Synthesis and biological evaluation of novel 123I-labeled 4-(4-iodophenyl)butanoyl-l-prolyl-(2S)-pyrrolidines for imaging prolyl oligopeptidase in vivo

Article abstract of DOI:10.1016/j.ejmech.2014.04.014

Prolyl oligopeptidase (POP) may be associated with neuromodulation and development of neurodegenerative diseases and it was recently shown to participate in the inflammatory cascade along with matrix metalloproteinases. Radiotracers, which can be used for

Full text of DOI:10.1016/j.ejmech.2014.04.014

European Journal of Medicinal Chemistry 79 (2014) 436e445
Contents lists available at ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Synthesis and biological evaluation of novel 123I-labeled
4-(4-iodophenyl)butanoyl- -prolyl-(2S)-pyrrolidines for
imaging prolyl oligopeptidase in vivo
L
,
a
b
c
Annukka Kallinen a , Boyan Todorov , Roope Kallionpää , Susanne Bäck ,
*
Mirkka Sarparanta a, Mari Raki d, Juan A. García-Horsman c, Kim A. Bergström d
,
,
e
Erik A.A. Wallén b, Pekka T. Männistö c, Anu J. Airaksinen a
,
*
a Laboratory of Radiochemistry, Department of Chemistry, P.O. Box 55, FI-00014 University of Helsinki, Finland
b Division of Pharmaceutical Chemistry, Faculty of Pharmacy, P.O. Box 56, 00014 University of Helsinki, Finland
c Division of Pharmacology and Toxicology, Faculty of Pharmacy, P.O. Box 56, 00014 University of Helsinki, Finland
d Centre for Drug Research, Faculty of Pharmacy, P.O. Box 56, 00014 University of Helsinki, Finland
e HUS Medical Imaging Center, Helsinki University Central Hospital, Finland
a r t i c l e i n f o
a b s t r a c t
Article history:
Prolyl oligopeptidase (POP) may be associated with neuromodulation and development of neurode-
generative diseases and it was recently shown to participate in the inflammatory cascade along with
matrix metalloproteinases. Radiotracers, which can be used for non-invasive imaging, are needed for
investigating the role of POP in normal physiology and in pathophysiological conditions in vivo. We
Received 16 October 2013
Received in revised form
2 April 2014
Accepted 4 April 2014
Available online 12 April 2014
synthesized two novel POP-specific 123I-radiolabeled 4-phenylbutanoyl-
L-prolyl-pyrrolidines of which 4-
(4-[123I]iodophenyl)butanoyl-
L
-prolyl-2(S)-cyanopyrrolidine ([123I]2f, Ki ¼ 4.2 nM) was selected. The
selected compound has an electrophilic cyano group that is known to increase the dissociation time of
Keywords:
POP inhibitors. [123I]2f was synthesized in high radiochemical yield and purity (87 ꢀ 4%, >99%,
123I-radiotracer
Prolyl oligopeptidase
Biodistribution
Inflammation
respectively) and with a specific activity of 456 ꢀ 98 GBq/
m
mol. [123I]2f was evaluated in healthy mice
(C57Bl/6JRccHsd) by ex vivo biodistribution studies and SPECT imaging. Pretreatment with the known
inhibitor 4-phenylbutanoyl-
L
-prolyl-(2S)-cyanopyrrolidine (KYP-2047, 2d, Ki ¼ 0.023 nM) showed that
binding of [123I]2f was POP specific. In addition, [123I]2f was evaluated in models of neuroinflammation
and acute localized inflammation. A minor increase in binding of [123I]2f was observed in the inflamed
region in the acute localized inflammation model. Similar increase in binding was not observed in the
neuroinflammation model.
Ó 2014 Elsevier Masson SAS. All rights reserved.
1. Introduction
soluble cytosolic enzyme, although evidence for the existence of a
membrane bound form of the enzyme has also been presented [3].
Prolyl oligopeptidase (POP, also known as prolyl endopeptidase,
PREP) (EC 3.4.21.26) is an endopeptidase that hydrolyzes -proline
peptide bonds on the carboxylic acid side of -proline residue in
peptides comprised of a maximum of 30 amino acids. It belongs to a
larger prolyl oligopeptidase family S9 which includes enzymes
dipeptidyl peptidase IV (DPPIV), acylaminoacyl peptidase and oli-
gopeptidase B [1,2]. Most of the studies indicate that POP is a
POP has been suggested to have a role in the metabolism of neu-
ropeptides. However, studies with POP inhibitors have ended up
with contradicting results, maintaining the role of the enzyme in
the metabolism of the neuropeptides elusive. [4,5] Alternatively,
POP has been shown to be involved in the cycle of inositol
L
L
triphosphate (IP3),
a secondary messenger transmitting the
receptor-mediated signaling of several neuropeptides [6]. Lately,
POP activity has been correlated to an increase in the accumulation
of
a-synuclein, an insoluble fibrillar inclusion typically found in
Parkinson’s disease (PD) and in Lewy body dementia [7,8]. More-
over, it was shown very recently that the inhibition of POP resulted
Abbreviations: POP, Prolyl oligopeptidase; PD, Parkinson’s disease; MMP, matrix
metalloproteinases; LPS, lipopolysaccharide.
* Corresponding authors.
E-mail addresses: annukka.kallinen@helsinki.fi, annukka.kallinen@ansto.gov.au
(A. Kallinen), anu.airaksinen@helsinki.fi (A.J. Airaksinen).
in a decrease of
a-synuclein accumulation in an a-synuclein-
expressing transgenic mouse model [9], and that POP co-localizes
http://dx.doi.org/10.1016/j.ejmech.2014.04.014
0223-5234/Ó 2014 Elsevier Masson SAS. All rights reserved.

A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
437
with
a
-synuclein in the brain of PD patients [10]. In the same study
tight-binding inhibitor [23], because it forms an imino ether adduct
with the enzyme [24]. The long-lasting inhibition would provide a
possibility for the signal detection over several hours [23]. The
para-position of the phenyl ring was chosen as the site for intro-
ducing the 123I-label, as different substituents have previously been
added to this position without substantial changes in the binding
affinity. The aromatic position should also result in a relatively
stable radiolabeling, when considering the in vivo stability of the
final structure.
In this study we have described the synthesis, radiosynthesis
and in vitro characterization of the 123I-labeled analogues of POP
inhibitors 2a and 2d. We have also presented the first evaluation of
a radiolabeled POP inhibitor in vivo in mice. Considering the
possible involvement of POP in inflammation and its potential use
as an imaging biomarker for inflammation, we made a preliminary
evaluation of the selected radiolabeled POP inhibitor in two animal
models of acute inflammation.
it was also shown that POP co-localizes with the tau protein.
Recently, several studies have indicated that POP may have a
role in inflammatory diseases [11]. POP has been found in granule-
like structures in the cytoplasm of neutrophils [12]. Other studies
implicate concerted action of POP and matrix metalloproteinases
(MMP) in the degradation of collagen at the site of inflammation.
The MMPs degrade the collagen strands thus exposing prolyl-
glycyl-proline (PGP) moieties that serve as substrates for POP. The
tripeptide PGP, specifically cleaved and released by POP, can sub-
sequently act as a chemoattractant for neutrophils [11]. In multiple
sclerosis, an autoimmune disease where matrix metalloproteinases
are upregulated, POP activity has been found consistently altered in
the forms of the disease with a strong inflammatory component
[13]. These findings suggest a rationale for studying the possible
involvement of POP in neurodegenerative diseases where an in-
flammatory component is typically found, and support its possible
role as a disease biomarker [14].
In vivo molecular imaging is an important tool in investigation of
biological processes and molecular interactions in preclinical ani-
mal models. A synthesis and use of a tight-binding fluorescent POP
inhibitor for the detection of POP in cell cultures via optical imaging
has been reported earlier [15,16]. By using imaging modalities
based on distribution of radiolabeled tracer molecules, such as
positron emission tomography (PET) and single photon emission
computed tomography (SPECT), the preclinical imaging method-
ology can be readily translated to clinical applications. However,
investigation of the role of POP in normal physiology and in path-
ophysiological conditions, especially in the CNS, has been
hampered by the lack of an imaging tracer specific for POP. By
developing radiolabeled POP inhibitors for in vivo imaging, new
information on the localization of the enzyme and its role in neu-
romodulation and development of neurodegenerative and inflam-
matory diseases could be obtained. To the best of our knowledge,
only one research group has published radiolabeled POP inhibitors.
The syntheses of radiolabeled POP inhibitors ((2-(8-(N,N-([11C]
methyl)methylamino)octylthio))-6-isopropyl)-3-pyridyl 2-thienyl
ketone ([11C]1b, Ki ¼ 0.95 nM, rat brain POP) [17] and N-((N-(-4-
(N-[11C]methylamino)phenyl)butyryl)-L-prolyl)pyrrolidine ([11C]
2c, IC50 ¼ 3.1 nM, mouse brain POP) [18] from the corresponding
desmethylated precursors 1a and 2b (Fig. 1) via 11C-methylation
have been described previously, although no biological evaluation
of these tracers has been reported [19,20].
2. Results & discussion
2.1. Chemistry
The iodide substituted analogues 2e and 2f of the known POP
inhibitors 2a (SUAM-1221) and 2d (KYP-2047), respectively, were
synthesized in good yields following a general synthetic procedure
originally reported by Arai et al. (Scheme 1) [22,25].
2.2. Evaluation of the binding properties with in vitro enzyme assay
To determine whether the p-iodo-substituent on the phenyl ring
affected the inhibitory activity, the apparent Ki values of 2d, 2e and
2f were determined. The obtained values for 2d, 2e and 2f were 0.1,
22 and 4.2 nM, respectively (Fig. 2). These results indicate that the
p-iodination decreased the inhibitory activity of the compounds,
however the constants still are at low nanomolar range. Because of
the higher inhibitory potency over 2e, 2f was chosen for further
in vitro characterization and in vivo studies.
2.3. Radiochemistry
For the radiosynthesis, 2e and 2f were treated with hexame-
thylditin in the presence of Pd(PPh3)4-catalyst to give the stannane
precursors 5a and 5b in 73% and 75% yield, respectively (Scheme 2).
The radiosynthetic step was performed according to the methods
described in the literature [26]. First, the oxidation was optimized
with respect to radiochemical yield and the purity by testing
different oxidation agents. The relatively mild oxidant IodogenÒ
was tested for precursor 5b resulting in a low radiochemical yield of
To add a new tool for the research on the physiological role of
POP, and in search of new diagnostic applications, we decided to
synthesize 123I-labeled analogues of 2a (SUAM-1221) [21] and 2d
(KYP-2047) [22] (Fig. 1) and to investigate the biodistribution in
healthy mice with SPECT/CT. The Ki values for 2a and 2d are
0.97 nM and 0.023 nM, respectively [23]. Inhibitor 2d is a slow,
[
123I]2f (15%) (Table 1). Chloramine-T (CAT) gave moderate to high
radiochemical yields (80e90%), however, an unwanted UV-active
side product eluting simultaneously to the tracer was detected in
the HPLC purification. The side product may have resulted from the
chlorination side reactions often encountered with CAT based ox-
idants. Lastly, peracetic acid was tested, which resulted in high
radiochemical yields (>99%) and purities (>99%) of [123I]2e and
[
123I]2f in a small-scale synthesis (radioactivity 10 MBq, 1.14 pmol
of 123I). The optimal reaction conditions proved to be the reaction
time of 10 min at room temperature in aqueous solution containing
20 equiv of peracetic acid in neutral conditions, and the use of
12 equiv of metabisulphite to quench the oxidant. On the basis of
in vitro enzyme assay studies, it was decided to continue to in vivo
evaluation with tracer [123I]2f showing a higher affinity to the
target enzyme. The optimized reaction conditions were used in a
larger scale synthesis with the starting activities ranging from 300
to 680 MBq. [123I]2f was obtained with an average radiochemical
Fig. 1. Structures of POP inhibitors 1a (desmethyl Y-29794) and 2a (SUAM-1221) and
their radiolabeled derivatives.

438
A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
Scheme 1. Synthesis of the stable reference compounds 2e and 2f. Reagents and conditions: (a) SOCl2, 65 ꢁC, 2 h; (b) L-proline, Na2CO3, Et2O/H2O, 0 ꢁC / room temperature (74%);
(c) for 2e, 1.) Pivaloyl chloride, TEA, DCM, 0 ꢁC, 1 h, 2.) Pyrrolidine, TEA, DCM, 0 ꢁC / room temperature (81%); for 2f, 1.) Pivaloyl chloride, DIPEA, DCM, 0 ꢁC, 1 h, 2.) 2(S)-cya-
nopyrrolidine (as TFA salt), DIPEA, DCM, 0 ꢁC / room temperature (57%).
was 2.37 ꢀ 0.11, indicating a suitable lipophilicity for the passive
CNS targeting.
2.5. Biodistribution studies
The biodistribution of [123I]2f was investigated in healthy male
mice by measuring the radioactivities of selected excised organs at
5, 20, 40, 80 and 160 min post injection, and non-invasively by
SPECT-imaging. The overall biodistribution of radioactivity after
intravenous injection of [123I]2f (2.86 ꢀ 1.26 MBq, 1% DMSO in 0.9%
NaCl) in mice in main organs as a function of time is presented in
Fig. 3. In the early time points, high accumulation of radioactivity in
lungs was seen. The radioactivity in lungs decreased steadily over
time, approximately half of the radioactivity being left at 40 min
post injection. Organs of moderate radiotracer uptake were heart,
kidney, liver and spleen, all of which are reported to have a high
Fig. 2. The kinetic studies of POP inhibition were conducted on the recombinant
expression level of the enzyme in mouse [29,30]. In addition, we
porcine POP enzyme and the inhibitors 2e, 2f and 2d (KYP-2047). Apparent Ki’s were
found binding in salivary gland and tongue. The uptake in thyroid
estimated from the intercepts of the dotted lines at the inhibitor concentration axis in
was initially 52 %ID/g at 5 min timepoint and decreased to 9.4 %ID/g
this graph. Apparent Km/Vmax values were obtained from the reciprocals varying
at 160 min (n ¼ 1), indicating that relatively little deiodination took
substrate concentrations at the given inhibitor concentrations.
place in vivo.
Previous studies have demonstrated that POP is widely
distributed not only in peripheral tissues but also in brain [30]. The
brain uptake of [123I]2f remained low over time. The brain uptake of
[
123I]2f was initially 0.43 ꢀ 0.14 %ID at 5 min, increased to
0.76 ꢀ 0.58 %ID at 40 min, and reached a plateau at 80 min
(0.27 ꢀ 0.08 %ID) (Fig. 3). The distribution of radioactivity in
different brain regions was found to be highly homogenous, which
was confirmed by ex vivo autoradiography studies (data not
shown). Consequently, the brain uptake of [123I]2f was found to be
insufficient for CNS imaging.
To determine the specific binding of [123I]2f to POP in vivo, the
effect of pretreatment with a known, high-affinity POP-inhibitor 2d
Scheme 2. Radiosynthesis of [123I]2e and [123I]2f. Reaction conditions: (a) hexame-
thylditin, PdPPh4, toluene (5a, 73%; 5b, 75%); (b) 1. [123I]NaI, peracetic acid (aq), 10 min;
2. Na2S2O5 (87 ꢀ 4%, decay corrected yield).
(KYP-2047, Ki ¼ 0.023 nM, 50
mmol/kg, 10 mL/kg) [31] on the bio-
distribution of radioactivity was examined. 2d was injected intra-
peritoneally 10 min prior the tracer administration, as described
earlier elsewhere [31]. Specific binding could clearly be seen starting
from the 40-min time point and it was noted to further enhance
during the later time points, indicating faster wash-out from the
non-specific binding sites. The results of the study at 80 min post
injection are compiled in Table 1. The binding was found to be
specific for POP in brain, liver, lung, heart, spleen, muscle, salivary
gland (p ꢂ 0.05) and quite surprisingly, in tongue (t ¼ 160 min,
blocking p ¼ 0.003). The higher radioactivity levels in the urine
samples indicated an increase in the renal elimination rate after 2d
pretreatment. Accordingly, the increase in the elimination rate of
incorporation yield of 82 ꢀ 4% (n ¼ 4) and with the radiochemical
purity of >99%. The specific activity was 456 ꢀ 98 GBq/
mmol at the
end of synthesis.
2.4. In vitro stability and LogP
The tracer stability was studied in human plasma at 37 ꢁC up to
120 min (2.5 MBq of tracer/1 mL of plasma). The tracer [123I]2f
showed excellent stability in vitro, no decomposition or radioactive
metabolites could be detected by HPLC (system A, method A1). The
average amount of 6.3% ꢀ 0.7% of the added radioactivity was found
to bind in plasma proteins in vitro over 2 h. In the development of
CNS radiotracers, moderate lipophilicity (LogP 1.5e2.7) is needed to
ensure the passive entry of radiotracer through the BBB into the
brain [27]. The LogP value of [123I]2f was determined experimen-
tally by the shake flask method [28]. The determined LogP value
[
123I]2f after 2d pretreatment could be clearly seen in the SPECT
images as highly active renal and hepatobiliary excretion (Fig. 4).
2.6. Radioactive metabolites of [123I]2f in plasma
Radioactive metabolites of [123I]2f were analyzed by HPLC from
plasma after i.v. injection of the tracer to mice.
A HPLC

A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
439
Table 1
radioactivity in blood cells, plasma proteins and free plasma at 20,
40, 80 and 160 min time points are presented in Table 2. Radio-
tracers that bind avidly to plasma proteins or blood cells have
reduced amount of freely available tracer in circulation, and as a
consequence, usually show a low brain uptake [33]. The plasma
protein bound fraction constituted w3% of the activity in blood.
However, [123I]2f was found to partition largely into blood cells; at
20 min, over 90% of the activity in blood resided in the cell fraction.
The activity of free, non-protein bound plasma fraction was found
to increase over time, at the expense of the decreasing activity in
blood cell fraction. The increased activity in the protein-free plasma
was mostly due to the accumulation of the main radiometabolite
over time. At the 80 min timepoint, the non-bound parent com-
pound accounted only for 0.8% of the activity in blood.
Interestingly, as the selective POP inhibitor 2d was administered
prior to the tracer injection, the activity in blood (%ID/g) was not
changed. However, the activity bound to blood cells decreased by
43.5% (at 20 and 160 min), indicating that the binding of the tracer
in the blood cells was partly POP specific. In humans, POP has been
found in thrombocytes and neutrophils [12,34]. In mice, POP has
been found in the peripheral, naive T-cells, although in less amount
than in immature thymocytes [35]. It is to be noted though that the
binding of [123I]2f in blood cells may increase the background
signal and hence hinder the use of the tracer for imaging purposes.
Biodistribution of [123I]2f in C57Bl/6JRccHsd male mice at 80 min in baseline con-
ditions and after pretreatment with 2d (KYP-2047, 50 mmol/kg, 10 mL/kg, i.p. 10 min
prior to the tracer injection).a
Organ
80 min control
80 min blocking
Blocking (%)
pb
Blood
Liver
Lung
2.10 ꢀ 0.93
20.10 ꢀ 3.81
14.67 ꢀ 2.77
12.20 ꢀ 5.05
10.02 ꢀ 4.33
10.56 ꢀ 3.26
0.55 ꢀ 0.18
2.23 ꢀ 2.54
2.15 ꢀ 0.60
7.77 ꢀ 2.21
2.69 ꢀ 0.68 (n ¼ 2)
5.05 ꢀ 1.69
28
ꢃ75
ꢃ89
ꢃ89
ꢃ53
ꢃ91
5
1933
ꢃ73
ꢃ80
nd
0.002
0.033
0.023
0.088
0.036
0.850
nd
1.54 ꢀ 0.46
Heart
Kidney
Spleen
Testis
Urine
Muscle
Salivary
gland
Tongue
Brain
1.34 ꢀ 0.38
4.73 ꢀ 1.22
0.92 ꢀ 0.24
0.57 ꢀ 0.15
45.24 ꢀ 3.41 (n ¼ 2)
0.59 ꢀ 0.19
0.010
0.010
1.52 ꢀ 0.28
4.43 ꢀ 0.65 (n ¼ 2)
0.62 ꢀ 0.21
1.44 ꢀ 0.31
0.08 ꢀ 0.02
ꢃ68
ꢃ88
nd
0.013
a
Data are expressed as percentage of injected dose per gram ꢀ SD, n ¼ 3.
b
p-values for the baseline versus blocking group at 80 min post injection by
Student’s t test (two-tailed, independent).
2.8. Evaluation of [123I]2f for the imaging of POP in the models of
acute neuroinflammation and acute localized inflammation
As we were looking into a possibility of using POP as a
biomarker for inflammation, we tested the tracer [123I]2f in models
of acute inflammation in the CNS and in the periphery. Lipopoly-
saccharide (LPS) induced acute neuroinflammation model [36,37]
was chosen for the evaluation of [123I]2f as a POP specific tracer
for the imaging of neuroinflammation. In the LPS model of neuro-
inflammation, BBB is reported to be compromised, thus increasing
its permeability in both hemispheres [38]. C57Bl/6JRccHsd mice
(n ¼ 3) were subjected to stereotaxic injections of LPS in the left
striatum 24 h before i.v. injection of [123I]2f. The biodistribution of
the tracer was studied up to 80 min after tracer injection. The right
hemisphere was intact and served as a control in each mouse. In the
LPS-induced neuroinflammation model, we did not observe any
difference in the uptake of the tracer between the affected
(1.34 ꢀ 0.37 %ID/g) and the control hemispheres (1.09 ꢀ 0.38 %ID/g)
(Fig. 6A). The overall brain uptake of the radiotracer, however, was
noted to increase from 0.27 ꢀ 0.08 %ID to 0.60 ꢀ 0.14 %ID in the LPS-
treated mice compared to the healthy mice due to the disrupted
BBB, as expected. Although an increase in the POP activity of the
neutrophils treated with LPS has been earlier reported [12], there is
no previous evidence on the increased POP enzyme levels in the
selected animal model. The insufficient binding in the inflamed
hemisphere may have been either due to the insensitivity of [123I]2f
to detect the increased POP enzyme levels in the brain or due to
lack of increase in the POP levels in this particular model.
Fig. 3. Biodistribution of radioactivity in an adult C57Bl/6JRccHsd mouse after intra-
venous injection of [123I]2f in selected organs. %ID/g is percentage of injected dose per
gram. At 5 min n ¼ 2; at 20, 40 and 160 min n ¼ 3; at 80 min n ¼ 4.
chromatogram of a plasma sample collected at 20 min post injec-
tion is shown in Fig. 5A. The activity is presented as a ratio cpm/
total cpm. The retention time (tr) of the parent compound [123I]2f
varied between 7.15 and 7.80 min. One less lipophilic radioactive
metabolite was detected in plasma (tr ¼ 4 min). The ratio of the
parent tracer and the metabolite in blood over time is shown in
Fig. 5B (5e40 min, n ¼ 1; 80 and 160 min, n ¼ 2). After 20 min, the
observed radioactive metabolite accounted for 61% of the radioac-
tivity found in blood.
We tried to analyze the possible radioactive brain metabolites,
however, in the extraction process approximately 70% of the
radioactivity was left unextracted in SPE cartridge, compromising
the validity of these results. Compounds that have a similar cya-
nopyrrolidine moiety found in [123I]2f structure have been shown
to decompose mainly via hydrolysis of the cyano group to a car-
boxylic acid in vivo [32]. The presence of a hydrophilic radio-
metabolite found in plasma could hence be explained by the cyano
group hydrolysis. However, the hydrophilic radiometabolite would
not likely enter the brain as readily as the more lipophilic parent
tracer.
To determine the sensitivity of [123I]2f for detection of POP in
acute peripheral inflammation, we studied the uptake of [123I]2f in
the inflamed tissue in turpentine-induced model of acute localized
inflammation in C57Bl/6JRccHsd mice (n ¼ 2). The mice were
subjected to i.m. injections of turpentine in the right musculus
gastrocnemius 24 h prior to tracer administration. The left m.
gastrocnemius was treated with i.m. saline injection as a control.
Mice were scanned (60 min) by SPECT-CT at 20 min, 4 h and 24 h
after tracer administration, after which the ex vivo biodistribution
studies were done. In the ex vivo studies (timepoint 24 h), we
observed a trend of increase in the uptake of [123I]2f in the inflamed
muscle as compared to the control area (2.0 ꢀ 0.6 %ID/g and
2.7. Binding to blood components in vivo
We also determined the binding of [123I]2f to blood components
in mice. The radioactivity in blood and the distribution of

440
A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
Fig. 4. SPECT-CT images of C57Bl/6JRccHsd male mice after injection of radiotracer [123I]2f. Uptake of [123I]2f is seen in liver, lungs and in tongue (top row). After pretreatment with
a known POP-inhibitor 2d (KYP-2047) (bottom row), binding of the radiotracer [123I]2f in these organs is blocked and the tracer is eliminated faster.
0.8 ꢀ 0.1 %ID/g, respectively) (Fig. 6B). However, the accumulation
of the tracer in the inflamed area was clearly not high enough for
imaging of inflammation.
performed under argon atmosphere in dry solvents. Flash chro-
matography was performed on Merck silica gel 60 (0.040e
0.063 mm) (Merck, Whitehouse Station, NJ, USA). Thin layer chro-
matography (TLC) was carried out on silica gel-coated aluminum
sheets (Merck, TLC silica gel 60 F254) using solvent mixtures of n-
hexane and ethyl acetate. Compounds were visualized by UV-light
(254 nm) or by I2 vapor. Melting points were measured using Büchi
510 melting point apparatus (Büchi Labortechnik AG, Flawil,
Switzerland). Optical rotations were determined using wavelength
589 nm at 25 ꢁC by Jasco DIP-1000 polarimeter (Jasco Inc. Mary’s
Court, MD, USA). NMR spectra were recorded in CDCl3 on Varian
300 MHz (1H NMR, 300 MHz; 13C NMR, 75 MHz) spectrometer
(Varian Inc. Palo Alto, CA, USA). TMS (1H) and CDCl3 (13C) were used
3. Conclusions
In vivo imaging of POP with g-emitting radiotracers offers a new
tool for studying the localization and the physiological role of this
pharmacologically interesting target. In this study, we successfully
synthesized two SPECT radiotracers [123I]2e and [123I]2f with high
radiochemical yields and purities and reported the first biological
evaluation of a radiolabeled POP inhibitor. [123I]2f showed high
inhibitory activity (Ki ¼ 4.2 nM) for the target enzyme and POP
specific binding in vivo. The slow dissociation of [123I]2f from the
enzyme would allow a selective detection of the POP protein for
several hours. [123I]2f is a potential radiotracer for studying pe-
ripheral POP levels in vivo by using SPECT imaging, however, its
insufficient brain uptake hampers its use in assessing changes in
the enzyme levels in the brain. Although [123I]2f was not successful
in the detection of inflammation in the chosen models of inflam-
mation in this preliminary study, the potential of POP enzyme as a
possible biomarker and pharmaceutical target certainly deserves
further investigation.
as internal standards. All recorded spectra were given in ppm (d).
Mass analyses were performed on electrospray ionization mass
spectrometer (Bruker Esquire 3000 plus) (Bruker Corporation,
Billerica, MA, USA). Microanalyses (C, H, N) were carried out on a
VarioMICRO apparatus (Elementar Analysensysteme GmbH,
Hanau, Germany). The chemical purities of non-radioactive com-
pounds were ꢄ95%, unless otherwise stated. The purities were
assessed by HPLC and CHN analysis. All HPLC-analyses and HPLC-
purifications used water (A) and MeCN (B) as eluents, unless
otherwise stated. HPLC system A consisted of two Shimadzu LC-
20AD pumps, a Shimadzu SPD-M20A diode array detector (Shi-
madzu Corporation, Kyoto, Japan) and NaI(Tl) radiodetector 925-
4. Experimental section
SCINT (Ortec, Oak Ridge, TN, USA). The HPLC method A1 used
ꢀ
analytical
mBondapak
C18-column
(10
mm,
125
A,
4.1. General
3.9 mm ꢅ 300 mm, Waters Corporation, Milford, MA, USA) and
gradient elution starting from B ¼ 40e50% (0e5 min), 50e60% (5e
6 min), 60e70% (6e7 min), 70e80% (7e8 min), 80% (8e15 min),
80e40% (15e17 min), 40% (17e22 min), flow rate 1.0 mL/min. The
All chemicals, reagents and solvents for the synthesis of the
compounds were analytical grade, purchased from SigmaeAldrich
chemical company (St. Louis, MO, USA) and used without further
purification. All air- and moisture-sensitive reactions were
HPLC method A2 used analytical Kinetex PFP-column (2.6
mm,

A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
441
4.2. Chemistry
4.2.1. 4-(4-Iodophenyl)butanoyl- -proline (4)
L
4-(4-Iodophenyl)butanoic acid (3) (1.0 g, 3.5 mmol) was dis-
solved in SOCl2 (1.3 mL, 17 mmol) and reacted for 2 h at 65 ꢁC.
Excess SOCl2 was evaporated under reduced pressure after the
reaction. The resulting 4-(4-iodophenyl)butanoyl chloride was
dissolved in a small volume of diethyl ether and slowly added to
a solution of Na2CO3 (0.93 g, 8.8 mmol) and L-proline (0.40 g,
3.5 mmol) in water (6.6 mL) at 0 ꢁC under vigorous stirring. The
mixture was stirred vigorously overnight at room temperature.
The aqueous layer was washed with diethyl ether, then acidified
using 2 M HCl, and extracted three times with diethyl ether. The
combined diethyl ether phases from the extractions were dried
with anhydrous Na2SO4 and evaporated under reduced pressure
yielding 4 (0.98 g, 2.6 mmol, 74%). M.p. 108 ꢁC. 1H NMR
(300 MHz, CDCl3) d/ppm 9.97 (br s, COOH, 1H), 7.60 (d, ArH,
3JH,H ¼ 8.1 Hz, 2H), 6.94 (d, ArH, 3JH,H ¼ 8.1 Hz, 2H), 4.61e4.55 (m,
NCHCOO, 1H), 3.53e3.30 (m, NCH2CH2, 2H), 2.64 (t, PhCH2CH2,
3JH,H ¼ 7.4 Hz, 2H), 2.50e1.84 (m, 2ꢅCH2CH2CH2; CH2CH2CH;
CH2CH2CO, 8H). 13C NMR (75 MHz, CDCl3)
d/ppm 174.35 (COO),
173.41 (CON), 141.23 (ArCCH2), 137.42 (ArCH, 2C), 130.68 (ArCH,
2C), 91.03 (ArCI), 59.31 (NCHCO), 47.50 (NCH2CH2), 34.48
(PhCH2), 33.29 (COCH2), 28.45 (CH2CH2CH), 25.78 (CH2CH2CH2),
24.71 (CH2CH2CH2).
4.2.2. 4-(4-Iodophenyl)butanoyl- -prolyl-pyrrolidine (2e)
L
A solution of pivaloyl chloride (0.097 mL, 0.79 mmol) in DCM
(0.6 mL) was slowly added to a solution of 4 (0.30 g, 0.79 mmol) and
Et3N (0.12 mL, 0.86 mmol) in DCM (1.7 mL) under argon at 0 ꢁC. After
1 h at 0 ꢁC a solution of Et3N (0.12 mL, 0.86 mmol) and pyrrolidine
(0.066 mL, 0.79 mmol) in DCM (0.6 mL) was added. The reaction
mixture was allowed warm to room temperature and react over-
night. DCM had to be added to dissolve precipitation that had
formed. The solution was washed once with water, twice with
saturated aqueous NaHCO3 and once with water. The organic phase
was dried with anhydrous Na2SO4 and the solvent was evaporated
under reduced pressure. Flash chromatography (5% MeOH in EtOAc)
Fig. 5. (A) HPLC-chromatogram of the plasma sample at 20 min post injection of [123I]
2f. (B) The ratio of the free parent compound [123I]2f to the main metabolite in plasma
over time.
ꢀ
100 A, 4.60 mm ꢅ 150 mm, Phenomenex, Torrance, CA, USA) and
gave 2e (0.28 g, 0.64 mmol, 81%). M.p. 81 ꢁC. [
a] D
25
¼ þ10.5ꢁ (c 1.0, toluene).
isocratic elution. The HPLC-method A3 used semipreparative
1H NMR (300 MHz, CDCl3)
d
/ppm 7.57 (d, ArH, 3JH,H ¼ 8.4Hz, 2H), 6.95
ꢀ
m
Bondapak C18-column (10
m
m, 125 A, 7.8 mm ꢅ 300 mm, Waters)
3
3
(d, ArH, JH,H ¼ 8.4 Hz, 2H), 4.63 (dd, NCHCO, JH,H ¼ 8.4 Hz,
3JH,H ¼ 4.0 Hz, 1H), 3.81 (m, NCH2CH2CH2CHCO, 1H), 3.59 (m,
NCH2CH2CH2CHCO and NCH2CH2CH2CH2, 2H), 3.40 (m,
and gradient elution starting from B ¼ 30% (0e0.3 min), 30e60%
(0.3e4 min), 60% (4e7 min), 60e80% (7e10 min), 80% (10e11 min),
80e30% (11e14 min), 30% (14e17 min), flow rate 4.0 mL/min. HPLC
system B consisted of two Hitachi LaChrom L-7100 pumps (Hitachi
Ltd., Tokyo, Japan) with Rheodyne 7010 injector, LKB 2151 UV-
detector (LKB, Bromma, Sweden) and PIN diode radiodetector.
3
NCH2CH2CH2CH2; NCH2CH2, 3H), 2.61 (t, PhCH2CH2, JH,H ¼ 7.5 Hz,
2H), 1.77e2.39 (m, CHCH2CH2; CH2CH2CO; 4ꢅCH2CH2CH2, 12H). 13
C
NMR (75 MHz, CDCl3) d/ppm 171.80 (CON), 170.90 (CON), 141.80
(ArCCH2), 137.64 (ArCH, 2C), 131.04 (ArCH, 2C), 91.18 (ArCI), 58.15
(NCHCO), 47.68 (NCH2CH2, 2C), 34.92 (NCH2), 33.55 (PhCH2), 29.15
(COCH2), 26.11 (CH2CH2CH2; CH2CH2CH, 2C), 25.12 (3ꢅCH2CH2CH2,
3C). ESI-MS m/z ¼ 463.2 [MþNa]þ. HPLC system A, method A2, sol-
The HPLC-method B1 used semipreparative mBondapak C18-column
ꢀ
(10
m
m, 125 A, 7.8 mm ꢅ 300 mm, Waters) and isocratic elution of
B ¼ 33%, flow rate 4.5 mL/min. All animal experiments were per-
formed according to European Community Guidelines for animal
experimentation and approved by the Finnish National Board of
Animal Experiments.
vent B ¼ 33%, flow rate 1.15 mL/min,
l
¼ 254 nm, tr ¼ 10.27 min
(>99%). Calc.: C 51.81%, H 5.72%, N 6.36% Found: C 51.89%, H 5.77%, N
6.28%.
Table 2
Radioactivity in blood after i.v. injection of [123I]2f in baseline (A) and after pretreatment with 2d (B) at 20, 40, 80 and 160 min timepoints (n ¼ 3). Distribution of radioactivity in
different blood components in the conditions A and B is presented as percentage of the total activity in the blood after injection of [123I]2f.
Time (min)
20
40
80
160
Condition
A
B
A
B
A
B
A
B
Blood (%ID/g)
Unbound (%)
Cell bound (%)
Protein bound (%)
8.5 ꢀ 4.0
10.9 ꢀ 4.9
7.6 ꢀ 4.8
8 ꢀ 2b
90 ꢀ 0b
2 ꢀ 1b
4.0 ꢀ 1.6
2.1 ꢀ 0.9
15 ꢀ 2b
84 ꢀ 0b
1 ꢀ 1b
2.7 ꢀ 0.7
2.1 ꢀ 0.9
26 ꢀ 7b
73 ꢀ 6b
1 ꢀ 2b
1.0 ꢀ 0.2
3a
47a
53a
0a
nd
nd
nd
nd
nd
nd
58a
95a
2a
42a
0a
a
n ¼ 1.
b
n ¼ 2.

442
A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
(NCH2), 46.66 (NCH2), 34.87 (COCH2), 33.55 (PhCH2), 30.02
(CNCHCH2), 29.03 (COCHCH2), 26.05 (PhCH2CH2CH2), 25.69
(NCH2CH2CH2), 25.19 (NCH2CH2CH2). ESI-MS m/z ¼ 488.3 [MþNa]þ.
HPLC system A, method A2, solvent B ¼ 33%, flow rate 1.15 mL/min,
l
¼ 254 nm, tr ¼ 10.65 min (>99%). Calc.: C 51.62%, H 5.20%, N 9.03%
Found: C 51.83%, H 5.46%, N 8.81%.
4.2.4. 4-(4-Trimethylstannylphenyl)butanoyl-L-prolyl-pyrrolidine
(5a)
To a solution of iodide (2e) (30 mg, 0.069 mmol) in toluene
(1.0 mL) were added hexamethylditin (57 mL, 0.276 mmol in toluene
0.3 mL) and Pd(PPh3)4 (8.3 mg, 0.007 mmol in toluene 1.0 mL). After
being refluxed at 110 ꢁC for 3 h, reaction mixture was cooled down
to room temperature and filtered through a pad of celite. The
filtrate was concentrated under reduced pressure to give oily crude
product. Column chromatography on silica gel (EtOAc) gave 5a as
ꢁ
colorless oil (24 mg, 0.050 mmol, 73%). [
a
]
25
D
¼ þ116.3 (c 0.05, MeCN).
3
1H NMR (300 MHz, CDCl3)
d
/ppm 7.39 (d, ArH, JH,H ¼ 7.9 Hz, 2H),
7.18 (d, ArH, 3JH,H ¼ 7.9 Hz), 4.63 (dd, 3JH,H ¼ 8.4 Hz, 3JH,H ¼ 3.7 Hz,
NCHCO, 1H), 3.83 (m, NCH2CH2CH2CHCO, 1H), 3.50e3.70 (m,
NCH2CH2CH2CHCO, 2H), 3.29e3.48 (m, NCH2CH2CH2CH2, 3H), 2.65
(t, PhCH2CH2, 3JH,H ¼ 7.4 Hz, 2H), 2.30 (m, CH2CH2CHCO, 2H), 1.73e
2.26 (m, 4ꢅCH2CH2CH2, CH2CH2CO, 10H), 0.26 (s, (CH3)3Sn, 9H). 13
C
NMR (75 MHz, CDCl3) d/ppm 171.73 (CH2CON), 171.03 (CHCON),
142.19 (ArCH), 139.19 (ArCCH2), 136.12 (ArCSn(CH3)3), 128.71
(ArCH), 57.94 (NCHCO), 47.53 (NCH2), 46.60 (NCH2), 46.23 (NCH2),
35.45 (CH2CO), 33.92 (PhCH2), 29.16 (CHCH2CH2), 26.50
(CH2CH2CH2), 26.14 (CH2CH2CH2), 25.10 (CH2CH2CH2), 24.44
(CH2CH2CH2), ꢃ9.28 ((CH3)3Sn). ESI-MS m/z ¼ 501.3 [MþNa]þ.
HPLC system A, method A2, solvent B ¼ 45%, flow rate 1.2 mL/min,
Fig. 6. The uptake of radioactivity in left and right brain hemispheres at 80 min post
i.v. injection of tracer [123I]2f administered at 24 h after LPS injection in left striatum of
C57Bl/6JRccHsd male mouse (A). The uptake of radioactivity in right and left m.
gastrocnemius at 24 h post i.v. injection of radiotracer [123I]2f administered at 24 h after
l
¼ 220 nm, tr ¼ 7.65 min (86%).
turpentine (5 mL, i.m.) injection to right hind leg of C57Bl/6JRccHsd male mouse (B).
4.2.5. 4-(4-Trimethylstannylphenyl)butanoyl-L-prolyl-2(S)-
cyanopyrrolidine (5b)
4.2.3. 4-(4-Iodophenyl)butanoyl-
(2f)
L
-prolyl-2(S)-cyanopyrrolidine
Compound 5b was synthesized and purified as above using 2f as
starting material. Purification gave 5b as colorless oil (75%).
ꢁ
Boc-2(S)-cyanopyrrolidine was synthesized according to
a
[a
]
25
D
1H NMR (300 MHz, CDCl3)
d/ppm 7.40 (d,
¼ ꢃ103.5 (c 0.05, MeCN).
3
3
literature procedure [39]. TFA (20 mL, 260 mmol) was added
dropwise to a solution of Boc-2(S)-cyanopyrrolidine (0.99 g,
5.0 mmol) in DCM (60 mL) under argon at 0 ꢁC. After 2 h 20 min the
solvent was evaporated yielding the TFA salt of 2(S)-cyanopyrroli-
dine, which was used directly in the reaction below.
A solution of pivaloyl chloride (0.31 mL, 2.5 mmol) in DCM
(5 mL) was slowly added to a solution of 4 (0.98 g, 2.5 mmol) and
DIPEA (0.49 mL, 2.8 mmol) in DCM (50 mL) under argon at 0 ꢁC.
After 1 h at 0 ꢁC a solution of DIPEA (2.9 mL, 17 mmol) in a small
volume of DCM was added. Directly after this, a solution of the TFA
salt of 2(S)-cyanopyrrolidine in a small volume DCM was added.
The reaction mixture was allowed warm to room temperature and
react overnight. The reaction mixture was diluted with DCM and
washed three times with 10% aqueous citric acid, once with brine
and twice with 10% aqueous NaHCO3. The organic phase was dried
using anhydrous Na2SO4 and the solvent was evaporated. Flash
chromatography (1% MeOH in DCM) yielded 2f (670 mg, 1.4 mmol,
ArH, JH,H ¼ 7.9 Hz, 2H), 7.17 (d, ArH, JH,H ¼ 7.9 Hz, 2H), 4.83 (m,
3
3
NCHCO, 1H), 4.56 (dd, JH,H ¼ 8.1 Hz, JH,H ¼ 4.0 Hz, NCHCN, 1H),
3.89 (m, NCH2CH2CH2CHCO, 1H), 3.63 (m, NCH2CH2CH2CHCO, 1H),
2
3.63 (m, NCH2CH2CH2CHCN, 1H), 3.45 (dt, JH,H
¼
9.7 Hz,
3JH,H
¼
7.0 Hz, NCH2CH2CH2CHCN, 1H), 2.64 (t, PhCH2CH2,
3JH,H ¼ 7.5 Hz, 2H), 2.22 (m, CH2CH2CO; (CN)CHCH2CH2; (CO)
CHCH2CH2; PhCH2CH2CH2, 8H), 1.96 (m, NCH2CH2CH2, 4H), 0.25 (s,
(CH3)3Sn, 9H). 13C NMR (75 MHz, CDCl3)
d/ppm 172.04 (CH2(CO)N),
171.69 (CH(CO)N), 141.98 (ArCCH2), 139.36 (ArCSn), 136.17 (ArCH,
2C), 128.67 (ArCH, 2C), 118.96 (CN), 57.69 (NCHCO), 47.57 (NCHCN),
46.83 (NCH2), 46.66 (NCH2), 35.37 (COCH2), 33.81 (PhCH2), 30.00
(CNCHCH2), 29.04 (COCHCH2), 26.15 (PhCH2CH2CH2), 25.68
(CH2CH2CH2), 25.19 (CH2CH2CH2), ꢃ9.27 ((CH3)3Sn, 3C). ESI-MS m/
z ¼ 526.3 [MþNa]þ. HPLC system A, method A2, solvent B ¼ 45%,
flow rate 1.2 mL/min,
l
¼ 220 nm, tr ¼ 10.29 (92%).
4.2.6. Evaluation of the binding properties with in vitro enzyme
assay
57%). M.p. 128 ꢁC. [
CDCl3)
a
]
25
D
3). 1H NMR (300 MHz,
¼
ꢃ71.9ꢁ (c 1.0, CHCl
3
d
/ppm 7.58 (d, ArH, JH,H ¼ 8.3 Hz, 2H), 6.94 (d, ArH,
The effect of 2e and 2f on the enzymatic cleavage of benzylox-
ycarbonyl-glycyl-alanyl-prolyl-7-amino-4-methyl-coumarine (Z-
Gly-Ala-Pro-AMC) by POP, to free the fluorescent AMC, was per-
formed by monitoring the fluorescence change on a reader (Victor-
2, PerkinElmer, Waltham MA, USA). Reactions were performed at
3JH,H ¼ 8.3 Hz, 2H), 4.81 (m, NCHCO, 1H), 4.55 (dd, NCHCN,
3JH,H ¼ 8.1 Hz, 3JH,H ¼ 4.0 Hz, 1H), 3.86 (m, NCH2CH2CH2CHCO, 1H),
3.62 (m, NCH2CH2CH2CHCO, 1H), 3.62 (m, NCH2CH2CH2CHCN, 1H),
3.44 (dt, 2JH,H ¼ 13.6 Hz, 3JH,H ¼ 7.0 Hz, NCH2CH2CH2CHCN,1H), 2.60
3
(t, PhCH2CH2, JH,H ¼ 7.1 Hz, 2H), 2.22 (m, CH2CH2CO; (CN)
37 ꢁC for 60 min in a 96-well plates in a final volume of 100
mL
CHCH2CH2; (CO)CHCH2CH2; PhCH2CH2CH2, 8H), 1.95 (m,
which contained 0.2, 0.5 or 1 mM Z-Gly-Pro-AMC, 10 fmol of re-
combinant porcine POP, in assay buffer (75 mM phosphate buffer
pH ¼ 7, and 5 mM DTT), in the absence or presence of increasing
concentrations of the parent KYP-2047 or iodinated compounds
NCH2CH2CH2, 4H). 13C NMR (75 MHz, CDCl3)
d/ppm 171.75 (CON),
171.59 (CON),141.64 (ArCCH2), 137.67 (ArCH, 2C), 130.98 (ArCH, 2C),
118.92 (CN), 91.23 (ArCI), 57.75 (NCHCO), 47.60 (NCHCN), 46.83

A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
443
(0.5 nMe1 mM). The reaction was stopped by adding 100
ammonium acetate (1 M, pH ¼ 4). Fluorescence was read at exci-
tation 360 nm and emission 460 nm. Fluorescence values were
interpolated on an AMC standard curve (0.5e10 nmol AMC in
100 L of assay buffer). Background fluorescence of the amounts
used of Z-Gly-Ala-Pro-AMC were subtracted. Ki was determined
from the intercepts in the double reciprocal plots of data fitted by
linear regression of the mean of the activities (triplicate), at every
condition, using data of three independent experiments.
m
L of
The radiochemical purity of the tracer [123I]2f was >99% at the end
of synthesis, and 98.9 ꢀ 2% after 24 h. Specific activity was 456 ꢀ 98
l
l
GBq/mmol at the end of synthesis.
m
4.3.3. Determination of the partition coefficient (LogP)
Lipophilicity of [123I]2f was assessed by the determination of the
partition of the compound between water and n-octanol phases
following a method described in the literature [27]. Briefly, [123I]2f
(40 mL,1% DMSO in 0.9% NaCl, 700 kBq) was added to the mixture of
phosphate buffer (10 mL, 0.02 M, pH 7.4) and n-octanol (10 mL) in a
separation funnel. The mixture was shaken for 3 min, after which
the aqueous phase was discarded. The octanol phase was pipetted
to four test tubes (2 mL to each) and equal amount of phosphate
buffer (0.02 M, pH 7.4) was added to each tube. The mixtures were
shaken for 10 min and centrifuged for 5 min at 1000 g. Samples
from the aqueous (0.5 mL) and octanol (0.5 mL) phases were taken
4.3. Radiochemistry
Radioiodine labeling solution [123I]NaI (in 0.1 M NaOH) was a
product of MAP Medical Technologies Oy (Helsinki, Finland). The
specific activity of the labeling solution was 1.3*10^ꢃ35 TBq/mg and
the radionuclidic purity >99.8% (123I). Human plasma was obtained
from the Finnish Red Cross Blood Service (Helsinki, Finland). Syn-
thesis, purification and formulation were conducted in lead shiel-
ded hot cells. The separation and purification of the tracer was
performed on the HPLC system B using method B1. The analyses of
the radioactive purity and specific activity were performed on the
HPLC system A using methods A1 and A2, respectively. The analyses
of the radioactive metabolites were performed on HPLC system A
using method A3. For the gamma-activity counting of the biological
samples Wizard 300 1480 gammacounter (PerkinElmer, Waltham,
MA, USA) was used. Digital autoradiography of tissue sections was
based on photostimulated luminescence detection on a Fujifilm
FLA-5100 system and Image Reader FLA-5000 series v1.0 software
(Fujifilm, Tokyo, Japan). The autoradiography data was processed
using AIDA Image Analyzer v.4.00 program (Raytest Isotopen
messgeräte GmbH, Straubenhardt, Germany). SPECT-CT imaging
was carried out with NanoSPECT/CT (Bioscan Inc., Washington, DC,
USA).
from each test tube for the activity measurement with g-counter
(PerkinElmer Wizard, 120 s). The partition coefficient was deter-
mined by calculating the logarithm of the ratio (cpm/mL in octa-
nol)/(cpm/mL in buffer) from the four replicates, and the data was
expressed as mean value ꢀ standard deviation.
4.3.4. In vitro stability in human plasma
Plasma stability of [123I]2f was determined by incubation in
human plasma at 37 ꢁC for up to 120 min. The tracer (5 MBq) was
formulated in PBS (2.0 mL) and this solution was divided equally
into 5 eppendorf tubes each containing human plasma (400
After incubation (15, 30, 60, 90 and 120 min) at 37 ꢁC, the proteins
were precipitated with MeCN (600 L) and the mixture was
mL).
m
centrifuged for 5 min at 10 000 g. The supernatant was collected
carefully and the activities of 123I in pellet and supernatant were
measured. The supernatant was analyzed with HPLC system A us-
ing gradient method A3 (tr parent ¼ 7.68 min).
4.3.1. 4-(4-[123I]iodophenyl)butanoyl-
L
-prolyl-pyrrolidine ([123I]2e)
4.3.5. Plasma metabolites in mouse
The small-scale radiosynthesis of [123I]2e (starting activity
10 MBq) was performed starting from the precursor 5a following
the method described below for the tracer [123I]2f that was selected
for in vivo studies. The HPLC system B, method B1 (MeCN/H2O, 33%;
3.6 MBq of [123I]2f was administered to C57Bl/6JRccHsd male
mice via the lateral tail vein. The animals were sacrificed at 5, 20,
40, 80 and 160 min post injection. Blood samples were immediately
obtained via cardiac puncture with heparin wetted syringe and
placed on ice. Similarly, blood samples were taken from the mice
4.5 mL/min;
mBondapak C-18 column, 7.8 ꢅ 300 mm;
l
¼ 220 nm,
Waters) was used for the purification of the product [123I]2e
(tr ¼ 17.5 min). The decay corrected, isolated radiochemical yield
and the radiochemical purity of [123I]2e were >99% and >99%,
respectively.
pretreated with 2d (KYP-2047, 50 mmol/kg, 10 mL/kg, DMSO (1%) in
NaCl (0.9%)). The sample was weighed and a small aliquot was
taken for the gamma-activity measurement (PerkinElmer Wizard
g-counter, 120 s). The rest of the blood was centrifuged for 5 min at
2000 g to separate plasma and the blood cells. The weight of the
plasma was measured, and a small aliquot of the plasma was taken
for the activity measurements. Acetonitrile (1.5 ꢅ the volume of
plasma) was added to the residual plasma and the mixture was
centrifuged for 5 min at 2000 g. Again, the supernatant was
4.3.2. 4-(4-[123I]iodophenyl)butanoyl-
L-prolyl-2(S)-
cyanopyrrolidine ([123I]2f)
To a glass vial, 0.1 mg (0.2
small amount of ethanol (0.1 mL). HCl (90
followed by the addition of [123I]NaI-labeling solution (90
0.05 M NaOH, 300e680 MBq). After this, peracetic acid solution
was added to the vial (8 L, 3.9% in water) and the reaction solution
was mixed gently and let to react for 10 min at room temperature.
Sodium metabisulphite (160 L, 2.85 mg/mL) was added to quench
m
mol) of precursor 5b was added in a
L, 0.05 M) was added,
L,
m
m
weighed and an aliquot of the supernatant was taken for g-mea-
surement. On the basis of these results, the binding of the tracer to
plasma proteins and blood cells was calculated. The rest of the
supernatant was analyzed by HPLC system A, gradient method A3.
Fractions of 0.65 min were collected and the activities were
m
m
the remaining unreacted oxidant. The reaction solution was diluted
with water (2 mL) and the resulting mixture was purified by using
the HPLC system B, method B1 (MeCN/H2O, 33%; 4.5 mL/min;
measured with a g-counter.
4.4. Animal experiments
mBondapak C-18 column, 7.8 ꢅ 300 mm;
l
¼ 220 nm, Waters). The
product fraction (tr ¼ 21 min) was collected, diluted with water (up
to 20 mL), and the resulting solution was passed through a C-18 SPE
cartridge (Sep-PakÒ Plus C18, Waters). The product was eluted out
with ethanol (1.5 mL) and evaporated to dryness under reduced
pressure. The resulting product was re-dissolved in a solution of
DMSO (1%) in 0.9% NaCl (v/v). The product [123I]2f was obtained
with radiochemical incorporation yield of 87 ꢀ 4% in 133 ꢀ 33 min.
4.4.1. Biodistribution
C57Bl/6JRccHsd male mice (aged 8e12 weeks, 18e29 g) were
anesthetized with isoflurane (IsoFloÒ vet, Orion Pharma, Espoo,
Finland) in medical oxygen carrier (4% for induction at 4 L/min, 2%
for upkeep at 2e3 L/min) and the lateral tail vein was cannulated
using a temporary catheter fabricated out of a 30G needle and a
PE10 polyethylene tubing (AgnTho’s, Lidingö, Sweden). [123I]2f

444
A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
(dose 2.86 ꢀ 1.26 MBq/mouse) in 1% DMSO in 0.9% NaCl was
injected as a bolus (an average of 65 L of compound flushed with
60 L of saline) via tail vein. For the determination of non-specific
binding of the tracer, KYP-2047 (50 mol/kg, 10 mL/kg, in DMSO
whole brain region, right side, and left side by using 3D ROI tool.
The VOIs were saved and used for all subsequent analyses. The
radioactivity in each VOI was calculated and results are presented
as a percentage from injected dose.
m
m
m
(1%)/0.9% NaCl) was injected i.p. 10 min prior to tracer adminis-
tration. Animals were sacrificed at 5, 20, 40, 80 and 160 min after
administration.
Acknowledgments
Academy of Finland (Grant Nos. 136805 and 140965), the
Finnish funding agency for technology and innovation (TEKES grant
No. 40044/11) and FinPharma Doctoral Program (drug discovery
section) are acknowledged for the financial support.
4.4.2. Animal models
LPS induced acute neuroinflammation model. C57BL/6JRccHsd
male mice were deeply anaesthetized with isoflurane (2e4%) and
placed in a stereotaxic frame (Stoelting, Wood Dale, IL, USA). Each
mouse received a stereotaxic microinjection of a 2
containing 5 g LPS (Escherichia coli serotype 055:B5, Sigmae
Aldrich, St. Louis, MO, USA) into the left striatum 24 h prior to
administration of [123I]2f. The injections were done using a 10
Hamilton-microsyringe connected to an injector (Stoelting) at
0.5
L/min to the coordinates A/P þ0.7; M/L þ1.8; D/V ꢃ2.7. After
mL solution
References
m
[1] A.J. Barret, N.D. Rawlings, Families and clans in serine peptidases, Archives of
Biochemistry and Biophysics 318 (1995) 247e250.
mL
[2] N.D. Rawlings, L. Polgar, A.J. Barret, A new family of serine-type peptidases
related to prolyl oligopeptidase, Biochemical Journal 279 (1991) 907e911.
[3] J. Tenorio-Laranga, J.I. Venäläinen, P.T. Männistö, J.A. García-Horsman, Char-
acterization of membrane-bound prolyl endopeptidase from brain, FEBS
Journal 275 (2008) 4415e4427.
[4] J.A. García-Horsman, P.T. Männistö, J.I. Venäläinen, On the role of prolyl oli-
gopeptidase in health and disease, Neuropeptides 41 (2007) 1e24.
[5] A.J. Jalkanen, K. Savolainen, M.M. Forsberg, Inhibition of prolyl oligopeptidase
by KYP-2047 fails to increase the extracellular neurotensin and substance P
levels in rat striatum, Neuroscience Letters 502 (2011) 107e111.
[6] R.S.B. Williams, M. Eames, W.J. Ryves, J. Viggars, A.J. Harwood, Loss of prolyl
oligopeptidase confers resistance to lithium by elevation of inositol (1,4,5)
trisphosphate, EMBO Journal 18 (1999) 2734e2745.
[7] I. Brandt, M. Gérard, K. Sergeant, B. Devreese, V. Baekelandt, K. Augustyns,
S. Scharpé, Y. Engelborghs, A.-M. Lambeir, Prolyl oligopeptidase stimulates the
m
the injection, the needle was kept in place for an additional 4 min
and then slowly retracted to prevent reflux. Mice received bupre-
norphine (0.1 mg/kg, s.c., TemgesicÒ, Reckitt Benckiser Healthcare
(UK) Ltd, East Yorkshire, Great Britain) for post-operative pain re-
lief, and were single-housed after the operation. The right hemi-
sphere of the mice were left intact and served as a control for LPS-
induced inflammation.
Turpentine induced model of acute localized inflammation. C57BL/
6JRccHsd male mice were anaesthetized with isoflurane in air
carrier (4% for induction, 2% for upkeep). The injection area was
carefully shaved with a scalpel blade and disinfected with iodine
aggregation of
[8] A.-M. Lambeir, Interaction of prolyl oligopeptidase with
Neurological Disorders e Drug Targets 10 (2011) 349e354.
[9] T.T. Myöhänen, M.J. Hannula, R. Van Elzen, M. Gerard, P. Van Der Veken,
J.A. García-Horsman, V. Baekelandt, P.T. Männistö, A.-M. Lambeir, A prolyl
a
-synuclein, Peptides 29 (2008) 1472e1478.
a
-synuclein, CNS &
solution. Turpentine oil (5
mL) was injected i.m. into m. gastrocne-
mius in the right hind leg with a 30G needle. Saline (0.9%, 5
mL) was
injected i.m. into the same area in the left hind leg as a control. The
pain due to operation was alleviated with TemgesicÒ (0.1 mg/kg,
0.015 mg/mL in 0.9% NaCl, s.c.), administered in every 12 h until the
SPECT/CT scan.
oligopeptidase inhibitor, KYP-2047, reduces
aggregates in cellular and animal models of parkinson’s disease, British
Journal of Pharmacology 166 (2012) 1097e1113.
a-synuclein protein levels and
[10] M.J. Hannula, T.T. Myöhänen, J. Tenorio-Laranga, P.T. Männistö, J.A. García-
Horsman, Prolyl oligopeptidase colocalizes with a-synuclein, b-amyloid, tau
protein and astroglia in the post-mortem brain samples with Parkinson’s and
Alzheimer’s diseases, Neuroscience 242 (2013) 140e150.
[11] A. Penttinen, J. Tenorio-Langa, A. Siikanen, M. Morawski, S. Roßner, J.A. García-
4.4.3. Small animal SPECT scanning
C57Bl/6JRccHsd male mice (aged 8e12 weeks, 18e29 g) were
anesthetized with isoflurane (4% for induction, 2% for upkeep) in
either air (injection) or 0.8 L/min medical oxygen (imaging) as the
Horsman, Prolyl oligopeptidase:
a rising star on the stage of neuro-
inflammation research, CNS & Neurological Disorders e Drug Targets 10
(2011) 340e348.
[12] P. O’Reilly, M.T. Hardison, P.L. Jackson, X. Xu, R.J. Snelgrove, A. Gaggar,
F.S. Galin, J.E. Blalock, Neutrophils contain prolyl endopeptidase and generate
the chemotactic peptide, PGP, from collagen, Journal of Neuroimmunology
217 (2009) 51e54.
[13] J. Tenorio-Laranga, I. Peltonen, S. Keskitalo, G. Duran-Torres, R. Natarajan,
P.T. Männistö, A. Nurmi, N. Vartiainen, L. Airas, I. Elovaara, J.A. García-Hors-
man, Alteration of prolyl oligopeptidase and activated a-2-macroglobulin in
multiple sclerosis subtypes and in the clinically isolated syndrome,
Biochemical Pharmacology 85 (2013) 1783e1794.
carrier gas. A dose of 47.6 ꢀ 4.1 MBq of [123I]2f in 100
mL was
administered intravenously by cannula 30G temporary catheter,
after which the cannula was flushed with 100 mL of 0.9% NaCl so-
lution. Radioactivity in the syringe was measured prior to and post-
injection with dose calibrator (CRC-25R, Capintec Inc., Ramsey, NJ,
USA). For the blocking studies, mice received i.p. injection of 2d
(KYP-2047) 10 min prior to tracer administration as above.
SPECT-CT imaging was carried out with NanoSPECT/CT (Bioscan
Inc., USA), a four-headed small animal scanner featuring 1.0 mm
multipinhole mouse apertures. The body temperature of mice was
maintained warm throughout the study by using a temperature-
controlled animal bed (Minerve, France). Brain SPECT images
were acquired 20 min, 40 min and 60 min post-injection in 24
projections using time per projection of 200 s resulting in acqui-
sition time of 20 min/scan. Alternatively, mice were imaged only
once starting 20 min post-injection by using time per projection of
600 s and total acquisition time of 60 min. CT imaging was per-
formed with 45 kVp tube voltage in 180 projections. SPECT images
were reconstructed with HiSPECT NG software (Scivis GmbH,
Göttingen, Germany) and fused with CT datasets with InVivoScope
software (Bioscan Inc., USA).
[14] J. Tenorio-Laranga, F. Coret-Ferrer, B. Casanova-Estruch, M. Burgal, J.A. García-
Horsman, Prolyl oligopeptidase is inhibited in relapsing-remitting multiple
sclerosis, Journal of Neuroinflammation 7 (2010) 23.
[15] J.I. Venäläinen, E.A.A. Wallén, A. Poso, J.A. García-Horsman, P.T. Männistö,
Synthesis and characterization of the novel fluorescent prolyl oligopeptidase
inhibitor 4-fluoresceinthiocarbamoyl-6-aminocaproyl-L-prolyl-2(S)-(hydrox-
yacetyl)pyrrolidine, Journal of Medicinal Chemistry 48 (2005) 7093e7095.
[16] M.-J. Moreno-Baylach, V. Felipo, P.T. Männistö, J.A. García-Horsman, Expres-
sion and traffic of cellular prolyl oligopeptidase are regulated during cere-
bellar granule cell differentiation, maturation, and aging, Neuroscience 156
(2008) 580e585.
[17] T. Nakajima, Y. Ono, A. Kato, J. Maeda, T. Ohe, Y-29794 e a nonpeptide prolyl
endopeptidase inhibitor that can penetrate into the brain, Neuroscience Let-
ters 141 (1992) 156e160.
[18] J.R. Atack, N. Suman-Chauhan, G. Dawson, J.J. Kulagowski, In vitro and in vivo
inhibition of prolyl endopeptidase, European Journal of Pharmacology 205
(1991) 157e163.
[19] A. Charalambous, T.J. Mangner, M.R. Kilbourn, Synthesis of (N-[11C]methyl) Y-
29794, a competitive inhibitor of prolyl endopeptidase, Journal of Labelled
Compounds and Radiopharmaceuticals 34 (1994) 499e503.
To quantify tracer uptake in brain areas, reconstructed SPECT
images were reoriented and analyzed with InVivoScope (Bioscan
Inc., USA) software by using CT data as a reference. Three-
dimensional volumes of interest (VOIs) were defined to cover
[20] M.E. Van Dort, M.R. Kilbourn, T.J. Mangner, Synthesis of N-(N-(-4-(N-[11C]
methylamino)phenyl)butyryl)-L-prolyl)pyrrolidine: a potential radiotracer for
prolyl endopeptidase, Journal of Labelled Compounds and Radiopharmaceu-
ticals 34 (1994) 447e452.

A. Kallinen et al. / European Journal of Medicinal Chemistry 79 (2014) 436e445
445
[21] M. Saito, M. Hashimoto, N. Kawaguchi, H. Shibata, H. Fukami, T. Tanaka,
N. Higuchi, Synthesis and inhibitory activity of acyl-peptidyl-pyrrolidine de-
rivatives toward post-proline cleaving enzyme; a study of subsite specificity,
Journal of Enzyme Inhibition 5 (1991) 51e75.
[22] E.M. Jarho, J.I. Venäläinen, J. Huuskonen, J.A. Christiaans, J.A. García-Horsman,
M.M. Forsberg, T. Järvinen, J. Gynther, P.T. Männistö, E.A.A. Wallén, Cyclopent-
2-enecarbonyl group mimics proline at the P2 position of prolyl oligopepti-
dase inhibitors, Journal of Medicinal Chemistry 47 (2004) 5605e5607.
[23] J.I. Venäläinen, J.A. García-Horsman, M.M. Forsberg, A. Jalkanen, E.A.A. Wallén,
E.M. Jarho, J.A.M. Christiaans, J. Gynther, P.T. Männistö, Binding kinetics and
duration of in vivo action of novel prolyl oligopeptidase inhibitors,
Biochemical Pharmacology 71 (2006) 683e692.
[24] K. Kaszuba, T. Róg, R. Danne, P. Canning, V. Fülöp, T. Juhász, Z. Szeltner, J.-F. St.
Pierre, A. García-Horsman, P.T. Männistö, M. Karttunen, J. Hokkanen,
A. Bunker, Molecular dynamics, crystallography and mutagenesis studies on
the substrate gating mechanism of prolyl oligopeptidase, Biochimie 94 (2008)
1398e1411.
[25] H. Arai, H. Nishioka, S. Niwa, T. Yamanaka, Y. Tanaka, K. Yoshinaga,
N. Kobayashi, N. Miura, Y. Ikeda, Synthesis of prolyl endopeptidase inhibitors
and evaluation of their structure-activity relationships: in vitro inhibition of
prolyl endopeptidase from canine brain, Chemical & Pharmaceutical Bulletin
41 (1993) 1583e1588.
[26] H.H. Coenen, J. Mertens, B. Mazière, Radioiodination Reactions for Pharma-
ceuticals, in: Compendium for Effective Synthesis Strategies, Springer, Dor-
drecht, The Netherlands, 2006.
[27] H. Pajouhesh, G.R. Lenz, Medicinal chemical properties of successful central
nervous system drugs, NeuroRx 2 (2005) 541e553.
[28] A.A. Wilson, L. Jin, A. Garcia, J.N. DaSilva, S. Houle, An admonition when
measuring the lipophilicity of radiotracers using counting techniques, Applied
Radiation and Isotopes 54 (2001) 203e208.
[29] S. Wilk, Prolyl endopeptidase, Life Sciences 33 (1983) 2149e2157.
[30] T.T. Myöhänen, J.I. Venäläinen, J.A. García-Horsman, M. Piltonen,
P.T. Männistö, Distribution of prolyl oligopeptidase in the mouse whole-body
sections and peripheral tissues, Histochemistry and Cell Biology 130 (2008)
993e1003.
[31] A.J. Jalkanen, T.P. Piepponen, J.J. Hakkarainen, I. De Meester, A.-M. Lambeir,
M. Forsberg, The effect of prolyl oligopeptidase inhibition on extracellular
acetylcholine and dopamine levels in the rat striatum, Neurochemistry In-
ternational 60 (2012) 301e309.
[32] H. He, P. Tran, H. Yin, H. Smith, Y. Batard, L. Wang, H. Einolf, H. Gu,
J.B. Mangold, V. Fischer, D. Howard, Adsorption, metabolism, and excretion of
[
14C]vildagliptin, a novel dipeptidyl peptidase 4 inhibitor, in humans, Drug
Metabolism and Disposition 37 (2009) 536e544.
[33] V.W. Pike, PET radiotracers: crossing the blood-brain barrier and surviving
metabolism, Trends in Pharmacological Sciences 30 (2009) 431e440.
[34] F. Goossens, I. De Meester, G. Vanhoof, S. Scharpé, Distribution of prolyl oli-
gopeptidase in human peripheral tissues and body fluids, European Journal of
Clinical Chemistry and Clinical Biochemistry 34 (1996) 17e22.
[35] C. Odaka, T. Mizuochi, T. Shirasawa, P. Morain, F. Checler, Murine T cells
expressing high activity of prolyl endopeptidase are susceptible to activation-
induced cell death, FEBS Letters 512 (2002) 163e167.
[36] R.L. Hunter, B. Cheng, D.-Y. Choi, M. Liu, S. Liu, W.A. Cass, G. Bing, Intrastriatal
lipopolysaccharide injection induced Parkinsonism in C57/B6 mice, Journal of
Neuroscience Research 87 (2009) 1913e1921.
[37] K.U. Tufekci, S. Genc, K. Genc, The endotoxin-induced neuroinflammation
model of Parkinson’s disease, Parkinsons Disease (2011) 487450.
[38] S. Aid, A.C. Silva, E. Candelario-Jalil, S.-H. Choi, G.A. Rosenberg, F. Bosetti,
Cyclooxygenase-1 and -2 differentially modulate lipopolysaccharide-induced
blood-brain barrier disruption through matrix metalloproteinase activity,
Journal of Cerebral Blood Flow and Metabolism 30 (2010) 370e380.
[39] E.A.A. Wallén, J.A.M. Christiaans, M.M. Forsberg, J.I. Venäläinen, P.T. Männistö,
J. Gynther, Dicarboxylic acid bis(L-prolyl-pyrrolidine) amides as prolyl oligo-
peptidase inhibitors, Journal of Medicinal Chemistry 45 (2002) 4581e4584.
Supporting Info.