A microwave radiosynthesis of the 4-[18F]-fluorobenzyltriphenylphosphonium ion

Article abstract of DOI:10.1002/jlcr.3241

The 4-[¹⁸F]-fluorobenzyltriphenylphosphonium cation was synthesized by a series of microwave reactions from no carrier added [¹⁸F]-fluoride. The microwave procedure reduced the quantity of reagents used and synthesis time when compared with the original synthesis. In addition, problematic solid phase extraction, sodium borohydride reduction by column and inconsistent yields with excessive precipitate formation during the bromination step were eliminated. The 4-[¹⁸F]-fluorobenzyltriphenylphosphonium cation was produced radiochemically pure in 8.3% yield with a specific radioactivity of 534.5 ± 371.4 GBq/μmole at end of synthesis.

Full text of DOI:10.1002/jlcr.3241

Research Article
Received 20 June 2014,
Revised 23 September 2014,
Accepted 24 September 2014
Published online 20 October 2014 in Wiley Online Library
(wileyonlinelibrary.com) DOI: 10.1002/jlcr.3241
18
A microwave radiosynthesis of the 4-[ F]-
fluorobenzyltriphenylphosphonium ion†
*
Hayden T. Ravert, Daniel P. Holt, and Robert F. Dannals
The 4-[18F]-fluorobenzyltriphenylphosphonium cation was synthesized by a series of microwave reactions from no carrier
added [18F]-fluoride. The microwave procedure reduced the quantity of reagents used and synthesis time when compared
with the original synthesis. In addition, problematic solid phase extraction, sodium borohydride reduction by column and
inconsistent yields with excessive precipitate formation during the bromination step were eliminated. The 4-[18F]-
fluorobenzyltriphenylphosphonium cation was produced radiochemically pure in 8.3% yield with a specific radioactivity
of 534.5 371.4 GBq/μmole at end of synthesis.
Keywords: phosphonium; fluorine-18; positron emission tomography; microwave
The [18F]-fluoride is unloaded from the target directly onto an
ion exchange resin column. After trapping, the column is eluted
with an aqueous acetonitrile (MeCN) solution of potassium
carbonate (K2CO3) and Kryptofix into the 5 mL reaction vial.
Introduction
Recent studies using the lipophilic 4-[18F]-fluorobenzyltriphenyl
phosphonium cation (18F-FBnTP) indicate that the radiotracer is
useful for studying mitochondrial dysfunction in numerous diseases
The solution is heated (in the heating block) to dryness with
including cardiac and cancer.1–4 18F-FBnTP was initially synthesized
nitrogen flow. Two additional aliquots of MeCN are added with
by manual manipulation involving multiple steps including two solid
heating to ensure azoetropic removal of the water. The reaction
phase extraction (SPE) procedures, a problematic on-column
vial is remotely transferred to the microwave cavity and cooled to
reduction, and a cumbersome bromination step resulting in
room temperature before microwave irradiation. The precursor,
inconsistent yields with excessive precipitate formation and
4-trimethylammoniumbenzaldehyde trifluoromethanesulfonate, in
significant radiation exposure.5 All these factors were major
MeCN is added, and the solution is microwave irradiated in the
sealed vial at 50 W for 30 s to produce 4-[18F]fluorobenzaldehyde
deterrents to performing the radiosynthesis on a routine basis. To
greatly reduce exposure, an initial automated and remote synthesis
using a modified dual-run FDG synthesis module was developed
(unpublished work). The dual-run synthesis module procedure
(18F-FBnO). Aqueous sodium borohydride is added, and the mixture
is microwave irradiated at 25 W for 30 s to produce 4-[18F]
fluorobenzyl alcohol (18F-FBnOH). Next, aqueous hydrobromic acid
following the original synthesis also proved problematic, frequently
(HBr) is added. The HBr solution is microwave irradiated at 25 W
for 60 s. The 4-[18F]fluorobenzyl bromide (18F-FBnBr) solution is
requiring regular manual intervention to avert a synthesis failure. In
order to synthesize 18F-FBnTP for routine research studies, a modified
immediately eluted by nitrogen pressure through a Waters Oasis
reaction sequence that could be automated and would reduce the
HLB Plus (Oasis SPE) cartridge. The reaction vial is rinsed with water
problems with the manual and module radiosynthesis was sought.
and washed quickly through the SPE. The Oasis SPE cartridge is
then eluted with an MeCN/acetic acid (HOAc) mixture into a second
5 mL reaction vessel containing triphenylphosphine (TP), and the
solution is thermally heated. After cooling and adding water, a
precipitate (triphenylphosphine oxide) forms. The mixture with
Results and discussion
A microwave radiosynthesis6 that greatly simplified the original
production of 18F-FBnTP involving a similar four-step synthesis
precipitate is filtered and injected directly onto the semi-preparative
HPLC column. The 18F-FBnTP product peak is collected into a
was developed (Figure 1). A hardware system (Figure 2) was built
to accommodate specifically the 18F-FBnTP radiosynthesis but is
reservoir containing HPLC water. The diluted product is loaded onto
adaptable to other fluorine-18 radiosyntheses. The custom
hardware contains a heating block, custom microwave cavity,
two Tecan Cavro syringe pumps, a multi-port cap constructed
for standard 5 mL v-vials, and valved reagent addition vials. The
synthesis module is controlled by a National Instruments
Compact Fieldpoint module linked to a laptop computer running
Labview Real-Time software (window examples are shown
Figure 3). The entire hardware–software system allows for the
complete control of every synthetic step from the collection of
the [18F]-fluoride to the injection of the reaction mixture onto
the semi-preparative HPLC. The general radiosynthetic
procedure for 18F-FBnTP is described in the succeeding texts.
a Waters Oasis HLB Plus cartridge. An automated SPE reformulation
Division of Nuclear Medicine, Department of Radiology, Johns Hopkins University
School of Medicine, 600 North Wolfe Street, Nelson B1-152, Baltimore, Maryland
21287, USA
*Correspondence to: Hayden T. Ravert, Division of Nuclear Medicine, Department
of Radiology, Johns Hopkins University School of Medicine, 600 North Wolfe
Street, Nelson B1-152, Baltimore, Maryland 21287, USA.
E-mail: htr@jhu.edu
†Additional supplementary material can be found in the online version of this
article at the publisher’s web-site.
J. Label Compd. Radiopharm 2014, 57 695–698
Copyright © 2014 John Wiley & Sons, Ltd.

H. T. Ravert et al.
Figure 1. 4-[18F]-Fluorobenzyltriphenylphosphonium cation microwave radiosynthesis.
to remove the acetonitrile prior to bromination resulted in a
substantial loss of 18F-FBnOH during the evaporation. Other
reaction solvents tested proved incompatible with either the
fluorination, reduction or bromination step.
A second issue was observed during the isolation of the 18F-FBnBr
on the SPE. HPLC analysis of the isolated 18F-FBnBr from the Oasis SPE
showed a large hydrophilic peak present (shown to be 18F-FBnOH).
Analysis prior to the SPE indicated little 18F-FBnOH. The SPE water
wash to remove the HBr apparently left a small amount of acidic water
on the SPE. The quantity was sufficient to hydrolyze the 18F-FBnBr
back to the 18F-FBnOH. The water wash could not be eliminated
because the majority of the aqueous HBr needed to be removed to
allow the reaction with a small quantity of triphenylphosphine
(3 mg) with a small amount of precipitate. By performing the wash
and elution of the SPE cartridge rapidly, the conversion back to
18F-FBnOH was significantly reduced (but not eliminated).
An HBr solution in acetic acid (HBr/HOAc) was examined in an effort
to prevent the hydrolysis of the 18F-FBnBr. The HBr/HOAc reaction
gave inconsistent results. As noted in the preceding texts, large
Figure 2. Custom automated microwave system.
includes a wash of the SPE with HPLC water, elution of the product amounts of the acid were needed to complete the reaction as a result
with ethanol and sterile saline through a sterile filter to a final of the acetonitrile. In addition, the 18F-FBnBr in the HBr/HOAc solution
product vial. The preparative and final product chromatograms (even diluted with water) could not be retained on the SPE. Finally, the
are presented in the Supporting Information.
simple addition of TP to the 18F-FBnBr/HBr/HOAc reaction mixture
The new microwave procedures allowed for the significant required a large quantity of TP (21 mg) to complete the reaction.
reduction in quantity and volume of precursor and other The excessive precipitate created was impossible to remove prior
reagents.6 Problems caused by larger volumes and precipitate to injection onto the semi-preparative column.
solids were virtually eliminated. During the optimization of the
A third issue (radiolysis) was noted when the final 18F-FBnTP
microwave syntheses, yields of each radiosynthetic step (n=3) tracer concentration was near or above 8 mCi/mL. Although not
were determined by analytical HPLC. All chromatograms are observed during initial quality control, a reduction in the
displayed in the Supporting Information. The observed (isolated radiochemical purity, below 80%, was noted during the 4-h stability
yields were not determined) HPLC yields were 57% for the check at the higher tracer concentrations. By reducing the starting
fluorination reaction, 95% for the reduction reaction, 85% for [18F]-fluoride quantity and thereby reducing amount of radiotracer
the bromination reaction, and 95% for the alkylation reaction. synthesized and concentration in solution to less than 8 mCi/mL,
These results are comparable to those observed in the manual the problem was resolved. The addition of an antioxidant to the
and module radiosyntheses5; however, the microwave procedures final radiotracer solution might also reduce radiolysis.
reduced the total radiosynthesis time by 40 min with an increase in
specific radioactivity and radiochemical yield.
During the development of the modified synthesis, some issues
Experimental
[
18F]-Fluoride was produced by 18 MeV proton bombardment of a high
were noted with the system and procedure. Initial bromination
reaction attempts used approximately 9 to 18-fold excess (880 to
1760 μmols HBr, 0.1 to 0.2 mL) of aqueous 48% HBr compared with
a total of sodium borohydride and potassium carbonate (nearly
100 μmols as unreacted base) added to the reaction vial. Little to
no reaction was observed. A larger excess, 0.8 mL (7070 μmols) of
HBr, did provide an almost complete conversion of the 18F-FBnOH
to 18F-FBnBr as observed by HPLC. Another confounding factor
was that acetonitrile reacts with the aqueous HBr. A large excess
pressure [18O]-water target using a GE PETtrace cyclotron. All reagents were
of ACS or HPLC purity. The precursor, 4-trimethylammoniumbenzaldehyde
trifluoromethanesulfonate, was purchased from ABX GmbH. Authentic
product, 4-fluorobenzyltriphenylphosphonium bromide, previously synthesized,5
was used as the standard to determine the identity, purity and specific
radioactivity of the final product. The semi-preparative chromatography
system consisted of an Agilent 1260 Prep pump with a VICI injector, a
Knauer 200 UV Detector (254 nm), a Bioscan Hot Cell interface with a
diode radioactivity detector. The semi-preparative HPLC column was a
of HBr reacted with both the acetonitrile and 18F-FBnOH. Attempts Waters XBridge C18 10 μm 10 × 150 mm column eluted with a mixture
www.jlcr.org
Copyright © 2014 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2014, 57 695–698

H. T. Ravert et al.
Figure 3. The software interface for the automated microwave system.
of 40:60 methanol : pH 3.2 triethylamine/phosphoric acid buffer at flow (9.5 mg K2CO3 :48mg Kryptofix in 0.6 mL of 50% aqueous MeCN) into a
rate of 10 mL/min. The analytical chromatography system included an
5 mL vial with a multi-port cap. After rinsing the cartridge with 0.25 mL MeCN,
Agilent 1260 Infinity System incorporating a quaternary pump, HiP ALS
the solution was dried at 110°C with controlled nitrogen flow (325 mL/min) for
autosampler, and DAD UV detector with a Max-Light flow cell set to 150 s. Two separate additions of 0.25 mL of MeCN were heated for 150 and
254 nm plus a Bioscan Flow-Count interface with a NaI radioactivity 180 s respectively. The vial was transferred remotely to the microwave cavity,
detector The analytical HPLC column was an Xbridge C18 3.5 μm cooled for 2 min using compressed air (flow approx. 6 L per minute).
4.6 × 100 mm column eluted with a mixture of 35:65 acetonitrile : pH 3.2 4-Trimethylammoniumbenzaldehyde trifluoromethanesulfonate (12.8 μmols,
triethylamine/phosphoric acid buffer at flow rate of 2 mL/min. 4 mg in 0.4 mL MeCN) was added from a valved reservoir attached through
a
Chromatographic data were acquired and analyzed on an Agilent OpenLAB the multi-port cap. The solution was microwave irradiated at 50 W for 30 s.
chromatography data system (Rev. A.04.02). Radioactivity measurements Next, sodium borohydride (31.7 μmols, 1.2 mg in 0.2 mL HPLC water) was
were made using a Capintec CRC-15R or Comcer model TALETE HC dose added from a valved reservoir, and microwave irradiation applied at 25 W for
calibrator. Synthesis automation equipment included a custom synthesis 30 s. Through a third valved reservoir, 0.8 mL of aqueous HBr (48%) was added
module controlled by National Instruments LABView (2009 SP1) software and the solution was microwave irradiated at 25 W for 60 s. The hot solution
and a National Instruments Compact Fieldpoint (CFP2110) system. The was rapidly eluted through a Waters Oasis HLB Plus LP cartridge using positive
microwave system was a Resonance Instruments Model 521 fitted with a pressure of nitrogen gas followed by 15 s of drying at approx. 125 mL/min. The
Model 521-5 reactor cavity for 5 mL v-vial modified for air cooling.
vessel was rinsed with 1 mL of water, and the solution eluted and dried in the
same manner. The Oasis cartridge was eluted with a mixture of MeCN (0.9 mL),
and glacial acetic acid (0.1 mL) into a new 5 mL reaction v-vial containing
triphenylphosphine (11.4 μmols, 3 mg). The vessel was heated to 100°C for
5 min followed by cooling for 2 min. HPLC water (1 mL) was added to the vial,
and the solution was injected onto the semi-preparative column through a
0.45 μm Teflon filter (17 mm diameter). The product, 18F-FBnTP (RT = 10.3 min,
k′= 6.2), was collected in 50 mL HPLC water for SPE reformulation. The water
Microwave radiosynthesis and purification of 4-[18F]-
fluorobenzyltriphenylphosphonium cation
Aqueous [18F]-fluoride was trapped on a Chromafix 30-PS-HCO3 cartridge
(ABX). The cartridge was eluted with 0.15 mL of a K2CO3 : Kryptofix solution
J. Label Compd. Radiopharm 2014, 57 695–698
Copyright © 2014 John Wiley & Sons, Ltd.
www.jlcr.org

H. T. Ravert et al.
solution was eluted through a Waters Oasis HLB Plus LP cartridge, and the
cartridge was washed with 10 mL saline. The cartridge was eluted with 1 mL
Conflict of Interest
absolute ethanol followed by 10 mL sterile saline through a 0.2 μm sterile The authors did not report any conflict of interest.
Millipore FG filter (25 mm) into a sterile vial containing 4 mL sterile saline. An
aliquot was removed for quality control analysis. Analytical HPLC was
performed to determine radiochemical identity (RT = 4.4 min, k′= 7.8), purity,
References
and specific activity.
[1] I. Madar, H. T. Ravert, Y. Du, J. Hilton, L. Volokh, R. F. Dannals, J. J. Frost,
J. M. Hare, J. Nucl. Med., 2006, 47, 1359.
[2] I. Madar, H. T. Ravert, A. DiPaula, Y. Du, R. F. Dannals, L. Becker, J. Nucl.
Med., 2007, 48, 1021.
[3] I. Madar, H. Ravert, B. Nelkin, M. Abro, M. Pomper, R. Dannals, J. J.
Frost, Eur. J. Nucl. Med. Mol. Im., 2007, 34, 2057.
[4] I. Madar, Y. Huang, H. Ravert, S. L. Dalrymple, N. E. Davidson, J. T.
Isaacs, R. F. Dannals, J. J. Frost, J. Nuc. Med., 2009, 50, 774.
[5] H. T. Ravert, I. Madar, R. F. Dannals, J. Label Compds. Radiopharm.,
2004, 47, 469.
[6] S. Stone-Elander, N. Elander, J. O. Thorell, A. Fredriksson. Ernst Schering
Res. Found. Workshop, 2007, 62, 243.
Conclusion
The microwave radiosynthesis of 18F-FBnTP provided an automated
method for the radiotracer production that simplified and eliminated
steps of the original synthesis that were burdensome. The new
radiosynthesis allowed the radiofluorination, reduction and
bromination to occur in one pot, reduced the need for large
quantities of reagents and changed an unreliable bromination step
into a reliable and consistent procedure. A radiochemically pure
product (>99%) was obtained in 8.3 2.4% yield (end of synthesis)
with an average synthesis time of 52.4 14 min (n=27) and an
average specific radioactivity of 534.5 371.4 GBq/μmole (14,446
10,039 mCi/μmole end of synthesis).
Supporting Information
Additional supporting information may be found in the online
version of this article at the publisher’s web-site.
www.jlcr.org
Copyright © 2014 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2014, 57 695–698