Cathepsin B inhibitors: Further exploration of the nitroxoline core

Article abstract of DOI:10.1016/j.bmcl.2018.02.042

Human cathepsin B is a cysteine protease with many house-keeping functions, such as intracellular proteolysis within lysosomes. Its increased activity and expression have been strongly associated with many pathological processes, including cancers. We present here the design and synthesis of novel derivatives of nitroxoline as inhibitors of cathepsin B. These were prepared either by omitting the pyridine part, or by modifying positions 2, 7, and 8 of nitroxoline. All compounds were evaluated for their ability to inhibit endopeptidase and exopeptidase activities of cathepsin B. For the most promising inhibitors, the ability to reduce extracellular and intracellular collagen IV degradation was determined, followed by their evaluation in cell-based in vitro models of tumor invasion. The presented data show that we have further defined the structural requirements for cathepsin B inhibition by nitroxoline derivatives and provided additional knowledge that could lead to non-peptidic compounds with usefulness against tumor progression.

Full text of DOI:10.1016/j.bmcl.2018.02.042

Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Contents lists available at ScienceDirect
Bioorganic & Medicinal Chemistry Letters
journal homepage: www.elsevier.com/locate/bmcl
Cathepsin B inhibitors: Further exploration of the nitroxoline core
a
b
a
b
b
ˇ a
´
´
´
ˇ
Izidor Sosic , Ana Mitrovic , Hrvoje Curic , Damijan Knez , Helena Brodnik Zugelj , Bogdan Štefane ,
Janko Kos a,c, Stanislav Gobec a,
⇑
a
ˇ
Faculty of Pharmacy, University of Ljubljana, Aškerceva 7, 1000 Ljubljana, Slovenia
Faculty of Chemistry and Chemical Technology, University of Ljubljana, Vecna pot 113, 1000 Ljubljana, Slovenia
Department of Biotechnology, Jozef Stefan Institute, Jamova 39, 1000 Ljubljana, Slovenia
b
c
ˇ
ˇ
a r t i c l e i n f o
a b s t r a c t
Article history:
Human cathepsin B is a cysteine protease with many house-keeping functions, such as intracellular pro-
teolysis within lysosomes. Its increased activity and expression have been strongly associated with many
pathological processes, including cancers. We present here the design and synthesis of novel derivatives
of nitroxoline as inhibitors of cathepsin B. These were prepared either by omitting the pyridine part, or by
modifying positions 2, 7, and 8 of nitroxoline. All compounds were evaluated for their ability to inhibit
endopeptidase and exopeptidase activities of cathepsin B. For the most promising inhibitors, the ability
to reduce extracellular and intracellular collagen IV degradation was determined, followed by their eval-
uation in cell-based in vitro models of tumor invasion. The presented data show that we have further
defined the structural requirements for cathepsin B inhibition by nitroxoline derivatives and provided
additional knowledge that could lead to non-peptidic compounds with usefulness against tumor
progression.
Received 12 December 2017
Revised 16 January 2018
Accepted 21 February 2018
Available online xxxx
Keywords:
Human cathepsin B
Nitroxoline derivatives
Structure-activity relationships
Tumor invasion
Ó 2018 Elsevier Ltd. All rights reserved.
Human cathepsin B (catB, EC 3.4.22.1) is an intracellular lysoso-
mal cysteine protease that is ubiquitously expressed in many tis-
sues and is involved in a number of physiological processes.1 The
enzyme possesses an endopeptidase and a dipeptidyl carboxypep-
tidase activity,2 a unique feature for cysteine cathepsins that is
attributed to the presence of the occluding loop.3 This is a flexible
20 amino acid insertion that in the so-called closed (i.e., exopepti-
dase) conformation spans from the left domain to cover the primed
subsites S30 and S20 of the active site cleft and prevents the access
of large endopeptidase substrates.4,5 The occluding loop is held in
this conformation via two salt bridges, His110-Asp22 and
Arg116-Asp224. Additionally, two histidine residues (His110 and
His111) are located at its tip providing positively charged anchors
for the negatively charged C-terminal carboxyl group of the sub-
strates.4,6 Besides the well-known exopeptidase activity that has
a pH optimum of around 5.0,7 the endopeptidase function of catB
greatly increases when the ionic contacts that bind the loop to
the body of the enzyme are weakened.6 It has been shown that
the endopeptidase activity of catB increases with a rising pH value,
reaching its maximum at neutral pH.7
CatB can actively participate in the majority of processes that
significantly contribute to cancer progression. Particularly, it can
modify the tumor microenvironment through the turnover and
degradation of the extracellular matrix (ECM) either directly via
proteolytic degradation of its components or indirectly via activa-
tion or amplification of other proteases in the proteolytic cas-
cade.8,9 This ECM breakdown is a crucial step that promotes
tumor invasion, and enables angiogenesis and metastasis.10,11
The degradation of metalloprotease inhibitors and the release of
growth factors that are bound to the ECM components are two
additional mechanisms through which catB contributes to angio-
genesis.12–14 CatB was also found to have a role in chemotherapy
resistance, as it was shown that lysosomal leakage of catB, caused
by chemotherapeutics 5-florouracil and gemcitabine, activates the
Nlrp3 inflammasome and promotes tumor growth.15 Importantly,
high pharmacological relevance of catB has also been established
in various tumor mouse models using catB-deficient mice16–20 ren-
dering catB as a validated and druggable target for the design of
new anti-tumor drugs.
Abbreviations: Abz, 2-aminobenzoyl; AMC, 7-amido-4-methylcoumarin; catB,
cathepsin B; DIPEA, N,N-diisopropylethylamine; DMSO, dimethyl sulfoxide; ECM,
extracellular matrix; EDC, 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide; HATU,
1-[bis(dimethylamino)methylene]-1H-1,2,3-triazolo[4,5-b]pyridinium
3-oxid
hexafluorophosphate; HOBt, hydroxybenzotriazole; Ki, inhibition constant;
TBTU, O-(benzotriazol-1-yl)-N,N,N0,N0-tetramethyluronium tetrafluoroborate; Z,
benzyloxycarbonyl.
⇑
Corresponding author.
E-mail address: stanislav.gobec@ffa.uni-lj.si (S. Gobec).
https://doi.org/10.1016/j.bmcl.2018.02.042
0960-894X/Ó 2018 Elsevier Ltd. All rights reserved.
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
2
Several types of exogenous inhibitors of catB have been identi-
in our previous series of 7-aminomethylated derivatives,28,30 we
focused also on amidoacetonitriles. These nitrile-based sub-
stituents are very often used as mildly electrophilic warheads in
inhibitors of different cathepsins.22 Furthermore, some of the most
recent and the most potent catB inhibitors known to date possess
substituents of this exact nature.23 Next, we synthesized a variety
of compounds with nitrile-based substituents, i.e., either
aminoacetonitriles or amidoacetonitriles at position 2 of the 8-
hydroxy-5-nitroquinoline scaffold. The underlying reason for the
preparation of these compounds was based on our recent molecu-
lar dynamics studies of the binding of nitroxoline and its deriva-
tives into the active site of catB.32 These results indicated that
the quinoline ring can rotate around the axis, which is represented
by the quinoline-NO2 bond; a characteristic that is not evident
from the crystal structure of the complex.28 In addition, docking
of a representative 2-amidoacetonitrile substituted 8-hydroxy-5-
nitroquinoline 22 further corroborated this approach (Figs. S1
and S2 in the Supplementary data). Thus, interactions of the nitrile
group of substituents at position 2 of the nitroxoline core with the
catalytic cysteine are also possible. Third, a series of truncated
compounds was prepared, in which the pyridine moiety was omit-
ted. Within this group of derivatives, the same substituents as in
the 7-carboxamido group were used and all were appended at
position ortho with respect to the hydroxyl group. By preparing
truncated derivatives we wanted to determine if the ortho substi-
tuted 4-nitrophenol represents a sufficient pharmacophore to be
recognized by catB and to enable binding in its active site. Finally,
we prepared a variety of compounds with diverse substituents
attached directly to position 8 of the 5-nitroquinoline ring. Despite
knowing the potential toxicity issues associated with NO2-substi-
tuted aryl derivatives, this functionality was present in the synthe-
sized molecules from all four classes. We established
previously28,30 that NO2 group is a prerequisite for the inhibition
of catB as it interacts with two histidines, His110 and His111, in
the occluding loop of catB.
fied and the majority of them have peptidyl backbones that contain
an electrophilic functionality in the position of the scissile peptide
bond. Various electrophilic warheads were explored in preclinical
studies and these form either an irreversible (epoxysuccinyl, vinyl
sulfone, acyloxymethyl ketone) or a reversible (ketone, nitrile)
covalent bond with the catalytic cysteine in the active site.21 Due
to their low reactivities towards other cellular nucleophiles, the
nitrile-containing inhibitors are receiving the most attention in
the development of inhibitors of catB,22 as well as other cysteine
cathepsins.23,24 The peptidic nature of currently available inhibi-
tors can be a cause of low bioavailability and poor metabolic stabil-
ity, which limit the use of a best part of such compounds to
research only.25 Given the fact that researchers involved in this
field continue to appreciate the importance of cathepsins in dis-
ease management26 and their involvement in cancer progression,27
there is a substantial need for the advances towards new catB-
inhibiting scaffolds that can bypass the limitations of peptidic catB
inhibitors.
Our work in the field of catB inhibitors has been focused on
nitroxoline and its derivatives and is summarized in Fig. 1. On
the basis of these previous results, we herein report a complemen-
tary and focused set of compounds used to further explore the
chemical space and the structure-activity relationships (SARs) of
nitroxoline-based derivatives. Our efforts have resulted in several
compounds that inhibit both catB activities in the low micromolar
range, inhibit degradation of the ECM, and concurrently reduce
invasiveness in cell-based in vitro models of tumor invasion.
The fragment-like characteristics of the 5-nitro-8-hydrox-
yquinoline core of nitroxoline enable numerous possibilities for
structural elaboration. In our design, we pursued four different
modifications to nitroxoline (Fig. 2).
First, several different 7-carboxamido substituted derivatives
were prepared. Besides incorporating substituents that were used
The nitroxoline carboxylic acid 130,33 was used as substrate for
the preparation of 7-carboxamido substituted nitroxoline deriva-
tives 2–14 by modification of the carboxylic acid via an acid chlo-
ride (Scheme 1: rectangle A). Interestingly, all attempts to
synthesize these derivatives using coupling reagents were unsuc-
cessful. To prepare 2-substituted derivatives, five different com-
mercially available quinolinol derivatives were used as starting
material (Scheme 1: rectangle B). An efficient nitration procedure
was applied in the first step using either a mixture of KNO3/97%
H2SO4 in acetic acid (for compounds 15 and 19), 65% HNO3 in
acetic acid (for compounds 16 and 17), or a mixture of 65%
HNO3/97% H2SO4 in acetic acid (for 18). The regioselective intro-
duction of the NO2 group was confirmed by two-dimensional
NMR experiments, such as homonuclear correlation spectroscopy
and Nuclear Overhauser effect spectroscopy. The 2-aminoacetoni-
trile derivative 20 was synthesized from 18 by reductive amination
procedure using Na(OAc)3BH as a reducing agent, whereas the ami-
doacetonitriles 21–24 were prepared by HATU-mediated coupling
of 2-carboxy-8-hydroxy-5-nitroquinoline (19) and the correspond-
ing amines (Scheme 1: rectangle B). Of note, several 2-substituted-
5,7-dinitro derivatives were prepared (compounds S1–S6,
Scheme S1 in the Supplementary data) and assayed for inhibition
of catB.
For the 2-aminomethylated 4-nitrophenols 25–31, either Man-
nich reaction conditions (25–29) or reductive amination (30 and
31) were used. The 2-carboxamide 4-nitrophenols 33–39 were
synthesized via nitration of the salicylic acid to obtain intermedi-
ate 32, followed by the reaction with SOCl2 to generate the acid
chloride in situ. Different amines were used subsequently for the
formation of an amide bond in good overall yields (Scheme 1: rect-
angle C).
Fig. 1. Nitroxoline (5-nitro-8-hydroxyquinoline) and its endopeptidase and
exopeptidase inhibition of catB.28,29 Two representative examples of a first set of
optimized derivatives30,31 are also represented.
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
3
Fig. 2. Schematic representation of structural modifications to nitroxoline.
Scheme 1. Synthesis of 7-Carboxamido Substituted Derivatives (A), 2-Substituted Derivatives (B), Truncated Nitroxolines (C), and 8-Substituted Derivatives (D). aReagents
and conditions: (a) 1. SOCl2, toluene, 80 °C, 4 h; 2. corresponding amine (R1R2NH, Table S1), THF, 0 °C to rt, 24 h; (b) KNO3/97% H2SO4, CH3COOH, 16 °C to rt, 1 h (for
compounds 15 and 19); 65% HNO3, CH3COOH, 16 °C to rt, 3 h (for 16) or 24 h (for 17); 65% HNO3/97% H2SO4, CH3COOH, 15–25 °C, 1 h (for 18) (c) methylaminoacetonitrile
hydrochloride, Na(OAc)3BH, DIPEA, 1,2-dichloroethane, 70 °C, 24 h; (d) corresponding amine (R1R2NH, Table S2), HATU, DIPEA, CH3CN, rt, 24 h; (e) corresponding amine
(R1R2NH, Table S3), HCHO, 2-propanol, 85 °C, 24 h; (f) corresponding amine (R1R2NH, Table S3), Na(OAc)3BH, DIPEA, 1,2-dichloroethane, rt, 24 h; (g) HNO3, CH3COOH, 50 °C,
15 min; (h) KNO3/97% H2SO4, 0 °C to rt, 1 h; (i) 97% H2SO4, K2Cr2O7, 80 °C, 1 h; (j) 1. ethyl chloroformate, Et3N, THF, ꢀ10 °C; 2. NH3 (g), rt, 1 h; (k) POCl3/imidazole, pyridine,
ꢀ30 °C, 30 min; (l) aminoacetonitrile hydrochloride, EDC, HOBt, Et3N, DMF rt, 24 h; (m) Ar-B(OH)2, Pd(PPh3)4, Cs2CO3, toluene, 100 °C, 24 h (for 46–49) or Ar-B(OH)2, Pd
(OAc)2, PPh3, K2CO3, 1,4-dioxane/H2O (4/1, v/v), 100 °C, 24 h (for 50 and 51) or HetAr–B(OH)2, Pd(OAc)2, PPh3, K2CO3, 1,4-dioxane/H2O (4/1, v/v), 100 °C, 24 h (for 52–59); (n)
Ar–Br, Pd(OAc)2, 120 °C, 24 h.
Diverse synthetic approaches were used to obtain 8-substituted
nitroxoline analogs (Scheme 1: rectangle D). The preparation of
compounds in which position 8 was substituted with a carboxylic
acid derivative (compounds 41–44) started from 8-methylquino-
line. The most appropriate nitration reagent to successfully obtain
compound 40 in a good yield (80%) was found to be an excess of
KNO3 in 97% H2SO4. The 8-carboxy nitroxoline derivative 41 was
prepared by oxidation of the aromatic methyl group of 40 using
potassium dichromate in 97% H2SO4. The 8-carboxamide derivative
42 was synthesized via generation of mixed anhydride by using
ethyl chloroformate, followed by bubbling of the reaction mixture
with ammonia. For the dehydration of the primary amide of 42 and
formation 8-cyano-5-nitroquinoline (43), the standard POCl3/imi-
dazole mixture was utilized. Compound 44 was prepared by
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
4
EDC-mediated coupling of 41 with aminoacetonitrile hydrochlo-
ride. To prepare a series of 8-(hetero)aryl substituted quinolines
46–67, an efficient methodology was developed by the Štefane
group.34 Compounds 46–59 were synthesized by Pd-catalyzed cou-
pling of organoboron reagents with 8-bromo-5-nitroquinoline
(45),34 whereas derivatives 60–67 were prepared by an original
one-pot sequential Suzuki-Miyaura cross-coupling/direct CAH ary-
lation reaction (Scheme 1: rectangle D). This methodology enabled
rapid access to a variety of polysubstituted molecular architectures
at position 8 of the quinoline ring.
The biochemical evaluation of the synthesized compounds was
initiated with the determination of their relative inhibition of catB
endopeptidase and exopeptidase activities, using the catB specific
substrates Z-Arg-Arg-7-amido-4-methylcoumarin (Z-Arg-Arg-
AMC) and 2-aminobenzoyl (Abz)-Gly-Ile-Val-Arg-Ala-Lys(Dnp)A
OH, respectively (results not shown). The compounds that showed
two compounds showed similar constants of inhibition for both
catB activities as were determined for nitroxoline (Table S2 and
Fig. S1). The nitrile-containing compounds 20–24 were poor inhi-
bitors of both catB activities, with only compound 22 showing
improved inhibition (endopeptidase activity: Ki0 = 34 1 mM;
exopeptidase activity: Ki = 156 48 mM, Ki0 = 18 5 mM, Table 1)
in comparison with nitroxoline.
Unsurprisingly, the majority of compounds from the truncated
series inhibited both catB activities poorly (Table S3), which indi-
cates that both rings are needed for efficient binding into the active
site of catB. Compounds 27 and 37 (Table S3) did, however, inhibit
the exopeptidase activity with Ki values below 100 mM rendering
these two molecules as interesting catB-inhibiting fragments.
In a previous study,30 we showed that 4-methylpiperidine moi-
ety attached at position 8 of the 5-nitroquinoline ring, results in
increased exopeptidase activity inhibition in comparison with
nitroxoline. Interestingly, the 8-substituted derivatives presented
herein showed a particularly weak or even absence of inhibition
of catB (Table S4). Only compounds 61 and 66 are of interest from
this series, as they showed uncompetitive inhibition of both catB
activities in the low micromolar range (Table 1). It is important
to note here that the SARs of compounds 60–63 (Table S4) seem
flat, because very small structural changes led to significant
decrease in inhibitory activities.
Addition of an NO2 group at position 7 of 15 (compound S1,
Table S5) resulted in a loss of catB inhibitory activity, whereas
there were no significant differences in inhibitory properties
between compound 16 and its 5,7-dinitro analogue S2 (Tables 1
and S5). The most potent dinitro nitroxoline derivative was com-
pound S5, with a Ki0 of 9 1 mM for endopeptidase activity and Ki
and Ki0 of 154 23 mM and 35 12 mM, respectively, for exopepti-
dase activity (Table 1).
Recently, it was shown that the pH value of the assay buffer
plays a more important role in the measurement of inhibitory
activities in comparison with the choice of the substrate, i.e., the
latter did not strongly affect the equilibrium between the
endopeptidase (open) or exopeptidase (closed) form of catB.22 To
address the effect of pH on the inhibitory activity and to shift the
equilibrium of catB to the closed conformation, we assayed the
compounds from Table 1 using the ‘small’ endopeptidase substrate
Z-Arg-Arg-AMC at pH 4.5 (Table S6). Given the fact that com-
pounds presented in Table 1 are small in size (with exceptions of
9 and 10), we were expecting improvements in inhibitory activities
as a result of the proximity of histidines in the occluding loop
(where the NO2 group of inhibitors binds28) and other enzyme’s
binding subsites. In contrast to our expectations and for reasons
currently unknown, all inhibitors exhibited higher Ki values under
acidic assay conditions. The most striking difference was observed
for compounds 9 and 10, which showed a 30-fold reduction in Ki
values (9: 8.2 mM at pH 6 versus 237 mM at pH 4.5; 10: 12.0 mM
at pH 6 versus 322 mM at pH 4.5; Table S6).
relative inhibitions similar to nitroxoline or better at 50 lM
(endopeptidase activity, 21.9% 3.3%; exopeptidase activity,
24.7% 2.8%, Fig. 1) were further characterized by determining
their kinetic parameters and their mode of inhibition. In Table 1,
a selection of the most potent derivatives with characterized mode
of inhibition is represented; whereas the assay results for all other
compounds are shown in the Supplementary data (Tables S1–S5).
Ki values represent inhibition constants for the dissociation of
the enzyme-inhibitor complex, whereas the Ki0 values are inhibi-
tion constants for the dissociation of the enzyme-substrate-inhibi-
tor complex.
Based on the data given in Tables 1 and in S1–S5, the structural
features needed for inhibition can be deduced. In order not to over-
interpret the SAR data, only the important findings are empha-
sized. Most of the 7-carboxamide derivatives 2–14 that showed
catB inhibitory activity were either uncompetitive inhibitors of
catB endopeptidase activity or showed mixed type of inhibition
with a predominantly uncompetitive component. Very similar data
was obtained when Abz-Gly-Ile-Val-Arg-Ala-Lys(Dnp)AOH was
used as a substrate and the influence on the exopeptidase activity
of catB was addressed, as these compounds showed the same
mode of inhibition and in the same concentration range (Tables 1
and S1). In comparison with their 7-aminomethylated counter-
parts,28,30 compounds 2–4 showed very similar inhibitory proper-
ties, whereas compounds 5–8 showed 2–10-fold decrease in
inhibition of the catB endopeptidase activity. Compound 7, on
the other hand, proved to be a better exopeptidase activity inhibi-
tor than its aminomethylated analogue (Ki = 426 39
constant of inhibition in the low micromolar range (Ki = 128 21
M, Ki0 = 30 10
M, Table 1). The assay results for compounds
l
M30) with
l
l
5–8 are indicating that the interaction of their mildly electrophilic
amidoacetonitrile warhead with the catalytic cysteine did not take
place, as the literature data show that the Ki values are substan-
tially lower when the covalent interaction occurs.23 The most
promising compounds from this series proved to be compounds
9 and 10 with a larger side chain comprised of 1,3-substituted
piperidine core at position 7 (Table 1). When using Z-Arg-Arg-
AMC as a substrate, compounds 9 and 10 acted as uncompetitive
Our next step was to explore the effects of selected inhibitors (6–
10, 15, and S5) on the proteolysis of the ECM and tumor cell inva-
sion in vitro. CatB is substantially involved in the progression of can-
cer, as it was shown that extracellular and intracellular catB
activities contribute substantially to the turnover of ECM.8,11,35 To
investigate the impact of inhibitors on ECM degradation, we evalu-
ated their effects on the degradation of DQ-collagen IV by MCF-10A
neoT cells. Collagen type IV is a major constituent of the ECM that
can be fluorescently tagged and gives after enzymatic cleavage
bright green fluorescence. Moreover, we have shown previously
that MCF-10A neoT cells degrade DQ-collagen IV inside cells and
extracellularly.29 To quantify the extracellular and intracellular
degradation of the DQ-collagen IV, spectrofluorimetry and flow
cytometry were employed (Fig. 3). Compounds 6 (5 mM), 9 (1.25
mM), and 10 (1.25 mM) showed the most potent reduction of the
inhibitors of catB, with Ki0 of 8
0
l
M and 12
1 lM, respectively.
In addition, both compounds inhibited exopeptidase activity of
catB with the same potency and in the same mode (9: Ki0 = 13
0
l
M, 10: Ki0 = 18
2 lM; Table 1). Interestingly, when we intro-
duced the 2,3-dihydro-1H-indene (compound 11) instead of the
benzoxazole (9) or benzothiazole (10) on the piperidine nitrogen,
no inhibition was observed. The same occurred when the methy-
lene group was inserted between the piperidine and aromatic
bicyclic system (compounds 12–14, Table S1).
Small substituents at position 2 led to diminished inhibitory
activity in comparison with nitroxoline, except for compounds 15
and 16 with Me and CN moiety, respectively, at position 2. These
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
5
Table 1
Cathepsin B inhibitory activities of selected nitroxoline derivatives.
Z-Arg-Arg-AMC
Abz-Gly-Ile-Val-Arg-Ala-Lys(Dnp)AOH
Ki
Ki0
Ki
Ki0
(l
M)a
(l
M)a
(l
M)a
(l
M)a
7-Carboxamido substituted derivatives
6
182 31b
64 13b
185 5b
32 7b
7
–
67 10c
128 21b
30 10b
8
9
–
–
124 14c
185 2d
8
0c
–
13 0c
10
–
12 1c
–
18 2c
2-Substituted derivatives
15
181 14b
99 18b
469 58b
77 20b
16
22
142 5b
68 0b
214 30d
156 48b
–
34 1c
18 5b
8-Substituted derivatives
44
240 35e
–
190 3d
49f
323 3d
118 0b
22 0b
61f
–
46 16c
–
11 1c
(continued on next page)
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
6
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
Table 1 (continued)
Z-Arg-Arg-AMC
Abz-Gly-Ile-Val-Arg-Ala-Lys(Dnp)AOH
Ki
Ki0
Ki
Ki0
(l
M)a
(
l
M)a
39 3c
(l
M)a
(l
M)a
66f
–
–
13 8c
2-Substituted-5,7-dinitro derivatives
S5
–
9
1c
154 23b
35 12b
S6
160 42d
201 89b
34 8b
a
Ki and Ki0 values are the mean SD (n = 3).
Mixed inhibition.
Uncompetitive inhibition.
Noncompetitive inhibition.
Competitive inhibition.
Poorly soluble compounds.
b
c
d
e
f
Fig. 3. Nitroxoline derivatives inhibit extracellular and intracellular DQ-collagen IV degradation. (A) Extracellular degradation of DQ-collagen IV by MCF-10A neoT cells (5 ꢁ
104) in the presence of DMSO or respective compound (5 mM for compounds 6, 7, 8, S5, and CA-074, 2.5 mM for compounds 15, or 1.25 mM for compounds 9 and 10) was
analyzed by monitoring the fluorescence intensity of extracellular degradation product using spectrofluorimetry. Data are presented as the mean SEM (n = 3) and
experiments were performed in six replicates. (B) Intracellular DQ-collagen IV degradation by MCF-10A neoT cells (6 ꢁ 104) after treatment with DMSO (solid black line) or
respective compound (50 mM for compounds 6, 7, 8, S5, CA-074, and CA-074Me, 25 mM for compound 15, or 12.5 mM for compounds 9 and 10, dotted black line) was
monitored using flow cytometry. Gray histograms denote unlabeled cells. (C) Reduction of intracellular DQ-collagen IV in the presence of DMSO or suitable compound (50 mM
for compounds 6, 7, 8, S5, CA-074, and CA-074Me, 25 mM for compound 15, or 12.5 mM for compounds 9 and 10) as assayed by flow cytometry. Data are presented as the
mean SEM (n = 3) and the experiments were performed in duplicates. *P < 0.05, **P < 0.01, ***P < 0.001.
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
7
Fig. 4. Impact of compounds 7–9 and 15 on invasion of MCF-10A neoT cells monitored in real-time. (A) MCF-10A neoT cells (3 ꢁ 104) were seeded on top of Matrigel (1 mg/
mL) coated upper compartments of CIM-plate 16. DMSO (0.05%) or respective compound (5 mM for compounds 7, 8, and CA-074, 2.5 mM for compounds 15, or 1.25 mM for
compound 9) was added to the growth medium in the upper and lower compartments of the CIM-plate 16. Cell invasion was monitored continuously for 72 h by measuring
impedance data (reported as CI) using the xCELLigence system. (B) The slopes (1/h) in the time interval between 4 and 14 h correlated with the ability of the cells to invade
and were used to calculate the percentage of invasion (%), presented as means SEM (n = 2). The experiments were performed in triplicates. *P < 0.05, **P < 0.01, ***P < 0.001.
extracellular DQ-collagen IV degradation by ꢂ 60%, ꢂ45%, and ꢂ
40%, respectively, whereas other compounds (7, 8, 15 and S5)
showed less pronounced effects (Fig. 3A). The results from intracel-
lular degradation assay showed that compounds 7 (50 mM), 8 (50
mM), 9 (12.5 mM), and 15 (25 mM) proved to be the most promising,
as they inhibited intracellular DQ-collagen IV degradation by ꢂ60%,
ꢂ15%, ꢂ35%, and ꢂ80%, respectively (Figs. 3B and 3C). Compounds
6 (50 mM), 10 (12.5 mM), and S5 (50 mM) inhibited this process by
less than 10% (6 and 10) or showed statistically non-significant
results (S5) (Figs. 3B and 3C). Similarly as determined previously,36
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
8
Fig. 5. Compound 9 impairs invasion of U-87 MG cells in a three-dimensional in vitro model of tumor invasion. (A) U-87 MG spheroids were implanted in Matrigel (5 mg/mL)
and covered with growth medium, both containing inhibitor 9 (1.25 mM) or DMSO (0.05%) as a control. The spheroid dimensions were measured for up to three days and the
spheroid volume was calculated. Data are presented as means SEM (n = 2). (B) Representative images of U-87 MG spheroids obtained at day three after implantation. Scale
bar, 100 mm. **P < 0.01, ***P < 0.001.
only the cell-permeable ester analogue of the positive control, CA-
074Me (50 mM), impaired intracellular degradation of the ECM by
ꢂ55%, whereas CA-074 (50 mM) failed to show activity inside cells.
The inhibition of both extracellular and intracellular catB is
needed for significant attenuation of tumor cell invasion.37 There-
fore, compounds 7–9 and 15 that showed cell-permeable charac-
teristics were selected to assess their impact on tumor-cell
invasion. As a first method, the invasion of tumor cells through
Matrigel, as a model of the ECM, was monitored in real time using
the xCELLigence system.38 This two-chamber-based system mea-
sures electrical impedance, expressed as cell index, across micro-
electrodes integrated in the membrane separating the top and
bottom compartments of tissue culture CIM-plates (Fig. 4A). In this
assay, compound 8 (5 mM) significantly abrogated the invasion of
MCF-10A neoT cells by ꢂ70% compared to the DMSO control
(Fig. 4B), followed by compound 9 (1.25 mM), which showed
ꢂ50% inhibition of the invasion. Interestingly, compound 15 was
not active despite showing potent inhibition of intracellular degra-
dation of DQ-collagen IV. Compound 7, on the other hand, which
was also effective intracellularly, reduced invasion by ꢂ20% (Fig.
4B); probably as a result of better inhibition of the extracellular
ECM degradation in comparison with 15. For comparison, an irre-
versible catB-specific inhibitor CA-074, which was used as a con-
trol compound, inhibited cell invasion by ꢂ40%.
during tumor progression16,27 could, in part, explain very potent
effects of compound 9 in both cell-based tumor models at concen-
trations that are lower than its constants of inhibition on isolated
enzymes. In addition, we cannot exclude possible effects on other
targets, such as type 2 methionine aminopeptidase and sirtuin,
which were identified to be targets for nitroxoline.40 Similarly,
compounds 7 and 8 also showed a preference for catB inhibition
over catH, but not over catL (Table S7), whereas compound 15
was non-selective, as it inhibited all three enzymes in a similar
concentration range, i.e., from approximately 80 to 120 mM.
To exclude the possibility that the inhibition of ECM degrada-
tion and tumor cell invasion resulted from the compound-induced
cell-toxicity, all effects of compounds in functional assays (6–10,
15, and S5) were studied at concentrations, which did not cause
a decrease in cell viability. These concentrations were selected by
performing cell viability studies that are presented in the Supple-
mentary data (Fig. S3).
In summary, we prepared a set of more than 60 nitroxoline-
based derivatives to complement previous results and to further
explore their SARs as a catB-inhibiting core. Despite the fact that
significant improvement in in vitro catB inhibition was not
achieved in comparison with previous results30 several derivatives
showed promising efficacy in tumor-cell-based assays, where they
inhibited extracellular and intracellular degradation of ECM and
tumor-cell invasion. Moreover, the fragment-like characteristics
of the majority of compounds represent solid foundation for fur-
ther ligand- and structure-based optimization of nitroxoline
derivatives as catB inhibitors (see also ligand efficiency changes
in comparison with previous compounds, Table S8). To achieve a
significant leap in the inhibitory activity, additional optimizations
of the substituents at positions 2 and 7 are needed, followed by
proper positioning of the mildly electrophilic functional group to
form reversible covalent interaction with the catalytic cysteine.
Furthermore, incorporation of other cysteine-selective elec-
trophilic warheads is possible and will represent a subject of our
following endeavors in the field of nitroxoline-based catB
inhibitors.
To further corroborate the results from two-dimensional assay,
compound 9 was analyzed by an in vitro three-dimensional inva-
sion assay. Multicellular tumor spheroids (MCTS) from U-87 MG
cells were prepared according to the hanging drop method39 and
implanted in ECM-mimicking matrix (Matrigel). For U-87 MG cells,
we have previously shown that they express high levels of catB,
efficiently form MCTS, and invade Matrigel.29 MCTS more closely
represent a tumor in vivo by mimicking the early, avascular stages
of tumor growth and are therefore, more representative as a model
of tumor cell invasion compared to those involving cell monolay-
ers.39 During the three day experiment, the U-87 MG MCTS showed
steady growth (Fig. 5A) and formed radially invasive strands
around the spheroid (Fig. 5B). Addition of compound 9 (1.25 mM)
significantly impaired tumor growth in a three-dimensional exper-
imental setting (Fig. 5A) confirming its encouraging anti-tumor
characteristics.
Acknowledgments
The authors thank all undergraduate students that helped with the
syntheses of compounds.
For compounds 7–9 and 15 that were assayed in models of
tumor invasion, selectivity against cathepsin H (catH) and cathep-
sin L (catL) was evaluated (Table S7). These results showed that
compound
9 had the same affinity for catL (Ki = 5 0 mM,
Funding sources
Table S7) as for catB (endopeptidase activity: Ki0 = 8 0 mM;
exopeptidase activity: Ki0 = 13 0 mM, Table 1), whereas lower
preference was shown for catH (Ki = 108 26 mM, Table S7). This
fact along with evidence that catL is involved in proteolytic events
This work was supported by the Slovenian Research Agency:
research core funding No. P1-0208, grant number J4-5529 (D)
to J.K., and grant number Z1-7181 (B) to I.S.: Optimization of
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042

ˇ
I. Sosic et al. / Bioorganic & Medicinal Chemistry Letters xxx (2018) xxx–xxx
9
18. Sevenich L, Schurigt U, Sachse K, et al. Synergistic antitumor effects of
combined cathepsin and cathepsin deficiencies on breast cancer
nitroxoline-based inhibitors of cathepsin B as potential drugs for
the treatment of cancer.
B
Z
progression and metastasis in mice. Proc Natl Acad Sci USA.
2010;107:2497–2502.
19. Vasiljeva O, Korovin M, Gajda M, et al. Reduced tumour cell proliferation and
delayed development of high-grade mammary carcinomas in cathepsin B-
deficient mice. Oncogene. 2008;27:4191–4199.
A. Supplementary data
Supplementary data (tables with inhibition data for all com-
pounds, assays results on cathepsins H and L, catB inhibitory activ-
ities of selected nitroxoline derivatives at pH 4.5 with Z-Arg-Arg-
AMC as a substrate, supporting schemes and figures, full experi-
mental section, synthetic procedures, and elemental analyses data
for final compounds) associated with this article can be found, in
the online version, at https://doi.org/10.1016/j.bmcl.2018.02.042.
20. Gopinathan A, Denicola GM, Frese KK, et al. Cathepsin
B promotes the
progression of pancreatic ductal adenocarcinoma in mice. Gut.
2012;61:877–884.
21. Kos J, Mitrovic´ A, Mirkovic´ B. The current stage of cathepsin B inhibitors as
potential anticancer agents. Future Med Chem. 2014;6:1355–1371.
22. Schmitz J, Li T, Bartz U, Gütschow M. Cathepsin
B inhibitors: combining
dipeptide nitriles with an occluding loop recognition element by click
chemistry. ACS Med Chem Lett. 2016;7:211–216.
23. Frizler M, Stirnberg M, Sisay MT, Gütschow M. Development of nitrile-based
peptidic inhibitors of cysteine cathepsins. Curr Top Med Chem.
2010;10:294–322.
24. Frizler M, Lohr F, Lülsdorff M, Gütschow M. Facing the gem-dialkyl effect in
enzyme inhibitor design: preparation of homocycloleucine-based azadipeptide
nitriles. Chem Eur J. 2011;17:11419–11423.
25. Turk B. Targeting proteases: successes, failures and future prospects. Nat Rev
Drug Discov. 2006;5:785–799.
References
´
1. Fonovic M, Turk B. Cysteine cathepsins and extracellular matrix degradation.
Biochim Biophys Acta, Gen Subj. 2014;1840:2560–2570.
2. Turk V, Stoka V, Vasiljeva O, et al. Cysteine cathepsins: from structure, function
and regulation to new frontiers. Biochim Biophys Acta. 2012;1824:68–88.
3. Illy C, Quraishi O, Wang J, Purisima E, Vernet T, Mort JS. Role of the occluding
loop in cathepsin B activity. J Biol Chem. 1997;272:1197–1202.
4. Musil D, Zucic D, Turk D, et al. The refined 2.15 Å X-ray crystal structure of
human liver cathepsin B: the structural basis for its specificity. EMBO J.
1991;10:2321–2330.
26. Kramer L, Turk D, Turk B. The future of cysteine cathepsins in disease
management. Trends Pharmacol Sci. 2017;38:873–898.
27. Olson OC, Joyce JA. Cysteine cathepsin proteases: regulators of cancer
progression and therapeutic response. Nat Rev Cancer. 2015;15:712–729.
28. Mirkovic´ B, Renko M, Turk S, et al. Novel mechanism of cathepsin B inhibition
by antibiotic nitroxoline and related compounds. ChemMedChem.
2011;6:1351–1356.
5. Krupa JC, Hasnain S, Nägler DK, Ménard R, Mort JS. S20 substrate specificity and
the role of His110 and His111 in the exopeptidase activity of human cathepsin
B. Biochem. J.. 2002;361:613–619.
6. Nägler DK, Storer AC, Portaro FCV, Carmona E, Juliano L, Ménard R. Major
increase in endopeptidase activity of human cathepsin B upon removal of
occluding loop contacts. Biochemistry. 1997;36:12608–12615.
7. Almeida PC, Nantes IL, Chagas JR, et al. Cathepsin B activity regulation. heparin-
like glycosaminogylcans protect human cathepsin B from alkaline pH-induced
inactivation. J Biol Chem. 2001;276:944–951.
8. Roshy S, Sloane BF, Moin K. Pericellular cathepsin B and malignant progression.
Cancer Metastasis Rev. 2003;22:271–286.
9. Skrzydlewska E, Sulkowska M, Koda M, Sulkowski S. Proteolytic-antiproteolytic
balance and its regulation in carcinogenesis. World
2005;11:1251–1266.
10. Koblinski JE, Ahram M, Sloane BF. Unraveling the role of proteases in cancer.
Clin Chim Acta. 2000;291:113–135.
11. Premzl A, Zavašnik-Bergant V, Turk V, Kos J. Intracellular and extracellular
cathepsin B facilitate invasion of MCF-10A NeoT cells through reconstituted
extracellular matrix. In Vitro Exp Cell Res. 2003;283:206–214.
12. Kostoulas G, Lang A, Nagase H, Baici A. Stimulation of angiogenesis through
cathepsin B inactivation of the tissue inhibitors of matrix metalloproteinases.
FEBS Lett. 1999;455:286–290.
13. Premzl A, Turk V, Kos J. Intracellular proteolytic activity of cathepsin B is
associated with capillary-like tube formation by endothelial cells in vitro. J Cell
Biochem. 2006;97:1230–1240.
14. Cavallo-Medved D, Rudy D, Blum G, Bogyo M, Caglic D, Sloane BF. Live-cell
imaging demonstrates extracellular matrix degradation in association with
active cathepsin B in caveolae of endothelial cells during tube formation. Exp
Cell Res. 2009;315:1234–1246.
15. Burchard M, Mignot G, Derangère V, et al. Chemotherapy-triggered cathepsin B
release in myeloid-derived suppressor cells activates the nlrp3 inflammasome
and promotes tumor growth. Nat Med. 2013;19:57–64.
´
29. Mirkovic B, Markelc B, Butinar M, et al. Nitroxoline impairs tumor progression
in vitro and in vivo by regulating cathepsin
B activity. Oncotarget.
2015;6:19027–19042.
ˇ
´
30. Sosic I, Mirkovic B, Arenz K, Štefane B, Kos J, Gobec S. Development of new
cathepsin B inhibitors: combining bioisosteric replacements and structure-
based design to explore the structure-activity relationships of nitroxoline
derivatives. J Med Chem. 2013;56:521–533.
´
ˇ
31. Mitrovic A, Sosic I, Kos Š, et al. Addition of 2-(ethylamino)acetonitrile group to
nitroxoline results in significantly improved anti-tumor activity in vitro and
in vivo. Oncotarget. 2017;8:59136–59147.
32. Ogrizek M, Turk S, Lešnik S, et al. Molecular dynamics to enhance structure-
J
Gastroenterol.
based virtual screening on cathepsin B.
2015;29:707–712.
J Comput Aided Mol Des.
33. Knez D, Brus B, Coquelle N, et al. Structure-based development of nitroxoline
derivatives as potential multifunctional anti-Alzheimer agents. Bioorg Med
Chem. 2015;23:4442–4452.
ˇ
34. Brodnik H, Pozgan F, Štefane B. Synthesis of 8-heteroaryl nitroxoline analogues
via one-pot sequential Pd-catalyzed coupling reactions. Org Biomol Chem.
2016;14:1969–1981.
35. Mohamed MM, Sloane BF. Cysteine cathepsins: multifunctional enzymes in
cancer. Nat Rev Cancer. 2006;6:764–775.
ˇ
36. Mitrovic´ A, Mirkovic´ B, Sosic I, Gobec S, Kos J. Inhibition of endopeptidase and
exopeptidase activity of cathepsin B impairs extracellular matrix degradation
and tumour invasion. Biol Chem. 2016;397:165–174.
37. Szpaderska AM, Frankfater A. An intracellular form of cathepsin B contributes
to invasiveness in cancer. Cancer Res. 2001;61:3493–3500.
38. Eisenberg MC, Kim Y, Li R, Ackerman WE, Kniss DA, Friedman A. Mechanistic
modeling of the effects of myoferlin on tumor cell invasion. Proc Natl Acad Sci
USA. 2011;108:20078–20083.
39. Kelm JM, Timmins NE, Brown CJ, Fussenegger M, Nielsen LK. Method for
generation of homogeneous multicellular tumor spheroids applicable to a wide
variety of cell types. Biotechnol Bioeng. 2003;83:173–180.
16. Gocheva V, Zeng W, Ke D, et al. Distinct roles for cysteine cathepsin genes in
multistage tumorigenesis. Genes Dev. 2006;20:543–556.
17. Vasiljeva O, Papazoglou A, Krüger A, et al. Tumor cell-derived and macrophage-
derived cathepsin B promotes progression and lung metastasis of mammary
cancer. Cancer Res. 2006;66:5242–5250.
40. Shim JS, Matsui Y, Bhat S, et al. Effect of nitroxoline on angiogenesis and growth
of human bladder cancer. J Natl Cancer Inst. 2010;102:1855–1873.
ˇ
Please cite this article in press as: Sosic I., et al. Bioorg. Med. Chem. Lett. (2018), https://doi.org/10.1016/j.bmcl.2018.02.042