Evaluation of [18F]BR420 and [18F]BR351 as radiotracers for MMP-9 imaging in colorectal cancer

Article abstract of DOI:10.1002/jlcr.3476

MMP-9 is a zinc-dependent endopeptidase that is involved in the proteolytic degradation of the extracellular matrix and plays an important role in cancer migration, invasion, and metastasis. The aim of this study was to evaluate the potential of MMP-trace

Full text of DOI:10.1002/jlcr.3476

Received: 19 July 2016
DOI 10.1002/jlcr.3476
Revised: 18 October 2016
Accepted: 24 October 2016
R E S E A R C H A R T I C L E
Evaluation of [18F]BR420 and [18F]BR351 as radiotracers for
MMP‐9 imaging in colorectal cancer
Naiara Vazquez1,2,3 | Stephan Missault1 | Christel Vangestel2,4 | Steven Deleye2 |
David Thomae2,3 | Pieter Van der Veken3 | Koen Augustyns3 | Steven Staelens2 |
Stefanie Dedeurwaerdere1 | Leonie wyffels2,4
1 Department of Translational Neurosciences,
MMP‐9 is a zinc‐dependent endopeptidase that is involved in the proteolytic degra-
University of Antwerp, Antwerp, Belgium
2 Molecular Imaging Center Antwerp, University
of Antwerp, Antwerp, Belgium
3 Laboratory of Medicinal Chemistry, University of
Antwerp, Antwerp, Belgium
dation of the extracellular matrix and plays an important role in cancer migration,
invasion, and metastasis. The aim of this study was to evaluate the potential of
MMP‐tracers [18F]BR420 and [18F]BR351 for MMP‐9 imaging in a colorectal
cancer xenograft model. [18F]BR420 and [18F]BR351 were synthesized using an
automated synthesis module. For [18F]BR420, a novel and improved radiosynthesis
was developed. Plasma stability and MMP‐9‐targeting capacity of both radiotracers
was compared in the Colo205 colorectal cancer model. MMP‐9 and MMP‐2 expres-
sion levels in the tumors were evaluated by immunohistochemistry and in situ
zymography. μPET imaging as well as ex vivo biodistribution revealed a higher
tumor uptake for [18F]BR420 (3.15% 0.03% ID/g vs 0.94% 0.18% ID/g for
[18F]BR351 at 2 hours pi) but slower blood clearance compared with [18F]BR351.
4 Department of Nuclear Medicine, University
Hospital Antwerp, Edegem, Belgium
Correspondence
Prof. Dr. Leonie wyffels, Department of Nuclear
Medicine, University Hospital Antwerp,
Wilrijkstraat 10, Antwerp 2650, Belgium.
Email: leonie.wyffels@uza.be
[18F]BR351 was quickly metabolized in plasma with 20.28%
5.41% of intact
tracer remaining at 15 minutes postinjection (PI). By contrast, [18F]BR420 displayed
a higher metabolic stability with >86% intact tracer remaining at 2 hours PI. Immu-
nohistochemistry revealed the presence of MMP‐9 and MMP‐2 in the tumor tissue,
which was confirmed by in situ zymography. However, an autoradiography analysis
of tracer distribution in the tumors did not correlate with MMP‐9 expression. [18F]
BR420 displayed a higher tumor uptake and higher stability compared with [18F]
BR351 but a low tumor‐to‐blood ratio and discrepancy between tracer distribution
and MMP‐9 immunohistochemistry. Therefore, both tracers will not be usefulness
for MMP‐9 imaging in colorectal cancer.
KEYWORDS
Colo205, MMP‐9, PET
1
|
INTRODUCTION
for Zn2+. The predomain acts as an internal inhibitor of MMP
activity, and activation occurs with cleavage of the predomain
Matrix metalloproteinases (MMPs) comprise a large family
of several structurally related zinc‐dependent endopeptidases,
which contribute to the extracellular matrix (ECM) proteoly-
sis.1,2 All MMP family members present a common domain
structure composed of at least a predomain, a catalytic
domain, and a highly conserved active site domain.3,4 The
catalytic core‐domain contains the active, protein‐degrading
ability of the proteinase, which includes a metal binding site
leading to a conformational change to the active site. MMPs
are synthesized as secreted or membrane‐associated
(MT‐MMPs) inactive zymogens and must be proteolytically
processed to an active state. This processing involves removal
of a cysteine residue that interacts with zinc ions from the
active site, thereby resulting in MMP activation. On the basis
of their structure and substructure specificity, human MMPs
are divided into
5 groups: collagenases, gelatinases,
J Label Compd Radiopharm 2016; 1–11
wileyonlinelibrary.com/journal/jlcr
Copyright © 2016 John Wiley & Sons, Ltd.
1

2
VAZQUEZ ET AL.
stromelysins, membrane‐type metalloproteinases, and
others.5
Recently, [18F]BR420 and [18F]BR351 (Figure 1) have
been described to be brain penetrant broad‐spectrum MMP
PET radiotracers with favorable affinities for activated
Because of their capacity to degrade the ECM, MMPs are
known to play an important role in several diseases.6–9 In par-
ticular in cancer, excess matrix degradation is one of the hall-
marks and an important component in the process of tumor
progression.10 Among the large number of MMPs, gelatinase
A (MMP‐2) and gelatinase B (MMP‐9) are the most fre-
quently investigated proteases for their involvement in cancer
migration, invasion, and metastasis.11,12 In particular, MMP‐
9 can show both anticancer and tumor‐promoting effects. A
disturbance of MMP‐9 function in animals has been corre-
lated with a decrease in tumor progression, whereas the over-
expression of MMP‐9 would induce angiogenesis and
increase malignant transformation.13,14 Colorectal cancer is
the third most common cancer worldwide and the forth most
common cause of death.15,16 Considerable evidence has
implicated MMPs in established colorectal cancer, showing
either an increased MMP expression or increased MMP
activity in this type of tumor. The overexpression of MMP‐
9 has been demonstrated in human colorectal cancers, where
the degree of overexpression has been shown to correlate
with stage of disease and/or prognosis.17
MMP‐2 and MMP‐9 ([18F]BR420, IC50 = 7
2 and
27 and
23
4
9 nM; and [18F]BR351, IC50 = 50
3 nM, for MMP‐9 and MMP‐2, respectively).23,24 Both
compounds belong to different families of MMP ligands with
BR420 being a barbiturate‐based MMP inhibitor and BR351,
a hydroxamate‐based MMP inhibitor. The in vivo evaluation
of [18F]BR420 has so far been limited to an in vivo
biodistribution study up to 2 hours postinjection (PI) in adult
C57/BL6 mice, indicating no tissue specific accumulation of
the radiotracer.23 By contrast, [18F]BR351 has recently suc-
cessfully been used to image MMP activity after transient
middle cerebral artery occlusion in mice.25 [18F]BR420 and
[18F]BR351 have not yet been evaluated in an oncology con-
text. Considering the prognostic value of MMP‐9 expression
in colorectal cancer, the aim of the present study was to eval-
uate and compare the in vivo behavior of [18F]BR420 and
[18F]BR351 in a mouse model of human colorectal cancer.
The tumor‐targeting properties were evaluated as well as
the in vivo stability at several time points after tracer injec-
tion. For [18F]BR420, a novel and improved automated
radiosynthesis was developed.
Because of their role in cancer, MMPs have gained signif-
icant attention as therapeutic target. The search for MMP
inhibitors has been largely conducted during the last
decades.18,19 There are several requirements for the design
of an appropriate MMP inhibitor: the presence of a functional
group providing a hydrogen bond with the enzyme, side
chains able to interact with the enzyme sub sites, and a zinc
binding group. The development of MMP inhibitors with a
high selectivity toward a particular MMP, however, remains
a big challenge. Most inhibitors are broad‐spectrum MMP
inhibitors, although some display favorable binding affinity
for a specific MMP type. Recently, a barbiturate‐based fluo-
rescent MMP inhibitor (Cy5.5‐AF443) with a high affinity
toward MMP‐2 and MMP‐9 has been successfully used for
the in vivo evaluation of MMP‐2/‐9 expression in mouse
models of colorectal cancer.20,21 The use of the fluorescent
label however requires the application of invasive endoscopic
techniques for imaging of MMP activity in the colon.
Targeting MMP with positron emission tomography (PET)
or single photon emission computed tomography (SPECT)
activity‐based imaging probes could offer a noninvasive tool
to early detect and diagnose MMP‐related diseases (eg,
colorectal cancer), to elucidate the exact role of specific
MMPs in normal and disease states or to understand the
mechanism of action of these enzymes. During the last
decades, several research groups have been developing PET
and SPECT probes for MMP‐targeted imaging (for a recent
review, see Matusiak et al22). Most of them are however
broad‐spectrum MMP inhibitors, and currently there are no
successful MMP‐selective radiotracers available. Conse-
quently, in vivo imaging of a particular MMP type activity
remains a challenge.
2
| EXPERIMENTAL
2.1 | General procedures and materials
Unless stated otherwise, all chemical reagents and solvents
were obtained from commercial sources and used without
further purification. The characterization of all compounds
was performed by H NMR, 13C NMR, and mass spectrom-
1
etry. H NMR and 13C NMR spectra were recorded on a
1
400‐MHz Bruker Advance DRX spectrometer and analyzed
using MestReNova analytical chemistry software. Chemical
shifts are in parts per million, and coupling constants are in
hertz. Purities were determined using 2 different UPLC
systems based either on mass detection or on UV detection:
a Waters Acquity UPLC system coupled to a Waters TUV
detector and a Waters TQD ESI mass spectrometer (Waters,
Milford, Massachusetts, USA). On both systems, a Waters
Acquity UPLC BEH C18 1.7 μm, 2.1 × 50 mm column
was used. Solvent A: water with 0.1% formic acid; Solvent
B: acetonitrile with 0.1% formic acid. Method I: 0.7 mL/
FIGURE 1 Structures of [18F]BR420 and [18F]BR351

VAZQUEZ ET AL.
3
min, 0.15 minutes 95% A, 5% B then in 1.85 minutes from
95% A, 5% B to 10 0% B, 0% A, then 0.25 minutes, 100%
B, 0% A, 0.75 minutes (0.35 mL/min) 95% A, 5% B. Method
II (purity method): 0.4 mL/min, 0.15 minutes 95% A, 5% B
then in 4.85 minutes from 95% A, 5% B to 100% B, 0% A,
then 0.25 minutes, 100% B, 0% A, 0.75 minutes 95% A,
5% B. Where necessary, flash purification was performed
with a Biotage ISOLERA One flash system equipped with
an internal variable dual‐wavelength diode array detector
(200‐400 nm). Biotage® SNAP cartridges were used for nor-
mal phase purification (KP‐Sil, 10‐100 g; flow rate, 10‐
100 mL/min). For radiosynthesis, no carrier‐added [18F] F−
was produced by the 18O(p,n)18F nuclear reaction in a Sie-
mens Eclipse HP cyclotron by bombardment of [18O]H2O
(Rotem Industries, Israel). The radiosynthesis of [18F]
BR420 and [18F]BR351 was conducted on an automated syn-
thesis module (Fluorsynthon I, Comecer Netherlands, The
Netherlands) specifically adapted for these radiosynthesis.
Radiochemical yields were calculated from the theoretical
initial amount of [18F]F− and decay corrected to end of bom-
bardment (EOB). The purification of [18F]BR420 and [18F]
BR351 after radiosynthesis was performed by reverse phase
semipreparative HPLC using a Knauer HPLC pump and a
Smartline UV detector (λ = 254 nm) in line with a Hi‐Rad
1000‐CD‐X CdWO4 scintillation detector (Scionix, The
Netherlands). Radiochemical and chemical purity was deter-
mined by analytical reverse phase HPLC using a Shimadzu
LC‐20AT HPLC pump equipped with an SPD‐20A UV/VIS
detector (Shimadzu, Japan) in series with a NaI scintillation
detector for radiation detection (Raytest, Germany). The
recorded data were processed by the GINA‐Star 5 software
(Raytest). Radioactivity in samples from animal studies was
measured in a Wizard2 2480 automatic gamma counter
(Perkin Elmer, Finland).
separated and washed with water and brine, dried over
Na2SO4, filtered, and concentrated under reduced pressure.
The crude product was purified by manual chromatography
(hexane/EtOAc, 1/2 v/v) to yield compound 1 (0.054 g,
20%). 1H NMR (400 MHz, CDCl3): δ = 9.18 (bs, 2H),
7.43‐7.39 (m, 4H), 7.21‐7.17 (m, 1H), 7.08‐7.02 (m, 4H),
4.25‐4.22 (m, 2H), 3.80‐3.78 (m, 2H), 3.40‐3.34 (m, 4H),
2.60‐2.56 (m, 4H) ppm. 13C NMR (100 MHz, CDCl3): δ
30.72, 42.94, 64.79, 74.30, 78.61, 78.98, 79.26, 118.19,
119.37, 124.19, 128.82, 129.23, 129.56, 130.19, 149.35,
153.55, 153.97, 155.70, 157.60, and 169.91 ppm. MS (ESI)
m/z = 487.2 [M + H + CH3CN]+.
2.3 | Radiochemistry
[18F]F− was transferred from the cyclotron to the synthesis
module and trapped on an activated anion exchange cartridge
(QMA Sep‐Pak Light, Waters). [18F]F− was eluted with 1 mL
of a solution of Kryptofix 222 (220 mg) and K2CO3 (40 mg)
in CH3CN/H2O (83/17 v/v) and collected in the reaction vial.
The solution was evaporated at 130°C with a stream of
helium (1 bar) and under vacuum for 8 minutes. To ensure
dryness, azeotropic distillation of 1 mL of acetonitrile was
performed twice for 6 minutes at 140°C with a stream of
helium (1 bar). The reaction was then cooled down to 90°C
with compressed air, and a solution of the precursor 1
(2 mg in 1 mL of DMSO) was added to the reaction vial.
The reaction mixture was heated at 120°C for 10 minutes,
cooled down to 80°C, and diluted with 1.9 mL of a solution
of 0.05 M NaOAc, pH 5.5/96% EtOH (50/50 v/v). This solu-
tion was passed through an alumina N light cartridge
(Waters) and injected into the HPLC loop. The crude reaction
mixture was purified on a Phenomenex Luna C18, 10 μm,
10 × 250 mm column and eluted with a solution of 0.05 M
NaOAc, pH 5.5/96% EtOH (50/50 v/v) at 4 mL/min. [18F]
BR420 was collected (Rt = 12‐13minutes), filtered through
a 0.22‐μm sterile filter (PES syringe filter, 25 mm, 0.2 μm),
and diluted with 0.9% sodium chloride solution to reduce
the EtOH concentration to <10%. Chemical and radiochemi-
cal purity of the tracer was measured by analytical HPLC
(Phenomenex Luna C18, 5 μm, 4.6 × 150 mm) eluted with
0.1% TFA in H2O/0.1% TFA in CH3CN (55/45 v/v) at a
flowrate of 1 mL/min. The specific activity was determined
using a UV calibration curve (λ = 272 nm).
2.2 | Chemistry
2.2.1
| 5‐[4‐(2‐Chloroethyl)piperazin‐1‐yl]‐5‐(4‐
phenoxyphenyl)pyrimidine‐2,4,6‐trione (1)
N‐(2‐Chloroethyl) piperazine (42%) was prepared as the
trifluoroacetic acid salt 3 by the reaction of N‐Boc piperazine
2 (97%) with bromo‐chloroethane. Barbiturate scaffold 4
(67%) was prepared as previously described.26 To a cooled
solution of 4 (0.186 g, 0.628 mmol) in dimethylformamide
(DMF) (2 mL) at 0°C, a solution of NBS (0.112 g,
0.628 mmol) in DMF (2 mL) was dropwise added for a
period of 15 minutes. Then, the addition funnel was rinsed
with DMF (2 mL), and the rinse was also added to the reac-
tion mixture. After stirring for 20 minutes, compound 3
(0.165 g, 0.628 mmol) and K2CO3 (0.174 g, 1.256 mmol)
were added and the reaction mixture was stirred at 0°C for
60 minutes and then allowed to reach room temperature and
stirred for additional 20 hours. The reaction mixture was then
diluted in EtOAc and an aqueous solution of citric acid
(0.628 mmol, 0.1 M) was added. The organic phase was
[18F]BR351 was reproduced according to the literature24
with slight modifications. A solution of the precursor 5
(2 mg in 1 mL of CH3CN) was added to the azeotropically
dried [18F]K222F and the reaction mixture was heated at
90°C for 20 minutes, cooled down to 80°C and diluted with
1.9 mL of a solution of 0.05 M NaOAc, pH 5.5/96% EtOH
(45/55 v/v). This solution was passed through an Alumina
N light cartridge (Waters) and injected into the HPLC loop.
The crude reaction mixture was purified on a Phenomenex
Luna C18, 10 μm, 10 × 250 mm column and eluted with a
solution of 0.05 M NaOAc, pH 5.5/96% EtOH (45/55 v/v)

4
VAZQUEZ ET AL.
at 3 mL/min. [18F]BR351 eluted with a retention time of 10
to 11 minutes, was collected, filtered through a 0.22‐μm
sterile filter (PES syringe filter, 25 mm, 0.2 μm) and diluted
with 0.9% sodium chloride solution to reduce the EtOH
concentration to <10%. Chemical and radiochemical purity
of the tracer was measured by analytical HPLC (Phenomenex
Luna C18, 5 μm, 4.6 × 150 mm [Phenomenex, Torrance,
California, USA]) eluted with 0.1% TFA in H2O/0.1% TFA
in CH3CN (52/48 v/v) at a flow rate of 1 mL/min. The
specific activity was determined using a UV calibration curve
(λ = 254 nm).
subsets of the 3‐dimensional ordered subset expectation max-
imization algorithm28 followed by 18 maximum a posteriori
iterations. Normalization, dead time, random, CT‐based
attenuation, and scatter corrections were applied. CT imaging
was done using a 220° rotation with 120 rotation steps.
Voltage and amperage were set to 80 keV and 500 μA,
respectively.
Volume of interests (VOIs) for the whole tumors, heart
(a measure of the blood pool), muscle, liver, and kidneys were
drawn manually on the coregistered CT images using PMOD
v3.3 (PMOD Technologies, Switzerland). Using this VOI, the
tracer uptake was then obtained and defined as the percent
injected dose per cubic centimeter (% ID/cc) = 100 × activity
concentration in the VOI (kBq/cc) / the injected dose (kBq).
2.4 | Cell cultures and animal model
The experimental protocol was approved by the Antwerp
University Ethical Committee for Animal Experiments
(ECD 2015‐13), and all applicable institutional and national
guidelines for the care and use of animals were followed.
Human colorectal cancer cell line Colo205 was used for
induction of xenografts because of its known MMP‐9 expres-
sion.27 Low‐passage Colo205‐luc2 cells (PerkinElmer,
Maltham, Massachusetts) were cultured in RPMI1640
medium enriched with 10% FBS, 2 mM L‐glutamine, 1 mM
sodium pyruvate, and 1% penicillin‐streptomycin
(Invitrogen, Merelbeke, Belgium) at 37°C and 5% CO2 in a
humidified incubator. For inoculation, cells were harvested
by trypsinization with trypsin‐EDTA (0.05%, Invitrogen)
and counted with the Muse™ Cell Count and Viability Assay
(Merck Millipore, Overijse, Belgium). Colo205 cells were
then resuspended in sterile phosphate‐buffered saline (PBS)
at a concentration of 2 × 106 viable cells per 100 μL for inoc-
ulation. Female CD1−/− athymic nude mice aged 6 to 8 weeks
(Charles Rivers, Calco, Italy) were subcutaneously inocu-
lated in both hind legs with 100 μL of cell suspension. Tumor
growth was evaluated 3 times a week with digital caliper
measurements from the moment tumors became palpable.
Tumor volume was calculated with the formula
V = 0.5 × (length × width2).
2.6 | Ex vivo biodistribution and blocking study
CD1−/− nude female mice bearing Colo205 xenografts (n = 4
for [18F]BR420 and n = 3 for [18F]BR351) were intrave-
nously injected with 18.76
1.02 MBq [18F]BR420 or
5.92 1.23 MBq [18F]BR351 via the lateral tail vein under
isoflurane anesthesia. At 2 hours PI, animals were sacrificed
by cervical dislocation to perform ex vivo biodistribution.
Blood was collected via cardiac puncture, and thereafter
tumors and main organs were rapidly removed, rinsed in
PBS, blotted dry, and measured. The uptake of radioactivity
in blood and organs was expressed as the percentage of the
injected dose per gram of tissue plus or minus the standard
deviation (% ID/g SD).
A blocking study was performed for [18F]BR420 (n = 4),
using
a different and more potent MMP‐9 ligand
(IC50 = 0.2 nM for MMP‐9, 20 nM for MMP‐3, and 30 nM
for MMP‐2) (Figure 2).29 The blocking agent was injected
30 minutes before the radiotracer at a dose of 1 mg/kg. Two
hours PI (6.66
1.68 MBq), the animals were sacrificed.
Blood, tumors, and main organs were collected, weighed,
and counted as described previously.
2.7 | Plasma metabolite analysis
2.5 | In vivo μPET imaging
Plasma radiometabolite analysis was performed at 2 hours PI
for [18F]BR420 (n = 4, 5.72 0.24 MBq) and at 5 minutes,
15 minutes, 30 minutes, and 2 hours PI for [18F]BR351
(n = 3 per time point, 14.41 5.72 MBq). At the indicated
time points, mice were sacrificed, and blood (500‐800 μL,
withdrawn by cardiac puncture) was centrifuged at 4°C for
7 min at 4000 rpm. The plasma was removed (200‐300 μL),
Full‐body PET/CT imaging was performed on an Inveon
μPET/CT scanner (Siemens Preclinical Solutions Inc, Knox-
ville, Tennessee). CD1−/− nude female mice bearing Colo205
xenografts were injected intravenously via a catheter in the
lateral tail vein with [18F]BR420 (37.8 MBq, n = 1) or
[18F]BR351 (5.6
3.4 MBq, n = 2) and underwent a
60‐min dynamic μPET scan followed by static scans at
2 hours PI (20 minutes) and 4 hours PI (46 minutes). For
the scan, mice were anesthetized with isoflurane (5% for
induction, 2% for maintenance) supplemented with oxygen.
PET data were recorded in list‐mode format, and the 60‐
minute scan was divided in a total of 33 frames: 12 × 10,
3 × 20, 3 × 30, 3 × 60, 3 × 150, and 9 × 300 seconds. The
PET images were reconstructed using 2 iterations with 16
FIGURE 2 Structure of the MMP‐9 blocking agent

VAZQUEZ ET AL.
5
and the same amount of ice‐cold CH3CN was added followed
by addition of 10 μL of a solution of the cold standard
(BR420 or BR351, 1 mg/mL). The sample was counted in
the automatic gamma counter and vortexed for 30 seconds.
After centrifugation for 4 minutes at 4000 rpm, the superna-
tant was separated from the pellet. Subsequently, both frac-
tions were counted in the gamma counter to calculate the
extraction efficiency (% recovery of radioactivity in CH3CN).
One hundred microliters of the supernatant was analyzed by
reverse phase HPLC using a Phenomex Luna (C18, 5 μm,
4.6 mm × 150 mm) column eluted with 0.1% TFA in
H2O/0.1% TFA in CH3CN (55/45 v/v for [18F]BR420;
52/48 v/v for [18F]BR351) at a flow rate of 1 mL/min. The
HPLC column was coupled to a Phenomenex guard column.
Fractions were automatically collected every 30 seconds and
measured with the gamma counter. A validation study for
metabolite analysis was also performed to exclude degrada-
tion of the radiotracer during workup. For this, an aliquot
of the radiotracer was added to blood in vitro, and the
samples were processed as described for the metabolite
analysis and analyzed by HPLC.
cresyl violet, dehydrated, and coverslipped. Images were
obtained using an Olympus CX31 light microscope
(Olympus Corporation, Tokyo, Japan) equipped with an
AxioCam Erc 5‐second camera (Zeiss, Oberkochen,
Germany). ZEN 2012 software (Zeiss) was used for image
acquisition.
In situ zymography was performed to assess gelatinase
activity (MMP‐9 and MMP‐2) in the same tumor tissue.
Ten‐micrometer fresh‐frozen sections were preincubated in
water at 37°C for 100 minutes and subsequently covered with
a fluorogenic substrate DQ gelatin (DQ gelation from pig
skin, fluorescein conjugate; Life Technologies/Molecular
Probes, Eugene, Oregon, USA), which was diluted 1:100 in
the buffer supplied by the manufacturer, at 37°C for
45 minutes. After washing, sections were mounted using
VECTASHIELD mounting medium with DAPI (Vector
Laboratories, Burlingame, California) and coverslipped. To
examine the specificity of the reaction, 10 mM 1,10‐
phenantroline (Life Technologies/Molecular Probes, USA),
a chelator and general MMP inhibitor, was added to the
preincubation and reaction steps in sections adjacent to those
used for the normal in situ zymography. Images were
acquired at room temperature using an Olympus BX51
fluorescence microscope equipped with PlanC‐N 4×/0.10
and UplanFL‐N 20×/0.5 objective lenses and an Olympus
DP71 digital camera. Olympus cellSens Dimension software
was used for image acquisition and processing.
2.8 | Ex vivo autoradiography, immunohistochemistry,
and in situ zymography
Immediately after gamma counting, the tumors were snap‐
frozen in cooled isopentane (−35°C) and sectioned. Serial
cryosections were collected in triplicate for ex vivo autoradi-
ography (20 μm), immunohistochemistry (20 μm), and in situ
zymography (10 μm).
3
| RESULTS AND DISCUSSION
For ex vivo autoradiography, 20 μm fresh‐frozen sections
were exposed overnight to high‐sensitivity phosphor imaging
plates and subsequently scanned using the Fujifilm image
reader FLA‐7000.
3.1 | Chemistry
Cold standard BR420 was synthesized in a moderate yield
(38%) following the published methodology.23 Because the
2‐step radiosynthesis of [18F]BR420 according to the
reported method23 failed in our hands, alternative 1‐step
radiolabeling procedures using different precursors were
explored. First the synthesis of a tosylate derivative was
evaluated. The synthesis of the labeling precursor, however,
failed most likely because tosylation/mesylation of the
primary alcohol results in a highly unstable aziridinium salt,
which undergoes a ring‐opening reaction to yield the starting
alcohol.30 We therefore decided to prepare chloroderivative 1
as alternative precursor for the 1‐step radiosynthesis of [18F]
BR420 following the synthetic pathway described in
Scheme 1.
Immunohistochemical staining of MMP‐9 and MMP‐2
was performed on 20 μm fresh‐frozen sections using, respec-
tively, a rabbit polyclonal anti‐MMP‐9 antibody (ab38898;
Abcam, Cambridge, UK) and a rabbit polyclonal anti‐
MMP‐2 antibody (ab37150; Abcam). Tumor sections were
fixed using 4% paraformaldehyde and washed with PBS.
Nonspecific binding was blocked using 3% hydrogen perox-
ide and 3% normal goat serum (Jackson ImmunoResearch,
Suffolk, UK). Three percent bovine serum albumin (Sigma‐
Aldrich, Belgium) was included in the blocking step for
immunohistochemical staining of MMP‐2. Sections were
incubated overnight with the primary antibody, diluted
1:1000 (anti‐MMP‐9) or 1:500 (anti‐MMP‐2) in antibody
diluent containing 0.1% bovine serum albumin, 0.2% Triton
X‐100 (Sigma‐Aldrich), and 2% normal goat serum. The
following morning, sections were washed and incubated for
1 hour with horse radish peroxidase–conjugated secondary
antibody (goat antirabbit; 1:500; Jackson ImmunoResearch,
UK). Finally, sections were exposed for 10 minutes to the
colorimetric substrate 3,3′‐diaminobenzidine (Dako,
Denmark). The tumor sections were counterstained with
Cold standard BR351 and the corresponding precursor 5
were reproduced as previously described in the literature24 in
moderate yields of 38% and 54%, respectively.
3.2 | Radiochemistry
While evaluating the 2‐steps radiosynthesis of [18F]BR420
described in literature,23 radio‐HPLC analysis of the sec-
ond step (reaction of the 1‐[18F]fluoro‐2‐tosyloxyethane

6
VAZQUEZ ET AL.
SCHEME 1 Synthesis of precursor 1. Reagents
and conditions: (a) (i) Cs2CO3, DMF, rt, 3d; (ii)
TFA, DCM, 0°C‐rt; (b) NBS, K2CO3, DMF, 5°C‐rt
SCHEME 2 Fully automated synthesis of [18F]
BR420 and [18F]BR351. Reagents and conditions:
(a) [18F]Fluoride, K2CO3, Kryptofix‐2.2.2,
DMSO, 120°C, 10 min; (b) [18F]Fluoride, K2CO3,
Kryptofix‐2.2.2, CH3CN, 90°C, 20 min
with the piperazine‐containing barbiturate scaffold) at
different time points indicated formation of several by‐
products while [18F]BR420 could not be obtained. By
implementing a 1‐step approach using direct displacement
of the chloroderivative 1 by [18F]fluoride (Scheme 2),
[18F]BR420 could be successfully obtained in a radio-
3.3 | Exploratory in vivo μPET imaging study
An ex vivo biodistribution study at different time points up to
60 minutes PI of [18F]BR351 as well as a μPET imaging
study up to 2 hours PI of [18F]BR420 in normal C57/BL6
mice have previously been published.23,24 These studies indi-
cated low uptake for both tracers in heart, brain, lung, and
muscle with clearance of radioactivity out of excretory organs
such as liver and kidneys with time after injection. Therefore,
in the current study, only a limited number of animals was
evaluated by μPET imaging to enable the selection of a time
point for a more elaborate ex vivo evaluation of tumor uptake
and tracer stability.
Representative μPET/CT images are presented in
Figure 3. For [18F]BR420, tumor uptake increased from
0.45% ID/cc at 1 minute PI to a maximum value of 1.94%
ID/cc at 2 hours after tracer injection with little decrease in
uptake up to 4 hours PI (1.78% ID/cc) (Figure 4).
chemical yield of 16.22%
3.12% (n = 10, decay
corrected to EOB). By using a biocompatible mobile
phase for HPLC purification of the reaction mixture, a
simple dilution and sterile filtration step could be used to
obtain an injection‐ready solution of [18F]BR420, thereby
reducing the total synthesis to 60 minutes (EOB),
including formulation. Chemical and radiochemical purity
were >98%, and the determined specific activity was
65.07
27.63 GBq/μmol (EOS, n = 10).
The radiosynthesis of [18F]BR351 was performed as pre-
viously reported24 by reaction of tosylate precursor 5 dis-
solved in CH3CN with dried [18F]F− for 20 minutes at
90°C (Scheme 2). Again, a biocompatible mobile phase was
used for the purification, omitting the need for an extra
Sep‐Pak purification step. Purified [18F]BR351 could be
obtained in a decay corrected radiochemical yield of
For [18F]BR351, an initial higher tumor uptake of
1.11% ID/cc was detected at 1 minute PI, which increased
to a maximum uptake of 1.23% ID/cc at 9 minutes PI.
Tumor uptake then decreased to 0.42% ID/cc at 2 hours
PI and remained stable up to 4 hours PI (0.41% ID/cc).
In contrast to [18F]BR420 for which tumor uptake exceeded
uptake in the muscle at 2 hours PI (1.66% ID/cc in the
muscle) and 4 hours PI (1.22% ID/cc in the muscle), tumor
22.42%
12.79% (n = 5), in a total synthesis time of
70 minutes (EOB), including formulation, and with a radio-
chemical purity >98%. The determined specific activity was
172.0 153.10 GBq/μmol (EOS, n = 5).

VAZQUEZ ET AL.
7
FIGURE 3 Representative μPET/CT images of CD1−/− athymic nude mice intravenously injected with 38 MBq [18F]BR420 (n = 1) (A) or 6 MBq [18F]BR351
(n = 2) (B). PET images are presented in sagittal and coronal orientation at 1 min, 30 min, and 2 h after tracer injection. Tumors are pointed with arrow heads
FIGURE 4 Time‐activity curve (% ID/cc) for
heart, tumor, and muscle (A) and kidneys and liver
(B) of CD1−/− athymic nude mice injected with
[
18F]BR420 (n = 1) or [18F]BR351 (n = 2)
uptake of [18F]BR351 did not exceed uptake in the muscle
at any of the investigated time points. [18F]BR420 demon-
strated a slow blood clearance with an initial peak uptake
of 17.78% ID/cc in the heart decreasing to 5.53% ID/cc
at 2 minutes PI, 3.90% ID/cc at 30 minutes PI, and
3.09% ID/cc at 2 hours PI. For [18F]BR351, an initial
higher peak uptake of 20.56% ID/cc was detected in the
heart followed by a faster washout compared with [18F]
BR420 (5.93% ID/cc at 2 minutes PI, 1.03% ID/cc at
30 minutes PI, and 0.76% ID/cc at 2 hours PI). For
[18F]BR420, the highest uptake could be detected in the
liver (26.27% ID/cc at 4 minutes PI and 15.69% ID/cc at
2 hours PI), indicating predominant hepatobiliary clear-
ance of the tracer. For [18F]BR351, a lower (13.24% ID/
cc at 4 minutes PI and 1.25% ID/cc at 2 hours PI) but also
predominant hepatobiliary clearance could be detected. In
the previously reported μPET study of [18F]BR420 in
C57/BL6 mice, an initial high liver uptake was demon-
strated, but in contrast to our study, the liver uptake
decreased and at 2 hours PI uptake was mainly visible
in intestines and bladder.23 The slower clearance from
the liver seen in our study might be related to a mouse
strain difference. For [18F]BR351, an ex vivo
biodistribution study in C57/BL6 mice indicated

8
VAZQUEZ ET AL.
predominant hepatobiliary clearance at 60 minutes after
tracer injection,24 which is comparable with the excretion
profile witnessed in our study.
T/B and T/M ratios for [18F]BR351 were 0.67 and 1.06,
respectively.
As the best tumor accumulation was obtained for [18F]
BR420, a blocking study was performed to evaluate MMP‐
9 selectivity of the tumor signal. Pretreatment of the animals
with a more potent MMP‐9 inhibitor resulted in a significant
decrease in [18F]BR420 uptake in the tumor (P < .01)
(Figure 5B). The administration of the blocking agent also
influenced the tracer uptake in all other organs with a signif-
icant reduction of tracer accumulation observed for liver
(P < .01) and brain (P < .05). This might be related to con-
stitutive physiological expression of MMPs in normal tissue,
reflecting the broad spectrum of both [18F]BR420 and our
blocking agent.
3.4 | Ex vivo evaluation of tumor uptake
For a more elaborate comparison of tumor uptake and
stability of [18F]BR420 and [18F]BR351, the 2‐hour PI time
point was selected as the exploratory μPET imaging study
indicated clearance of radioactivity from nontarget organs
at 2 hours PI while tumor uptake stabilized. The 2‐hour PI
time point was also the time point demonstrating the highest
tumor uptake, which was attained for [18F]BR420. The
distribution of 18F radioactivity in various organs at 2 hours
PI after intravenous administration of [18F]BR420 (with
and without blocking) and [18F]BR351 is presented in
Figure 5. As was also visible in the μPET imaging, [18F]
BR420 displays a slower blood clearance compared with
3.5 | Plasma metabolite analysis
A possible explanation for the higher tumor uptake of [18F]
BR420 compared with [18F]BR351 could be its lower bind-
ing affinity for MMP‐9 (IC50 = 50 27 nM vs 7 2 nM
for [18F]BR420), the faster blood clearance of [18F]BR351,
or a lower stability compared with [18F]BR420 resulting in
a reduced tumor accumulation. A plasma metabolite study
was therefore performed to analyze the in vivo stability of
both tracers. The metabolite validation assay demonstrated
high extraction yields (93.06% 0.69% for [18F]BR420 and
[18F]BR351 with 3.44%
0.93% ID/g remaining in the
blood at 2 hours PI for [18F]BR420 versus 1.40% 0.17%
ID/g for [18F]BR351. For [18F]BR420, the biodistribution
study showed a large accumulation of the tracer in the
liver (19.33%
4.76% ID/g) and large intestines
(9.49% 3.70% ID/g), whereas for [18F]BR351, the highest
uptake could be detected in large and small intestines
(27.90% 0.79% and 14.14% 3.22% ID/g, respectively)
(Figure 5A). Tumor uptake was significantly higher for
[18F]BR420 (3.15% 0.03% ID/g vs 0.94% 0.18% ID/g
for [18F]BR351, P < .05; Figure 5B). The slow blood clear-
ance of [18F]BR420 resulted in low a tumor‐to‐blood (T/B)
ratio of 0.92 and a tumor‐to‐muscle (T/M) ratio of 2.03.
94.75%
1.48% for [18F]BR351), indicating limited or
reversible binding to plasma proteins. No degradation of the
radiotracers was observed during workup.
The radio‐HPLC metabolite analysis for [18F]BR420
revealed high stability of the radiotracer in plasma at 2 hours
after intravenous injection, where the concentration of
FIGURE 5 Ex vivo biodistribution
(% ID/g SD) in selected organs (A) and tumors
(B) at 2 h PI of [18F]BR420 (n = 4 for non blocked
and blocked) or [18F]BR351 (n = 3). Statistical
analysis was performed by unpaired 2‐tailed t test,
*P < .05, ** P < .01, significantly different from
nonblocked

VAZQUEZ ET AL.
9
TABLE 1 In vivo plasma metabolite analysis of [18F]BR351 at different
time points after tracer injection
[18F]BR351 remaining at 2.5 minutes PI and 37% at
20 minutes PI with formation of 2 polar metabolites. The much
lower plasma stability observed in our study might be mouse
strain related or might be attributed to different HPLC condi-
tions used for the analysis. Wagner et al did not identify the
major metabolite that was formed but assumed phase‐I metab-
olism and thus O‐dealkylation of the parent compound,
resulting in 2‐[18F]fluoroethanol to be the mechanism of
metabolite generation. This was confirmed in the current study
by coelution with 2‐fluoroethanol.
Retention time on HPLC
Time
PI (min)
M1
2 min
M2
3.5 min
[
18F]BR351
7 min
5 min
42.08% 2.45%
67.24% 6.06%
75.75% 4.31%
80.45% 3.24%
6.53 0.51%
9.40 0.44%
ND
48.34% 2.60%
20.28% 5.41%
12.17% 2.86%
6.95% 0.95%
15 min
30 min
120 min
ND
Values are expressed as % total radioactivity, mean SD (n = 3). ND indicates
not detected.
3.6 | Ex vivo autoradiography, immunohistochemistry
and in situ zymography
unmetabolized [18F]BR420 (tR = 8.5 minutes) represented
86.18%
radiometabolite peaks could be detected eluting with reten-
tion times of 2 minutes (M1, 7.02% 2.76%), 4 minutes
(M2, 1.96% 0.32%), and 10 minutes (M3,
5.60% 0.68%), respectively. Extraction yield for plasma
3.12% of the total radioactivity. Three minor
As both [18F]BR351 and [18F]BR420 are broad‐spectrum
MMP inhibitors, yet with favorable affinities for activated
MMP‐2 and MMP‐9, the presence of MMP‐9 and MMP‐2
in the tumor tissue was assessed by immunohistochemistry.
MMP‐9 was sparsely present in the stroma of the tumor
(Figure 6B‐C and F‐G). By contrast, MMP‐2 immunostain-
ing demonstrated a higher expression and homogeneous dis-
tribution of MMP‐2 throughout the tumor tissue (Figure 6D
and H).
In situ zymography revealed gelatinolytic activity in the
tumor tissue (Figure 7A). Gelatinolytic activity could be
blocked by adding 1,10‐phenanthroline, a zinc chelator
(Figure 7B).
remained high (91.32% 1.10%).
By contrast, for [18F]BR351, the radio‐HPLC metabolite
analysis of plasma indicated only 6.95% 0.95% of intact
tracer remaining at 2 hours PI. Additional time points were
therefore analyzed indicating a fast metabolism with almost
50% of the tracer being metabolized as early as 5 minutes
PI (Table 1). At 15 min PI, the concentration of unmetabo-
lized [18F]BR351 (tR = 7 minutes) represented only
20.28%
5.41% of the total radioactivity. Two polar
radiometabolites were detected in the plasma, eluting at
2 minutes (M1) and 3.5 minutes (M2). The presence of
metabolite M1 increased with time PI, and it became the
main radiometabolite peak at 2 hours PI (80.45% 3.24%).
M1 was identified as 2‐[18F]fluoroethanol, a known
brain penetrant metabolite, by injection of a solution of
2‐fluoroethanol under the same HPLC conditions. Extraction
yields for plasma remained relatively constant and high
(85.50% 5.27%) at all investigated time points. A plasma
metabolite study by Wagner et al24 reported 86% of intact
Ex vivo autoradiography of tumor slices revealed a homo-
geneous distribution of both radiotracers in tumor tissue
(Figure 6A and E). Although in the blocking study a signifi-
cant reduction in [18F]BR420 tumor uptake could be obtained
by pretreatment with a more potent MMP‐9 inhibitor, indicat-
ing possible selective binding of [18F]BR420 to MMP‐9, the
discrepancy between radioactivity distribution of [18F]
BR420 in the autoradiography analysis and the distribution
of MMP‐9 in the IHC analysis does not support selective bind-
ing of the tracer to MMP‐9. As the tracer distribution in the
FIGURE 6 Ex vivo autoradiography and immunohistochemichal staining of MMP‐9 and MMP‐2 in the tumor tissue. A–E, Representative ex vivo
autoradiographs of, respectively, [18F]BR420 (A) and [18F]BR351 (E) uptake in Colo205 tumor tissue at 2 hours PI. Ex vivo autoradiography reveals a
homogeneous distribution of both radiotracers in the tumors. B‐C‐F‐G, MMP‐9 immunostaining reveals sparse expression of MMP‐9 in the stroma of the tumor.
B‐F, 4× magnification; C‐G, 10× magnification; scale bar = 10 μm. D‐H, MMP‐2 immunostaining reveals a high and homogeneous expression of MMP‐2
throughout the tumor

10
VAZQUEZ ET AL.
FIGURE 7 Gelatinolytic activity in Colo205 tumor tissue. Fluorescence micrographs of in situ zymography with DQ gelatin as fluorogenic substrate (green) in
the absence (A) and presence (B) of 10 mM 1,10‐phenanthroline, a general MMP inhibitor. Scale bar = 200 μm
2. Yong VW, Power C, Forsyth P, Edwards DR. Metalloproteinases in biology
and pathology of the nervous system. Nat Rev Neurosci. 2001;7:502–511.
tumor is more correlated with MMP‐2 IHC and because the
blocking agent also has a relative affinity for MMP‐2
(IC50 = 30 nM versus 24 nM for [18F]BR420), the reduction
3. Stamenkovic I. Matrix metalloproteinases in tumor invasion and metastasis.
Semin Cancer Biol. 2000;10:415–433.
in [18F]BR420 uptake in the tumor might also be related to
4. Manello F, Tonti G, Papa S. Matrix metalloproteinase inhibitors as anticancer
therapeutics. Curr Cancer Drug Targets. 2005;4:285–298.
MMP‐2 binding. However, the homogeneous distribution
might as well just reflect a broad‐spectrum binding profile.
5. Curran S, Murray GI. Matrix metalloproteinases: molecular aspects of their
roles in tumor invasion and metastasis. Eur J Cancer. 2000;36:1621–1630.
6. Mobasheri A. Osteoarthritis year 2012 in review: biomarkers. Osteoarthr
Cartilage. 2012;20:1451–1464.
4
| CONCLUSION
7. Folgueras AR, Pendas AM, Sanchez LM, Lopez‐Otin C. Matrix metallopro-
teinases in cancer: from new functions to improved inhibition strategies. Int
J Dev Biol. 2004;48:411–424.
MMP‐9 plays an important role in cancer migration, inva-
sion, and metastasis. The availability of an MMP‐9‐selective
radiotracer could aid in drug development, diagnosis, and
staging of tumors with known MMP‐9 expression, including
colorectal cancer. Here we present the first evaluation of [18F]
BR420 and [18F]BR351 as potential MMP‐9 PET radio-
tracers in a colorectal cancer model expressing MMP‐9.
The preparation of [18F]BR420 was improved using a novel,
faster, and fully automated strategy. [18F]BR351 was quickly
metabolized and displayed only low tumor uptake. These
unpromising results make [18F]BR351 not suited for MMP
imaging in mouse models of cancer. [18F]BR420 showed a
significantly higher tumor uptake compared with [18F]
BR351 and was found to be stable in plasma up to 2 hours
PI. Nevertheless, due to a low T/B ratio and discrepancy of
tracer distribution in the tumor with MMP‐9 immunostain-
ing, [18F]BR420 will not be useful for selective in vivo imag-
ing of MMP‐9 imaging in colorectal cancer.
8. Vihinen P, Ala‐aho R, Kahari V M. Matrix metalloproteinases as therapeutic
targets in cancer. Curr Cancer Drug Targets 2005;5:203–220.
9. Tziakas DN, Chalikias GK, Stakos D, et al. Independent and additive prog-
nostic ability of serum carboxy‐terminal telopeptide of collagen type‐I in
heart failure patients: a multi‐marker approach with high‐negative predictive
value to rule out long‐term adverse events. Eur J Prev Card. 2012;19:62–71.
10. Mc Cawley L J, Matrisian L. Matrix metalloproteinases: multifunctional
contributors to tumor progression. Mol Med Today. 2000;4:149–156.
11. Yoshizaki T, Sato H, Furukawa M. Recent advances in the regulation of
matrix metalloproteinase 2 activation: from basic research to clinical applica-
tion. Oncol Rep. 2002;9:607–611.
12. Klein G, Vellenga E, Fraaije M W, Kamps WA, de Bont ES. The possible role
of matrix metalloproteinase MMP‐2 and MMP‐9 in cancer, e.g. acute leuke-
mia. Crit Rev Oncol Hematol. 2004;2:87–100.
13. Konstantinopoulos PA, Karamouzis MV, Papatsoris AG, Papavassiliou AG.
Matrix metalloproteinase inhibitors as anticancer agents. Int J Biochem Cell
Biol. 2008;40:1156–1168.
14. Leeman MF, Curran S, Murray G I. New insights into the roles of matrix
metalloproteinases in colorectal cancer development and progression. J
Pathol. 2003;4:528–534.
ACKNOWLEDGMENTS
This work has been funded by the ERA‐NET NEURON and
the University of Antwerp with a BOF research grant (KP
2014 Vazquez N) and Stephan Missault's PhD fellowship of
the Research Foundation Flanders (FWO 11K3714N). The
authors are thankful to Caroline Berghmans and Philippe
Joye for their technical support.
15. Gonzalez‐Pons M, Cruz‐Correa M. Colorectal cancer biomarkers: where are
we now? Biomed Res Int. 2015;1–14.
16. Siegel R, DeSantis C, Jemal A. Colorectal cancer statistics, 2014. CA Cancer
J Clin. 2014;64:104–117.
17. Zucker S, Vacirca J. Role of matrix metalloproteinases (MMPs) in colorectal
cancer. Cancer Metastasis Rev. 2004;23:101–117.
18. Matter H, Schudok M. Recent advances in the design of matrix metallopro-
teinase inhibitors. Curr Opin Drug Discov Devel. 2004;7:513–535.
REFERENCES
19. Cuniasse P, Devel L, Makaritis A, et al. Future challenges facing the develop-
ment of specific active‐site‐directed synthetic inhibitors of MMPs. Biochimie.
2005;87:393–402.
1. Birkedal‐Hanse H, Moore WGI, Bodden MK, et al. Matrix metalloprotein-
ases: a review. Crit Rev Oral Biol Med. 1993;4:197–250.

VAZQUEZ ET AL.
11
20. Schwegmann K, Bettenworth D, Hermann S, et al. Detection of early murine
colorectal cancer by MMP‐2/‐9‐guided fluorescence endoscopy. Inflamm
Bowel Dis. 2016;22:82–91.
27. Liang W, Kujawaski M, Wu J, et al. Antitumor activity of targeting Src
kinases in endothelial and myeloid cell compartments of the tumor microen-
vironment. Clin Cancer Res. 2010;16:924–935.
21. Bettenworth D, Mücke M M, Schwegmann K, et al. Encoscopy‐guided
orthotopic implantation of colorectal cancer cells results in metastatic colorec-
tal cancer in mice. Clin Exp Metastasis. 2016;33:551–562.
28. Hudson HM, Larkin RS. Accelerated image reconstruction using ordered sub-
sets of projection data. IEEE Trans Med Imag. 1994;13: 601–609.
29. Hanessian S, Moitessier N, Gauchet C, Viau M. N‐Aryl sulfonyl homocyste-
ine inhibitors of matrix metalloproteinases: further probing of the S1, S1' and
S2' pockets. J Med Chem. 2001;44:3066–3073.
22. Matusiak N, van Waarde A, Bischoff R, et al. Probes for non‐invasive matrix
metalloproteinases‐targeted imaging with PET and SPECT. Curr Pharm Des.
2013;19:4647–4672.
30. Peng Z, Maxwell DS, Sun D, et al. Imatinib analogs as potential agents for
PET imaging of Bcr‐Abl and c‐KIT expression at a kinase level. Bioorg
Med Chem. 2014;1:623–632.
23. Breyholz HJ, Wagner S, Faust A, et al. Radiofluorinated pyrimidine‐2,4,6‐
triones as molecular probes fro noninvasive MMP‐targeted imaging. Chem
Med Chem. 2010;5:777–789.
24. Wagner S, Breyholz H J, Law M P, et al. Novel fluorinated derivatives of the
broad‐spectrum MMP inhibitors N‐hydrxy‐2(R)‐[[(4‐methoxyphenyl)sulfo-
nyl](benzyl)‐ and (3‐picolyl)‐amino]‐3‐methyl‐butanamide as potential tools
for the molecular imaging of activated MMPs with PET. J Med Chem.
2007;50:5752–5764.
How to cite this article: Vazquez N, Missault S,
Vangestel C, Deleye S, Thomae D, Van der Veken P,
Augustyns K, Staelens S, Dedeurwaerdere S, wyffels L.
Evaluation of [18F]BR420 and [18F]BR351 as
radiotracers for MMP‐9 imaging in colorectal cancer.
J Label Compd Radiopharm. 2016. doi: 10.1002/
jlcr.3476
25. Zinnhardt B, Viel T, Wachsmuth L, et al. Multimodal imaging reveals tempo-
ral and spatial microglia and matrix metalloproteinase activity after
experimental stroke. J Cereb Blood Flow Metab. 2015;35:1711–1721.
26. Hutchings S, Liu W, Radinov R Palladium‐catalyzed arylation of diisopropyl
malonate applied to the efficient synthesis of the selective MMP inhibitor
5‐(4‐phenoxyphenyl)‐5‐[4‐(2‐pyrimidinyl)‐1‐piperazinyl]barbituric
Heterocycles. 2006;67: 763–768.
acid.