UNIVERSITÉ FRANÇOIS – RABELAIS DE TOURS École Doctorale « Santé - Sciences Biologiques - Chimie du Vivant » and UNIVERSITY OF LJUBLJANA, FACULTY OF PHARMACY «Department of Pharmaceutical Chemistry»
A cotutelle thesis submitted in fulfillment of the requirements for the degree of «Docteur» at the University François Rabelais of Tours (France) and Doctor of Pharmacy at the University of Ljubljana (Slovenia) In Pharmaceutical Chemistry Publicly defended on the 1st of March 2013 by
Mitja KOVAČ in Ljubljana
FLUORATION DE DERIVES DU BENZOVESAMICOL POUR L'OBTENTION DE RADIOLIGANDS POTENTIELS DU TRANSPORTEUR VESICULAIRE DE L'ACETYLCHOLINE
Under the co-direction of: Associate Professor Sylvie Mavel (MCU, Tours) and Associate Professor Marko Anderluh (Ljubljana) ------
JURY for Oral Defense:
Ms MAVEL Sylvie – Associate Professor, University François-Rabelais, Tours, France Mr ANDERLUH Marko – Associate Professor, University of Ljubljana, Slovenia Mr DOLLÉ Frédéric – Service Hospitalier Frédéric Joliot, Institut d'Imagerie BioMédicale - CEA, Orsay, France (Reviewer) Mr EMOND Patrick – Professor, University François-Rabelais, Tours, France Ms GMEINER STOPAR Tanja – Assistant Professor, University of Ljubljana, Slovenia (Reviewer) Mr GOBEC Stanislav – Professor, University of Ljubljana, Slovenia (Chairman)
This cotutelle PhD was carried out with the collaboration between the University of Tours (Laboratoire de Biophysique Médicale et Pharmaceutique, Unité INSERM U930 - FRANCE) and the University of Ljubljana (Faculty of Pharmacy, Department of Pharmacutical Chemistry - SLOVENIA).
The work was supported by a grant from the Slovene Human Resources Development and Scholarship Fund, by a grant from the University of Ljubljana (Inovativna shema za sofinanciranje doktorskega študija za spodbujanje sodelovanja z gospodarstvom in reševanja aktualnih družbenih izzivov - generacija 2010 Univerza v Ljubljani), and by a Slovenia- French bilateral collaboration project (project n° BI-FR/12-13-PROTEUS-007).
AKNOWLEDGEMENTS
I would like to extend my most sincere gratitude to my supervisors Dr. Sylvie Mavel and Dr. Marko Anderluh for their continued direction, training, and encouragement. Their guidance has been inspirational and instructive throughout my journey along this path, and I will continue to draw on the wisdom they have imparted as I move forward.
Special thanks go to Dr. Johnny Vercouille and his team for their training and supervision in the radiopharmaceutical laboratory CERRP (Centre d'Etude et de Recherche sur les Radiopharmaceutiques).
I would like to extend my thanks to Dr. Patrick Emond and Dr. Frédérick Dollé for their valuable advices and reading the thesis.
Thanks are given to Dr. Tanja Gmeiner Stopar and Dr. Stanislav Gobec for reading the thesis.
Thanks also to Dr. Sylvie Chalon, and Dr. Mohamed Abarbri.
Finally, I would like to thank my mother and all the family who have always been loving, supportive, and a guiding light throughout my life.
RÉSUMÉ
La maladie d’Alzheimer (MA) est une maladie neurodégénérative progressive et l’une des principales démences. Les plaques amyloïdes extracellulaires, les dégénérescences neurofibrillaires intracellulaires, et la dégénérescence synaptique sont des marqueurs neurophysiopathologiques de la MA. Il a été montré que la diminution en transporteur vésiculaire de l’acétylcholine (VAChT) est un paramètre neurologique précoce, précédant les signes cliniques de la maladie, et fortement corrélée avec la démence associée à la maladie. L’utilisation de radiotraceur sélectif et spécifique pouvant être utilisé en tomographie par émission de positrons (TEP) ou par tomographie d'émission monophotonique (TEMP) offre la possibilité d’identifier de subtils changements neurologiques aux stades précoces de la maladie, et ainsi aider au diagnostic différentiel de la MA avec d’autres démences corticales ou sous-corticales.
Le (2R,3R)-5-IBVM, dérivé du benzovésamicol, est un ligand de haute affinité et sélectivité du VAChT, et est le seul radiotraceur utilisé en imagerie humaine par TEMP pour le diagnostic de la MA. Or, la TEP présente des avantages par rapport à la TEMP telle qu'une meilleure détection, meilleure résolution de l'image et possibilité de quantification. Sachant que l’analogue fluoré du 5-IBVM devrait présenter une affinité et sélectivité pour le VAChT du même ordre, nous avons donc synthétisé les énantiomères du 5-FBVM et nous avons développé des méthodes d'introduction régiosélective d'ion fluorure en position 5 du benzovésamicol, qui est une position non activée. Pour cela, nous avons choisi comme précurseur, la fonction triazène (Ar-N=N-NR2) en tant que «groupe partant » pour l'obtention d'aryles fluorés.
En partant d'études théoriques de fluoro-dediazénation, nous avons synthétisé les énantiomères du 5-FBVM, dans un rendement de 25%, en utilisant le t-butyle de nitrite comme agent de diazotation et l'éthérate de trifluorure de bore en tant qu'agent de fluoration. Des études de modélisation moléculaire (QSAR), faites sur 32 composés de type vésamicol, ont été réalisées en tenant compte de la stéréospécificité du site de fixation du VAChT. L'évaluation in vitro a montré la très bonne affinité pour le VAChT des 2 énantiomères, du même ordre que le 5-IBVM, comme le prédisaient les études QSAR. Le (2S,3S)-5-FBVM présente une meilleure sélectivité vis-à-vis des récepteurs 1 et pourrait être un traceur potentiel pour l'imagerie in vivo des neurones cholinergiques.
Actuellement un des facteurs limitants du développement de la technologie TEP est l'introduction d'un ion fluorure sur un système aromatique non activé. Nous nous sommes donc focalisés dans un deuxième temps sur la fluoration d'aryles non activés possédant un triazène. Nous avons recherché un acide n'interférant pas lors d'un potentiel radiomarquage. Nous avons étudié différentes conditions expérimentales (acide, solvant, et agent de fluoration) pour la fluoration du 3,3-diéthyl-1-naphtyltriazène choisi comme modèle d'étude. À partir des résultats obtenus en chimie froide, l'acide polyphosphorique (PPA) dans un solvant chloré est le plus prometteur et de plus innovant dans ce type de réaction. De plus, en se basant sur la chimie de coordination des triazènes avec le trifluorure de bore, nous proposons que la fluoro-detriazénation pourrait être obtenue, avec uniquement de l'éthérate de trifluorure de bore, sans addition d'acide protique, à partir d'une température suffisante. Nous avons confirmé cette hypothèse sur le naphtyltriazène ainsi que sur des phényltriazènes para- substitués en comparant chauffage traditionnel et chauffage par micro-ondes. Nous avons validé notre méthode en fluorant un système plus complexe, à savoir le 5-TBV qui a conduit au 5-FBVM dans un rendement de 72%, sous micro-onde, dans le tétrachlorure de carbone.
Des tests préliminaires de radiomarquage du 5-TBV, en utilisant le PPA dans le chloroforme sous micro-onde ont donné des résultats prometteurs.
Mots Clés: fluorination, triazène, benzovésamicol, VAChT, TEP
ABSTRACT
Alzheimer's disease (AD), a progressive neurodegenerative and terminal disorder, is the most common cause of dementia in the elderly. Extracellular amyloid plaques, intracellular neurofibrillary tangles, and degeneration of the synaptic terminals are the most characteristic neuropathophysiological hallmarks of AD. It has been shown that deficiencies in vesicular acetylcholine transporter (VAChT) are among the earliest neuronal changes, preceding clinical symptoms of the disease, and showing a strong correlation with the severity of dementia. Thus, the use of selective and specific radiotracer functional imaging modalities, such as positron emission tomography (PET) and single photon emission computed tomography (SPECT), offers non-invasive in vivo identification of subtle neurological changes in the early stages of AD and, therefore, offers value in the differential diagnosis of AD from other cortical and subcortical dementias.
Benzovesamicol-related ligand (2R,3R)-5-IBVM has a high affinity and enough selectivity for the VAChT, and is the only SPECT VAChT radiotracer used in human to obtain an early diagnosis of AD. Regarding physico-chemical properties of fluorine-19 and fluorine-18, and PET advantages over SPECT in terms of higher detection efficiency, better spatial resolution and possibility for quantification, it is expected that the fluoro analog of 5-IBVM should be of the same order of affinity and selectivity for the VAChT. Thus, we have prepared pure enantiomers of 5-FBVM and developed method for regioselective introduction of fluorine into the 5-position of non-activated benzovesamicol scaffold. We have chosen as a precursor system triazene function (Ar-N=N-NR2) as a leaving group for arylfluorination.
Firstly, we built theoretical model by studying characteristics of fluoro-de-diazoniation process. Accordingly, we have synthesized pure enantiomers of 5-FBVM from amino analog 5-ABV in around 25% yield by using t-butyl nitrite as diazonating agent and boron trifluoride etherate as fluorinating agent. Furthermore, to demonstrate the suitability of a triazene as a leaving group, the fluoro-de-triazenation of the corresponding triazene precursor (5-TBV), using triflic acid to trigger triazene moiety decomposition and boron trifluoride etherate, afforded 5-FBVM in reasonable yield (25%). QSAR studies based on 32 vesamicol derivatives taking into account the stereoselectivity of the VAChT binding site were performed. Both enantiomers exhibited high in vitro VAChT binding affinities determined by radioligand displacement studies and were in the same range as 5-IBVM as predicted by 3D
QSAR studies. (2S,3S)-5-FBVM was more selective over σ1 receptors and could be a potential PET radioligand for in vivo mapping of cholinergic terminals.
Actually, one of the main limitations in aromatic nucleophilic fluorination is that arylfluorides are only satisfactory obtained on “activated system”. As new techniques to incorporate fluoride are needed for PET technology, we focused our research in the second step on fluorination of non-activated aryl skeleton from triazene precursor. We sought for the appropriate acid to trigger triazene decomposition but with no interference in the radiofluorination step. We studied different conditions (acid, solvent, and fluorinating agent) for the fluorination of 3,3-diethyl-1-naphthyltriazene (1-NT), chosen as precursor model. According to the results in the non-radioactive attempts, polyphosphoric acid (PPA) proved to be the most suitable one in chlorinated solvents, although had never been used in this type of reaction. Furthermore, from coordination chemistry of triazene derivatives with boron trifluoride, we proposed that fluoro-de-triazenation can be successfully accomplished by the only presence of boron trifluoride without any protic acid at elevated temperature. Our hypothesis was first confirmed on 1-NT. This methodology was also extended on several para-substituted 3,3-diethyl-1-aryltriazenes by conventional and microwave heating. To prove that the method is applicable to obtain more complex arylfluorides too, 5-FBVM was accomplished in high yield (72%) with microwave heating in tetrachloromethane.
Preliminary tests were transposed to F-18 radiolabelling. Encouraging results were obtained by radiofluorination of 5-TBV using PPA in chloroform with microwave heating.
Keywords: fluorination, triazene, benzovesamicol, VAChT, PET
ABBREVIATIONS
A- conjugate base ACh acetylcholine AChE acetylcholine esterase AChEI acetylcholinesterase inhibitor AD Alzheimer's disease Aβ amyloid beta peptide APP amyloid precursor protein ATP adenosine triphosphate ChAT choline acetyltransferase ChT choline transporter CSF cerebrospinal fluid CT computerized tomography d deuteron D.c. decay corrected DIPEA diisopropylethylamine DLB dementia with Lewy body EWG electron-withdrawing group FTD frontotemporal dementia GC gas chromatography HPLC high-pressure liquid chromatography LG leaving group MCI mild cognitive impairment MRI magnetic resonance imaging n neutron N.d.c. non-decay corrected NFT neurofibrilary tangles NMR nuclear magnetic resonance p proton PA phosphoric anhydride PDD Parkinson's disease dementia
PP phosphatase PET positron emission tomography PPA polyphosphoric acid PPAR peroxisome proliferator-activated receptor PS presenilin RCY radiochemical yield SA specific activity sMRI structural magnetic resonance imaging SNAP synaptosome-associated proteins SP senile plaque SPECT single photon emission computed tomography tR retention time TEA triethylamine TLC thin-layer chromatography VAChT vesicular acetylcholine transporter VaD vascular dementia VAMP vesicle-associated membrane proteins ν neutrino
TABLE OF CHEMICAL NAMES AND STRUCTURES
Chemical name Structure
5-amino-3-(4-phenyl-piperidin-1-yl)-1,2,3,4- HO tetrahydro-naphthalen-2-ol: 5- N aminobenzovesamicol
5-ABV H2N
5-(3,3-diethyltriaz-1-enyl)-3-(4- HO phenylpiperidin-1-yl)-1,2,3,4- N tetrahydronaphthalen-2-ol: 5-ABV- diethyltriazene Et2N N N 5-TBV 5-fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4- HO tetrahydro-naphthalen-2-ol: 5- N fluorobenzovesamicol
5-FBVM F
2,2-diethyl-1-(naphthalen-5- N N NEt2 ylimino)hydrazine: 3,3-diethyl-1- naphthyltriazene
1-NT
1-fluoronaphthalene F
1-NF
N N NEt2
1-(4-tolylimino)-2,2-diethylhydrazine
CH3
F
1-fluoro-4-methylbenzene
CH3
N N NEt2
1-(4-nitrophenylimino)-2,2-diethylhydrazine
NO2
F
1-fluoro-4-nitrobenzene
NO2
N N NEt2
1-(4-butoxyphenylimino)-2,2- diethylhydrazine O
F
1-butoxy-4-fluorobenzene
O
N N NEt2
1-(4-iodophenylimino)-2,2-diethylhydrazine
I
F
1-fluoro-4-iodobenzene
I
N N NEt2
1-(4-cyanophenylimino)-2,2-diethylhydrazine
CN
F
4-fluorobenzonitrile
CN
TABLE OF CONTENTS 1. INTRODUCTION ...... 1 1.1. Positron Emision Tomography ...... 2 1.2. Strategies for 18F-labelling ...... 7 1.2.1. Direct electrophilic 18F-fluorination...... 8 1.2.2. Direct nucleophilic 18F-fluorination ...... 9 1.2.2.1. Direct aliphatic 18F-nucleophilic substitution reactions ...... 11 1.2.2.2. Direct aromatic 18F-nucleophilic substitution reactions...... 12 1.2.3. Indirect 18F-labelling reactions...... 16 1.3. 18F-labelled Aryl-Tracers through Direct Introduction of [18F]fluoride into Electron-Rich Arenes ...... 18 1.4. Alzheimer's disease ...... 57 1.4.1. Epidemiology and risk factors of Alzheimer’s disease ...... 57 1.4.2. Neurophysiology and pathology of Alzheimer’s disease ...... 58 1.4.3. Pharmaceutical management and research directions ...... 62 1.4.4. The diagnosis of Alzheimer’s disease by PET ...... 65 1.5. Vesicular acetylcholine transporter (VAChT) and the most promising PET imaging tracers ...... 68 1.6. Synthesis of 5-aminobenzovesamicol (5-ABV) and its enantiomers ...... 71
2. AIMS AND SCOPE ...... 73
3. RESULTS AND DISCUSSION ...... 76 3.1. Theoretical model for efficient one-pot fluoro-de-diazoniation ...... 77 3.2. Synthesis of 5-FBVM and its enantiomers via fluoro-de-diazoniation ...... 82 3.3. Theoretical model for efficient fluoro-de-triazenation and synthesis of 5-FBVM from the corresponding triazene precursor ...... 86 3.4. 3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential PET radioligand for the vesicular acetylcholine transporter (VAChT) ...... 88 3.5. Aromatic fluoro-de-triazenation with boron trifluoride diethyl etherate under non protic acid conditions ...... 99 3.6. Examination of reaction parameters for radio-fluoro-de-triazenation of 5-TBV ...... 106 3.7. Radiochemistry ...... 110
3.7.1. Preparation of [18F]TBAF using TRACERlab™ FX F-N synthesizer ...... 110 3.7.2. N.c.a. 18F-radiofluoro-de-triazenation of 5-TBV ...... 112
4. EXPERIMENTAL ...... 116 4.1. General information ...... 117 4.2. Chemistry ...... 117 4.2.1. One-pot fluoro-de-diazoniation of 5-aminobenzovesamicol (5-ABV) in 1,2-dichlorobenzene ...... 117 4.2.2. One-pot fluoro-de-diazoniation of 5-aminobenzovesamicol (5-ABV) in ionic liquid ...... 118 4.2.3. Fluoro-de-triazenation of 3,3-diethyl-1-naphthyltriazene (1-NT) using KF/Kryptofix® complex and triflic acid (TfOH) in dichloromethane ...... 118 4.2.4. Fluoro-de-triazenation of 3,3-diethyl-1-naphthyltriazene (1-NT) using tetra-n- butylammonium fluoride (TBAF) and triflic acid (TfOH) in chloroform ...... 119 4.2.5. Fluoro-de-triazenation of 3,3-diethyl-1-naphthyltriazene (1-NT) using tetra-n- butylammonium fluoride (TBAF) and polyphosphoric acid (PPA) in chloroform ...... 119 4.2.6. General procedure for the fluoro-de-triazenation of 3,3-diethyl-1-naphthyltriazene (1-NT) with different amounts of polyphosphoric acid (PPA) ...... 120 4.2.7. General procedure for the fluoro-de-triazenation of 3,3-diethyl-1-naphthyltriazene (1-NT) with different fluoride sources ...... 120 4.2.8. General procedure for the de-triazenation of 5-TBV with polyphosphoric acid (PPA) ...... 121 4.3. Radiochemistry...... 121 4.3.1. Preparation of [18F]TBAF using TRACERlab™ FX F-N synthesizer ...... 121 4.3.2. General procedure for the thermal n.c.a. 18F-radiofluoro-de-triazenation of 5-TBV ...... 121 4.3.3. General procedure for the microwave-assisted n.c.a. 18F-radiofluoro-de-triazenation of 5-TBV ...... 122
5. CONCLUSION ...... 123
6. REFERENCES ...... 126
1. INTRODUCTION
1
1.1. Positron Emision Tomography
Positron emission tomography (PET) is a very powerful non-invasive in vivo quantitative molecular and functional imaging technique that is used to study and visualize human and animal physiology and biochemical events by monitoring the distribution and concentration of positron-emitting radiopharmaceuticals in the body over time.1 Information about metabolism, receptor or enzyme function, and biochemical mechanisms in living tissue can be obtained directly from PET experiments in a quantitative manner. Unlike conventional anatomic imaging techniques (e.i. computerized tomography/CT), which mainly provide detailed anatomical images, PET can provide an early diagnosis (chemical changes that occur before macroscopic signs of a disease are observed), more accurate staging, monitoring response to therapy, and assessment of recurrence of disease. For all these reasons PET is increasingly used in oncology,2 cardiology,3,4 neurology,5 drug development6,7 and therapy.8 It is anticipated that PET studies will improve the selection of potential drug candidates in early stages of development, give a greater understanding of drug's mechanism of action, and aid in guiding dose selection.
One of the main challenges for radiochemists is the development of rapid synthetic methods for introducing short-lived positron-emitting isotopes into the molecule of interest. The radiolabelled probe has to be synthesized, purified, analyzed, and formulated roughly within three isotope half-lives to ensure there is enough radiolabelled material to be administered to a subject (animal or human) undergoing the PET scan. Except for fluorine-18, the extremely short-lives of the isotopes shown in Table 1 necessitate that the labelled probes be prepared in proximity to where the isotopes are produced and used almost immediately after their synthesis. A number of modern PET facilities house cyclotrons for radioisotope production, radiosynthetic laboratories, and PET scanners are under one roof to allow efficient production and transport of short-lived PET probes from the laboratory to the PET scanner.1
2
Table 1. The most important and commonly used short-lived positron-emitting radionuclides for PET imaging.
Half-life, t1/2 Nuclear Decay Radionuclide (min) reaction Target Product product
20 18 18 18 Ne(d,α) F Ne(+F2) [ F]F2 18 F 109.8 18 18 18 18 - O O(p,n) F [ O]H2O [ F]F
11 11 14 11 N2(+O2) [ C]CO2 11 C 20.4 N(p,α) C 11 B N2(+H2) [ C]CH4
13 13 16 13 H2O [ N]NOx 13 N 9.97 O(p,α) N 13 C H2O + EtOH [ N]NH3
15 15 15 15 15 O 2.07 N(d,n) O N2(+O2) [ O]O2 N
Radioisotope production begins in a cyclotron. This is a (compact) particle accelerator, which is capable of producing proton (p) or deuteron (d) beams of the required energy range to generate 18F, 11C, 13N, and 15O species.9 The beam is directed onto a target system at the exit of the cyclotron, which contains the target material suitable for the production of the required radioisotope (Table 1).
Dealing with high-energy short lived radioactive compounds safely and effectively is a priority, and traditional bench-top synthetic chemistry is clearly not an option. State-of-the-art PET radiosynthesis laboratories use “hot cells”, which are basically enclosed lead-lined versions of a fumehood with lead glass windows many inches thick, to carry out radiolabelling procedures. The radioactive isotope is transferred to the hot cell (usually under an inert gas stream) where it is converted by a series of chemical steps into the final radiolabelled product. Typically, computer-controlled robotic or automated systems are used for such labelling to restrict, as much as possible, exposure of the user to radiation.1
3
Figure 1. The process of PET radiotracer production begins at a cyclotron and end at the PET scanner. The whole process typically takes few hours. From left to right: (A.) commercially available biomedical cyclotron; (B.) “Hot-Cell” - automated radiolabelling system controlled from outside the hot cell; (C.) combined PET/CT scanner and processed PET image (bottom right).
An important advance in scanner technology and evolution in imaging technology has been the physical integration of PET and CT within the same device (hardware fusion approach). The combined PET/CT scanners allow to acquire co-registered anatomic and functional images in a single scan session and therefore allows accurate localization of functional abnormalities.10
Since PET labelling reactions are performed with nanomolar amounts of radioisotopes, there is normally a vast stoichiometric excess of “cold” reagents which results in pseudo-first- order reaction kinetics with respect to the radioisotope concentration. Advantageously, reactions, which may normally need hours or days to reach completion on a macroscopic scale, can often be performed in minutes using PET radioisotopes. It is also worth noting that even minor impurities found in precursors, solvents and reagents can become significant when performing such small-scale reactions.
Before a PET radiotracer can be administered to a patient, usually by intravenous injection in the form of saline solution, it must be suitable and rapidly characterized, and sterilized. PET radiotracers are often characterized by using a combination of high-pressure liquid chromatography (HPLC), thin-layer chromatography (TLC), and gas chromatography (GC) in conjuction with suitable radioactivity and mass detectors. Quality control procedures for radiotracers and radiopharmaceuticals are similar to those applied to non-radioactive pharmaceuticals. There are two categories for quality control tests: physiochemical test and biological test. Physiochemical tests give the level of radioisotope and radiochemical 4
impurities, chemical impurities, pH value, ionic strength, osmolality, and physical state of the sample, while the biological tests determine the sterility, apyrogenicity, and toxicity of the sample.1
The radiochemist has to consider the radiochemical yield (RCY) of the radiosynthesis and the specific activity (SA) of the final radiolabelled compound. The RCY is a function of both the chemical yield and half-life of the radioisotope, and is expressed as a fraction of the radioactivity originally present in the sample following a radiochemical separation. The value is often quoted as being either non-decay corrected (n.d.c.) or decay corrected (d.c.). Decay corrected figures are mathematically adjusted measurements that take into account radioactive decay that has occurred between two different times to give a single value (decay equation:
-0.693Δt/t1/2 A(t2) = A(t1)e ). It is desirable, but not always essential, to have a high RCY as it is a useful gauge to measure the efficiency of the radiolabelling procedure. The specific activity is a measure of the radioactivity per unit of mass of the labelled compound, and is commonly expressed as giga-Becquerel (SI unit) per micromol (GBq/μmol) or Curies per micromol (Ci/μmol). The theoretical maximum values of the specific activity (e.g. for fluorine-18 is 63344 GBq/μmol or 1712 Ci/μmol) are never reached for radiolabelled compounds, because of unavoidable isotopic dilution by the naturally occurring stable isotope. This effect is particularly apparent when a “cold” fluorine-19 gas is added to the target to allow recovery of 18 molecular F[F2] for electrophilic fluorinations. Typical specific activities of PET-labelled products are in the order of 50-500 GBq/μmol (ca. 1-15 Ci/μmol). Since a small amount of radioactivity can lead to a good quality PET image, only very low amounts (tracer dose) of compound need to be administered - typically in the sub-micro-gram level. This implies that the fate of labelled molecules can be studied in vivo without perturbing the biological system being measured and that very potent or toxic compounds can be studied in human at subpharmacological and subtoxicological doses.1,11
Due to the characteristic physico-chemical properties, fluorine-18 as artificial radionuclide is among the available PET radionuclides the most favoured and widely used radiolabel for in vivo imaging. Fluorine-18 emits quite low energy positron (β+ particle; max. 0.635 MeV; 97% positron emission and 3% electron capture), which has short path in vivo (~0.5 mm in water, max. 2.3 mm in tissue) before it annihilates with an electron (e-) giving rise to two opposed and coincidentally detected by PET camera 511 keV γ-rays. This is the physical basis to reconstruct the highest resoluted PET images of all the available positron emitters (Figure 5
2).1,11 The decay of fluorine-18 gives innocuous 18O as the product atom (Table 1, Figure 2). It can be produced as the single-atom species [18F]fluoride ion (no-carrier-added) in very high specific radioactivities (~37-370 GBq/μmol or ~1-10 Ci/μmol ) by irradiation of [18O]- enriched water (available from commercial vendors) with 11-19 MeV energy proton (p) beam from small compact cyclotrons according to the 18O(p,n)18F reaction.1,11 The ease of production of high amounts of fluorine-18 coupled with its almost 2 hour half-life (t1/2 = 109.8 min) allows either more time consuming multistep labelling reactions or longer lasting in vivo investigation or commercial distribution of a tracer over reasonable distances from remote cyclotron to the clinical PET centers that lack radiochemistry facilities. Furthermore, fluorine may to some extent mimic a hydrogen atom or hydroxy group in an organic molecule and in parallel can lead to favourable conformational and physico-chemical changes (such as pKa, logD) with improved pharmacokinetic (e.g. bioavailability), pharmacodynamic (e.g. enhanced target binding affinity and selectivity) and toxicologic profiles of organofluorine compounds.12–15
Figure 2. The physical principles of PET imaging shown schematically. After intravenous administration to the patient or animal, 18F-labelled tracer distributes to target tissue(s) and selectively binds to particular biochemical target (e.g. neuroreceptor). Fluorine-18 is neutron-
6
deficient isotope that achieves stability through the nuclear transmutation of a proton (p) into a neutron (n). This process involves emission of positron (β+) and neutrino (ν). Positron quickly collides with a free or loosely bound electron (e-) in surrounding tissue, and both are annihilated to form two 180 degrees separated 511 keV gamma-ray photons (γ-rays). PET camera/scanner (circular configuration of scintillation detectors around the subject) detects a large number of pairs of annihilation photons in coincidence (“annihilation coincidence detection”) in all the lines by opposing detectors as the basis for the approximate determination of location of the fluorine-18 nuclei/PET probe in the patient. Millions of individual annihilation events are required to give enough data to reconstruct a high- resolution PET image.
In comparison with SPECT that utilizes the single photons emitted by gamma-emitting radionuclides, PET has some essential advantages in comparison to SPECT: higher detection efficiency, better three-dimensional resolution of a studied image, possibility to quantify the studied target, and after all, minimization of radiation dose to the patient due to low positron energy of fluorine-18.
1.2. Strategies for 18F-labelling
The main synthetic strategies behind 18F-labelling can be generally divided into two main distinct areas: (1) direct fluorination and (2) indirect fluorination. In direct fluorination fluorine-18 is introduced directly into the target molecule of interest in one step, and in indirect fluorination fluorine-18 is introduced via so-called prosthetic groups and requires a multistep synthetic approach. These prosthetic groups are typically small 18F-labelled alkyl or aryl compounds that have reactive functional groups. They are used to react with more complex biological molecules which may not be suitable or stable enough to tolerate direct fluorination methods. The direct 18F-labelling strategies can be subdivided into two categories: electrophilic and nucleophilic. Of these two methods, nucleophilic 18F- fluorination has dominated in importance because of its greater selectivity and capability to give highly specific radioactive compounds suitable for PET imaging.
7
Synthetic methods for the introduction of fluorine-18 into organic molecules need to be convenient, rapid and of reasonable high RCY and SA. High SA also enables radiotracers to be administered to subjects in low mass doses (1-10 nmol or sub-microgram level) to avoid any toxic or pharmacological effects and perturbation of the biological target or process.
1.2.1. Direct electrophilic 18F-fluorination
Electrophilic 18F-fluorinations are less favoured nowadays for two reasons: (A.) they generally give labelled products with low specific activity, because of the carrier-added 18 method of [ F]F2 production, and (B.) labelling with electrophilic reagents such as very 18 18 reactive [ F]F2 is generally non-regioselective and can result in mixtures of F-labelled products. However, electrophilic fluorination has played an important and historical role in the development of 18F-labelled molecules for PET imaging. For example, the first synthesis of the hugely important PET tracer 2-deoxy-2-[18F]fluoro-D-glucose ([18F]FDG) was carried out by electrophilic fluorination.16 Other more recent examples of important PET tracers, which still rely on electrophilic 18F-fluorination methods of synthesis, include [18F]fluoro-L- DOPA and 2-L-[18F]fluorotyrosine.
18 The most common reagent for electrophilic fluorination is [ F]F2, which is obtained from the nuclear reactions of 20Ne(d,α)18F or 18O(p,n)18F. It can be used as it is or converted into 18 the less reactive but more selective derivative, such as acetyl hypofluorite ([ F]CH3COOF) (Scheme 1, Example A.).1,11 Electrophilic 18F-reagents can be used to fluorinate electron-rich substrates by either direct electrophilic substitution or by more regioselective direct demetallation reactions using organometallic intermediates precursors, such as organomercury and organotin preursors (Scheme 1, Example B.).
A. OAc OH 1. [18F]CH COOF AcO O 3 HO O AcO HO 2. HCl OH 18F 18 [ F]FDG
8
B. OH OH 1. CCl3F/AcOH, 18 BocO NHBoc [ F]F2 HO NH2 2. HBr 18 Me3Sn F 18 4-[ F]FMR
18 18 Scheme 1. (A.) Electrophilic F-fluorination using [ F]CH3COOF was the early synthetic method for the production of [18F]FDG, a powerful imaging agent for determining cerebral and myocardial glucose metabolism, tumour localization and determining its respons to therapy.1,16 (B.) Synthesis of (1R,2S)-4-[18F]fluorometaraminol (4-[18F]FMR), a radiotracer for in vivo mapping of adrenergic nerves in the heart, from Boc-protected organotin precursor 18 17 by electrophilic aromatic substitution with [ F]F2.
1.2.2. Direct nucleophilic 18F-fluorination
Nucleophilic 18F-fluorination reactions are routinely used to efficiently produce some of the most important 18F PET radiotracers: [18F]FDG widely used in oncology investigations, [18F]fallypride,18 [18F]haloperidol,19 and [18F]N-methylspiperone20 used in dopamine receptor studies, and [18F]fluoroazomycin arabinoside ([18F]FAZA)21 and [18F]fluoromisonidazole ([18F]FMISO)22 for non-invasive tumour hypoxia imaging in vivo with PET.
As [18F]fluoride ion ([18F]F-) is in most cases produced by the proton irradiation of highly enriched [18O]water target (> 95%) which renders the anion poorly nucleophilic due to its 23 high degree and strength of hydration in aqueous solution (ΔHhydr = 506 kJ/mol), the bulk [18O]water is removed and can be recovered for reuse by adsorption of [18F]fluoride onto an anion exchange resin.24 [18F]Fluoride is eluted from the anion-exchange resin using either an alkali carbonate, hydrogencarbonate and phase transfer catalyst in a water/acetonitrile solution or using dilute aqueous solution of tetra-n-butylammonium bicarbonate (n-TBAHCO3). The most commonly used cryptand in combination with potasium carbonate (K2CO3) in order to 18 activate [ F]fluoride ion is Kryptofix 2.2.2. (K222). Water is then removed via repeated cycles of azeotropic evaporation with anhydrous acetonitrile (CH3CN) to form the reactive (“naked”) fluoride.1,25 Finally, a solution of precursor mostly dissolved in polar aprotic anhydrous organic solvent, such as acetonitrile (CH3CN), dimethyl sulfoxide (DMSO) and
9
dimethylformamide (DMF), is added for the radiolabelling reaction (Scheme 2). Completely anhydrous or “naked” fluoride is never obtained by this procedure and its alternative variants.25,26 This can be troublesome for some reactions such as aromatic nucleophilic substitution, where highly anhydrous conditions are required for reactions to proceed successfully.
18 18 18 O(p,n) F on H2 O
(A.) Cyclotron (11-19 MeV protons)
18 - 18 18 [ F]F x(H2 O)n in H2 O
(B.) Adsorption on O O anion exchange resin N O O N + aq. K2CO3 O O
K222 in CH3CN/H2O [18F]F- 18 - N (C.) Elution of [ F]F with aq. K2CO3/K222
O K+ O 18 - N N [ F]F x(H2O)n in CH3CN/H2O O O O O
(D.) Azeotropic evaporation with anhydrous CH3CN (2-3x) to form ''naked'' [18F]F-
O K+ O N [18F]F-x(H O) N O O 2 m O O
Scheme 2. General radiosynthetic scheme to obtain no-carrier-added [18F]fluoride ion ready for radiolabelling reaction. The azacryptand K222 forms a strong complex with potassium cation and leaves [18F]fluoride exposed and highly nucleophilic when dissolved in a polar aprotic anhydrous solvent (m is expected to be less than n).
10
Direct 18F nucleophilic labelling can be subdivided into aliphatic and aromatic 18F-labelling strategies.
1.2.2.1. Direct aliphatic 18F-nucleophilic substitution reactions
Direct aliphatic nucleophilic 18F reactions are generally straightforward. Unlike aromatic substitution reactions, activating groups are not required. The only requirement for aliphatic nucleophilic 18F reactions is a good leaving group, such as triflate (TfO) (Scheme 3), tosylate (TsO), mesylate, iodo, or bromo group. Labelling at aliphatic carbon atoms using sulfonate ester (mesylate, tosylate) or halide leaving groups can be accomplished very efficiently even in the presence of trace water27 or in sterically hindered alcohols as a protic reaction medium (e.g. tert-butyl alcohol, t-BuOH).28,29 The main drawback of aliphatic method is the need to protect any potentially competing sites of nucleophilic attack in the molecule (principally acid, alcohol, or amine groups). Additionally, fluorine-18 bound to an aliphatic carbon atom can be prone to de-fluorination in vivo, giving rise to [18F]F-. The latter binds avidly to bone, including skull, and compromise PET measurements with failure to image specific target in vivo.30–32
OAc SO2CF3 OH O 1. [18F]KF/K O 222 HO O AcO OAc AcO 2. HCl or NaOH HO OH 18F 18 [ F]FDG
Scheme 3. Direct nucleophilic n.c.a. 18F-fluorination and subsequent deprotection (acidic or basic hydrolysis) of acetyl-protected mannose triflate for the preparation of [18F]FDG.33
Direct aliphatic nucleophilic 18F reactions have been used to effectively label a number of complex organic molecules in either one step where no protecting groups are necessary (Scheme 4, Example A.) or two synthetic steps which involves a deprotection (Scheme 4, Example B.).
11
A. Cl 18F N N COOCH3 COOCH3 18 [ F]KF/K222 , DMSO
165 oC, 10 min Chlorinated precursor [18F]LBT-999 CH3 CH3
B. O O S O H3CO O SO -K+ O 3 18 [ F]KF/K222 , CH3CN 18F 110-115 oC, 15 min O O
F OH
Protected cyclic sulfone precursor H3CO 18F acidic ethanol
110-115 oC, 15 min HO
F 18 [ F]4F-MFES
Scheme 4. (A.) One-step radiosynthesis of [18F]LBT, 18F-labelled cocaine derivative as a highly selective dopamine transporter radioligand.34 (B.) Two-step radiosynthesis of [18F]4F- MFES, PET tracer for imaging estrogen receptor densities in breast cancer patients.35
1.2.2.2. Direct aromatic 18F-nucleophilic substitution reactions
Direct nucleophilic aromatic 18F-fluorination is in general restricted to electron-deficient arenes. Thus, an electron-withdrawing group (EWG: NO2, CN, CHO, COR, COOR, CF3) in + - - - para and/or ortho position to the good leaving group (LG: NO2; Halides; N Me3 X , X = TfO - - - , TsO , ClO4 , I ) is indispensably required on the ring to be fluorinated successfully, reproducibly and in acceptable radiochemical yields (Scheme 5).
12
18 NO2 or -Cl F
N N
HO O 18 HO O [ F]KF/K222
[18F]haloperidol Nitro- or chloro-precursor Cl Cl
Scheme 5. Direct synthesis of [18F]haloperidol (butyrophenone neuroleptic) from the corresponding nitro-36 or chloro-precursor.19
In spite of the presence of EWG quite high reaction temperatures are still required for n.c.a. 18F-labelling.
Synthesis of [18F]fluoropyridine derivatives proceeds in a similar fashion since these compounds are reactive toward nucleophilic substitution at the C(2) and C(4) positions to the 37 + ring nitrogen (Scheme 6). Thus, only a good leaving group (a halogen, NO2 or N Me3) is required, except if fluorination on C(3) to the ring nitrogen is desired. As direct nucleophilic aromatic 18F-fluorination is restricted to electron-deficient arenes, extra steps after labelling have to be performed occasionally to modify or eliminate the activating group completely on account of a certain loss of overall RCY and specific activity (SA).
X 18F Boc 18 Boc N [ F]KF/K222, DMSO N N N 150-180 oC, 10 min or MW (100 W), 1-2.5 min
+ - X = Br, NO2, N Me3 I
18F H N CF COOH 3 N
CH2Cl2, rt, 2-5 min
18 [ F]norchlorofluoroepibatidine
Scheme 6. Synthesis of [18F]norchlorofluoroepibatidine as a selective central nicotinic cholinergic α4β2 PET radioligand. The high radiochemical yield for the first step (up to 70%) and the quantitative conversion in the deprotection with trifluoroacetic acid afforded overall d.c. radiochemical yields of up to 65%.38,39 13
There are only a few available methods, namely, Balz-Schiemann reaction,40 Wallach reaction,41 and more recently diaryliodonium salt method,42 to allow direct introduction of n.c.a. [18F]fluoride into arenes without the need for further activating groups (Scheme 7).
A I
R R'
[18F]F- Diaryliodonium salt method
Balz-Schiemann Alkyl reaction 18 Wallach reaction N2 BF4 F N N N 18 18 [ F]F [ F]F Alkyl
R R R
Classical aromatic nucleophilic [18F]F substitution (SNAr)
LG
EWG
+ LG = N Me3, NO2, halogen EWG = o- or p-NO2, -CN, -CF3, CHO, RCO, COOR
Scheme 7. Direct regioselective one-step n.c.a. methods to the radiosyntheses of [18F]fluoroarenes using [18F]fluoride anion.
Since only one of the four fluorine atoms is retained in the final product after thermal + - decomposition of diazonium tetrafluoroborate salt (ArN2 BF4 ), Balz-Schiemann reaction is very inefficient from a radiochemical point. Therefore, the yield of [18F]fluoride labelling is theoretically limited to only 25%, and more importantly, the dilution of SA by the non- 18 - 18 - 18 radioactive fluorine in the counterion [ F]BF4 (carrier added [ F]F in the form of F- labelled tetrafluoroborate anion) is high with the consequential great possibility of failure to localize the biochemical target in the desired tissue(s). More precisely, low specific radioactivity would saturate targets with the co-administered non-radioactive tracer, and so 14
annul any signal from radiotracer binding. In spite of described limitations, Balz-Schiemann reaction was the first method used in nucleophilic 18F-labelling of arenes.43
Aryl-3,3-dialkyltriazenes (Ar-N=N-NR'R'') are regarded as protected form of anilines and stable surrogates of aryldiazonium ions.44 Consequently, the decomposition of aryltriazenes proceeds via a diazonium ion or corresponding radical, and cause the same yield and reproducibility problems as Balz-Schiemann method or its modifications. For this reason there are very few examples of the successful application of the Balz-Schiemann and fluoro- de-triazenation reaction (Wallach reaction) to the preparation of complex [18F]fluoroarenes usually with low RCYs.45–47 Because of limited RCYs, [18F]fluoro-de-triazenation is rarely applied for the production of 18F-labelled tracers nowadays.
On the other hand, diaryliodonium salts have been proven to be useful precursors for the introduction of n.c.a. [18F]fluoride into simple as well as more complex arenes (Scheme 8) and heterocycles. Thus, they gain more and more interest for direct radiofluorination of otherwise unfavourable electron-rich arenes.25,48 The prominent effect of diaryliodonium precursors in radiofluorination is the so-called ortho-effect.49–52 Generally higher Ar18F yields compared to Balz-Schiemann and Wallach reactions probably arise from the direct vicinity of [18F]fluoride to the proximal equatorial aryl ring due to [18F]fluoride “fixationˮ to the hypervalent iodine, and because of short-lived nature of the subsequently formed transition state at elevated temperature all together limit the formation of reactive intermediates (e.g. aryl cations, aryl radicals).51,53 The principal limitation of the method is the preparation of highly pure complex precursors suitable for radiolabelling.
N O N O
N N H3C O 18 O [ F]KF/K222 CH3 CH3 DMF, 150 oC, 5 min I N 18F N CH CH TsO O 3 O 3 [18F]flumazenil 67% RCY
Scheme 8. Preparation of [18F]flumazenil, chemically indistinguishable from its non- radioactive counterpart, using 4-methylphenyl-mazenil iodonium tosylate precursor.54
15
For more information of n.c.a. nucleophilic 18F-fluorination of electron-rich arenes see Chapter 1.3 (pages 18-56): 18F-labelled Aryl-Tracers through Direct Introduction of [18F]fluoride into Electron-Rich Arenes.
1.2.3. Indirect 18F-labelling reactions
Direct 18F-fluorination methods are not always suitable or possible for the synthesis of 18F- target compound, because either the compound is not sufficiently activated or it can't tolerate harsh reaction conditions (high reaction temperatures, basic conditions, polar organic solvents). In these cases indirect introduction of the 18F radioisotope by reaction with small 18F-labelled reactive precursors or suitable 18F-labelled prosthetic groups under milder reaction conditions (e.g. room temperature) is used.
Simple mono[18F]fluoroalkyl halide or sulfonate derivatives with methyl, ethyl, propyl, and butyl carbon backbones are important synthetic precursors for the introduction of [18F]fluoroalkyl groups into complex target molecules, and thus provide alternative synthetic path to label target biological compounds. These derivatives are prepared by the reaction of [18F]fluoride with dihalo or disulfonate alkyl starting materials where the vast excess of the alkyl starting material compared to [18F]fluoride allows exclusive formation of the mono[18F]fluoroalkyl halide or sulfonate reagent (Scheme 9).
A. 18 - X [ F]KF/K222 X Nu Nu X 18F 18F n = 0-3 n = 0-3 n = 0-3
X = tosylate, Br, I
B. H 18F N N 18 COOCH3 F COOCH OTs 3 KI, DMF, 90 oC, 20 min
CH3 CH 18 3 [ F]LBT-999 Scheme 9. (A.) Synthesis and general reaction of simple [18F]fluoroalkyl halide/sulfonate derivatives. (B.) Alternative two-step synthesis of [18F]LBT-999 using the reactive (E)- [18F]fluoro-4-tosyloxybut-2-ene group and the corresponding secondary amine precursor.55
16
[18F]-1-Bromo-4-fluorobenzene and 1-[18F]fluoro-4-iodobenzene have been used as 18F- fluorinated synthons in palladium-mediated cross-coupling (Suzuki reaction,56 Stille reaction,57 Sonogashira reaction58) for the synthesis of complex 18F-labelled target molecules. Both synthons can be efficiently prepared by the thermal decomposition of the corresponding diphenyliodonium salts in the presence of [18F]fluoride.53,59
18F 18F OH OH
I 18F-fluorinated synthon
Pd[PPh3]4 , CuI, RO THF, TEA RO
R = H, CH3
Scheme 10. Sonogashira cross-coupling with terminal alkyne (17α-ethynyl-3,17β-estradiol) in the presence of CuI and Pd[PPh3]4 as catalysts and triethylamine (TEA) as the base to give corresponding cross-coupled compound: [18F]-17α-(fluorophenylethynyl)-3,17β-estradiol and [18F]-17α-(fluorophenylethynyl)-3,17β-estradiol-3-methylether.58
The labelling of proteins and peptides for PET relies on the indirect methods using 18F- prosthetic reagents that target amino, carboxylic, or thiol functional groups within the peptide.1 The most commonly used 18F-labelling reagent to label peptides via primary amino groups at the N-terminus or lysine residues is the active ester [18F]-N-succinimidyl-4- fluorobenzoate ([18F]SFB) (Scheme 11).60
O O O H2N peptide 18F 18F O N DIPEA HN peptide
[18F]SFB [18F]SFB-labeled peptide O
Scheme 11. Synthesis of 18F-labelled peptide by reaction of [18F]SFB and free peptide amino group in the presence of diisopropylethylamine (DIPEA).
17
1.3. 18F-labeled Aryl-Tracers through Direct Introduction of [18F]fluoride into Electron-Rich Arenes
Mitja Kovača,b, Sylvie Mavela, Marko Anderluhb,* a Université François-Rabelais de Tours, INSERM U930, 37000 Tours, France b University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Aškerčeva 7, 1000 Ljubljana, Slovenia
Accepted in: Current Organic Chemistry
18
SUMMARY:
Radiolabelling of electron-rich arenes directly with no-carried-added [18F]fluoride is a challenging problem for radiochemists specialized in labelling of organic molecules with fluorine-18. In this review we described critically chemical methods, namely Balz-Schiemann reaction with its modifications, Wallach reaction, and more recently developed diaryliodonium salt methodology that are used to label non-activated arenes with no-carried- added [18F]fluoride anion regioselectively. The review is focused on diaryliodonium salts, because they have been shown to be the most suitable precursors for direct single-step nucleophilic [18F]fluorination of simple as well as more complex electron-rich arenes recently.
STATEMENT: I declare, that nobody of co-authors has used the article 18F-labelled Aryl- Tracers through Direct Introduction of [18F]fluoride into Electron-Rich Arenes, which was submitted in the Current Organic Chemistry, for his/her own thesis.
19
ABSTRACT
Rapid and efficient methods using no-carried-added [18F]fluoride as the source of fluorine-18 for nucleophilic aromatic fluorination play an important role in the development of new radiopharmaceuticals for positron emission tomography (PET). Molecules that bear electron-rich aromatic moieties are especially difficult to label by direct single-step nucleophilic no-carrier-added radiofluorination. Classical Balz-Schiemann reaction with its modifications, Wallach reaction and diaryliodonium salts methodology are a few methods to enable this. The present review provides a critical overview of these chemical methods with the emphasis on diaryliodonium salt as precursors for the direct introduction of [18F]fluoride into electron-rich arenes in synthesis of 18F-labeled molecules for PET scanning.
GRAPHICAL ABSTRACT
I
A
R R' Diaryliodonium salt method [18F]F-
Balz-Schiemann reaction Wallach reaction Alkyl 18 N2 BF4 [18F]F- F [18F]F- N N N Alkyl
R R R
Keywords: aromatic fluorination, arylfluoride, Balz-Schiemann reaction, diaryliodonium salts, 18F-labeled molecules, PET, triazene.
20
1. INTRODUCTION
Positron emission tomography (PET) is a very powerful non-invasive in vivo molecular imaging technique that allows visualization, characterization, and quantification of biochemical target function and physiopathological processes at the cellular or molecular levels even before macroscopic anatomical and clinical signs of a disease are observed in animal and human subjects [1, 2, 3]. Thus, it is increasingly applied in clinical research and diagnosis, as well as in drug discovery, development, and therapy [1, 4–6]. Expansion of PET utility depends on the development and availability of selective and specific positron-emitting radionuclide- labeled molecular probes for particular biochemical targets or pathways that enable their non-invasive imaging and quantification in vivo [4, 7, 8]. The development of receptor-specific probes is far from trivial and represents an important challenge for synthetic and medicinal chemists. Fluorine-18 is the most attractive and favored radiolabel for in vivo imaging among the available PET radionuclides due to its characteristic physical and chemical properties [1-3, 9–15]. No-carrier-added (n.c.a.) [18F]fluoride ([18F]F-) is nowadays mostly produced via the proton irradiation of an [18O]-enriched cyclotron water target, and according to the 18O(p,n)18F reaction
[9, 11, 12] which renders the anion, due to its high degree and strength of hydration in aqueous solution, poorly nucleophilic [16]. Still, a variety of rapid phase-transfer-type protocols have been developed based on trapping and subsequent elution of [18F]F- from the anion-exchange resin [17], with the addition of either bulky counter-
+ - cations (e.g. of Bu4N HCO3 ) or cryptands (e.g. diazacryptand Kryptofix 2.2.2., K222) with alkali metal salts
18 - (e.g. K2CO3) in order to obtain (after azeotropic drying step(s)) a highly nucleophilic [ F]F system, such as
+ 18 - 18 18 18 - Bu4N [ F]F ([ F]TBAF) and [ F]KF/K222 complex [9, 18]. Although completely anhydrous or “naked” [ F]F reagents are never obtained by these procedures, their degree of dryness are high enough to perform difficult reactions, such as aromatic nucleophilic substitution in polar aprotic anhydrous organic solvents (e.g. DMSO,
CH3CN, DMF).
A major challenge in PET radiotracer development is to find an efficient and rapid method for n.c.a. incorporation of cyclotron-produced [18F]F- into an organic molecule. This may be achieved at aliphatic and aromatic sites by substitution reactions [9-12]. Labeling at aliphatic carbon atoms using sulfonate ester
(mesylate, tosylate) or halide leaving groups can be accomplished very efficiently, even in the presence of trace water [19] or in sterically hindered alcohols such as tert-butyl alcohol (t-BuOH) as a protic reaction medium [9,
18, 20]. However, fluorine-18 bound to an aliphatic carbon atom is often prone to de-fluorination in vivo, giving rise to [18F]F-, which binds avidly to bone, including the skull, and compromises PET measurements with the
21
failure to image specific target in vivo [21–23]. However, attachment of fluorine-18 to an aromatic carbon atom through a stronger C-F bond compared to fluoroalkyl bond greatly reduces the tendency for radio-de- fluorination. Consequently, methods for the introduction of fluorine-18 into aromatic ring systems play an important role in the development of new radiopharmaceutical for PET.
There are principally two common strategies for the direct 18F-labeling of the arenes: (1) electrophilic; and (2) nucleophilic 18F-substitution, among which the latter dominates in importance of researches as will be discussed in the following sections. Several nucleophilic 18F-substitution methods to obtain 18F-labeled aryl fluorides have been established, evaluated, and applied [2, 3, 9, 11, 12]. As nucleophilic aromatic substitution is an energetically demanding reaction and not all biological molecules or drug candidates contain a suitably activated aryl ring for fluorination by the addition-elimination mechanism, the direct incorporation of [18F]F- into electron-rich arenes represents a significant challenge in the synthesis of PET tracers. In order to perform radiofluorination successfully, reproducibly, and in acceptable to high radiochemical yields (RCYs), radiotracers are usually designed so that the electron-withdrawing group (EWG), such as NO2, CN, CHO, COR, COOR, and
CF3, is easily incorporated in the para and/or ortho position to the good leaving group (LG) (NO2, Halides,
+ ------Me3N X ; X = TfO , TsO , ClO4 , I ) [9, 24–29]. In spite of the presence of EWG, quite high reaction temperatures are still required for n.c.a. [18F]-labeling. Synthesis of [18F]fluoropyridine derivatives proceeds in a similar fashion since these compounds are reactive toward nucleophilic substitution at the C(2) and C(4) positions [30]. In some cases, extra steps after labeling have to be performed occasionally to modify or completely eliminate the activating group on account of a certain loss of overall RCY and specific activity (SA)
[31, 32]. Electron-rich aromatic rings can in principle be more conveniently directly radiolabeled by electrophilic
18 18 18 18 19 18 F-substitution using [ F]fluorine gas ([ F]F2 = F- F) or less reactive but more selective electrophilic F-
18 18 fluorination reagents derived from it, such as acetyl [ F]hypofluorite (CH3COO[ F]F) [1, 2, 33, 34]. Fluoro- demetalation reactions using organomercuric or preferably less toxic organostannane precursors afford more
18 18 18 regioselective aromatic F-fluorination with [ F]F2 and [ F]CH3COOF as electrophilic radiofluorinating agents. In this manner some important radiopharmaceuticals such as 6-[18F]fluoro-L-3,4-dihydroxyphenylalanine
(6-[18F]fluoro-L-DOPA) [35–37], 2-[18F]fluoro-L-tyrosine [38] (1R,2S)-4-[18F]fluorometaraminol [39] have been prepared. Important consideration of using an organometallic approach is to ensure that there are no residual amounts of the metals in the final product which would complicate the quality control analysis. However, electrophilic radiofluorination of organic compounds has several significant shortcomings [1, 2, 11, 12]. Firstly,
22
18 the theoretical maximum achievable RCY can be only 50% because only half the radioactivity of [ F]F2 can be utilized for mono-radiofluorination of an organic compound (only one of the fluorine atoms in molecular
[18F]fluorine carries the 18F label, the other 50% of the input activity is lost in the form of fluoride) which has not been realized in practice. Secondly, electrophilic radiofluorination is not applicable to n.c.a. labeling, because
18 19 [ F]F2 is produced along with non-radioactive fluorine gas ( F2) as a carrier in order to increase the recovering
18 19 efficiency of [ F]F2 from the cyclotron target after its production. Thus, the addition of F2 significantly lowers
18 18 - 18 (100-1000x) the specific radioactivity (SA) of [ F]F2 compared to the SA of [ F]F even when [ F]F2 is generated via 18O(p,n)18F reaction [40]. Consequently, the SA of radiotracers prepared by the electrophilic approach are typically less than 0.4 GBq/μmol (~ 0.011 Ci/μmol) and usually too low for PET investigations of low density in vivo imaging. High SA also enables radiotracers to be administered to subjects in low mass doses
(1-10 nmol or sub-microgram level) to avoid any toxic or pharmacological effects and perturbation of the biological target or process [9].
Scheme 1. Direct regioselective n.c.a. methods for the radiosyntheses of [18F]fluoroarenes using [18F]fluoride anion.
Given the above reasons, the ultimate goal is to perform direct regioselective n.c.a. [18F]F- incorporation into complex electron-rich arenes as late as possible in the synthetic sequence to obtain a radiotracer of high SA.
23
This is a particular challenge in arenes of high electron density, where an electrophilic aromatic carbon or intermediate should be generated. Only a limited number of available methods proceed via generation of mentioned electrophilic species, namely, Balz-Schiemann reaction [41], Wallach reaction [42], and more recently with the use of diaryliodonium salt precursors (Scheme 1) [43, 44]. This review focuses on radiolabeling strategies applied in the synthesis of 18F-labeled aryl-tracers from electron rich aryl precursors.
Furthermore, pros and cons of each method are highlighted, and an overview of the successful and most recent examples with an emphasis on diaryliodonium salt precursors is provided.
2. RADIOFLUORINATION OF ELECTRON-RICH ARNES VIA A BALZ-SCHIEMANN REACTION
Although known for almost a century, the Balz-Schiemann reaction [41] is still the broadest substrate scope method for the regioselective nucleophilic introduction of fluorine into aromatic ring. This is a deaminative fluorination type of reaction composed of three sequential steps: (1) diazonitation of primary
o aromatic amine in aqueous medium with sodium nitrite (NaNO2) and fluoroboric acid (HBF4) at 0-5 C to
+ - + - produce arenediazonium tetrafluoroborate (ArN2 BF4 ); (2) isolation and drying of ArN2 BF4 to avoid side
+ - formations of phenols and biaryl ethers [45]; and (3) thermal fluorinated decomposition of ArN2 BF4 (fluoro-de- diazoniation) [46, 47]. However, this method suffers from yield reproducibility problems because isolation and
+ - complete drying can be tedious and unsafe, and controlled thermal decomposition of ArN2 BF4 is problematic
[45, 48]. To overcome reproducibility problems, simplify the procedure, broaden substrate tolerance, improve safety, and to increase the yields, alternative approaches based mostly on one-pot methodology (in situ fluoro- de-diazoniation) in non-aqueous solvents have been developed during the last few decades [48–56].
Decomposition of aryldiazonium cations can occur by an ionic (heterolytic) pathway via aryl cation intermediates or by a homolytic pathway that generates aryl radical intermediates which quickly react with fluoride or any other nucleophile due to their high reactivity and non-selectivity via a SNAr1 type of reaction mechanism [57–59]. The delicate decomposition pathway balance is crucially dependent on the substituents in the aromatic ring and reaction conditions. More precisely, substituents and their substitution pattern affect stability of the aryldiazonium ion, its redox potential, and consequently its decomposition temperature and pathway [47, 59, 60]. For successful fluoro-de-diazonitation, conditions should be carefully chosen to promote aryl cation formation. The solvent, pH of the medium, the nature of the counterion, and the presence of reducing
24
agents and/or radical sources decisively influence arylfluoride yields. The choice of the solvent is one of the most important parameters [59, 60] and so to facilitate fluoro-de-diazonitation it should possess the following properties: (1) it should dissolve all the reagents with minimal solvation of fluoride anion; (2) it should be non- nucleophilic; (3) it should have suitably high redox potential to avoid reduction of the aryldiazonium ion and consequently suppress the homolytic decomposition pathway; (4) it should be aprotic; and (5) it should have a high enough boiling point as radical decomposition pathway is kinetically and thermodynamically favored to an ionic pathway. Chlorinated solvents (e.g. CCl4) have been reported to have a beneficial effect on arylfluoride yields via probable enhancement of the ionic decomposition pathway [60]. Selection of the suitable counter- anion with non-nucleophilic and non-reducing properties is also an important consideration to avoid its interference with fluorination. Since only one of the four fluorine atoms is retained in the final product after thermal decomposition of diazonium tetrafluoroborate salt, the Balz-Schiemann reaction is from a radiochemical point very inefficient (Scheme 2). Therefore, the yield of [18F]F- labeling is theoretically limited to only 25%,
18 - 18 - and more importantly, the dilution of SA by the non-radioactive fluorine in the counterion [ F]BF4 / F-BF3 /
(carrier added 18F- in the form of 18F-labeled tetrafluoroborate anion) is high with the consequential great possibility of failure to localize the biochemical target in the desired tissue(s) [61]. Namely, co-administered non-radioactive tracer would saturate targets by competitive displacement of radiotracer and thereby remove most of the localized signal from the radiotracer binding. In spite of the described limitations, the Balz-
Schiemann reaction was the first method used in nucleophilic 18F-labeling of arenes [62] and adapted to prepare some of the electron-rich 18F-labeled amino acids, such as 18F]fluorophenylalanine isomers [63–65], 5- and 6-
[18F]fluorotryptophan [66] as potential pancreas scanning agents, and 3,4-dihydroxy-5-[18F]fluorophenylalanine
(5-[18F]fluoro-DOPA) as a potential brain capillaries scanner [67–71].
Scheme 2. The Balz-Schiemann reaction for the preparation of [18F]fluoroarenes; 18F is introduced as 18F-labeled tetrafluoroborate anion (carrier-added) via exchanged reaction.
25
The syntheses of typical butyrophenone antipsychotic [18F]haloperidol (1) [70] and antifungal 4-
[18F]fluconazole (1-2% decay non-corrected yield) [71] were performed by the modified Balz-Schiemann reaction. Also noteworthy, a 35.5% incorporation of 18F into the [18F]haloperidol product from the corresponding diazonium fluoroborate precursor was reported (Scheme 3) [70]. A higher RCY than the maximal theoretical yield of 25% might be explained by the fact that 18F introduced into the arene can come from [18F]F- which exchanges into the fluoroborate salt (carrier-added) as well as coming from unexchanged [18F]F- (non-carrier- added) [66]. Nevertheless, [18F]haloperidol can be more conveniently prepared with higher SA using a single- step 18F-for-LG exchange reaction [72] because the butyrophenone system is moderately activated toward a direct aromatic nucleophilic substitution.
Scheme 3. Synthesis of [18F]haloperidol (1) by Balz-Schiemann reaction [70].
Knöchel and Zwernemann investigated several reaction parameters (solvents, reaction temperature and time, pH of the reaction medium, counter-anions, phase transfer catalysts, and the fluoride sources) influencing the [18F]F- fluorination RCYs of p-tolyldiazonium ion used as a model precursor in a modified Balz-Schiemann reaction [73, 74]. The best result was obtained with p-tolyldiazonium 2,4,6-tri-isopropylbenzenesulfonate in p- chlorotoluene [74]. Under optimized conditions n.c.a. labeling with [18F]fluoride gave a decay corrected (EOB) radiochemical yield of 39% (60 minutes) 4-[18F]fluorotoluene (2) with a calculated specific radioactivity around
1GBq/μmol in a total synthesis time of 48 minutes (Scheme 4) [74]. In comparison, this method gave approximately 104 times higher specific radioactivity of 4-[18F]fluorotoluene than fluoro-de-diazonitation using
18 18 [ F]-BF3 as counter-anion in preparation of [ F]haloperidol [70]. It should be noted that during optimization much lower or no fluorination yields of non-radioactive 4-fluorotoluene were obtained using K222/KF and/or 18-
Crown-6/KF complexes of the solubilized fluoride ion. The authors believe that the poor fluorination yields and slow reaction rates were related to ability of K222 and 18-Crown-6 to complex the aryldiazonium ions [75–77]. In this regard solubilization of the fluoride was reduced, since parts of the cryptand reacted with aryldiazonium ion
26
instead of mobilizing the potassium fluoride, thus enhancing the thermal stability of the complexed aryldiazonium ion and giving much lower fluorination yields.
total synthesis time 48 min
o 18 N2 X 110 C, 30 min F 15-Crown-5 / Na18F
4-chlorotoluene
CH3 CH3 dryed 2 39% decay corrected RCY (60 minutes) SA ~ 1 GBq/mol X = SO3
Scheme 4. No-carrier-added 18F-labeling of p-tolyldiazonium 2,4,6-triisopropylbenzenesulfonate by a modified
Balz-Schiemann reaction [74].
3. RADIOFLUORINATION OF ELECTRON-RICH ARENES BY WALLACH REACTION
1-Aryl-3,3-dialkyltriazenes (Ar-N=N-NR'R''), compounds having a diazoamino group are regarded as a protected form of anilines and stable surrogates of aryldiazonium ions [78]. Aryltriazenes are safely and mostly readily prepared by the coupling of diazotized aniline with amine, or by the action of Grignard reagents on aryl azides [79], and can be stored for a long period of time in cold temperatures protected from light. They are adaptable to numerous synthetic transformations with wide applicability in the chemical, medical, and technological fields [80, 81]. An essential advantage over aryldiazonium ions (Balz-Schiemann reaction) is their solubility in a number of (anhydrous) organic and ionic solvents that allows in situ generation of the corresponding aryldiazonium ion upon reacting with acids, and therefore enabling a one-pot fluoro-de- triazenation reaction [60, 82]. In this respect laborious isolation, drying, and the accumulation of the potential hazardous and thermally unstable diazonium intermediates is avoided. Moreover, aryltriazenes can be easily isolated, chromatographically purified, introduced in the early stages of the synthesis, and functionalized and thermally decomposed in the presence of protic acid in the latest stages if the preceding reactions have been performed under non-acidic conditions [83, 84]. Thus, fluoro-de-triazenation can be an attractive means of forming 18F-labeled fluoroarenes by direct nucleophilic substitution with [18F]F- due to both the one-pot
27
methodology and rapid nature of triazene transformation to fluoroarenes in order to obtain the tracers with good
SA. Nevertheless, the decomposition of aryltriazenes proceeds via diazonium ions and consequently leads to the same yield and reproducibility problems as (modified) Balz-Schiemann method (Scheme 5).
Pages and Langlois [84] investigated the acidic decomposition of simple, non-activated 1-aryl-3,3- dialkyltriazenes in the presence of non-radioactive fluoride anions (19F-) by examining different parameters such as solvent, acid, fluoride source, stoichiometry, reaction temperature, time, and introduction order of the reaction components, in order to build a good model for radiofluorination so the radical decomposition pathway could be suppressed and the ionic pathway maximized. According to their findings, they have successfully radiofluorinated protected (S)-[18F]-3-fluoro-α-methylphenylalanine (1) under optimized conditions through acidic decomposition of the corresponding 1-aryl-3,3-dimethyltriazene precursor with 15 % RCY (decay corrected (d.c.) RCY , Scheme 6) [85].
Scheme 5. General competitive processes during n.c.a. [18F]fluoro-de-triazenation and [18F]fluoro-de- diazonitation in the protic acid mediated decomposition of 1-aryl-3,3-dialkyltriazenes [84].
Scheme 6. Radiofluorination of protected form (S)-[18F]-3-fluoro-α-methylphenylalanine (3) through acidic decomposition of the corresponding 1-aryl-3,3-dimethyltriazene precursor [85].
The choice of the suitable protic acid is an important consideration to transform triazeno moiety via heterolytic decomposition into a diazonium group in order to obtain arylfluorides in reasonable yields. As the 28
phenyl cation is highly reactive and a non-selective species, fluoro-de-triazenation is often accompanied by the formation of a substantial amount of the acid counterion-substituted byproduct (Ar-A). This is especially true when the protic acid is used in excess, even if its conjugate base (A-) is considered non-nucleophilic (e.g. triflate anion) [84–87]. Since PET labeling reactions are performed with nanomolar amounts of [18F]F-, there is a vast stoichiometric excess, typically about 103-104-fold, of unlabeled precursor and acid [85]. Thus, the use of an acid constitutes a severe limitation to obtain considerable yields of Ar18F by fluoro-de-triazenation. The competitive reactions during fluoro-de-triazenation are the main reason for very few examples of the successful application of the Wallach reaction in the preparation of complex [18F]fluoroarenes with RCYs not exceeding 2%: [18F]-1- methyl-4-(2-fluorophenyl)-1,2,3,6-tetrahydropyrine (2'-18F-MPTP) [88], [18F]haloperidol, and [18F]spiroperidol
[89].
Recently, Riss et al. [90] reported successful solid phase supported [18F]fluoro-de-triazenation of 2- phenoxy-1-(aryldiazenyl)piperazine to afford 1-[18F]fluoro-2-phenoxybenzene (4) in up to 14% RCYs using reaction conditions similar to that of Pages et al. (Scheme 7) [85]. It is however questionable whether the solid phase synthesis offers a significant advantage over the synthesis in solution, since the last method yielded the
18F-labeled product under essentially the same conditions in a higher RCY of 23%.
Scheme 7. (i.) N.c.a. [18F]fluoro-de-triazenation of solid phase supported 1-(aryldiazenyl)piperazine and (ii.)
''soluble'' n.c.a. [18F]fluoro-de-triazenation of 1-(aryldiazenyl)piperazine to obtain 1-[18F]fluoro-2- phenoxybenzene (4) [90].
29
Because of limited RCYs, [18F]fluoro-de-triazenation is rarely applied for the production of 18F-labeled tracers nowadays.
4. RADIOFLUORINATION OF ELECTRON-RICH ARENES BY DIARYLIODONIUM SALTS
Hypervalent iodine compounds have more than eight electrons in their valence shell [91]. The fundamental feature of these compounds is the highly polarized three-centered-four-electron (3c-4e) bond, in which the central iodine atom is electron-deficient or bears a positive charge, and monovalent ligands (L) share the corresponding negative charge. Thus, hypervalent iodine compounds react as electrophiles or/and oxidants at the iodine center. They belong to two general structural types: (1) iodine (III) compounds A and B, also named
λ3-iodanes; and (2) iodine (V) compounds C and D, termed λ5-iodanes according to IUPAC nomenclature
(Scheme 8) [91, 92]. In 10-I-3 species the interchange of axial and equatorial ligands via Berry pseudorotation as well as turnstile rotation is rapid, while such fluxional processes in 12-I-5 species is slower [93–95]. Hypervalent iodine compounds are used as mild, non-toxic (compare to heavy metals) and selective reagents due to exploitation of their electrophilic and excellent leaving-group character in a wide range of applications [91, 92,
96–99].
Scheme 8. General oxidation of iodine compounds. Polyvalent iodine compounds differ in the number of valence electrons surrounding the central iodine atom, the number of ligands and their chemical structure. In terms of the Martin-Arduengo N-X-L designation/notation [100], 8-I-2 (A) and 10-I-3 (B) species are derivatives of trivalent iodine and are termed according to IUPAC λ3-iodanes. 10-I-4 (C) and 12-I-5 (D) species are derivatives of pentavalent iodine and are termed λ5-iodanes (periodanes). According to the hypervalent model 30
[101], the apical hypervalent 3c-4e bond in 10-I-3 species is close to linear, longer, and weaker compared to a regular covalent and equatorial bond and is responsible for high electrophilic reactivity of λ3-iodanes [96].
Scheme 9. Structure and configurations of diaryliodonium salts in solution and in the solid state.
The most common, stable, and well established class among polyvalent iodine compounds are diaryliodonium salts (ArI+Ar' X-). Their chemistry, preparative methods, and synthetic applications have been covered in previous reviews [43, 91, 92, 96]. According to conventional classification, they are defined as diaryl-
λ3-iodanes or as positively charged 8-I-2 species with two aryl ligands and a closely associated negatively charged counterion. In a solid state, the overall X-ray experimentally determined geometry is pseudo-trigonal bipyramidal or approximately T-shaped with characteristic 3c-4e linear hypervalent bond (Schemes 8 and 9) [91,
92, 102, 103]. The least electronegative aromatic carbon or the most sterically demanding aryl group and both electron pairs reside in equatorial positions. On the other hand, the configuration of λ3-iodanes in solution is still debated, but a certain amount of dissociation is expected depending on the aryl substituents, anion of the salt, and the type of the solvent used [103]. Diaryliodonium salts are air- and moisture-stable, mild, nontoxic, and versatile, selective arylating agents delivering one of the aryl moieties to the nucleophile under polar, catalytic, or photo-chemical conditions. The salt is referred to as a symmetrical salt if R1 = R2, and as an unsymmetrical salt if R1 ≠ R2 (Scheme 9).
When utilizing diaryliodonium salts in reactions with nucleophiles, the nucleophile has a choice of displacing either of the two aryl groups due to the exceptional leaving group ability of the -IAr fragment, which has been estimated to be roughly six orders of magnitude greater than that of triflate [104]. In this instance, one of the more recent application areas of diaryliodonium salts, firstly introduced in 18F-radiochemistry by Pike and
31
Aigbirhio in 1995 [105], is their use as suitable precursors for the preparation of n.c.a. 18F-labeled arenes as potential radiotracers for PET imaging. Diaryliodonium salts allow direct introduction of [18F]fluoride ion in a single step into aromatic systems without the need for further activating groups and with little or no restriction on the nature of the functionality present [9]. Thus, they gain more and more interest for direct radiofluorination of otherwise unfavorable electron-rich arenes. At this, symmetric diaryliodonium salts are generally preferred over unsymmetrical salts as no regioselectivity problems arise (Scheme 10). However, the use of unsymmetrical salts is in some situations desirable and necessary, such as when the starting materials are expensive or if one of the aryl substituent is very complex.
Scheme 10. Regioselectivity problems of direct n.c.a. [18F]-fluorination of unsymmetrical diarylidonium salts.
Table 1. Synthesis of simple n.c.a. [18F]fluoroarenes via nucleophilic displacement of corresponding 4- substituted diaryliodonium salt by n.c.a. [18F]fluoride. Selectivity for the product 6 increases with relative increase of electron density on the partner para-substituted phenyl ring.
Compound Ref. R R' X- RCY 5 (%) RCY 6 (%) Selectivity for 6
- a a 5a/6a [105] CH3 H TfO ~13 ~26 2 - b a 5b/6b [105] OCH3 H Br ~62 > 62 - a 5c/6c [105] OCH3 4-OCH3 CF3CO2 ~55 - 5d/6d [107] OCH3 H TfO 0 96 > 96 5e/6e [107] t-BuO H TfO- 0 95 > 95 - 5f/6f [107] OCH3 3-CH3 TfO 0 66 > 66 - 5g/6g [28] OCH3 3-Br TfO 3±1 43±5 ~ 14 - 5h/6h [108] CH3 H TsO 10 43 ~ 4 - a a 5i/6i [108] OCH3 H TsO ~3 ~77 ~ 26 - 5j/6j [108] OCH3 4-CH3 TsO 9 36 4 - 5k/6k [108] OCH3 4-Cl TsO 6 80 ~ 13 - 5l/6l [108] OCH3 4-OCH3 TsO 90 - 5m/6m [109] OCH3 3-CN TsO 11 82 ~ 7 - 5n/6n [109] OCH3 3-NO2 TsO 7 58 ~ 8 - 5o/6o [109] OCH3 3-CF3 TsO < 1 53 > 53 a Average RCY, b Not detected 32
The outcome and regioselectivity of the fluorination has been shown to be strongly dependent on three parameters: (1) the electronic density (inductive and resonance effects of substituent groups on the aryl ring); (2) the substitution pattern; and (3) the steric structure of the diaryliodonium precursor [106]. Following the trend of
18 SNAr, the n.c.a. [ F]fluoride attack occurs preferably on the most electron-deficient arene in the absence of an ortho effect [9]. The strategy is therefore to make one of the aryl rings more electron-rich. Iodonium salts containing p-methoxyphenyl (p-anisyl) (Table 1) [28, 105, 107–109] and 2-thienyl (Table 2) [106, 109–112] as highly electron rich arenes were found to lead to highly regioselective nucleophilic 18F-fluorination of the partner aryl ring.
Table 2. Radiochemical yields of n.c.a. 18F-fluorination of 4-phenyl(2-thienyl)iodonium bromides to obtain simple electron-rich and electron-poor 4-substituted 18F-fluoroarenes [106].
R H CH3 OCH3 OBn Cl Br I
RCY(%) 64 ± 4 32 ± 2 29 ± 3 36 ± 3 62 ± 4 70 ± 5 60 ± 8
Carroll et al. [113] have performed fluorination of simple 2-thienyliodonium salts and found that the process is not selective, and the 2-thienyl group, although highly electron-rich, may not be the ideal non- participating ring for production of fluoroarenes by the fluorination of diaryliodonium salts as had been described the same year by Ross et al. [106]. They concluded that it is the analysis, characterization, and isolation of 2-fluorothiophene that may be (extremely) problematic. Iodonium compounds with bulky aryl rings tend to undergo nucleophilic substitution on the bulky ring. However, the bulkiness alone is not an exclusive factor for stereoselective fluorination. The decisive factor is one (hydrophobic) group (e.g. methyl) at the ortho- position that shows a directing steric effect and induces an attack in the ortho-substituted aromatic ring [114–
117], even though it is more electron-rich than the partner ring (Table 3, Entries 1, 3, 9-11, 15-16) [107, 118].
This preference increases further along with RCYs for doubly ortho- and/or more alkyl/methyl-substituted aryl rings in spite of increasing steric hindrance and electronic deactivation (Table 3, Entries 9-11, 15-17) [107, 118–
120].
33
Table 3. Radiochemical yields and reaction selectivities of [18F]fluoroarenes obtained by radiofluorination of ortho-substituted diaryliodonium salts.
[18F]F- + - 18 18 ArI Ar'X Ar F + Ar' F
RCY of [18F]fluoroarene (%) Selectivity Entry Ref. Ar Ar' X- Ar18F Ar'18F for Ar18F
- 1 [118] 2-MeC6H4 Ph Cl 57 25 ~2 - 2 [118] 2-MeC6H4 2-MeC6H4 Cl 83 - 3 [118] 2-MeOC6H4 Ph Cl 6.5 60 ~0.1 - 4 [118] 2-MeOC6H4 2-MeOC6H4 Cl 51 - 5 [118] 2-BrC6H4 Ph Cl 68 25 ~3 - 6 [118] 2-MeC6H4 2-MeOC6H4 Cl 75 4 ~19 - 7 [118] 2-MeC6H4 2-EtC6H4 Cl 52 43 ~1 - 8 [118] 2-MeC6H4 2-i-PrC6H4 Cl 48 40 ~1 - 9 [118] 2,6-di-MeC6H3 2-MeC6H4 Cl 62 11 ~6 - 10 [118] 2,4,6-tri-MeC6H2 Ph TsO 63 19 ~3 - 11 [118] 2,4,6-tri-MeC6H2 2-MeC6H4 Cl 59 33 ~2 - 12 [118] 2,4,6-tri-MeC6H2 2,6-di-MeC6H3 TsO 21 61 ~0.3 - 13 [107] 2-MeC6H4 4-t-BuC6H4 TfO 48 12 4 - 14 [107] 2-MeC6H4 4-MeOC6H4 TfO 64 0 > 64 - 15 [107] 2,4,6-tri-MeC6H2 Ph TfO 96 0 > 96 - 16 [107] 2,4,6-tri-MeC6H2 2-MeC6H4 TfO 52 13 4 - 17 [107] 2,4,6-tri-MeC6H2 4-MeOC6H4 TfO 67 0 > 67 - a a 18 [106] 2-MeOC6H4 2-thienyl Br 61±5 a Not detected.
The so-called ‘ortho-effect,’ first mentioned in 1967 by Le Count et al. [114], can be explained by examining the arrangement of the aryl groups around the iodine-centered pseudo trigonal bypiramidal intermediate; sterically more demanding ortho-substituted aromatic ring is preferentially in the equatorial position to decrease the steric strain to a greater extent, and is therefore more proximal (syn) for 18F-fluorination compared to other less bulky axial positioned arene (Scheme 11) [116, 117, 119–121].
34
Scheme 11. Suggested outline mechanism for the radiofluorination of an unsymmetrical substituted diaryliodonium salt through transition states TS1 and TS2 to give 18F-labeled products P1 and P2, respectively
[118].
As studied and rationalized by Chun et al. [118] the electronic features of the ortho-substituent and transition state stabilities are important in determining product selectivity, because the ortho-effect is not purely related to substituent bulkiness (Table 3, Entries 3, 6-8). Opposing (Table 3, Entry 3) and reinforcing (Table 3, Entry 5) electronic properties of ortho substituent appeared to be important in determining product selectivity. Thus, when both of the ortho substituents are of similar effective bulkiness, [18F]fluoride together with the more electron-deficient ligand would eliminate in order to decrease the positive charge on the iodine atom (Table 3,
Entries 6-8). They suggested that ortho-hydrophobic groups create lipophilic micro-environment in which the incoming [18F]fluoride can act as a powerful nucleophile (even in moderately hydrated state), by first loosely binding to the hypervalent iodine atom and then attacking the locally lipophilic ortho-substituted ring (Scheme
11). They also found evidence that the reactions of [18F]fluoride with unsymmetrical diaryliodonium salt (2-
35
methylphenyl)(phenyl)iodonium chloride) comply with the Curtin-Hammett principle [118, 122]. Accordingly, they proposed that radiofluorination of an unsymmetrical diaryliodonium salt involves an attack of [18F]fluoride onto either of the two rapidly interconverting conformers [94, 95] relative to the rate of product formation, yielding two transition states (TS1 and TS2) that each give a single radiofluorinated product (P1, P2), whereas the products do not undergo interconversion (Scheme 11). Consequently, the ratio of radiofluorinated products
(P1/P2) was not in direct proportion to the relative concentrations of the conformational isomers in the substrate, but is dependent only on the difference in standard free energies of the respective transition states
(P1/P2=e-(GTs1-GTs2)/RT), which is relatively small compared to the activation energies for the same reactions. In other words, P1/P2 will increase with increasing difference between free energies of the respective transition states. As depicted in the Scheme 11, generally higher Ar18F yields compared to Balz-Schiemann and Wallach reactions probably arise from a [18F]fluoride preferential attack on the proximal equatorial aryl ring, and subsequent decomposition of “trigonal” transition state which limits the formation of reactive intermediates, e.g. aryl cations or aryl radicals. This might explain why even the addition of water is not detrimental for successful radiofluorination [118, 123–125]. Other mechanisms for reactions of diaryliodonium salts with [18F]fluoride have also been proposed, such as ʽturnstileʼ mechanism by Grushin [117] and SNAr by Ross et al. [106] on the basis of a reasonable good linear fit between reaction rates of para-substituted aryl(2-thienyl)iodonium bromides and Hammett substituent constants. The application of the Hammett constants failed for the ortho-OMe substituted precursor, as claimed by Authors, due to the strong ortho-effect which could not be taken into account by that approach. Meta-derivatives are known to be electron-rich in comparison to corresponding para- and ortho-derivatives, and are therefore more problematic for nucleophilic fluorination. Unsymmetrical diaryliodonium salts have also been shown to be effective precursors to obtain simple 3-[18F]fluoroarenes bearing meta electron-withdrawing or meta electron-donating substituents [109, 112, 126] 3-
[18F]fluoroheteroarenes [127] in moderate to high RCYs. Also noteworthy is that pyridyliodonium salts are the most appropriate precursors for the preparation of fluorine-18 labeled 3-fluoropyridine that is more stable in vivo, but less easily available via conventional SNAr reaction than 2-fluoro and 4-fluoropyridines. The use of 4- methoxyphenyl moiety as a partner aromatic ring in 4-methoxyphenyl(3-pyridine) iodonium salt gave 3-
[18F]fluoropyridine in radiochemical yields of about 60%. The same approach has been employed to yield 3-
[18F]fluoroquinoline in about 25% RCY via a 4-methoxyphenyl(3-quinoline) iodonium salt precursor [127].
36
Besides the substitution pattern and electronic properties, parameters such as solvent, counter-anion(s),
[18F]fluoride source, stoichiometry, reaction temperature, and time, strongly affect the outcome (product distribution) and RCYs of nucleophilic 18F-labeling reaction with diaryliodonium salts. An appropriate solvent should be non-nucleophilic with weak power of cation and anion solvation, and with suitable redox potential to exclude or limit any redox processes between iodine (III) and solvent molecules [28, 106, 128]. Organic aprotic solvents like CH3CN and DMF have appeared to be the most beneficial, according to the RCYs in contrast to
DMSO which is very useful for direct SNAr. To avoid the homolytic aryl-iodine bond fission [117], and consequently to improve RCYs and reproducibility, addition of radical scavengers (e.g. TEMPO) has been shown to be advantageous [125, 129–131]. According counter-anionʼs influence on RCYs, stability, reaction rates, and selectivity, inorganic and organic counter-anions such as bromide [105, 106], tosylate (TsO-) [108,
109, 124, 125, 131, 132], and triflate (TfO-) [105, 107] have appeared attractive considering their low nucleophilicity and good leaving group ability. Lee et al. determined X-ray structure of a representative unsymmetrical iodonium salt, 2-methylphenyl(2ʼ-methoxy-phenyl)iodonium chloride (7), in order to achieve a deeper understanding of its structure, and to assist in understanding radiofluorination mechanism and similar reactions of diaryliodonium salts with nucleophiles in organic media [133]. Their X-ray study unveiled that the hypervalent iodine in 7 acts as a previously unrecognized stereogenic center within a dimeric structure as the unit cell in a centrosymmetric crystal, composed of conformational M and P enantiomers (Scheme 12). They investigated racemization process of 7 in CH3CN solution with the ab initio replica path method, thereby revealing two additional pairs of conformational enantiomers. All identified six conformers of 7 were calculated to be comparable in energy and thus, all are likely to exist in CH3CN at room temperature due to fast interconversion via two essentially isoenergetic transition states (calculate energy barrier of 9,1 kcal/mol in
CH3CN). In addition, their quantum chemical and dimerization energy calculations together with LC-MS observations of clusters of 7 suggested that it predominantly exists as dimers (dimeric anion-bridge clusters) in
CH3CN due to the secondary bonding interaction between I and the Cl atoms within an enantiomeric pair. The evidence of the existence of dimeric solution clusters of 7 further indicates that the reactions of diaryliodonium salts similar to 7 with nucleophiles (e.g. [18F]F-) in organic solvents may require dissociation of dimers or possibly even higher-ordered clusters (e.g. tetramers), preceding replacement of chloride ion with [18F]F- and subsequent attack of the bound fluoride onto an aryl carbon atom to give either of the two possible
[18F]fluorarene products (Scheme 12).
37
i Cl Cl ii
Cl Ar Ar' disssociation I I I ArI+Ar'Cl- I Ar' Ar Cl 18 - O O [ F]F Cl-bridged dimer in organic solvent (square planar configuration) Ar18F/Ar'18F + Ar'I/Ar
M 7 P
Scheme 12. (i.) M and P are conformational enantiomers of 2-methylphenyl(2ʼ-methoxy-phenyl)iodonium chloride (7). (ii.) Favorable dimerization energy calculation and LC-MS observations of clusters suggest that dissociation of dimeric Cl-bridge cluster, held together by iodine-chloride ionic bonds, of diaryliodonium salt similar to 7 is preceding necessary step to allow formation of the two possible [18F]fluorarene products [133].
So far, diaryliodonium salts have mainly been proven to be useful precursors for the introduction of
[18F]F- onto simple aromatic rings via straightforwardly prepared precursors where comparison with the Wallach reaction has shown much greater efficiency for the former methodology [28, 134]. It was assumed that the 18F- labeling via iodonium precursors is somehow limited by the molecule size and complexity of the structure [120].
With further mechanistic, stability, permutational, and intramolecular interference studies, and also with an improvement of existent and development of new synthetic methods to obtain stable and highly pure complex diaryliodonium precursors, this approach will likely find wider application in preparing more complex 18F- labeled tracers for PET imaging. More complex 18F-labeled radiopharmaceuticals from corresponding diaryliodonium salts are represented in Schemes 13 and 14.
38
Scheme 13. Successful examples of 18F-labeling using complex diaryliodonium precursors. (i.) Radiosynthesis of metabolically stable 4-[18F]fluorophenyl pyrazolo steroid (9) as high affinity ligand for brain glucocorticoid receptors [132]. (ii.) Radiosynthesis of 11 ([18F]F-ADTQ), classified as non-competitive AMPA receptor antagonist [120].
Radiosynthesis of metabolically stable 4-[18F]fluorophenyl pyrazolo steroid 9 as high affinity ligand for brain glucocorticoid receptors was accomplished by Wüst et al. [132]. Since the aromatic ring is not sufficiently activated by a strong electron-withdrawing group, diaryliodonium salts 8a and 8b were used as precursors for the incorporation of n.c.a. [18F]F- to obtain 9 in low decay-corrected radiochemical yields of 0.2 and 2.0%, respectively. The use of the more electron-donating tolyl-functionalized iodonium salt 8b favored the formation of corticosteroid 9 which is in accordance of the para-substituted electronic effects of the counter rings. Authors
39
also detected by radio-TLC analysis the formation of large amounts of [18F]fluorobenzene and [18F]fluorotoluene as by-products. Ross reported radiosynthesis of 11 ([18F]F-ADTQ), classified as a putative, non-competitive
AMPA receptor antagonist, from iodonium precursors 10a, 10b, and 10c [120]. Comparing the precursors 10a and 10b, the RCYs showed an increase from the phenyliodonium group (1.2 %) to the 2-thienyliodonium group
(2.9 %), as expected for the electronic differences between the iodonium precursors and the corresponding aryl iodides as leaving groups. It should be noted that the synthesis starting from iodonium precursor 10c with a bromide as a counter-anion showed the best RCY of about 3.6%, and was obtained in only 3% yield.
Consequently, 11 was not isolated or prepared for pharmacological evaluation studies because of a very low
RCY.
Preparation of 13 ([18F]DAA1106) by Zhang et al., a PET ligand for imaging peripheral-type benzodiazepine receptor in the brain, was accomplished in much higher d.c. RCY of 46 % than in previous noted attempts and was therefore the first report of a complex and electron rich practical PET ligand synthesized by the reaction of diphenyliodonium salt with n.c.a. [18F]F- in high RCY [108]. Based on consideration that the
[18F]fluoride attacks the diphenyliodonium salt preferably at the electron-deficient benzene ring, p-iodoanisole as a leaving group was designed to increase the regioselectivity of 18F into a desired ring. Since 12 was unstable, it was used for radiosynthesis after the respective coupling reaction without further purification. Another successful radiofluorination in high RCY via iodonium salts is the preparation of [18F]flumazenil (15) using 4- methylphenyl-mazenil iodonium tosylate precursor 14 without any structural modifications of the parent molecule [131]. Interestingly, 14 was superior to other precursors, in spite of the fact that 2-thienyl-, 3- thienyl-, and 4-methoxyphenyl-mazenil iodonium tosylate have relatively high electron densities. The authors observed that the more electron-rich diaryliodonium tosylate precursors have lower stability and selectivity for the desired product formation and that these correspond with the tendency of the [18F]fluoride incorporation yield. Thus, the
18 best result was obtained for the reaction of 14 with n.c.a. [ F]fluoride and K222/K2CO3 (0.6 equiv. of K2CO3 relative to the precursor) in the presence of TEMPO in DMF at 150 °C for 5 min. Under these conditions, d.c.
RYC was 67 ± 2.7 % with more than 99% radiochemical purity after HPLC purification. The total synthesis time for 15 was about 55 min, including HPLC purification and the specific activity was in the range of 370-450
GBq/μmol (10-12 Ci/μmol). Further studies showed that the optimized reaction conditions were well adapted to the reproducible (n = 26 with no failure) high-scale automatic production of [18F]flumazenil (RCY 63.5 ± 3.2 % in total synthesis time 60 ± 1.1 min) in a commercial automated device (Scheme 14).
40
Scheme 14. Successful examples of 18F-labeling using complex diaryliodonium precursors. (i.) Preparation of 13
([18F]DAA1106), a PET ligand for imaging peripheral-type benzodiazepine receptor in the brain [108]. (ii.)
Preparation of 15 ([18F]flumazenil), chemically indistinguishable from its non-radioactive counterpart, using 4- methylphenyl-mazenil iodonium tosylate precursor 14 [131]. (iii.) Radiochemical synthesis of benzophenone- tyrosine PPARγ ligand 17 [124]. (iv.) Radiosynthesis of 6-[18F]fluoro-labeled benzothiazole analogue 19 as a promising PET probe for Aβ plaque imaging [125].
41
Lee et al. developed radiochemical synthesis of 18F-labeled analog of the potent and selective PPARγ agonist farglitazar (17), a tyrosyne-benzophenone class of PPARγ regulators) by radiofluorination of a diaryliodonium tosylate precursors 16a and 16b [124]. The radiosynthesis of 17 was accomplished in approximately 90 minutes with a good d.c. RCY of up to 42% and the SA after decay correction of approximately 37 GBq/μmol (1 Ci/μmol). Authors concluded that although the compound had high and selective
PPARγ binding affinities and also good metabolic stability, its nonselective target-tissue (brown and/or white fat) biodistribution uptake versus non-target tissues in rats indicated that it was likely to be unspecific for effective imaging of breast cancer or vascular disease in humans. It should be noted that the addition of some water significantly increased RCYs as suggested due to increased solubility of Cs[18F]F- salt in the studied solvents (DMF and CH3CN). It is also interesting to note that they were unable to obtain radiolabeled product 17 from the p-methoxyphenyl-based iodonium salt precursor, even though this precursor worked quite well to produce the unlabeled fluoroproduct [124, 135]. Similarly, the same authors (Lee et al.) have recently synthesized various diaryliodonium tosylate precursors (containing the more electron-rich p-methoxyphenyl, p- methylphenyl, 2- and 3-thienyl compare to phenyl ring on 6-position of benzothiazole ring) to allow aromatic
18F-labeling at 6-position of benzothiazole ring [125]. The highest d.c. RYC 40.5% with SA 110 GBq/μmol
(2.97 Ci/μmol) of one of the most promising Aß plaque-specific PET imaging probes 19 was obtained via one- pot radiofluorination and deprotection of benzothiazole iodonium tosylate precursor 18 in the presence of
+ 18 TEMPO using nBu4N [ F]fluoride as the fluoride source.
5. CONCLUSIONS
A direct, one-step, no-carrier-added synthesis of inactivated or electron-rich [18F]fluoroaromatic compounds with high specific activity represent an important challenge for radiochemists. This is limited to only a few methods in preparative organic chemistry. The traditional Balz-Schiemann reaction with its modification and Wallach reaction are not particularly efficient due to the competing reactions of the highly reactive and non- selective intermediates leading to low radiochemical yields and specific activities of 18F-labeled arenes. On the other hand, diaryliodonium salts have been shown to be the most suitable precursors for direct single-step nucleophilic [18F]fluorination of simple arenes with little or no restriction on the nature of the functionality present. The regioselectivity of this reaction has been found to be controlled electronically as well as by the steric bulkiness of the substituents. The so-called ortho-effect is the most prominent feature of this methodology
42
and almost quantitatively radiochemical yields of small electron-rich arenes have been reported due to this effect. Higher radiochemical yields compared to Balz-Schiemann and Wallach reactions have also been explained by the synchronous reductive elimination of the transition state which limits formation of reactive intermediates. The principal drawback of this method is that radiochemical yield is limited by the additional formation of undesirable counter [18F]fluoroarene. Thus, to improve the radiochemical yield, symmetrically substituted diaryliodonium salts as precursors are preferred, which is not the case for complex arenes. So far, this approach has been limited for simple readily prepared diaryliodonium salts. However, very promising for future investigation are a few recent successful examples of direct aromatic nucleophilic 18F-labeling of complex non- activated radiopharmaceuticals using corresponding diaryliodonium salts as precursors. With further mechanistic studies and with an improvement of existent synthetic methods and development of new to obtain stable and highly pure complex diaryliodonium precursors, this approach will likely find wider application in radiofluorination.
List of abbreviations: AMPA - 2-amino-3-(3-hydroxy-5-methyl-isoxazol-4-yl)propanoic acid, EWG - electron- withdrawing group, LG - leaving group, N.C.A. - no-carrier-added, PET - positron emission tomography, PPAR
- peroxisome proliferator-activated receptor, RCY - radiochemical yield, SA - specific activity, TEMPO -
2,2,6,6-tetramethylpiperidin-1-yl)oxyl.
ACKNOWLEDGEMENTS
This work was supported by Egide for graduate grant (PHC PROTEUS, 2012, n°26502QF). The authors would also like to acknowledge Slovenian Research Agency for financial support of Slovenian-French bilateral collaboration (project n° BI-FR/12-13-PROTEUS-007).
REFERENCES
[1] Ametamey, S.M.; Honer, M.; Schubiger, P.A. Molecular Imaging with PET. Chem. Rev., 2008, 108(5), 1501-
1516.
[2] Antoni, G.; Långström, B. Radiopharmaceuticals: molecular imaging using positron emission tomography.
Handb. Exp. Pharmacol., 2008, 185, 177-201.
43
[3] Dollé, F.; Roeda, D.; Kuhnast, B.; Lasne, M.-C. In: Fluorine and Health, Molecular Imaging, Biomedical
Materials and Pharmaceuticals; Alain Tressaud and Günter Haufe Eds.; Elsevier Science B. V: Amsterdam,
2008; pp. 3-65.
[4] Paans, A.M.J.; Vaalburg, W. Positron Emission Tomography in Drug Development and Drug Evaluation.
Curr. Pharm. Des., 2000, 6(16), 1583-1591.
[5] Lin, M.; Shon, I.H.; Lin, P. Positron Emission Tomography: Current Status and Future Challenges. Intern.
Med. J., 2010, 40(1), 19-29.
[6] Zimmer, L.; Luxen, A. PET Radiotracers for Molecular Imaging in the Brain: Past, Present and Future.
NeuroImage, 2012, 61(2), 363-370.
[7] Lee, C.-M.; Farde, L. Using Positron Emission Tomography to Facilitate CNS Drug Development. Trends
Pharmacol. Sci., 2006, 27(6), 310-316.
[8] Vallabhajosula, S. 18F-Labeled Positron Emission Tomographic Radiopharmaceuticals in Oncology: An
Overview of Radiochemistry and Mechanisms of Tumor Localization. Semin. Nuc. Med., 2007, 37(6), 400-419.
[9] Cai, L.; Lu, S.; Pike, V.W. Chemistry with [18F]Fluoride Ion. Eur. J. Org. Chem., 2008, 2008(17), 2853-
2873.
[10] Cai, L.; Lu, S.; Pike, V.W. Chemistry with [18F]Fluoride Ion. Eur. J. Org. Chem., 2008, 2008(17), 2843-
2848.
[11] Coenen, H.H. Fluorine-18 labelling methods: features and possibility of basic reactions. In: PET
Chemistry, The Driving Force in Molecular Imaging; P. A. Schubiger, L. Lehmann, M. Friebe, Eds.; Springer-
Verlag: Berlin Heidelberg, 2007; pp. 15-50.
[12] Kilburn, M.R.; Huizenga J.R. Fluorine-18 labelling of radiopharmarmaceuticals; Nuclear Science Series,
Natural Academic Press, Washington, DC, 1990.
[13] O’Hagan, D. Understanding Organofluorine Chemistry. An Introduction to the C-F Bond. Chem. Soc. Rev.,
2008, 37, 308-319.
44
[14] Zhou, P.; Zou, J.; Tian, F.; Shang, Z. Fluorine Bonding - How Does it Work in Protein-Ligand Interactions?
J. Chem. Inf. Model., 2009, 49(10), 2344-2355.
[15] Shah, P.; Westwell, A.D. The Role of Fluorine in Medicinal Chemistry. J. Enzyme Inhib. Med. Chem.,
2007, 22(5), 527-540.
[16] Kemp, D.D.; Gordon, M.S. Theoretical Study of the Solvation of Fluorine and Chlorine Anions by Water. J.
Phys. Chem. A, 2005, 109(34), 7688-7699.
[17] Schlyer, D.J.; Bastos, M.A.V.; Alexoff, D.; Wolf, A.P. Separation of [18F]Fluoride from [18O]Water Using
Anion Exchange Resin. Int. J. Rad. Appl. Instrum. A., 1990, 41(6), 531-533.
[18] Kim, D.W.; Jeong, H.J.; Lim, S.T.; Sohn, M.-H. Recent Trends in the Nucleophilic [18F]-radiolabeling
Method with No-carrier-added [18F]fluoride. Nucl. Med. Mol. Imaging, 2010, 44(1), 25-32.
[19] Briard, E.; Pike, V.W. Substitution-reduction: an Alternative Process for the [18F]N-(2-fluoroethylation) of
Anilines. J. Labelled Compd. Radiopharm., 2004, 47(4), 217-232.
[20] Kim, D.W.; Ahn, D.-S.; Oh, Y.-H.; Lee, S.; Kil, H.S.; Oh, S.J.; Lee, S.J.; Kim, J.S.; Ryu, J.S.; Moon, D.H.;
Chi, D.Y. A New Class of SN2 Reactions Catalyzed by Protic Solvents: Facile Fluorination for Isotopic
Labeling of Diagnostic Molecules. J. Am. Chem. Soc., 2006, 128(50), 16394-16397.
[21] Pike, V.W. PET Radiotracers: Crossing the Blood-Brain Barrier and Surviving Metabolism. Trends
Pharmacol. Sci., 2009, 30(8), 431-440.
[22] Baumann, C.A.; Mu, L.; Wertli, N.; Krämer, S.D.; Honer, M.; Schubiger, P.A.; Ametamey, S.M. Syntheses and Pharmacological Characterization of Novel Thiazole Derivatives as Potential mGluR5 PET Ligands. Bioorg.
Med. Chem., 2010, 18(16), 6044-6054.
[23] Al Hussainy, R.; Verbeek, J.; van der Born, D.; Molthoff, C.; Booij, J.; Herscheid, J.D.M. Synthesis,
Biodistribution and PET Studies in Rats of 18F-Labeled Bridgehead Fluoromethyl Analogues of WAY-100635.
Nucl. Med. Biol., 2012, 39(7), 1068-1076.
[24] Attiná, M.; Cacace, F.; Wolf, A.P. Displacement of a Nitro-Group by [18F] fluoride ion. A New Route to
Aryl Flourides of High Specific Activity. J. Chem. Soc., Chem. Commun., 1983, 108-109.
45
[25] Angelini, G.; Speranza, M.; Wolf, A.P.; Shiue, C.-Y. Nucleophilic Aromatic Substitution of Activated
Cationic Groups by 18F-labeled Fluoride. A Useful Route to No-Carrier-Added (NCA) 18F-labeled Aryl
Fluorides. J. Fluorine Chem., 1985, 27(2), 177-191.
[26] Haka, M.S.; Kilbourn, M.R.; Watkins, G.L.; Toorongian, S.A. Aryltrimethylammonium
Trifluoromethanesulfonates as Precursors to Aryl [18F]Fluorides: Improved Synthesis of [18F]GBR-13119. J.
Labelled Compd. Radiopharm., 1989, 27(7), 823-833.
[27] Karramkam, M.; Hinnen, F.; Bramoullé, Y.; Jubeau, S.; Dollé, F. Ortho-[18F]Fluoronitrobenzenes by No-
Carrier-Added Nucleophilic Aromatic Substitution with K[18F]F-K222-A Comparative Study. J. Labelled
Compd. Radiopharm., 2002, 45(13), 1103-1113.
[28] Ermert, J.; Hocke, C.; Ludwig, T.; Gail, R.; Coenen, H.H. Comparison of Pathways to the Versatile Synthon of No-Carrier-Added 1-bromo-4-[18F]fluorobenzene. J. Labelled Compd. Radiopharm., 2004, 47(7), 429-441.
[29] Sun, H.; DiMagno, S.G. Competitive Demethylation and Substitution in N,N,N-trimethylanilinium fluorides. J. Fluorine Chem., 2007, 128(7), 806-812.
[30] Dollé, F. Fluorine-18-Labelled Fluoropyridines: Advances in Radiopharmaceutical Design. Curr. Pharm.
Des., 2005, 11(25), 3221-3235.
[31] Sobrio, F.; Amokhtari, M.; Gourand, F.; Dhilly, M.; Dauphin, F.; Barré, L. Radiosynthesis of [18F]Lu29-
024: A Potential PET Ligand for Brain Imaging of the Serotonergic 5-HT2 Receptor. Bioorg. Med. Chem., 2000,
8(10), 2511-2518.
[32] Langer, O.; Dollé, F.; Valette, H.; Halldin, C.; Vaufrey, F.; Fuseau, C.; Coulon, C.; Ottaviani, M.; Någren,
K.; Bottlaender, M.; Maziére, B.; Crouzel, C. Synthesis of High-Specific-Radioactivity 4- and 6-
[18F]Fluorometaraminol- PET Tracers for the Adrenergic Nervous System of the Heart. Bioorg. Med. Chem.,
2001, 9(3), 677-694.
[33] Ishiwata, K.; Ishii, S.-I.; Senda, M.; Tsuchiya, Y.; Tomimoto, K. Electrophilic Synthesis of 6-[18F]fluoro-l- dopa: Use of 4-O-pivaloyl-l-dopa as a Suitable Precursor for Routine Production. Appl. Radiat. Isot., 1993,
44(4), 755-759.
46
[34] Hatano, K.; Ishiwata, K.; Yanagisawa, T. Co production of 2,6-[18F] difluoroDOPA during Electrophilic
Synthesis of 6-[18F]fluoro-l-DOPA. Nucl. Med. Biol., 1996, 23(1), 101-103.
[35] Luxen, A.; Perlmutter, M.; Bida, G.T.; Van Moffaert, G.; Cook, J.S.; Satyamurthy, N.; Phelps, M.E.; Barrio,
J.R. Remote, Semiautomated Production of 6-[18F]fluoro-l-dopa for Human Studies with PET. Int. J. Rad. Appl.
Instrum. A., 1990, 41(3), 275-281.
[36] Namavari, M.; Bishop, A.; Satyamurthy, N.; Bida, G.; Barrio, J.R. Regioselective radiofluorodestannylation with [18F]F2 and [18F]CH3COOF: A High Yield Synthesis of 6-[18F]fluoro-l-dopa. Int. J. Rad. Appl. Instrum.
A., 1992, 43(8), 989-996.
[37] Luxen, A.; Guillaume, M.; Melega, W.P.; Pike, V.W.; Solin, O.; Wagner, R. Production of 6-[18F]fluoro-l-
DOPA and its Metabolism in vivo-A Critical Review. Int. J. Rad. Appl. Instrum. B., 1992, 19(2), 149-158.
[38] Hess, E.; Sichler, S.; Kluge, A.; Coenen, H.H. Synthesis of 2-[18F]Fluoro-L-Tyrosine via Regiospecific
Fluoro-De-Stannylation. Appl. Radiat. Isot., 2002, 57(2), 185-191.
[39] Eskola, O.; Grönroos, T.; Bergman, J.; Haaparanta, M.; Marjamäki, P.; Lehikoinen, P.; Forsback, S.;
Langer, O.; Hinnen, F.; Dollé, F.; Halldin, C.; Solin, O. A Novel Electrophilic Synthesis and Evaluation of
Medium Specific Radioactivity (1R,2S)-4-[18F]Fluorometaraminol: A Tracer For The Assessment Of Cardiac
Sympathetic Nerve Integrity with PET. Nucl. Med. Biol., 2004, 31(1), 103-110.
[40] Bergman, J.; Solin, O. Fluorine-18-labeled fluorine gas for synthesis of tracer molecules. Nucl. Med. Biol.,
1997, 24(7), 677-683.
[41] Balz, G.; Schiemann, G. Über aromatische Fluorverbindungen, I.: Ein neues Verfahren zu ihrer Darstellung.
Ber. Dtsch. Chem. Ges., 1927, 60(5), 1186-1190.
[42] Wallach, O. Ueber das Verhalten einiger Diazo- und Diazoamidoverbindungen. Justus Liebigs Ann. Chem.,
1886, 235(3), 233-255.
[43] Merritt, E.A.; Olofsson, B. Diaryliodonium Salts: A Journey from Obscurity to Fame. Angew. Chem. Int.
Ed., 2009, 48(48), 9052-9070.
47
[44] Coenen, H.H.; Ermert, J. Direct Nucleophilic 18F-Fluorination of Electron Rich Arenes: Present Limits of
No-Carrier-Added Reactions. Curr. Radiopharm., 2010, 3(3), 163-173.
[45] Roe, A. In: Organic Reactions; Roger Adams, Ed.; John Wiley & Sons Inc.: New York, 1949; Vol. 5, pp.
193-228.
[46] Swain, C.G.; Rogers, R.J. Mechanism of Formation of Aryl Fluorides from Arenediazonium Fluoborates. J.
Am. Chem. Soc., 1975, 97(4), 799-800.
[47] Koval’chuk, E.P.; Reshetnyak, O.V.; Kozlovs’ka, Z.Y.; Błażejowski, J.; Gladyshevs’kyj, R.Y.; Obushak,
M.D. Mechanism of the Benzenediazonium Tetrafluoroborate Thermolysis in the Solid State. Thermochim. Acta,
2006, 444(1), 1-5.
[48] Yoneda, N.; Fukuhara, T. Facile Preparation of Aromatic Fluorides by Deaminative Fluorination of
Aminoarenes using Hydrogen Fluoride Combined with Bases. Tetrahedron, 1996, 52(1), 23-36.
[49] Doyle, M.P.; Bryker, W.J. Alkyl Nitrite-Metal Halide Deamination Reactions. 6. Direct Synthesis of
Arenediazonium Tetrafluoroborate Salts from Aromatic Amines, tert-butyl Nitrite, and Boron Trifluoride
Etherate in Anhydrous Media. J. Org. Chem., 1979, 44(9), 1572-1574.
[50] Yong, H.K.; Chun, H.L.; Ki, Y.C. Facile Direct Conversion of Amino-Heterocycles to Fluoro-Heterocycles using t-butylthionitrite or t-butylthionitrate with Sodium Tetrafluoroborate. Tetrahedron Lett., 1990, 31(21),
3019-3022.
[51] Yoneda, N. The Combination of Hydrogen Fluoride with Organic Bases as Fluorination Agents.
Tetrahedron, 1991, 47(29), 5329-5365.
[52] Milner, D.J. Fluoroaromatics from Arylamines, a Convenient One-Pot Conversion Using Nitrosonium
Tetrafluoroborate. Synth. Commun., 1992, 22(1), 73-82.
[53] Shinhama, K.; Aki, S.; Furuta, T.; Minamikawa, J.-I. Facile Conversion of Arenediazonium Salts to the
Corresponding Fluoroarenes Using Boron Trifluoride Diethyl Ether Complex. Synth. Commun., 1993, 23(11),
1577-1582.
48
[54] Laali, K.K.; Gettwert, V.J. Fluorodediazoniation in ionic liquid solvents: new life for the Balz-Schiemann reaction. J. Fluorine Chem., 2001, 107(1), 31-34.
[55] Tamura, M.; Shibakami, M.; Sekiya, A. Deaminative Fluorination of Anilines with Silicon Tetrafluoride:
Utility of Silicon Tetrafluoride as a Fluorine Source. Eur. J. Org. Chem., 1998, 1998(4), 725-727.
[56] Garel, L.; Saint-Jalmes, L. One-pot fluoro-de-diazoniation of anilines in Organic Medium. Tetrahedron
Lett., 2006, 47(32), 5705-5708.
[57] Swain, C.G.; Sheats, J.E.; Harbison, K.G. Evidence for Phenyl Cation as an Intermediated in Reactions of
Benzenediazonium Salts in Solution. J. Am. Chem. Soc., 1975, 97(4), 783-790.
[58] Galli, C. Radical Reactions of Arenediazonium Ions: An Easy Entry into the Chemistry of the Aryl Radical.
Chem. Rev., 1988, 88(5), 765-792.
[59] Canning, P.S.J.; McCrudden, K.; Maskill, H.; Sexton, B. Rates and Mechanisms of the Thermal Solvolytic
Decomposition of Arenediazonium Ions. J. Chem. Soc., Perkin Trans. 2, 1999, 2735-2740.
[60] Satyamurthy, N.; Barrio, J.R.; Schmidt, D.G.; Kammerer, C.; Bida, G.T.; Phelps, M.E. Acid-catalyzed
Thermal Decomposition of 1-aryl-3,3-Dialkyltriazenes in the Presence of Nucleophiles. J. Org. Chem., 1990,
55(15), 4560-4564.
[61] Eakins, M.N.; Palmer, A.J.; Waters, S.L. Studies in the Rat with 18F-4-Fluoro-Oestradiol and 18F-4-
Fluoro-Oestrone as Potential Prostate Canning Agents: Comparison with 125I-2-iodo-oestradiol and 125I-2,4-
Di-iodo-oestradiol. Int. J. Appl. Radiat. Isot., 1979, 30(11), 695-700.
[62] Nozaki, T.; Tanaka, Y. The preparation of F18-labelled aryl fluorides. Int. J. Appl. Radiat. Isot., 1967,
18(2), 111-119.
[63] Hoyte, R.M.; Lin, S.S.; Christman, D.R.; Atkins, H.L.; Hauser, W.; Wolf, A.P. Organic
Radiopharmaceuticals Labeled with Short-Lived Nuclides. III. 18F-Labeled Phenylalanines. J. Nucl. Med., 1971,
12(6), 280-286.
[64] Goulding, R.W.; Palmer, A.J. The Preparation of Fluorine-18 labelled p-Fluorophenylalanine for Clinical
Use. Int. J. Appl. Radiat. Isot., 1972, 23(3), 133-137.
49
[65] Goulding, R.W.; Gunasekera, S.W. Fluorine-18 labelled L-p-Fluorophenylalanine. Int. J. Appl. Radiat. Isot.,
1975, 26(9), 561-564.
[66] Atkins, H.L.; Christman, D.R.; Fowler, J.S.; Hauser, W.; Hoyte, R.M.; Klopper, J.F.; Lin, S.S.; Wolf, A.P.
Organic Radiopharmaceuticals Labeled with Isotopes of Short Half-Life. V. 18F-Labeled 5- and 6-
Fluorotryptophan. J. Nucl. Med., 1972, 13(10), 713-719.
[67] Firnau, G.; Nahmias, C.; Garnett, S. Synthesis of 3,4-dihydroxy-5-fluoro-DL-Phenylalanine and 3,4-
Dihydroxy-5-[fluorine-18]fluoro-DL-phenylalanine. J. Med. Chem., 1973, 16(4), 416-418.
[68] Firnau, G.; Nahmias, C.; Garnett, S. The Preparation of [18F]5-fluoro-DOPA with Reactor-Produced
Fluorine-18. Int. J. Appl. Radiat. Isot., 1973, 24(3), 182-184.
[69] Argentini, M.; Wiese, C.; Weinreicht, R. Syntheses of 5-fluoro-d/l-dopa and [18F]5-fluoro-l-dopa. J.
Fluorine Chem., 1994, 68(2), 141-144.
[70] Kook, C.S.; Reed, M.F.; Digenis, G.A. Preparation of Fluorine-18-labeled Haloperidol. J. Med. Chem.,
1975, 18(5), 533-535.
[71] Livni, E.; Fischman, A.J.; Ray, S.; Sinclair, I.; Elmaleh, D.R.; Alpert, N.M.; Weiss, S.; Correia, J.A.; Webb,
D.; Dahl, R. Synthesis of 18F-labeled Fluconazole and Positron Emission Tomography Studies in Rabbits. Int. J.
Rad. Appl. Instrum. B, 1992, 19(2), 191-199.
[72] Hashizume, K.; Tamakawa, H.; Hashimoto, N.; Miyake, Y. Single-step Synthesis of [18F]Haloperidol from the Chloro-precursor and its Applications in PET Imaging of a Cat’s Brain. Appl. Radiat. Isot., 1997, 48(9),
1179-1185.
[73] Knöchel, A.; Zwernemann, O. Aromatic n.c.a. labelling with 18F- by modified Balz-Schiemann- decomposition. Int. J. Radiat. Appl. Instrum. A., 1991, 42(11), 1077-1080.
[74] Knöchel, A.; Zwernemann, O. Development of a No-Carrier-Added Method for 18F-Labelling of Aromatic
Compounds by Fluorodediazonation. J. Labelled Compd. Radiopharm., 1996, 38(4), 325-336.
[75] Bartsch, R.A.; Chen, H.; Haddock, N.F.; Juri, P.N. Stabilization of Aryldiazonium Ions by Crown Ether
Complexation. J. Am. Chem. Soc., 1976, 98(21), 6753-6754.
50
[76] Krane, J.; Skjetne, T. Studies of Crown Ether Complexes Aryldiazonium Ion Complexes. Tetrahedron Lett.,
1980, 21(18), 1775-1778.
[77] Nakazumi, H.; Kitao, T.; Zollinger, H. Dediazoniation of Arenediazonium Ions. 24. Dual and Triple
Substituent Parameter Evaluation of Competitive Heterolytic and Homolytic Dediazoniations of Diazonium Ions
Complexed with 18-Crown-6 Ether. J. Org. Chem., 1987, 52(13), 2825-2830.
[78] Gross, M.L.; Blank, D.H.; Welch, W.M. The Triazene Moiety as a Protecting Group for Aromatic Amines.
J. Org. Chem., 1993, 58(8), 2104-2109.
[79] Nifontov, V.I.; Bel’skaya, N.P.; Shtokareva, E.A. Manufacture Methods of Synthesizing Triazenes
(Review). Pharm. Chem. J., 1993, 27(9), 652-665.
[80] Nifontov, V.I.; Bel’skaya, N.P.; Shtokareva, E.A. The Reactivity and Mechanism of Action of Triazenes
(Review). Pharm. Chem. J., 1994, 28(10), 687-706.
[81] Kimball, D.B.; Haley, M.M. Triazenes: A Versatile Tool in Organic Synthesis. Angew. Chem. Int. Ed.,
2002, 41(18), 3338-3351.
[82] Chu, C.-K.; Kim, J.-H.; Kim, D.W.; Chung, K.-H.; Katzenellenbogen, J.A.; Chi, D.Y. Aromatic
Fluorination by Decomposition of Triazenes in Ionic Liquids. Bull. Korean Chem. Soc., 2005, 26(4), 599-602.
[83] Threadgill, M.D.; Stevens, M.F.G. Selective Reactions in the Triazene Series; I. Reduction of 1-(4-
Acetylphenyl)-triazenes. Synthesis, 1983, 1983(4), 289-291.
[84] Pages, T.; Langlois, B.R. Fluorination of Aromatic Compounds from 1-aryl-3,3-Dimethyltriazenes and
Fluoride Anions in Acidic Medium: 1. A model for 18F labelling. J. Fluorine Chem., 2001, 107(2), 321-327.
[85] Pages, T.; Langlois, B.R.; Le Bars, D.; Landais, P. Fluorination of aromatic compounds from 1-aryl-3,3- dimethyltriazenes and fluoride anions in acidic medium: 2. Synthesis of (S)-[18F]-3-fluoro-α- methylphenylalanine. J. Fluorine Chem., 2001, 107(2), 329-335.
[86] Yoneda, N.; Fukuhara, T.; Mizokami, T.; Suzuki, A. A Facile Preparation of Aryl Triflates. Decomposition of Arenediazonium Tetrafluoroborate Salts in Trifluoromethanesulfonic Acid. Chem. Lett., 1991, 20(3), 459-
460.
51
[87] Picherit, C.; Wagner, F.; Uguen, D. The Sequel to a Carbocyclic Nucleoside Synthesis: A Divergent Access to Both Arenediazonium Ions and Aryl Triflates. Tetrahedron Lett., 2004, 45(12), 2579-2583.
[88] Berridge, M.S.; Sayre, L.M.; Arora, P.K.; Terris, A.H.; Riachi, N.J.; Harik, S.I. Design and Synthesis of
18F-Labeled Neurotoxic Analogs of MPTP. J. Med. Chem., 1993, 36(9), 1284-1290.
[89] Kilbourn, M.R.; Welch, M.J.; Dence, C.S.; Tewson, T.J.; Saji, H.; Maeda, M. Carrier-Added and No-
Carrier-Added Syntheses of [18F]Spiroperidol and [18F]Haloperidol. Int. J. Appl. Radiat. Isot., 1984, 35(7),
591-598.
[90] Riss, P.J.; Kuschel, S.; Aigbirhio, F.I. No Carrier-Added Nucleophilic Aromatic Radiofluorination using
Solid Phase Supported Arenediazonium Sulfonates and 1-(aryldiazenyl)Piperazines. Tetrahedron Lett., 2012,
53(14), 1717-1719.
[91] Stang, P.J.; Zhdankin, V.V. Organic Polyvalent Iodine Compounds. Chem. Rev., 1996, 96(3), 1123-1178.
[92] Zhdankin, V.V.; Stang, P.J. Chemistry of Polyvalent Iodine. Chem. Rev., 2008, 108(12), 5299-5358.
[93] Ugi, I.; Marquarding, D.; Klusacek, H.; Gillespie, P.; Ramirez, F. Berry pseudorotation and turnstile rotation. Acc. Chem. Res., 1971, 4(8), 288-296.
[94] Cass, M.E.; Hii, K.K.; Rzepa, H.S. Mechanisms That Interchange Axial and Equatorial Atoms in Fluxional
Processes: Illustration of the Berry Pseudorotation, the Turnstile, and the Lever Mechanisms via Animation of
Transition State Normal Vibrational Modes. J. Chem. Educ., 2006, 83(2), 336.
[95] Couzijn, E.P.A.; Slootweg, J.C.; Ehlers, A.W.; Lammertsma, K. Stereomutation of Pentavalent Compounds:
Validating the Berry Pseudorotation, Redressing Ugi’s Turnstile Rotation, and Revealing the Two- and Three-
Arm Turnstiles. J. Am. Chem. Soc., 2010, 132(51), 18127-18140.
[96] Zhdankin, V.V.; Stang, P.J. Recent Developments in the Chemistry of Polyvalent Iodine Compounds.
Chem. Rev., 2002, 102(7), 2523-2584.
[97] Wirth, T. Hypervalent Iodine Chemistry in Synthesis: Scope and New Directions. Angew. Chem. Int. Ed.,
2005, 44(24), 3656-3665.
[98] Zhdankin, V.V. Hypervalent iodine(III) reagents in organic synthesis. ARKIVOC, 2009, 2009, 1-62.
52
[99] Zhdankin, V.V. Organoiodine(V) Reagents in Organic Synthesis. J. Org. Chem., 2011, 76(5), 1185-1197.
[100] Perkins, C.W.; Martin, J.C.; Arduengo, A.J.; Lau, W.; Alegria, A.; Kochi, J.K. An Electrically Neutral
.Sigma.-Sulfuranyl Radical from the Homolysis of a Perester with neighboring Sulfenyl Sulfur: 9-S-3 species. J.
Am. Chem. Soc., 1980, 102(26), 7753-7759.
[101] Musher, J.I. The Chemistry of Hypervalent Molecules. Angew. Chem. Int. Ed., 1969, 8(1), 54-68.
[102] Alcock, N.W.; Countryman, R.M. Secondary bonding. Part 1. Crystal and molecular structures of (C6H5)2
IX (X = Cl, Br, or I). J. Chem. Soc., Dalton Trans., 1977, 217-219.
[103] Ochiai, M. In: Topics in Current Chemistry; Thomas Wirth, Ed.; Springer-Verlag: Berlin Heidelberg New
York, 2003; Vol. 224, pp. 5-68.
[104] Okuyama, T.; Takino, T.; Sueda, T.; Ochiai, M. Solvolysis of Cyclohexenyliodonium Salt, a New
Precursor for the Vinyl Cation: Remarkable Nucleofugality of the Phenyliodonio Group and Evidence for
Internal Return from an Intimate Ion-Molecule Pair. J. Am. Chem. Soc., 1995, 117(12), 3360-3367.
[105] Pike, V.W.; Aigbirhio, F.I. Reactions of Cyclotron-produced [18F]Fluoride with Diaryliodonium Salts-A
Novel Single-Step Route to No-Carrier-Added [18]Fluoroarenes. J. Chem. Soc., Chem. Commun., 1995, 2215-
2216.
[106] Ross, T.L.; Ermert, J.; Hocke, C.; Coenen, H.H. Nucleophilic 18F-Fluorination of Heteroaromatic
Iodonium Salts with No-Carrier-Added [18F]Fluoride. J. Am. Chem. Soc., 2007, 129(25), 8018-8025.
[107] Shah, A.; Pike, V.W.; Widdowson, D.A. The Synthesis of [18F]Fluoroarenes from the Reaction of
Cyclotron-Produced [18F]Fluoride Ion with Diaryliodonium Salts. J. Chem. Soc., Perkin Trans. 1, 1998, 2043-
2046.
[108] Zhang, M.-R.; Kumata, K.; Suzuki, K. A Practical Route for Synthesizing a PET ligand Containing
[18F]Fluorobenzene Using Reaction of Diphenyliodonium Salt with [18F]F-. Tetrahedron Lett., 2007, 48(49),
8632-8635.
53
[109] Chun, J.-H.; Lu, S.; Pike, V.W. Rapid and Efficient Radiosyntheses of meta-Substituted [18F]Fluoroarenes from [18F]Fluoride Ion and Diaryliodonium Tosylates within a Microreactor. Eur. J. Org. Chem., 2011,
2011(23), 4439-4447.
[110] Ross, T.L.; Ermert, J.; Coenen, H.H. N.C.A. 18F-Fluorination of Various Arenes via Aryl(2- thieny)iodonium Salts. J. Label. Compd. Radiopharm., 2005, 48(S1), S153.
[111] Yan, R;, Brichard, L.; Soloviev, D.; Aigbirhio, F.I.; Carroll, M.A. The First Single-Step-Single-Post
Synthesis of 4-[18F]SFB. J. Label. Compd. Radiopharm., 2009, 52(6), 216.
[112] Yan, R;, Brichard, L.; Soloviev, D.; Aigbirhio, F.I.; Carroll, M.A. The First Synthesis of Ethyl 3-
[18F]fluorobenzoate Using [18F]fluoride. J. Label. Compd. Radiopharm., 2009, 52(6), 217.
[113] Carroll, M.A.; Jones, C.; Tang, S.-L. Fluoridation of 2-Thienyliodonium Salts. J. Label. Compd.
Radiopharm., 2007, 50(5-6), 450-451.
[114] Le Count, D.J.; Reid, J.A.W. Iodonium Salt Synthesis. Evidence for the Formation of Isomers in the
Synthesis of Diaryliodonium Salts. J. Chem. Soc. C, 1967, 1298-1301.
[115] Yamada, Y.; Okawara, M. Steric Effect in the Nucleophilic Attack of Bromide Anion on Diaryl- and Aryl-
2-thienyliodonium Ions. Bull. Chem. Soc. Jpn., 1972, 45(6), 1860-1863.
[116] Lancer, K.M.; Wiegand, G.H. The Ortho Effect in the Pyrolysis of Iodonium Halides. A Case for a
Sterically Controlled Nucleophilic Aromatic (SN) Substitution Reaction. J. Org. Chem., 1976, 41(21), 3360-
3364.
[117] Grushin, V.V. Carboranylhalonium ions: from Striking Reactivity to a Unified Mechanistic Analysis of
Polar Reactions of Diarylhalonium Compounds. Acc. Chem. Res., 1992, 25(11), 529-536.
[118] Chun, J.-H.; Lu, S.; Lee, Y.-S.; Pike, V.W. Fast and High-Yield Microreactor Syntheses of ortho-
Substituted [18F]Fluoroarenes from Reactions of [18F]Fluoride Ion with Diaryliodonium Salts. J. Org. Chem.,
2010, 75(10), 3332-3338.
54
[119] Gail R., Hocke C., Coenen H.H. Direct N.C.A. 18F-fluorination of Halo- and Alkylarenes via
Corresponding Diphenyliodonium Salts. J. Label. Compd. Radiopharm., 1997, 40(1), 50-52.
[120] Ross, T.L. Direct no-carrier-added 18F-labelling of arenes via nucleophilic substitution on aryl(2- thienyl)iodonium salts. PhD Thesis, The German National University: Köln, December 2005.
[121] Martín-Santamaría, S.; Carroll, M.A.; Carroll, C.M.; Carter, C.D.; Pike, V.W.; Rzepa, H.S.; Widdowson,
D.A. Fluoridation of Heteroaromatic Iodonium Salts-Experimental Evidence Supporting Theoretical Prediction of the Selectivity of the Process. Chem. Commun., 2000, 649-650.
[122] Seeman, J.I. Effect of Conformational Change on Reactivity in Organic Chemistry. Evaluations,
Applications, and Extensions of Curtin-Hammett Winstein-Holness Kinetics. Chem. Rev., 1983, 83(2), 83-134.
[123] Wadsworth, H.J.; Devenish, T. Fluoridation Method. International Patent WO2005097713, October 20,
2005.
[124] Lee, B.C.; Dence, C.S.; Zhou, H.; Parent, E.E.; Welch, M.J.; Katzenellenbogen, J.A. Fluorine-18 Labeling and Biodistribution Studies on Peroxisome Proliferator-Activated Receptor-Γ Ligands: Potential Positron
Emission Tomography Imaging Agents. Nucl. Med. Biol., 2009, 36(2), 147-153.
[125] Lee, B.C.; Kim, J.S.; Kim, B.S.; Son, J.Y.; Hong, S.K.; Park, H.S.; Moon, B.S.; Jung, J.H.; Jeong, J.M.;
Kim, S.E. Aromatic Radiofluorination and Biological Evaluation of 2-aryl-6-[18F]Fluorobenzothiazoles as a
Potential Positron Emission Tomography Imaging Probe for β-amyloid Plaques. Bioorg. Med. Chem., 2011,
19(9), 2980-2990.
[126] Basuli, F.; Wu, H.; Griffiths, G.L. Syntheses of Meta-[18F]Fluorobenzaldehyde and Meta-
[18F]Fluorobenzylbromide from Phenyl(3-Formylphenyl) Iodonium Salt Precursors. J. Labelled Compd.
Radiopharm., 2011, 54(4), 224-228.
[127] Carroll, M.A.; Nairne, J.; Woodcraft, J.L. Diaryliodonium salts: a solution to 3-[18F]fluoropyridine. J.
Label. Compd. Radiopharm., 2007, 50(5-6), 452-454.
[128] Wüst, F.R.; Kniess, T. Synthesis of 4-[18F]Fluoroiodobenzene and its Application in Sonogashira Cross-
Coupling Reactions. J. Labelled Compd. Radiopharm., 2003, 46(8), 699-713.
55
[129] Wadsworth, H.J.; Widdowson, D.A.; Wilson, E.; Carroll, M.A. Radical Trap in Fluoridation of Iodonium
Salt. International Patent WO2005061415, July 07, 2005.
[130] Carroll, M.A.; Nairne, J.; Smith, G.; Widdowson, D.A. Radical Scavengers: A Practical Solution to the
Reproducibility Issue in the Fluoridation of Diaryliodonium Salts. J. Fluorine Chem., 2007, 128(2), 127-132.
[131] Moon, B.S.; Kil, H.S.; Park, J.H.; Kim, J.S.; Park, J.; Chi, D.Y.; Lee, B.C.; Kim, S.E. Facile Aromatic
Radiofluorination of [18F]Flumazenil from Diaryliodonium Salts with Evaluation of their Stability and
Selectivity. Org. Biomol. Chem., 2011, 9(24), 8346-8355.
[132] Wüst, F.; Carlson, K.E.; Katzenellenbogen, J.A. Synthesis of novel Arylpyrazolo Corticosteroids as
Potential Ligands for Imaging Brain Glucocorticoid Receptors. Steroids, 2003, 68(2), 177-191.
[133] Lee, Y.-S.; Hodošček, M.; Chun, J.-H.; Pike, V.W. Conformational Structure and Energetics of 2-
Methylphenyl(2′-methoxyphenyl)iodonium Chloride: Evidence for Solution Clusters. Chem. Eur. J., 2010,
16(34), 10418-10423.
[134] Huiban, M.; Gee, A.D.; Martarello, L.; Passchier, J. Labelling Methods and Radiosynthesis–Organic
Positron Emitters. J. Label. Compd. Radiopharm. 2007, 50(S1), S130.
[135] Lee, B.C.; Lee, K.C.; Lee, H.; Mach, R.H.; Katzenellenbogen, J.A. Strategies for the Labeling of Halogen-
Substituted Peroxisome Proliferator-Activated Receptor γ Ligands: Potential Positron Emission Tomography and Single Photon Emission Computed Tomography Imaging Agents. Bioconjugate Chem., 2007, 18(2), 514-
523.
56
1.4. Alzheimer's disease
1.4.1. Epidemiology and risk factors of Alzheimer's disease
Deterioration of intellectual ability and cognitive decline was and is considered an inevitable and natural consequence of aging that was recognized as early as the 7th century BC.61,62 Alzheimer’s disease (AD) has become one of the leading health concerns as aging of the population has now become a worldwide phenomenom and is no longer limited to the developed Western societies. The number of people surviving into their 80s and 90s and beyond is expected to grow dramatically due to advances in medical treatment and medical technology, as well as socio-economic conditions, further contributing to the increase in prevalence of AD.63,64 The increasing number of people with AD have a marked impact on health care systems, not to mention families and caregivers.
Alzheimer’s disease (AD) is a neurodegenerative and terminal disease characterised with insidious onset, gradual memory loss and progressive decline in other cognitive functions over time. The term AD is today use for mental dementia (without antecedent causes as stroke, head trauma, alcohol, …) irrespective of the age of onset. It is the most common type of dementia in the elderly and affects approximately half of all patients with dementia worldwide. An estimated 5.4 million Americans of all ages have AD in 2012 and one in eight people age 65 and older (13 percent) has AD in the United States.63 As demonstrated by the number of incidence and prevalence studies, advancing age is the greatest and most consistent risk factor for the disease, where incidence and prevalence for AD show approximately exponential increase with increasing age.63,64 Most of the prevalence and incidence studies are consistent with the hypothesis that genetic factors predispose a person to develop AD, while other factors modulate the age of onset of clinical dementia.
Once AD has been diagnosed, the average life expectancy is approximately 4 to 8 years, while very small percent of the patients live as long as 15 years. As the disease progresses, the individual’s cognitive and functional abilities decline. In advanced AD, people need help with basic activities of daily living and those in the final stages of the disease become bed-bound and reliant on caregivers.
Analytic epidemiologic studies are important for identification of risk factors, which can be in general divided into non-modifiable and modifiable ones. Modifiable factors are obvious
57
targets for preventive treatment, while non-modifiable ones (gender, level of education, genetic factors and family history) are important in understanding the pathogenesis of the disease. Although in numerous studies women were found to be at greater risk for AD than men, they are not more likely than men to develop dementia at any given age. The larger proportion of women with AD or other dementias is primarily explained by the fact that women live longer on average than men. Data from a number of studies show that a low education level is associated with a greater risk of developing and greater likelihood of having dementia and AD.63,64 These findings led to the hypothesis that lifetime cognitive experience and engagement in leisure activities of intellectual and social nature may influence the number of neurons and synapses that survive into adult life (cognitive reserve hypothesis) - the greater the number of synapses formed between neurons, the longer the time needed for the synapses to degenerate, and later the onset of dementia.65 However, other believe that the increased risk of dementia among those with lower educational level may be explained by other factors common to people in lower socioeconomic groups, such as increased risk for disease in general and less access to medical care.63 Many studies have also shown that family history is another risk factor for AD, especially for occurrence in a first-degree relative (parent or sibling). Currently four genes have been found to be definitely associated with AD. Mutations in the amyloid precursor protein (APP), presenilin 1 (PS1), and presenilin 2 (PS2) are known to cause early onset of AD, with PS1 mutations being the most frequent.66 Molecular studies of these three mutations are all associated with an increased production of the Aβ42 amyloid peptide, and provide support for the amyloid cascade hypothesis of AD pathogenesis. By contrast, individuals who inherit one or two genes of the apolipoprotein-ε4 gene (ApoE-ε4 gene) are at increased risk to develop AD and to develop it at an earlier age than those who inherit the ε2 or/and ε3 forms of the ApoE gene.67 Nevertheless, ApoE-ε4 is a susceptibility risk factor, as it is neither necessary nor sufficient that an individual will develop AD.63,64
1.4.2. Neurophysiology and pathology of Alzheimer’s disease
AD pathology can be characterized on a macro level as the progressive loss of brain tissue due to localised degeneration of neurons and synapses, which roughly correlates with the severity of cognitive decline. Diffused cerebral atrophy is manifested by narrowing of the cerebral gyri, widening of sulci, thinning of the cortical ribbon, and by enlargement of the 58
volume of the ventricles, especially the temporal horn. As the disease progresses, neurons die in a particular pattern over time and cerebrospinal fluid fills in the space previously occupied by brain tissue.68 Preclinical AD begins in the transentorhinal region followed by the entorhinal cortex, which connects the hippocampus to the cerebral cortex. Several studies suggest that progressive neuronal loss in medial temporal lobe may start years before signs of dementia emerge.69,70 Thus, in the early stages of AD, short-term memory begins to fade and disorientation appears when the cells in the hippocampus, transentorhinal region, and entorhinal cortex degenerate. The ability to perform routine tasks also declines. As the disease progresses, atrophy extends to other areas of neocortex, leading to judgment decline, and changes in behavior (emotional outburst, wandering, agitation). At the later stages of the disease, the neocortex atrophies in areas that control speech, reasoning, sensory processing, and conscious thought.71 Therefore, people lose the ability to recognize faces, communicate, control bodily functions and require constant care.
Figure 3. Characteristic volumetric brain changes during progression of Alzheimer's disease: diffused cerebral atrophy with narrowing of gyri, widening of sulci, ventricular dilatation and atrophy of the hippocampus.68
At microscopic level, AD is characterised by the development of two neuropathological hallmarks, namely extracellular amyloid plaques or senile plaques (SPs), and intracellular neurofibrilary tangles (NFTs) (Figure 4).72 These neuropathologic lesions likely begin to form years prior to the full clinical expression of clinical dementia, particularly within stages of mild cognitive impairment.73 Although both SPs and NFTs are generally considered to be characteristic pathologic changes of AD, they are not specific. However, controversy remains in the relationship between these lesions and which process is central to disease pathogenesis. Over the past few decades, support has grown for the amyloid cascade hypothesis of AD, in
59
which accumulation of neurotoxic amyloid beta (Aβ) peptide in brain tissue is believed to be an early and necessary step. This triggers a series of events including inflammatory response, free radical formation, oxidative stress, lipid peroxidation, excessive excitotoxicity of glutaminergic neurons, and formation of neurofibrillary tangles that lead to neurodegeneration, neurotransmitter dysfunction and dementia (Figure 5 on page 64).74,75 While there is substantial evidence supporting the amyloid cascade hypothesis, there are also limitations because SPs and NFTs may develop independently, and they may be a protective response against inflammatory cascade rather than the cause of neurodegeneration in AD. Furthermore, plaque pathology and spread do not always correlate with clinical findings in AD and can exist in normal individuals.76,77 Thus, the amount of defficient nerve cells and synaptic loss in the hippocampus and neocortex much better correlates with the decline of cognitive function in AD than the number of amyloid plaques.78
Axon terminal Phosphorylated tau in tangles Neuron
Astrocyte
Capillary
A42 in senile plaque
Figure 4. Schematic drawing of a neuron with an adjacent astrocyte and capillary. Intracellular hyperphosphorylated tau in tangles and extracellular insoluble aggregates of Aβ42 peptide within senile plaques are the characteristic hallmarks of AD.79
The core of SPs is consisted mainly of focal insoluble aggregates of amyloid beta peptide 42 (Aβ42) and is surrounded by dystrophic neurites, activated microglia, reactive astrocytes and immune system proteins, to name a few. Aβ42 is a neurotoxic peptide and is the result of alternate proteolytic cleavage of transmembrane amyloid precursor protein (APP) due to increased activity of β- and γ-proteases towards α-proteases (Figure 5 on page 64).80 It is believed that this altered processing of APP is the primary causative factor in the pathogenesis of AD (amyloid cascade hypothesis). Aβ42 initially forms monofibrils, but then quickly 60
aggregates into protofibrils and finally into insoluble extracellular Aβ pleated sheet structures.81 β-Amyloid peptides (39-43 aminoacid residues) activate immune system (complement, T-lymhocytes, microglia) and stimulate the release of chemokines and cytokines. Inflammation surrounding Aβ plaques is therefore consistent feature of the AD in brain and add to the pathologic cascade of the disease.82 Affected neurons are susceptible to ischaemia, excitotoxicity and oxidative stress too, which quickens their apoptotic ruin. Aged individual without AD have some density of SPs, though most often these are diffuse, not compact as in AD, and are not disease-specific. The anatomic distribution of SP in AD is widespread and higher order association cortex tends to have the highest density of SPs, and primary motor and sensory cortex the least. It has been shown that soluble Aβ oligomer intermediates are key contributors to Aβ mediated neurodegeneration. Thus, soluble Aβ aggregates appeared to correlate much better with neuron loss and severity of AD than insoluble high molecular weight Aβ fibrils within amyloid plaques.83–85
Neurofibrillary tangles (NFTs) consist of cross-linked paired helical filaments of pathologically hyperphoshorylated form of a microtubule-associated protein, namely Tau protein. This probably results from an imbalance in the activity and regulation of tau kinases (like glycogensynthase kinase-3 and cyclin dependent kinase-5) and phosphatases (PP1, PP2A, PP2B).86,87 Abnormally high levels of neurofibrillary tangles inside the neuron destabilizes microtubules and disrupts axonal transport, which is believed to be directly associated with neuronal death and disease progression.88 Braak H. and Braak E. have shown that neurofibrillary pathology in the brain accumulates in a hierarchical topographic fashion, with the transentorhinal cortex affected first (stages I and II), followed by the entorhinal cortex, hippocampus and other limbic structures of the medial temporal lobe (stages III and IV), and finally the neocortex (stages V and VI).89,90 The neocortex is involved in a hierarchical fashion, the associative regions affected the most, whereas the primary sensory and motor cortex are relatively spared. This corresponds to the clinical features of marked impairment of memory and abstract reasoning, with preservation of vision and movement. NFTs are also frequently present in certain subcortical nuclei such as the nucleus basalis, limbic nuclei of the thalamus, amygdala, locus coeruleus, substantia nigra, and raphe nuclei of the brainstem. The distribution, progression, and abundance of tangles are better proportional to the severity of cognitive impairment and severity of AD than senile plaque pathology (tau hypothesis). Most patients with AD are diagnosed at stage V or VI pathology, while those at
61
lower stages are usually not demented. Interestingly, amyloid plaques and neurofibrillary tangles are also observed in intellectually normal individuals too, but they are far more abundant in patients with AD. Thus, although NFTs and SPs are considered to be the characteristic pathologic changes of AD, they are not specific.
As stated above, AD is characterised by marked atrophy of the cerebral cortex, hippocampus and subcortical brain regions. Subcortical brain regions include the basal forebrain, the locus coeruleus, and the dorsal raphe nuclei, thus generating deficits of acetylcholine, norepinephrine, and serotonin which contribute to the impairment of attention, memory, mood, and behaviour. Numerous measurements on post-mortem AD brain tissues have revealed a relatively selective degeneration of subcortical cholinergic neurons; particularly nucleus basalis of Meynart, which is part of the magnocellular forebrain nuclei in the basal forebrain that provide cholinergic innervation to the whole cortex. Choline acetyltransferase (ChAT; acetylcholine synthesising enzyme) and acetylcholine esterase (AChE; an enzyme within synaptic cleft that hydrolyses acetylcholine to choline and acetic acid, thus preparing the synapse for the passage of a new impulse) activity in the neocortex and hippocampus are reduced considerably (30 % - 70 %) in AD, which correlates positively with the severity of dementia but not in other disorders such as a depression or schizophrenia. Muscarinic receptor (mAChR; G-protein coupled) density is “notˮ affected, but pentameric ionotropic nicotinic receptors (nAChR; mainly located presynaptically), particularly in the cortex and hippocampus, are reduced. This discovery (1979) lead to the cholinergic hypothesis which is the mainstay of the current symptomatic therapy approach. It states that deficiency of acetylcholine (ACh) in affected areas is critical in the genesis of the symptoms. Moreover, cholinergic neuronal loss correlates well with the severity of the disease. Therefore, enhancement of cholinergic transmission in affected areas might compensate for the cholinergic deficit.
1.4.3. Pharmaceutical management and research directions
For AD there is no known cure. Current drug treatments are just palliative in nature (symptomatic treatment) and do not halt or even reverse the progression of the disease; hence, they are far from ideal. The therapies under experimental and evaluation phases for the treatment of AD have disease modifying and neuroprotective approaches. There are currently
62
two licensed symptomatic treatments for AD: (A.) acetylcholinesterase inhibitors (AChEIs), and (B.) memantine, the uncompetitive N-methyl-D-aspartate (NMDA) receptor antagonist.
AChEIs augment acetylcholine (ACh) levels and its intrasynaptic residence time via inhibition of acetylcholinesterase in the synaptic cleft, and therefore facilitates interaction between ACh and the postsynaptic cholinergic receptors. Donepezil, rivastigmine, and galantamine are approved in most countries worldwide for mild to moderately severe AD, because clinical trials have demonstrated their benefits in cognitive functioning, activities of daily living and behaviour.91 However, results in the long term have generally been disappointing. Only patients with stable medical or psychiatric illnesses who adhere, tolerate and respond to AChEI might experience modest cognitive improvements and small symptomatic benefit.92
Memantine has been licensed for the treatment of moderate to severe AD. It is an uncompetitive, moderate-affinity, and voltage dependent NMDA antagonist that selectively blocks glutamate overactivation of these receptors. Interestingly, it exits the channel pore to allow normal physiologic neurotransmission under conditions of learning and memory formation. Thus, it protects neuron against glutamate-mediated excitotoxicity, because an increase in extracellular glutamate can lead to overactivation of NMDA receptors, which results in excessive Ca2+ influx through the receptor associated ion channel.91,92 Intracellular calcium accumulation can induce a cascade of evens resulting in neuronal death by necrosis or apoptosis.93
Enormous research attention is currently being directed at the strategies designed to modify AD pathogenesis. These approaches are based on the recent advances in understanding AD pathogenesis and are directed at slowing or halting the progression of the disease. Such disease-modifying treatments, which target physiopathological biochemical pathways in AD, include β-amyloid-lowering approaches, tau-based treatments (e.g. Tau phosphorylation inhibitors), as well as neuroprotective and neurorestorative approaches. Anti-amyloid therapies target reduction of amyloid production through inhibition or modulation of β- and γ- secretases or enhance α-secretase activity. Other compounds aim to prevent the oligomerization and fibrillization of Aβ (β-amyloid antiaggregants, and β-sheet breakers), whereas other strategies include immunotherapies (β-amyloid vaccines and passive immunization with anti-β-amyloid humanized monoclonal antibodies) endeavor to remove
63
neurotoxic Aβ from the brain. According to the one of the most current versions of the amyloid cascade hypothesis, amyloid-directed therapeutic strategies will only be effective early in the disease, whereas latter in the disease tau-directed therapies may be a better strategy of choice. Peroxisome proliferator-activated receptor γ agonists (PPAR-γ agonists), metal chelators, and muscarinic M1 agonists also appear to diminish amyloid production. Neuroprotective and neuroregenerative approaches include neurotrophic factors (e.g. nonpeptidic neurotrophic factor enhancers), antioxidants (e.g. Coenzyme Q10, vitamin C), anti-apoptotic agents (e.g. caspase inhibitors), astrocyte-modulating agents, NMDA-receptor antagonists, and anti-inflammatory drugs, to name a few.94–96
Figure 5. Diagram of the cascade of events currently hypothesized to compromise the physiopathology of AD. Sites for potential therapeutic interventions are designated with red colour.
64
The goal of disease-modifying therapy is to preserve neuronal integrity and synaptic plasticity by reducing amyloid toxicity and neuronal vulnerability. The focus of disease- modifying treatments is towards early and incipient AD with a view of preventing or attenuating disease before symptoms occur. According to the current knowledge of the disease physiopathology, interventions at the stage of mild to moderate AD may already be too late. Furthermore, slowing progression in more severe stages of the disease might not be desirable for patients, family members, or society as a whole. Though none are yet proven, it might be foreseen that such treatments are on the edge of potential breakthrough. If successful, they will be used in conjunction with existing symptomatic therapies.
1.4.4. The diagnosis of Alzheimer’s disease by PET
Though AD is the most common type of dementia, it still must be distinguished from other causes of dementia, including vascular dementia (VaD), frontotemporal dementia (FTD), corticobasal degeneration (CBD), dementia with Lewy body (DLB), and Parkinson's disease dementia (PDD) among others. The detection of AD early in its clinical course can be quite challenging, while identification later in its course is often more obvious.
Today, the diagnosis for possible and probable AD is based on clinical examination. A number of common clinical features are specified within sets AD diagnostic criteria: memory decline and impairment of at least one non-memory cognitive function (language, motor and executive function, visuo-spatial skills) that is sufficiently severe to interfere with daily function.63,79 The disease is confirmed by neuropsychological testing (memory testing and assessment of intellectual functioning) and neurological examinations, which is crucial in differential diagnosis of the disease. But for definitive diagnosis, histopathologic confirmation is needed (microscopic examination of brain tissue pathologic changes), which is up to now possible just by post-mortem examination (autopsy). Since there is no absolute qualitative difference that distinguishes the brains of demented patients from those of non-demented elderly individuals, the definitive diagnosis of AD depends on identifying quantitative differences. Post-mortem neuropathologic examination of demented individuals frequently demonstrates AD in combination with other pathologic changes most frequently with cerebrovascular dementia (cerebral amyloid angiopathy), Lewy body disease or both (“Mixed dementiaˮ). 65
In preclinical AD individuals have measurable changes in the brain (injured or degenerated nerve cells), cerebrospinal fluid (CSF) and/or blood biomarkers that indicate the earliest signs of disease, but they not yet have developed symptoms such as memory lost.63 Thus, there is a growing interest in measuring CSF biomarkers as their levels reflect the severity of symptoms or disease progression in AD.97,98 The most promising biomarkers in CSF (β-amyloid and tau) have high sensitivity for differentiating early AD, but lower specificity against other dementias, which ranges between 60 and 90%. Combining a battery of CSF biomarkers may improve diagnostic specificity.
Advances in a variety of structural and functional neuroimaging techniques allow the greatest in vivo and non-invasive insight into brain structure and function, respectively. To date, the best established methods for the detection and tracking AD include structural magnetic resonance imaging (sMRI) measurements of regional and whole brain tissue shrinkage, [18F]FDG PET measurements of decline in the regional cerebral metabolic rate for glucose, and PET measurements of fibrillar β-amyloid burden.99 For example, PET radiotracers Pittsburgh compound-B (PIB)100,101, [18F]FDDNP102–104 and AmyvidTM (vials containing 500-1900 MBq/mL Florbetapir F18)105 which label SPs have been used in humans and have demonstrated the characteristic retention in AD subjects that mirrors the pattern of this hallmark known from post-mortem studies. Florbetapir F18 is the first and only FDA- approved diagnostic PET tracer for estimation of β-amyloid neuritic plaque density in adult patients with cognitive impairment who are being evaluated for AD and other causes of cognitive decline. It is an adjunct to other diagnostic evaluations, because positive scan which indicates moderate to frequent amyloid neuritic plaques does not establish a diagnosis of AD or other cognitive disorder.
66
NC CN OH
HO O HO OH 18F 18F N [18F]FDG 18 [ F]FDDNP Figure 6. PET images comparing temporal lobe uptake of [18F]FDDNP, an β-amyloid binding tracer, and [18F]FDG, a marker of glucose metabolism, in a patient with AD (left) and a control subject (right).106 [18F]FDDNP and [18F]FDG uptake are presented on a heat scale. Note increased uptake and retention of [18F]FDDNP (arrowheads) in temporal lobes of the patient with AD, compared with those of the control subject. AD patient also demonstrates typical decrease of [18F]FDG or glucose metabolism in temporal (arrows) and parietal lobes (not shown).
Thus, the potential of PET technique lies in its ability of in vivo non-invasive quantification of AD pathology in preclinical phase. The combination of neuroimaging techniques with blood or CSF biomarkers may in the future play a major part in establishing the diagnosis on their own. However, although the new criteria and guidelines identify preclinical disease as a stage of AD, they do not establish diagnostic criteria that doctors could use now. Rather, they state that additional biomarker research and development in specific radiotracers are needed before preclinical AD can be diagnosed.
67
1.5. Vesicular acetylcholine transporter (VAChT) and the most promising PET imaging tracers
Evidence from biopsy and autopsy samples suggests that degeneration of the cholinergic neurons located in the basal forebrain nuclei (e.g. in the nucleus basalis of Meynert), the major cholinergic output of the central nervous system (CNS), and their synapses in the cerebral cortex and hyppocampus are among the earliest neurochemical changes identified in AD, and precedes for several years clinical onset of the disease.107–110 Consequently, different cholinergic synaptic elements are depleted in the cortex and subcortical brain areas, such as
α4β2 nicotinic ACh receptor, acetylcholine esterase (AChE), choline acetyltransferase (ChAT), choline transporter (ChT) and vesicular acetylcholine transporter (VAChT).
VAChT is a presynaptic vesicular transmembrane glycoprotein and is responsible for loading ACh into secretory vesicles making ACh available for secretion into synaptic cleft (Figure 7).111,112 VAChT contains 12 transmembrane domains, and the gene coding for the VAChT is embeded within the first intron of the ChAT's gene in all species examined, suggesting that the expression of ChAT, regarded as reference standard for cholinergic markers, is tightly coupled to that of VAChT.112–114 VAChT gained increasing interest in the last years as a reliable cholinergic marker for in vivo imaging of cholinergic deficiencies using PET or SPECT, because its reduced density and activity change in parallel fashion with ChAT,115,116 and show a strong correlation with the onset, progression, and severity of the AD.117 Thus, in vivo qualitative and quantitative non-invasive measurements of the mentioned target with PET can provide valuable information on disease onset and progression with the aim to obtain an early diagnosis and to better understand AD.
68
Figure 7. Schematic illustration of a generalized cholinergic junction (ChT, choline transporter; AcCoA, acetyl coenzyme A; ChAT, choline acetyltransferase; VAChT, vesicular acetylcholine transporter; P, peptides; ATP, adenosine triphosphate; SNAPs, synaptosome- associated proteins; VAMPs, vesicle-associated membrane proteins)118
Lipophilic aminoalcohol 2-(4-phenylpiperidino)cyclohexanol (vesamicol) is a well known high affinity VAChT ligand with neuromuscular blocking properties as a result of stereoselective, non-competitive (allosteric binding site) and reversible blockage of VAChT (Figure 7 and 8).119–121 Vesamicol also binds with low affinity to α-122 and moderate to high affinity to σ-receptors.123 Because of this low selectivity, vesamicol itself is not suitable for PET or SPECT imaging in brain. However, it has been a useful lead for developing more potent and selective ligands as potential SPECT and PET imaging probes aimed for in vitro, ex vivo, and in vivo studies of AD by accurate mapping distribution and amount of VAChT in the brain regions of interest (VAChT density: striatum > cortex > hippocampus > hypothalamus > cerebellum).121,124–149 Results obtained from these numerous studies confirmed that the VAChT binding site is stereoselective and that the (2R,3R)-enantiomers are
69
141,143 generally more potent (lower Ki) than (2S,3S)-enantiomers. Benzovesamicol has been found to be nearly equipotent to vesamicol, and is therefore one of the most explored class of compounds.136 The different groups can be incorporated at the 5-position of the benzovesamicol scaffold to modulate affinity, selectivity and pharmacokinetic properties of potential VAChT tracers.136 PET radioligands for VAChT has been recently reviewed by Giboureau et al.150 Poor selectivity over σ receptors (due to the structural similarity of σ receptors and VAChT pharmacophores and the similar tissue distribution profiles), and unfavourable pharmacokinetics (e.g. fast metabolism, low extraction from the blood, slow brain kinetics) are the main reasons why only a few VAChT ligands are currently promising radiotracers for the VAChT, though further validation is required to confirm their clinical usefulness (Figure 8).150
HO HO
N N
(1R,2R)-vesamicol (2R,3R)-benzovesamicol
HO HO
N N
(2R,3R)-[11C]MABV (2R,3R)-[18F]FEOBV HN O 13 18 CH3 H2C CH2 F HO HO H O N O N
H N O
18 (4aR,6R,7R,8aR)-[ F]FBMV 18F (2R,3R)-[18F]-2-Hydroxy-3-(4-(4-fluorobenzoyl)piperidino)tetralin
18 F HO
N
123I 123 (2R,3R)-5-[ I]IBVM 70
Figure 8. Chemical structures of (1R,2R)-vesamicol, (2R,3R)-benzovesamicol and the most promising PET and SPECT VAChT tracers, namely, (2R,3R)-[11C]-5-(N- methylamino)benzovesamicol ([11C]MABV),151 (2R,3R)-5-[18F]fluoroethoxybenzovesamicol ([18F]FEOBV),134,135,143 (4-[18F]fluorophenyl)((4aR,6R,7R,8aR)-7-hydroxy-6-(4- phenylpiperidin-1-yl)hexahydro-2H-benzo[b][1,4]oxazin-4(3H)-yl)methanone ([18F]FBMV),152 (2R,3R)-[18F]-2-hydroxy-3-(4-(4-fluorobenzoyl)piperidino)tetralin,147 and (2R,3R)-5-[123I]iodobenzovesamicol ([123I]IBVM).153
However, (2R,3R)-5-[123I]IBVM is to date the only VAChT radioligand widely used in human, and is therefore the lead VAChT tracer for SPECT.153–155
1.6. Synthesis of 5-aminobenzovesamicol (5-ABV) and its enantiomers
It should be noted that the synthesis of 5-aminobenzovesamicol (5-ABV) and its enantiomers (Scheme 12), which have been previously described,134,136 were not performed in this thesis, because there were enough of them left from the previous work by Giboureau et. al.143
The first step of the synthesis of 5-ABV is the reduction of the commercial available aminonaphthalene (1) by Birch reaction to 1,4-dihydronaphthalene-1-amine (2). The second step is the protection of the amine function by trifluoroacetic anhydride to obtain corresponding amide 3. The first two steps are almost quantitative. The third step is epoxidation of 3 by reaction with m-chloroperoxybenzoic acid (m-CPBA) to produce the corresponding epoxide 4 in ~75% yield. In the fourth step, the epoxide 4 is treated with the 4- phenylpiperidine to obtain after in situ deprotection and subsequent chromatographic purification two positional isomers 5-ABV and 8-ABV in ~25% and ~30% yield, respectively. In the fifth step, enantiomers of racemic 5-ABV are separeted by using the Mosher's acid chloride (MTPA-Cl), affording the corresponding diastereomers to allow their chromatographic separation. After reductive cleavage of the MTPA groups using diisobutylaluminium hydride (DIBAL-H), (2R,3R)-5-ABV and (2S,3S)-5-ABV are obtained in 6-8 % overall yield in enantiomeric purity greater than 98% as determined by chiral HPLC.
71
O O NH2 NH2 O NHCOCF3 Na, NH3 CF CF tBuOH, Et2O 3 3 benzene
1 2 3 (98 %) (100 %) O
OOH NHCOCF3 HO 2 1 NH 3 Cl N 8 O + 8-ABV Et O 7 2 EtOH, Et3N 4 5-ABV (30 %) 4 5 6 CF (25 %) (75 %) H3CO 3 H2N Cl 1. , DMAP/Et3N O (S)-MTPA-Cl 2. Enantiomeric purification by chromatography
MTPA O MTPA O
N N
(50 %) (65 %) HN HN MTPA MTPA
DIBAL-H DIBAL-H
HO HO
N N
H N H N (2R,3R)-5-ABV 2 (2S,3S)-5-ABV 2 (70 %) (65 %)
Scheme 12. Synthesis of 5-ABV and its enantiomers.134,136
72
2. AIMS AND SCOPE
73
As discussed in Chapter 1.5, one of the best selective and specific high affinity radiotracer to explore VAChT and one of the few experimental radiopharmaceuticals used in vivo in human to obtain an early diagnosis of AD is 5-[123I]-iodobenzovesamicol (5-[123I]IBVM). 5- [123I]IBVM is SPECT radiotracer (Figure 8, page 70). We suppose that the fluorine derivative 5-fluorobenzovesamicol (5-FBVM) should be of similar affinity and selectivity for the VAChT as 5-IBVM. This statement is based on the characteristic physico-chemical properties of fluorine. Its small atomic size should not impose substantial steric hindrance at binding of 5-FBVM on VAChT, and its electronegativity is roughly similar to iodine's electronegativity. In the case of successfully developed synthetic route towards 5-[18F]FBVM, biological studies could be done to prove the usefulness of 5-[18F]FBVM as a radiotracer with the appropriate pharmacokinetic and pharmacodynamic properties to selectively maps cholinergic brain areas in vivo; providing a non-invasive means of safely and accurately evaluate the functional integrity of cholinergic synapses in human using PET to obtain an early diagnosis of AD.
The main goal of the present PhD thesis is the synthesis of VAChT-selective ligand 5- FBVM labelled with radioactive fluorine isotope F-18. To reach this goal, we have established a plan based on following systematic and consecutive stages:
1. Choose appropriate synthetic method and find conditions via established theoretical model to introduce non-radioactive fluorine isotope F-19 at the C-5 position of the 5- aminobenzovesamicol (5-ABV).
HO HO 2 1
3 8 N ? N 7 4
5 6 5-ABV 5-FBVM H2N F
2. QSAR study and in vitro VAChT binding affinity determination of 5-FBVM by radioligand displacement study.
3. Choose and synthesize a suitable precursor of 5-FBVM for radiolabelling, and establish theoretical model to introduce non-radioactive fluorine isotope F-19 at the C- 5 position of the benzovesamicol under conditions which can be transposed to radiofluorination.
74
HO Under conditions which HO can be transposed to radiofluorination N N
5-FBVM Precursor moiety F
Benzovesamicol precursor
4. Develop improved method in order to acquire a variety of arylfluorides via one-pot fluorination of the chosen type of the precursor under rapid and operationally simple conditions.
Precursor moiety F Improved fluorination method R R
e.g. 5-FBVM
5. Examination of reaction parameters for radiofluorination of the developed benzovesamicol precursor.
6. Synthesis of 5-[18F]FBVM via radiofluorination of the developed benzovesamicol precursor.
HO HO Appropriate N reaction conditions N
[18F]F-
18 Precursor moiety 5-[ F]FBVM 18F
Benzovesamicol precursor
The major challenge will be the radiofluorination of the corresponding benzovesamicol precursor via aromatic nucleophilic fluorination, because benzovesamicol is a non-activated system towards aromatic nucleophilic fluorination. Furthermore, the synthesis and isolation of the target compound must be performed rapidly due to use of short-lived positron emitting 18 radionuclide F-18 (t1/2 ~ 110 min), which in the form of n.c.a. [ F]fluoride is used in great deficiency towards precursor and other reagents.
75
3. RESULTS AND DISCUSSION
76
3.1. Theoretical model for efficient one-pot fluoro-de-diazoniation
Balz-Schiemann reaction40 is a representative method and a broad scope method for the regioselective nucleophilic introduction of fluorine into aromatic rings in preparative organic syntheses. This is a deaminative fluorination type of reaction composed of three sequential steps: (A.) diazoniation of primary aromatic amine in aqueous medium with sodium nitrite o (NaNO2) and fluoroboric acid (HBF4) at 0-5 C to produce arenediazonium tetrafluoroborate + - + - (ArN2 BF4 ), (B.) isolation and drying of ArN2 BF4 to avoid side formations of phenols and 156 + - biaryl ethers, and (C.) thermal fluorinated decomposition of ArN2 BF4 (fluoro-de- diazoniation).157,158 However, the method suffers from yield reproducibility problems, because isolation and complete drying can be tedious and unsure (the method works only for those diazonium tetrafluoroborates which can precipitate from aqueous medium), and 156,159 controlled thermal decomposition of ArN2BF4 is problematic. To overcome reproducibility problems, simplify procedure, broaden substrate tolerance (e.g. anilines with hydrophilic substituents, heterocyclic amines), improve safety (diazonium tetrafluoroborates are toxic and potential explosive when perfectly dried), and to increase the yields, alternative approaches based mostly on one-pot methodology (in situ fluoro-de-diazoniation without isolation of arenediazonium salt) in organic and ionic solvents have been developed during the last few decades.159–165
77
A. 0-5 oC Ar-NH + NaNO + 2 HX ArN + X- + H O + NaX 2 2 (aq.) 2 2
B. - NaX HX (aq.) - H2O + - + + + Na O-N=O + HX (aq.) HO-N=O H2O -N=O N =O N O Sodium nitrite Nitrous acid Nitrosonium ion (unstable) (nitrosyl cation) C. NH 2 H2N N O HN N O
- H+ + N+ O N O+ + H+ R R R N-Nitrosamine (unstable)
N N OH N N OH2 N N N N
+ + H - H2O - H+ R R R R Diazohydroxide Aromatic diazonium ion
Scheme 13. (A.) Summary equation of formation of diazonium salt in acidified water solution at 0-5 oC. (B.) Formation of nitrosonium ion/nitrosyl cation as diazonating agent. Nitrosonium ion is a very weak electrophile due to resonance stabilisation. (C.) Mechanism of diazoniation (N-nitrozination) of aniline.
Arenediazonium ions can undergo three types of reaction: (A.) Reactions of nucleophile at
N(2)-atom of diazonium moiety; (B.) Unimolecular nucleophilic substitution (SN1) via aryl cation intermediate; (C.) One-electron reduction of diazonium moiety with further homolytic dissociation of arenediazonium ion into aryl radical (Scheme 14). Understanding these processes is pivotal to construct theoretical model for successful fluorination of 5-ABV via modified Balz-Schiemann procedure.
A. Arenediazonium ions are soft electrophiles and coexist in cold solution with soft nucleophiles. Stable diazo compound can be formed through attacking N(2)-atom by soft nucleophile at enough elevated temperature. On the other hand, reaction with hard nucleophile (e.g. hydroxide anion) gives corresponding unstable diazo compound that undergoes decomposition to finally yield more stable product(s).
78
B. Fluorination of anilines by Balz-Schiemann reaction is an example of unimolecular nucleophilic aromatic substitution that proceeds via highly unstable and non-selective aryl cation intermediate.
C. Arenediazonium ion can be subject to one-electron reduction if any present reagents or solvent have lower redox potential than arenediazonium group. Consequently, arendiazonium ion is decomposed via aryl radical, which is capable of abstracting a hydrogen atom (e.g. from protic solvent) or another atom from a covalent bond (e.g. Sandmeyer reaction).
Taken together, the diazonium ions can decompose both by ionic (heterolytic) as well as radical (homolytic) paths.
+ Ar-N2 Arenediazonium ion Ionic path (heterolytic decomposition) Radical path - N2 (g) X
+ Ar + Ar N(1) N(2) X Aryl cation Ion pair X One-electron transfer Ar-X + . Ar N N. Ar N N . X
Diazonium radical
- N2(g) Homolytic decomposition
. Ar . Protic solvent Ar Ar-H Aryl radical Ar-Ar . X
Ar-X Other radical reactions Scheme 14. At elevated temperature aryldiazonium ion can decompose by ionic and/or radical paths.166,167
The delicate decomposition path balance is crucially dependent on the reaction conditions and the substitutents in the aromatic ring(s). More precisely, substituents (their mesomeric and inductive effects) and their substitution pattern affect stability of the aryldiazonium ion
79
and its decomposition intermediates, its redox potential and consequently its decomposition temperature and paths.158,166,167 Substituents with electron withdrawing character (e.g. nitro group) generally increase activation energy for de-diazoniation and are therefore rate retarding, and vice versa for substituents with electron donating character. Moreover, the redox potentials of the substituted aryl diazonium ions are highly useful in predicting the nature of the de-diazoniation path. Aryl diazonium ions having electron-donating groups increase electron density of the N(2)-diazonium atom and stabilize the diazonium cation, thereby supressing their tendency to undergo one-electron reduction. On the other hand, diazonium ions having electron-withdrawing substituents have a stabilising influence by resonance on the diazenyl radical and increase the ease of their reduction.167
Redox potential of molecules is also strongly dependent on the selected solvent.
Table 2. Redox potentials vs normal hydrogen electrode (NHE) for halide ions in acetonitrile 167 (CH3CN) and water (H2O).
Eo, V reduced form of the redox couple CH3CN H2O
F - 2.4 3.6 Br - 1.2 2.0 I - 1.1 2.2 As shown in Table 2, all halide ions have lower redox potential (easier to donate electron and become oxidized) in acetonitrile than in water. This could be due to lower solvation strength of CH3CN compared to H2O. The more is halide anion solvated less nucleophilic it is and less efficiently it donates electron and become oxidized (higher redox potential). According to the Table 2, the redox potential of fluoride is most likely not enough low for diazonium ion to undergo a one-electron reduction by fluoride.
Thus, reaction conditions for successful (one-pot) fluoro-de-diazoniation should be carefully chosen to promote aryl cation formation. The solvent, pH of the reaction medium, nature of the counter ion and the presence of reducing agents and/or radical sources influence arylfluorides yields. Choice of the solvent is one of the most important parameter to facilitate 166,167 fluoro-de-diazoniation. Chlorinated solvents such as chloroform (CHCl3, tetrachloromethane (CCl4), and 1,2-dichlorobenzene have been reported to have beneficial effect on arylfluride yields via probable enhancement of the ionic decomposition path.167
80
Accoding to the studied properties of aryldiazonium ions, we can write parameters with optimal properties for efficient one-pot fluoro-de-diazoniation reaction:
A. Solvent: Dissolving all the reagents with minimal solvation of fluoride anion in order to preserve its nucleophilicity, non-nucleophilic, suitable high redox potential to avoid reduction of the aryldiazonium ion and consequently suppress homolytic decomposition path, to avoid formation of Ar-H (aprotic solvent) which can be problematic to separate from Ar-F, high enough boiling point to decompose aryldiazonium salt and to suppress radical decomposition path which is kinetically and thermodynamically favoured to ionic path.
B. Temperature: Decomposition temperature of diazonium ion is unique to particular diazonium ion and is strongly dependent on the substitution of aromatic ring(s).
C. Diazonating agent (source of nitrosonium ion): The best choice are alkyl nitrites (n- butyl nitrite or t-butyl nitrite), although it is well known that they are mild reagents for diazonation in organic solvents. They are soluble in organic solvents.
D. Source of fluoride anion: Should be soluble in organic solvent as boron trifluoride- + - diethyl etherate (BF3xE2O), nitrosonium tetrafluoroborate (NO BF4 ) or nitrosonium + - hexafluorophosphate (NO PF6 ). Alkali metal fluorides (CsF, KF, …) are not good choice, because of their hygroscopicity and low solubility in organic solvents (long reaction time and use of very high temperature can lead to thermal run-away and multiplicity of products). Quaternary ammonium fluorides are very hygroscopic, but have better solubility in organic dipolar aprotic solvents, such as N,N-
dimethylformamide (DMF), dimethyl sulfoxide (DMSO), and acetonitrile (CH3CN), than their alkali metal counterparts. They are generally commercially available in hydrated states and for this reason are not the best choice. Only tetra-n-butyl ammonium fluoride (TBAF) is commercially available as 1M solution in
tetrahydrofuran (THF). Silicium fluoride (SiF4) is toxic and thus not reagent of choice.
81
3.2. Synthesis of 5-FBVM and its enantiomers via fluoro-de-diazoniation
We first performed one-pot fluoro-de-diazoniation of 5-ABV according to the literature165 using 1,2-dichlorobenzene as the solvent, 1.2 eq. of t-butylnitrite as the diazonating agent and
1.5 eq. of boron trifluoride diethyl etherate (BF3xEt2O) as the fluoride source. As suggested, t- butylnitrite was added portionwise to the hot (105 oC) reaction mixture of the solvent, 5-ABV, and fluoride source, and left stirring for one hour. It is assumed that under this condition amount of diazonium ion of starting aniline is very low, because the reaction temperature is above the decomposition of diazonium salt, which probably have beneficial effect on the arylfluoride (ArF) yields.165 After extraction and chromatographic purification, 5-FBVM was obtained as a white powder in 16% yield. The product was analyzed by 1H-, 13C-nuclear magnetic resonance (13C-NMR) and mass spectrometry (MS).
HO HO 1. 5-ABV 2. 1.5 eq. BF3xEt2O N 3. 1.2 eq. t-butylnitrite N 1,2-dichlorobenzene, 105 oC, 1h H N F 5-ABV 2 16 % 5-FBVM
Scheme 15. Synthesis of 5-FBVM via one-pot fluoro-de-diazoniation of 5-ABV.
In Balz–Schiemann reaction, the mineral acid is used to generate nitrosonium ion ([NO]+), and to protonate diazohydroxide, followed by elimination of water and formation of aryl- diazonium ion (Scheme 13). In one-pot fluoro-de-diazoniation reaction using t-butylnitrite
(even weaker electrophile than nitrosonium ion) and BF3xEt2O is no protic acid to generate nitrosonium ion from alkyl nitrite and to protonate diazohydroxide to form an aryldiazonium ion. Accordingly, we proposed that the first rate-limiting step is nucleophilic attack of primary aromatic amine on alkyl nitrite followed by the formation of diazohydroxide which is complex with boron trifluoride molecule. Formation of the proposed complex probably lowers activation energy (Ea) for breaking the CAr-N(1) bond at elevated temperature (second rate-limiting step), and ArF is obtained via formation of diazonium fluoride and highly unstable diazonium cation. Chlorinated solvents might stabilize this complex through weak δ+ intermolecular electrostatic interaction, between the C Ar-atom (diazohydroxide group exerts negative inductive effect upon CAr-atom) and an electron pair on the chlorine atom of the chlorinated solvent (Scheme 16). Besides their non-nucleophilic nature and high redox
82
potential, this stabilization of the complex might be one of the reasons why chlorinated solvents are the best choice for one-pot fluoro-de-diazoniation.
First Ar NH rate-limiting step H Ar NH + R O N O - ROH 2 N OH Ar N N O Aniline R O N-Nitrosamine Alkyl nitrite
Chlorinated solvent weak intermolecular Second R Cl electrostatic interaction H rate-limiting step Ar N N O : H - BF OH - N2 (g) .... N N O 2 ArN Ar Diazohydroxide _ 2 F forming I B Aryldiazonium Aryl cation F fluoride F Complex F F F B F F
breaking Ar-F ...... Aryl fluoride O Et Et Scheme 16. Proposed mechanism of formation arylfluoride (ArF) using aniline, alkylnitrite and boron trifluoride etherate in chlorinated solvent under reflux.
Increasing the amount of t-butylnitrite and boron trifluoride (2 equivalents) did not increase the 5-FBVM yield (12%).
We left reaction mixture of 5-ABV, 1.5 equivalent of BF3xEt2O, and 1.5 equivalent of t- butylnitrite in dichloromethane (CH2Cl2) at room temperature overnight and afterwards refluxed for 90 minutes. After the usual workup and chromatographic purification, (rac)-5- FBVM was obtained in 25% yield. We also obtained (2R,3R)-5-FBVM and (2S,3S)-5-FBVM by this procedure as white powder in 25% and 27% yield in two consecutive experiments, respectively, starting from the corresponding (2R,3R)-5-ABV and (2S,3S)-5-ABV (Chapter 3.4, pages 88-98). The reaction mixture should be left at least four hours at cold to room temperature, otherwise some unreacted 5-ABV was always detected after refluxing the reaction mixture for 60-90 minutes.
Similarly, we detected considerable amount of unreacted 5-ABV with no formation of 5- FBVM (thin layer chromatography, 1H-, and 13C-NMR) when 1.5 eq. of nitrosonium tetrafluoroborate (NOBF4; diazonating agent and fluoride source), or 1.5 eq. of sodium tetrafluoroborate (NaBF4) with 1.5 equivalent of t-butylnitrite, or 1.5 eq. tetra-n-
83
butylammonium fluoride (TBAF 1M soln. in THF) with 1.5 eq. t-butylnitrite were used under similar reaction conditions as described in previous paragraph. Failed attempts are probably due to insolubility of NOBF4 and NaBF4 in chlorinated solvents, and much lower fluorinating efficiency of TBAF compared to BF3xEt2O in this type of reaction.
Ionic liquid solvents are alternative reaction media of increasing interest in numerous types of reactions; even in fluorination (Scheme 17).164,168–173 They are regarded as an environmentally friendly reaction media (“green solvents”), in contrast to the volatile and many times toxic organic solvents widely used in organic reactions. They can act as catalysts by accelerating reaction rates (shorter reaction time and lower reaction temperature), enhance the reactivity of reagents, improve reaction selectivity, reduce the formation of by-products and facilitate catalyst recovery. According to the literature, fluorination method using ionic solvents does not require strictly anhydrous conditions, which is in contrast what is generally considered to be required for fluorination reactions. It is sometimes enough to use just catalytic amounts of the solvent to complete reaction in relative short time. However, a major disadvantage is their high cost that makes their regeneration an important issue.
Our attempt to perform one-pot fluoro-de-diazoniation of 5-ABV in ionic liquid, namely [emim]x[OTf] (1-ethyl-3-methylimidazolium trifluoromethanesulfonate/triflate), using nitrosonium tetrafluoroborate failed.164 It is important to note that 5-ABV was not soluble in [emim]x[OTf] and that anion exchange between aryl diazonium cation and the cationic part of the ionic solvent most probably prevented formation of 5-FBVM (Scheme 17).164,173
84
Et Et Et Et N N N CF SO N 3 3 BF4 N2 BF4 N2 CF3SO3
R R Aryl-diazonium Aryl-diazonium tetrafluoroborate Et Et Et Et triflate N N N N CF3SO3 BF4
- N heat - N2 (g) heat 2 (g) emim x OTf emim x BF4
O CF F 3 S O O
R Arylfluoride R Trifluoro-methanesulfonic acid aryl ester Scheme 17. Anion exchange (metathesis) between an aryl-diazonium ion and a cation of the ionic solvent can (drastically) decrease yield of the desired arylfluoride, inspite of the low nucleophilicity of the anion (e.g. triflate) of the ionic solvent. This phenomenom has special importance in reactions involving short-lived reactive intermediates (aryl cation).
According to the anion exchange and the fact that ionic liquid is used in excess to fluoride source, [emim]x[BF4] or [emim]x[PF6] should be the best choice. However, these solvents can not be used in radiofluorination due to source of “cold’’ fluoride.
85
3.3. Theoretical model for efficient fluoro-de-triazenation and synthesis of 5-FBVM from the corresponding triazene precursor
1-Aryl-3,3-dialkyltriazenes (Ar-N=N-NR'R''), compounds having a diazoamino group, are regarded as protected form of anilines and stable surrogates of aryldiazonium ions.44 After treatment with the appropriate reagents, they are adaptable to numerous synthetic transformations with wide applicability in chemical, medical and technological fields.174,175 Triazene group is stable towards electrophilic reagents, some oxidants and reductants, but not towards acids at room or elevated temperature. Thus, their acid-triggered thermal decomposition parallels that of diazonium ion reactions (Scheme 18).166,167,176 Consequently the theoretical model for efficient fluoro-de-triazenation is almost the same as for fluoro-de- diazoniation, except that the acid is present (Scheme 18).
HA Ar N(1) N(2) N(3)R2 Ar N N NHR2 A
Ar Ar + ... heat HA R2NH R2NH2 A Ox Red . Ar F Protic solvent, HA, Ar Ar N N A - N F triazene 2 (g) Ionic path A - N2 (g) Ar A Ar H Radical path Solvent Ar Ar Solv Scheme 18. General competitive processes during fluoro-de-triazenation and fluoro-de- diazoniation in the protic acid mediated decomposition of 1-arly-3,3-dialkyltriazenes.176 In order to obtain arylfluoride in satisfactory yields, reaction conditions should be carefully chosen that the radical decomposition path is suppressed and the ionic path maximized.
Aryltriazenes are safely and mostly readily prepared by coupling of a diazotized aniline with amine or by the action of Grignard reagents on aryl azides. They can be stored for a long period of time at cold (0-5 oC) protected from light.177 Additionally, aryltriazenes can be easily isolated, chromatographically purified, introduced in the early stages of the synthesis, functionalized and thermally decomposed in the presence of protic acid in the latest stages if the preceding reactions have been performed under non-acidic conditions.176,178 Thus, fluoro- de-triazenation can be attractive means of forming 18F-labelled fluoroaromatics by direct nucleophilic substitution with [18F]fluoride due to both the one-pot and rapid nature of triazene transformation to fluoroarenes in order to obtain PET tracers with good specific
86
activity. Although diaryliodonium salts have mainly been proven to be much more efficient precursors for the introduction of [18F]fluoride onto mostly simple non-activated aromatic rings than fluoro-de-triazenation, preparation of complex diaryliodonium salt, such as from 5- ABV, in a highly pure and stable form needed for radiofluorination can be very problematic. Due to the reasons stated above, we decided to prepare corresponding 5-ABV triazene precursor (5-TBV) and search for the fluoro-de-triazenation conditions that can be transposed to radiolabelling.
The acid-triggered thermal decomposition of 1-aryl-3,3-dialkyltriazenes involves N(3) protonation by a strong protic acid, followed by the rate-limiting heterolytic N(2)-N(3) bond- cleavage at elevated temperature giving the corresponding diazonium ion and dialkylamine. The N(3) protonation is a crucial step, because it competes with N(1) protonation but both are more favourable than N(2) protonation. However, N(1) protonation does not induce triazene decomposition, because it allows charge delocalisation over the three nitrogen atoms. In contrast, N(3) protonation lengthens and destabilizes the N(2)-N(3) bond by inhibiting delocalization across the triazene linkage. Electron-donating substituents on aryl ring facilitate the N(3) protonation because electron-donation stabilizes the positive charge on N(3) and therefore promote the N(2)-N(3) bond-breaking or diazonium ion formation.167,179,180
According to the Scheme 18 and chemical properties of the aryltriazenes and aryldiazonium ions, the optimal properties of the used acid in fluoro-de-triazenation are the following: (A.) strong acid to protonate and induce the cleavage of the triazene moiety; (B.) solubility in chosen solvent (e.g. chlorinated solvents); (C.) non-nucleophilic conjugated base (A-) with high redox potential to avoid the acid counterion substitueted byproduct (Ar-A) formation and inducing radical decomposition pathway via reduction of the corresponding diazonium ion intermediate.
We prepared the suitable triazene precursor 5-ABV-diethyltriazene (5-TBV) by the standard synthetic procedure via diazoniation of the 5-ABV in the cold acidified water mixture using sodium nitrite with subsequent addition of diethylamine to provide 5-TBV as a white powder in almost quantitative yield (95%). (2R,3R)-5-TBV and (2S,3S)-5-TBV were obtained by the same procedure as white powders in 96% and 92% yield, respectively, starting from the corresponding (2R,3R)-5-ABV and (2S,3S)-5-ABV. We also prepared triazene precursor (rac)-5-TBV in 88% yield by modified method using boron trifluoride etherate and t-butylnitrite in CH2Cl2 with subsequent addition of diethylamine (Chapter 3.4).
87
3.4. 3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential PET radioligand for the vesicular acetylcholine transporter (VAChT)
Mitja Kovac a, Sylvie Mavel a,*, Winnie Deuther-Conrad b, Nathalie Méheux a, Jana Glöckner b, Barbara Wenzel b, Marko Anderluh c, Peter Brust b, Denis Guilloteau a, Patrick Emond a a Université François-Rabelais de Tours, INSERM U930, CHRU, Hôpital Bretonneau, Service de Médecine Nucléaire, 37000 Tours, France b Forschungszentrum Dresden-Rossendorf, Research Site Leipzig, Institute of Radiopharmacy, Permoserstr. 15, 04318 Leipzig, Germany c University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Aškerčeva 7, 1000 Ljubljana, Slovenia
Published in: Bioorganic & Medicinal Chemistry 18 (2010) 7659-7667
88
SUMMARY:
In this report we described the synthesis of 5-FBVM and its enantiomers via fluoro-de- diazoniation from the corresponding 5-ABV. To demonstrate the suitability of triazene as leaving group, the non-radioactive fluoro-de-triazenation of 5-TBV resulting in reasonably 25% yield of 5-FBVM was accomplished. We also performed three QSAR studies based on 32 vesamicol and benzovesamicol derivatives taking into account for the first time the stereoselectivity of the VAChT binding site in order to predict the binding affinity of (2R,3R)- and (2S,3S)-5-FBVM. Both enantiomers exhibited high in vitro VACHT binding affinites determined by radioligand displacement studies, and were in the same range as 5-IBVM as predicted by 3D QSAR studies. Only (2S,3S)-FBVM was selective enough over σ1 receptors to warrant further investigation as a potential PET radioligand for in vivo mapping of cholinergic nerve terminals.
STATEMENT: I declare, that nobody of co-authors has used the article 3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential PET radioligand for the vesicular acetylcholine transporter (VAChT), published in the Bioorganic & Medicinal Chemistry, for his/her own thesis.
89
Bioorganic & Medicinal Chemistry 18 (2010) 7659–7667
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
3D QSAR study, synthesis, and in vitro evaluation of (+)-5-FBVM as potential PET radioligand for the vesicular acetylcholine transporter (VAChT)
Mitja Kovac a, Sylvie Mavel a,*, Winnie Deuther-Conrad b, Nathalie Méheux a, Jana Glöckner b, Barbara Wenzel b, Marko Anderluh c, Peter Brust b, Denis Guilloteau a, Patrick Emond a a Université François-Rabelais de Tours, INSERM U930, CHRU, Hôpital Bretonneau, Service de Médecine Nucléaire, 37000 Tours, France b Forschungszentrum Dresden-Rossendorf, Research Site Leipzig, Institute of Radiopharmacy, Permoserstr. 15, 04318 Leipzig, Germany c University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Aškercˇeva 7, 1000 Ljubljana, Slovenia article info abstract
Article history: Located in presynaptic cholinergic nerve terminals, the vesicular acetylcholine transporter (VAChT) rep- Received 18 May 2010 resents a potential target for quantitative visualization of early degeneration of cholinergic neurons in Revised 26 May 2010 Alzheimer’s disease using PET. Benzovesamicol derivatives are proposed as radioligands for this purpose. Accepted 12 August 2010 We report QSAR studies of vesamicol and benzovesamicol derivatives taking into account the stereose- Available online 17 August 2010 lectivity of the VAChT binding site. Use of different data sets and different models in this study revealed that both enantiomers of 5-fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tetrahydro-naphthalen-2-ol Keywords: (5-FBVM) are promising candidates, with predicted VAChT affinities between 6.1 and 0.05 nM. The syn- Benzovesamicol derivative thesis of enantiopure (R,R)- and (S,S)-5-FBVM and their corresponding triazene precursors for future VAChT Triazene radiofluorination is reported. Both enantiomers exhibited high in vitro affinity for VAChT [(+)-5-FBVM: Fluoro-dediazoniation Ki = 6.95 nM and (À)-5-FBVM: Ki = 3.68 nM] and were selective for r2 receptors (70-fold), only (+)-5- 3D QSAR FBVM is selective for r1 receptors (fivefold). These initial results suggest that (+)-(S,S)-5-FBVM warrants further investigation as a potential radioligand for in vivo PET imaging of cholinergic nerve terminals. Ó 2010 Elsevier Ltd. All rights reserved.
1. Introduction ing by PET have focused on compounds such as benzovesamicols,3–6 trozamicols,7,8 or morpholino vesamicols,9,10 all based on the struc- The degeneration of cholinergic neurons in the brain is one of the ture of (À)-vesamicol, a drug that binds to a side allosteric to the act- most significant neuropathological features in Alzheimer’s disease eylcholine transport side with low-nanomolar affinity.11–14 (AD) synapse disorder has been shown to precede the neurofibrillary However, (À)-vesamicol possesses moderate affinity to a-adreno- and neuritic aspects of AD during the course of the disease,1 and the ceptors and nanomolar affinity to r-receptors.15 In combination 16 17 loss of synaptic terminals correlates better with cognitive decline with similar expression patterns of VAChT and r1-receptors in than extracellular plaque load or loss of neurons.2 Located in presyn- the brain the latter cause marginal signal-to-background ratios in aptic cholinergic nerve terminals, the vesicular acetylcholine trans- in vivo imaging by PET, one reason why so many VAChT radioligands porter (VAChT) is postulated to be a valuable target for in vivo have failed in pre-clinical evaluation to date. On the other hand, the quantitative visualization of early neurodegenerative processes in successful application of the benzovesamicol-related SPECT ligand AD by using molecular imaging techniques such as SPECT (Single (À)-5-[123I]IBVM (5-iodo-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tet- 18–21 Photon Emission Computed Tomography) and PET (Positron Emis- rahydro-naphthalen-2-ol, Kd = 0.30 nM) has encouraged fur- sion Tomography), PET being regarded as superior in terms of detec- ther research into the design of VAChT-specific PET tracers. tion efficiency, spatial resolution, and quantification. However, Structure-affinity studies assessing the potential of iodobenzove- VAChT-specific tracer compounds labeled with the short-lived PET samicol derivatives for visualization of cholinergic terminals have radionuclides 18F and 11C have not been developed for clinical use revealed that there is considerable bulk tolerance at different to date. Approaches to the development of tracers for VAChT imag- structural positions of the benzovesamicol molecule.18,22 Therefore we have recently synthesized and evaluated new benzovesamicol derivatives21 and aza-analogs of trozamicol deriv- atives both in vitro and in vivo in animals models.23–25 Several 18F- * Corresponding author. Address: Faculté de Pharmacie, Laboratoire de Biophy- labeled benzovesamicol derivatives have now been synthesized sique Médicale et Pharmaceutique, 31 avenue Monge, 37200 Tours, France. Tel.: +33 18 18 6 18 5 18 3 2 47 36 72 40; fax: +33 2 47 36 72 24. including [ F]NEFA, [ F]FAA, [ F]FEOBV, and [ F]FPOBV 18 E-mail address: [email protected] (S. Mavel). (Table 1). However, except for [ F]FEOBV, which is currently
0968-0896/$ - see front matter Ó 2010 Elsevier Ltd. All rights reserved. doi:10.1016/j.bmc.2010.08.028 7660 M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
Table 1 Vesamicol analogs I and benzovesamicols II used for the QSAR studies
HO HO 9 8 N A 11 N C 1 13 R R R3 D 2 I II R 5
R I Refa R1 R2 R3 II Refa
H (À)-vesamicol 21 H 5-I H (À)-5-IBVM 21 (+)-vesamicol (+)-5-IBVM
o-CH3 (À)-oMV 30 H 5-N(Et)COCH2FH (À)-NEFA 50 (+)-oMV
p-CH3 (À)-pMV 30 H 5-O(CH2)2FH (À)-FEOBV (+)-pMV (+)-FEOBV 52 o-I (À)-oIV H 5-(CH2)3FH (À)-FPOBV 21 (+)-oIV (+)-FPOBV 52 m-I (À)-mIV H 5-OCH2CHCHI H (À)-AOIBV 21 (+)-mIV (+)-AOIBV 52 p-I (À)-pIV H 5-I and 8-OCH3 H (À)-MOIBV 21 (+)-MOIBV
H 5-I CH2NH2 (À)-MAIBV 21 (+)-MAIBV H 6-I H (À)-6-IBVM 22 H 7-I H (À)-7-IBVM 22 H 8-I H (À)-8-IBVM 22 p-I H H (À)-pI-BVM 22 m-I H H (À)-mI-BVM 53
H 5-NH2 H (À)-ABV 50 (+)-ABV HH H (À)-BVM 50
a The data of binding affinities used for QSAR study were taken from these references.
under investigation in PET studies,4 these PET tracers have not synthesis and provided the basis for the synthesis of enantiopure been suitable for in vivo applications, possibly due to the metabolic (+)-(S,S) and (À)-(R,R)-5-FBVM as well as (S,S) and (R,R)-5-TBV. susceptibility of the F-carrying substituents, for example, amide in NEFA and alcoholate in FEOBV. In previous paper, FPOBV instability 2. Results and discussion in vivo,3 possibly arising from the weak strength of the C(sp3)–F Ò bond (0.00079 kcal/mol, DiscoveryStudio2.5 , Accelrys Inc.) caus- 2.1. QSAR study ing de-fluorination is reported. To increase the metabolic stability 2 and to reinforce the C–F bond strength by a C(sp )–F bond The QSAR study is based on 32 derivatives which belong to the Ò (0.0628 kcal/mol, DiscoveryStudio2.5 , Accelrys Inc.), we have classes of vesamicols (I) and benzovesamicols (II) (Table 1). Six developed the novel benzovesamicol analog 5-FBVM as a fluoro derivatives were synthesized and evaluated regarding VAChT affin- analog of 5-IBVM (Table 1). ity and specificity by in-house in vitro assays, the Ki values of the 26 In 2008, Szymoszek et al. have published a first Comparative remaining compounds were taken from literature. Since binding Molecular Field Analysis (CoMFA) study using a partial least to the VAChT is known to be highly enantioselective (generally, squares (PLS) algorithm for a set of 37 vesamicol derivatives, cov- the in vitro affinity for the VAChT of (À)-enantiomers is about 10 ering three different structural types, to predict the binding affinity times greater) we made three sets from the same overlay: Set 1 of vesamicol-type ligands from their respective molecular struc- for all 32 derivatives, Set 2 for the (À)-enantiomers (20 deriva- ture. To expand these efforts, we have performed a further 3D tives), and Set 3 for the (+)-enantiomers (12 derivatives). All the QSAR study, which is based on 32 vesamicol and benzovesamicol structures were minimized under a CHARMm forcefield with a root derivatives (Table 1) in order to predict the binding affinity for mean squared (RMS) difference of energy gradient reached the new compound 5-FBVM. Furthermore, this study considered 0.1 kcal/mol Å (Discovery StudioÒ 2.5, Accelrys Inc., San Diego, for the first time the stereoselectivity of the binding of vesamicol CA). According to crystallographic findings on vesamicol deriva- derivatives to the VAChT protein. tives10,27 and ABV (data not shown, structure see Table 1), the Moreover, we also wished to improve the accessibility of the piperidine ring is in a chair conformation, and is almost perpendic- radiolabeling procedure by radiofluorination of non-activated aro- ular to the cyclohexanol ring, with the hydroxyl in trans position as matic cycles (i.e., without an electron withdrawing group in ortho an equatorial conformer. The carbons C-9, C-11, and C-13 of piper- or para position to the leaving group). Therefore, in this report we idine (see Table 1) and C-OH were chosen for the overlay. Ki values describe not only the synthesis of 5-FBVM via fluoro-dediazonia- were taken from the literature, and pKi = Àlog Ki was used for the tion based on the secondary amine 5-ABV, but also the synthesis Genetic Function Approximation (GFA) algorithm.28 The QSAR pro- 18 of a suitable triazene precursor (5-TBV) for future F-labeling. To gram presents 179 physicochemical descriptors as different elec- demonstrate the suitability of triazene as leaving group, the non- tronic, spatial, shadow, shape, and thermodynamic indices. After radioactive fluorination of 5-TBV resulting in 5-FBVM was several filters (analyzing the correlation matrix, eliminating the accomplished. Because of the stereoselective binding of vesamicol highly correlated descriptors, and eliminating descriptors with derivatives, 5-ABV was enantioseparated via Mosher ester too wide a range of training data (>700)), only 26 2D or 3D M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667 7661
descriptors were retained as dependent on pKi. The definitions of 2.1.3. Set 3 mean descriptors are given in Table 2. In the GFA model, linear N = 12: (+)-enantiomers. or quadratic terms of descriptors were allowed in the selection. Linear model form: pKi(+)pred= 27.60(±0.02) + 3.42(±0.003)  For the statistical parameters, the cross validated r2 is not the most CHI_2 À 1.73(±0.001)  CHI_V_3_PÀ 0.059(±0)  E_ADJ_mag + 19.66 important, and the Friedman’s lack-of-fit (LOF) score, which evalu- (±0.05)  JX À 30.19(±0.05)  JY À 0.654(±0.0003)  Dipole_Z À 9.64 ates the QSAR model by considering the number of descriptors as (±0.01)  Jurs_RPSA À 3.73(±0.007)  Shadow_XYfrac + 3.67(±0.002) well as the quality of fitness, is chosen: the lower the LOF, the less  Bond Energy À 0.067(±0)  Dihedral Energy. likely it is that the GFA model will fit the data. The significant regression is given by F, and the higher the value, the better the model. The standard errors of regression coefficients are given in r2 = 1.0000 r2 (adj) = r2 (pred) = RMS residual Friedman parentheses. GFA models were tested with 5-FBVM, with the excel- 1.0000 1.0000 error = 0.0004352 L.O.F. = 1.844eÀ006 lent predictive sub-nanomolar Ki.
2.1.1. Set 1 Predicted affinity: pKi(+)-5-FBVMpred = 8.22. N = 32: (+) and (À)-enantiomers. The major contributing factors are JY > JX > Jurs_RPSA > Sha- dow_XYfrac Bond Energy CHI_2 > CHI_V_3_P > Dipole_Z > Di- Linear model form: pKi(+)/(À)pred = 79.61(±10.49) À 39.86 (±5.82) Â JY + 0.99(±0.22) Â Dipole_Z + 0.0185(±0.004) Â Jurs_PN- hedral Energy E_ADJ_mag. SA_2 À 4.61(±1.45) Â Shadow_nu + 3.28(±1.07) Â Bond Energy À The best predictive values for 5-FBVM were obtained with the 0.894(±0.348) Â Van der Waals Energy. linear model rather than the quadratic model. Furthermore, consid- ering the (À)-enantioselectivity of VAChT binding site, we focused on (+)/(À) and (À) linear models. For set 1 ((+)/(À) model), the main r2 = 0.7425 r2 (adj) = 0.6807 r2 (pred) = 0.6101 RMS residual Friedman descriptor was JY (Balaban index which characterizes the shape of a error = 0.7419 L.O.F. = 1.155 molecule taking into account the relative covalent radius of the atoms of the model) with a mean value for the 32 compounds of Predicted affinity: pK (+)-5-FBVM = 8.91 pK ( )-5-FBVM = i pred i À pred 1.50. As the coefficient in the equation was negative, the smaller 9.34. the value, the better the affinity for the VAChT (JY( )-5-FBVM = 1.460 The major contributing factors are JY > Shadow_nu > Bond En- À compared to JY = 1.447 and JY(À)-vesamicol = 1.579). For set 2 ((À) ergy > Van der Waals Energy > Dipole_Z > Jurs_PNSA_2. (À)-IBVM model), the main descriptor was Shadow_XZfrac (a steric descrip- Quadratic model form: pK (+)/( ) = 71.22(±9.53) 32.92 i À pred À tor) with a mean value for the 20 compounds of 0.606. The negative (±4.57) JY + 1.03(±0.21) Dipole_Z + 0.0160(±0.003) Jurs_PNSA_2 Â Â Â coefficient of shadow in the XZ plane indicated that a decrease in 4.10(±1.28) Shadow_nu 0.73(±0.23) Bond Energy Van À Â À Â Â the area of the molecular shadow in the XZ plane (Shadow_XZfrac der Waals Energy. (À)-5-FBVM = 0.577 compared to Shadow_XZfrac(À)-IBVM = 0.581 and Shadow_XZfrac(À)-vesamicol = 0.611) was favorable for affinity for r2 = 0.7469 r2 (adj) = 0.6982 r2 (pred) = 0.6454 RMS residual Friedman the VAChT. The equations showed multiple occurrence of Jurs error = 0.7213 L.O.F. = 1.028 descriptors,29 suggesting the importance of charge distribution and surface areas, in particular for the Jurs_RPSA descriptor (total polar surface area divided by the total molecular solvent-accessible Predicted affinity: pKi(+)-5-FBVMpred = 9.21 pKi(À)-5-FBVMpred = surface area) with a mean value for the 20 compounds of 0.0789. 9.75. Since Jurs_RPSA depends on polar surface, fluorine derivatives had the highest values and the coefficient in the equation was 2.1.2. Set 2 negative. For 5-FBVM the descriptor was the smallest of the N = 20: (À)-enantiomers. fluorine derivatives (Jurs_RPSA(À)-5-FBVM = 0.128 compared to Linear model form: pK i(À)pred = 14.06(±3.67) + 0.0759 Jurs_RPSA = 0.152 and Jurs_RPSA = 0.130). (±0.0120) Â Molecular_SASA À 0.0766(±0.011) Â E_ADJ_mag + 0.9205 (À)-NEFA (À)-FEOBV The 3D QSAR model (Discovery StudioÒ 2.5, Accelrys), defines (±0.0834) Â Jurs_PPSA_3 À 38.88(±7.73) Â Jurs_RPSA + 2.89(±1.13) Â the critical regions (steric or electrostatic) affecting binding affin- RadofGyration À 5.20(±0.76) Â Shadow_nu À 66.45(±10.20)Â ity. A contour plot of the electrostatic field region favorable Shadow_XZfrac + 27.27(±4.31) Â Shadow_YZfrac À 0.335(±0.051) Â (in blue) or unfavorable (red) for the VAChT affinity is shown in Dihedral Energy. Figure 1 with the superposition of both stereoisomers. As reported by the 3D QSAR modeling shown that substitutions near the 5-po- 2 2 2 r = 0.9591 r (adj) = 0.9223 r (pred) = 0.8523 RMS residual Friedman sition should increase the affinity for VAChT (as it was previously error = 0.3716 L.O.F. = 0.6274 suggested by Szymoszek et al.26) as well as an electropositive sub- stituent (near the 8-position) as shown in Figure 2. A good ligand
Predicted affinity: pKi(À)-5-FBVMpred = 8.64. should have strong Van der Waals attraction in the green area The major contributing factors are Shadow_XZfrac > Jurs_ and a polar group in the blue electrostatic potential area. Figure RPSA>Shadow_YZfrac>Shadow_nu>RadofGyration Jurs_PP- 3 shows the differences obtained for the 3D QSAR model with po- SA_3 > Dihedral Energy > Molecular_SASA E_ADJ_mag. sitive coefficients (in green) on a Van der Waals grid between the
Quadratic model form: pKi(À)pred = 26.21(±3.37) À 5.32 (À)-enantiomers model and the (+)-enantiomers model. Moreover (±0.85)JY Â Shadow_nu À 0.0031(±0.0006)Dipole_Z Â Jurs_PNSA_2 we observed an overlap of a negative yellow area near the 5-posi- À 0.0053(±0.001) Jurs_DPSA_3 Â Dihedral Energy + 0.145(±0.02) tion for (+)-enantiomer FBVM in the favorable positive green area Jurs_PPSA_3 Â Shadow_nu. of (À)-enantiomers model, this could explain partially the stereo- specificity of VAChT binding site. With this good 3D model
(N = 32) pKi(À)-5-FBVM = 9.72 and pKi(+)-5-FBVM = 8.24 with r = 0.922 2 r2 = 0.8346 r2 (adj) = 0.7904 r2 (pred) = 0.7359 RMS residual Friedman and r = 0.851 could be predicted, corresponding to the experimen- error = 0.6105 L.O.F. = 0.7662 tal finding reported in Table 3 (pKi = 8.43 and 8.16, respectively). A third QSAR study was performed based on Bayesian modeling Predicted affinity: pKi(À)-5-FBVMpred = 10.26. (Discovery StudioÒ 2.5, Accelrys) which distinguishes ‘active’ 7662 M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
Table 2 Description of the molecular properties used as descriptors in QSAR studies
Property Description Molecular_SASA 2D surface area Total solvent accessible surface area CHI-2 2D topological Unmodified molecular connectivity indices. This type of emphasizes different aspects of atom connectivity within a descriptors molecule—the amount of branching, ring, structures present and flexibility CHI-V_3_P 2D topological descriptors E_ADJ_mag 2D topological descriptors JX 2D topological Highly discriminating descriptor. This Balaban indices characterize the shape of a molecule taking into account descriptors electronegativity of the atoms of the model JY 2D topological Highly discriminating descriptor. This Balaban indices characterize the shape of a molecule taking into account relative descriptors covalent radius of the atoms of the model Shadow XZ frac Spatial Projection of molecular surface on a plan (XZ) Shadow XY frac Spatial Projection of molecular surface on a plan (XY) Shadow YZ frac Spatial Projection of molecular surface on a plan (YZ) Shadow_nu Spatial Ratio of largest to smallest dimension of projection of molecular surface on a plan Dipole Z 3D electronic Dipole moment descriptors Jurs_PNSA_2 Total charge weighted negative surface area Jurs_PPSA_3 Atomic charge weighted positive surface area Jurs_RPSA Relative polar surface area: total polar surface area divided by the total molecular solvent-accessible surface area RadofGyration 3D molecular properties Angle energy Molecular Conformational properties Bond energy Molecular Conformational properties Electrostatic Molecular energy properties Dihedral energy Molecular properties Van der Waals Molecular Conformational energy properties
Figure 1. Isosurface of the 3D QSAR model coefficients on Electrostatic Potential grids with positive (in blue) and negative (in red) coefficients for the aligned Figure 2. Isosurface of the 3D QSAR model positive coefficients on Van der Waals molecular structures of 32 (+)/(À)-enantiomers I and II in solid representation (A, C, grid (in green) and Electrostatic Potential grid (in blue) for (+)/(À)-enantiomer and D visualize cycles as precise in Table 1). model in solid representation with (À)-5-FEOBV.
ligands from baseline ligands. The Bayesian statistics assign the distinguish these two enantiomers, they were both predicted as probability for each individual descriptor of a molecule to be a ‘active’ with a probability of 82%. From the listing scores obtained member of an ‘active’ class. From the data set of 32 known ligands in this Bayesian model, the closest compound of FBVM was 30 (Table 1), 81% were defined as ‘active’ or ‘inactive’, with a cutoff va- (À)oMV, which presented a Ki = 6.7 nM, that is, was close to the lue at a pKi = 7.69 (Ki <25 nM). From this effective model we tested experimental findings (Ki(À)-5-FBVM = 3.7 nM and Ki(+)-5-FBVM = (À) and (+)-5-FBVM to predict their activity. The results did not 6.9 nM, Table 3). M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667 7663
Table 3
Affinities and selectivity of FEOBV, (rac)-5-IBVM, (rac)-5-ABV, and 5-FBVM for rVAChT, r1 and r2 (Ki values are given in means ± SD)
rVAChT Ki values, in nM Ratio of Ki values
r1 r2 r1/VAChT r2/VAChT (À)-FEOBV 19.6 ± 1.1a 209 ± 94a n.d. 10.7 n.c. (+)-FEOBV 56.9 ± 4.8a 269 ± 37a n.d. 4.7 n.c. (rac)-5-IBVM 15.8 ± 5.42 266 ± 113a n.d 17 n.c. (rac)-5-ABV 40.6 ± 1.69 652 ± 60 n.d 16 n.c. (rac)-5-FBVM 10.9 ± 2.78 13.2 ± 5.84 229 ± 60a 1.21 21 b (+)-(S,S)-5-FBVM 6.95 ± 0.62 (pKi = 8.16) 38.1 ± 4.85 526 5.48 76 a (À)-(R,R)-5-FBVM 3.68 ± 0.48 (pKi = 8.43) 3.57 ± 0.86 252 ± 13 0.97 68
3 Ki values were determined by competition of derivatives against bound (À)-[ H]vesamicol under equilibrium at 22 °C, unless stated. n.d.: not determinated. n.c.: not calculated. All experiments were performed in triplicate (n P 3; an =2;bn = 1).
The Balz-Schiemann reaction (1927) is a well known process of
deamine fluorination of aniline via the SN1 mechanism by the ac- tion of sodium nitrite (NaNO2), followed by thermal decomposition with fluoroboric acid (HBF4). The mechanism of halogeno-dediazo- tization may take place via an ionic pathway (heterolytic decom- position)32,33 or via a radical pathway.34 An alkyl nitrite such as n-butylnitrite or t-butylnitrile, that is, soluble in organic solvents could be a good diazotizating agent according to the literature.35,36 Moreover they are very weak electrophiles (even weaker than the nitrosonium ion). The solvent, which is one of the key parameters for a successful reaction, should be a chlorinated organic solvent because it is aprotic and presents non-nucleophilic and non-oxi- dizing properties. Ionic liquid solvents are known to improve the yield of the Balz-Schiemann reaction.37,38 Because of the high re- À dox potential of fluoride anion (e.g., F2/F in CH3CN E = 2.4 V), the formation of a fluoroaryl bond proceeds generally via a hetero- lytic mechanism.39 Furthermore, the redox potential could be in- creased by solvation.39 The source of the fluorine atom is
variable: silicium fluoride (SiF4) is fairly toxic; alkali metal fluo- rides (CsF or KF) are hygroscopic with poor solubility; and quater- nary ammonium fluorides are very hygroscopic but are soluble in aprotic solvents. Other salts such as boron trifluoride-diethyl 35 etherate (BF3ÁEt2O) are excellent fluorinating agents and are sol- Figure 3. Isosurface of the 3D QSAR model positive coefficients on Van der Waals uble in chlorinated solvents in contrast to nitrosonium tetrafluoro- grids in green and negative coefficients on Van der Waals grid in yellow: in solid borate (NOBF ) or nitrosonium hexafluorophosphate (NOPF ). The representation for (+)-enantiomers model, in quad mesh representation for (À)- 4 6 enantiomers model with (+) and (À)-5-FBVM. decomposition temperature of the diazonium group depends on further substituents on the aromatic ring.40 We tested several fluorinating agents and reaction conditions
From these three QSAR studies, 5-FBVM is predicted to be a li- such as (i) 1.5 equiv of NOBF4,CH2Cl2, 1 h reflux; (ii) 1.5 equiv of gand with nanomolar VAChT binding affinity, without significant NaBF4, 1.5 equiv t-BuONO, CH2Cl2, 1 h, rt; (iii) 1.5 equiv of differences between the two enantiomers. (t-Bu)4NF, 1.5 equiv t-BuONO, CH2Cl2, 1 h, reflux; and (iv) 1.5 equiv of BF3ÁEt2O, 1.2 equiv t-BuONO, C6H4Cl2, 1 h, 60 °C. Scheme 1 2.2. Chemistry presents the best results of one pot fluoro-dediazoniation of the pri- mary aromatic amine 5-ABV to obtained 5-FBVM in moderate yield Synthesis of the 5-ABV enantiomers which were the basis for all (around 25% yield). The chlorinated organic solvents dichloroben- 5,27,31 35 fluorinations has already been described. zene (method A) or CH2Cl2 (method B) were suitable for such
Method A ; Yield 16% HO HO 1.5 eq. BF3.Et2O, 1.2 eq. t-BuONO,C 6H4Cl2 105°C, 1h N N
Method B ; Yield 25% H2N F 5-ABV 1) 1.5 eq. BF3.Et2O, 1.5 eq. t-BuONO, 5-FBVM CH2Cl2 , RT, overnight 2) reflux, 1.5h
Scheme 1. Synthesis of 5-FBVM from 5-ABV by one pot fluoro-dediazotization. 7664 M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667 one pot fluoro-dediazoniation. Increasing of the amount of nitrite or and non-nucleophilic properties. However, we observed the forma- fluorine agent (2 equiv) did not increase the yield (12% yield). tion of the corresponding aryl-p-toluenesulfonic ester analog. In
Since this method is not useable for radiolabeling, we developed contrast, the reaction with triflic acid in CH2Cl2 was more success- a further procedure to introduce a fluorine atom, in particular for ful, and we obtained 5-FBVM with 25% yield (Scheme 3). A chlori- 18 future F-radiofluorination. Aryl-dialkyltriazenes (as a protected nated organic solvent, especially CH2Cl2 (compared to C6H4Cl2), form of aryldiazonium ions) have been extensively studied for dif- was a good solvent, possibly due to the potential stabilization of ferent purposes (for review, see Ref. 41). We synthesized triazene the diazonium-boron trifluoride complex intermediate. We also 5-TBV precursors using two methods: the typical method via a dia- tested several fluorinating agents, including tetra-n-butylammo- zonium salt quenched with diethylamine to provide 5-TBV in nium fluoride (TBAF), which is soluble in chlorinated solvents. quantitative yield (method A, Scheme 2), and method B started Reaction of TBAF, CH2Cl2, and CF3SO3H under reflux was not suc- by the synthesis of a diazonium salt obtained by using t-butylni- cessful (data not shown). The use of CsF in carbon tetrachloride 49 trite in CH2Cl2 then quenched with diethylamine to provide 88% with triflic acid, as previously described for dimethyltriazene, yield of 5-TBV. Triazenes can be decomposed by acidic-thermal- failed in our case (data not shown), possibly due to the insolubility 42,43 decomposition through diazonium ion to provide fluorinated of CsF and triflic acid in CCl4. Using KF in C2H4Cl2 at reflux was also or iodinated44 derivatives as described for certain radiolabelings. not successful. Two pathways are possible: photoinduced decomposition of 1- The enantiomeric purity and the optical rotation of (S,S)- and aryl-3,3-dialkyltriazenes or thermal decomposition via ionic (R,R)-5-FBVM were checked by chiral HPLC by using an amylose 39 paths. Both processes start with cleavage of the N(2)–N(3) based column in RP mode (91% CH3CN/20 mM NH4OAc aq) and a bond.34 Most studies agree on strong acid-catalyzed decomposi- chiral detector. Under these conditions, (S,S)-5-FBVM was a (+)- tion of triazenes, involving fast and reversible protonation of enantiomer and (R,R)-5-FBVM a (À)-enantiomer. N(3), followed by the ‘slow’ heterolytic cleavage of the N(2)–N(3) bond to yield the corresponding diazonium ion and amine.39,45–47 2.3. In vitro evaluation Protonation of N(3) is a crucial step to decompose the triazeno moiety (competing with N(1) protonation): the partial atom Vesamicol-derived ligands are generally insufficiently selective charges of N(1) and N(3) for 5-TBV were À0.138 and À0.0912, towards VAChT due to a non-negligible affinity to r receptors. respectively (Discovery StudioÒ 2.5, CFF forcefield, Accelrys Inc), Since these receptors are distributed in cholinergic brain areas, corresponding to the dipolar charge distribution of the triazene specific imaging of cholinergic deficiency is almost impossible with functional group. unselective compounds. Because the triazene moiety is decomposed through a diazo- Binding affinities of 5-FBVM and reference compounds to nium ion, the theoretical model for acid-catalyzed thermal decom- VAChT and r receptor sites were determined in vitro with Ki val- position to yield the desired arylfluoride (fluoro-de-triazenation) is ues presented in Table 3. 5-FBVM presents very good affinity for the same as for fluoro-dediazoniation.48 This acid should not pres- VAChT and, surprisingly, the (+)-(S,S) enantiomer showed almost À ent a nucleophilic conjugated base (A ) to prevent competition the same affinity as the (À)-(R,R) enantiomer (Ki(S,S)FBVM = 6.95 nM with fluoride anion for aryl cation. Additionally, an acid with a and Ki(R,R)FBVM = 3.68 nM), the affinity being better than for ABV 50 low redox potential (e.g., trifluoroacetic acid) is desirable, to pre- (Ki = 6.95 nM), that is, known to be a good ligand. Furthermore, vent reduction of the aryldiazonium ion via the radical pathway (+)-(S,S)-5-FBVM is selective for VAChT towards the r1 receptor and the formation of radicals. (Ki(S,S) = 38.1 nM and Ki(R,R) = 3.75 nM), and both isomers are selec- For the synthesis of (±)-5-FBVM from the aryl-dialkyltriazene tive for VAChT towards r2 receptors (Ki >200 nM). The affinity of 5- (±)-5-TBV we first tested TBFA in C6H4Cl2 or in CH2Cl2 at reflux FBVM is in the same range as the affinity of 5-IBVM which is cur- and p-toluenesulfonic acid (PTSA) with a high redox potential rently regarded as the standard radiotracer.
Method A; Yield 95%
1) 1.5 eq. NaNO2,HClaq,0°C,0.5h HO HO 2) NaHCO3 aq,1.5eq.Et2NH, 20°C, 1h N N
Method B ;Yield88% H2N Et2N-N=N
1) 1.5 eq. BF3.Et2O, CH 2Cl2 , 5-ABV 1.5 eq. t-BuONO, 0°C, 2 h 5-TBV
2) NaHCO3 aq, 1.5 eq. Et2NH, 20°C, 1h
Scheme 2. Synthesis of (S,S) and (R,R)-5-TBV.
HO HO
N 1) 3 eq. CF3SO3,CH2Cl2 N
2) 1.5 eq. BF3.Et2O, 40°C, 1h Et2N-N=N 25% F (rac)-5-TBV (rac)-5-FBVM
Scheme 3. Synthesis of (rac)-5-FBVM from triazene precursor (rac)-5-TBV. M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667 7665
3. Conclusion 4.2.15-Fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4-tetrahydro- naphthalen-2-ol: 5-FBVM from 5-ABV We obtained a good linear GFA model and a 3D QSAR model 4.2.1.1 Method A. which confirmed the spatial impact on affinity for VAChT via steric To a cold (0 °C) solution of (rac)-5-ABV (0.161 mg, 0.5 mmol) in descriptors and the Van der Waals coefficient. Each study predicted anhydrous 1,2-dichlorobenzene (4 mL), boron trifluoride diethyl a good affinity of both 5-FBVM enantiomers. etherate (BF3ÁO(CH2CH3)2) (0.064 mL, 0.75 mmol) was added. The All one pot acid-catalyzed thermal fluoro-dediazoniation reac- stirred reaction mixture was warmed to 105 °C and n-butylnitrite tions confirmed that parameters such as good solubility, non- (0.07 mL, 0.6 mmol) was added. The mixture was warmed for nucleophilicity, and high redox potential of the reagents used, such 1 h. After cooling to room temperature, the solution was quenched as triflic acid, are the most important conditions for successful flu- with water and extracted with EtOAc. The water phase was made oro-detriazenation via the formation of diazonium ion. Using bor- basic by Na2CO3 and was extracted once again with EtOAc. The on trifluoride etherate we succeeded in introducing fluorine on an combined organic extracts were dried over MgSO4 and concen- aromatic nucleus. Dichloromethane proved to be the most oppor- trated under reduced pressure. The crude product was purified tune solvent for one pot fluoro-dediazoniation. The in vitro evalu- by gradient flash chromatography (Al2O3, n-hexane/EtOAc 4/1 to ations confirmed the QSAR model where both enantiomers n-hexane/EtOAc 1/1). (rac)-5-FBVM was obtained as a white pow- exhibited high affinity for VAChT [(+)-5-FBVM: Ki = 6.95 nM and der in 16% yield. (À)-5-FBVM: Ki = 3.68 nM]. The stereoisomers were selective to- wards r2 receptors (70-fold), however, only (+)-5-FBVM is also 4.2.1.2. Method B. To a cold (0 °C) solution of (rac)-5-ABV selective for r1 receptors (fivefold). Further experiments are (0.161 mg, 0.5 mmol) in anhydrous dichloromethane (4 mL) boron needed to improve the characterization of the pharmacological trifluoride diethyl etherate (BF3ÁO(CH2CH3)2) (0.064 mL, and pharmacokinetics profiles of this compound in order to deter- 0.75 mmol) and n-butylnitrite (0.095 mL, 0.75 mmol) were added. mine its potential use as an F-18-labeled imaging agent for studies The mixture was stirred at room temperature overnight and after- involving the cholinergic system. ward heated to 60 °C for drying. After cooling, water was added to this reaction residue. The water phase was made basic by NaHCO3 and was extracted with EtOAc for two times. The combined organic 4. Experimental section extracts were dried over MgSO4 and concentrated under reduced pressure. The crude product was purified by gradient flash chroma- 4.1. Molecular modeling tography (SiO2, n-hexane/EtOAc 4/1 to n-hexane/EtOAc 1/2). (rac)- 5-FBVM was obtained as a white powder in 25% yield. Computational results were obtained using software programs (+)-5-FBVM and (À)-5-FBVM were obtained by the same proce- from Accelrys Software Inc. The molecules were built and mini- dure as white powders in 27% and 25% yield, respectively, starting mized in molecular package (Discovery StudioÒ 2.5.5, Accelrys, from the corresponding (+)-5-ABV and (À)-5-ABV. San Diego, CA) by CHARMm with CFF partial charge estimation 1 H NMR (CDCl3): d 1.75–1.96 (m, 4H, 4H-10), 2.53–3.08 (m, 9H, method. The GFA model in QSAR protocol was used with a popula- 1H-11, 1H-1, 2H-4, 1H-3, 4H-9), 3.25 (dd, J = 5.4 Hz, J = 16 Hz, 1H- tion size of 100 and 5000 maximum generations. All the parame- 1), 3.85–3.99 (m, 1H-2), 4.34 (br s, OH), 6.86–6.97 (m, 1HAr), 7.10– ters have been left to the system defaults. Two model forms have 7.17 (m, 2HAr), 7.26–7.41 (m, 5HAr). been used: linear or full quadratic. For the 3D QSAR model, the grid 13 C NMR (CDCl3): d 19.0 (C-4), 33.7, 34.2 (2C-10), 37.6 (C-1), spacing was 1 Å. 2 42.8 (C-11), 44.8, 53.4 (2C-9), 65.1 (C-2), 65.9 (C-3), 112.1 ( JC–F = 2 4 22 Hz, C-6), 122.5 ( JC–F = 18 Hz, C-4a), 124.5 ( JC–F = 3 Hz, C-8), 3 4.2. Chemistry 126.2 (CHAr), 126.7 (2CHAr), 127.1 ( JC–F = 8 Hz, C-7), 128.4 (2CHAr), 3 1 136.5 ( JC–F = 8 Hz, C-8a), 145.9 (CAr), 160.9 ( JC–F = 243 Hz, C-5). NMR spectra were recorded on a Bruker DPX Avance 200 spec- MS: m/z = 325 (14), 228 (74), 174 (35), 161 (18), 155 (16), 146 1 13 trometer (200 MHz for H, 50.3 MHz for C). CDCl3 was used as (100), 133 (32), 56 (31). solvent; chemical shifts are expressed in ppm relative to TMS as Anal. Calcd for rac-C21H24FNO: C, 77.51; H, 7.43. Found: C, an internal standard. Mass spectra were obtained on a CG–MS 77.90; H, 3.91. Hewlett Packard 5989A spectrometer (electronic impact at 70 eV). The thin-layer chromatographic (TLC) analyzes were per- 4.2.25-(3,3-Diethyltriaz-1-enyl)-3-(4-phenylpiperidin-1-yl)- formed using Merck 60-F254 silica gel plates. Flash chromatography 1,2,3,4-tetrahydronaphthalen-2-ol: (rac)-5-ABV-diethyltriazene was used for routine purification of reaction products using silica (5-TBV) gel (230–400 mesh). Visualization was accomplished under UV or 4.2.2.1 Method A. in an iodine chamber. All chemicals and solvents were of commer- (rac)-5-ABV (256 mg, 0.8 mmol) was dissolved in 0.1 mL of 12 N cial quality and were purified following standard procedures. Ele- HCl and the flask was immersed in an ice-bath (0–5 °C). A 1.5 equiv mental analyzes of new compounds were within ±0.5% of of NaNO2 (82 mg) was added to the solution and the reaction mix- theoretical values. ture was stirred for 30 min. After neutralization with saturated
4-Phenylpiperidine, used in the synthesis of 5- and 8-amino- aqueous solution of NaHCO3, 1.5 equiv of diethyl amine (0.12 mL) benzovesamicol (ABV),27 was obtained by alkaline fusion from 4- was added to react for 1 h. The mixture was extracted three times 31 cyano-4-phenylpiperidine. Enantiomeric resolution of 5-ABV with CH2Cl2. Combined organic extracts were dried over MgSO4 was done by chromatography separation of diastereomeric N,O- and concentrated under reduced pressure. The triazene (rac)-5- bis-(À)-a-methoxy-a-trifluoromethylphenylacetyl (MTPA) deriva- TBV was obtained as a white powder and was pure enough to be tives followed by hydrolysis of Mosher esters.5 The optical purity used without any purification (95% yield). of both (À)-ABV and (+)-ABV was checked on Chiracel OD column (+)-5-TBV and (À)-5-TBV were obtained by the same procedure (4.6 Â 250 nm, 10 lm particle, Daicel Chemical Industries Ltd, Ill- as white powders in 92% or 96% yield, respectively, starting from kirch France) with n-hexane/isopropanol (80/20) as eluent at a flow the corresponding (+)-5-ABV and (À)-5-ABV. rate of 1.5 mL/min ((+)-ABV; tR = 11 min, (À)-ABV); tR = 13.5 min). (À)-5-IBVM18,27 and (À)-FEOBV5 were synthesized as previ- 4.2.2.2. Method B. (rac)-5-ABV (256 mg, 0.8 mmol) was dis- ously described. solved in CH2Cl2 (2 mL) and the flask was immersed in an ice-bath 7666 M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667
(0–5 °C). A 1.5 equiv of boron trifluoride diethyl etherate (0.15 mL) 7.4 at room temperature. Incubation was terminated after 60 min and 1.5 equiv of t-butylnitrite (0.14 mL) were added to the stirred by filtration (GF-B filter, pre-incubated in 0.3% PEI at room temper- solution. After 2 h, the reaction mixture was neutralized with sat- ature for 90 min; Brandel Cell harvester). Non-specific binding was urated aqueous solution of NaHCO3 and 1.5 equiv of diethyl amine determined in the presence of 10 mM (±)-vesamicol. (0.12 mL) was added and stirred for 1 h. The mixture was r1 affinity was determined by radioligand displacement studies quenched with water and extracted three times with CH2Cl2. Com- on homogenates of rat cortical membranes by using (À)-[3H]pen- bined organic extracts were dried over MgSO4 and concentrated tazocine (Perkin Elmer; specific activity: 1070 GBq/mmol). Assays under reduced pressure. The triazene (rac)-5-TBV was pure enough were incubated in 50 mM TRIS–HCl, pH 7.4 at room temperature. to be used without any purification (88% yield). Incubation was terminated after 120 min by filtration (GF-B filter, 1 H NMR (CDCl3): d 1.34 (t, J = 7 Hz, 6H, 2CH3), 1.76–2.01 (m, 4H, pre-incubated in 0.3% PEI at room temperature for 90 min; Brandel 2H-10, 2H-12), 2.48–3.08 (m, 8H, 1H-11, H-1, 2H-4, 2H-9, 2H-13), Cell harvester). Non-specific binding was determined in the pres- 3.37 (dd, J = 5.6 Hz, J = 16 Hz, 1H-1), 3.45 (dd, J = 3.4 Hz, J = 15.5 Hz, ence of 10 lM haloperidol.
1H-1), 3.83 (q, J = 7 Hz, 4H, 2CH2), 3.91–3.97 (m, 1H-2), 6.95 (d, r2 affinity was determined by radioligand displacement studies J = 6.5 Hz, 1HAr), 7.15–7.37 (m, 7HAr). on homogenates of rat liver membranes by using (À)-[3H]DTG 13 C NMR (CDCl3): d 12.6 (br s, 2CH3); 21.5 (C-4), 33.8, 34.3 (2C- (Perkin Elmer; specific activity: 1147 GBq/mmol) in the presence 3 10), 38.1 (C-1), 42.8 (C-11), 45.0, 53.5 (2C-9), 65.4 (C-2), 66.7 (C-3), of 1 lM dextrallorphan (Roche) to block r1 binding of [ H]DTG. As- 114.0 (CHAr), 125.9, 126.1, 126.3 (3CHAr), 126.7 (2CHAr), 128.4 says were incubated in 50 mM TRIS–HCl, pH 7.4 at room tempera- (2CHAr), 129.4 (CAr), 134.6 (CAr), 146.1 (CAr), 148.9 (CAr). ture. Incubation was terminated after 120 min by filtration (GF-B MS: m/z = 307 (84), 174 (100), 129 (26), 117 (44), 115 (41), 91 filter, pre-incubated in 0.3% PEI at room temperature for 90 min; (49), 70 (33). Brandel Cell harvester). Non-specific binding was determined in the presence of 10 lM haloperidol. 4.2.3. (rac)-5-Fluoro-3-(4-phenyl-piperidin-1-yl)-1,2,3,4- All assays were performed in triplicates at least three times. The tetrahydro-naphthalen-2-ol: (rac)-5-FBVM from 5-triazene 5- IC50-values were estimated by computational non-linear regres- TBV sion analysis. Ki-Values were calculated according to Cheng and To a solution of (rac)-5-TBV (0.162 g, 0.4 mmol) in anhydrous Prusoff.51
CH2Cl2 (5 mL) trifluoromethanesulfonic acid monohydrate (CF3SO3HÁH2O) (0.18 mg, 1.2 mmol) dissolved in CH3CN (0.5 mL) Acknowledgment was added. BF3ÁEt2O (boron trifluoride diethyl etherate, 0.075 mL, 0.6 mmol) diluted in CH2Cl2 (0.2 mL) was then added to the stirred This work was supported by INSERM. This study was funded in reaction mixture. The reaction mixture was heated to 60 °C for dry- part by FEDER: ImAD project. We thank the ‘Département d’analy- ing over 1 h. After cooling, water was added to this reaction resi- ses Chimiques et S.R.M. biologique et médicale’ (Tours, France) for due. The water phase was made basic by NaHCO3 and was chemical analyzes. extracted with CH2Cl2. The combined organic extracts were dried over MgSO4 and concentrated under reduced pressure. The crude References and notes product was purified by gradient flash chromatography (SiO2, n- hexane/EtOAc 4/1 to n-hexane/ EtOAc 1/2). (rac)-5-FBVM was ob- 1. Bauer, J.; Hull, M.; Berger, M. Z. Gerontol. Geriatr. 1995, 28, 155. tained as a white powder in 25% yield. 2. Terry, R. D.; Masliah, E.; Salmon, D. P.; Butters, N.; DeTeresa, R.; Hill, R.; Hansen, L. A.; Katzman, R. Ann. Neurol. 1991, 30, 572. 3. Giboureau, N.; Emond, P.; Fulton, R. R.; Henderson, D. J.; Chalon, S.; Garreau, L.; 4.3. Determination of optical rotation of (R,R)-5-FBVM and Roselt, P.; Eberl, S.; Mavel, S.; Bodard, S.; Fulham, M. J.; Guilloteau, D.; Kassiou, (S,S)-5-FBVM M. Synapse 2007, 61, 962. 4. Kilbourn, M. R.; Hockley, B.; Lee, L.; Sherman, P.; Quesada, C.; Frey, K. A.; Koeppe, R. A. Nucl. Med. Biol. 2009, 36, 489. The analytical separation of the 5-FBVM enantiomers by chiral 5. Mulholland, G. K.; Jung, Y. W.; Wieland, D. M.; Kilbourn, M. R.; Kuhl, D. E. J. J. HPLC was performed on a Reprosil Chiral-AM-RP column Labelled Compd. Radiopharm. 1993, 33, 583. (250 Â 4.6 mm), which is based on amylose-tris-(3,5-dimethyl- 6. Rogers, G. A.; Stone-Elander, S.; Ingvar, M.; Eriksson, L.; Parsons, S. M.; Widén, L. Nucl. Med. Biol. 1994, 21, 219. phenyl)-carbamate as chiral selector (Dr. Maisch-GmbH, Ger- 7. Efange, S. M.; Langason, R. B.; Khare, A. B.; Low, W. C. Life Sci. 1996, 58, many). The optical rotation was determined by using a chiral 1367. detector (OR 2090 model from JASCO, Germany). In general, the 8. Efange, S. M.; Khare, A.; Parsons, S. M.; Bau, R.; Metzenthin, T. J. Med. Chem. 1993, 36, 985. OR detector operated under following conditions: range: 0.05, re- 9. Sorger, D.; Scheunemann, M.; Vercouillie, J.; Grossmann, U.; Fischer, S.; Hiller, sponse: SLOW, gain: 10. The polarity of signal amplitudes obtained A.; Wenzel, B.; Roghani, A.; Schliebs, R.; Steinbach, J.; Brust, P.; Sabri, O. Nucl. by the chiral detector was checked with (À)-vesamicol as reference Med. Biol. 2009, 36, 17. 10. Sorger, D.; Scheunemann, M.; Großmann, U.; Fischer, S.; Vercouille, J.; Hiller, compound. By using 91% CH3CN/20 mM NH4OAc aq and a flow rate A.; Wenzel, B.; Roghani, A.; Schliebs, R.; Brust, P.; Sabri, O.; Steinbach, J. Nucl. of 1 mL/min the enantiomers were separated. With tR = 15.5 min Med. Biol. 2008, 35, 185. the (À)-(R,R)-5-FBVM eluted in front of the (+)-(S,S)-5-FBVM with 11. Parsons, S. M.; Bahr, B. A.; Rogers, G. A.; Clarkson, E. D.; Noremberg, K.; Hicks, B. W. Prog. Brain Res. 1993, 98, 175. tR = 23.5 min. 12. Bahr, B. A.; Clarkson, E. D.; Rogers, G. A.; Noremberg, K.; Parsons, S. M. Biochemistry 1992, 31, 5752. 4.4. Receptor binding studies 13. Altar, C. A.; Marien, M. R. Synapse 1988, 2, 486. 14. Bahr, B. A.; Parsons, S. M. Proc. Natl. Acad. Sci. U.S.A. 1986, 83, 2267. 15. Efange, S. M.; Mach, R. H.; Smith, C. R.; Khare, A. B.; Foulon, C.; Akella, S. K.; For radioligand displacement studies, the test compounds were Childers, S. R.; Parsons, S. M. Biochem. Pharmacol. 1995, 49, 791. solved in DMSO at 10 mM stock solutions. Serial dilutions in the 16. Ichikawa, T.; Ajiki, K.; Matsuura, J.; Misawa, H. J. Chem. Neuroanat. 1997, 13, range of 0.01 nM to 1 lM were obtained by further dilution in 23. 17. Phan, V. L.; Miyamoto, Y.; Nabeshima, T.; Maurice, T. J. Neurosci. Res. 2005, 79, incubation buffer. 561. VAChT affinity was determined by radioligand displacement 18. Jung, Y. W.; Van Dort, M. E.; Gildersleeve, D. L.; Wieland, D. M. J. Med. Chem. studies on homogenates of PC12 cells stably transfected with 1990, 33, 2065. 19. Kuhl, D. E.; Minoshima, S.; Fessler, J. A.; Frey, K. A.; Foster, N. L.; Ficaro, E. P.; rVAChT (Ali Roghani, Texas Tech University, Lubbock, TX, USA) Wieland, D. M.; Koeppe, R. A. Ann. Neurol. 1996, 40, 399. by using (À)-[3H]vesamicol (Perkin Elmer; specific activity: 20. Sorger, D.; Schliebs, R.; Kampfer, I.; Rossner, S.; Heinicke, J.; Dannenberg, C.; 1296 GBq/mmol). Assays were incubated in 50 mM TRIS–HCl, pH Georgi, P. Nucl. Med. Biol. 2000, 27, 23. M. Kovac et al. / Bioorg. Med. Chem. 18 (2010) 7659–7667 7667
21. Zea-Ponce, Y.; Mavel, S.; Assaad, T.; Kruse, S. E.; Parsons, S. M.; Emond, P.; Chalon, 34. Galli, C. Chem. Rev. 1988, 88, 765. S.; Giboureau, N.; Kassiou, M.; Guilloteau, D. Bioorg. Med. Chem. 2005, 13, 745. 35. Garel, L.; Saint-Jalmes, L. Tetrahedron Lett. 2006, 47, 5705. 22. Jung, Y. W.; Frey, K. A.; Mulholland, G. K.; del Rosario, R.; Sherman, P. S.; Raffel, 36. Doyle, M. P.; Bryker, W. J. J. Org. Chem. 1979, 44, 1572. D. M.; Van Dort, M. E.; Kuhl, D. E.; Gildersleeve, D. L.; Wieland, D. M. J. Med. 37. Laali, K. K.; Gettwert, V. J. J. Fluorine Chem. 2001, 107, 31. Chem. 1996, 39, 3331. 38. Hubbard, A.; Okazaki, T.; Laali, K. K. J. Org. Chem. 2008, 73, 316. 23. Assaad, T.; Mavel, S.; Emond, P.; Chalon, S.; Drossard, M. L.; Vergote, J.; Bodard, 39. Satyamurthy, N.; Barrio, J. R.; Schmidt, D. G.; Kammerer, C.; Bida, G. T.; Phelps, S.; Allouchi, H.; Besnard, J.-C.; Guilloteau, D. J. Labelled Compd. Radiopharm. M. E. J. Org. Chem. 1990, 55, 4560. 2007, 50, 139. 40. Yoneda, N.; Fukuhara, T. Tetrahedron 1996, 52, 23. 24. Assaad, T.; Mavel, S.; Parsons, S. M.; Kruse, S.; Galineau, L.; Allouchi, H.; 41. Kimball, D. B.; Haley, M. M. Angew. Chem., Int. Ed. 2002, 41, 3338. Kassiou, M.; Chalon, S.; Guilloteau, D.; Emond, P. Bioorg. Med. Chem. Lett. 2006, 42. Berridge, M. S.; Sayre, L. M.; Arora, P. K.; Terris, A. H.; Riachi, N. J.; Harik, S. I. J. 16, 2654. Med. Chem. 1993, 36, 1284. 25. Emond, P.; Mavel, S.; Zea-Ponce, Y.; Kassiou, M.; Garreau, L.; Bodard, S.; 43. Pages, T.; Langlois, B. R. J. Fluorine Chem. 2001, 107, 321. Drossard, M. L.; Chalon, S.; Guilloteau, D. Nucl. Med. Biol. 2007, 34, 967. 44. Harmata, M.; Hong, X.; Zheng, P. Tetrahedron Lett. 2006, 47, 7343. 26. Szymoszek, A.; Wenzel, B.; Scheunemann, M.; Steinbach, J.; Schüürmann, G. J. 45. Schmidt, B. F.; Snyder, E. J.; Carroll, R. M.; Farnsworth, D. W.; Michejda, C. J.; Med. Chem. 2008, 51, 2128. Smith, R. H. J. Org. Chem. 1997, 62, 8660. 27. Rogers, G. A.; Parsons, S. M.; Anderson, D. C.; Nilsson, L. M.; Bahr, B. A.; 46. Smith, R. H.; Wladkowski, B. D.; Mehl, A. F.; Cleveland, M. J.; Rudrow, E. A.; Kornreich, W. D.; Kaufman, R.; Jacobs, R. S.; Kirtman, B. J. Med. Chem. 1989, 32, Chmurny, G. N.; Michejda, C. J. J. Org. Chem. 1989, 54, 1036. 1217. 47. Tewson, T. J.; Welch, M. J. J. Chem. Soc., Chem. Commun. 1979, 1149. 28. Rogers, D.; Hopfinger, A. J. J. Chem. Inf. Comput. Sci. 1994, 34, 854. 48. Barrio, J. R.; Satyamurthy, N.; Ku, H.; Phelps, M. E. J. Chem. Soc., Chem. Commun. 29. Stanton, D. T.; Jurs, P. C. Anal. Chem. 1990, 62, 2323. 1983, 443. 30. Shiba, K.; Ogawa, K.; Ishiwata, K.; Yajima, K.; Mori, H. Bioorg. Med. Chem. 2006, 49. Pages, T.; Langlois, B. R.; Bars, D. L.; Landais, P. J. Fluorine Chem. 2001, 107, 329. 14, 2620. 50. Rogers, G.; Kornreich, W.; Hand, K.; Parsons, S. Mol. Pharmacol. 1993, 44, 633. 31. Berkoff, C. E.; Rivard, D. E.; Kirkpatrick, D.; Ives, J. L. Synth. Commun. 1980, 10, 51. Cheng, Y.; Prusoff, W. H. Biochem. Pharmacol. 1973, 22, 3099. 939. 52. Shiba, K.; Yano, T.; Sato, W.; Mori, H.; Tonami, N. Life Sci. 2002, 71, 1591. 32. Hashida, Y.; Landells, R. G. M.; Lewis, G. E.; Szele, I.; Zollinger, H. J. Am. Chem. 53. Efange, S. M. N.; von Hohenberg, K.; Khare, A. B.; Tu, Z.; Mach, R. H.; Parsons, S. Soc. 1978, 100, 2816. M. Nucl. Med. Biol. 2000, 27, 749. 33. Szele, I.; Zollinger, H. J. Am. Chem. Soc. 1978, 100, 2811. 3.5. Aromatic fluoro-de-triazenation with boron trifluoride diethyl etherate under non protic acid conditions
Mitja Kovac a,b, Marko Anderluh a, Johnny Vercouillie a, Denis Guilloteau a, Patrick Emond a, Sylvie Mavel a a Université François-Rabelais de Tours, INSERM U930, CHRU, Hôpital Bretonneau, Service de Médecine Nucléaire, 37000 Tours, France b University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Aškerčeva 7, 1000 Ljubljana, Slovenia
Published in: Journal of Fluorine Chemistry 147 (2013) 5-9
99
SUMMARY:
A strong protic acid is required to transform triazeno moiety into diazonium group at elevated temperature. Since the phenyl cation intermediate is highly reactive non- discriminating species, and even the conjugate base of the used acid is considered non- nucleophilic (e.g. triflate), fluoro-de-triazenation is often accompanied by the formation of the acid counterion substituted byproduct (Ar-A);173,176 especially if the acid is used in excess.181 To circumvent this limitation, we examined the coordination chemistry of triazene derivatives,182–184 and electron donor-acceptor (EDA) complexes between boron trifluoride and methylated ammonia derivatives.185–189 Accordingly, we propose fluoro-de-triazenation can be successfully accomplished by the only presence of boron trifluoride via complexation at elevated temperature. Our hypothesis was first confirmed on the model precursor, 3,3- diethyl-1-naphthyltriazene and after on simple para substitueted 3,3-diethyl-1-aryltriazenes, by conventional and microwave heating. To prove that the method is applicable to obtain more complex arylfluorides too, 5-FBVM was accomplished in high 72% yield under appropriate microwave conditions in tetrachloromethane as the most opportune solvent.
STATEMENT: I declare, that nobody of co-authors has used the article Aromatic fluoro-de- triazenation with boron trifluoride diethyl etherate under non protic acid conditions, published in the Journal of Fluorine Chemistry, for his/her own thesis.
100
Journal of Fluorine Chemistry 147 (2013) 5–9
Contents lists available at SciVerse ScienceDirect
Journal of Fluorine Chemistry
jo urnal homepage: www.elsevier.com/locate/fluor
Aromatic fluoro-de-triazenation with boron trifluoride diethyl etherate under non-protic acid conditions
a,b b a a
Mitja Kovac , Marko Anderluh , Johnny Vercouillie , Denis Guilloteau ,
a a,
Patrick Emond , Sylvie Mavel *
a
Universite´ Franc¸ois-Rabelais de Tours, INSERM U930, CHRU, Hoˆpital Bretonneau, Service de Me´decine Nucle´aire, 37000 Tours, France
b
University of Ljubljana, Faculty of Pharmacy, Department of Pharmaceutical Chemistry, Asˇkercˇeva 7, 1000 Ljubljana, Slovenia
A R T I C L E I N F O A B S T R A C T
Article history: Fluoro-de-triazenation of 3,3-diethyl-1-aryltriazenes can be achieved by conventional or under
Received 19 October 2012
microwave heating in carbon tetrachloride, in the presence of boron trifluoride diethyl etherate without
Received in revised form 9 January 2013
any protic acid to avoid corresponding unwanted byproduct formation.
Accepted 12 January 2013
ß 2013 Elsevier B.V. All rights reserved.
Available online 21 January 2013
Keywords: Fluorination Triazene Diazonium
Boron trifluoride
Arylfluoride
1. Introduction triazenes [10,14,15]. The 3,3-dialkyl-1-aryltriazenes are
regarded as protected form of aryldiazonium ions and therefore
Most of the pharmacologically active fluorinated drugs are their acid-triggered thermal decomposition parallels that of the
aromatic, bearing a fluoro or a trifluoromethyl substituent [1]. corresponding diazonium ionic reactions. The reactivity of
The efficient regioselective introduction of fluorine in electron- aryltriazenes is well described, especially the competition
rich arenes under mild conditions continues to be a challenge between ionic (heterolytic) and radical (homolytic) de-triazena-
[2,3]. Fluoro-de-triazenation (so called Wallach reaction) repre- tion and de-diazoniation pathways during fluoro-de-diazonia-
sents one of the few regioselective nucleophilic routes yielding tion, mechanism strongly dependents upon the reaction
arylfluorides (Ar-F) [4]. The triazenes have been known for more conditions [4,7–10]. As protic acid is usually required to
than a century and have been studied for their versatility in decompose aryltriazene, and since the formed phenyl cation
organic synthesis, especially after their biological activities were intermediate is highly reactive, fluoro-de-triazenation is often
first reported in the beginning of the 1960s [5,6]. The 3,3- accompanied by the formation of a substantial amount of the
0 00
dialkyl-1-aryltriazenes (Ar–N55N–NR R ) are regarded as pro- acid counterion substituted byproduct (Ar-A) [9,11].
tected form of aryldiazonium ions and therefore their acid- We have been investigating the fluoro-de-triazenation reaction
triggered thermal decomposition parallels that of the corre- as part of a program toward the synthesis of 5-fluoro-3-(4-
sponding diazonium ionic reactions [4,7–10]. The use of 3,3- phenylpiperidin-1-yl)-1,2,3,4-tetrahydro-naphthalen-2-ol (5-
0 00
dialkyl-1-aryltriazenes (Ar–N55N–NR R ) has an essential ad- FBVM) [16]. Therein we have proposed that 3,3-diethyl-1-
vantage over the aryldiazonium ions because of their solubility aryltriazene could be thermally decomposed and subsequently
in a number of anhydrous organic and ionic solvents [4,9,11]. regioselectively fluorinated using boron trifluoride diethyl ethe-
Moreover, they can be safely and readily prepared in moderate rate (BF3 Et2O) as both Lewis acid and fluorinating agent. Thus, the
to high yields [5,12,13]. In parallel, solid-phase methodologies main advantage of the protocol is the avoidance of the competitive
were also applied for the synthesis and reactivity of resin-bound formation of the unwanted compound Ar-A. Using 3,3-diethyl-1-
naphthyltriazene as a model, the influence of different organic
solvents was studied. Furthermore, to demonstrate the versatility
of the method, a fluoro-de-triazenation was successfully per-
* Corresponding author at: Faculte´ de Pharmacie, INSERM U930, 31 Avenue
Monge, 37200 Tours, France. Tel.: +33 2 47 36 72 40; fax: +33 2 47 36 72 24. formed using conventional and microwave heating on several 4-
E-mail address: [email protected] (S. Mavel). substituted 3,3-diethyl-1-aryltriazenes.
0022-1139/$ – see front matter ß 2013 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jfluchem.2013.01.012
6 M. Kovac et al. / Journal of Fluorine Chemistry 147 (2013) 5–9
Table 1
2. Results and discussion
Synthesis of 3,3-diethyl-1-aryltriazenes 2a-g. .
1. HCl 37 %, 2. NaNO aq. , 0 - 5 oC, 45 min
2.1. Preparation of 3,3-diethyl-1-aryltriazenes 2
ArNH2 Ar-N=N-NEt 2
o 3. NHEt ,NaCO aq. , 0 - 5 Ctort,1 h
1a-g 2 2 3 2a-g
All 3,3-diethyl-1-aryltriazenes 2a-g were prepared according to
known procedure [13], involving diazotization of aromatic amines Ar-NH2 Triazene (isolated yield %)
1a-g with sodium nitrite in acidic aqueous medium at 0–5 8C
1a [16] HO 2a (69)
followed by addition of diethylamine to yield the corresponding
triazenes 2a-g (Table 1).
N
2.2. Fluoro-de-triazenation with polyphosphoric acid and boron trifluoride diethyl etherate
H2N
Triflic acid (TfOH) is not an optimal acid for aromatic fluoro-de-
1b 1-Naphthylamine 2b (34)
triazenation. Besides its incompatibility with the acid-sensitive 1c 4-Me-C6H4NH2 2c (76)
functional groups, substantial amounts of aryl triflate (Ar-OTf) is 1d 4-NO2-C6H4NH2 2d (47)
1e 4-O-C4H9-C6H4NH2 2e (53)
usually produced [9,11]. Polyphosphoric acid (PPA) presents a