Radiosynthesis and in vivo evaluation of a fluorine-18 labeled pyrazine based radioligand for PET imaging of the adenosine A2B receptor

Article abstract of DOI:10.1016/j.bmc.2018.07.045

On the basis of a pyrazine core structure, three new adenosine A_2B receptor ligands (7a–c) were synthesized containing a 2-fluoropyridine moiety suitable for ¹⁸F-labeling. Compound 7a was docked into a homology model of the A_2B<

Full text of DOI:10.1016/j.bmc.2018.07.045

Accepted Manuscript
Radiosynthesis and in vivo evaluation of a fluorine-18 labeled pyrazine based
radioligand for PET imaging of the adenosine A2B receptor
Marcel Lindemann, Sonja Hinz, Winnie Deuther-Conrad, Vigneshwaran
Namasivayam, Sladjana Dukic-Stefanovic, Rodrigo Teodoro, Magali
Toussaint, Mathias Kranz, Cathleen Juhl, Jörg Steinbach, Peter Brust, Christa
E. Müller, Barbara Wenzel
PII:
S0968-0896(18)30987-8
https://doi.org/10.1016/j.bmc.2018.07.045
BMC 14482
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
23 May 2018
23 July 2018
26 July 2018
Please cite this article as: Lindemann, M., Hinz, S., Deuther-Conrad, W., Namasivayam, V., Dukic-Stefanovic, S.,
Teodoro, R., Toussaint, M., Kranz, M., Juhl, C., Steinbach, J., Brust, P., Müller, C.E., Wenzel, B., Radiosynthesis
and in vivo evaluation of a fluorine-18 labeled pyrazine based radioligand for PET imaging of the adenosine A2B
receptor, Bioorganic & Medicinal Chemistry (2018), doi: https://doi.org/10.1016/j.bmc.2018.07.045
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Radiosynthesis and in vivo evaluation of a fluorine-18 labeled pyrazine based
radioligand for PET imaging of the adenosine A2B receptor
Marcel Lindemann1, Sonja Hinz2, Winnie Deuther-Conrad1, Vigneshwaran Namasivayam2,
Sladjana Dukic-Stefanovic3, Rodrigo Teodoro1, Magali Toussaint1, Mathias Kranz1, Cathleen
Juhl3, Jörg Steinbach1, Peter Brust1, Christa E. Müller2, Barbara Wenzel1#
1 Helmholtz-Zentrum Dresden-Rossendorf, Institute of Radiopharmaceutical Cancer Research,
Department of Neuroradiopharmaceuticals, Leipzig, Germany;
2
Pharma Center Bonn, Pharmaceutical Institute, Pharmaceutical Chemistry I, University of
Bonn, Bonn, Germany;
3 ROTOP Pharmaka GmbH, Dresden, Germany
# corresponding author: b.wenzel@hzdr.de
Abstract
On the basis of a pyrazine core structure, three new adenosine A2B receptor ligands (7a – c)
18
were synthesized containing a 2-fluoropyridine moiety suitable for F-labeling. Compound
7a was docked into a homology model of the A2B receptor based on X-ray structures of the
related A2A receptor, and its interactions with the adenosine binding site were rationalized.
Binding affinity data were determined at the four human adenosine receptor subtypes.
Despite a rather low selectivity regarding the A1 receptor, 7a was radiolabeled as the most
suitable candidate (Ki(A2B) = 4.24 nM) in order to perform in vivo studies in mice with the aim
to estimate fundamental pharmacokinetic characteristics of the compound class. Organ
distribution studies and a single PET study demonstrated brain uptake of [18F]7a with a
standardized uptake value (SUV) of ≈1 at 5 min post injection followed by a fast wash out.
Metabolism studies of [18F]7a in mice revealed the formation of a blood-brain barrier
penetrable radiometabolite, which could be structurally identified. The results of this study
provide an important basis for the design of new derivatives with improved binding
properties and metabolic stability in vivo.
Keywords: adenosine A2B receptor, 18F-labeling, pyrazines, metabolism, micellar
chromatography
1

1. Introduction
Adenosine is a purine nucleoside that is released from cells or generated extracellularly
following cleavage of adenosine 5′-triphosphate (ATP) by ectonucleotidases.1 It acts as a
signaling molecule and activates four subtypes of G protein-coupled receptors namely,
adenosine A1, A2A, A2B, and A3 receptors. Each of these receptors exhibits distinct
pharmacological properties, cell and tissue distribution, and intracellular signaling.2, 3
Activation of the adenosine A2B receptor subtype mediates a stimulation of the adenylate
cyclase through coupling to the Gs protein and results in an increase in intracellular cAMP. In
addition to Gs coupling, A2B receptors also activate phospholipase C by coupling to Gq
proteins leading to an increase in intracellular calcium concentration.4 Compared to the
other adenosine receptors, the A2B receptor is a low-affinity receptor and only activated at
adenosine concentrations in the micromolar range.5, 6 It might therefore not be activated
under physiological conditions. However, pathological events (e.g. inflammation, cancer,
infection) can lead to elevated adenosine concentrations and increased expression of this
receptor subtype.4 As an example, the contribution of the A2B receptor in cancer progression
via mechanisms such as tumor growth, angiogenesis, metastasis and immunomodulation
was demonstrated, and its potential as a therapeutic target has recently been reviewed.4, 7-10
The tissue distribution and cell expression of A2B receptors have already been investigated by
detecting of corresponding mRNA as reviewed elsewhere.6, 11, 12 Accordingly, a widespread
distribution of the A2B receptor gene expression in rats was suggested being highest in colon,
cecum, lung, uterus, eye, and bladder.13, 14 A ubiquitous, low level of mRNA was observed for
the brain areas investigated in these studies. In the last decade, A2B receptor expression was
also studied in various human tumor cell lines and was found to be high e.g. in prostate,
colon carcinoma, neuroendocrine, breast cancer and glioblastoma U87MG cells.15-19
However, a closer examination of those reports reveals, that there are only very few
quantitative data available regarding the actual density of the A2B receptor protein, neither
for peripheral organs nor for the brain. To our knowledge, Gessi et al. reported in 2005 on
saturation binding experiments using the A2B receptor specific radioligand [3H]MRE 2090-
F20.20 With the determination of Bmax and KD values they characterized the expression of the
A2B receptor on neutrophils and lymphocytes, cells which are strongly involved in
2

inflammation. Recently, Hinz et al. reported on A2B receptor expression determined by the
radioligand [3H]PSB-603 in native lymphocytes and cancer cell lines.21
The main objective of our work is to characterize the expression of the A2B receptor by using
positron emission tomography (PET) as a non-invasive molecular imaging technique. The
development of a specific radioligand labeled with the PET radionuclide fluorine-18 would
support the investigation of this promising target in vitro and in vivo. In particular, it would
also be possible to examine the contribution of the A2B receptor in brain tumors and could
therefore serve as a basis for a targeting based therapy.
To develop a suitable PET radioligand for this purpose, a lead compound is needed which has
i) a high affinity and selectivity for the A2B receptor and ii) the potential to cross the blood-
brain barrier (BBB). Several A2B receptor antagonists have been developed in the past and
were reviewed elsewhere.22-24 Most of them are based on the xanthine scaffold such as
MRS175425, PSB-111526, CVT-6883 (GS 6201)27 and PSB-60328 (Figure 1). Based on a
pyrimidine core, Vidal et al. developed LAS38096, a potent non-xanthine based compound
which was later structurally optimized leading to the discovery of the pyrazine derivative
LAS101057.29, 30 Recently, the group of Sotelo et al. identified with ISAM-140 another novel
class of A2B receptor antagonists.31 In 2016, a carbon-11 labeled triazinobenzimidazole
based compound was published,32 the only A2B receptor targeting PET tracer developed so
far ([11C]2 in Figure 1). With an IC50 value of 210 nM obtained with a functional assay by
measuring cAMP accumulation, [11C]2 demonstrates moderate binding toward the A2B
receptor. First PET studies in rats showed highest uptake in brown adipose tissue, lungs and
testicles. The authors concluded that this compound needs to be further characterized, but
may represent an appropriate lead compound for further developments.
In the present study, the pyrazine-based compound class published by Eastwood & Vidal et
al.29 was chosen, due to the compounds' potential to penetrate the blood-brain barrier.
Accordingly, derivative 1 was selected as a lead compound due to its high affinity and
selectivity toward the A2B receptor (Figure 1 & Table 1). Since both fluorine atoms are in
positions, which are not activated for a straightforward nucleophilic 18F-labeling, it was
decided to substitute the 3-fluoropyridine by a 2-fluoropyridine ring at C5 of the pyrazine
ring. Thus, with the new derivative 7a (Scheme 1) a regioisomer of 1 was developed, in
which the fluorine atom is in the ortho-position to the pyridine nitrogen. Moreover, the
aromatic groups at C6 of the pyrazine ring were modified in order to investigate their
3

influence on the binding profile. In compound 7b a more polare pyridine and in 7c a
nonpolar phenyl ring was substituted (Scheme 1).
Figure 1. Selected known A2B receptor antagonists, PET radioligand [11C]2 and lead compound 1
2. Results and discussion
2.1. Ligand synthesis
Three novel derivatives 7a - c were synthesized which differ regarding their aromatic
substituents in position 6 of the pyrazine core. In addition, lead compound 1 was synthesized
as a reference29 to investigate the influence of structural modifications on the in vitro
binding affinity toward the four adenosine receptor subtypes.
The syntheses of 1 and the new derivatives 7a - c (Scheme 1) were performed based on the
literature.29, 33 Accordingly, the synthesis route started with the formation of 429 by applying
a regioselective iodination of chloropyrazine 3 under mild conditions as described.29 To
obtain 5a29 and 5b, the following metal-catalyzed cross-coupling reaction was performed by
two different methods. First, the Stille reaction was applied by using 3-fluoro-4-
(tributylstannyl)pyridine and [Pd(PPh3)2Cl2]/CuI as described.29 Second, the Suzuki reaction
4

was used, because it is beneficial regarding handling and toxicity aspects. Both reactions
showed similar yields in the range of 50-60%. Thereafter, the amides 6a29 and 6b were
synthesized in pyridine using cyclopropylcarbonyl chloride. In the last step a second Suzuki
reaction of 6a and 2-fluorophenylboronic acid was performed leading to compound 1 in
good yields. The amide 6b served as a precursor molecule for the coupling of different
aromatic rings resulting in the new derivatives 7a - c with moderate to high yields.
Scheme 1. Synthesis of lead compound 1 and new derivatives 7a - c
Reagents and conditions: (i) N-iodosuccinimide, DMSO/H2O, 88%; (ii) 5a: 3-fluoro-4-(tributylstannyl)pyridine,
[Pd(PPh3)2Cl2], CuI, DMF, 150 °C, 61% or 3-fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolane-2-yl)pyridine, 2 M
Cs2CO3, [Pd(dppf)Cl2], 1,4-dioxane, reflux, 60%; 5b: 2-fluoropyridine-4-boronic acid, 2 M Cs2CO3, [Pd(dppf)Cl2],
1,4-dioxane, 90 °C, 2 h, 99%; (iii) cyclopropanecarbonyl chloride, pyridine, 80 °C; (iv) 2 M Cs2CO3, [Pd(dppf)Cl2],
1,4-dioxane, 1: 2-fluorophenylboronic acid, 85 °C, 5.5 h, 64%; 7a: 2-fluorophenylboronic acid, 90 °C, 7 h, 94%;
7b: pyridine-4-boronic acid, 90 °C, 2.5 h, 51%; 7c: phenylboronic acid, 90 °C, 6 h, 61%.
2.2. Determination of in vitro binding affinities
The binding affinities of 1 and 7a – c toward the four adenosine receptor subtypes A2B, A2A,
A1, and A3 were determined in competition binding assays using membrane preparations of
Chinese hamster ovary (CHO) cells expressing the corresponding human receptor subtype
and appropriate radioligands (data in Table 1).
5

In our assay, lead compound 1 showed a high affinity for the A2B receptor (Ki = 1.09 nM)
which is similar to the Ki value of 0.6 nM reported by Eastwood et al.29 In general, except for
the A1 receptor, the obtained binding data of 1 are similar to the published data. However,
the binding affinity of 1 for the A1 receptor was 4.6-fold higher than reported (Ki = 28.3 nM
versus published 130 nM29), which could be due to the different binding assays used. This
unexpected drop in selectivity questions the general suitability of 1 as a lead compound.
Compared to 1, the new derivative 7a differs in the position of the fluorine substitution at
the pyridine ring (3-position in 1 versus 2-position in 7a). This modification resulted in slight
differences in the binding affinities toward the four receptor subtypes. While the affinity of
7a to the A2B receptor remains high (Ki = 4.24 nM), the selectivities toward A2A and A1 are
decreased resulting in Ki ratios of 1:13 (A2B/A2A) and 1:5 (A2B/A1).
Eastwood et al.29 already performed a structure-activity relationship study by substituting
different aromatic and heteroaromatic ring systems at the 5- and 6-position of the pyrazine
core. By using a pyridine substituent in the 6-position a considerable decrease in affinity for
A2A, A1, and A3 receptors was observed resulting in an increase in A2B selectivity. This
prompted us to also modify 7a by exchanging the 2-fluorophenyl moiety by a pyridine ring to
obtain 7b. However, this modification caused not only the desired decrease in affinities
toward the A2A, A1 and A3 receptors, but also a considerable loss in affinity for the A2B
receptor (Ki = 296 nM). Assuming that the rather polar heteroaromatic ring caused this
reduction in binding, which was supported by docking studies (see chapter 2.3.), we
synthesized derivative 7c with a more hydrophobic phenyl ring. With this modification we
could indeed obtain a high affinity toward the desired target (Ki = 1.78 nM), however, also
the binding affinities to the A2A and A1 receptors increased considerably resulting in lower
selectivities compared to 7a.
Table 1: Ki values of 1 and 7a - c toward the four adenosine receptor subtypes.
Ki in nM a
R1 R2 R3
X
A2B
A2A
A1
A3
compound
6

1
H
F
F
C
1.09 ± 0.20
0.6 ± 0.0b
88.9 ± 25.5
101 ± 33 b
55.0 ± 6.1
1930 ± 400
10.2 ± 1.5
28.3 ± 8.2
130 ± 44 b
19.0 ± 5.2
644 ± 62
937 ± 158
b, c
8 ± 6%
7a
7b
7c
F
F
F
H
H
H
F
H
H
C
N
C
4.24 ± 0.04
296 ± 9.0
796 ± 26
>1000
1.78 ± 0.25
4.29 ± 0.15
680 ± 118
a Data are means ± SEM of three independent assays performed in duplicates. Competition binding assays with
[3H]PSB-603 (A2B), [3H]MSX-2 (A2A), [3H]CCPA (A1), and [3H]PSB-11 (A3) as radioligands and membranes obtained
from CHO cells stably expressing the corresponding human adenosine receptor. b Data are taken from
literature29 , c (%) inhibition at 1 μM of [3H]NECA29.
2.3. Molecular modelling studies
In order to propose binding modes for the pyrazine derivatives and to rationalize their
affinity and selectivity data we selected compound 7a for performing molecular docking
studies using a recently created homology model of the A2B receptor.34 The putative binding
mode of 7a and the important residues in the binding pocket of the A2B receptor are shown
in Figure 2A and 2B, respectively. The model shows that the compound is anchored inside
the binding cleft by two key interactions found in all adenosine receptor subtypes. The
pyrazine moiety of 7a likely forms one of the key π-π stacking interactions with F173 and
utilizes the hydrophobic surface provided by I276. The second plausible key interaction is the
formation of two hydrogen bonds, between the N1 of the pyrazine moiety and the NH of the
carboxamido group of 7a with N254. The pyridine ring attached to the pyrazine (in position
5) is directed towards a pocket formed by the amino acid residues A64, V85, I276 and H280.
The nitrogen atom of the pyridine ring possibly introduces a hydrogen bond interaction with
H280. The 2-fluorophenyl residue attached in the 6-position of the pyrazine ring is expected
to be in proximity to the highly conserved W247 and is also found to occupy the subpocket
formed by V85, L86, T89, M182, W247, V250, and H251. The residues L86, M182, W247 and
V250 possibly contribute to hydrophobic interactions with the phenyl ring. The hydrophobic
residues in the subpocket favor interactions with the unsubstituted phenyl ring (in 7c), which
showed increased binding affinity in comparison to the more polar pyridine ring (in 7b). The
docking results also indicated that the cyclopropylcarboxamide residue of 7a further extends
to a binding pocket formed by another set of residues comprised of F173, E174, M179,
N254, T257, N266, K269 and M272. Among the amino acid residues interacting with 7a
7

which were discussed above, V250, N266 and K269 are unique for the A2B receptor subtype
and could play an important role in determining A2B selectivity (see sequence alignment of
the adenosine receptors in Supporting Information, Figure S1).
A
B
7a
TM4
TM3
TM2
TM1
7a
TM5
TM6
TM7
Figure 2. Putative binding mode of 7a in a homology model of the A2B receptor. A) The A2B receptor is
represented as cartoon model, and atoms of the docked ligand (7a, purple) are depicted as sticks. B) The atoms
of the docked ligand (7a, purple) with important amino acid residues (orange) are depicted as sticks in the
binding pocket of the A2B receptor. The amino acids which are unique for the A2B receptor are framed in blue
and important interactions are indicated by yellow dotted lines. The oxygen atoms are colored in red, nitrogen
atoms in blue, sulfur atoms in yellow and fluorine in cyan.
2.4. Precursor synthesis and radiochemistry
Despite the rather moderate in vitro binding profile of the new derivatives, we decided to
radiolabel 7a as the most promising candidate. The main aim was to examine the general
suitability of this substance class by investigating brain uptake and in vivo metabolism.
2.4.1. Synthesis of the precursor
8

Based on our own experiences,35 we decided to use a precursor with a nitro leaving group
18
for the intended F-labeling. The precursor 13 was generated according to the three step
synthesis route depicted in Scheme 2. Starting from compound 429, the amide 8 was
synthesized using the same method as for the amides 6a/6b. In the next step, the
nitropyridine ring was introduced via a Suzuki reaction of 8 with the dioxaborolane 1136 to
obtain compound 12. The borolane 11 was generated in a two-step synthesis, starting with
the oxidation of 9 to obtain the known 4-bromo-2-nitropyridine 10 according to a procedure
already described.37 The following Miyaura borylation was performed using [Pd(dppf)Cl2] as
catalyst as reported.35, 36 In the last step, the 2-fluorophenyl moiety was substituted into
compound 12 by using a second Suzuki reaction to obtain the precursor 13.
Scheme 2. Synthesis of the precursor 13
Reagents and conditions: (i) cyclopropanecarbonyl chloride, pyridine, rt, 5.5 h, 90%; (ii) conc. H2SO4, 30% H2O2,
rt, 1 d, 38%; (iii) 4,4,4',4',5,5,5',5'-octamethyl-2,2'-bi-1,3,2-dioxaborolane, KOAc, [Pd(dppf)Cl2], 1,4-dioxane,
reflux, 45 min, 85% (iv) 2 M K2CO3, [Pd(dppf)Cl2], 90 °C, 5 h, 49%; (v) 2-fluorophenylboronic acid, 2 M Cs2CO3,
[Pd(dppf)Cl2], 90 °C, 3 h, 61%.
2.4.2. Radiosynthesis and characterization of [18F]7a
Manual synthesis - optimization
For the synthesis of the new radioligand [18F]7a, a nucleophilic aromatic substitution
reaction of the nitro group of precursor 13 and [18F]fluoride as azeotropically dried kryptofix-
carbonate complex was performed. Based on our former experiments with similar aromatic
NO2/[18F]F- systems,35, 38 we selected dimethylsulfoxide (DMSO) and N,N-dimethylformamide
9

(DMF) as the most promising solvents. In a first screening, [18F]7a was obtained with
comparable results using DMSO and DMF. We went further with DMSO achieving
radiochemical yields (RCY) for the labeling process of 58 ± 7% after 20 min at 150 °C (Scheme
3). A considerable decrease of the RCY was observed when reducing the temperature to 120
°C. According to radio-TLC analysis, two radioactive by-products were formed in the reaction
mixture accounting for less than 5% of total radioactivity. The precursor remained stable
under all conditions tested as proven by HPLC analysis.
Scheme 3. 18F-labeling of [18F]7a
All labeling experiments were performed with a very low amount of precursor (1.0 - 1.2 mg),
which was important for the subsequent isolation of the radiotracer by semi-preparative
18
HPLC. The difficulties in the separation arise from the chromatographic similarity of the F-
labeled compound and its nitro precursor. For the isolation of [18F]7a a cyanopropyl-
modified column was found to be appropriate using a mixture of acetonitrile (ACN)/water
and ammonium acetate as buffer (Figure 3A). By contrast, no separation could be achieved
using common strong hydrophobic or polar C18-phases.39
Automated radiosynthesis of [18F]7a
On the basis of the manual experiments, the radiosynthesis of [18F]7a was translated to an
automated procedure using the TRACERlab FX2 N synthesis module (GE Healthcare). The
synthesizer setup is described in the experimental part. Briefly, after trapping and elution of
[18F]fluoride from an anion exchange cartridge, the labeling reaction of the azeotropically
dried [18F]F-/K2.2.2./K2CO3 complex with the precursor 13 was performed in DMSO for 15 min
at 150 °C. For isolation of [18F]7a, the crude reaction mixture was diluted with aqueous
acetonitrile and directly applied to a semi-preparative HPLC system. The radiotracer fraction
10

was collected at a retention time of 30 - 35 min and subjected to a solid phase extraction
(SPE) using a C18-cartridge. The obtained radiotracer eluate was transferred out of the hot
cell, concentrated and formulated in sterile isotonic saline containing 10% of ethanol. The
entire process lasts about 80 - 90 min. Activity balance showed that less than 10% of the
radioactivity was lost in the anion exchange cartridge, the reactor and during SPE. Finally,
[18F]7a could be reproducibly produced with a high radiochemical purity of ≥ 99%, a
radiochemical yield of 38.9 ± 1.5% (n = 3) and molar activities between 6 and 8 GBq/µmol (n
= 3) at starting activities of 2-3 GBq. Co-injection of the corresponding reference compound
7a confirmed the identity of [18F]7a by analytical radio-HPLC (Figure 3B).
Figure 3. A) Semi-preparative radio- and UV-HPLC chromatograms of [18F]7a (conditions: Reprosil-Pur CN,
250x20 mm, 40% ACN/20 mM NH4OAcaq., 8.0 mL/min). B) Analytical radio- and UV-HPLC chromatograms of the
final product of [18F]7a spiked with the non-radioactive reference 7a (conditions: Reprosil-Pur C18-AQ, 250 x
4.6 mm, gradient with an eluent mixture of ACN/20 mM NH4OAc aq., 1.0 mL/min).
Stability and logD value of [18F]7a
The stability of the radiotracer was investigated by incubation at 40 °C in n-octanol and pig
plasma samples. [18F]7a proved to be stable in these media, and no defluorination or
degradation was observed within 60 min of incubation time.
To estimate the lipophilicity of [18F]7a, the logD value was determined by the shake flask
method using n-octanol and phosphate-buffered saline as partition system. With an
obtained logD value of 2.07 ± 0.22 (n = 4) the radiotracer shows a promising physicochemical
characteristic with respect to potential PET imaging of the brain.40-42
2.5. Specific and selective binding of [18F]7a in vitro
11

To evaluate if [18F]7a binding is (i) specific and (ii) selective to adenosine A2B receptors,
mouse brain and heart homogenates were used for binding studies. We observed
substantial binding of [18F]7a at 4 nM which was significantly blocked by 1 µM 7a and 1 µM
PSB-603 (brain: -80% and -30%; heart: -50% and -20%, respectively). While the blocking
observed with 7a indicates specific binding of [18F]7a in rat brain and heart, the
comparatively weak efficacy of the selective A2B inhibitor PSB-603 indicates that A2B
receptors account for only a fraction of the specific binding sites of the radioligand. Taking
into account the nanomolar affinity of 7a to A1 and A2A receptors, further binding sites might
be related to these receptors which are both highly expressed in brain43, 44 as well as in
cardiomyocytes and cardiac fibroblasts.45, 46 A supplementary radioligand displacement
study (data not shown) using A1- and A2A-specific tritiated radioligands as well as membrane
homogenates of mouse brain and heart supported this assumption as 7a was able to
compete with the respective radioligand for the respective adenosine receptor binding site.
So far and mainly due to the herein detected lack of selectivity, we did not test [18F]7a for
relevant interaction with additional binding sites not related to adenosine receptors.
2.6. Preliminary in vivo studies of [18F]7a in mouse
2.6.1. Brain uptake and specific binding
Despite the insufficiency of selectivity observed in vitro, we decided to evaluate the ability of
[18F]7a to penetrate the blood-brain barrier to more thoroughly characterize the suitability
of this compound class for the development of PET radiotracers. Therefore, we initially
performed organ distribution studies and could show uptake of activity in whole brain
followed by washout (SUV at 5 min p.i.: 0.91 (n=2: 0.81 and 1.00); at 15 min p.i.: 0.60 (n=2:
0.50 and 0.69); and at 30 min p.i.: 0.22 ± 0.04 (n=3)). A preliminary dynamic PET scan
performed in one mouse accords with this result (data not shown). However, the
distribution of activity at scanning times corresponding to specific binding was rather
homogenous (Figure 4A), as also indicated by ex vivo autoradiography performed at 30 min
p.i. (Figure 4B). We assume that this result is related to both a low density of A2B receptors
under physiological conditions13 as well as to the low selectivity of the radioligand.
12

Figure 4. Analysis of the distribution of activity in mouse brain after i.v. injection of [18F]7a. A) Summed PET
image (2 – 15 min p.i. of 3.2 MBq [18F]7a) of a mouse brain with atlas superimposed (left). B) Autoradiographic
image of a sagittal brain slice (10 µm) obtained at 30 min p.i. of 4 MBq [18F]7a. C) Corresponding Nissl staining.
2.6.2. Metabolism studies
The metabolic stability of [18F]7a in vivo was analysed in mice. Blood samples and brains
were obtained at 5, 15, and 30 min post i.v. injection of the radiotracer (n = 2 per time
point), and plasma and brain samples acquired by centrifugation and homogenization,
respectively. Analysis of the samples was performed by using micellar chromatography
(MLC) and RP-HPLC. MLC allows to directly inject the samples into the HPLC system without
eliminating the tissue matrix due to the ability of micellar aggregates to dissolve the proteins
and other components.47 For RP-HPLC analysis, aliquots of plasma and brain homogenates
were treated with a mixture of acetone/water to precipitate the proteins resulting in
recoveries of about 90% and 95% of activity for the plasma and brain samples, respectively.
In general, the results obtained with both methods are comparable. A diagram summarizing
the chromatographic MLC data is shown in Figure 5A; Figure 5B shows the MLC and RP-HPLC
chromatogram of a plasma sample at 30 min p.i. Accordingly, intact radiotracer accounts for
~35% of total activity in plasma at this time point and four radiometabolites are detected.
Only small amounts of [18F]fluoride ([18F]M1) are formed, which can be seen in the radio-
MLC chromatogram. By contrast, in the corresponding RP-HPLC chromatogram almost no
[18F]fluoride could be detected, which is presumably caused by a loss of [18F]fluoride during
the extraction process due to a partially non-reversible binding on plasma proteins (lower
recovery in plasma samples, see exp. part). The retention of metabolite [18F]M2 is quite
13

different when comparing the two chromatograms, a phenomenon which is caused by the
differences of the separation mechanism of the two chromatographic systems.
Figure 5. A) Percentage of intact radiotracer in dependence of time after injection into mice (n = 2) obtained
from radio-MLC chromatographic data. B) Exemplary MLC and RP-HPLC radio-chromatograms of a mouse
plasma sample at 30 min p.i. of [18F]7a (conditions MLC: Reprosil-Pur C18-AQ, 250 x 4.6 mm, gradient with an
eluent mixture of THF/100 mM SDS/10 mM NaHPO4aq., 1.0 mL/min; conditions RP-HPLC: Reprosil-Pur C18-AQ,
250 x 4.6 mm, gradient with eluent mixture of ACN/20 mM NH4OAc aq., 1.0 mL/min).
The most hydrophobic radiometabolite [18F]M4 is obviously able to penetrate the BBB,
because it could be also detected in the brain samples (see B in Figure 6). Thus, 82% of the
measured radioactivity in brain at 15 min p.i. and 70% at 30 min p.i. can be assigned to
[18F]7a.
Eastwood et al.29 already investigated the in vitro metabolism of several pyrazine based
derivatives by using rat liver microsomes. They determined a turnover of 75% when
incubating lead compound 1 for 30 minutes, which corresponds to our in vivo results of
[18F]7a in plasma. The authors also described that N- and C-oxidation of the pyridine rings
are the major metabolic pathways in vitro and only minor amounts of metabolite are derived
from hydrolysis of the amide bond. Since we previously observed in own studies, that the
metabolic degradation in vitro is often faster compared to the in vivo metabolism, the rapid
conversion of [18F]7a was unexpected. Moreover, the ability of [18F]M4 to pass the blood-
brain barrier prompted us to investigate the molecular structure of this degradation product
in order to use this information for the future design of more stable compounds. Therefore,
we synthesized the corresponding non-radioactive hydrolysis product M4 (A in Figure 6). A
brain sample obtained at 30 min p.i. of [18F]7a in a mouse was spiked with M4 and subjected
to MLC measurement. As it can be seen in the corresponding chromatogram B in Figure 6,
14

this amine corresponds to the radiometabolite [18F]M4. The calculated logD value of M4 is
1.71 at physiological pH and could explain the observed permeation of [18F]M4 across the
BBB, since this is within a range which was described to be beneficial for good brain
penetration.40 With a Ki value of 63.0 nM, this radiometabolite still shows affinity for A2B
receptors although 15-fold lower than the parent compound 7a (Ki = 4.24 nM).
Figure 6. A) Synthesis of the nonradioactive metabolite compound M4 and its Ki values for the adenosine
receptor subtypes (reagents and conditions: (i) (2-fluorophenyl)boronic acid, 2 M Cs2CO3, [PdCl2(dppf)], 90 °C,
7 h, 72%). B) Analytical radio- and UV-MLC chromatogram of a mouse brain sample at 15 min p.i. of [18F]7a
spiked with the nonradioactive references M4 and 7a (conditions: Reprosil-Pur C18-AQ, 250 x 4.6 mm, gradient
with an eluent mixture of THF/100 mM SDS/10 mM NaHPO4 aq., 1.0 mL/min).
3. Summary and conclusion
The objective of this work was to develop an 18F-labeled radiotracer as an imaging probe for
characterizing the expression pattern of the adenosine A2B receptor in vitro and in vivo. As
our future intention is the use of such radiotracer for investigating brain tumors by non-
invasive imaging, we additionally turned our attention to the selection of compounds
promising sufficient blood-brain barrier penetration.
29
Starting with the pyrazine-based derivative 1 , which was described to have a high affinity
and selectivity for the A2B receptor, we developed the novel derivatives 7a-7c containing a 2-
fluoropyridine moiety suitable for nucleophilic 18F-labeling. However, the binding affinity
data of the lead compound 1 determined in our radioligand binding assays, did not entirely
correspond to the published data29; in particular with respect to the affinity toward the A1
receptor which was considerably higher than reported. This trend was also observed for the
new derivatives 7a-7c. Despite the resulting rather low selectivity, we radiolabeled 7a as
most suitable candidate in order to perform initial in vivo studies with the aim to get more
15

information on this substance class regarding brain uptake and metabolic stability. The
metabolism studies in mice revealed the formation of a brain penetrable radiometabolite,
which could be identified as the corresponding deacylated amine of [18F]7a generated by
hydrolysis of the amide bond. Preliminary organ distribution studies demonstrated a brain
uptake of [18F]7a at 5 min p.i. of which we assume to be sufficient for PET imaging.
Our findings show the potential of this substance class regarding brain penetration,
physicochemical properties, and pharmacokinetics. However, structural modifications are
necessary to improve the selectivity of the compounds toward the A2B receptor subtype as
well as the metabolic stability in vivo. Molecular modeling studies based on a recently
established homology model of the A2B receptor and published X-ray structures of adenosine
A1 and A2A receptors will support the future design of suitable candidates.
4. Material and methods
4.1. Organic chemistry
4.1.1. General
Analysis of all compounds was performed by MS, HPLC, TLC and NMR spectroscopy. The
spectra for the final compounds are given in the supplementary material.
High resolution mass spectra were recorded on an ESI-qTOF Impact II (Bruker Daltonik
GmbH) and on an ESI-TOF micrOTOF (Bruker Daltonik GmbH) using ElectroSpray Ionization
(ESI). NMR spectra (1H, 13C, 13C-APT, 19F, H,H-COSY, HSQC, HMBC) were recorded on
spectrometers from Varian (MERCURY plus 300 MHz, MERCURY plus 400 MHz) and Bruker
(AVANCE III HD 400 MHz, DRX-400 400 MHz, Fourier 300 300 MHz). Splitting patterns have
been designated as follows: s: singlet, d: doublet, bs: broad singlet, m: multiplet, t: triplet,
dd: doublet of doublet, td: triplet of doublet, dt: doublet of triplet.
Analytical thin-layer chromatography was performed on silica gel coated plates (Macherey-
Nagel, ALUGRAM SIL G/UV254). The spots were identified by using an UV lamp or by spraying
a solution of 0.1% ninhydrin in ethanol/water 1/10 or dipping into a KMnO4-solution (3 g
KMnO4, 20 g K2CO3, 0.25 ml glacial acid, 300 ml water).
For purification of final products flash column chromatography was used with silica gel
ZEOsorb 60/40-63 µm from Apollo Scientific Ltd. and silica gel 40-63 µm from VWR
Chemicals.
The chemical purity of the final compounds is ≥ 96% and was controlled by HPLC using a
150 x 3 mm Reprosil-Pur Basic HD – 3 µm column (Dr. Maisch GmbH, Germany). These
analytical chromatographic separations were performed on a Dionex Ultimate 3000 system,
incorporating a LPG-3400SD pump, an autosampler WPS-3000 TSL, a column compartment
TCC-3000SD, a diode array detector DAD3000 (monitoring from 254-720 nm) and a low
resolution mass spectrometer MSQ 3000 (Thermo Fisher Scientific Inc., Waltham, USA). A
mixture of acetonitrile (ACN) and aqueous 20 mM NH4OAc was used as eluent in a linear
16

gradient system (0-5 min at 25% ACN, 5-45 min up to 95% ACN, 45-55 min at 95% ACN, 55-
57 min up to 25% ACN, 57-65 min at 25%) with a flow of 0.6 mL/min.
Chemical names of compounds were generated by ChemDraw Professional 16.0.
4.1.2. Syntheses
Compound 3 was purchased from Alfa Aesar, Thermo Fisher, Germany. Compound 4 was
synthesized according to the literature.29
4.1.2.1. 6-Chloro-5-(3-fluoropyridine-4-yl)pyrazine-2-amine 5a
Under argon atmosphere, 100 mg of 4 (0.393 mmol, 1.0 eq.) and 130 mg (0.589 mmol,
1.5 eq.)
of
3-fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)pyridine
(from
Fluorochem, UK) were suspended in a mixture of 5 mL 1,4-dioxane and 1.2 mL of a 2 M
Cs2CO3 solution Afterwards, 28 mg (0.039 mmol, 0.1 eq.) [1,1′-bis(diphenyl-
phosphino)ferrocene]dichloropalladium(II) Pd(dppf)Cl2 (from Fluorochem, UK) was added
and the reaction mixture was refluxed for 10 h. Further 44 mg (0.196 mmol, 0.5 eq.) of 3-
fluoro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolan-2-yl)pyridine was added and the mixture
was again heated for 2 h. After pouring the mixture into 100 mL water and extraction with
ethyl acetate (4 × 25 mL), the combined organic layers were dried over sodium sulfate,
filtered and the solvent was evaporated under reduced pressure. The residue was purified
by column chromatography (SiO2, dichloromethane/methanol, 39/1, v/v) to obtain 5a with a
yield of 60% (53 mg, 0.235 mmol).
1
Rf (SiO2, dichloromethane/methanol, 19/1, v/v): 0.32; H NMR (DMSO-d6, 400 MHz) : 7.31
(s, 2 H); 7.55 (dd, J = 6.1 Hz, J = 5.2 Hz, 1 H); 7.97 (s, 1 H); 8.53 (d, J = 4.8 Hz, 1 H); 8.67 (d, J =
13
1.1 Hz, 1 H) ppm; C NMR (DMSO-d6, 75 Hz) : 126.2; 130.3; 131.1; 132.8 (d, J = 12.2 Hz);
138.7 (d, J = 24.3 Hz); 145.3; 146.3 (d, J = 5.1 Hz); 155.9; 156.5 (d, J = 256.6 Hz) ppm; 19F NMR
(DMSO-d6, 282 MHz) : -128.3 ppm; ESI-MS m/z calcd. for C9H7N4FCl+ [M+H]+ 225.0, found
225.1.
4.1.2.2. 6-Chloro-5-(2-fluoropyridine-4-yl)pyrazine-2-amine 5b
Under argon atmosphere, 50 mg of 4 (0.196 mmol, 1.0 eq.) and 41 mg of 2-fluoropyridine-4-
boronic acid (0.294 mmol, 1.5 eq., from OXCHEM Corporation, USA) was suspended in a
mixture of 3.6 mL 1,4-dioxane and 0.3 mL of a 2 M Cs2CO3 solution. Afterwards, 7.2 mg
(0.010 mmol, 0.05 eq.) Pd(dppf)Cl2 was added and the mixture was heated for 2 h at 90 °C.
After pouring the mixture into 100 mL water and extraction with ethyl acetate (3 × 25 mL),
the combined organic layers were dried over sodium sulfate, filtered and the solvent was
evaporated under reduced pressure. The residue was purified by column chromatography
(SiO2, dichloromethane/methanol, 39/1, v/v) to obtain 5b with a yield of 99% (44 mg,
0.196 mmol).
1
Rf (SiO2, dichloromethane/methanol, 39/1, v/v): 0.52; H NMR (DMSO-d6, 300 MHz) : 7.36
(s, 2 H), 7.41 (m, 1 H), 7.66 (ddd, J = 5.3 Hz, J = 2.1 Hz, J = 1.4 Hz, 1 H), 7.95 (s, 1 H), 8.27 (d,
J = 5.3 Hz, 1 H) ppm; 13C NMR (DMSO-d6, 100 Hz) : 108.9 (d, J = 39.3 Hz), 121.9 (d,
17

J = 3.8 Hz), 131.2, 133.1 (d, J = 4.2 Hz), 144.7, 147.8 (d, J = 15.6 Hz), 150.5 (d, J = 8.8 Hz),
19
155.6, 163.7 (d, J = 234.2 Hz) ppm; F NMR (DMSO-d6, 376 MHz) : -68.9 ppm; ESI-MS m/z
calcd. for C9H7N4FCl+ [M+H]+ 225.03, found 225.03, m/z calcd. for C9H6N4FClNa+ [M+Na]+
247.02, found 247.02.
4.1.2.3. General procedure for 6a and 6b:
Amine 5a or 5b (1 eq.) was dissolved in 0.5 mL dried pyridine and cyclopropanecarbonyl
chloride (3.0 eq.) was dropwise added to the solution under ice cooling. After heating for 1 h
at 80 °C, the reaction was monitored by TLC until it was completed. Then the solvent was
removed under reduced pressure, the residue dissolved in dichloromethane and washed
with 5% aqueous sodium hydrogen carbonate solution (1 × 50 mL). The organic layer was
separated, dried over sodium sulfate, filtered and concentrated under reduced pressure. The
desired product was purified by column chromatography (SiO2) using a solvent mixture of
CHCl3/ethyl acetate, 3/2 (v/v) to get 6a with a yield of 95% (25 mg, 0.085 mmol). A solvent
mixture of dichloromethane/methanol, 49/1 (v/v) was used to get 6b with a yield of 62% (13
mg, 0.044 mmol).
4.1.2.3.1. N-(6-Chloro-5-(3-fluoropyridine-4-yl)pyrazine-2-yl)cyclopropanecarboxamide 6a:
Rf (SiO2, CHCl3/ethyl acetate, 3/2, v/v): 0.51; 1H NMR (CDCl3, 400 MHz) : 1.00-1.06 (m, 2 H),
1.19-1.24 (m, 2 H), 1.65-1.71 (m, 1 H), 7.49 (t, J = 5.4 Hz, 1 H), 8.55-8.69 (m, 3 H), 9.57 (s, 1 H)
ppm; 13C NMR (CDCl3, 75 MHz) : 9.4 (2 C), 15.7, 125.2, 131.8 (d, J = 12.4 Hz), 134.2, 139.0 (d,
J = 24.4 Hz), 140.8, 145.2, 145.8 (d, J = 5.2 Hz), 147.1, 156.4 (d, J = 260.4 Hz), 172.5 ppm; 19F
NMR (CDCl3, 376 MHz) : -126.5 ppm; ESI-MS m/z calcd. for C13H11ON4FCl+ [M+H]+ 293.1,
found 293.1, calcd. for C13H11ON4FClNa+ [M+Na]+ 315.0, found 315.0.
4.1.2.3.2.. N-(6-Chloro-5-(2-fluoropyridine-4-yl)pyrazin-2-yl)cyclopropanecarboxamide 6b:
1
Rf (SiO2, dichloromethane/methanol, 39/1, v/v): 0.53; H NMR (DMSO-d6, 300 MHz) : 0.90-
0.92 (m, 4 H), 2.01-2.08 (m, 1 H), 7.53 (s, 1 H), 7.74 (d, J = 5.2 Hz, 1 H), 8.39 (d, J = 5.2 Hz,
1 H), 9.39 (s, 1 H), 11.64 (s, 1 H) ppm; 13C NMR (DMSO-d6, 100 Hz) : 9.1 (2 C), 14.7, 109.9 (d,
J = 39.5 Hz), 122.5 (d, J = 4.0 Hz), 134.2, 142.2 (d, J = 3.8 Hz), 143.9, 148.2 (d, J = 2.1 Hz),
148.3, 149.5 (d, J = 8.5 Hz), 163.5 (d, J = 235.1), 173.8 ppm; 19F NMR (DMSO-d6, 376 MHz) : -
68.2 ppm; ESI-MS m/z calcd. for C13H9ON4FCl- [M-H]- 291.0, found 291.1.
4.1.2.4. N-(6-(2-Fluorophenyl)-5-(3-fluoropyridine-4-yl)pyrazine-2-
yl)cyclopropanecarboxamide 1:
Amide 6a (91.7 mg, 0.313 mmol, 1.0 eq.) and 2-fluorophenylboronic acid (65.7 mg,
0.470 mmol, 1.5 eq. (from Fluorochem, UK) were added to a solution of 4 mL 1,4-dioxane
and 2 mL of a 2 M Cs2CO3 solution (4.0 mmol, 12.8 eq.). Under argon atmosphere, 22.7 mg
(0.031 mmol, 0.1 eq.) of Pd(dppf)Cl2 was added and the reaction mixture was heated for
5.5 h at 85 °C. Afterwards, the mixture was poured into 100 mL water and filtered over
celite. The filtrate was extracted with ethyl acetate (3 × 25 mL) and the combined organic
layers were dried over sodium sulfate, filtered and concentrated under reduced pressure.
18

The residue was purified by column chromatography (SiO2, ethyl acetate/n-hexane, 2/3, v/v)
to give the desired product 1 with a yield of 64% (71.0 mg, 0.201 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 2/3, v/v): 0.17; Mp: 189.8-191.7 °C; H NMR (CDCl3, 300
MHz) : 0.91-0.97 (m, 2 H), 1.14-1.19 (m, 2 H), 1.58-1.66 (m, 1 H), 6.94 (t, J = 10.0 Hz, 1 H),
7.16 (td, J = 7.6 Hz, J = 1.1 Hz, 1 H), 7.32-7.44 (m, 3 H); 8.33 (d, J = 1.7 Hz, 1 H), 8.43 (d, J = 4.9
Hz, 1 H), 8.72 (bs, 1 H), 9.63 (s, 1 H) ppm; 13C NMR (CDCl3, 75 Hz) : 9.0 (2 C), 15.6, 115.9 (d, J
= 21.7 Hz), 124.4 (d, J = 3.5 Hz), 125.1 (d, J = 14.4 Hz), 125.2, 131.1 (d, J = 1.9 Hz), 131.5 (d, J =
8.3 Hz), 133.7 (d, J = 12.0 Hz), 135.3, 138.6 (d, J = 24.8 Hz), 141.9, 145.7 (d, J = 5.2 Hz), 147.1,
19
147.2, 155.9 (d, J = 242.8 Hz), 159.3 (d, J = 233.5 Hz), 172.6 ppm; F NMR (CDCl3, 376 MHz)
: -115.1, -129.4 ppm; ESI-MS m/z calcd. for C19H14ON4F2Na+ [M+Na]+ 375.10279, found
375.10250.
4.1.2.5. General procedure for 7a – c:
Amide 6b (1.0 eq.) and 2-fluorophenylboronic acid for 7a (1.2 eq.), pyridine-4-boronic acid
for 7b (1.2 eq., from Apollo Scientific, UK) or phenylboronic acid for 7c (1.3 eq., from Merck
KGaA, Germany) were added to a solution of 3 mL 1,4-dioxane and 6.0 eq. of a 2 M aqueous
Cs2CO3 solution. Under argon atmosphere, 0.05 eq. [Pd(dppf)Cl2] was added and the reaction
mixture was heated at 90 °C (7a: 7 h, 7b: 2.5 h and 7c: 6h). Afterwards, the mixture was
poured into ~50 mL water and extracted with ethyl acetate (3 × ~25 mL). The combined
organic layers were washed with 50 mL brine, dried over anhydrous sodium sulfate, filtered
and then concentrated under reduced pressure. The crude product was purified by column
chromatography.
4.1.2.5.1. N-(6-(2-Fluorophenyl)-5-(2-fluoropyridine-4-yl)pyrazine-2-yl)cyclo-
propanecarboxamide 7a:
Column chromatography: SiO2, chloroform/ethyl acetate, 4/1, v/v. Yield: 94% (42.0 mg,
1
0.119 mmol). Rf (SiO2, ethyl acetate/n-hexane, 2/3, v/v): 0.52; Mp: 156.8-159.8 °C; H NMR
(DMSO-d6, 300 MHz) : 0.86-0.90 (m, 4 H), 2.04-2.08 (m, 1 H), 7.04 (s, 1 H), 7.15-7.24 (m,
2 H), 7.35 (td, J = 7.5 Hz, J = 0.9 Hz, 1 H), 7.53 (tdd, J = 7.3 Hz, J = 5.4 Hz, J = 1.8 Hz, 1 H), 7.64
13
(td, J = 7.5 Hz, J = 1.8 Hz, 1 H), 8.15 (d, J = 5.2 Hz, 1 H), 9.45 (s, 1 H), 11.48 (s, 1 H) ppm; C
NMR (DMSO-d6, 100 Hz) : 8.8 (2 C), 14.6, 108.9 (d, J = 39.4 Hz), 116.2 (d, J = 21.1 Hz), 121.7
(d, J = 3.7 Hz), 125.4 (d, J = 3.0 Hz), 125.4 (d, J = 15.1 Hz), 132.4 (d, J = 2.3 Hz), 132.4 (d, J =
11.3 Hz), 135.2, 143.9 (d, J = 3.8 Hz), 145.9, 148.0 (d, J = 15.5 Hz), 148.4, 151.7 (d, J = 9.2 Hz),
19
159.0 (d, J = 247.0 Hz), 163.5 (d, J = 234.8 Hz), 173.8 ppm; F NMR (DMSO-d6, 282 MHz)
+
: -115.7, -68.5 ppm; ESI-MS m/z calcd. for C19H15ON4F2 [M+H]+ 353.12084, found
353,12091, m/z calcd. for C19H14ON4F2Na+ [M+Na]+ 375.10279, found 375.10274.
4.1.2.5.2. N-(5-(2-Fluoropyridine-4-yl)-6-(pyridine-4-yl)pyrazine-2-yl)cyclopropane-
carboxamide 7b:
Column chromatography: SiO2, chloroform/methanol, 12/1, v/v. Yield: 51% (39 mg,
0.116 mmol). Rf (SiO2, chloroform/methanol, 12/1, v/v): 0.44; Mp: 239.7-240.5 °C
(decomposition); 1H NMR (DMSO-d6, 300 MHz) : 0.90-0.92 (m, 4 H), 2.06-2.13 (m, 1 H), 7.16
(s, 1 H), 7.25 (d, J = 5.1 Hz, 1 H), 7.44 (d, J = 4.4 Hz, 2 H), 8.19 (d, J = 5.2 Hz, 1 H), 8.63 (d,
19

13
J = 4.2 Hz, 2 H), 9.46 (s, 1 H), 11.50 (s, 1 H) ppm; C NMR (DMSO-d6, 100 Hz) : 8.9 (2C),
14.6, 110.1 (d, J = 39.1 Hz), 122.8 (d, J = 3.7 Hz), 124.5 (2C), 135.3, 143.2 (d, J = 3.5 Hz), 145.0,
148.1 (d, J = 16.2 Hz), 148.2, 148.6, 150.3 (2C), 151.5 (d, J = 8.3 Hz), 163.5 (d, J = 235.2 Hz),
19
173.8 ppm; F NMR (DMSO-d6, 376 MHz) : -68.2 ppm; ESI-MS m/z calcd. for C18H15ON5F+
[M+H]+ 336.12551, found 336.12527.
4.1.2.5.3. N-(5-(2-Fluoropyridine-4-yl)-6-phenylpyrazine-2-yl)cyclopropanecarboxamide 7c:
Column chromatography: SiO2, ethyl acetate/n-hexane, 1/3, v/v. Yield: 61% (35 mg,
1
0.105 mmol). Rf (SiO2, ethyl acetate/n-hexane, 1/3, v/v): 0.29; Mp: 160.9-162.2 °C; H NMR
(DMSO-d6, 300 MHz) : 0.91-0.97 (m, 2 H), 1.14-1.19 (m, 2 H), 1.58-1.63 (m, 1 H), 7.02 (s,
1 H), 7.13 (dt, J = 5.2 Hz, J = 1.6 Hz, 1 H), 7.35-7.41 (m, 5 H), 8.09 (d, J = 5.2 Hz, 1 H), 8.50 (s,
13
1 H), 9.54 (s, 1 H) ppm; C NMR (DMSO-d6, 75 Hz) : 9.0 (2 C), 15.7, 109.9 (d, J = 38.8 Hz),
121.8 (d, J = 4.1 Hz), 128.7 (2 C), 129.35 (2 C), 129.59, 134.4, 136.9, 143.5 (d, J = 3.6 Hz),
146.8, 147.4 (d, J = 15.0 Hz), 150.6, 151.5 (d, J = 3.6 Hz), 163.9 (d, J = 238.5 Hz), 172.5 ppm;
19F NMR (DMSO-d6, 282 MHz) : -67.5 ppm; ESI-MS m/z calcd. for C19H16ON4F+ [M+H]+
335.13027, found 335.13023.
4.1.2.6. N-(6-Chloro-5-iodopyrazine-2-yl)cyclopropanecarboxamide 8:
Amine 4 (200 mg, 0.786 mmol, 1.0 eq.) was dissolved in dried pyridine and
cyclopropanecarbonyl chloride (85 µL, 98 mg, 0.943 mmol, 1.2 eq.) was dropwise added to
the solution under ice cooling. After stirring for 2.5 h at room temperature, the reaction was
monitored by TLC and further cyclopropanecarbonyl chloride (58 µL, 66 mg, 0.631 mmol,
0.78 eq.) was added until the reaction was completed. Then the solvent was removed under
reduced pressure, the residue suspended in 50 mL water and extracted with ethyl acetate (3
× 50 mL). The organic layers were separated, dried over sodium sulfate, filtered and
concentrated under reduced pressure. The desired product 8 was purified by column
chromatography (SiO2) using a solvent mixture of dichloromethane/methanol, 39/1, v/v to
get 8 with a yield of 90% (280 mg, 0.711 mmol).
1
Rf (SiO2, dichloromethane/methanol, 99/1, v/v): 0.79; Mp: 186.7-187.8 °C; H NMR (DMSO-
13
d6, 400 MHz) : 0.86-0.90 (m, 4 H), 1.94-2.02 (m, 1 H), 9.08 (s, 1 H), 11.45 (s, 1 H) ppm; C
NMR (DMSO-d6, 100 Hz) : 9.0 (2 C), 14.7, 106.4, 135.2, 147.6, 150.5, 173.5 ppm; ESI-MS m/z
calcd. for C8H8N3PClI+ [M+H]+ 323.940, found 323.939, m/z calcd. for C8H7N3OClINa+ [M+Na]+
345.921, found 345.921, m/z calcd. for C8H6N3OClI- [M-H]- 321.925, found 321.925.
4.1.2.7. 4-Bromo-2-nitropyridine 10:
A mixture of 2.5 mL of a 30% hydrogen peroxide solution and 5.0 mL of 96% sulfuric acid
was added to a solution of 1.5 mL of 96% sulfuric acid and 0.52 g (3.006 mmol) 9 (2-amino-4-
bromopyridine from Fluorochem, UK). After one day stirring at room temperature the
solution was diluted with 100 mL water and adjusted to a pH value of 12 using 6 M aqueous
NaOH. The water phase was extracted with ethyl acetate (3 × 50 mL) and the combined
organic layers were washed with brine, dried over sodium sulfate, filtered and the solvent
was evaporated under reduced pressure. The crude product was purified by column
20

chromatography (SiO2, ethyl acetate/n-hexane, 2/1, v/v), to get 10 with a yield of 38%
(0.23 g, 1.143 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 1/3, v/v): 0.43; Mp: 49.5-50.6 °C; H NMR (DMSO-d6,
400 MHz) : 8.20 (dd, J = 5.1 Hz, J = 1.6 Hz, 1 H), 8.56 (d, J = 1.5 Hz, 1 H), 8.58 (d, J = 5.1 Hz,
13
1 H) ppm; C NMR (DMSO-d6, 100 Hz) : 122.0, 133.3, 135.7, 150.2, 157.3 ppm; ESI-MS m/z
calcd. for C5H3O2N2BrNa+ [M+Na]+ 224.9, found 224.9.
4.1.2.8. 2-Nitro-4-(4,4,5,5-tetramethyl-1,3,2-dioxaborolane-2-yl)pyridine 1136:
Under argon atmosphere, 0.20 g (0.985 mmol, 1.0 eq.) of 10, 0.25 g (0.985 mmol, 1.0 eq.)
4,4,4',4',5,5,5',5'-octamethyl-2,2'-bi-1,3,2-dioxaborolane (from Frontier Scientific, USA) and
0.48 g (4.925 mmol, 5.0 eq.) potassium acetate are mixed in 12 mL 1,4-dioxane. Then 0.03 g
(0.039 mmol, 0.04 eq.) Pd(dppf)Cl2 was added and the reaction mixture was refluxed for 45
min. After cooling to room temperature, the solution was filtered over celite and washed
with 50 mL 1,4-dioxane. The solvent was evaporated under reduced pressure and the crude
product was purified by column chromatography (SiO2, ethyl acetate/n-hexane, 1/1, v/v), to
get 11 with a yield of 85% (0.21 g, 0.840 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 1/1, v/v): 0.05; Mp: 111.0-117.4 °C; H NMR (CDCl3,
400 MHz) : 1.37 (s, 12 H), 7.98 (dd, J = 4.5 Hz, J = 0.7 Hz, 1 H), 8.56 (s, 1 H), 8.65 (dd,
J = 4.5 Hz, J = 0.8 Hz, 1 H) ppm; 13C NMR (CDCl3, 100 Hz) : 24.8 (4 C), 85.3 (2 C), 122.7, 134.2,
148.4, 156.6 ppm; ESI-MS m/z calcd. for C11H15O4N2BNa+ [M+Na]+ 273.1, found 273.1, m/z
calcd. for C22H30O8N4B2Na+ [2M+Na]+ 523.2, found 523.2.
4.1.2.9. N-(6-Chloro-5-(2-nitropyridine-4-yl)pyrazine-2-yl)cyclopropanecarboxamide 12:
Under argon atmosphere, 200 mg (0.618 mmol, 1.0 eq.) of 8 and 170 mg (0.680 mmol,
1.1 eq.) of 11 are dissolved in a mixture of 5 mL 1,4-dioxane and 1.24 mL (2.472 mmol,
4.0 eq.) 2 M aqueous potassium carbonate solution. Then 9.0 mg (0.012 mmol, 0.02 eq.)
Pd(dppf)Cl2 was added and the reaction mixture was heated for 5 h at 90 °C. After cooling to
room temperature, the reaction mixture was quenched with 100 mL water and extracted
with ethyl acetate (3 × 25 mL). The combined organic layers were dried over sodium sulfate,
filtered and the solvent was evaporated under reduced pressure. The crude product was
purified by column chromatography (SiO2, ethyl acetate/n-hexane, 2/3, v/v) to give 12 with a
yield of 49% (97 mg, 0.303 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 2/3, v/v):0.50 °C; Mp: 172.5-175.5 °C; H NMR (DMSO-d6,
400 MHz) : 0.89-0.91 (m, 4 H), 1.99-2.05 (m, 1 H), 8.28 (dd, J = 4.9 Hz, J = 1.4 Hz, 1 H), 8.63
(d, J = 0.9 Hz, 1 H), 8.80 (d, J = 5.0 Hz, 1 H), 9.37 (s, 1 H), 11.64 (s, 1 H) ppm; 13C NMR (DMSO-
d6, 100 Hz) : 9.1 (2 C), 14.7, 118.2, 129.7, 134.3, 141.2, 144.1, 147.9, 148.4, 149.6, 157.1,
173.8 ppm; ESI-MS m/z calcd. for C13H11N5O3Cl+ [M+H]+ 320.054, found 320.054, m/z calcd.
for C13H10N5O3ClNa+ [M+Na]+ 342.036, found 342.036, m/z calcd. for C26H20N10O6Cl2Na+
[2M+Na]+ 661.084, found 661.084, m/z calcd. for C13H9N5O3Cl- [M-H]- 318.040, found
318.040.
21

4.1.2.10. N-(6-(2-Fluorophenyl)-5-(2-nitropyridine-4-yl)pyrazine-2-yl)cyclo-
propanecarboxamide 13:
Amide 12 (70 mg, 0.219 mmol, 1.0 eq.) and 2-fluorophenylboronic acid (40 mg, 0.285 mmol,
1.3 eq.) were added to a mixture of 5 mL 1,4-dioxane and 660 µl of a 2 M Cs2CO3 solution
(1.314 mmol, 6.0 eq.). Under argon atmosphere, 8.0 mg (0.011 mmol, 0.05 eq.) Pd(dppf)Cl2
was added and the reaction mixture was heated 3 h at 90 °C. After cooling to room
temperature, the mixture was poured into 50 mL water and extracted with ethyl acetate (3 ×
30 mL). The combined organic layers were dried over anhydrous sodium sulfate, filtered and
concentrated under reduced pressure. The crude product was purified by column
chromatography (SiO2, ethyl acetate/n-hexane, 2/3, v/v) to give the desired product 13 with
a yield of 61% (64 mg, 0.158 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 2/3, v/v): 0.35; Mp: 134.7-137.0 °C; H NMR (DMSO-d6,
400 MHz) : 0.85-096 (m, 4 H), 2.04-2.14 (m, 1 H), 7.21 (t, J = 8.0 Hz, 1 H), 7.40 (t, J = 7.5 Hz,
1 H), 7.57 (m, 1 H), 7.69 (t, J = 7.5 Hz, 1 H), 7.77 (dd, J = 5.0 Hz, J = 1.4 Hz, 1 H), 8.21 (d,
J = 0.8 Hz, 1 H), 8.60 (d, J = 5.0 Hz, 1 H), 9.51 (s, 1 H), 11.54 (s, 1 H) ppm; 13C NMR (DMSO-d6,
100 Hz) : 8.89 (2 C), 14.7, 116.4 (d, J = 21.3 Hz), 117.3, 125.2 (d, J = 14.5 Hz), 125.6 (d,
J = 3.0 Hz), 128.9, 132.4 (d, J = 1.4 Hz), 132.7 (d, J = 8.2 Hz), 135.4, 142.9, 146.2, 148.7, 149.6,
150.1, 157.1, 158.9 (d, J = 247.0 Hz), 173.8 ppm; 19F NMR (DMSO-d6, 376 MHz) : -
115.9 ppm; ESI-MS m/z calcd. for C19H15O3N5F+ [M+H]+ 380.11534, found 380.11577, m/z
calcd. for C19H14O3N5FNa+ [M+Na]+ 402.09729, found 402.09768.
4.1.2.11. 6-(2-Fluorophenyl)-5-(2-fluoropyridine-4-yl)pyrazine-2-amine M4:
Amine 5b (50 mg, 0.222 mmol, 1.0 eq.) and 2-fluoropyridine-2-boronic acid (37 mg,
0.266 mmol, 1.2 eq.) were added to a mixture of 2 mL 1,4-dioxane and 350 µl of a 2 M
Cs2CO3 solution (0.700 mmol, 3.1 eq.). Under argon atmosphere, 16.0 mg (0.022 mmol,
0.1 eq.) Pd(dppf)Cl2 was added and the reaction mixture was heated 3.5 h at 90 °C and
further 4 h at 110 °C. After cooling to room temperature, the reaction mixture was quenched
with 70 mL water and extracted with ethyl acetate (4 × 25 mL). The combined organic layers
were dried over sodium sulfate, filtered and the solvent was evaporated under reduced
pressure. The crude product was purified by column chromatography (SiO2, ethyl acetate/n-
hexane, 1/2, v/v) to get M4 with a yield of 54% (34 mg, 0.120 mmol).
1
Rf (SiO2, ethyl acetate/n-hexane, 1/1, v/v): 0.35; Mp: 210.5-212.7 °C; H NMR (CDCl3,
400 MHz) : 5.12 (s, 2 H), 6.95 (s, 1 H), 7.02 (ddd, J = 9.6 Hz, J = 8.3 Hz, J = 1.1 Hz, 1 H), 7.10
(dt, J = 5.4 Hz, J = 1.7 Hz, 1 H), 7.26 (d, J = 6.9 Hz, 1 H), 7.43 (tdd, J = 7.2 Hz, J = 5.1 Hz,
J = 1.8 Hz, 1 H), 7.50 (dt, J = 7.4 Hz, J = 1.8 Hz, 1 H), 8.05 (d, J = 5.3 Hz, 1 H), 8.10 (s, 1 H) ppm;
13C NMR (CDCl3, 100 Hz) : 108.5 (d, J = 38.7 Hz), 116.2 (d, J = 21.7 Hz), 120.6 (d, J = 3.9 Hz),
124.8 (d, J = 3.6 Hz), 126.1 (d, J = 14.8 Hz), 131.18, 131.2 (d, J = 4.4 Hz), 131.4 (d, J = 8.2 Hz),
139.2 (d, J = 3.7 Hz), 145.8, 147.1 (d, J = 15.2 Hz), 152.0 (d, J = 8.3 Hz), 153.4, 159.2 (d,
19
J = 249.4 Hz), 163.9 (d, J = 237.6 Hz) ppm; F NMR (CDCl3, 376 MHz) : -114.7, -68.4 ppm;
+
ESI-MS m/z calcd. for C15H11N4F2 [M+H]+ 285.0946, found 285.0945, m/z calcd. for
C15H10N4F2Na+ [M+Na]+ 307.0762, found 307.0766.
22

The logD value of M4 was calculated with ACD/ChemSketch, Vers. 12.5.
4.2 Radiochemistry
4.2.1. General
No-carrier-added [18F]fluoride was produced via the [18O(p,n)18F] nuclear reaction by
irradiation of an [18O]H2O target (Hyox 18 enriched water, Rotem Industries Ltd, Israel) on a
Cyclone 18/9 (iba RadioPharma Solutions, Belgium) with fixed energy proton beam using
Nirta [18F]fluoride XL target.
Radio thin layer chromatography (radio-TLC) was performed on silica gel (Polygram® SIL
G/UV254) pre-coated plates with a mixture of EtOAc/cyclohexane 2/1 (v/v) as eluent. The
plates were exposed to storage phosphor screens (BAS-TR2025, FUJIFILM Co., Tokyo, Japan)
and recorded using the Amersham Typhoon RGB Biomolecular Imager (GE Healthcare Life
Sciences). Images were quantified with the ImageQuant TL8.1 software (GE Healthcare Life
Sciences).
Analytical chromatographic separations were performed on a JASCO LC-2000 system,
incorporating a PU-2080Plus pump, AS-2055Plus auto injector (100 L sample loop), and a
UV-2070Plus detector coupled with a radioactivity HPLC flow monitor (Gabi Star, raytest
Isotopenmessgeräte GmbH). Data analysis was performed with the Galaxie chromatography
software (Agilent Technologies) using the chromatograms obtained at 254 nm. A Reprosil-
Pur C18-AQ column (250 x 4.6 mm; 5 µm; Dr. Maisch HPLC GmbH; Germany) with ACN/20
mM NH4OAc aq. (pH 6.8) as eluent mixture and a flow of 1.0 mL/min was used (gradient:
eluent A 10% ACN/20 mM NH4OAc aq.; eluent B 90% ACN/20 mM NH4OAc aq.; 0–5 min
100% A, 5–30 min up to 100% B, 30–35 min 100% B , 35–37 min up to 100% A, 37–40 min
100% A).
Semi-preparative HPLC separations in the manual syntheses were performed on a JASCO LC-
2000 system, incorporating a PU-2080-20 pump, an UV/VIS-2075 detector coupled with a
radioactivity HPLC flow monitor whose measurement geometry was slightly modified (Gabi
Star, raytest Isotopenmessgeräte GmbH) and a fraction collector (Advantec CHF-122SC).
Data analysis was performed with the Galaxie chromatography software (Agilent
Technologies) using the chromatograms obtained at 254 nm.
The ammonium acetate and the SDS concentrations stated as 20 mM NH4OAc aq. and 100
mM aq., respectively, correspond to the concentration in the aqueous component of an
eluent mixture.
4.2.2. Radiosyntheses
4.2.2.1. Manual syntheses
-
No carrier added [18F]fluoride in 1.5 mL water was trapped on a Chromafix® 30 PS-HCO3
cartridge (MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany). The activity was eluted
with 300 µL of an aqueous solution of potassium carbonate (K2CO3, 1.8 mg, 13 µmol) into a 4
mL V-vial and Kryptofix 2.2.2 (K2.2.2, 11 mg, 29 µmol) in 1 mL ACN was added. The aqueous
[18F uoride was a eotropically dried under vacuum and nitrogen ow within 7-1 min using
a single mode microwave (75 , at 5 °C, power cycling mode). Two aliquots of (
23

1. m ) were added during the drying procedure and the nal comple was dissolved in 5
M or MF ready for labeling. Therea er, a solu on of 1 1.2 mg of precursor in 500 µL
18
DMSO or DMF was added, and the F-labeling was performed at different temperatures
(120 and 155 °C). To analyze the reaction mixture and to determine radiochemical yields,
samples were taken for radio-HPLC and radio-TLC at different time points (5, 10, 15, and 20
min).
After cooling to < 30 °C, the reaction mixture was diluted with 2.5 mL water and 0.5 mL ACN
and directly applied to an isocratic semi-preparative RP-HPLC for isolation of [18F]7a (40%
ACN/20 mM NH4OAcaq., 8 mL/min, Reprosil-Pur 120 CN, 250 x 20 mm; 10 µm; Dr. Maisch
HPLC GmbH; Germany). The collected radiotracer fraction was diluted with 35 mL water to
perform final purification by sorption on a Sep-Pak C18 light cartridge (Waters, Milford,
MA, USA) and successive elution with 0.75 mL of ethanol.
4.2.2.2. Automated syntheses
Remote controlled radiosynthesis of [18F]7a was performed using a TRACERlab FX2 N
synthesizer (GE Healthcare, USA) equipped with a Laboport vacuum pump N810.3FT.18 (KNF
Neuburger GmbH, Freiburg, Germany), a BlueShadow UV detector 10D (KNAUER GmbH,
Berlin, Germany) and the TRACERlab FX Software.
-
[18F]Fluoride (3-4 GBq) was trapped on a Chromafix® 30 PS-HCO3 cartridge (Figure 7, entry 1,
MACHEREY-NAGEL GmbH & Co. KG, Düren, Germany) and eluted into the reactor with 400
µL of an aqueous solution of potassium carbonate (K2CO3, 1.8 mg, 13 µmol, entry 2). After
addition of Kryptofix 2.2.2. in 1 mL ACN (11 mg, 29 µmol, entry 3), the mixture was
azeotropically dried for approximately 10 min. Thereafter, 1.0 1.2 mg of the nitro precursor
(13) dissolved in 1 mL DMSO (entry 4) was added, and the reaction mixture was stirred at
150 ⁰C for 15 min. After cooling, the reaction mixture was diluted with 2.5 mL H2O and 0.5
mL ACN (entry 5) and transferred into the injection vial (entry 6). Semi-preparative HPLC was
performed using a Reprosil-Pur 120 CN column (250 x 20 mm; 10 µm; Dr. Maisch HPLC
GmbH; Germany) with a solvent composition of 40% ACN/20 mM NH4OAcaq. at a flow rate of
8.0 mL/min (entry 7). [18F]7a was collected in the dilution vessel (entry 8) previously loaded
with 35 mL H2O. Final purification was performed by passing the solution through a Sep-Pak
C18 light cartridge (entry 9), followed by washing with 2 mL water (entry 10) and elution of
[18F]7a with 1.2 mL EtOH (entry 11) into the product vial (entry 12). The ethanolic solution
was transferred out of the hot cell and the solvent was reduced under a gentle argon stream
at 70 ⁰C to a final volume of 10-50 µL. Afterwards the radiotracer was diluted in isotonic
saline to obtain a final product containing 10% of EtOH (v/v).
The molar activity was determined on the base of a calibration curve carried out under
isocratic HPLC conditions (26% ACN/20 mM NH4OAcaq; Reprosil-Pur 120 CN 250 x 4.6 mm)
using chromatograms obtained at 232 nM as an appropriate maximum of UV absorbance.
24

Figure 7. Scheme of the synthesis module TRACERlab FX2 N for the radiosynthesis of [18F]7a. (1) Chromafix 30-
-
PS-HCO3 , (2) K2CO3 (1.8 mg in 400 µL water), (3) K2.2.2. (11 mg in 1 mL ACN), (4) precursor (1 mg of 13 in 1 mL
DMSO), (5) 2.5 mL water and 0.5 mL ACN, (6) injection vial, (7) Reprosil-Pur 120 CN (40% ACN/20 mM
NH4OAcaq., flow 8.0 mL/min), (8) 35 mL water, (9) Sep-Pak C18 light, (10) 2 mL water, (11) 1.2 mL EtOH, (12)
product vial.
4.2.2.3. In vitro stability and determination of logD value
The in vitro stability of [18F]7a was investigated by incubation of small tracer amounts (~ 5
MBq) at 40 °C in i) 500 µL n-octanol and ii) 500 µL pig plasma samples. After 30 and 60 min,
aliquots were taken and analyzed by radio-TLC and radio-HPLC. The plasma samples were
processed as described in 4.4.3.
The partition coefficient of [18F]7a was experimentally determined for the n-octanol/PBS
system by the shake-flask method. Small tracer amounts (~ 800 kBq) were added to a
mixture of 3.0 mL of n-octanol and 3.0 mL of PBS. After shaking for 20 min at room
temperature, the samples were centrifuged (10,000 rpm, 5 min) and 1 mL aliquots of the
organic as well as the aqueous layer were taken and measured in a gamma counter
(PerkinElmer Wallac Wizard 1480 Gamma Counter, manufactured by WALLAC, Turku,
Finland). Another 1 mL aliquote of the organic layer was mixed with 2.0 mL n-octanol and 3.0
mL of PBS and subjected to the same procedure until constant partition coefficient values
had been obtained. All measurements were done in quadruplicate.
4.3. In vitro binding assays
4.3.1. Determination of Ki values
Membrane preparations of recombinant CHO or HEK (human embryonic kidney) cells
expressing the respective human adenosine receptor subtype were obtained according to
28
Borrman et al. or purchased from Perkin Elmer. Radioligand binding assays at human A1,
A2A, A2B and A3 receptors were performed according to Alnouri et al.48
25

The following radioligands were employed: [3H]2-chloro-N6cyclopentyladenosine ([3H]CCPA,
1 nM49) for the A1 receptor, [3H]3-(3-hydroxypropyl)-7-methyl-8-(m-methoxystyryl)-1-
propargylxanthine ([3H]MSX-2,
1
nM50) for the A2A receptor, [³H]8-(4-[4-(4-
chlorophenyl)piperazine-1-sulfonyl]phenyl)-1-propyl-2,3,6,7-tetrahydro-1H-purine-2,6-dione
([³H]PSB-603, 0.3 nM28) for the A2B receptor, and [3H](R)-8-ethyl-4-methyl-2-(phenyl)1,4,7,8-
tetrahydro-5H-imidazo[2,1-i]purin-5-one ([3H]PSB-11, 1 nM51) for the A3 receptor. Three
separate experiments were performed for determination of Ki values. All data were analyzed
with GraphPad Prism, Version 4.1 (GraphPad Inc., La Jolla, CA).
4.3.2. Binding of [18F]7a to mouse tissue
Mouse tissue samples were prepared by homogenizing the brain and heart from female CD-
1 mice (10-12 weeks) with homogenization buffer (50 mM TRIS-HCl, pH 7.4) on ice using a
Potter S homogenizer (B. Braun Biotech, Göttingen, Germany). 10 mL buffer were added to 1
g tissue in a borosilicate vessel with a PTFE plunger and the biological material disrupted by
up and down movement of the pestle (10 repetitions), rotating at 1500 rpm. The suspension
was centrifuged at 12000 rpm for 10 min at 4 °C (3-30KS, Sigma Laborzentrifugen GmbH,
Osterode am Harz, Germany). The pellet was re-suspended in homogenization buffer and
centrifuged as described. After repeating this washing step, the sediment was suspended in
the initial volume of homogenization buffer and stored in aliquots of 1 mL at -25 °C. For the
experiment, the tissue homogenates were thawed and incubated at 560 µg protein/mL
(Pierce BCA Protein Assay; Thermo Fisher Scientific, Dreieich, Germany) with 0.275 MBq/mL
of [18F]7a (4 nM at start of incubation) in incubation buffer (50 mM TRIS-HCl, pH 7.4, 100
mM NaCl, 5 mM MgCl2, 1 mM EDTA) with continuous shaking at room temperature. The
binding experiments were performed in the presence of recombinant human adenosine
deaminase at a working concentration of 1 U/L. To assess non-specific binding of [18F]7a, co-
incubations with either 1 µM 7a or 1 µM PSB-603 were performed. To assess filter-binding of
[18F]7a, incubations without tissue homogenates were performed. The incubation was
terminated after 60 min by filtration via a glass microfibre filter (Whatman grade GF-B; GE
Healthcare via Fisher Scientific GmbH, Schwerte, Germany), pre-incubated for 60 min with
freshly prepared 0.3% polyethyleneimine, using a 48-well cell harvester (Brandel Harvesters,
Brandel, Gaithersburg, MD, USA). Activity trapped on the filters was counted using a gamma
counter ( 48 i ard ™, PerkinElmer LAS GmbH, Germany).
4.4. In vivo / ex vivo studies
Animals for in vivo studies were obtained from the Medizinisch-Experimentelles Zentrum,
Universität Leipzig. All procedures that include animals were approved by the respective
State Animal Care and Use committee and conducted in accordance with the German Law
for the Protection of Animals (TVV 08/13).
4.4.1. Ex vivo autoradiography of [18F]7a in mice
26

[18F]7a was administered as bolus in one awake female CD-1 mouse (12 weeks, 32.5 g) via
the tail vein (4 MBq and 31 nmol/kg at time of injection). The brain was isolated after
cervical dislocation of the anesthetized animal at 30 min p.i., and the two hemispheres
separated and frozen immediately in pre-cooled isopentane at -25°C for 10 min.
Cryosections of 12 µm were obtained using a cryostat microtome (Microm HM 560 Cryostat,
Microm International GmbH, Walldorf, Germany) and mounted on microscope slides. After
drying for 10 min under a stream of cold air, the sections were exposed to an imaging plate
(FUJIFILM Corporation, Tokyo, Japan) for 120 min. Afterwards, the imaging plate was read
out (HD-CR 35 NDT, DÜRR NDT GmbH & Co. KG, Bietigheim-Bissingen, Germany) and the
digital image analysed using AIDA Image Analysis software (V 4.27; Elysia-raytest GmbH,
Straubenhardt, Germany). Eventually the brain sections were Nissl-stained and scanned
(Super COOLSCAN 9000 ED, Nikon GmbH, Düsseldorf, Germany) to delineate brain regions
using a mouse brain atlas.52
4.4.2. Dynamic PET scan of [18F]7a in mice
One female CD-1 mouse (12 weeks, 33 g) was initially anesthetized with 4 % of isoflurane
and was positioned prone in a small-animal PET/MRI system (nanoScan®, MEDISO,
Budapest, Hungary) equipped with respiratory monitoring, heated mouse bed (37 °C), and
inhalation anesthesia (1.8 % isoflurane in a 60 % oxygen/40 % air gas mixture at 250 ml/min
airflow; Anesthesia Unit U-41 , agntho’s, idingö, weden; Gas blender 1 series, M Q
Instruments, Rome, Italy). Prior to the PET scan a scout image MR sequence was done to
outline the animal dimensions. The animal was injected via the tail vein with 3.2 MBq [18F]7a
(15 nmol/kg at time of injection) in a volume of 200 µl saline. The injected dose was
calculated by the difference of the activity in the syringe before and after the injection. A
dynamic whole body animal PET scan of 60 min length was started simultaneously. The list-
mode data was sorted into sinograms using a framing scheme of 6 × 10 s, 6 × 30 s, 5 × 60 s,
10 × 300 s frames and reconstructed by OSEM3D with 4 iterations and 6 subsets. A mouse
brain atlas (PMOD (v.3.7, PMOD technologies LLC, Zurich, Switzerland) was used to delineate
regions of interest.
4.4.3. Metabolism studies of [18F]7a in mice
[18F]7a was administered as bolus in awake female CD-1 mice (10-12 weeks, 25-32 g) via the
tail vein (10-80 MBq and 1-50 nmol/kg at time of injection). At 5 (n=2), 15 (n=2), and 30
(n=3) min p.i., blood samples were obtained retro-orbitally from the anesthetized animals.
Blood plasma was obtained as supernatant after centrifugation of the whole blood samples
(14000 rpm, 1 min; Centrifuge 5418, Eppendorf Vertrieb Deutschland GmbH, Wesseling-
Berzdorf, Germany). After cervical dislocation, the brains were isolated immediately, and
urine samples obtained. All samples were weighed and the respective activity measured in a
dose calibrator (ISOMED 2010; MED Nuklear-Medizintechnik Dresden GmbH, Dresden,
Germany). Afterwards, 1 mL of deionized water was added to the brain in a borosilicate
vessel with a PTFE plunger and the biological material was homogenized on ice by repeated
up and down movement of the pestle, rotating at 1500 rpm.
27

MLC: For preparation of the MLC injection samples, mouse plasma (20 – 50 µL) was
dissolved in 100 – 300 µL of 200 mM aqueous SDS. Homogenized brain material (100 – 200
µL) was dissolved in 500 µL of 200 mM aqueous SDS, stirred at 75 °C for 5 minutes, diluted
with further 200 µL of 200 mM aqueous SDS and injected into the MLC system after cooling
to room temperature. The MLC system was built up of a JASCO PU-980 pump, an AS-
2055Plus auto injector with a 2000 L sample loop, and a UV-1575 detector coupled with a
gamma radioactivity HPLC detector (Gabi Star, raytest Isotopenmessgeräte GmbH). Data
analysis was performed with the Galaxie chromatography software (Agilent Technologies). A
Reprosil-Pur C18-AQ column (250 x 4.6 mm, particle size: 10 µm) coupled with a pre-column
of 10 mm length was used. Separations were performed by using an eluent mixture of
THF/100 mM aqueous SDS/10mM Na2HPO4 aq. in gradient mode (0 – 10 min at 100% 100
mM aqueous SDS, 10 – 25 min up to 50% THF, 25 – 30 min at 50% THF, 30 – 31 min up to
100% 100 mM aqueous SDS; 31 – 40 min at 100% 100 mM aqueous SDS) at a flow rate of 1.0
mL/min.
RP-HPLC: Protein precipitation was performed by addition of an ice-cold mixture of
acetone/water (9/1; v/v) in a ratio of 4 : 1 (v/v) of organic solvent to plasma or brain
homogenate, respectively. The samples were vortexed for 3 min, equilibrated on ice for 5
min, and centrifuged for 5 min at 10,000 rpm. The precipitates were washed with 100 µL of
the solvent mixture and subjected to the same procedure. The combined supernatants (total
volume between 1.0 –1.5 mL) were concentrated at 65 °C under argon flow to a final volume
of approximately 100 µL and analyzed by analytical radio-HPLC. To determine the
percentage of activity in the supernatants compared to total activity, aliquots of each step as
well as the precipitates were quantified by  counting (PerkinElmer Wallac Wizard 1480
Gamma Counter, manufactured by WALLAC, Turku, Finland). For plasma samples recoveries
of 97% of total activity for the 5 min samples and 87-85% for the 15 and 30 min samples
were obtained indicating the formation of [18F]fluoride which could not be entirely extracted
from the protein precipitates. Independent from the sampling time, 96-97% of recoveries
were obtained for the brain samples. To analyze the samples, the same HPLC method was
used as described in the radiochemistry part.
4.5. Molecular modelling studies
4.5.1. Homology model of the A2B receptor
A homology model of the human adenosine A2B receptor based on the X-ray crystal structure
of the human adenosine A2A receptor was recently described.34 The A2A receptor structure
was used as a template for generating a homology model of the A2B receptor using
Modeller9.53, 54 The selected A2A receptor template co-crystallized with the antagonist
ZM241385 (PDB ID: 4EIY, resolution: 1.8 Å) was downloaded from the Protein Data Bank.55
The amino acid sequence of the A2B receptor with the accession number of P29275 was
retrieved from the UniProt sequence database (http://www.uniprot.org/). The sequence
similarity of 73.1% and sequence identity of 58.3% between the A2B and the A2A receptors
and high resolution of the A2A receptor crystal structure justified the choice of this structure
as a template for the homology model. For each model, the variable target function method
28

with conjugate gradients in Modeller was applied to optimize the model, followed by
refinement using molecular dynamics with a simulated annealing method. Discrete
Optimized Protein Energy (DOPE) score was utilized to select the best model of A2B from the
100 generated models.54 The protonation of the selected model was achieved using the
Protonate3D algorithm applying Molecular Operating Environment (MOE 2016.08) followed
by minimization with a root mean square of 0.5 Å.56
4.5.2. Molecular docking
The selected homology model of the A2B receptor was applied for flexible ligand docking
using the molecular docking program, AutoDock 4.2.57 During the docking simulations, the
ligands were fully flexible while the residues of the receptor were treated as rigid. In order to
predict the binding modes of the compounds, selected compounds were docked into the
active site of the A2B receptor. The atomic partial charges were added using
AutoDockTools.57, 58 Fifty independent docking calculations using the varCPSO-ls algorithm
from PSO@Autodock implemented in AutoDock4.2 were performed and terminated after
500,000 evaluation steps.59 Parameters of varCPSO-ls algorithm, the cognitive and social
coefficients c1 and c2 were set at 6.05 with 60 individual particles as swarm size. All the
other parameters of the algorithm were set at their default values. Possible binding modes
of the compounds were explored by visual inspection of the resulting docking poses.
Acknowledgements
The authors would like to thank the BMBF (German Federal Ministry for Education and
Research) for the financial support within the BioPharma initiative “ euroallian ” (D11B
project).
We also thank Dr. K. Franke, Dr. A. Mansel and Dr. S. Fischer for providing [18F]fluoride and T.
Spalholz for her help with the in vitro studies. We are thankful to Dr. L. Hennig, Dr. S. Billig
and R. Oehme from the University of Leipzig for measuring the NMR and high resolution
mass spectra. We thank A. Fischer and A. Püsche for performing radioligand binding assays.
We are very thankful to Dr. D. Thimm for his support in the project management.
Supplementary data
Supplementary data related to this article can be found in the online version at...
These data include NMR and MS spectra, LC-MS chromatograms and the sequence
alignment of human adenosine receptor subtypes.
References
1.
inflammation. Trends Mol. Med. 2013, 19, 355.
2. Fredholm, B. B.; IJzerman, A. P.; Jacobson, K. A.; Linden, J.; Müller, C. E. International
Antonioli, L.; Pacher, P.; Vizi, E. S.; Hasko, G. CD39 and CD73 in immunity and
Union of Basic and Clinical Pharmacology. LXXXI. Nomenclature and Classification of
Adenosine Receptors-An Update. Pharmacol. Rev. 2011, 63, 1.
29