Small Prosthetic Groups in18F-Radiochemistry:
`Useful Auxiliaries for the Design of18F-PET
`Tracers
`Ralf Schirrmacher, PhD,*Björn Wängler , PhD,† Justin Bailey, PhD,*V adim Bernard-Gauthier, PhD,*
`Esther
`Schirrmacher, PhD,*and Carmen Wängler, PhD‡
`Prosthetic group (PG) applications in18F-radiochemistry play a pivotal role among current18F-labeling
`techniques for the development and availability of18F-labeled imaging probes for PET (Wahl, 2002)
`(1). The introduction and popularization of PGs in the mid-80s by pioneers in18F-radiochemistry
`has profoundly changed the landscape of available tracers for PET and has led to a multitude of
`new imaging agents based on simple and efficiently synthesized PGs. Because of the chemical
`nature of anionic
`18F− (apart from electrophilic low specific activity18F-fluorine), radiochemistry
`before the introduction of PGs was limited to simple nucleophilic substitutions of leaving group
`containing precursor molecules. These precursors were not always available, and some target
`compounds were either hard to synthesize or not obtainable at all. Even with the advent of re-
`cently introduced “late-stage fluorination” techniques for the
`18F-fluorination of deactivated aromatic
`systems, PGs will continue to play a central role in18F-radiochemistry because of their robust
`and almost universal usability. The importance of PGs in radiochemistry is shown by its current
`significance in tracer development and exemplified by an overview of selected methodologies for
`PG attachment to PET tracer molecules. Especially, click-chemistry approaches to PG conjuga-
`tion, while furthering the historical evolution of PGs in PET tracer design, play a most influential
`role in modern PG utilization. All earlier and recent multifaceted approaches in PG development
`have significantly enriched the contingent of modern
`18F-radiochemistry procedures and will con-
`tinue to do so.
`Semin Nucl Med 47:474–492 © 2017 Elsevier Inc. All rights reserved.
`Introduction
`B
`efore discussing a prosthetic group (PG)’s purpose in
`radiochemistry ,1 its definition needs to be clarified. The
`use of the term “prosthetic group” has been traditionally em-
`ployed in enzymology and defined as a cofactor that is “either
`tightly or loosely bound to the enzyme. If tightly con-
`nected, the cofactor is referred to as a prosthetic group.”
`2 How
`did this definition migrate into the field of radiochemistry and
`what is the significance of it in this particular context? It should
`be recalled that for the development of a new radiotracer,
`usually (in almost all cases), a nonradioactive lead com-
`pound devised by medicinal chemistry is taken as a chemical
`template or scaffold and translated into a radiotracer for in
`vivo imaging by introducing a radiolabel. This general strat-
`egy holds true for tracers intended for PET, where so-called
`organic radionuclides such as carbon-11 and fluorine-18 (
`18F)
`are commonly used, and where the size of compounds varies
`between small tracers for brain imaging, and peptide- and
`protein-based probes for cancer imaging.
`3-8 If the original lead
`structure does not contain, for example, fluorine, the radio-
`chemist has to add fluorine at a position in the molecule that
`most likely does not interfere with its binding properties to
`the target. The smaller the change in the molecular struc-
`ture of the lead compound, the less probable the binding
`behavior to the target structure will change. Therefore, ra-
`diochemists and also medicinal chemists (addition of fluorine
`to a molecule can have very positive effects on its bio-
`availability and stability in vivo
`9) often simply exchange a
`*Medical Isotope and Cyclotron Facility , Cross Cancer Institute, University
`of Alberta, Alberta, Canada.
`†Molecular Imaging and Radiochemistry , Department of Clinical Radiology
`and Nuclear Medicine, Medical Faculty Mannheim of Heidelberg
`University , Germany.
`‡Biomedical Chemistry , Department of Clinical Radiology and Nuclear
`Medicine, Medical Faculty Mannheim of Heidelberg University , Germany.
`Address reprint requests to Ralf Schirrmacher, PhD, Medical Isotope and
`Cyclotron Facility , Cross Cancer Institute, University of Alberta, AB,
`Canada. E-mail: schirrma@ualberta.ca
`http://dx.doi.org/10.1053/j.semnuclmed.2017.07.001
`0001-2998/© 2017 Elsevier Inc. All rights reserved.
`474
`Petitioner GE Healthcare – Ex. 1019, p. 474
`
`
`
`
`
`
`
`hydrogen atom in a lead with a fluorine atom based on the
`similar van der Waals radii of hydrogen and fluorine. Un-
`fortunately , it is not always possible to introduce radioactive
`18F at the same position in a molecule where medicinal chem-
`ists are able to add stable fluorine in their multistep synthetic
`schemes. This ostensible contradiction can be explained easily
`by the different requirements for nonradioactive and radio-
`active fluorine in preparative chemistry . Although
`18F is still
`fluorine in chemical terms, the radioactive nature of 18F in-
`troduces the problem of radioactive decay to the synthetic
`equation. In general, multistep syntheses up to this day are
`not desirable in
`18F radiochemistry because such laborious
`and lengthy procedures will never yield the desired radio-
`tracer in high radiochemical yields (RCYs) and high specific
`activities, not to mention the cumbersome purification of many
`intermediates on the way toward the final product.
`What Is a Prosthetic Group?
`These ramifications are still the reasons why PGs have been in-
`vented and used in tracer development ever since. In principle, a
`PG overcomes the inability to radiolabel certain compounds with
`18F in 1-2 steps via nucleophilic18F− substitution of a leaving group
`in a timely manner. Comparable with carbon-11 methylations using
`11C-methyliodide, small reactive18F bearing labeling synthons were
`envisioned as analogous labeling reagents to engage nucleophilic moi-
`eties such as amino, hydroxy , carboxy , and thiol groups. This strategy
`made it possible to easily transfer a
`11C-methylation to an equiva-
`lent18F-fluoro methylation or ethylation (or alkylation in general),
`taking advantage of the already existing labeling precursor for
`11C-labeling. This concept has been recently extended to a multi-
`tude of chemical18F-bioconjugations based on click chemistry ,
`significantly supporting the use of PG applications in radiotracer de-
`velopment despite the development of novel metal catalyzed late-
`stage
`18F-fluorination techniques that have substantially enriched
`existing18F-labeling techniques.10-14 Referring back to the above-
`introduced definition of a PG in enzymology , a PG in
`radiopharmaceutical chemistry is a serviceable auxiliary that can easily
`be attached to a precursor molecule in analogy to existing
`11C-labeling
`techniques (alkylation, acylation, amination, etc) or click-chemistry
`methodologies (triazole-, dihydropyrazine-, oxime formations, etc).
`However, the use of PGs for tracer development carries an impor-
`tant limitation. Usually , a first-in-kind radiotracer (especially for small
`compounds of low molecular weight) is directly derived from the
`lead structure that has no functionalities that could be easily modi-
`fied to introduce
`18F by a PG. As a result, in most cases, a methyl
`group that is usually abundantly available in most small organic com-
`pounds is substituted for a
`11C-methyl group. This strategy yields
`a radiotracer that is in every regard exactly the same as the nonra-
`dioactive lead compound. In contrast, the structural exchange of a
`methyl group with a higher
`18F-labeled homolog such as an
`18F-fluoroethyl group or even a closely related18F-fluoromethyl group
`can significantly alter the biological and chemical properties of the
`resulting molecule. It is the responsibility of the researcher to prove
`that this very minor chemical-radiochemical alteration in struc-
`ture does not compromise important parameters such as toxicity ,
`metabolism, and binding affinity to the target. This of course adds
`a new layer to the development process of
`18F-radiopharmaceuticals,
`a noticeable detriment in comparison with an unaltered11C-tracer
`that has most often already been tested for toxicity , binding affin-
`ity , and metabolism in the course of commercial drug development.
`This handicap, however, is offset by the advantages of
`18F over11C
`in terms of nuclide properties, and longer half-life, which allows for
`a prolonged duration of PET measurement and thus a higher quality
`of corresponding PET images.
`Many of the abovementioned concepts of
`18F-PG introduction
`can be illustrated by the synthesis of 6-O-(2-[18F]fluoroethyl)-6-O-
`desmethyl diprenorphine, a PET tracer for opioid in vivo imaging,
`where the original methyl group in 6 position is substituted by an
`18F-fluoroethyl moiety (Fig. 1).
`The
`methoxy group can be easily converted into a hydroxy group
`(the desmethyl precursor) that can be labeled either by [11C]methyl
`iodide or 2-[18F]fluoroethyl tosylate ([18F]FETos).15 This particular
`position is the most suited to make small changes to the structure
`of diprenorphine without compromising its binding to the opioid
`receptors. Another group had previously chosen the N-17 posi-
`tion to replace the cyclopropyl methyl group with an
`18F-fluoro ethyl
`O
`OH
`N-17
`HO O-6CH3
`favorable derivatization site
`O
`OH
`N
`HO O
`18F
`18F-PG labeling
`at N-17 position
`preserved binding to opioid receptors
`18F-PG labeling
`at O-6 position
`O
`OH
`N
`HO OCH3
`unfavorable derivatization site
`18F
`n=1-2
`reduced binding to opioid receptors
`Figure 1Derivatization sites of the opioid receptor ligand diprenorhine. Structural changes of diprenorphine at the O-6 position (right) yield
`preserved binding to the receptor, whereas changes at the N-17 position (left) lead to a reduced binding to the receptor.
`475Small prosthetic groups in18F-radiochemistry
`Petitioner GE Healthcare – Ex. 1019, p. 475
`
`
`
`
`
`
`
`and -propyl moiety , but the in vivo evaluation of the resulting com-
`pounds revealed reduced binding in opioid receptor-rich regions
`in comparison with the
`11C-original.16 Fortunately , the O-6 derivatized
`18F variant of diprenorphine overcomes the shortcoming of the origi-
`nal11C-labeled compound of a very limited PET scan duration and
`even enables shipment of the compound to PET centers without a
`cyclotron. This review does not claim to summarize the entire lit-
`erature with regard to reported PGs over the last 30 years but will
`rather explain conceptual strategies behind PG utilization using de-
`monstrative examples.
`The Early Days of Prosthetic
`Group Chemistry
`Early research was reported in the late 1970s and early 1980s,
`when the synthesis of halo- 18F-fluoromethanes was de-
`scribed in the context of hot-atom recoil labeling where the
`precursor substrate was directly added to the neon gas target
`and bombarded with 14-MeV deuterons.
`17 The “hot” 18F−
`formed via the 20Ne(d,α)18F reaction could react directly
`with in-target components to form the corresponding
`halo-
`18F-fluoromethanes. Unfortunately , the high energy trans-
`fer rates within a target during bombardment led to an increased
`formation of radioactive impurities impairing the produc-
`tion of higher activities. The irradiation of halo-alkanes in the
`presence of
`18F generates halo-18F-fluoro alkanes, which were
`characterized but not used for any kind of follow-up label-
`ing. Another experimental study with more detailed information
`but still limited practical applicability for preparative
`18F-radiochemistry was reported by Root and Manning.18 At
`the same time,18F-fluorinated surface oxidized silver wool as
`a curious source of 18F was used for the synthesis of
`halo-18F-fluoro alkanes. The authors did not have the so-
`phisticated labeling utilities at their disposal, as modern
`radiochemists and consequently the reaction mechanism using
`silver wool is still not well understood.
`19 Another early con-
`tribution from Coenen and coworkers in 1985 fully
`characterized for the first time 1-bromo-[
`18F]fluoromethane
`with the intent to use this synthon analogously to [11C]methyl
`iodide.20 The labeling conditions to 18F-fluorinate the pre-
`cursor dibromomethane via nucleophilic substitution were
`different from previous approaches. Almost at the same time,
`Dae and coworkers reported on a mechanistic study of
`halofluorination of olefins that paved the way towards other
`investigations.
`21 The full automation of the synthesis of
`bromo-18F-fluoromethane was achieved by automated dis-
`tillation of the reaction mixture over 3 Sep-Pak Plus silica
`cartridges to remove the dibromomethane precursor and
`unreacted
`18F−.22 The introduction of the aminopolyether
`Kryptofix222 and potassium carbonate system as well as the
`tetrabutylammonium
`18F-fluoride complex23,24 constituted im-
`portant improvements enhancing the nucleophilicity of the
`18F−anion, via complexation of the potassium cation through
`Kryptofix222, leaving the18F− un-solvatized and thus “naked”
`or through the formation of highly soluble nBu4N[18F]F (due
`to the phase transfer capabilities of nBu4N+), respectively . The
`implementation of the Kryptofix222/K[18F]F labeling system
`marked a turning point in18F-radiochemistry . It is still an im-
`portant pillar of modern 18F-labeling methodology and had
`a catalytic effect on the further development of PGs and ra-
`diotracers in general. Previously , hard to label compounds
`suddenly became available through this innovation, invigo-
`rating and intensifying the research on
`18F-labeled compounds
`and making 18F the most important diagnostic radionuclide
`today .25 Besides bromo- and iodo-[18F]fluoromethane, the more
`reactive [18F]fluoromethyl triflate was introduced in 2002 via
`online conversion of bromo-[18F]fluoromethane using silver
`triflate in a gas-solid phase reaction.26 The year 1986 was most
`memorable with regard to PG development and PG applica-
`tion in tracer research. Several pioneering researchers in
`radiochemistry such as Kilbourn, Katzenellenbogen, Welch,
`Shiue, Wolf, Barrio, Coenen, and Stöcklin presented their early
`or continuing research on
`18F-fluorinated alkyl tosylates,
`-mesylates, and -halides at the Sixth International Sympo-
`sium on Radiopharmaceutical Chemistry in Boston,
`Massachusetts, USA, initializing a new important chapter in
`radiochemistry that still continues to have an effect on tracer
`conception. It cannot be overstated that putting forward the
`Kryptofix222/K
`2CO3 labeling system for anionic18F-fluorination
`sustainably invigorated and accelerated progress in18F-tracer
`development. Although some new methods have completely
`omitted toxic Kryptofix222 from the synthesis and also omitted
`the azeotropic drying procedure,
`27 this18F-nucleophilicity-enhancing
`auxiliary is still in frequent use. Shiue and Wolf for example applied
`this labeling system in the synthesis of higher
`18F-fluoroalkylated
`analogs of spiroperidol (spiperone4, Fig. 2) (at the same confer-
`ence
`there were several contributions from different groups featuring
`the synthesis of18F-N-alkyl labeled spiroperidols5, Fig. 2), an im-
`portant
`ligand to investigate dopamine receptors in humans.23 The
`motivation to use PG chemistry was clearly motivated by the dif-
`ficult multistep synthesis of the original compound, making a routine
`application for human PET imaging unlikely . The obtained
`N-[
`18F]fluoroalkyl spiroperidol derivatives5, despite being struc-
`turally slightly different from the lead, could be synthesized consistently
`from their corresponding
`18F-fluoroalkyl halides in radiochemical
`yields of up to 60% (decay corrected) (Fig. 2). Generally for con-
`venience’s
`sake, a 1-step fluorination is always preferable and the
`chosen route if a labeling precursor is available. However, if the syn-
`thesis of a direct 1-step labeling precursor is cumbersome or the
`direct 1-step labeling results in the formation of many side-
`products because of harsh labeling conditions, the PG labeling
`approach is preferred for reasons of more efficient synthesis and faster
`and immediate evaluation of the tracer. The pioneering work from
`all those early groups provided the necessary innovative momen-
`tum to advance radiochemistry to the next level and set the state for
`numerous applications, further developments, and, finally , click chem-
`istry in PG application.
`28-31
`2-[18F]fluoroethyl tosylate: The PG
`Workhorse
`If there is one PG that almost every radiochemist has at least
`used once in his or her professional career, it would most prob-
`ably be [
`18F]FETos, first introduced by Block et al at the
`476 R. Schirrmacher et al.
`Petitioner GE Healthcare – Ex. 1019, p. 476
`
`
`
`
`
`
`
`Forschungszentrum Juelich, Germany , in 1987. 32 The
`easy and reliable synthesis of [18F]FETos, its simple and con-
`venient purification, its perfect balance between reactivity
`and stability in solution, and its unprecedented number of
`labeling applications attesting to its usefulness make it the
`most stand-out PG to date (originally cited from Block et al:
`“This compound appears optimal as fluoroalkylation agent
`with respect to size, stability and ease of its preparation”)
`32
`(Fig. 2). Countless applications of 18F-FETos have been
`published in the literature and it would be beyond the scope
`of a review to just count them all. Although many other similar
`2-[
`18F]fluoroethyl sulfonates (and triflates cf. above) have
`been the subject of systematic investigations 33 (eg, for the
`synthesis of the dopamine reuptake ligand [ 18F]FECNT 7
`from precursor 6, Fig. 3), [ 18F]-FETos remains the most
`prominent labeling synthon among18F-fluoro-ethylating agents
`targeting amino and deprotonated hydroxy functions. [18F]FETos
`functions as an 18F-surogate PG for [ 11C]-methyl iodide 34
`although bromo-[ 18F]fluoromethyl or [ 18F]trifluoromethyl
`synthons structurally more closely resemble [11C]methyl iodide
`in terms of spatial and steric demand. A drawback of these
`smaller PGs is their more intricate preparation (see above).
`Especially for less-experienced radiochemists, the synthesis
`of 1-bromo-
`18F-fluoro methane bears some preparative
`pitfalls, which can translate into low radiochemical yields
`or an impure PG. Furthermore,
`18F-trifluoromethyl-chemistry ,
`despite its high synthetic potential, has only recently been
`introduced and has not been widely applied yet.
`35 In stark
`contrast, the synthesis of [18F]FETos is almost foolproof. The
`precursor ethylene 1,2-ditosylate is commercially available,
`constitutes a non-volatile solid (unlike dibromo- or diiodo
`ethane, which are liquids but also greenhouse gases and
`should be used with care) and has a high molecular weight
`allowing accurate, easy , and convenient weighing control.
`The
`18F-fluorination of ethylene 1,2-ditosylate reliably
`yields the PG between 50% and 80% RCY. The [18F]FETos is
`easily separated by High Performance Liquid Chromatogra-
`phy (HPLC) from its more lipophilic precursor, and final
`workup is equally simple via solid phase extraction. Even pro-
`cedures without HPLC separation have been reported.
`36 As
`for many PGs, one drawback of [18F]FETos should not be dis-
`regarded, namely the fact that the exchange of a 11C-methyl
`group with a 18F-fluoro ethyl group means a change in mo-
`lecular structure. Albeit small, this structural dissimilarity
`necessitates that any resulting radiotracer has to be tested for
`toxicity and change in biological behavior irrespective of whether
`these tests have already been performed for the original
`11C-tracer that is most often structurally equal to a fully evalu-
`ated pharmaceutical drug. [18F]FETos has been challenged over
`time by similar PGs such as bromo-18F-fluoro ethane.37,38 The
`efficient alkali iodide promoted 18F-fluoroethylation of
`p-anisidine8, a very unreactive substrate for alkylations; using
`both 18F-FETos and bromo-18F-fluoroethane demonstrates the
`usefulness of both PGs for radiolabeling (Fig. 3). Several
`radioligands
`of relevance in nuclear medicine were labeled
`applying this methodology , for example,18F-fluoroethyl-choline
`9 (Fig. 3). Although strategies have been reported to
`r
`outinely prepare bromo-18F-fluoroethane by various meth-
`odologies, 18F-FETos seems continuously to assert itself as the
`most applied PG in the history of PG applications. An inter-
`esting metabolically stable alternative to
`18F-FETos has been
`proposed in 2013 where a 18F-fluorocyclobutyl PG serves as
`an 18F-labeling agent 10 to provide an18F-labeled amino acid
`11 (Fig. 3). The labeling yields using this synthon unfortu-
`nately
`did not compare favorably with other 18F-alkylating
`PGs despite its higher in vivo metabolic stability .39
`X OTf
`n
`n=1, X=Br
`n=2, X=Br
`n=2, X=I
`n=3, X=Br
`nBu4N[18F]
`X
`18F
`n
`Chi et al. (28)
`Br Br
`n
`I I
`n
`n=1-2
`Br,I
`18F
`n
`Shiue et al. (23)
`K[18F]F/Kryptofix 222
`TosO
`OTos
`K[18F]F/Kryptofix 222
`TosO
`18F
`Block et al. (29)
`N
`N
`HN
`O
`O
`F
`N
`N
`N
`O
`O
`F
`spiperone
`18F
`n n=1-3
`n
`n=1
`18F-fluoroalkyl-spiperone
`Figure 2Synthesis of various 18F-fluoro alkylation PGs for the synthesis of an 18F-derivative of the D2 receptor ligand spiperone.
`477Small prosthetic groups in18F-radiochemistry
`Petitioner GE Healthcare – Ex. 1019, p. 477
`
`
`
`
`
`
`
`18F-Active Esters: Unspecific but
`Highly Reactive Labeling Agents
`In spite of [ 18F]FETos’s usefulness in labeling organic mol-
`ecules of low to medium molecular weight (<1000 u), it is
`less suited to react with complex molecules such as pep-
`tides and proteins for diverse reasons. First of all, [
`18F]FETos
`is not “super” reactive (OTos is a nucleofuge of medium re-
`activity) and requires some energy (eg, higher reaction
`temperatures) to react efficiently and in high radiochemical
`yields. Proteins in particular do not tolerate high tempera-
`tures and denature quickly . Peptides, if not fully side-chain
`protected, can easily yield more than just 1
`18F-fluoroalkylation
`product, demanding laborious purification via HPLC.
`[
`18F]FETos has never found widespread application for peptide
`(or other larger compounds such as oligonucleotides) label-
`ing, and conditions for labeling have not been optimized
`because there are far better PGs for that purpose. Highly re-
`active chemically activated esters, radiolabeled with
`18F , have
`early on been envisioned as reactive secondary labeling pre-
`cursors for peptide and protein labeling under preferably
`aqueous conditions because water perfectly dissolves both.
`Direct
`18F-labeling approaches in water were not available at
`that time and were only introduced later in 2005-2006 using
`unconventional radiochemistries based on B-
`18F and Si- 18F
`formation.40 The most prominent example of an activated ester-
`based PG is N-succinimidyl-4-[18F]fluorobenzoate ([18F]SFB)
`introduced by Vaidyanathan and Zalutsky in the 1990s. 41,42
`Its publication has set a field-wide search in motion for new
`and even more efficient active ester-based PGs despite the
`report of significant radioactive side-product formation. The
`initial synthetic procedure (Fig. 4) where 4-formyl-N,N,N-
`trimethyl
`anilinium trifluoromethane sulfonate 12 is
`radiolabeled with 18F− to yield 4-[18F]fluorobenzaldehyde 13,
`which is oxidized into 4-[18F]fluoro benzoic acid14 and finally
`converted in situ into [18F]SFB 15, has been improved, modi-
`fied, and automatized to make it more convenient and
`guarantee high RCYs for routine applications. Depending on
`the setup and method used, the overall synthesis can take
`approximately 2 hours, which is decidedly at the higher end
`of the preferable time line for PG syntheses. Automation of
`this usually 3-step preparation includes the initial labeling
`of the triflate salt of ethyl 4-(N,N,N-trimethylammonium)benzoate,
`its ester hydrolysis, and final in situ active ester formation. It was
`always a challenge for radiochemists to automatize the
`18F-SFB
`synthesis to make its use more attractive for routine radiochemis-
`try . The introduction of other ester moieties such astert-butyl ester
`into the labeling precursor 16 that is easily cleavable after
`18F-fluorination via trifluoroacetic acid yielding18F-fluorobenzoic
`acid14, has improved the original procedure for [18F]SFB synthesis
`H
`N
`COOMe
`Cl
`18F
`O S
`O
`O
`X
`Y
`leaving group
`X=Y=H
`X=Br, Y=H
`X=NO
`2, Y=H
`X=Y=Br
`X=Me, Y=H : tosylate moiety
`N
`COOMe
`Cl
`18F
`NH2
`MeO
`NaI or KI or CsI
`18F OTos/Br
`low reactivity
`N
`OH
`18F OTos/Br
`no reaction
`70-80%
`H
`N
`MeO
`18F
`NaI or KI or CsI
`18F OTos/Br
`N
`OH18F
`OTos-/Br-
`[18F]FECH
`TosO
`OTos
`18F-/Kryptofix 222
`acetonitrile
`TosO
`18F
`H2N COOH
`O
`18F+tyrosine
`base
`Figure 3Application of several different18F-fluoroalkylating PGs for the 18F-fluoroalkylation of different precursor mol-
`ecules such as the dopamine re-uptake radioligand 18F-FECNT 7, 18F-FECH 9, and the tyrosine derivative 11.
`478 R. Schirrmacher et al.
`Petitioner GE Healthcare – Ex. 1019, p. 478
`
`
`
`
`
`
`
`(Fig. 4).43 Besides technical advancements in [18F]SFB synthesis
`such as a recently reported 3-step 1-pot synthesis for automation,44
`other active esters were devised to challenge [18F]SFB with regard
`to ease of preparation, synthesis time, and labeling efficiency . The
`most desirable synthetic route toward an18F-active ester would be
`a single-step fluorination. One complication on the way toward a
`1-step labeling procedure is caused by the inherent lability of
`the active ester moiety . This strategy was tried very early in 1988
`and proved to be unsuccessful at that time.
`45 Although desirable
`for the intended labeling of proteins, this chemical reactiveness
`hinders the 1-step
`18F-fluorination of an active ester precursor.
`Even if possible, the reactiveness of the active ester moiety leads
`to side products that have to be removed before the subsequent
`labeling with this PG. The synthesis of N-succinimidyl
`4-([
`18F]fluoromethyl)benzoate ([18F]SFMB)18by Lang and Eckelman
`was the first example of a 1-step active ester labeling that yielded
`the reactive PG in 18% RCY at the end of a 30- to 35-minute
`synthesis demonstrating impressively the reduction in synthesis
`time in comparison with the originally published
`18F-SFB synthesis.46
`The relatively low yields of the18F-fluorination of precursor17
`were explained through experimentally confirmed formation of
`radioactive side products19 and products of active ester hydroly-
`sis20-21(Fig. 5). [
`18F]Fluoromethyl benzoate synthesis was further
`improved by changing the leaving group of the precursor mol-
`ecule from Tos to Nos (4-nitobenzene sulfonate) as demonstrated
`in a 1-step labeling reaction of peptides.
`47,48 In 2010, Olberg et al
`and Malik et al reported on the 1-step synthesis of the18F-labeled
`nicotinic acid derived active ester [18F]Fpy-TFP23, where the
`2,3,5,6-tetrafluorophenyl active ester group survived the labeling
`of the trimethylammonium moiety at the 6-position of nicotinic
`acid22.
`49,50 The 18F-fluorination surprisingly takes place at room
`temperature (Fig. 5). This PG23 has been used in the synthesis of
`a PSMA targeting ligand. The purification of the PG was conve-
`niently achieved by simple Sep-Pak cartridge separation, omitting
`the need for an HPLC system. The yields of the labeling reaction
`were surprisingly high with 60%-70% at just 40°C, and the con-
`secutive reaction with several peptides (eg, arginylglycylaspartic
`acid (RGD)) was almost quantitative at relatively low peptide con-
`centrations. The recently introduced non-canonical labeling
`approaches based on B-
`18F51,52 and Si-18F40,53 that are character-
`ized by an easy 1-step introduction of18F into small as well as
`complex organic compounds at very mild temperatures, led to
`the introduction of exotic PGs such as [
`18F]SiFB24,54,55 which is
`structurally very similar to [18F]SFB. One caveat, however, is the
`higher lipophilicity of18F-SiFB in comparison with [18F]SFB, lim-
`iting its use to protein labeling only , which can sufficiently compensate
`for the higher lipophilic character of the introduced silicon fluo-
`ride acceptor (SiFA) group.
`55 However, the synthesis of [18F]24is
`outstandingly convenient, relying on the concept of isotopic ex-
`change and one simple cartridge purification step only as a result
`of the chemical identity of precursor and labeled product (Fig. 5).
`Another
`PG bearing an active ester group has been published in
`2013, tackling the high lipophilicity of the transferred18F-fluoro
`benzyl group that can sometimes cause trouble when attached to
`small peptides.
`56 Although not based on a convenient 1-step reac-
`tion, this sulfonated hetero-bifunctional18F-crosslinking PG28 is
`based on the ring opening of a 1,3-propane sultone25 with18F−
`yielding26, its deprotection to compound27, and its final con-
`version into an N-hydroxysuccinimidyl ester28 (Fig. 6). The
`18F-ring opening yielded only low amounts of product under
`conventional18F-labeling conditions (acetonitrile, Kryptofix222,
`base) but works exceptionally well in tertiary alcohols (up to 90%
`radiochemical yield) at 90°C, a new labeling paradigm introduced
`by Kim et al.
`57 The sultone ring opening leads to the liberation of
`a sulfonic acid group, which imparts hydrophilicity to the PG.
`The PG 28 has only been used so far in the synthesis of an
`18F-labeled cyanine dye, and its usefulness for peptide and protein
`labeling has still to be proven. Shortly after, the same group applied
`a sultone-based labeling approach featuring a bis-sultone cross-
`linker to the synthesis of a peptide without implication of an
`active ester moiety .
`58 Another well-tolerated and synthetically proven
`18F-active ester was introduced by Guhlke et al in 1994,59 and
`(H3C)3N
`CF3SO3-
`O
`H
`18F-/Kryptofix 222
`DMSO
`O
`H
`18F
`KMnO4, NaOH O
`OH
`18F
`O
`O N
`O
`O
`18FNHS, DCC, THF
`(H3C)3N
`CF3SO3-
`O
`O
`Vaidyanathan et al. (41, 42)
`1. 18F-/Kryptofix 222
`DMSO
`2. TFA
`18F
`O
`OH
`TSTU
`active ester
`formation
`active ester
`formation
`Wüst et al (43)
`[18F]SFB
`H
`N
`R'
`O
`N
`H
`HOOC
`R
`NH2
`R'
`O
`N
`H
`HOOC
`R
`O
`18F
`H2O
`18F-SFB reaction with a simple di-peptide
`Figure 4Synthesis of [18F]SFB 15 and its application in radiolabeling of a dipeptide (box).
`479Small prosthetic groups in18F-radiochemistry
`Petitioner GE Healthcare – Ex. 1019, p. 479
`
`
`
`
`
`
`
`TosO O
`O N
`O
`O
`18F-
`leaving group
`activated ester
`a) b)
`a) b)
`18F O
`O N
`O
`O
`TosO O
`18F
`18F O
`OH
`hydrolysis hydrolysis
`HO O
`18F
`N(H3C)3N
`TfO-
`O
`O
`F
`F
`F
`F
`activated ester
`activated carbon
`18F-
`[18F]Fpy-TFP
`N18F
`O
`O
`F
`F
`F
`F
`Kryptofix 222
`acetonitrile
`RT
`[18F]SFMB
`O
`ON
`O
`O
`Si
`F
`SiFB
`Kryptofix 222
`acetonitrile
`RT
`isotopic exchange labeling
`O
`ON
`O
`O
`Si
`18F
`[18F]SiFB
`Si
`18F
`SiFA group
`high lipohilicity
`Figure 5Synthesis of different active ester PGs (18, 22, and 24) and possible side reactions (hydrolysis of 18 to 20 and 19 to 21).
`O
`OtBu
`O
`O
`SO O
`O
`OtBu
`O
`HO
`SO O
`R-OH, 90ºC
`Kryptofix 222/18F-
`Cs2CO3
`18F
`deprotection
`HCl, 80ºC
`O
`OH
`O
`HO
`SO O
`18Fin situ formation
`of active ester
`TSTU, DIEA
`acetonitrile, 50ºC, 5 min
`O
`O
`O
`HO
`SO O
`18F
`H3C
`OTos/Br
`O
`OR
`R= various ester groups
`Kryptofix 222/18F-
`K2CO3
`acetonitrile
`H3C
`18F
`O
`OR
`H3C
`18F
`O
`O-
`N OH-
`N,N'-disuccinimidyl carbonateH3C
`18F
`O
`O N
`O
`O
`N
`O
`O
`Figure 6Synthesis of active esters PGs 28 (sultone based) and N-succinimidyl-2-[18F]fluoropropionate 32.
`480 R. Schirrmacher et al.
`Petitioner GE Healthcare – Ex. 1019, p. 480
`
`
`
`
`
`
`
`used in the synthesis of 18F-peptides.60-62 The small PG
`N-succinimidyl-2-[18F]fluoropropionate 32 was used in the
`labeling of RGD dimer peptides, confirming its usefulness for
`peptide-based PET tracer development.63 The synthesis of32 starts
`from either the bromo or the tosyl precursor29, which is18F-labeled
`under standard conditions, yielding ester compound30. The ester
`moiety is cleaved by a base, and the resulting18F-acid31 is con-
`verted into the final active ester PG 32 (Fig. 6). A simplified
`1-step
`synthesis for32 has been added recently to the portfolio of
`active ester PGs.64
`Click Chemistry: Modern Tools for
`PG Development
`Besides recent metal-mediated late-stage18F-fluorination (see
`above), the facilitation and perfecting of click chemistry for
`18F-radiolabeling (and other radioactive tags) has been seri-
`ously changing radiochemistry over the last 2 decades after
`K. Barry Sharpless introduced the terminology to organic chem-
`istry in 2001.
`65-67 Click chemistry applications in all fields of
`radiochemistry have extensively been reviewed previously .68-73
`Especially ,18F-PG development has tremendously benefit-
`ted from the concept of click chemistry . The chemical efficiency
`of
`18F-PG bioconjugation has been maximized through the
`utilization of a wide contingent of click-chemistry reactions
`such as oxime formation, Michael addition (eg, thiol-
`maleimide conjugation), 1,3-dipolar Huisgen cycloaddition
`(eg, (Cu(I) assisted as well as catalyst-free triazole forma-
`tion), and more recently Diels-Alder-like reactions (eg, tetrazine-
`dienophile reactions). One outstanding feature of click
`chemistry is its orthogonality , which adds an unprec-
`edented chemoselectivity to PG conjugation to organic
`compounds that contain a large number of different func-
`tional chemical groups. A click chemistry reaction should be
`simple, fast, and modul