A dual inhibitor of matrix metalloproteinases and a disintegrin and metalloproteinases, [18F]FB-ML5, as a molecular probe for non-invasive MMP/ADAM-targeted imaging

Article abstract of DOI:10.1016/j.bmc.2014.11.013

Background Numerous clinical studies have shown a correlation between increased matrix metalloproteinase (MMP)/a disintegrin and metalloproteinase (ADAM) activity and poor outcome of cancer. Various MMP inhibitors (MMPIs) have been developed for therapeutic purposes in oncology. In addition, molecular imaging of MMP/ADAM levels in vivo would allow the diagnosis of tumors. We selected the dual inhibitor of MMPs and ADAMs, ML5, which is a hydroxamate-based inhibitor with affinities for many MMPs and ADAMs. ML5 was radiolabelled with ¹⁸F and the newly obtained radiolabelled inhibitor was evaluated in vitro and in vivo. Materials and methods ML5 was radiolabelled by direct acylation with N-succinimidyl-4-[¹⁸F]fluorobenzoate ([¹⁸F]SFB) for PET (positron emission tomography). The resulting radiotracer [¹⁸F]FB-ML5 was evaluated in vitro in human bronchial epithelium 16HBE cells and breast cancer MCF-7 cells. The non-radioactive probe FB-ML5 and native ML5 were tested in a fluorogenic inhibition assay against MMP-2, -9, -12 and ADAM-17. The in vivo kinetics of [¹⁸F]FB-ML5 were examined in a HT1080 tumor-bearing mouse model. Specificity of probe binding was examined by co-injection of 0 or 2.5 mg/kg ML5. Results ML5 and FB-ML5 showed high affinity for MMP-2, -9, -12 and ADAM-17; indeed IC₅₀ values were respectively 7.4 ± 2.0, 19.5 ± 2.8, 2.0 ± 0.2 and 5.7 ± 2.2 nM and 12.5 ± 3.1, 31.5 ± 13.7, 138.0 ± 10.9 and 24.7 ± 2.8 nM. Radiochemical yield of HPLC-purified [¹⁸F]FB-ML5 was 13-16% (corrected for decay). Cellular binding of [¹⁸F]FB-ML5 was reduced by 36.6% and 27.5% in MCF-7 and 16HBE cells, respectively, after co-incubation with 10 μM of ML5. In microPET scans, HT1080 tumors exhibited a low and homogeneous uptake of the tracer. Tumors of mice injected with [¹⁸F]FB-ML5 showed a SUV_mean of 0.145 ± 0.064 (n = 6) which decreased to 0.041 ± 0.027 (n = 6) after target blocking (p < 0.05). Ex vivo biodistribution showed a rapid excretion through the kidneys and the liver. Metabolite assays indicated that the parent tracer represented 23.2 ± 7.3% (n = 2) of total radioactivity in plasma, at 90 min post injection (p.i.). Conclusion The nanomolar affinity MMP/ADAM inhibitor ML5 was successfully labelled with ¹⁸F. [¹⁸F]FB-ML5 demonstrated rather low binding in ADAM-17 overexpressing cell lines. [¹⁸F]FB-ML5 uptake showed significant reduction in the HT1080 tumor in vivo after co-injection of ML5. [¹⁸F]FB-ML5 may be suitable for the visualization/quantification of diseases overexpressing simultaneously MMPs and ADAMs.

Full text of DOI:10.1016/j.bmc.2014.11.013

Bioorganic & Medicinal Chemistry 23 (2015) 192–202
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
A dual inhibitor of matrix metalloproteinases and a disintegrin
and metalloproteinases, [18F]FB-ML5, as a molecular probe for
non-invasive MMP/ADAM-targeted imaging
Nathalie Matusiak a, Riccardo Castelli b, Adriaan W. Tuin b, Herman S. Overkleeft b, Rosalina Wisastra c,
Frank J. Dekker c, Laurette M. Prély d, Rainer P. M. Bischoff d, Aren van Waarde a, Rudi A. J. O. Dierckx a,
Philip H. Elsinga a,
⇑
a Department of Nuclear Medicine and Molecular Imaging, University Medical Center Groningen, University of Groningen, Hanzeplein 1, 9700 RB Groningen, The Netherlands
b Leiden Institute of Chemistry, Leiden University, Leiden, The Netherlands
c Department of Pharmaceutical Gene Modulation, Groningen Research Institute of Pharmacy, University of Groningen, Groningen, The Netherlands
d Department of Analytical Biochemistry, Groningen Research Institute of Pharmacy, University of Groningen, Groningen, The Netherlands
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 22 July 2014
Revised 6 November 2014
Accepted 6 November 2014
Available online 15 November 2014
Background: Numerous clinical studies have shown a correlation between increased matrix metallopro-
teinase (MMP)/a disintegrin and metalloproteinase (ADAM) activity and poor outcome of cancer. Various
MMP inhibitors (MMPIs) have been developed for therapeutic purposes in oncology. In addition, molec-
ular imaging of MMP/ADAM levels in vivo would allow the diagnosis of tumors. We selected the dual
inhibitor of MMPs and ADAMs, ML5, which is a hydroxamate-based inhibitor with affinities for many
MMPs and ADAMs. ML5 was radiolabelled with 18F and the newly obtained radiolabelled inhibitor was
evaluated in vitro and in vivo.
Keywords:
MMPs
ADAMs
MMP/ADAM inhibitor
Hydroxamate
Materials and methods: ML5 was radiolabelled by direct acylation with N-succinimidyl-4-[18F]fluor-
obenzoate ([18F]SFB) for PET (positron emission tomography). The resulting radiotracer [18F]FB-ML5
was evaluated in vitro in human bronchial epithelium 16HBE cells and breast cancer MCF-7 cells. The
non-radioactive probe FB-ML5 and native ML5 were tested in a fluorogenic inhibition assay against
MMP-2, -9, -12 and ADAM-17. The in vivo kinetics of [18F]FB-ML5 were examined in a HT1080 tumor-
bearing mouse model. Specificity of probe binding was examined by co-injection of 0 or 2.5 mg/kg ML5.
Results: ML5 and FB-ML5 showed high affinity for MMP-2, -9, -12 and ADAM-17; indeed IC50 values were
respectively 7.4 2.0, 19.5 2.8, 2.0 0.2 and 5.7 2.2 nM and 12.5 3.1, 31.5 13.7, 138.0 10.9 and
24.7 2.8 nM. Radiochemical yield of HPLC-purified [18F]FB-ML5 was 13–16% (corrected for decay).
Cellular binding of [18F]FB-ML5 was reduced by 36.6% and 27.5% in MCF-7 and 16HBE cells, respectively,
PET
HT1080 xenograft mouse model
after co-incubation with 10 lM of ML5. In microPET scans, HT1080 tumors exhibited a low and
homogeneous uptake of the tracer. Tumors of mice injected with [18F]FB-ML5 showed a SUVmean of
0.145 0.064 (n = 6) which decreased to 0.041 0.027 (n = 6) after target blocking (p < 0.05). Ex vivo
biodistribution showed a rapid excretion through the kidneys and the liver. Metabolite assays indicated
that the parent tracer represented 23.2 7.3% (n = 2) of total radioactivity in plasma, at 90 min post
injection (p.i.).
Conclusion: The nanomolar affinity MMP/ADAM inhibitor ML5 was successfully labelled with 18F. [18F]FB-
ML5 demonstrated rather low binding in ADAM-17 overexpressing cell lines. [18F]FB-ML5 uptake showed
significant reduction in the HT1080 tumor in vivo after co-injection of ML5. [18F]FB-ML5 may be suitable
for the visualization/quantification of diseases overexpressing simultaneously MMPs and ADAMs.
Ó 2014 Elsevier Ltd. All rights reserved.
Abbreviations: ACN, acetonitrile; AcOH, acetic acid; ADAM, a disintegrin and metalloproteinase; CD, catalytic domain; DCC, N,N0-dicyclohexylcarbodiimide; DCM,
dichloromethane; DMF, N,N-dimethylformamide; ECM, extracellular matrix; EOB, end of bombardment; EOS, end of synthesis; EtOAc, ethyl acetate; FBA, 4-fluorobenzoic
acid; MMP, matrix metalloproteinase; MMPI, matrix metalloproteinase inhibitor; NHS, N-hydroxysuccinimide; PDB, Protein Data Bank; PET, positron emission tomography;
p.i., post injection; PMA, phorbol 12-myristate 13-acetate; SFB, N-succinimidyl-4-fluorobenzoate; SPE, solid phase extraction; SPPS, solid phase peptide synthesis; TIMP,
tissue inhibitor of matrix metalloproteinase; TNF, tumor necrosis factor; TSTU, O-(N-succinimidyl)-N,N,N’,N’-tetramethyluronium tetrafluoroborate; ZBG, zinc binding group.
⇑
Corresponding author. Tel.: +31 50 3613247; fax: +31 50 3611687.
E-mail address: p.h.elsinga@umcg.nl (P.H. Elsinga).
http://dx.doi.org/10.1016/j.bmc.2014.11.013
0968-0896/Ó 2014 Elsevier Ltd. All rights reserved.

N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
193
on both sides of the zinc binding group (ZBG) or a left hand-side
inhibitor where the amino acids are present on only the left side
have shown modest inhibition.29
1. Introduction
The metzincins are a superfamily of multi-domain zinc(II)-
dependent endopeptidases which comprise the matrix metallopro-
teinases (MMPs) and the a disintegrin and metalloproteinases
(ADAMs).1,2 These proteins3 possess similar domain structure for
the signal peptide, the pro-domain and the catalytic domain. In
addition, the catalytic activity exhibited by the metalloproteinase
domain of the ADAMs is highly homologous with the MMP
catalytic domain. MMPs and ADAMs are expressed as inactive
pro-proteins and require enzymatic removal of the pro-domain
to become proteolytically active.4 In addition, active MMPs and
ADAMs can be inhibited by endogenous inhibitors such as tissue
inhibitors of metalloproteinases (TIMPs).5
The inhibitor ML5 consists of a succinylhydroxamate attached
N-terminally to a phenylalanine—lysine dipeptide (Fig. 1). The
hydroxamate functions as a Zn-chelating group whereas the
dipeptide part serves both as additional recognition element and
as substitution handle (through the lysine residue). An isobutyl
group was chosen as a P10 substituent, a benzyl group for the P20
substituent and a lysine for the P30 substituent to fill, respectively,
the S10, S20 and S30 pocket of the active site of MMPs/ADAMs. The
choice of a relatively small aliphatic substituent (isobutyl) as a
P10 residue results in a broad-spectrum affinity of ML5 for MMPs
and ADAMs. The benzyl group, by its hydrophobicity, induces more
affinity for MMPs and ADAMs. The need to incorporate a substitu-
ent in the third loop is still under debate considering that the S30
pocket is an ill-defined solvent-exposed pocket. However, a lysine
as a P30 substituent can be employed for other purposes such as
fluorescent labelling, biotinylation or radiolabelling through the
The diversity of biological functions implies that MMPs and
ADAMs are simultaneously involved in a wide range of physiologi-
cal and pathological processes.6 To date, MMP-2 (gelatinase A,
72 kDa type IV collagenase), MMP-9 (gelatinase B, 92 kDa type IV
collagenase), MMP-12 (macrophage metalloelastase) and ADAM-
17 (TNF alpha converting enzyme) have attracted the most interest.
MMP-2 and MMP-9 are the most extensively studied metallopro-
teinases because dysregulation of MMP-2 and MMP-9 is related
to poor prognosis in oncology.7–9 Many publications involving
increased gelatinase expression and activity have been reported
for instance in brain tumors,10 breast cancer,11 lung cancer,12 skin
cancer.13 MMP-2 and MMP-9 are also upregulated in some inflam-
matory diseases.14,15 MMP-12, as well as MMP-8 and MMP-14,
exert anti-cancer effects. More particularly, MMP-12 is the primary
protease responsible for the proteolytic liberation of angiostatin
from plasminogen. Angiostatin is a protein with anti-angiogenic
properties by the selective inhibition of endothelial prolifera-
tion.16–18 Finally, ADAM-17 or TNF alpha converting enzyme (TACE)
is definitely the ADAM which has attracted the most research since
the discovery as the protease responsible for the release of soluble
e-amino group.
2.2. Synthesis of ML5
The overall yield of the synthesis of the building block 9 (over 7
steps) was around 16%. The coupling by SPPS leading to ML5 (after
HPLC purification) was obtained in a yield of around 5–8%.
2.3. Strategy for radiolabelling ML5
Our strategy was to radiolabel the free amine of the lysine of
ML5 with the positron-emitter fluorine-18 (T1/2 = 109.8 min) for
PET. The [18F]labelling of small molecules or peptides proceeds
generally by conjugation of [18F]prosthetic groups by [18F]fluor-
oalkylation, [18F]fluoroacylation or [18F]fluoroamidation, or via
click chemistry. To avoid the need of further modifications of
ML5, by the incorporation of an azido or an alkyno-moiety for
instance, we adopted the direct acylation in solution using N-succ-
inimidyl-4-[18F]fluorobenzoate ([18F]SFB). [18F]SFB was chosen as
the [18F]acylating agent because of its high in vivo stability and
labelling efficiency.30–33
TNF
numerous tumors,22 identically as MMP-2 and MMP-9. Besides, as
TNF is an important pro-inflammatory mediator, ADAM-17 activ-
a
from cells.19–21 ADAM-17 expression is upregulated in
a
ity has been implicated in many diseases involving inflammation
such as rheumatoid arthritis,23 Crohn’s disease24 and inflammatory
bowel disorder,25 similar as MMP-2 and MMP-9.
As both MMP and ADAM families have members which are
related to some inflammatory and tumor progression processes,
the ability to exploit possible synergic effects of ‘cross-reactivity’
of a dual inhibitor of these enzymes would be of significant interest
for the characterization of inflammatory lesions or tumors.
Positronemission tomography(PET) is a molecularimaging tech-
nique which produces a detailed image of biochemical and physio-
logical processes in living organisms by injection of a radiotracer.26
To achieve detection of the proteolytic activity of MMPs and
ADAMs, we selected the peptidomimetic inhibitor ML5, synthesized
originally by Leeuwenburgh et al.,27,28 which is a broad-spectrum
MMP/ADAM inhibitor. We radiolabelled ML5 with N-succinimidyl-
4-[18F]fluorobenzoate ([18F]SFB) for PET. The binding of the obtained
radiotracer [18F]FB-ML5 was assessed in vitro, using human bron-
chial epithelium 16HBE or breast cancer MCF-7 cells. Finally,
2.4. Molecular modeling of ML5 and FB-ML5
The docking of ML5 and FB-ML5 into MMP-2, MMP-9, MMP-12
and ADAM-17 are reported in the Supplementary information.
According to the docking studies, the incorporation of a fluor-
obenzoyl substituent to the lysine moiety resulted to a moderate
loss of affinity of FB-ML5 compared to the parent compound ML5
into MMP-2, MMP-9, MMP-12 and ADAM-17 (Table 1).
S3'
NH2
S1'
ZBG
[
18F]FB-ML5 was tested in vivo in mice bearing HT1080 tumors.
O
O
H
N
HO
NH2
2. Results and discussion
N
N
H
H
O
O
2.1. Design of ML5—(R)-N1-((S)-1-(((S)-1,6-diamino-1-
oxohexan-2-yl)amino)-1-oxo-3-phenylpropan-2-yl)-N4-
hydroxy-2-isobutylsuccinimide
ML5
S2'
In the design of ML5, a right hand-side inhibitor was selected.
Indeed, a combined inhibitor where amino acid residues are linked
Figure 1. Structure and design of ML5.

194
N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
Table 1
The IC50 of ML5, FB-ML5 and 2FB-ML5 against MMP-2, -9, -12
and ADAM-17 are reported in (Table 2). The incorporation of the
fluorobenzoyl moiety did not result in a significant modification
of the affinity for MMP-2, -9 and ADAM-17. However, FB-ML5
showed an almost 100-fold loss of affinity for MMP-12 compared
to ML5. 2FB-ML5 exhibited affinity in the low micromolar range.
The incorporation of a fluorobenzoyl moiety on the hydroxamic
acid decreased considerably the inhibition potency of ML5.
MolDock scores of ML5 and FB-ML5 with MMP-2, MMP-9, MMP-12 and ADAM-17
MMP-2
MMP-9
MMP-12
ADAM-17
ML5
FB-ML5
ꢀ133.4
ꢀ126.1
ꢀ115.6
ꢀ118.9
ꢀ128.0
ꢀ118.6
ꢀ137.2
ꢀ112.9
2.5. Synthesis of the acylating agent SFB
2.8. Radiolabelling of ML5 with [18F]SFB
N-Succinimidyl 4-fluorobenzoate (SFB) was prepared for further
coupling with ML5 and for HPLC standard during the optimisation
of the purification of [18F]FB-ML5. This amide was synthesized
from 4-fluorobenzoic acid (FBA) according to the literature in
two different ways. In the first approach,33 4-fluorobenzoic acid
was activated with N,N0-dicyclohexylcarbodiimide (DCC) and the
resulting O-acylisourea was treated with N-hydroxysuccinimide
(NHS) to give SFB. In the second method,31 4-fluorobenzoic acid
was treated with sodium carbonate and then with O-(N-succinim-
idyl)-N,N,N0,N0-tetramethyluronium tetrafluoroborate (TSTU). Both
methods led to SFB in good yields of 79% and 85%, respectively.
The prosthetic group [18F]SFB was prepared from 4-trimethy-
lammonium ethylbenzoate and was synthesized in a three step
procedure as described in the literature30–33 with some modifica-
tions. The automated system produced [18F]SFB with a good
radiochemical yield (about 35%) with a synthesis time of about
90 min from the end of bombardment (EOB). The radiochemical
purity after SPE-formulation was always more than 95%, as
determined by HPLC.
The radiochemical yield of HPLC-purified [18F]FB-ML5 was 13–
16% based on [18F]SFB (corrected for decay), the specific activity
2.6. Synthesis of the reference compound FB-ML5
was 41–66 GBq/lmol at the end of synthesis (EOS) and the
radiochemical purity >95%. The total synthesis time of [18F]FB-
ML5 was about 3 h (EOB).
MMPI FB-ML5 was prepared as HPLC standard and for in vitro
fluorogenic inhibition assays. FB-ML5 was prepared either by SPPS
or by acylation in solution using SFB (Scheme 3). The coupling in
solution using an excess of SFB led to the formation of the doubly
acylated ML5: 2FB-ML5 (Fig. 2).
2.9. In vitro stability of [18F]FB-ML5 in human plasma and saline
After 1 h and 3 h of incubation, 99% of the radioactivity still
corresponded to the intact tracer both in human plasma and saline.
This indicates that [18F]FB-ML5 is highly stable in vitro.
2.7. In vitro evaluation of ML5, FB-ML5 and 2FB-ML5 in a
fluorogenic inhibition assay
2.10. Octanol/water partition coefficient of [18F]FB-ML5
The affinity of ML5, FB-ML5 and 2FB-ML5 to four selected
recombinant metzincins was determined in a fluorogenic inhibi-
tion assay against MMP-2, MMP-9, MMP-12 and ADAM-17.
The measured logP of [18F]FB-ML5 was 1.11 0.01, indicating
intermediate lipophilicity.
O
O
O
O
O
O
a
b
HN
N
N
O
Cl
O
O
3
60 %
90 %
O
Bn
2
Bn
Bn
1
OtBu
81 %
4
c
O
O
6
d
e
O
OBn
OBn
TBSO
OBn
f
quant.
O
N
5
O
Boc
O
7
OtBu
78 %
OH
g
O
F
F
O
h
TBSO
O
F
F
F
TBSO
OH
N
N
Boc
85 % over 2 steps
Boc
O
O
9
8
Scheme 1. Synthesis of the building block 9. (a) 1.2 eq. n-BuLi, THF, 0 °C ; (b) 1.2 eq. LiHMDS, 1.8 eq. tBu-bromoacetate, THF, ꢀ78 °C; (c) 2 eq. BnOH, 1.2 eq. n-BuLi, THF, 0 °C;
(d) TFA, DCM, RT; (e) 5 eq. (COCl)2, cat. DMF, DCM, 0 °C; (f) 1.1 eq. TBSONHBoc, 2 eq. DMAP, ACN, 0 °C; (g) H2, Pd/C, MeOH; (h) 1 eq. PFP-OH, 1 eq. EDC, DCM, RT.

N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
195
NH2
NH2
O
(1) 20 % piperidine / DMF
(2) Fmoc-Phe-OH (5 eq), HCTU (5 eq), DIPEA (10 eq), NMP
O
O
H
N
Fmoc
H
N
HO
N
H
Rink
N
H
N
H
(3) 20% piperidine / DMF
10
O
O
Building block 9 (5 eq), DIPEA (2 eq), NMP
(4)
(5) 95 % TFA / H2O
ML5
NHBoc
Scheme 2. Solid phase peptide synthesis towards ML5.
F
O
NH2
NH2
NH
O
O
O
O
O
O
H
N
H
N
i
HO
NH2
HO
F
N
H
N
H
N
N
O
N
H
H
O
O
O
O
SFB
O
ML5
FB-ML5
(i) ACN/phosphate buffer 0.01M pH=8.5 (1/1) (v/v), 50°C, 30 min
Scheme 3. Synthesis of FB-ML5 (with 1 equiv of SFB).
found to not express MT1-MMP (MMP-14).34 In addition, Köhr-
mann et al.35 found that MCF-7 cells express inactive MMP-14,
inactive MMP-15 (MT2-MMP) but not their active forms. They also
found that MCF-7 cells do not express MMP-24 (MT5-MMP). How-
ever, our aim was not to use a subtype specific cell line but to use
cells as a tool to estimate overall specific binding to MMPs/ADAMs.
The association kinetics of [18F]FB-ML5 was performed with
MCF-7 cells by adding the radioligand at 0, 1, 2, 5, 17, 27 and
57 min (Fig. 3). At 57 min, 6.85 times more radioactivity were
bound to the cells than at 0 min. The half-life of association was
7.85 min.
F
O
NH
F
O
O
H
N
NH2
O
N
H
N
H
O
O
O
[
18F]FB-ML5 (Fig. 4) showed binding to MCF-7 cells which could
be competed with non-radioactive ML5. The obtained reduction of
2FB-ML5
the cellular binding was 36.6% by using 10 lM of ML5. PMA stim-
ulation did not increase the level of binding of [18F]FB-ML5 in
agreement with previous work of our group.
Figure 2. Structure of 2FB-ML5.
Preliminary evaluation of [18F]FB-ML5 with 16HBE cells (Fig. 5)
demonstrated binding to the cells, with a binding decrease of 27.5%
2.11. In vitro evaluation of [18F]FB-ML5
after blocking with 10
lM of ML5. The reduction of uptake with
10 M of ML5 was expected to be higher, which may suggest that
l
Contrary to the majority of MMPs which are largely excreted
proteins, ADAMs are membrane bound proteins. Therefore, the
in vitro studies of [18F]FB-ML5 were focused on ADAMs. We quan-
tified the binding of the latter analogue in vitro on two human cell
lines. Two cell lines, which are both overexpressing ADAM-17,
were selected: the human breast cancer cell line MCF-7 and the
human bronchial epithelium cell line 16HBE. MCF-7 cells were
the non-specific binding of the tracer in 16HBE cells is apparently
relatively high. As the lysine residue is solvent exposed and is not
involved in the interaction with the MMP active site (Fig. 8S), the
loss of affinity of [18F]FB-ML5 compared to the parent compound
ML5 was not anticipated. PMA and LPS stimulation did not increase
the level of binding of [18F]FB-ML5.
Table 2
IC50 values of ML5, FB-ML5 and 2FB-ML5 with MMP-2, MMP-9, MMP-12 and ADAM-17
IC50
MMP-2
MMP-9
MMP-12
ADAM-17
ML5
FB-ML5
2FB-ML5
7.4 2.0 nM
12.5 3.1 nM
19.5 2.8 nM
31.5 13.7 nM
2.0 0.2 nM
138.0 10.9 nM
nd
5.7 2.2 nM
24.7 2.8 nM
3.26 1.70 lM
1.49 0.08
l
M
2.42 0.40
lM

196
N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
2.12. Preclinical evaluation of [18F]FB-ML5 in a HT1080
xenograft mouse model
150
Control
PMA + LPS
10 uM ML5
PMA + LPS + 10 uM ML5
100
50
0
The radiotracer [18F]FB-ML5 was evaluated in a mouse model of
cancer overexpressing many matrix metalloproteinases: the
HT1080 xenograft mouse model. This model was already used for
the evaluation of other MMP probes in fluorescent imaging
notably.36
*
Plasma samples at 90 min p.i. were analysed for parent and
metabolite levels by HPLC. Metabolite assays showed that the
parent tracer represented 23.2 7.3% (n = 2) of total radioactivity
in plasma in control mice, at 90 min p.i. (Table 3). The metabolism
of [18F]FB-ML5 was relatively fast in plasma. According to HPLC,
only more polar radio-metabolite(s) was observed (Fig. 9S). This
suggests that the radiometabolite(s) is structurally different from
Figure 5. Specificity study of [18F]FB-ML5 with 16HBE cells.
to muscle ratio. Therefore, the binding of [18F]FB-ML5 in the
HT1080 xenograft mouse model was target mediated.
For the SUVmean
, the following organs were statistically
different between the control and the block mice: the bone
[
18F]FB-ML5 and therefore probably unlikely to retain affinity for
0.041 0.007 versus 0.017 0.001
0.111 0.020 versus 0.043 0.013
(
(
⁄⁄⁄p < 0.0001), the heart
⁄⁄⁄p < 0.0001), the large
active MMPs/ADAMs.
The microPET/CT images (Fig. 6) demonstrated a homogeneous
uptake throughout the tumor volume, suggesting tracer binding
was not only to membrane-bound ADAMs but also to extracellular
MMPs. Autoradiography of a tumor slice confirmed the regular
uptake on the tumor (Fig. 10S). A high kidney uptake was also
observed in the microPET/CT images.
The radioactivity uptake in the selected tissues (SUVmean data
presented in mean SD) is reported in (Fig. 7). The SUVmean norma-
lised to plasma and SUVmean normalised to muscle are reported in
(Figs. 8 and 9). The uptake of the radioligand in the tumor substan-
tially (⁄⁄p = 0.0043) decreased after co-injection of non-radioactive
intestine 1.800 0.424 versus 2.750 0.636 (⁄p = 0.0124), the liver
4.582 0.989 versus 1.995 0.148
(
⁄⁄⁄p < 0.0001), the lung
0.254 0.072 versus 0.106 0.045 (⁄⁄p = 0.0016) and the plasma
0.238 0.039 versus 0.185 0.020 (⁄p = 0.0142).
For the tissue–plasma ratio, the bone 0.179 0.060 versus
0.090 0.006
0.235 0.093
(
⁄⁄p = 0.0047), the heart 0.482 0.163 versus
(
⁄⁄p = 0.0091), the large intestine 7.830 3.069
versus 14.758 1.870 (⁄⁄⁄p = 0.0008) and the liver 19.208 1.017
versus 10.883 1.959 (⁄⁄⁄p < 0.0001) were statistically different.
Finally, for the tissue–muscle ratio, the bone 0.769 0.176 ver-
sus 0.478 0.136 (⁄⁄p = 0.0094), the heart 2.069 0.462 versus
1.178 0.065 (⁄⁄⁄p = 0.0009), the large intestine 33.065 5.514
versus 81.219 37.415 (⁄p = 0.0109) and the lung 4.616 0.546
versus 2.897 0.559 (⁄⁄⁄p = 0.0003) were statistically different.
ML5 from
a SUVmean of 0.145 0.064 in control mice to
0.041 0.027 in block mice at 90 min p.i. The tumor to plasma
ratio was 0.597 0.170 versus 0.231 0.171 (⁄⁄p = 0.0039) and
the tumor to muscle ratio was 3.035 2.329 versus 1.084 0.487
(p = 0.0724) at 90 min p.i. The change of tumor to plasma ratio
was statistically significant in contrast to the change of the tumor
Table 3
Percentage of parent compound in plasma from two control mice at 90 min p.i of
[
18F]FB-ML5
Association of [18F]FB-ML5 to MCF-7 cells
Mouse #
% of parent [18F]FB-ML5
800
Mouse # 1
Mouse # 2
28.3
18.1
600
400
200
0
0
5
10 15 20 25 30 35 40 45 50 55 60
Incubation time (min)
Figure 3. Time course of [18F]FB-ML5 binding to MCF-7 cells.
150
Control
PMA
100 nM ML5
PMA + 100 nM ML5
10 uM ML5
100
50
0
*
*
***
PMA + 10 uM ML5
Figure 6. In vivo [18F]FB-ML5 microPET/CT images of a HT1080 tumor bearing
mouse shown in coronal (left) and transaxial (right) views. The microPET images
correspond to the sum of all the frames from 2 to 90 min p.i. of [18F]FB-ML5.
Figure 4. Specificity study of [18F]FB-ML5 with MCF-7 cells.

N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
197
SUVmean
Tissue / Muscle ratio
Control
Block
150
125
10
8
*
Control
Block
100
75
6
50
4
*
25
***
2
10
8
0.5
0.4
0.3
0.2
0.1
0.0
6
4
***
*
2
**
***
**
**
0
***
***
t
r
e
a
s
ll
e
d
n
e
h
ey
testin
as
ne
Bo
ng
Pa
ar
He
ain
Br
in
est
ve
Li
e
l
ac
mor
Tu
sm
re
Lu
t
r
s
e
d
Bloo
le
ng
ls
d ce
ey
a
nc
t
n
om
Sp
ne
Bo
en
le
ch
Tu
Kidn
e In
ar
He
Pla
ain
Br
ine
st
ma
as
ve
Li
el
l I
mor
sc
re
St
oo
Lu
l I
c
Bloo
l
e
d c
oo
n
testin
oma
Sp
Kidn
Mu
g
r
Pl
n
d bl
St
a
Pa
e Int
Sma
L
Re
rg
La
d bl
Re
Smal
Figure 9. Tissue/muscle ratio of tumor-bearing mice scanned with [18F]FB-ML5 and
tumor-bearing mice scanned with [18F]FB-ML5 and co-injection of 2.5 mg/kg of
ML5, at 90 min p.i. of [18F]FB-ML5 ML5. Bars represent average and error bars SD,
n = 6 for each group, ⁄p < 0.05, ⁄⁄p < 0.01 and ⁄⁄⁄p < 0.001.
Figure 7. Ex vivo biodistribution data of tumor-bearing mice scanned with [18F]FB-
ML5 and tumor-bearing mice scanned with [18F]FB-ML5 and co-injection of 2.5 mg/
kg of ML5, at 90 min p.i. of [18F]FB-ML5 ML5. Bars represent average and error
bars SD, n = 6 for each group, ⁄p < 0.05, ⁄⁄p < 0.01 and ⁄⁄⁄p < 0.001.
Tissue / Plasma ratio
Control
Block
50
40
30
20
10
***
***
2.0
1.5
1.0
0.5
0.0
Figure 10. Time-activity curve of tumor-bearing mice scanned with [18F]FB-ML5
and tumor-bearing mice scanned with [18F]FB-ML5 and co-injection of 2.5 mg/kg of
ML5, from 2 to 90 min p.i. of [18F]FB-ML5 ML5. Points represent average and error
bars SD, n = 6 for each group.
**
**
**
t
r
e
s
s
l
l
e
n
d
n
ey
ne
Bo
ch
ma
ng
Lu
ar
He
ain
Br
ine
est
i
ea
e
ee
iver
mo
Tu
scl
l
L
c
u
Bloo
est
ncr
t
o
Sp
Kidn
M
od
n
I
St
o
l
Pa
l
e Int
l
The average time-activity curves in the tumor of the control and
block animals are depicted in (Fig. 10). PET-SUVmean showed a sig-
nificant decrease (p = 0.0406) of the tracer accumulation in the
tumor: 0.125 0.087 versus 0.037 0.029 at 90 min p.i. [18F]FB-
ML5 demonstrated a relatively fast wash out in the tumor.
A similar MMPI based on a peptidomimetic scaffold, the radio-
labelled Marimastat-ArB[18F]F3, was preliminary evaluated in a
67NR breast tumor mice.37 A low uptake in the tumor was
obtained as well and a significant reduction was obtained after
pre-treatment with a non-radioactive inhibitor, which suggests
that our in vivo results are quite satisfactory.
Finally, based on the multifactorial nature of numerous disease
processes, the ability of a tracer to simultaneously target two path-
ways associated to the same disease might represent an asset in
PET. As a result, the dual MMP/ADAM inhibitor [18F]FB-ML5 might
afford a more efficient approach to overcome the complex pro-
cesses of cancer.
b
rg
La
d
ma
e
S
R
Figure 8. Tissue/plasma ratio of tumor-bearing mice scanned with [18F]FB-ML5 and
tumor-bearing mice scanned with [18F]FB-ML5 and co-injection of 2.5 mg/kg of
ML5, at 90 min p.i. of [18F]FB-ML5 ML5. Bars represent average and error bars SD,
n = 6 for each group, ⁄⁄p < 0.01 and ⁄⁄⁄p < 0.001.
A low uptake in the bone was obtained which suggests no deflu-
orination of the tracer. High uptakes of the radioactivity in the kid-
neys and to a lesser extent in the liver 90 min p.i. were obtained.
This is probably due to the excretion of the radiotracer and radiom-
etabolites. The amount of activity (i.e., tracer and metabolites)
excreted by the liver into the small and large intestines was very
high. Indeed, for the control mice, the SUVmean obtained in the
liver, small intestine and large intestine were, respectively:
4.582 0.989, 5.207 4.058 and 1.800 0.424.

198
N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
Virtual Docker (www.molegro.com). Docking solutions were
3. Conclusions
selected based on the MOLDOCKSCORE and the docking solutions
were evaluated manually, followed by energy minimization of
the ligand.
The MMPI ML5 was successfully radiolabelled with [18F]SFB.
[
[
18F]FB-ML5 showed rather low binding in MCF-7 and 16 HBE cells.
18F]FB-ML5 retention showed significant reduction in the HT1080
tumor after co-injection of ML5. [18F]FB-ML5 may be suitable for
the visualization/quantification of pathologies overexpressing
simultaneously MMPs and ADAMs.
4.4. Synthesis of ML5
The peptidyl MMP inhibitor ML5 (Fig. 1) was synthesized fol-
lowing slight modifications of the literature procedures.27,28
Briefly, the Evan’s template 1 was prepared in a 3-step reaction
4. Materials and methods
4.1. General
from the Boc-(D)-phenylalanine. The latter analogue was treated
with triethylamine and ethyl chloroformate in order to get the cor-
responding acid anhydride which was then reduced with sodium
borohydride to give the corresponding alcohol. After treatment
with thionyl chloride, the resulting tert-Bu-ester was transformed
in its acyl chloride, which was then intramolecularly attacked by
the primary alcohol to obtain the Evan’s template 1. 4-Methyl-pen-
tanoyl chloride 2 was obtained after treatment of 4-methyl-penta-
noic acid with thionyl chloride. N-Boc-O-TBS-hydroxylamine was
prepared in a 2-step procedure from the hydroxylamine hydro-
chloride. The amino group of the latter analogue was protected
with di-tert-butyl dicarbonate. The resulting compound was trea-
ted with tert-butyl dimethylsilyl chloride to give N-Boc-O-TBS-
hydroxylamine. The building block 9 was prepared in 7 steps
(Scheme 1), and started by deprotonation of the Evan’s template
1 with n-butyllithium, followed by acylation with 4-methyl-penta-
noyl chloride 2 to give 3. Removal of the acidic alpha-proton with
lithium bis(trimethylsilyl)amide followed by reaction of the result-
ing enolate with tert-butyl-bromoacetate gave 4. Treatment of 4
with lithium benzyl alcoholate led to the benzyl ester 5. The tert-
butyl-ester 5 was then deprotected to give the intermediate 6.
The obtained carboxylic acid 6 was transformed into the corre-
sponding acyl chloride, which was treated with N-Boc-O-TBS-
hydroxylamine to give 7. The benzyl ester in 7 was hydrolyzed
by catalytic hydrogenation to the acid 8. The carboxylic acid 8,
after activation with N-(3-dimethylaminopropyl)-N0-ethylcarbodi-
imide, was then acylated with pentafluorophenol to give the build-
ing block 9. Dipeptide 10 was synthesized on Rink amide resin
using standard protocols (Scheme 2). After deprotection of the
Fmoc group, the Fmoc-protected phenylalanine was incorporated.
Subsequent removal of the Fmoc was followed by the coupling of
the building block 9 under the agency of iPr2EtN. The resulting
compound was cleaved from the resin and simultaneously
deprotected with 95% aqueous TFA. The obtained compound was
purified and characterized by LCMS.
All chemicals, reagents, and solvents for the (radio)synthesis of
the compounds were analytical grade, purchased from commercial
suppliers (Aldrich, Fluka, Sigma and Merck) and were used without
further purification, unless otherwise specified. Solid phase extrac-
tion cartridges were obtained from Waters Chromatography Divi-
sion, Millipore Corporation (for SepPak Light Accell plus QMA
anion exchange cartridge and SepPak C18 plus cartridge) or Merck
(for LiChrolut EN cartridge and LiChrolut SCX cartridge). Flash
chromatography was performed on silica gel 60 (0.040–0.063,
Merck). All reactions were monitored by thin layer chromatogra-
phy on Merck F-254 silica gel plates. Detection of the compounds
on the TLC plates was performed with UV light (254 nm). 1H, 13C,
and 19F NMR spectra were recorded on a Varian AMX400 spec-
trometer (400, 100 and 200 MHz, respectively). Chemical shifts
were determined relative to the signal of the solvent, converted
to the TMS (tetramethylsilane) scale, and expressed in d units
(ppm) downfield from TMS (for chloroform-d: d 7.261 for 1H and
d 76.98 for 13C, for DMSO-d6: d 2.504 for 1H and d 39.98 for 13C
and for MeOH: d 3.312, 4.867 for 1H and d 47.84 for 13C). Data
are reported as follows: chemical shifts, multiplicity (s = singlet,
d = doublet, t = triplet, q = quartet, dd = doublet of doublets,
dt = doublet of triplets, td = triplet of doublets, m = multiplet,
br = broad), coupling constants (Hz), and integration. Mass spec-
trometry was recorded on an AEI-MS-902 mass spectrometer by
EI (70 eV) measurements. Radioactivity measurements for logP
determination, plasma/saline stability, biodistribution and metab-
olites were performed using an automated gammacounter (LKB
Wallac, Turku, Finland).
4.2. Statistical analysis
Calculations were performed using Excel 2007 (Microsoft) and
GraphPad prism 5.0 for Windows (GraphPad Software, San Diego,
USA). Results are expressed as mean SD. Comparisons between
different experimental groups were made using unpaired two-
sided student t-test. Data were considered statistically significant
when p values were smaller than 0.05.
4.5. Synthesis of the acylating agent SFB
4.5.1. Synthesis of SFB by the first approach
To a solution of 4-fluorobenzoic acid (1.0 g, 7.1 mmol) and N-
hydroxysuccinimide (0.9 g, 7.8 mmol) in 20 mL of DCM and 1 mL
of DMF at 0 °C was added dropwise a solution of N,N0-dicyclohex-
ylcarbodiimide (1.6 g, 7.8 mmol) in 10 mL of DCM at 0 °C. The
reaction was stirred overnight at room temperature. The reaction
was filtered to remove the N,N0-dicyclohexyl urea and the obtained
filtrate was concentrated under reduced pressure. The resulting
residue was partitioned between EtOAc and water. The organic
layer was washed with water (3 times), brine, dried over Na2SO4,
filtered and concentrated under reduced pressure to give the crude
product. The latter was recrystallized from EtOAc–hexane to afford
SFB (0.67 g) as a white solid.
4.3. Molecular modeling of ML5 and FB-ML5
The crystal structures of MMPs and ADAM-17 were down-
loaded from the Protein Data Bank (PDB) with MMP-2 (PDB code
1HOV), MMP-9 (PDB code 2OW1), MMP-12 (PDB code 1JK3) and
ADAM-17 (PDB code 2I47). All molecules were drawn using
ChemaxonMarvinSketch (www.chemaxon.com) and prepared
(structure recognition and protonation) using SPORES
(www.tcd.uni-konstanz.de/research/spores.php). Molecular dock-
ing simulations were performed using PLANTS v1.6.140,141. The
docking site center was determined by applying a constraint for
the hydroxamic group to be able to form a coordination with the
zinc in the active site. Fifteen poses were generated for each
compound and the docking results were analyzed using Molegro
4.5.2. Synthesis of SFB by the second approach
To a solution of 4-fluorobenzoic acid (0.1 g, 0.7 mmol) and
sodium carbonate (0.07 g, 0.66 mmol) in 3 mL of ACN, stirred for
5 min, was added TSTU (0.21 g, 0.7 mmol). The mixture was stirred

N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
199
for 1 h at 50 °C and was then filtered. The resulting filtrate was
diluted with 10 mL 1% acetic acid and was passed through a Lichro-
lut EN cartridge. The cartridge was washed with 10 mL 0.01%
AcOH/ACN (70/30) (v/v) and was dried with a stream of nitrogen.
The cartridge was eluted with 3 mL of ACN and subsequent evap-
oration gave SFB (0.14 g) as a white product.
in Escherichia coli as described in Refs. 39,40) were a kind gift from
AstraZeneca R&D (Lund & Moelndal, Sweden).
This competitive enzyme activity assay was performed by
monitoring the conversion of the fluorogenic substrate Mca-PLA-
QAV-Dpa-RSSSR-NH2 (R&D systems) by recombinant ADAM-17 in
the presence of increasing concentrations of ML5, FB-ML5 or
2FB-ML5. For MMP-2, -9 and -12, the conversion of the fluorogenic
1H NMR (chloroform-d) d 8.24–8.13 (m, 2H), 7.48–7.37 (m, 2H),
2.81 (s, 4H).
substrate
Mca-Pro-Leu-Gly-Leu-Dpa-Ala-Arg-NH2
(Bachem,
Bubendorf, Switzerland) was followed. Measurements were per-
formed in Costar White 96-well plates (Corning, Schiphol-Rijk,
The Netherlands), where each well contained 10 ng ADAM-17
4.6. Synthesis of the reference compound FB-ML5
4.6.1. Synthesis of FB-ML5 by SPPS
and a final concentration of 10
lM substrate in a final volume of
The SPPS resulting to the reference compound FB-ML5 followed
a procedure identical to the synthesis of ML5.
100 L ADAM assay buffer (25 mM Tris pH 9.0, 2.5 l
0.005% w/v Brij-35). Inhibition of MMP proteolytic activity was
l
M ZnCl2,
determined with 10 ng of MMP-2, MMP-9 or MMP-12 per well
4.6.2. Synthesis of FB-ML5 in solution
with a final concentration of 2
buffer (50 mM Tris pH 7.4, 0.2 M NaCl, 10 mM CaCl2, 2.5
0.05% v/v Brij-35). Proteolysis rates were followed by measuring
fluorescence (kex,em = 320, 440 nm) increase using Fluostar
Optima plate reader (BMG Labtech, Offenburg, Germany), at
20 °C for recombinant MMPs and at 37 °C for recombinant
ADAM-17, for 15 min (conditions of the experiments not in the sta-
tionary phase). For ML5, eleven-point inhibition curves (0, 0.5, 1, 2,
6, 12, 25, 50, 100, 500 and 1000 nM) were plotted in GraphPadP-
rism 5.0 for Windows (GraphPad Software, San Diego, USA). For
FB-ML5, nine-point inhibition curves (0, 5, 10, 20, 60, 125, 250,
500 and 1000 nM) were plotted. For 2FB-ML5, six-point inhibition
l
M substrate in 100
l
L MMP assay
The acylation in solution with NHS-coupling agents proceeds
generally in borate buffer at pH 8.5. However, it was shown that
boronic acid complex can be formed with hydroxamic acid moie-
ties.38 As ML5 contains a hydroxamic acid, we considered phos-
phate buffer as a substitute of borate buffer. This buffer was
prepared by a Na2HPO4ꢁ12H2O 1 M/NaH2PO4ꢁH2O 1 M (93.2/6.8)
(v/v) solution and by adjustment of the pH to 8.5 with sodium
lM ZnCl2,
a
hydroxide 1 M. ML5 (0.50 mg, 1.08
phosphate buffer pH 8.5 (0.01 M) was transferred to SFB
(1.08 mol, 500 L) in ACN. The reaction mixture was allowed to
react for 30 min at 50 °C. The reaction was quenched with HCl
(1 mL, 0.25 M). Then an aliquot of 100 L was injected through
an analytical HPLC, using Phenomenex Luna C18 column
(4.6 mm ꢂ 250 mm, m) from Waters, preceded of
lmol, 500 lL) dissolved in
l
l
l
curves (0, 0.1, 1, 5, 10 and 100 lM) were plotted. IC50 values were
determined by sigmoidal fitting. Each experiment was performed
in triplicate.
a
5
l
a
20 ꢂ 4.6 mm2 precolumn. Gradient elution was performed using
a mixture of H2O + 0.1% TFA (solvent A) and CH3CN + 0.1% TFA (sol-
vent B). A linear gradient (overall time = 60 min) starting from 95%
solvent A in solvent B to 100% solvent B at 60 min was employed at
a flow rate of 1 mL minꢀ1. The column effluent was monitored
using an Elite Lachrom VWR Hitachi L-2400 UV detector
(k = 254 nm, AUFS = 0.010) and a Bicron frisk-tech radioactivity
detector. Sample injection was carried out using an injector block
4.8. Synthesis of 4-trimethylammonium ethylbenzoate triflate
To
a solution of ethyl-4-dimethylaminobenzoate (0.65 g,
3.36 mmol) dissolved in 12 mL of anhydrous benzene was added
dropwise methyltrifluoromethanesulfonate (0.42 mL, 3.70 mmol).
The mixture was heated to reflux for 6 h and was then stirred over-
night at room temperature. The reaction mixture was concentrated
under reduced pressure and the resulting crude precipitate was
recrystallized with diethyl ether to yield 1.01 g (84%) of white
product.
with a loop of 100 lL. Fractions of 1 min were collected and the
formed products were identified by mass spectrometry.
The retention time of FB-ML5 was 37 min.
ESI-MS (m/z): 586.5 [M+H]+, calcd 586.7; 608.5 [M+Na]+, calcd
608.7.
1H NMR (chloroform-d) d 8.06–7.97 (m, 2H), 7.78–7.71 (m, 2H),
4.31 (q, J = 6.0 Hz, 2H), 3.70 (s, 9H), 1.32 (t, J = 5.8 Hz, 3H).
4.6.3. Synthesis of FB-ML5 in solution by using an excess of
acylating agent
4.9. Radiosynthesis of no-carrier-added N-succinimidyl-4-
[
18F]fluorobenzoate ([18F]SFB)
ML5 (0.50 mg, 1.08
pH 8.5 (0.01 M) was transferred to SFB (5.40
l
mol, 500
l
L) dissolved in phosphate buffer
mol, 500 L) in ACN.
l
l
Aqueous [18F]fluorine was produced by irradiation of [18O]-
The reaction mixture was allowed to react for 30 min at 50 °C. The
reaction was quenched with HCl (1 mL, 0.25 M). Then an aliquot of
100 lL was injected through an analytical HPLC, as previously
described. Fractions of 1 min were collected and the formed prod-
ucts were identified by mass spectrometry.
The obtained product was 2FB-ML5 with a retention time of
25.7 min.
ESI-MS (m/z): 708.4 [M+H]+, calcd 708.8; 730.4 [M+Na]+, calcd
730.8.
water with a Scanditronix MC-17 cyclotron via the 18O(p,n) 18F
nuclear reaction. The [18F]fluorine solution was passed through a
SepPak Light Accell plus QMA anion exchange cartridge (precondi-
tioned with 5 mL of sodium bicarbonate 1.4% and 100 mL of H2O
and then dried under a flow of Argon) to recover the [18O]-enriched
water. The [18F]fluorine was eluted from the QMA anion exchange
cartridge with 1 mg potassium carbonate in 1 mL of water and col-
lected into a vial containing 5 mg kryptofix[2.2.2]. Subsequently,
1 mL of anhydrous acetonitrile was added and the solvents were
removed at 130 °C under an argon stream. The [18F]KF/krypto-
fix[2.2.2] complex was then dried by azeotropic distillation with
3 times addition of 0.5 mL anhydrous acetonitrile at 130 °C. A solu-
tion of 10 mg 4-trimethylammonium ethylbenzoate triflate in
4.7. In vitro evaluation of ML5, FB-ML5 and 2FB-ML5 in a
fluorogenic inhibition assay
Recombinant ADAM-17 (ectodomain) was purchased from R&D
Systems (Minneapolis, MN, USA). Recombinant catalytic domain
(CD) of human MMP-2 was from Biomol International (Butler Pike,
PA, USA). Recombinant human MMP-12 CD and recombinant
human MMP-9 CD without fibronectin type II insert (expressed
250 l
L anhydrous DMF was added to the [18F]KF/kryptofix[2.2.2]
complex. The radiofluorination was performed at 90 °C for
12 min. Thereafter, 0.5 mL of 1 M HCl was added and the hydroly-
sis of 4-[18F]fluoroethylbenzoate was performed for 5 min at
100 °C. After cooling, crude 4-[18F]fluorobenzoic acid ([18F]FBA)

200
N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
was passed through a SepPak C18 plus cartridge (preconditioned
with 10 mL of EtOH and 10 mL of H2O). The reaction vial was
rinsed with 2 mL H2O which were passed through the SepPak
C18 plus cartridge. To remove unreacted 4-trimethylammonium
ethylbenzoate, a cation exchange extraction SCX cartridge (precon-
ditioned with 3 mL of HCl 0.1 M and 100 mL of H2O and then dried
under a flow of Argon) was connected to the SepPak C18 plus car-
tridge. Purified [18F]FBA was eluted from the SepPak C18 plus/SCX
cartridges with 2 mL acetonitrile in a vial containing 10 mg krypto-
[
18F]FB-ML5 was dissolved in 1 mL human plasma and incubated
at 37 °C for 3 h. After 1 h and 3 h of incubation, aliquots of
250 L were taken. 750 L of ACN were added in order to deprote-
l
l
inize the plasma, and the mixture was centrifuged for 3 min at
3000 rpm. The supernatant was passed through a Millex Filter
(0.22
by semi-preparative HPLC using a Phenomenex reversed-phase
m), preceded by a
lm), diluted with 600 lL ACN and 600 lL H2O and analysed
Luna C18 column (10 mm ꢂ 250 mm,
5
l
20 ꢂ 4.6 mm2 precolumn. Gradient elution was performed using
a mixture of 0.01 M monosodium phosphate buffer (NaH2PO4)
pH 6.0 (solvent A) and ACN (solvent B). The following gradient pro-
file (overall time = 47 min) at a flow rate of 2.5 mL minꢀ1 was used:
30% of ACN in solvent A over 5 min, followed by a linear gradient
from 30% to 60% of ACN in solvent A over 40 min and followed
by a linear gradient from 60% to 10% of ACN in solvent A over
2 min. Fractions of the eluate were collected every minute and
radioactivity in the fractions was determined with a gammacount-
er (LKB Wallac, Turku, Finland).
fix[2.2.2], 5 mg K2CO3 and 50 l
L H2O. As [18F]FBA is highly volatile,
it was converted to [18F]fluorobenzoate which was coevaporated 3
times with 0.5 mL acetonitrile in the presence of kryptofix[2.2.2]
and potassium carbonate. The resulting salt was treated with
freshly dissolved O-(N-succinimidyl)-N,N,N’,N’-tetramethyluroni-
um tetrafluoroborate (TSTU) in 0.5 mL acetonitrile for 5 min at
90 °C to allow the formation of [18F]SFB. After cooling, the produc-
tion of [18F]SFB was completed by addition of 1 mL HCl 0.03 M.
4.10. Purification of N-succinimidyl-4-[18F]fluorobenzoate by
SPE
4.12.2. Saline stability
The stability of [18F]FB-ML5 was evaluated in vitro in saline.
Formulated [18F]FB-ML5 was dissolved in 1 mL saline and incu-
bated at 37 °C for 3 h. After 1 h and 3 h of incubation, aliquots of
[
18F]SFB was diluted with 15 mL of water for injection and
passed through an Oasis HLB 30 mg (1 cc) cartridge for solid phase
extraction. The cartridge was washed with 10 mL of water and
250 lL were taken and diluted with 1 mL ACN and 1 ml H2O and
eluted with 500
4.11. Radiosynthesis of [18F]FB-ML5
18F]SFB dissolved in 500
l
L acetonitrile to give the purified [18F]SFB.
analysed by semi-preparative HPLC as described above. Fractions
of the eluate were collected every minute and radioactivity in
the fractions was determined with a gammacounter (LKB Wallac,
Turku, Finland).
[
lL acetonitrile was added to a solution
of ML5 (0.50 mg, 1.08 mol) in 500
8.5. The reaction was performed at 50 °C for 30 min. After cooling,
the reaction mixture was diluted with 600 L acetonitrile and
600 L H2O and was purified by semi-preparative reverse phase
HPLC. HPLC was performed with Elite LaChrom Merck Hitachi L-
7100 pump system using a Phenomenex reversed-phase Luna
l
l
L 0.01 M phosphate buffer pH
4.13. Octanol/water partition coefficient of [18F]FB-ML5
l
About 5 kBq of formulated [18F]FB-ML5 diluted in 5
was diluted in 495 L PBS (pH = 7.4) and 500 L n-octanol in an
Eppendorf cup. The mixture was vortexed for 5 min and the cup
was centrifuged at 3000 rpm for 5 min. Radioactivity in 100
lL saline
l
l
l
lL
C18 column (10 mm ꢂ 250 mm,
5
lm), preceded of
a
aliquots of the water and n-octanol layers was determined in a
gammacounter (LKB Wallac, Turku, Finland). The experiment was
performed in triplicate. The partition coefficient (logP) was
calculated as:
20 ꢂ 4.6 mm2 precolumn, equipped with both UV (Elite LaChrom
VWR Hitachi L-2400 UV detector set at 254 nm, AUFS = 0.5) and
a Bicron radioactivity monitor. Sample injection was carried out
using an injector block with a loop of 1 mL. Gradient elution was
performed using a mixture of 0.01 M monosodium phosphate buf-
fer (NaH2PO4) pH 6.0 (solvent A) and ACN (solvent B). The follow-
ing gradient profile (overall time = 47 min) at a flow rate of
2.5 mL minꢀ1 was used: 30% of ACN in solvent A over 5 min, fol-
lowed by a linear gradient from 30% to 60% of ACN in solvent A
over 40 min and followed by a linear gradient from 60% to 10% of
ACN in solvent A over 2 min. The retention time of [18F]FB-ML5
was about 20.3 min (18.1–22.6 min). The HPLC-collected fraction
was diluted with about 100 mL of water for injection and passed
through an Oasis HLB 30 mg (1 cc) cartridge. The cartridge was
washed with 10 mL of water for injection and eluted with 0.7 mL
of EtOH. The obtained product was redissolved in saline to
decrease the percentage of EtOH to less than 10% for the subse-
quent cell/animal experiments.
!
cpmoctanol layer
cpmaqueous layer
logP ¼ log10
4.14. In vitro evaluation of [18F]FB-ML5
Human breast cancer MCF-7 cells and human bronchial
epithelium 16HBE cells were obtained from American Type Culture
Collection, Manassas, USA. MCF-7 and 16HBE cells were
maintained in 15 mL Eagle’s Minimum Essential Medium (EMEM)
(Lonza, Walkersville, USA) supplemented with 10% fetal calf serum
(FCS) in a T75 culture flask. Cells were grown in a humidified atmo-
sphere containing 5% CO2 and were passaged twice per week.
For 16HBE cells: 16HBE cells were seeded in a 24-well plate at
50.000 cells/mL 6 days before the experiment. An equal number
of cells were dispensed in each well in 0.5 mL of serum-containing
medium: EMEM supplemented with 10% FCS. Cells were grown to
confluency and serum-starved overnight. One day before the
experiment, the medium was changed to low serum medium:
EMEM supplemented with 0.5% FCS.
Quality control was performed as described for FB-ML5. The
retention time of [18F]FB-ML5 was 37 min.
4.12. In vitro stability of [18F]FB-ML5 in human plasma and
saline
For MCF-7 cells: MCF-7 cells were seeded in a 12-well plate 48 h
before the experiment. An equal number of cells were dispensed in
each well in 1 mL of serum-containing medium: EMEM
supplemented with 10% FCS. Cells were grown to confluency and
serum-starved overnight. One day before the experiment, the
4.12.1. Human plasma stability
The stability of [18F]FB-ML5 was evaluated in vitro in human
plasma. Whole blood from
temperature for 15 min, was centrifuged at 3000 rpm for 5 min,
subsequently the supernatant was taken. 100 L of formulated
a healthy donor, kept at room
l

N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
201
medium was changed to low serum medium: EMEM supple-
mented with 0.5% FCS.
seven days. The study was approved by the Animal Ethics Commit-
tee of the University of Groningen, The Netherlands (DEC 6058B).
Binding/specificity studies were performed when confluency
had reached 80–90%.
4.15.2. HT1080 inoculation
For binding study with MCF-7 cells: 0.5 MBq of [18F]FB-ML5 in
Human fibrosarcoma HT1080 cells were obtained from Ameri-
can Type Culture Collection, Manassas, USA. HT1080 cells were
maintained in 15 mL EMEM supplemented with 10% FCS in a T75
culture flask. Cells were grown in a humidified atmosphere con-
taining 5% CO2 and were passaged twice per week. (2–2.5) ꢂ 106
HT1080 cells, in a 1:1 (v/v) mixture of Matrigel (extracellular
matrix compound, Becton Dickinson) and EMEM with 10% FCS,
were subcutaneously injected into the right shoulder of the
BALB/c nude mice (7–8 weeks old).
<50 lL of saline (containing maximum 10% of absolute ethanol)
were added to each well. After 57 min of incubation, the medium
was quickly removed and the monolayer was washed 3 times with
PBS. Cells were then treated with 0.2 mL of trypsin. When the
monolayer had detached from the bottom of the well, 0.8 mL of
EMEM supplemented with 10% FCS was added to stop the proteo-
lytic action. Cell aggregates were resolved by repeated pipetting of
the trypsin/EMEM mixture. Radioactivity in the cell suspension
(1 mL) was assessed using a gamma counter for 15 s. A sample of
the suspension was mixed with trypan blue solution (1:1 v/v)
and was used for cell counting. Cell numbers were determined
manually, using a phase contrast microscope (Olympus, Tokyo,
Japan), a Bürker bright-line chamber (depth 0.1 mm; 0.0025 mm2
squares) and a hand tally counter. All experiments were performed
as a quadruplicate study with at least two repeats.
4.16. MicroPET studies
Animals were scanned when the tumors reached an adequate
size (0.3–0.6 mL), 14 to 21 days after inoculation. The mice were
randomly divided into two groups: tumor-bearing mice scanned
with [18F]FB-ML5 and tumor-bearing mice scanned with [18F]FB-
ML5 and coinjection of 2.5 mg/kg of ML5.
[
18F]FB-ML5
4.14.1. For specificity study
(6.1 2.3 MBq, 0.15 0.09 nmol), dissolved in saline, was intrave-
nously injected into the penile vein of mice anesthetized with 5%
isoflurane (Pharmacie BV, The Netherlands) in medical air at a flow
rate of 2 mL minꢀ1, after which anesthesia was maintained with 2%
isoflurane. Following induction of anesthesia, the mice were posi-
tioned on the bed of the microPET camera (Focus 220, Siemens
Medical Solutions USA, Inc.) in transaxial position. The body tem-
perature of the mice was maintained by electronically regulated
heating pads. Data acquisition of the microPET camera was initiated
and continued for a period of 90 min. After completion of the
dynamic emission scan, a 515 s transmission scan with a Co-57
point source was performed for correction of attenuation of
511 keV photons by tissue. After microPET scanning, the mice were
terminated by administering a high dose of isoflurane (5%) for
about 20 min.
For 16HBE cells: four different experimental conditions were
examined in quadruplicate: non-stimulated cells, non-stimulated
cells + 10
(25 ng/mL PMA and 100 ng/mL LPS added 60 min before [18F]FB-
ML5 addition) and stimulated cells + 10 M of ML5 (10 L). Blocker
was added 2 min before tracer addition. 0.5 MBq of [18F]FB-ML5 in
<50 L of saline (containing maximum 10% of absolute ethanol)
was added to each well and incubated for 60 min. After washing
lM of non-radioactive ML5 (10 lL), stimulated cells
l
l
l
3 times with 500
trypsin and transferred to test tubes. After addition of 400
l
L PBS, the cells were detached with 100
l
l
L of
L of
EMEM + 10% FCS and resuspension, radioactivity in the cell sus-
pension (0.5 mL) was assessed using a gamma counter. A sample
of the suspension was mixed with trypan blue solution (1:1 v/v)
and was used for cell counting. Cell numbers were determined
manually, using a phase contrast microscope, a Bürker bright-line
chamber and a hand tally counter.
4.17. MicroCT
For MCF-7 cells: six different experimental conditions were
examined in quadruplicate: non-stimulated cells, non-stimulated
After the microPET scan, a computed tomography (CT) scan was
performed for attenuation correction and to provide anatomical
localization. The sacrificed mouse attached to the bed was inserted
to the microCT scanner (MicroCAT II, CTI Siemens) and a microCT
image (exposure time = 1050 ms; X-ray voltage = 55 kvp; anode
current = 500 lA; number of rotation steps = 500; total rota-
tion = 360°) was acquired for 15 min.
cells + 100 nM of ML5 (10
ML5 (10 L), stimulated cells (100 nM PMA added 2.5 h before
18F]FB-ML5 addition), stimulated cells + 100 nM of ML5 (10
L)
and stimulated cells + 10 M of ML5 (10 L). Blocker was added
2 min before tracer addition. 0.5 MBq of [18F]FB-ML5 in <50
L of
lL), non-stimulated cells + 10 lM of
l
[
l
l
l
l
saline (containing maximum 10% of absolute ethanol) was added
to each well and incubated for 60 min. After washing 3 times with
1 mL PBS, the cells were detached with 200
lL of trypsin and trans-
4.18. Ex vivo biodistribution
ferred to test tubes after addition of 800 L of EMEM + 10% FCS and
l
resuspension. Radioactivity in the cell suspension (1 mL) was
assessed using a gamma counter. A sample of the suspension
was mixed with trypan blue solution (1:1 v/v) and was used for cell
counting. Cell numbers were determined manually, using a phase
contrast microscope, a Bürker bright-line chamber and a hand tally
counter.
After the microPET scan, the ex vivo biodistribution was
performed on the sacrificed mice. The following organs were taken:
bladder, bone, brain, heart, kidney, large intestine, liver, lungs, mus-
cle, pancreas, small intestine, spleen, stomach, tumor and urine. A
small drop of infusate was taken for data correction. The blood
was centrifuged in order to collect plasma and red blood cells. All
the organs and infusate were weighed and analyzed for the amount
of radioactivity, using a gammacounter (LKB Wallac, Turku, Fin-
land). Tracer uptake was expressed as the standardized uptake
value (SUVmean), defined as [tissue activity concentration (MBq/
g)/(injected activity (MBq)/mouse body weight (g))]. The tumor-
to-plasma and tumor-to-muscle ratios were also determined.
4.15. HT1080 fibrosarcoma xenograft mouse model
4.15.1. Animals
Male BALB/c nu/nu (BALB/cOlaHsd-Foxn1nu) mice (nude mice)
were obtained from Harlan (Lelystad, The Netherlands). The mice
were housed in IVC cages with paper bedding on a layer of wood
shavings in a room with constant temperature (ꢃ20 °C) and fixed,
12-h light–dark regime. Food (standard laboratory chow, RMH-B,
Hope Farms, The Netherlands) and water were available ad libi-
tum. After arrival, the mice were allowed to acclimatize for at least
4.19. MicroPET image analysis
Emission sinograms were iteratively reconstructed (OSEM2d)
after being normalized, corrected for attenuation, scatter and

202
N. Matusiak et al. / Bioorg. Med. Chem. 23 (2015) 192–202
12. Somiari, S. B.; Somiari, R. I.; Heckman, C. M.; Olsen, C. H.; Jordan, R. M.; Russell,
S. J.; Shriver, C. D. Int. J. Cancer 2006, 119, 1403.
13. Nawrocki, B.; Polette, M.; Marchand, V.; Monteau, M.; Gillery, P.; Tournier, J.-M.
Int. J. Cancer 1997, 72, 556.
14. Zhou, M.; Qin, S.; Chu, Y.; Wang, F.; Chen, L.; Lu, Y. Int. J. Clin. Exp. Pathol. 2014,
7, 3048.
15. Meijer, M. J.; Mieremet-Ooms, M. A.; van der Zon, A. M.; van Duijn, W.; van
Hogezand, R. A.; Sier, C. F.; Hommes, D. W.; Lamers, C. B.; Verspaget, H. W. Dig.
Liver Dis. 2007, 39, 733.
16. Pyke, C.; Ralfkiær, E.; Huhtala, P.; Ralfkiaer, E.; Hurskainen, T.; Dane, K.;
Tryggvason, K. Cancer Res. 1992, 52, 1336.
17. Chandler, S.; Cossins, J.; Lury, J.; Wells, G. Biochem. Biophys. Res. Commun. 1996,
228, 421.
18. Dong, Z.; Kumar, R.; Yang, X.; Fidler, I. J. Cell 1997, 88, 801.
19. Houghton, A. M.; Grisolano, J. L.; Baumann, M. L.; Kobayashi, D. K.; Hautamaki,
R. D.; Nehring, L. C.; Cornelius, L. A.; Shapiro, S. D. Cancer Res. 2006, 66, 6149.
20. Black, R. A.; Rauch, C. T.; Kozlosky, C. J.; Peschon, J. J.; Slack, J. L.; Wolfson, M. F.;
Castner, B. J.; Stocking, K. L.; Reddy, P.; Srinivasan, S.; Nelson, N.; Boiani, N.;
Schooley, K. A.; Gerhart, M.; Davis, R.; Fitzner, J. N.; Johnson, R. S.; Paxton, R. J.;
March, C. J.; Cerreti, D. P. Nature 1997, 385, 729.
21. Arribas, J.; Esselens, C. Curr. Pharm. Des. 2009, 15, 2319.
22. Moss, M. L.; White, J. M.; Lambert, M. H.; Andrews, R. C. Drug Discovery Today
2001, 6, 417.
23. Jones, B. A.; Riegsecker, S.; Rahman, A.; Beamer, M.; Aboualaiwi, W.; Khuder, S.
A.; Ahmed, S. Arthritis Rheum. 2013, 65, 2814.
24. Tursi, A.; Elisei, W.; Principi, M.; Inchingolo, C. D.; Nenna, R.; Picchio, M.;
Giorgio, F.; Ierardi, E.; Brandimarte, G. J Gastrointestin. Liver Dis. 2014, 23,
261.
decay of radioactivity. The list-mode data of the emission scans
were separated into 22 frame sinograms (15 frames of 2 min, 3
frames of 5 min, 2 frames of 7.5 min, 2 frames of 15 min). PET
image analysis was performed using Inveon Research Workplace
(Siemens) software. Regions of interest were drawn around the
tumor. Tracer uptake in the region of interest was determined as
Bq cmꢀ3 and converted to PET-SUVmean using the following for-
mula: [tissue activity concentration (MBq/cc)/(injected activity
(MBq)/mouse body weight (g))].
4.20. Metabolite analysis (plasma) of [18F]FB-ML5 in a HT1080
xenograft mouse model
Metabolite analysis was performed on plasma collected after the
ex vivo biodistribution study. 750
plasma, the mixture was then centrifuged for 3 min at 3000 rpm.
The supernatant was passed through a Millex Filter (0.22 m),
diluted with 600 L acetonitrile and 600 L H2O, and analysed by
lL of ACN was added to 250 lL of
l
l
l
semi-preparative HPLC. Fractions of the eluate were collected every
minute and radioactivity in these fractions was then determined
with a gammacounter (LKB Wallac, Turku, Finland).
25. Kirkegaard, T.; Pedersen, G.; Saermark, T.; Brynskov, J. Clin. Exp. Immunol. 2004,
135, 146.
Acknowledgements
26. Miller, P. W.; Long, N. J.; Vilar, R.; Gee, A. D. Angew. Chem., Int. Ed. 2008, 47,
8998.
The authors wish to thank the Dutch Technology Foundation
(STW) for financial support (project 08008).
27. Leeuwenburgh, M. A.; Geurink, P. P.; Klein, T.; Kauffman, H. F.; van der Marel,
G. A.; Bischoff, R.; Overkleeft, H. S. Org. Lett. 2006, 8, 1705.
28. Geurink, P.; Klein, T.; Leeuwenburgh, M.; van der Marel, G.; Kauffman, H.;
Bischoff, R.; Overkleeft, H. Org. Biomol. Chem. 2008, 6, 1244.
29. Whittaker, M.; Floyd, C. D.; Brown, P.; Gearing, A. J. Chem. Rev. 1999, 99,
2735.
30. Johnström, P.; Clark, J. C.; Pickard, J. D.; Davenport, A. P. Nucl. Med. Biol. 2008,
35, 725.
31. Wester, H. J.; Hamacher, K.; Stöcklin, G. Nucl. Med. Biol. 1996, 23, 365.
32. Wüst, F.; Hultsch, C.; Bergmann, R.; Johannsen, B.; Henle, T. Appl. Radiat. Isot.
2003, 59, 43.
Supplementary data
Supplementary data associated with this article can be found, in
the online version, at http://dx.doi.org/10.1016/j.bmc.2014.11.013.
References and notes
33. Marik, J.; Sutcliffe, J. L. Appl. Radiat. Isot. 2007, 65, 199.
34. Sounni, N. E.; Devy, L.; Hajitou, A.; Frankenne, F.; Munaut, C.; Gilles, C.;
Deroanne, C.; Thompson, E. W.; Foidart, J. M.; Noel, A. FASEB J. 2002, 16,
555.
35. Köhrmann, A.; Kammerer, U.; Kapp, M.; Dietl, J.; Anacker, J. BMC Cancer 2009, 9,
188.
36. Bremer, C.; Tung, C. H.; Weissleder, R. Nat. Med. 2001, 7, 743.
37. Bellac, C. L.; Li, Y.; Lou, Y.; Lange, P. F.; Harwig, C.; Kappelhoff, R.; Dedhar, S.;
Adam, M. J.; Ruth, T. J.; Bénard, F.; Perrin, D.; Overall, C. Cancer Res. 2010, 70,
7562.
1. Seals, D. F.; Courtneidge, S. A. Genes Dev. 2003, 17, 7.
2. Nagase, H.; Woessner, J. F. J. Biol. Chem. 1999, 274, 21491.
3. White, J. M. Curr. Opin. Cell Biol. 2003, 15, 598.
4. MacFadyen, R. J. Curr. Opin. Pharmacol. 2007, 7, 171.
5. Hu, J.; Van den Steen, P. E.; Sang, Q.-X. A.; Opdenakker, G. Nat. Rev. Drug Disc.
2007, 6, 480.
6. Hidalgo, M.; Eckhardt, S. G. J. Natl. Cancer Inst. 2001, 93, 178.
7. Wiley, J. P.; Hughes, K. A.; Kaiser, R. J.; Kesicki, E. A.; Lund, K. P.; Stolowitz, M. L.
Bioconjugate Chem. 2001, 12, 240.
8. Konstantinopoulos, P. A.; Karamouzis, M. V.; Papatsoris, A. G.; Papavassiliou, A.
G. Int. J. Biochem. Cell Biol. 2008, 40, 1156.
38. Temma, T.; Sano, K.; Kuge, Y.; Kamihashi, J.; Takai, N.; Ogawa, Y.; Saji, H. Biol.
Pharm. Bull. 2009, 32, 1272.
9. Björklund, M.; Koivunen, E. Biochim. Biophys. Acta 2005, 1755, 37.
10. Klein, G.; Vellenga, E.; Fraaije, M. W.; Kamps, W. A.; de Bont, E. S. J. M. Crit. Rev.
Oncol. Hematol. 2004, 50, 87.
11. Forsyth, P. A.; Wong, H.; Laing, T. D.; Rewcastle, N. B.; Morris, D. G.; Muzik, H.;
Leco, K. J.; Johnston, R. N. Br. J. Cancer 1999, 79, 1828.
39. Shipley, J. M.; Doyle, G. A.; Fliszar, C. J.; Ye, Q. Z.; Johnson, L. L.; Shapiro, S. D.;
Welgus, H. G.; Senior, R. M. J. Biol. Chem. 1996, 271, 4335.
40. Parkar, A. A.; Stow, M. D.; Smith, K.; Panicker, A. K.; Guilloteau, J. P.; Jupp, R.;
Crowe, S. J. Protein Expr. Purif. 2000, 20, 152.