Design, synthesis and biological evaluation of bifunctional inhibitors of membrane type 1 matrix metalloproteinase (MT1-MMP)

Article abstract of DOI:10.1016/j.bmc.2018.11.041

Collagen degradation and proMMP-2 activation are major functions of MT1-MMP to promote cancer cell invasion. Since both processes require MT1-MMP homodimerization on the cell surface, herein we propose that the use of bifunctional inhibitors of this enzym

Full text of DOI:10.1016/j.bmc.2018.11.041

Accepted Manuscript
Design, Synthesis and Biological Evaluation of Bifunctional Inhibitors of Mem-
brane Type 1 Matrix Metalloproteinase (MT1-MMP)
Doretta Cuffaro, Elisa Nuti, Valentina Gifford, Noriko Ito, Caterina Camodeca,
Tiziano Tuccinardi, Susanna Nencetti, Elisabetta Orlandini, Yoshifumi Itoh,
Armando Rossello
PII:
S0968-0896(18)31373-7
https://doi.org/10.1016/j.bmc.2018.11.041
BMC 14645
DOI:
Reference:
Bioorganic & Medicinal Chemistry
To appear in:
Received Date:
Revised Date:
Accepted Date:
31 July 2018
27 November 2018
28 November 2018
Please cite this article as: Cuffaro, D., Nuti, E., Gifford, V., Ito, N., Camodeca, C., Tuccinardi, T., Nencetti, S.,
Orlandini, E., Itoh, Y., Rossello, A., Design, Synthesis and Biological Evaluation of Bifunctional Inhibitors of
Membrane Type 1 Matrix Metalloproteinase (MT1-MMP), Bioorganic & Medicinal Chemistry (2018), doi: https://
doi.org/10.1016/j.bmc.2018.11.041
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers
we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and
review of the resulting proof before it is published in its final form. Please note that during the production process
errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

Design, Synthesis and Biological Evaluation of Bifunctional Inhibitors of Membrane Type 1
Matrix Metalloproteinase (MT1-MMP)
Doretta Cuffaro,1 Elisa Nuti,1 Valentina Gifford,2 Noriko Ito,2 Caterina Camodeca,1 Tiziano
Tuccinardi,1 Susanna Nencetti,1 Elisabetta Orlandini,3 Yoshifumi Itoh,2 and Armando Rossello1,*
1 Department of Pharmacy, Università di Pisa, via Bonanno 6, 56126 Pisa, Italy.
2 Kennedy Institute of Rheumatology, Nuffield Department of Orthopaedics, Rheumatology and
Musculoskeletal Sciences, University of Oxford, Oxford OX3 7FY, United Kingdom
3 Department of Earth Sciences, University of Pisa, Via Santa Maria 53, 56126 Pisa, Italy
*Corresponding author: Armando Rossello, phone, +39 050 2219562, fax, +39 050 2219605, e-mail:
armando.rossello@farm.unipi.it
Keywords: MT1-MMP homodimerization, bifunctional inhibitors, arylsulfonamide hydroxamates,
MMP inhibitors.
1

Abstract
Collagen degradation and proMMP-2 activation are major functions of MT1-MMP to promote cancer
cell invasion. Since both processes require MT1-MMP homodimerization on the cell surface, herein we
propose that the use of bifunctional inhibitors of this enzyme could represent an innovative approach to
efficiently reduce tumor growth. A small series of symmetrical dimers derived from previously
described monomeric arylsulfonamide hydroxamates was synthesized and tested in vitro on isolated
MMPs. A nanomolar MT1-MMP inhibitor, compound 6, was identified and then submitted to cell-
based assays on HT1080 fibrosarcoma cells. Dimer 6 reduced MT1-MMP-dependent proMMP-2
activation, collagen degradation and collagen invasion in a dose-dependent manner with better results
even compared to its monomeric analogue 4. This preliminary study suggests that dimeric MT1-MMP
inhibitors might be further developed and exploited as an alternative tool to reduce cancer cell invasion.
Introduction
Matrix metalloproteinases (MMPs) are a family of zinc-dependent endopeptidases responsible for
extracellular matrix (ECM) degradation and remodeling. In humans, there are 23 MMPs, classified into
soluble and membrane-type MMPs (MT-MMPs). Soluble-types are secreted into the extracellular space
and can diffuse into the ECM, while MT-MMPs are attached to the plasma membrane via a
transmembrane (TM) domain (MT1-MMP, MT2-MMP, MT3-MMP, and MT5-MMP) or a
glycosylphosphatidylinositol (GPI) anchor (MT4-MMP and MT6-MMP).1 MT1-MMP is one of the
most studied enzymes in the MMP family.2 It is a proteolytic enzyme that has a broad substrate
specificity, especially against ECM components: fibrillar collagens (type I, II, III), laminin-1 and -5,
fibronectin, vitronectin, fibrin and aggrecan. 3 MT1-MMP effectively promotes cell migration by
degrading pericellular ECM at migration front of cell.
MT1-MMP can also degrade ECM components indirectly by activating proMMP-24 and proMMP-135.
Active MMP-2 cleaves type IV collagen, one of the major components of the basement membrane,
whose hydrolysis represents an important step for cancer invasion, growth and angiogenesis.6,7 MMP-2
activation process on the cell surface requires two MT1-MMP molecules forming a dimeric interaction
through their hemopexin (Hpx)8 and transmembrane (TM) domain.9 One of the dimer MT1-MMP is
bound to its endogenous inhibitor, TIMP-2 (Tissue Inhibitor of Metalloproteinases), resulting in a
trimeric complex. Once the trimolecular complex has been formed, exposed TIMP-2 C-terminal
domain attracts proMMP-2 through binding its Hpx domain, and the other MT1-MMP that is not
bound to TIMP-2 can cleave the pro-domain of MMP-2, initiating activation. The disruption of the
2

dimer formation effectively inhibits the MT1-MMP mediated proMMP-2 activation.8 Dimerization
through its Hpx domain was also found to be crucial for MT1-MMP to degrade collagen on the cell
surface.10
Given its involvement in cell migration, tumor growth and angiogenesis, many attempts have been
made to develop selective inhibitors of MT1-MMP. Recent examples of inhibitors targeting the
catalytic domain of MT1-MMPs are represented by a Val-PO2-Leu phosphoroamidate scaffold
developed by Mendes et al 11 and by a cyclic peptide reported by Suojanen et al12. Accomplishing the
required target specificity with small-molecule MMP-inhibitors (MMPIs) is exceedingly difficult, and
so far, the success has been limited due to the high homology of the catalytic domain in the MMP
family members. An alternative strategy to inhibit MT1-MMP could be targeting the hemopexin
domain which is less conserved among MMPs than the catalytic site and was shown to be necessary for
MT1-MMP dimerization. In this regards, Remacle et al13 in 2012 reported the first MT1-MMP
hemopexin-targeting small molecule, NSC405020, able to inhibit in vivo the tumor growth of MCF7
cells in animal model of breast cancer. In the same years, other research groups identified anti-
hemopexin domain inhibitors for MMP-9 able to reduce cell migration and tumor growth,14 thus
validating this strategy. Our previous studies have shown that bifunctional ligands able to interact with
the catalytic site of two MMP-9 monomers could be used to impair MMP-9 homodimerization by
preventing the Hpx domains to approach each other. The best biological results were obtained with a
dimeric dicarboxylate inhibitor (1, Figure 1), a nanomolar inhibitor of MMP-9, which was able to arrest
cancer cell invasiveness in a human glioblastoma cell line.15 Moreover, we have already reported two
potent MMP-2, MMP-9 and MT1-MMP dimeric inhibitors 2 and 3 (Figure 1), endowed with an
appreciably reduced cell toxicity with respect to their corresponding monomeric analogues.16
3

Figure 1. Structure of the previously reported dimeric MMP inhibitors 1, 2 and 3.
Dimeric inhibitors have already been studied in several fields of medicinal chemistry, such as kinase
inhibitors, HIV protease inhibitors, cholinesterase inhibitors and MMP inhibitors, with promising
results.17,18
Based on these considerations, the aim of this work was to develop a new class of MT1-MMP
bifunctional inhibitors able to potentially inhibit dimer MT1-MMP on tumor cell surface in efficient
manner through the binding to the catalytic site of two MT1-MMP molecules. Looking into our in-
house library of compounds, we chose two potent inhibitors of MT1-MMP, the N-isopropoxy-
arylsulfonamide hydroxamate 419 (Figure 2) and the arylsulfonamide 520 (Figure 3), as hit compounds
to be modified. Both these derivatives had a nanomolar affinity for the target enzyme and 4 was able to
potently arrest angiogenesis in vitro and also in vivo in the matrigel sponge assay in mice. X-ray
crystallographic and docking studies performed on compound 4 have confirmed the importance of the
phthalimidic group in P1 position to allow a good positioning of the molecule in the catalytic domain
of MT1-MMP.19 On the basis of these findings, compounds 4 and 5 were taken as starting points to
4

design new selective bifunctional inhibitors of MT1-MMP. We decided to synthesize and test
compounds 6 and 7 (Figure 2) as dimeric derivatives of hydroxamate 4. They maintained the
phthalimidic group in P1 position and the p-methoxybiphenyl sulfonamido substituent in P1’ necessary
for a good fitting in the MT1-MMP catalytic domain. Moreover, they presented the dimerization in P2’
position on the basis of the good results obtained with the bifunctional ligand of MMP-9, compound 1.
In addition, we developed the bifunctional derivatives of 5, compounds 8 and 9 (Figure 3), dimerized in
P1 position as reported for our previously published dimers 2 and 3.16 Dimers 6-9 were synthesized
using both diethyleneisophthalamide (DEI) and N-N’-bis-(3-ethylencarbamoyl-propyl)-isophthalamide
(BEPI) as linkers between the two chelating heads, in order to investigate the effect of dimers of
different length on biological activity.
Figure 2. Structure of compound 4 and its dimeric derivatives 6 and 7.
5

Figure 3. Structure of compound 5 and its dimeric derivatives 8 and 9.
All new compounds were first assayed on human recombinant MMPs and then the most promising
were tested on HT1080 cells to prove their ability to block MT1-MMP collagenolytic activity.
2. Results and Discussion
2.1. Chemistry.
The synthesis of the amine salt 16, key intermediate to obtain the bifunctional hydroxamates 6 and 7, is
shown in Scheme 1. The tert-butyl ester 10 was obtained from the commercially available (S)-(+)-α-
hydroxy-1-3-diaxo-2-isoindolinebutyric acid as previously reported.21 The secondary alcohol group of
10 was converted to the corresponding mesylate as a leaving group to provide 11 in quantitative yield.
The azido derivative 12 was prepared by conversion of 11 through SN2 reaction (NaN3, DMF, 65 °C) in
high yield (98%). Compound 12 was reduced to the corresponding amine 13 by Pd-catalyzed
hydrogenation (92% yield) followed by sulfonylation with the commercial p-methoxy-biphenyl
sulfonyl chloride affording derivative 14 (49% yield). A further SN2 reaction of compound 14 with the
commercial 2-(Boc-amino)ethyl bromide gave ester 15 (64% yield). Finally, amine 16 was obtained as
trifluoroacetate salt by selective hydrolysis of Boc group in the presence of the tert-butyl group upon
treatment of compound 15 with trifluoroacetic acid (TFA) under controlled conditions (4 h at 0 °C
under a nitrogen atmosphere, 56% yield).
6

Scheme 1. Reagents and conditions: i) MsCl, Et3N, DCM, rt (quantitative); ii) NaN3, DMF, 65 °C,
98%; iii) H2/Pd MeOH, rt, 92%; iv) p-methoxy-biphenylsulfonyl chloride, NMM,THF/H2O, rt, 49%; v)
2-(Boc amino)ethylbromide, K2CO3, DMF, rt, 64%; vi) TFA, DCM, 0°C, 4h, 56%.
The final dimeric hydroxamates 6 and 7 were synthesized as reported in Scheme 2. The di-NHS(N-
hydroxysuccinimido) activated esters 1715 and 2116 were added to the corresponding solution of 16 in
DMF and in the presence of N,N-diisopropylethylamine (DIPEA), affording respectively the tert-butyl
esters 18 (30% yield) and 22 (25% yield). The carboxylic acids 19 and 23 were obtained by tert-butyl
group acidic cleavage with TFA. These carboxylates were converted into the corresponding
hydroxamates 6 and 7 by condensation with O-(tetrahydro-2H-pyran-2-yl)hydroxylamine (THPONH2)
to give the tetrahydropyranyl intermediates 20 and 24. Subsequently, the deprotection of the
hydroxamate moiety was conducted by acid hydrolysis with HCl 4N affording 6 and 7 (78% and 48%
yield).
7

Scheme 2: Reagents and conditions: i) 17 or 21, DIPEA, DMF, rt (compound 18: 30%, compound 22:
25%); ii) TFA, DCM, rt, (compound 19: 86%, compound 23 90%; iii) THPONH2, EDC, HOBt, Et3N,
DMF, rt; iv) HCl 4N, MeOH, dioxane, rt, (compound 6: 48% from 19, compound 7: 48% from 23).
Homodimers 8 and 9 have been synthesized following the procedure reported in Scheme 3. The
reported dichloride 2616 was added to a DMF solution of 2520 in the presence of DIPEA to give the di-
tert butyl esters (R,R)-27 and 28, which were separated by flash chromatography. The corresponding
carboxylic acids derivatives 29 and 30 were obtained by tert-butyl group cleavage with TFA. The
carboxylate 29 was converted into the corresponding hydroxamate 8 by condensation with O-
(tetrahydro-2H-pyran-2-yl)hydroxylamine (THPONH2) to give the tetrahydropyranyl intermediate 31
followed by acid hydrolysis with HCl 4N. The carboxylate 28 was converted into the corresponding
hydroxamate 9 by condensation with O-(tert-butyldimethylsilyl)hydroxylamine (TBDMSiONH2) to
give the silyl intermediates 32 and subsequent deprotection of the hydroxamate moiety with TFA.
8

Scheme 3: Reagents and conditions: i) 26, DIPEA, DMF, rt, (compound 27: 15%, compound 28:
23%); ii) TFA, DCM, rt (compound 29: 86%, compound 30: 46%); iii) THPONH2, EDC, HOBt, Et3N,
DMF, rt; iv) HCl 4N, MeOH, dioxane, rt (10% from 29); v) (CH3)3CSi(CH3)2ONH2, EDC, DMF
(15%); vi) TFA, DCM (46%).
2.2. Biological activity on isolated enzymes
The newly synthesized compounds 6-9 were tested in vitro on human recombinant MT1-MMP, MMP-
1, MMP-2, and MMP-9 by a fluorometric assay.22 MMP-2 and MMP-9 are validated targets in cancer
therapy so they are useful to investigate the selectivity profile. MMP-1 is considered an anti-target
because its inhibition seems to be responsible for side effects such as musculoskeletal syndrome,
clinically observed for broad spectrum MMPIs.23 Results are reported in Table 1 as IC50 values (nM)
and monomers 4 and 5 were used as reference compounds together with CGS 27023A,
. Compounds 6, 7 and 8, 9 reveled a nanomolar activity towards MT1-MMP
respectively comparable to 4 and 5. These data suggest that the dimerization in P1 and P2’ does not
substantially affect the affinity for MT1-MMP. The length of the linker does not seem to have a great
impact on MT1-MMP inhibitory activity as well (see 6 vs 7 and 8 vs 9). Unfortunately, these four
9

dimeric inhibitors have not shown an improved selectivity for the target enzyme over the other MMPs
tested with respect to the parent compounds.
Since compounds 8 and 9 were less active than 6 and 7 against MT1-MMP while compound 7 showed
a low water solubility, probably due to the high molecular weight, we decided to perform the following
cell-based studies on the hydroxamate 4 and its shorter dimeric derivative 6. In fact dimer 6 was
endowed with a nanomolar activity on MT1-MMP (IC50 =14 nM) and an appreciable water solubility
(> 1 mM).
Table 1. In vitroa inhibitory activity (IC50, nM) of compounds 4-9 on MMPs.
Compd
IC50 nM
MMP-1 MMP-2 MMP-9 MT1-MMP
4
170
150
45
0.67
13
0.43
14
3.9
14
7
6
7
4.1
0.70
1.2
1.6
25
2.1
8.6
1.25
8.7
4.8
5
196
870
155
56
34
80
41
23
8
9
CGS 27023A
aEnzymatic data are mean values for three independent
experiments performed in duplicate. SD were generally
within ±10%.
2.3. Biological activity on cells
2.3.1. Monomeric compound 4 significantly decreases cell growth compared to dimeric
compound 6
To examine the potential toxicity of compounds 4 and 6, inhibitor treated HT1080 fibrosarcoma cells
were subjected to MTS (3-(4,5 dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)2-(4-sulfophenyl)-
2H-tetrazolium) colorimetric assay. As shown in Figure 4, there was no significant difference
comparing the optical density of the samples treated with compound 6 at any concentration and DMSO,
on the other hand, compound 4 significantly decreased cell growth when applied at 50 µM, 20 µM or
10

10 µM compared to DMSO (P values were respectively < 0.001, < 0.01, < 0.05, and they are shown in
the graph). These data suggest a positive effect of dimerization to prevent toxicity, confirming our
previous data.16
Figure 4. Toxicity evaluation of 4 and 6. HT1080 cells were incubated overnight with compounds 4
or 6 and then MTS was added to the media. Optical density was measured every hour for 4 h using an
ELISA plate reader. Average optical densities were calculated and plotted with standard deviations
(SDs) using Prism (n=8). Data were compared using a one-way ANOVA followed by a Tukey’s
multiple correlation (***P < 0.001, **P < 0.01,*P < 0.05).
2.3.2. Effect of MMPIs 4 and 6 on MT1-MMP-dependent proMMP-2 activation
Gelatin zymography was used to test compounds 4 and 6 inhibitory activity on MT1-MMP-medited
proMMP-2 activation. ProMMP-2 activation is one of the major means MT1-MMP uses to promote
cell invasion into basement membrane.24 We used HT1080 cells for this assay, since these cells
endogenously express MT1-MMP and it has been shown that proMMP-2 activation in this cells is
MT1-MMP-dependent.8, 25 HT1080 cells were cultured in DMEM for 24 h and in serum-free DMEM
in the presence of GM6001, a well-known MMP inhibitor, compounds 4 or 6 at different
concentrations for another 24 h. For each treatment, cells were stimulated with type I collagen, which
has been shown to promote MT1-MMP-mediated proMMP-2 activation.26 As shown in Figure 5, when
HT1080 cells were cultured in DMEM or DMEM containing DMSO, proMMP-2 (P, 72 KDa) was
effectively processed to the active form (A, 66 KDa), GM6001 abolished proMMP-2 activation except
at the concentration of 0.01 M where we could detect a band corresponding to the active MMP-2.
11

Dimer 6 seemed to prevent proMMP-2 activation at the concentrations of 10 M and 1 M, while it
did not have any detectable effect for concentrations lower than 0.1 M, on the other hand, monomer 4
inhibited proMMP-2 processing only when the concentration was 10 M, since active MMP-2 could be
detected in the other cases. GM6001 treatment noticeably increased the level of proMT1-MMP (65
KDa) and active MT1-MMP (55 KDa), compared to the DMSO negative control, while processed form
of MT1-MMP that has lost the catalytic domain (40 KDa) was slightly decreased when GM6001 was
applied at the concentrations of 10 M and 1 M. Compound 6 also dose-dependently increased band
intensities of proMT1-MMP and notably increased active MT1-MMP at the concentrations of 10 M
or 1 M, compared to the DMSO negative control while it did not affect the amount of processed MT1-
MMP at any of the reported concentrations. Compound 4 did not alter the level of proMT1-MMP, the
active and the processed form at any of the tested concentrations.
Figure 5. Effect of 4 and 6 on proMMP-2 activation. HT1080 cells were cultured for 24 h with
collagen (Col-I, 50 µg/mL) and in the presence of GM6001, compounds 4 or 6 at different
concentrations. Conditioned media were collected and analyzed by zymography for proMMP-2
activation (zymo), cell lysates were subjected to WB using anti-MT1-MMP (Hpx) 222-1D8 antibody.
12

Arrow heads point to bands of interest. P indicates proMMP-2, while A the active enzyme form. The
experiments were repeated twice, and we confirmed that the data are reproducible.
2.3.3. Effect of MMPIs 4 and 6 on collagen degradation
To test compounds 4 and 6 inhibitory activity on MT1-MMP collagenolytic activity, a collagen film
degradation assay was carried out. Again we have used HT1080 cells since they express MT1-MMP
endogenously and its collagen degradation activity is MT1-MMP-dependent.25 HT1080 cells were
cultured for 48 h on a thin collagen type-I layer in the presence of compounds 4 or 6, GM6001 and
DMSO, as a negative control, after the incubation, remaining collagen was stained with Coomassie
Blue. As shown in Figure 6A, cells degraded the collagen layer when treated with DMSO, while
GM6001 (10 M) completely abolish MT1-MMP mediated collagen degradation. Compounds 6 and 4
seemed to completely abolish collagen degradation when they were used at the concentration of 10
M, while small degradation areas could be detected when the concentration was only 1 M. Collagen
degradation was quantified measuring the integrated density of the collagen layer though ImageJ
software, five pictures, corresponding to five different areas of the same sample, were quantified for
each treatment. Average integrated densities were calculated and plotted with error bars, referring to
the standard deviations (SDs). As shown in Figure 6B, collagen degradation was significantly
decreased for all the treatments, in comparison to the DMSO control (****P < 0.0001, shown in the
graph), significant difference could not be detected comparing the two tested concentrations of
compound 6, while there was statistical difference between applying 10 M or 1 M of compound 4 (P
< 0.001, not shown in the graph).
13

Figure 6. Effect of 4 and 6 on collagen degradation. A) HT1080 cells were subjected to collagen
film degradation assay in the presence of GM6001, compounds 4 or 6 at different concentrations.
Representative images are shown. Collagen is grey and degraded areas are white. Scale bars are 150
µm. B) Collagen degradation was quantified using ImageJ software on 5 representative areas of
collagen film in each treatment (n= 5). Mean values ± SD were plotted for each treatment. Each mean
integrated density was compared to the negative control (DMSO) using a one-way ANOVA test, ****P
< 0.0001. Experiments are repeated twice and the data are confirmed to be reproducible.
2.3.4. Effect of MMPIs 4 and 6 on MT1-MMP-dependent collagen invasion
To investigate activity of compounds 4 and 6 on MT1-MMP-dependent collagen invasion, a
microcarriers beads invasion assay was employed. We have previously shown that HT1080 cells
invade into collagen matrix in MT1-MMP-dependent manner.25 Once HT1080 cells were attached to
gelatin-coated microcarrier beads, the beads/cells were suspended in a type-I collagen matrix and they
were cultured overnight. Cells migrated out from the bead surface into the collagen matrix and
representative pictures are shown in Figure 7A. Migrated distance of 50 cells per treatment was
measured using ImageJ software and average migrated distances were plotted with red error bars
(showing  SDs) with Prism (Figure 7B). All the treatments diminished significantly MT1-MMP
dependent collagen invasion compared to DMSO (****P < 0.0001, shown in the graph). Collagen
invasion was not significantly different when compared the two tested concentrations of inhibitor 6
while cells migrated significantly more upon exposure to 1 M than to 10 M of compound 4 (P <
0.001, not shown in the graph). MT1-MMP-dependent invasion was statistically inhibited more when
14

cells were incubated with 10 M of compound 6 than 10 M of compound 4 (P < 0.01, not shown) and
the same was observed comparing the 1 M concentrations of the two compounds (P < 0.00001, not
shown). It has been shown that MT1-MMP is the only MMP that is capable of promoting cellular
invasion within 3D collagen matrix.27 Thus, this difference in inhibitory activity reflects effectiveness
of the dimeric compound 6 to inhibit MT1-MMP on the cell surface.
Figure 7. Effect of 4 and 6 on collagen invasion. A) HT1080 cells were subjected to a microcarrier
bead invasion assay in the presence of GM6001, compounds 4 or 6. A representative image of cells
migrating out from the bead surface is shown for each treatment. Scare bars are 150 µm. B)
Quantitative analysis of microcarrier bead invasion assay: migratory distance was measured for 50 cells
per treatment (n=50), using ImageJ software (see methods) and mean migratory distances were plotted
using Prism software. Each mean migratory distance was compared to the negative control (DMSO)
using a one-way ANOVA test, ****P < 0.0001. Experiments are repeated twice and the data are
confirmed to be reproducible.
Considering these data we can hypothesize a mechanism of action of dimeric MT1-inhibitor 6 on tumor
cell surface that explains the differences found between enzymatic assays and cell-based assays. When
compounds 4 and 6 were compared using catalytic domain of MT1-MMP (enzymatic assay), 6 showed
a slightly lower inhibitory activity (higher IC50) as shown in Table 1. Under these assay conditions, the
catalytic domain does not have any restriction and is soluble in the assay solution. On the other hand,
15

when the effects of these inhibitors on the cell surface were compared, compound 6 had a slightly
higher inhibitory activity (Figure 5, 6, and 7). ProMMP-2 activation, collagen degradation and collagen
invasion in HT1080 cells are all dependent on dimerization of MT1-MMP on the cell surface, and it is
possible that the distance of two catalytic domains within dimeric MT1-MMP is similar to the distance
of two zinc binding groups within dimeric compound 6, and this provides steric advantage to 6 with
respect to monomer 4. A molecular modeling study was performed to support this hypothesis.
2.4. Molecular modeling
In order to provide a possible binding disposition of compound 6, a hypothetical dimeric complex of
two MT1-MMP catalytic domains has been developed. The crystallographic structure of the MMP-9
dimer in complex with a bifunctional inhibitor (PDB id: 4H82)28 was used as a template for generating
a MT1-MMP dimeric structure. This structure was then used as a starting point for 50 ns of molecular
dynamic (MD) simulation. Compound 6 was thus docked into the resulting dimer and the best ranked
solution was taken into account. The corresponding MT1-MMP-6-MT1-MMP trimer thus obtained was
subjected to 100 ns of MD simulation. As shown in Figure 8, the geometry of the ligand is in an
agreement with the possible interaction within this dimeric structure, with both 4-methoxybiphenyl
moieties that are inserted in the S1’ cavity and the two hydroxamic acid groups able to chelate the zinc
ions that show a trigonal bipyramidal chelation geometry. Interestingly, each of the two phthalimidic
groups shows lipophilic interactions with both MT1-MMP chains, reducing their exposition to water.
16

Figure 8. Proposed trimeric adduct MT1-MMP-6-MT1-MMP (catalytic domains).
3. Conclusions
In the present study we reported the design and synthesis of the first bifunctional MT1-MMP inhibitors,
compounds 6-9, potentially able to impair MT1-MMP homodimerization on tumor cell surface. These
were symmetrical homodimers derived from previously described monomeric arylsulfonamide
hydroxamate-based MMP inhibitors, characterized by different linker lengths and branch points.
Among these compounds, a promising derivative, 6, was identified by fluorometric assay on human
recombinant MMPs and was selected for further studies on HT1080 fibrosarcoma cell line. Dimer 6
was a nanomolar inhibitor of MT1-MMP (IC50 = 14 nM) and also of MMP-2 and MMP-9, widely
known as targets in cancer therapy. Moreover 6 showed a lower cell cytotoxicity (by MTS assay up to
50 µM) with respect to its monomeric analogue, compound 4, indicating that dimerization can have a
positive effect to prevent compounds toxicity. To prove the collagenolytic activity of dimer 6 with
respect to 4, they were tested on HT1080 cells that express endogenous MT1-MMP. Both compounds
reduced MT1-MMP-dependent proMMP-2 activation, collagen degradation and collagen invasion in a
dose-dependent manner but dimer 6 showed a higher potency in all assays. In particular, 6 prevented
proMMP-2 activation already at the concentration of 1 M, completely abolished collagen degradation
at the concentration of 10 µM and diminished significantly MT1-MMP dependent collagen invasion at
17

10 µM concentration. Finally, a theoretical model of compound 6 binding to dimeric MT1-MMP
(catalytic domains) have been proposed.
Taken together, our preliminary results supported the use of dimeric inhibitors as alternative approach
to reduce cancer cell invasion by inhibiting dimeric MT1-MMP. Further studies will be necessary to
improve dimer 6 selectivity for MT1-MMP and to fully demonstrate the binding mode and mechanism
of action of dimeric MT1-inhibitors on tumor cell surface.
4. Materials and methods
4.1 Chemistry
1
13
Melting points were determined on a Kofler hotstage apparatus and are uncorrected. H and C
NMR spectra were recorded on a Bruker Avance III HD 400 MHz spectrometer. Chemical shifts (δ)
are reported in parts per million and coupling constants (J) are reported in hertz (Hz). 13C NMR spectra
were fully decoupled. The following abbreviations were used to explain molteplicities: singlet (s),
doublet (d), triplet (t), double doublet (dd), broad (br), and multiplet (m). Chromatographic separations
were performed on silica gel columns by flash column chromatography (Kieselgel 40, 0.040−0.063
mm, Merck) or using ISOLUTE Flash Si II cartridges (Biotage). Reactions were followed by thin-layer
chromatography (TLC) on Merck aluminum silica gel (60 F254) sheets that were visualized under a
UV lamp, and hydroxamic acids were visualized with FeCl3 aqueous solution. Evaporation was
performed in vacuo (rotating evaporator). Sodium sulfate was always used as the drying agent.
Commercially available chemicals were purchased from Sigma-Aldrich. Elemental analysis was used
to determine the purity of target compounds. Analytical results are within ± 0.40% of the theoretical
values. CGS 27023A was synthesized as previously described.29
4.1.1 (S)-tert-butyl 4-(1,3-dioxoisoindolin-2-yl)-2-((methylsulfonyl)oxy)butanoate (11)
To a solution of 1016, 21 (210 mg, 0.687 mmol) in DCM (5 mL) and Et3N (190 µL), MsCl was added
dropwise. The solution was stirred for 1.5 h at rt. The reaction mixture was dissolved in DCM and
washed (3 × 15 mL) with water. The combined organic phases were dried over Na2SO4 and evaporated
under reduced pressure. The mesylate 11 was obtained as a pure colorless oil (261 mg, quantitative
yield). 1H-NMR (400 MHz, CDCl3): δ 7.86 (m, 2H, Ar), 7.71 (m, 2H, Ar), 4.95 (dd, J1= 4.4 Hz, J2= 8
Hz, 1H), 3.90-3.79 (m, 2H), 3.21 (s, 3H), 2.38-2.08 (m, 2H), 1.49 (s, 9H). Elemental Analysis for
C17H21NO7S calculated: % C, 53.25; % H, 5.52; % N 3.65; found: % C, 53.30; % H, 5.67; % 3.71.
18

4.1.2 (R)-tert-butyl 2-azido-4-(1,3-dioxoisoindolin-2-yl)butanoate (12)
To a solution of the mesylate 11 (732 mg, 1.90 mmol) in 4.7 mL of DMF, sodium azide (744 mg, 11.45
mmol) was added and the mixture was stirred overnight at 65 °C. The solvent was removed and the
residue was dissolved in EtOAc and washed with water (3× 20 mL). The organic phase was dried over
Na2SO4 and evaporated under reduced pressure. The azido derivative 12 was isolated pure as white
1
solid (622 mg, 98% yield) without further purifications. H NMR (400 MHz, CDCl3): δ 7.86 (m, 2H,
Ar), 7.73 (m, 2H, Ar), 3.87-3.81 (m, 3H), 2.21-2.12 (m, 1H), 2.04-2.00 (m, 1H), 1.50 (s, 9H).
Elemental Analysis for C16H18N4O4 calculated: % C, 58.17; % H, 5.49; % N 16.96; found: % C, 58.22;
% H, 5.53; % N 16.99.
4.1.3 (R)-tert-butyl 2-amino-4-(1,3-dioxoisoindolin-2-yl)butanoate (13)
Compound 12 (600 mg, 1.81 mmol) was dissolved in MeOH (10 mL) and the solution was stirred
under hydrogen atmosphere in the presence of 10% Pd-C (575 mg) overnight. The resulting mixture
was filtered on celite and the filtrate was evaporated under reduced pressure to give the deprotected salt
1
13 (507 mg, 92% yield). H-NMR (400 MHz, DMSO-d6): δ 7.87-7.81 (m, 4H, Ar), 3.69-3.65 (t, J= 8
Hz, 1H), 3.26-3.24 (m, 2H), 1.89-1.98 (m, 1H), 1.77-1.65 (m, 1H), 1.38 (s, 9H). Elemental Analysis for
C16H20N2O4 calculated: % C, 63.14; % H, 6.62; % N 9.20; found: % C, 63.24; % H, 6.78; % N 9.31.
4.1.4 (R)-tert-butyl
4-(1,3-dioxoisoindolin-2-yl)-2-(4'-methoxy-[1,1'-biphenyl]-4-
ylsulfonamido)butanoate (14)
To a solution of amine 13 (500 mg, 1.64 mmol) in dry THF, N-methylmorpholine (0.36 mL) and 4-
methoxy-biphenylsulforylchloride (557 mg, 1.97 mmol) were added under nitrogen atmosphere. The
solution was stirred at rt for 48h and then the solvent was removed. The residue was dissolved in
EtOAc, washed (3 x 15 mL) with water, dried over Na2SO4 and evaporated. The crude was purified by
flash chromatography (in gradient from n-hexane/EtOAc 6:1 to only EtOAc) to give 14 (440 mg, 45%
1
yield) as white solid. H-NMR (400 MHz, CDCl3): δ 7.89-7.81(m, 4H, Ar), 7.72-7.68 (m, 2H, Ar),
7.65-7.62 (m, 2H, Ar), 7.52-7.49 (m, 2H, Ar), 7.00-6.98 ( m, 2H, Ar),5.48 (d, J=9.2 Hz, 1H), 3.97-3.90
(m, 1H), 3.88 (s, 3H), 3.79-3.74 (m, 1H), 2.29-2.15 (m, 2H), 1.17 (s, 9H). Elemental Analysis for
C29H30N2O7S calculated: % C, 63.26; % H, 5.49; % N 5.09; found: % C, 63.34; % H, 5.62; % N 5.22.
4.1.5 (R)-tert-butyl
2-(N-(2-((tert-butoxycarbonyl)amino)ethyl)-4'-methoxy-[1,1'-
biphenyl]-4-ylsulfonamido)-4-(1,3-dioxoisoindolin-2-yl)butanoate (15)
To a solution of sulfonamide 14 (250 mg, 0.45mmol) in dry DMF (6.85 mL), K2CO3 (627 mg, 4.54
mmol) and 2-(Boc amino) ethylbromide (104 mg, 0.68 mmol) were added under Argon atmosphere.
19

The solution was stirred at rt for 48h, then the solvent was removed. The residue was dissolved in
EtOAc and washed (2 × 15 mL) with water, dried over Na2SO4 and evaporated under reduced pressure.
The crude was purified by flash chromatography using Isolute Si II cartridge (n-hexane/EtOAc 5:1) to
1
give 15 (203 mg, 63% yield) as white solid. H-NMR (400 MHz, CDCl3): δ 7.91-7.81 (m, 4H, Ar),
7.74-7.69 (m, 2H, Ar), 7.64 (m, 2H, Ar), 7.52 (m, 2H, Ar), 6.99 (m, 2H, Ar), 5.47-5.43 (m, 1H), 4.52-
4.50 (m, 1H), 3.86-3.77 (m, 2H), 3.86 (s, 3H), 3.48-3.53 (m, 2H), 3.31-3.25 (m, 1H), 2.36-2.26 (m,
1H), 2.04-1.98 (m, 2H), 1.37 (s, 9H), 1.27 (s, 9H). Elemental Analysis for C36H43N3O9S calculated: %
C, 62.32; % H, 6.25; % N 6.06; found: % C, 62.47; % H, 6.38; % N 6.23.
4.1.6 (R)-tert-butyl-4'-methoxy-[1,1'-biphenyl]-4-ylsulfonamido)-4-(1,3-dioxoisoindolin-
2-yl)butanoate ethyl ammonium salt (16)
Trifluoroacetic acid (TFA) (0.91 mL, 11.87 mmol) was added dropwise, under argon atmosphere, to a
stirred solution of tert-butyl ester 15 (275 mg, 0.39 mmol) in dry DCM (4.37 mL) cooled to 0 °C. The
solution was stirred for 3 h at 0 °C and then the solvent was removed in vacuum. The residue was co-
evaporated (3 x 10 mL) with a mixture of toluene and DCM. The crude product was purified by flash
chromatography (CHCl3/MeOH 40:1) using Isolute Flash Si II cartridge 5g to yield 16 as a white solid
(117 mg, 56% yield). 1H-NMR (400 MHz, CDCl3): δ 8.14 (brs, 3H), 7.90-7.82 (m, 4H, Ar), 7.72-7.69
(m, 4H, Ar), 7.53 (m, 2H, Ar), 6.99 (m, 2H, Ar), 4.50 (dd, J1= 4.4 Hz, J2=10.4 Hz, 1H), 3.86 (s, 3H),
19
3.78-3.66 (m, 4H), 3.57-3.45 (m, 3H), 2.37-2.28 (m, 1H), 2.00-1.96 (m, 1H), 1.39 (s, 9H). F-NMR
(376.46 MHz, CDCl3): δ -75.81. Elemental Analysis for C33H36F3N3O9S calculated: % C, 56.00; % H,
5.13; % N 5.94; found: % C, 56.28; % H, 5.28; % N 5.19.
4.1.7 General procedure for the preparation of dimeric di-tert-butyl esters 18 and 22.
A solution of the suitable NHS ester (1 eq) and N,N diisopropylethylamine (60 µL, 6 eq) in dry DMF
(1.5 mL) was slowly added to a solution of the amine salt 16 (2.5 eq) in dry DMF (1.0 mL) cooled to 0
°C under Argon atmosphere. The resulting solution was stirred overnight at rt and then the solvent was
removed in vacuum.
4.1.7.1 Dimeric di-tert-butyl ester 18
The title compound was prepared from NHS ester 17 (32 mg, 0.09 mmol) and the amine salt 16 (129
mg, 0.18 mmol) following the general procedure. The crude product was purified by flash
chromatography (CHCl3) using Isolute Si II cartridge to afford the di-tert-butyl ester 18 as a colorless
oil (60 mg, 30% yield). 1H-NMR (400 MHz, CDCl3): δ 8.34 (brs, 1H), 7.95 (m, 2H, Ar), 7.84-7.80 (m,
20

8H, Ar), 7.72-7.68 (m, 4H, Ar), 7.58-7.56 (m, 4H, Ar), 7.47-7.39 (m, 5H, Ar), 6.96 (m, 4H, Ar), 4.55
(dd, J1= 5.6 Hz, J2= 9.2 Hz, 2H), 3.85 (s, 6H), 3.85-3.53 (m, 12H), 2.41-2.32 (m, 2H), 2.11-2.01 (m,
2H), 1.29 (s, 18H). 13C-NMR (100 MHz, CDCl3): δ 169.2 (2×COO), 168.2 (2×CON), 167.1
(2×CONH), 160.0 (Ar-C-OCH3), 145.3 (Ar-C-SO2), 137.0, 134.5, 132.0, 131.5, (Ar-C), 134.01, 130.3,
128.4, 128.0, 127.0, 123.4, 114.4 (Ar-CH), 83.0 (C(CH3)3), 58.9 (CHNSO2), 55.4 (OCH3 ), 45.3
(CH2NSO2 ), 40.5 (CH2NHCO), 35.5 (CH2NCO), 29.7 (CH2CHNSO2) 27.7 (C(CH3)3). Elemental
Analysis for C70H72N6O16S2 calculated: % C, 63.81; % H, 5.51; % N 6.38; found: % C, 63.96; % H,
5.67; % N 6.69.
4.1.7.2 Dimeric di-tert-butyl ester 22
The title compound was prepared from NHS ester 21 (31.48 mg, 0.0593 mmol) and the amine salt 16
(105 mg, 0.15 mmol) following the general procedure. The crude product was purified by flash
chromatography (CHCl3 /MeOH 20:1) to give the di-tert-butyl ester 22 as a colorless oil (54 mg, 25%
1
yield). H-NMR (400 MHz, CDCl3): δ 8.25 (brs, 1H), 7.97 (m, 2H, Ar), 7.80-7.43 (m,8H, Ar), 7.69-
7.66 (m, 4H, Ar), 7.62-7.57 (m, 6H, Ar), 7.49 (m, 4H, Ar), 7.42 (m, 1H, Ar), 7.20 (brs, 1H), 6.97 (m,
4H, Ar), 4.51 (dd, J1=5.6 Hz, J2=9.6 Hz, 2H), 3.84 (s, 6H), 3.84-3.75 (m, 4H), 3.63-3.48 (m, 10H),
13
3.35-3.31(m, 2H), 2.41 (t, J= 6.4 Hz, 4H), 2.33-2.27 (m, 2H), 2.00-1.97 (m, 6H), 1.26 (s, 18H). C-
NMR (100 MHz, CDCl3): δ 173.8 (2×CONH), 169.1 (2×COO), 168.4(2×CON), 166.8 (2×CONH),
160.2 (2× Ar-C-OCH3), 145.5 (2× Ar-C-SO2), 137.0, 134.7, 132.0, 131.6 (Ar-C), 134.2, 130.4, 128.9,
128.5, 128.1, 127.2, 123.5, 114.6 (Ar-CH), 83.2 (2×C(CH3)3), 59.0 (2×CHNSO2), 55.5 (2×OCH3), 51.3
(2×CH2NSO2), 45.6, 39.9, 35.6 (2×CH2NSO2, 4×CH2NHCO, 2× CH2CO), 33.9 (CH2CH2NHCO), 27.9
(C(CH3)3) , 24.6 (2×CH2CHNHCO). Elemental Analysis for C78H86N8O18S2 calculated: % C, 62.97; %
H, 5.83; % N 7.53; found: % C, 63.03; % H, 5.98; % N 4.70.
4.1.8 Dimeric tert-butyl esters 27 and 28
To a solution of acetate salt 2520 (331 mg, 0.73 mmol) and dichloride 2616 (178 mg, 0.48 mmol) in dry
DMF (5 mL), N,N-diisopropylethylamine (DIPEA) (0.29 mL, 1.69 mmol) was added dropwise. The
resulting mixture was stirred at rt overnight. The solvent was removed and the crude was dissolved in
EtOAc, washed with H2O (1 × 25 mL) and brine (1 × 25 mL). The organic phase was dried over
Na2SO4, filtered and evaporated in vacuo. Compounds 27 (83 mg, 15% yield) and 28 (91 mg, 23%
yield) were isolated after flash chromatography (EtOAc/n-hexane 1.5/1) as white solids. Compound 27:
1H NMR (400 MHz, CDCl3): δ 8.36 (brs, 1H), 8.13 (dd, J1= 7.6 Hz, J2= 1.6 Hz, 2H), 7.96-7.94 (m, 4H,
Ar), 7.72-7.69 (m, 4H, Ar), 7.60 (t, J= 8 Hz, 1H, Ar), 7.56-7.53 (m, 4H), 7.49-7.37 (m, 8H), 5.64 (d, J=
21

9.2 Hz, 2H), 3.95-3.87 (m, 4H), 3.65-3.57 (m, 2H), 2.28-2.18 (m, 2H), 1.80-1.71 (m, 2H), 1.16 (s,
18H). Elemental Analysis for C48H54N4O10S2 calculated: % C, 63.28; % H, 5.97; % N 6.15; found: %
C, 63.48; % H, 6.28; % N 6.19.
Compound 28: 1H NMR (400 MHz, CDCl3): δ 8.25 (brs, 1H), 8.12 (m, 4H, Ar), 7.92 (m, 4H, Ar), 7.71
(m, 4H, Ar), 7.58-7.40 (m, 11H), 5.36 (d, J= 9.1 Hz, 2H), 3.97 (t, J= 7 Hz, 4H), 3.84-3.78 (m, 2H), 3.5-
3.44 (m, 2H), 2.59 (t, J= 7.8 Hz, 4H), 2.33-2.24 (m, 2H), 2.24-2.10 (m, 4H), 1.15 (s, 18H). Elemental
Analysis for C56H68N6O12S2 calculated: % C, 62.20; % H, 6.34; % N 7.77; found: % C, 62.42; % H,
6.48; % N 7.99.
4.1.9 General procedure for the preparation of dimeric di-carboxylic acids 19, 23, 29 and
30
Trifluoroacetic acid (TFA) (114 eq) was added dropwise to a solution of the appropriate tert-butyl ester
(18 or 22 or 27 or 28) (1 eq) in dry DCM (1.5 mL), cooled to 0 °C. The reaction was stirred at rt
overnight. The mixture was co-evaporated (3 x 10 mL) with toluene/DCM and the crude was purified
by trituration with Et2O/n-hexane.
4.1.9.1 Dimeric di-carboxylic acid 19
The title compound was prepared from tert-butyl ester 18 (56 mg, 0.042 mmol) following the general
procedure. The dicarboxylic acid 19 is a white solid (44 mg, 86% yield). 1H NMR (400 MHz, CDCl3):
δ 7.80-7.77 (m, 10H, Ar), 7.70-7.68 (m, 4H, Ar), 7.52-7.43 (m, 6H, Ar), 7.39-7.33 (m, 4H, Ar), 7.19-
7.11 (m, 2H), 6.93-6.90 (m, 4H, Ar), 4.65-4.75 (m, 2H), 3.83 (s, 6H), 3.83-3.44 (m, 12H), 2.51-2.42
(m, 2H), 2.10-1.95 (m, 2H). Elemental Analysis for C62H56N6O16S2 calculated: % C, 61.78; % H, 4.68;
% N 6.97; found: % C, 61.96; % H, 4.78; % N 6.99.
4.1.9.2 Dimeric di-carboxylic acid 23
The title compound was prepared from tert-butyl ester 22 (55 mg, 0.04 mmol) following the general
procedure. The dicarboxylic acid 23 is a white solid (46 mg, 90% yield). 1H NMR (400 MHz, CDCl3):
δ 8.12-8.10 (m, 1H), 7.95-7.36 (m, 23H, Ar), 6.99-6.95 (m, 4H, Ar), 4.55-4.52 (m, 2H), 3.84 (s, 6H),
3.66-3.73 (m, 16H), 2.36-2.22 (m, 6H), 2.00-1.79 (m, 6H). Elemental Analysis for C70H70N8O18S2
calculated: % C, 61.12; % H, 5.13; % N 8.15; found: % C, 61.48; % H, 5.28; % N 8.19.
4.1.9.3 Dimeric di-carboxylic acid 29
The title compound was prepared from tert-butyl ester 27 (83 mg, 0.09 mmol) following the general
procedure. The dicarboxylic acid 29 is a white solid (63 mg, 86% yield). 1H NMR (400 MHz, CD3OD):
22

δ 8.24-8.23 (br t, 1H), 7.94-7.91 (m, 6H, Ar), 7.75 (m, 4H, Ar), 7.63 (m, 4H, Ar), 7.53-7.37 (m, 7H),
4.03 (dd, J1=4.8 Hz, J2= 8.4 Hz, 2H), 3.61-3.55 (m, 2H), 3.51-3.44 (m, 2H), 2.19-2.11 (m, 2H), 1.99-
1.90 (m, 2H). Elemental Analysis for C40H38N4O10S2 calculated: % C, 60.14; % H, 4.79; % N 7.01;
found: % C, 60.34; % H, 4.89; % N 7.31.
4.1.9.4 Dimeric di-carboxylic acid 30
The title compound was prepared from tert-butyl ester 28 (16 mg, 0.01 mmol) following the general
procedure. The crude product was purified using Isolute cartridge C18 (H2O-MeOH 2:1) obtaining the
1
dicarboxylic acid 30 is a white solid (46% yield). M.p.: 119-121 °C; H NMR (400 MHz, CD3OD): δ
8.04-8.03 (m, 2H), 7.94-7.88 (m, 6H), 7.74-7.71 (m, 6H), 7.64-7.62 (m, 4H), 7.48-7.44 (m, 6H), 7.41-
7.39 (m, 2H), 3.94 (t, J=7.2 Hz, 4H), 3.83-3.80 (m, 2H), 3.48-3.45 (m, 2H), 2.60 (t, J=8 Hz, 4H), 2.18-
2.11 (m, 4H), 2.01-1.96 (m, 2H), 1.85-1.80 (m, 2H). Elemental Analysis for C48H52N6O12S2 calculated:
% C, 59.49; % H, 5.41; % N 8.67; found: % C, 59.66; % H, 5.62; % N 8.75.
General procedure for the preparation of dimeric di-hydroxamic acids 6, 7 and 8
To a stirred solution of the appropriate dicarboxylic acid 19 or 23 or 29 (1 eq) in dry DMF (1 mL) 1-
hydroxybenzotriazole (HOBt) (2.4 eq), O-(tetrahydro-2h-pyran-2-yl)hydroxylamine (THPONH2) (6.2
eq), N-methylmorpholine (6 eq) and 1-[3-(Dimethylamino)propyl]-3-ethyl carbodiimide hydrochloride
(EDC) (2.8 eq) were added under nitrogen atmosphere. After stirring at rt overnight, the mixture was
dissolved in EtOAc and washed with water, NaHCO3 saturated solution and brine. Then the organic
phase was dried over Na2SO4 and evaporated in vacuum. The crude O-tetrahydropyranyl derivatives
20, 24 and 31 were used in the next reaction without further purification. Tetrahydropyranyl precursor
(1 eq) was then dissolved in a mixture of dioxane (0.9 mL) and methanol (0.6 mL) and a solution of
HCl 4N (1.8 mL/mmol) was added dropwise at 0 °C. After 2 h of stirring, the solvents were evaporated
and the crude products were purified by flash chromatography using a Isolute Flash Si II cartridge to
give the desired hydroxamic acids 6, 7 and 8.
4.1.9.5 Dimeric di-hydroxamic acid 6
The title compound 6 was prepared starting from the carboxylic acid 19 (44 mg, 0.04 mmol) obtaining
the tetrahydropyranyl derivative 20 (42 mg, 0.03 mmol) following the general procedure. The crude
product was purified by flash chromatography (in gradient from CHCl3 to CHCl3 /MeOH 20:1) to give
1
the hydroxamic acid 6 as white solid (20 mg, 48% yield). M.p.: 158-160 °C; H NMR (400 MHz,
23

DMSO-d6): δ 10.84 (s, 2H), 9.03 (s, 2H) 8.68 (brs, 2H), 8.28 (s, 1H), 7.87-7.78 (m, 18H, Ar), 7.70-7.68
(m, 4H, Ar), 7.39-7.35 (m, 1H, Ar), 7.10-7.04 (m, 4H, Ar), 4.34 (t, J = 7.6 Hz, 2H), 3.94-3.78 (m, 2H),
3.81 (s, 3H), 3.46 (m, 10H), 2.17-2.08 (m, 2H), 1.94-1.92 (m, 2H). 13C NMR (100 MHz, DMSO-d6): δ
168.0 (2×CON, 2×CONHOH), 166.1 (2×CONH), 160.2(2×Ar-C-OCH3), 144.4 (2×Ar-C-SO2), 137.0, ,
132.1, 131.0, 130.0,(Ar-C), 134.8, 128.8, 128.1, 127.2, 123.5, 115.0, (Ar-CH) 55.7 (OCH3), 54.5
(2×CHNSO2), 43.5, 35.09,( 2×CH2NCO, 2×CH2NHCO, 2×CH2NSO2), 29.8(CH2CHNSO2). Elemental
Analysis for C62H58N8O16S2 calculated: % C, 60.28; % H, 4.73; % N 9.07; found: % C, 60.30; % H,
4.78; % N 9.10.
4.1.9.6 Dimeric di-hydroxamic acid 7
The title compound 7 was prepared starting from the di-carboxylic acid 23 (46 mg, 0.03 mmol)
obtaining the tetrahydropyranyl derivative 24 (42 mg, 0.03 mmol) following the general procedure. The
crude product was purified by flash chromatography (in gradient from CHCl3 to CHCl3 /MeOH 50:1) to
1
give the hydroxamic acid 7 as a white solid (17 mg, 48% yield). M.p.: 128-131 °C; H NMR (400
MHz, DMSO-d6): δ 10.82 (s, 2H), 9.02 (s, 2H) 8.57 (t, J = 5.2 Hz, 2H), 8.28 (s, 1H), 7.98-7.92 (m, 4H,
Ar), 7.86-7.77 (m, 15H, Ar), 7.69 (m, 4H, Ar), 7.53-7.51 (m, 1H, Ar), 7.05 (m, 4H), 4.29 (t, J = 7.6 Hz,
2H), 3.81 (s, 6H), 3.68-3.20 (m, 16H), 2.11-2.08 (m, 6H), 1.86-1.83 (m, 2H), 1.70 (t, J = 7.2 Hz, 4H).
13C NMR (100 MHz, DMSO-d6): δ 171.9 (2×CH2CONH), 169.6 (2×CONHOH), 167.6 (2×CON),
165.8 (2×CONH) 159.8 (Ar-C-OCH3), 143.9 (Ar-C-SO2), 136.6 , 130.5, 129.5, 134.8, (Ar-C), 131.7,
128.3, 127.6, 126.7, 123.0, 114.5, (Ar-CH), 55.2 (2×OCH3), 54.5 (2×CHNSO2), 43.4, (2×CH2NSO2),
34.6- 32.8 (4×CH2NHCO, 2×CH2NCO, 4×CH2CO), 29.6 (2×CH2CH2NHCO), 25.1(2×CH2CHNHCO).
Elemental Analysis for C70H72N10O18S2 calculated: % C, 59.82; % H, 5.16; % N 9.97; found: % C,
59.86; % H, 5.20; % N 9.94.
4.1.9.7 Dimeric di-hydroxamic acid 8
The title compound 8 was prepared starting from the carboxylic acid 29 (63 mg, 0.08 mmol) obtaining
the tetrahydropyranyl derivative 31 (42 mg 0.03 mmol) following the general procedure. The crude
product was purified by flash chromatography using Isolute cartridge C18 (H2O-MeOH 2:1) to give the
hydroxamic acid 8 as white solid (6 mg, 10 % yield). M.p.: 152-153 °C; 1H NMR (400 MHz, DMSO-
d6):δ 10.65 (s, 2H), 8.94 (s, 2H), 8.44 (br t, 2H), 8.23 (br s, 3H, Ar), 7.89-7.77 (m, 10H, Ar), 7.68-7.66
(m, 4H, Ar), 7.50-7.41 (m, 7H, Ar), 3.73-3.72 (m, 2H), 3.17-3.16 (m, 4H), 1.83-1.80 (m, 2H), 1.68-
1.64 (m, 2H). 13C NMR (100 MHz, DMSO-d6): δ 166.8, 165.7 (2×CONH 2×CONHOH), 143.6 (Ar-C-
SO2), 140.0, 138.4, 134.4, 129.6, 129.0, 127.1,( Ar-CH), 126.9, 126.9, 128.3, 128.1, 126.1,( Ar-C) 52.1
24

(2×CHNSO2), 48.5, 45.6, 36.1, 32.6 (2×CH2CHNHSO2, 2×CH2NHCO). Elemental Analysis for
C40H40N6O10S2 calculated: % C, 57.96; % H, 4.86; % N, 10.14; found: % C, 57.99; % H, 4.90; % N,
10.10.
4.1.10 Dimeric hydroxamic acid 9
1-[3-(Dimethylamino)propyl]-3-ethyl carbodiimide hydrochloride (EDC) was added portionwise
(49.85 mg, 0.26 mmol) to a stirred and cooled solution (0 °C) of the carboxylic acid 30 (84 mg, 0.09
mmol) and O-(tert-butyldimethylsilyl)-hydroxylamine (38 mg, 0.26 mmol) in dry DCM (3 mL). After
stirring at rt overnight, the mixture was washed with water, and the organic phase was dried and
evaporated in vacuo. The crude was purified by flash chromatography trough Isolute cartridge SI II
1
(CHCl3) to afford the silyl derivative 32 as a white solid (16 mg, 15% yield). H NMR (400 MHz,
CDCl3): δ 10.34 (s, 2H), 8.14-7.36 (m, 22H), 6.13 (d, J = 8 Hz, 2H), 4.00-3.84 (m, 6H), 3.48-3.22 (m,
2H), 2.61-2.53 (m, 4H), 2.16-1.99 (m, 8H), 0.81 (s, 18H), 0.07 (s, 12H).
Silyl precursor 32 (16 mg, 0.01 mmol) was then dissolved in dry CH2Cl2 (1 mL), and trifluoroacetic
acid (TFA) (0.1 mL, 1.49 mmol, 114 eq) was added dropwise at 0 °C. After 5 h of stirring at rt, the
solvent was evaporated. The crude product was purified by flash chromatography through Isolute
cartridge C18 (2:1 H2O-MeOH) affording hydroxamic acid 9 (6 mg, 46% yield) as a white solid. M.p.:
1
119-120 °C; H NMR (400 MHz, CD3OD): δ 8.03 (br t, 2H), 7.94-7.88 (m, 6H), 7.74-7.71 (m, 6H),
7.64-7.62 (m, 4H), 7.48-7.44 (m, 6H), 7.41-7.39 (m, 2H), 3.94 (t, J = 7.2 Hz, 4H), 3.81 (dd, J1 = 6 Hz,
J2 = 8 Hz 2H), 3.48-3.45 (m, 2H), 2.60 (t, J = 8 Hz, 4H), 2.18-2.11 (m, 4H), 2.01-1.96 (m, 2H), 1.85-
1.80 (m, 2H). 13C NMR (100 MHz, CD3OD): δ 175.88, 170.03, 167.5 (4×CONH, 2×CONHOH) 145.4
(2×Ar-C-SO2), 135.2, 133.5, 131.3, 129.8 (Ar-C) 139.2, 138.9, 128.6, 128.0, 127.7, 127.3, 127.2, 126.9
(Ar-CH), 52.3(2×CHNSO2), 47.6, 46.2 (4×CH2NHCO), 36.1(2×NHCO CH2), 32.5
(2×CH2CH2NHCO), 4×CH2NHCO,), 17.12 (2×CH2CHNHSO2). Elemental Analysis for
C48H54N8O12S2 calculated: % C, 57.70; % H, 5.45; % N 11.22; found: % C, 57.78; % H, 5.50; % N
11.18.
4.2 MMP Inhibition Assays.
Recombinant human proMMP-2, proMMP-9 and proMMP-1 were purchased from Calbiochem (Merck
Millipore). MMP-14 catalytic domain was a kind gift of Prof. Gillian Murphy (Department of
Oncology, University of Cambridge, UK). Proenzymes were activated immediately prior to use with p-
aminophenylmercuric acetate (APMA 2 mM for 1h at 37 °C for MMP-2, 1 mM for 1 h at 37 °C for
MMP-9, 2 mM for 2 h at 37 °C for MMP-1). For assay measurements, the inhibitor stock solutions
25

(DMSO, 10 mM) were further diluted in the fluorimetric assay buffer (FAB: Tris 50 mM, pH = 7.5,
NaCl 150 mM, CaCl2 10 mM, Brij 35 0.05%, and DMSO 1%). Activated enzyme (final concentration
0.5 nM for MMP-2, 1.3 nM for MMP-9, 1.0 nM for MMP-14 cd, 2.0 nM for MMP-1) and inhibitor
solutions were incubated in the assay buffer for 3 h at 25 °C. After the addition of 200 μM solution of
the fluorogenic substrate Mca-Lys-Pro-Leu-Gly-Leu-Dap(Dnp)-Ala-Arg-NH2 (Calbiochem) in DMSO
(final concentration 2 μM), the hydrolysis was monitored every 10 s for 15 min recording the increase
in fluorescence (λex = 325 nm, λem = 400 nm) using a Molecular Devices SpectraMax Gemini XPS
plate reader. The assays were performed in a total volume of 200 μL per well in 96-well microtiter
plates (Corning, black, NBS). Control wells lack inhibitor. The MMP inhibition activity was expressed
in relative fluorescent units (RFU). Percent of inhibition was calculated from control reactions without
the inhibitor. IC50 was determined using the formula: vi/v0 = 1/(1 + [I]/IC50), where vi is the initial
velocity of substrate cleavage in the presence of the inhibitor at concentration [I] and v0 is the initial
velocity in the absence of the inhibitor. Results were analyzed using SoftMax Pro software and GraFit
software.
4.3 Materials and methods of the biological experiments
4.3.1 Cell Culture
HT1080 cells were cultured in DMEM (Lonza) supplemented with 10% of FBS (GIBCO) and
penicillin/streptomycin (PAA) at 37 °C, 5% CO2.
4.3.2 Antibodies and Reagents
The following primary antibody were used: mouse anti-MT1-MMP monoclonal antibody (clone 222-
1D8, WB: 1:1000), mouse anti-actin monoclonal antibody (clone C4, sc47778, Santa Cruz, WB:
1:1000). The following secondary antibody was used: IRDye 800 CW conjugated goat anti-mouse
IgG (H + L, P/N 926-32210, Li-Cor biosciences, WB: 1:10000).
GM6001 was from Elastin Products Company, PureCol bovine collagen type I was from Inamed
Biomaterials, Cellmatrix type I-A collagen was from Nitta, Cytodex 3 microcarrier beads (Cytodex 3:
133-215 m) were from Sigma. MTS was from abcam.
4.3.3 MTS colorimetric assay
HT1080 cells were seeded in the amount of 8 x 103 cells/well in a 96-well plate and they were cultured
overnight in the presence of compounds 4 and 6. MTS (3-(4,5 dimethylthiazol-2-yl)-5-(3-
carboxymethoxyphenyl)2-(4-sulfophenyl)-2H-tetrazolium) in PMS (phenazine methosulphate) was
added to each well and cells were incubated for 4 h, the absorbance at 490 nm was measured every h
using an ELISA plate reader (FLUOstar Omega, BMG Labtech).
26

4.3.4 Western Blotting
Cell lysates were prepared using SDS-PAGE loading buffer, containing 2-mercaptoethanol and
subjected to SDS-PAGE, they were then transferred to PVDF membranes using the Trans-Blot Turbo
Transfer system (Bio-Rad). After blocking with 10% dry milk in TBS (20 mM Tris-HCl, pH 7.5, 150
mM NaCl), membranes were probed with mouse anti-MT1-MMP monoclonal antibody or mouse anti-
actin monoclonal antibody, protein bands were visualized using IRDye-near-infrared secondary
antibodies.
4.3.5 Gelatin Zymography
HT1080 cells were seeded in the amount of 5 X 104 cells/well in a 12 well plate, after 24 h, media was
replaced with serum-free DMEM containing collagen (PureCol, final concentration 50 g/mL) and
GM6001 (10 M) or compounds 4 or 6 (1 M, 10 M). DMSO-containing media was used as negative
control. After 24 h, samples were recovered, mixed with SDS-PAGE loading buffer without 2-
mercaptoethanol and subjected to electrophoresis using an SDS-polyacrylamide gel containing gelatin
(0.8 mg/mL). Enzyme activity was visualized as negative staining with Coomassie Blue.
4.3.6 Collagen Film Degradation Assay
Collagen film degradation assay was performed as previously described.30 In brief, a thin layer of ice
cold neutralized collagen (1:1 mixture of PureCol and Cellmatrix type I-A collagen, final concentration
2mg/mL) was used to coat 12-well culture plates, HT1080 cells were seeded on the film in the amount
of 1 X 105 cells/ well and cultured in the presence of GM6001 (10 M) or compounds 4 or 6 (1 M, 10
M). DMSO-containing media was used as negative control. After 48 h, cells were trypsinized and
remaining collagen was fixed with 4 % formaldehyde. Collagen was visualized by staining with
Coomassie Blue.
Image acquisition and analysis
Representative images were acquired using the 10x dry lens (NA = 0.3) of a Nikon TE2000-E
microscope equipped with an ORCA-ER CCD camera (Hammamatsu Photonics) operated by Volocity
Acquisition module software (Improvision, PerkinElmer). To quantify collagen film degradation,
integrated density of the collagen layer was measured using ImageJ software (National Institutes of
Health, Bethesda, MD, USA). Mean integrated densities were calculated and plotted with error bars
representing the standard deviation (SD), using Prism 6 (GraphPad Software, La Jolla, CA, USA). Data
were compared using one-way ANOVA statistical analysis.
4.3.7 Microcarrier Beads Invasion Assay
27

Microcarrier beads invasion assay was carried out as previously described.31 HT1080 cells were
attached to gelatin-coated Cytodex 3 microcarrier beads by preparing a cells/beads suspension which
was incubated on a shaker for 6 h. Beads coated with cells were suspended in neutralized collagen
(Cellmatrix type I-A, final concentration 2 mL/mL) and incubated overnight in the presence of either
GM6001 (10 M) or inhibitor 4 (1 M, 10 M) or inhibitor 6 (1 M, 10 M). DMSO-containing
media was used as negative control.
Image acquisition and analysis
Images of the beads/cells were acquired using the 10x dry lens (NA = 0.3) of a Nikon TE2000-E
microscope equipped with an ORCA-ER CCD camera (Hammamatsu Photonics) operated by Volocity
Acquisition module software (Improvision, PerkinElmer). To quantify collagen invasion, the distance
between a cell nucleus and the surface of the bead was measured using the line tool of ImageJ software
(National Institutes of Health, Bethesda, MD, USA). Only well-focused cells were included in the
calculation and mean migrated distances were calculated taking into consideration 50 cells per
treatment. Mean migrated distances were calculated and plotted with SD, using Prism 6 (GraphPad
Software, La Jolla, CA, USA). Data were compared using one-way ANOVA statistical analysis.
4.4. Molecular modeling studies
All simulations were performed using AMBER, version 14.32 A high resolution structure of a MMP-9-
inhibitor-MMP-9 trimer has been already reported (PDB id: 4H82). On this basis, two MT1-MMP
catalytic domains have been superimposed to the MMP-9-inhibitor-MMP-9 trimer in order to create a
MT1-MMP dimer. The system was then solvated with a 20 Å water cap. Sodium ions were added as
counterions to neutralize the system. Prior to MD simulations, two steps of energy minimization were
carried out. In the first stage, we kept the dimer fixed with a position restraint of 100 kcal mol-1 Å-2 and
we just energy minimized the positions of the water molecules. In the second stage, we energy
minimized the whole system without applying constraints. The two energy minimization stages
consisted of 10000 steps. The first 1000 steps were Steepest Descent, and the last 9000 were Conjugate
Gradient (CG). The MD trajectory was run using the energy minimized structure as the input, and
particle mesh Ewald electrostatics and periodic boundary conditions were used in the simulation. The
time step of the simulations was 2.0 fs with a cutoff of 10 Å for the non-bonded interaction. SHAKE
was employed to keep all bonds involving hydrogen atoms rigid. A constant-volume was carried out
for 500 ps, during which time the temperature was raised from 0 to 300 K (using the Langevin
dynamics method). Then, 50 ns of constant-pressure MD were carried out at 300 K. The final structure
was obtained as the average of the last 25 ns of MD energy minimized with the CG method until a
28

convergence of 0.05 kcal mol-1 Å-2 was reached. Automated docking of compound 6 was carried out by
means of the Gold program (version 5.1).33 The region of interest used by Gold was defined in order to
contain the residues within 20 Å from the catalytic zinc ions. The “allow early termination” option was
deactivated while the remaining Gold default parameters were used, and the ligand was submitted to
200 genetic algorithm runs by applying the ChemScore fitness function. The best docked conformation
was taken into account. The General Amber Force Field (GAFF) parameters were assigned to the
ligand. The partial charges were calculated using the AM1-BCC method, as implemented in the
Antechamber suite of AMBER 14. The corresponding MT1-MMP-6-MT1-MMP trimer thus obtained
was subjected to 100 ns with the same procedure described above. The final structure was obtained as
the average of the last 50 ns of MD energy minimized with the CG method until a convergence of 0.05
kcal mol-1 Å-2 was reached.
Acknowledgements
This work has been supported by funding from University of Pisa (PRA_2017_51 and PRA_2018_20)
and from Kennedy Trust Prize Studentships for V. Gifford.
Supplementary Data
Supplementary data associated with this article can be found, in the online version, at
References
1 Nagase H, Visse R, Murphy G. Structure and function of matrix Metalloproteinases and TIMPs.
Cardiovascular Research, 2006;69:562-573.
2
Itoh Y. Membrane-type matrix metalloproteinases: Their functions and regulations. Matrix Biol.
2015;44-46:207-223.
3
Barbolina MV, Stack SS. Membrane type 1-matrix metalloproteinase: Substrate diversity in
pericellular proteolysis. Semin Cell Dev Biol. 2008;19:24-33.
29