Synthesis and in vitro evaluation of targeted tetracycline derivatives: Effects on inhibition of matrix metalloproteinases

Article abstract of DOI:10.1016/j.bmc.2007.01.026

Among other non-antibiotic properties, tetracyclines inhibit matrix metalloproteinases and are currently under study for the treatment of osteoarthritis. Quaternary ammonium conjugates of tetracyclines were synthesized by direct alkylation of the amine function at the 4-position with methyl iodide. When tested in vitro, they inhibited cytokine-induced MMP expression to a lesser extent than parent tetracyclines. This was compensated by an improved inhibition of MMP catalytic activity. Since inhibition of collagen degradation was maintained these derivatives could be potent drug candidates for cartilage-targeted chondroprotective treatment.

Full text of DOI:10.1016/j.bmc.2007.01.026

Bioorganic & Medicinal Chemistry 15 (2007) 2368–2374
Synthesis and in vitro evaluation of targeted tetracycline derivatives:
Effects on inhibition of matrix metalloproteinases
a,b,c,
d
d
d
*
´
Aurelien Vidal,
Massimo Sabatini, Gae¨lle Rolland-Valognes, Pierre Renard,
Jean-Claude Madelmonta,b,c and Emmanuelle Mounetoua,b,c
aINSERM 484, Clermont-Ferrand F-63500, France
bUniv d’Auvergne, Clermont-Ferrand F-63500, France
cCentre Jean Perrin, Clermont-Ferrand F-63500, France
dInstitut de Recherches Servier, Suresnes F-92150, France
Received 13 July 2006; revised 22 December 2006; accepted 17 January 2007
Available online 19 January 2007
Abstract—Among other non-antibiotic properties, tetracyclines inhibit matrix metalloproteinases and are currently under study for
the treatment of osteoarthritis. Quaternary ammonium conjugates of tetracyclines were synthesized by direct alkylation of the amine
function at the 4-position with methyl iodide. When tested in vitro, they inhibited cytokine-induced MMP expression to a lesser
extent than parent tetracyclines. This was compensated by an improved inhibition of MMP catalytic activity. Since inhibition of
collagen degradation was maintained these derivatives could be potent drug candidates for cartilage-targeted chondroprotective
treatment.
ꢀ 2007 Elsevier Ltd. All rights reserved.
1. Introduction
The anti-MMP activity of tetracyclines slows down the
degradation of articular cartilage, which is generally
increased in arthritic pathologies. A first clinical trial
of doxycycline (DC) on patients with osteoarthritis
showed positive effects on both the symptoms and clin-
ical signs of cartilage loss.11,12 However, long-term
administration of tetracyclines, which is essential in
treating slow-developing degenerative pathologies such
as osteoarthritis, can lead to severe adverse effects
(kidney and liver damage), partly due to the antibiotic
properties of those molecules.13,14
Matrix metalloproteinases (MMPs) play a major role in
cartilage metabolism as degradative enzymes of the
extracellular matrix, which is mainly formed by proteo-
glycans and collagen. Both synthesis as pro-enzymes
and activation to the mature form are upregulated by
pro-inflammatory cytokines, including interleukin-1b
(IL-1b).1 Activated MMPs can be inhibited by endoge-
nous tissue inhibitors of metalloproteinases (TIMPs),
which non-covalently bind MMPs with high affinity.2,3
Due to the central role of MMPs in a wide number of
pathological processes, MMP inhibition is regarded as
a promising therapeutic approach, particularly for the
treatment of osteoarthritis.4 Aside from their well-
known anti-microbial activity, tetracyclines inhibit
MMPs through actions at three levels: mRNA expres-
sion, pro-enzyme activation, and catalytic activity of
the mature enzyme.5–10
Our strategy was based on drug targeting toward carti-
lage via the introduction of a quaternary ammonium
function (QA), which displays a high affinity for the
anionic sites (carboxylate and ester sulfate groups) of
proteoglycans.15 This concept has already led to the
development of cartilage-selective radiopharmaceuti-
cals16,17 and drugs.18–22 In the present study, it was
applied to tetracycline (TC) and DC by quaternization
of the C-4 dimethylamino group. This approach was
supported by structure–activity studies demonstrating
that the antibiotic properties of tetracyclines are depen-
dent on this C-4 dimethylamino group, whereas the inhi-
bition of MMP catalytic activity is more due to the
oxygens of the lower part of the molecule.23 Our
Keywords: Tetracyclines; Quaternary ammonium; Osteoarthritis;
Metalloproteinases.
*
Corresponding author. Tel.: +33 4 73 15 08 18; fax: +33 4 73 15 08
01; e-mail: vidal@inserm484.u-clermont1.fr
0968-0896/$ - see front matter ꢀ 2007 Elsevier Ltd. All rights reserved.
doi:10.1016/j.bmc.2007.01.026

A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
2369
objective was to develop more selective anti-MMP drugs
that would concentrate strongly in cartilaginous tissues
and thus allow a significant decrease of the effective dose
and, as a consequence, of general side effects. The results
of this structural modification on the anti-MMP proper-
ties of TC and DC were analyzed in vitro with respect to
mRNA expression of MMP-3 and MMP-13, the
production of MMP-13 in chondrosarcoma cells, the
catalytic enzymatic activity of MMPs on a synthetic
substrate and on articular cartilage explants, and finally,
collagen degradation on articular cartilage explants.
13C NMR spectra, the three trimethyl carbons of the
ammonium function were observed as one signal at
54.75 ppm for QA-TC and 55.56 ppm for QA-DC. A
HMQC sequence allowed to clearly distinguish the C-4
signal from C-6 and C-12a signals for QA-TC. The sig-
nal at 68.79 ppm was assigned to the C-6 and the signal
at 73.09 ppm to the C-12a based on comparison with the
TC spectrum.24 For QA-DC, the same HMQC sequence
allowed to distinguish the C-4 signal (74.92 ppm) from
C-5 (69.10 ppm) and the C-4a signal (46.56 ppm) from
C-5a (44.33 ppm).
2.2. Biological activity
2. Results and discussion
2.1. Chemistry
2.2.1. mRNA expression and production of MMPs. TC,
DC, and QA derivatives were first tested for their ability
to inhibit IL-1b-induced mRNA expression of MMP-3
and MMP-13 in human SW-1353 chondrosarcoma cells
(Fig. 1). RT-PCR analysis showed that IL-1b strongly
stimulated the mRNA expression of MMP-3 and -13
compared with the control. This effect of the cytokine
was totally blocked by 10À4 M TC or DC. QA-TC
and QA-DC were more effective in inhibiting the mRNA
expression of MMP-13 than MMP-3, but this selective
activity against MMP-13 remained weaker than that
of the parent tetracyclines at the same concentration.
Moreover, QA-DC was more potent than QA-TC. On
one hand, this loss of activity could be explained by a
reduced cellular uptake due to the positive charge intro-
duced by the quaternary ammonium function. On the
other hand, a region of the tetracycline involved in the
inhibition of MMP expression may have been modified
by the quaternization. However, this second hypothesis
could not be confirmed, since no structure–activity data
on the effects of tetracyclines on MMP expression at
mRNA level have been reported in the literature. This
study on MMP expression at the mRNA level was com-
plemented by experiments on MMP-13 protein levels,
which were measured by ELISA. The activation of
the mRNA expression of MMP-13 by IL-1b in the
SW-1353 cells was paralleled by a large increase in
enzyme release in the culture media (from 0.24 0.09
to 70.11 4.73 ng/mL; P < 0.001). At the highest tested
The quaternized conjugates of TC (QA-TC) and DC
(QA-DC) were prepared by direct methylation of the
C-4 dimethylamino group of TC and DC free bases
according to Scheme 1. Physical data and spectral char-
acterizations of QA-TC and QA-DC are presented in
the Section 4.
NMR proved to be particularly useful in clarifying the
chemical structures of various forms of tetracyclines
(hydrochloride, free base).24,25 1D (1H, 13C) and 2D
(COSY, HMQC) experiments performed at 500 MHz
enabled complete proton and carbon assignment for
each QA derivative in DMSO-d6, in accordance with
the data reported by Asleson and Casy for TC and
1
DC.24,25 In the H NMR spectra, two intense signals
at 1.50 (s) and 3.40 (s) ppm for QA-TC and at 1.42
(d) and 3.35 (s) ppm for QA-DC were assigned to the
C-6 methyl and C-4 trimethylammonium groups,
respectively. The presence of a hydroxyl group at C-5
in QA-DC did not influence the chemical shift of the
C-4 proton, as previously noted for DC.25 Moreover,
the two H-5 methylene protons were non-equivalent in
QA-TC. The H-5 methylene signal at 1.72 ppm that dis-
played a broad quadruplet pattern was assigned to the
pseudo-axial b-proton, whereas the broad doublet at
2.25 ppm was assigned to the pseudo equatorial a-pro-
ton, in accordance with Casy.25 Broad resonances at
5.02, 7.71, 9.43, 9.51, 11.74, and 15.24 ppm for QA-TC
and at 5.69, 7.67, 9.09, 9.22, 11.45, and 15.26 ppm for
QA-DC were assigned to amino and hydroxy groups,
since they disappear in samples that have been deuteri-
um exchanged. Moreover, the two NH2 protons were
non-equivalent in QA-TC and QA-DC. Finally, the sig-
nals of the aromatic protons were observed, as expected,
as a triplet and two doublets at 7.55, 6.91, 7.12 and 7.53,
6.85, 6.89 ppm for QA-TC and QA-DC, respectively. In
Figure 1. Effects of DC, TC, and their quaternized conjugates (QA-
DC, QA-TC) on the IL-1b-induced mRNA expression of MMP-3 and
MMP-13 in SW1353 human chondrosarcoma cells.
Scheme 1. Synthesis of QA derivatives.

2370
A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
concentration of 50 lM, TC and DC but not their QA
derivatives significantly inhibited IL-1b-induced MMP-
13 production in SW-1353 cells (data not shown).
expression show that different interactions take place be-
tween tetracyclines and (a) MMP proteins and (b) tar-
gets, still unknown, controlling MMP expression.
Also, they suggest that introduction of QA in C-4 can
modulate the anti-MMP activity known to be exerted
by the lower part of tetracycline molecules.23 We used
as a positive control the reference compound AG3340
(prinomastat), a potent broad-spectrum inhibitor of
MMPs, with Ki values between 3 · 10À11 and 3 · 10À10
for MMP-2, -3, -9, -13, and -14.25 In this study,
AG3340 inhibited MMP-2, -3, -8, and -13 with IC50
values of 4 · 10À10, 1 · 10À9, 2 · 10À10 and 9 · 10À10 M,
respectively.
2.2.2. MMP catalytic activity. We evaluated the inhibi-
tory effect of the tetracyclines on the catalytic activity
of the following MMPs: gelatinase A/MMP-2, stromely-
sin-1/MMP-3, collagenase-2/MMP-8, and collagenase-3/
MMP-13 (Table 1). The IC50 values were in the order of
10À4 M for TC and QA-TC, and 10À5 M for DC and
QA-DC. QA-TC and QA-DC increased MMP-2, -8,
and -13 inhibition by 50–64% and 182–716%, respective-
ly. Conversely, the quaternized conjugates were less
efficient inhibitors of MMP-3 than their parent tetracy-
clines. These diverging results on MMP activity and
2.2.3. Proteoglycan degradation. The anti-degradative
effects of tetracyclines were examined on cultured frag-
ments of rabbit articular cartilage. In order to activate
the pro-MMPs induced by IL-1b, tissue fragments
already stimulated by the cytokine were further incubat-
ed in the presence of the MMP activator APMA. Two
protocols were used: one to examine the effects of tetra-
cyclines on degradation depending on de novo MMP
production, the other to study the effects on degradation
depending on already active MMPs. In the first case
(ability of tetracyclines and their QA conjugates to
inhibit the production of IL-1b-induced MMPs) the
fragments were stimulated for 24 h by IL-1b in the
absence or presence of tetracyclines, and then exposed
to APMA for another 24 h (Fig. 2a). IL-1b combined
Table 1. Effects of DC, TC, and their quaternized conjugates (QA-DC,
QA-TC) on MMP-2, -3, -8, and -13 activity
a
IC50 (lM)
MMP-2 MMP-3 MMP-8 MMP-13
TC
QA-TC
220
140
140
150
À7
180
110
+64
180
120
+50
Activity enhancementb (%) +57
DC
QA-DC
24
7.7
25
81
31
11
40
4.9
Activity enhancementb (%) +212
À69
+182
+716
a Dose required to inhibit MMP activity by 50%.
b See Section 4.
Figure 2. Effects of DC, TC, and their quaternized conjugates (QA-DC, QA-TC) on proteoglycan degradation induced by IL-1b and APMA, in
fragments of rabbit articular cartilage. (a) Effects of the tetracyclines and their conjugates on IL-1b-induced MMP catabolic activity. (b) Effects of the
same tetracyclines and their conjugates on the catabolic activity of MMPs previously induced by IL-1b. Degradation is expressed as percentage of
radioactivity released from fragments previously labeled with 35SO4. Data are expressed as means sem; n = 8/group. The asterisks indicate a
significant difference between treated and IL-1b alone. *P < 0.05; ***P < 0.001.

A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
2371
with APMA strongly stimulated proteoglycan degrada-
tion. A significant reduction of proteoglycan degrada-
tion induced by IL-1b plus APMA was observed for
TC at 200 lM, with levels decreasing from 49% (control)
to 29%, that is, a 45% (P < 0.05) inhibition of proteogly-
can degradation. The effect was even stronger with 100
lM DC, with only 8% of proteoglycan degradation
compared to 49% for the control, that is, a 93% inhibi-
tion of proteoglycan degradation. However, no inhibi-
tion was observed in the presence of the two QA
derivatives used at the same concentrations as their par-
ent compounds. In order to investigate whether tetracy-
clines could inhibit the catalytic activity of the MMPs
already produced, cartilage fragments were stimulated
by IL-1b for 24 h and then exposed to APMA in the ab-
sence or presence of tetracyclines (Fig. 2b). When tissue
fragments were exposed to IL-1b followed by APMA,
proteoglycan degradation was significantly stimulated.
Nevertheless, no significant inhibition of proteoglycan
degradation by MMPs was observed at concentrations
reaching 400 lM, whatever the nature of the tetracy-
cline. Conversely, in a separate experiment, the reference
MMP-inhibitor AG3340 significantly inhibited proteo-
glycan degradation from a value of 87 1% for IL-1b
plus APMA, to 14 1% at the concentration of
0.1 lM (P < 0.001).
Figure 3. Effects of DC, TC, and their quaternized conjugates (QA-
DC, QA-TC) on IL-1b-induced collagen degradation, in the presence
of plasmin, in fragments of rabbit articular cartilage. The degradation
is expressed as percentage of collagen released. Data are expressed as
means sem; n = 8/group. Statistical significance: **P < 0.01;
***P < 0.001 between treated and IL-1b. §P < 0.05; §§§P < 0.001
between DC and QA-DC at the same concentration.
MMP-13/collagenase-3, particularly in situations of
increased tissue catabolism. MMP-13 has been shown
to degrade type II collagen, which is the main compo-
nent in cartilage, more rapidly than MMP-1 and
MMP-8, and to be upregulated in chondrocytes from
patients with osteoarthritis.26,27 Moreover, overexpres-
sion of MMP-13 alone could cause osteoarthritis-type
lesions in transgenic mice.28 In contrast with the two as-
says of proteoglycan degradation, where inhibition is
due to decrease of MMP production in one case, and
of MMP activity in the other, the assay of collagen deg-
radation can detect the effect of an inhibitor acting on
either or both of MMP activity and MMP production.
Compared with DC, QA-DC partially lost the ability
to inhibit MMP-13 expression, while improving the ef-
fect against the catalytic activity of MMP-13, as well
as MMP-8 and MMP-2, which also participate in colla-
gen breakdown.1 The balance of the two effects can ex-
plain why QA-DC, while still inhibiting collagen
degradation, is less effective than DC. Combination of
the effects against MMP production and activity can
also explain why QA-DC can antagonize collagen deg-
radation while being inactive in the assays of proteogly-
can breakdown.
This study on the inhibition of proteoglycan degrada-
tion depending on previously produced MMPs showed
that none of the tetracyclines, either parent molecules
or derivatives, could significantly decrease the catabolic
action of the pre-produced enzymes, at concentrations
ranging up to 400 lM. The very intense proteoglycan
degradation observed with this protocol could neverthe-
less be inhibited, but only with a very potent MMP
inhibitor such as AG3340, with IC50 values in the
10À10–10À9 M range, compared with IC50 values for tet-
racyclines between 10À6 and 10À4 M. This means that, at
least for proteoglycan degradation, the action of tetracy-
clines at concentrations in the range of 100 lM depends
more on the blockage of MMP production than on the
inhibition of the catalytic activity of pre-existing
enzymes.
2.2.4. Collagen degradation. This experiment evaluated
the effect of tetracyclines on both the production and
catabolic activity of MMPs. In this assay, collagen deg-
radation was dependent on both the induction of pro-
MMPs by IL-1b and their activation by plasmin. When
cartilage fragments were cultured in the presence of IL-
1b and plasmin, collagen degradation was stimulated.
DC and QA-DC significantly inhibited IL-1b- and plas-
min-induced collagen degradation by 89% and 66%,
respectively, at a concentration of 100 lM (Fig. 3). In
addition, AG3340 significantly inhibited collagen degra-
dation from a value of 54 1% for IL-1b plus plasmin,
to 15 1% at the concentration of 10À7 M (P < 0.001)
(data not shown).
3. Conclusion
The inhibition of collagen degradation by QA-DC is a
promising result for the development of a selective chon-
droprotective agent for the treatment of osteoarthritis.
Moreover, since we have previously shown in vivo that
QA derivatives exhibited an increased affinity for carti-
lage,15,16 we can hypothesize that the QA conjugates dis-
play a similar profile and preferentially concentrate in
the chondral matrix. In addition, it is already known
that quaternization of the C-4 dimethylamino group of
tetracyclines results in the loss of their antibiotic proper-
ties.29 These developments should allow a significant
decrease of the efficient dose administered and hence
an attenuation of adverse side effects. Future studies
In contrast to the results on proteoglycan degradation,
QA-DC was effective at inhibiting collagen degradation,
even though it was slightly less active than DC. Collagen
degradation in cartilage appears mainly to depend on

2372
A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
have been planned in order to investigate the biodistri-
bution of 14C-labeled QA-TC and QA-DC, and confirm
that these compounds can effectively target cartilage.
5.63 mmol) was added under argon to a stirred solution
of doxycycline (100 mg, 0.22 mmol) in anhydrous tetra-
hydrofuran (10 mL). The mixture was left at room tem-
perature, and a brown precipitate of QA-DC appeared
within a few hours. The reaction was completed after
24 h. The residue was filtered, washed with water and
ether, and dried under vacuum to afford QA-DC as a
brown solid (95 mg, 72%). Mp 204–206 ꢁC (dec); IR
(KBr) m (cmÀ1) 3500–2500 (NH2, OH); 1H NMR
(DMSO-d6) d 1.44 (d, J = 5.2, 3H, CH3), 2.48 (m, 1H,
H-5a), 2.62 (m, 1H, H-4a), 3.10 (m, 1H, H-6), 3.35 (s,
9H, N+(CH3)3), 4.55 (s, 1H, H-4), 5.69 (d, J = 10.4,
1H, OH-5), 6.85 (d, J = 7.9, 1H, H-9), 6.89 (d, J = 7.9,
1H, H-7), 7.53 (t, J = 7.9, 1H, H-8), 7.67 (s, 1H, OH-
10), 9.09 (se, 1H, NH2), 9.22 (se, 1H, NH2), 11.45 (s,
1H, OH-3), 15.26 (se, 1H, OH-12); 13C NMR
(DMSO-d6) d 16.87 (CH3), 39.54 (C-6), 44.33 (C-5a),
46.56 (C-4a), 55.56 (N+(CH3)3), 69.10 (C-5), 73.14 (C-
12a), 74.92 (C-4), 98.59 (C-2), 107.59 (C-11a), 116.31,
116.58, 116.81, 137.62, 148.66, 161.92 (aromatic car-
bons), 173.00 (C@O amide), 186.90, 192.89, 193.11 (C-
1, C-3, C-11); ESI-MS m/z 459 (MÀI)+; elemental anal-
ysis: Calcd for C23H27N2O8IÆH2O, C, 45.70; H, 4.84; N,
4.63. Found: C, 45.83; H, 4.96; N, 4.37.
4. Experimental
4.1. General methods
All chemicals were obtained from commercial suppliers.
Proton nuclear magnetic resonance (1H NMR) and car-
bon nuclear magnetic resonance (13C NMR) spectra were
recorded on a Bruker AM 500 spectrometer at 500 and
¨
50 MHz, respectively, and autocalibrated to the deuterat-
ed solvent reference peak. Chemical shift values (d) are
quoted in parts per million (ppm), and coupled constants
(J) in hertz (Hz). Infrared spectra (IR) were recorded as
KBr disks on a Bruker Vector 22 FTIR system. Electro-
¨
spray ionization mass spectra (ESI-MS) were performed
on a Bruker ESQUIRE-LC spectrometer. Melting points,
¨
uncorrected, were taken in glass capillary tubes on an
Electrothermal digital melting point apparatus.
4.2. 1,4,4a,5,5a,6,11,12a-Octahydro-2-aminocarbo-nyl-3,
6,10,12,12a-hexahydroxy-6-methyl-1,11-dioxo-naphtha-
cene-4-trimethylammonium iodide (QA-TC)
4.4. Inhibition of MMP activity
Tetracycline hydrochloride dissolved in water was neu-
tralized by NaOH (1 N) and the resulting precipitate
was filtered and dried. Methyl iodide (1.36 mL,
21.83 mmol) was added under argon to a stirred solu-
tion of tetracycline (388 mg, 0.87 mmol) in anhydrous
tetrahydrofuran (10 mL). The mixture was left at room
temperature, and a yellow precipitate of QA-TC ap-
peared within a few hours. The reaction was completed
after 4 days. The residue was filtered, washed with water
and ether, and dried under vacuum to afford QA-TC as
a yellow solid (291 mg, 57%). Mp 212–214 ꢁC (dec); IR
(KBr) m (cmÀ1) 3500–2500 (NH2, OH); 1H NMR
(DMSO-d6): d 1.50 (s, 3H, CH3), 1.72 (br q, J = 12.0,
1H, H-5b), 2.25 (br d, J = 12.1, 1H, H-5a), 2.88 (dd,
J = 4.6, J = 10.9, 1H, H-5a), 3.08 (d, J = 12.5, 1H, H-
4a), 3.40 (s, 9H, N+(CH3)3), 4.40 (s, 1H, H-4), 5.02 (s,
1H, OH-6), 6.91 (d, J = 8.3, 1H, H-9), 7.12 (d, J = 8.3,
1H, H-7), 7.55 (t, J = 8.3, 1H, H-8), 7.71 (s, 1H, OH-
10), 9.43 (se, 1H, NH2), 9.51 (se, 1H, NH2), 11.74 (s,
1H, OH-3), 15.24 (se, 1H, OH-12); 13C NMR
(DMSO-d6): d 23.36 (CH3), 27.74 (C-5), 36.84 (C-4a),
46.72 (C-5a), 54.75 (N+(CH3)3), 68.79 (C-6), 73.09 (C-
12a), 76.44 (C-4), 98.30 (C-2), 107.45 (C-11a), 115.32,
116.21, 118.02, 137.56, 148.73, 162.29 (aromatic
carbons), 173.60 (C@O amide), 187.03, 192.96, 193.79
(C-1, C-3, C-11); ESI-MS m/z 459 (MÀI)+; elemental
analysis: Calcd for C23H27N2O8IÆ2H2O, C, 44.38; H,
5.02; N, 4.50. Found: C, 44.77; H, 4.87; N, 4.49.
The test was adapted from Chollet et al.30 Reference
compound AG3340 (prinomastat), a broad-spectrum
MMP inhibitor,31 was synthesized at the Institut de
Recherches Servier. Human pro-MMPs were dissolved
in developing buffer (Novex, Cat. No. LC2671) at the
following concentrations: pro-MMP-2 (Boehringer) at
300 lg/mL; pro-MMP-3 (AbCys) and pro-MMP-8 (Pr.
G. Murphy, Univ. East Anglia) at 1 lg/mL; pro-
MMP-13 (Univ. East Anglia) at 2 lg/mL. Pro-enzymes
were activated by 2 mM p-aminophenyl mercuric ace-
tate (APMA, Sigma) at 37 ꢁC for 30 min (MMP-2) or
1 h (MMP-3, -8, -13). Activation was stopped by trans-
ferring the samples to ice. Tetracyclines and derived
molecules were dissolved at 10À2 M in dimethylsulfoxide
(DMSO), then serially diluted (1/10) in developing buffer
at concentrations ranging from 10À3 to 10À6 M. The
fluorogenic substrate (Bachem) for MMP-3 was (7-meth-
oxycoumarine-4-yl)-Arg-Pro-Lys-Pro-Tyr-Ala-Nva-
Trp-Met-Lys(Dnp)-NH2.32 The fluorogenic substrate
(Bachem) for MMP-2, -8, and -13 was Dnp-Pro-
Cha-Gly-Cys(ME)-His-Ala-Lys(Nma)-NH2.33
These
substrates were dissolved in DMSO at 10À2
and 2 · 10À3 M, respectively, and then diluted to
2 · 10À4 M in water. Assays were performed in 96-well
plates by adding to each well 70 lL of developing buffer,
10 lL of inhibitor (or buffer for the control), and 10 lL
of enzyme (or buffer for the blank). Each point was run
in triplicate. After a 30-min preincubation at 37 ꢁC,
10 lL of substrate was added and the plates were incu-
bated for 6 h at 37 ꢁC. Reading was then performed
using a fluorimeter at excitation and emission wave-
lengths of 340 and 440 nm, respectively. Substrate
degradation in the presence of inhibitor at a given
concentration was calculated as % fluorescence of con-
trol wells. IC50 on each enzyme was calculated using
EXCEL software from 3 points in the central linear
4.3. 1,4,4a,5,5a,6,11,12a-Octahydro-2-aminocarbo-nyl-
3,5,10,12,12a-hexahydroxy-6-methyl-1,11-dioxo-naph-
thacen-4-trimethylammonium iodide (QA-DC)
Doxycycline hydrochloride dissolved in water was neu-
tralized by NaOH (1 N) and the resulting precipitate
was filtered and dried. Methyl iodide (350 lL,

A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
2373
range of fluorescence inhibition. Activity enhancement
was calculated using the formula (IC50(TC) À
IC50(QA À TC))/IC50(TC) · 100, and the same formula
was used for QA-DC. According to this formula, an
IC50 from, for example, 100 to 50 lM corresponds to
a 100% gain in activity. Conversely, a change from
100 to 200 lM corresponds to a 50% decrease in
activity.
MMP-13 (pro-enzyme plus mature form) was assayed
using the ELISA Biotrak kit (Amersham).
4.7. Cartilage degradation
The detailed protocols of proteoglycan and collagen
degradation have been published previously.37 The tests
are based on the culture of fragments of rabbit articular
cartilage, as outlined below.
4.5. Expression of MMP-3 and MMP-13 mRNAs by
SW-1353 cells
4.7.1. Proteoglycan degradation. Fragments were first
incubated in the presence of 35SO4 in order to label neo-
synthesized proteoglycans, then washed to eliminate un-
bound radiolabel and used for either of the following
two assays. (1) MMP-production-dependent degrada-
tion: radiolabeled fragments were transferred to 96-well
plates, in the absence (control) or presence of 1 ng/mL
rmIL-1b, and supplemented or not with tetracyclines.
After 24 h, the media were discarded and the fragments
were transferred to plates containing 5 · 10À4 M APMA
as pro-MMP activator. After another 24 h, the fragments
were collected and digested in 0.6 mg/mL papain. Radio-
activity in the culture media and in the tissue digest was
measured by liquid scintillation using a b-counter (Beck-
man). (2) Activated-MMP-dependent degradation: radi-
olabeled fragments were first batch-stimulated with
10 ng/mL rmIL-1b for 24 h, then transferred to 96-well
plates, in the absence (control) or presence of
5 · 10À4 M APMA, and supplemented or not with tetra-
cyclines. After another 24 h, the media and fragments
were collected and processed as above for measurement
of proteoglycan degradation. For both assays, proteogly-
can degradation in each fragment was expressed as the
percentage of released radioactivity using the following
formula: degradation = media radioactivity/(media
radioactivity + tissue radioactivity) · 100.
Human chondrosarcoma SW-1353 cells (ATCC) were
seeded in 6-well culture plates at
a density of
2.5 · 105 cells/well/2 mL of Ham’s F12 media (Invitro-
gen) supplemented with 10% fetal bovine serum (FBS)
and 1% of a stock solution of 10 IU/mL penicillin plus
10 mg/mL streptomycin (PS, Invitrogen). The media
were changed every 2–3 days until confluence was
attained. Cells were then washed with Hanks’ balanced
salt solution (HBSS, Invitrogen), then incubated in ser-
um-free Ham’s F12 media supplemented with 0.1%
bovine serum albumin (BSA, Sigma) and 1% PS. After
24 h, the cells were treated with vehicle (DMSO) or tet-
racyclines (solubilized in DMSO at 10À2 M) at a final
concentration of 10À4 M. Thirty minutes later, vehicle
(basal control) or 1 ng/mL recombinant murine (rm)
IL-1b (Sigma) was added. After 24 h, the media were
discarded and the mRNA was analyzed according to
the previously published protocol.34 Briefly, total RNA
was extracted using the Rneasy Mini Kit (Qiagen) fol-
lowing the manufacturer’s instructions. Five micro-
grams of total RNA was loaded onto a 1% agarose gel
containing ethidium bromide in order to check for
RNA integrity and loading. Then, 2.5 lg of total
RNA was reverse-transcribed (RT) for 1 h at 37 ꢁC,
using oligo d(T)12-18 and superscript II, in a total vol-
ume of 20 lL. Two microliters of the RT reaction was
used for each PCR run. Gene-specific oligonucleotide
primers were designed from the reported cDNA
sequences of human MMP-335 and MMP-13,36 and
18S rRNA (Clontech) was used as a housekeeping gene.
The amplification profile used 35 cycles of denaturation
at 94 ꢁC for 30 s, annealing at 55 ꢁC for 30 s, and
extension at 72 ꢁC for 1 min in a Gene Amp PCR
system 2400 (Perkin-Elmer) using VENT polymerase
(Biolabs). The PCR products were then separated by
electrophoresis on a 1% agarose gel and photographed
under UV light.
4.7.2. Collagen degradation. Fragments (not radiola-
beled) were transferred to 96-well plates, in the absence
(control) or presence of 10 ng/mL IL-1b plus 0.1 U/mL
EACA- and Lys-free plasmin (Calbiochem), and supple-
mented or not with tetracyclines. After 48 h, the tissue
fragments and the media were hydrolyzed in 6 M HCl
at 110 ꢁC for 12 h. Hydroxyproline (OH-Pro) in media
and tissue hydrolysates was assayed by colorimetry.38
Collagen degradation in each fragment was expressed
as the percentage of OH-Pro released using the follow-
ing formula: degradation = media OH-Pro/(media OH-
Pro + tissue OH-Pro) · 100.
4.6. Production of MMP-13 by SW-1353 cells
4.8. Statistics
SW-1353 cells were seeded in 48-well culture plates at a
density of 1.5 · 104 cells/well/0.5 mL of Ham’s F12 med-
ia supplemented with 10% FBS and 1% PS. The media
were renewed every 2–3 days until cells were confluent.
After three washes with HBSS, cells were incubated in
serum-free media supplemented with 0.1% BSA and
1% PS. After 24 h, the cells were treated with vehicle
(DMSO) or tetracyclines at concentrations ranging from
6 · 10À6 to 5 · 10À5 M. After 30 min, vehicle (basal con-
trol) or rmIL-1b (10 ng/mL) was added. After another
24 h, the media were collected and total secreted
Data are expressed as means standard error of the
mean (sem). Control and treated groups were statistical-
ly compared by analysis of variance followed by
Dunnett’s t test. Significance levels are shown as follows:
*
***P < 0.001; **P < 0.01; P < 0.05.
Acknowledgments
We thank Dr. Guillaume De Nanteuil (Division of
Medicinal Chemistry, Institut de Recherches Servier)

2374
A. Vidal et al. / Bioorg. Med. Chem. 15 (2007) 2368–2374
19. Giraud, I.; Rapp, M.; Maurizis, J.-C.; Madelmont, J.-C.
Bioconjug. Chem. 2000, 11, 212–218.
20. Rapp, M.; Giraud, I.; Maurizis, J.-C.; Madelmont, J.-C.
Bioorg. Med. Chem. 2003, 11, 5007–5012.
for synthesis of AG3340. This work was supported by
ADIR, Servier (France).
21. Rapp, M.; Giraud, I.; Maurizis, J.-C.; Galmier, M.-J.;
Madelmont, J.-C. Bioconjug. Chem. 2003, 14, 500–
506.
References and notes
22. Giraud, I.; Rapp, M.; Maurizis, J.-C.; Madelmont, J.-C.
J. Med. Chem. 2002, 45, 2116–2119.
23. Smith, G. N.; Hasty, K. A. In Tetracyclines in Biology,
Chemistry and Medicine; Nelson, M., Hillen, W., Green-
wald, R. A., Eds.; Birkhauser Verlag: Basel, 2001; pp
283–293.
1. Goldring, M. B. Arthritis Rheum. 2000, 43, 1916–1926.
2. Murphy, G.; Willenbrock, F. Methods Enzymol. 1995,
248, 496–511.
3. Hutton, M.; Willenbrock, F.; Brocklehurst, K.; Murphy,
G. Biochemistry 1998, 37, 10094–10098.
4. Sabatini, M.; Lesur, C.; Thomas, M.; Chomel, A.; Anract,
P.; de Nanteuil, G.; Pastoureau, P. Arthritis Rheum. 2005,
52, 171–180.
24. Asleson, G. L.; Franck, C. W. J. Am. Chem. Soc. 1975, 97,
6246–6249.
25. Casy, A. F.; Yasin, A. Magn. Reson. Chem. 1985, 23,
767–770.
5. Cole, A. A.; Chubinskaya, S.; Luchene, L. L.; Chlebek,
K.; Orth, M. W.; Greenwald, R. A.; Kuettner, K. E.;
Schmid, T. M. Arthritis Rheum. 1994, 37, 1727–1734.
6. Orth, M. W.; Chlebek, K. A.; Cole, A. A.; Schmid, T. M.
Res. Vet. Sci. 1997, 63, 11–14.
26. Reboul, P.; Pelletier, J.-P.; Tardif, G.; Cloutier, J.-M.;
Martel-Pelletier, J. J. Clin. Invest. 1996, 97, 2011–2019.
27. Shlopov, B. V.; Lie, W.-R.; Mainardi, C. L.; Cole, A. A.;
Chubinskaya, S.; Hasty, K. A. Arthritis Rheum. 1997, 40,
2065–2074.
7. Steinmeyer, J.; Daufeldt, S.; Taiwo, Y. O. Biochem.
Pharmacol. 1998, 55, 93–100.
28. Neuhold, L. A.; Killar, L.; Zhao, W.; Sung, M.-L. A.;
Warner, L.; Kulik, J.; Turner, J.; Wu, W.; Billinghurst, C.;
Meijers, T.; Poole, A. R.; Babij, P.; De Gennaro, L. J.
J. Clin. Invest. 2001, 107, 35–44.
29. Nelson, M. L. In Tetracyclines in Biology, Chemistry and
Medicine; Nelson, M., Hillen, W., Greenwald, R. A., Eds.;
Birkhauser Verlag: Basel, 2001; pp 3–63.
30. Chollet, A.-M.; Le Diguarher, T.; Kucharczyk, N.; Loy-
nel, A.; Bertrand, M.; Tucker, G. C.; Guilbaud, M.;
Burbridge, M.; Pastoureau, P.; Fradin, A.; Sabatini, M.;
8. Shlopov, B. V.; Smith, G. N.; Cole, A. A.; Hasty, K. A.
Arthritis Rheum. 1999, 42, 719–727.
9. Shlopov, B. V.; Stuart, J. M.; Gumanovskaya, M. L.;
Hasty, K. A. J. Rheumatol. 2001, 28, 835–842.
10. Hanemaaijer, R.; Sorsa, T.; Konttinen, Y. T.; Ding, Y.;
Sutinen, M.; Visser, H.; Van Hinsbergh, W. M.; Helaako-
ski, T.; Kainulainen, T.; Ro¨nka, H.; Tschesche, H.; Salo,
T. J. Biol. Chem. 1997, 272, 31504–31509.
11. Brandt, K. D.; Mazzuca, S. A.; Katz, B. P.; Lane, K. A.
Arthritis Rheum. 2003, 48(9-Suppl), LB22, abstract.
12. Brandt, K. D.; Mazzuca, S. A.; Katz, B. P.; Lane, K. A.
Arthritis Rheum. 2003, 48(9-Suppl), LB1, abstract.
13. Marzo-Ortega, H.; Strauss, R. M.; Emery, P. In Tetracy-
clines in Biology, Chemistry and Medicine; Nelson, M.,
Hillen, W., Greenwald, R. A., Eds.; Birkhauser Verlag:
Basel, 2001; pp 311–324.
`
Fauchere, J.-L.; Casara, P. Bioorg. Med. Chem. 2002, 10,
531–544.
31. Sorbera, L. A.; Castan˜er, J. Drugs Future 2000, 25, 150–
158.
32. Nagase, H.; Fields, C. G.; Fields, G. B. J. Biol. Chem.
1994, 269, 20952–20957.
33. Bickett, D. M.; Green, M. D.; Berman, J.; Dezube, M.;
Howe, A. S.; Brown, P. J.; Roth, J. T.; McGeehan, G. M.
Anal. Biochem. 1993, 212, 58–64.
14. Zerler, B. R. In Tetracyclines in Biology, Chemistry and
Medicine; Nelson, M., Hillen, W., Greenwald, R. A., Eds.;
Birkhauser Verlag: Basel, 2001; pp 325–332.
34. Rolland-Valognes, G. Methods Mol. Med. 2004, 100,
69–78.
35. Witham, E.; Murphy, G.; Angel, P.; Rahmsdorf, H.-J.;
Smith, B. J.; Lyons, A. J. Biochem. 1986, 240, 913–
916.
36. Freije, J. M.; Diez-Itza, I.; Balbin, M.; Sanchez, L. M.;
Blasco, R.; Tolivia, J. J. Biol. Chem. 1994, 269, 16766–
16773.
37. Lesur, C.; Sabatini, M. Methods Mol. Med. 2004, 100,
219–236.
15. Maurizis, J.-C.; Ollier, M.; Nicolas, C.; Madelmont, J.-C.;
Guarrigue, H.; Veyre, A. Biochem. Pharmacol. 1992, 44,
1927–1933.
16. Maurizis, J.-C.; Rapp, M.; Nicolas, C.; Ollier, M.; Verny,
M.; Madelmont, J.-C. Drug Metabolism and Disposition
2000, 28, 418–422.
17. Nicolas, C.; Borel, M.; Maurizis, J.-C.; Gallais, N.; Rapp,
M.; Ollier, M.; Verny, M.; Madelmont, J.-C. J. Labelled
Compd. Radiopharm. 2000, 43, 585–594.
18. Nicolas, C.; Verny, M.; Giraud, I.; Ollier, M.; Rapp, M.;
Maurizis, J.-C.; Madelmont, J.-C. J. Med. Chem. 1999, 42,
5235–5240.
38. Woessner, J. F. Arch. Biochem. Biophys. 1961, 93,
440–447.