Radiofluorinated pyrimidine-2,4,6-triones as molecular probes for noninvasive MMP-targeted imaging

Article abstract of DOI:10.1002/cmdc.201000013

Matrix metalloproteinases (MMPs) are zinc- and calcium-dependent endopeptidases. Representing a subfamily of the metzincin superfamily, MMPs are involved in the proteolytic degradation of components of the extracellular matrix. Unregulated MMP expression,

Full text of DOI:10.1002/cmdc.201000013

DOI: 10.1002/cmdc.201000013
Radiofluorinated Pyrimidine-2,4,6-triones as Molecular
Probes for Noninvasive MMP-Targeted Imaging
Hans-Jçrg Breyholz,*[a] Stefan Wagner,[a] Andreas Faust,[a] Burkhard Riemann,[a]
Carsten Hçltke,[b] Sven Hermann,[c] Otmar Schober,[a] Michael Schꢀfers,[c, d] and
Klaus Kopka[a, d]
Matrix metalloproteinases (MMPs) are zinc- and calcium-depen-
dent endopeptidases. Representing a subfamily of the metzin-
cin superfamily, MMPs are involved in the proteolytic degrada-
tion of components of the extracellular matrix. Unregulated
MMP expression, MMP dysregulation and locally increased
MMP activity are common features of various diseases, such as
cancer, atherosclerosis, stroke, arthritis, and others. Therefore,
activated MMPs are suitable biological targets for the specific
visualization of such pathologies, in particular by using radiola-
beled MMP inhibitors (MMPIs). The aim of this work was to de-
velop a radiofluorinated molecular probe for noninvasive
in vivo imaging for the detection of up-regulated levels of acti-
vated MMPs in the living organism. Fluorinated MMPIs (26, 31
and 38) based on the pyrimidine-2,4,6-trione lead structure
RO 28-2653 (1) were synthesized, and their MMP inhibition po-
tency was evaluated in vitro. The radiosynthesis and the in vivo
biodistribution of the first 18F-labeled prototype, MMP-targeted
tracer [18F]26, suitable for molecular imaging by means of posi-
tron emission tomography (PET) were realized.
Introduction
Molecular imaging of locally up-regulated and activated extra-
cellular matrix (ECM)-degrading peptidases in vivo, such as
matrix metalloproteinases (MMPs), is a clinical challenge. MMPs
have been implicated in the proteolytic degradation of ECM
components, but are also capable of processing and initiating
the release of bioactive molecules, such as growth factors, pro-
teinase inhibitors, cytokines and chemokines.[1] MMPs are sub-
divided into collagenases, gelatinases, stromelysins and matri-
lysins, according to common structural elements and substrate
specificities. Another subclass of the MMPs is represented by
the membrane-type MMPs (MT-MMPs).[2,3] In pathophysiologi-
cal processes, an increased cytokine and growth-factor-stimu-
lated MMP-gene transcription, followed by an abnormal pro-
MMP secretion and zymogen activation, might become preva-
lent, triggering the progression of inflammation, cancer or car-
diovascular diseases, such as atherosclerosis.[1,4–10]
activity in vascular remodeling.[13,14] The hydroxamic acid-based
broad-spectrum MMP inhibitor 111In-RP782 was applied in com-
bination with microSPECT/CT in an ApoEꢀ/ꢀ mouse model of
carotid artery injury. An increased and specific tracer uptake
was observed in the arterial wall of the guidewire-injured left
common carotid artery. MicroSPECT imaging and CT angiogra-
phy precisely located the MMP-rich lesion within the artery.[15]
Because of the inferior selectivity profile of hydroxamates, a
new trend can be recognized in contemporary MMPI design in
proceeding from inhibitors with a broad inhibition spectrum
to subtype-selective hydroxamate and non-hydroxamate repre-
sentatives.[16–19] Further shortcomings experienced with typical
hydroxamates, such as metabolic instability and high transi-
tion-metal binding potential—and thus possible interactions
with the metals of other metalloproteases—might additionally
be avoided. Pyrimidine-2,4,6-triones (barbiturates) based on
In 2004, our group succeeded in the development of an 123I-
labeled nonselective high-affinity MMP-targeted radiotracer
based on the hydroxamic acid-based MMP inhibitor CGS
27023A.[11,12] The hydroxamate moiety of CGS27023A functions
as zinc-binding group and chelates the zinc ion of the MMP
active site. Using the corresponding nonpeptidic radioligand
and planar scintigraphy, local MMP activity was specifically
imaged in apolipoprotein E-deficient (ApoEꢀ/ꢀ) mice that devel-
op macrophage- and MMP-rich vascular lesions after carotid
artery ligation and cholesterol-rich diet. For the first time, this
study highlighted the feasibility of imaging activated MMPs
in vivo and emphasized the value of radiolabeled MMPIs for
noninvasive scintigraphic in vivo visualization of MMP-related
diseases. This was also verified by a recent study, which suc-
ceeded in the visualization of endothelial injury-induced MMP
[a] Dr. H.-J. Breyholz, Dr. S. Wagner, Dr. A. Faust, Prof. Dr. B. Riemann,
Prof. Dr. O. Schober, Priv.-Doz. Dr. K. Kopka
Department of Nuclear Medicine, University of Mꢀnster
Albert-Schweitzer-Str. 33, 48149 Mꢀnster (Germany)
Fax: (+49)251-83-47363
E-mail: breyholz@uni-muenster.de
[b] Dr. C. Hçltke
Department of Clinical Radiology, University of Mꢀnster
Albert-Schweitzer-Str. 33, 48149 Mꢀnster (Germany)
[c] Dr. S. Hermann, Prof. Dr. M. Schꢁfers
European Institute for Molecular Imaging (EIMI), University of Mꢀnster
Mendelstr. 11, 48149 Mꢀnster (Germany)
[d] Prof. Dr. M. Schꢁfers, Priv.-Doz. Dr. K. Kopka
Interdisciplinary Center of Clinical Research (IZKF) Mꢀnster (Germany)
Supporting information for this article is available on the WWW under
http://dx.doi.org/10.1002/cmdc.201000013.
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
777

H.-J. Breyholz et al.
MED
the lead structure of RO 28-2653 (1; Figure 1) are prominent
examples of such non-hydroxamate MMPIs, exhibiting specific
activities for subgroups of secreted MMPs, including the gelati-
nases A (MMP-2) and B (MMP-9), neutrophil collagenase (MMP-
8), as well as the membrane-bound MMPs: MT-1-MMP (MMP-
14) and MT-3-MMP (MMP-16).[20–23]
Results
Synthesis of the piperazines
Piperazines 6, 9–11, 14, 18 and 21, required for the synthesis
of pyrimidine-2,4,6-triones 25–32, were prepared as described
in Schemes 1–5.
N-(2-Fluoroethyl)piperazine (6) was prepared as the trifluoro-
acetic acid (TFA) salt by the reaction of N-Boc piperazine (4)[26]
with 2-fluoroethyl tosylate (5). Removal of the Boc group was
achieved with TFA to give compound
(Scheme 1).
6 in 46% yield
Figure 1. Structures of RO 28-2653 (1) and preclinical research tracer 2.[25]
Scheme 1. Synthesis of piperazine derivative 6. Reagents and conditions:
a) 1. Cs2CO3, DMF, RT, 2 d, 33%; 2. 50% TFA in CH2Cl2, RT, o/n, 46%.
The selective targeting of MMP-2, MMP-9 and MMP-14 may
help to visualize tumor angiogenesis, since these three en-
zymes are most consistently detected in malignant tissues and
are associated with tumor aggressiveness, metastatic potential,
and poor prognosis.[6,24] Furthermore, RO 28-2653 was found
to exhibit strong antitumor and antiangiogenic efficacy and
thus is able to target MMP-rich tumor tissues.[23] These prereq-
uisites support the approach of using barbiturates as molecu-
lar imaging agents.
Bromo derivative 9 was prepared from tetraethylene glycol
(7) (Scheme 2), which was first converted to the ditosylate and
then heated in the presence of lithium bromide to give dibro-
mide 8. Coupling with piperazine 4 (similar to the synthesis of
6) was accomplished in 40% yield. Cleavage of the Boc group
gave 9 in 47% yield.
Therefore, in recent years we started exploiting derivatives
of RO 28-2653 as a new class of potential selective MMPI radio-
tracers, resulting in compound 2 (Figure 1), a preclinical re-
search tracer[25] of the first generation, radio-iodinated by
[
125I]iodine. In more recent studies,[26,27] the amino precursor 29
(shown later in Scheme 6) was conjugated to the near-infrared
fluorescent (NIRF) dye Cy 5.5-NHS ester via a short polyethy-
lene glycol (’mini-PEG’) linker to give a barbiturate-based NIRF
imaging agent, the first of its kind to have been developed so
far. MMP-positive tumor xenografts could be successfully vi-
sualized in tumor-bearing mice 1–2 h post injection (p.i.) by
optical techniques, using this novel imaging agent.
Scheme 2. Synthesis of piperazine derivative 9. Reagents and conditions:
a) 1. TsCl, Et3N, CH3CN, RT, 20 h, 85%; 2. LiBr, acetone, reflux, o/n, 69%;
b) 1. 4, Cs2CO3, DMF, RT, 2 d, 40%; 2. 50% TFA in CH2Cl2, RT, o/n, 47%.
Azide 10 and amino derivative 11 (shown later in Scheme 6)
were prepared according to our previously described proce-
dures.[26] Hydroxy derivative 14 was synthesized by reaction of
7 with one equivalent triphenylmethanol (Scheme 3). Subse-
Having successfully proven the feasibility of the chosen ap-
proach, herein, we outline the progress of our continuing work
towards the development of an MMP imaging agent, based on
RO 28-2653, designed for preclinical and clinical imaging.
Three exemplary mini-PEG-spaced, fluoro-PEGylated com-
pounds were synthesized as non-radioactive analogues of PET-
compatible MMPI radiotracers 26 (n=1), 31 (n=4) and 38 (n=
8), where n is the number of ethylene glycol (EG) units in the
spacer). All of these lipophilicity-reduced compounds retained
good to excellent MMP inhibitory activity. The radiosynthesis
and the in vivo biodistribution of [18F]26 are reported.
Scheme 3. Synthesis of piperazine derivative 14. Reagents and conditions:
a) 1. Ph3COH, TsOH·H2O, toluene, reflux, 2 h; 2. MsCl, Et3N, CH2Cl2, 08C!RT,
o/n, 97%;[28] b) piperazine (3), Cs2CO3, DMF, 1008C, 2 d, 86%; c) HCl, MeOH,
RT, 16 h, 56%.
778
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
quent mesylation yielded 12,[28] which was coupled to pipera-
zine (3) in DMF to give 13. Finally the trityl protecting group
of 13 was removed to yield the target compound 14.
Fluoro derivative 18 was prepared as shown in Scheme 4.
Triethylene glycol (15) was converted into the corresponding
tritylether and subsequently reacted with 2-fluoroethyl tosylate
18 or 21, respectively (Scheme 6). To assemble the barbiturate
scaffold, a palladium-catalyzed (Pd(OAc)2) coupling reaction of
diisopropyl malonate and 4-bromophenyl phenyl ether (22)
was used.[29] Condensation of malonic ester 23 with urea, and
a one-pot reaction involving the N-bromosuccinimide (NBS)-
mediated bromination at the 5-position of the pyrimidine-
Scheme 4. Synthesis of piperazine derivative 18. Reagents and conditions:
a) 1. Ph3COH, TsOH·H2O, toluene, reflux, 2 h; 2. 2-fluoroethyl tosylate (5),
tBuOK, tBuOH, reflux, 24 h, 54%; b) 1. TsOH·H2O, MeOH, reflux, 2 h; 2. MsCl,
Et3N, CH2Cl2, 08C!RT, o/n, 52%; c) Piperazine (3), Cs2CO3, DMF, 1008C, 2 d,
94%.
(5) to give the fluoro compound 16. After cleavage of the trityl
protecting group and mesylation, intermediate 17 was ob-
tained and coupled to piperazine (3) to give the target com-
pound 18.
The synthesis of alkyne derivative 21 began with the forma-
tion of intermediate 12 (Scheme 5). Conversion with propargyl
alcohol led to alkyne 19. Exchange of the trityl group for mesyl
was accomplished in 67% yield as described above. Coupling
to N-Boc piperazine (4) and deprotection was performed in
70% and 86% yield, respectively, to give the target compound
21.
Scheme 6. Syntheses of phenoxyphenyl pyrimidine-2,4,6-triones 25-32. Re-
agents and conditions: a) Pd(OAc)2, tBu3P, CuF2, tBuONa, diisopropyl malo-
nate, 678C, 15 h; b) urea, tBuOK, THF, 808C, 17 h, 84% (two steps),[29] this
work: 30–40%; c) piperazines 3, 6, 9–11, 14, 18 and 21 (1.0 equiv), NBS,
K2CO3, DMF, 5–108C, 2 h then RT, 20 h, 43% (25), 45% (26) 36% (27), 43%
(28),[26] 71% (29),[26] 55% (30), 32% (31), 78% (32).
2,4,6-trione and the introduction of the piperazine subunit,
yielded barbiturates 25–32 in three steps. Hutchings et al. ob-
tained intermediate 24 in 84% yield (over two steps).[29] The
average yield of our preparations ranged between 30 and 40%
with a maximum of 62%. Barbiturates 25–32 were obtained in
yields between 36% and 78%. The overall yields (including the
synthesis of the corresponding piperazine derivatives) ranged
between 0.5% and 17% for the three- to nine-step syntheses:
25, 17% (3 steps); 26, 1% (7 steps); 27, 9% (7 steps); 28, 0.8%
(8 steps);[26] 29, 0.6% (8 steps);[26] 30, 15% (7 steps); 31, 4% (8
steps); 32, 6% (9 steps).
Scheme 5. Synthesis of piperazine derivative 21. Reagents and conditions:
a) 1. Ph3COH, TsOH·H2O, RT, 2 h, toluene; 2. MsCl, Et3N, CH2Cl2, RT, o/n,
97%;[28] b) 1. HCꢁC-CH-OH, Cs2CO3, DMF, 1008C, 1.5 h, 58%; c) 1. TsOH·H2O,
MeOH, reflux, 2 h; 2. MsCl, Et3N, CH2Cl2, 08C!RT, o/n, 67%; d) 1. 4, Cs2CO3,
DMF, RT, 2 d, 70%; 2. 50% TFA in CH2Cl2, RT, o/n, 86%.
Synthesis of 5-(4-phenoxyphenyl) pyrimidine-2,4,6-triones
Synthesis and in vitro data of pyrimidine-2,4,6-triones
25–32
42–45
The 5-(4-phenoxyphenyl) pyrimidine-2,4,6-triones 25–32 were
prepared from the corresponding piperazines 3, 6, 9–11, 14,
The 3-bromo-4-methoxy-phenoxyphenyl barbituric acid deriva-
tives 42–45 were synthesized as depicted in Scheme 7. Inter-
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
779

H.-J. Breyholz et al.
MED
Scheme 8. Synthesis of the 1,2,3-triazoles 36–38. Reagents and conditions:
a) azides 33, 34 or 35 (1.0 equiv), CuSO4·5H2O (30 mol%), sodium ascorbate
(40 mol%), tBuOH/H2O (4:1), RT, o/n, 19% (36), 30% (37), 32% (38).
Scheme 7. Syntheses of compounds 41–45. Reagents and conditions: a) NBS
(1.2 equiv), dibenzoyl peroxide, CCl4, reflux, 4 h, 58%; b) MeOH, RT, o/n
(3.0 equiv Et3N added for 10), 30% (42), 54% (43), 84% (44); c) H2, Pd/C,
EtOAc, RT, 16 h, 54%.
identified by analytical HPLC and MS (see Experimental sec-
tion) turned out to be inseparable from the main products 36–
38 by silica gel column chromatography. Therefore, the 1,2,3-
triazoles were isolated by preparative HLPC. The synthesis of
the azides 33–35 is outlined in Scheme 9. Tetraethylene glycol
mediate 40[25] was intended to be brominated at the barbituric
acid 5-position by using NBS and dibenzoyl peroxide in tetra-
chloromethane. However, instead of one bromine—as expect-
ed and demonstrated in our previous work[25]—two bromine
atoms were introduced to give the dibromo product 41 in
58% yield. One bromine was incorporated at the aliphatic 5-
position and the other at the anisole ring, ortho to the me-
thoxy group, as demonstrated by two-dimensional NMR and
NOE experiments (see Supporting Information). Compound 41
was reacted with piperazines 14 and 18 (Schemes 3 and 4) in
methanol to generate the fluoro derivative 42 (yield: 30%) and
the and hydroxy derivative 43 (yield: 54%). In case of the TFA
salt 10, three equivalents of triethylamine were added to the
reaction mixture to give azide 44 in 84% yield. The azide was
subsequently reduced by catalytic hydrogenation to give
amine 45 in moderate conversion.
Scheme 9. Mini-PEG intermediates 33, 34 and 35, synthesized as azide com-
ponents for 1,3-dipolar cycloadditions leading to 1,2,3-triazoles 36–38. Re-
agents and conditions: a) MsCl or TsCl, Et3N, CH2Cl2, RT, 20 h, 85% (39),[26]
80% (40); b) NaN3, CH3CN, 1208C, 5 h, 36% (33),[26] 34% (34); c) TBAF (1m)
in THF, reflux (2 h)!RT (o/n), 78%.
Synthesis of the triazoles 36–38
Pyrimidine-2,4,6-trione 32 was used as the alkyne component
in a copper(I)-catalyzed, 1,3-dipolar Huisgen cycloaddition
using azides 33, 34 or 35, respectively. The click reaction was
realized in a solvent mixture of tBuOH and water containing
the catalytic system sodium ascorbate and copper sulfate pen-
tahydrate at room temperature (see Scheme 8). The 1,4-disub-
stituted 1,2,3-triazole click products 36–38 were formed in
yields ranging from 19% to 32% (see Experimental Section for
further details).
(7) was subjected either to twofold mesylation or tosylation,
resulting in the dimesylate 39 or ditosylate 40, respectively,
which were then converted to the mono azides 33 and 34.
Mesylate 33 was subsequently submitted to fluorination by
tetra-n-butylammonium fluoride to give the fluorinated azide
35.
Enzyme assays and logD values
The overall yields for this twelve- to thirteen-step procedures
ranged between 0.3% and 0.5% (0.3% (36), 0.5% (37), 0.4%
(38)). The by-product (30) of all click reactions (Scheme 6), as
The MMP inhibition potencies of the barbituric acid derivatives
25–32, 38 and 42–46 were measured in fluorogenic in vitro in-
780
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
hibition assays as described previously.[30] The resulting IC50
values for the investigated MMPs were calculated from a non-
linear regression fit of the concentration-dependent reaction
rates. The results are displayed in Table 1, which also contains
Table 1. Inhibition of various MMPs by the pyrimidine-2,4,6-triones 25–
32, 38, and 42–46 with corresponding logD values.
Compd
IC50 [nm][a]
clogD[b]
MMP-2
MMP-8
MMP-9
MMP-13
Scheme 10. Two-step radiosynthesis of [18F]fluoroethyl radioligand [18F]26 as
a potentially PET-compatible barbituric acid-derived MMP-targeted imaging
probe. Reagents and conditions: a) [18F]K(K222)F, K2CO3, CH3CN, 848C, 4 min;
b) DMSO, 1208C, 15 min.
25[c]
26[d]
27
28[f]
29
30
31
32
38
42
43
44
45
37ꢂ3
23ꢂ9
58ꢂ8
138ꢂ12
55ꢂ24
71ꢂ8
44ꢂ5
7ꢂ2
83ꢂ4
645ꢂ17
207ꢂ22
57ꢂ12
177ꢂ20
333ꢂ60
86ꢂ4
2.62
2.88[e]
2.54
1.84
ꢀ0.34
0.84
1.72
1.75
0.67
1.84
10ꢂ4
13ꢂ8
38ꢂ1
10ꢂ4
31ꢂ3
118ꢂ10
122ꢂ56
108ꢂ7
91ꢂ25
80ꢂ30
395ꢂ42
600ꢂ233
n.d.
85ꢂ14
6ꢂ2
chemical yield of this two-step radiosynthesis was 11.3ꢂ2.2%
(decay-corrected, based on cyclotron-derived [18F]fluoride ions).
The radiochemical purity was 98.5ꢂ1.5% after 142ꢂ12 min
from the end of radionuclide production. The determined spe-
cific radioactivities were 23ꢂ10 GBqmmolꢀ1 at the end of radi-
osynthesis (n=6).
70ꢂ14
62ꢂ5
93ꢂ6
32ꢂ15
81ꢂ6
52ꢂ4
77ꢂ7
75ꢂ5
237ꢂ32
2950ꢂ360
5480ꢂ970
n.d.
292ꢂ53
259ꢂ102
136ꢂ20
240ꢂ75
23ꢂ5
36ꢂ18
928ꢂ262
157ꢂ65
864ꢂ217
17ꢂ5
0.96
1.97
ꢀ0.22
2.99
The PET-compatible PEG4-tracer variant [18F]31 can be pre-
pared by direct radiofluorination of the tosylate precursor,
which could be generated either from the bromo derivative 27
(e.g., by reaction with silver tosylate) or the hydroxy derivative
30 (e.g., by reaction with tosyl chloride and pyridine). Ap-
proaches to realize this lipophilictiy-reduced tracer variant
(clogD (31)=1.72 versus clogD (26)=2.88) are currently un-
derway. Different strategies for the radiosynthesis of PET-com-
patible PEG8-tracer variant [18F]38 can be conceived. Starting
from precursors 28 or 32, the radiosynthesis of [18F]38 might
be performed either as two-step- or as one-pot reaction, i. e.
radiofluorination of labeling synthons, such as 33 or 34, giving
the 18F-labeled building block [18F]35. This intermediate could
then be used for subsequent 1,2,3-triazole formation in the
same reaction vessel without intermediate purification.[31] Alter-
natively, direct one-step radiofluorination of precursors already
containing the 1,2,3-triazole moiety, such as 36 or 37, might
also be feasible. Regarding the one-pot procedure, only traces
of [18F]38 could be isolated thus far, most likely because a in-
tramolecular cyclization product (see Discussion section and
scheme S1 in the Supporting Information) was formed. To cir-
cumvent this unintended side reaction, we are currently inves-
tigating the two-step process for the generation of [18F]38, i.e.,
via click reaction between alkyne 32 and azide [18F]35. Accord-
ing to calculated logD values, [18F]38 should be more hydro-
philic by more than two logD units (clogD (38)=0.67 versus
clogD (26)=2.88) compared to non-PEG8-PEGlyated derivative
26.
n.d.
n.d.
28ꢂ11
46
146ꢂ48
[a] Values are the mean ꢂSD of three experiments. [b] calculated logD
(clogD) values were calculated by ACD/Chemsketch version ACD/
Labs 6.00 (clogD=clogP at physiological pH (7.4) with consideration of
charged species). [c] IC50 (MMP-14)=37ꢂ3 nm [d] IC50 (MMP-1)>50 mm,
IC50 (MMP-3)=760 mm. [e] Experimental logD=2.15ꢂ0.02. [f] IC50 (MMP-
14)=120ꢂ18 nm. n.d.=not determined.
the calculated logD values (clogD) to indicate the changes in
lipophilicity caused by the structural modifications. Additional-
ly, the logD value of the key radioligand [18F]26 was deter-
mined experimentally. The calculated and experimental logD
values differ by 0.73 logD units (clogD (26)=2.88 compared
with experimental logD ([18F]26)=2.15).
All unlabeled analogues of tracer candidates 26, 31 and 38
retained good MMP inhibition potencies. A detailed reflection
on the IC50 values can be found in the discussion section. Be-
cause of its higher gelatinase selectivity, derivative 26 ap-
peared to be an ideal starting point for the radiosynthesis of
the corresponding fluoroethyl tracer [18F]26. Its radiosynthesis
is described in the following paragraph. Our current research
focus is on the realization of PEGylated tracer variants [18F]31
(PEG =4 EG units) and [18F]38 (PEG =8 EG units).
Radiosynthesis of fluorinated pyrimidne-2,4,6-trione
The radiosynthesis of the barbituric acid-based [18F]fluoroethyl
radioligand [18F]26 is shown in Scheme 10. This semi-automat-
ed two-step procedure starts with the synthesis of
1-[18F]fluoro-2-tosyloxyethane [18F]5 from ethylene glycol di-p-
tosylate and [18F]K(K222)F. [18F]5 was obtained in a radiochemical
yield of 53.3ꢂ2.3% (decay-corrected, n=6). The second step
was performed outside the automated synthesizer. [18F]5 was
coupled to the piperazine moiety of precursor 25 by heating
in dimethyl sulfoxide. [18F]26 was produced in a radiochemical
yield of 21.3ꢂ4.0% (decay-corrected, n=6). The overall radio-
Biodistribution
For biodistribution studies, adult C57/BL6 mice were intrave-
nously injected with the prototype MMP-targeted radiotracer
[18F]26. Shortly after completion of the [18F]26 scan (5 min), the
PET tracer 2-deoxy-2-[18F]fluoro-d-glucose ([18F]FDG) was addi-
tionally injected via the tail vein to outline the left ventricular
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
781

H.-J. Breyholz et al.
MED
myocardium for placement of an arterial region-of-interest
(ROI).
analogues (reference compounds) of prospective radiotracers
that may be useful for the in vivo imaging of locally up-regu-
lated and activated MMPs by means of SPECT or PET.[25] We
have suggested a radioiodinated variant 2 (Figure 1), labeled
with the radionuclide [125I]iodine, as a model compound for in
vitro and ex vivo applications.
Dynamic PET scans were performed throughout the studies
and reconstructed in to images at different time points after
injection of [18F]26 and [18F]FDG (Figure 2). The coronal whole-
Our ongoing research efforts
are now aimed at the design of
radiofluorinated
variants
of
model tracer 2 for preclinical
and clinical PET studies. In partic-
ular, we aim to increase the hy-
drophilicity and water solubility
of the commonly lipophilic bar-
biturates, which are known to
be highly metabolically stable, to
reduce plasma-protein binding.
Finally, efforts will be made to
direct the pharmacokinetics of
the modified radiolabeled deriv-
atives towards early and high
in vivo
target-to-background
ratios of the radiotracer-associat-
ed scintigraphic signal.
Figure 2. Biodistribution of radioactivity in an adult C57/BL6 mouse after intravenous injection of [18F]26 and
[18F]FDG visualized by small-animal PET. Panels A–C and E show maximum intensity projections. PET scans 5–
10 min (panel A), 30–60 min (panel B) and 90–120 min (panel C) after injection of [18F]26 are displayed. Panel D:
ROI overview. Panel E: [18F]FDG scan after completion of the [18F]26 scan. Labeled structures: liver (lv), bladder (bl),
kidneys (kd), gastrointestinal tract (git), brain (br), lung (lu), muscle (m), myocardium (mc).
Tracer modification strategy
PEGylation with polyethylene
glycol (PEG) units, consisting of
body small-animal PET images showed that the tracer uptake
in nontarget organs, such as the brain, lung, heart and muscle,
was low over all time points. The initial liver radioactivity 5–
10 min after injection of [18F]26 was high but rapidly decreased
up to 90–120 min p.i. There was also notable uptake in the
bladder, indicative of a predominantly renal excretion with
some hepatobilliary clearance. The low accumulation of [18F]26
in the skeleton suggests that the extent of tracer defluorina-
tion and generation of free [18F]fluoride is negligible. Figure 3
shows the averaged time–activity curves (TACs) for tissues in-
volved with metabolic and excretory processes (bladder,
kidney, liver and small/large bowel; Figure 3A) and for possible
target tissues (brain, lung and muscle; Figure 3B) where dysre-
gulated MMP activity might occur in MMP-associated diseases.
The data were determined from a multiple time-point PET
study and demonstrate the favorable pharmacokinetic behav-
ior of model tracer [18F]26, corresponding to the physical half-
life of the positron emitter 18F (~110 min).
several thousands of daltons, is known to significantly influ-
ence the pharmacokinetic behavior of a drug.[32] Another ap-
proach involves the conjugation of drugs with short PEG
chains, containing only a few PEG units (mini-PEGs), which in-
crease the hydrophilicity and the water solubility and can
direct the tracer towards the renal excretion route. This would
result in a higher signal-to-noise ratio of the tracer accumula-
tion in target tissues compared to the non-PEGylated deriva-
tives. This approach has been successfully shown in other radi-
opharmaceutical studies on peptides[33,34] and small-molecule
(nonpeptide) drugs.[35,36]
We have recently shown that by introducing the NIRF dye
Cy 5.5 into the mini-PEGylated precursor 29 (see Scheme 6 for
structure) a dramatic increase of water solubility, achieved here
by introducing negative charges into the molecule, can be a
crucial element in the design of barbiturate-based MMP-target-
ed tracers.[26,27] As MMPs act in the extracellular space, water-
soluble MMPIs should reach and interact with their biological
target more efficiently, show a reduced intracellular MMPI ac-
cumulation and consequently reduced interaction with metal-
loenzymes, which are abundant in cells.
Discussion
Consequently, we fixed the 5-(4-phenoxyphenyl) substituent
and initially introduced the mini-PEG spacer with four EG units
at the piperazine unit, which appeared to be an ideal PEGyla-
tion site since it functions as a P2’ substituent directed towards
the solvent-exposed S2’ pocket of the enzyme. This allows an
arbitrary elongation of the PEG chains of azide 28 and alkyne
32, which may be optionally elongated by attaching PEG moi-
Prerequisites
Based on the work of Grams et al.[20,21] and Foley et al.[22], who
introduced pyrimidine-2,4,6-triones, such as RO 28-2653
(Figure 1), as a new class of non-hydroxamate small-molecule
MMP inhibitors, we modified this lead structure to gain access
to a complete set of four pairs of precursors/non-radioactive
782
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
Figure 3. Representative biodistribution of radioactivity in an adult C57/BL6 mouse after intravenous injection of [18F]26. A) time–activity curves of the blood,
brain, lung and muscle. B) time–activity curves of the blood, bladder, kidney, liver and small/large bowel. ROI: region-of-interest. %ID: % injected dose.
eties with the desired number of EG units using click chemis-
try. Thus, the effect of the number of the EG units on the phar-
macokinetic behavior of the molecule can be explored.
Instead of the hitherto used standard eight-step barbiturate
synthesis,[20,25] a much more effective three-step synthesis, pro-
posed by Hutchings et al.[29] was employed for their prepara-
tion.
With a clogD value of 2.88, the fluorinated analogue 26,
which we define as PEG1-modified derivative, is already four-
to five-times more hydrophilic than the first generation iodo
radioligand 2 (clogD=3.53). The deviation of the calculated
from the experimental logD values for compound 26 was ap-
proximately 25%. The MMP inhibitory potency of compound 2
(MMP-2, IC50 =7 nm; MMP-9, IC50 =2 nm) was maintained in
the modified analogue 26, which is still a one- to two-digit
nanomolar MMP inhibitor of gelatinases (MMP-2, IC50 =23 nm;
MMP-9, IC50 =7 nm. Inhibition of MMP-8 and MMP-13 by com-
pound 26 was only observed in the upper nanomolar range.
The marked gelatinase selectivity becomes even more appar-
ent when the micromolar IC50 values of 26 for the inhibition of
MMP-1 and MMP-3 are considered (Table 1). Similar to com-
pound 2, derivative 26 exhibits a threefold MMP-9 selectivity
relative to MMP-2. As might not necessarily be expected, the
biodistribution study shows excretion of the PEG1-variant
[18F]26 through the kidneys with only moderate uptake in, and
fast clearance from, the liver (Figures 2 and 3).
PEG1- and PEG4-tracer variants
A series of pyrimidine-2,4,6-triones 25–32 and 42–46 was syn-
thesized via strategies given in Schemes 6 and 7. These com-
pounds were prepared for the following purposes: Piperazine
derivative 25 and fluoroethyl piperazine derivative 26 were
synthesized as a labeling precursor and the corresponding
non-radioactive analogue of the PET-compatible MMP-targeted
radiotracer candidate [18F]26, respectively (Scheme 10). A two-
step radiosynthesis consisting of the preparation off
1-[18F]fluoro-2-tosyloxyethane ([18F]5) and subsequent fluoro-
ethylation of piperazine precursor 25 successfully yielded
target compound [18F]26. In a conceivable one-step process
consisting of the direct nucleophilic radiofluorination of the
corresponding tosyl precursor (system: piperazine-CH2CH2-OTs),
an intramolecular cyclization reaction is likely and accounts for
the failure of the synthesis of a toslyate precursor for direct ra-
diofluorination.
The azido derivative 28 and the amino derivative 29 were
originally synthesized as intermediate and labeling precursor
for the aforementioned MMP-targeted optical imaging
probe.[26] Beyond this application, we intend to use azide 28 in
further click approaches, and amine 29 for coupling reactions
with radionuclide-containing prosthetic groups (such as N-suc-
cinimidyl-4-[18F]-fluorobenzoate, [18F]SFB). As depicted in
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
783

H.-J. Breyholz et al.
MED
Table 1, the introduction of the PEG4 chain is well tolerated in
28 and 29 in terms of retained MMP inhibitory potencies.
Compounds 28 (IC50 values: 10–71 nm) and 29 (IC50 values: 31–
177 nm) represent two- and two- to three-digit nanomolar
MMP inhibitors, respectively. Azido compound 28 displayed a
1.5 to twofold MMP-2 selectivity (versus MMP-8 and MMP-13)
and a six- to sevenfold MMP-9 selectivity (versus MMP-8 and
MMP-13). Amino derivative 29 displayed a four- to sixfold
MMP-2 selectivity (versus MMP-8 and MMP-13) and a 1.5- to
twofold MMP-9 selectivity (versus MMP-8 and MMP-13). Com-
pound 28 showed a threefold selectivity for MMP-9 relative to
MMP-2, similar to compound 26. In contrast, compound 29 ex-
hibited a threefold MMP-2 selectivity relative to MMP-9.
The PEG4-variants 27 (bromo derivative) and 30 (hydroxy de-
rivative) were prepared to serve as advanced intermediates
(one step each) to realize a tosyl precursor for radiofluorina-
tion, which might ultimately result in the PET-compatible PEG4-
tracer variant [18F]31. Both intermediates 27 and 30 also
showed notable gelatinase selectivity relative to MMP-8 and
MMP-13 (Table 1).
By using this click approach, the formal PEG8 derivatives 36
and 37 (mesyl-/tosyl-radiofluorination precursors), as well as
the unlabeled analogue 38 of the intended PET-compatible
MMP-targeted imaging probe, could easily be prepared
(Scheme 8 and 9). A further significant logD shift of one logD
unit (tenfold) was accomplished by the ’transition’ from deriva-
tive 31 to 38 (Table 1). This enhancement in hydrophilicity can
be explained by a synergistic effect arising from both, PEGyla-
tion and triazole formation (see above). The MMP inhibition
profiles of 31 and 38 are similar in that both fluoro-PEGylated
compounds retain acceptable MMP inhibitory potencies. How-
ever, the MMP selectivity profile, as shown for compound 26,
is completely lost in both PEGylated compounds (Table 1).
Radiosyntheses were initiated to achieve the radiofluorinat-
ed analogue of 38, [18F]38. Unfortunately, due to nearly identi-
cal HPLC retention times of the mesylate precursor 36 and the
non-radioactive tracer analogue 38 under the chosen HPLC-
conditions (data not shown), mesylate 36 emerged as an un-
suitable precursor for the radiofluorination of the PEG8 variant
[18F]38. Therefore, the tosyl precursor 37, showing a different
HPLC retention time compared with 38, was applied in prelimi-
nary radiosynthesis approaches. The HPLC-purification step
was then feasible, however, only traces of the desired radiola-
beled product [18F]38 were obtained. A major non-radioactive
by-product was isolated by HPLC. We hypothesize, based on
mass spectrometry data ([M]+ =796 Da), that a rearrangement
product (see compound 39 in scheme S1 of the Supporting In-
formation) is formed by intramolecular ring closure either at
the pyrimidine-2,4,6-trione oxygen or nitrogen. Future refine-
ment of the radiofluorination conditions; for example, the ex-
change of the [18F]K(K222)F system for [18F]TBAF, is aimed at the
suppression of this undesired side reaction, which may arise
from a template effect of the potassium ion.
Considering the clogD values of the PEG4-variant 31
(clogD=1.72) and PEG1-variant 26 (clogD=2.88), we succeed-
ed to further reduce the lipophilicity of the fluorinated mini-
PEG-barbiturates by more than one logD unit (14-fold). Using
compounds 42–46, further fluoro-, hydroxy-, azido- and
amino-functionalized PEG4 derivatives were prepared, bearing
bromo, iodo or methoxy groups, respectively, at the 3- and/or
4-(4-phenoxyphenyl) positions.
PEG8-tracer variant by click approach
Like azide 28, alkyne 32 allows for the application of the cop-
per(I)-catalyzed 1,3-dipolar Huisgen cycloaddition (click reac-
tion[37–40]) to generate 1,4-disubstituted 1,2,3-triazoles, such as
36–38, by reaction with azides 33–35 (Scheme 8). The 1,2,3-tri-
azole unit has been shown to possess some valuable charac-
teristics with respect to the intended in vivo application. Apart
from its function as a linker, it is also able to act as part of a
pharmacophore. As such, it can contribute to the binding of
the molecule to biological targets through hydrogen-bonding,
dipole (dipole moment >5D) and p-stacking interactions.[41,42]
Furthermore, it can improve pharmacokinetic parameters, hy-
drophilicity and solubility beyond the level achieved by PEGy-
lation alone. Moreover, 1,2,3-triazoles are relatively robust to-
wards metabolic degradation.
Conclusions
Three mini-PEG-spaced, fluoro-PEGylated non-radioactive ana-
logues of barbituric acid-based PET-compatible MMP-targeted
radiotracers 26 (n=1), 31 (n=4) and 38 (n=8) (n=number of
EG units) have been synthesized to systematically explore the
effect of PEGylation on the lipophilicity, as well as MMP inhibi-
tory potency and MMP-subgroup selectivity of the resulting
barbiturate-based MMP inhibitors. A two-step-radiosynthesis of
[18F]26 (n=1), which in analogy to model tracer 2 was found
to be a highly potent and gelatinase-selective tracer candidate,
was successfully established.
The number of studies on bioactive molecules that are la-
beled with [18F]fluoro-PEGylated radiosynthons by click chemis-
try[43] is increasing continuously. Recently, Chen et al. success-
fully applied a similar click chemistry approach by labeling an
RGD peptide with a [18F]fluoro-PEG3 spacer, yielding the tracer
18F-FTPA-RGD2. Preclinical microPET studies using tumor-bear-
ing mice with an increased avb3 integrin density in the tumor
tissue showed that the tracer displayed good tumor-targeting
efficacy, metabolic stability, as well as favorable in vivo pharma-
cokinetics (i.e., predominantly renal excretion, fast clearance
rate, low liver uptake).[44]
Preliminary biodistribution studies of [18F]26 indicated that
there was no tissue specific accumulation in wild-type mice.
The clearance of any 18F-labeled radiotracer candidate from a
living subject should correspond with the physical half-life of
1
the PET radionuclide (t =109.7 min), which is indeed realized
2
in the case of model tracer [18F]26, warranting its future evalu-
ation in various murine models displaying pathological MMP
dysregulation. Low uptake in the normal state is advantageous
for the study of locally enhanced MMP activity in the diseased
state. The non-radioactive analogues of the PEGylated tracer
variants 31 (n=4) and 38 (n=8) maintained a high MMP in-
784
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
the mixture was extracted with CH2Cl2 (3ꢃ). The combined organic
extracts were washed with brine (1ꢃ), dried (Mg2SO4), filtered and
concentrated in vacuo. Purification by silica gel chromatography
gave the N-Boc-protected piperazines: EtOAc/cyclohexane, 2:1 (6);
EtOAc/MeOH, 19:1 (9); EtOAc/MeOH, 2:1 (21). The reaction prod-
ucts were then subjected to N-Boc removal by action of an excess
of 50% TFA in CH2Cl2. The mixture was stirred at RT overnight and
finally evaporated to remove the excess of TFA. Purification by
silica gel column chromatography (EtOAc/MeOH, 2:1) gave pipera-
zinium trifluoroacetates 6, 9 and 21.
hibitory potency, however, at the expense of MMP-subgroup
selectivity, which was still found for derivative 26 (n=1). Ad-
vanced precursors (e.g., 32 and 37) were prepared, but further
work remains necessary to ultimately establish corresponding
radiolabeling procedures for the preparation of the [18F]fluoro-
PEG4- and -PEG8-ylated variants of barbiturate-based MMP-tar-
geted radioligands.
Finally, the approach of barbiturate-based MMP-targeted ra-
diotracers considered here should be applicable in noninvasive
in vivo imaging of MMP-2 and MMP-9-associated diseases by
means of PET, ideally in combination with the necessary mor-
phological overlay for precise anatomic localization of the mo-
lecular signal, which is nowadays realized by the state-of-the-
art hybrid imaging systems, such as PET-CT and MR-PET.
4-(2-Fluoroethyl)piperazin-1-ium trifluoroacetate (6): N-Boc-pro-
tected intermediate: yield=33%, yellow oil; 1H NMR (300 MHz,
2
3
CDCl3): d=4.57 (dt, JH,F =47.9 Hz, J=4.9 Hz, 2H, N-CH2CH2F), 3.45
(t, 3J=4.9 Hz, 4H, CH2), 2.70 (dt, 3JH,F =28.4 Hz, J=4.9 Hz, 2H, N-
3
CH2CH2F), 2.48 (t, 3J=4.9 Hz, 4H, CH2), 1.46 ppm (s, 9H, CH3);
13C NMR (75.5 MHz, CDCl3): d=154.68, 81.85 (j JC,F j=168.3 Hz),
1
79.61, 58.23 (j JC,F j=20.2 Hz), 53.27, 48.90, 28.40 ppm; 19F NMR
2
(282 MHz, CDCl3): d=ꢀ218.25 ppm; HRMS-ESI: m/z [M+Na]+
calcd for C11H21FN2O2Na: 255.1479, found: 255.1455; Anal. calcd for
C11H21FN2O2: C 56.87, H 9.11, N 12.06, found: C 57.09, H 9.32, N
Experimental Section
Chemistry
12.30. TFA salt: yield=46%, colorless solid; mp: 112–1138C;
General Methods: All chemicals, reagents and solvents for the syn-
thesis of the compounds were analytical grade, purchased from
commercial sources and used without further purification, unless
otherwise specified. Melting points were determined in capillary
tubes on a Stuart Scientific SMP3 capillary melting point apparatus
2
1H NMR (300 MHz, CDCl3): d=9.28 (br s, 2H, NH2+), 4.72 (d, JH,F
=
47.7 Hz, 2H, N-CH2CH2F), 3.31–3.16 ppm (m, 10H, CH2); 13C NMR
1
2
(75.5 MHz, CDCl3): d=79.50 (j JC,F j=165.9 Hz), 56.09 (j JC,F j=
19.1 Hz), 48.69, 41.01 ppm; 19F NMR (282 MHz, CDCl3): d=ꢀ218.46;
HRMS-ESI: m/z [M]+ calcd for C6H14FN2: 133.1136, found: 133.1146;
Anal. calcd for C6H14FN2·2TFA: C 33.44, H 3.93, N 7.80, found: C
33.39, H 3.97, N 7.70.
1
and are uncorrected. H NMR, 13C NMR and 19F NMR spectra were
recorded on Bruker ARX 300 and/or AMX 400 spectrometers. Two-
dimensional and NOE-NMR spectra were determined on a Varian
600 MHz Unity Plus instrument. CDCl3 contained tetramethylsilane
(TMS) as an internal standard. Mass spectra were obtained on a
Varian MAT 212 (EI=70 eV) spectrometer and a Bruker MALDI-TOF-
MS Reflex IV instrument (matrix: DHB). Exact mass analyses were
conducted on a Waters Quattro LC and a Bruker MicroTof appara-
tus. Elemental analyses were carried out on a Vario EL III analyzer.
All aforementioned spectroscopic and analytical investigations
were done by staff members of the Institute of Organic Chemistry,
University of Mꢂnster (Germany). All compound purifications and
determinations of purity by HPLC were performed using a Knauer
gradient RP-HPLC system, equipped with two K-1800 pumps, an S-
2500 UV detector and RP-HPLC Nucleosil Eurosphere 100–10 C-18
4-(2-(2-(2-(2-Bromoethoxy)ethoxy)ethoxy)ethyl)piperazin-1-ium
trifluoroacetate (9): N-Boc-protected intermediate: yield=40%,
light yellow oil; 1H NMR (300 MHz, CDCl3): d=3.83–2.43 (m, 24H,
CH2), 1.46 ppm (s, 9H, CH3); 13C NMR (75.5 MHz, CDCl3): d=154.70,
79.53, 71.18, 70.65, 70.56, 70.51, 70.36, 68.81, 57.80, 53.33, 48.92,
30.31, 28.41 ppm; HRMS-ESI: m/z [M+H]+ calcd for C17H34BrN2O5:
425.1645, found: 425.1649. TFA salt: yield=47%, light yellow oil;
1H NMR (300 MHz, CDCl3): d=8.15 (br s, 2H, NH2+), 3.80–2.99 ppm
(m, 24H, CH2); 13C NMR (75.5 MHz, CDCl3): d=71.21, 71.06, 70.57,
70.39, 70.31, 66.56, 60.40, 56.96, 49.43, 30.48 ppm; HRMS-ESI: m/z
[M]+ calcd for C12H26BrN2O3: 325.1121, found: 325.1129.
columns for analytical (250 mm
ꢃ 4.6 mm), semipreparative
4-(2-(2-(2-(2-Hydroxyethoxy)ethoxy)ethoxy)ethyl)piperazine (14):
Compound 13 (10.0 g, 19.8 mmol; see Supporting Information)
was stirred in MeOH (60 mL) containing concd HCl (10 mL) for
16 h. The solution was adjusted to pH 7.6 by adding solid NaHCO3
(~10 g) and the resulting solution was diluted with water (200 mL).
The precipitate (tritylmethylether) that formed upon addition of
water was filtered off and the filtrate was evaporated to dryness.
After the addition of CH2Cl2 (100 mL) the newly formed precipitate
was removed by filtration. Concentration of the filtrate resulted in
(250 mm ꢃ 8 mm) and preparative (250 mm ꢃ 20 mm) purposes
(unless specified otherwise). The following eluents were used:
eluent A: water (0.1% TFA), eluent B: acetonitrile (0.1% TFA)
(unless specified otherwise). The following conditions were used
(unless specified otherwise):
a) analytical: 90% A!20% A (30 min), 20% A (5 min), 20% A!
90% A (5 min); flow rate=1.5 mLminꢀ1; detection at l=254 nm.
b) semipreparative: 90% A!20% A (17 min), 20% A (2.5 min),
20% A!90% A (1.5 min); flow rate=5.5 mLminꢀ1; detection at
l=254 nm. c) preparative: 80% A!30% A (12 min), 30% A
(1 min); flow rate=18.5 mLminꢀ1; detection at l=254 nm.
1
an orange oil; yield=2.89 g (11.0 mmol, 56%); H NMR (300 MHz,
CDCl3): d=3.78–2.43 ppm (m, 24H, CH2); 13C NMR (75.5 MHz,
CDCl3): d=72.83, 70.59, 70.51, 70.25, 70.21, 68.48, 61.49, 58.38,
54.73, 45.80 ppm; MS (MALDI-TOF): m/z 263 [M+H]+; Anal. calcd
for C12H26N2O4: C 54.94, H 9.99, N 10.68, found: C 52.62, H 9.86, N
10.35.
Preparation of piperazine derivatives: Compounds 10 and 11
were prepared according to previously described procedures.[25] Pi-
perazine derivates 14 and 18 were prepared as specified below.
The piperazinium trifluoroacetates 6, 9 and 21 were prepared fol-
lowing this general procedure: N-Boc piperazine (4)[26] (1.45 equiv)
was added to a solution of 2-fluoroethyl tosylate (5) (prepared by
adding TsCl to a solution of 2-fluoroethanol in pyridine), com-
pound 8 (see Supporting Information) or mesylate 20 (see Sup-
porting Information) and Cs2CO3 (1 equiv) in dry DMF
(10 mLmmolꢀ1). After stirring at RT for 2 d, water was added and
4-(2-(2-(2-(2-Fluoroethoxy)ethoxy)ethoxy)ethyl)piperazine (18): A
solution of 17 (2.75 g, 10.0 mmol; see Supporting Information) and
Cs2CO3 (4.88 g, 15.0 mmol) in abs DMF (60 mL) was treated with pi-
perazine (3) (1.72 g, 20.0 mmol) and heated at 1008C for 2 d. The
solvent was then evaporated and the residue was taken up in
EtOAc (120 mL). The precipitate was filtered off and the solvent
was removed to give an orange oil, which was used in the next
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
785

H.-J. Breyholz et al.
MED
step without further purification. The product contained traces of
the dialkylated by-product (as seen by MS); yield=2.48 g
(1ꢃ), dried (Na2SO4), filtered and concentrated to give an orange
oily residue. The residue was purified by silica gel column chroma-
tography (EtOAc/MeOH, 19:1) to give pure 27; yield=36%; mp:
1508C; 1H NMR (300 MHz, [D6]DMSO): d=11.59 (br s, 2H), 7.44–
6.91 (m, 9H, HAryl), 3.68–2.32 ppm (m, 24H, CH2); 13C NMR
(75.5 MHz, [D6]DMSO): d=169.22, 157.22, 155.74, 149.36, 130.13,
129.62, 129.56, 124.00, 119.22, 118.00, 73.86, 70.30, 69.70, 69.57,
69.52, 68.22, 68.10, 57.05, 53.65, 47.23, 30.71 ppm; HRMS-ESI: m/z
[M+H]+ calcd for C28H35BrN4O7: 619.1764, found: 619.1762; Anal.
calcd for C28H35BrN4O7: C 54.29, H 5.69, N 9.04, found: C 55.59, H
5.78, N 8.99.
2
(9.38 mmol, 94%); 1H NMR (300 MHz, CDCl3): d=4.55 (dt, JH,F
=
3
47.6 Hz, J=8.5 Hz, 2H, O-CH2CH2F), 3.81–2.45 ppm (m, 22H, CH2);
13C NMR (75.5 MHz, CDCl3): d=83.11 (j JC,F j=166.0 Hz), 70.84,
1
2
70.66, 70.56, 70.29, 69.70 (j JC,F j=20.3 Hz), 68.88, 58.47, 55.08,
46.78 ppm; 19F NMR (282 MHz, CDCl3): d=ꢀ221.55; MS (MALDI-
TOF): m/z 265 [M+H]+.
4-(3,6,9,12-Tetraoxapentadec-14-ynyl)piperazin-1-ium 2,2,2-tri-
fluoroacetate (21): N-Boc-protected intermediate: yield=70%,
yellow oil; 1H NMR (300 MHz, CDCl3): d=4.20 (d, 2H, 4J=2.5 Hz,
CH2CꢁCH), 3.70–2.44 (m, 25H, CH2, CH), 1.46 ppm (s, 9H, CH3);
13C NMR (75.5 MHz, CDCl3): d=154.70, 79.64, 79.55, 74.57, 70.62,
70.59, 70.57, 70.41, 70.38, 69.08, 68.84, 58.39, 57.82, 53.36, 49.03,
28.42 ppm; HRMS-ESI: m/z [M+Na]+ calcd for C20H36N2O6Na:
423.2655, found: 423.2466; Anal. calcd for C20H36N2O6: C 59.98, H
9.06, N 6.99, found: C 59.06, H 9.07, N 7.02. TFA salt: yield=86%,
light yellow oil; 1H NMR (300 MHz, CDCl3): d=4.17 (d, 2H, 4J=
5-[4-(2-(2-(2-(2-Hydroxyethoxy)ethoxy)ethoxy)ethyl)piperazin-1-
yl]-5-(4-phenoxyphenyl)pyrimidine-2,4,6-trione (30): Compound
14 was used to prepare this compound. For work up, the solvent
was evaporated and the residue purified by silica gel column chro-
matography (EtOAc/MeOH, 2:1) to give a highly viscous oil; yield=
1
55%; H NMR (300 MHz, [D6]DMSO): d=11.53 (br s, 2H), 7.38–6.93
(m, 9H, HAryl), 3.51–2.22 ppm (m, 24H, CH2); 13C NMR (75.5 MHz,
[D6]DMSO): d=169.98, 157.29, 155.75, 149.46, 130.52, 130.13,
129.63, 123.99, 119.21, 118.00, 73.84, 72.27, 69.74, 69.69, 69.57,
68.11, 60.12, 57.08, 53.67, 47.25, 29.46 ppm; HRMS-ESI: m/z [M+
H]+ calcd for C28H37N4O8: 557.2606, found: 557.2599. The purity of
30 was determined by analytical HPLC to be 95%, tR =20.93ꢂ
0.73 min (n=5).
4
2.4 Hz, CH2CꢁCH), 3.82–3.34 (m, 24H, CH2, CH), 2.52 ppm (t, J=
2.4 Hz, 1H, CH); 13C NMR (75.5 MHz, CDCl3): d=79.37, 75.04, 70.26,
70.24, 70.20, 70.00, 69.95, 69.85, 69.83, 58.20, 56.56, 51.90,
48.86 ppm; HRMS-ESI: m/z [M]+ calcd for C15H29N2O4: 301.2122,
found 301.2134; Anal. calcd for C15H29N2O4·3TFA: C 39.38, H 4.56, N
4.37, found: C 39.47, H 5.04, N 4.58.
General procedure for the preparation of pyrimidine-2,4,6-tri-
ones 24 and 25–32: Compounds 24 and 25–32 were prepared ac-
cording to the previously described procedures.[26,29]
5-[4-(2-(2-(2-(2-Fluoroethoxy)ethoxy)ethoxy)ethyl)piperazin-1-
yl]-5-(4-phenoxyphenyl)pyrimidine-2,4,6-trione (31): Compound
18 was used to prepare this compound. For work up, the solvent
was evaporated and the residue purified by silica gel column chro-
matography (EtOAc/MeOH, 2:1) to give a highly viscous oil; yield=
5-(4-Phenoxyphenyl)-5-piperazinyl-pyrimidine-2,4,6-trione (25):
Piperazine (3) was used to prepare this compound. For work up,
the reaction mixture was taken up in EtOAc and water was added.
The colorless precipitate was collected by suction; yield=43%, col-
orless solid; mp: 275–2778C (decomp); 1H NMR (300 MHz,
[D6]DMSO): d=7.45–6.93 (m, 9H, HAryl), 2.68–2.19 ppm (m, 8H,
CH2); 13C NMR (75.5 MHz, [D6]DMSO): d=170.52, 157.10, 155.89,
150.30, 130.12, 129.81, 123.92, 119.87, 119.15, 117.93, 74.18, 48.69,
46.03 ppm; HRMS-ESI: m/z [M+H]+ calcd for C20H21N4O4: 381.1563,
found 381.1559; Anal. calcd for C20H20N4O4·0.5H2O: C 61.69, H 5.43,
N 14.38, found: C 61.74, H 5.12, N 14.29.
1
32%; H NMR (300 MHz, [D6]DMSO): d=11.53 (br s, 2H), 7.36–6.91
(m, 9H, HAryl), 4.52 (d, 2JH,F =48.5 Hz, 2H, O-CH2CH2F), 3.63–
2.19 ppm (m, 22H, CH2); 13C NMR (75.5 MHz, [D6]DMSO): d=169.86,
157.39, 155.65, 149.36, 130.15, 129.54, 126.94, 124.05, 119.21,
1
118.04, 83.00 (j JC,F j=165.4 Hz), 73.82, 72.27, 70.15, 69.73, 69.64
2
(j JC,F j=21.9 Hz), 69.46, 56.63, 53.36, 46.66, 30.11 ppm; HRMS-ESI:
m/z [M+H]+ calcd for C28H36FN4O7: 559.2563, found: 559.2557. The
purity of 31 was determined by analytical HPLC to be 91%, tR =
22.48ꢂ0.07 min (n=3).
5-[4-(3,6,9,12-Tetraoxapentadec-14-ynyl)piperazin-1-yl]-5-(4-phe-
noxyphenyl)pyrimidine-2,4,6-trione (32): Compound 21 was used
as the piperazine component to prepare this compound. For work
up, water was added to the reaction mixture, which was subse-
quently extracted with EtOAc (3ꢃ). The combined extracts were
washed with brine (1ꢃ), dried (Mg2SO4) and filtered. The solvent
was removed in vacuo and the residue purified by silica gel
column chromatography (EtOAc/MeOH, 19:1) to give pure 32;
yield=78%; mp: 155–1578C; 1H NMR (400 MHz, [D6]DMSO): d=
1.61 (br s, 2H), 7.42–6.99 (m, 9H, HAryl), 4.11 (s, 2H, CH2-CꢁCH),
3.53–2.48 (m, 25H, CH, CH2); 13C NMR (100 MHz, [D6]DMSO): d=
169.90, 157.33, 155.78, 149.35, 130.12, 129.64, 126.75, 124.00,
119.22, 118.02, 80.29, 77.02, 73.89, 69.97, 69.72, 69.69, 69.59, 69.45,
68.47, 68.11, 57.44, 57.07, 53.66, 47.25 ppm; HRMS-ESI: m/z [M+
H]+ calcd for C31H39N4O8: 595.2562, found: 595.2759; Anal. calcd
for C31H38N4O8: C 62.61, H 6.44, N 9.42, found: C 62.33, H 6.40, N
9.29.
5-[4-(2-Fluoroethyl)piperazin-1-yl]-5-(4-phenoxyphenyl)pyrimi-
dine-2,4,6-trione (26): Compound 6 was used to prepare this com-
pound. For work up, the reaction mixture was diluted with EtOAc
and aqueous citric acid (0.1 mmolmLꢀ1) was added. The aqueous
phase was extracted with EtOAc (2ꢃ) and the combined organic
phases were washed with water, dried (Na2SO4), filtered and con-
centrated. The residue was purified by silica gel column chroma-
tography (EtOAc/MeOH, 19:1) and the concentrated eluent was
subsequently stirred in diisopropylether/acetone (99:1 v/v) to give
pure 26; yield=45%; mp: 222–2248C (decomp); 1H NMR
(300 MHz, [D6]DMSO): d=11.46 (br s, 2H), 7.37–6.93 (m, 9H, HAryl),
4.43 (dt, 2JH,F =48.0 Hz, 3JH,H =9.6 Hz, 2H, N-CH2CH2F), 2.59–
2.38 ppm (m, 10H, CH2); 13C NMR (75.5 MHz, [D6]DMSO): d=169.92,
157.33, 155.74, 149.35, 130.13, 129.62, 129.56, 124.01, 119.23,
1
2
118.02, 81.65 (j JC,F j=164.8 Hz), 73.87, 57.39 (j JC,F j=19.7 Hz),
53.46, 47.23 ppm; HRMS-ESI: m/z [M+H]+ calcd for C22H24FN4O4:
427.1776, found: 427.1779. The purity of 26 was determined by an-
alytical HPLC to be >97%, tR =20.64ꢂ0.08 min (n=3).
General procedure for the preparation of triazoles 36–38:
Alkyne 32 (100 mg, 168 mmol) was dissolved in a mixture of tBuOH
(6 mL) and H2O (1.5 mL) under heating. The solution was cooled to
RT and mesylate 33 (50 mg, 168 mmol), tosylate 34 (90 mg,
168 mmol), or fluoro compound 35 (37 mg, 168 mmol), CuSO4·5H2O
(16.8 mg, 67.2 mmol, 40 mol% in case of 36 and 38 or 12.6 mg,
5-[4-(2-(2-(2-(2-Bromoethoxy)ethoxy)ethoxy)ethyl)piperazin-1-
yl]-5-(4-phenoxyphenyl)pyrimidine-2,4,6-trione (27): Compound
9 was used to prepare this compound. For work up, water was
added to the reaction mixture, which was subsequently extracted
with EtOAc (3ꢃ). The combined extracts were washed with brine
786
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
50.4 mmol, 30 mol% in case of 37, respectively) as well as sodium
ascorbate (26.6 mg, 13.4 mmol, 80 mol% in case of 36 and 38 or
13.3 mg, 67.2 mmol, 40 mol%, in case of 37, respectively) were
added in succession. After stirring at RT overnight, the solvents
were removed and the residue purified by silica gel column chro-
matography (36: EtOAc/MeOH (4:1); 37: EtOAc/MeOH (9:1) !
EtOAc/MeOH (4:1); 38: EtOAc/MeOH (1:1)). All eluted product frac-
tions still contained 16–50% by-product (see below). Further purifi-
cation was performed by preparative HPLC. The appropriate frac-
tions (36: tR =10.50 min; 37: tR =10.48ꢂ0.03 min (n=2); 38: tR =
10.38ꢂ0.08 min (n=2)) were collected and evaporated to give the
pure (>95%, see below) triazoles 36–38 as colorless oils. The by-
product (yields not determined), which could not be separated
from the main products 36–38 by silica gel column chromatogra-
phy turned out to be hydroxy compound 30, as determined by an-
alytical HPLC by co-injection with a sample of 30. tR(30)=21.49ꢂ
0.12 min (n=4), tR(36)=23.67ꢂ0.02 min (n=2), tR(37)=27.37ꢂ
0.07 min (n=5), tR(38)=22.95ꢂ0.04 min (n=2); HRMS-ESI: m/z
[M+H]+ calcd for C28H37N4O8: 557.2606, found: 557.2599 (see
above for 30).
Radiochemistry
General methods: Radiofluorinations were carried out on a modi-
fied PET tracer radiosynthesizer (TRACERLab FxFDG, GE Healthcare,
http://www.gehealthcare.com/eude/fun_img/products/radiophar-
macy/products/tracercenter_tracerlabs.html). The recorded data
were processed by the TRACERLab Fx software (GE Healthcare).
Separation and purification of the radiosynthesized compounds
were—unless specified otherwise—performed by a gradient radio-
HPLC chromatography system (HPLC A) using a Knauer K-500 and
a Latek P 402 pump, a Knauer K-2000 UV-detector (l=254 nm) a
Crismatec NaI(Tl) Scintibloc 51 SP51 g-detector and a RP-HPLC Nu-
cleosil column 100–10 C-18 (250 mm ꢃ 8 mm), a corresponding
precolumn (20 mm ꢃ 4.0 mm). Sample injection was carried out
using a Rheodyne injector block (type 7125 including a 200 mL
loop). The recorded data were processed by the NINA version 4.9
software (GE Medical Systems). The radiochemical purities and the
specific activities were determined using a radio-HPLC system
(HPLC B) composed of a Syknm S1021 pump, a Knauer K-2501 UV-
detector (l=254 nm), a Crismatec NaI(Tl) Scintibloc 51 SP51 g-de-
tector, a RP-HPLC Nucleosil 100–3 C-18 column (200 mm ꢃ 3 mm),
a VICI injector block (type C1 including a 20 mL loop) and the GINA
Star version 4.07 radiochromatography software (Raytest Isotopen-
meßgerꢀte GmbH). Radio-TLC plates were analyzed using a miniGI-
TA TLC-Scanner (Raytest Isotopenmeßgerꢀte GmbH).
2-(2-(2-(2-(4-(13-(4-(2,4,6-Trioxo-5-(4-phenoxyphenyl)-hexa-hy-
dropyrimidin-5-yl)piperazin-1-yl)-2,5,8,11-tetraoxatridecyl)-1H-
1,2,3-triazol-1-yl)ethoxy)ethoxy)ethoxy)ethyl-methane-sufonate
1
(36): yield=29 mg (32 mmol, 19%); H NMR (300 MHz, CDCl3): d=
Production of [18F]fluoride and synthesis of [18F]K(K222)F: No-car-
rier-added aqueous [18F]fluoride was produced on a RDS 111e cy-
clotron (CTI-Siemens) by irradiation of a 1.2 mL water target using
10 MeV proton beams on 97.0% enriched 18O-water by the
18O(p,n)18F nuclear reaction. Typical batches of [18F]fluoride ions
were 8.1 GBq (in the case of the generation of [18F]26) and 2.7 GBq
(in the case of radiofluorination of tosylate 37) at the end of radio-
nuclide production for currents of 32 mA and irradiation times of
10 or 4 min, respectively. To recover the 18O-water, the batch of
aqueous [18F]fluoride was passed through an anion exchange resin
(Sep-Pak Light Waters Accell Plus QMA cartridge, preconditioned
with 5 mL 1mK2CO3 and 10 mL water). [18F]Fluoride was eluted
from the resin with a mixture of 40 mL 1m K2CO3, 200 mL water for
injection and 800 mL DNA-grade CH3CN containing 18 mg K222. Sub-
sequently, the aqueous [18F]K(K222)F solution was carefully evaporat-
ed to dryness in vacuo.
7.74 (s, 1H, CHTriazole), 7.50–6.94 (m, 9H, HAryl), 4.61 (s, 2H, OCH2 Tria-
zole), 3.87–3.63 (m, 30H, CH2), 3.04 (s, 3H, CH3), 2.81–2.65 ppm (m,
10H, CH2); 13C NMR (75.5 MHz, CDCl3): d=169.64, 158.09, 156.35,
149.59, 144.75, 130.03, 129.85, 128.89, 123.96, 123.85, 119.46,
118.19, 85.49, 70.62, 70.60, 70.52, 70.51, 70.46, 70.44, 70.40, 70.32,
70.21, 70.06, 69.53, 69.43, 69.01, 64.21, 53.50, 53.46, 50.21, 47.05,
37.65 ppm; HRMS-ESI: m/z [M+H]+ calcd for C40H58N7O14:
892.3757, found: 892.3705. The purity was >95% as determined
by analytical HPLC, tR =23.65 min.
2-(2-(2-(2-(4-(13-(4-(2,4,6-Trioxo-5-(4-phenoxyphenyl)hexa-hydro-
pyrimidin-5-yl)piperazin-1-yl)-2,5,8,11-tetraoxatridecyl)-1H-1,2,3-
triazol-1-yl)ethoxy)ethoxy)ethoxy)ethyl-4-methyl-benzene-sulfo-
1
nate (37): yield=34 mg (36 mmol, 30%); H NMR (300 MHz, CDCl3):
d=7.80–6.94 (m, 14H, HAryl, HTriazole), 4.68 (s, 2H, OCH2 Triazole), 4.12–
3.60 (m, 30H, CH2), 2.82–2.65 (m, 10H, CH2) 2.43 ppm (s, 3H, CH3);
13C NMR (75.5 MHz, CDCl3): d=169.50, 158.14, 156.36, 149.35,
144.88, 144.72, 132.93, 130.07, 129.88, 129.87, 124.05, 123.89,
123.88, 119.52, 119.50, 118.22, 86.34, 70.72, 70.60, 70.47, 70.43,
70.39, 70.33, 70.28, 70.25, 70.19, 70.14, 69.47, 69.40, 69.26, 68.68,
67.71, 57.43, 53.48, 50.21, 47.09, 21.66 ppm; HRMS-ESI: m/z [M+
H]+ calcd for C46H62N7O14S: 968.4070, found 968.3837. The purity
was >95% as determined by semipreparative HPLC, tR =14.30 min.
5-[4-(2-[18F]Fluoroethyl)piperazin-1-yl]-5-(4-phenoxyphenyl)pyri-
midine-2,4,6-trione ([18F]26): Ethylene glycol di-p-tosylate (5.0 mg,
13.5 mmol) in DNA-grade CH3CN (1 mL) was added to the carefully
dried [18F]K(K222)F residue (see above) and the mixture was heated
at 848C for 4 min. The mixture was cooled to 408C, diluted with
water (10 mL) and passed through a Waters Sep-Pak Light C18 car-
tridge. The cartridge was washed with additional water (10 mL)
and dried under a flow of He for 10–20 min, followed by elution of
1-[18F]fluoro-2-tosyloxyethane ([18F]5) with hot DMSO (400 mL,
heated to 1208C prior to elution) into a 5 mL flask containing pre-
cursor 25 (4.0 mg, 10.5 mmol) and DMSO (100 mL). [18F]5 was ob-
tained in a radiochemical yield of 53.3ꢂ2.3% (decay-corrected, n=
6). The following steps were performed outside the synthesizer.
5-(4-(1-(1-(2-(2-(2-(2-Fluoroethoxy)ethoxy)ethoxy)ethyl)-1H-1,2,3-
triazol-4-yl-)-2,5,8,11-tetraoxa-tridecan-13-yl)piperazin-1-yl)-5-(4-
phenoxyphenyl)pyrimidine-2,4,6-trione
(38):
yield=44 mg
1
(54 mmol, 32%); H NMR (300 MHz, CDCl3): d=7.75 (s, 1H, CHTriazole),
7.37–6.94 (m, 9H, HAryl), 4.54 (dt, 2JH,F =47.7 Hz, 2H, O-CH2CH2F),
4.51 (s, 2H, OCH2Triazole), 3.87–2.68 ppm (m, 38H, CH2); 13C NMR
(75.5 MHz, CDCl3): d=169.62, 158.10, 156.39, 149.53, 144.77,
The reaction mixture was stirred at 1208C for 15 min, diluted with
water (400 mL) and applied to the HLPC column (see general meth-
ods section above). HPLC A; conditions: eluent A: CH3CN/H2O/TFA
(950:50:1), eluent B: CH3CN/H2O/TFA (50:950:1); gradient: 75%
eluent B, after elution of the product 30% eluent B for 20 min at a
flow rate of 4.0 mLminꢀ1, l=254 nm. The appropriate fraction
([18F]26: tR =27.1 min) was collected, evaporated to dryness in va-
cuo and redissolved in 1 mL H2O/EtOH (9:1). [18F]26 was synthe-
1
129.87, 128.94, 124.00, 123.85, 119.48, 118.21, 88.16, 83.15 (j JC,F j=
2
169.6 Hz), 70.63 (j JC,F j=20.2 Hz), 70.61, 70.54, 70.50, 70.48, 70.45,
70.42, 70.38, 70.32, 70.31, 70.27, 69.51, 69.43, 64.27, 57.35, 53.47,
50.24, 42.11 ppm; 19F NMR (282 MHz, CDCl3): d=ꢀ222.67 ppm;
HRMS-ESI: m/z [M+H]+ calcd for C39H55FN7O11: 816.3938, found:
816.3898. The purity was >95% as determined by semipreparative
HPLC, tR =12.17ꢂ0.01 min (n=2).
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
787

H.-J. Breyholz et al.
MED
sized with radiochemical yields of 21.3ꢂ4.0% (decay-corrected,
n=6) starting from [18F]5. The overall radiochemical yields of this
two-step radiosynthesis were 11.3ꢂ2.2% (decay-corrected, based
on cyclotron-derived [18F]fluoride ions). The radiochemical purities
were 98.5ꢂ1.5% after 142ꢂ12 min from the end of radionuclide
production. The determined specific radioactivity (As) was 23ꢂ
10 GBqmmolꢀ1 at the end of radiosynthesis (n=6). The As of prod-
uct [18F]26 was estimated by comparing the peak area of purified
[18F]26 in the UV channel (tR =8.4 min, HPLC B) with a standard
curve of known concentrations of reference compound 26 mea-
sured with the HPLC B system (flow rate: 0.3 mLminꢀ1; eluent:
CH3CN/H2O/TFA (400:600:1)). The chemical identity of [18F]26 was
proven by two different chromatographical methods. 1) HPLC: Co-
injection of [18F]26 and non-radioactive counterpart 26 resulted in
co-elution on the HPLC A and B systems. 2) Radio-TLC: Both [18F]26
and non-radioactive counterpart 26 possessed the same Rf value
(Rf =0.69 (EtOAc/MeOH, 4:1); silica gel UV254).
1 mm) in Tris·HCl (50 mm), pH 7.5, containing NaCl (0.2m), CaCl2
(5 mm), ZnSO4 (20 mm) and 0.05% Brij 35 at 378C for 30 min. An ali-
quot of substrate (10 mL of a 50 mm solution) was then added to
the preincubated MMP/inhibitor mixture (90 mL), and the fluores-
cence was determined at 378C by following product release over
time. The fluorescence changes were monitored using a Fusion
Universal Microplate Analyzer (Packard Bioscience, Massachusetts,
USA) with excitation and emission wavelengths of 330 and
390 nm, respectively. Reaction rates were measured from the initial
10 min of the reaction profile where product release was linear
with time and plotted as a function of inhibitor concentration.
From the resulting inhibition curves, the IC50 values for each inhibi-
tor were calculated by nonlinear regression analysis, performed
using the Grace 5.1.8 software (Linux).
Animal studies
Reaction of tosylate precursor 37 with [18F]K(K222)F: Tosylate pre-
cursor 37 (4.2 mg, 4.34 mmol) and carefully dried [18F]K(K 222)F (see
above) residue were heated at 848C in DNA-grade CH3CN (1 mL)
for 20 min. The mixture was cooled to 408C, diluted with water for
injection (10 mL) and passed through a Waters Sep-Pak Light C18
cartridge. The cartridge was washed with additional water for in-
jection (10 mL) and eluted with hot DMF (2.0 mL, heated to 1208C
prior to elution). The eluent was evaporated to dryness, diluted
with CH3CN/water (1:1; 1.0 mL) for injection, vortexed and purified
by semipreparative gradient HPLC (see the general methods under
the Chemistry section). By collecting and evaporating of the appro-
priate fractions, only a small amount of radiofluorinated product
[18F]38 (tR =12.35 min) was discovered. Interestingly, a non-radioac-
tive product fraction (tR =12.43 min) was isolated. The radiofluori-
nated product was presumably [18F]38. As the radiochemical yield
was ꢃ1% the identity of this product with its unlabeled analogue
38 was not verified. The tosylate precursor 37 (tR =14.30 min) was
not recovered in the UV channel of the radio-HPLC system. The
non-radioactive product was examined by MS: HRMS-ESI: m/z [M+
H]+ calcd for C39H54N7O11: 796.3876, found: 796.3863. The identity
of this product is delineated in the Discussion section.
Studies were carried out according to protocols approved by the
local ethics committee and were performed in accordance with in-
stitutional guidelines for health and care of experimental animals.
Biodistribution of [18F]26: Adult C57/BL6 mice (male, ~29 g) were
anaesthetized by isoflurane/O2 and one lateral tail vein was cannu-
lated using a 27 G needle connected to 15 cm polyethylene cathe-
ter tubing (o.d.=1.09 mm, i.d.=0.38 mm). [18F]26 (315 MBqkgꢀ1
)
was injected as a bolus (200 mL) via the tail vein, and subsequent
PET scanning was performed. To outline the position of the heart,
[18F]FDG (4 MBq in 200 mL per mouse) was injected via the tail vein
5 min after completion of the scan with [18F]26, and data were ac-
quired for a further 30 min.
Small-animal PET scanning: PET experiments were carried out using
a sub-millimeter high-resolution (0.7 mm full width at half maxi-
mum) dedicated small-animal scanner (32 module quadHIDAC,
Oxford Positron Systems Ltd., Oxford, UK), which uses wire-cham-
ber detectors and offers uniform spatial resolution (<1 mm) over a
large cylindrical field (165 mm diameter, 280 mm axial length).[47–49]
List-mode data were acquired for 120 min and reconstructed into
dynamic time frames using an iterative reconstruction algorithm.[50]
PET images were analyzed using in-house software algorithms in
MATLAB (The MathWorks Company) and C programming languag-
es.[48,51] For in vivo biodistribution measurements of [18F]26 regions-
of-interest (ROIs) were drawn over selected tissues/organs in five
coronal planes. For precise placement of ROIs, representing arterial
blood pool image data was resliced according to heart coordinates
and ROIs were defined on five short axis planes of the left ventri-
cle.
Measurement of logD value: Determination of the logD value of
[18F]26 was performed in 1-octanol and 0.02m phosphate buffer at
pH 7.4 as described previously.[45,46] The liquid layers were counted
in an automated gamma counter (Wallac Wizard 3’’, Perkin–Elmer
Life Sciences, Boston, USA), and the measured counts were back
corrected for decay. The measurement was repeated four times.
The logD value was determined by calculating the 1-octanol-to-
buffer ratios:
cpm mLꢀ1 octanol
cpm mLꢀ1 buffer
Glossary
log D ¼ logð
Þ
EG, ethylene glycol; EM, exact mass; ESI, electron spray ionization;
[18F]FDG, 2-deoxy-2-[18F]fluoro-d-glucose; K222, Kryptofix 2.2.2; MMP,
matrix metalloproteinase; MMPI, matrix metalloproteinase inhibi-
tor; NHS, N-hydroxysuccinimidyl; NIRF, near-infrared fluorescence;
NOE, Nuclear Overhauser Effect; PEG, polyethylene glycol; PET,
positron emission tomography; p.i., post injection; ROI, region of
interest; SD, standard deviation; SPECT, single photon emission
computed tomography.
Biology
Enzyme inhibition assays
The synthetic fluorogenic substrate (7-methoxycoumarin-4-yl)ace-
tyl-pro-Leu-Gly-Leu-(3-(2,4-dinitrophenyl)-l-2,3-diamino-propionyl)-
Ala-Arg-NH2 (R&D Systems, Minneapolis, USA) was used to assay
activated MMP-2, MMP-8, MMP-9 and MMP-13 as described previ-
ously.[30] The inhibition of human active MMPs by the barbituric
acid derivatives 25–32 and 38 were assayed by preincubating
MMP-1, MMP-2, MMP-3, MMP-8, MMP-9, MMP-13 or MMP-14 (each
at 2 nm) and test compounds at varying concentrations (10 pm–
Acknowledgements
This work was supported by grants from the Deutsche For-
schungsgemeinschaft (DFG; Mꢀnster, Germany; Projects SFB
788
www.chemmedchem.org
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
ChemMedChem 2010, 5, 777 – 789

Molecular Probes for Noninvasive Imaging
[27] a) H.-J. Breyholz, A. Faust, B. Waschkau, S. Wagner, C. Hçltke, J. Waldeck,
W. Brandau, C. Bremer, O. Schober, M. Schꢀfers, K. Kopka, Q. J. Nucl.
Med. Mol. Imaging 2008, 52, N. 1, 23–25; b) H.-J. Breyholz, A. Faust, B.
Waschkau, S. Wagner, C. Hçltke, J. Waldeck, W. Brandau, C. Bremer, O.
Schober, M. Schꢀfers, K. Kopka, Eur. J. Nucl. Med. Mol. Imaging 2008, 35,
Suppl. 2, S316.
[28] C. Tahtaoui, I. Parrot, P. Klotz, F Guillier, J.-L. Galzi, M. Hibert, B. Ilien, J.
Med. Chem. 2004, 47, 4300–4315.
[29] S. Hutchings, W. Liu, R. Radinov, Heterocycles 2006, 67, 763–768.
[30] W. Huang, Q. Meng, K. Suzuki, H. Nagase, K. Brew, J. Biol. Chem. 1997,
272, 22086–22091.
656A2, B1, C6 and Z5) and the Interdisciplinary Center of Clinical
Research (IZKF; Mꢀnster, Germany; Project Schꢁ2/020/09 & core
unit SmAP). We thank Sandra Schrçer for her excellent technical
assistance, K. Bergander and H. Luftmann (Department of Organ-
ic Chemistry, WWU Mꢀnster) for performing NMR spectroscopy
and mass spectrometry, respectively, and S. Fatum for further
technical assistance.
Keywords: fluorinated ligands · inhibitors · metalloenzymes ·
[31] U. Sirion, H. J. Kim, J. H. Lee, J. W. Seo, B. S. Lee, S. J. Lee, S. J. Oh, D. Y.
Chi, Tetrahedron Lett. 2007, 48, 3953–3957.
positron emission tomography
radiochemistry
· pyrimidine-2,4,6-triones ·
[32] a) J. M. Harris, R. B. Chess, Nat. Rev. Drug Discovery 2003, 2, 214–221;
b) J. M. Harris, N. E. Martin, M. Modi, Clin. Pharmacokinet. 2001, 40, 539–
551; c) M. Hamidi, A. Azadi, P. Rafiei, Drug Delivery 2006, 13, 399–409.
[33] a) Z. Wu, Z. B. Li, K. Chen, W. Cai, L. He, F. T. Chin, F. Li, X. Chen, J. Nucl.
Med. 2007, 48, 1536–1544; b) Z. Liu, S. Liu, F. Wang, S. Liu, X. Chen, Eur.
J. Nucl. Med. Mol. Imaging 2009, 36, 1296–1307; c) Z. Liu, G. Niu, J. Shi,
S. Liu, F. Wang, S. Liu, X. Chen, Eur. J. Nucl. Med. Mol. Imaging 2009, 36,
947–957; d) L. Wang, J. Shi, Y. S. Kim, S. Zhai, B. Jia, H. Zhao, Z. Liu, F.
Wang, X. Chen, S. Liu, Mol. Pharm. 2009, 6, 231–245.
[34] Q. Wang, K. Graham, T. Schauer, T. Fietz, A. Mohammed, X. Liu, J. Hof-
fend, U. Haberkorn, M. Eisenhut, W. Mier, Nucl. Med. Biol. 2004, 31, 21–
30.
[35] R. B. Pepinsky, W.-C. Lee, M. Cornebise, A. Gill, K. Wortham, L. L. Chen,
D. R. Leone, K. Giza, B. M. Dolinski, S. Perper, C. Nickerson-Nutter, D.
Lepage, A. Chakraborty, E. T. Whalley, R. C. Petter, S. P. Adams, R. R.
Lobb, D. M. Scott, J. Pharmacol. Exp. Ther. 2004, 312, 742–750.
[36] W. Zhang, M. P. Kung, S. Oya, C. Hou, H. F. Kung, Nucl. Med. Biol. 2007,
34, 89–97.
[1] A. R. Folgueras, A. M. Pendas, L. M. Sanchez, C. Lꢄpez-Otꢅn, Int. J. Dev.
Biol. 2004, 48, 411–424.
[2] C. M. Overall, C. Lꢄpez-Otꢅn, Nat. Rev. Cancer 2002, 2, 657–672.
[3] M. H. Tayebjee, G. Y. Lip, R. J. MacFadyen, Curr. Med. Chem. 2005, 12,
917–925.
[4] I. Stamenkovic, J. Pathol. 2003, 200, 448–464.
[5] P. Vihinen, R. Ala-aho, V. M. Kahari, Curr. Cancer Drug Targets 2005, 5,
203–220.
[6] M. Egeblad, Z. Werb, Nat. Rev. Cancer 2002, 2, 161–174.
[7] S. Heeneman, J. P. Cleutjens, B. C. Faber, E. E. Creemers, R.-J. van Suylen,
E. Lutgens, K. B. Cleutjens, M. J. Daemen, J. Pathol. 2003, 200, 516–525.
[8] Z. S. Galis, J. J. Khatri, Circ. Res. 2002, 90, 251–262.
[9] Z. S. Galis, Circulation 2004, 110, 244–246.
[10] A. C. Newby, Physiol. Rev. 2005, 85, 1–31.
[11] K. Kopka, H.-J. Breyholz, S. Wagner, M. P. Law, B. Riemann, S. Schrçer, M.
Trub, B. Guilbert, B. Levkau, O. Schober, M. Schꢀfers, Nucl. Med. Biol.
2004, 31, 257–267.
[12] M. Schꢀfers, B. Riemann, K. Kopka, H.-J. Breyholz, S. Wagner, K. P. Schꢀ-
fers, M. P. Law, O. Schober, B. Levkau, Circulation 2004, 109, 2554–2559.
[13] H. Su, F. G. Spinale, L. W. Dobrucki, J. Song, J. Hua, S. Sweterlitsch, D. P.
Dione, P. Cavaliere, C. Chow, B. N. Bourke, X. Y. Hu, M. Azure, P. Yalaman-
chili, R. Liu, E. H. Cheesman, S. Robinson, D. S. Edwards, A. J. Sinusas, Cir-
culation 2005, 112, 3157–3167.
[37] G. C. Tron, T. Pirali, R. A. Billington, P. L. Canonico, G. Sorba, A. Genazza-
ni, Med. Res. Rev. 2008, 28, 278–308.
[38] Y. L. Angell, K. Burgess, Chem. Soc. Rev. 2007, 36, 1674–1689.
[39] A. D. Moorhouse, J. E. Moses, ChemMedChem 2008, 3, 715–723.
[40] C. D. Hein, X. M. Liu, D. Wang, Pharm. Res. 2008, 25, 2216–2230.
[41] W. S. Horne, M. K. Yadav, C. D. Stout, M. R. Ghadiri, J. Am. Chem. Soc.
2004, 126, 15366–15367.
[42] A. D. Moorhouse, A. M. Santos, M. Gunaratnam, M. Moore, S. Neidle,
J. E. Moses, J. Am. Chem. Soc. 2006, 128, 15972–15973.
[43] C. Mamat, T. Ramenda, F. R. Wuest, Mini-Rev. Org. Chem. 2009, 6, 21–34.
[44] Z. B. Li, Z. Wu, K. Chen, F. T. Chin, X. Chen, Bioconjugate Chem. 2007, 18,
1987–1994.
[45] A. A. Wilson, S. Houle, J. Labelled Compd. Radiopharm. 1999, 42, 1277–
1288.
[46] A. A. Wilson, L. Jin, A. Garcia, J. N. DaSilva, S. Houle, Appl. Radiat. Isot.
2001, 54, 203–208.
[14] J. Zhang, L. Nie, M. Razavian, M. Ahmed, L. W. Dobrucki, A. Asadi, D. S.
Edwards, M. Azure, A. J. Sinusas, M. M. Sadeghi, Circulation 2008, 118,
1953–1960.
[15] S. Ohshima, A. Petrov, S. Fujimoto, J Zhou, M. Azure, D. S. Edwards, T.
Murohara, N. Narula, S. Tsimikas, J. Narula, J. Nucl. Med. 2009, 50, 612–
617.
[16] J. F. Fisher, S. Mobashery, Cancer Metastasis Rev. 2006, 25, 115–136.
[17] E. Nuti, T. Tuccinardi, A. Rossello, Curr. Pharm. Des. 2007, 13, 2087–
2100.
[47] A. P. Jeavons, R. A. Chandler, C. A. R. Dettmar, IEEE Trans. Nucl. Sci. 1999,
46, 468–473.
[48] K. P. Schꢀfers, A. J. Reader, M. Kriens, C. Knoess, O. Schober, M. Schꢀfers,
J. Nucl. Med. 2005, 46, 996–1004.
[49] J. Missimer, Z. Madi, M. Honer, C. Keller, P. A. Schubiger, S.-M. Ametamey,
Phys. Med. Biol. 2004, 49, 2069–2081.
[18] E. Breuer, J. Frant, R. Reich, Expert Opin. Ther. Pat. 2005, 15, 253–269.
[19] C. A. Corbitt, J. Lin, M. L. Lindsey, Recent Pat. Anti-Cancer Drug Discovery
2007, 2, 135–142.
[20] F. Grams, H. Brandstetter, S. D’Alo, D. Geppert, H. W. Krell, H. Leinert, V.
Livi, E. Menta, A. Oliva, G. Zimmermann, Biol. Chem. 2001, 382, 1277–
1285.
[50] A. J. Reader, S. Ally, F. Bakatselos, R. Manavaki, R. J. Walledge, A. P. Jea-
vons, P. J. Julyan, Z. Sha, D. L. Hastings, J. Zweit, IEEE Trans. Nucl. Sci.
2002, 49, 693–699.
[51] S. Biedenstein, M. Schꢀfers, L. Stegger, T. Kuwert, O. Schober, Eur. J.
Nucl. Med. 1999, 26, 201–207.
[21] H. Brandstetter, F. Grams, D. Glitz, A. Lang, R. Huber, W. Bode, H. W.
Krell, R. A. Engh, J. Biol. Chem. 2001, 276, 17405–17412.
[22] L. H. Foley, R. Palermo, P. Dunten, Bioorg. Med. Chem. Lett. 2001, 11,
969–972.
[23] E. Maquoi, N. E. Sounni, L. Devy, F. Olivier, F. Frankenne, H. W. Krell, F.
Grams, J. M. Foidart, A. Noꢆl, Clin. Cancer Res. 2004, 10, 4038–4047.
[24] B. Fingleton, Expert Opin. Ther. Targets 2003, 7, 385–397.
[25] H.-J. Breyholz, M. Schꢀfers, S. Wagner, C. Hçltke, A. Faust, H. Rabeneck,
B. Levkau, O. Schober, K. Kopka, J. Med. Chem. 2005, 48, 3400–3409.
[26] A. Faust, B. Waschkau, J. Waldeck, C. Hçltke, H. J. Breyholz, S. Wagner, K.
Kopka, W. Heindel, M. Schꢀfers, C. Bremer, Bioconjugate Chem. 2008, 19,
1001–1008.
Received: January 13, 2010
Revised: March 5, 2010
Published online on April 1, 2010
ChemMedChem 2010, 5, 777 – 789
ꢁ 2010 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
www.chemmedchem.org
789