Synthesis, in vitro and in vivo evaluation of 18F-fluoronorimatinib as radiotracer for Imatinib-sensitive gastrointestinal stromal tumors

Article abstract of DOI:10.1016/j.nucmedbio.2017.11.004

Introduction Gastrointestinal stromal tumors (GIST) have a wide range of mutations, but can mostly be treated with Imatinib, until eventually resistance towards this tyrosine kinase inhibitor is acquired. Early and non-invasive determination of the sensitivity of the tumor and its metastases towards Imatinib by positron emission tomography (PET) would be beneficial for therapy planning and monitoring. Methods We developed a synthesis strategy towards the precursor molecule, performed the ¹⁸F-synthesis and in the following evaluated the radioligand in vitro regarding its lipophilicity, stability and biological activity (KIT binding properties) as well as its in vivo properties in GIST tumor-bearing mice. Results [¹⁸F]fluoronorimatinib could be obtained in an overall radiochemical yield of 22.2 ± 3.3% within 90 min. The radioligand showed high GIST cell uptake and was able to distinguish between Imatinib-sensitive and resistant tumor cell lines (GIST-T1, GIST882, GIST430) in vitro. Further biological evaluations of the ligand towards 9 different GIST-relevant KIT mutations showed comparable binding affinities compared to the structural lead Norimatinib (65 nM vs. 53 nM for wt-KIT). The in vivo evaluation of the newly developed radioligand showed tumor-to-background-ratios comparable to previously described, similar radiotracers. Conclusions Thus, [¹⁸F]fluoronorimatinib is able to distinguish between Imatinib-resistant and sensitive KIT mutations. Although no improvement of in vivo tumor-to-background ratios could be achieved compared to formerly described radioligands, the hepatic uptake could be considerably reduced, being advantageous for the imaging of GIST. Advances in knowledge and implications for patient care: We were able to show that it is possible to significantly reduce the unfavorably high hepatic uptake of small-molecule radioligands applicable for GIST PET imaging. This work can thus be the basis for further work intending to develop a PET-radioligand for Imatinib-dependent GIST imaging.

Full text of DOI:10.1016/j.nucmedbio.2017.11.004

Nuclear Medicine and Biology 57 (2018) 1–11
Contents lists available at ScienceDirect
Nuclear Medicine and Biology
journal homepage: www.elsevier.com/locate/nucmedbio
Synthesis, in vitro and in vivo evaluation of 18F-fluoronorimatinib as
radiotracer for Imatinib-sensitive gastrointestinal stromal tumors
Martin Prause a, Sabrina Niedermoser a, Carmen Wängler b, Clemens Decristoforo c, Uwe Seibold a,b
,
Stephanie Riester a, Takahiro Taguchi d,e, Ralf Schirrmacher f, Gert Fricker g, Björn Wängler a,
⁎
a
Molecular Imaging and Radiochemistry, Department of Clinical Radiology and Nuclear Medicine, Medical Faculty Mannheim of Heidelberg University, Mannheim, Germany
Biomedical Chemistry, Department of Clinical Radiology and Nuclear Medicine, Medical Faculty Mannheim of Heidelberg University, Mannheim, Germany
Department of Nuclear Medicine, Medical University Innsbruck, Innsbruck, Austria
b
c
d
Laboratory of Human Health and Medical Science, Graduate School of Kuroshio Science, and Department of Molecular Biology and Cellular Biology, Kochi Medical School, Kochi University,
Nankoku, Japan
e
Department of Molecular and Cellular Biology, Kochi Medical School, Nankoku, Japan
Division of Oncological Imaging, Department of Oncology, University of Alberta, Edmonton, Canada
Department of Pharmaceutical Technology, Institute of Pharmacy and Molecular Biotechnology, University of Heidelberg, Heidelberg, Germany
f
g
a r t i c l e i n f o
a b s t r a c t
Article history:
Introduction: Gastrointestinal stromal tumors (GIST) have a wide range of mutations, but can mostly be
treated with Imatinib, until eventually resistance towards this tyrosine kinase inhibitor is acquired. Early and
non-invasive determination of the sensitivity of the tumor and its metastases towards Imatinib by positron
emission tomography (PET) would be beneficial for therapy planning and monitoring.
Received 8 September 2017
Received in revised form 3 November 2017
Accepted 15 November 2017
Available online xxxx
Methods: We developed a synthesis strategy towards the precursor molecule, performed the 18F-synthesis and
in the following evaluated the radioligand in vitro regarding its lipophilicity, stability and biological activity
(KIT binding properties) as well as its in vivo properties in GIST tumor-bearing mice.
Keywords:
GIST
Results: [18F]fluoronorimatinib could be obtained in an overall radiochemical yield of 22.2 3.3% within 90 min.
The radioligand showed high GIST cell uptake and was able to distinguish between Imatinib-sensitive and
resistant tumor cell lines (GIST-T1, GIST882, GIST430) in vitro. Further biological evaluations of the ligand
towards 9 different GIST-relevant KIT mutations showed comparable binding affinities compared to the
structural lead Norimatinib (65 nM vs. 53 nM for wt-KIT). The in vivo evaluation of the newly developed
radioligand showed tumor-to-background-ratios comparable to previously described, similar radiotracers.
Conclusions: Thus, [18F]fluoronorimatinib is able to distinguish between Imatinib-resistant and sensitive
KIT mutations. Although no improvement of in vivo tumor-to-background ratios could be achieved compared
to formerly described radioligands, the hepatic uptake could be considerably reduced, being advantageous for
the imaging of GIST.
18F-fluoronorimatinib
KIT
PET
Small molecule radioligand
TKI
Advances in knowledge and implications for patient care: We were able to show that it is possible to significantly
reduce the unfavorably high hepatic uptake of small-molecule radioligands applicable for GIST PET imaging.
This work can thus be the basis for further work intending to develop a PET-radioligand for Imatinib-
dependent GIST imaging.
© 2017 Elsevier Inc. All rights reserved.
1. Introduction
therapy is important for the treatment of GIST, since this tumor type is
known to be refractory against chemotherapy and radiation therapy
Imatinib was approved by the FDA as a tyrosine kinase inhibitor for
treatment of chronic myeloid leukemia (CML) in 2001, as well as of gas-
trointestinal stromal tumors (GIST) in 2002, increasing the five year sur-
vival rate in both tumor entities significantly [1–7]. This targeted
[8]. Furthermore, surgery is challenging due to a high recurrence rate
in case of intraperitoneal rupture or bleeding during the resection [9].
Although 85% of GIST bear Imatinib-sensitive mutations of KIT [10],
secondary resistances are often acquired within 2 years [1,11–13]. As
a consequence, second or third line treatments using unspecific tyrosine
kinase inhibitors (TKI) have to be applied which in general result in
increased side-effects [14,15]. Another problem is the in most cases
late detection of resistances which are often not discovered until rapid
disease progression or metastatic spread [9].
⁎
Corresponding author at: Department of Clinical Radiology and Nuclear Medicine,
Medical Faculty Mannheim of Heidelberg University, Theodor-Kutzer-Ufer 1-3, 68167
Mannheim, Germany.
E-mail address: Bjoern.Waengler@medma.uni-heidelberg.de (B. Wängler).
https://doi.org/10.1016/j.nucmedbio.2017.11.004
0969-8051/© 2017 Elsevier Inc. All rights reserved.

2
M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
Therefore, the non-invasive and early determination of the
we intended to decrease the lipophilicity of the radiotracer to potentially
reduce the metabolic turnover by CYP3A4 [21,23]. This should result in a
lower accumulation of the tracer in liver and small intestines, which
would be advantageous for PET imaging of the gastrointestinal tract.
Additionally, the metabolic stability should also benefit from a decreased
lipophilicity and thus reduced interaction with CYP3A4 and therefore
lead to a lower amount of metabolites (especially when removing the
methyl group of the piperazine ring of Imatinib, which is most prone to
metabolism). Therefore, we chose Norimatinib as our structural
lead which is known to show a similar biological activity against KIT as
Imatinib itself [17].
sensitivity of the tumor and the metastases towards Imatinib would
be an important improvement regarding monitoring of the effective-
ness of a TKI-related Imatinib-based therapy. Achieving this objective
would require the possibility to functionally image the Imatinib-
sensitivity of the malignancy to avoid a problematic multiple biopsy
sampling. So far, no functional imaging technique is available giving ap-
propriate results [16–18] which prompted us to develop an 18F-labelled
radiotracer for positron emission tomography (PET) being able to dis-
tinguish between Imatinib-sensitive and resistant GIST.
To obtain a radiotracer being able to efficiently and successfully
delineate Imatinib-sensitive malignancies, an appropriate molecular
design is of crucial importance. Considering the structure of Imatinib,
the substitution of the methyl group at the piperazine ring with a [18F]
fluoroethyl label (SKI696) was found to be advantageous regarding ra-
diochemistry while preserving the binding affinity of the respective
substance (Fig. 1) [16,17,19,20].
However, when applied in vivo, a high accumulation of the radioac-
tivity was found in the liver and the small intestines, limiting the func-
tional imaging information being accessible using this radiotracer. This
unfavorable biodistribution most probably originated from binding of
the ligand to cytochrome P450 (CYP) enzymes being present in the
liver as well as the intestines [21] and is driven by the high lipophilicity
of SKI696 as these enzymes show a higher affinity towards lipophilic
substrates than to hydrophilic ones due to their lipophilic drug binding
sites [21,22]. This impedes imaging GIST using the available 18F-labeled
radioligands within the gastrointestinal tract and GIST liver metastases
due to the high background signal in these organs. Furthermore, a
high lipophilicity causes reabsorption in the kidney tubules and there-
fore highly lipophilic 18F-labeled substances are not excreted via the
urine, but remain in the blood for further circulation until the hydrophi-
licity is increased by metabolism via CYP enzymes [21], resulting in a
high background activity further impeding imaging.
2. Materials and methods
2.1. General
All used commercially available chemicals and solvents were at least
of analytical grade and used without further purification (ABCR,
Activate Scientific, Alfa Aesar, Carl Roth, Chempur, Fluorochem, Sigma
Aldrich). For analytical and semipreparative HPLC chromatography, a
Dionex UltiMate 3000 system equipped with a Chromolith Performance
(RP-18e, 100–4.6 mm, Merck) and a Chromolith SemiPrep (RP-18e,
100–10 mm, Merck) column was used, operated with a flow rate of
4 mL/min. As mobile phase, water and acetonitrile (MeCN) (both
solvents contained 0.1% TFA) were used and different gradients were
applied. Column chromatography was carried out using silica gel 60
(Merck) as stationary phase. For analytical radio-HPLC chromatography,
a
Dionex UltiMate 3000 system equipped with a Chromolith
Performance (RP-18e, 100–4.6 mm, Merck) and a GabiStar radioactivity
detector (Raytest) was used. For semipreparative radio-HPLC purifica-
tion, a Knaur BlueShadow system equipped with a Chromolith Perfor-
mance (RP-18e, 100–10 mm, Merck) and a GabiStar radioactivity
detector (Raytest) was used. As solvents, water and acetonitrile
(MeCN) (both solvents contained 0.1% TFA) were used and different
gradients were applied. ESI-MS measurements were performed on a
Finnigan LTQ FT Ultra FT-ICR (Thermo) mass spectrometer. NMR
spectra were recorded on a Varian Mercury Plus 300 MHz (Bruker)
and a Varian 500 MHz (Bruker) NMR system. NMR signals were refer-
enced to the respective solvent signal: acetone-d6: 2.050 (1H), 29.84
As a possible alternative, the more hydrophilic metabolite of Imatin-
ib, Norimatinib (Fig. 1), was shown to be still biologically active [23], im-
plying that an 18F-labeled Norimatinib analog could be of interest for
GIST imaging.
Therefore, we developed the new radioligand, [18F]fluoronorimatinib
([18F]9, Fig. 1). Compared to SKI696, the position of the radiolabel was
changed in [18F]fluoronorimatinib from the piperazine ring to the pyri-
dine ring to decrease the probability of metabolism and thus cleavage
of the radiolabel. To prevent the cleavage of the radiolabel without affect-
ing the binding affinity of the substance, the following considerations
were made and resulted in the chosen molecular design: Substitutions
on the phenyl rings were described to be unfavorable due to changes
in selectivity of the molecule, resulting in an inhibition of not only Bcr-
Abl, KIT and PDGFRA, but also other kinases [24,25], whereas changes
at the pyridine moiety have not yet been described. Considering the in-
troduction of 18F into Imatinib, a nitropyridine structure enables an effi-
cient nucleophilic introduction of the radiolabel [26–28] while the
resulting [18F]fluoropyridine should exhibit a higher metabolic stability
than the aliphatic [18F]fluoroethyl label of SKI696 [29–31]. Furthermore,
(
13C); DMSO-d6: 2.500 (1H), 39.52 (13C). γ-Counter measurements
were performed on a 2480 Wizard system (Perkin Elmer).
2.2. Chemical syntheses
2.2.1. Synthesis of 2-chloro-4-(6-nitropyridin-3-yl)pyrimidine (1)
2-Nitro-5-pyridinylboronic acid pinacol ester (4.00 mmol, 1.00 g,
1.0 eq.), 2,4-dichloropyrimidine (4.00 mmol, 596 mg, 1.0 eq.), sodium
carbonate (10.0 mmol, 1.06 g, 2.5 eq.) and tetrakis(triphenylphosphine)
palladium(0) (0.08 mmol, 92 mg, 0.02 eq.) were suspended in a degassed
solvent mixture (THF/water = 9:1, 60 mL) and stirred under argon
atmosphere for 9 h at 80 °C. After neutralization with hydrochloric acid
(1 M, 10 mL), the solvent was removed in vacuo. The residue was
Fig. 1. Structures of Imatinib, Norimatinib, SKI696 and fluoronorimatinib (9).

M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
3
redissolved in dichloromethane and filtered. The filtrate was concentrat-
ed in vacuo, added dropwise to cyclohexane (30-fold volume) and the
precipitate was filtered. After drying under reduced pressure, the brown
solid was purified by column chromatography (DCM/MeOH = 99:1) to
obtain the product as yellowish solid (623 mg, 2.63 mmol, 66%). 1H
NMR (300 MHz, (CD3)2CO) δ 9.40 (dd, J = 2.3, 0.5 Hz, 1H), 9.00 (dd,
J = 8.5, 2.3 Hz, 1H), 8.96 (d, J = 5.3 Hz, 1H), 8.50 (dd, J = 8.5, 0.6 Hz,
1H), 8.32 (d, J = 5.2 Hz, 1H). 13C NMR (75 MHz, (CD3)2CO) δ 163.59,
162.63, 159.45, 148.86, 140.30, 137.18, 134.34, 119.27, 118.13. ESI-MS:
m/z [M + H]+ (calc.) 236.81 (237.02).
directly used for the following reaction step without further purification
(350 mg, 0.64 mmol, 51%). ESI-MS: m/z [M + H]+ (calc.) 545.30
(545.19).
2.2.5. Synthesis of 4-((4-(tert-butoxycarbonyl)piperazin-1-yl)methyl)
benzoic acid (5)
A solution of 4-(chloromethyl)-benzoic acid (5.86 mmol, 1.00 g,
1.0 eq.), 1-Boc piperazine (5.86 mmol, 1.09 g, 1.0 eq.) and
trimethylamine (29.3 mmol, 4.06 mL, 5.0 eq.) in acetonitrile (30 mL)
was stirred for 22 h at room temperature, afterwards the solvent was
removed in vacuo. By use of repeated recrystallization (n = 3) from
methanol, the product was obtained as an off-white solid (1.22 g,
3.81 mmol, 65%). 1H NMR (500 MHz, (CD3)2SO) δ 12.64 (s, 1H), 7.90
(dt, J = 8.4, 1.8 Hz, 2H), 7.42 (d, J = 8.4 Hz, 2H), 3.54 (s, 2H), 2.32 (d,
J = 5.0 Hz, 2H), 2.30 (d, J = 5.0 Hz, 2H), 1.38 (s, 9H). 13C NMR (126
MHz, (CD3)2SO) δ 167.19, 153.76, 143.17, 129.62, 129.26, 128.79,
78.73, 61.46, 52.38, 45.52, 28.04. ESI-MS: m/z [M + H]+ (calc.) 321.07
(321.18).
2.2.2. Synthesis of (9H-fluoren-9-yl)methyl(4-methyl-3-nitrophenyl)carbamate
(2)
A solution of 4-methyl-3-nitroaniline (4.00 mmol, 609 mg, 1.0 eq.)
in dioxane (20 mL) was added to a solution of sodium carbonate
(10.0 mmol, 1.06 g, 2.5 eq.) in water (12 mL) and cooled to 5 °C. A solu-
tion of 9-fluorenylmethyl chloroformate (4.00 mmol, 1.04 g, 1.0 eq.) in
dioxane (8 mL) was added dropwise to the reaction mixture under stir-
ring. After 45 min at 5 °C, the reaction was allowed to warm to room
temperature and stirred for another 2 h. After the reaction was com-
plete, the product could be precipitated by adding water (20 mL). The
remaining organic solvent was removed in vacuo, followed by an ex-
traction of the residue with ethyl acetate. The combined organic phases
were washed with brine, dried over magnesium sulfate and the solvent
was removed in vacuo to obtain the product as light brown solid (1.462-
g, 3.91 mmol, 98%). 1H NMR (300 MHz, (CD3)2SO) δ 10.06 (s, 1H), 8.18
(s, 1H), 7.91 (dt, J = 7.6, 0.9 Hz, 2H), 7.77–7.71 (m, 2H), 7.61 (d, J =
8.4 Hz, 1H), 7.49–7.31 (m, 5H), 4.54 (d, J = 6.5 Hz, 2H), 4.32 (t, J =
6.5 Hz, 1H), 2.44 (s, 3H). 13C NMR (75 MHz, (CD3)2SO) δ 153.34,
148.65, 143.63, 140.78, 138.05, 133.11, 127.68, 127.10, 126.31, 125.03,
122.83, 120.18, 113.30, 65.78, 46.54, 19.03. ESI-MS: m/z [M + H]+
(calc.) 375.05 (375.13).
2.2.6. Synthesis of 6-methyl-N1-(4-(6-nitropyridin-3-yl)pyrimidin-2-
yl)benzene-1,3-diamine (6)
A solution of crude 4 (1.66 mmol, 904 mg, 1.0 eq.) and piperidine
(33.2 mmol, 3.30 mL, 20.0 eq.) in DMF (20 mL) was stirred at room tem-
perature for 30 min. Afterwards, the solvent was removed in vacuo. Pu-
rification by column chromatography (DCM/MeOH = 98:2) yielded the
product - which still contained some impurities - as red to brown oil
which was used for the following reaction without further purification
(455 mg, 1.41 mmol, 85%). ESI-MS: m/z [M + H]+ (calc.) 323.20
(323.13).
2.2.7. Synthesis of tert-butyl-4-(4-((4-methyl-3-((4-(6-nitropyridin-3-
yl)pyrimidin-2-yl)amino)phenyl)carbamoyl)benzyl)piperazine-1-carboxylate
(7)
2.2.3. Synthesis of (9H-fluoren-9-yl)methyl
(3-amino-4-methylphenyl)carbamate (3)
A solution of 5 (0.76 mmol, 244 mg, 1.0 eq.), HBTU (0.76 mmol,
289 mg, 1.0 eq.) and DIPEA (0.76 mmol, 133 μL, 1.0 eq.) in DMF
(3.5 mL) was stirred for 5 min at room temperature, followed by the ad-
dition of a solution of 6 (0.76 mmol, 245 mg, 1.0 eq.) in DMF (3.75 mL).
After complete reaction, the solvent was removed in vacuo and the
crude product was purified by column chromatography (DCM/
MeOH = 95:5). After multiple recrystallizations (n = 3) from methanol,
the product was obtained as an orange-yellow solid (161 mg,
0.26 mmol, 34%). 1H NMR (300 MHz, (CD3)2SO) δ 10.17 (s, 1H, H-11),
9.33 (dd, J = 2.2, 0.5 Hz, 1H, H-3), 9.13 (s, 1H, H-6), 8.88 (dd, J = 8.5,
2.3 Hz, 1H, H-2), 8.61 (d, J = 5.1 Hz, 1H, H-5), 8.45–8.38 (m, 1H, H-1),
8.10 (d, J = 2.0 Hz, 1H, H-7), 7.91 (dd, J = 8.1, 1.8 Hz, 2H, H-12), 7.56
(d, J = 5.1 Hz, 1H, H-4), 7.47 (dd, J = 8.2, 2.2 Hz, 1H, H-8), 7.46–7.43
(m, 2H, H-13), 7.21 (d, J = 8.6 Hz, 1H, H-9), 3.56 (s, 2H, H-14), 3.32
(s, 4H, H-16), 2.33 (s, 4H, H-15), 2.23 (s, 3H, H-10), 1.39 (s, 9H, H-17).
13C NMR (75 MHz, (CD3)2SO) δ 165.25 (C-9), 161.21 (C-17), 159.42 (C-
6), 157.18 (C-1), 153.78 (C-25), 147.49 (C-5), 141.66 (C-21), 138.91 (C-
2), 137.73 (C-10), 137.53 (C-4), 137.20 (C-12), 133.84 (C-18), 131.75
(C-15), 130.08 (C-14), 128.69 (C-20), 127.61 (C-19), 124.40 (C-8),
118.39 (C-2), 117.22 (C-11), 116.89 (C-13), 108.44 (C-7), 78.74 (C-
26), 61.47 (C-22), 52.37 (C-23), 43.16 (C-24), 28.04 (C-27), 17.63 (C-
16). ESI-MS: m/z [M + H]+ (calc.) 625.29 (625.29). Note: Assignments
of NMR signals can be found in the supplementary information.
Tin(II)chloride (53.1 mmol, 10.1 g, 14 eq.) was added under argon
atmosphere to a solution of 2 (3.79 mmol, 1.42 g, 1.0 eq.) in ethyl ace-
tate/methanol (50 mL, 9:1) and stirred at room temperature for 27 h.
Then, the reaction mixture was neutralized using concentrated sodium
carbonate solution. The precipitating tin compounds were removed by
filtration and washed repeatedly with ethyl acetate. The filtrate was ex-
tracted with ethyl acetate, the combined organic phases were washed
with brine, dried over magnesium sulfate and the solvent was removed
in vacuo to obtain the product as a light yellow solid (1.23 g, 3.57 mmol,
94%). 1H NMR (300 MHz, (CD3)2SO) δ 9.34 (s, 1H), 7.91 (d, J = 7.3 Hz,
2H), 7.75 (d, J = 7.4 Hz, 2H), 7.46–7.39 (m, 2H), 7.34 (td, J = 7.4,
1.2 Hz, 2H), 6.81 (s, 1H), 6.77 (d, J = 8.1 Hz, 1H), 6.53 (d, J = 7.5 Hz,
1H), 4.78 (s, 2H), 4.40 (d, J = 6.9 Hz, 2H), 4.28 (t, J = 6.8 Hz, 1H), 1.97
(s, 3H). 13C NMR (75 MHz, (CD3)2SO) δ 153.29, 146.62, 143.84,
140.74, 137.39, 129.76, 127.65, 127.09, 125.19, 120.14, 115.52, 109.53,
106.36, 65.37, 46.64, 16.83. ESI-MS: m/z [M + H]+ (calc.) 345.16
(345.16).
2.2.4. Synthesis of (9H-fluoren-9-yl)methyl (4-methyl-3-((4-(6-nitropyridin-
3-yl)pyrimidin-2-yl)amino)phenyl)carbamate (4)
A suspension of 1 (1.27 mmol, 300 mg, 1.0 eq.), 3 (1.52 mmol,
524 mg, 1.2 eq.) and acetic acid (1.27 mmol, 72.5 μL, 1.0 eq.) in dioxane
(3 mL) was heated to 100 °C in sealed Eppendorf tubes (1.5 mL, 12
pieces) for 32 h. The reaction mixture was transferred into a glass
flask containing dichloromethane (48 mL) and the solvent was
removed in vacuo. The dark green residue was suspended in aqueous
sodium hydrogen carbonate solution (25.5 mM, 50 mL, 1 eq.), filtered
and washed with water. The residue was dissolved in dichloromethane
(100 mL), dried over sodium sulfate and the solvent was removed in
vacuo. After purification by column chromatography (DCM/MeOH =
99:1), the obtained product which still contained some impurities was
2.2.8. Synthesis of tert-butyl 4-(4-((3-((4-(6-fluoropyridin-3-yl)pyrimidin-
2-yl)amino)-4-methylphenyl)carbamoyl)benzyl)piperazine-1-carboxylate
(8)
A solution of 13 (0.95 mmol, 200 mg, 1.0 eq.), 12 (1.14 mmol,
486 mg, 1.2 eq.) and acetic acid (0.76 mmol, 43.7 μL, 0.8 eq.) in dioxane
(1.92 mL) was shaken in sealed Eppendorf tubes at 80 °C for 4 days.
After cooling to room temperature, the reaction mixtures were added
to dichloromethane and the formed solid was filtered. The solid was
washed with acetone and ethyl acetate and the combined organic

4
M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
filtrates were evaporated in vacuo. After purification by column chro-
matography (DCM/MeOH/HOAc = 92:7:1), the product was obtained
as a yellow solid (30 mg, 0.05 mmol, 5%). ESI-MS: m/z [M + H]+
(calc.) 598.21 (598.29).
127.51, 124.74, 115.38, 78.75, 61.42, 52.36, 43.58, 28.04, 19.22. ESI-
MS: m/z [M + H]+ (calc.) 455.17 (455.23).
2.2.12. Synthesis of tert-butyl 4-(4-((4-methyl-3-nitrophenyl)carbamoyl)
benzyl)piperazine-1-carboxylate (12)
2.2.9. Synthesis of N-(3-((4-(6-fluoropyridin-3-yl)pyrimidin-2-yl)amino)-
4-methylphenyl)-4-(piperazin-1-ylmethyl)benzamide (9)
To a suspension of 11 (6.00 mmol, 2.73 g, 1.0 eq.) in a ethyl acetate/
methanol (55 mL, 9:1) was added pyridine (42.0 mmol, 3.39 mL,
7.0 eq.), followed by tin(II)chloride (42.0 mmol, 7.96 g, 7.0 eq.). After
stirring for 3.5 h at 30 °C, the reaction mixture was diluted with ethyl ac-
etate (40 mL) and filtered. The solid was washed repeatedly with ethyl
acetate and the filtrates were combined. After removing the solvent in
vacuo, the residue was suspended in sodium hydroxide solution
(0.1 M, 200 mL) and lyophilized to remove remaining pyridine. The ob-
tained solid was redissolved in acetone and filtered to remove remain-
ing salts. The obtained solution was evaporated to dryness in vacuo,
giving the product in quantitative yield (2.55 g, 6.00 mmol, 100%). 1H
NMR (300 MHz, (CD3)2SO) δ 9.85 (s, 1H), 7.87 (d, J = 8.3 Hz, 2H),
7.42 (d, J = 8.2 Hz, 2H), 7.12 (d, J = 2.0 Hz, 1H), 6.85 (d, J = 8.2 Hz,
1H), 6.80 (dd, J = 8.0, 2.0 Hz, 1H), 4.83 (s, 2H), 3.55 (s, 2H), 3.32 (m,
4H), 2.35–2.30 (m, 4H), 2.02 (s, 3H), 1.39 (s, 9H). 13C NMR (75 MHz,
(CD3)2SO) δ 164.87, 153.78, 146.45, 141.35, 137.58, 134.11, 129.60,
128.60, 127.52, 116.69, 108.75, 106.37, 78.74, 61.48, 52.36, 43.67,
28.04, 16.98. ESI-MS: m/z [M + H]+ (calc.) 425.18 (425.25).
Trifluoroacetic acid (2 mL) was added to a solution of 8 (0.05 mmol,
30 mg, 1.0 eq.) and triisopropyl silane (50 μL) in dichloromethane
(8 mL) and stirred at room temperature for 20 min. After evaporation
of the solvent in vacuo, the residue was dissolved in water/acetonitrile
(1:1) and purified by semipreparative HPLC to obtain the product as a
yellow solid (0.036 mmol, 17.9 mg, 72%) in a purity of 99%. 1H NMR
(500 MHz, (CD3)2SO) δ 10.23 (s, 1H, H-11), 9.00 (s, br, 2H, H-6), 8.98
(d, J = 2.4 Hz, 1H, H-3), 8.68 (td, J = 8.3, 2.5 Hz, 1H, H-2), 8.52 (d,
J = 5.1 Hz, 1H, H-5), 8.11 (d, J = 1.6 Hz, 1H, H-7), 8.00 (d, J = 8.2 Hz,
2H, H-12), 7.57 (d, J = 8.1 Hz, 2H, H-13), 7.47 (dd, J = 8.2, 2.0 Hz, 1H,
H-8), 7.43 (d, J = 5.2 Hz, 1H, H-4), 7.32 (dd, J = 8.6, 2.6 Hz, 1H, H-1),
7.21 (d, J = 8.4 Hz, 1H, H-9), 4.05 (s, 2H, H-14), 3.25 (s, 4H, H-16),
2.96 (s, 4H, H-15), 2.22 (s, 3H, H-10). 13C NMR (126 MHz, (CD3)2SO) δ
164.84 (C-17), 164.04 (d, J = 201.4 Hz, C-1), 161.04 (C-9), 160.43 (C-
6), 159.52 (C-8), 146.83 (d, J = 15.9 Hz, C-5), 140.69 (d, J = 8.8 Hz, C-
3), 137.71 (C-10), 137.04 (C-12), 136.63 (C-21), 134.96 (C-18), 130.97
(d, J = 4.3 Hz, C-4), 130.05 (C-20), 129.95 (C-14), 127.86 (C-19),
127.66 (C-15), 117.19 (C-11), 116.79 (C-13), 109.73 (d, J = 37.8 Hz,
C-2), 107.37 (C-7), 59.70 (C-22), 48.43 (C-23), 41.73 (C-24), 17.62 (C-
16). ESI-MS: m/z [M + H]+ (calc.) 498.24 (498.24). Note: Assignments
of the NMR signals can be found in the supplementary information.
2.2.13. Synthesis of 2-chloro-4-(6-fluoropyridin-3-yl)pyrimidine (13)
A solution of 5-bromo-2-fluoropyridine (5.60 mmol, 573 μL, 1.4 eq.)
in anhydrous diethyl ether (16 mL) was cooled to −80 °C in an argon
atmosphere and n-butyllithium (1.6 M, 6.00 mmol, 3.75 mL, 1.5 eq.)
was added dropwise. After 1 h at −80 °C, a pre-cooled solution of zinc
bromide (6.00 mmol, 1.35 g, 1.5 eq.) in anhydrous diethyl ether
(22 mL) was added to the yellow reaction mixture and stirred for 1 h.
The mixture was allowed to warm slowly to room temperature and
the solvent was removed in vacuo. The pale pink residue was
redissolved in anhydrous THF (15 mL) under argon atmosphere and a
solution of 2,4-dichloropyrimidine (4.00 mmol, 596 mg, 1.0 eq.) and
tetrakis(triphenylphosphine)palladium(0) (0.20 mmol, 231 mg,
0.05 eq.) in anhydrous THF (15 mL) were added. After stirring for 2 h
at room temperature followed by 1.5 h at 50 °C, the solvent was re-
moved in vacuo. The residue was taken up in ethyl acetate and the or-
ganic phase was washed with brine, dried over magnesium sulfate
and the solvent was removed in vacuo. The product was obtained
after purification by column chromatography (EtOAc/cyclohexane =
2:5) as an off-white solid (556 mg, 3.71 mmol, 66%). 1H NMR
(500 MHz, (CD3)2SO) δ 9.05 (dd, J = 1.8, 0.8 Hz, 1H), 8.90 (d, J =
5.3 Hz, 1H), 8.73 (ddd, J = 8.6, 7.8, 2.6 Hz, 1H), 8.25 (d, J = 5.3 Hz,
1H), 7.42 (ddd, J = 8.7, 2.8, 0.6 Hz, 1H). 13C NMR (126 MHz,
(CD3)2SO) δ 164.72 (d, J = 241.3 Hz), 163.18, 161.54, 160.47, 147.50
(t, J = 16.5 Hz), 141.26 (d, J = 9.3 Hz), 129.20, 116.44, 110.28 (d, J =
37.8 Hz). ESI-MS: m/z [M + H]+ (calc.) 209.81 (210.02).
2.2.10. Synthesis of 4-(chloromethyl)-N-(4-methyl-3-nitrophenyl)benzamide
(10)
DMF (0.90 mmol, 69 μL, 0.1 eq.) was added to a solution of 4-
(chloromethyl)benzoic acid (9.00 mmol, 1.54 g, 1.0 eq.) in thionyl chlo-
ride (18 mL) under nitrogen atmosphere. After stirring for 4 h at 80 °C,
the reaction mixture was cooled to 50 °C and remaining thionyl chloride
was removed in vacuo. The residue was dissolved in anhydrous THF
(13 mL) and cooled to 0 °C. 4-Methyl-3-nitroaniline (9.00 mmol,
1.37 g, 1.0 eq.) and pyridine (18.0 mmol, 1.45 mL, 2.0 eq.) were slowly
added and the reaction mixture was stirred at room temperature for
2 h. After neutralization with hydrochloric acid (1 M), the reaction mix-
ture was extracted with ethyl acetate. The combined organic phases
were washed with brine, dried over magnesium sulfate and the solvent
was evaporated in vacuo. After purification by column chromatography
(EtOAc/cyclohexane = 1:3), the product was obtained as a yellow solid
(2.62 g, 8.60 mmol, 96%). 1H NMR (300 MHz, (CD3)2SO) δ 10.59 (s, 1H),
8.54 (d, J = 2.2 Hz, 1H), 7.99 (dt, J = 8.3 Hz, 2.3 Hz, 1H), 7.98 (d, J =
8.3 Hz, 2H), 7.60 (d, J = 8.3 Hz, 2H), 7.48 (d, J = 8.5 Hz, 1H), 4.84 (s,
2H), 2.49 (s, 3H). 13C NMR (75 MHz, (CD3)2SO) δ 165.80, 148.93,
141.90, 138.49, 134.52, 133.45, 129.29, 128.54, 128.12, 125.23, 115.90,
45.79, 39.97, 19.70. ESI-MS: m/z [M + H]+ (calc.) 304.95 (305.07).
2.2.11. Synthesis of tert-butyl 4-(4-((4-methyl-3-nitrophenyl)carbamoyl)
benzyl)piperazine-1-carboxylate (11)
2.3. Radiochemistry
A solution of 10 (0.66 mmol, 200 mg, 1.0 eq.), tert-butyl piperazine-
1-carboxylate (0.98 mmol, 368 mg, 3.0 eq.) and N,N-
diisopropylethylamine (1.64 mmol, 572 μL, 5.0 eq.) in ethyl acetate
(5.4 mL) was stirred at 50 °C for 34 h. After extraction of the obtained
mixture with ethyl acetate, the combined organic phases were washed
with brine and dried over magnesium sulfate. The crude product was
purified by column chromatography (DCM/MeOH = 97:3) to obtain
the product as a pale yellow solid (273 mg, 0.60 mmol, 92%). %). 1H
NMR (300 MHz, (CD3)2SO) δ 10.54 (s, 1H), 8.55 (d, J = 2.2 Hz, 1H),
8.00 (dd, J = 8.4, 2.2 Hz, 1H), 7.94 (d, J = 8.2 Hz, 2H), 7.48 (dd, J =
8.5, 2.6 Hz, 3H), 3.57 (s, 2H), 3.32 (d, J = 15.4 Hz, 4H), 3.31 (s, 3H),
2.38–2.30 (m, 4H), 1.39 (s, 9H). 13C NMR (75 MHz, (CD3)2SO) δ
165.64, 153.77, 148.49, 142.24, 138.15, 133.00, 132.96, 128.79, 127.70,
2.3.1. Preparation of nucleophilic [18F]fluoride
Aqueous [18F]fluoride was purchased from Zyklotron AG (Karlsruhe,
Germany) or EuroPET (Freiburg, Germany) and produced by irradiation
of enriched [18O]H2O via the 18O(p,n)18F nuclear reaction. The obtained
[
18F]fluoride (400–4500 MBq) was loaded on a pre-conditioned Sep-
Pak Light QMA cartridge (Waters) for purification and eluted with a
freshly prepared solution of Kryptofix 2.2.2 (12.5 mg), potassium
carbonate (12.5 μL, 1 M), water (187.5 μL) and acetonitrile (800 μL)
into a 5 mL Wheaton vial. The solvent was evaporated under reduced
pressure (600 mbar) using a stream of helium and heating to 90 °C.
This azeotropic drying step was repeated twice using acetonitrile
(800 μL), followed by heating under reduced pressure (b50 mbar)
until complete dryness was reached.

M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
5
2.3.2. Radiosynthesis of [18F]fluoronorimatinib ([18F]9)
concurrent addition of radioligand (1 nM [18F]9 in phosphate buffered
saline, pH 7.4, equaling about 20,000 cpm/well) and competitor (0–
5 μM - varying concentrations of Imatinib) the incubation was per-
formed at 37 °C for 1 h. The supernatant was removed by suction and
the cells were washed three times with cold PBS. The membrane filters
were punched and collected in tubes for activity measurement in the
gamma-counter. Standards of the radioligand were used to determine
the total activity added. The raw data were corrected for residual unspe-
cific binding to the membrane filter and the counts from the blocked
fraction (5 μM Imatinib) were subtracted from all data points for nor-
malization. The assay was performed thrice, each experiment in qua-
druplicate. IC50 values were calculated by GraphPad Prism 6.03 using
non-linear regression and the fitting method least-squares (ordinary).
A solution of precursor 7 (5 mg) in anhydrous DMSO (500 μL) was
added to the freshly purified and dried [18F]fluoride/Kryptofix mixture
and heated to 150 °C for 8 min. After cooling the reaction mixture to
below 90 °C, a solution of ascorbic acid (10 mg/mL) in hydrochloric
acid (2 M, 200 μL) was added and the deprotection of the intermediate
took place at 85 °C for 10 min. The reaction mixture was further cooled
to room temperature, diluted with water (1.5 mL) and purified via
semipreparative HPLC. The obtained product was loaded onto a C-18
cartridge, washed with water and eluted with a solution of ascorbic
acid (0.125 mg/mL) in ethanol/water (1:1). After discarding the first
250 μL, the purified product was obtained in a volume of 500 μL.
2.4. In vitro evaluations
2.4.6. Radiometric kinase assay
2.4.1. General
The radiometric protein kinase assay was performed by ProQinase
(Freiburg, Germany) according to the following procedure. In a 96
well FlashPlate (Perkin Elmer) were added assay buffer (70 mM
All GIST cell lines (T1, 882 and 430) were obtained by the MITIGATE
consortium and cultured using the corresponding protocols. K562 cells
were purchased from ECACC and cultured using the recommended
media and additives. All cell lines were incubated at 37 °C in humidified
air containing 5% CO2. For determination of IC50 values, a Millipore vac-
uum manifold system and 96-well DV filter membrane plates (Merck
Millipore, Darmstadt) were used.
HEPES-NaOH pH 7.5,
3 mM MgCl2, 3 mM MnCl2, 3 μM Na-
orthovanadate, 1.2 mM DTT, 50 μg/mL PEG20000, ATP (amount corre-
sponding to the apparent ATP-Km of the respective kinase)), [γ-33P]-
ATP (6 × 105 cpm/well), protein kinase, substrate and test compound.
After incubation at 30 °C for 1 h, the enzyme reaction was stopped by
addition of H3PO4 (2%). The supernatant was removed and the wells
were washed two times with saline. The incorporation of [33P]phos-
phate was determined by a microplate scintillation counter (Microbeta,
Wallac). Activity measurements were corrected for unspecific binding
to the plate in the absence of a protein kinase with the presence of sub-
strate (“low control”) and the full activity in the absence of any inhibitor
(“high control”) according to the following formula:
2.4.2. Determination of the lipophilicity of [18F]9
A phosphate-buffered solution of [18F]9 (2 μL, pH 7.0–7.4) was
added to a mixture of n-1-octanol (800 μL) and sodium phosphate buff-
er (0.05 M, pH 7.4, 800 μL) and shaken for 5 min. After centrifugation, a
sample (100 μL) of each phase was measured for radioactivity in the
gamma counter.
Res. Activity (%) = 100 × [(cpm of compound − low control) / (high
control − low control)]
2.4.3. Determination of the serum stability of [18F]9
A phosphate-buffered solution of [18F]9 (100 μL, pH 7.0–7.4) was
added to blood serum (500 μL) and incubated at 37 °C. After 10, 20,
30, 60, 90 and 120 min, an aliquot (75 μL) of this solution was added
to acetonitrile (75 μL), cooled on ice for 5 min and centrifuged. The
phases were separately measured for radioactivity and the supernatant
was analyzed via HPLC.
The IC50 values were calculated using Quattro Workflow V3.1.1
(Quattro Research GmbH, Germany). The fitting model for the IC50 de-
terminations was “Sigmoidal response” and the fitting method used
was a least-squares fit. The determination of the IC50 values has been
performed once. To ensure reliable results, the Z′-factor was used as a
parameter for assay quality indicating excellent assay quality (Z′-factor
N0.52) [32,33], making the repetition of the tests dispensable. Addition-
ally, a reference compound was used for validation.
2.4.4. Cell permeability assay
1x106 cells were seeded into 6-well-plates 24 to 48 h prior to the ex-
periment. The cells were washed using cold DMEM + 1% FCS three
times before the experiment. The incubation of the radioligand (1 nM
per well with and without 1 μM blocking substance) was performed
in DMEM + 1% FCS at 37 °C for one to 3 h. Subsequently, the superna-
tants were removed and collected in gamma-counter tubes (including
the supernatants of 2 washing steps using cold DMEM+1% FCS) to ob-
tain the unbound fraction. The cells were incubated twice with cold gly-
cine buffer (pH 2.8) for 5 min and the supernatants were collected in
tubes to measure the membrane bound fraction. Finally, the cells were
lysed using 1 M NaOH (2 times) and collected in tubes to determine
the internalized fraction. All tubes were measured in the gamma-
counter using standards of the radioligand to determine the total activ-
ity. To obtain the specific internalized or membrane bound fractions, the
counts obtained by the measurements of the blocked radioligand frac-
tions (=unspecific binding) were subtracted from the counts obtained
by the measurements of the non-blocked radioligand fractions. The
added counts of unbound, membrane bound and internalized fractions
were compared with the counts of the standards to ensure consistent
assay conditions. The assay was performed thrice, each experiment in
triplicate.
2.5. In vivo evaluations
2.5.1. General
All animal studies were performed in compliance with the National
Guidelines for Animal Protection, Germany, and with the approval of
the animal care committee. Female Balb/c SCID mice (Balb/cJHan®-
Hsd-Prkdcscid, Envigo RMS GmbH) were inoculated subcutaneously
with 5 × 106 tumor cells (in a mixture of 50 μL PBS and 50 μL Matrigel)
into the left hind leg and the tumors were grown to a diameter of 5–
8 mm within 7–11 weeks (GIST882) or 5–7 weeks (GIST430).
2.5.2. Small animal PET/CT imaging studies
Small-animal PET imaging experiments were conducted on the PET/
SPECT/CT system Albira II (Bruker Biospin MRI GmbH). The mice were
anesthetized with isoflurane (2–3%) and the radiotracer (5–12 MBq in
50–100 μL isotonic phosphate buffer, pH 7.0–7.4) was applied via the
lateral tail vein. The animals were imaged over 120 min (90 min PET
and 30 min CT). The analysis of the PET data was performed using
PMOD 3.611 (PMOD Technologies).
2.4.5. Competitive binding assay
2.5.3. Biodistribution studies
After incubation of a 96 well Multiscreen filter plate with PBS + 1%
BSA for 1 h, the solution was removed and 1 × 105 cells suspended in
assay buffer (OptiMEM + 0.5% BSA) were pipetted in each well. After
The mice were instantly sacrificed after PET/CT imaging and the or-
gans (tumor, pancreas, adrenal glands, blood, heart, lungs, muscle, bone,
brain, large intestines, small intestines, spleen, stomach, tail, kidneys

6
M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
and liver) were excised, weighed and the activity in each organ was de-
termined by gamma counter measurement.
1-yl)uranium hexafluorophosphate) and conjugated to amine 6 under
mild conditions. Purification of the product was performed by column
chromatography and subsequent recrystallization provided the final
radiolabeling precursor in a purity of 98%.
3. Results and discussion
The non-radioactive reference compound 9 was obtained in a 6-step
synthesis (Scheme 2), starting with the formation of an amide bond be-
tween 4-(chloromethyl)benzoic acid and 4-methyl-3-nitroaniline, giv-
ing the benzamide intermediate 10.
3.1. Synthesis of the labeling precursor and the cold reference compound
Although several reaction paths to obtain Imatinib and its deriva-
tives are already known [16,17,25,34], we had to develop a new synthe-
sis strategy for the introduction of the nitro group we intended to use as
leaving group for the 18F-introduction into the pyridine ring. Nitro
groups in general exhibit strong electron withdrawing characteristics
and are known as good leaving groups. Therefore, the introduction of
a nitro group during a multi-step synthesis influences the possible reac-
tion paths, limiting the available reaction conditions. To use a protected
amino group during the synthesis which could be oxidized at the end of
the preparation [35,36] was found to be equally unfavorable as intro-
ducing the nitro group at the end of the synthesis which was expected
to result in a high number of by-products like nitro arenes, entailing
an ineffective purification and limited synthesis yields. Furthermore,
within the common synthesis strategy towards Imatinib, the reduction
of a nitro group to an amine at the phenyl ring for further conversion is
described [16,34]. Since this step would also affect the nitropyridine
when introduced early during the multi-step synthesis, the amine func-
tion at the phenyl ring would consequently have to be created using an-
other strategy.
First, the benzoic acid derivative was transformed into the respective
activated benzoyl chloride by addition of thionyl chloride and catalytic
amounts of N,N-dimethylformamide. The resulting acyl chloride was
used directly after evaporation of excess thionyl chloride and reacted
with 4-methyl-3-nitroaniline under basic conditions to give 10 in an ex-
cellent yield of 96%. The following alkylation with 1-Boc-piperazine
under basic conditions afforded the nitro intermediate 11 in high yields
of 92%. The reduction of the nitro group, however, proved to be chal-
lenging. Palladium-catalyzed hydrogenation was not possible due to
the piperazine ring being in a benzylic position (the mentioned
reaction parameters would have resulted in the cleavage of the
piperazine ring). Another method for the reduction of nitro groups is
the use of tin(II) chloride under strongly acidic conditions. However,
the acid lability of the Boc group forced us to modify this approach in
order to preserve the Boc protecting group. The addition of pyridine to
tin(II) chloride in equimolar amounts proved to be the best choice
since the Boc protecting group was stable under these conditions
and no precipitation of tin(II) hydroxide was observed. After multiple
filtration steps, being necessary to remove tin compounds as well as
pyridine salts, the amine 12 could be obtained in high purity and
quantitative yields.
The fluoropyridinylpyrimidine 13 was synthesized via a Negishi
coupling of 5-bromo-2-fluoropyridine and 2,4-dichloropyrimidine.
First, 5-bromo-2-fluoropyridine underwent halogen-lithium-exchange
with n-butyllithium, followed by addition of zinc bromide to generate
the zinc intermediate. After changing the solvent to THF [39], addition
of 2,4-dichloropyrimidine and tetrakis(triphenylphosphine)palladium(0) gave
13 in a good yields of 66%. The condensation of the two intermediates
12 and 13 took place under acidic conditions and gave the Boc-
protected reference compound 8 in a poor yield of 5% due to the
formation of several by-products. Since however only minute amounts
of this reference substance were needed, no optimization of the
reaction parameters was performed. The final step of the synthesis
was the removal of the Boc protecting group by addition of
trifluoroacetic acid and triisopropyl silane at room temperature which
afforded the cold reference compound 9 in a good yield of 72% and an
excellent purity of 99%.
Therefore, we developed the a new reaction route to obtain our
radiolabeling precursor (Scheme 1).
This synthesis strategy included seven reaction steps: intermediate 1
was obtained via a Suzuki coupling [37] of a nitropyridinylboronic acid
ester with 2,4-dichloropyrimidine, giving the product in 66% yield. To
obtain the Fmoc-protected phenylenediamine derivative 3, 4-methyl-
3-nitroaniline was initially protected with an Fmoc group (almost quan-
titative reaction, 98% yield) and the following reduction with tin(II)
chloride gave 3 in 94% yield. The coupling of the intermediates 1 and 3
took place under acidic conditions [38] and yielded the key intermedi-
ate 4 in good yields of 51%. Since column chromatography could not re-
move the formed by-products completely, this pre-purified product
was used for the following deprotection reaction. The removal of the
Fmoc group with piperidine gave 6 in good yields of 85%. 6 unfortunate-
ly could not be isolated in high purity, being however unproblematic as
its further chemical conversion was not impaired by the impurities. To
form the final product 7, the carboxylic acid 5, resulting from an alkyl-
ation reaction of 1-Boc-piperazine with (4-chloromethyl)benzoic acid,
was activated by HBTU ((N,N,N′,N′-Tetramethyl-O-(1H-benzotriazol-
Scheme 1. Schematic depiction of the synthesis pathway towards the nitro radiolabeling precursor 7.

M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
7
Scheme 2. Schematic depiction of the synthesis pathway towards the cold reference compound 9.
3.2. Radiochemistry
The deprotection of [18F]8 to the final product [18F]9 proved to be
rather challenging since the removal of the tert-butyloxycarbonyl
group required relatively harsh reaction conditions. Interestingly, the
addition of trifluoroacetic acid to the reaction mixture did not succeed
in removing the protecting group even at elevated temperatures of
80 °C within 30 min. Although the desired product could be obtained by
use of aqueous hydrochloric acid (2 M), significant amounts of a radioac-
tive by-product, as well as free [18F]fluoride were formed under these
conditions (80 °C, 30 min). We intended to minimize these side-
reactions using the following approaches: First, we attempted to suppress
the supposed substitution reaction of [18F]fluoride by using anhydrous
deprotection conditions using hydrochloric acid (2–4 M) in dioxane
(80 °C, 30 min) which however proved to be less efficient with regard
to deprotection while also releasing [18F]fluoride from [18F]8. Second, in
order to suppress the formation of the radioactive by-product which
was assumed to originate from oxidation by DMSO [42], ascorbic acid
was added to the radiolabeling reaction mixture under otherwise identi-
cal conditions (80 °C, 30 min, addition of 2 M HCl). The ascorbic acid was
found to be able prevent the formation of the radioactive by-product al-
most completely. Furthermore, it was able to accelerate the deprotection
reaction with hydrochloric acid. As a result, this shorter reaction time led
to a decreased release of [18F]fluoride, giving the raw product [18F]9 after
optimization in yields of 58 5% (n = 15) within 10 min at 80 °C.
The radiosynthesis of
[
18F]9 (Scheme 3) started with the
radiofluorination of the nitro precursor 7 at 150 °C in DMSO using a
[
18F]fluoride-kryptand mixture prepared according to standard proce-
dures [40]. Concordant to former studies, we observed a high depen-
dency of [18F]fluoride incorporation on the amount of precursor as
well as the reaction solvent volume used [41].
Although 76–81% of [18F]fluoride were incorporated within 2.5 min
under all tested conditions (n = 1.6–8.0 μmol, V = 300–500 μL), solvent
volumes of 300 μL resulted in a faster substitution of already incorporat-
ed [18F]fluoride (and thus product disintegration) compared to larger
solvent volumes. This effect was observed to be the stronger, the
lower the amount of precursor used. Furthermore, we noticed an almost
complete cleavage and liberation of [18F]fluoride from the radiolabeled
intermediate when using 1.6 μmol of precursor in a reaction volume of
300 μL. Kügler et al. also described this effect examining chloropyridines
and attributed it to the strong polarization of the C\\F bond by activat-
ing groups which facilitated a substitution even by weak nucleophiles
like halides [41]. The optimal conditions to obtain a constantly stable
fluoride incorporation as well as high radiolabeling yields of 83 2%
(n = 15) were found to be 8 μmol precursor in 500 μL DMSO within
8 min reaction time (Fig. S5).
Scheme 3. Two-step radiosynthesis towards [18F]9 starting from precursor 7: incorporation of [18F]fluoride, followed by deprotection and purification. Total yield of the radiosynthesis:
22 3% after 90 min overall synthesis time.

8
M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
The purification of [18F]9 was performed by semipreparative HPLC
using 17% acetonitrile / 83% H2O +0.1% TFA (isocratic), followed by C-
18 cartridge purification to remove organic solvents. Purified [18F]9
was eluted from the cartridge with EtOH:H2O (1:1, +0.125 mg/mL
GIST430 cells, demonstrating the specific binding of [18F]9. The calculation
of valid IC50 values however proved to be difficult due to the high
unspecific accumulation of [18F]9 within the cells (Fig. 2A), resulting
in large errors (Fig. S6). This high unspecific interaction most likely
resulted from the cell uptake by passive diffusion through the cell
membrane of both [18F]9 and Imatinib.
To circumvent this problem and to obtain valid information about
the biological activity of [18F]9, a radiometric kinase assay was per-
formed using 9 different GIST-relevant KIT mutants and 9. Norimatinib
and SKI696 [16] were also evaluated under identical conditions as refer-
ence compounds. The advantage of this radiometric kinase assay is the
possibility to determine the kinase activity depending on the concentra-
tion of inhibitor while eliminating interfering processes like unspecific
uptake and binding using viable cells. Based on this approach, we ob-
served that 9 shows a comparable to slightly lower affinity towards
the tested KIT analogs compared to Norimatinib but is nevertheless ab-
solutely able to distinguish between Imatinib-sensitive and resistant KIT
mutations (e.g. sensitive V559D mutant vs. V559D/T670I mutant featur-
ing Imatinib resistance by the gatekeeper mutation [45–48]) (Fig. 3).
ascorbic acid) and obtained in an overall yield of 22.2
3.3% (n =
15) and a molar activity of 3.1
1.7 GBq/μmol starting from 3.8
0.1 GBq of [18F]fluoride. Sample spiking with the non-radioactive refer-
ence [9] identified the eluted product as [18F]9 via HPLC analysis (see
Fig. S4).
3.3. Determination of stability and lipophilicity of [18F]9
[
18F]9 proved to be stable in phosphate buffered saline solution
(1 mM, pH 7.4) at room temperature, as well as in human blood
serum at 37 °C for at least 2 h, demonstrating no detectable degradation
products. The distribution coefficient (logD, determined in 1-octanol/
phosphate buffer (0.05 M, pH 7.4)) of [18F]9 was determined to be
1.16 0.01 (n = 6). Thus, the aimed reduction of lipophilicity of [18F]
9 compared to Imatinib (2.34 0.17 [43]) was successful.
3.4. Characterization of the in vitro binding parameters of the new
radioligand
3.5. In vivo evaluation of the new radioligand
To determine if the biological activity of the newly developed tracer
The promising results of the kinase assay prompted us to investigate
the in vivo behavior of [18F]9 in GIST tumor-bearing Balb/c SCID mice.
The biodistribution at 2 h post injection (Fig. 4) showed a moderate up-
take of the tracer in Imatinib-sensitive GIST882 tumors of 1.48 0.03 %
ID/g, resulting in a tumor-to-muscle-ratio (TMR) of 2.21 0.28 which
slightly exceeds that of [18F]SKI696 in K562 tumors [16]. Interestingly,
the Imatinib-resistant GIST430 tumors also showed an uptake of [18F]
9 of 1.22 0.24 % ID/g, resulting in a TMR of 1.81 0.26. As described
in the in vitro evaluations section, GIST430 cells show a residual specific
binding of Imatinib, which partially explains these observations.
Furthermore, GIST430 tumors have a far better vascularization
(more than 2.5-fold higher microvessel permeability and plasmatic vol-
ume [49]) compared to GIST882 tumors, also contributing to the ob-
served, relatively high unspecific tumor uptake of the tracer. Thus, a
discrimination between Imatinib-resistant and sensitive tumors was
possible under in vitro but unfortunately not under in vivo conditions.
However, the liver uptake (5.54 0.66 % ID/g in the GIST882 tumor
model and 5.99 1.71% ID/g in the GIST430 tumor model) was signifi-
cantly reduced compared to that of [18F]SKI696 in female nude mice
(11.84 3.1% ID/g) [16] and at the same time, the kidney uptake was
[
18F]9 is retained despite the performed structural modification com-
pared to Imatinib, its cellular uptake as well as KIT binding parameters
were assessed. The cell uptake assays demonstrated a still adequate
cell permeability of the radiotracer (Fig. 2A), which is a prerequisite to
enable the tracer to address its intracellular target.
Furthermore, we could show that the uptake of [18F]9 into the GIST
cells is in part specific and depends on the sensitivity of the respective
cell line towards Imatinib (Fig. 2B). The Imatinib-sensitive GIST-T1
and GIST882 cells showed a specific binding of 9.3 1.0% and 11.0
0.4%, respectively, whereas the Imatinib-resistant GIST430 cells showed
a much smaller specific uptake. Although GIST430 is known as Imatinib-
resistant cell line, this observed residual specific binding of [18F]9 is in
accordance with previous studies showing that GIST430 cells are not
completely resistant towards Imatinib in an in vitro setting [44].
Furthermore, we also determined the affinity of [18F]9 towards KIT by
different approaches. Initially, a competitive binding assay using Imatinib-
sensitive (GIST-T1, GIST882) and Imatinib-resistant (GIST430) cell lines
was performed. As expected, a displacement of [18F]9 by its competitor I-
matinib (5 μM) could be observed in GIST-T1 and GIST882 but not in
Fig. 2. Results of the evaluation of the cell permeability and specific binding of [18F]9. A: Cellular uptake and distribution of [18F]9 across the different cell fractions (S: supernatant, M:
membrane bound, C: intracellular). In lighter color are displayed the uptakes without blocking. In darker color are displayed the uptakes under blocking and otherwise identical
conditions (using 1 μM Imatinib as competitor). B: Differences in cellular uptake between non-blocking and blocking conditions are depicted for each cell line. GIST-T1, GIST882:
Imatinib-sensitive, GIST430: Imatinib-resistant.

M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
9
Fig. 5. PET/CT images (coronal slices of PET images) of tumor-bearing mice at 18 min post
injection of [18F]9. The tumor is marked with the blue crosshair and the arrow. The mice
were anesthetized with isoflurane (2–3%) and the radiotracer (6.11–9.57 MBq) was
applied via the lateral tail vein. The animals were imaged over 120 min (90 min PET and
30 min CT).
development of radiolabeled kinase inhibitors and narrow the clinical
applications of these radiotracers [50]. Although our approach to use
the main metabolite as lead structure did not give satisfactory results,
we addressed some of these issues and thus improved the pharmacoki-
netic behavior of our newly developed radioligand.
Fig. 3. Binding affinities of Norimatinib, 9 (FNI) and SKI696 towards different GIST-
relevant KIT mutants (n = 1, Z′-factor was N0.52 ensuring excellent assay quality).
increased, resulting in a more pronounced and more favorable renal
excretion of [18F]9 compared to [18F]SKI696.
Also, dynamic in vivo PET imaging experiments were conducted
using [18F]9, showing a moderate tumor uptake of the tracer (Fig. 5).
Some of the limitations like a low tumor-to-background ratio, high
lipophilicity and high hepatic uptake are common issues in the
Unfortunately, the strong difference in vascularization of the exam-
ined tumors, the only partial Imatinib-resistance of the used GIST430
cell line as well as the still relevant lipophilicity of the ligand, resulting
in an unspecific uptake of the tracer might have influenced the tumor
accumulation of the radiotracer and prevented a discrimination be-
tween GIST882- and GIST430-based tumors in vivo although giving
Fig. 4. Biodistribution of [18F]9 in Balb/c SCID mice bearing GIST430 (n = 5) and GIST882 (n = 5) tumors at 2 h post injection of [18F]9. The mice were sacrificed directly after PET/CT
measurements having received 5–12 MBq of the radioligand and having been imaged over 120 min.

10
M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
promising results in vitro. However, [18F]9 might be a promising
candidate for the functional in vivo imaging of other Imatinib-
sensitive diseases such as CML and should provide an interesting basis
for the development of new radioligands for the imaging of Imatinib-
sensitivity with PET.
[12] Verweij J, Casali PG, Zalcberg J, LeCesne A, Reichardt P, Blay JY, et al. Progression-free
survival in gastrointestinal stromal tumours with high-dose imatinib: randomised
trial. Lancet 2004;364:1127–34.
[13] Li B, Wang A, Liu J, Qi Z, Liu X, Yu K, et al. Discovery of N-((1-(4-(3-(3-((6,7-
Dimethoxyquinolin-3-yl)oxy)phenyl)ureido)-2-(trifluoromethyl)phenyl)piperidin-
4-yl)methyl)propionamide (CHMFL-KIT-8140) as a highly potent type II inhibitor
capable of inhibiting the T670I “Gatekeeper” mutant of cKIT kinase. J Med Chem
2016;59:8456–72.
[14] Knight ZA, Shokat KM. Features of selective kinase inhibitors. Chem Biol 2005;12:
4. Conclusions
621–37.
[15] Zuccotto F, Ardini E, Casale E, Angiolini M. Through the “Gatekeeper Door”:
exploiting the active kinase conformation. J Med Chem 2010;53:2681–94.
[16] Glekas AP, Pillarsetty NK, Punzalan B, Khan N, Smith-Jones P, Larson SM. In vivo im-
aging of Bcr-Abl overexpressing tumors with a radiolabeled imatinib analog as an
imaging surrogate for imatinib. J Nucl Med 2011;52:1301–7.
[17] le Coutre P, Kreuzer K-A, Pursche S, Bonin Mv, Leopold T, Baskaynak G, et al. Pharma-
cokinetics and cellular uptake of imatinib and its main metabolite CGP74588. Cancer
Chemother Pharmacol 2003;53:313–23.
In summary, we developed a novel 18F-labeled small molecule
radiotracer targeting KIT. Furthermore, we characterized the new radio-
tracer regarding its in vitro and in vivo properties. The new radioligand
[
18F]fluoronorimatinib shows similar KIT binding properties in vitro
than Norimatinib and SKI696, comparable tumor-to-background-ratios
than previously published Imatinib-derived radiotracers in vivo but ex-
hibits a considerably reduced hepatic uptake compared to former radio-
tracers which is highly favorable for in vivo GIST imaging.
[18] Veach DR, Namavari M, Pillarsetty N, Santos EB, Beresten-Kochetkov T, Lambek C,
et al. Synthesis and biological evaluation of a fluorine-18 derivative of dasatinib. J
Med Chem 2007;50:5853–7.
[19] Mol CD. Structural basis for the autoinhibition and STI-571 inhibition of c-Kit tyro-
sine kinase. J Biol Chem 2004;279:31655–63.
[20] Peng ZH, Maxwell DS, Sun DL, Prasad BAB, Pal A, Wang SM, et al. Imatinib analogs as
potential agents for PET imaging of Bcr-Abl and c-KIT expression at a kinase level.
Bioorg Med Chem 2014;22:623–32.
Acknowledgements
[21] van de Waterbeemd H, Smith DA, Beaumont K, Walker DK. Property-based design:
optimization of drug absorption and pharmacokinetics. J Med Chem 2001;44:
1313–33.
This work was performed within the MITIGATE project, co-founded
by the European Community's 7th Framework Programme (FP7/2007-
2013) under grant agreement no. 602306. This research project is part
of the Research Campus M2OLIE and funded by the German Federal
Ministry of Education and Research (BMBF) under the funding code
13GW0091E. The authors acknowledge financial support of the German
Research Foundation for the small animal PET/SPECT/CT imaging
system Albira Bruker II (funding code: INST 91027/11-1-FUGG).
The authors thank Dr. M. Roscher and A.-M. Suhr for their excellent
support during the PET/CT experiments, Dr. J.A. Fletcher (Dana-Farber,
Boston) for the kind gift of Cancer Center GIST cell lines (GIST430,
GIST882) and Dr. W. Spahl (LMU, Munich) for mass spectrometry
measurements.
[22] Arnott JA, Kumar R, Planey SL. Lipophilicity indices for drug development. J Appl
Biopharm Pharmacokinet 2013;1:31–6.
[23] Gschwind HP. Metabolism and disposition of imatinib mesylate in healthy volun-
teers. Drug Metab Dispos 2005;33:1503–12.
[24] Deadman BJ, Hopkin MD, Baxendale IR, Ley SV. The synthesis of Bcr-Abl inhibiting
anticancer pharmaceutical agents imatinib, nilotinib and dasatinib. Org Biomol
Chem 2013;11:1766–800.
[25] Zimmermann J, Buchdunger E, Mett H, Meyer T, Lydon NB. Potent and selective in-
hibitors of the Abl-kinase: phenylamino-pyrimidine (PAP) derivatives. Bioorg Med
Chem Lett 1997;7:187–92.
[26] Dolci L, Dolle F, Jubeau S, Vaufrey F, Crouzel C. 2-[F-18]fluoropyridines by no-carrier-
added nucleophilic aromatic substitution with [F-18]FK-K-222 - a comparative
study. J Label Compd Radiopharm 1999;42:975–85.
[27] Jacobson O, Kiesewetter DO, Chen X. Fluorine-18 radiochemistry, labeling strategies
and synthetic routes. Bioconjug Chem 2015;26:1–18.
[28] Kuhnast B, de Bruin B, Hinnen F, Tavitian B, Dollé F. Design and synthesis of a new
[18F]fluoropyridine-based haloacetamide reagent for the labeling of oligonucleo-
tides: 2-bromo-N-[3-(2-[18F]fluoropyridin-3-yloxy)propyl]acetamide. Bioconjug
Chem 2004;15:617–27.
Appendix A. Supplementary data
Supplementary data to this article can be found online at https://doi.
org/10.1016/j.nucmedbio.2017.11.004.
These data contain HPLC chromatograms, additional figures, NMR
spectra and NMR assignments.
[29] Pike VW. PET radiotracers: crossing the blood–brain barrier and surviving metabo-
lism. Trends Pharmacol Sci 2009;30:431–40.
[30] Dolle F. Fluorine-18-labelled fluoropyridines: advances in radiopharmaceutical de-
sign. Curr Pharm Des 2005;11:3221–35.
[31] Kuchar M, Mamat C. Methods to increase the metabolic stability of 18F-radiotracers.
Molecules 2015;20:16186–220.
[32] Iversen PW, Eastwood BJ, Sittampalam GS, Cox KL. A comparison of assay perfor-
mance measures in screening assays: signal window, Z' Factor, and assay variability
ratio. J Biomol Screen 2006;11:247–52.
References
[33] Zhang JH. A simple statistical parameter for use in evaluation and validation of high
[1] Blanke CD, Demetri GD, von Mehren M, Heinrich MC, Eisenberg B, Fletcher JA, et al.
Long-term results from a randomized phase II trial of standard- versus higher-dose
imatinib mesylate for patients with unresectable or metastatic gastrointestinal stro-
mal tumors expressing KIT. J Clin Oncol 2008;26:620–5.
[2] Blanke CD, Rankin C, Demetri GD, Ryan CW, von Mehren M, Benjamin RS, et al.
Phase III randomized, intergroup trial assessing imatinib mesylate at two dose levels
in patients with unresectable or metastatic gastrointestinal stromal tumors express-
ing the kit receptor tyrosine kinase: S0033. J Clin Oncol 2008;26:626–32.
[3] Call J, Walentas CD, Eickhoff JC, Scherzer N. Survival of gastrointestinal stromal
tumor patients in the imatinib era: life raft group observational registry. BMC Cancer
2012;12:90.
throughput screening assays. J Biomol Screen 1999;4:67–73.
[34] Kompella A, Adibhatla BRK, Muddasani PR, Rachakonda S, Gampa VK, Dubey PK. A
facile total synthesis for large-scale production of imatinib base. Org Process Res
Dev 2012;16:1794–804.
[35] Agosti A, Bertolini G, Bruno G, Lautz C, Glarner T, Deichtmann W. Handling hydrogen
peroxide oxidations on a large scale: synthesis of 5-bromo-2-nitropyridine. Org Pro-
cess Res Dev 2017;21:451–9.
[36] Brown EV. Preparation and reactions of 2-nitropyridine-1-oxides. JACS 1957;79:
3565–6.
[37] Parry PR, Wang CS, Batsanov AS, Bryce MR, Tarbit B. Functionalized pyridylboronic
acids and their Suzuki cross-coupling reactions to yield novel heteroarylpyridines.
J Org Chem 2002;67:7541–3.
[4] De Giorgi U, Verweij J. Imatinib and gastrointestinal stromal tumors: where do we
go from here? Mol Cancer Ther 2005;4:495–501.
[38] Bursavich MG, Lombardi S, Gilbert AM. Expedient parallel synthesis of 2-amino-4-
[5] Druker BJ, Guilhot F, O'Brien SG, Gathmann I, Kantarjian H, Gattermann N, et al. Five-
year follow-up of patients receiving imatinib for chronic myeloid leukemia. N Engl J
Med 2006;355:2408–17.
[6] Gold JS, van der Zwan SM, Gonen M, Maki RG, Singer S, Brennan MF, et al. Outcome
of metastatic GIST in the era before tyrosine kinase inhibitors. Ann Surg Oncol 2007;
14:134–42.
heteroarylpyrimidines. Org Lett 2005;7:4113–6.
[39] Simkovsky NM, Ermann M, Roberts SM, Parry DM, Baxter AD. Some regioselective
cross-coupling reactions of halopyridines and halopyrimidines. J Chem Soc Perkin
1 2002;1:1847–9.
[40] Schirrmacher R, Wängler C, Schirrmacher E. Munich Molecular Imaging Handbook
Series. Pharmaceutical Radiochemistry (I), volume 1. Fürstenfeldbruck: Scintomics
Print Media and Publishing; 2010.
[7] Morphy R. Selectively nonselective kinase inhibition: striking the right balance. J
Med Chem 2010;53:1413–37.
[41] Kügler F, Ermert J, Coenen HH. Labeling of benzodioxin piperazines with fluorine-18
as prospective radioligands for selective imaging of dopamine D4 receptors. J Label
Compd Radiopharm 2013;56:609–18.
[8] Overton LC, Heinrich MC. Regorafenib for treatment of advanced gastrointestinal
stromal tumors. Expert Opin Pharmacother 2014;15:549–58.
[9] Beham AW, Schaefer I-M, Schüler P, Cameron S, Michael Ghadimi B. Gastrointestinal
stromal tumors. Int J Color Dis 2011;27:689–700.
[10] Linch M, Claus J, Benson C. Update on imatinib for gastrointestinal stromal tumors:
duration of treatment. Onco Targets Ther 2013;6:1011–23.
[42] Epstein WW, Sweat FW. Dimethyl sulfoxide oxidations. Chem Rev 1967;67:
247–60.
[43] Kil K-E, Ding Y-S, Lin K-S, Alexoff D, Kim SW, Shea C, et al. Synthesis and positron
emission tomography studies of carbon-11-labeled imatinib (Gleevec). Nucl Med
Biol 2007;34:153–63.
[11] Gramza AW, Corless CL, Heinrich MC. Resistance to tyrosine kinase inhibitors in gas-
trointestinal stromal tumors. Clin Cancer Res 2009;15:7510–8.

M. Prause et al. / Nuclear Medicine and Biology 57 (2018) 1–11
11
[44] Bauer S, Duensing A, Demetri GD, Fletcher JA. KIT oncogenic signaling mechanisms
in imatinib-resistant gastrointestinal stromal tumor: PI3-kinase/AKT is a crucial
survival pathway. Oncogene 2007;26:7560–8.
[45] Gajiwala KS, Wu JC, Christensen J, Deshmukh GD, Diehl W, DiNitto JP, et al. KIT ki-
nase mutants show unique mechanisms of drug resistance to imatinib and sunitinib
in gastrointestinal stromal tumor patients. PNAS 2009;106:1542–7.
[46] Antonescu CR. The GIST paradigm: lessons for other kinase-driven cancers. J Pathol
2011;223:252–62.
[48] Vincenzi B, Nannini M, Fumagalli E, Bronte G, Frezza AM, De Lisi D, et al.
Imatinib dose escalation versus sunitinib as a second line treatment in
KIT exon 11 mutated GIST: a retrospective analysis. Oncotarget 2016;7:
69412–9.
[49] Consolino L, Longo DL, Sciortino M, Dastrù W, Cabodi S, Giovenzana GB, et al.
Assessing tumor vascularization as a potential biomarker of imatinib resistance in
gastrointestinal stromal tumors by dynamic contrast-enhanced magnetic resonance
imaging. Gastric Cancer 2016;20:629–39.
[47] Richters A, Ketzer J, Getlik M, Grütter C, Schneider R, Heuckmann JM, et al. Targeting
gain of function and resistance mutations in Abl and KIT by hybrid compound
design. J Med Chem 2013;56:5757–72.
[50] Bernard-Gauthier V, Bailey JJ, Berke S, Schirrmacher R. Recent advances in the devel-
opment and application of radiolabeled kinase inhibitors for PET imaging. Molecules
2015;20:22000–27.