Radiosynthesis of 4-[18F]fluoromethyl-L-phenylalanine and [ 18F]FET via a same strategy and automated synthesis module

Article abstract of DOI:10.1002/jlcr.1701

Currently there is still a need for more potent amino acid analogues as tumour imaging agents for peripheral tumour imaging with PET as it was recently reported that the success of O-(2'-[¹⁸F]fluoroethyl)-L-tyrosine ([¹⁸F]FET) is limited to brain, head and neck tumours. As the earlier described 2-Amino-3-(2-[¹⁸F]fluoromethyl-phenyl)-propionic acid (2-[¹⁸F]FMP) suffered from intramolecular-catalysed defluorination, we synthesized 2-Amino-3-(4-[¹⁸F]fluoromethyl-phenyl)-propionic acid (4-[¹⁸F]FMP) as an alternative for tumour imaging with PET. Radiosynthesis of 4-[¹⁸F]FMP, based on Br for [¹⁸F] aliphatic nucleophilic exchange, was performed with a customized modular Scintomics automatic synthesis hotboxthree system in a high overall yield of 30% and with a radiochemical purity of gt 99%. 4-[¹⁸F]FMP was found to be stable in its radiopharmaceutical formulation, even at high radioactivity concentrations. Additionally, for a comparative study, [¹⁸F]FET was synthesized using the same setup in 40% overall yield, with a radiochemical purity gt 99%. The described automated radiosynthesis allows the production of two different amino acid analogues with minor alternations to the parameter settings of the automated system, rendering this unit versatile for both research and clinical practice. Copyright

Full text of DOI:10.1002/jlcr.1701

Research Article
Received 29 July 2009,
Revised 20 October 2009,
Accepted 23 October 2009
Published online 30 November 2009 in Wiley Interscience
(www.interscience.wiley.com) DOI: 10.1002/jlcr.1701
Radiosynthesis of 4-[18F]fluoromethyl-
L-phenylalanine and [18F]FET via a same
strategy and automated synthesis module
Ã
Ken Kersemans,a John Mertens,a and Vicky Caveliersb
Currently there is still a need for more potent amino acid analogues as tumour imaging agents for peripheral tumour
imaging with PET as it was recently reported that the success of O-(2’-[18F]fluoroethyl)-L-tyrosine ([18F]FET) is limited to
brain, head and neck tumours. As the earlier described 2-Amino-3-(2-[18F]fluoromethyl-phenyl)-propionic acid (2-[18F]FMP)
suffered from intramolecular-catalysed defluorination, we synthesized 2-Amino-3-(4-[18F]fluoromethyl-phenyl)-propionic
acid (4-[18F]FMP) as an alternative for tumour imaging with PET. Radiosynthesis of 4-[18F]FMP, based on Br for [18F]
aliphatic nucleophilic exchange, was performed with a customized modular Scintomics automatic synthesis hotboxthree
system in a high overall yield of 30% and with a radiochemical purity of \gt 99%. 4-[18F]FMP was found to be stable in its
radiopharmaceutical formulation, even at high radioactivity concentrations. Additionally, for a comparative study, [18F]FET
was synthesized using the same setup in 40% overall yield, with a radiochemical purity \gt 99%. The described automated
radiosynthesis allows the production of two different amino acid analogues with minor alternations to the parameter
settings of the automated system, rendering this unit versatile for both research and clinical practice.
Keywords: 4-[18F]fluoromethyl-L-phenylalanine; radiosynthesis; automation; PET
study, meanwhile illustrating the versatility of the described
system.
Introduction
At present, the world wide best established fluorinated amino
acid analogue for brain tumour imaging with PET is O-(2’-
[18F]fluoroethyl)-L-tyrosine ([18F]FET).1,2 However, its success was
limited to brain, head and neck tumours as [18F]FET did not
seem to accumulate significantly in other peripheral tumours.3,4
This means that more potent tumour-specific radiofluorinated
amino acid analogues, which can be synthesized for routine
clinical use are still required.
Results and discussion
Synthesis of the precursor and the nonradioactive reference
4-FMP
The ester formation of commercially available N-Boc-4-methyl-l-
phenylalanine with tert-butyl 2,2,2-trichloroacetimidate by the
described method yielded L-4-tert-butoxycarbonylamino-3-o-
tolyl-propionic acid tert-butyl ester in 99% yield.
Radical bromination, using N-Bromosuccinimide and azobisi-
sobutyronitrile as a radical initiator, yielded 66% of 3-(4-
Bromomethyl-phenyl)-2-l-tert-butoxycarbonylamino-propionic
acid tert-butyl ester. The main side-products of the synthesis of 3-
(4-Bromomethyl-phenyl)-2-l-tert-butoxycarbonylamino-propionic
acid tert-butyl ester were di- and tri-brominated species. The
amounts of side products could be limited using dichloro-
methane as solvent instead of the generally used CCl4.
2-Amino-3-(2-[18F]fluoromethyl-phenyl)-propionic acid (2-
[18F]FMP) described earlier by our group had good tumour
targeting properties but suffers from radiodefluorination in
aqueous solutions as well as in vivo, related to an intra-molecular
interaction of the –CH2–F group on the 2-position of the aromatic
ring with the amino acid ammonium entity, leading to rapid
hydrolysis.5–8 Placing the fluoromethyl side chain in the para
position avoids the unfavourable interaction with the ammonium
entity as the intra molecular distance between both groups is too
large to allow interaction. Therefore, it was decided to synthesize
2-Amino-3-(4-[18F]fluoromethyl-phenyl)-propionic acid (4-[18F]-
fluoromethyl-L-phenylalanine or 4-[18F]FMP, Figure 1), as a new
and more stable alternative to 2-[18F]FMP.
The radiosynthesis was automated using a customized
modular Scintomics automatic synthesis hotboxthree system to
allow routine productions of 4-[18F]FMP for the clinical evalua-
tion and application of the compound for PET acquisition of
peripheral tumours but also different types of brain tumours,
linked to blood-brain barrier damage. Furthermore [18F]FET was
synthesisd on the same automatic system for a comparative
aRadiopharmaceutical Chemistry, BEFY, Vrije Universiteit Brussel, Laarbeeklaan
103, 1090 Brussels, Belgium
bLaboratory of In Vivo Cellular and Molecular Imaging, Vrije Universiteit Brussel,
Laarbeeklaan 103, Brussels 1090, Belgium
*Correspondence to: Ken Kersemans, Radiopharmaceutical Chemistry, BEFY,
Vrije Universiteit Brussel, Laarbeeklaan 103, 1090 Brussels, Belgium.
E-mail: kkersema@vub.ac.be
J. Label Compd. Radiopharm 2010, 53 58–62
Copyright r 2009 John Wiley & Sons, Ltd.

K. Kersemans et al.
For the purpose of comparative studies in the clinic, we also
synthesized [18F]FET using the same experimental setup.
Although a settled method for the synthesis of [18F]FET is used
by several other groups,9 for practical reasons we opted to apply
our custom synthesis setup for the production of [18F]FET as this
allowed us to synthesize both molecules on the same module
without the need for an extra experimental setup. The principle
of the labelling reaction is the same with some different time
and temperature settings.
O
OH
H2N
H
H
H
H
H
18
For purification only minor alternations were required as only
the mobile phase for semi preparative HPLC needed to be
adjusted, because of the difference in lipofilicity of fully
protected [18F]FET and 4-[18F]FMP. The reaction time (100 min)
using our method is somewhat longer in comparison to that of
Hamacher et al.9
In the described conditions [18F]FET was synthesized in 40%
yield (comparable to the method described by Hamacher et al.9)
with a radiochemical purity of at least 99%, determined by HPLC
as well as by TLC and was enantiomerically pure.
The radiopharmaceutical formulations of 4-[18F]FMP and
[18F]FET were tested for the presence of residual AcN and
Kryptofix2.2.2 and were both found to be smaller than
1.10ꢀ3 mg/mL, which is far below the limits of the European
Pharmacopee (0.44 mg/mL for Kryptofix2.2.2 and 0.8 mg/mL for
AcN), proving that the described method allows the production
of high quality 4-[18F]FMP and [18F]FET.
As discussed above, the described setup allows for both
[18F]FET and 4-[18F]FMP to be synthesized on the same
commercially available module (Scintomics hotboxthree), without
the requirement of changing any hardware on the module,
rendering it a versatile unit, well suited for both research and
clinical routine.
F
Figure 1. The structure of 4-[18F]FMP.
The AgF based nonradioactive fluorination coupled to column
chromatography yielded 65% of pure 3-(4-fluoromethyl-phenyl)-
2-l-tert-butoxycarbonylamino-propionic acid tert-butyl ester. The
yield of the deprotection reaction was 87%, allowing the
recovery of pure 4-FMP with an overall yield of 37%.
Fluoride acting as a strong base in the fluorination reaction
did not cause racemisation of the L-a-carbon as Chiral HPLC
showed only 4-fluoromethyl-L-phenylalanine to be present.
Radiosynthesis of 4-[18F]FMP and [18F]FET
The radiolabelling of the precursor with [18F]fluoride provided
Boc-4-[18F]fluoromethyl-L-phenylalanine-tButyl ester with a re-
producible radiochemical yield of at least 90%. It is important
that the volume of AcN during the labelling exceeds the initial
volume of the aqueous [18F]fluoride solution, recovered from
the QMA, as only under these conditions an optimal contact
(and optimal reaction) with the dry [18F]fluorine layer on the wall
of the vial could be guaranteed. Furthermore, the reaction
should not be continued for longer than 5 min as from this point
loss of the Boc-protecting group was observed, leading to a
lower overall yield as only fully protected compounds are
collected in the subsequent HPLC purification step.
The deprotection step proved to be the most critical step in the
whole procedure as extensive defluorination is observed if water is
present in the deprotection medium. Hence, it is of major
importance to 11 dry online the acetonitrile, here using the tandem
minicolumn system (silicagel cartridge-CaCl2 minicolumn) and 21 to
use fresh and dry trifluoroacetic acid and dry dichloromethane. The
loss of free [18F]fluoride that is still observed is due to small traces of
water still present, despite the drying procedures, as experiments
with dry solvents showed a loss of only 7% of radiofluoride in the
same reaction conditions. Therefore, we opted to use 5 mL aliquots
of trifluoroacetic acid (Sigma Aldrich), rather than the larger
available quantities which lost the required properties rapidly over
time. We suspect that the TFA not only accumulates water but also
seems to dissolve some components of the septum of the
container, as a film like residu is observed on the wall of the vial
when evaporating ‘older’ TFA/DCM: 1/4 (v/v) mixtures.
Efforts are currently being made to reduce the overall
synthesis time by switching to a more convenient one-pot
synthesis and deprotection, without the need of a preconcen-
tration step from an aqueous medium thus avoiding the
presence of traces of water and with the HPLC purification at
the end.
Shelf life stability
Moving the CH2F– group from ortho to para position in
phenylalanine decreased considerably the hydrolysis of the
noncarrier-added [18F]fluoromethyl-L-phenylalanine analogue,
confirming our hypothesis that the fast radio-defluorination of
the noncarrier-added 2-[18F]FMP was caused by an intramole-
cular interaction of the CH2F– on the 2-position and amino acid
amine group.6 The activity concentration of the noncarrier-
added 4-[18F]FMP in the final formulation (pH 6, ethanol/0.9%
NaCl solution: 1/99 (v/v)) was generally about 2.5 GBq/5 mL. At
room temperature after 3 h at least 99.5% of the parent
compound was still present. In case of [18F]FET the amount of
[18F]fluoride was below the detection limit of the radio-HPLC
method.
The use of chloroform instead of dichloromethane, as proposed
by Hamacher (private communication), did not decrease defluor-
ination as expected, certainly because in our conditions the
physico-chemical properties of the solvent are of less influence
than the presence of traces of water in the deprotection mixture.
The described radiosynthesis procedure resulted in an overall
yield of about 30% noncarrier-added 4-[18F]FMP within 120 min
and a radiochemical purity of at least 99% as determined
by HPLC and TLC. Chiral HPLC showed only the 4-[18F]FMP
to be present.
This means that in our case 1% of ethanol reduces sufficiently
the radiolytic decomposition, even of the fluorobenzyl analogue
that could be suspected to undergo more radiolytic hydrolysis.
Materials and methods
All the conventional products mentioned were at least analytical
grade. The solvents were of HPLC quality or better. All NMR data
J. Label Compd. Radiopharm 2010, 53 58–62
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org

K. Kersemans et al.
were obtained on an AVANCE DRX 250 instrument and all MS sonication for 10min and the suspension stored at ꢀ201C
data on a Fisons VG II Quattro Mass Spectrometer.
overnight. The solvents were removed by filtration and the
2-Amino-3-(4-fluoromethyl-phenyl)-propionic acid (4FMP) was
obtained as a white solid (87%). 1H-NMR (CD3OD, 250 MHz):
7.26–7.37 (4H, m), 5.46 (1H, dd, J1 = 11.0 Hz, J2 = 19.0Hz,
J[19F-1H]= 45.0 Hz), 5.37 (1H, dd, J1 = 11.0 Hz, J2 = 19.0Hz,
J[19F-1H]= 45.0 Hz), 3.97 (1H, dd, J1 = 7.2 Hz, J2 = 14.2 Hz), 3.26
(1H, dd, J1 = 7.5 Hz, J2 = 14.2Hz), 3.10 (1H, dd, J1 = 7.5 Hz,
J2 = 14.2 Hz). MS (EI) m/z 198 (MH1).
Synthetic procedures
Synthesis of the precursor: 3-(4-bromomethyl-phenyl)-2-L-tert-
butoxycarbonyl-amino-propionic acid tert-butyl ester
The free carboxyl group of commercially available (Peptech
corp., USA) L-4-tert-butoxycarbonylamino-3-o-tolyl-propionic
acid (N-Boc-4-methyl-L-phenylalanine) (1 g, 3.58 mmol) is pro-
tected with a tert-butyl ester by reacting it with tert-butyl 2,2,2-
trichloroacetimidate (3.10 g, 14.19 mmol) in dichloromethane
(36 mL) at room temperature for 24 h. After silica gel column
chromatography (50 g Si-gel, id 2 ꢁ 25 cm, petroleumether/
diethylether: 2/8) 3-(4-methyl-phenyl)-2-L-tert-butoxycarbonyla-
mino-propionic acid tert-butyl ester was obtained as an
Radiosynthesis
The radiosynthesis has been automated using a customized
modular Scintomics hotboxthree system (Scintomics, Fu¨rstenfeld-
bruck, Germany). A standard Scintomics hotboxthree was chosen
as the core of the system, equipped with preparative Variopump
LC1, VarioHPLC, Variodetect and a PS1 powersupply unit. All
connections between the valves are made by means of the
standard Scintomics PEEK tubing. The radiosynthesis is summar-
ized in Figure 2 while Figure 3 shows the flow scheme of the
automated synthesis module.
1
amorphous white solid (99% yield). H-NMR (CD3OD, 250 MHz):
7.14–7.20 (4H, m), 4.53 (1H, dd, J1 = 10.0 Hz, J2 = 12.5 Hz), 3.11
(2H, dd, J1 = 6.5, J2 = 14.5 Hz), 2.41 (3H, s), 1.52 (9H, s), 1.51 (9H, s).
MS (EI) m/z 336 (MH1).
Radiosynthesis of 2-Amino-3-(4-[18F]fluoromethyl-phenyl)-propio-
nic acid (4-[18F]FMP)
For the radical bromination of the methyl side chain, 3-
(4-methyl-phenyl)-2-L-tert-butoxycarbonylamino-propionic acid
tert-butyl ester (500 mg, 1.49 mmol) was reacted with
N-bromosuccinimide (245 mg, 1.38 mmol), using azobisisobutyr-
onitrile (28 mg, 0.17 mmol) as a radical initiator, in dichlor-
omethane (50 mL) at 501C for 2 h. The crude product was
purified via silica gel column chromatography (50 g Si-gel, id
2 ꢁ 25 cm, petroleumether/diethylether: 1/9) to provide the 3-(4-
bromomethyl-phenyl)-2-L-tert-butoxycarbonylamino-propionic
acid tert-butyl ester as an amorphous white solid in 66% yield.
1H-NMR (CDCl3, 500 MHz): 7.42 (2H, d, J = 7.3), 7.26 (2H, d,
J = 7.8), 4.56 (2H, dd, J1 = 6.5, J2 = 18.7 Hz), 3.15 (2H, s), 1.53 (9H,
s), 1.50 (9H, s). MS (EI) m/z 413 (M1).
[18F]fluoride recovery and radiolabelling. About 9.2 GBq of
[18F]fluoride, produced with a CGR 560 multiple particle variable
energy AVR Cyclotron using the 18O (p,n) 18F nuclear reaction,
was separated from the 18O-enriched water by trapping on a
SepPak Light Waters AccellTM Plus QMA (Waters), preconditioned
with 10 mL of a 0.01 M solution of K2CO3 (Merck) and 20 mL of
de-ionized water. Elution of the activity was achieved with 0.5 mL
of an AcN/water (9:1) mixture containing 1 mg of K2CO3 (Merck)
and 10 mg of Kryptofix2.2.2 (Acros) into a 5 mL reaction vial
(Pierce) by means of a dropwise flowrate. This mixture was
azeotropically dried at 1401C (6 min) and the vial was subse-
quently cooled to 1201C by external airflow. Thereafter, 1 mL of
dry AcN containing 10 mmol (4 mg) of precursor 3-(4-bromo-
methyl-phenyl)-2-L-tert-butoxycarbonylamino-propionic acid tert-
butyl ester was added and N2 pressure was applied. The labelling
reaction was allowed to occur for 5 min, after which the vial was
cooled to room temperature. The AcN phase was transferred into
a second 5 mL conical glass vial containing 1 mL of 1 mM NaF
solution. The first vial was rinsed with another 0.5mL AcN that is
also transferred to the second vial.
Synthesis of the nonradioactive 2-Amino-3-(4-fluoromethyl-
phenyl)-propionic acid (4-fluoromethyl-L-phenylalanine or 4FMP)
In the first step, 3-(4-bromomethyl-phenyl)-2-L-tert-butoxycarbo-
nylamino-propionic acid tert-butyl ester (140 mg, 0.339 mmol)
was treated with an excess of AgF (180 mg, 1.42 mmol) in dry
acetonitrile (5 mL) for 3 h at 651C after which the crude product
was purified via silica gel column chromatography (50 g Si-gel,
id 2 ꢁ 25 cm) using a gradient of petroleum ether (40–601C) and
ethyl acetate.
Semi-preparative HPLC separation. The content of this vial
(AcN/H2O: 60/40 (v/v)) was mixed by nitrogen bubbling and
injected for semi-preparative RP-HPLC separation using an Apollo
1
The H-NMR and EIMS data for 3-(4-fluoromethyl-phenyl)-2-L-
tert-butoxycarbonylamino-propionic acid tert-butyl ester, ob-
1
tained as a white amorphous solid in 65% yield, were: H-NMR
(CD3OD, 250 MHz): 7.24–7.40 (4H, m), 5.58 (1H, dd, J1 = 11.0 Hz,
J2 = 19.0 Hz, J [19F-1H] = 47.8 Hz), 5.39 (1H, dd, J1 = 11.0 Hz, J2 =
19.3 Hz, J [19F-1H] = 47.8 Hz), 5.08 (1H, dd, J1 = 7.2 Hz, J2 =
14.2 Hz), 3.17 (1H, dd, J1 = 6.3 Hz, J2 = 14.0 Hz), 3.06 (1H, dd,
J1 = 7.5 Hz, J2 = 14.3 Hz) 1.40 (18H, s). MS (EI) m/z 354 (MH1).
The final deprotected fluorinated amino acid was obtained
after deprotection of 3-(4-fluoromethyl-phenyl)-2-L-tert-butoxycar-
bonylamino-propionic acid tert-butyl ester (200mg, 0.541 mmol)
in a dichloromethane/trifluoroactic acid: 1/1 mixture (20mL) for
90min at room temperature. After reaction, the solvents were
removed by rotatory evaporation. The residue was suspended in
5 mL dichloromethane/ethanol: 1/1, stirred and subsequently
evaporated by a gentle N2 flow. This process was repeated twice.
The residue was suspended in hexane/dichloromethane: 4/1 by
O
OH
O
O
O
O
O
O
H
H
H2N
H
H
O
N
H
H
O
N
H
H
H
H
H
B,C,D
A
H
18
H
H
18
H
H
H
F
F
Br
Figure 2. Summarized radiosynthesis of 4-[18F]FMP: (A) 5 min, 1201C, AcN, K222
,
K2CO3, 18Fꢀ; (B) semipreparative HPLC: Apollo C18 250 ꢁ 10 mm (5 m) column, AcN/
H2O: 65/35 (v/v) 6 mL/min; (C) preconcentration and elution from C18 minicolumn;
(D) Deprotection: TFA/DCM: 1/4 (v/v), 501C, 20 min. Final elution over tandem C18
and Al2O3 minicolumns.
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2010, 53 58–62

K. Kersemans et al.
Figure 3. Flow scheme of the custom scintomics synthesis module setup for the radiosynthesis of 4-[18F]FMP. R1: 0.5 mL AcN/water: 9/1 (v/v) containing 1 mg of K2CO3
and 10 mg of Kryptofix2.2.2.; R2: 500 mL dry AcN ; R3: 500 mL AcN; R4: 1 mL AcN containing 4 mg of 3-(4-bromomethyl-phenyl)-2-L-tert-butoxycarbonylamino-propionic acid
tert-butyl ester; R7: 1 mL AcN; R10: 1 ml dichloromethane/trifluoroacetic acid: 4/1 (v/v); R11: 1 mL dichloromethane; R12:, 5 mL of a 1% Ethanol/water (v/v).
[18F]fluoride recovery and radiolabelling. [18F]fluoride recovery
C18 250 ꢁ 10 mm (5 m) column and AcN/water: 65/35 (v/v) as
and drying was performed in exactly the same way as for 4FMP.
Subsequently 1 mL of dry AcN containing 10 mmol (7 mg of
precursor: L-Tyrosine, O-[2-[[(4-methylphenyl)sulfonyl]oxy]ethyl]-
N-(triphenylmethyl)-, 1,1-dimethylethyl ester) (ABX, Germany)
was added to the reaction vial and N2 pressure applied. The
labelling reaction was allowed to occur for 8 min at 1201C, after
which the vial was cooled to room temperature The AcN phase
was then transferred into a second 5 mL conical glass vial
containing 0.625 mL of 1 mM NaF solution. The first vial was
then rinsed with 0.875 mL of AcN and transferred to the
second vial.
mobile phase at a flow rate of 6 mL/min. The retention times of
[18F]fluoride and of [18F]-NBoc4FMPtbutyl ester were, respectively,
2.5 and 15 min. The protected n.c.a. radiofluorinated compound
was collected in a 100 mL vial containing 54 mL of de-ionized water.
Preconcentration. After mixing the collected [18F]-NBoc4FMPt-
butyl ester and the deionized water by N2 bubbling, this
solution was passed through a 50 mg Extract-Clean C18 mini
column (Alltech) at a flowrate of 10 mL/min, hereby trapping the
n.c.a. 18F-labeled compound on the C18 phase. The n.c.a. [18F]-
NBoc4FMPtbutyl ester, adsorbed on the C18 column, was eluted
with 1 mL of dry AcN through a tandem of a custom made dry
CaCl2 (300 mg, Sigma-Aldrich) mini-column and a 300 mg
Extract-clean Silica mini-column (Alltech) into a 5 mL vial that
was kept at 501C during the entire process. The AcN was
evaporated using a gentle N2 flow for 10 min.
Semi-preparative HPLC separation. The content of the vial was
mixed by nitrogen bubbling resulting in a AcN/H2O: 75/25
(v/v) solution and injected for semi-preparative RP-HPLC
separation, using an Apollo C18 250 ꢁ 10 mm (5 m) column
and AcN/H2O: 70/30 (v/v) as mobile phase at a flow rate of
6 mL/min. The retention times of [18F]fluoride and of [18F]-
Deprotection and work-up of 4-[18F]FMP. One millilitre of a
dichloromethane/trifluoroacetic acid: 4/1 (v/v) mixture was
added and deprotection applied for 20 min under N2 pressure.
After the removal of the solvents by evaporation (gentle N2 flow
at room temperature), another 1 mL of dichloromethane was
added and evaporated (5 min.). Finally, 5 mL of a 1% ethanol/
water (v/v) mixture was added to the vial. After mixing by N2
flow for 10 s the solution was passed through a tandem of a
50 mg Extract-Clean C18 mini column (Alltech) and a 50 mg
Alumina-N (Alltech) mini-column (to remove traces of respec-
tively 4-[18F]FMP-tButyl ester and [18F]fluoride) and the 4-
[18F]FMP was recovered in a sterile vial containing 45 mg of
trityl4FETtbutyl
ester
([18F]-O-[2-fluoroethyl]-N-(triphenyl-
methyl)-1,1-dimethylethyl ester) were, respectively, 2.5 and
32.2 min. The protected n.c.a. radiofluorinated compound was
collected over 3 min in a 100 mL vial containing 54 mL of
deionized water.
Preconcentration. After mixing by N2 bubbling the solution was
passed through a 50 mg Extract-Clean C18 mini column (Alltech)
at a flow rate of 10 mL/min through and the 18F-labeled
compound trapped. The C18 column was rinsed with 10 mL of
deionized water and purged with a dry N2 flow. The [18F]-
trityl4FETtbutyl ester adsorbed on the C18 column was eluted
with 1 mL of dry AcN through a tandem of a custom made CaCl2
(300 mg, Sigma-Aldrich) mini-column and a 300 mg Extract-clean
Silica mini-column (Alltech) into a 5 mL vial kept at 901C during
the entire process.
`
saline after passing through a 0.22 ım Cathivex filter (Millipore).
Radiosynthesis of O-(2’-[18F]fluoroethyl)-L-tyrosine ([18F]FET)
In our hands [18F]FET was synthesized following almost the
same steps as described for 4-[18F]FMP.
J. Label Compd. Radiopharm 2010, 53 58–62
Copyright r 2009 John Wiley & Sons, Ltd.
www.jlcr.org

K. Kersemans et al.
Deprotection and work-up of [18F]FET. To the reaction vial was
added 0.5 mL of trifluoroacetic acid and the deprotection reaction
was allowed to proceed for 15min. After removal of the solvents
by evaporation (gentle N2 flow at room temperature), another
0.250 mL of acetonitrile was added and evaporated (5min).
Finally, after cooling to room temperature 1 mL of a 10% ethanol/
water (v/v) solution was added to the vial. After mixing by N2 flow
over 10s the solution was passed through a tandem of a 50mg
Extract-Clean C18 mini column (Alltech) and a Alumina-N (Alltech,
50mg) mini-column and the product was recovered after passing
through a 0.22mm Cathivex filter (Millipore) in a sterile vial
containing 4 mL of a 1.125% NaCl solution.
Conclusion
The switching of the fluoromethyl side on the ring of
phenylalanine resulted in a compound, namely noncarrier-added
4-[18F]FMP, that was stable in its radiopharmaceutical formulation,
even at high radioactivity concentrations. The described auto-
mated radiosynthesis for this compound makes it possible to
produce 4-[18F]FMP in a high yield (30% overall) and with a high
radiochemical purity (>99%). Moreover, the developed setup is
also suited to synthesize [18F]FET in yields (40%) comparable to
earlier described methods, without having to change any
hardware components. The versatility of the method and the
automatic system makes the new compound readily accessible
for widespread use in both research and clinical practice.
Quality control
Quality control of the final n.c.a. 18F-labeled compounds was
Acknowledgement
achieved by HPLC analysis, performed on
a Vydac C18
˚
Monomeric 120 A (125 ꢁ 4 mm) (Grace) column using an AcN/
1 mM solution of CH3COONH4 and NaF: 5/95 (v/v) mixture of pH
6.5 as mobile phase with a flow rate of 1 mL/min while
monitoring both radioactivity (NaI(Tl), Harshaw Chemie) and
UV absorption at 254 nm (Shimadzu). The k0 values were
respectively: 0.6 for [18F]fluoride and 3.9 for 4-[18F]FMP. The k0
value of [18F]FET was 5.5. Chiral analysis was performed using an
The authors thank FWO-Vlaanderen (project G.0299.04N) and GOA-
Vrije Universiteit Brussel (GOA28) for financial support. K. Kerse-
mans wishes to thank the Research Unit High Resolution NMR and
the department of Organic Chemistry of the Vrije Universiteit
1
Brussel for performing the H NMR and ESI-MS analysis.
`
Astec Chirobiotic T column (5 ım, 125 mm ꢁ 4.6 mm) (Alltech)
and ethanol/water: 80/20 (v/v) as mobile phase with a flow of
1 mL/min, using the same detection system.
References
[1] R. Michael, PET in Oncology, Springer, Berlin, Heidelberg, 2008.
[2] K. J. Langen, K. Hamacher, M. Weckesser, F. Floeth, G. Stoffels,
D. Bauer, H. H. Coenen, D. Pauleit, Nucl. Med. Biol. 2006, 33, 287–294.
[3] D. Pauleit, A. Zimmermann, G. Stoffels, D. Bauer, J. Risse, M. O. Fluss, K.
Hamacher, H. H. Coenen, K. J. Langen, J. Nucl. Med. 2006, 47, 256–261.
[4] D. Pauleit, G. Stoffels, W. Schaden, K. Hamacher, D. Bauer,
L. Tellmann, H. Herzog, S. Broeer, H. H. Coenen, K. J. Langen,
J. Nucl. Med. 2005, 46, 411–416.
[5] K. Kersemans, M. Bauwens, T. Lahoutte, A. Bossuyt, J. Mertens,
Nucl. Instrum. Methods A. 2007, 571(1–2), 151–154.
[6] K. Kersemans, M. Bauwens, J. Mertens, Nucl. Med. Biol. 2008, 35(4),
425–432.
TLC control was performed using RP-18 F254S (5 ꢁ 7.5 cm)
plates with MeOH/H2O [50:50 (v/v)] as mobile phase. The Rf
values were, respectively, 0.0 for [18F], 0.5 for 4-[18F]FMP and 0.4
for [18F]FET. The radioactivity in both spots was measured.
The amount of Kryptofix2.2.2 still present in the final
radiopharmaceutical formulations was assessed using the colour
spot test that was described by Mock et al.10
The amount of residual AcN present in the final solution was
determined by means of the GC-method developed by
Channinga et al.,11 using a J&W HP-INNOWax 30 m ꢁ 250 mm ꢁ
0.25 mm column.
[7] V. Caveliers, K. Kersemans, T. Lahoutte, M. Bauwens, C. Vanhove,
M. Keyaerts, J. Mertens, A. Bossuyt, H. Everaert, J. Nucl. Med.
MEETING ABSTRACTS 2008, 49, 282.
Shelf-life stability of 4-[18F]FMP and [18F]FET
[8] H. Everaert, C. Chaskis, V. Caveliers, T. Lahoutte, J. Mertens, K.
Kersemans, C. Vanhove, M. Bauwens, A. Bossuyt, B. Neyns, J. Nucl.
Med. MEETING ABSTRACTS 2008, 49, 357.
[9] K. Hamacher, H. H. Coenen, Appl. Rad. Isot. 2002, 57, 853–856.
[10] B. H. Mock, W. Winkle, M. T. Vawrek, Nucl. Med. Biol. 1997, 24, 193–195.
[11] M. A. Channinga, B. X. Huang, W. C. Eckelman, Nucl. Med. Biol.
2001, 28(4), 469–471.
The radiopharmaceutical formulation containing 0.9% NaCl and
1% v/v of ethanol at pH 6 was kept at 251C. At appropriate times
a sample was taken and analysed by HPLC and TLC (QC
procedure).
www.jlcr.org
Copyright r 2009 John Wiley & Sons, Ltd.
J. Label Compd. Radiopharm 2010, 53 58–62