Synthesis and PET evaluation of (R)-[S-methyl-11C] thionisoxetine, a candidate radioligand for imaging brain norepinephrine transporters

Article abstract of DOI:10.1002/jlcr.1128

Introduction: (R)-3-(2-(methylthio)phenoxy)-N-methyl-3-phenylpropan-1-amine [(R)-thionisoxetine; 1] is a potent inhibitor of the norepinephrine transporter (NET). We aimed to label 1 with carbon-11 (t_1/2 = 20.4 min) for evaluation as a radioligand for imaging NET in living brain with positron emission tomography (PET). Methods: Methyl 3-(2-((R)-3-(methylamino)-1- phenylpropoxy)phenylthio)-propanoate (MPPP) and 1 were each prepared from o-hydroxythiophenol in three steps. Treatment of MPPP with potassium t-butoxide and [¹¹C]methyl iodide in tetrahydrofuran gave [5-methyl- ¹¹C]thionisoxetine ([¹¹C]1), which was purified with HPLC. The distribution of radioactivity in brain after intravenous injection of [¹¹C]1 into cynomolgus monkey was followed with PET and the appearance of radiometabolites in plasma monitored with radio-HPLC. Results: [¹¹C]1 was obtained in high yield from [¹¹C]methyl iodide. Of the radioactivity injected into monkey, 2.4% entered brain. Ratios of radioactivity in thalamus, mesencephalon, occipital cortex and caudate to that in cerebellum at 93 min were 1.3, 1.2, 1.2 and 1.1, respectively. The radioactivity in plasma corresponding to unchanged radioligand decreased to 53% at 45 min, with the remainder represented by hydrophilic radiometabolites. Conclusions: MPPP is an effective precursor for ¹¹C-methylation to [¹¹C]1, suggesting that the S-γ-propionic acid methyl ester protecting group may have wider value in the ¹¹C-labeling of aryl methyl sulfides. However, the relatively low ratios of radioactivity to the cerebellum together with an unexpected accumulation of radioactivity in the caudate, makes [¹¹C]1 an unpromising NET radioligand. Copyright

Full text of DOI:10.1002/jlcr.1128

JOURNAL OF LABELLED COMPOUNDS AND RADIOPHARMACEUTICALS
J Label Compd Radiopharm 2006; 49: 1007–1019.
Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/jlcr.1128
Research Article
Synthesis and PET evaluation of (R)-[S-
methyl-11C]thionisoxetine, a candidate radioligand for
imaging brain norepinephrine transporters
Magnus Schou1,2,*, Victor W. Pike2, Andrea Varrone1, Bala
zs Gulya
´ ´
s1,
Lars Farde1 and Christer Halldin1
1 Karolinska Institutet, Department of Clinical Neuroscience, Psychiatry Section,
Karolinska Hospital, S-17176 Stockholm, Sweden
2 Molecular Imaging Branch, National Institute of Mental Health, Building 10, Rm. B3
C346A, National Institutes of Health, Bethesda, MD 20892-1003, USA
Summary
Introduction: (R)-3-(2-(methylthio)phenoxy)-N-methyl-3-phenylpropan-1-amine [(R)–
thionisoxetine; 1] is a potent inhibitor of the norepinephrine transporter (NET). We
aimed to label 1 with carbon-11 (t1/2 = 20.4 min) for evaluation as a radioligand for
imaging NET in living brain with positron emission tomography (PET).
Methods: Methyl 3-(2-((R)-3-(methylamino)-1-phenylpropoxy)phenylthio)-propanoate
(MPPP) and 1 were each prepared from o-hydroxythiophenol in three steps. Treatment of
MPPP with potassium t-butoxide and [11C]methyl iodide in tetrahydrofuran gave
[S-methyl-11C]thionisoxetine ([11C]1), which was purified with HPLC. The distribution of
radioactivity in brain after intravenous injection of [11C]1 into cynomolgus monkey was
followed with PET and the appearance of radiometabolites in plasma monitored with
radio-HPLC.
Results: [11C]1 was obtained in high yield from [11C]methyl iodide. Of the
radioactivity injected into monkey, 2.4% entered brain. Ratios of radioactivity in
thalamus, mesencephalon, occipital cortex and caudate to that in cerebellum at 93 min
were 1.3, 1.2, 1.2 and 1.1, respectively. The radioactivity in plasma corresponding to
unchanged radioligand decreased to 53% at 45 min, with the remainder represented by
hydrophilic radiometabolites.
Conclusions: MPPP is an effective precursor for 11C-methylation to [11C]1,
suggesting that the S-g-propionic acid methyl ester protecting group may have wider
*Correspondence to: M. Schou, Karolinska Institutet, Department of Clinical Neuroscience, Psychiatry
Section, Karolinska Hospital, S-17176 Stockholm, Sweden. E-mail: magnus.schou@ki.se
Contract/grant sponsor: National Research Council of Italy
Contract/grant sponsor: National Institute of Mental Health
Copyright # 2006 John Wiley & Sons, Ltd.

1008
M. SCHOU ET AL.
value in the 11C-labeling of aryl methyl sulfides. However, the relatively low ratios
of radioactivity to the cerebellum together with an unexpected accumulation of
radioactivity in the caudate, makes [11C]1 an unpromising NET radioligand.
Copyright # 2006 John Wiley & Sons, Ltd.
Received 30 March 2006; Revised 7 August 2006; Accepted 7 August 2006
Key Words: norepinephrine transporter; PET; monkey; radioligand; carbon-11
Introduction
Although the norepinephrine transporter (NET) has long been recognized as a
target in the treatment of several neuropsychiatric disorders,1–5 the detailed
mapping of NET in the human brain, including the determination of regional
concentrations (Bmax), has been hindered through the lack of an effective
radioligand. [3H]nisoxetine,6 after initial studies with rodent tissue,7, 8
appeared to be suitable for imaging NET in human brain in vitro. Never-
theless, only the binding of this radioligand to the two highest NET density
regions in post mortem human brain, the locus coeruleus and the raphe nuclei,
has been described.1, 9 Furthermore, ex vivo studies with [11C]nisoxetine (11C,
t
1/2=20.4 min) in mice10 and positron emission tomography (PET) measure-
ments with its eutomer, (R)-[11C]nisoxetine, in baboons11 have failed to show a
useful NET-specific signal.
Among several other evaluated radioligands,11–21 analogs of reboxetine
(Edronax1) so far show the best properties for imaging primate brain NET
in vivo. These include the O-methyl, O-fluoromethyl and O-fluoroethyl
21
analogs, namely (S,S)-[11C]MeNER,18–20 (S,S)-[18F]FMeNER-D2
and
(S,S)-[18F]FRB-D4 11, respectively. In human studies with (S,S)-[11C]MeNER,
measurements of NET suffered from slow pharmacokinetics rendering the
PET data vulnerable to noise.22 Test-retest variabilities were therefore too
large (17–34%) for assessing discrete changes in NET inhibition following
reboxetine treatment. (S,S)-[18F]FMeNER-D2 gives a signal of similar
magnitude to that of (S,S)-[11C]MeNER in primates and, because of the
longer half-life of fluorine-18 (t1/2 = 109.8 min), also allows PET data to be
acquired at the time of specific binding equilibrium, an asset to biomathema-
tical modeling of the PET data and derivation of NET-radioligand binding
parameters.23 Although (S,S)-[18F]FMeNER-D2 could be used to image NETs
in the locus coeruleus on post mortem human brain hemispheres in vitro, the
radioligand also failed to detect NET in low density regions.24 Therefore,
(S,S)-[18F]FMeNER-D2 may have limited sensitivity for investigations of
regional NET densities in human subjects with PET.
We aim to develop a more sensitive radioligand for PET imaging of NET
across the entire living human brain. (R)-thionisoxetine (1; Figure 1) is
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1009
XMe
OR
O
O
NHMe
NH
O
Figure 1. Structures of the NET inhibitors, (R)-thionisoxetine (X = S; 1),
(R)-nisoxetine (X = O; 2), (S,S)-MeNER (R = Me) and (S,S)-FMeNER-D2
(R = FCD2)
reportedly about three-fold more potent (IC50, 1.9 nM) than (R)-nisoxetine
(IC50, 5.8 nM, 2; Figure 1).25 Such an increase in binding affinity, though
modest, may be critical for achieving sensitivity for imaging NET with PET,
since on theoretical grounds, binding potential (BP) is inversely proportional
to binding affinity. 1 is also 68-fold selective for binding to NET versus the
serotonin transporter (SERT).25 Due to its very close structural similarity to
the highly NET-selective 2,26 1 is also not expected to have any significant
affinity towards the dopamine transporter (DAT) or any other transporter or
neurotransmitter. By inspection, the lipophilicity of 1 is expected to be only
slightly greater than that of 2. On this basis, we considered 1 to be worthy of
labeling with carbon-11 and evaluation as a new PET radioligand for NET.
1 appeared amenable to labeling through either N- or S-methylation
with, for example, [11C]methyl iodide. We chose to synthesize a precursor for
S-11C-methylation, since such a precursor might also allow preparation of the
S-[18F]fluoromethyl analog of 1. S-[18F]fluoromethyl analogs are known
to be chemically stable, as exemplified by [18F]FMcN5652,27 whereas few
N-fluoromethyl compounds are stable enough to survive isolation.28 A longer-
lived radioligand might be valuable if the 11C-labeled ligand showed
unacceptably slow pharmacokinetics in vivo. Given the susceptibility of thiols
to oxidize, we aimed to prepare an S-protected precursor that might be stored
and then deprotected in situ for radiomethylation with [11C]methyl iodide,
such as, for example, an S-acetyl ester.29 We further aimed to evaluate [11C]1
as a radioligand for brain NET in cynomolgus monkey with PET.
Results and discussion
Chemistry
We were interested to obtain an S-protected precursor that would be stable to
storage and directly useful for preparing [11C]1. In view of a previous report, in
which an S-acetyl ester was used as a precursor for S-11C-methylation,29
we attempted to protect the thiol group of the starting material,
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

1010
M. SCHOU ET AL.
o-hydroxy-thiophenol, as a thioester by treatment with acetyl chloride or
acetic anhydride. However, consistent with a previous report,30 very low
regioselectivity was observed under all attempted conditions. In contrast, we
found methylation and benzylation of o-hydroxy-thiophenol to be thiol-
selective. We therefore re-directed our attention to removal of S-methyl,
S-benzyl and S-p-methoxybenzyl groups from analogs of 1 with sodium
thiomethoxide, sodium metal in ethanol or mercury(II) trifluoroacetate.31–33
However, these reactions failed to proceed cleanly, as previously observed in
the particular case of demethylations of some aryl methyl sulfides with sodium
thiomethoxide.34
Aryl sulfides may be protected as their sulfanyl g-propionic acid
methyl esters and deprotected by b-elimination under basic conditions.35
We therefore decided to explore the Michael addition of methyl acrylate to
o-hydroxy-thiophenol as the first step in the synthesis of 1. Alkylation of
o-hydroxy-thiophenol with methyl acrylate was found to be highly thiol-
selective; for example, a four-fold molar excess of methyl acrylate gave only
S-alkylation. To obtain enantiomerically pure 1 and 5, commercially available
(S)-3-chloro-1-phenylpropan-1-ol was first coupled with the respective
o-hydroxy-thiophenol derivative, under Mitsunobu conditions. This type of
reaction proceeds with complete inversion of configuration.36, 37 Treatment of
the chloro products with methylamine gave the enantiomerically pure free
bases of 1 and 5 in 19 and 7% overall yields, respectively (Scheme 1). The
synthetic sequence for 1 was similar to that described earlier for related
compounds.36
OH
SR
O
SR
O
Cl
SH
OH
SR
OH
MeNH2
A
Cl
NHMe
OMe
DEAD
PPh3
KHCO3
DMF
A= MeI or
methyl acrylate
6, R = Me
3, R=
7, R = Me
4, R=
1, R = Me
5, R=
OMe
OMe
O
O
O
Scheme 1. Synthesis of (R)-thionisoxetine (1) and its sulfanyl g-propionic acid
methyl ester precursor (MPPP, 5)
Radiochemistry
[11C]1 was obtained in >99% decay-corrected radiochemical yield (as assessed
by analytical radio-HPLC of the radiomethylation mixture) from the
intermediate thiol 8 and [11C]methyl iodide (Scheme 2) within 30 min from
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1011
S
SH
O
O
11
11
CH3
[
C]MeI
t.BuOK
O
+
5
O
THF, RT
RT
N
H
N
H
11
[
C]1
8
Scheme 2. Radiolabeling of 1 via alkylation of the intermediate thiol (8) formed
in situ by b-elimination of the c-propionic acid methyl ester moiety in 5
the end of radionuclide production and with a radiochemical purity exceeding
99%. Base-promoted in situ b-elimination of the sulfanyl g-propionic
acid methyl ester moiety in 5 to generate the free thiolate ion (8) was
thus effective for accomplishing 11C-methylation. The high radiochemical
yield of [11C]1 indicates a wider potential utility for this protective group
in the radiosyntheses of other S-[methyl-11C] or S-[18F]fluoromethyl
radiotracers. The specific radioactivity of [11C]1 was 8.2 Ci/mmol (304 GBq/
mmol) at the time of administration to monkey. Precursor (5) was absent
from the product formulation, which was radiochemically stable for at
least 90 min.
c Log P calculations
Both neutral and ionized species of 1 were calculated to be more lipophilic
than 2. The c Log P values of 1 and 2 were 3.94 and 3.52, respectively, and the
c Log D values were 1.47 and 1.02, respectively.
PET measurements
After injection of [11C]1 into cynomolgus monkey (1.38 mCi; 51 MBq; 0.06 mg
carrier), radioactivity readily entered brain (maximally 2.4% of injected dose
at 14 min after injection) with a regional distribution approximately matching
that expected for NET; however, radioactivity also accumulated in striatum
(caudate), a region shown to be low in NET density in rat,8 cat38 and human.39
This accumulation is most likely non-specific, but it cannot be ruled out
that [11C]1 binds to another non-identified binding site. The rank order for
radioactivity accumulation in different brain regions was: thalamus
>mesencephalon ꢀoccipital cortex >caudate >cerebellum (Figure 2).
Cerebellum, has a NET density similar to striatum in these species,8,38,39
Also, the cerebellum shows radioactivity accumulation similar to striatum in
PET measurements with (S,S)-[11C]MeNER in cynomolgus monkeys (Schou
et al., unpublished data). Therefore, in this study we decided to use cerebellum
as an alternative reference region. Ratios of radioactivity concentration in
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

1012
M. SCHOU ET AL.
700
600
500
400
300
200
100
0
thalamus
mesencephalon
occipital cortex
caudate
cerebellum
0
20
40
60
80
100
Time (min)
Figure 2. Regional distribution of radioactivity in brain following intravenous
injection of [11C]1 into a cynomolgus monkey
thalamus, mesencephalon and occipital cortex to that in caudate were 1.15, 1.0
and 1.0 at 93 min, respectively. Corresponding ratios to cerebellum were 1.3,
1.2 and 1.2. The ratio of caudate to cerebellum radioactivity at 93 min was 1.1.
The maximal ratio of radioactivity in any ROI to the cerebellum in PET
measurements with [11C]1 was not greater than 1.3 (for thalamus). This ratio is
lower than that observed with the NET radioligands, (S,S)-[11C]MeNER and
(S,S)-[18F]FMeNER-D2, which show thalamus to striatum ratios of about
1.5–1.6.19, 21 This lower ratio, together with the accumulation of radioactivity
in the striatum, therefore severely limits the utility of [11C]1 as a PET
radioligand for NET. Since with [11C]1 there was no sign of wash-out of
radioactivity from the striatum nor any sign that specific binding in thalamus
increased with time, we expended no effort in preparing the S-[18F]fluor-
omethyl analog of this radioligand.
Ex vivo studies with (S,S)-[11C]MeNER in the rodent brain gave a signal to
noise ratio (hypothalamus to striatum) of about 2.5,20 whereas the
corresponding ratio in non-human primates was about 1.5–1.6.19 It has been
reported that the density of NETs in human cortex is about nine times lower
than in rodents.40 Although this may not be the case for all brain regions, a
new structural class of radioligands displaying very high affinity for NET will
probably be required for detailed mapping of NET throughout the entire
primate brain.
Analysis of radiometabolites in plasma
The overall recovery of radioactivity from the analytical procedure was
between 88 and 91%. After injection of [11C]1 into monkey, unchanged
radioligand (tR, 5.6 min) represented 95% of the radioactivity in plasma at
4 min, 70% at 15 min, 40% at 30 min and 53% at 45 min. Radiometabolites of
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1013
[11C]1 in monkey plasma were all highly hydrophilic based on their lack of
retention on the reverse phase HPLC column (i.e. elution at the void volume)
under the employed conditions. It is thus unlikely, although not impossible,
that radiometabolites cross the blood–brain barrier and contribute to brain
radioactivity.
Conclusions
The thiol moiety of o-hydroxy-thiophenol, a starting material in the synthesis
of 1 and [11C]1, was regioselectively protected by reaction with methyl
acrylate. This protective group was removable in situ under mild conditions
during 11C-methylation of precursor to [11C]1 in high radiochemical yield.
Following intravenous injection of [11C]1 into cynomolgus monkey, there was
uptake of radioactivity into NET-rich regions, but only to a slightly greater
extent than into NET-poor regions. Hence, [11C]1 is inferior to existing PET
radioligands, such as (S,S)-[11C]MeNER or (S,S)-[18F]FMeNER-D2, for the
study of NET. Nevertheless, the protecting group applied in the synthesis of
[11C]1 may find wider useful application in the labeling of other aryl alkyl
sulfides.
Experimental
Chemistry
Materials and general procedures. Diethyl azodicarboxylate (DEAD) (40% in
toluene) was obtained from TCI (USA). All other reagents and solvents (e.g.
N,N-dimethylformamide (DMF), tetrahydrofuran (THF)), including anhy-
drous solvents, were obtained from Sigma-Aldrich (USA) and used without
further purification. Reactions were carried out under a dry inert atmosphere
with dry solvents, unless otherwise stated. Thin layer chromatography was
performed on silica gel layers (60 F254; Merck) with compound detection
under UV light (l = 254 nm). Flash column chromatography was performed
on silica gel (60 A; 70–230 mesh; Aldrich). 1H-NMR (400 MHz) and 13C-NMR
spectra (100 MHz) were recorded on a Bruker drx400 spectrometer for
solutions in CDCl3 or MeOD-d4 [both with TMS (d=0.00) as internal
standard] at 300 K. Mass spectrometry data were acquired on a quadrupole-
orthogonal injection time-of-flight (TOF) mass spectrometer (QTOF1,
Waters/Micromass Plc, Manchester, UK). Samples were sprayed from gold-
plated borosilicate capillaries (Protana AS, Odense, Denmark) using a
capillary voltage of +1 kV.
Methyl 3-(2-hydroxyphenylthio)propanoate (3). To a solution of o-hydroxy-
thiophenol (4 g; 31.7 mmol) in DMF (30 ml) was added potassium bicarbonate
(2.73 g; 31.7 mmol). The solution was stirred at room temperature (RT) for
30 min, after which methyl acrylate (3.1 g; 31.7 mmol) was added. The reaction
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

1014
M. SCHOU ET AL.
mixture was stirred for 4 h, diluted with water and extracted with ether. The
organic layer was dried (MgSO4), solvents were removed in vacuo and the
crude product purified by flash column chromatography (hexanes-EtOAc;
1
20: 1, v/v) to yield 3 as a yellow oil (5.17 g; 24.7 mmol, 78%). H-NMR
(CDCl3): 7.45 (dd, J = 1.6, 7.7 Hz, 1 H), 7.26 (m, 1 H), 7.03 (s, 1 H), 6.99 (dd,
J = 1.4, 8.1 Hz, 1 H), 6.86 (dt, J = 1.3, 7.2 Hz, 1 H), 3.68 (s, 3 H), 2.96 (t, J =
7.1 Hz, 2 H), 2.56 (t, J = 7.1 Hz, 2 H). 13C-NMR (CDCl3): 172.2, 157.3, 135.9,
131.2, 120.6, 118.1, 115.3, 51.9, 34.0, 30.8. MS (TOF ES+): 213.08
[analytically (M+1)] 212.05 (calculated.)
Methyl 3-(2-((R)-3-chloro-1-phenylpropoxy)phenylthio)propanoate (4). To a
solution of (S)-3-chloro-1-phenylpropan-1-ol (3.9 g, 22.6 mmol) and 3 (2.4 g;
11.3 mmol) in THF (75 ml) at 08C was added triphenyl phosphine (5.9 g;
22.6 mmol) and DEAD (3.9 g; 22.6 mmol). The reaction mixture was allowed
to attain RT and was stirred overnight, after which solvents were removed
in vacuo. The sticky crude product was purified by flash column chromato-
graphy (hexanes–EtOAc; 10: 1, v/v) to yield 4 as a thick yellow oil (1.4 g;
1
3.5 mmol; 31%). H-NMR (CDCl3): 7.22–7.40 (6 H), 6.97 (dq, J = 0.8, 1.7,
7.7 Hz, 1 H), 6.84 (dd, J = 1.2, 7.5 Hz, 1 H), 6.66 (dd, J = 1.1, 8.3 Hz, 1 H),
5.42 (m, 1 H), 3.89 (m, 1 H), 3.68 (s, 3 H), 3.60 (m, 1 H), 3.18 (m, J = 1.1 Hz,
2 H), 2.67 (t, J = 7.5 Hz, 2 H), 2.51 (m, 1 H), 2.23 (m, 1 H). 13C-NMR
(CDCl3): 172.3, 156.8, 142.1, 131.2, 120.6, 118.1, 115.3, 51.9, 34.0, 30.8. MS
(TOF ES+): 365.14 [analytically (M+1)] (calculated) 364.09.
Methyl 3-(2-((R)-3-(methylamino)-1-phenylpropoxy)phenylthio)propanoate
(MPPP, 5). A stirred solution of 4 (1.0 g; 2.74 mmol) and methylamine
(10 ml; 2 M in THF) was heated at 908C in a sealed vessel for 4 h, after which
solvents were evaporated off. The crude product was dissolved in hydrochloric
acid (0.1 M) and washed with dichloromethane. The aqueous phase was then
basified with potassium hydroxide pellets and extracted with ether. Purifica-
tion with flash column chromatography (CH2Cl2–MeOH–NH4OH; 9:1:0.1 by
1
vol.) gave 5 as a sticky white wax (300 mg; 0.82 mmol; 30%). H-NMR
(MeOD-d4): 7.17–7.41 (6 H), 6.97 (m, 1 H), 6.83 (m, 1 H), 6.63 (m, 1 H), 5.33
(m, 1 H), 3.68 (s, 3 H), 3.21 (m, 2 H), 2.85 (m, 2 H), 2.67 (t, J = 7.5 Hz, 2 H),
2.46 (s, 3 H), 2.2 (m, 2 H). 13C-NMR (MeOD-d4): 172.3, 156.8, 142.1, 131.2,
120.6, 118.1, 115.3, 51.9, 34.0, 30.8. MS (TOF ES+): 360.19 [analytically
(M+1)] 359.16 (calculated). [a]D = 84.08 (c. = 5, MeOH).
2-(Methylthio)phenol (6). Aryl methyl sulfide 6 was obtained from
o-hydroxy-thiophenol (2.15 g; 17.0 mmol), potassium bicarbonate (1.35 g;
17.0 mmol) and methyl iodide (2.42 g; 17.0 mmol) by the same method as for
3. Purification with flash column chromatography (CHCl3) gave 6 as a yellow
1
liquid (1.8 g; 12.9 mmol, 76%). H-NMR (CDCl3): 7.47 (dd, J = 1.6, 7.7 Hz,
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1015
1 H), 7.24 (m, 1 H), 6.98 (dd, J = 1.3, 8.3 Hz, 1 H), 6.87 (dt, J = 1.3, 7.54 Hz,
1 H), 6.67 (s, 1 H), 2.31 (s, 3 H). 13C-NMR (CDCl3): 154.2, 132.7, 128.6, 119.1,
112.4, 17.3. MS (TOF ES+): 141.18 [analytically (M+1)] 140.03 (calculated).
(2-((R)-3-chloro-1-phenylpropoxy)phenyl)(methyl)sulfane (7). Compound 7
was obtained from 6 (1.00 g; 7.1 mmol), (S)-3-chloro-1-phenylpropan-1-ol
(2.44 g; 14.3 mmol), triphenyl phosphine (3.75 g, 14.3 mmol) and DEAD
(2.44 g; 14.3 mmol) under Mitsunobu conditions similar to those described for
4. Purification with flash column chromatography (hexanes–EtOAc; 10: 1,
v/v), gave 7 as a thick colorless oil (1.0 g; 3.4 mmol; 48%). 1H-NMR (CDCl3):
7.18–7.38 (5 H), 7.05 (m, 1 H), 6.86 (m, 1 H), 6.61 (m, 1 H), 5.40 (m, 1 H), 3.87
(m, 1 H), 3.87 (m, 1 H), 3.61 (m, 1 H), 2.47 (m, 1 H), 2.37 (s, 3 H), 2.19 (m, 1 H).
13C-NMR (CDCl3): 172.8, 156.5, 142.4, 132.8, 129.2, 128.6, 128.1, 127.7,
126.6, 125.4, 125.4, 125.1, 121.9, 112.8, 80.1, 52.2, 48.5, 34.9, 37.5, 36.4, 29.1.
MS (TOF ES+): 293.23 [analytical (M+1)] 292.07 (calculated).
(R)-3-(2-(Methylthio)phenoxy)-N-methyl-3-phenylpropan-1-amine ((R)-thio-
nisoxetine, 1). 7 (900 mg; 3.1 mmol) was heated with methylamine (10 ml; 2 M
in THF) as described for the preparation of 5. Purification by flash column
chromatography (CH2Cl2–MeOH–NH4OH; 9: 1: 0.1 by vol.) gave 8 as a sticky
1
brown oil (350 mg; 1.2 mmol, 51%). H-NMR (MeOD-d4): 7.16–7.39 (5 H),
7.07 (m, 1 H), 6.86 (m, 1 H), 6.59 (m, 1 H), 5.32 (m, 1 H), 3.61 (s, 1 H), 2.83
(m, 2 H), 2.42 (s, 3 H), 2.40 (s, 3 H), 2.24 (m, 1 H), 2.10 (m, 1 H). 13C-NMR
(MeOD-d4): 172.3, 156.0, 141.1, 130.0, 128.7, 128.6, 127.4, 127.2, 126.0, 125.8,
125.8, 123.9, 121.2, 113.7, 79.2, 51.9, 48.1, 37.8, 36.5, 35.9, 27.2. MS (TOF
ES+): 288.16 [analytically (M+1)] 287.13 (calculated). [a]D = 80.78 (c. = 0.5,
MeOH).
Radiochemistry
Materials and general procedures. [11C]Methane was produced at the
Karolinska Hospital with a PETtrace cyclotron (GE Medical Sciences) using
16 MeV protons in the 14N(p,a)11C reaction on nitrogen containing hydrogen
(10%). [11C]methane was isolated from the target gas on a cooled [N2 (l)]
Porapak Q trap and subsequently converted into [11C]methyl iodide by radical
iodination in a recirculation system.41
Preparative HPLC was performed using a reverse phase m-Bondapak C-18
column (300 Â 7.8 mm, 10 mm; waters) eluted with mobile phase system A
(MeCN–aq. HCO2NH4, (0.1 M), 35:65 v/v). The column outlet was connected
with an UV absorbance detector (l = 254 nm) in series with a GM-tube for
radiation detection.
The radiochemical purity of [11C]1 was determined by reverse phase HPLC
equipped with a m-Bondapak C-18 column (300 Â 3.9 mm; 10 mm; Waters) and
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

1016
M. SCHOU ET AL.
an absorbance detector (l = 254) in series with a b-flow detector (Beckman)
for radiation detection. The HPLC column was eluted with mobile phase
system B (MeCN–aq. H3PO4 (10 mM); 35:65 v/v). The retention time of 1 was
5 min.
The specific radioactivity was determined by HPLC using a m-Bondapak
C-18 column (300 Â 3.9 mm; 10 mm; Waters) eluted with mobile phase system
C (MeCN–H3PO4 (50 mM); 35:65 v/v). The level of non-radioactive precursor
(5) present in the formulated solution was determined under the same
conditions.
(R)-[S-methyl-11C]-3-(2-(methylthio)phenoxy)-N-methyl-3-phenylpropan-1-
amine. To a suspension of 5 (1 mg, 2.8 mmol) in THF (400 ml) at RT was added
potassium t-butoxide (6ml; 1.0 M in THF). The reaction vial was sealed, mixed
and kept at RT for 15–45 min until [11C]methyl iodide was delivered to the vial
in a helium stream. After entrapment of [11C]methyl iodide was complete,
mobile phase (system A; 600 ml) was added to the crude reaction mixture which
was then injected onto the preparative HPLC column. The product fraction
which eluted at 9 min was collected and concentrated to dryness. The residue
was taken up in sterile disodium phosphate-buffered saline (pH 7.4; 8 ml) and
filtered through a sterile Millipore filter (0.22 mm), yielding [11C]1 in sterile
solution free from pyrogens.
c Log P calculations
c Log P and c Log D values of 1 and 2 were calculated with Pallas 3.0 for
Windows software (CompuDrug International Inc., San Fransisco CA, USA).
PET measurements
The PET experimental procedure was similar to that employed for other
radioligands. An ECAT EXACT HR PET system (Siemens) was run in 3D
mode. The spatial resolution is about 3.8 mm full width half maximum. Images
were displayed as 47 brain sections with a separation of 3.3 mm.42
One cynomolgus monkey (4900 g) was supplied by the National Institute for
Infectious Disease Control (Solna, Stockholm). The study was approved by
the Animal Ethics Committee of Northern Stockholm. The cynomolgus
monkey was anaesthetized treated as previously described before and during
the PET measurement,19, 43 in which the monkey was injected with [11C]1
(1.38 mCi; 51 MBq) as a bolus into the left sural vein. Radioactivity in brain
was measured according to a pre-programmed sequence of frames during
93 min.
Regions of interest (ROI’s) (cerebellum, caudate, mesencephalon, occipital
cortex, thalamus and whole brain) were delineated as previously described in
detail elsewhere.19 Caudate, which is a region almost devoid of NET,1, 38, 39
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1017
was initially used as a reference region for free radioligand concentration and
non-specific binding in brain. The cerebellum was subsequently used as an
alternative reference region, since it has also often been reported to have low
NET density.1, 38, 39 Binding ratios were thus calculated by dividing radio-
activity in a given ROI with radioactivity in the caudate or cerebellum. To
calculate specific binding, radioactivity in a reference region was subtracted
from the radioactivity in the ROI.
Analysis of radiometabolites in plasma
The HPLC method used to determine the percentages of radioactivity in
monkey plasma corresponding to unchanged radioligand and radiometabo-
lites at set times during the course of a PET experiment was adapted from a
method established for other PET radioligands.19, 44
Mixtures of phosphoric acid (10 mM) (D) and acetonitrile (E) were used as
the mobile phase at 6.0 ml/min in the following elution program; 0–5.0 min,
(D/E) 80/20–30/70; 5.0–7.5 min, (D/E) 30/70–80/20; 7.5–10 min (D/E) 80/20
isocratic. The radioactive peak having a retention time corresponding to 1 was
integrated and its area expressed as a percentage of the sum of the areas of all
detected radioactive peaks (decay-corrected).
Acknowledgements
The authors are grateful to the National Institutes of Health-Karolinska
Institutet Graduate Training Partnership in Neuroscience for a Studentship to
Dr Magnus Schou. We also thank Mr Arsalan Amir and Mr Phong Truong
and other members of the PET group at the Karolinska Institutet. Dr Andrea
Varrone was supported by the Short-term Mobility Program of the National
Research Council of Italy. This work was also partly supported by the
Intramural Research Program of the NIH (National Institute of Mental
Health).
References
1. Tejani-Butt SM, Yang JX, Zaffar H. Brain Res 1993; 631: 147–150.
2. Klimek V, Stockmeier C, Overholser J, Meltzer HY, Kalka S, Dilley G,
Ordway GA. J Neurosci 1997; 17: 8451–8458.
3. Ressler KJ, Nemeroff CB. Biol Psychiatry 1999; 46: 1219–1233.
4. Spencer T, Heiligenstein JH, Biederman J, Faries DE, Kratochvil CJ,
Conners CK, Potter WZ. J Clin Psychiatry 2002; 63: 1140–1147.
5. Brunello N, Kasper S, Leonard B, Montgomery S, Nelson JC, Paykel E,
Versiani M, Racagni G. Eur Neuropsychopharmacol 2002; 12: 461–475.
6. Tejani-Butt SM, Brunswick DJ, Frazer A. Eur J Pharmacol 1990; 191: 239–243.
7. Cheetham SC, Viggers JA, Butler SA, Prow MR, Heal DJ. Neuropharmacology
1996; 35: 63–70.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

1018
M. SCHOU ET AL.
8. Tejani-Butt SM. J Pharm Exp Ther 1992; 260: 427–436.
9. Ordway GA, Stockmeier CA, Cason GW, Klimek V. J Neurosci 1997; 17:
1710–1719.
10. Haka MS, Kilbourn MR. Nucl Med Biol 1989; 16: 771–774.
11. Ding YS, Lin KS, Logan J, Benveniste H, Carter P. J Neurochem 2005; 94:
337–351.
12. Schou M, Sovago J, Pike VW, Gulyas B, Bꢀgesꢀ K, Farde L, Halldin C. Mol
´ ´ ´
Imaging Biol 2005; 7: 1–8.
13. Kiyono Y, Kanegawa N, Kawashima H, Kitamura Y, Iida Y, Saji H. Nucl Med
Biol 2004; 31: 147–153.
14. Kung M-P, Choi S-R, Hou C, Zhuang Z-P, Foulon C, Kung HF. Nucl Med Biol
2004; 31: 533–541.
15. McConathy J, Owens MJ, Kilts CD, Malveaux EJ, Camp VM, Votaw JR,
Nemeroff CB, Goodman MM. Nucl Med Biol 2004; 31: 705–718.
16. Lin K-S, Ding Y-S, Betzel T, Quandt G. J Label Compd Radiopharm 2005; 48:
S152.
17. McConathy J, Owens MJ, Kilts CD, Malveaux EJ, Votaw JR, Nemeroff CB,
Goodman MM. Nucl Med Biol 2005; 32: 593–605.
18. Ding YS, Lin KS, Garza V, Carter P, Alexoff D, Logan J, Shea C, Xu YW,
King P. Synapse 2003; 50: 345–352.
19. Schou M, Halldin C, Sovago J, Pike VW, Gulyas B, Mozley PD, Johnson DP,
´ ´ ´ ´
Hall H, Innis RB, Farde L. Nucl Med Biol 2003; 30: 707–714.
20. Wilson AA, Johnson DP, Mozley P, Hussey D, Ginovart N, Nobrega J.
Garcia A, Meyer J, Houle S. Nucl Med Biol 2003; 30: 85–92.
21. Schou M, Halldin C, So
Dobson D, Shchukin E, Innis RB, Farde L. Synapse 2004; 53: 57–67.
22. Andree B, Seneca N, Schou M, Mozley PD, Potter WZ, Farde L, Gulyas B,
vago J, Pike VW, Hall H, Gulyas B, Mozley PD,
´ ´ ´
´
Halldin C. Eur J Nucl Med Mol Imaging 2004; 31: S208.
23. Farde L, Eriksson L, Blomquist G, Halldin C. J Cereb Blood Flow Metab 1989; 9:
696–708.
24. Schou M, Halldin C, Pike VW, Mozley PD, Dobson D, Innis RB, Farde L,
Hall H. Eur Neuropsychopharmacol 2005; 15: 517–520.
25. Gehlert DR, Hemrick-Luecke SK, Schober DA, Krushinski J, Howbert JJ,
Robertson DW, Wong DT, Fuller RW. Life Sci 1995; 56: 1915–1920.
26. Wong DT, Threlkeld PG, Best KL, Bymaster FP. J Pharmacol Exp Ther 1982;
222: 61–65.
27. Marjamaki P, Zessin J, Eskola O, Gronroos T, Haaparanta M, Bergman J,
Lehikoinen P, Forsback S, Brust P, Steinbach J, Solin O. Synapse 2003; 47: 45–53.
28. Zhang M-R, Ogawa M, Furutsuka K, Yoshida Y, Suzuki K. J Fluorine Chem
2004; 125: 1879–1886.
29. Zessin J, Gucker P, Ametamey SM, Steinbach J, Brust P, Vollenweider FX,
Johannsen B, Schubiger PA. J Label Compd Radiopharm 1999; 42: 1301–1312.
30. McKinnon DM. Can J Chem 1980; 58: 2761–2764.
31. Nishimura O, Kitada C, Fujino M. Chem Pharm Bull 1978; 26: 1576–1585.
32. Hoffman K, Bridgewater A, Axelrod AE. J Am Chem Soc 1949; 71: 1253–1257.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr

(R)-[S-METHYL-11C]THIONISOXETINE—A CANDIDATE NET RADIOLIGAND 1019
33. Suehiro M, Wang TS, Yatabe T, Van Heertum RL, Mann JJ. J Label Compd
Radiopharm 1998; 41: 725–730.
34. Pinchart A, Dallaire C, Bierbeek AV, Gingras M. Tetrahedron Lett 1999; 40:
5479–5482.
35. Becht J-M, Wagner A, Mioskowski C. J Org Chem 2003; 68: 5758–5761.
36. Chumpradit S, Kung MP, Panyachotipun C, Prapansiri V, Foulun C, Brooks BP,
Szabo SA, Tejani-Butt S, Frazer A, Kung HF. J Med Chem 1992; 35: 4492–4497.
37. Mitsunobu O. Synthesis 1981; 1: 1–30.
38. Charnay Y, Leger L, Vallet P, Hof P, Juovet M, Bouras C. Neuroscience 1995; 69:
259–270.
39. Donnan G, Kaczmarczyk S, Paxinos G, Chilco P, Kalnins R, Woodhouse D,
Mendelsohn F. J Comp Neurol 1991; 304: 419–434.
40. Smith H, Smith HR, Beveridge TJ, Porrino LJ. Neuroscience 2006; 138: 703–714.
41. Nagren K, Truong P, Helin S, Amir A, Halldin C. J Label Compd Radiopharm
2003; 46: S76.
42. Wienhard K, Dahlbom M, Eriksson L, Michel C, Bruckbauer T, Pietrzyk U,
Heiss W. J Comput Assist Tomogr 1994; 18: 110–118.
43. Karlsson P, Farde L, Halldin C, Swahn C-G, Sedvall G, Foged C, Hansen K,
Skrumsager B. Psychopharmacology 1993; 113: 149–156.
44. Halldin C, Swahn C-G, Farde L, Sedvall G. Kluwer Academic Publishers:
Dodrecht Comar D. (ed.). PET for Drug Development and Evaluation 1995; 55–65.
Copyright # 2006 John Wiley & Sons, Ltd.
J Label Compd Radiopharm 2006; 49: 1007–1019
DOI: 10.1002/jlcr