC for 20 min, 6 (200 lL) in
o-dichlorobenzene (200 lL) was added into the radioac-
tive mixture. Under a He flow (50 mL/min), the resulted
[18F]FCD2I ([18F]9) was immediately distilled at 110 ꢁC
for 2 min and bubbled in another reactor containing 7
(1.5 mg) and NaH (10 lL, 1.5 g/20 mL DMF) in anhy-
drous DMF (300 lL) at À15 ꢁC. After maximum radio-
activity was trapped into the solution, the reaction was

M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
1817
terminated by adding CH3CN/H2O (6/4, 500 lL) and
the reaction mixture was applied to a semi-preparative
HPLC system. HPLC purification was performed on
YMC JÕsphere ODS-H80 column (10 mm ID · 250 mm)
using a mobile phase of CH3CN/H2O (60/40) at a flow
rate of 6.0 mL/min. The retention time (tR) for [18F]5
was 12.8 min, while that for 7 was 6.7 min. The radioac-
tive fraction corresponding to [18F]5 was collected in a
sterile flask containing polysorbate (80) (75 lL) and eth-
anol(150 lL), evaporated to dryness under vacuum, re-
dissolved in sterile normal saline (7 mL) and passed
through a 0.22 lm Millipore filter to obtain the final
product. At the end of synthesis, 100–180 MBq (n = 4)
of [18F]5 was obtained as an i.v. injectable solution at
a beam current of 10–15 lA and 20–25 min proton
bombardment.
mined, respectively. The specific binding for PBR or
CBR was defined as totalbinding minus nonspecific
binding of [11C]1 or [11C]flumazenil. PSL data corre-
sponding to specific binding at each compound concen-
tration were calculated as a percentage in relation to the
controlspecific binding, and were converted to probit
values to determine the IC50 of each compound.
4.6. Metabolite analysis for mouse plasma and brain
tissue
After i.v. injection of [18F]5 or [18F]2 (5–10 MBq/
100 lL) into ddy mice (n = 3), these mice were sacrificed
by cervical dislocation at 5, 15, 30, 60 or 120 min. Blood
(0.7–1.0 mL) and whole brain samples were removed
quickly, respectively. The blood sample was centrifuged
at 15,000 rpm (3615 Model, Kubota, Tokyo) for 2 min
at 4 ꢁC to separate plasma, which (250 lL) was collected
in a test tube containing CH3CN (500 lL) and 5 or 2
(10 lL, 1.5 mg/5.0 mL of CH3CN). After the tube was
vortexed for 15 s and centrifuged at 15,000 rpm for
2 min for deproteinization, the supernatant was col-
lected. The extraction efficiency of radioactivity into
the CH3CN supernatant ranged from 70–92% of the to-
talradioactivity in the palsma. On the other hand, the
cerebellum and forebrain including the olfactory bulb
were dissected from the mouse brain and homogenized
together in an ice-cooled CH3CN/H2O (1/1, 1.0 mL)
solution. The homogenate was centrifuged at
15,000 rpm for 2 min at 4 ꢁC and the supernatant was
collected. The recovery percentage of radioactivity into
the supernatant was 74–90% of the totalradioactivity
in the brain homogenate.
4.4. Radiochemical purity and specific activity
determinations
Radiochemicalpurity was assayed by anaylticalHPLC
(column: CAPCELL PAK C18, 4.6 mm ID · 250 mm,
UV at 254 nm; mobile phase: CH3CN/H2O = 6/4). The
tR for [18F]5 was 6.1 min at a flow rate of 2.0 mL/min.
The specific activity of [18F]5 was calculated by compar-
ing the assayed radioactivity (GBq) with the mass mea-
sured from a calibration curve obtained by the carrier
UV peaks at 254 nm corresponding to five known
concentrations.
4.5. In vitro binding assays
Male Sprague–Dawley rats (n = 3) weighing 220–250 g
were sacrificed by decapitation under ether anesthesia,
and their brains were quickly removed and frozen on
powdered dry ice. Brain sagittalsections (20 lm) were
cut on a cryostat microtome (HM560, Leica, Bensheim)
and thaw-mounted on glass slides (Matsunami Glass
Ind., Tokyo), which were then dried at 25 ꢁC and stored
at À18 ꢁC untilused for experiments. The brain sections
were pre-incubated at 25 ꢁC for 20 min in 50 mM Tris–
HCl(pH 7.4) buffer. After pre-incubation, these sections
were incubated at 37 ꢁC for 30 min in assay buffer con-
taining [11C]1 (about 1 nM, specific activity: 100 GBq/
lmol) or [11C]flumazenil(about 1 nM, specific activity:
190 GBq/lmol). To determine the IC50 values of 5 and
8 for the [11C]1 (for PBR) or [11C]flumazenil(for
CBR) binding, the brain sections were incubated with
[11C]1 or [11C]flumazenilusing 5 and 8 at nine different
concentrations (0.1–1000 nM), respectively. Ten
micrometers of 1 or flumazenilwas used to determine
the nonspecific binding for PBR or CBR in the brains.
After incubation, the brain sections were washed three
times for 2 min each time with cold assay buffer, dipped
into cold distilled water for 10 s, and dried with a warm
air flow (about 50 ꢁC). These sections were then placed
in contact with imaging plates (BAS-SR 127, Fuji Photo
Film, Tokyo) for 60 min to analyze their radioactivity
distribution with a FUJIX BAS 1800 bioimaging ana-
lyzer (Fuji). Region of interest (ROI) in the sections
was identified in the cerebellum. PSL data correspond-
ing to radioactivity in the cerebellum in the presence
and absence of the displacement 5 and 8 were deter-
An aliquot of the supernatant (100–500 lL) obtained
from the plasma or brain homogenate was injected into
the HPLC with a highly sensitive detector25 for radioac-
tivity, and analyzed under the same HPLC conditions
described above except the mobile phase of CH3CN/
H2O with a ratio of 1/1. The percent ratio of [18F]ligand
to the totalradioactivity (corrected for decay) on the
HPLC chromatogram was calculated as % = (peak area
for [18F]ligand/total radioactive peak area) · 100.
4.7. Biodistribution in mice
A saline solution of [18F]5 or [18F]2 (average of 8 MBq/
200 lL, specific activity: 55 GBq/lmol) was injected into
ddy mice (30–40 g, 9 weeks, male; n = 3) through the tail
vein. Three mice for each time point were sacrificed by
cervicaldisolcation at 5, 15, 30, 60, and 120 min after
injection, respectively. The whole brain, bone and blood
samples were quickly removed and weighed. The radio-
activity present in the various tissues were measured in
a Packard autogamma scintillation counter, and ex-
pressed as a percentage of the injected dose per gram
of wet tissue (% ID/g). All radioactivity measurements
were corrected for decay.
4.8. Ex vivo autoradiography
A saline solution of [18F]5 (20 MBq/200 lL, specific
activity: 65 GBq/lmol) or
a
mixture of [18F]5

1818
M.-R. Zhang et al. / Bioorg. Med. Chem. 13 (2005) 1811–1818
6. Batra, S.; Iosif, C. S. Int. J. Oncol. 1998, 12, 1295.
7. Miyazawa, N.; Hamel, E.; Diksic, M. J. Neurooncol. 1998,
38, 19.
8. Venturini, I.; Zeneroli, M. I.; Corsi, L.; Avallone, R.;
Farina, F.; Alho, H.; Baraldi, C.; Ferrarese, C.; Pecora,
N.; Frigo, M.; Ardizzone, G.; Arrigio, A.; Pellicci, R.;
Baraldi, M. Life Sci. 1998, 63, 1269.
9. Camsonne, R.; Crouzel, C.; Comar, D.; Maziere, M.;
Prenant, C.; Sastre, J.; Moulin, M. A.; Syrota, A. J.
Labelled Comp. Radiopharm. 1984, 21, 985.
10. Pike, V. W.; Halldin, C.; Crouzel, C.; Barre, L.; Nutt, D.
J.; Osman, S.; Shah, F.; Turton, D. R.; Waters, S. L. Nucl.
Med. Biol. 1993, 20, 503.
(25 MBq/200 lL) and 1 (1 mg/kg, 200 lL) was injected
into a male Sprague–Dawley rat (220–250 g, 9 weeks,
male) through the tail vein. At 30 min after injection,
the rat was sacrificed by decapitation under ether anes-
thesia, and the brain was quickly removed and frozen
on powdered dry ice. Brain sagittalsections (20 lm)
were cut on a cryostat microtome (HM560) and thaw-
mounted on glass slides (Matsunami Glass Ind., Tokyo),
dried with warm air flow (about 50 ꢁC). These sections
were then placed in contact with imaging plates (BAS-
SR 127) for 60 min to analyze the radioactivity distribu-
tion with the FUJIX BAS bioimaging analyzer (Fuji).
11. Cappelli, A.; Anzini, M.; Vomero, S.; De Benedetti, P. G.;
Menziani, M. C.; Giorgi, G.; Manzoni, C. J. Med. Chem.
1997, 40, 2910.
4.9. Monkey PET
12. Matarrese, M.; Moresco, R. M.; Cappelli, A.; Anzini, M.;
Vomero, S.; Simonelli, P.; Verza, E.; Magni, F.; Sudati, F.;
Soloviev, D.; Todde, S.; Carpinelli, A.; Kienle, M. G.;
Fazio, F. J. Med. Chem. 2001, 44, 579.
13. Zhang, M.-R.; Kida, T.; Noguchi, J.; Furutsuka, K.;
Maeda, J.; Suhara, T.; Suzuki, K. Nucl. Med. Biol. 2003,
30, 513.
14. Maeda, J.; Suhara, T.; Zhang, M.-R.; Okauchi, T.;
Ichimiya, T.; Inajii, M.; Ohbayashi, S.; Suzuki, K. Synapse
2004, 52, 283.
15. Pappata, P.; Cornu, Y.; Samson, C.; Prenant, J.; Bena-
vides, B.; Scatton, C.; Crouzel, J. J.; Hauw, A.; Syrota, A.
J. Nucl. Med. 1991, 32, 1608.
16. Cagnin, A.; Brooks, D. J.; Kennedy, A. M.; Gunn, R. N.;
Myers, R.; Turkheimer, F. E.; Jones, T.; Banati, R. B.
Lancet 2001, 358, 461.
17. Zhang, M.-R.; Maeda, J.; Furutsuka, K.; Yoshida, Y.;
Ogawa, M.; Suhara, T.; Suzuki, K. Bioorg. Med. Chem.
Lett. 2003, 13, 201.
PET scan was performed using a high-resolution SHR-
7700 PET camera (Hamamatsu Photonics, Hamamatsu)
designed for laboratory animals, which provides 31
transaxialsilces 3.6 mm (center-to-center) apart and a
33.1 cm field of view. A male rhesus monkey (Macaca
mulatta) weighing about 5 kg was repeatedly anesthe-
tized with ketamine (Ketalarꢂ, 10 mg/kg/h, i.m.) every
hour throughout the session. After a transmission scan
for attenuation correction for 1 h using 74 MBq
68Ge–68Ga source, a dynamic emission scan in 3D acqui-
sition mode was performed for 180 min (2 min · 5 scans,
4 min · 10 scans, 10 min · 4 scans). All emission scan
images were obtained by treatment using an image
analysis software.28,29 A saline of [18F]5 (80 MBq) was
injected i.v. into the monkey, and time-sequential
tomographic scanning was performed on a transverse
section of the brain for 180 min.
18. Zhang, M.-R.; Maeda, J.; Ogawa, M.; Noguchi, J.; Ito, T.;
Yoshida, Y.; Okauchi, T.; Obayashi, S.; Suhara, T.;
Suzuki, K. J. Med. Chem. 2004, 47, 2228.
Acknowledgements
19. Okubo, T.; Yoshikawa, R.; Chaki, S.; Okuyama, S.;
Nakazato, A. Bioorg. Med. Chem. 2004, 12, 423.
20. Zhang, M.-R.; Ogawa, M.; Furutsuka, T.; Yoshida, Y.;
Suzuki, K. J. Fluoro. Chem. 2004, 126, 1879.
21. Zhang, M.-R.; Tsuchiyama, A.; Haradahira, T.; Yoshida, Y.;
Furutsuka, T.; Suzuki, K. Appl. Radiat. Isot. 2002, 57,
335.
22. Melander, L.; Saunder, W. H., Jr. Reaction Rates of
Isotopic Molecules; John Wiley & Sons, 1980.
23. Flowler, J. S.; Wang, G. J.; Logan, J.; Xie, S.; Volkow, N.
D.; Schlyer, D. J.; Pappas, N.; Alexoff, D. L.; Patlak, C. J.
Nucl. Chem. 1995, 36, 1255.
The authors are gratefulto Taisho PharmaceuticalCo.,
Ltd for giving us the samples (DAA1106: 1 and precur-
sor: 7) and helpful suggestions. We also thank the staff
of the Cyclotron Operation Section, Radiopharmaceuti-
calChemistry Section and Brain Imaging Project of Na-
tionalInstitute of RadioolgicalSciences (NIRS) for
their support in the operation cyclotron operation,
radioisotope production and PET experiment.
References and notes
24. Staley, J. K.; Van Dyck, C. H.; Tan, P.-Z.; Staley, J. K.;
Tikriti, M.; Ramsby, Q.; Klump, H.; Ng, C.; Grag, P.;
Soufer, R.; Baldwin, R. M.; Innis, R. B. Nucl. Med. Biol.
2001, 28, 271.
1. Braestrop, C.; Squires, R. F. Proc. Natl. Acad. Sci. U.S.A
1977, 74, 1839.
2. Papadopoulos, V.; Amri, H.; Li, H.; Yao, Z.; Brown, R.
C.; Vidic, B.; Culty, M. Structure, function and regulation
of the mitochondrialperipher-atlype benzodiazepine
receptor. J. Pharmacol. Exp. Ther. 2001, 299, 793–800.
3. Diorio, D.; Welner, S.; Butterworth, R.; Meaney, M.;
Suranyl-Cadotte, R. Neurobiol. Aging 1991, 12, 255.
4. Banati, R. B.; Newcombe, J.; Gunn, R. N.; Cagnin, A.;
Turkheimer, F.; Heppner, F.; Price, G.; Wegner, F.;
Giovannoni, G.; Miller, D. H.; Perkin, G. D.; Smith, T.;
Hewson, A. K.; Bydder, G.; Kreutzberg, G. W.; Jones, T.;
Cuzner, M. L.; Myers, R. Brain 2000, 123, 2321.
5. Raghavendra Rao, V. L.; Dogan, A.; Bowen, K. K.;
Dempsy, R. J. Exp. Neurol. 2000, 161, 102.
25. Zhang, M.-R.; Haradahira, T.; Maeda, J.; Okauchi, T.;
Kawabe, K.; Kida, T.; Obayashi, S.; Suzuki, K.; Suhara,
T. Nucl. Med. Biol. 2002, 29, 469.
26. Takei, M.; Kida, T.; Suzuki, K. Appl Radiat Isot. 2001, 55,
229.
27. Chaki, S.; Funakoshi, T.; Yoshikawa, R.; Okuyama, S.;
Okubo, T.; Nakazato, A.; Nagamine, M.; Tomisawa, K.
Eur. J. Pharmacol. 1999, 371, 197.
28. Okauchi, T.; Suhara, T.; Maeda, J.; Kawabe, K.; Obay-
ashi, S.; Suzuki, K. Synapse 2001, 41, 87.
29. Maeda, J.; Suhara, T.; Kawabe, K.; Okauchi, T.; Obay-
ashi, S.; Hojo, J.; Suzuki, K. Synapse 2003, 47,
200.