Functional in vivo imaging of cysteine cathepsin activity in murine model of inflammation

Article abstract of DOI:10.1016/j.bmc.2010.10.028

Near-infrared fluorophore (NIRF)-labeled imaging probes are becoming increasingly important in bio-molecular imaging applications, that is, in animal models for tumor imaging or inflammation studies. In this study we showed that the previously introduced

Full text of DOI:10.1016/j.bmc.2010.10.028

Bioorganic & Medicinal Chemistry 19 (2011) 1055–1061
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry
journal homepage: www.elsevier.com/locate/bmc
Functional in vivo imaging of cysteine cathepsin activity in murine model
of inflammation
b
b
c
b
b
ˇ a
Dejan Caglic , Anja Globisch , Maik Kindermann , Ngee-Han Lim , Volker Jeske , Hans-Paul Juretschke ,
Eckart Bartnik b, K. Ulrich Weithmann b, Hideaki Nagase c, Boris Turk a, K. Ulrich Wendt b,
⇑
a
ˇ
Jozef Stefan Institute, Department of Biochemistry and Molecular and Structural Biology, Jamova 39, 1000 Ljubljana, Slovenia
b Sanofi-Aventis Deutschland GmbH R&D Chemical, Analytical Sciences and R&D Thrombosis & Angiogenesis, Industriepark Höchst, 65926 Frankfurt, Germany
c Kennedy Institute of Rheumatology Division, Imperial College London, London W6 8LH, United Kingdom
a r t i c l e i n f o
a b s t r a c t
Article history:
Received 23 July 2010
Revised 29 September 2010
Accepted 12 October 2010
Available online 19 October 2010
Near-infrared fluorophore (NIRF)-labeled imaging probes are becoming increasingly important in
bio-molecular imaging applications, that is, in animal models for tumor imaging or inflammation studies.
In this study we showed that the previously introduced chemical concept of ‘Reverse Design’ represents
an efficient strategy for the generation of selective probes for cysteine proteases from chemically opti-
mized protease inhibitors for investigations in proteomic lysates as well as for in vivo molecular imaging
studies. The newly developed activity-based probe AW-091 was demonstrated to be highly selective for
cathepsin S in vitro and proved useful in monitoring cysteine cathepsin activity in vivo, that is, in zymo-
san-induced mouse model of inflammation. AW-091 showed higher signal-to-background ratios at
earlier time points than the commercially available polymer-based ProSense680 (VisEn Medical) and
thus represents an efficient new tool for studying early proteolytic processes leading to various diseases,
including inflammation, cancer, and rheumatoid arthritis. In addition, the fluorescent signal originating
from the cleaved AW-091 was shown to be reduced by the administration of an anti-inflammatory drug,
dexamethasone and by the cathepsin inhibitor E-64, providing a valuable system for the evaluation of
small-molecule inhibitors of cathepsins.
Keywords:
Solid-phase synthesis
Activity-based probe
Cysteine cathepsin
Zymosan
Inflammation
In vivo imaging
Ó 2010 Published by Elsevier Ltd.
1. Introduction
of target enzymes (covalent labeling ABPs) for the subsequent
identification of specific enzyme proteomes from different physio-
The fields of activity-based proteomics and functional in vivo
imaging both apply reactive chemical probes to detect enzymatic
activities in native proteomic extracts, cells, and whole organisms.
Because increased protease activity has been shown to be associ-
ated with a number of human diseases, including cancer, osteoar-
thritis, rheumatoid arthritis, much effort has been put into
development of agents for monitoring of protease activity. Proteo-
mic methods utilize activity-based probes (ABPs) which function
as irreversible inhibitors, covalently modifying the catalytic site
logical samples.1 While such covalent labeling ABPs were originally
designed to localize and identify activated enzymes in complex
proteomic lysates, they have also been applied successfully to
in vivo bio-imaging studies.2,3 Nevertheless, monitoring the activ-
ity of low-abundant enzymes in vivo demands a different type of
ABPs which are not irreversible inhibitors but substrates (substrate
ABPs) and exhibit high specificity towards the enzyme under
investigation. Selectivity for a given protease is controlled by the
peptide recognition sequence. While some proteases show a high
degree of selectivity for a given sequence (aspartic acid at P1 site
of caspase), others show high promiscuity and thus represent
difficult targets for selective substrates. As more than 500 prote-
ases are encoded in the human genome, it remains challenging
to control selectivity of substrate cleavage in vivo. To address the
issue of selectivity of substrates for a subset of proteases, namely
cysteine cathepsins, we recently proposed that substrate ABPs with
high selectivity for cysteine cathepsins can be obtained by a
straightforward chemical concept, which we referred to as ‘Reverse
Design’ (Ref. 4). We utilized two chemically optimized cathepsin
inhibitors to design selective and cell-permeable ABPs. This was
Abbreviations: ABP, activity-based probe; Boc, tert-butyloxycarbonyl; DIPEA,
diisopropyl-ethyl amine; DMF, dimethylformamide; DMSO, dimethylsulfoxide;
DTT, dithiothreitol; ESI-MS, electrospray ionization mass spectrometry; equiv,
equivalents; Fmoc, 9-fluorenylmethoxycarbonyl; HBTU, O-benzotriazole-N,N,N0,N0-
tetramethyluronium hexafluorophosphate; HEPES, 4-(2-hydroxyethyl)-1-piperazi-
neethanesulfonic acid; HOBt, N-hydroxybenzotriazole; HPLC, high performance
liquid chromatography; LC–MS, liquid chromatography–mass spectrometry; NMR,
nuclear magnetic resonance; PBS, phosphate buffered saline; RT, room tempera-
ture; TFA, trifluoroacetic acid; SPPS, solid-phase peptide synthesis.
⇑
Corresponding author. Fax: +49 69 305 80169.
E-mail address: ulrich.wendt@sanofi-aventis.com (K.U. Wendt).
0968-0896/$ - see front matter Ó 2010 Published by Elsevier Ltd.
doi:10.1016/j.bmc.2010.10.028

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1056
achieved by converting their substrate-mimicking warheads into
cleavable peptide bonds and subsequently attaching appropriate
reporter groups. The selectivity profile of the inhibitors was thus
transferred to the corresponding ABPs. Here we expand these stud-
ies in order to investigate whether ‘Reverse Design’ as a chemical
concept is applicable to near-infrared (NIR) quenched fluorescent
in vivo ABPs.
2.2. General procedure for the solid-phase synthesis
The molecular scaffold 3 was synthesized using standard solid-
phase peptide synthesis methods. The 2-chlorotrityl-chloride resin
(Novabiochem, loading 1.4 mmol/g) was used as solid support. For
loading of the resin (100 mg, 0.14 mmol) 2 equiv Fmoc-protected
amino acid and 3 equiv DIPEA (74 ll, 0.42 mmol) were dissolved
Cathepsins belong to a family of cysteine proteases, the major-
ity of which reside in the endosomal/lysosomal system and are
thus termed lysosomal cysteine cathepsins. There are 11 members
present in humans, including cathepsins B, C, F, H, K, L, O, S, V, W,
and X (Ref. 5). Cysteine cathepsins were long believed to be pri-
marily involved in intracellular protein turnover, but a growing
body of evidence suggests that at least some of them are involved
in specific cellular processes as well as in numerous pathologies
including cancer, cardiovascular diseases, osteoarthritis, rheuma-
toid arthritis, and other diseases linked with the immune system.5
Cathepsin S, a potent cysteine peptidase, has been recognized to be
the critical enzyme in invariant chain degradation, and antigen
processing and presentation.6 The enzyme has recently been con-
sidered as a pharmacological target for immune disorders and
inflammatory diseases, including rheumatoid arthritis, atheroscle-
rosis, and myasthenia gravis.7–10
We describe here the development and application of AW-
091, a near-infrared quenched fluorescent substrate ABP, that is
cleaved efficiently by cathepsin S. A direct comparison with a
commercially available cathepsin-selective polymer-based sub-
strate ProSense680 has been performed. AW-091 was shown to
give higher signal-to-noise ratios at earlier time points after
probe injection and thus offered readouts at earlier time points
after induction of inflammation than the ProSense680 substrate.
Moreover, comparable activity ratios (i.e., ratio between the sig-
nal in ipsilateral and contralateral limbs) were observed, whereas
our probe showed shorter residence time in the tissue, rendering
it attractive for repeated frequent activity determinations. In
vitro profiling of AW-091 on recombinant enzymes together with
in vivo inhibition studies suggests that the fluorescent signal of
cleaved AW-091 recorded in mice correlates with the activity
of cathepsin S.
in 2 ml CH2Cl2 and the reaction mixture was added to the resin.
The reaction mixture was shaken overnight at room temperature.
The resin was washed three times with 2 ml CH2Cl2 and 2 ml
DMF. For Fmoc-deprotection the resin was treated two times for
15 min with 2 ml 30% piperidine/DMF. A standard protocol was
used for solid-phase peptide synthesis: 4 equiv Fmoc-protected
amino acid, 4 equiv HBTU (212 mg, 0.56 mmol), 4 equiv HOBt
(76 mg, 0.56 mmol), and 8 equiv DIPEA (196 ll, 1.12 mmol) were
dissolved in 2 ml CH2Cl2/DMF (1/1; v/v). The reaction mixture
was stirred 20 min at room temperature and then added to the re-
sin. The reaction mixture was shaken for 2 h at room temperature.
For the cleavage of the peptide 3 from the solid-phase the resin
was treated two times for 15 min with 2 ml 2% TFA/CH2Cl2 (v/v).
The solvent was co-evaporated with toluene under reduced pres-
sure and the product was purified by preparative HPLC
(H2O + 0.1% TFA; 10–95% CH3CN, 15 min, 120 ml/min, column:
Sun Fire 50 ꢀ 100 mm, Waters).
2.3. Synthesis of AW-091
AW-091 was synthesized following Scheme
1 with the
fluorophore Cy7 and the quencher BHQ-3. ESI-MS: calculated:
[M+H]+ = 1620.0, found: [M+H]+ = 1619.8.
2.4. Fluorescence assay for AW-091
For the in vitro enzyme activity assay, the active cathepsins
were dissolved in AHNP-buffer (150 mM acetate/HEPES, pH 6.5,
300 mM NaCl; 0.001% Pluronic; 5–100 mM
on the enzyme) at a final concentration of 10 nM. The DMSO-dis-
solved AW-091 was added at 0.8–40 M (final concentrations), fol-
L-cysteine, depending
l
lowed by fluorescence measurements with a Tecan Safire2 plate
reader (kex = 710 nm, kem = 767 nm) at 25 °C. The final DMSO con-
centration in the assay did not exceed 1% (v/v). Steady-state kinetic
data were fitted by non-linear least-squares regression analysis
using the following relationship:
AW-091 enabled the monitoring of cathepsin S activity in vitro
as well as in a mouse model of inflammatory paw edema. The
probe thus represents a promising tool for experimental investiga-
tion of inflammation pathophysiology and for monitoring efficacy
of small-molecule inhibitors.
v
¼ ½SꢁVmax=ðKm þ ð1 þ ð½Sꢁ=KsiÞÞ½SꢁÞ;
where is the initial velocity, [S] the equilibrium concentration of
v
substrate, Vmax the maximal rate, Km the Michaelis–Menten con-
stant and Ksi is the constant for substrate inhibition.
2. Material and methods
2.1. General material and methods
2.5. In vivo experiments with mice
Unless otherwise noted, all reagents were purchased from com-
mercial suppliers and used without further purification. Recombi-
nant human cathepsins L and S were purchased from R&D Systems
and cathepsin B from Sigma–Aldrich, whereas cathepsin K was
produced as previously described.11 Active protein concentrations
were determined by active-site titration using E-64 (Peptide Insti-
tute). All solvents used were of HPLC grade. Reactions were ana-
lyzed by thin-layer chromatography on Merck 50 ꢀ 100 mm
silica gel 60 aluminum sheets with fluorescent indicator or
LC–MS. Column chromatography was carried out with Merck silica
gel 60 (0.040–0.063 mm). Reverse-phase HPLC was performed on a
C18 column Sun Fire (50 ꢀ 100 mm, Waters) or XBridge Prep C18
2.5.1. Zymosan model
The animal study was performed according to the regulations of
the German Animal Protection Law. Male CD-1 mice, 10 weeks of
age, were obtained from Charles River Laboratories. Paw edema
was induced by intraplantar injection of 10 ll of 10 mg/ml
b-zymosan in sterile 0.9% NaCl-solution.
2.5.2. Administration of imaging agents, and treatments with
dexamethasone and E-64
To study the effect of dexamethasone and E-64 on protease
activity in inflamed paws, mice were distributed randomly into
four groups (n = 3). One group was treated orally with dexametha-
sone (10 mg/kg) in vehicle (0.6% (w/v) methylcellulose–0.5% (v/v)
Tween80) 1 day before zymosan, shortly after zymosan injection
and 1 h before injecting AW-091. The second and third groups re-
(5
l
m, 10 ꢀ 100 mm, Waters). LC/MS data were acquired using
the HP-Agilent 1100 MSD system. NMR-data were recorded on a
Bruker DRX-400 system in d6-DMSO and calibrated to the residual
solvent peak.

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1057
Cl
Cl
Solid Phase
Peptide Synthesis
O
O
O
O
H
N
H
N
NHFmoc
a
b, c
N
N
H
N
H
N
N
H
OH
O
O
O
O
HN
O
HN
O
O
O
3
4
Chloro-trityl resin
SO3-
N+
SO3-
N+
N
N
N
SO3-
SO3-
d, e
HN
O
O
HN
O
O
O
O
O
H
N
H
N
O
O
O
N
H
H
N
O
N
H
N
H
N
N
N
N
O
H
N
N+
N
N
O
N
5
2
Scheme 1. Synthesis of AW-091 (molecule 2) using a combination of solid-support and solution-phase synthesis as previously described (Ref. 4). Reagents and conditions:
(a) 1.2 equiv HOBt, 1.3 equiv HBTU, 2 equiv N-Fmoc-ethane-1,2-diamine hydrochloride, 3 equiv DIPEA, DCM/DMF (1/1), 12 h; (b) 50% TFA/DCM, 10 min; (c) 1 equiv Cy7–OSu
(GE Healthcare), 6 equiv DIPEA, DMF, 12 h; (d) Et2NH/DMF (1/4), 30 min, rt; (e) 1 equiv BHQ-3–OSu (Biosearch Technologies), 6 equiv DIPEA, DMF, 12 h.
ceived four ip injections (2 days and 1 day before zymosan injec-
tion, shortly after zymosan injection and 1 h before injecting
AW-091) of 10 and 30 mg/kg E-64 in 40% DMSO/PBS, respectively.
15 min at 0, 8, and 24 h after injection of the probe using the Kodak
system. Fluorescence detection was performed with image acquisi-
tion times of 3 s or 2 min for the Berthold system and the Kodak
system, respectively. Images were analyzed with Winlight 32 soft-
ware version 2.93. Regions of interest (ROIs) were placed identi-
cally over the ipsi- and contralateral hind paws.
For groups one to three, all mice were administered a 20 lM solu-
tion of AW-091 in PBS iv through tail vein injection (800 nmol/kg
body weight of AW-091) 24 h after the zymosan injection. When
animals (n = 3) were imaged with ProSense680 (VisEn Medical),
13.3 lM solution of the probe was administered iv through tail
vein 24 h after zymosan treatment.
3. Results and discussion
Because of the limited bioavailability of E-64, another control
experiment was performed in the fourth group of animals. For
these animals, zymosan, AW-091 and E-64 were injected directly
into the foot pads (n = 3 for each group). Specifically, the right paws
The synthesis of quenched fluorescent activity-based probe
AW-091 was based on previous findings that the selectivity profile
of an optimized cathepsin inhibitor could be transferred into a
selective and cell-permeable ABP by replacing the substrate-mim-
icking warhead of such an inhibitor by a cleavable peptide bond
and subsequently attaching appropriate reporter groups for optical
imaging.4 Here we expand these studies in order to demonstrate
that the chemical concept of Reverse Design is applicable to near-
infrared (NIR) quenched fluorescent ABPs, suitable for in vivo
bio-molecular imaging studies.
of 42-day-old C57BL/6 mice were injected with 30
lg of zymosan.
After 2 days, 10 l of 1 M AW-091 with or without 1 l
l
l
l of 1 mM
E-64 was injected directly into ipsi- and contralateral paws of
mice.
2.5.3. In vivo optical imaging and image analysis
Crystal structures of cathepsin S suggested that the P1 and P10
moieties are openly accessible to the surrounding solvent.12,13
When compared to our previous cathepsins S probe (Ref. 4), two
modifications have been applied here. In order to design a red-
shifted NIR probe, we exchanged the positions of quencher and
NIRF in vivo imaging was performed with a Berthold Night Owl
CCD imaging system with an excitation band-pass filter at 700/
20 nm and an emission band-pass filter at 780/20 nm, or with a
Kodak In Vivo FX Pro CCD imaging system with an excitation
band-pass filter 710/20 nm and an emission band-pass filter at
790/20 nm emission filter. For imaging, the animals were anesthe-
tized with 1.5–2% isofluran/N2O/O2. The animals were imaged in
the supine position on a template ensuring a defined and reproduc-
ible position of the hind feet and care was taken to provide a uni-
form illumination at both positions of the feet. The tail and the
upper portion of the body were covered with non-fluorescent plas-
tic to obscure unspecific signals originating from the tail vein injec-
tion site, bladder, and liver. Throughout anesthesia and imaging,
the body temperature of the mice was kept constant with heating
pads and rectal measurement of their body temperature. Fluores-
cent images were taken at 10 min, 1, 3, 6, and 24 h after injection
of the AW-091 using the Berthold system and every 3 min for
fluorophore in the P1 and P0 position. This was mainly due to sta-
1
bility problems of the Cy7 fluorophore during the required depro-
tection conditions of the Boc-group. Furthermore, the composition
of both linker groups in P1 and P0 showed to have an influence on
1
the performance of the probe. Especially unspecific cleavage could
be reduced by changing the original lysine group in P1 to the non-
natural 2,4-diaminobutyric acid and from 2,4-diaminobutane in P10
to the shorter 1,2-diamino ethane. The cleavable peptide bond
between the P1 butyric acid derivative functionalized with the cor-
responding fluorophore Cy7 and the diamino ethane carrying the
BHQ-3 quencher in P01 should be precisely oriented into the active
site to be cleaved by cathepsin S.

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1058
The core scaffold of quenched fluorescent AW-091 (Fig. 1) was
synthesized by standard solid-phase peptide chemistry followed
by further derivatization of orthogonally protected amino groups
in solution with Cy7 as a fluorophore and BHQ-3 as a quencher
(for details see Reaction Scheme 1 and Supplementary Fig. S1).
The core molecular scaffold, tripeptide 3, was synthesized by con-
ventional solid-phase peptide synthesis on chlorotrityl based res-
ins by using Fmoc-protected amino-acid building blocks.
Coupling of the Fmoc-mono protected 1,2-diaminoethane lead to
the orthogonal protected building block 4. First deprotection of
the Boc-group and subsequent coupling to the Cy7–OSu ester led
to intermediate 5 which after purification on a preparative HPLC
was deprotected from the Fmoc group and a last coupling step with
BHQ-3–OSu ester yielded AW-091 (molecule 2).
In the first step, we set out to assess the efficiency of fluores-
cence quenching by comparing equimolar solutions (10 lM) of
the free fluorescent dye Cy7 and of AW-091. The relative fluores-
cence intensities were 1: 0.22 ꢀ 10ꢂ3 (data not shown). Assuming
similar brightness of the free fluorophores, this result suggests that
AW-091 is efficiently quenched unless activated by enzymatic
cleavage and thus represents a good in vivo imaging agent.
Enzymatic turnover and selectivity of AW-091 were tested
in vitro against a panel of recombinant cysteine cathepsins (B, K,
L, and S; seeFig. 2a). AW-091 exhibited good catalytic efficiency to-
wards cathepsin S (kcat = 0.36 sꢂ1; Km = 2.5 ꢀ 10ꢂ5 M) together
with pronounced selectivity for cathepsin S in comparison to re-
lated cysteine cathepsins B, L, and K (kcat/Km for cathepsin
Figure 2. In vitro profiling of AW-091. (a) Recombinant cathepsins B, K, L, and S
(10 nM final concentration) were incubated with 22.5 M AW-091 and fluorescence
was measured. Steady-state kinetic data were fitted by non-linear least-squares
regression analysis. (b) In vitro selectivity of AW-091 towards other families of
proteases. Recombinant enzymes (10–40 nM final concentration) were incubated
l
S = 1.4 ꢀ 104 Mꢂ1 sꢂ1
,
for cathepsin L = 1.9 ꢀ 102 Mꢂ1 sꢂ1
,
no
with 60.0 lM AW-091 and fluorescence was measured. Cathepsin D belongs to the
family of aspartic proteases, cathepsin G to serine proteases, cathepsin S is a
cysteine proteases, whereas MMPs and ADAMTSs belong to the family of
metalloproteases.
turnover detected for cathepsins B and K; see Fig. 2a). The kinetic
properties of AW-091 show enhanced selectivity for cathepsin S
over cathepsin L, another cysteine cathepsin involved in inflamma-
tion, as compared to our previous cathepsin-selective ABP 4 (see
Table 1 and Ref. 4). Namely, the kcat/Km-value of AW-091 for
cathepsin S is ꢃ100-fold higher than for cathepsin L, whereas
ABP 4 showed only ꢃ2-fold preference for cathepsin S (Table 1).
The results thus imply that the newly developed probe AW-091
is cathepsin S-selective and could be used to monitor cathepsin S
activity in complex proteomic systems or in vivo models where
activity of cathepsin S has been found to be upregulated.
Before testing the applicability of AW-091 in vivo, we wanted to
rule out any proteolytic cleavages of AW-091 by unrelated families
of proteases. Selectivity of AW-091 was therefore tested in vitro
against cathepsins D and G, members of aspartic and serine prote-
ase families, respectively. As expected, none of the two non-cys-
teine cathepsins exhibited any proteolytic activity towards the
probe (Fig. 2b). In addition, cross-reactivity of AW-091 towards
metalloproteases was ruled out in a separate experiment using
matrix metalloproteinase-3 (MMP-3), MMP-8, MMP-9, and
MMP-14, which are involved in the degradation of the ECM in var-
ious pathologies including inflammation (reviewed in Ref. 14), as
none of the MMPs cleaved the fluorescent probe (Fig. 2b). Collec-
tively, these data demonstrated that AW-091 was cleaved prefer-
entially by cathepsin S and only to a minor extent by cathepsin
L, whereas unrelated serine cathepsin G and aspartic cathepsin D
did not exhibit any activity towards the probe. Moreover, as none
of the tested metalloproteases, which are involved in the patho-
genesis of inflammation, cleaved AW-091, this probe could serve
as a useful tool for in vivo imaging of cysteine cathepsin activity
with minimal, if any, cross-reactivity with unrelated classes of
proteases.
In vivo applicability of AW-091 was evaluated in a mouse mod-
el of inflammatory paw edema, introduced by unilateral, intraplan-
Figure 1. Conversion of the selective cathepsin inhibitor 1 into an activity-based probe AW-091 (molecule 2) by the ‘Reverse Design’. The quenched fluorescent probe is
designed for in vivo studies containing Cy7 and BHQ-3 as fluorophore/quencher pair.

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1059
Table 1
Kinetic parameters of AW-091 and ABP 4 for related cysteine cathepsins4
Cathepsin B
nd
Cathepsin K
Cathepsin L
Cathepsin S
AW-091
ABP 4
Km [M]
nd
nd
2.8 ꢀ 10ꢂ4
187
2.5 ꢀ 10ꢂ5
14,160
kcat/Km [Mꢂ1 sꢂ1
Km [M]
]
nd
3.2 ꢀ 10ꢂ3
332
8.9 ꢀ 10ꢂ5
187
8.4 ꢀ 10ꢂ4
7070
1.3 ꢀ 10ꢂ6
15,100
kcat/Km [Mꢂ1 sꢂ1
]
tar injection of zymosan.15 Zymosan induces a rapid chronic
inflammatory arthritis with mononuclear cell infiltration, synovial
hypertrophy, and pannus formation. Joint swelling and infiltration
of mononuclear cells appear already within 1 day after injection
and subside within 14 days. Infiltration of immune cells in turn re-
sults in release of lysosomal enzymes and subsequent local in-
crease in proteolytic activity.15 In addition, expression and
activity of cathepsin S, a protease with potent inflammatory activ-
ity, are known to be increased in various types of inflammatory
arthritis.10,16 Paw swelling in the ipsilateral hind paws was ob-
served within hours after zymosan injection, while the contralat-
eral paws served as controls. AW-091 was injected via tail vein
and animals were imaged using a charge-coupled device (CCD)
camera-based imaging system at various times after probe injec-
tion. Upon iv administration of the probe a differential fluores-
cence signal developed between ipsi- and contralateral paws
(Fig. 3). A differential signal was evident as early as 10 min after
probe application, reaching a ratio of 2–3 between ipsi- and con-
tralateral paws already 1 h after probe application and remained
above the signal in contralateral paws until 50 h after probe injec-
tion (Fig. 4a and c). Presence of BHQ-3 quencher group resulted in a
low background signal (i.e., very low signal was observed in the
contralateral paw).
Intensity, kinetics and tissue distribution of fluorescent signal
generated by cleaved AW-091 was compared with a commercially
available substrate ProSense680 (VisEn Medical). This substrate is
a polymer-based reagent, where the relatively high density of
loaded fluorophores onto the backbone structure leads to an inter-
nally quenched imaging agent.17 Two groups of mice (n = 3), pre-
treated with zymosan, were imaged shortly after probe injection
and at certain time intervals after injection. For AW-091 a differen-
tial signal between ipsi- and contralateral paws (ratio ipsi/contra)
was observed as early as 10 min after AW-091 application and
peaked at ꢃ180 min, whereas ProSense680 yielded the maximum
differential signal only after 24 h (Fig. 4). This faster signal onset fa-
vors AW-091 over ProSense680 for faster imaging and more fre-
quent activity determinations. However, overall signal recorded
Figure 4. Quantification of paw fluorescence in mice treated with AW-091 and
ProSense680. (a) 20 M AW-091 was injected iv through tail vein. (b) 13.3 lM
l
ProSense680 was injected iv through tail vein. Fluorescence (counts per second; cts/
s) from inflamed ipsilateral and non-inflamed contralateral paws was measured at
various time points after probe injection. Mean fluorescence and standard error is
plotted relative to time after probe injection. Background levels were around
1.3 ꢀ 105 cts/s at all times. (c) Quantification of signal-to-background ratios
expressed as ratios between fluorescent signals in ipsi- and contralateral paws for
each time point.
in ProSense680-injected mice was more intense than signal from
the cleaved AW-091 at all time points (see Fig. 4a and b for com-
parison). This higher signal probably stems from the higher density
of fluorophores loaded onto the substrate backbone resulting in
more intense fluorescence after cleavage by cathepsins and/or
from the prolonged tissue retention of the polymer-based
ProSense680 probe as compared to the smaller AW-091. Pro-
Sense680 substrate is cleaved by a range of cysteine cathepsins,
including cathepsins B, K, L, S, aspartic cathepsin D, serine cathep-
sin G, and other proteases such as plasmin, urokinase plasminogen
activator (uPA), CD10 and kallikrein-related peptidase 5 (http://
www.visenmedical.com/products/fluorescence_agents/activat-
able/technical_support/index.html; information obtained from the
VisEn Medical product datasheet), several of them involved in
Figure 3. Optical imaging in a zymosan-induced paw inflammation mouse model.
Signals from cleaved AW-091 and ProSense680 were compared using non-invasive
NIRF in vivo imaging. Shown are colorimetric fluorescence images of the inflamed
ipsilateral paws and the non-inflamed contralateral paws of three mice for each
probe (tail and abdomen covered; images were taken 6 h after iv injection of the
probe).

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1060
inflammation (for review see Refs. 10,18–20), whereas AW-091 is
preferentially cleaved by cathepsin S and exhibits no cleavage by
proteases belonging to other families (Fig. 2).
In the next set of experiments, we chose to test if the anti-
inflammatory agent dexamethasone could suppress inflammation
and consequently reduce fluorescent signal originating from the
cleavage of AW-091 in inflamed paw. Oral application of dexa-
methasone just before zymosan injection lowered the differential
signal between ipsi- and contralateral paws significantly, demon-
strating a downregulation of proteolytic activity in the inflamed
paws (Fig. 5). As AW-091 was shown to be cleaved only by cysteine
cathepsins, and not by other unrelated proteases, in vitro, we
tested if the fluorescent signal in zymosan-injected paws could
be suppressed by the general cysteine cathepsin inhibitor E-64.
The inhibitor was applied intraperitoneally 2 days and 1 day before
zymosan treatment, as well as during and after the treatment. Al-
Figure 6. Intraplantar injection of E-64 significantly reduced fluorescent signal in
inflamed paws. E-64 (1 l of 1 mM solution) was injected directly into the inflamed
paw and fluorescence from inflamed ipsilateral and non-inflamed contralateral
paws was measured at various time points after probe injection.
l
most no inhibition was observed when low (10
lg/kg) or high con-
lated with dexamethasone, enabling the use of this model for com-
pound testing. Intraplantarly-injected E-64 reduced the fluorescent
signal, suggesting a significant portion of the fluorescent signal
stemmed from cathepsin-mediated cleavage of AW-091. The abil-
ity to dynamically follow changes in proteolytic activity of cathep-
sin S suggests that AW-091 could be applied to monitor an
inflammatory process in vivo and provides a model system for
evaluation of small-molecule inhibitors against cathepsin S.
centration (30 g/kg) of E-64 was used (Fig. 5). These data are in
l
line with findings of Schurigt et al. who concluded that insignifi-
cant inhibition of cysteine cathepsins in mammary cancers and
in lungs resulted from the poor pharmacokinetic properties of
JPM-OEt, an E-64 derivative, and poor bioavailability when
JPM-OEt was administered intraperitoneally.21 Therefore, we
decided to test a different type of administration, where E-64
and AW-091 were administered directly into the inflamed paw.
This time, intraplantar injection of 1 ll 1 mM E-64 drastically de-
4. Conclusions
creased the signal of AW-091 (Fig. 6), suggesting that the observed
fluorescent signal was indeed cathepsin-specific. The fluorescent
signal from ipsilateral paw with injected E-64 (‘ipsi + E-64’ on
Fig. 6) was significantly lower than the fluorescent signal from
uninhibited inflamed paws (‘ipsi’ on Fig. 6) already 5 min after
injection of the inhibitor and remained below the level from the
uninhibited inflamed paw as long as 26 h after injection. The sig-
nificant reduction of fluorescent signal by the cell-impermeable
version of E-64 was anticipated as cathepsins are known to be
the dominantly secreted enzymes in inflammation (for review
see Ref. 10). Secreted cathepsins are thus expected to co-localize
with E-64, which in turn efficiently inhibits their proteolytic activ-
ity. However, the fluorescent signal could be significantly reduced
only when the inhibitor was administered directly into the in-
flamed paw, assuring that the inhibitor reached the site of the in-
creased proteolytic activity.
Near-infrared fluorophore (NIRF)-labeled imaging probes are
becoming increasingly important in bio-molecular imaging appli-
cations, that is, in animal models for tumor imaging or inflamma-
tion studies, due to the associated low auto-fluorescence of the
surrounding tissue and deeper tissue penetration. The design of
internally quenched protease probes allows imaging of enzyme
activities in vitro or in vivo in real time and offers valuable infor-
mation in monitoring and understanding the role of these enzymes
in living systems. Not only activities and regulations but also spe-
cific inhibition of proteolytic enzymes by small molecules can now
be made visible, providing new test modalities for medicinal chem-
istry research. The request for highly specific molecular imaging
tools has prompted us to extend the design principle of such
probes beyond peptide-based scaffolds and rather utilize the com-
bination of chemically optimized protease inhibitors out of medic-
inal chemistry research together with data from structural biology
for the design and synthesis of the presented fluorescent imaging
activity-based probe AW-091.
In this study we show with newly developed AW-091 that the
previously introduced chemical concept of ‘Reverse Design’ repre-
sents an efficient strategy for the generation of selective molecular
probes for cysteine proteases using chemically optimized protease
inhibitors. The concept thus allows the fine tuning of ABPs for
activity-based investigations in proteomic lysates as well as for
in vivo molecular imaging studies. Contrary to the known covalent
labeling ABPs, ‘Reverse Design’ ABPs are substrates of the target
enzyme. This property (i) renders them attractive for the monitor-
ing of low-abundant enzymes and (ii) may provide a basis to ulti-
mately establish quantitative (real time) investigations of the
activity and inhibition of proteases in pharmacologically relevant
in vivo models.
In conclusion, an animal model was found to provide increased
cysteine cathepsin activity which was successfully imaged with a
newly developed and characterized activity-based probe
AW-091. The imaged protease activity was localized in the foot
pad around the site of zymosan injection. The fluorescent signal
originating from the cleaved AW-091 was shown to be downregu-
In specific, we demonstrate that AW-091 is highly selective for
cathepsin S in vitro and proved useful in monitoring cysteine
cathepsin activity in vivo, that is, in zymosan-induced mouse
model of inflammation. AW-091 showed higher signal-to-back-
ground ratios at earlier time points than the commercially avail-
able polymer-based ProSense680 and thus represents an efficient
new tool for studying proteolytic processes leading to various dis-
Figure 5. Dexamethasone reverted zymosan-induced inflammation in murine
paws. Oral administration of 10 mg/kg dexamethasone (Dexamet.) prior to zymo-
san injections significantly suppressed the differential signal between ipsi- and
contralateral paws. Intraperitoneal administration of 10 mg/kg or 30 mg/kg E-64
only slightly reduced the fluorescent signal in inflamed paws.

ˇ
D. Caglic et al. / Bioorg. Med. Chem. 19 (2011) 1055–1061
1061
4. Watzke, A.; Kosec, G.; Kindermann, M.; Jeske, V.; Nestler, H. P.; Turk, V.; Turk,
B.; Wendt, K. U. Angew. Chem., Int. Ed. 2008, 47, 406.
5. Turk, B.; Turk, V. J. Biol. Chem. 2009, 284, 21783.
6. Nakagawa, T. Y.; Brissette, W. H.; Lira, P. D.; Griffiths, R. J.; Petrushova, N.;
Stock, J.; McNeish, J. D.; Eastman, S. E.; Howard, E. D.; Clarke, S. R.; Rosloniec, E.
F.; Elliott, E. A.; Rudensky, A. Y. Immunity 1999, 10, 207.
7. Link, J. O.; Zipfel, S. Curr. Opin. Drug Discov. Devel. 2006, 9, 471.
8. Riese, R. J.; Wolf, P. R.; Bromme, D.; Natkin, L. R.; Villadangos, J. A.; Ploegh, H. L.;
Chapman, H. A. Immunity 1996, 4, 357.
eases, including inflammation and rheumatoid arthritis. In addi-
tion, the fluorescent signal originating from the cleaved AW-091
was shown to be reduced by the administration of the anti-inflam-
matory drug dexamethasone and the cathepsin inhibitor E-64, pro-
viding
a valuable system for evaluation of small-molecule
inhibitors against cathepsin S.
9. Thurmond, R. L.; Sun, S.; Sehon, C. A.; Baker, S. M.; Cai, H.; Gu, Y.; Jiang, W.;
Riley, J. P.; Williams, K. N.; Edwards, J. P.; Karlsson, L. J. Pharmacol. Exp. Ther.
2004, 308, 268.
10. Vasiljeva, O.; Reinheckel, T.; Peters, C.; Turk, D.; Turk, V.; Turk, B. Curr. Pharm.
Des. 2007, 13, 387.
11. D’Alessio, K. J.; McQueney, M. S.; Brun, K. A.; Orsini, M. J.; Debouck, C. M. Protein
Expr. Purif. 1999, 15, 213.
12. Pauly, T. A.; Sulea, T.; Ammirati, M.; Sivaraman, J.; Danley, D. E.; Griffor, M. C.;
Kamath, A. V.; Wang, I. K.; Laird, E. R.; Seddon, A. P.; Menard, R.; Cygler, M.;
Rath, V. L. Biochemistry 2003, 42, 3203.
13. Turkenburg, J. P.; Lamers, M. B.; Brzozowski, A. M.; Wright, L. M.; Hubbard, R.
E.; Sturt, S. L.; Williams, D. H. Acta Crystallogr., D Biol. Crystallogr. 2002, 58, 451.
14. Tallant, C.; Marrero, A.; Gomis-Ruth, F. X. Biochim. Biophys. Acta 2010, 1803, 20.
15. Keystone, E. C.; Schorlemmer, H. U.; Pope, C.; Allison, A. C. Arthritis Rheum.
1977, 20, 1396.
Acknowledgements
This work was supported by the European Commission
Framework 6 Program (CAMP project, LSHG-2006-018830) and
the European Commission Framework 7 Program (FP7-Health-
2009-241919-LIVIMODE project) and by the grant P1-040 from
the Slovenian Research Agency to B.T. The authors declare no
conflict of interest.
Supplementary data
Supplementary data (Supplementary Fig. S1 showing LC–MS
spectrum of AW-091) associated with this article can be found in
the online version, at doi:10.1016/j.bmc.2010.10.028.
16. Pozˇgan, U.; Caglicˇ, D.; Rozman, B.; Nagase, H.; Turk, V.; Turk, B. Biol. Chem.
2010, 391, 571.
17. Weissleder, R.; Tung, C. H.; Mahmood, U.; Bogdanov, A., Jr. Nat. Biotechnol.
1999, 17, 375.
18. Blasi, F.; Sidenius, N. FEBS Lett. 2010, 584, 1923.
19. Henke, P. K. Vascular 2007, 15, 366.
20. Heutinck, K. M.; ten Berge, I. J.; Hack, C. E.; Hamann, J.; Rowshani, A. T. Mol.
Immunol. 2010, 47, 1943.
21. Schurigt, U.; Sevenich, L.; Vannier, C.; Gajda, M.; Schwinde, A.; Werner, F.;
Stahl, A.; von Elverfeldt, D.; Becker, A. K.; Bogyo, M.; Peters, C.; Reinheckel, T.
Biol. Chem. 2008, 389, 1067.
References and notes
1. Evans, M. J.; Cravatt, B. F. Chem. Rev. 2006, 106, 3279.
2. Blum, G.; Mullins, S. R.; Keren, K.; Fonovic´, M.; Jedeszko, C.; Rice, M. J.; Sloane,
B. F.; Bogyo, M. Nat. Chem. Biol. 2005, 1, 203.
3. Blum, G.; von Degenfeld, G.; Merchant, M. J.; Blau, H. M.; Bogyo, M. Nat. Chem.
Biol. 2007, 3, 668.