Spirocyclic Iodonium Ylides for Fluorine‐18 Radiolabeling of Non‐Activated Arenes: From Concept to Clinical Research

Positron emission tomography (PET) is a powerful imaging tool for drug discovery, clinical diagnosis, and monitoring of disease progression. Fluorine‐18 is the most common radionuclide used for PET, but advances in radiotracer development have been limited by the historical lack of methodologies and precursors amenable to radiolabeling with fluorine‐18. Radiolabeling of electron‐rich (hetero)aromatic rings remains a long‐standing challenge in the production of PET radiopharmaceuticals. In this personal account, we discuss the history of spirocyclic iodonium ylide precursors, from inception to applications in clinical research, for the incorporation of fluorine‐18 into complex non‐activated (hetero)aromatic rings.


Introduction
Positron emission tomography (PET) is a molecular imaging technique often used in the diagnosis and management of disease in neurology, oncology, and cardiology. [1] PET enables clinicians and researchers to investigate interactions between a molecule and its target (e. g., receptor, enzyme, protein aggregates, etc.) by introducing a positron-emitting radionuclide into the compound of interest. [2] Radioactive decay can be detected with a high degree of sensitivity, meaning that it is feasible to administer a sub-pharmacological dose of a radiotracer and image a biological system without perturbing its natural state -a phenomenon referred to as the "tracer principle". Moreover, the non-invasive nature of PET makes it an invaluable tool for drug discovery programs because it permits in vivo investigations of target engagement, pharmacokinetics, pharmacodynamics and metabolism. [3] Between aca-demic research, industry pursuits, and clinical applications, there is an ever-growing demand for novel PET radiotracers.
Many positron-emitting radionuclides are routinely used for clinical and research purposes (e. g., 11 C, 13 N, 15 O, 68 Ga, 89 Zr, 64 Cu), but fluorine-18 remains the most widely used PET radionuclide. [4] Fluorine-18 is most commonly prepared in medical cyclotrons by proton bombardment of an 18 Oenriched H 2 O target to produce an 18 O(p,n) 18 F nuclear reaction. This method provides no-carried added [ 18 F]fluoride ([ 18 F]F À ) in high molar activity (A m ), which is defined as the amount of radioactivity per mole of material (GBq/μmol or Ci/μmol). [5] Fluorine-18 has many advantageous physical characteristics, including its simple decay profile (97 % β + emission) and low positron energy (0.635 MeV), which contribute to high-resolution images. [4] Fluorine-18 also has a physical half-life (t 1/2 = 109.8 min) that allows for multi-step radiosyntheses, extended imaging protocols, and shipment of [ 18 F]fluoride and 18 F-radiotracers to sites without a cyclotron. It is estimated that approximately 20 % of commercial pharmaceuticals contain at least one fluorine atom, making them theoretically adaptable for 18 F-radiotracer development. [6] Unfortunately, most fluorination reactions used in organic chemistry are not suitable for 18 F-PET radiochemistry due to their relatively long reaction times, harsh reaction conditions, and limited substrate scopes. Fluorine-18 radiotracer synthesis relies upon the mild and selective fluorination of advanced intermediates to cater to the limited physical half-life of fluorine-18. Therefore, specialized radiofluorination methods have evolved over the past few decades to enable the production of diverse 18 F-radiotracers ( Figure 1).
Historically, fluorine-18 PET radiochemistry was restricted to nucleophilic aliphatic substitutions through halogen exchange or hydroxy-derived leaving groups (e. g., tosylates, triflates, mesylates), or nucleophilic aromatic substitution (S N Ar) on activated aromatic/heteroaromatic rings. During S N Ar, fluorine performs a nucleophilic attack at a carbonbearing leaving group (LG), leading to negatively charged intermediates called Meisenheimer complexes, which can only be sufficiently stabilized by electron-withdrawing substituents. [7] As such, radiofluorination of most non-activated (hetero)aromatic rings has been a long-standing challenge in the field. Traditional nucleophilic 18 F-fluorination reactions (e. g. Balz-Schiemann, Wallach, etc.) are seldom used due to low radiochemical yields (RCYs), elaborate precursor syntheses, harsh reaction conditions, isotopic exchange with 19 F often leading to low A m , and poor substrate scopes. [7] Other methods for 18 F-labeling of non-activated arenes include diarylsulfoxides, triarylsulfonium salts, and post-S N Ar functional group manipulations; however, these methods are limited by their substrate scope and/or utility. [7] For example, sydnones are primarily used as click reagents. Within the past few decades, a variety of efficient and selective methods for the nucleophilic radiofluorination of non-activated aromatics have been developed, including oxidative/deoxyfluorinations, dithiophene sulfonium salts precursors, and transition metalmediated reactions of boronic acids/esters or organostannane precursors. [8] These methods have demonstrated excellent reactivity, and some have been implemented for radiopharmaceutical production in clinical research, but none are without limitations (e. g., poor substrate scopes, air-sensitive and/or toxic components, unstable precursors, etc.). Hypervalent iodonium precursors are another relatively recent development for the preparation of electron-rich 18 F-arenes. Diaryliodonium salts were adopted early on and applied for 18 F-labeling of several compounds. [8] However, due to competitive radio-labeling on the ancillary arene ring, this method suffers from poor regioselectivity. Precursor instability and the generation of volatile radioactive by-products also contribute to limited RCYs.
In this personal account, we will discuss work from our laboratories and others on the development of spirocyclic iodonium(III) ylides (SCIDY) precursors for radiolabeling electron-rich aromatic rings and their successful translation for clinical use. This account is intended to build upon recent reviews by our group and others on advancements in PET radiochemistry methodologies. [8]

Early Iodonium Ylide Radiochemistry
Radiofluorination of non-activated aromatic rings using iodonium ylide precursors bearing β-dicarbonyl auxiliaries was conceptualized by Satyamurthy and Barrio in their 2010 patent application. [9] Synthesis of these hypervalent iodonium(III) ylide precursors proceeds through oxidation of the iodoarene, followed by coupling to the appropriate auxiliary under basic conditions. The patent claims a wide range of auxiliaries, with the most promising being derived from Meldrum's acid due to their high reactivity and comparative stability (Figure 2A). The radiolabeling conditions detailed in the patent are similar to those typically used in Melissa Chassé is currently pursuing her Ph.D. at the University of Toronto (Toronto, Canada) in Dr. Vasdev's lab. Her research focuses on developing PET radiotracers for novel targets, including rare diseases. She is also interested in new methodologies for radiolabeling with carbon-11 and fluorine-18.  Figure 2B). [10] The method was able to adequately label both precursors with A m of~50 GBq/ μmol and RCY of 20 % relative to starting [ 18 F]fluoride. However, significant amounts of an undesirable regioisomer (4-11 % RCY) are produced with this method through aryne decomposition products, thereby complicating purification and lowering yields (Figure 2A). The Meldrum's acid-based precursors were subsequently applied to the synthesis of a neuronal nitric oxide synthase inhibitor ([ 18 F]1, RCY 15 %, A m 48 GBq/μmol) and a dopamine D4 receptor antagonist ([ 18 F]FAUC-316, RCY 10 %, A m 90 GBq/μmol), but they were not pursued for biological evaluation. [11] Concurrently with Cardinale and colleagues, our lab was investigating the use of various spirocyclic iodonium ylide (SCIDY) auxiliaries for fluorine-18 labeling (Figure 3). [12] Synthesis of this range of precursors can be attained through oxidation of the iodine atom, followed by coupling with the

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
desired auxiliary acid. We reported six oxidative conditions suitable for precursor synthesis, each with its own merits ( Figure 3A). The 3-chloroperoxybenzoic acid (mCPBA), NaBO 3 , and urea-hydrogen peroxide methods oxidize simple substrates extremely well. The dimethyldioxirane (DMDO)/ acetone and Selectfluor methods function efficiently at neutral pH, making them well-suited for acid-sensitive substrates. The Oxone conditions can be applied to substrates containing tertiary amine or pyridines in the presence of trifluoroacetic acid (TFA). Iodonium ylide precursors with Meldrum's acid and barbituric acid-based auxiliaries provided radiochemical conversions (RCCs) of up to 47 % and 53 %, respectively, when applied to radiofluorination of a simple model compound ( Figure 3B). [12] However, barbiturate derivatives bearing substituents on the nitrogen atoms and diesters bearing nonmethyl alkyl groups on the quaternary carbon exhibited poor stability and/or were thick oils that were difficult to handle. The inability to derivatize upon these structures pushed us to explore more exotic auxiliary structures, which led to the discovery of a series of stable and easily purified crystalline spirocyclic auxiliaries. The radiolabeling efficiency of these novel auxiliaries was highest with the spirocyclopentyl derivative (SPI5), which gave an RCC of 85 % for the model compound. Further substrate scope investigation demonstrated consistent one-step radiolabeling of a variety of non-activated 18 F-aromatics, including hindered alkyl substituents, benzyl azides, anisoles, amides, heterocycles, and halogenated aromatics in moderate-to-excellent RCYs ( Figure 4A). To demonstrate the practical utility of this approach for radiopharmaceutical productions, we applied the new method to the synthesis of several radiotracers that were previously inaccessible by [ Figure 4B). [12,13] Subsequent studies from our group and others also demonstrated the utility of this method for the synthesis of radiolabeled bioconjugation reagents, including 4-[ 18 [14,15,16] Importantly, this technique overcame the regioselectivity issues observed with the original Meldrum's and barbituric acid auxiliaries, making this methodology suitable for routine use. Mechanistic studies using density functional theory (DFT) and multi-NMR revealed that the regioselectivity of the spirocyclic iodonium(III) ylide precursors arises from the charge distribution created by the carbanionic character formed during reductive elimination of the auxiliary, producing a charge distribution which disfavours radiofluorination at the auxiliary site. [13] Not only do these results further support the experimental results observed in the initial paper, but the DFT results were consistent with those reported by Hill and Holland. [12,17] Our modeling of the reaction pathways suggested that reductive elimination is the rate-determining step in iodonium-mediated radiofluorination, with electron-withdrawing groups dramatically lowering the barrier to reductive elimination of the corresponding fluoroarenes ( Figure 3B). Computational modelling illustrated that ortho-substituted arenes force the arene to rest outside of the FÀ IÀ C plane due to steric interactions with the ortho-substituent. As the reductive elimination proceeds through an out-of-plane tran-

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
sition state, the ortho-substituted aromatic does not pay the energetic cost of arene rotation prior to reductive elimination, which explains the comparatively higher yields of orthosubstituted arenes compared to their alternately substituted counterparts. Our computational model also demonstrated that auxiliary effects on radiofluorination of arenes do not stem from the reductive elimination step, and likely arise from rates of precursor decomposition during radiolabeling. From this information, we devised a new, bulkier spiroadamantyl-1,3-dioxane-4,6-dione (SPIAd) auxiliary, where the insensitivity of reductive elimination to the bulky auxiliary improved precursor stability during radiolabeling of challenging, electron-rich iodonium ylides ( Figure 5A, 5B). Substrate scope of SPIAd is shown in Figure 5C. In addition to the easy automation of this radiolabeling method on commercial automated radiosynthesis modules, iodonium ylide precursors bearing SPI5 auxiliaries were adopted for 18 F-labeling via continuous-flow microfluidics. Microfluidics methodologies can benefit radiochemistry by enabling low precursor consumption, fast kinetics, improved reproducibility, enhanced RCYs, and facile automation. In order to demonstrate the suitability of SCIDY precursors for continuous-flow microfluidics, we published a proof-of-concept study demonstrating the automated synthesis of radiopharmaceuticals and 18 F-labeled building blocks. [18]

Radiotracer Applications of SCIDY
The ability to perform efficient, regioselective radiolabeling of non-activated (hetero)aromatics unveiled new radiochemistry possibilities for a whole class of previously inaccessible compounds. This led to the rapid application of SCIDY chemistry for preclinical imaging studies in various indications. Many aspects of our spirocyclic iodonium ylide methodology make it suitable for clinical translation, including the lack of toxic transition metals (metal-free conditions), small amounts of precursor required, excellent regioselectivity, the high isolated RCYs, and precursor shelf-stability. As such, many centers have adopted this method for routine radiopharmaceutical production.

Neuroimaging
Nerve signals are transmitted between neurons via neurotransmitters, which get released from the presynaptic neuronal terminals into the synapse, where they then act upon postsynaptic receptor sites to excite or inhibit the postsynaptic neuron. Signaling effects of the neurotransmitters will persist until the extracellular excess is removed from the synapse through reuptake transporters on the presynaptic membrane. PET radiotracers for imaging within the central nervous system (CNS) have historically focused on neuroreceptors due to their prominent role in many neuropsychiatric conditions. [2] Radiotracers targeted towards presynaptic transporters and/or postsynaptic receptors continue to be an invaluable tool for clinical researchers and the pharmaceutical industry to quantify these targets in living human brains in order to interrogate their role in the pathophysiology of neurological illness. The first portion

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
of this section will focus on recent efforts to label neuroreceptor-targeted radiotracers using SCIDY chemistry. Our group was the first to validate SCIDY methodology for human use with the metabotropic glutamate receptor subtype 5 (mGluR 5 ) targeted radiotracer [ 18 Figure 6A). [20] While [ 18 F]FPEB had been applied for clinical research prior to the advent of SCIDY chemistry, traditional radiofluorination of the insufficiently activated nitro-precursor requires elevated temperatures and produces several radiochemical impurities, resulting in low isolated RCYs (< 5 %). Coppermediated radiofluorination methods using boronic acid and ester precursors were reported to provide [ 18 F]FPEB in 8 � 2% RCC and 13 � 5% RCY, respectively. [21] Alternative methods, including 18 F-deoxyfluorination of phenols and Cu-mediated labeling of organostannane precursors, demonstrate RCCs of 11-75 %, but these methods of synthesizing [ 18 F]FPEB were not automated, did not provide values for isolated yields, and have not been validated for human use. [21] SCIDY radiolabeling conditions enable radiofluorinations to be conducted under much milder conditions compared to radiolabeling of the nitro-precursor, which minimizes the amount of radiochemical impurities. We first reported the synthesis of [ 18 F]FPEB using an iodonium ylide precursor with a SPI5 auxiliary to achieve non-decay corrected, isolated RCYs of 20 � 5% with A m of 666 GBq/μmol in an automated, one-pot synthesis. [20] The radiosynthesis of [ 18 F]FPEB using SCIDY chemistry resulted in vastly improved RCY and more reliable productions compared to the reported nitro-precursor, so this method was validated for human use at our site. Notably, the same SPI5-based precursor was also radiolabeled via continuous-flow microfluidics ([ 18 F]TEAF, DMF at 200°C, 1 min) with a RCC of 68 � 5%, although the product was not isolated. [18] While performing routine re-validation of the [ 18 F]FPEB synthesis at our site, we decided to perform a headto-head comparison of five different metal-free precursors; chloro-, nitro-, SPI5, SPIAd, and the sulfonium salt. [20] Consistent with previous reports, both the chloro-and nitroprecursors were radiolabeled in RCYs of < 5 %. During the head-to-head comparison, radiolabeling of the SPI5-SCIDY precursor was further optimized, resulting in RCYs of 25 � 2% and A m of 37 � 13 GBq/μmol. The SPIAd precursor performed similarly, with 24 % RCY and a A m of 21 GBq/μmol, however semi-preparative HPLC conditions were not optimized to separate the adamantyl precursor from the radiolabeled product. Interestingly, the sulfonium salt first reported by Gendron and colleagues performed the best in our hands, with RCY of 36 � 6% and A m of 77 � 35 GBq/μmol. [22] Thus, the SPI5 auxiliary and the sulfonium salt precursors were best suited for routine radiopharmaceutical production of [ 18 F]FPEB, although to the authors' knowledge, the sulfonium salt precursor has yet to be validated for human use.
The k-opioid receptor (KOR) is well-known to play a crucial role in analgesia, but it has more recently been found to affect mood regulation, reward systems, general cognition, and tumour progression. [23] There has been increasing interest in the KOR as a potential pharmaceutical target for various   Figure 6B). [19] The nitro-and iodo-precursors were assessed first due to the comparatively easier precursor synthesis and commercial availability of the building blocks. However, RCYs were consistently low (< 1 %). The group was able to synthesize the SPI5 precursor through oxidation and auxiliary coupling, which proceeded in 45 % yield. After extensive optimization, it was found that radiolabeling and hydrolysis conditions were most effective when heating the mixture of has the added benefit of a longer half-life, which enables better imaging at later time points and distribution to off-site locations for use in multicenter clinical trials. The first-inhuman results using this tracer have been recently disseminated at scientific symposia, but have yet to be published. The sigma-1 (σ 1 ) receptor is a chaperone protein that regulates Ca 2 + secondary signaling at the surface of the endoplasmic reticulum. [24] Disruptions in σ 1 function have been associated with neurodegenerative conditions, cardiovascular function, neuropsychiatric disorders, and several cancers. For many years, no ideal radioligand for CNS imaging of σ 1 had advanced to the clinic, primarily due to common issues with slow washout kinetics. [ 18 F]FBFP demonstrates high σ 1 affinity and selectivity over other CNS receptors, as well as favourable rodent PET imaging characteristics, including high binding specificity, good brain uptake, minimal metabolism, and vastly improved tissue kinetics compared to previously reported radiotracers for this target. In order to simplify the synthesis, which was originally a 2-step procedure, [ 18 F]FBFP was synthesized using SCIDY chemistry ( Figure 6C). [25] The SPI5-precursor was synthesized using the oxone method in 63 % yield. Radiolabeling was performed by heating a mixture of [ 18 F]fluoride, TEAB, and the SCIDY precursor in DMF to 150°C for 15 min. [ 18 F]FBFP was prepared in RCYs of 5.6-9.2 % and with A m of 21.5-1300 GBq/μmol. RCYs were lower than the previous two-step method (up to 30 % RCY), however, the SCIDY labeling conditions were not optimized. Notably, [ 18 F]FBFP was produced as a racemate. Enantiomeric differences in washout kinetics have previously been observed with σ 1 radiotracers. [26] In a subsequent report, the 18 F-labeled enantiomers were separately synthesized for further characterization. [25] No difference was observed in the radiosynthesis of the enantiomers from their enantiopure SCIDY precursors (RCY 24.4 � 2.6 %, A m 86-214 GBq/μmol). The only notable difference discovered during biological characterizations was that (S)-[ 18 F]FBFP cleared from the brain significantly faster than (R)-[ 18 F]FBFP during rodent PET imaging studies. Given the improved washout kinetics, the (S)-[ 18 F]FBFP enantiomer has been validated for human use and currently being used for clinical trials (NCT05335200).
Serotonin 5-HT 2A receptors (5-HT 2A R) are not only involved in several brain disorders (e. g. depression, Alzheimer's disease, schizophrenia), but 5-HT 2A agonism is a hallmark of understudied psychedelic compounds. [27] As such, there has been growing interest in PET imaging of 5-HT 2A receptors for clinical research. Petersen and coworkers have synthesized 18 Flabeled derivatives of their lead 5-HT 2A R agonist radioligand, [ 11 C]Cimbi-36. [28,29] Four radiotracers in their fluorine-18 series have been labeled using SPI5 precursors with decaycorrected RCYs of 8-15 %, in some cases using multi-step radiosyntheses to access structures that could not be directly labeled. [28] Interestingly, this structural class, including [ 18 F]Cimbi-198 and similar derivatives, are reported to have improved RCCs in the presence of the radical scavenger 2,2,6,6-Tetramethylpiperidin-1-yl)oxyl (TEMPO) (Figure 6D). This is in contrast to our original paper, where both TEMPO and butylated hydroxytoluene (BHT) were assessed and demonstrated no improvement on radiolabeling efficiency. [12] Radical scavengers may be advantageous to radiofluorinate SCIDY precursors and should be determined on a case-by-case basis.
Another instance of SCIDYs being applied to the preclinical development of neuroreceptor-targeted radiotracer is [ 18 F]N2B-0518 ( Figure 6E). [30] PET imaging of N-methyl-D-aspartate receptors (NMDARs) is of interest because these receptors are known to play a role in various neurological and neuropsychiatric conditions, including cerebral ischemia, acute and chronic pain, schizophrenia, and Alzheimer's disease. [31] NMDARs are comprised of two GluN1 subunits and two GluN2 subunits. The GluN2 subunits include four subtypes (GluN2 A-GluN2D), which have distinct expression patterns and functional properties. Fu and coworkers developed a series of GluN2B antagonists based on a new structural class disclosed in a patent by Janssen Pharmaceuticals. [30,32] [ 18 F]N2B-0518 was chosen for radiolabeling based on its high potency and selectivity in vitro, in addition to its optimal calculated lipophilicity and topological polar surface area. The SPIAd precursor was pursued for [ 18  The authors conclude by stating that further work will be pursued in higher species to investigate if interspecies differences may play a role in brain uptake and metabolic rate in the plasma. Synaptic vesicle glycoprotein 2 A (SV2A) is a synaptic vesicle membrane protein involved in neurotransmission from presynaptic to postsynaptic neurons. Disturbances in synaptic signaling involving SV2A has been implicated in various neurological and neuropsychiatric conditions, including dementia, epilepsy, autism spectrum disorder, depression, and schizophrenia. [33] Studies on changes in synaptic density have historically been limited to histological evaluation of excised or post-mortem brain tissue. However, a recent surge in interest on imaging synaptic density has led to the development of several radiotracers that target SV2A. [34] Among these are [ 18 F]SynVesT-1 and [ 18 F]SynVesT-2, two structurally related 1-((3-methylpyridin-4-yl)methyl)pyrrolidin-2-ones (Figure 7). [35,36] The SPI5 precursors for both radiotracers were derived from their respective iodinated analogs using the oxone method (20-30 % yield). Radiolabeling of both precursors required heating to 150°C, which gradually racemized the enantiopure starting material. Radiolabeling was performed on the racemic precursor, followed by HPLC purification on a C18 column, followed by a reverse phase chiral column to separate the two enantiomers. While the final products were confirmed to be enantiomerically pure, the tandem HPLC purification resulted in more than 75 % of the radioactivity being lost during purification of [ 18 F]SynVesT-1; the radioactivity lost during chiral purification was not reported for [ 18 F]SynVesT-2. [34,36] Ultimately, the isolated RCY was < 1 % in both cases, which was not sufficient for clinical research studies. In comparison, the racemic boronic ester precursor of SynVesT-1 and the trimethylin precursors of both radiotracers can be readily radiolabeled at mild temperatures that do not induce racemization and, therefore, do not require chiral separation. Moreover, the organostannane precursors demonstrated significantly improved RCY for the synthesis of both Since the turn of the millennium, there has been a surge of interest in using high-affinity antagonists, and occasionally agonists, for imaging neuroreceptors. [2] However, relatively few enzymes have been the target of PET radiotracer development. The remaining portion of the neuroimaging section will highlight the impact that SCIDY chemistry has had on the development of fluorine-18 radiotracers for imaging CNS enzymes.
Phosphodiesterase 10 A (PDE10A) is one isoform of the many phosphodiesterase enzymes that regulate the intracellular concentrations of certain secondary messengers, such as cyclic adenosine monophosphate (cAMP) and cyclic guanosine monophosphate (cGMP). [37] PDE10A is highly expressed on medium spiny neurons in the striatum. The enzyme is involved in dopaminergic and glutamatergic signal transduction through the striatum, so its inhibition holds therapeutic potential in disorders caused by dysfunction of the basal ganglia, including Parkinson's disease, obsessive-compulsive disorder, and Tourette's syndrome. Several carbon-11 and fluorine-18 PET tracers for imaging PDE10A have been pursued previously, but most suffer from one of the following: low specific binding, poor metabolic stability, overly slow pharmacokinetics, or difficulty automating the radiosynthesis. In hopes of overcoming these challenges, Xiao and colleagues developed the PDE10A PET radiotracer, [ 18 F]P10A-1910, a fluoroaryl derivative of the leading radiotracer at the time, [ 18 F]MNI659 ( Figure 8A). [38] From the iodinated intermediate, the SPIAd precursor was formed in 72 % yield using oxonemediated oxidation and coupling. In this case, radiolabeling was optimal in N,N-dimethylacetamide (DMAc) at 140°C for 5 min with small amounts of base (TEAHCO 3 ) to minimize precursor degradation. These conditions resulted in RCYs of 25 � 5% with A m of over 111 GBq/μmol. Through in vitro binding assays, digital autoradiography, and in vivo PET imaging, the novel derivative was consistently shown to be highly specific, selective and metabolically stable. Doseresponse studies in nonhuman primates showed high radiotracer uptake in the striatum with a dose-dependent signal reduction in response to escalating doses of a structurally dissimilar PDE10A inhibitor. This radiotracer is currently being translated for human use (personal communication: Dr. Yiyun (Henry) Huang).
11β-Hydroxysteroid dehydrogenase isoform 1 (11β-HSD1) regulates glucocorticoid interactions with steroid receptors by catalyzing glucocorticoid conversion to their inactive 11-keto forms. [39] Due to its association with diabetes, metabolism, and cognitive decline, 11β-HSD1 is a target of

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
interest for investigating disease mechanisms and developing novel therapeutic treatments. Promising human imaging data was acquired with [ 11 C]AS2471907, but the synthesis was generally unreliable, low yielding, and unsuited for widespread clinical research due to the carbon-11 radiolabel. [40] The first fluorine-18 PET radiotracers for imaging 11β-HSD1, para-[ 18 F]AS2471907 and ortho-[ 18 F]AS2471907, were developed as alternatives to the carbon-11 isotopologue ( Figure 8B). [41] Baum and coworkers reported the synthesis of two SPI5 precursors to enable radiolabeling at the para-and orthopositions of the trifluoroaryl moiety. Both precursors were synthesized in yields of 58 % by oxone-mediated oxidation of the iodine intermediates to their corresponding iodo(III) intermediates, which were then coupled to their respective auxiliary. Optimal radiolabeling conditions for para- Monoacylglycerol lipase (MAGL) is an enzyme primarily known for degrading 2-arachidonoylglycerol, a prominent lipid signaling molecule within the endocannabinoid system. Liang's group recently reported the synthesis of an irreversible 18 F-labeled MAGL PET tracer containing a novel azabicyclo[3.1.0]hexane scaffold as a more accessible alternative to the leading 11 C-labeled radiotracers for this target. [42] The fluorine-containing lead, PF06795071, was identified through in vitro activity-based protein profiling assays and molecular docking studies. Radiolabeling was achieved through the nucleophilic substitution of the SPIAd precursor with [ 18 F]TEAF, followed by acid deprotection of the PMB group ( Figure 8C). This provided [ 18  demonstrated high brain uptake and specific binding, with appropriate distribution in the brain. This radiotracer is currently under investigation in nonhuman primates.
Safinamide is a reversible monoamine oxidase B inhibitor with secondary ion channel blocking effects that is approved

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D by the Food and Drug Administration and the European
Commission as an additive treatment for Parkinson's disease in conjunction with levodopa. [43] We reported the first synthesis of [ 18 F]Safinamide through a two-step labeling, whereby an N-Boc-protected SPIAd precursor was labeled with [ 18 F]fluoride and deprotected with acid, resulting in a RCC of 15 % ( Figure 5C). [13] A recent paper by Xu Figure 8D). [ F]FDG PET may be unsuitable for a given indication include high background noise in a given tissue (e. g. brain), low metabolic activity of certain cancers (e. g. prostrate), and non-specific radiotracer uptake resulting from certain treatments which may cloud image interpretations (e. g. inflammatory lesions). [45] It logically follows that there is a strong desire for radioligands targeting tumour-type specific features, which has led to an increase in novel fluorine-18 radiotracers for oncology imaging. This section outlines instances in which SCIDY precursors have been utilized to support novel fluorine-18 radiotracer developments in oncology. Many CNS and neuroendocrine tumours over-express the norepinephrine transporter (NET), a presynaptic norepinephrine re-uptake protein. The most frequently used radiotracer for imaging NETs is the single photon emission computed tomography (SPECT) agent norepinephrine derivative meta-[ 123 I]iodobenzylguanidine ([ 123 I]MIBG). [46] While [ 123 I]MIBG is considered the current gold standard for imaging NETs in oncology, imaging with this radiopharmaceutical has several practical limitations, including a 2-day imaging protocol, requisite thyroid blockade, and the limited resolution of SPECT compared to PET imaging. All of these issues would be easily addressed through the use of a suitable fluorine-18 analogue of [ 123 I]MIBG. Meta-[ 18 F]fluorobenzylguanidine ([ 18 F]mFBG) was originally synthesized through a 3-step nitroexchange process with overall, decay-corrected RCYs of 10-15 % in a synthesis time of 60 min. [47] Zhang and colleagues reported an improved 3-step synthesis in 2014 based on S N Ar with a trimethylammonium salt precursor, resulting in decaycorrected RCYs of 11 � 2% over 3 hours. [47] Both of these multi-step processes require several manual extractions, which complicates their automation. Subsequently, Hu and coworkers established an automated synthesis of [ 18 F]mFBG using a two-step synthesis from a diaryliodonium salt precursor. [47] This method resulted in decay-corrected RCYs of 21.7 % � 3.5 % and a synthesis time of~1 hour. [47] A one pot, two-step method using copper-mediated radiofluorination of the boronic ester precursor was investigated more recently, achieving non-decay corrected RCYs from 7-25 %. [21,47] Of these methods, the methods by Zhang and Hu have been applied for human use; the latter method is preferable due to the automated synthesis, moderate yields, manageable synthesis times, and commercially-available precursor. The first report of a SCIDY-based radiosynthesis for meta-[ 18 F]fluorobenzylguanidine ([ 18 F]mFBG) was reported by our group to demonstrate the utility of SPIAd-SCIDY chemistry. [13] The SPIAd precursor was synthesized using the Selectfluor oxidative route, which, when followed by auxiliary coupling, resulted in isolated yields of 54 %. The automated radiosynthesis was initially optimized for preclinical cardiac imaging (section 3.4), and recently validated for human use in clinical neuro-oncology research. [48] The one-pot, two-step synthesis involves direct radiolabeling of the SPIAd precursor using [ 18 F]TEAF in DMF, followed by an acid deprotection step for non-decay corrected RCY of 10.2 � 2.7 % and A m of 128 � 53 GBq/μmol ( Figure 9A). Preclinical applications have reported non-decay corrected RCYs up to 24 % with the SCIDY precursor (section 3.4), which is higher than those reported for the previously described method of Hu and coworkers.
Our lab has also applied SCIDY chemistry for the development of novel small molecule radiotracers for oncology imaging. Lorlatinib is a second-generation inhibitor of the anaplastic lymphoma kinase (ALK) and the orphan receptor tyrosine kinase c-ros oncogene 1 (ROS1) intended as a brainpenetrant therapeutic for ROS1-driven fusion cancers and related brain metastases. [49] In 2017, we published in collaboration with Pfizer on the synthesis of both carbon-11 and fluorine-18 isotopologues of lorlatinib to aid ongoing clinical trials and image non-small cell lung cancer (NSCLC) brain metastases. [50] Initial PET studies with the carbon-11 isotopo-

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
logue confirmed that the radiotracer penetrated the bloodbrain barrier in nonhuman primates, and warranted the development of a fluorine-18 radiotracer for widespread use. To synthesize [ 18 F]lorlatinib, we initially attempted S N Ar reaction of [ 18 F]fluoride with the nitro precursor, but found several issues, including poor RCC (~1 %), nitrile decomposition, and difficult purification. Fortunately, the iodonium ylide precursor bearing an SPI5 auxiliary was successfully synthesized by Boc-protecting the nitro-precursor, followed by sodium dithionite reduction, and diazotization/iodination. The aryl iodide was then converted to the SPI5 precursor as previously described, generating the desired, enantiopure product in 36 % isolated yield. Radiolabeling of the SCIDY precursor resulted in simplified purification and non-decay corrected RCY of 14 % relative to starting [ 18 F]fluoride ( Figure 9B). A recent study by Sardana and coworkers synthesized [ 18 F]crizotinib, [51] a fluorine-18 isotopologue of the firstgeneration ALK/mesenchymal-epithelial transition factor/ROS inhibitor crizotinib, which is an approved chemotherapy for treating NSCLC ( Figure 9C). [52] Crizotinib is unable to treat NSCLC brain metastases, in part because it is a substrate for the P-glycoprotein drug efflux transporter. By assessing radiotracer uptake in response to experimental treatments intended to improve brain penetrance, radiolabeled crizotinib could help develop and validate new strategies to improve its therapeutic effects on brain metastases. The precursor bearing the SPI5 auxiliary was synthesized in a yield of 15 % over two steps, using the comparatively milder reaction conditions attainable through 3,3-dimethyldioxirane (DMDO) oxidation to avoid side reactions on the pyridine core. The moderate isolated yield was, in part, due to difficulty in separating the product from the iodinated starting material. Optimized radiolabeling conditions led to RCCs of 64 % and isolated RCYs of 13 � 3%, with no observable racemization. It should be noted that TEMPO did not affect the yields observed in this case. In vivo imaging in nonhuman primates demonstrated pharmacokinetics and pharmacodynamics that align with previous reports, including the expected poor brain penetration.
In a series of collaborations led by the Liang and Chen labs, we applied SPI5-auxiliaries to the labeling of single-strand DNA aptamers targeting protein tyrosine kinase 7 (PTK7). [15] PTK7 over-expression has been correlated to proliferation, invasion, and migration in various cancers. Our first study in this series focused on the efficient, one-step radiolabeling of [ 18 F]FBA and its subsequent conjugation to the alkynylated aptamer, Sgc8. [15] Radiosynthesis of [ 18 F]FBA was historically synthesized in a 75 min, 4-step process that gave decay corrected RCYs of 34 %, which could be raised to 60 % after automation using solid supports. [53] Further efforts to make this radiolabeled click reagent in a single-step employed microfluidic chemistry to radiolabel the diaryliodonium salt precursor with RCCs of 35-45 %, although the product was never isolated. [53] We first reported the use of SCIDY chemistry for radiosynthesis of [ 18 F]FBA (section 2), but this one-step radiolabeling of [ 18 F]FBA saw its first application in this Sgc8 study. Using the respective SPI5-SCIDY precursor, [ 18 F]FBA was produced in RCYs of 35 � 3% and A m up to 148 GBq/μmol ( Figure 10A). The isolated [ 18 F]FBA intermediate was then conjugated to the alkynylated Sgc8 using classic click-chemistry, which proceeded with non-decaycorrected RCYs of 62 � 2% (based on the starting amount of [ 18 F]FBA) to produce [ 18 F]FÀ Tr-Sgc8 ( Figure 10C). The radiolabeled aptamer demonstrated high affinity binding in vitro and readily visualized high PTK7 expression HCT116 cells in subcutaneous tumour (0.76 � 0.09 %ID/g) and liver metastasis (1.5-2.0 %ID/g) models. In this study, [ 18 F]FBA was produced from the SPI5-SCIDY precursor in high RCYs, and subsequent bioconjugation proceeded in sufficient yields for biological evaluation. However, the high volatility of [ 18 F]FBA is not ideal for radiolabeling procedures.

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
Our subsequent publication focused on the development of improved radiolabeled prosthetic groups for bioconjugation via click chemistry. [16] In light of the previously described ortho-stabilizing effect on the hypervalent iodine, we designed a range of radiolabeling precursors with varying linker lengths, positions, and ortho-coordinating substituents. Additionally, substituents para to the iodonium ylide group were investigated as a means to decrease the volatility of these prosthetic groups. The results were highly thermostable, non-volatile crystalline precursors which could be readily radiolabeled in high RCYs under standard one-step [ 18 F]fluoride displacement reactions from the corresponding SPI5-SCIDY precursor. The lead azide chosen for further investigation ([ 18 F]2, Figure 10B) demonstrated excellent RCYs of 82 � 5%, and good RCCs in both copper-mediated click reactions and strained-induced DBCO cycloadditions. The new radiolabeled click agent [ 18 F]2 was bioconjugated with two alkyne-linked aptamers (TsC and Sgc8) in high coupling efficiencies and RCYs (5 0 %), readily outperforming previous attempts at aptamer radiolabeling using [ 18 F]SFB and [ 18 F]FBA ( Figure 10C). [15,54] The 18 F-labeled Sgc8 progressed to in vivo evaluation, clearly visualizing PTK-7-positive tumors and demonstrating good radiometabolic stability with negligible defluorination.
In spite of the advantages of the novel ortho-oxygenstabilized iodonium derivatives, many labs may prefer to continue using [ 18 F]FBA due to its established use. For example, a recent study applied SPI5 SCIDY precursors for the efficient synthesis of [ 18 F]FBA and subsequent click chemistry to epidermal growth factor receptor (EGFR) targeted aptamers for imaging glioblastoma. [55] This work created two novel aptamer radiotracers, [ 18 F]FB-GR20 and [ 18 F]FB-GOL1, in non-decay corrected RCYs calculated from starting [ 18 F]FBA of up to 67 % and 74 %, respectively. According to the authors, further research for evaluating these tracers in malignant glioma animal models are underway. Another example is the development of HER2-targeted aptamers, named Heraptamers, although the RCC, RCY and A m of these materials were not reported. [56]

Inflammation
Inflammation is the immune system's front-line defense against harmful pathogens and tissue damage. Inflammatory responses trigger the recruitment and/or activation of relevant immune cells (e. g. neutrophils, macrophages, microglia, astrocytes), which may produce proinflammatory cytokines and chemokines. Inappropriate inflammatory responses are implicated in various human diseases, including ischemic stroke, Alzheimer's disease, atherosclerosis, and autoimmune diseases. [57] A major target of interest for imaging inflammation, especially within the CNS, is the translocator protein (18 kDa) (TSPO). TSPO is a mitochondrial protein that may be involved in immunomodulation and inflammatory response. Several TSPO-targeted radioligands have been developed for imaging inflammation, including many radiopharmaceuticals used in clinical research. [58] The TSPO radiotracer [ 18 F]DPA-714 has shown promise in clinical research studies for imaging neuroinflammation in amyotrophic lateral sclerosis, stroke, and Alzheimer's disease. [ [59] Nucleophilic radiofluorination of the aforementioned precursors resulted in poor RCCs (< 3 %). On the other hand, electrophilic radiofluorination proceeded with adequate RCYs of 15 � 3%, but because [ 18 F]Selectfluor must be synthesized from carrieradded [ 18 F]F 2 gas, the final A m was not ideal (7.8 � 0.5 GBq/ μmol). To remedy this issue, we collaborated with Wang and coworkers to develop a SCIDY precursor that would enable high RCY and A m for neuroimaging TSPO ( Figure 11A). [60] [61] In 2003, Zhang and coworkers synthesized and evaluated [ 11 C]DAA1106, a second-generation PET agent for imaging TSPO in clinical research studies of dementia and schizophrenia. [62] The fluorine atom native to the electron-rich arene ring of DAA1106 made it a perfect candidate for radiofluorination. The group was able to manually synthesize [ 18 F]DAA1106 through the diphenyliodonium salt (46 % RCC), but did not isolate the product, and found the precursor to be unstable. [63] Other methods for synthesizing [ 18 F]DAA1106 have been investigated, but they rely on the use of copper complex catalysts (59 % RCY), toxic organostanne precursors (31 % RCC), large amounts of precursor, and manual labeling techniques. [21,64] In an effort to translate [ 18 F]DAA1106 for human use, the group employed a SPIAd precursor for radiolabeling with [ 18 F]TBAF ( Figure 11B). [65] The automated radiosynthesis was achieved with A m of 60-100 GBq/μmol and non-decay corrected RCYs of 6 � 2 % based on starting [ 18 F]fluoride. Biodistribution studies with [ 18 F]DAA1106 showed no signs of in vivo defluorination, and small amounts of radiolabeled metabolite in the brain (< 5 %). PET imaging in rodent models of ischemic stroke confirmed the distribution and specificity of [ 18 F]DAA1106 for TSPO. The authors state that this method would be used to produce [ 18 F]DAA1106 for clinical research at their imaging center.

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
It should be noted that, while many TSPO radiotracers have been advanced for human use, the majority have limitations (e. g. metabolic instability, high non-specific binding, sensitivity to a common single nucleotide polymorphism). As such, the search for novel stable fluorine-18 tracers for TSPO is ongoing. [58] A study by Fujinaga and colleagues sought to improve the metabolic stability of their previous 18 Fleads for imaging TSPO, which suffered from 18 F-defluoroalkylation in vivo, by modifying the structures to contain a more stable [ 18 F]fluorobenzene moiety. [66] The authors proceeded with the SPIAd precursors on account of their improved chemical stability and regiospecific radiolabeling with [ 18 F]fluoride. The radiosynthesis of [ 18 F]F3 was successfully automated in decay-corrected RCYs of 23 � 10 % (Figure 11C). In vivo analysis in healthy mice revealed significantly improved radiometabolism of [ 18

Miscellaneous
Sympathetic cardiac nerve dysfunction is characteristic in several cardiological conditions, including cardiomyopathy, heart disease, and heart failure. Cardiac denervation predisposes patients to ventricular arrhythmia, which can go on to cause ischemic cardiomyopathy. Therefore, regional cardiac denervation can be used as a biomarker to assess the risk of sudden heart-related death. [67] One of the few PET agents that has been used to image this phenomenon in humans is 4- , which is taken up into neurons by norepinephrine transporters and sequestered into storage vesicles by vesicular monoamine transporter 2 (VMAT2). The radiosynthesis of 4-[ 18 F]F-MHPG currently used in clinical research involves nucleophilic labeling of an iodonium salt precursor, with decay-corrected RCYs of 7.0 � 3.5 % and A m of 58 � 25 GBq/ μmol. [68] In a pilot study, the SPIAd precursor for 4-[ 18 F]F-MHPG was prepared through oxidation of the iodinated starting material with dimethyldioxirane (DMDO) and coupling to the SPIAd auxiliary in 69 % yield. [69] Using the radiolabeling methods reported in our original publication on the SPIAd auxiliary, 4-[ 18 F]F-MHPG was reliably synthesized with RCYs more than three-fold higher than the iodonium salt approach (non-decay corrected RCYs 7.8 � 1.4 %) (Figure 12A).
We previously discussed another phenethylguanidine radiotracer, [ 18 F]mFBG, for neuro-oncology applications (see section 3.2). Both [ 123 I]MIBG and [ 18 F]mFBG is a NETtargeted radiotracer, which can also be used in cardiology for imaging myocardial presynaptic nerve density. [70] Our recent collaboration with the Rotstein lab described the optimized and automated radiosynthesis of [ 18 F]mFBG in RCYs of 24 � 1% with A m of 30-95 GBq/μmol. [48] Preclinical cardiac PET studies in healthy rats revealed favourable uptake (~3 SUV) and monoexponential washout kinetics. [ 18 F]mFBG was discovered to have unique NET-dependent neuronal uptake, in addition to extraneuronal uptake mechanisms. Moreover, the radiotracer appeared to experience insignificant myocardial re-uptake. The favourable imaging characteristics and unique uptake/clearance mechanisms distinguish [ 18 F]mFBG from its NET imaging counterparts, and may allow for accurate PET quantification of neuronal dysfunction.
Para-[ 18 F]fluorohippurate ([ 18 F]PFH) is one of the three tubular secretion PET agents designated as suitable for clinical renography. [71] [ 18 F]PFH is a promising candidate for clinical translation because it is rapidly and exclusively cleared by the kidneys and it provides higher quality images through dynamic PET compared to dynamic planar imaging with the clinicallyused SPECT agent [ 99m Tc]MAG3. [ 18 F]PFH was previously synthesized through a four-step radiosynthesis using [ 18 F]SFB, requiring two separate reaction vessels, lengthy synthesis times, and manual synthesis steps. These drawbacks made the synthesis inaccessible for many sites and incentivized Knepang and coworkers to devise a simplified synthesis. [72] To this end, the authors developed a synthetic strategy based on a SPI5 precursor, which was radiolabeled with [ 18 F]TBAF (Figure 12B). Using SCIDY chemistry, the group was able to synthesize [ 18 F]PFH in a one-pot, three-step reaction. The new technique provided decay-corrected RCYs of 46.5 � 2.9 % in less than half the time required for the original [ 18 F]SFB synthesis (A m was not reported). Subsequent in vivo evaluations in healthy rats confirmed that there was no observable difference between [ 18 F]PFH synthesized using the SCIDY method compared to the previous method. Due to the simplicity of this one-pot strategy, the authors state that this synthesis might be readily automated and adopted for future clinical use.

Summary and Outlook
There is a continued need for practical, reliably high-yielding, metal-free radiofluorinations to access a wider range of (hetero)aromatic rings for clinical PET imaging. Despite the ongoing success of our SCIDY methodology in the preclinical and clinical research space, limitations remain. For instance,

P e r s o n a l A c c o u n t T H E C H E M I C A L R E C O R D
SCIDY precursors can only be synthesized from iodinated intermediates that are stable under oxidative and basic conditions. Moreover, synthesis of these precursors requires multi-step syntheses, and they are not readily derived from commercially available starting materials, unlike boronate or (pseudo)halogen precursors. The most challenging aspect of SCIDY chemistry is the efficient oxidation of the aryl iodide moiety, especially those bearing multiple nitrogen atoms Further methods to oxidize aryl iodides would improve the utility of this method to encompass a wider substrate scope. As for advanatages of this methodology, SCIDY chemistry provides shelf-stable precursors that enable the direct, onestep, regioselective incorporation of 18 F-atoms into nonactivated and otherwise difficult-to-label arenes (e. g. sterically hindered). A wide scope of 18 F-PET tracers, relevant building blocks, and model compounds have been successfully radiolabeled in high yield and A m . The metal-free reactions facilitate routine production for human use by obviating the need for additional quality control of formulated radiopharmaceuticals to ensure that trace metals are not present, and offer simplified cleaning validations for automated synthesis units. Perhaps most importantly, several laboratories beyond our own have applied SCIDY chemistry for radiopharmaceutical production in academic and pharmaceutical laboratories. The continued use of SCIDY chemistry nearly a decade after its inception demonstrates the utility and versatility of this method.