GE-180: A novel fluorine-18 labelled PET tracer for imaging Translocator protein 18 kDa (TSPO)

Article abstract of DOI:10.1016/j.bmcl.2011.12.084

A series of tricyclic compounds have been synthesised and evaluated in vitro for affinity against Translocator protein 18 kDa (TSPO) and for preferred imaging properties. The most promising of the compounds were radiolabelled and evaluated in vivo to determine biodistribution and specificity for high expressing TSPO regions. Metabolite profiling in brain and plasma was also investigated. Evaluation in an autoradiography model of neuroinflammation was also carried out for the best compound, 12a ([¹⁸F]GE-180).

Full text of DOI:10.1016/j.bmcl.2011.12.084

Bioorganic & Medicinal Chemistry Letters 22 (2012) 1308–1313
Contents lists available at SciVerse ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
[18F]GE-180: A novel fluorine-18 labelled PET tracer for imaging Translocator
protein 18 kDa (TSPO)
Harry Wadsworth a, Paul A. Jones a, Wai-Fung Chau a, Clare Durrant a, Naghmeh Fouladi a,
Joanna Passmore a, Dennis O’Shea a, Duncan Wynn a, Veronique Morisson-Iveson a
Amanda Ewan a, Mikkel Thaning b, Dimitrios Mantzilas b, Ingvil Gausemel b,
Imtiaz Khan a, Andrew Black a, Michelle Avory a, William Trigg a,
⇑
a GE Healthcare, The Grove Centre, White Lion Road, Amersham, Bucks HP9 7LL, UK
b GE Healthcare AS, Nycoveien 1-2 Postboks 4220 Nydalen, N-0401 Oslo, Norway
a r t i c l e i n f o
a b s t r a c t
Article history:
A series of tricyclic compounds have been synthesised and evaluated in vitro for affinity against Translo-
cator protein 18 kDa (TSPO) and for preferred imaging properties. The most promising of the compounds
were radiolabelled and evaluated in vivo to determine biodistribution and specificity for high expressing
TSPO regions. Metabolite profiling in brain and plasma was also investigated. Evaluation in an autoradi-
ography model of neuroinflammation was also carried out for the best compound, 12a ([18F]GE-180).
Ó 2011 Elsevier Ltd. All rights reserved.
Received 17 November 2011
Revised 14 December 2011
Accepted 16 December 2011
Available online 27 December 2011
Keywords:
TSPO
PBR
PET
Neuroinflammation
Radiosynthesis
Translocator protein 18 kDa (TSPO) formerly known as the
peripheral benzodiazepine receptor (PBR) is an established target
for the imaging of inflammation using PET and SPECT. [11C]
PK11195 has been used clinically for more than 20 years.1,2 Whilst
Our initial studies (see Fig. 3) focused on generating a derivative
of compound 8 in which fluorine could be introduced in a manner
suitable for future radiolabelling without impacting the affinity for
TSPO. It was discovered that the introduction of a fluoroethyl group
on to the indole nitrogen both retained affinity (9, Ki = 0.67 nM
measured in a rat heart tissue assay10) for TSPO and improved
the stability of the molecules to oxidation. As part of our efforts
to reduce lipophilicity, we also discovered that the 4th ring in
the tetracyclic structure and the sulphur atom could also be re-
moved without affecting affinity (10).
Compound 10 was synthesised using the route described in
Scheme 1. This generic route was used to rapidly evaluate the
impact of substitution on the aromatic ring and varying the amide
substituents (Table 1). Both substitution on the phenyl ring and the
amide substituents were found to have an impact on the affinity
for the receptor. It was also observed that subtle variations around
the core structure often had a significant impact on the biodistribu-
tion and metabolic profile of the compounds. Using the generic
route shown in Scheme 1 we were able to rapidly and systemati-
cally probe the SAR of this class of molecule.
[
11C]PK11195 has provided valuable insight into the manifestation
of inflammation in a wide variety of diseases, it is not an ideal PET
tracer due to poor pharmacokinetics and the carbon-11 radiolabel
that limits its use to PET centres with an on-site cyclotron. The need
to discover improved PET tracers for TSPO imaging has led to the
development of many new PET tracer candidates including the car-
bon-11 labelled DAA1106,3 DPA7134 and PBR285 as well as the fluo-
rine-18 labelled FE-DAA1106,6 DPA7147 and PBR06 (Fig. 1).8
For our own discovery programme we aimed to find a high
affinity pharmacophore which was suitable for radiolabelling with
fluorine-18 using a direct labelling strategy, which would be both
high yielding and compatible with an automated radiosynthesis
platform such as FASTlab™.
In 2004, Okubo et al.9 described a tetracyclic indole based phar-
macophore (Fig. 2) with a robust synthetic route and high affinity
for TSPO. This pharmacophore has been used as the starting point
of our synthetic chemistry programme.
A direct radiolabelling approach was used to synthesise the
fluorine-18 compounds starting from the corresponding mesylate
precursor. The generic route used to prepare the precursors is
described in Scheme 2.
⇑
Corresponding author.
E-mail address: william.trigg@ge.com (W. Trigg).
0960-894X/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmcl.2011.12.084

H. Wadsworth et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1308–1313
1309
O
O
O
O
N
N
N
O
F
N
O
N
O
N
N
O
O
O
N
N
2.DAA1106
3.DPA-713
4.PBR28
F
Cl
O
F
O
O
N
O
N
N
N
O
F
1.PK11195
N
O
O
O
F
O
O
N
5.FE-DAA1106
6.DPA-714
7.PBR06
Figure 1. Example PET ligands for TSPO.
the olfactory bulbs which have a high expression compared to
other brain areas.16 We used this difference in tracer retention be-
tween the olfactory bulbs (OB) and the low TSPO expressing striata
(Str) as the basis for screening our novel compounds.
R
N
R
Y
O
X1
X2
Y = S, SO2 or CH2
X1/X2 = H, F, Cl, Me
R = alkyl or cycloalkyl
The biodistribution of the 18F-labelled compounds was evalu-
ated in naïve Sprague–Dawley rats.17 The key brain data for the
compounds in our screening programme (those which relate to
ability to image neuroinflammation in the brain) are detailed in
Table 2. The data for the original tetracyclic lead compound (9),
N
H
8
[
11C]PK11195 and [18F]PBR06 are included for comparison.
In vivo metabolic profile was also assessed in brain and plasma
Figure 2. Pharmacophore from Okubo et al. 2004.
using HPLC. For an imaging agent targeted at the CNS it is impor-
tant that the signal in the brain is attributable to the parent mole-
cule and not a radiolabelled metabolite. Table 3 shows the data
from the compounds which showed promise in the naïve rat
biodistribution.
From these studies, compound 12 was selected as the most
promising lead candidate as it demonstrated high brain uptake, a
high uptake and retention in the high TSPO expressing region
(olfactory bulbs) and good clearance from the chosen low expres-
sion region (striata). This high signal to background ratio combined
with the data from the metabolic stability study which shows that
the uptake in the brain is almost exclusively derived from the par-
ent compound and not affected by the presence of radiolabelled
metabolites. Compound 17 had a comparable in vivo performance
but was shown to have lower affinity for TSPO.
N
N
N
O
O
S
O
S
N
N
N
H
F
9
F
8a
10
Figure 3. Initial molecules prepared.
In order to investigate the effect of chirality on affinity and bio-
logical activity of the most promising compound 12 was resolved
into the corresponding single enantiomers using super critical fluid
chromatography.18 For compounds 12a and 12b the absolute ste-
reochemistry was identified using vibrational circular dichroism
(VCD) spectroscopy.19 Table 4 shows the difference in affinity be-
tween the enantiomers which is approximately 4.4-fold.
Table 5 shows the comparative data from the regional brain bio-
distribution of the racemate and the 2 enantiomers of compound
12. As would be expected there is no significant difference between
the initial brain uptake but there is a significant difference between
the olfactory bulbs signal at 30 min and the ratio between the high
TSPO expressing olfactory bulbs and the low expressing striata.
Compound 12a was also evaluated in the in vivo metabolic stability
Using the routes above the precursors and cold compounds
were synthesised as racemates and our initial evaluations focused
on using the racemic compounds to probe the SAR and select the
lead candidate.
Radiolabelling of the compounds was achieved from the
mesylates as described above. The typical procedure is shown in
Scheme 3.
End of synthesis yields for this radiolabelling were typically
25–35% (non-decay corrected).14 The radiolabelled molecules were
purified by HPLC prior to the biological evaluations.
In the naïve human CNS the expression of TSPO is very low15
and the same is generally true for the rat with the exception of

1310
H. Wadsworth et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1308–1313
O
O
O
O
i
Br
O
O
O
O
*
R1
iii
N
F
ii
R1
R1
F
NH2
N
iv
H
R3
N
O
R4
*
R1
N
F
Scheme 1. General synthetic route to tricyclic compounds. Reagents and conditions: (i) Br2, Et2O; (ii) 2-fluoroethyl 4-methylbenzenesulfonate, K2CO3, DMF; (iii) (a) heat
neat; (b) ZnCl2, IPA, reflux; (iv) (a) NaOH, ethanol; (b) oxalyl chloride, DCM; (c) amine, DCM.11
Table 1
Structures of tricyclic compounds and measured Ki data for TSPO in rat heart and LogD 7.4
R2
N
O
R3
1
2
R1
3
N
4
F
Compound
R1
R2
R3
Ki nM
LogD 7.4
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
H
2-OH
1-OMe
2-OMe
3-OMe
1,3-Dimethoxy
2-OEt
1-F
2-F
3-F
2-OMe, 3-F
2-F, 3-OMe
H
H
H
H
H
H
H
H
H
H
H
H
H
H
2-F
2-F
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
Et
5.4
135
2.4
1.9
5.5
4.34
3.1
7.6
3.8
1.05
1.5
2.78a
1.70
2.95
2.71
2.41
1.95
2.56
NTb
2.37
NT
2.38
2.08
1.53
2.18
2.53
2.35
1.99
2.20
2.28
NT
2.42
2.98
2.13
1.70
2.84
2.47
3.10
2.72
3.1
R2/R3 = morpholinyl
R2/R3 = pyrrolidinyl
R2/R3 = piperidinyl
607
503
18.3
4.4
2.2
4.0
1.7
0.38
1.3
0.7
0.7
502
0.73
28.6
1.45
8.23
R2/R3 = homopiperidinyl
Me
Me
iPr
nPr
iPr
iPr
nPr
Bn
nPr
n-Hex
Me
Bn
Et
Me
nPr
n-Hex
Me
Et
H
Me
Me
Bn
Bn
iPr
a
eLogP derived using HPLC method.12
NT = not tested.
b

H. Wadsworth et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1308–1313
1311
Table 2
H
N
Regional brain data from the biodistribution data of 18F-compounds naïve Sprague–
OH
OBn
Dawley rats (n = 3, pi refers to time post injection)
R1
CO2Et
Br
CO2Et
Compound number
Whole brain
2 min pi
(% id/g)
OB 30 min
pi (% id/g)
OB:Str
30 min
pi ratio
Str
2:30 min
pi ratio
R1
i
N
9
0.17
0.39
0.36
0.31
0.39
0.25
0.51
0.41
0.48
0.21
0.25
0.12
0.25
0.33
0.25
0.35
0.25
0.46
0.34
0.40
0.10
0.34
0.45
0.23
0.49
0.21
1.42
3.25
3.00
3.88
4.33
2.50
3.19
3.72
1.92
3.50
2.78
1.33
2.50
4.13
2.50
3.18
2.78
2.09
3.40
2.20
1.11
3.40
3.21
2.88
3.50
2.10
2.16
4.08
3.67
4.25
4.00
3.00
2.56
4.27
1.80
4.17
3.00
3.88
4.70
5.38
3.50
4.54
5.11
1.77
4.70
2.39
0.44
4.30
3.07
7.38
4.64
3.20
10
12
13
14
16
17
18
19
20
21
23
24
25
26
27
28
29
30
31
32
34
12a
12b
0.53
0.52
0.39
0.36
0.33
0.50
0.56
0.52
0.29
0.33
0.34
0.51
0.45
0.39
0.53
0.48
0.45
0.55
0.49
0.06
0.48
0.53
0.53
0.74
0.28–0.43
OBn
ii
R2
O
N
CO2H
R3
R1
iii
R1
N
N
OBn
OBn
iv
[
[
18F]PBR06
11C]PK11195
R2
N
R2
N
O
R3
O
R3
Table 3
v
Brain data from in vivo metabolism study of 18F-compounds naïve Sprague–Dawley
rats (n = 1). Data collected at 10, 30 and 60 min post injection from brain and plasma
to evaluate metabolic stability and fate of radiolabelled metabolites. Only data from
60 min brain samples shown for clarity
R1
R1
N
N
OMs
Compound
number
Brain 60 min pi (%
parent)
Compound
number
Brain 60 min pi
(% parent)
OH
Scheme 2. General synthetic route to radiolabelling precursors. Reagents and
conditions: (i) (a) heat neat; (b) ZnCl2, IPA, reflux; (ii) NaOH, H2O, EtOH; (iii) method
C: EDC, NHS, DMF; method D: (a) oxalyl chloride, DCM; (b) amine, DCM; method E:
(a) oxalyl chloride, DCM, isolated; (b) amine, DCM; (iv) H2, Pd/C, MeOH; (v) MsCl,
pyridine, DCM.13
9
10
12
13
17
18
24
95% @ 30 min
25
27
28
30
79
74
43
81
76
62
90
76
93
84
83
34
PBR06
91
100
[
11C]PK11195
18F (in target
water)
R3
N
R3
N
[K(K2.2.2)]HCO3
Table 4
O
O
R2
N
R2
N
Resolved enantiomer affinities
Compound
Assignment
Affinity (nM)
R1
R1
12a
12b
S
R
0.87
3.87
Acetonitrile
100 °C, 10 mins
18F
OMs
Scheme 3. Radiosynthesis of 18F-tricyclic indoles.
Table 5
Regional brain data from the biodistribution data from the racemate and individual
enantiomers of compound 12
assay and was shown to have a 94% parent in the brain at 60 min
(n = 3). This data enables the selection as compound 12a as our lead
candidate for further evaluation.
Axotomy of the facial nerve was performed in rats using the
method of Banati et al.20 Coronal sections of the brainstem at the
level of the facial nucleus were obtained from animals culled
7 days post-lesion. Application of compound 12a on to the
brain slices enabled visualisation by autoradiography of the
Compound
number
Whole brain
2 min pi (% id/g)
OB 30 min
pi (% id/g)
OB:Str
30 min pi
ratio
Str
2:30 min pi
ratio
12
12a
12b
0.52
0.53
0.53
0.36
0.45
0.23
3.00
3.21
2.88
3.67
3.07
7.38

1312
H. Wadsworth et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1308–1313
Naive FN
Ipsilateral FN
Ipsilateral FN + PK11195
Ipsilateral FN + 12a
Contralateral FN
Contralateral FN + PK11195
Contralateral FN + 12a
0.7
0.6
0.5
0.4
0.3
0.2
0.1
0.0
86%
86%
64% 64%
Figure 4. A–C: In vitro autoradiography images of the facial nucleus sections from rats 7 days after axotomy of facial nerve. Representative autoradiograph images of total
binding of [18F]12a (A) with co-incubation with excess PK11195 (1 mM) (B) and unlabelled12a (1 mM) (C) from brain stem sections. Ipsilateral facial nucleus is highlighted in
red and the contralateral nucleus in yellow. D: Photomicrograph of OX-42 immunostained activated microglia from ipsilateral facial nucleus 7 days after axotomy. Graph
shows binding to ipilsateral and contralateral facial nucleus in the absence or presence of unlabelled TSPO ligand. Specific binding (total minus non-specific) was 86% in the
ipsilateral facial nucleus following co-incubation with either ligand.
5. Imaizumi, M.; Kim, H.-J.; Zoghbi, S. S.; Briard, E.; Hong, J.; Musachio, J. L.;
inflammation in the facial nucleus corresponding to the axoto-
mised nerve (ipsilateral) with low uptake in the contralateral facial
nucleus. Blocking studies with PK11195 and non-radiolabelled
12a showed high specific binding (Fig. 4).
We have developed a novel tricyclic indole series of TSPO ligands.
The best compounds from the series have high affinity, good brain
uptake and high specific binding in an established neuroinflamma-
tion model. The most preferred compound 12a ([18F]GE-180) will
be further developed for clinical studies in neuroinflammatory con-
ditions such as MS.
Ruetzler, C.; Chuang, D.-M.; Pike, V. W.; Innis, R. B.; Fujita, M. Neurosci. Lett.
2007, 411, 200.
6. Zhang, M.-R.; Maeda, J.; Furutsuka, K.; Yoshida, Y.; Ogawa, M.; Suhara, T.;
Suzuki, K. Bioorg. Med. Chem. Lett. 2003, 13, 201.
7. James, M. L.; Fulton, R. R.; Vercoullie, J.; Henderson, D. J.; Garreau, L.;
Chalon, S.; Dollé, F.; Selleri, S.; Guilloteau, D.; Kassiou, M. J. Nucl. Med. 2008,
49, 814.
8. Imaizumi, M.; Briard, E.; Zoghbi, S. S.; Gourley, J. P.; Hong, J.; Musachio, J. L.;
Gladding, R.; Pike, V. W.; Innis, R. B.; Fujita, M. Synapse 2007, 61, 595.
9. Okubo, T.; Yoshikawa, R.; Chaki, S.; Okuyama, S.; Nakazato, A. Bioorg. Med.
Chem. 2004, 12, 3569.
10. Assay carried out using homogenated rat heart tissue with [3H]PK11195 as the
radioligand. Ki measurements are averaged from an 2 experiments across 6
concentrations (see Supplementary data for details).
11. Example analytical data (compound 12): 1H NMR (300 MHz, CDCl3) dH 1.13
(3H, t, J = 6 Hz, N(CH2CH3)2), 1.25 (3H, t, J = 6 Hz, N(CH2CH3)2), 1.78–2.25 (4H,
m, 2- and 3-CH2), 2.58–2.84 (2H, m, 4-CH), 3.39–3.64 (4H, m, N(CH2CH3)2), 3.79
(3H, s, OCH3), 4.02–4.36 (3H, m, NCH2CH2F and 1-CH2), 4.56–4.71 (2H, m,
NCH2CH2F), 6.43 (1H, d, J = 6 Hz, ArH), 6.82 (1H, d, J = 6 Hz, ArH) and 7.00 (1H, t,
J = 6 Hz, CHCHCHC(OCH3)). See Supplementary data for additional analytical
data.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at doi:10.1016/j.bmcl.2011.12.084.
References and notes
12. Donovan, S. F.; Pescatore, M. C. J. Chromatogr., A 2002, 952, 47.
13. Example analytical data (radiolabelling precursor of compound 12): 1H NMR
(300 MHz, CDCl3) dH1.12 (3H, t, J = 9 Hz, N(CH2CH3)a), 1.34 (3H, t, J = 9 Hz,
N(CH2CH3)b), 1.80–2.21 (4H, m, 2- and 3-CH2), 2.48 (3H, s, SO2CH3), 2.58–2.89
(2H, m, 1-CH2), 3.27–3.59 (4H, m, N(CH2CH3)2), 3.78 (3H, s, OCH3), 4.31 (3H, m,
4-CH and NCH2CH2O), 4.39–454 (2H, m, NCH2CH2O), 6.44 (1H, d, J = 6 Hz,
6-CH), 6.85 (1H, d, J = 6 Hz, 8-CH) and 7.02 (1H, t, J = 6 Hz, 7-CH). See
Supplementary data for additional analytical data.
1. Chauveau, F.; Boutin, H.; Van Camp, N.; Dollé, F.; Tavitian, B. Eur. J. Nucl. Med.
Mol. Imaging 2008, 35, 2304.
2. Katsifis, A.; Mattner, F.; Zhang, Z.; Dikic, B.; Papazian, V. J. Labelled Compd.
Radiopharm. 2000, 43, 385.
3. Zhang, M.-R.; Kida, T.; Noguchi, J.; Furutsuka, K.; Maeda, J.; Suhara, T.; Suzuki,
K. Nucl. Med. Biol. 2003, 30, 513.
4. James, M. L.; Fulton, R. R.; Henderson, D. J.; Eberl, S.; Meikle, S. R.; Thomson, S.;
Allan, R. D.; Dollé, F.; Fulham, M. J.; Kassiou, M. Bioorg. Med. Chem. 2005, 13,
6188.
14. Example radiolabelling procedure (Compound 12): [18F]Fluoride was supplied
from GE Healthcare on a GE PETracecylcotron.KryptofixÒ222(K2.2.2) (4 mg,
10.6 l lmol) and
mol), potassium bicarbonate (0.1 mol dmÀ3, 0.1 ml, 10 mg, 10

H. Wadsworth et al. / Bioorg. Med. Chem. Lett. 22 (2012) 1308–1313
1313
acetonitrile (0.5 ml) was added to [18F]FÀ/H2O (ca. 400 MBq, 0.1–0.3 ml) in a
cyclic olefin copolymer (COC) reaction vessel. The mixture was dried by heating
at 100 °C under a stream of nitrogen for 20–25 min. After drying and without
cooling, 2-(4-(diethylcarbamoyl)-5-methoxy-3,4-dihydro-1H-carbazol-9(2H)-
17. Male Wistar rats were intravenously injected test compound (approximately
1–3 MBq at 10 MBq/mL) under anaesthesia (4% isoflurane). At 2, 10, 30 or
60 min post-injection animals were sacrificed by cervical dislocation under
anaesthesia. Blood, bone (whole femur, tibia and fibula from both hind legs),
muscle (from both hind legs), skin sample, fat sample, faeces, whole organs/
tissues (kidneys, bladder (including urine and voided urine), lung, liver, spleen,
stomach (including contents), small and large intestine (including contents),
heart, adrenals), injection site (whole tail) and remaining carcass were
assessed for radioactivity by a gamma counter. The brain was dissected into
striatum, cerebellum, hippocampus, pre-frontal cortex, remaining cortex, pons
and medulla, thalamus, olfactory bulbs, hypothalamus and remaining brain
and activity determined by a gamma counter. Whole body and regional brain
biodistribution was generated. All animal studies were performed in
accordance with the Animals (Scientific Procedures) Act 1986 act.
yl)ethyl methanesulfonate, (0.5–1 mg, 1.5–3 lmol) in acetonitrile (1 ml) was
added to the COC reaction vessel and heated at 100 °C for 10 min. The reaction
mixture was cooled and diluted with water and then purified using semi-
preparative HPLC and finally formulated in Dulbecco’s phosphate buffered
saline. Radiochemical yield 25.4 6% (n = 14) non-decay corrected calculated
from starting activity, (synthesis time 90–120 min), radiochemical purity
P99%. Identity was confirmed by subsequent analysis of a sample spiked
with the corresponding unlabeled compound and correlation of the radiotrace
and UV trace. Specific activity from a radiosynthesis starting from <1 GBq were
typically >20 GBq/
>200 GBq/ mol.
lmol and when starting from ꢀ30 GBq were typically
l
18. Taylor, L. T. J. Supercrit. Fluids 2009, 47, 566.
15. Scarf, A. M.; Kassiou, M. J. J. Nucl. Med. 2011, 52, 677.
16. Casellas, P.; Galiegue, S.; Basile, A. S. Neurochem. Int. 2002, 40, 475.
19. Freedman, T. B.; Cao, X.; Dukor, R.; Nafie, L. A. Chirality 2003, 15, 743.
20. Banati, R. B.; Myers, R.; Kreutzberg, G. W. J. Neurocytol. 1997, 26, 77.