Comparison of in Silico, Electrochemical, in Vitro and in Vivo Metabolism of a Homologous Series of (Radio)fluorinated σ1Receptor Ligands Designed for Positron Emission Tomography

Article abstract of DOI:10.1002/cmdc.201600366

The imaging of σ₁receptors in the brain by fluorinated radiotracers will be used for the validation of σ₁receptors as drug targets as well as for differential diagnosis of diseases in the central nervous system. The biotransformation of four homologous fluorinated PET tracers 1′-benzyl-3-(ω-fluoromethyl to ω-fluorobutyl)-3H-spiro[2]benzofuran-1,4′-piperidine] ([¹⁸F]1–4) was investigated. In silico studies using fast metabolizer (FAME) software, electrochemical oxidations, in vitro studies with rat liver microsomes, and in vivo metabolism studies after application of the PET tracers [¹⁸F]1–4 to mice were performed. Combined liquid chromatography and mass spectrometry (HPLC–MS) analysis allowed structural identification of non-radioactive metabolites. Radio-HPLC and radio-TLC provided information about the presence of unchanged parent radiotracers and their radiometabolites. Radiometabolites were not found in the brain after application of [¹⁸F]2–4, but liver, plasma, and urine samples contained several radiometabolites. Less than 2 % of the injected dose of [¹⁸F]4 reached the brain, rendering [¹⁸F]4 less appropriate as a PET tracer than [¹⁸F]2 and [¹⁸F]3. Compounds [¹⁸F]2 and [¹⁸F]3 possess the most promising properties for imaging of σ₁receptors in the brain. High σ₁affinity (K_i=0.59 nm), low lipophilicity (logD_7.4=2.57), high brain penetration (4.6 % of injected dose after 30 min), and the absence of radiometabolites in the brain favor the fluoroethyl derivative [¹⁸F]2 slightly over the fluoropropyl derivative [¹⁸F]3 for human use.

Full text of DOI:10.1002/cmdc.201600366

DOI: 10.1002/cmdc.201600366
Full Papers
Comparison of in Silico, Electrochemical, in Vitro and
in Vivo Metabolism of a Homologous Series of
(Radio)fluorinated s1 Receptor Ligands Designed for
Positron Emission Tomography
Christian Wiese,[a] Eva Große Maestrup,[a] Fabian Galla,[a] Dirk Schepmann,[a] Achim Hiller,[b]
Steffen Fischer,[b] Friedrich-Alexander Ludwig,[b] Winnie Deuther-Conrad,[b]
Cornelius K. Donat,[b] Peter Brust,[b] Lars Bꢀter,[c, d] Uwe Karst,[c, d] and Bernhard Wꢀnsch*[a, d]
The imaging of s1 receptors in the brain by fluorinated radio-
tracers will be used for the validation of s1 receptors as drug
targets as well as for differential diagnosis of diseases in the
central nervous system. The biotransformation of four homolo-
gous fluorinated PET tracers 1’-benzyl-3-(w-fluoromethyl to w-
about the presence of unchanged parent radiotracers and
their radiometabolites. Radiometabolites were not found in the
brain after application of [18F]2–4, but liver, plasma, and urine
samples contained several radiometabolites. Less than 2% of
the injected dose of [18F]4 reached the brain, rendering [18F]4
less appropriate as a PET tracer than [18F]2 and [18F]3. Com-
pounds [18F]2 and [18F]3 possess the most promising properties
for imaging of s1 receptors in the brain. High s1 affinity (Ki =
0.59 nm), low lipophilicity (logD7.4 =2.57), high brain penetra-
tion (4.6% of injected dose after 30 min), and the absence of
radiometabolites in the brain favor the fluoroethyl derivative
[18F]2 slightly over the fluoropropyl derivative [18F]3 for human
use.
fluorobutyl)-3H-spiro[2]benzofuran-1,4’-piperidine]
([18F]1–4)
was investigated. In silico studies using fast metabolizer (FAME)
software, electrochemical oxidations, in vitro studies with rat
liver microsomes, and in vivo metabolism studies after applica-
tion of the PET tracers [18F]1–4 to mice were performed. Com-
bined liquid chromatography and mass spectrometry (HPLC–
MS) analysis allowed structural identification of non-radioactive
metabolites. Radio-HPLC and radio-TLC provided information
Introduction
The s receptor was first postulated in 1976 by Martin et al.,
who found that various opioid analgesics exert different phar-
macological profiles.[1] This observation led to the sub-differen-
tiation of the opioid receptor into three opioid receptor sub-
types, which were termed according to their prototypical li-
gands m (morphine), k (ketocyclazocine), and s receptors (SKF-
10,047).[1,2] Because typical opioid antagonists such as nalox-
one and naltrexone are unable to inhibit the effects caused by
s ligands, the s receptor was removed from the class of opioid
receptors.[3] Today s receptors represent a unique class of non-
opioid, non-NMDA, and non-dopamine receptors, which are
classified into s1 and s2 receptor subtypes.[4,5] Herein we focus
on the s1 receptor subtype.
Cloning of the s1 receptor led to a protein with 223 amino
acids and a molecular weight of 25.3 kDa, which is not related
to any other mammalian protein. It shares a 30% homology
with the yeast enzyme sterol-D8/D7-isomerase.[6,7] The s1 recep-
tor is composed of two transmembrane helices, an extracellu-
lar loop (50 residues), a short N-terminal domain (10 residues)
and a long C-terminal domain containing the ligand binding
site. Both the N- and C-terminal domains are located intracellu-
larly.[8] Very recently a 3D homology model of the s1 receptor
was reported showing the 3D organization of the protein and
its interaction with ligands. A combined site-directed mutagen-
esis and theoretical analysis of the receptor protein provided
profound insight into the receptor–ligand interactions at the
molecular level.[9,10]
[a] Dr. C. Wiese, Dr. E. Große Maestrup, Dr. F. Galla, Dr. D. Schepmann,
Prof. Dr. B. Wꢀnsch
Institut fꢀr Pharmazeutische und Medizinische Chemie, Westfꢁlische Wil-
helms-Universitꢁt Mꢀnster, Corrensstraße 48, 48149 Mꢀnster (Germany)
E-mail: wuensch@uni-muenster.de
[b] Dr. A. Hiller, Dr. S. Fischer, Dr. F.-A. Ludwig, Dr. W. Deuther-Conrad,
Dr. C. K. Donat, Prof. Dr. P. Brust
Helmholtz-Zentrum Dresden-Rossendorf, Institut fꢀr Radiopharmazeutische
Krebsforschung, Forschungsstelle Leipzig, Permoserstraße 15, 04318 Leipzig
(Germany)
[c] Dr. L. Bꢀter, Prof. Dr. U. Karst
Institut fꢀr Anorganische und Analytische Chemie, Westfꢁlische Wilhelms-
Universitꢁt Mꢀnster, Corrensstraße 30, 48149 Mꢀnster (Germany)
The s1 receptor is located within the membrane of the en-
doplasmic reticulum, but also within the cell membrane and
the membrane of mitochondria. It controls the Ca2+ transfer
from the endoplasmic reticulum to mitochondria via interac-
tion with the Ca2+ binding chaperone BiP (binding immuno-
globulin protein) and the inositol triphosphate receptor.[11–13]
[d] Dr. L. Bꢀter, Prof. Dr. U. Karst, Prof. Dr. B. Wꢀnsch
Cells-in-Motion Cluster of Excellence (EXC 100–CiM), Westfꢁlische Wilhelms-
Universitꢁt Mꢀnster, 48149 Mꢀnster (Germany)
Supporting information for this article can be found under http://
dx.doi.org/10.1002/cmdc.201600366.
ChemMedChem 2016, 11, 1 – 15
1
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Moreover, it has been shown that s1 receptors are able to reg-
ulate K+ and Na+ channels.[14,15] In addition to the influence on
ion channels, s1 receptors modulate various neurotransmitter
systems including the cholinergic, dopaminergic and glutama-
tergic neurotransmission.[16–18]
alkyl fluorides can be defluorinated by elimination, substitution
or oxidation losing the [18F]label. Moreover, radioactive metab-
olites penetrating the blood–brain barrier (BBB) will lead to dis-
torted images.
In this manuscript the biotransformation of the homologous
fluorinated s1 receptor ligands 1–4 is reported. After prediction
of metabolically labile positions using the software package
Fast Metabolizer (FAME), the products formed by electrochemi-
cal oxidation were analyzed. Then, the in vitro metabolism
during incubation with rat liver microsomes was investigated
and the formed metabolites were compared with the predict-
ed and electrochemically formed metabolites. Finally, the radio-
metabolites formed after injection of the [18F]-labeled tracers
[18F]1–[18F]4 in mice were characterized.
A high density of s1 receptors is not only found in the cen-
tral nervous system (CNS), but also in various tissues of the pe-
riphery, including heart, kidney, liver, endocrine organs, and
placenta. In the CNS, a high density of s1 receptors is found in
the frontal cortex, hypothalamus, hippocampus, and cerebel-
lum regions associated with memory and emotion as well as
motor and sensory control.[19–22]
It has been shown that the s1 receptor is involved in several
psychiatric and neurodegenerative disorders including depres-
sion, psychosis, anxiety, neuropathic pain and Alzheimer’s and
Parkinson’s disease.[22–25] The naphthyl substituted pyrazole de-
rivative S1RA is currently being investigated in phase II clinical
trials of neuropathic pain.[26–28] Due to the important role of
the s1 receptor in the above mentioned psychiatric and neuro-
degenerative disorders, imaging of the s1 receptor in positron
emission tomography (PET) studies is of particular interest. A
PET tracer suitable for imaging of s1 receptors in the brain will
be used for target validation with the aim of developing novel
therapeutic concepts. Moreover, a correlation of s1 receptor
density and activity with a particular disease could be exploit-
ed for diagnosis and prognosis.[29]
Results and Discussion
Prediction of metabolically labile positions by FAME
Data about phase 1 and phase 2 metabolites of a large
number of organic molecules are collected in FAME.[39] Based
on this information, heavy atoms of unknown compounds are
ranked according to the probability of being involved in a bio-
transformation reaction. Values between 0 and 1 are attributed
to the heavy atoms; a value of 0 indicates a low and a value of
1 a high probability of the heavy atom being involved in a bio-
transformation reaction. In this study, heavy atoms showing
a value greater than 0.22 were considered. In Figure 2 the
heavy atoms of 1–4 predicted by FAME as potentially metabol-
ically labile positions are marked with arrows.
Recently, a set of four homologous fluoroalkyl-substituted
spirocyclic s1 receptor ligands 1–4 has been synthesized and
pharmacologically evaluated (Figure 1). Due to the high s1 re-
ceptor affinity (Ki =0.59–1.4 nm) and selectivity against the s2
subtype and other related receptors, the compounds 1–4 were
labeled with [18F]fluoride and investigated as PET tracers for
imaging of s1 receptors in the brain. With respect to organ dis-
tribution and pharmacokinetic properties the fluoroethyl deriv-
ative 2 (fluspidine) represents the most promising radioligand
and is currently used in clinical studies.[30–38]
Figure 2. Metabolically labile positions predicted by FAME using the algo-
rithm for rat phase 1 reactions; black arrows indicate metabolically labile po-
sitions; the dashed arrow indicates a position which was only predicted for
the phenylbutyl derivative 4.
Figure 1. Homologous fluorinated spirocyclic s1 receptor ligands 1–4 devel-
oped for imaging of s1 receptors in the brain.
According to the predictions of FAME the N atom and the
adjacent benzylic C atom have the highest probability for met-
abolic reactions. Additionally, the para position of the benzyl
moiety was identified as metabolically labile. For all homologs,
C atoms of the side chain and the terminal fluorine atom were
marked as metabolically labile positions. Unexpectedly, the
benzene ring of the benzofuran was predicted only for the flu-
orobutyl derivative 4. Here, the C atom at position 5 was pre-
dicted to be biotransformed.
For studies with living organisms the biotransformation of
the respective ligand has to be considered. During the devel-
opment of a PET tracer the metabolism is of particular interest,
since the tracer should have a biological half-life longer than
the period of observation, usually at least 30–60 min. In gener-
al, aliphatic CÀF bonds are rather stable. Nevertheless, primary
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
2
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
Electrochemical oxidation of 1–4
ence of oxidation products resulting from N-debenzylation and
hydroxylation indicates that hydroxylation probably has taken
place at the positions labeled in the respective structures
shown in Figure 4. For the homologous fluoroalkyl derivatives
1–4, very similar oxidation products were detected. However,
in the case of the fluoroethyl derivative 2, the hydroxylated
and N-debenzylated product 2i was not detected, indicating
that further dehydrogenation of 2i to form 2j or dehydrogena-
tion of the corresponding precursors 2a and 2g have already
taken place quantitatively.
The CYP-catalyzed oxidative biotransformation of compounds
1–4 can also be simulated by means of electrochemical oxida-
tion. Subsequent mass spectrometric detection and the ab-
sence of biological matrices allow the identification of reactive
and short-lived species such as radical cations as well as to
evaluate their binding tendency toward selected biomole-
cules.[40–44] To investigate the oxidative biotransformation of
compounds 1–4 electrochemically, a potential ramp was ap-
plied to the electrochemical cell and the detected mass spectra
were plotted depending on the applied oxidation potential.
The obtained mass voltammograms allow an easy identifica-
tion of possible biotransformation products, since electro-
chemically generated oxidation products typically show in-
creasing signal intensities during the potential ramp, while the
signal intensity of the starting compound is decreasing.
A comparison of in silico predicted (FAME) and electrochemi-
cally generated metabolites shows that the N atom and the
adjacent benzylic C atom are the most labile positions. In con-
trast to the in silico prediction, electrochemical oxidation of
the fluoroalkyl side chain was not detected. However, the ben-
zene ring of all four homologs was oxidized electrochemically,
although this biotransformation reaction was predicted only
for the fluorobutyl derivative 4.
In Figure 3, the mass voltammogram of 1 is shown. The
signal intensity of 1 (m/z 312) starts to decrease at a potential
of ~1.7 V. At the same time, several signals appear with in-
creasing intensities. The signal of m/z 328, which can be traced
back to a hydroxylated oxidation product of 1, is already pres-
ent without applying a potential. This can be explained by the
fact that electrochemical oxidations can also occur during the
electrospray ionization process. However, this is the only oxida-
tion product present in the mass spectrum of the initial solu-
tion. Based on the detected exact masses of all indicated oxi-
dation products, sum formulae were determined and chemical
structures were deduced.
In vitro metabolism using rat liver microsomes
Analysis of metabolites formed upon incubation of 1–4 with
rat liver microsomes and NADPH
To analyze the metabolic labile positions, the fluoroalkyl-substi-
tuted s1 ligands 1–4 were incubated with rat liver microsomes
and Nicotinamide adenine dinucleotide phosphate (NADPH).
After precipitation of the proteins with acetonitrile, the super-
natant containing the metabolites was analyzed by HPLC–
HRMS methods.
In Figure 4, the suggested structures of the oxidation prod-
ucts of compounds 1–4 are shown. Hydroxylation of the ben-
zofuran moiety (e.g., 1g–4g), dehydrogenation of the piperi-
dine ring (e.g., 1k–4k), N-debenzylation (e.g., 1a–4a), and all
possible combinations of these reactions (e.g., 1h–4h, 1i–4i,
1m–4m) have taken place. The position of the hydroxy group
cannot be specified based on the exact masses, but the pres-
In general six metabolites of the homologous fluoroalkyl-
substituted derivatives 1–4 could be identified (Figure 5). The
secondary amines 1a–4a belong to the main metabolites.
Their exact mass indicates the removal of the N-benzyl group
by cytochrome P450 (CYP) catalyzed dealkylation. Moreover,
the structures of the fluoroethyl and fluoropropyl derivatives
Figure 3. Mass voltammogram obtained from the electrochemical oxidation of compound 1. The signal intensity of 1 (m/z 312) is decreasing, while signals
for oxidation products appear with increasing signal intensities. For clarity only normal mass values of oxidation products are shown.
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
3
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Figure 4. Possible oxidation pathway of compounds 1–4. Structural formulae were deduced based on the determined sum formulae, which were obtained
from the detected exact masses. (a) Electrochemical N-debenzylation.
Figure 5. Metabolites formed by incubation of 1–4 with rat liver microsomes and NADPH. Metabolites of 2 and 3 have already been reported.[33, 36]
2a and 3a were confirmed unequivocally by comparison of
the retention times and mass spectra with independently syn-
thesized compounds.[32,36]
Three monooxygenated products were formed by CYP cata-
lyzed oxidation, respectively. Compounds 1b,c–4b,c bearing
the additional oxygen atom in the fluoroalkyl side chain or the
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
4
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
benzyl moiety represent important metabolites. In the MS
analyses of the hydroxybenzyl metabolites 1c–4c fragments
with a mass decreased by 106 Da were detected. It is assumed
that the M+H+À106 fragment results from the cleavage of
the hydroxybenzyl moiety. Additionally, the fragments M+H+
À106 of the metabolites 1c–4c are identical with the frag-
ments M+H+À91 of the parent compounds 1–4 and produce
the same fragmentation pattern. The 4-hydroxybenzyl substi-
tuted compound 2c had been synthesized to prove the struc-
ture of this metabolite and furthermore the position of the OH
moiety in the para position.[34]
the four homologous fluoroalkyl derivatives 1–4 after incuba-
tion with rat liver microsomes and electrochemical oxidation.
In addition to N-debenzylation, rat liver microsomes oxidized
compounds 1–4 at the fluoroalkyl side chain (1b–4b), the
benzyl moiety (1c–4c), and the N atom (1d–4d). These posi-
tions were also predicted by FAME. In contrast, the electro-
chemical oxidation did not modify the fluoroalkyl side chain
and the phenyl ring of the benzyl moiety. Nevertheless, dehy-
drogenation of the piperidine ring (1k–4k) was observed
during electrochemical oxidation, which could be the first step
in oxidative transformations around the tertiary amine. The
predicted and detected (electrochemical oxidation) products
with a hydroxy moiety in the benzofuran ring (1g–4g) could
not be found after incubation with rat liver microsomes.
The fragmentation pattern of the metabolites 1b–4b ex-
cludes hydroxylation in the benzyl moiety and the piperidine
ring. The stepwise fragmentation of the metabolite 2b showed
elimination of the benzyl moiety (M+H+À91) and subsequent
loss of two molecules of water (M+H+À91À18 and M+H+
À91À18À18). The removal of two molecules of water is only
possible if the OH moiety is in an aliphatic position, that is, in
the fluoroalkyl side chain. The first H2O molecule resulted from
removal of the additional hydroxy moiety, the second one
stems from the benzofuran subunit. Moreover, in the chroma-
togram, the metabolite 2b shows two peaks close together
with the same exact mass and the same fragmentation pat-
tern. These peaks could result from two diastereomeric metab-
olites, since hydroxylation in the fluoroalkyl side chain led to
an additional center of chirality.
Kinetics of metabolic degradation
Next, the stability of the four homologous fluoroalkyl substitut-
ed spirocyclic s1 ligands 1–4 during incubation with rat liver
microsomes and NADPH was determined. For this purpose, in-
cubations were stopped after 5, 10, 30, 40, 60 and 90 min, by
addition of cold (À208C) acetonitrile. The amount of remaining
parent compound 1–4 was determined by HPLC. A matrix
matched calibration was performed without addition of
NADPH using the compound concentration in the assay as
highest calibration concentration.
The retention times during RP-HPLC of the third monooxy-
genated metabolites 1d–4d were longer than those of the
parent compounds 1–4 indicating higher lipophilicity of the
metabolites. This observation together with the fragmentation
pattern points to the corresponding N-oxides 1d–4d. The N-
oxides 2d and 3d derived from the fluoroethyl and fluoro-
propyl derivatives 2 and 3 had been prepared independently
confirming the structures 1d–4d.[34,37]
The metabolic degradation of the ligands 1–4 is shown in
Figure 6. It can be seen that the fluoropropyl and fluorobutyl
derivatives 3 and 4 with the elongated side chains are more
stable than the corresponding fluoroethyl and fluoromethyl
derivatives 2 and 1 with shorter side chains. After an incuba-
tion period of 90 min, ~25–30% of the parent compounds 3
and 4 were still detected, but 90% of 1 and almost 100% of 2
were metabolized by the microsome preparation.
The secondary amines 1e–4e with a hydroxy moiety in the
fluoroalkyl side chain represent the most polar metabolites.
After elimination of water (aliphatic OH group) in MSn frag-
mentation experiments, the further fragmentation patterns
were very similar to the fragmentation patterns of the metabo-
lites 1a–4a. Additionally, the twin peaks in the chromatogram
indicate the existence of diastereomers originating from the
additional OH group in the fluoroalkyl side chain.
The amount of unchanged parent compound after an incu-
bation period of 30 min is of particular interest, since PET stud-
ies are usually performed during this period of time. The fluo-
ropropyl derivative 3 represents the most stable compound of
this series of fluorinated ligands with 46% of the parent com-
In addition to monooxygenated metabolites 1b,c,d–4b,c,d,
incubation with rat liver microsomes produced one dihydroxy-
lated metabolite 1 f–4 f, respectively. The mass of the main
fragment after cleavage of the hydroxybenzyl moiety (M+H+
À106) was identical to the mass of the metabolites 1e–4e.
Moreover, the fragmentation pattern of the main fragments
and the metabolites 1e–4e were very similar pointing to the
metabolites 1 f–4 f bearing an N-(hydroxybenzyl) moiety and
an additional OH moiety in the fluoroalkyl side chain. This
comparative study leads to the conclusion that incubation of
the four homologous fluoroalkyl substituted spirocyclic s1 li-
gands 1–4 with rat liver microsomes produced very similar pat-
terns of metabolites.
As predicted by FAME, the secondary amines 1a–4a formed
by oxidative debenzylation represent the main metabolites of
Figure 6. Degradation of the four homologous fluoroalkyl derivatives 1–4
over time during incubation with rat liver microsomes and NADPH.
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
5
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
pound remaining unchanged after 30 min. The longer fluoro-
butyl derivative 4 showed almost the same stability (36% un-
changed after 30 min). However, the metabolic stability of the
fluoromethyl and fluoroethyl derivatives 1 and 2 was consider-
ably lower than the stability of 3 and 4, since only 20% and
13% of the parent compounds remained unchanged after an
incubation period of 30 min.
parent compounds from biological matrices was investigated.
In the majority of radiometabolite experiments, brain and liver
homogenates (tissues) as well as blood plasma and urine (bio-
fluids) were analyzed. Investigation of radiometabolites in the
brain is of major interest, since it represents the target organ
of the tracer. Analysis of plasma samples is also crucial for
studying metabolic pathways and for modelling of PET data.
The liver is the central organ responsible for the biotransforma-
tion of exogenous compounds. Polar metabolites are preferen-
tially excreted via the kidney and urine and, therefore, the radi-
ometabolites in the urine were determined.
It can be concluded that the fluoropropyl derivative 3 is
metabolically the most stable and the fluoroethyl derivative 2
is metabolically the least stable ligand of this series of fluori-
nated compounds. However, these results have to be interpret-
ed very carefully, since the rate of degradation in vivo depends
also on parameters such as lipophilicity, plasma protein bind-
ing, penetration into the tissue, etc. In order to image s1 re-
ceptors in the brain the penetration of the ligands into the
brain and the absence of radiometabolites in the brain are im-
portant features.
The first essential step in studying radiometabolites in tis-
sues and biofluids is the separation of low molecular weight
analytes from the biological matrix components in particular
macromolecular components such as proteins. Generally more
than 85% of the activity of the parent radiotracer and its me-
tabolites should be extracted. Due to the complexity of matrix
components, radiotracers and radiometabolites as well as their
interactions, a general method for separation does not exist.
However, protein precipitation is the most commonly used
method.[45,46]
In vivo metabolism of [18F]1–[18F]4 in mice during PET
studies
Despite their strong precipitation efficiency, strong acids
such as HClO4, and trichloroacetic acid are not suitable, since
they are chemically too aggressive and can lead to decomposi-
tion of acid labile metabolites. Organic solvents (CH3CN,
CH3OH, EtOH, acetone) and solvent mixtures (CH3CN/CH3OH,
CH3OH/CH2Cl2, CH3OH/EtOH, CH3CN/H2O, CH3OH/CH3CN/ace-
tone) are preferred precipitation agents.
Radiosynthesis
The fluorinated PET tracers [18F]1–[18F]4 were synthesized as
previously described.[30,33,36,38] In brief the tosylates 5–8 were
reacted with [18F]fluoride in the presence of the cryptand Kryp-
tofix K2.2.2 (Figure 7). Heating of the acetonitrile solutions of
6–8 for 15 min provided the fluorinated PET tracers [18F]2–
[18F]4 in reproducible high labeling yields of ~55–70%. Howev-
er, the synthesis of the fluoromethyl derivative [18F]1 required
heating of the components in DMSO (1508C). The sterically de-
manding 2-benzofuran system close to the reaction center is
the reason for the sluggish reaction of the tosylate 5. In gener-
al radiochemical yields of 35–40%, radiochemical purity of
ꢀ99%, high chemical purity (free from non-radioactive com-
pounds) and a specific activity of ꢀ200 GBqmmolÀ1 within
a total synthesis time of 90–120 min were obtained for all radi-
otracers.
To investigate the recovery of fluorinated PET tracers from
complex biological matrices, [18F]2 (fluspidine) was selected as
model compound. A defined amount of [18F]2 was added to
freshly prepared biomaterials obtained from different sources
and various extracting agents were used. Plasma (rabbit), brain
homogenates (mouse) and liver homogenates (pig) were em-
ployed as biomaterials and CH3CN (100%, 90% and 80%) as
well as CH3OH (100%, 90% and 80%) were used for the pre-
cipitation of proteins and further matrix components. The first
precipitation step already provided ꢀ90% of the total 18F ac-
tivity in the supernatant (radio-HPLC). A second extraction step
increased the extracted activity less than 5% indicating that
the first extracting step was already very efficient.
Extraction of brain, plasma and liver samples with different
precipitation media resulted in very high amounts of the intact
parent radiotracer [18F]2 (96–99%) as monitored by radio-TLC.
These results indicate very low protein binding of the PET
tracer and negligible matrix effects. However, the UV signals
(254 nm) of the HPLC chromatograms of samples prepared
using different extraction media showed significant differences.
Whereas CH3OH as well as aqueous CH3OH and CH3CN re-
vealed some UV-active impurities, samples extracted with pure
CH3CN provided very clean chromatograms. Therefore, CH3CN
was chosen as the preferred precipitating agent in further ex-
periments.
Figure 7. Radiosynthesis of [18F]1–[18F]4: (a) K[18F]F, K2.2.2, DMSO, 1508C,
25 min, for [18F]1; (b) CH3CN, 858C, 15 min for [18F]2–[18F]4.
Recovery of PET tracer [18F]2 from complex biological
matrices
Additionally to these experiments, bovine serum albumin
was incubated with [18F]2 and subsequently the sample was
precipitated with cold CH3CN. The recovery of the PET tracer
was >98%. According to the method of Bradford residual pro-
To analyze the metabolites formed in vivo after application of
the fluorinated tracers [18F]1–[18F]4 to mice, the recovery of the
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
6
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
tein components were not found in the extract indicating
complete precipitation of the proteins. The ratio of biomaterial
and precipitating agent was crucial. Whereas a ratio of bioma-
terial: precipitating agent of 1:2 was somewhat critical to give
a good pellet, a ratio of 1:3 or 1:4 led to formation of good
pellets.
experiments. The recovery was lower than expected (60–90%).
As the amount of intact radiotracer was always >90% (excep-
tion [18F]F1 after 60 min), it is difficult to determine whether
the extraction of the parent compounds or their radiometabo-
lites is incomplete.
The relative activity in mouse brain at 30 min and at 60 min
after injection derived from organ distribution studies of
[18F]1–4 is depicted in Figure 8. The relative amount of the
metabolically more stable radiotracers [18F]2–4 decreased in
the order [18F]2>[18F]3>[18F]4. The logD values at pH 7.4 in-
crease in the same order indicating that the uptake of the
intact radiotracer decreases with increasing lipophilicity. It is as-
sumed that the more lipophilic fluorobutyl derivative [18F]4 is
trapped partly within membranes.[38]
In summary, cold CH3CN used in three- to four-fold excess
over the biomaterial represents the preferred procedure to ex-
tract the fluorinated PET tracers [18F]1–[18F]4 and their metabo-
lites almost quantitatively from the biomaterial.
Analysis of radiometabolites of [18F]1–[18F]4 in different
tissues and biofluids
To analyze radiometabolites in different tissues and biofluids,
brain and liver samples were homogenized and urine and
plasma samples were used directly. The proteins of the sam-
ples were precipitated by cold CH3CN.
Radiometabolites of [18F]1–4 found in mouse brain
In brain, more than 95% of intact radiotracers [18F]2–4 were
present at 30 min and 60 min after injection (Table 1). At these
time points, radiometabolites could not be detected indicating
that radiometabolites were not formed in the brain and radio-
metabolites potentially formed in the periphery could not pass
the BBB.
Figure 8. Relative amounts of activity in mouse brain 30 min and 60 min
after injection of radiotracers [18F]1–4. Doses were standardized related to
wet weight; data are mean values Æstandard deviation; number of experi-
ments n=3–5; distribution coefficients at pH 7.4 (logD7.4) were recorded by
distribution of the radiotracers between n-octanol and phosphate buffer
pH 7.4 and subsequent quantification of activity.[30,33,36, 38]
Table 1. Amount of parent compounds [18F]1–4 and recovery in mouse
brain samples 30 and 60 min after injection (extraction agent: CH3CN).
Compd
30 min[a]
60 min[a]
intact
recovery [%][c]
intact
recovery [%][c]
radiotracer [%][b]
radiotracer [%][b]
A similar trend was reported for N-([18F]fluoroalkyl)spiperone
derivatives, where increasing lipophilicity led to decreased
amounts of radiotracers in the brain. However, the logD7.4
values of fluorinated spiperone derivatives are higher than the
logD7.4 values of the radiotracers [18F]1–4 and differ more from
each other. Within this series of fluoroalkyl spiperone deriva-
tives the preference was given to metabolically more stable
compounds.[47]
[18F]1
[18F]2
[18F]3
[18F]4
91Æ2.5
97Æ1.1
96Æ1.5
96Æ1.0
75–80
50–60
85–90
80–85
81Æ2.0
98Æ0.5
98Æ1.4
95Æ2.7
60–68
65–70
80–85
85–95
[a] Time after radiotracer injection. [b] Mean valuesÆSD (n=3) from radio-
TLC and radio-HPLC experiments. [c] Amount of activity in the CH3CN ex-
tract relative to the total activity in the respective brain sample.
The absence of radiometabolites in the brain together with
the high relative amount of radiotracer in the target tissue
brain (4.6% of injected dose) render the fluoroethyl derivative
[18F]2 the favored radiotracer of this series of homologs.
Notably, the fluoromethyl compound [18F]1 differed signifi-
cantly from [18F]2–4 (Table 1, and Supporting Information (SI)
Figure SI1). HPLC analysis resulted in two radioactive com-
pounds at shorter retention times of 35 min (compound 1-A)
and 26 min (compound 1-B) indicating higher polarity of these
radiometabolites than the parent compound [18F]1. The
amount of both radiometabolites increased over the time. At
30 min after injection of [18F]1, the amount of radiometabolites
was 6% (1-A) and 4% (1-B), whereas at 60 min, 9% and 12%
of the radiometabolites 1-A and 1-B were found, respectively
(Figure SI1). In parallel, the amount of intact radiotracer [18F]1
decreased from 90% (30 min) to 79% (60 min). HPLC and TLC
data were in agreement (Æ3%). The reproducibility of the re-
sults was excellent, considering the natural variability of animal
Radiometabolites of [18F]1–4 found in mouse liver
Because the liver plays a central role in the biotransformation
of xenobiotics and their excretion, the presence of radiotracers
[18F]1–4 and their radiometabolites in mouse liver homoge-
nates was analyzed; 30 min after injection, the amounts of un-
changed radiotracers [18F]2–4 in the liver sample were rather
similar (57–69%), but the amount of intact fluoromethyl deriv-
ative [18F]1 was only 27% (Table SI1). Clearly, the biotransfor-
mation of the fluoromethyl derivative [18F]1 is faster than the
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
7
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
biotransformation of the longer chained homologs [18F]2–4.
Surprisingly, the amounts of [18F]1, [18F]3 and [18F]4 remained
constant between 30 and 60 min after injection, whereas the
fluoroethyl derivative [18F]2 was further biotransformed.
All detected radiometabolites were more polar than the cor-
responding parent compounds [18F]1–4. However, neither in
radio-TLC nor in radio-HPLC was a signal for [18F]fluoride ob-
served. At 30 min after injection, radio-HPLC of liver samples
showed three major metabolites 1-A, 1-B and 1-D for the fluo-
romethyl derivative [18F]1. At the same time, the chromato-
grams of liver samples of mice treated with the homologous
fluoroalkyl derivatives [18F]2–4 were very similar considering
the increasing retention times of homologs with longer alkyl
chains (Figure 9). The corresponding radio-HPLC chromato-
grams of liver samples at 60 min after injection of the radio-
tracers [18F]1-4 were very similar to the chromatograms ob-
tained at 30 min.
Figure 10. Radiochromatograms of mouse liver samples using different ex-
tracting agents. Top: Precipitation with CH3CN; bottom: Precipitation with
CH3OH; insets: chromatograms with UV detection (254 nm). Liver homoge-
nates were obtained at 30 min after injection of the radiotracer [18F]1; the
additional radiometabolite is marked in the second chromatogram; reten-
tion times of parent compound [18F]1 and its radiometabolites: [18F]1:
tR =44.5 min; 1-A: tR ~35 min; 1-B: tR ~27 min; 1-C: tR ~24.1 min (additional
radiometabolite); 1-D: tR ~22 min.
or aqueous Ch3CN (90% CH3CNaq) resulted in lower recovery of
activity than extraction with CH3CN. However, two very polar
radiometabolites with very short retention times (tR <4 min)
were found after precipitation with aqueous CH3CN and aque-
ous CH3OH. Extraction of liver samples containing the higher
homologous fluorobutyl radiotracer [18F]4 with CH3OH led to
~10% higher amounts of activity.
Figure 9. Chromatograms of radioactive metabolites of tracers [18F]1–4 de-
tected by radio-HPLC in samples of mouse liver at 30 min after injection (ex-
traction agent CH3CN). Retention times of parent compounds: [18F]1:
tR =44.5 min; [18F]2: tR =49.9 min; [18F]3: tR =53.7 min; [18F]4: tR =54.7 min.
Radiometabolites of [18F]1–4 found in mouse plasma
At 30 min after injection, more than 84% of the activity of
plasma samples was represented by the unchanged radiotrac-
ers [18F]2–4. Only for the fluoromethyl derivative [18F]1 was
a decreased amount (75%) of intact radiotracer detected.
Treatment of the mice for 60 min led to reduced amounts of
unchanged radiotracers in the plasma, which can be attributed
to increased amounts of radioactive metabolites in the plasma.
In the radio-HPLC chromatograms displayed in Figure 11
(30 min samples) only two small peaks were observed in addi-
tion to the unchanged fluorobutyl derivative [18F]4 (parent ra-
diotracer 84%). These peaks appear at retention times of
In Figure SI2 the chromatograms of liver samples at 30 min
after injection of the fluoromethyl and fluoroethyl radiotracers
[18F]1 and [18F]2 are compared. For the fluoroethyl derivative
[18F]2 a few trace radiometabolites and two main radiometabo-
lites 2-A and 2-B were detected. On the contrary, three major
radiometabolites 1-A, 1-B and 1-D in considerably higher
amounts were observed for the fluoromethyl derivative [18F]1.
The number of minor radiometabolites, however, increased at
60 min after injection. In liver samples of mice treated with flu-
oropropyl and fluorobutyl homologs [18F]3 and [18F]4, several
radiometabolites (>10) were detected by radio-HPLC. Two of
these radiometabolites can be considered as major metabo-
lites.
As documented in Table SI1 the recovery of activity after ex-
traction of liver homogenates with CH3CN was generally rather
low. In some experiments, other extracting agents were investi-
gated. Precipitating liver samples containing the fluoromethyl
radiotracer [18F]1 with CH3OH, led to the same amount of un-
changed radiotracer [18F]1 as found with CH3CN. However, an
additional radiometabolite 1-C at a retention time of 24.1 min
was detected after extraction with CH3OH, which was not visi-
ble after extraction with CH3CN (Figure 10).
Figure 11. Radiochromatograms of plasma samples 30 min after injection of
the radiotracers [18F]1–4 (extraction agent CH3CN). The peaks of the parent
For the fluoroethyl radiotracer [18F]2, extraction with more
polar solvents such as CH3OH, aqueous CH3OH (90% CH3OHaq)
radiotracers are marked with . The peak for radiometabolite 2-B of tracer
*
[18F]2 is increased by a factor of 3.
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
8
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
27.1 min (4-D, 14.3%) and 3.5 min (4-E, 3.6%). A few further
metabolites were detected in traces. A very similar radiochro-
matogram was obtained for the fluoropropyl derivative [18F]3
(86% parent) showing two major metabolites at retention
times of 25.6 min (3-D, 10.3%) and 3.5 min (3-E, 3.8%). For the
smallest homolog, the fluoromethyl derivative [18F]1, several
small and two main radioactive metabolites were detected,
which correlate nicely with the metabolites 1-A (37.1 min,
9.5%) and 1-D (21.7 min, 12.0%) found in tissue samples (only
1-A). The plasma sample of the fluoroethyl derivative [18F]2
(89% parent) however reveals a large peak for the radiometa-
bolite 2-D at a retention time of 28.2 min (5.6%) and a smaller
peak at 3.5 min (2-E, 5.4%). The formation of the metabolite 2-
D is the reason for decreased amounts of unchanged radio-
tracer in plasma (Table 2).
decreased to ~50% of the activity obtained at 30 min. Clearly
the more polar radiotracers [18F]1–3 are distributed faster into
tissues and organs, whereas the more lipophilic fluorobutyl ra-
diotracer [18F]4 remains in the plasma for a longer time, which
could be due to higher plasma protein binding.
Metabolites of [18F]1–4 found in mouse urine
Since urine samples were only available in some experiments,
the analysis of radiometabolites was not as exhaustive as for
tissue and plasma samples. In Table SI2 the amounts of un-
changed radiotracers in urine samples at 30 min and at 60 min
after injection of [18F]1–4 is summarized. In the urine samples,
only a small amount (<4%) of parent radiotracers was found.
Clearly, the intact radiotracers were not eliminated via the
kidney and urine.
Table 2. Amount of recovered parent compounds [18F]1–4 and recovery in
mouse plasma 30 min and 60 min after injection (extraction agent:
CH3CN).
On the contrary, up to 10 radiometabolites were detected
by analytical HPLC. The majority of radiometabolites were
found in trace amounts only, but a few radiometabolites oc-
curred in large amounts (Figure 13, Table 3). For all tracers
[18F]1–4, rather polar radiometabolites 1-E to 4-E with a reten-
tion time of 3.5 min were detected, which represent major me-
tabolites, in the case of fluoroethyl and fluorobutyl derivatives
[18F]2 and [18F]4. The metabolites 1-E to 4-E appeared directly
after the exclusion volume of the radio-HPLC. It is possible that
this peak is caused by [18F]fluoride. However, separate analyses
of femur, that is, activity in bone versus marrow, did not indi-
cate any defluorination of the tracers [18F]1–4. The short reten-
tion time in the HPLC analysis indicates the presence of
[18F]fluoroethanol or [18F]fluoroacetic acid.
Compd
30 min[a]
60 min[a]
intact
recovery [%][c]
intact
recovery [%][c]
radiotracer [%][b]
radiotracer [%][b]
[18F]1
[18F]2
[18F]3
[18F]4
75Æ10
89Æ3
86Æ2
84Æ9
76–80
60–75
55–65
60–70
52Æ7
75Æ8
83Æ5
86Æ5
60–65
50–60
60–68
67–75
[a] Time after radiotracer injection. [b] Data are mean valuesÆSD (n=3–5)
from radio-TLC and radio-HPLC experiments. [c] Amount of activity in the
CH3CN extract relative to the total activity in the respective plasma
sample.
The main metabolites for the four homologous fluoroalkyl
derivatives [18F]1–4 are the metabolites 1-C to 4-C with reten-
tion times of 22–29 min. The increasing retention times from
1-C to 4-C are due to increasing chain length. Interestingly, the
radiometabolite pattern in the urine of mice treated with [18F]1
differs considerably from the pattern of the homologous trac-
ers [18F]2–4. Although the urine is already a solution of the ra-
diometabolites, extraction of the solution with CH3CN and sub-
sequent analysis of activity led to a recovery of 81% for [18F]1,
66% for [18F]2, 59% for [18F]3 and 60% for [18F]4.
Figure 12 shows the amount of activity in CH3CN extracts of
mouse plasma related to the injected dose obtained during
organ distribution studies.[30,33,36,38] The following trend was ob-
served for plasma samples obtained at 30 min: [18F]3>[18F]4>
[18F]1~[18F]2. The most lipophilic fluorobutyl derivative [18F]4
revealed the smallest difference between the amount of activi-
ty recorded at 30 min and at 60 min as both values are almost
identical. On the contrary, the activity amount determined for
the more polar radiotracers [18F]1–3 60 min after injection was
Figure 12. Amount of activity in mouse plasma in relation to the injected
dose at 30 min and at 60 min after injection of the radiotracers [18F]1–4; wet
weight normalized dose; data are mean values Æstandard deviation, n=3–
5.
Figure 13. Radiochromatograms of urine samples 30 min after injection of
the radiotracers [18F]1–4 (extraction agent CH3CN). The peaks of the parent
radiotracers are marked with
.
*
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
9
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
Table 3. Relative amounts of major radiometabolites 60 min after injection.[a]
Compd
Metab.
tR [min]
Amount [%]
Metab.
tR [min]
Amount [%]
Metab.
tR [min]
Amount [%]
[18F]1
[18F]2
[18F]3
[18F]4
1-E
2-E
3-E
4-E
3.5
3.5
3.5
3.6
2–9
31–48
4–6
1-D
2-D
3-D
4-D
22.6
24.1
24.3
24.3
14–29
5–7
6–10
0.10
1-C
2-C
3-C
4-C
24.2
26.3
26.9
28.6
47–64
42–59
74–89
23–34
41–61
[a] Values are the mean (n=2) from radio-TLC and radio-HPLC experiments.
Comparison of metabolite spectra from brain, liver, plasma
and urine using [18F]2 as example
in vivo studies with mice were correlated. The most lipophilic
metabolites formed as major metabolites with rat liver micro-
somes are the hydroxybenzyl derivatives 1c–4c. In vivo, radio-
metabolites 1-A to 4-A represent the most lipophilic radiome-
tabolites, which are also formed as major radiometabolites (see
Figure 10–12). Therefore, it can be speculated that the radio-
metabolites 1-A to 4-A formed in vivo bear a hydroxy moiety
in the benzyl ring as shown for metabolites 1c–4c.
Debenzylation of the tertiary amines (1a–4a) represents the
second important metabolic pathway catalyzed by CYP en-
zymes of rat liver microsomes and NADPH. The secondary
amines 1a–4a were also detected as major products after elec-
trochemical oxidation and predicted by FAME. As the secon-
dary amines 1a–4a are considerably more polar than metabo-
lites bearing a hydroxy moiety in the phenyl ring (1c–4c) or in
the fluoroalkyl side chain (1b–4b), we speculate that the radi-
ometabolites 1-D to 4-D are the secondary amines [18F]1a–4a.
According to this hypothesis, the less polar radiometabolites
marked with 1-B to 4-B could bear the hydroxy moiety in the
fluoroalkyl side chain corresponding to metabolites 1b–4b.
After analyzing the organ-specific radiometabolites separately,
the radiometabolite spectra of the homologous tracers [18F]1–
4 were compared. Here the fluoroethyl derivative [18F]2 is dis-
cussed. In Figure 14, the radiochromatograms recorded from
tissue samples and body fluids at 30 min after application of
[18F]2 are displayed. In the chromatogram of brain samples,
only one peak for the unchanged parent compound [18F]2 is
shown indicating that radiometabolites are not formed in the
brain and radiometabolites are not able to pass the BBB. In
contrast, the urine sample does not contain the parent com-
pound [18F]2, but two major very polar radiometabolites 2-D
and 2-E with retention times of 26.4 min and 3.5 min, respec-
tively. Due to their high polarity these radiometabolites are
eliminated via kidney and urine. In the liver and plasma sam-
ples the parent compound [18F]2 represents the major compo-
nent indicating high metabolic stability. In both samples only
trace amounts of the radiometabolites 2-A (tR =38.2 min) and
2-B (tR =31.0 min) were detected.
The highly polar radiometabolite 2-E was present predomi-
nantly in urine and in low amounts in plasma. Its structure has
not yet been identified, but the fast renal elimination inhibits
the penetration into various tissues and in particular the CNS.
Conclusions
Different levels of animal experiments provided complementa-
ry information about the biotransformation of homologous flu-
oroalkyl substituted s1 receptor ligands 1–4. Theoretical analy-
sis of metabolites using databases (e.g., FAME) does not re-
quire any lab experiments. Electrochemical oxidation can be
performed without any animal material and provides valuable
ideas about oxidatively labile positions in the s1 ligands. Al-
though structural information can be derived from these ex-
periments and computational studies, the biological relevance
remains to be elucidated. In vitro metabolism studies using rat
liver microsomes require liver from rats as native biological ma-
terial. The degradation rate and the structure of metabolites
formed by oxidation with CYP enzymes can be analyzed with
this artificial system. However, studies with mice resemble best
the fate of a PET tracer in a living organism. Disadvantages of
in vivo studies are the requirement of animals and the missing
structural information about radiometabolites due to the very
low amount of PET tracers used in animal experiments. These
drawbacks are compensated by in vitro studies with liver mi-
crosomes and electrochemical oxidation.
Figure 14. Comparison of radio HPLC chromatograms of samples obtained
from different tissues and biofluids at 30 min after injection of the fluoroeth-
yl derivative [18F]2 (extraction agent CH3CN); major metabolites are marked
*
with 2-A, 2-B, 2-D and 2-E; minor metabolites are marked with
.
Proposed structures of radiometabolites
For in vivo studies, receptor affinity and selectivity as well as
radiochemical availability represent important features of PET
tracers. However, pharmacokinetic parameters, such as logD7.4
value determining the passage of membranes, penetration
In order to speculate about the nature of the in vivo formed
radiometabolites, the information gained in in vitro studies
with rat liver microsomes, electrochemical oxidations and
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
10
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
ferred into an ultracentrifuge (Beckmann L8M with rotor Ti 70.1)
and centrifuged for 60 min at 100000 g and 48C. The resulting su-
pernatant (cytosol) was discarded, the pellet carefully washed with
buffer, resuspended and the centrifugation repeated. Finally the
supernatant was removed, the pellet resuspended in a small
amount of phosphate buffer (pH 7.4) and the microsome suspen-
sion was stored at À808C. The protein concentration was deter-
mined according to the method of Bradford[48] using bovine serum
albumin as standard.
into the CNS and biotransformation are also important proper-
ties of PET tracers. This study showed that the database
(FAME), electrochemical oxidation experiments, in vitro bio-
transformation with rat liver microsomes and in vivo metabo-
lism studies with mice provided complementary information
supporting the final decision, which compound of this series
of homologous fluoroalkyl derivatives should be further devel-
oped.
In the case of the fluoroalkyl substituted s1 ligands [18F]1–
[18F]4, the fluoroethyl and fluoropropyl derivatives [18F]2 and
[18F]3 revealed a good brain uptake, while their metabolites
could not be detected in the brain. Clearly, the radiometabo-
lites detected in liver, plasma and urine are not able to pass
the BBB. Considering the high amount in the brain (4.6% of in-
jected dose), the high s1 affinity (Ki =0.59 nm) and the low
logD7.4 value (2.57), [18F]2 is the favored radioligand to be fur-
ther developed as PET tracer.
Incubation of 1–4 with rat liver microsomes: The incubation was
carried out in phosphate buffer (pH 7.4) at room temperature in
a circular shaker (IKA vibrax VXR, Staufen, Germany) and contained
rat liver microsomes (1.5 mgmLÀ1 protein), MgCl2 (0.86 mm) and
NADPH (2.6 mm). The concentration of 1–4 was 260 mm in a final
volume of 0.9 mL. Usually, the incubation was stopped after
90 min by addition of cold acetonitrile (À208C). The samples were
stored for 10 min on ice to complete protein precipitation. After
thawing, the samples were centrifuged for 8 min (10000 g in Het-
tich Mikro 20 benchtop centrifuge). The supernatant was decanted,
filtered with a 0.45 mm pore size syringe filter made from regener-
ated cellulose and analyzed.
Experimental Section
Structure elucidation by HPLC–MS: The LC–MS system consisted
of a LTQ Orbitrapꢂ XL with Accelaꢂ HPLC pump and autosampler
(Thermo Fisher). The prepared incubation solutions (20 mL) were in-
jected onto a LiChrospherꢂ RP Select B column (5 mm, 250ꢃ
4.6 mm) with guard column LiChroCARTꢂ RP Select B (4ꢃ4.0 mm)
at a flow rate of 1.0 mLminÀ1. The mobile phase was composed of
(A) 15% CH3CN in water and (B) 60% CH3CN in water; both com-
ponents contained formic acid (0.1%). The following gradient was
applied (A%): 0–20 min: gradient from 100% to 0%, 20–23 min:
0%, 23–24 min: gradient from 0% to 100%, 24–30 min: 100%. For
MS analysis the eluent was reduced to 25% by a post-column split-
ter (Acurate, LC Packings, Dionex). The MS parameters were as fol-
lows: ion spray voltage: 4 kV in positive mode, sheath gas flow: 40
arbitrary units, aux gas flow: 15 arbitrary units, sweep gas flow: 10
arbitrary units, capillary temperature: 2758C, capillary: voltage
40.5 V. First a total ion current (TIC) spectrum was recorded. In
order to elucidate the structures of the metabolites, collision in-
duced dissociation (CID), fragmentation experiments were carried
out. The Orbitrap MS system generated data with fragmentations
up to MS3.
In silico prediction of metabolically labile positions
ChemDraw was used to draw the basic amines 1–4 as protonated
species as they exist at pH 7.4. Then, sdf files were created. These
files were used by FAME for the prediction of metabolically labile
positions. The applied algorithm included phase 1 reactions from
rats (“ratPhase1”).
Electrochemical oxidation of 1–4
Compounds 1–4[30,32,36,38] were dissolved in ammonium formate
(50/50 (v/v) 10 mm, pH 7.4, adjusted with aqueous ammonia) and
CH3CN to a final concentration of 10 mm. The solutions were con-
tinuously pumped through an amperometric electrochemical thin-
layer cell (ReactorCell, Antec Leyden, Netherlands) equipped with
a boron-doped diamond working electrode, a graphite-doped
Teflon counter electrode and a Pd/H2 reference electrode, to which
all potentials mentioned in this work refer. The flow rate was set to
10 mLminÀ1, and a potential ramp between 0 and 2.5 V with a scan
rate of 10 mVsÀ1 was applied. The effluent of the electrochemical
cell was directly analyzed by means of ESI-HRMS using an Orbitrap
mass analyzer (Exactive, Thermo Fisher Scientific, Bremen, Germa-
ny). Mass spectra were recorded in the positive ionization mode
with the following MS parameters: ESI(+); resolution, high; AGC
target, balanced; sheath gas flow (N2), 10 au; capillary temperature,
2758C; spray voltage, 4.0 kV; capillary voltage, 95 V; tube lens volt-
age, 120 V; skimmer voltage, 38 V; maximum injection time,
100 ms; m/z range, 100–1000.
Metabolic stability: Six incubations of the compounds were car-
ried out in phosphate buffer (pH 7.4) at room temperature in a cir-
cular shaker containing rat liver microsomes (1.5 mgmLÀ1 protein),
MgCl2 (0.86 mm) and NADPH (2.6 mm). The concentration of com-
pounds 1–4 was 320 mm in a final volume of 900 mL. After periods
of 5, 10, 30, 40, 60 and 90 min, the incubations were stopped by
addition of cold CH3CN (500 mL, À208C). Then praziquantel solu-
tion (200 mL, 0.6 mgmLÀ1 in CH3CN) was added as internal standard
resulting in a total volume of 1.6 mL. All incubations were carried
out as duplicates. The samples were kept at À208C for 30 min to
complete protein precipitation. After thawing, the samples were
centrifuged at 10000 g for 10 min, then the supernatant was deca-
nted, filtered with a 0.45 mm pore size syringe filter made from re-
generated cellulose (Macherey–Nagel, Dꢀren, Germany) and ana-
lyzed. A calibration was carried out with the same microsomal
matrix except NADPH. All calibration standards were treated in the
same way (90 min on the shaker, protein precipitation with aceto-
nitrile, centrifugation, etc.). The HPLC analysis was carried out with
a Merck Hitachi equipment consisting of UV detector: L-7400; au-
tosampler: L-7200; pump: L-7150; interface: D-7000. Twenty micro-
liters of the prepared incubation solutions were injected onto
In vitro metabolism using rat liver microsomes
Preparation of rat liver microsomes: Frozen rat livers from Wistar
rats were thawed in phosphate buffer (pH 7.4) with sucrose
(0.25m) and EDTA (5 mm), cut into pieces and homogenized with
a Potter (Elvehjem Potter, B. Braun Biotech International, Melsun-
gen, Germany) at 48C. The resulting suspension was centrifuged
(high-speed cooling centrifuge model Sorvall, RC-5C plus, Thermo
Finnigan, Langerwehe, Germany) at 10000 g for 15 min at 48C. Fat
was removed by absorption with cellulose. The pellet was resus-
pended in buffer, the mixture centrifuged again and afterward
both supernatants were combined. This suspension was trans-
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
11
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
a Phenomenex Geminiꢂ C18 column (5 mm, 250ꢃ4.6 mm) at a flow
rate of 1.0 mLminÀ1 and UV detection wavelength of 210 nm. The
mobile phase was composed of (A) 15% CH3CN in water and
(B) 60% CH3CN in water. Trifluoroacetic acid (0.05%) was added to
both components. The following gradient was applied (A%): 0–
20 min: gradient from 100% to 0%, 20–23 min: 0%, 23–24 min:
gradient from 0% to 100%, 24–30 min: 100%.
Fragmentation of the metabolites of the fluoropropyl deriva-
tive 3
Metabolite 3a: HPLC: tR =12.6 min. MS1 (ESI+ exact mass): m/z
calcd for [C15H20FNO+H] 250.1602, found 250.1602. MS2 (ESI+
exact mass): m/z calcd for [C15H18FN+H] 232.1496, found 232.1493
[MÀH2O].
Metabolite 3b: HPLC: tR =15.1 min and 16.9 min (diastereomers).
MS1 (ESI+ exact mass): m/z calcd for [C22H26FNO2 +H] 356.2020,
found 356.2016. MS2 (ESI+ exact mass): m/z calcd for [C12H13FO2 +
H] 209.0972, found 209.0969 [MÀCH2C6H5ÀN(CH2)3].
Data of the metabolites: Exact masses of metabolites and corre-
sponding fragments were determined using LTQ Orbitrapꢂ XL. The
exact masses confirm the elemental composition of the postulated
metabolites.
Metabolite 3c: HPLC: tR =18.3 min and 18.6 min (regioisomers).
MS1 (ESI+ exact mass): m/z calcd for [C22H26FNO2 +H] 356.2020,
found 356.2016. MS2 (ESI+ exact mass): m/z calcd for [C15H20FNO+
H] 250.1602, found 250.1598 [MÀCH2C6H4OH].
Fragmentation of the metabolites of the fluoromethyl deriva-
tive 1
Metabolite 1a: HPLC: tR =8.3 min. MS1 (ESI+ exact mass): m/z
calcd for [C13H16FNO+H] 222.1294, found 222.1291. MS2 (ESI+
exact mass): m/z calcd for [C12H14FN+H] 204.1189, found 204.1183
[MÀH2O].
Metabolite 1b: HPLC: tR =12.0 min and 12.7 min (diastereomers).
MS1 (ESI+ exact mass): m/z calcd for [C20H22FNO2 +H] 328.1713,
found 328.1711. MS2 (ESI+ exact mass): m/z calcd for [C10H9FO2 +
H] 181.0665, found 181.0659 [MÀCH2C6H5ÀN(CH2)3].
Metabolite 3e: HPLC: tR =8.1 min. MS1 (ESI+ exact mass): m/z
calcd for [C15H20FNO2 +H] 266.1551, found 266.1551.
Metabolite 3 f: HPLC: tR =12.6 min. MS1 (ESI+ exact mass): m/z
calcd for [C22H26FNO3 +H] 372.1969, found 372.1969. MS2 (ESI+
exact mass): m/z calcd for [C22H24FNO2 +H] 354.1864, found
354.1863 [MÀH2O]. MS3 (ESI+ exact mass): m/z calcd for
[C15H18FNO+H] 248.1445, found 248.1445 [MÀH2OÀCH2C6H4OH].
Metabolite 1c: HPLC: tR =13.5 min. MS1 (ESI+ exact mass): m/z
calcd for [C20H22FNO2 +H] 328.1713, found 328.1711. MS2 (ESI+
exact mass): m/z calcd for [C13H16FNO+H] 222.1294, found
222.1290 [MÀCH2C6H4OH].
Fragmentation of the metabolites of the fluorobutyl deriva-
tive 4
Metabolite 4a: HPLC: tR =14.9 min. MS1 (ESI+ exact mass): m/z
calcd for [C16H22FNO+H] 264.1764, found 264.1758. MS2 (ESI+
exact mass): m/z calcd for [C16H20FN+H] 246.1658, found 246.1654
[MÀH2O].
Metabolite 4b: HPLC: tR =20.2 min and 20.6 min (diastereomers).
MS1 (ESI+ exact mass): m/z calcd for [C23H28FNO2 +H] 370.2182,
found 370.2176. MS2 (ESI+ exact mass): m/z calcd for [C12H13FO2 +
H] 223.1134, found 223.1129 [MÀCH2C6H5ÀN(CH2)3].
Metabolite 4c: HPLC: tR =14.6 min and 18.8 min (regioisomers).
MS1 (ESI+ exact mass): m/z calcd for [C23H28FNO2 +H] 370.2182,
found 370.2176. MS2 (ESI+ exact mass): m/z calcd for [C15H20FNO+
H] 264.1764, found 264.1758 [MÀCH2C6H4OH].
Metabolite 4e: HPLC: tR =9.8 min. MS1 (ESI+ exact mass): m/z
calcd for [C16H22FNO2 +H] 280.1713, found 280.1709.
Metabolite 1e: HPLC: tR =6.0 min. MS1 (ESI+ exact mass): m/z
calcd for [C13H16FNO2 +H] 238.1234, found 238.1238.
Metabolite 1 f: HPLC: tR =9.4 min and 9.87 min (isomers). MS1
(ESI+ exact mass): m/z calcd for [C20H22FNO3 +H] 344.1662, found
344.1658. MS2 (ESI+ exact mass): m/z calcd for [C20H20FNO2 +H]
326.1556, found 326.1551 [MÀH2O]. MS3 (ESI+ exact mass): m/z
calcd
for
[C13H14FNO+H]
220.1138,
found
220.1133
[MÀH2OÀCH2C6H4OH].
Fragmentation of the metabolites of the fluoroethyl derivative
2
Metabolite 2a: HPLC: tR =11.1 min. MS1 (ESI+ exact mass): m/z
calcd for [C14H18FNO+H] 236.1451, found 236.1444. MS2 (ESI+
exact mass): m/z calcd for [C14H16FN+H] 218.1345, found 218.1338
[MÀH2O].
Metabolite 2b: HPLC: tR =13.4 min and 15.4 min (diastereomers).
MS1 (ESI+ exact Mass): m/z calcd for [C21H24FNO2 +H] 342.1869,
found 342.1865. MS2 (ESI+ exact mass): m/z calcd for [C11H11FO2 +
H] 195.0821, found 195.0813 [MÀCH2C6H5ÀN(CH2)3].
Metabolite 2c: HPLC: tR =16.6 min and 17.0 min (regioisomers).
MS1 (ESI+ exact mass): m/z calcd for [C21H24FNO2 +H] 342.1869,
found 342.1865. MS2 (ESI+ exact mass): m/z calcd for [C14H18FNO+
H] 236.1451, found 236.1444 [MÀCH2C6H4OH].
Metabolite 2e: HPLC: tR =6.6 min. MS1 (ESI+ exact mass): m/z
calcd for [C14H18FNO2 +H] 252.1400, found 252.1392.
Metabolite 4 f: HPLC: tR =14.5 min. MS1 (ESI+ exact mass): m/z
calcd for [C23H28FNO3 +H] 386.2131, found 386.2128 MS2 (ESI+
exact mass): m/z calcd for [C23H26FNO2 +H] 368.2026, found
368.2023 [MÀH2O]. MS3 (ESI+ exact mass): m/z calcd for
[C16H20FNO+H] 262.1607, found 262.1604 [MÀH2OÀCH2C6H4OH].
In vivo metabolism of radiofluorinated s1 ligands [18F]1–4 in
mice
Animal experiments: All procedures involving animals were car-
ried out according to the Principles of Laboratory Animal Care (NIH
publication No. 86-23, revised 1985) and following the national
laws on the protection of animals. For the HZDR group, the studies
were performed based on an authorized approval (Landesdirektion
Leipzig, No. 24-9168, 11/12/23; TVV 22/10).
Metabolite 2 f: HPLC: tR =10.7 min. MS1 (ESI+ exact mass): m/z
calcd for [C21H24FNO3 +H] 358.1818, found 358.1810. MS2 (ESI+
exact mass): m/z calcd for [C21H22FNO2 +H] 340.1713, found
340.1704 [MÀH2O]. MS3 (ESI+ exact mass): m/z calcd for
[C14H16FNO+H] 234.1284, found 234.1284 [MÀH2OÀCH2C6H4OH].
Radiosynthesis: Aqueous [18F]fluoride solution was transformed
into the reactive K[18F]F–K2.2.2–carbonate complex using K2CO3
(1.8 mg, 12.9 mmol) and K2.2.2 (Kryptofix, 11.2 mg, 29.7 mmol). This
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
12
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

Full Papers
complex was reacted with tosylates 5–8 (2.5–3 mg) dissolved in an-
hydrous CH3CN (1 mL). The mixture was heated at 858C for 15 min.
For the synthesis of [18F]1 the transformation was performed in
DMSO (1 mL) and the mixture was heated at 1508C for 25 min. The
resulting mixture was diluted with water (3 mL) and directly ap-
plied to an isocratic semi-preparative HPLC for purification (Multis-
pherꢂ 120 RP18 AQ column, 150ꢃ10 mm, particle size 5 mm with
guard column, 50ꢃ10 mm, 5 mm); eluent: CH3CN (55%) with
NH4OAc (20 mm), flow rate 4.0 mLminÀ1. The fractions containing
the respective radiotracer were combined, diluted with water,
passed through a pre-conditioned Sep-Pak C18 cartridge, and the
purified radiotracer was eluted with CH3OH (~1.5 mL). For in vivo
experiments, the solvent was evaporated and the radiotracer was
re-dissolved in saline.
again at 23500 g for 20 min at 48C. This procedure was repeated
twice. The final pellet was resuspended in 5–6 volumes of buffer.
The precipitation of proteins in brain, liver, plasma and urine sam-
ples was performed by addition of four volumes of ice-cold aceto-
nitrile. The precipitate was removed by centrifugation and the su-
pernatant was concentrated (608C, air flow).
Recovery of PET tracer [18F]2 from complex biological matrices:
Plastic tubes were filled with extracting agent (500 mL). Then, the
biological sample was added in a ratio of about 4:1 to 7:1 (extract-
ing agent/biological sample). The samples were vortexed and kept
in an ice bath for 15 min. After centrifugation, the supernatants of
all aliquots were combined and concentrated with argon gas and
gentle warming (~608C) up to a final residual volume of ~150 mL.
The activity of the extract was recorded and related to the activity
of the biological sample. Recovery (%) in this article is defined as
the activity of the extract divided by the activity of the original bio-
logical sample multiplied by 100%. The percentages of parent radi-
otracer and radiometabolites were analyzed by radio-HPLC and
radio-TLC. Data obtained for all radiotracers are based on repetitive
analyses (n=2–5). In general, good agreement was achieved be-
tween radio-HPLC and radio-TLC data (deviations for brain and
urine ~1%, for plasma ꢁ5–8% and liver ꢁ5%). Here, both inter-in-
dividual variability of mice, methodical discrepancies and the ana-
lytical error have to be considered.
Radio-thin-layer chromatography (radio-TLC):
Radio-thin-layer chromatography (radio-TLC) was performed using
SIL G/UV254 precoated plates using different mixtures of ethyl ace-
tate/petroleum ether/NH3; (v/v/v) as optimized standard procedure.
All data are based on duplicate values. Spots were visualized under
UV light (254 nm) and by radioluminescence recording (BAS-1800 II
Bioimaging Analyzer, Fuji film, Tokyo, Japan), and quantitatively
evaluated with AIDA 2.31 software (raytest, Straubenhardt, Germa-
ny). Alternative eluents were applied: ethyl acetate/CH3OH varying
between 95:5 to 80:20 (v/v) and CH2Cl2/CH3OH/NEt3 =8:2:0.1. Rf
values of the reference compounds of the corresponding radiotrac-
ers: 1: Rf =0.74 (ethyl acetate/petroleum ether/NH3 5:5:0.1); 2: Rf =
0.54 (ethyl acetate/petroleum ether/NH3 3:7:0.1); 3: Rf =0.61 (ethyl
acetate/petroleum ether/NH3 3:7:0.1); 4: Rf =0.52 (ethyl acetate/pe-
troleum ether/NH3 3:7:0.1).
In general acetonitrile was used as standard precipitating and ex-
tracting agent. In experiments regarding the recovery of activity,
additional precipitating agents such as pure methanol, aqueous
methanol (90% CH3OHaq) and aqueous acetonitrile (90% CH3CNaq)
were investigated.
For comparative metabolite studies, non-native biological samples
preserved as Tris buffer solutions and stored in a deep freezer or
obtained from fresh ex vivo samples were applied. They included
plasma (rabbit), brain homogenate (mouse), and liver (pig). For this
purpose, tissue material (50–75 mL) or biofluids was mixed with ra-
diotracer in saline (~10 mL, ~1–5 MBq) and extractive agent
(500 mL) and processed as described before for in vivo metabolite
experiments.
Radio-HPLC
Method 1: Agilent device (HPLC 1100 ChemStation; quaternary
pump, degasser, multiwavelength detector), coupled with a radio-
activity-HPLC flow monitor (FAP 100k/HV, bte, Germany); Multos-
pher 120 RP18 AQ 250ꢃ4.6 mm, 5 mm with guard column (10ꢃ
4.6 mm, 5 mm) (CS Chromatographie Service, Germany); gradient:
5–80% CH3CN/20 mmol NH4OAc; 1.0 mLminÀ1; run time: ~60–
70 min; solution A: 5% CH3CN+20 mm NH4OAc; solution B: 80%
CH3CN+20 mm NH4OAc; gradient: 0–10 min (100% A) 10–50 min
(solution B increasing from 0% to 100%), 50–60 min. In few cases
an isocratic method using 42.5% CH3CN+20 mm NH4OAc as
mobile phase was used.
g-counting: The extraction efficiency of the total 18F activity (“activ-
ity balance” recovery) was monitored permanently for each step
and aliquots were measured by gamma counting of the superna-
tants and precipitates using a calibrated gamma counter (Wallac
WIZARD 3, PerkinElmer, USA).
Protein determination according to the method of Bradford:
The amount of residual proteins in supernatants was determined
by photometric analysis with the BCA method[48] using a bovine
serum albumin standard (PerkinElmer) and a microplate reader
(BIORAD Model 680 microplate reader, Bio-Rad Laboratories, Ger-
many); Windows based Microplate Manager software; 96-well mi-
crotiter plate; Instruction Thermo Scientific Pierce BCA Protein
Assay Kit [23225/23227] (Pierce Biotechnology, Rockford, USA).
Method 2: Jasco HPLC Serie 2000 (interface LC-NET II ADC; pumps:
binary p., quaternary HP gradient system p.; 4-line-degassers; auto-
sampler; UV/Vis detectors; coupled with radioactivity-HPLC-low-
monitors (Gabi Star, raytest GmbH, Germany), software Chrompass;
Reprosil-Pur C18-AQ 250ꢃ4.6 mm, 5 mm with guard column (Dr.
Maisch GmbH, Ammerbuch-Entringen, Germany); gradient 10–90%
CH3CN/20 mm NH4OAc.
Preparation of tissue samples: For the investigation of radiometa-
bolites, radiotracers [18F]1–[18F]4 (~200–300 MBq) were adminis-
tered to female CD-1 mice by tail vein injection. After 30 or 60 min,
brain and liver samples were homogenized and plasma and urine
were directly used. Mouse brain or liver was homogenized with
the Potter (500–800 rpm, 10 up-and-down strokes) in six volumes
of cold sucrose (0.3m). The suspension was centrifuged at 1000 g
for 10 min at 48C. The supernatant was separated and centrifuged
at 23500 g for 20 min at 48C. The pellet was resuspended in 5–6
volumes of buffer (500 mm Tris buffer, pH 7.4) and centrifuged
Author contributions
C.W., D.S., P.B., U.K., and B.W. participated in the research design;
C.W., E.G.M., A.H., S.F., F.A.L., W.D.-C., C.D., and L.B. contributed ex-
periments; F.G. contributed to the reagents or analytical tools; F.G.
and C.D. performed data analysis; A.H., S.F., U.K., and B.W. wrote or
contributed to the writing of the manuscript. The authors declare
no conflicts of interest.
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
13
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ

Full Papers
[22] E. J. Cobos, J. M. Entrena, F. R. Nieto, C. M. Cendꢄn, E. Del Pozo, Curr.
Neuropharmacol. 2008, 6, 344–366.
Acknowledgements
[23] T. Hayashi, T. P. Su, CNS Drugs 2004, 18, 269–284.
[24] T. Maurice, J. Meunier, B. Feng, J. Ieni, D. T. Monaghan, J. Pharmacol.
Exp. Ther. 2006, 317, 606–614.
[25] X. Guitart, X. Codony, X. Monroy, Psychopharmacology 2004, 174, 301–
319.
[26] J. L. Diaz, D. Zamanillo, J. Corbera, J. M. Baeyens, R. Maldonado, M. A.
Pericas, J. M. Vela, A. Torrens, Cent. Nerv. Syst. Agents Med. Chem. 2009,
9, 172–183.
[27] J. M. Entrena, E. J. Cobos, F. R. Nieto, C. M. Cendan, G. Gris, E. Del Pozo,
D. Zamanillo, M. Baeyens, Pain 2009, 143, 252–261.
This work was performed within the framework of the Interna-
tional Research Training Group (IRTG) ‘Complex Functional Sys-
tems in Chemistry: Design, Synthesis and Applications’ in collabo-
ration with Nagoya University. Financial support of the IRTG and
this project by the Deutsche Forschungsgemeinschaft (DFG) is
gratefully acknowledged. The authors thank Prof. Dr. J. Kirchmair
for providing the FAME software.
[28] B. Wꢀnsch, J. Med. Chem. 2012, 55, 8209–8210.
[29] P. Brust, W. Deuther-Conrad, K. Lehmkuhl, H. Jia, B. Wꢀnsch, Current
Med. Chem. 2014, 21, 35–69.
[30] A. Maisonial, E. Große Maestrup, C. Wiese, A. Hiller, D. Schepmann, S.
Fischer, W. Deuther-Conrad, J. Steinbach, P. Brust, B. Wꢀnsch, Bioorg.
Med. Chem. 2012, 20, 257–269.
Keywords: biotransformation · electrochemistry · fluorination ·
metabolism · oxidation · positron emission tomography · s1
receptor ligands
[31] K. Holl, D. Schepmann, C. G. Daniliuc, B. Wꢀnsch, Tetrahedron: Asymme-
try 2014, 25, 268–277.
[32] E. Große Maestrup, C. Wiese, D. Schepmann, P. Brust, B. Wꢀnsch, Bioorg.
Med. Chem. 2011, 19, 393–405.
[33] S. Fischer, C. Wiese, E. Große Maestrup, A. Hiller, W. Deuther-Conrad, M.
Scheunemann, D. Schepmann, J. Steinbach, B. Wꢀnsch, P. Brust, Eur. J.
Nucl. Med. Mol. Imaging 2011, 38, 540–551.
[34] K. Holl, E. Falck, J. Kçhler, D. Schepmann, H.-U. Humpf, P. Brust, B.
Wꢀnsch, ChemMedChem 2013, 8, 2047–2056.
[1] G. P. Martin, C. Marriott, I. W. Kellaway, J. Pharm. Pharmacol. 1976, 28,
76P.
[2] T. P. Su, J. Pharmacol. Exp. Ther. 1982, 223, 284–290.
[3] D. B. Vaupel, Eur. J. Pharmacol. 1983, 92, 269–274.
[4] R. Quirion, W. D. Bowen, Y. Itzhak, L. Junien, J. M. Musacchio, R. B. Roth-
man, T. P. Su, S. W. Tam, D. P. Taylor, Trends Pharmacol. Sci. 1992, 13, 85–
86.
[5] S. B. Hellewell, A. Bruce, G. Feinstein, J. Orringer, W. Williams, W. D.
Bowen, Eur. J. Pharmacol. 1994, 268, 9–18.
[35] A. Maisonial-Besset, U. Funke, B. Wenzel, S. Fischer, K. Holl, J. Steinbach,
P. Brust, B. Wꢀnsch, Appl. Radiat. Isot. 2014, 84, 1–7.
[36] E. Große Maestrup, S. Fischer, C. Wiese, D. Schepmann, A. Hiller, W.
Deuther-Conrad, J. Steinbach, B. Wꢀnsch, P. Brust, J. Med. Chem. 2009,
52, 6062–6072.
[37] C. Wiese, E. Große Maestrup, D. Schepmann, S. Grimme, H.-U. Humpf, P.
Brust, B. Wꢀnsch, Chirality 2011, 23, 148–154.
[38] A. Maisonial, E. Große Maestrup, S. Fischer, A. Hiller, M. Scheunemann,
C. Wiese, D. Schepmann, J. Steinbach, W. Deuther-Conrad, B. Wꢀnsch, P.
Brust, ChemMedChem 2011, 6, 1401–1410.
[39] J. Kirchmair, M. J. Williamson, A. M. Afzal, J. D. Tyzack, A. P. Choy, A.
Howlett, P. Rydberg, R. C. Glen, J. Chem. Inf. Model. 2013, 53, 2896–
2907.
[40] L. Bꢀter, M. Vogel, U. Karst, TrAC Trends Anal. Chem. 2015, 70, 74–91.
[41] U. Jurva, H. V. Wikstrçm, L. Weidolf, A. P. Bruins, Rapid Commun. Mass
Spectrom. 2003, 17, 800–810.
[42] H. Faber, M. Vogel, U. Karst, Anal. Chim. Acta 2014, 834, 9–21.
[43] E. Nouri-Nigjeh, R. Bischoff, A. P. Bruins, H. P. Permentier, Curr. Drug
Metab. 2011, 12, 359–371.
[44] F. T. van den Brink, L. Bꢀter, M. Odijk, W. Olthuis, U. Karst, A. van den
Berg, Anal. Chem. 2015, 87, 1527–1535.
[6] R. Kekuda, P. D. Prasad, Y. J. Fei, F. H. Leibach, V. Ganapathy, Biochem.
Biophys. Res. Commun. 1996, 229, 553–558.
[7] F. F. Moebius, R. J. Reiter, M. Hanner, H. Glossmann, Br. J. Pharmacol.
1997, 121, 1–6.
[8] S. Brune, S. Pricl, B. Wꢀnsch, J. Med. Chem. 2013, 56, 9809–9819.
[9] E. Laurini, V. Dal Col, M. G. Mamolo, D. Zampieri, P. Posocco, M. Ferme-
glia, V. Vio, S. Pricl, ACS Med. Chem. Lett. 2011, 2, 834–839.
[10] S. Brune, D. Schepmann, K.-H. Klempnauer, D. Marson, V. Dal Col, E.
Laurini, M. Fermeglia, B. Wꢀnsch, S. Pricl, Biochemistry 2014, 53, 2993–
3003.
[11] T. Hayashi, S. Y. Tsai, T. Mori, M. Fujimoto, T. P. Su, Expert Opin. Ther. Tar-
gets 2011, 15, 557–577.
[12] T. Hayashi, T. P. Su, Cell 2007, 131, 596–610.
[13] T. P. Su, T. Hayashi, T. Maurice, S. Buch, A. E. Ruoho, Trends Pharmacol.
Sci. 2010, 31, 557–566.
[14] P. D. Lupardus, R. A. Wilke, E. Aydar, C. P. Palmer, Y. Chen, A. E. Ruoho,
M. B. Jackson, J. Physiol. 2000, 526, 527–539.
[15] M. Johannessen, S. Ramachandran, L. Riemer, A. Ramos-Serrano, A. E.
Ruoho, M. B. Jackson, Am. J. Physiol. Cell Physiol. 2009, 296, C1049–
C1057.
[16] T. Maurice, T. P. Su, Pharmacol. Ther. 2009, 124, 195–206.
[17] G. Navarro, E. Moreno, M. Aymerich, D. Marcellino, P. J. McCormick, J.
Mallol, A. Cortes, V. Casado, E. I. Canela, J. Ortiz, K. Fuxe, C. Lluis, S.
Ferre, R. Franco, Proc. Natl. Acad. Sci. USA 2010, 107, 18676–18681.
[18] S. J. Nuwayhid, L. L. Werling, J. Pharmacol. Exp. Ther. 2003, 304, 364–
369.
[45] H. Pham-Tuan, L. Kaskavelis, C. A. Daykin, H.-G. Janssen, J. Chromatogr.
B 2003, 789, 283–301.
[46] H. Mascher, Klinische Analytik mit HPLC: Ein Ratgeber fꢀr die Praxis,
Wiley-VCH, Weinheim, 2010.
[47] M. J. Welch, D. Y. Chi, C. J. Mathias, M. R. Kilbourn, J. W. Brodack, J. A.
Katzenellenbogen, Nucl. Med. Biol. 1986, 13, 523–526.
[48] M. M. Bradford, Anal. Biochem. 1976, 72, 248–254.
[19] C. Ela, J. Barg, Z. Vogel, Y. Hasin, Y. Eilam, J. Pharmacol. Exp. Ther. 1994,
269, 1300–1309.
[20] S. A. Wolfe, Jr., S. G. Culp, E. B. De Souza, Endocrinology 1989, 124,
1160–1172.
Received: July 19, 2016
[21] S. A. Wolfe, Jr., C. Kulsakdinun, G. Battaglia, J. H. Jaffe, E. B. De Souza, J.
Pharmacol. Exp. Ther. 1988, 247, 1114–1119.
Revised: August 29, 2016
Published online on && &&, 0000
&
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
14
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ÝÝ These are not the final page numbers!

FULL PAPERS
C. Wiese, E. Große Maestrup, F. Galla,
D. Schepmann, A. Hiller, S. Fischer,
F.-A. Ludwig, W. Deuther-Conrad,
C. K. Donat, P. Brust, L. Bꢀter, U. Karst,
B. Wꢀnsch*
&& – &&
Taking a close look: The biotransforma-
tion of a homologous series of fluo-
roalkyl-substituted s1 receptor ligands
designed as PET tracers was investigat-
ed at different levels: in silico using
FAME software, electrochemically, in
vitro, and in vivo. Electrochemical oxida-
tion and in vitro studies with rat liver
microsomes combined with LC–MS
analysis allowed identification of the
metabolite structure. The nature and
amount of radiometabolites in various
tissues were analyzed by radio-HPLC
and radio-TLC after application of the
[18F]-labeled PET tracers to mice.
Comparison of in Silico,
Electrochemical, in Vitro and in Vivo
Metabolism of a Homologous Series of
(Radio)fluorinated s1 Receptor Ligands
Designed for Positron Emission
Tomography
ChemMedChem 2016, 11, 1 – 15
www.chemmedchem.org
15
ꢁ 2016 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
&
These are not the final page numbers! ÞÞ