Identification of novel molecular scaffolds for the design of MMP-13 inhibitors: A first round of lead optimization

Article abstract of DOI:10.1016/j.ejmech.2011.10.035

Osteoarthritis (OA) is the leading cause of joint pain and disability in middle-aged and elderly patients, and is characterized by progressive loss of articular cartilage. Among the various matrix metalloproteinases (MMPs), MMP-13 is specifically expresse

Full text of DOI:10.1016/j.ejmech.2011.10.035

European Journal of Medicinal Chemistry 47 (2012) 143e152
Contents lists available at SciVerse ScienceDirect
European Journal of Medicinal Chemistry
journal homepage: http://www.elsevier.com/locate/ejmech
Original article
Identification of novel molecular scaffolds for the design of MMP-13 inhibitors:
A first round of lead optimization
,
, ,
Valeria La Pietra a 1, Luciana Marinelli a 1, Sandro Cosconati b, Francesco Saverio Di Leva c, Elisa Nuti d,
*
Salvatore Santamaria d e, Isabella Pugliesi d, Matteo Morelli d, Francesca Casalini d, Armando Rossello d,
,
Concettina La Motta d, Sabrina Taliani d, Robert Visse e, Hideaki Nagase e, Federico da Settimo d,
,
1
Ettore Novellino a
a Dipartimento di Chimica Farmaceutica e Tossicologica, Università di Napoli “Federico II” Via D. Montesano 49, 80131 Napoli, Italy
b Dipartimento di Scienze Ambientali, Seconda Università di Napoli, Via Vivaldi 43, 81100 Caserta, Italy
c Department of Drug Discovery and Development, Istituto Italiano di Tecnologia (IIT), Via Morego 30, 16163 Genova, Italy
d Dipartimento di Scienze Farmaceutiche, Università di Pisa, via Bonanno 6, 56126 Pisa, Italy
e Department of Matrix Biology, The Kennedy Institute of Rheumatology Division, Imperial College, 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:
Osteoarthritis (OA) is the leading cause of joint pain and disability in middle-aged and elderly patients,
and is characterized by progressive loss of articular cartilage. Among the various matrix metal-
loproteinases (MMPs), MMP-13 is specifically expressed in the cartilage of human OA patients and is not
present in normal adult cartilage. Thus, MMP-13-selective inhibitors are promising candidates in oste-
oarthritis therapy. Recently, we designed an N-isopropoxy-arylsulfonamide-based hydroxamate inhib-
itor, which showed low nanomolar activity and high selectivity for MMP-13. In parallel to further studies
aiming to assess the in vivo activity of our compound, we screened the Life Chemicals database through
computational docking to seek for novel scaffolds as zinc-chelating non-hydroxamate inhibitors.
Experimental evaluation of 20 selected candidate compounds verified five novel leads with IC50 in the
Received 11 July 2011
Received in revised form
29 September 2011
Accepted 18 October 2011
Available online 25 October 2011
Keywords:
MMPs
Virtual screening
Molecular docking
Osteoarthritis
Cancer
low mM range. These newly discovered inhibitors are structurally unrelated to the ones known so far and
provide useful scaffolds to develop compounds with more desirable properties. Finally, a first round of
structure-based optimization on lead 1 was accomplished and led to an increase in potency of more than
5 fold.
Ó 2011 Elsevier Masson SAS. All rights reserved.
1. Introduction
OA [2], and inhibition of its activity has proven to be efficacious in
a variety of models of experimentally induced as well as sponta-
Osteoarthritis (OA) is the leading cause of joint pain and
disability in middle-aged and elderly patients. It is characterized by
progressive loss of articular cartilage that eventually leads to
denudation of the joint surface. The cartilage loss observed in OA is
the result of a complex process involving degradation of various
components of the cartilage matrix. Particularly, degradation of
cartilage-specific type II collagen by mammalian collagenases
(MMPs) is a key step in the loss of structural and functional
integrity of cartilage [1]. Among all known MMPs, MMP-13 is
considered the principal target in OA. Indeed, today there are
overwhelming data on the role of MMP-13 in the pathogenesis of
neously occurring OA [3]. Unfortunately, none of the known MMP
inhibitors (MMPIs) have been successfully utilized as therapeutic
agents so far. This was due to the lack of selectivity for a specific
isozyme, leading to heavy dose- and duration-dependent muscu-
loskeletal side effects [4]. Therefore, current drug development
strategies for treatment of OA are focused on selective inhibition of
MMP-13, although recent evidences suggest that other MMPs, such
as MMP-1, may also contribute to the collagen degradation process
[5]. However, the design of a selective MMPI is not a trivial task, as
MMPs share an high similarity in the overall three-dimensional fold
and many conserved amino acids exist in the inhibitor binding site,
besides the conserved catalytic zinc ion. The major structural
difference observed between the MMP enzymes resides in the
relative size and shape of the S10 subsite, which is located in
proximity of the catalytic metal. From a structural point of view,
* Corresponding author. Tel./fax: þ39 081 678644.
E-mail address: lmarinel@unina.it (L. Marinelli).
1
Fax: þ39 081 678619.
0223-5234/$ e see front matter Ó 2011 Elsevier Masson SAS. All rights reserved.
doi:10.1016/j.ejmech.2011.10.035

144
V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
almost all MMPIs known so far are based on a zinc-binding group
(ZBG) and a hydrophobic portion protruding into the hydrophobic
S10 subsite. These compounds behave as competitive inhibitors
since the ZBG can mimic one of the transition states occurring
during the substrate hydrolysis. Currently, two successful strategies
to confer selectivity of action to an MMP13 inhibitor are known: the
first resides in the proper modification of the P10 substituent on
MMPI to take advantage of the differences between the diverse
MMPs; the second is the finding of an allosteric inhibitor [6], which
binds tightly to the S10* site, a unique side pocket adjacent to S10
that hasn’t been observed for other MMPs, without chelating the
metal that is thought to contribute to the promiscuous inhibition of
multiple MMPs. Recently, as a result of the first strategy, we
2.3. Virtual screening calculations
Docking calculations were performed with version 4.0 of the
AutoDock software package as implemented through the graphic
user interface AutoDockTools (ADT 1.4.6). All compounds of the Life
Chemical diversity set together with the 830C structure of MMP-13
were converted to AutoDock format files (.pdbqt) using ADT. The
docking area was defined by a box, centered on the catalytic zinc.
Grids (dimension of 60 Å ꢁ 65 Å ꢁ 60 Å) were then generated for 13
ligand atom types (sufficient to describe all atoms in the selected
database) with the help of AutoGrid4 using a grid spacing of 0.375 Å.
For each ligand of the Life Chemical diversity set, 100 separate
docking calculations were performed. Each docking calculation
consisted of 1 ꢁ107 energyevaluations using the Lamarckian genetic
algorithm local search (GALS) method. A low-frequency local search
in accordance with the method of Solis and Wets was applied to
docking trials to ensure that the final solution represents a local
minimum. Each docking run was performed with a population size
of 150, and 300 rounds of Solis and Wets local search were applied,
with a probability of 0.06. A mutation rate of 0.02 and a crossover
rate of 0.8 were used to generate new docking trials for subsequent
generations. The docking results from each of the 100 calculations
were clustered on the basis of root-mean square deviation (rmsd
2 Å) betweenthe Cartesian coordinatesof the ligand atoms and were
ranked on the basis of the free energy of binding. The top-ranked
compounds were visually inspected for good chemical geometry.
Finally, as a last criterion of selection, we introduced the visual
inspection of the putative best ranking ligand/receptor complexes.
In this regard, we decided to discard all the molecules for which AD4
did not predict coordination of the catalytic zinc in order to obtain
compounds of a certain potency. Another selection criterion resided
in the occupancy of the S10 pocket, in the attempt to obtain a selec-
tivity of action towards the MMP-13. Pictures of the modelled
ligand/enzyme complexes together with graphic manipulations
were rendered with UCSF chimera package from the Resource for
Biocomputing, Visualization, and Informatics at the University of
California, San Francisco [13].
designed
a N-isopropoxy-arylsulfonamide-based hydroxamate
inhibitor, which showed low nanomolar activity for MMP-13 and
high selectivity over some other tested MMPs [7]. In parallel to
further studies aiming to assess the activity of this promising
compound using in vivo models of OA, we decided to seek for novel
scaffolds as zinc-chelating non-hydroxamate inhibitors. In fact,
a debate is still open on the advisability of using hydroxamates as
ZBG due to toxicity and metabolic stability issues [8,9].
In this respect, we have taken advantage of the availability of
several MMP-13 crystal structures and have used computational
docking to screen the Life Chemicals database. Experimental tests
of a limited selection of candidate compounds (20) verified five
novel leads, structurally unrelated to the known MMPIs. A first
round of structure-based optimization on lead 1 was accomplished
and led to an increase in potency of more than 5 fold.
2. Experimental section
2.1. Database preparation
For the in silico screening, the Life Chemicals database [10] was
used. This library is a collection of small compounds carefully
selected to provide the broadest pharmacophore coverage for
a total of six thousands non-redundant molecules. The database
was uploaded on ZINC server [11] as 1D smiles strings and pro-
cessed with the ZINC protocol. This protocol filters-out molecules
with molecular weight greater than 700, calculated LogP greater
than 6 and less than ꢀ4, number of hydrogen-bond donors,
hydrogen-bond acceptors, and rotatable bonds greater than 6, 11,
and 15 respectively. It also removes all molecules containing
“exotic” atoms (i.e. different from H, C, N, O, F, S, P, Cl, Br, or I).
Moreover it allows the creation of all stereoisomers, tautomers and
correctly protonated forms of the molecules between pH 5 and 9.5.
The protocol outcome from the server was a file containing 7769
compounds.
2.4. Chemistry
The purity of the five hits that were essential to the conclusions
drawn in the text were determined by HPLC on a Merck Hitachi D-
7000 liquid chromatograph equipped with a Discovery C18 column
(250 mm ꢁ 4.6 mm, 5
mm particle size) and a UV/vis detector
¼ 250 nm. All compounds were eluted with the solvent
setting at
l
systems listed in Supporting Information Table 1.
As regards the newly synthesized compounds 1aeh and their
intermediates, melting points were determined using a Reichert
Köfler hot-stage apparatus and are uncorrected. Routine nuclear
magnetic resonance spectra were recorded in DMSO-d6 solution on
a Varian Gemini 200 spectrometer operating at 200 MHz. Analyt-
ical TLC was carried out on Merck 0.2 mm precoated silica gel
aluminium sheets (60 F-254). Elemental analyses were performed
by our Analytical Laboratory and agreed with theoretical values to
within ꢂ0.4%.
2.2. Selection of the MMP-13 X-ray structure for VS experiment and
protein preparation
Several X-ray structures of MMP-13 have been released in the
Protein Data Bank (PDB) so far. A superposition of all X-ray struc-
tures on the alpha carbon atoms, using 830C as reference structure,
shows that the protein folding and the catalytic loops shape are
highly superimposable and that in the catalytic site, the large
majority of the residues are all preserved in the side chain
conformations. Thus, only the enzyme structure 830C, which has
the higher resolution (1.60 Å), was selected for our VS experiment.
From this structure, all water molecules, ions and the inhibitor
were removed from the binding site. All hydrogen atoms were
added to the protein structure using ADT [12], and to the catalytic
Zn ion present in the active site a þ2 charge was assigned.
2-hydroxy-3-methoxybenzaldehyde, 2-hydroxy-5-methoxyben
zaldehyde, 2,3-dihydroxybenzaldehyde, 5-chloro-2-hydroxyben
zaldehyde, 5-chloro-2-hydroxy-3-methoxybenzaldehyde, 5-allyl-
2-hydroxy-3-methoxybenzaldehyde, 3-(2-bromoacetyl)-2H-chro-
men-2-one (6), ethyl acetoacetate, aniline, 3-aminobenzoic acid,
ethyl 3-aminobenzoate, reagent and solvents were from Sigma-
Aldrich.
3-Acetyl-8-methoxy-2H-chromen-2-one (21) [14], 3-acetyl-6-
methoxy-2H-chromen-2-one (22) [15], were obtained according
to reported procedures.

V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
145
3. General procedure for the synthesis of 6,8-substituted 3-
acetyl-2H-chromen-2-one derivatives 23e26
4.3. 3-(2-Bromoacetyl)-8-hydroxy-2H-chromen-2-one (9)
Yield: 62%; mp: 223e225 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d
:
A mixture of equimolar quantities (0.006 mol) of the appro-
priate salicylaldehyde (2,3-dihydroxybenzaldehyde, 5-chloro-2-
4.89 (s, 2H); 7.23e7.29 (m, 2H); 7.37e7.44 (m. 1H); 8.76 (s, 1H);
10.47 (bs, 1H, OH, exch. with D2O). Anal. Calcd. for C11H7BrO4: C,
46.67; H, 2.49; found: C, 46.79; H, 2.42.
hydroxybenzaldehyde,
5-chloro-2-hydroxy-3-methoxybenzalde
hyde, 5-allyl-2-hydroxy-3-methoxybenzaldehyde) and ethyl ace-
toacetate in 15 mL of absolute ethanol was stirring in the presence
of a few drops of piperidine for 5 h at room temperature (TLC
analysis). After cooling, the precipitate which formed was collected
to give the 3-acethyl-2H-chromen-2-one derivatives 23e26 which
were purified by crystallization from ethanol.
4.4. 3-(2-Bromoacetyl)-6-chloro-2H-chromen-2-one (10)
Yield: 60%; mp: 217e219 ꢃC (ref. [20], mp: 221e222 ꢃC). Anal.
Calcd. for C11H6BrClO3: C, 43.83; H, 3.01; found: C, 43.98; H, 2.92.
4.5. 3-(2-Bromoacetyl)-6-chloro-8-methoxy-2H-chromen-2-one
(11)
3.1. 3-Acetyl-8-hydroxy-2H-chromen-2-one (23)
Yield: 59%; mp: 218e220 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
Yield: 63%; mp: 248e250 ꢃC (ref. [16], mp: 253 ꢃC). Anal. Calcd.
3.97 (s, 3H); 4.90 (s, 2H); 7.54 (s, 1H); 7.67 (s, 1H); 8.76 (s, 1H). Anal.
Calcd. for C12H8BrClO4: C, 43.47; H, 2.43; found: C, 43.67; H, 2.35.
for C11H8O4: C, 64.71; H, 3.95; found: C, 64.99; H, 3.89.
3.2. 3-Acetyl-6-chloro-2H-chromen-2-one (24)
4.6. 6-Allyl-3-(2-bromoacetyl)-8-methoxy-2H-chromen-2-one (12)
Yield: 69%; mp: 198e200 ꢃC (ref. [17], mp: 204 ꢃC). Anal. Calcd.
for C11H7ClO3: C, 59.35; H, 3.17; found: C, 59.59; H, 3.09.
Yield: 66%; mp: 156e158 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
3.38e3.55 (m, 2H); 3.92 (s, 3H); 4.88 (s, 2H); 5.08e5.17 (m, 2H);
5.95e6.02 (m, 1H); 7.32e7.38 (m, 2H); 8.74 (s, 1H). Anal. Calcd. for
C15H13BrO4: C, 53.43; H, 3.89; found: C, 53.27; H, 3.93.
3.3. 3-Acetyl-6-chloro-8-methoxy-2H-chromen-2-one (25)
Yield: 87%; mp: 189e191 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
2.58 (s, 3H); 3.96 (s, 3H); 7.46 (s, 1H); 7.63 (s, 1H); 8.59 (s, 1H). Anal.
Calcd. for C12H9ClO4: C, 57.05; H, 3.59; found: C, 57.29; H, 3.43.
5. General procedure for the synthesis of 30,6,8-substituted 3-
(2-phenylaminoacetyl)-2H-chromen-2-one derivatives 13e20
A solution of the opportune amine (aniline, 3-aminobenzoic
acid, ethyl 3-aminobenzoate) (0.0021 mol), NaHCO3 (0.176 g,
3.4. 3-Acetyl-6-allyl-8-methoxy-2H-chromen-2-one (26)
0.0021 mol) and the appropriately substituted a-bromocetyl-2H-
Yield: 81%; mp: 147e149 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
2.57 (s, 3H); 3.37e3.45 (m, 2H); 3.96 (s, 3H); 5.08e5.17 (m, 2H);
5.95e6.023 (m, 1H); 7.27e7.29 (m, 2H); 8.58 (s, 1H). Anal. Calcd. for
C15H14O4: C, 69.76; H, 5.46; found: C, 69.99; H, 5.29.
chromen-2-ones 6e12 (0.0021 mol) in 10 mL of ethanol was
refluxed for 5 h (TLC analysis). After cooling, the precipitate which
formed was collected and derivatives 13e20 resulted sufficiently
pure to be used in the next reaction without further purification.
4. General procedure for the synthesis of 6,8-substituted 3-(2-
bromoacetyl)-2H-chromen-2-one derivatives 7e12
5.1. 3-[2-(Phenylamino)acetyl]-2H-chromen-2-one (13)
Yield: 70%; mp: 177e179 ꢃC (ref. [21], mp ¼ 180e185 ꢃC). 1H
A hot solution of the appropriate 3-acetyl-2H-chromen-2-ones
21e26 (0.0022 mol) in chloroform was added to a suspension of
copper (II) bromide (0.900 g, 0.0044 mol) in 10 mL of ethyl acetate.
The resulting reaction mixture was refluxed with vigorous stirring
to ensure complete exposure of the copper (II) bromide to the
reaction medium until the reaction was complete, as judged by
colour change of the solution from green to amber, disappearance
of all black solid, and cessation of hydrogen bromide evolution. The
suspension obtained was filtered to remove the copper (I) bromide.
After cooling, the solid which separated was collected to give the 3-
(2-bromoacetyl)-2H-chromen-2-ones 7e12, which were purified
by crystallization from ethanol.
NMR (200 MHz, DMSO-d6),
d: 4.59 (d, 2H, J ¼ 5.8 Hz); 5.94 (t, 1H,
J ¼ 5.6 Hz, exch. with D2O); 6.52e6.63 (m, 3H); 7.04e7.12 (m, 2H);
7.40e7.53 (m, 2H); 7.68e7.82 (m, 1H); 7.97e8.01 (m, 1H); 8.75 (s,
1H). Anal. Calcd. for C17H13NO3: C, 73.11; H, 4.69; N, 5.02; found: C,
73.27; H, 4.73; N, 4.97.
5.2. 3-[2-(3-Ethoxycarbonyl)phenylaminoacetyl]-2H-chromen-2-
one (14)
Yield: 76%; mp: 175e177 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
1.29 (t, 3H, J ¼ 7.0 Hz); 4.27 (q, 2H, J ¼ 7.1 Hz); 4.65 (d, 2H,
J ¼ 4.2 Hz); 6.20 (t, 1H, J ¼ 4.0 Hz, exch. with D2O); 6.85e6.88 (m,
1H); 7.18e7.22 (m, 1H); 7.41e7.54 (m, 2H); 7.74e7.82 (m, 1H);
7.98e8.02 (m, 1H); 8.77 (s, 1H). Anal. Calcd. for C20H17NO5: C, 68.37;
H, 4.88; N, 3.99; found: C, 68.27; H, 4.73; N, 3.85.
4.1. 3-(2-Bromoacetyl)-8-methoxy-2H-chromen-2-one (7)
Yield: 68%; mp: 199e201 ꢃC (ref. [18], mp: 206 ꢃC); 1H NMR
(200 MHz, DMSO-d6), d: 3.95 (s, 3H); 4.91 (s, 2H); 7.38e7.56 (m,
5.3. 3-[2-(3-Hydroxycarbonyl)phenylaminoacetyl]-8-methoxy-2H-
3H); 8.82 (s, 1H). Anal. Calcd. for C12H9BrO4: C, 48.51; H, 3.05;
found: C, 48.73; H, 2.98.
chromen-2-one (15)
Yield: 66%; mp: 205e207 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
4.2. 3-(2-Bromoacetyl)-6-methoxy-2H-chromen-2-one (8)
3.95 (s, 3H); 4.64 (d, 2H, J ¼ 5.6 Hz); 6.27 (t, 1H, J ¼ 5.6 Hz, exch.
with D2O); 6.84e6.88 (m, 1H); 7.16e7.19 (m, 3H); 7.32e7.56 (m,
3H); 8.74 (s, 1H); 12.69 (bs, 1H, exch. with D2O). Anal. Calcd. for
C19H15NO6: C, 64.59; H, 4.28; N, 3.96; found: C, 64.71; H, 4.52; N,
4.06.
Yield: 70%; mp: 135e137 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
3.83 (s, 2H); 4.91 (s, 2H); 7.40e7.58 (m, 3H); 8.79 (s, 1H) [19]. Anal.
Calcd. for C12H9BrO4: C, 48.51; H, 3.05; found: C, 48.68; H, 3.00.

146
V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
5.4. 3-(2-(3-(Hydroxycarbonyl)phenylamino)acetyl)-6-methoxy-
6.1. 3-(2,3-Dihydro-3-phenyl-2-thioxo-1H-imidazol-5-yl)-2H-
2H-chromen-2-one (16)
chromen-2-one (1a)
Yield: 63%; mp: 205e207 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
Yield: 73%; mp: 255e257 ꢃC (ref. [22]), mp: 154 ꢃC. Anal. Calcd.
for C18H12N2O2S: C, 67.48; H, 3.78; N, 8.74; found: C, 67.65; H, 3.59;
N, 8.77.
3.82 (s, 3H); 4.65 (d, 2H, J ¼ 5.2 Hz); 6.29 (t, 1H, J ¼ 5.4 Hz, exch.
with D2O) 6.85e6.88 (m,1H); 7.16e7.23 (m, 3H); 7.35e7.59 (m, 3H);
8.73 (s, 1H); 12.73 (bs, 1H, exch. with D2O). Anal. Calcd. for
C19H15NO6: C, 64.59; H, 4.28; N, 3.96; found: C, 64.41; H, 4.36; N,
3.84.
6.2. 3-[2,3-(dihydro-3-(3-ethoxycarbonylphenyl)-2-thioxo-1H-
imidazol-5-yl)]-2H-chromen-2-one] (1b)
Yield: 90%; mp: 260e262 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
5.5. 8-(Hydroxy-3-(2-(3-hydroxycarbonyl)phenylamino)acetyl)-
2H-chromen-2-one (17)
1.34 (t, 3H, J ¼ 7.0 Hz); 4.36 (q, 2H, J ¼ 7.1 Hz); 7.40e7.50 (m, 2H);
7.63e7.74 (m, 3H); 7.91e8.05 (m, 3H); 8.25 (s, 1H); 8.59 (s, 1H);
13.18 (s, 1H, exch. with D2O). Anal. Calcd. for C21H16N2O4S: C, 64.27;
H, 4.11; N, 7.14; found: C, 64.45; H, 4.03; N, 7.27.
Yield: 65%; mp: 217e219 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
4.60 (d, 2H, J ¼ 4.6 Hz); 6.22 (t, 1H, J ¼ 4.4 Hz, exch. with D2O)
6.80e6.83 (m, 5H); 7.12e7.20 (m, 1H); 7.34e7.39 (m, 1H); 8.66 (s,
1H); 10.42 (s, 1H, exch. with D2O); 12.64 (bs, 1H, exch. with D2O).
Anal. Calcd. for C18H13NO6: C, 63.72; H, 3.86; N, 4.13; found: C,
63.91; H, 3.72; N, 4.16.
6.3. 3-[2,3-(dihydro-3-(3-hydroxycarbonylphenyl)-2-thioxo-1H-
imidazol-5-yl)]-8-methoxy-2H-chromen-2-one (1c)
Yield: 62%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d: 3.94
(s, 3H); 7.19e7.24 (m, 1H); 7.34e7.37 (m, 2H); 7.64e7.72 (m, 1H);
7.91e8.03 (m, 3H); 8.23 (s, 1H); 8.57 (s, 1H); 13.16 (bs, 2H, exch.
with D2O). Anal. Calcd. for C20H14N2O5S: C, 60.91; H, 3.58; N, 7.10;
found: C, 61.05; H, 3.71; N, 7.06.
5.5. 6-(Chloro-3-(2-(3-hydroxycarbonyl)phenylamino)acetyl)-2H-
chromen-2-one (18)
Yield: 80%; mp: 210e212 ꢃC; 1H NMR (200 MHz, DMSO-d6):
d:
4.66 (d, 2H, J ¼ 4.6 Hz); 6.29 (t, 1H, J ¼ 4.6 Hz, exch. with D2O)
6.87e6.88 (m, 1H); 7.16e7.20 (m, 3H); 7.54e7.61 (m, 1H); 7.76e7.87
(m, 1H); 8.14 (s, 1H); 8.73 (s, 1H); 12.70 (bs, 1H, exch. with D2O).
Anal. Calcd. for C18H12ClNO5: C, 60.43; H, 3.38; N, 3.92; found: C,
60.61; H, 3.47; N, 4.05.
6.4. 3-[2,3-(dihydro-3-(3-hydroxycarbonylphenyl)-2-thioxo-1H-
imidazol-5-yl)]-6-methoxy-2H-chromen-2-one (1d)
Yield: 87%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d: 3.85
(s, 3H); 7.06e7.08 (m, 1H); 7.21e7.27 (m, 1H); 7.41e7.45 (m, 1H);
7.63e7.71 (m, 1H); 7.89e8.03 (m, 3H); 8.22 (s, 1H); 8.52 (s, 1H);
13.17 (bs, 1H, exch. with D2O); 13.28 (bs, 1H, exch. with D2O). Anal.
Calcd. for C20H14N2O5S: C, 60.91; H, 3.58; N, 7.10; found: C, 60.85; H,
3.51; N, 7.22.
5.6. 6-(Chloro-3-(2-(3-hydroxycarbonyl)phenylamino)acetyl)-8-
methoxy-2H-chromen-2-one (19)
Yield: 72%; mp: 213e215 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
3.98 (s, 3H); 4.64 (d, 2H, J ¼ 4.2 Hz); 6.17 (t,1H, J ¼ 4.4 Hz, exch. with
D2O); 6.83e6.87 (m, 1H); 7.16e7.18 (m, 3H); 7.51 (s, 1H); 7.66 (s,
1H); 8.68 (s, 1H); 12.60 (bs, 1H, exch. with D2O). Anal. Calcd. for
C19H14ClNO6: C, 58.85; H, 3.64; N, 3.61; found: C, 58.69; H, 3.53; N,
3.70.
6.5. 3-[2,3-(dihydro-3-(3-hydroxycarbonylphenyl)-2-thioxo-1H-
imidazol-5-yl)]-8-hydroxy-2H-chromen-2-one (1e)
Yield: 67%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
7.06e7.28 (m, 3H, Ar-H); 7.63e7.71 (m,1H); 7.89e8.03 (m, 3H); 8.23
(s, 1H); 8.53 (s, 1H); 10.37 (bs, 1H, exch. with D2O); 13.13 (bs, 2H,
exch. with D2O). Anal. Calcd. for C19H12N2O5S: C, 59.99; H, 3.18; N,
7.36; found: C, 60.13; H, 3.23; N, 7.14.
5.7. 6-(Allyl-3-(2-(3-hydroxycarbonyl)phenylamino)acetyl)-8-
methoxy-2H-chromen-2-one (20)
Yield: 69%; mp: 193e195 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
6.6. 6-Chloro-3-[2,3-(dihydro-3-(3-hydroxycarbonylphenyl)-2-
3.42 (d, 2H, J ¼ 6.2 Hz); 3.95 (s, 3H); 4.61 (d, 2H, J ¼ 4.0 Hz);
5.10e5.19 (m, 2H); 6.01e6.09 (m, 2H, 1H exch. with D2O);
6.82e6.85 (m, 1H); 7.16e7.27 (m, 6H); 8.63 (s, 1H); 12.72 (bs, 1H,
exch. with D2O). Anal. Calcd. for C22H19NO6: C, 67.17; H, 4.87; N,
3.56; found: C, 67.31; H, 4.73; N, 3.61.
thioxo-1H-imidazol-5-yl)]-2H-chromen-2-one (1f)
Yield: 78%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d:
7.50e7.54 (m, 1H); 7.64e7.71 (m, 3H); 7.92e8.04 (m, 3H); 8.22 (s,
1H); 8.48 (s, 1H); 13.23 (bs, 2H, exch. with D2O). Anal. Calcd. for
C19H11ClN2O4S: C, 57.22; H, 2.78; N, 7.02; found: C, 57.43; H, 2.77; N,
7.13.
6. General procedure for the synthesis of 30’,6,8-substituted 3-
(2,3-dihydro-3-phenyl-2-thioxo-1H-imidazol-5-yl)-2H-
chromen-2-one derivatives 1a-h
6.7. 6-Chloro-3-[2,3-dihydro-3-(3-hydroxycarbonylphenyl)-2-
thioxo-1H-imidazol-5-yl)]-8-methoxy-2H-chromen-2-one (1g)
A mixture of compounds 13e20 (0.001 mol) and an excess of
ammonium thiocyanate (1.14 g, 0.015 mol) in acetic acid (2 mL) was
stirred for 1.5 h at 60e65 ꢃC (TLC analysis). After cooling, the
insoluble product was filtered off, washed with water, cold MeCN
and cold diethyl ether to afford derivatives 1a-h in the desired
purity degree (ꢄ95%)
Yield: 71%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6):
d: 3.96
(s, 3H); 7.26 (s, 1H); 7.40 (s, 1H); 7.67e7.71 (m, 1H); 7.93e8.03 (m,
3H); 8.22 (s, 1H); 8.45 (s, 1H); 13.23 (bs, 2H, exch. with D2O). Anal.
Calcd. for C20H13ClN2O5S: C, 56.01; H, 3.06; N, 6.53; found: C, 56.23;
H, 3.11; N, 6.42.

V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
147
6.8. 6-Allyl-3-[2,3-(dihydro-3-(3-hydroxycarbonylphenyl)-2-
thioxo-1H-imidazol-5-yl)]-8-methoxy-2H-chromen-2-one (1h)
8. Results and discussion
8.1. Virtual screening
Yield: 77%; mp: >300 ꢃC; 1H NMR (200 MHz, DMSO-d6),
d: 3.49
(d, 2H, J ¼ 6.0 Hz); 3.93 (s, 3H); 5.02e5.22 (m, 2H); 5.98e6.10 (m,
1H); 7.01 (s, 1H); 7.19 (s, 1H); 7.64e7.71 (m, 1H); 7.91e8.04 (m, 3H);
8.23 (s, 1H); 8.52 (s, 1H); 13.16 (bs, 2H, exch. with D2O). Anal. Calcd.
for C23H18N2O5S: C, 63.58; H, 4.18; N, 6.45; found: C, 63.39; H, 4.27;
N, 6.71.
To date, several X-ray structures of MMP-13 have been released
in the Protein Data Bank. Besides that co-crystallized with TIMP-2
(PDB code: 2E2D), all the others were co-crystallized with organic
inhibitors such as the diphenylether sulfone RS-130830 (PDB code
830C). A superposition of all X-ray structures on the alpha carbon
atoms, using 830C as reference structure, shows that the protein
folding and the catalytic loops shape are highly superimposable,
and that in the catalytic site the large majority of the residues are all
preserved in the side chain conformations. Thus, only the enzyme
structure 830C, which has the higher resolution (1.60 Å), was
selected for our VS experiment. As docking program for the VS, we
used the Autodock program (AD4) [25], which has been extensively
and successfully employed in multiple VS campaigns undertaken
by our research group [26]. AD4 was applied to virtually screen the
Life Chemicals database, a collection of six thousands non-
redundant drug-like compounds selected to provide the broadest
pharmacophore coverage. Prior to docking experiments, the entire
Life Chemicals database was processed with the ZINC protocol
leading to a total of 7769 molecules (see Experimental section for
details). The results of the VS on the Life Chemical database, were
then sorted on the basis of the predicted binding free energies
7. Biology. Materials and methods
Recombinant human MMP-14 catalytic domain was a kind gift
of Prof. Gillian Murphy (Department of Oncology, University of
Cambridge, UK). Pro-MMP-1, pro-MMP-2, pro-MMP-3, and pro-
MMP-13 were purchased from Calbiochem. APMA was from
Sigma-Aldrich. All compounds were subjected to combustion
analysis prior to be tested for their inhibitory activity, to verify their
consistence with a purity of at least 95%. ARP100 was synthesized at
Department of Pharmaceutical Sciences (Pisa, Italy) according to
the previously described procedure [23]. All other chemicals were
of reagent grade.
(
DGAD4) which in our case ranged from ꢀ3.93 to ꢀ15.61 kcal/mol. A
7.1. Enzyme activation
scoring filter was set arbitrarily to ꢀ10.5 kcal/mol so as to retain
23% of the docked solutions. The top 1800 compounds in their
predicted binding poses were selected for visual inspection. In
order to obtain compounds endowed with an inhibitory potency
against MMP-13, we discarded all the molecules for which AD4 did
not predict coordination of the catalytic zinc. Then, in the attempt
to find leads with a certain selectivity of action, for each inspected
compound, the occupancy of the S10 pocket has been evaluated,
although it was not expected to be total due to the small size of the
docked compounds. As last criterion of choice, we evaluated the
attitude of each molecule to be chemically optimized. At the end of
this process, a total of 24 compounds of the Life Chemical Data Set
were selected for further analysis. Two products were not available
from the vendor, and two were not soluble at the test concentra-
tion, so a total of twenty compounds were used for biochemical
assays. Initially, all compounds were screened at a concentration of
Proenzymes were activated immediately prior to use with p-
aminophenylmercuric acetate (APMA 2 mM for 1 h at 37 ꢃC for
MMP-2, APMA 2 mM for 2 h at 37 ꢃC for MMP-1, 1 mM for 30 min at
37 ꢃC for MMP-13). Pro-MMP-3 was activated with trypsin 5
for 30 min at 37 ꢃC followed by soybean trypsin inhibitor 62
mg/mL
mg/mL.
7.2. Enzyme inhibition assays
For assay measurements, the purchased compound stock solu-
tions (10 mM in DMSO) were further diluted for each MMP in the
fluorimetric assay buffer (FAB: Tris 50 mM, pH ¼ 7.5, NaCl 150 mM,
CaCl2 10 mM, Brij 35 0.05% and DMSO 1%). Activated enzyme (final
concentration 0.56 nM for MMP-2, 0.3 nM for MMP-13, 5 nM for
MMP-3, 1 nM for MMP-14cd, and 2.0 nM for MMP-1) and inhibitor
solutions were incubated in the assay buffer for 4 h at 25 ꢃC. After
100 mM by fluorometric assay on recombinat enzyme. ARP100 [23],
a hydroxamate-based MMP inhibitor previously developed by our
research group, was used in the same assay conditions as reference
compound. To exclude any possible nonspecific/promiscuous
inhibition of MMP-13 due to aggregate formation, we performed all
the assays pertaining the active compounds in the presence of
0.05% Brij-35, a nonionic detergent similar to Triton X-100, as
suggested by Shoichet et al. [27]. Five ligands, out of the twenty
tested, provided considerable inhibition of MMP-13 activity and
were characterized in detail (see experimental methods). All other
the addition of 200 mM solution of the fluorogenic substrate Mca-
Arg-Pro-Lys-Pro-Val-Glu-Nva-Trp-Arg-Lys(Dnp)eNH2 (Sigma) for
MMP-3 and Mca-Lys-Pro-Leu-Gly-Leu-Dap(Dnp)-Ala-ArgeNH2
(Bachem) for all the other enzymes in DMSO (final concentration
2
mM), the hydrolysis was monitored every 15 s for 15 min
recording the increase in fluorescence lex 325 nm,
(
¼
lem ¼ 395 nm) using a Molecular Devices SpectraMax Gemini XS
plate reader. The assays were performed in triplicate in a total
volume of 200
mL per well in 96-well microtitre plates (Corning,
compounds that did not cause detectable inhibition at 100 mM
black, NBS). The MMP inhibition activity was expressed in relative
fluorescent units (RFU). Percent of inhibition was calculated from
control reactions without the inhibitor. The inhibitory effect of the
tested compounds was routinely estimated at a concentration of
concentrations were not further investigated (see SI for chemical
structures). Table 1 lists structures, Life Chemicals codes, AD4
binding free energies and the MMP-13 IC50 of the novel inhibitors
which ranges from 9 to 140 mM. The IC50 values were deduced from
100
m
M towards MMP-13. Those derivatives found to be active were
the non linear regression analysis of the log dose response curves
(see Supporting Information).
tested at additional concentrations and IC50 was determined using
at least five concentrations of the inhibitor causing an inhibition
between 10% and 90%, using the formula: Vi/Vo ¼ 1/(1 þ [I]/IC50),
where Vi is the initial velocity of substrate cleavage in the presence
of the inhibitor at concentration [I] and Vo is the initial velocity in
the absence of the inhibitor. Results were analyzed using SoftMax
Pro software [24] and Origin 6.0 software.
As shown in Table 1, all inhibitors scaffolds are structurally
diverse from each other and from most of the known MMPIs. With
the exception of 5 (and maybe 4, see paragraph “Active Compounds
Binding Modes and Hints for Lead Optimization”), all active
compounds possess a carboxylate function as ZBG. Compound 5
which holds a dimethoxybenzene as ZBG retains a certain activity

148
V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
Table 1
Structures, labels, AD4 binding free energies and IC50 of MMP-13 inhibitors identified with VS experiments.
Chemical structure
Life chemicals code
D
GAD4 (Kcal/mol)
IC50
14
(m
M)a
S
N
HN
1
F0920-6501
ꢀ13.33
OH
O
O
O
O
S
N
2
F1074-0280
ꢀ13.12
22
O
OH
O
H
N
S
3
F1204-0078
ꢀ10.96
67
O
HO
O
NH
O
O
O
O
4
F1542-0089
ꢀ12.11
120
N
S
COOH
S
5
F0807-0342
ꢀ10.5
140
N
N
O
N
O
S
a
IC50 values represent the concentration required to produce 50% enzyme inhibition.
although his IC50 (140
m
M) is higher than all of the carboxylate-
ClogD, and TPSA, were in silico calculated for our five leads and
compared to those of 24f (Table 2) [29]. As shown in Table 2, with
the exception of 5, which seems to be the least drug-like
compound, all other inhibitors possess an average value of ClogP
ranging from 0.89 to 3.32 and a ClogD and a TPSA very similar to
that of 24f. Thus, with the exception of 5, all the others seem to be
suitable leads, for which the S10 substituent could be easily
extended and/or modified.
containing inhibitors. Recently, Novartis researchers reported that
a series of carboxylic acids such as the MMP-13 inhibitor 24f [28]
were orally available and equipotent to the most potent hydroxa-
mic acid based inhibitors in in vivo models of cartilage protection.
Thus, some key physicochemical properties, such as pKa, ClogP,
Table 2
In silico calculated physicochemical properties of compounds 1e5 and 24f.
b
d
Compd
pKa a
ClogP
ClogD c
TPSA (Å)
Table 3
1
2
3
4
3.91
3.89
3.68
3.62
e
3.32
2.6
0.89
2.91
5.55
3.39
0.11
ꢀ0.61
ꢀ2.56
ꢀ0.42
5.55
119.96
99.98
119.75
137.37
71.71
In Vitroa activity (IC50
MMPs.
mM values) of the novel MMP-13 inhibitors towards diverse
Compd Life chemicals MMP-1
code
MMP-2
MMP-3
MMP-13 MMP-14
5
24fe
2.55
ꢀ0.13
158.86
1
2
3
4
5
F0920-6501
F1074-0280
F1204-0078
F1542-0089
F0807-0342
400 ꢂ 150 67 ꢂ 3.0 110 ꢂ 15
93 ꢂ 8.0 2.7 ꢂ 0.2 110 ꢂ 26
14 ꢂ 0.5 51 ꢂ 7.0
22 ꢂ 0.6 21 ꢂ 2.0
67 ꢂ 10 55 ꢂ 4.0
a
pKa predictions refers to the ZBG.
114 ꢂ 23
61 ꢂ 7.0 77 ꢂ 21
b
c
Calculated n-octanol/water partition coefficient.
Calculated distribution coefficient at pH ¼ 7.4.
Topological polar surface area.
860 ꢂ 110 350 ꢂ 38 850 ꢂ 200 120 ꢂ 8 310 ꢂ 18
360 ꢂ 46 120 ꢂ 14 230 ꢂ 24 140 ꢂ 10 150 ꢂ 18
d
e
a
Orally active carboxylic acid-derived MMP-13 inhibitor used for comparison
Assays were run in triplicate. The final values given here are the mean ꢂSD of
purpose [19].
three independent experiments.

V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
149
Fig. 1. Docked conformations of 1 in the MMP-13 catalytic site. Hydrogen atoms are
omitted for the sake of clarity. Ligands carbon atoms are displayed in golden, and key
binding site residues as cyan sticks.
8.2. Biological evaluation
The inhibitory activity of the five novel leads was evaluated
(Table 3) against a panel of MMP isozymes (MMP-1, -2, -3, -13, -14),
some of which are implicated in cartilage degradation. Over the five
inhibitors, two (1 and 4) are definitely more active on MMP-13
showing appreciably weaker activity on all the other tested
enzymes (Table 3). In this respect, both compounds represent
appealing leads amenable of structural modification to develop
selective MMP-13 inhibitors. Inhibitors 3, and 5 are equally active
on MMP-13 and MMP-14. The two compounds show inhibitory
activity also towards MMP-2. In this respect, it is not clear whether
this inhibitory profile is beneficial in terms of protecting cartilage
degradation. Actually, the role of MMP-2 activity itself in the
pathogenesis of OA is unclear. Interestingly, mRNA levels of MMP-2
are increased in OA patients compared to normal controls, sug-
gesting that MMP-2 may play a role in this disease [30]. On the
other hand, MMP-2-null mice exhibit a more severe arthritic
phenotype than wild type mice in antigen-induced arthritis, sug-
gesting that the total loss of MMP-2 activity is unfavourable [31].
Differently, compound 2 shows a certain preference for MMP-2
Fig. 2. Docked conformations of 2 (a) and 3 (b) In the MMP-13 catalytic site.
with Q262 and M264 in MMP-14. This hypotheses is confirmed
by inhibitor 2 (Fig. 2a), which shows the same activity on MMP-13
and MMP-14 possessing a thin olefinic chain ending with a phenyl
ring which is unable to fill the roomy S10 pocket. Differently from
inhibitor 2, compound 3 has a small and polar P10 group, and this is
the reason for the lower activity and selectivity for MMP-13 with
respect to
1 and 2. However in 3 (Fig. 2b), the thio-
xoimidazolidinone ring could be substituted with groups featuring
shapes and electrostatic propertied able to favourably interact with
the peculiar S10 tunnel of MMP-13. Especially for this compound,
the extension of the P10 group is certainly a priority step.
As regards compound 4 (Fig. 3a), molecular docking unambig-
uously indicate that the ZBG would be the carboxylate group and
not the rhodanine ring via the thiazolidine sulphur atom, as
(IC50 ¼ 2.7
mM) and could be developed as novel antitumor agent.
8.3. Active compounds binding modes and general hints for leads
optimization
Besides the carboxylate function, which, with the exception of 5,
is a conserved feature of all active inhibitors, the five compounds
deeply differ in their chemical structures. Indeed, in 1, the
carboxylate moiety is directly attached to a benzene ring, in 2 this
portion is linked to a thiazolidindione nucleus by a propyl-linker,
while in 3 and 4 a oxymethylene and a methylene bridge link the
carboxylate group to a benzene and thioxothiazolidinone ring,
respectively. Regardless the structural dissimilarities among the
aforementioned ligands, all of them are characterized by a small
number of rotatable bonds (ranging from 0 to 4). Indeed, the
rigidity of 1 allows the proper orientation of the ZBG to chelate the
catalytic zinc ion and the P10 group into the S10 pocket (see Fig. 1).
The imidazolethione ring is in a suitable position to establish a pep
interaction with H222 side chain. The micromolar IC50 for this
compound might be due to the non-optimized interaction between
the P10 group and the S10 pocket and to a too high rigidity. The
selectivity of 1 towards the MMP-13 is surely ascribable to the
bulky chromenone nucleus located into the unusually large S10
specificity pocket. In fact, although MMP-13 and -14 possess a S10
specificity loop of the same length, the latter has a narrower shaped
S10 pocket, due to the substitution of T245 and T247 in MMP-13
Fig. 3. Docked conformations of 4 (a) and 5 (b) In the MMP-13 catalytic site.

150
V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
Table 4
previously found for anthrax lethal factor inhibitors [32], which
have in common with compound 4 both the rhodanine ring and the
carboxylate group. However, a search in the Cambridge Structural
Databases shows that, at least in absence of any receptor structure,
the carboxylate moiety prevails onto the rhodanine ring in the
coordination of metal ions. Thus, prior of any rational optimization,
further studies have to be conducted in order to assess the real
binding mode of 4 in the MMP-13 active site.
Chemical structure and inhibitor activity of compound
compounds.
1 and the prepared
S
HN
N
Z
Inhibitor 5 (Fig. 3b), is the weakest inhibitor (IC50 on MMP-
O
O
13 ¼ 140
mM) on the entire panel of MMPs tested, although it is the
X
only one whose P10 group is able to make some contacts with the
W
entrance residues of the S10 pocket like P242, V219, as well as
a parallel
p-stacking with the H222. The thiazolidine ring makes
some lipophilic contacts with the S10 pocket floor residues (L184
and L185), while the N-benzylidene group projects itself towards
the beta carbons of the Y244 and the I243.
Chemical structure
W
Z
X
IC50 (mM)
1
H
H
H
OCH3
H
OH
H
OCH3
OCH3
H
H
H
H
OCH3
H
Cl
COOH
H
COOCH2CH3
COOH
COOH
COOH
COOH
COOH
COOH
14 ꢂ 1.4
177 ꢂ 14
>300
5.2 ꢂ 1.1
3.2 ꢂ 0.2
5.5 ꢂ 1.1
2.6 ꢂ 0.6
4.9 ꢂ 0.6
4.9 ꢂ 0.8
1a
1b
1c
1d
1e
1f
In this case, the low activity is ascribable to the presence of
a putative weak zinc ion chelator (dimethoxybenzene) and to the
fact that it has been tested as a mixture of diastereoisomers. Thus,
separation and testing of each single diastereoisomer, together
with the substitution of the weak chelator moiety with a stronger
one, could be the first step of the lead optimization process of this
inhibitor. Subsequent steps could include proper substitutions of
both phenyl rings to enhance the interaction with the S10 and S30
pockets.
1g
1h
Cl
CH2CH]CH2
O
O
H
Z
Br
Z
N
H2N
X
O
O
O
O
NaHCO3
EtOH/Δ
W
W
X
6-12
13-20
6
7
8
9
W = H, Z = H
NH4SCN
W = OCH3, Z = H
W = H, Z = OCH3
W = OH, Z = H
CH3COOH
60-65°C
10 W = H, Z = Cl
11 W = OCH3, Z = Cl
X
12 W = OCH3, Z = CH2CH=CH2
N
S
Z
N
H
O
O
W
1a-h
13, 1a W = H, Z = H, X = H
14, 1b W = H, Z = H, X = COOCH2CH3
15, 1c W = OCH3, Z = H, X = COOH
16, 1d W = H, Z = OCH3, X = COOH
17, 1e W = OH, Z = H, X = COOH
18, 1f W = H, Z = Cl, X = COOH
19, 1g W = OCH3, Z = Cl, X = COOH
20, 1h W = OCH3, Z = CH2CH=CH2, X = COOH
Scheme 1.

V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
151
O
O
O
O
O
Z
CHO
OH
Z
CuBr2
CHCl3/ACOEt
Δ
Z
Br
O
O
O
O
Piperidine
EtOH, rt
W
W
W
21-26
7-12
7, 21 W = OCH3, Z = H
8, 22 W = H, Z = OCH3
9, 23 W = OH, Z = H
10, 24 W = H, Z = Cl
11, 25 W = OCH3, Z = Cl
12, 26 W = OCH3, Z = CH2CH=CH2
Scheme 2.
Thus, generally speaking, none of these compounds has such
an extended P10 group to occupy the whole S10 tunnel of the
MMP-13, neither the P10 are well-optimized to interact with the
pocket. This is certainly the reason for the inhibitory activities in
which has been fully confirmed by this first round of optimization,
suggests that for further activity optimization, more flexibility has
to be conferred to 1. Thus, design and synthesis of novel, more
flexible, derivatives will be the next step of our research program
and it is expected to led to the desired activity.
the range of mM. However, a rationally designed lead optimization
project will surely increase the experimental IC50. In fact, even if
less potent than hydroxamate-based inhibitors, carboxylates
could be a valid alternative to this moiety. This weaker zinc-
binder could allow to have selective inhibition if present in
properly optimized structures. For 1 and 4, which show a pretty
good selectivity profile, lead optimizations are already an ongoing
projects. Herein we reported the first round of structure-based
optimization on lead 1.
8.5. Chemistry
The general procedure for the synthesis of the target inhibitors
1aeh involved the reaction of the 3-bromoacetylcoumarin deriva-
tives 6e12 with the appropriate arylamine (aniline, 3-aminobenzoic
acid, ethyl 3-aminobenzoate) in ethanol at room temperature for 5 h
(TLC analysis), in the presence of NaHCO3, to yield products 13e20
(Scheme 1), which crystallized pure by cooling the reaction mixture.
Derivatives 1aeh were then obtained by treatment of compounds
13e20 with a largeexcess of ammonium thiocyanate in acetic acid at
60e65 ꢃC for 1.5 h (TLC analysis). After work-up of the reaction,
washing with water, cold acetonitrile, and cold diethyl ether,
derivatives 1aeh were directly obtained with the desired purity
degree (ꢄ95%). The 3-bromoacetylcoumarin 6 is commercially
available. The substituted coumarin derivatives 7e12 were easily
prepared as outlined in Scheme 2. Following a reported procedure
with slight modifications, the 3-acetylcoumarins 21e26 were ob-
tained by a Knoevenagel condensation of the appropriate salicy-
laldehyde (2-hydroxy-3-methoxybenzaldehyde, 2-hydroxy-5-
methoxybenzaldehyde, 2,3-dihydroxybenzaldehyde, 5-chloro-2-
8.4. Receptor-based lead optimization of 1
As first step in our lead optimization process we verified the
reliability of the proposed binding mode, synthetizing and testing
the decarboxylated analogue (1a) and the ethyl ester derivative
(1b) of compound 1. The IC50 obtained of 1a and 1b (Table 4) fully
substantiate that the carboxylate group binds the catalytic zinc ion
as predicted by the docking program. As next step, the first round of
lead optimization was rationally carried out by taking into account
the theoretical binding pose of 1 in the MMP-13 binding site. As
aforesaid, the rigidity of 1 allows the proper orientation of the ZBG
to chelate the catalytic zinc ion and the P10 group into the S10
hydroxybenzaldehyde,
5-chloro-2-hydroxy-3-methoxybenz
pocket, while the imidazolethione ring establishes a
pep interac-
aldehyde, 5-allyl-2-hydroxy-3-methoxybenzaldehyde) and ethyl
acetoacetate, using piperidine as a catalyst in absolute ethanol at
room temperature for 5 h (TLC analysis) (Scheme 2), Crystallization
from ethanol gave the pure products 21e26. These latter
compounds were then transformed into the corresponding 2-
bromoacetyl derivatives 7e12 by treatment with copper (II)
bromide and reflux in chloroform-ethyl acetate. The reaction was
complete as judged by a colour change of the solution from green to
amber, disappearance of all black solid, and cessation of hydrogen
bromide evolution. Products 7e12 were finally purified by crystal-
lization from ethanol.
tion with H222 side chain. Thus, the micromolar IC50 for this
compound might be ascribed to the non-optimized interaction
between the P10 group and the S10 pocket. In the first step of lead
optimization we decided to explore positions 6- and 8- of the
chromenone ring, while positions 4- and 5- will be object of a next
round of optimization. However, it has to be said that close to the
inhibitor core there is the L218 (distance between hydrogen in
position 4- and L218
in position 5- and
b
carbon is 3.60 Å, distance between hydrogen
d
1 carbon of L218 is 3.20 Å) which would not
tolerate bulky substituents. On the basis of the proposed binding
mode, substitutions on position 7 should not help in increasing the
number of interactions with the enzyme as this position is solvent
exposed. Thus, in the attempt to allow 1 to establish further
contacts with the upper part of the S10-specificity loop, eg. an H-
bond with the A238 carbonyl oxygen or an hydrophobic interaction
with the P255 side chain the 8-OH (1e) and the 8-OCH3 (1c)
derivatives were respectively prepared (Table 4). These small
modifications to 1 were successful in lowering the IC50 (Table 4).
Furthermore, with the aim of enhancing the lipophilic interactions
with F252 and T247, at the bottom of the S10 pocket, the 6-OCH3
(1d) and 6-Cl (1f) derivatives were synthesized. In this case the
improvement in the inhibition toward the MMP-13 was even
9. Conclusions
This paper reports the identification of structurally non-classic
MMP-13 inhibitors by means of an in silico screening of 7769
compounds to the active site of the enzyme. Experimental evalua-
tionof20candidates, whichwereselectedbyvisualinspectionof the
poses predicted for the best scoring compounds, led to the identi-
fication of five novel zinc-chelating non-hydroxamate inhibitors,
structurally distinct from those already reported. All five may
provide scaffolds upon which to develop compounds with more
desirable properties, such as selectivity of action and oral avail-
ability. A first round of structure-based optimization on lead 1 was
higher with an IC50 of 3.2
mM and 2.6 mM, correspondingly.
However, a close inspection of the binding mode predicted for 1,

152
V. La Pietra et al. / European Journal of Medicinal Chemistry 47 (2012) 143e152
(b) E.C. Horning, M.G. Horning, Coumarins from 2-hydroxy-3-
methoxybenzaldehyde, J. Am. Chem. Soc. 69 (1947) 968e969.
[15] I.Y. Flores-Larios, L. Lopez-Garrido, F.J. Martinez-Martinez, J. Gonzalez,
E.V. Garcia-Baez, A. Cruz, I.I. Padilla-Martinez, Thermal [4þ2] cycloadditions of
3-acetyl-, 3-carbamoyl-, and 3-ethoxycarbonyl-coumarins with 2,3-dimethyl-
1,3-butadiene under solventless conditions: a structural study, Molecules 15
(2010) 1513e1530.
[16] X. Liu, J. Fan, Y. Liu, Z. Shang, L-Proline as an efficient and reusable promoter
for the synthesis of coumarins in ionic liquid, J. Zhejiang Univ. Sci. B 9 (2008)
990e995.
accomplished and led to an increase in potency of more than 5 fold.
The next round of optimization, where more flexible derivatives will
be designed and synthesized is expected to lead to the desired
activity. Finally the herein reported work demonstrated that the use
of virtual screening can represent a successful method for the
discovery of novel MMPIs with a chemical structure diverse from
each other and from most of the known MMPIs.
[17] X. Liu, J. Fan, Y. Liu, Z. Shang, L-proline as an efficient and reusable promoter
for the synthesis of coumarins in ionic liquid, J. Zhejiang Univ. Sci. B 9 (2008)
990e995.
Appendix. Supplementary material
[18] V.R. Rao, T.V.P. Rao, Studies on coumarin derivatives. Part I. Synthesis of some
substituted thiazolyl and benzoxazinylcoumarins, Indian J. Chem. Sect. B. Org.
Chem. Including Med. Chem. 25 (1986) 413e415.
[19] C. Murata, T. Masuda, Y. Kamochi, K. Todoroki, H. Yoshida, H. Nohta,
M. Yamaguchi, A. Takadate, Improvement of fluorescence characteristics of
coumarins: syntheses and fluorescence properties of 6-methoxycoumarin and
benzocoumarin derivatives as novel fluorophores emitting in the longer
wavelength region and their application to analytical reagents, Chem. Pharm.
Bull. 53 (2005) 750e758.
[20] G.K. Rao, K.N. Venugopala, P.N.S. Pai, Microwave-assisted synthesis of some 6-
chloro-3-[2-(substitutedanilino)-1,3-thiazol-4-yl]-2H-1-benzopyran-2-ones
as antibacterial agents, Indian J. Heterocyclic Chem. 17 (2008) 397e400.
[21] C.F. Koelsch, Bromination of 3-acetocoumarin, J. Am. Chem. Soc. 72 (1950)
2993e2995.
[22] S.S. Soman, T.H. Thaker, Synthesis of new 3-(1-aryl-2-thioxo-3,4-dihydro-2H-
imidazol-4-yl)chromen-2-ones, Indian J. Heterocyclic Chem. 19 (2009) 55e58.
[23] A. Rossello, E. Nuti, E. Orlandini, P. Carelli, S. Rapposelli, M. Macchia,
F. Minutolo, L. Carbonaro, A. Albini, R. Benelli, G. Cercignani, G. Murphy,
A. Balsamo, New N-arylsulfonyl-N-alkoxyaminoacetohydroxamic acids as
selective inhibitors of gelatinase A (MMP-2), Bioorg. Med. Chem. 12 (2004)
2441e2450.
Supplementary data related to this article can be found online at
doi:10.1016/j.ejmech.2011.10.035.
References
[1] J.S. Mort, C.J. Billington, Articular cartilage and changes in arthritis: matrix
degradation, Arthritis Res. 3 (2001) 337e341.
[2] (a) A. Jüngel, C. Ospelt, M. Lesch, M. Thiel, T. Sunyer, O. Schorr, B.A. Michel,
R.E. Gay, C. Kolling, C. Flory, S. Gay, M. Neidhart, Effect of the oral application
of a highly selective MMP-13 inhibitor in three different animal models of
rheumatoid arthritis, Ann. Rheum. Dis. 69 (2010) 898e902;
(b) N.G. Li, Z.H. Shi, Y.P. Tang, Z.J. Wang, S.L. Song, L.H. Qian, D.W. Qian,
J.A. Duan, New hope for the treatment of osteoarthritis through selective
inhibition of MMP-13, Curr. Med. Chem. 18 (2011) 977e1001.
[3] (a) M. Sabatini, C. Lesur, M. Thomas, A. Chomel, P. Anract, G. deNanteuil,
P. Pastoureau, Effect of inhibition of matrix metalloproteinases on cartilage
loss in vitro and in a guinea pig model of osteoarthritis, Arthritis Rheum. 52
(2005) 171e180.(b) V.M. Baragi, G. Becher, A.M. Bendele, R. Biesinger, H. Bluhm,
J. Boer, H. Deng, R. Dodd, M. Essers, T. Feuerstein, B.M. Gallagher, C. Gege,
M. Hochgürtel, M. Hofmann, A. Jaworski, L. Jin, A. Kiely, B. Korniski, H. Kroth,
D. Nix, B. Nolte, D. Piecha, T.S. Powers, F. Richter, M. Schneider, C. Steeneck,
I. Sucholeiki, A. Taveras, A. Timmermann, J. Van Veldhuizen, J. Weik, X. Wu,
B. Xia, A new class of potent matrix metalloproteinase 13 inhibitors for potential
treatment of osteoarthritis: evidence of histologic and clinical efficacy without
musculoskeletal toxicity in rat models, Arthritis Rheum. 60 (2009) 2008e2018.
[4] R. Renkiewicz, L. Qiu, C. Lesch, X. Sun, R. Devalaraja, T. Cody, E. Kaldjian,
H. Welgus, V. Baragi, Broad-spectrum matrix metalloproteinase inhibitor
marimastat-induced musculoskeletal side effects in rats, Arthritis Rheum. 48
(2003) 1742e1749.
[24] SoftMax Pro 4.7.1 by Molecular Devices.
[25] S.F. Sousa, P.A. Fernandes, M.J. Ramos, Proteineligand docking: current status
and future challenges, Proteins 65 (2006) 15e26.
[26] (a) S. Cosconati, L. Marinelli, R. Trotta, A. Virno, L. Mayol, E. Novellino,
A.J. Olson, A. Randazzo, Tandem application of virtual screening and NMR
experiments in the discovery of brand new DNA quadruplex groove binders,
J. Am. Chem. Soc. 131 (2009) 16336e16337;
(b) S. Cosconati, J.A. Hong, E. Novellino, K.S. Carroll, D.S. Goodsell, A.J. Olson,
Structure-based virtual screening and biological evaluation of Mycobacterium
tuberculosis adenosine 50-phosphosulfate reductase inhibitors, J. Med. Chem.
51 (2008) 6627e6630;
[5] H. Wu, J. Du, Q. Zheng, Expression of MMP-1 in cartilage and synovium of
experimentally induced rabbit ACLT traumatic osteoarthritis: immunohisto-
chemical study, Rheumatol. Int. 29 (2008) 31e36.
[6] J.J. Li, J. Nahra, A.R. Johnson, A. Bunker, P. O0Brien, W.S. Yue, D.F. Ortwine,
C.F. Man, V. Baragi, K. Kilgore, R.D. Dyer, H.K. Han, Quinazolinones and pyrido-
(c) S. Cosconati, L. Marinelli, C. La Motta, S. Sartini, F. Da Settimo, A.J. Olson,
E. Novellino, Pursuing aldose reductase inhibitors through in situ cross-docking
and similarity-based virtual screening, J. Med. Chem. 52 (2009) 5578e5581.
[27] (a) S.L. McGovern, E. Caselli, N. Grigorieff, B.K. Shoichet, A common mecha-
nism underlying promiscuous Inhibitors from virtual and high-throughput
screening, J. Med. Chem. 45 (2002) 1712e1722;
[3,4-d]pyrimidin-4-ones
as
orally
active
and
specific
matrix
metalloproteinase-13 inhibitors for the treatment of osteoarthritis, J. Med.
Chem. 51 (2008) 835e841.
(b) S.L. McGovern, B.T. Helfand, B. Feng, B.K. Shoichet, A specific mechanism of
non-specific inhibition, J. Med. Chem. 46 (2003) 4265e4272.
[7] E. Nuti, F. Casalini, S.I. Avramova, S. Santamaria, G. Cercignani, L. Marinelli,
V. La Pietra, E. Novellino, E. Orlandini, S. Nencetti, T. Tuccinardi, A. Martinelli,
N.H. Lim, R. Visse, H. Nagase, A. Rossello, NeOeisopropyl sulfonamido-based
hydroxamates: design, synthesis and biological evaluation of selective matrix
metalloproteinase-13 inhibitors as potential therapeutic agents for osteoar-
thritis, J. Med. Chem. 52 (2009) 4757e4773.
[8] M. Flipo, J. Charton, A. Hocine, S. Dassonneville, B. Deprez, R. Deprez-Poulain,
Hydroxamates: relationships between structure and plasma stability, J. Med.
Chem. 52 (2009) 6790e6802.
[28] L.G. Monovich, R.A. Tommasi, R.A. Fujimoto, V. Blancuzzi, K. Clark,
W.D. Cornell, R. Doti, J. Doughty, J. Fang, D. Farley, J. Fitt, V. Ganu, R. Goldberg,
R. Goldstein, S. Lavoie, R. Kulathila, W. Macchia, D.T. Parker, R. Melton,
E. O’Byrne, G. Pastor, T. Pellas, E. Quadros, N. Reel, D.M. Roland, Y. Sakane,
H. Singh, J. Skiles, J. Somers, K. Toscano, A. Wigg, S. Zhou, L. Zhu, W.C. Shieh,
S. Xue, L.W. McQuire, Discovery of potent, selective, and orally active
carboxylic acid based inhibitors of matrix metalloproteinase-13, J. Med. Chem.
52 (2009) 3523e3538.
[29] The pKa, CLogP, CLogD calculations were performed with the chemsilico
server on the web site: http://chemsilico.com. The TPSA calculation was
performed with marvin sketch and calculator plugins on the web-site: http://
www.chemaxon.com/demosite/marvin/index.html.
[30] L. Kevorkian, D.A. Young, C. Darrah, S.T. Donell, L. Shepstone, S. Porter,
S.M. Brockbank, D.R. Edwards, A.E. Parker, I.M. Clark, Expression profiling of
metalloproteinases and their inhibitors in cartilage, Arthritis Rheum. 50
(2004) 131e141.
[31] T. Itoh, H. Matsuda, M. Tanioka, K. Kuwabara, S. Itohara, R. Suzuki, The role of
matrix metealloproteinase-2 and matrix metalloproteinase-9 in antibody
induced arthritis, J. Immunol. 169 (2002) 2643e2647.
[9] E. Nuti, L. Panelli, F. Casalini, S.I. Avramova, E. Orlandini, S. Santamaria,
S. Nencetti, T. Tuccinardi, A. Martinelli, G. Cercignani, N. D’Amelio, A. Maiocchi,
F. Uggeri, A. Rossello, Design, synthesis, biological evaluation, and NMR
studies of
a new series of arylsulfones as selective and potent matrix
metalloproteinase-12 inhibitors, J. Med. Chem. 52 (2009) 6347e6361.
[10] www.lifechemicals.com
[11] http://zinc.docking.org
[12] R. Huey, G.M. Morris, A.J. Olson, D.S. Goodsell, A semiempirical free energy force
field with charge-based desolvation, J. Comput. Chem. 28 (2007) 1145e1152.
[13] E.F. Pettersen, T.D. Goddard, C.C. Huang, G.S. Couch, D.M. Greenblatt,
E.C. Meng, T.E. Ferrin, UCSF, Chimeraea visualization system for exploratory
research and analysis, J. Comput. Chem. 25 (2004) 1605e1612.
[14] (a) S. Stanchev, G. Momekov, F. Jensen, I. Manolov, Synthesis, computational
study and cytotoxic activity of new 4-hydroxycoumarin derivatives, Eur. J.
Med. Chem. 43 (2008) 694e706;
[32] S.L. Johnson, D. Jung, M. Forino, Y. Chen, A. Satterthwait, D.V. Rozanov,
A.Y. Strongin, M. Pellecchia, Anthrax lethal factor protease inhibitors:
synthesis, SAR, and structure-based 3D QSAR studies, J. Med. Chem. 49 (2006)
27e30.