One-step synthesis of [18F]cabozantinib for use in positron emission tomography imaging of c-Met

Article abstract of DOI:10.1002/jlcr.3564

Cabozantinib is an FDA-approved kinase inhibitor for the treatment of medullary thyroid cancer and advanced renal cell carcinoma, which exerts its therapeutic effect by inhibiting, among others, the tyrosine kinase c-Met. Noninvasive imaging techniques are becoming increasingly important clinically to ensure drug efficacy, staging, monitoring, and patient stratification. PET isotope labelled tyrosine kinase inhibitors have, for the same reason, potential as PET tracers for imaging of various cancers. On the basis of cabozantinib, we synthesized the novel boronic acid pinacol ester 4 as a labelling precursor, where the boronic ester moiety replaces the fluorine native to this kinase inhibitor. By this, we wanted to explore whether recently developed Cu-mediated fluorination methods are adaptable to more complex substrates and thereby provide easy access to [¹⁸F]cabozantinib directly. Hydrolysis was implemented before preparative purification due to challenges with on-column hydrolysis of the precursor 4, and [¹⁸F]cabozantinib was obtained in ≥99% radiochemical purity and in 2.8?±?0.05% (n?=?4) isolated decay corrected yield in a synthesis time of 90?minutes. The molar activity of representative batches was determined to be 17?±?8?GBq/μmol.

Full text of DOI:10.1002/jlcr.3564

One-step synthesis of [18F]cabozantinib for use in positron emission tomography
imaging of c-Met.
Vegard Torp Lien
- Universitetet i Oslo
Pharmacy, Sem Sælands vei 3
6305 Oslo, Norway
- v.t.lien@farmasi.uio.no
Jo Klaveness
- Universitetet i Oslo
Pharmacy, Sem Sælands vei 3
6305 Oslo, Norway
- jo.klaveness@farmasi.uio.no
Dag Erlend Olberg
- Norwegian Medical Cyclotron Center
Research&Development, Oslo, Norway
- dag.erlend.olberg@syklotronsenteret.no
Abstract
Cabozantinib is an FDA approved kinase inhibitor for the treatment of medullary thyroid
cancer and advanced renal cell carcinoma, which exerts its therapeutic effect by inhibiting,
among others, the tyrosine kinase c-Met. Non-invasive imaging techniques are becoming
increasingly important clinically to ensure drug efficacy, staging, monitoring and patient
stratification. PET isotope labelled tyrosine kinase inhibitors have, for the same reason,
potential as PET tracers for imaging of various cancers. Based on cabozantinib, we
synthesized the novel boronic acid pinacol ester 4 as a labelling precursor, where the boronic
ester moiety replaces the fluorine native to this kinase inhibitor. By this we wanted to explore
whether recently developed Cu-mediated fluorination methods are adaptable to more
complex substrates and thereby provide easy access to [18F]cabozantinib directly. Hydrolysis
was implemented before preparative purification due to challenges with on-column
hydrolysis of the precursor 4, and [18F]cabozantinib was obtained in ≥ 99 % radiochemical
purity and in 2.8±0.05 % (n=4) isolated decay corrected yield in a synthesis time of 90
minutes. The molar activity of representative batches was determined to be 17±8 GBq/μmol.
This article has been accepted for publication and undergone full peer review but has not
been through the copyediting, typesetting, pagination and proofreading process which may
lead to differences between this version and the Version of Record. Please cite this article as
doi: 10.1002/jlcr.3564
This article is protected by copyright. All rights reserved.

Introduction
Kinase alterations have emerged as one of the most important factors for why cancers
proliferate and metastasize, and are therefore viewed as a central oncological drug target.1
This has led to substantial research activity to understand the function and to inhibit the
action of various kinases.2 To date, 34 kinase inhibitors are approved by the U.S. Food and
Drug Administration (FDA),3,4 and despite challenges regarding resistance and relapse in
patients, the field is developing and expanding. Although only a small portion of the human
kinome has been studied in detail, many kinases have been shown to be key regulators of
important oncogenes. The tyrosine kinase c-Met plays a central role in many cancer types, e.g.
lung, colon and gastric cancer,5 by affecting cell proliferation, motility and invasion.6
Hepatocyte growth factor (HGF) binds extracellular to the c-Met receptor, which then leads
to several downstream events. Importantly, it has been observed that c-Met is frequently
amplified in metastatic tumors.5 This suggest its role in the proliferation of cancer cells
throughout the body, and further emphasize the importance of therapeutics inhibiting the c-
Met pathway, as metastasis is the primary cause of cancer related deaths. Therefore,
considerable efforts have been made towards inhibiting this target.7 To date, two c-Met drugs
have been clinically approved, which are crizotinib (to treat subtypes of non-small cell lung
cancer) and cabozantinib (to treat medullary thyroid cancer and advanced renal cell
carcinoma). The structures of these drugs are shown in Figure 1.
Positron emission tomography (PET) has a valuable role to play in making personalized
medicine a clinical reality. To be able to exclude a patient population that is unlikely to
respond to a specific therapeutic treatment through the use of companion diagnostics might
be of great value, since it enables individually tailored therapy regimen to be identified. Due
to potential severe side-effects of kinase inhibitors, high development cost, and patient-
unique drug resistance, the correct kinase inhibitor regimen is imperative. By using PET, the
target expression for a certain drug can be assessed in vivo, and this could give vital clinical
information resulting in a more targeted and effective treatment. Several of the FDA-
approved kinase inhibitors have in recent years been labelled with positron emitting nuclides
and investigated as PET tracers.8,9 Overexpression of c-Met in several tumors5 motivates the
use of PET, and studies have been performed with a small-molecule,10 a monoclonal
antibody11 and HGF itself.12 Cabozantinib is an attractive candidate for labelling with either
fluorine-18 or carbon-11 as it contains two positions that are readily available for labelling, a
methoxy group and a fluorine, as shown in Figure 2. An advantage of using exactly the same
structure for the radiopharmaceutical as the approved drug is that pharmacology, toxicity and
metabolism are already known, which should promote more rapid clinical translation.
For cabozantinib, 11C-labelling of the methoxy group is feasible from the demethylated
precursor using e.g. [11C]CH3I. However, with its short half-life (20 min), carbon-11 is not as
attractive for regular clinical applications and off-site distribution. Introduction of fluorine-18
into the aromatic ring is expected to be impractical by a direct SNAr reaction due to the
electron donating nature of the anilide. An alternative strategy could be a multiple step
procedure starting with 18F-labelling of dinitrobenzene, subsequent reduction to the aniline
and acylation. This method was recently reported in the patent literature.13 However, such a
multi-step radiosynthesis is time consuming and inherently challenging to automate.
This article is protected by copyright. All rights reserved.

Methods for introducing fluorine-18 directly in electron-rich aryls, as is the case for
cabozantinib, has recently been developed.14 One method utilizes boronic acid pinacol esters
as substrates in copper mediated reactions with [18F]fluoride to give the 18F-labelled products
in one step, both with electron releasing and electron donating groups present.15,16 Many
boronic acid pinacol esters are commercially available, aryl borylation chemistry is well
established and the utility of these precursors for [18F]fluorination chemistry should be
straight-forward to implement. This one-step procedure simplifies and opens new chemical
avenues for both old and new PET-tracers, including cabozantinib. So far, this method has
seen limited use outside the inventing research group17,18,19, and it remains to see whether it is
robust enough to be applicable to a wider selection of substrates using GBqs of [18F]fluoride.
Herein, we report the chemical synthesis of the boronic acid pinacol ester precursor 4 and a
one-step, automated radiosynthesis of [18F]cabozantinib.
Results and discussion
The desired BPin precursor 4 was synthesized over four steps, in an overall yield of 14 %, as
shown in Scheme 1. Carboxylic acid 1 was obtained from reaction of commercial
cyclopropane-1,1-dicarboxylic acid and 4-aminophenylboronic acid pinacol ester. 6,7-
dimethoxyquinolin-4-ol was reacted with 1-fluoro-4-nitrobenzene to yield quinoline 2. This
was followed by reduction of the nitro group to the aniline 3, which was then coupled with
the carboxylic acid 1 to give the precursor 4.
The reaction of 4 with [18F]fluoride was performed as shown in Scheme 2a, using the same
protocol as previously reported.15 Initially, a test reaction was conducted by adding an aliquot
(40 μL, 13 MBq) of [18F]KF/K222 in acetonitrile to a reaction vial containing 4 and
Cu(OTf)2(py)4 in N,N-dimethylformamide (DMF). Analytical HPLC of the reaction mixture
after 20 minutes at 120 °C showed that the reaction conditions indeed gave the radiolabelled
product, confirmed by co-elution with non-radioactive cabozantinib. Radio-TLC indicated a
conversion yield of 36 %, and radio-HPLC revealed that [18F]cabozantinib was the only
radioactive species besides unreacted [18F]fluoride. The observed yield was in the expected
range, and no further optimization was attempted at this scale.
To be useful as a method to produce [18F]cabozantinib for small-animal PET-studies, and
eventually for clinical use, the chemistry must allow for higher amounts of radioactivity and
be amenable to automation. Therefore, a full-scale experiment using 15 GBq of activity was
performed. This time, a change from DMF to N,N-dimethylacetamide (DMA) as solvent and
a higher catalyst loading (1.4 equivalents) was used, consistent with optimisations made to
the method in later publications.16 After semi-preparative HPLC the isolated yield was only
0.7 % (decay corrected yield). As much as 40 % of the activity was remaining in the reaction
vial or on the Chromafix® PS-HCO3 column. Modifying the preconditioning (using oxalate
solution) and the elution (using more K2CO3) of the column, resulted in only 13 % loss of
activity, with the decay corrected yield increasing to 2.8 %. HPLC analysis of the purified
fraction containing [18F]cabozantinib revealed a relatively high contamination of several non-
radioactive impurities (see supporting information, figure S8). The identity of the main
contamination was identified as 5 (Figure 2b), the boronic acid of 4, using LC-MS. Boronic
acid pinacol esters are in general considered to a be stable class of compounds, but hydrolysis
has been shown to occur for some derivatives, and possibly yield problems in
purifications.20,21 The precursor 4 has the longest retention time of all species in the reaction
This article is protected by copyright. All rights reserved.

mixture, and on-column hydrolysis to the more hydrophilic and faster eluting 5 will result in
contamination throughout the semi-preparative run. Different eluents (gradients with
acetonitrile or ethanol) and additives (NH4COOH) were investigated to improve the on-
column stability of 4, however, with little success. A pre-purification hydrolysis strategy of 4
was then considered, and heating in phosphate buffer (pH 7.4) at 100 ºC for 10 minutes
indicated a complete hydrolysis. When implementing hydrolysis before preparative HPLC in
the automated synthesis, [18F]cabozantinib was obtained in ≥ 99 % radiochemical purity with
a decay corrected yield of 2.8±0.05 % and molar activity 17±8 GBq/μmol (n=4). As can be
seen in Figure 3, trace amounts of 5 was still present, however this was quantified using
HPLC to only 15 μg (0.17 % of the amount 4 added). These results indicate that also more
complex 18F-labelled tracers are available in a one-step reaction with this protocol. However,
the observed hydrolysis of the precursor in this case, with subsequent challenges during
purification, is a potential drawback with the method.
Conclusions
The FDA-approved tyrosine kinase inhibitor cabozantinib was synthesized as its 18F-labelled
analogue using a one-step procedure from the boronic acid pinacol ester precursor 4 in
2.8±0.05 % yield and with a molar activity of 17±8 GBq/μmol (n=4). The precursor was
synthesized over four steps. The applied method proved to produce [18F]cabozantinib as the
sole radioactive species, although in low yield. Non-radioactive impurities due to hydrolysis
of 4 prompted us to explore a pre-column hydrolysis strategy which improved the chemical
purity of the final product. It is apparent that some boronic acid pinacol ester substrates are
prone to hydrolysis using this method, which may lead to challenges during HPLC
purification. Despite the low yield, a total of around 280±15 MBq (n=4) of [18F]cabozantinib
was isolated, which are suitable for applications in small-animal PET-studies. After
reformulation, [18F]cabozantinib will be tested as a potential companion PET tracer for
cabozantinib in relevant tumor models.
Experimental
General
All chemicals were purchased from Sigma–Aldrich or Fluorochem and used without further
purification. Air and/or moisture sensitive reactions were performed under argon atmosphere
with dried solvents and reagents. TLC was performed on Merck silica gel 60 F254 plates, and
visualized using UV light at 312 nm or 365 nm, a phosphomolybdic acid solution (12 g phos-
phomolybdic acid in 250 mL EtOH) or a potassium permanganate solution (1,5 g KMnO4, 10
g K2CO3, 2,5 mL 5M NaOH/H2O, 200 mL H2O). Column chromatography was performed
with silica gel (pore size 60 Å, 230–400 mesh particle size) purchased from Fluka. 1H and 13C
NMR spectra were obtained on a Bruker AVIII HD 400 instrument (400/101 MHz). Chemi-
cal shifts (δ) are reported in parts per million, and coupling constants are reported in Hertz
(Hz).The residual proton solvent resonance in 1H NMR (CDCl3 at δ 7.27, DMSO-d6 at δ 2.50)
and the residual carbon solvent resonance in 13C NMR (CDCl3 at δ 77.16 ppm and DMSO-d6
at δ 39.52) are used as reference. Accurate mass determination (HRMS) in positive or nega-
tive mode was performed on a Waters Prospec Q instrument, ionized by electrospray (ESI).
LC-MS was performed on a Thermo Finnigan LCQ Deca XP Plus and preparative HPLC was
This article is protected by copyright. All rights reserved.

performed on a Waters Delta Prep 4000, using a gradient from 20% to 80 % acetonitrile in
water, collecting fractions of 10 mL/minute.
Precursor and cabozantinib synthesis
1-((4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-
yl)phenyl)carbamoyl)cyclopropanecarboxylic acid (1)
Triethylamine (0.30 mL, 2.2 mmol) was added via syringe to a solution of cyclopropane-1,1-
dicarboxylic acid (0.210 g, 1.6 mmol) in tetrahydrofuran (THF) (6 mL) at 0 ºC. The solution
was then stirred for 15 minutes at 0 ºC, before SOCl2 (0.12 mL, 1.6 mmol) was added via
syringe. After another 15 minutes of stirring, 4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-
yl)aniline (390 mg, 1.78 mmol) in THF (3 mL) was added via cannula at 0º C, and the
solution was then stirred at ambient temperature for 20 h. The crude mixture was diluted with
ethyl acetate (EtOAc) (20 mL), washed with water (2x10 mL) and brine (10 mL), dried over
MgSO4 and concentrated on a rotary evaporator. Work-up was performed under neutral
conditions, and with no further purifications, to prevent possible hydrolysis of the pinacol
1
ester. The title compound was achieved as an off white solid (0.53 g, 99 % crude yield). H
NMR (CDCl3, 400 MHz): δ 10.62, (s, 1H), 7.77 (d, 2H, J = 8.4 Hz), 7.55 (d, 2H, J = 8.4 Hz),
1.94-1.92 (m, 2H), 1.83-1.80 (m, 2H), 1.34 (s, 12H). 13C NMR (101 MHz, CDCl3): δ 177.75,
166.83, 140.49, 135.90, 119.63, 119.35, 83.94, 26.54, 25.00, 22.18. HRMS (ESI-) m/z calcd.
for C17H21NO5B [M-H]-: 330.1518 , found 330.1519.
6,7-dimethoxy-4-(4-nitrophenoxy)quinolone (2)
Cs2CO3 (3.68 g, 11.3 mmol) and 6,7-dimethoxyquinolin-4-ol (0.94 g, 4.57 mmol) were
weighed out in a round-bottom flask, added 10 mL DMF and 5 mL MeCN, and stirred for 15
minutes. 1-fluoro-4-nitrobenzene (1.49 g, 10.5 mmol) in 10 mL 1:1 DMF:MeCN was then
added, and the reaction mixture was stirred for 20 h at 55 ºC. The crude mixture was diluted
with EtOAc (80 mL), washed with water (4x40 mL) and brine (40 mL), dried over MgSO4
and concentrated on a rotary evaporator, and then purified by column chromatography
(Hep:EtOAc (4:1) → Hep:EtOAc:MeOH (10:10:1). The title compound was achieved as a
1
light yellow solid (0.486 g, 33%). H NMR (CDCl3, 400 MHz): δ 8.62 (d, 1H, J = 5.2 Hz),
8.32 (d, 2H, J = 9.2 Hz), 7.47 (s, 1H), 7.36 (s, 1H), 7.26, (d, 2H, J = 9.2 Hz), 6.69 (d, 1H, J =
5.2 Hz), 4.06 (s, 3H), 4.0 (s, 3H).
4-((6,7-dimethoxyquinolin-4-yl)oxy)aniline (3)
Iron (0.488 g, 8.74 mmol), NH4Cl (0.511 g, 9.53 mmol) and 2 (0.233, 0.714 mmol) were
weighed out in a round-bottom flask, and added water (10 mL) and EtOH (15 mL). The
reaction mixture was stirred for 3 h at 70 ºC, and then cooled, filtered through Celite and
washed with EtOAc (20 mL). The aqueous phase was extracted with EtOAc (2x15 mL), and
the combined organic phases were washed with water (10 mL) and brine (10 mL), dried over
MgSO4 and concentrated on a rotary evaporator. The title compound was achieved as a light
1
brown solid (0.176 g, 83 %), and used without further purification. H NMR (CDCl3, 400
MHz): δ 8.45 (d, 1H, J = 5.5 Hz), 7.58 (s, 1H), 7.42 (s, 1H), 6.99 (d, 2H, J = 8.8 Hz), 6.77 (d,
2H, J = 8.8 Hz), 6.43 (d, 1H, J = 5.5 Hz)
N-(4-((6,7-dimethoxyquinolin-4-yl)oxy)phenyl)-N-(4-(4,4,5,5-tetramethyl-1,3,2-
dioxaborolan-2-yl)phenyl)cyclopropane-1,1-dicarboxamide (4)
The acid 1 (0.135 g, 0.407 mmol), 1-[Bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-
b]pyridinium 3-oxid hexafluorophosphate (HATU) (0.184 g, 0.484 mmol) and DMF (2 mL)
were placed in a round bottom flask, and then added N,N-diisopropylethylamine (DIPEA)
This article is protected by copyright. All rights reserved.

(0.13 mL, 0.746 mmol). The aniline 3 (0.128 mg, 0.432 mmol) dissolved in DMF (4 mL) was
added after 10 minutes, and the mixture was stirred for 20 h at room temperature. The
mixture was then diluted with EtOAc (30 mL), washed with water (4x10 mL) and brine (10
mL), dried over MgSO4 and concentrated on a rotary evaporator. This solid was further
purified on a preparative HPLC, and the title compound was achieved as a white solid (123
mg, 50 %). 1H NMR (CDCl3, 400 MHz): δ 9.42 (s, 1H), 8.87 (s, 1H), 8.47 (d, 1H, J = 5.2 Hz),
7.79 (d, 2H, J = 8.4 Hz), 7.63 (d, 2H, J = 8.8 Hz), 7.54 (s, 1H), 7.53 (d, 2H, J = 8.4 Hz), 7.41
(s, 1H), 7.16 (d, 2H, J = 8.8 Hz), 6.45 (d, 2H, J = 5.2 Hz), 4.04 (s, 3H), 4.03 (s, 3H), 1.73-
1.70 (m, 2H), 1.67-1.64 (m, 2H), 1.34 (s, 12H). 13C NMR (101 MHz, CDCl3) δ 169.23,
168.87, 160.79, 153.01, 151.12, 149.69, 148.99, 147.04, 139.92, 135.98, 134.94, 122.61,
121.77, 119.59, 116.24, 107.97, 103.57, 99.63, 83.98, 56.32, 56.27, 29.67, 25.01, 17.67.
HRMS (ESI+) m/z calcd. for C34H37N3O7B [M+H]+: 610.2719, found 610.2722.
N-(4-((6,7-dimethoxyquinolin-4-yl)oxy)phenyl)-N-(4-fluorophenyl)cyclopropane-1,1-
dicarboxamide (cabozantinib)
1-((4-fluorophenyl)carbamoyl)cyclopropanecarboxylic acid22 (0.146 g, 0.654 mmol), HATU
(0.415 g, 1.09 mmol) and DMF (5 mL) were placed in a round-bottom flask, and then added
DIPEA (0.26 mL, 1.49 mmol). 3 (0.157 g, 0.53 mmol) dissolved in DMF (5 mL) was added
after 10 minutes. The mixture was stirred for 20 h at room temperature. The mixture was then
diluted with EtOAc (30 mL), washed with water (4x10 mL) and brine (10 mL), dried over
MgSO4 and concentrated on a rotary evaporator. This solid was further purified on a
1
preparative HPLC. The title compound was achieved as a white solid (0.112 g, 42 %). H
NMR (CDCl3, 400 MHz): δ 10.20 (s, 1H), 8.74 (s, 1H), 8.46 (d, 1H, J = 6.4 Hz), 7.99 (s, 1H),
7.79 (d, 2H, J = 8.8 Hz), 7.63 (s, 1H), 7.50-7.47 (m, 2H), 7.18 (d, 2H, J = 8.8 Hz), 7.02 (t, 2H,
J = 8.8 Hz), 6.68 (d, 1H, J = 6.4 Hz), 4.13 (s, 3H), 4.09 (s, 3H), 1.83-1.81 (m, 2H), 1.69-1.67
(m, 2H). 13C NMR (101 MHz, CDCl3): δ 170.08, 169.04, 165.83, 159.99 (d, J=246 Hz),
156.86, 151.90, 148.70, 141.46, 138.70, 137.12, 133.15 (d, J=3 Hz), 123.24 (d, J=8 Hz),
122.95, 121.65, 116.20, 115.87 (d, J=23 Hz), 102.67, 101.13, 100.05, 57.37, 56.74, 29.21,
18.23.
Radiochemistry
General
18
[18F]fluoride was produced from the O(p,n)18F nuclear reaction on a GE PETtrace 880 cy-
clotron. Radiosynthesis and azeotropic drying was performed on a ScanSys chemistry module
(Værløse, Denmark). Radio-HPLC analyses were performed using an Agilent system (1100
series) with UV detection equipped in series with radiodetector (10 to 90 % acetonitrile in
water (0.05 % TFA) over 10 minutes, ACE 3 µm C18-AR 50×4.6mm column, flow
1mL/min). Radio-TLC was performed using a Raytest MiniGita (Straubenhardt, Germany).
Small-scale reaction
[18F]fluoride (~1 GBq) was separated from 18O-enriched-water using a Chromafix® PS-HCO3
18F cartridge and released with 900 μL of a solution of K222/K2CO3 (Kryptofix® 222 (145 mg)
and K2CO3 (29 mg) in 10 mL of acetonitrile/H2O (4:1)) into a 3 mL V-vial. The Chromafix®
PS-HCO3 18F cartridge was preconditioned with 1 mL of 1 M oxalic acid solution, 10 mL of
H2O and 10 mL of air. The solution was dried with two cycles of azeotropic drying with ace-
tonitrile (200 μl) under a flow of helium at 100 °C, and redissolved in anhydrous MeCN (0.6
mL). A 3 mL vial containing precursor 4 (21.8 mg, 0.036 mmol) and Cu(OTf)2(py)4 (3.9 mg,
0.0058 mmol) was added [18F]KF/K222 in MeCN (40 μL). Anhydrous DMF (300 μL) was
This article is protected by copyright. All rights reserved.

added and the vial was purged with 10 mL of air using a syringe, and then heated at 120 °C
for 20 min. The reaction was quenched by adding water (300 μL), and analysed by radio-TLC
and HPLC.
Automated synthesis
[18F]fluoride (~15 GBq) was separated from 18O-enriched-water using a Chromafix® PS-
HCO3 18F cartridge and released with 900 μL of a solution of K222/K2C2O4/K2CO3 (Krypto-
fix® 222 (63 mg), K2C2O4 (12 mg) and K2CO3 (2.5 mg) in 10 mL of acetonitrile/H2O (4:1))
into a 3 mL V-vial placed in the reactor. The Chromafix® PS-HCO3 18F cartridge was pre-
conditioned with 10 mL of 0.06 M K2C2O4 solution, 10 mL of H2O and 5 mL of air. The so-
lution was dried with two cycles of azeotropic drying with acetonitrile (200 μl) under a flow
of helium at 100 °C. The 3 mL vial containing the dried [18F]KF/K222 complex was added a
solution of 4 (10 mg, 0.016 mmol) and Cu(OTf)2(py)4 (15 mg, 0.022 mmol) in anhydrous
DMA (400 μL) and purged with 10 mL of air using a syringe. The mixture was heated at
120 °C for 20 min, after which, the reaction mixture was added 1.3 mL phosphate buffer (pH
7.4), and heated at 100 °C for 10 minutes. The crude mixture was then diluted with 1.3 mL of
40 % acetonitrile in water and loaded directly onto a HPLC loop and injected on a semi-prep
HPLC column (Phenomenex Onyx 100x10 mm, 4 mL/min) using a gradient from 10 to 40 %
EtOH in water (0.05 % TFA) over 20 minutes, monitored with UV- and radiodetector. An
aliquot from the fraction collected around 20 minutes was analysed by analytical HPLC for
product identity and molar activity measurements. Molar activity was calculated by compar-
ing the UV absorption to a calibration curve of cabozantinib. Radio-TLC was used for further
radiochemical purity measurements.
References
1.
2.
3.
4.
Lemmon MA, Schlessinger J. Cell signaling by receptor-tyrosine kinases. Cell.
2010;141:1117-1134.
Zhang J, Yang PL, Gray NS. Targeting cancer with small molecule kinase inhibitors.
Nat Rev. 2009;9:28-39.
Wu P, Nielsen TE, Clausen MH. FDA-approved small-molecule kinase inhibitors.
Trends Pharmacol Sci. 2015;36:422-439.
U.S. Food and Drug Administration. Approved kinase inhibitors.
https://www.fda.gov/Drugs/InformationOnDrugs/ApprovedDrugs/ucm279174.htm.
Published 2017.
5.
6.
7.
Sierra JR, Tsao M-S. c-MET as a potential therapeutic target and biomarker in cancer.
Ther Adv Med Oncol. 2011;3:S21-S35.
Ma PC, Tretiakova MS, Mackinnon AC, et al. Expression and Mutational Analysis of
MET in Human Solid Cancers. Genes, Chromosom cancer. 2008;47:1025-1037.
Underiner TL, Herbertz T, Miknyoczki SJ. Discovery of small molecule c-Met
inhibitors: Evolution and profiles of clinical candidates. Anticancer Agents Med Chem.
2010;10:7-27.
8.
9.
Bernard-Gauthier V, Bailey J, Berke S, Schirrmacher R. Recent Advances in the
Development and Application of Radiolabeled Kinase Inhibitors for PET Imaging.
Molecules. 2015;20:22000-22027.
Slobbe P, Poot AJ, Windhorst AD, Van Dongen G a MS. PET imaging with small-
molecule tyrosine kinase inhibitors: TKI-PET. Drug Discov Today. 2012;17(21-
22):1175-1187. doi:10.1016/j.drudis.2012.06.016.
10. Wu C, Tang Z, Fan W, et al. In vivo positron emission tomography (PET) imaging of
This article is protected by copyright. All rights reserved.

mesenchymal-epithelial transition (MET) receptor. J Med Chem. 2010;53:139-146.
11. Pool M, van Scheltinga AGT, Kol A, Giesen D, de Vries EGE, Lub-de Hooge MN.
89Zr-Onartuzumab PET imaging of c-MET receptor dynamics. Eur J Nucl Med Mol
Imaging. 2017:1-9.
12. Luo H, Hong H, Slater MR, et al. PET of c-Met in cancer with 64Cu-labeled
Hepatocyte Growth Factor. J Nucl Med. 2015;56:758-763.
13. Xu W, Donnelly DJ, Chow PL, Henley BJ. Method of preparing fluorine-18 labeled
cabozantinib and its analogs. 2016:WO 2016/019285 A1.
14. Preshlock S, Tredwell M, Gouverneur V. 18F‑Labeling of Arenes and Heteroarenes for
Applications in Positron Emission Tomography. Chem Rev. 2016;116:719-766.
15. Tredwell M, Preshlock SM, Taylor NJ, et al. A General Copper-Mediated Nucleophilic
18F Fluorination of Arenes. Angew Chemie Int Ed. 2014;53:7751-7755.
16. Preshlock S, Calderwood S, Verhoog S, et al. Enhanced copper-mediated 18F-
fluorination of aryl boronic esters provides eight radiotracers for PET applications.
Chem Commun. 2016;52:8361-8364.
17. Zhang Z, Zhang C, Lau J, Colpo N, Bénard F, Lin K-S. One-step synthesis of 4-
[18F]fluorobenzyltriphenylphosphonium cation for imaging with positron emission
tomography. J Label Compd Radiopharm. 2016;59:467-471.
18. Zlatopolskiy BD, Zischler J, Krapf P, et al. Copper-mediated aromatic
radiofluorination revisited: efficient production of PET tracers on a preparative scale.
Chem - A Eur J. 2015;21:5972-5979.
19. Zhang Z, Lau J, Kuo H, et al. Synthesis and evaluation of 18F-labeled CJ-042794 for
imaging prostanoid EP4 receptor expression in cancer with positron emission
tomography. Bioorg Med Chem Lett. 2017;27:2094-2098.
20. Achilli C, Ciana A, Fagnoni M, Balduini C, Minetti G. Susceptibility to hydrolysis of
phenylboronic pinacol esters at physiological pH. Cent Eur J Chem. 2013;11:137-139.
21. Xu J, Duran D, Mao B. On-Column Hydrolysis Kinetics Determination of Boronic
Pinacol Ester Intermediates for Use in Optimization of Fast HPLC Methods. J Liq
Chromatogr Relat Technol. 2006;29:2795-2806.
22. Zhan Z, Ai J, Liu Q, et al. Discovery of Anilinopyrimidines as Dual Inhibitors of c-Met
and VEGFR-2: Synthesis, SAR, and Cellular Activity. ACS Med Chem Lett.
2014;5:673-678.
This article is protected by copyright. All rights reserved.

Figure 1: The structure of crizotinib and cabozantinib.
Figure 2: Cabozantinib with its two sites for isotopic labelling with either carbon-11 or fluorine-18.
This article is protected by copyright. All rights reserved.

Scheme 1: The synthesis of precursor 4. i) NEt3, SOCl2, THF, rt, 99 %; ii) Cs2CO3, DMF, MeCN, 20
h, 55 ºC, 33 %; iii) Fe, NH4Cl, EtOH, H2O, 3 h, 70 ºC, 83 %; iv) 1, HATU, DIPEA, DMF, rt, 20 h,
50 %.
This article is protected by copyright. All rights reserved.

Scheme 2: a) The radiochemical reaction yielding [18F]cabozantinib from 4. b) observed hydrolysis
during purification produced the boronic acid 5.
This article is protected by copyright. All rights reserved.

Figure 3: HPLC chromatograms of the purified [18F]cabozantinib; a) Radioactive trace, with
[18F]cabozantinib at tR = 5.8 min; b) UV-trace. The impurity from boronic acid 5 can be seen eluting at
tR = 4.6 min.
This article is protected by copyright. All rights reserved.