A solid-phase route to 18F-labeled tracers, exemplified by the synthesis of [18F]2-fluoro-2-deoxy-D-glucose

Article abstract of DOI:10.1002/anie.200603394

(Chemical Equation Presented) Scintillating synthesis: ¹⁸F- Containing radiopharmaceuticals can be prepared by using [¹⁸F] fluoride ions to displace a sulfonate linker and release a radiotracer from a solid support (see scheme; EOM =

Full text of DOI:10.1002/anie.200603394

Angewandte
Chemie
DOI: 10.1002/anie.200603394
Synthetic Methods
A Solid-Phase Route to 18F-Labeled Tracers, Exemplified by the
Synthesis of [18F]2-Fluoro-2-deoxy-d-glucose**
Lynda J. Brown, Denis R. Bouvet, Sue Champion, Alex M. Gibson, Yulai Hu, Alex Jackson,
Imtiaz Khan, Nianchun Ma, Nicholas Millot, Harry Wadsworth, and Richard C. D. Brown*
Positron emission tomography (PET) is an important diag-
nostic tool in modern medicine,[1] where it is routinely used to
locate and assess abnormalities in oncology,[2] neurology,[3]
and cardiology.[4] In addition, PET and microPET are also
finding increased applications in the areas of drug discovery
poor (based on 1 in the stoichiometric reaction), the use of a
massive excess of triflate 1 and a limiting amount of the
[18F]fluoride ion in the radiolabeling step ensures high yields
of the radiolabeled product (> 60%).
Current methods of [18F]FDG preparation have the
disadvantage that the FDG is produced as a mixture with a
large stoichiometric excess of d-glucose and other degrada-
tion products that arise from side reactions of the starting
material 1 during the fluoridation and deprotection steps.[11]
In the case of FDG, the presence of the by-products is not
considered to be harmful to the patient or to affect the
efficacy of the injected product.However, the above may not
hold true for other 18F-containing radiotracers under devel-
opment for research and diagnostic applications.In addition,
the short half-life of 18F requires the radiotracer to be
synthesized and purified as rapidly as possible, ideally
within one hour for clinical use.This makes reverse-phase
HPLC purification undesirable for commercial applications.
In recognition of the issues discussed above we set out to
develop a platform technology based on the nucleophilic
cleavage of solid-supported precursors using the [18F]fluoride
ion which would allow the production of 18F-labeled com-
pounds, as exemplified by the synthesis of [18F]FDG.[12,13]
Preparation of radiopharmaceuticals by this method would
be advantageous in terms of the purity of the cleaved
fluorinated products, whilst facilitating automation of the
process and thus offering increased reproducibility and
further protection from radiation exposure.[8,14] The new
technology would produce 18F-labeled tracers quickly and
with high specific activity, as well as minimize time-consuming
purification steps.
and
development.[5]
[18F]Fluoro-2-deoxy-d-glucose
([18F]FDG, [18F]-3) is the most widely used radiochemical
tracer in PET applications that possesses the positron-emit-
ting radionuclide 18F (half-life = 110 min).[6] [18F]FDG is used
to measure glucose uptake by tissue, and produces real-time
images for diagnosis, management, and study of diseases such
as cancer.
The radiosynthesis of [18F]FDG was first achieved by
electrophilic fluorination of d-glucal with [18F]F2, but rela-
tively low yields and limited stereoselectivity were obtained.[7]
Since that time, various synthetic routes have been
reported,[6,8–10] including the one used in PET centers
throughout the world: this method involves the reaction of
excess (105 equiv) tetra-O-acetyl-2-O-trifluoromethanesul-
fonyl d-mannose (1) with the [18F]fluoride ion in the presence
of a phase-transfer agent 4 (Scheme 1).Deprotection of the
FDG precursor 2 thus obtained, under basic or acidic
conditions, produces [18F]FDG.Although the chemical
yields for the nucleophilic fluoridation reaction are rather
Scheme 1. Conventional radiosynthesis of [18F]FDG ([18F]-3). Tf=
triflate.
Our chosen strategy was to attach the FDG precursor to
the solid-support through a sulfonate linker that would allow
specific cleavage of the radiotracer into solution by using the
[18F]fluoride ion.[15] Any precursor present at the end of the
reaction would remain attached to the resin, thus permitting
its separation from the product by simple filtration, and
avoiding the presence of the excess reactive triflate in the
deprotection step.Following removal of the protecting
groups, the 18F tracer could be purified by ion-exchange
chromatography to leave pure product ready for administra-
tion.
[*]Dr. L. J. Brown, Dr. Y. Hu, Dr. N. Ma, Dr. R. C. D. Brown
School of Chemistry
University of Southampton
Highfield, Southampton, SO171BJ (UK)
Fax: (+44)238-059-6805
E-mail: rcb1@soton.ac.uk
Dr. D. R. Bouvet, Dr. S. Champion, Dr. A. M. Gibson, Dr. A. Jackson,
Dr. I. Khan, Dr. N. Millot, Dr. H. Wadsworth
GE Healthcare
Initial efforts were focussed on adapting conventional
FDG solution-phase chemistry to produce a solid-supported
analogue of triflate 1.However, major limitations in this
approach became apparent: fluoridation of triflate 1 using a
stoichiometric amount of the [19F]fluoride ion afforded
[19F]FDG tetraacetate ([19F]-2) only as a minor product
(<5%; Scheme 2).[16]
The Grove Centre
White Lion Road, Amersham, Bucks, HP79LL (UK)
[**]R.C.D.B. thanks The Royal Society for the award of a University
Research Fellowship. We acknowledge the use of the EPSRC’s
Chemical Database Service at Daresbury.[23]
Supporting information for this article is available on the WWW
under http://www.angewandte.org or from the author.
Angew. Chem. Int. Ed. 2007, 46, 941 –944
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
941

Communications
Scheme 2. Synthesis of [19F]FDG by using stoichiometric amounts of
[19F]fluoride ions.
Furthermore, synthesis of the nonaflate 8 as a model for a
solid-supported sulfonate ester could only be effected in
reasonable yield using nonaflic anhydride (Nf2O;
Scheme 3);[17] the corresponding sulfonyl fluorides and chlor-
ides failed to react effectively with the hindered 2-hydroxy
group of 5 under similar conditions.One curious observation
was the formation of perfluorobutanoate 9 when perfluo-
roalkylsulfonation was attempted in the presence of pyridine
and AgOTf in THF (Scheme 3).
Scheme 4. Synthesis of [19F]FDG by using the sugar-linker model 16.
CSA=camphorsulfonic acid, EOM=ethoxymethyl.
occurred under basic conditions.It is important to recognize
that any 19FÀ ions released from the resin, through elimination
or 18FÀ exchange with residual resin-bound sulfonyl fluoride
(for example, 21), would be detrimental to the yield and
specific activity of the labeled product.A different linker was
therefore sought, which would be less prone to elimination
and that would be suitable for the formation of a sugar-linker
conjugate in solution ready for coupling with an aminomethyl
resin (Scheme 6).
Scheme 3. Reactions of 5 with F9C4SO2Cl and (F9C4SO2)2O. pyr=
pyridine.
At this point the high reactivity of the perfluoroalkylsul-
fonic anhydrides prevented us from preparing a suitably
functionalized linker, while the base-sensitive acetyl protect-
ing groups would not allow the formation of a more reactive
alkoxide from 5 to facilitate its coupling with a sulfonyl halide
linker.Fortunately, by changing to base-stable acetal protect-
ing groups,[18] it was possible to couple a b-d-mannose
derivative 15 with a perfluoroalkylsulfonyl halide to give
the sugar-linker model compound 16 in good yield
(Scheme 4).[17,19] The use of the b anomer was crucial to
obtain the desired stereoselectivity in the hydride reduction
(d.r. = ca.9:1), [18] and to ultimately open up C2 for nucleo-
philic attack by the fluoride ion.Our change in protecting-
group strategy was further vindicated when the reaction of
nonaflate 16 with [19F]KF afforded the desired protected
FDG 18 in a greatly improved 53% yield along with a smaller
quantity of an eliminated by-product 17.Global deprotection
of 18 was achieved by refluxing it in 6n HCl for 10 minutes to
afford unlabeled FDG [19F]-3, which was characterized
following conversion into the a- and b-tetraacetates 19 and
[19F]-2.[16a,20]
Scheme 5. Previous syntheses of perfluoroalkylsulfonate linkers.[15]
DMF=N,N-dimethylformamide.
Linker-sugar conjugates 29–32 were prepared by coupling
d-mannose derivative 15 to 5-iodooctafluoro-3-oxapentane-
sulfonyl fluoride, followed by radical-mediated coupling
reactions of the resulting iodide 25 with a series of enoic
acids of various chain lengths.[21] The yield of this reaction
using l6-heptadecenoic acid was lower because of the poor
solubility of this olefin.Reaction of acrylic acid with iodide 25
provided the reduced product 29 directly, whilst radical
coupling products 26–28 required de-iodination using zinc
powder in refluxing Et2O/AcOH.[22] The resulting acids were
coupled to aminomethyl-functionalized polystyrene to give
resins 33–36 in high yields (estimated from the masses of dry
resins).Successful formation of the supported sulfonate esters
Our attention then turned to the synthesis of a sugar-
linker construct.Recent publications described a perfluo-
roalkylsulfonyl fluoride linker, which was used to prepare aryl
triflate equivalents (Scheme 5).[15a,b] However, poor yields
were obtained for the coupling of d-mannose derivative 15 to
this linker in solution or after attachment to the resin.In
addition, elimination of the [19F]fluoride ion from the linker
942
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 941 –944

Angewandte
Chemie
traces (205 and 210 nm) showed that the protected 18F-labeled
sugar obtained from the resin precursor 35 contained
significantly reduced levels of chemical impurities.The
absence of a significant UV-active peak corresponding to
the protected FDG from the solid-phase synthesis indicated
very little leaching of 19FÀ ions from the linker during labeling,
and a high specific radioactivity of [18F]-18.Five successive
labeling experiments re-using the same sample of resin 35 all
led to the formation of the protected [18F]FDG with
consistently high radiochemical yields.These experiments
clearly demonstrated that the majority of the protected
d-mannose derivative remained attached to the resin through
the fluoroalkylsulfonyl linker during the 18F-labeling experi-
ments.
The effect of altering the chain length of the linker upon
the radiochemical yield was also investigated for resins 33–36
(Figure 1).There was a steady rise in the radiochemical yield
with increasing length of the alkyl chain up to four methylene
groups, after which it leveled off and then began to fall.
Sulfonate ester 35 was ultimately considered to be the linker
of choice based on radiochemical yield, economy, and
efficiency of synthesis.
Scheme 6. Synthesis of solid-supported FDG precursors 33–36.
HMDS=1,1,1,3,3,3-hexamethyldisilazane.
was confirmed by on-bead IR spectroscopy, elemental
analysis, and 19F NMR as well as MAS 1H and 13C NMR
spectroscopy.
Low-activity labeling experiments were carried out in a
carbon glass vessel using [18F]KF (185–350 MBq) and
kryptofix[2.2.2] (4) in CH3CN.Radiochemical yields were
calculated from the reversed-phase HPLC chromatograms of
the protected product [18F]-18 recorded with g-ray detection
(Scheme 7).Typically, solution-phase 18F-labeling results in
Figure 1. Effect of the linker chain length on the radiochemical yield.
(C1 resin synthesized by coupling 15 to resin 21.)
To identify the optimal conditions required for removal of
the protecting groups a sample of the protected [18F]fluoro-
sugar [18F]-18 was subjected to a range of acidic conditions,
with Dionex anion-exchange HPLC fitted with g-ray detec-
tion used to quantify the [18F]FDG produced.The preferred
conditions for the generation of [18F]FDG required depro-
tection of [18F]-18 with 6m HCl for five minutes at 1258C,
which gave a radioactive ion-exchange chromatogram with
one main peak ([18F]FDG) and one small peak (< 3%)
corresponding to partially deprotected [18F]FDG.This result
produced an average radiochemical yield of 73% (decay
corrected) and activity losses on the resin between 3–8%,
with 91–97% of the activity collected.
In conclusion, we have presented a reliable route for the
synthesis of a solid support that liberated protected [18F]FDG
in high radiochemical yield on treatment with 18FÀ ions.The
product [18F]-18 was deprotected to give [18F]FDG with
excellent chemical purity.The method lends itself to auto-
mation, as the solid-supported precursor could be provided as
Scheme 7. Synthesis of [18F]FDG from resin-bound precursors.
80–90% incorporation after 2 minutes at 868C in a reaction
volume of 1–2 mL.[8] For heterogeneous reactions involving
the solid-supported precursors 33–36 it was advantageous to
have the fluoride ion as concentrated as possible (reaction
volume: 0.2 mL). Using resin 35, labeling times of 3–4
minutes at 868C resulted in 70–91% incorporation of 18F
into [18F]-18.Furthermore, high-activity labeling studies (up
to 6.16 Ci) resulted in similar incorporation yields (68–77%).
The chemical purity of the crude protected FDG [18F]-18
from the resin precursor 35 was compared against the crude
protected FDG product [18F]-2 obtained by conventional
solution-phase synthesis.HPLC analyses, with g-ray detec-
tion, indicated that the desired protected fluorinated sugars
and 18FÀ were the only significant 18F-containing components
in both reactions.By contrast, the corresponding UV activity
Angew. Chem. Int. Ed. 2007, 46, 941 –944
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.angewandte.org
943

Communications
HC. .Kolb, ON. .Witte, N.Satyamurthy, JR. .Heath, ME. .
Phelps, S.R.Quake, H.R.Tseng, Science 2005, 310, 1793.
[11] Y.Kuge, K.Nishijima, K.Nagatsu, K.Seki, K.Ohkura, A.
Tanaka, M.Sasaki, E.Tsukamoto, N.Tamaki, Nucl. Med. Biol.
2002, 29, 275.
part of a kit to a radiopharmacy.The production of a wide
range of other 18F-labeled radiopharmaceuticals will be
facilitated by this new platform technology, thus making
such compounds more widely available for future applica-
tions.
[12] LJ..Brown, RC. D. .Brown, HJ..Wadsworth, A.Jackson,
WO 012319A1, 2005.
[13] The solid-phase synthesis of 11C-labeled N-methylated sulfona-
mides and amines has been reported: D.Maclean, J.Zhu, M.Y.
Received: August 18, 2006
Revised: October 31, 2006
Chen, R.Hale, N.Satymurthy, J.R.Barrio,
J. Am. Chem. Soc.
2003, 125, 10168.
Keywords: fluorine · positron emission tomography ·
radiopharmaceuticals · solid-phase synthesis ·
synthetic methods
.
[14] a) HC. .Padgett, DG. .Schmidt, A.Luxen, GT. .Bida, N.
Satyamurthy, J.R. Barrio, Appl. Radiat. Isot. 1989, 40, 433;
b) S.M. Moerlein, J.W. Brodack, B.A. Siegel, M.J. Welch,
Appl. Radiat. Isot. 1989, 40, 741; c) K.Hamacher, G.Blessing, B.
Nebeling, Appl. Radiat. Isot. 1990, 41, 49; d) J.S.Fowler, R.R.
MacGregor, A.P.Wolf, A.A.Farrell, K.I.Karlstrom, T.J.Ruth,
J. Nucl. Med. 1981, 22, 376.
[1] a) H.R.Herschman, Science 2003, 302, 605; b) J.Czernin, M.E.
Phelps, Annu. Rev. Med. 2002, 53, 89; c) R.G. Blasberg, J.G.
Tjuvajev, J. Clin. Invest. 2003, 111, 1620; d) R.Nutt, Mol. Imag.
Biol. 2002, 4, 11.
[2] a) H.Anderson, P.Price, Eur. J. Cancer 2000, 36, 2028; b) S.S.
Gambhir, Nat. Rev. 2002, 2, 683.
[3] M. Iacoboni, J.C. Baron, R.S.J. Frackowiak, J.C. Mazziotta,
G.L.Lenzi, Clin. Neurophysiol. 1999, 110, 2.
[4] M.N.Maisey, Nucl. Med. Commun. 2000, 21, 234.
[5] J.Wang, L.Maurer, Curr. Top. Med. Chem. 2005, 5, 1053.
[6] B.Beuthien-Baumann, K.Hamacher, F.Oberdorfer, J.Stein-
bach, Carbohydr. Res. 2000, 327, 107.
[15] a) Y.J.Pan, C.P.Holmes, Org. Lett. 2001, 3, 2769; b) Y.Pan, B.
Ruhland, C.P.Holmes, Angew. Chem. 2001, 113, 4620; Angew.
Chem. Int. Ed. 2001, 40, 4488.
[16] Data for [19F]-2, -6, and -7 were consistent with the literature.
[19F]-2: a) P.Kovac, H.J.C.Yeh, C.P.J.Glaudemans,
Carbo-
hydr. Res. 1987, 169, 23; 6: b) H.Shimizu, J.M.Brown, S.W.
Homans, R.A. Field, Tetrahedron 1998, 54, 9489. 7: c) M.
Hayashi, K.Yamada, O.Arikita, Tetrahedron 1999, 55, 8331.
[17] P.J.Stang, M.Hanack, L.R.Subramanian,
[18] a) S.Levy, E.Livni, D.Elmaleh, W.Curatolo,
Chem. Commun. 1982, 972; b) B.Doboszewski, GW. .Hay,
W.A.Szarek, Can. J. Chem. 1986, 65, 412.
[19] R.K.Sehgal, B.Almassian, D.P.Rosenbaum, R.Zadrozny, S.K.
Sengupta, J. Med. Chem. 1987, 30, 1626.
Synthesis 1982, 85.
J. Chem. Soc.
[7] T.Ido, C.Wan, V.Casella, J.S.Fowler, A.P.Wolf, M.Reivich,
D.E.Kuhl, J. Labelled Compd. Radiopharm. 1978, 14, 175.
[8] For a detailed discussion of methods for the synthesis of
[18F]FDG, see a) T.J. Tewson, J. Nucl. Med. 1983, 24, 718;
b) HH. .Coenen, VW. .Pike, G.Stöcklin, R.Wagner,
Radiat. Isot. 1987, 38, 605.
Appl.
[20] K.Dax, B.I.Glanzer, G.Schulz, H.Vyplel,
Carbohydr. Res.
1987, 162, 13.
[9] K.Hamacher, H.H.Coenen, G.Stöcklin, J. Nucl. Med. 1986, 27,
235.
[21] G.B.Rong, R.Keese, Tetrahedron Lett. 1990, 31, 5615.
[22] N.O.Brace, J. Org. Chem. 1975, 40, 766.
[23] D.A. Fletcher, R.F. McMeeking, D. Parkin,
[10] For the synthesis of [18F]FDG by using continuous-flow micro-
reactors, see C.C.Lee, G.D.Sui, A.Elizarov, C.Y.J.Shu, Y.S.
Shin, A.N.Dooley, J.Huang, A.Daridon, P.Wyatt, D.Stout,
J. Chem. Inf.
Comput. Sci. 1996, 36, 746.
944
www.angewandte.org
ꢀ 2007 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
Angew. Chem. Int. Ed. 2007, 46, 941 –944