Synthesis and evaluation of two fluorine-18 labelled phenylbenzothiazoles as potential in vivo tracers for amyloid plaque imaging

Article abstract of DOI:10.1002/jlcr.1662

Uncharged derivatives of thioflavin-T have known in vitro and in vivo affinity for amyloid β. We synthesized and evaluated two derivatives with a fluorine-18 labelled fluoropropoxy substituent either at the 6-position or at the 2′-position of the 2-(4′-am

Full text of DOI:10.1002/jlcr.1662

Research Article
Received 16 January 2009,
Revised 16 April 2009,
Accepted 12 June 2009
Published online 11 August 2009 in Wiley Interscience
(www.interscience.wiley.com) DOI: 10.1002/jlcr.1662
Synthesis and evaluation of two fluorine-18
labelled phenylbenzothiazoles as potential
in vivo tracers for amyloid plaque imaging
Ã
K. Serdons,a D. Vanderghinste,b M. Van Eeckhoudt,a P. Borghgraef,c
H. Kung,d F. Van Leuven,c T. de Groot,b G. Bormans,a and A. Verbruggena
Uncharged derivatives of thioflavin-T have known in vitro and in vivo affinity for amyloid b. We synthesized and evaluated
two derivatives with a fluorine-18 labelled fluoropropoxy substituent either at the 6-position or at the 20-position of the
2-(40-aminophenyl)-1,3-benzothiazole core with the aim to get suitable radiotracers to perform amyloid plaque imaging.
The fluorine-18 labelled compounds were obtained by nucleophilic substitution of the corresponding tosyl precursors with
[
18F]fluoride with a radiochemical yield of 50%, yielding 6-(300-[18F]fluoropropoxy)-2-(40-aminophenyl)-1,3-benzothiazole
([18F]2) and 2-[20-(300-[18F]fluoropropoxy)-40-aminophenyl]-1,3-benzothiazole ([18F]3) with a specific activity between 33 and
51 GBq/lmol. The identity of the radiolabelled compounds was confirmed using radio-LC-MS and by comparing retention
times on RP-HPLC. Biodistribution studies in healthy mice showed for both compounds a relatively high initial brain uptake,
which was significantly higher for [18F]2 than for [18F]3 (4.5% ID/g versus 3.0% ID/g, po0.05). Wash-out from control brain
was faster for [18F]3. In vitro binding affinity tests using human AD brain homogenates revealed that only compound 2 has
affinity for fibrillar amyloid b (Ki = 14.5 nM). This was confirmed by the incubation of transgenic APP mouse brain sections
with the cold compounds, where 3 did not stain any structure whereas 2 stained amyloid plaques present in APP mouse
brain. These data suggest that [18F]2 may be a useful tracer for in vivo visualization of fibrillar amyloid b.
Keywords: amyloid; diagnosis; PET; fluorine-18; phenylbenzothiazole
compounds bind not only to senile plaques but also to
neurofibrillary tangles, another hallmark of AD. Patient studies
Introduction
Fibrillar amyloid b (Ab) is an aggregated protein found in brain
of patients suffering from Alzheimer’s disease (AD). Therefore, in
vivo visualization of the presence of this protein in the brain
using single photon emission computed tomography (SPECT) or
positron emission tomography (PET) with an appropriate
radiopharmaceutical would allow early diagnosis of this disorder
and evaluation of therapies aiming to reduce the amyloid b
burden.
Several radiolabelled imaging agents binding specifically to
Ab have already been reported (Figure 1).1 The first synthesized
compounds were derived from Congo red or chrysamine G, dyes
useful for staining fibrillar amyloid in vitro,2–5 and derivatives of
styrylbenzene such as X-34,3 BSB,4–9 ISB and IMSB.7 These agents
have a rather high molecular mass and are ionized at
physiological pH, impairing their ability to cross the blood–brain
barrier (BBB) by passive diffusion. Smaller derivatives consisting
of about half the structure of X-34 and lacking the ionizable
groups have been developed. These are labelled with either
carbon-11 (N-[11C]methylamino-40-hydroxystilbene, SB-13) or
iodine-125 (N,N-dimethylamino-3’-[125I]iodostilbene).8,9 Contrary
to the chrysamine G and X-34 derivatives, these stilbenes show a
high initial brain uptake in healthy mice and a nanomolar affinity
for fibrillar amyloid b in vitro. Barrio et al. investigated fluorine-18
labelled derivatives of FDDNP as potential tracers for in vivo
visualization of fibrillar amyloid b.10 They found that these
have shown the ability of these compounds to detect
pathological deposits in the brain of AD patients.11 Other
promising tracer agents for in vivo imaging of amyloid
pathology are an iodine-125 labelled N,N-dimethylamino-
fluorene,12 the carbon-11 labelled PIB (see infra) and the
fluorine-18 labelled styrylbenzoxazole BF-168,13 acridine orange
analog BF-10814 and stilbene 18F-BAY94-9172.15
Over the last years, uncharged derivatives of thioflavin-T (ThT)
have been synthesized and labelled with a variety of radio-
isotopes in the search for a radiolabelled probe for diagnosis of
AD. ThT itself is positively charged due to the presence of a
quaternary nitrogen atom and is not able to cross the BBB by
aLaboratory for Radiopharmacy, K. U. Leuven, O&N2, Herestraat 49 - bus 821,
3000 Leuven, Belgium
bRadiopharmacy, U. Z. Gasthuisberg, Herestraat 49, 3000 Leuven, Belgium
cLaboratory of Experimental Genetics and Transgenesis, K. U. Leuven, O&N1,
Herestraat 49 - bus 602, 3000 Leuven, Belgium
dDepartment of Radiology, University of Pennsylvania, Market Street 3700 -
room 305, Philadelphia, United States
*Correspondence to: K. Serdons, Laboratory for Radiopharmacy, Faculty of
Pharmaceutical Sciences, K. U. Leuven, O&N2, Herestraat 49 - bus 821, 3000
Leuven, Belgium.
E-mail: Kim.serdons@pharm.kuleuven.be
J. Label Compd. Radiopharm 2009, 52 473–481
Copyright r 2009 John Wiley & Sons, Ltd.

K. Serdons et al.
Figure 1. Structures of reported potential tracer agents for imaging of amyloid b.
passive diffusion. Klunk et al. reported that uncharged thioflavin- using a modification of a method described by Shi and co-
T derivatives bind with high affinity to fibrillar amyloid.16 workers.27
Derivatives labelled with iodine-125 (125I-TZDM10 and 2-(30-
For the synthesis of the non-radioactive 3-fluoropropoxy
show derivatives 6-(300-fluoropropoxy)-2-(4’-aminophenyl)-1,3-benzo-
[
125I]iodo-40-aminophenyl)-6-hydroxybenzothiazole17)
high binding affinity for fibrillar amyloid b and are useful for thiazole (2) and 2-[20-(300-fluoropropoxy)-40-aminophenyl]-1,
in vitro experiments. Other compounds labelled with carbon-11 3-benzothiazole (3), 1,3-propanediol di-p-tosylate was mono-
have also been developed, such as [11C]6-Me-BTA-1, [11C]6-Me- fluorinated via a nucleophilic substitution to get 3-fluoropro-
BTA-2 and [11C]6-OH-BTA-1 with the latter, also known as pyltosylate (1), followed by reaction of this intermediate with
Pittsburgh Compound-B or PIB, being the most promising tracer the appropriate hydroxy substituted 2-(40-aminophenyl)-1,3-
agent.18,19 The benzothiazole moiety is apparently not strictly benzothiazole (Scheme 1). For the synthesis of 3-fluoropropyl-
necessary for binding affinity, as similar structures have been tosylate 1, tetraethylammonium fluoride was reacted with
reported in which the benzothiazole core has been replaced, 1,3-propanediol di-p-tosylate in acetonitrile at reflux.28 1 was
including a benzoxazole (IBOX),20 a benzofuran21 and an then used for reaction with the 6- or 20-hydroxy-substituted
imidazo[1,2-a]pyridine (IMPY)22,23 that bind with high affinity precursors to yield 2 or 3. The compounds were purified using
to fibrillar amyloid b. Mathis et al. recently reported an 18F- silica gel column chromatography, followed by reversed phase
labelled ThT derivative and are evaluating this compound as medium pressure liquid chromatography (RP-MPLC), resulting in
potential amyloid imaging agent.24 We already reported several a chemical purity of more than 99%.
18F-labelled phenylbenzothiazoles with the 18F-atom directly
The synthesis of the precursors for fluorine-18 labelling, i.e.
attached to the 2-phenyl ring.25,26 In the present study, we 2-[20-(300-p-tosyloxypropoxy)-40-aminophenyl]-1,3-benzothiazole
report the synthesis, labelling and (biological) evaluation of two (4) and 6-(300-p-tosyloxypropoxy)-2-(40-aminophenyl)-1,3-ben-
phenylbenzothiazoles with a fluorine-18 labelled fluoropropoxy zothiazole (5) is depicted in Scheme 2. 1,3-Propanediol
substituent as potential PET-tracers for fibrillar amyloid b.
di-p-tosylate was reacted with 6-hydroxy-2-(40-aminophenyl)-
1,3-benzothiazole or 2-(20-hydroxy-40-aminophenyl)-1,3-benzothia-
zole in the presence of sodium methanolate. A two-fold excess
of 1,3-propanediol di-p-tosylate was used to avoid reaction at
both tosyl groups and this was further inhibited by adding
6-hydroxy-2-(40-aminophenyl)-1,3-benzothiazole or 2-(20-hydro-
Results and discussion
Chemistry
6-Hydroxy-2-(40-aminophenyl)-1,3-benzothiazole and 2-(20- xy-4’-aminophenyl)-1,3-benzothiazole slowly to the solution of
hydroxy-40-aminophenyl)-1,3-benzothiazole were synthesized 1,3-propanediol di-p-tosylate.
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2009, 52 473–481

K. Serdons et al.
OTs
TsO
N
S
N
S
(Et)4N+F-, MeCN
12 h, reflux
NH2
NH2
HO
HO
NaOMe, MeCN, MeOH
12 h, 50 °C
NaOMe, MeCN, MeOH
12 h, 50 °C
OTs
F
1
N
N
NH2
NH2
S
S
O
O
2
3
F
F
Scheme 1. Synthesis of 6-(3’’-fluoropropoxy)-2-(4’-aminophenyl)-1,3-benzothiazole (2) and 2-[2’-(3’’-fluoropropoxy)-4’-aminophenyl]-1,3-benzothiazole (3).
N
S
N
S
NH2
NH2
HO
HO
Na, MeCN, MeOH
12 h, 50 °C
Na, MeCN, MeOH
12 h, 50 °C
OTs
TsO
N
N
NH2
NH2
S
S
O
O
4
5
OTs
OTs
Scheme 2. Synthesis of 6-(3’’-p-tosyloxypropoxy)-2-(4’-aminophenyl)-1,3-benzothiazole (4) and 2-[2’-(3’’-p-tosyloxypropoxy)-4’-aminophenyl]-1,3-benzothiazole (5).
303.0981 Da, corresponding to the theoretical molecular
ion mass (303.0962 Da) of (Figure 3). The single ion
Radiolabelling
2
Labelling of the tosylated precursors 4 and 5 was done using
the mixture 18F-/K2CO3/kryptofixs 222 in anhydrous acetonitrile.
The decay corrected radiochemical yield was about 50% for
both compounds and the specific activity after purification using
high performance liquid chromatography (HPLC) varied typically
between 33 and 51 GBq/mmol at the end of synthesis (EOS) in
both cases.
Strong evidence of the identity of the radiolabelled com-
pounds was obtained by comparing their retention times on
HPLC with the retention times of the cold reference compounds
and by using radio-LC-MS. Because of the low amounts of
fluorine-18 labelled compound in the labelling reaction mixture,
it was not possible to directly detect the radiolabelled
compound. Instead, we detected the fluorine-19 labelled
analogue, which is present in 1200 – 2000-fold excess in the
labelling mixture. Nevertheless, this indirect method provided us
confirmation of the identity of the radiolabelled compound, as
the molecular mass of the fluorinated compound was not found
in the mixture before, but only after labelling. A radio-LC-MS
chromatogram of the labelling reaction mixture of [18F]2
after purification over two Seppaks Plus silica cartridges is
shown in Figure 2. The accurate mass spectrum corresponding
mass chromatogram of mass range 303.533 – 303.573 Da
(Figure 2a) shows that the retention time of these ions is
identical to that of the main peak in the radiometric signal, also
indicating that the peak in the radiometric signal corresponds to
[
18F]2. The small peak at 7.35 min in the single ion mass
chromatogram corresponds to the hydrolysed precursor. The
tosylated precursor 4 elutes at 15.37 min. Similar results were
obtained for the identity confirmation of [18F]3 using radio-LC-
MS (data not shown). Identity confirmation was also done
by comparison of the retention times of the authentic analogues
2 and 3 and 18F-labelled [18F]2 and [18F]3, respectively, on
HPLC.
Partition coefficient
The logP value, which was determined using a modification of
the method described by Yamauchi and co-workers,29 was
2.0770.06 for [18F]2 and 2.9670.05 for [18F]3, which is within
the optimal range for free diffusion over the BBB.30
Biodistribution in control mice
to the main peak in the radiometric signal (retention time Table 1 shows the results of the biodistribution study of
12.64 min, Figure 2b) shows molecular ion mass of
18F]2 and [18F]3 in control mice. Both compounds show a
a
[
J. Label Compd. Radiopharm 2009, 52 473–481
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org

K. Serdons et al.
Figure 2. Radio-LC-MS analysis of the labeling reaction mixture of 6-(3’’-[18F]fluoropropoxy)-2-(4’-aminophenyl)-1,3-benzothiazole ([18F]2) after elution over two Seppaks
Plus silica cartridges. (a) Ion mass chromatogram of the ions in the mass range between 303.533 – 303.573 Da; (b) radiometric signal; and (c) UV signal at 365 nm.
In vitro affinity for human fibrillar amyloid b
In order to obtain information on the in vitro affinity of these
tracer agents for amyloid b fibrils, determination of their binding
to homogenates of post mortem human AD brain is a useful test.
The Ki value obtained for the non-radioactive compound 2 was
14.574 nM, which is in the range (Kip20 nM) for good binding
affinity for Ab plaques. Surprisingly, position isomer 3 did not
show any affinity for fibrillar Ab in the tested concentration
range. The obtained Ki value was more than 4000 nM for 3. This
lack of binding potential of 3 may be due to the fact that a
derivatization at the 20-position causes a torsion of the bond
between carbon-2 and carbon-10, leading to a conformation of
the phenylbenzothiazole that is not able to bind to amyloid. To
our knowledge, all uncharged derivatives of thioflavin T with
affinity for fibrillar amyloid b reported up to now are derivatized
either at the 30-position or at the benzothiazole moiety itself and
not at the 20-position.
In vitro affinity for amyloid present in mouse brain sections
Figure 3. Accurate mass spectrum of [18F]2 with kryptofixs 222 (Mr = 377.2646 Da)
as lock mass. (a) Theoretical mass spectrum and (b) experimental mass spectrum
with at 377.2646 the peak originating from kryptofixs 222.
We evaluated the binding potential of the two non-radioactive
fluorinated compounds (2 and 3) for fibrillar amyloid b on post
mortem brain sections of transgenic AD mice. The results of the
significant initial brain uptake, which is the highest for [18F]2 incubation of 6 mm microtome sections of transgenic APP
(2 min p.i.: [18F]2: 4.5% ID/g and [18F]3: 3.0% ID/g). Wash-out mouse brain with either 2 or 3 in a 1 mM concentration together
from brain is faster for [18F]3 than for [18F]2 and the ratio of with the immunohistochemical staining of neighbouring sec-
the brain activity at 2 min p.i. versus 60 min p.i. is 4.5 for tions is shown in Figure 4. The region of the mouse brain shown
[
18F]3 and 3.8 for [18F]2, but the difference is not significant is the subiculum, a structure located in the hippocampic region
(p40.05). The rate of excretion through the renal pathway is and rich in plaques in AD mice.
for both tracer agents about the same, but the excretion
From Figure 4 it can be seen that 2 (top left) stains a few
through the hepatobiliary pathway is 3 times higher for [18F]3 plaques in this region of the brain while 3 (top right) does not.
compared with [18F]2 (60 min p.i.: [18F]2: 15.5% ID and [18F]3: This is consistent with the findings of the in vitro binding affinity
46.9% ID).
test on homogenates of human AD brian. Comparison of the
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2009, 52 473–481

K. Serdons et al.
images from the immunohistochemical staining (bottom) with plaques that were stained with 2 at this concentration, no co-
the fluorescence staining shows that 2 stains less fibrillar Ab localization was possible.
than the pan Ab does, probably because of the higher affinity of
the antibodies for fibrillar Ab. Owing to the few potential
Experimental section
Instruments and general conditions
Table 1.
[
Biodistribution in control mice of [18F]2 and
18F]3 at 2 and 60 min p.i. (n = 4 at each time point)
% ID7s % ID/g7s
All reagents and solvents used in synthesis were purchased from
Acros Organics (Geel, Belgium), Fluka (Bornem, Belgium) or
Aldrich (Sigma-Aldrich, Bornem, Belgium) and were used with-
out further purification. MgSO4 was used as drying agent.
Column purification of reaction mixtures was done using silica
gel with a particle size between 0.04 and 0.063 mm (230 – 400
mesh) (MN Kieselgel 60 M, Macherey Nagel, Du¨ren, Germany) or
by medium pressure liquid chromatography (MPLC). Thin layer
chromatography (TLC) was carried out using precoated silica
TLC plates (DC-Alufolien-Kieselgel, Fluka, Buchs, Switzerland).
The structure of the synthesized products was confirmed with
1H-nuclear magnetic resonance (NMR) spectroscopy on a Gemini
200 MHz spectrometer (Varian, Palo Alto, CA, USA). Chemical
shifts are reported as d-values (parts per million) relative to
tetramethylsilane (d = 0). Coupling constants are reported in
hertz. Splitting patterns are defined by s (singlet), d (doublet), dd
(double doublet), t (triplet), q (quartet) and m (multiplet). For
melting point determination, an Electrothermal IA9000 digital
melting point apparatus was used (Electrothermal, Southend-
on-Sea, England).
2 min p.i. 60 min p.i. 2 min p.i. 60 min p.i.
[
18F]2
Urine
Kidneys
Liver
0.170.1
5.070.8
20.671.5
7.270.5
1.370.4
4.170.8
—
—
8.370.6
10.771.5
—
2.270.4
1.970.4
—
Intestines 10.370.7 11.471.7
Stomach
Cerebrum
Cerebellum 0.570.1
Blood
1.670.8
1.570.4
0.770.5
0.470.0
0.270.0
3.170.6
—
—
4.570.8
4.670.8
2.770.3
1.270.0
1.470.2
1.170.2
6.970.8
[
18F]3
Urine
Kidneys
Liver
Intestines
Stomach
Cerebrum
0.370.2
5.872.5
17.271.6
5.573.4
1.871.0
7.571.0
—
—
9.073.3
7.671.2
—
3.571.9
4.570.6
—
8.670.7 39.4710.6
1.670.3
0.970.1
2.472.5
0.270.1
0.170.1
1.170.6
—
—
Batches of [18F]fluoride were prepared by an 18O(p,n)18F
reaction by irradiation of 500 ml of 97% 18O-enriched water
(Rotem HYOX18, Rotem Industries, Beer Sheva, Israel) in a
titanium target with 10 MeV protons using an IBA Cyclone 10/5
3.070.3
3.270.2
3.470.9
0.670.3
0.770.3
0.570.3
Cerebellum 0.370.1
Blood 9.472.5
Figure 4. Fluorescence microscopic images after incubation of transgenic AD mouse brain microtome sections with 2 (top left) and 3 (top right). Neighbouring sections
were immunohistochemically stained with pan Ab (bottom images). Bar = 100 mm.
J. Label Compd. Radiopharm 2009, 52 473–481
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org

K. Serdons et al.
cyclotron (IBA, Louvain-la-Neuve, Belgium). Quantitative deter- was slowly added to a vial containing 1 (232 mg, 1 mmol) in 2 ml
mination of radioactivity in samples was done using an acetonitrile/methanol (9:1 v/v) at 501C. After 12 h the solvent was
automatic gamma counter coupled to a multi-channel analyzer evaporated under reduced pressure. The residue was purified
(Wallac 1480 Wizards 3’’, Wallac, Turku, Finland). The data were using silica gel column chromatography with CH2Cl2 as eluent,
corrected for physical decay and background radiation.
followed by RP-MPLC purification using LiChroprep RP 18 (Merck)
Combined liquid chromatography and mass spectrometry as stationary phase and acetonitrile/0.1 % ammonium acetate
(LC–MS) was performed on a system consisting of a Waters (50:50 v/v) as mobile phase at a flow rate of 17ml/min. UV
Alliance 2690 separation module (Waters, Milford, MA, USA) absorption was monitored at 254 nm. Acetonitrile was evapo-
coupled to an XTerras MS C18 3.5 mm 2.1 mm ꢀ 50 mm reverse rated under reduced pressure. The white precipitate formed was
phase high performance liquid chromatography (RP-HPLC) filtered off and dried in a vacuum oven. 94 mg (0.31 mmol,
column (Waters). The mobile phase consisted of a mixture of 31.1%) of the desired compound 2 was directly obtained as a
acetonitrile (MeCN, A) and 0.1% ammonium acetate (B) with a white powder. 1H-NMR (CDCl3, 200 MHz): d 2.21 (2H, dq,
linear gradient of 30 A:70 B v/v to 80 A:20 B v/v in 20 min at a 4J(H,F) = 26 Hz, CH2–CH2–CH2); d 4.16 (2H, t, FCH2–CH2–CH2O); d
3
flow rate of 300 ml/min. The eluate was first led through a UV- 4.68 (2H, dt, J(H,F) = 47.4 Hz, FCH2–CH2–CH2O); d 6.73 (2H, d,
spectrometer (Waters 2487 Duallabsorbance detector) and 3J = 8.4Hz, 30-H 50-H); d 7.04 (1H, d, 3J = 8.8Hz, 5-H); d 7.33 (1H, s,
subsequently over a 3-inch NaI(Tl)-crystal coupled to a single 7-H); d 7.84 (2H, d, 3J = 8.4Hz, 20-H 60-H); d 7.87 (1H, d, 3J = 9.2 Hz,
channel analyzer (The Nucleus, Oak Ridge, TN, USA) for 4-H). Accurate MS ES1 m/z [M1H]1 303.0943 (calculated for
radiometric detection before it was injected in a time-of-flight C16H16FN2OS 303.0962). Mp: 149.5–150.51C.
mass spectrometer (LCT, Micromass, Manchester, England)
equipped with a standard electrospray ionization (ESI) probe. 2-[20-(300Fluoropropoxy)-40-aminophenyl]-1,3-benzothiazole (3)
Accurate mass determination was done by co-injection of a
A similar procedure as for the synthesis of 2 was used, starting
from 2-(20-hydroxy-40-aminophenyl)-1,3-benzothiazole (242 mg,
kryptofixs 222-solution by means of a Harvard 22 syringe pump
(Harvard Instruments, Holliston, USA). Acquisition and proces-
sing of data was done with Masslynx software (Version 3.5).
RP-HPLC was done with a system consisting of a Merck-
Hitachi ternary pump connected with a Waters 2487 dual
1 mmol) and yielding 59.6% of 3. 1H-NMR (CDCl3, 200 MHz):
4
d 2.39 (2H, dq, J(H,F) = 26 Hz, CH2–CH2–CH2); d 4.27 (2H, t,
3
FCH2–CH2–CH2O); d 4.81 (2H, dt, J(H,F) = 47 Hz, FCH2–CH2–
3
CH2O); d 6.29 (1H, s, 30-H); d 6.42 (1H, d, J = 8.8 Hz, 50-H);
wavelength absorbance detector and a 3-inch NaI(Tl) crystal
3
3
connected with an EG&G Ortec ACEMateTM model 925-scint
single channel analyzer (EG&G Ortec, Oak Ridge, TN, USA).
Gas chromatography was performed with a DI 200 gas
chromatograph (Delsi Instruments, Suresnes, France) with a
Porapaks QS 80/100 column (Alltech, Deerfield, IL, USA) of
180 cm ꢀ 0.25 inch.
d 7.31 (1H, dd, J = 7.5 Hz, 6-H); d 7.45 (1H, dd, J = 7.7 Hz, 5-H);
d 7.87 (1H, d, J = 7.8 Hz, 60-H); d 8.01 (1H, d, 3J = 8.0 Hz,
3
7-H); d 8.39 (1H, d, 3J = 8.4 Hz, 4-H). Accurate MS ES1 m/z
[M1H]1 303.0943 (calculated for C16H16FN2OS 303.0962). Mp:
140.5–1411C.
6-(300-p-Tosyloxypropoxy)-2-(40-aminophenyl)-1,3-benzothiazole
(4)
Animal experiments were performed according to the Belgian
code of practice for the care and use of animals, after approval
from the university ethics committee for animals.
To a dispersion of 6-hydroxy-2-(40-aminophenyl)-1,3-benzothia-
zole (1.817 g, 7.5 mmol) in 15 ml of a mixture of acetonitrile and
methanol (9:1 v/v) was added Na (190 mg, 8.25 mmol) dissolved
in a mixture of 3 ml methanol and 12 ml acetonitrile. The
resulting mixture was added to a dispersion of 1,3-propanediol
di-p-tosylate (768 mg, 15 mmol) in 75 ml of a mixture of
acetonitrile and methanol (9:1 v/v) and the reaction mixture
was heated to 501C for 12 h. The formed suspension was filtered
off and the solvent removed from the filtrate by vacuum
evaporation. The residue was purified using silica gel column
chromatography with a mixture of CH2Cl2 and methanol (99:1
v/v) as eluent to obtain 4 (896 mg, 1.97 mmol, 26.3%) as a
yellow–brown powder. 1H-NMR (CDCl3, 200 MHz): d 2.14 (2H, m,
CH2–CH2–CH2); d 2.30 (3H, s, CH3); d 3.98 (2H, t, OCH2–CH2–CH2);
Syntheses
3-Fluoropropyltosylate (1)
A solution of dry tetraethylammonium fluoride (745mg, 5 mmol)
(dried via repeated azeotropic evaporation of a solution in
acetonitrile) in 20 ml anhydrous acetonitrile was added to a
refluxing solution of 1,3-propanediol di-p-tosylate (1.92 g,
5 mmol) in 20 ml acetonitrile. Refluxing was continued for 12 h.
The obtained suspension was filtered and the filtrate was
evaporated under reduced pressure. The residue was purified
using silica gel column chromatography with a mixture of
CH2Cl2/hexane (50:50 v/v) as the eluent to yield 315 mg
(1.36 mmol, 26.7%) of pure compound 1 in the form of an oil.
3
d 4.27 (2H, t, CH2–CH2OTs); d 6.73 (2H, d, J = 8.8 Hz, 30-H 50-H);
3
d 6.87 (1H, d, J = 8.8 Hz, 5-H); d 7.16 (1H, s, 7-H); d 7.20 (2H, d,
3
3J = 8.0 Hz, 200-H 600-H); d 7.75 (2H, d, J = 7.6 Hz, 300-H 500-H);
4
1H-NMR (CDCl3, 200 MHz): d 2.04 (2H, dq, J(H,F) = 25.8 Hz,
d 7.83 (1H, d, 3J = 9.0 Hz, 4-H); d 7.84 (2H, d, 3J = 8.8 Hz, 20-H 60-H).
Accurate MS ES1 m/z [M1H]1 455.1067 (calculated for
C23H23N2O4S2 455.1094). Mp: 129.5–130.51C.
CH2–CH2–CH2); d 2.46 (3H, s, CH3); d 4.16 (2H, t, FCH2–CH2–CH2O);
3
d 4.49 (2H, dt, J(H,F) = 46.8 Hz, FCH2–CH2–CH2O); d 7.36 (2H, d,
3
3J = 8.6 Hz, 3-H 5-H); d 7.80 (2H, d, J = 8.0Hz, 2-H 6-H).
6-(300-Fluoropropoxy)-2-(40-aminophenyl)-1,3-benzothiazole (2)
2-[20-(300-p-Tosyloxypropoxy)-40-aminophenyl]-1,3-benzothiazole
(5)
To a dispersion of 6-hydroxy-2-(40-aminophenyl)-1,3-benzothia-
zole (242 mg, 1 mmol) in 2 ml of a mixture of acetonitrile and The same procedure as for the synthesis of 4 was used starting
methanol (MeOH) (9:1 v/v) was added NaOMe (1.1 mmol) in 2 ml from 2-(20-hydroxy-40-aminophenyl)-1,3-benzothiazole (1.21 g,
of a mixture of acetonitrile and methanol (8:2 v/v). The mixture 5 mmol) and yielding 64.4% of 5. 1H-NMR (CDCl3, 200 MHz): d
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2009, 52 473–481

K. Serdons et al.
2.17 (3H, s, CH3); d 2.31 (2H, m, CH2CH2CH2); d 4.01 (2H, t,
OCH2CH2CH2); d 4.43 (2H, t, CH2CH2OTs); d 6.16 (1H, s, 30-H); d
6.40 (1H, d, 3J = 8.0 Hz, 50-H); d 6.98 (2H, d, 3J = 8.2 Hz, 300-H 500-H);
d 7.31 (1H, dd, 3J = 7.6 Hz, 6-H); d 7.45 (1H, dd, 3J = 6.9 Hz, 5-H); d
7.64 (2H, d, 3J = 8.4 Hz, 200-H 600-H); d 7.81 (1H, d, 3J = 7.8 Hz, 60-H);
d 7.99 (1H, d, 3J = 7.6 Hz, 7-H); d 8.30 (1H, d, 3J = 8.4 Hz, 4-H).
Accurate MS ES1 m/z [M1H]1 455.1064 (calculated for
C23H23N2O4S2 455.1094). Mp: 130.5–1321C.
Biodistribution in control mice
The radiometric peak on RP-HPLC containing the desired
radiolabelled compound was collected and heated in an oil
bath at 401C under a nitrogen flow for 10 min. The residue was
diluted with saline to adjust the radioactivity concentration to
about 3.7 MBq/ml. Gas chromatography was used to verify that
the concentration of ethanol was below 10%. Male NMRI mice
were sedated by intraperitoneal injection of 0.75 ml of a 1/4
dilution of Hypnorms (fentanyl citrate 63 mg/ml1fluanisone
2 mg/ml after dilution, Janssen-Cilag, Beerse, Belgium). An
aliquot of 0.1 ml of one of the fluorine-18 labelled compounds
(370 kBq) was injected via a tail vein. After decapitation at 2 or
60 min post-injection (p.i., four mice at each time point), blood
was collected in tared tubes and weighed. All organs and other
body parts were dissected and weighed and their activity was
counted using an automatic gamma counter. Results were
corrected for background activity and are expressed as
percentage of the injected dose (% ID), equal to the sum of
the activities in all organs except the tail, and as percentage of
the injected dose per gram tissue (% ID/g). For calculation of the
total activity in blood, blood mass was assumed to be 7 % of
total body mass.31
Radiolabelling
A Seppaks Light AccellTM plus QMA anion exchange column
(Waters) was rinsed with 10 ml of a 0.5 M K2CO3 solution,
followed by two 10 ml washes with H2O. After irradiation of
500 ml of [18O]water with 10 MeV protons, the contents of the
target were passed over the cartridge. The cartridge was rinsed
with 3 ml water and [18F]fluoride was eluted with a solution of
7.5 mg kryptofixs 222 and 750 mg K2CO3 in 0.75 ml of methanol/
H2O (90:10 v/v). The eluate was heated under argon flow at 981C
for 5 min to evaporate H2O and methanol. Anhydrous acetoni-
trile (1 ml) was added twice and after each addition the
acetonitrile was evaporated at 981C under argon flow for
5 min. To the residue was added a solution of 5 mg of the
precursor for labelling in 1.7 ml of anhydrous acetonitrile. The
reaction mixture was heated at 851C for 5 min in a closed vial.
After cooling to room temperature, 10 ml of diethylether was
added and the solution was passed over two Seppaks Plus silica
cartridges (Waters). After evaporation of the solvent under
argon flow, the residue was redissolved in 1 ml ethanol (EtOH).
For RP-HPLC purification, the solution was applied on an
Econosphere C18 10 mm column (10 mm ꢀ 250 mm, Alltech),
eluted with a mixture of ethanol/0.1% ammonium acetate (50:50
v/v) at a flow rate of 3 ml/min. Specific activity was determined
using a calibration curve obtained from the UV absorption
(365 nm) of the cold compounds. For the calibration curve,
different concentrations of a solution of 2 or 3 in acetonitrile
were analyzed by RP-HPLC and the area under the curve was
determined.
In vitro affinity for human fibrillar amyloid b
Preparation of brain tissue homogenates
Post mortem brain tissues were obtained from four AD patients
and four age-matched control persons at autopsy, and
neuropathological diagnosis was confirmed by current criteria
(NIA-Reagan Institute Consensus Group, 1997). The cerebellum
and the affected temporal or parietal cortex of AD patients
together with the corresponding regions from control persons
were isolated. Gray matter was carefully separated from white
matter from the cortical tissues. Homogenates were then
prepared in phosphate buffered saline (PBS, pH 7.4) at a
concentration of approximately 100 mg wet tissue/ml (motor-
driven glass homogenizer for 30 s). The homogenates were
aliquoted into 1 ml portions and stored at ꢁ701C for 3–6 months
without loss of binding signal.
Partition coefficient determination
The lipophilicity of the RP-HPLC isolated fluorine-18 labelled
compounds was determined using a modification of the
method described by Yamauchi and co-workers.29 In a test
tube, 25 ml of the RP-HPLC isolated fluorine-18 labelled
compound was added to a mixture of 3 ml each of 1-octanol
and 0.025 M phosphate buffer pH 7.4. The test tube was
vortexed at room temperature for 2 min and then centrifuged at
2700 g for 10 min. A 100 ml aliquot was taken from the octanol
phase and a 900 ml aliquot from the aqueous phase, taking care
to avoid cross-contamination between the phases, and the
aliquots were weighed. Their activity was counted using an
automatic gamma counter and the partition coefficient P was
calculated using the equation:
Binding studies
Binding assays were carried out according to a method
described previously32 using [125I]IMPY (6-iodo-2-(4’-dimethyla-
mino)phenyl-imidazo1,2 pyridine) with a specific activity of
81.4 TBq/mmol and more than 95% radiochemical purity.23
Binding assays were carried out in 12 mm ꢀ 75 mm borosilicate
glass tubes. For saturation studies, the reaction mixture
contained 50 ml of post mortem AD brain homogenate
(20–50 mg), 50 ml of an [125I]IMPY solution in PBS (0.02–0.04 nM)
and 40 ml of a solution of the test compound (2 and 3;
10ꢁ7–10ꢁ10 M diluted serially in PBS containing 0.1 % bovine
serum albumin) in a final volume of 1 ml. Non-specific binding
was defined in the presence of 600 nM non-radioactive IMPY in
the same assay tubes. For the competition binding studies,
concentrations of 10ꢁ5–10ꢁ10 M of the compounds and 0.06 nM
cpm=ml octanol
P ¼
cpm=ml buffer
[
125I]IMPY were used. The mixtures were incubated at 371C for
2 h, and the bound and the free radioactivity were separated by
vacuum filtration through Whatman GF/B filters using a Brandel
M-24R cell harvester (London, England) followed by two 3 ml
with cpm ¼ counts per minute
Experiments were performed at least in triplicate.
J. Label Compd. Radiopharm 2009, 52 473–481
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org

K. Serdons et al.
washes with PBS at room temperature. Filters containing the was done using radio-LC-MS and by comparison of retention
bound iodine-125 ligand were counted in a gamma counter times on RP-HPLC. The partition coefficient of the two
(Packard 5000, Ramsey, Minnesota, USA) with 70% counting compounds is compatible with a free passive diffusion over the
efficiency. Under the assay conditions, the specifically bound BBB. In control mice both compounds have a relatively high
fraction was less than 15% of the total radioactivity. Protein initial brain uptake, which is significantly higher for [18F]2 (2 min
determinations were performed with Lowry’s method using p.i.: [18F]2: 4.5% ID/g and [18F]3: 3.0% ID/g, po0.05), whereas
bovine serum albumin as a standard.33 The results of saturation wash-out from control brain is faster for [18F]3. In vitro binding
and inhibition experiments were subjected to non-linear affinity tests using post mortem human AD brain homogenates
regression analysis using EBDA by which Kd and Ki values were revealed that only compound 2 has affinity for fibrillar amyloid b.
calculated.
This was confirmed by incubation of APP mouse brain sections
with the cold compounds, where 3 did not stain any structure
In vitro affinity for amyloid present in mouse brain sections while 2 stained amyloid plaques present in APP mouse brain.
These data suggest that [18F]2 may be a useful tracer for in vivo
Transgenic mice overexpressing human AbPP (London mutation,
visualization of fibrillar amyloid b.
412 months old) were anaesthetized with pentobarbital and the
brain was flushed transcardially with icecold saline. The brain was
removed and post-fixed overnight with 4% paraformaldehyde in
PBS. One part of the brain was stored in 0.1% sodium azide in PBS
at 41C. The other part was processed for paraffin embedding. 6 mm
sagittal sections were cut using a Microm HM 340 E microtome
REFERENCES
[1] L. Cai, R. B. Innis, V. W. Pike, Curr Med Chem 2007, 4,
19–52.
[2] W. E. Klunk, M. L. Debnath, J. W. Pettegrew, Neurobiol Aging 1994,
15, 691–698.
(Microm, Walldorf, Germany). Prior to incubation, paraffin sections
were taken through two 5min washes with xylol, followed by 1min
sequential washes with 100, 100, 90, 70, 50% ethanol and PBS
before they were blow-dried. After rehydration, the sections were
incubated for 1 h with one of the cold compounds (2 or 3) in a
concentration range between 10 mM and 10 nM. After 1 h, the brain
sections were rinsed with tap water and coverslipped with a
Mowiol-DABCO solution (Calbiochem, Darmstadt, Germany). For
immunohistochemical staining, the same deparaffinization protocol
was used. The sections were pretreated with 96% formic acid for
7 min to enhance signal. After rinsing with PBS, the sections were
treated with 1.5% H2O2 in PBS/methanol 50:50 v/v to inhibit
endogenous peroxidase. Non-specific binding was blocked by
treatment with PBS containing 0.05% tweens 20 (PBST) and 10%
fetal calf serum (FCS, Gentaur, Brussels, Belgium). The sections were
incubated overnight with pan b-amyloid (AB-1) polyclonal anti-
bodies (Oncogene Research Products, San Diego, CA, USA) diluted
1/1000 in PBST 10% FCS at 41C. After rinsing in PBST, the sections
were incubated for 1 h with 1/500 diluted goat anti-rabbit biotin
(Vector Laboratories, Burlingame, CA, USA) in PBST 10% FCS. Then,
sections were incubated with avidin biotin complex (Vectastain Elite
ABC-peroxidase kit, Vector Laboratories) dissolved in PBST 10% FCS
for 30 min at room temperature. Peroxidase activity was developed
with 3,30-diaminobenzidine, after which the sections were counter-
stained with hematoxylin. In case no immunohistochemical staining
was done, the presence of plaques in the transgenic mice was
confirmed by incubation of neighbouring sections with X-34.
Fluorescence microscopy was performed using a Leica DMR
microscope equipped with a digital Leica DC480 camera (Leica
Microsystems, Groot Bijgaarden, Belgium) and a UV filter set with
following specifications: excitation: 340–380 nm bandpass filter,
dichromatic mirror 400 nm; emission: 425 nm longpass filter. The
images were acquired and processed with Leica IM500 image
processing software.
[3] S. D. Styren, R. L. Hamilton, G. C. Styren, W. E. Klunk, J Histochem
Cytochem 2000, 48, 1223–1232.
[4] D. M. Skovronsky, B. Zhang, M. P. Kung, H. F. Kung, J. Q. Trojanowski,
VM-Y. Lee, Proc Natl Acad Sci USA. 2000, 97, 7609–7614.
[5] M. L. Schmidt, T. Schuck, S. Sheridan, M. P. Kung, H. Kung,
Z. P. Zhuang, Am J Pathol 2001, 159, 937–943.
¨
[6] C-W. Lee, Z-P. Zhuang, M-P. Kung, K. Plossl, D. Skovronsky,
T. L. Gur, J Med Chem 2001, 44, 2270–2275.
[7] Z-P. Zhuan, M-P. Kung, C. Hou, D. Skovronsky, T. L. Gur, K. Plossl,
¨
J Med Chem 2001, 44, 1905–1914.
[8] M. Ono, A. Wilson, J. Nobrega, D. Westaway, P. Verhoeff,
Z-P. Zhuang, Nucl Med Biol 2003, 30, 565–571.
¨
[9] H. Kung, C. Lee, Z. Zhuang, M. Kung, C. Houc, K. Plossl, J Am Chem
Soc 2001, 123, 12740–12741.
[10] E. D. Agdeppa, V. Kepe, J. Liu, S. Flores-Torres, N. Satyamurthy,
A. Petric, J Neurosci 2001, 21, RC189.
[11] K. Shoghi-Jadid, G. W. Small, E. D. Agdeppa, V. Kepe, L. M. Ercoli,
P. Siddarth, Am J Geriatr Psychiatry 2002, 10, 24–35.
[12] C-W. Lee, M-P. Kung, C. Hou, H. F. Kung, Nucl Med Biol 2003, 30,
573–580.
[13] N. Okamura, T. Suemoto, H. Shimadzu, M. Suzuki, T. Shiomitsu,
H. Akatsu, J Neurosci 2004, 24, 2535–2541.
[14] T. Suemoto, N. Okamura, T. Shiomitsu, M. Suzuki, H. Shimadzu,
H. Akatsu, Neurosci Res 2004, 48, 65–74.
[15] C. C. Rowe, U. Ackerman, W. Browne, R. Mulligan, K. L. Pike,
G. O’Keefe, H. Tochon-Danguy, G. Chan, S. U. Berlangieri, G. Jones,
K. L. Dickinson-Rowe, H. P. Kung, W. Zhang, M. P. Kung,
D. Skovronsky, F. Dyrks, G. Holl, S. Krause, M. Friebe, L. Lehman,
S. Lindeman, L. M. Dinkelborg, C. L. Masters, V. L. Villemagne,
Lancet Neurol 2008, 7, 129–135.
[16] W. E. Klunk, Y. Wang, G-F. Huang, M. L. Debnath, D. P. Holt,
C. A. Mathis, Life Sci 2001, 69, 1471–1484.
[17] Y. Wang, W. E. Klunk, G-F. Huang, M. L. Debnatg, D. P. Holt,
C. A. Mathis, J Mol Neurosci 2002, 19, 11–16.
[18] C. A. Mathis, B. J. Bacskai, S. T. Kajdasz, M. E. McLellan, M. P. Frosch,
B. T. Hyman, Bioorg Med Chem Lett 2002, 12, 295–298.
[19] C. A. Mathis, Y. Wang, D. P. Holt, G-F. Huang, M. L. Debnath,
W. E. Klunk, J Med Chem 2003, 46, 2740–2754.
¨
[20] Z-P. Zhuang, M-P. Kung, C. Hou, K. Plossl, D. Skovronsky, T. L. Gur,
Nucl Med Biol 2001, 28, 887–894.
[21] M. Ono, M-P. Kung, C. Hou, H. F. Kung, Nucl Med Biol 2002, 29,
633–642.
Conclusion
[22] M-P. Kung, C. Hou, Z-P. Zhuang, B. Zhang, D. Skovronsky,
J. Q. Trojanowski, Brain Res 2002, 956, 202–210.
[23] M-P. Kung, C. Hou, Z-P. Zhuang, A. J. Cross, D. L. Maier, H. F. Kung,
Eur J Nucl Med Mol Imaging 2004, 31, 1136–1145.
[24] C. Mathis, B. Lopresti, N. Mason, J. Price, N. Flatt, W. Bi, S. Ziolko,
S. DeKosky, W. J. Klunk, J Nucl Med 2007, 48, 56P.
Two 2-(4’-aminophenyl)-1,3-benzothiazoles were derivatized
either at the 6-position or at the 2’-position with a tosyloxypro-
poxy moiety and then labelled with 18F in a 50% yield, resulting
in an average specific activity of about 42GBq/mmol EOS. Identity
confirmation of the 18F-labelled phenylbenzothiazole derivatives
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2009, 52 473–481

K. Serdons et al.
[25] K. Serdons, T. Verduyckt, D. Vanderghinste, J. Cleynhens,
P. Borghgraef, P. Vermaelen, C. Terwinghe, F. Van Leuven,
K. Van Laere, H. Kung, G. Bormans, A. Verbruggen, Bioorg Med
Chem Lett 2009, 19, 602–605.
[26] K. Serdons, C. Terwinghe, P. Vermaelen, K. Van Laere, H. Kung,
L. Mortelmans, G. Bormans, A. Verbruggen, J Med Chem 2009, 52,
1428–1437.
[27] D-F. Shi, T. D. Bradshaw, S. Wrigley, C. J. McCall, P. Lelieveld,
I. Fichtner, J Med Chem 1996, 39, 3375–3384.
[28] Y. Zea-Ponce, S. Mavel, T. Assaad, S. E. Kruse, S. M. Parsons,
P. Emond, S. Chalon, N. Giboureau, M. Kassiou, D. Guilloteau,
Bioorg Med Chem 2005, 13, 745–753.
[29] H. Yamauchi, J. Takahashi, S. Seri, H. Kawashima, H. Koike,
M. Kato-Azuma In Technetium and Rhenium in Chemistry and
Nuclear Medicine, Vol. 3 (Eds.: M. Nicolini, G. Bandoli, U. Mazzi),
Cortina International, Verona, Italy, 1989, pp. 475–502.
[30] D. D. Dishino, M. J. Welch, M. R. Kilbourn, M. E. Raichle, J Nucl Med
1983, 24, 1030–1038.
[31] A. Fritzberg, W. Whitney, C. Kuni, W. Klingensmith, Int J Nucl Med
Biol 1982, 9, 79–82.
[32] M. P. Kung, C. Hou, Z. P. Zhuang, D. Skovronsky, H. F. Kung, Brain
Res 2004, 1025, 89–105.
[33] P. J. Munson, D. Rodbard, Anal Biochem 1980, 107,
220–239.
J. Label Compd. Radiopharm 2009, 52 473–481
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org