MINIREVIEW
Regulation of matrix metalloproteinase activity in health
and disease
Elin Hadler-Olsen, Bodil Fadnes, Ingebrigt Sylte, Lars Uhlin-Hansen and Jan-Olof Winberg
Department of Medical Biology, Faculty of Health Sciences, University of Tromsø, Norway
Introduction
Matrix metalloproteinases (MMPs) are a subfamily of
zinc- and calcium-dependent enzymes belonging to the
metzincin superfamily. Characteristic for this super-
family is the HEXXHXXGXXH zinc-binding motif
and a conserved methionine located C-terminal to the
zinc-ligands, which forms a Met-turn [1]. In humans,
there are 24 MMP genes, but only 23 MMP proteins
because MMP-23 is coded by two identical genes at
chromosome 1. MMPs are built up by various
domains (Fig. 1). All MMPs contain an N-terminal
signal peptide that directs the enzymes to the secretory
pathway, a prodomain with a conserved PRCGXPD
sequence that confers the latency of the enzymes and a
catalytic domain with the catalytic zinc localized in the
large and relatively shallow active site cleft. In addi-
tion, all MMPs except the two matrilysins (MMP-7
and -26) and MMP-23 contain a C-terminal hemopex-
in (HPX)-like domain that is linked to the catalytic
domain through a hinge region. In most MMPs, this
hinge region consists of 10–30 amino acids, whereas, in
MMP-9, this linker contains approximately 64 amino
acids and is heavily O-glycosylated [2]. In six of the
membrane-anchored members of the MMP family, the
HPX region ends in either a type I transmembrane
domain with a short intracellular sequence or a glycosyl-

Knowledge about regulation of MMP activity is essential for understanding
various physiological processes and pathogenesis of diseases, as well as for
the development of new MMP targeting drugs.
Abbreviations
APMA, p-aminophenylmercuric acetate; CS, chondroitin sulfate; FnII, fibronectin II; GAG, glycosaminoglycan; GSH, glutathione;
Hp, haptoglobulin; HNL, human neutrophil lipocalin; HPX, hemopexin; MMP, matrix metalloproteinases; MMPI, metalloproteinase inhibitor;
MT-MMP, membrane-type matrix metalloproteinase; NuMAP, nuclear MMP-3 associated protein; PG, proteoglycan; SIBLING, small integrin-
binding ligand N-linked glycoprotein; TIMP, tissue inhibitor of metalloproteinase; TnI, troponin I.
28 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
other MMPs by lacking the HPX domain, which is
replaced by a C-terminal cystein array region and an
immunoglobulin G-like domain, and, instead of the
N-terminal signal peptide, this enzyme contains an
N-terminal type II transmembrane domain. In two of
the secreted MMPs, MMP-2 and MMP-9, the catalytic
domain also contains a module of three fibronectin II
(FnII)-like inserts. A recent review described the molec-
ular interactions of the HPX domains of the various
human MMPs [3]. The various domains, modules and
motifs in the MMPs are involved in interactions with
other molecules, and hence affect or determine the activ-
ity, substrate specificity, and cell and tissue localization,
as well as activation, of MMPs.
Activation of proMMPs requires physical delocaliza-
tion of the prodomain from the catalytic site, the
so-called cystein-switch model [4]. Two main mecha-
nisms are involved in the activation of MMPs. One is
proteolytic cleavage and removal of the prodomain
[5,6] and the other is allosteric activation where the
prodomain is displaced from the catalytic site of

Membrane-anchored MMPs
Fibronectin II-like domain
MMP-2
MMP-9
Minimal domain
MMP-7, -26
MMP-11, -21,
-28
Simple HPX domain
MMP-1, -3, -8,
-10, -12, -13,
-19, -20, -27
Furin-activated
MMP-23
MT1, -2, -3,
-5-MMP
MT4, -6-MMP
Zn
N
C
Zn
N
Zn
N
Zn
N
Zn
N
Zn
N

domain. The two minimal domain MMPs and MMP-23 lack the HPX domain and, in the latter enzyme, this domain is replaced by a C-termi-
nal cystein array (Ca) and an immunoglobulin-like (Ig) domain.
E. Hadler-Olsen et al. Regulation of MMP activity
FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS 29
MMP activities. Previous reviews have described new
techniques that can be used in the search for exosites
and examples of exosites derived from the use of these
techniques [9,10]. Although our knowledge of specific
exosites in the various MMPs is still very limited, these
sites will become of increasing importance as targets
for future drug development. Hopefully, future drugs
will not affect all substrate degradation by a given
enzyme, but only the processing of selected substrates.
Some of the substrates that MMPs are known to
process are localized intracellularly [7]. Although all
MMPs contain a signal peptide that directs them to
the secretory pathway, an increasing number of reports
have found various MMPs localized also inside cells.
This may partly explain the ability of some MMPs to
process intracellular proteins and further demonstrates
the complex roles of MMPs under physiological and
pathological conditions. Among the earliest intracellu-
lar MMP substrates detected are troponin I (TnI) [11],
aB-crystallin [12] and lens bB1 crystallin [13]. In vivo,
MMP cleavage of these substrates was linked to health
and disease. MMP-2 degradation of TnI is associated
with diminished contractive function of the heart [11],
MMP-9 degradation of aB-crystallin with multiple
sclerosis [12] and MMP-9 cleavage of lens bB1 crystal-
lin with cataract [13].

4
Zn
SH
N
1a
SH
Zn
2a
Auto-
cleavage
Zn
3a
Auto-
cleavage
Proteases
Mercurials
SH-reactive agents
Chaotropic agents
ROS
Detergents
Zn
N
SH
Zn
N
SH
1
Zn
SH
N

or APMA to proMMP-9 results in a conformational change (step 1a) followed by autocleavage
that did not remove the conserved PRCGV sequence from the enzyme (step 2a). This truncated enzyme had a low specific activity. Neutro-
phil gelatinase associated lipocalin (NGAL) ⁄ human neutrophil lipocalin (HNL) bound to the new N-terminus without further processing of the
enzyme (step 3a), resulting in a fully active enzyme (see text).
Regulation of MMP activity E. Hadler-Olsen et al.
30 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
formation of heterodimers and complexes, compart-
mentalization, and the role of exosites in substrate deg-
radation and enzyme inhibition.
Activation mechanisms
To induce activation of a proMMP, the prodomain
must be physically delocalized from the catalytic site
(Fig. 2). There are various ways to achieve such a
delocalization followed by activation. One is through
S-reactive agents, organomercurials and reactive oxy-
gen species, interacting with the conserved cysteine in
the prodomain. Another is the induction of conforma-
tional changes through binding of chaotropic agents
and detergents such as SDS. In all cases, the confor-
mational changes (Fig. 2A, step 1a) are followed by an
autocatalytic stepwise degradation of the prodomain
(Fig. 2A, steps 2a and 3a) [5,6]. Proteinases can cleave
the prodomain in one or several steps, producing an
active MMP with reduced molecular size. A large
number of proteinases such as serine and metallopro-
teinases are involved in the activation of proMMPs. In
some cases, one enzyme generate a partly active
enzyme that can be fully activated by a second enzyme
removing one or more amino acids from the prodo-
main, as described for MMP-1 [5,6]. Thus, it is not

[22]. At 37 °C, the
presence of 0.05% Brij-35 and 10 mm Ca
2+
mainly
prevented both activation and degradation, whereas a
lack of these two compounds resulted in trypsin-
induced degradation. However, at intermediate concen-
trations of Brij-35 and Ca
2+
, trypsin induced the
activation of proMMP-2. Different modes of activa-
tion can have implications for the biochemical proper-
ties of the enzymes depending on the cleavage site. In
the trypsin-activated MMP-2, the N-terminal residue
was either Lys87 or Trp90 [22], with the former being
identical to the cleavage site generated by human tryp-
sin-2 [23]. The N-terminal residue was Tyr81 in mem-
brane-type1 MMP (MT1-MMP) or APMA-activated
enzyme [6]. The slightly shorter N-terminus in the
trypsin-activated enzyme resulted in reduced catalytic
efficiency and weaker tissue inhibitor of metallopro-
teinase (TIMP)-1-binding compared to the enzyme
activated by MT1-MMP or APMA [22]. Docking stud-
ies of TIMP-1 revealed that the slightly weaker binding
of the inhibitor to the trypsin-activated MMP-2 could
be attributed to its shorter N-terminus (Lys87 ⁄ Trp90
versus Tyr81) because Phe83 and Arg86 interacted
directly with the inhibitor.
Activation through domain specific interactions
A proMMP can be presented to its activator protein-

degrade DQ-gelatin in the presence of low concentra-
tions of the triple-helical domain of the a2 chain of the
microfilamentous collagen VI [25]. The above examples
are illustrated in Fig. 2B (step 1).
Interactions with small integrin-binding ligand
N-linked glycoprotein (SIBLING)
Individual members of the SIBLING family are known
to bind strongly to both pro- and active forms of
specific MMPs. Bone sialoprotein binds MMP-2,
osteopontin binds MMP-3 and dentin matrix protein-1
binds MMP-9, all with a 1 : 1 stoichiometric ratio and
binding constants in the nanomolar range [26]. These
SIBLINGs and MMPs are also co-expressed and colo-
calized in salivary glands of humans and mice [27].
The interaction between the SIBLING and its partner
proMMP resulted in an active MMP without autocata-
lytic removal of the propeptide [26]. Studies indicated
that binding of a SIBLING to a proMMP induced
large conformational changes in the enzyme, suggest-
ing that the propeptide is physically removed from the
catalytic site, thereby allow substrate binding (Fig. 2B,
step 1). Furthermore, the three SIBLINGs have a ten-
to 100-fold higher affinity for the complement regula-
tor factor H than for their partner MMPs. The proM-
MP ⁄ SIBLING complex was dissociated in the presence
of factor H and a re-inactivation of the catalytic
activity by the still attached propeptide [26]. The same
research group also showed that the amino-terminal
region, especially exon 4, is essential for bone sialopro-
tein-mediated activation of proMMP-2 [28]. It appears

an activity comparable to trypsin activated MMP-9
[33]. Despite the full activity of the HgCl
2
and HNL
activated enzyme, this had an N-terminus identical to
the HgCl
2
activated enzyme (Fig. 2B, step 3a) [33],
whereas trypsin activation of MMP-9 caused removal
of the entire propeptide, with Phe88 as the N-terminal
residue (Fig. 2A, steps 1 and 2) [6]. Similar results were
obtained with isolated proMMP-9 homodimer and
proMMP-9 ⁄ HNL heterodimer when activated with
HgCl
2
and an excess of HNL. By contrast, HNL had
no effect on trypsin-activated MMP-9. Kallikrein is a
plasma proteinase that can partially activate proMMP-
9, and the presence of an excess of HNL resulted in a
synergistic effect with a 30–50% increase in activity
compared to kallikrein activation alone. Altogether,
this suggested that the N-terminus of the partially acti-
vated proenzyme is entrapped in the hydrophobic-
binding pocket of HNL and the propeptide–HNL
complex is thereby detached from the catalytic site,
generating a fully active enzyme without further trun-
cation (Fig. 2B, step 3a) [33].
Activation by peroxynitrite and glutathione
Enzymatic activity of intact proMMPs against physio-
logical substrates has also been detected in the pres-

activation of MMPs by modulating the conserved Cys
in the enzyme prodomains [39].
Activation through proteolytic removal
of the prodomain
TIMP regulation of MT1-MMP-induced activation
of proMMP-2
MT1-MMP-induced activation of proMMP-2 is a two-
step process involving the MMP inhibitor, TIMP-2,
which has been described in detail in several reviews.
Briefly, it has been shown that the TIMP-2 enhancement
of the MT1-MMP-induced activation of proMMP-2 is a
result of the formation of a ternary complex where the
inhibitor acts as a link between the two enzymes. In this
complex, the MT1-MMP is inactive as a result of its
interaction with the N-terminal part of TIMP-2,
whereas the C-terminal part of the inhibitor binds to the
HPX domain of proMMP-2. Another MT1-MMP mol-
ecule can now cleave the proMMP-2 in the complex and
generate a 64 kDa inactive intermediate. This intermedi-
ate is further autocatalytically processed into the fully
active 62 kDa form of MMP-2 [40,41]. The step at
which TIMP-2 is involved when it enhances the
MT1-MMP-induced activation of proMMP-2 has been
questioned because studies have shown that the inhibi-
tor enhances the autoactivation step, but is not neces-
sary for the first cleavage step [42–44]. The other
TIMPs, TIMP-1, -3 and -4, can also regulate the MT1-
MMP-induced activation of proMMP-2 [43], where
TIMP-1 only prevents the second step (autoactivation)
and locks the enzyme in an inactive intermediate form

but also enhanced the activation of proMMP-2 by all
MT-MMPs. Claudin-1 binds to both MT1-MMP and
proMMP-2, and this binding appears to involve only
the catalytic domains of the two enzymes. Another
membrane protein shown to enhance MT1-MMP-
induced activation of proMMP-2 was avb3 integrin
[49–51]. MMP-2 binds through its HPX domain to an
MT1-MMP-cleaved and activated form of avb3 inte-
grin [49–52]. This activated integrin enhanced the sec-
ond autocatalytic step of the activation by binding to
the 64 kDa intermediate form of MMP-2 [52]. Binding
and activation of MMP-2 was abrogated in the pres-
ence of avb3 integrin-binding macromolecules such as
vitronectin and HKa (two-chain high molecular weight
kinogen) [50,52,53]. Binding of MMP-2 to avb3 inte-
grin appears to be controversial because the findings
of another study did not support an interaction
between MMP-2 ⁄ PEX and avb3 integrin [54].
Activation through interactions with elastin,
heparin and CD151
ProMMP-2 and active MMP-2 binds to soluble and
insoluble elastin through the FnII module of the cata-
lytic domain [55]. When proMMP-2 binds to insoluble
elastin, this induces a fast autoactivation of the proen-
zyme followed by inactivation [56]. A similar phenom-
enon of enhanced autolysis has also been observed
when proMMP-2 binds to heparin, although this bind-
ing involves the enzymes C-terminal HPX domain [57].
These are just two examples of how an interaction of
E. Hadler-Olsen et al. Regulation of MMP activity

-induced activation
resulted in a proteolytic removal of the propeptide
from the complex bound proMMP-9. In the presence
of Ca
2+
, activated enzyme forms were also released
from the complex. This was the result of cleavage of a
part of the PG core protein and at least a part of the
C-terminal HPX domain of proMMP-9, leaving the
hinge region bound to the enzyme [64]. A large reduc-
tion of the HPX is likely to alter substrate specificity
because several specific substrate exosites in the HPX
domain may have been removed. Only the proMMP-9
in the CSPG complex was activated when Ca
2+
was
added to a mixture of purified proMMP-9 and proM-
MP-9 ⁄ CSPG complex. Furthermore, a mixture of
Ca
2+
and APMA did not activate the proMMP-
9 ⁄ CSPG complex [64], although Ca
2+
is known to
participate and enhance APMA induced activation of
proMMP-9 [62,65–67].
During hemolysis and ⁄ or hemorrhage, Hb is released
into the circulation and ⁄ or into surrounding tissues.
Heme, the prostetic group of Hb, is released from the
protein and converted to hemin, the Fe

different MMP-9 complexes results in altered biochem-
ical properties of the enzyme. In cells that produce
both proMMP-9 and TIMP-1, these two molecules are
bound together through their C-terminal domains, and
the presence of TIMP-1 affects the activity of the
enzyme [69]. When proMMP-9 forms a dimer with col-
lagenase, binding to TIMP-1 is prevented [70]. There
are conflicting data concerning whether the proMMP-9
homodimer is able to form a complex with TIMP-1
[59,61,71].
In its heterodimer form with neutrophil gelatinase
associated lipocalin, proMMP-9 can bind TIMP-1 and
form a ternary complex [72] and the enzyme is pro-
tected from degradation [73]. Two members of the
cystatin family, fetuin-A and cystatin C, bind to
MMP-9 and protect the enzyme from autolytic degra-
dation [74]. The above examples show that there are
different ways by which the activity of an MMP can
be regulated and preserved.
Both MMP-9 and MMP-2 interact with gelatin as
well as collagen through the three FnII-like modules in
their catalytic domain [70,75–84]. This interaction is
important for the ability of these enzymes to degrade
these physiological substrates, although it has no effect
on their degradation of several other physiological sub-
strates or chromogenic peptide substrates. ProMMP-9
forms a complex with one or several CSPG core
Regulation of MMP activity E. Hadler-Olsen et al.
34 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
proteins through its HPX domain [64]. When proM-

the clearance of Hb. The HPX region of MMP-9 has
been shown to form a strong reduction sensitive com-
plex with Hp [88]. Gelatin was reported to bind more
strongly to the proMMP-9 ⁄ Hp complex than to either
proMMP-9 monomer or homodimer, although the spe-
cific activity against gelatin was similar for the active
MMP-9 ⁄ Hp complex and the active MMP-9 monomer.
Furthermore, binding of proMMP-9 to Hp did not
influence the activation of the enzyme by MMP-3.
Binding of MMP-9 to Hp may comprise a method of
regulating MMP-9 activity because Hp is known to
bind cellular receptors followed by internalization and
degradation.
Role of exosites in regulation of activity
The complex substrate specificity of individual MMPs
is not only determined by their substrate-binding sub-
sites on each side of the catalytic zinc, but also by sub-
strate-binding to motifs outside this region (exosites).
The role of exosites has been recognized for a long
time for enzymes acting on polymer biomolecules such
as the restriction endonucleases [89], although was not
reported until 1989 for MMPs [90]. It was observed
that stored MMP-1 was autocatalytically truncated,
which resulted in a processed enzyme lacking the
C-terminal HPX domain. This truncated enzyme was
no longer able to cleave triple helical collagen I, but was
able to degrade gelatin (denatured collagen). The HPX
region was also found to be necessary for the cleavage
of the triple helical region in interstitial collagen by
other collagen-degrading MMPs (MMP-2, -8, -13 and -

for the degradation of some substrates.
Many potent small molecule MMP inhibitors
(MMPIs) have been entered into clinical trials for can-
cer treatment, although most of them have been dis-
continued as a result of a lack of specificity and
selectivity. Successful cancer therapy based on MMPIs
must not only be selective against MMPs validated as
targets, but also spare MMPs validated as antitargets
[97,98]. To develop new therapeutic MMPIs, it is of
pivotal importance to understand the structural basis
of recognition, binding and cleavage of substrates, as
well as the recognition and binding of natural inhibi-
tors (TIMPs). Recent data indicate that subtype spe-
cific inhibitors may also lead to new treatment of acute
and chronic inflammatory and vascular diseases [99].
Most known MMPIs are targeting the catalytic
region and the catalytic zinc, which are very similar
between the MMPs. Designing specific small molecular
MMPIs targeting the catalytic site is therefore prob-
lematic [99]. MMPIs targeting less conserved binding
sites outside the prime subsites of MMPs are consid-
ered to be more specific. Within the MMP family, dis-
tinct preferences for collagen types are seen, which
must reflect structural differences in MMP collagen-
binding [100]. Exosites are considered to be important
E. Hadler-Olsen et al. Regulation of MMP activity
FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS 35
determinants for these differences in specificity by
introducing contact regions between the substrate and
the MMP outside the primary specificity subsites. Exo-

region shows diversity among MMPs. A recent study
has further elucidated the role of the specificity loop for
selective MMP-13 inhibition by indentifying the steric
requirements for binding to this region [108]. Other
studies have also described selective MMP-13 inhibitors
that do not interfere with the catalytic zinc [101,109].
Regulation of activity through
compartmentalization
Through their motifs and modules, the secreted MMPs
are directed to various compartments in the extracellu-
lar environment as well as to cell membranes. Among
their binding partners in these compartments are colla-
gens, laminins, fibronectin, elastin, core proteins and
GAG-chains of PGs. This compartmentalization regu-
lates the MMP activity by locating and concentrating
them close to or on potential substrates. The interac-
tion with their binding partners varies in strength,
which has implications for the ability to extract a given
enzyme from a tissue. Examples are the binding of
MMP-1, -2, -7, -8, -9 and -13 to heparin and heparan
sulfate [57,72,110–117], where the interaction with hep-
arin occurs through the HPX domain of MMP-1, -2
and 9 [110,112,116]. MMP-7 lacks the HPX domain
and interacts through the catalytic and the prodomain.
This MMP binds much stronger to the GAG-chains
than the other MMPs [117]. MMP-7 could be
extracted from tissues by heparinase digestion or by
using extraction buffer containing heparin, heparan
sulfate or protamin [117]. Similarily, it was necessary
to use various extraction conditions to quantify the

MMP-3 was absent from the nucleus, suggesting that
processing is required to expose the nuclear localiza-
tion signal for nuclear transport [125]. For MT1-
MMP, a caveolae-mediated endocytosis has been sug-
gested as a mechanism of internalization and nuclear
translocation as a result of the colocalization of caveo-
lin-1 and MT1-MMP in perinuclear regions [126].
Nuclear localization of MMPs has been associated
with apoptosis in several studies. Increased nuclear
Regulation of MMP activity E. Hadler-Olsen et al.
36 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
gelatinolytic activity, colocalized with MMP-2, has
been demonstrated in pulmonary endothelial cells
undergoing apoptosis. MMP-2 activation in these cells
was suggested to be induced by reactive oxygen and
nitrogen species produced by cigarette smoke [127]. In-
tranuclear gelatinolytic activity has also been observed
in rat brain neurons after post-ischemic reperfusion,
and this activity was associated with DNA fragmenta-
tion. Furthermore, this gelatinolytic activity colocalized
with MMP-2 and MMP-9, and was reported to be
markedly reduced in the presence of a general MMP
inhibitor or by MMP-2 and MMP-9 antibodies. MT1-
MMP as well as furin, a MT1-MMP activator, was also
found in the nucleus of the ischemic rat brain neurons,
suggesting a possible mechanism for intracellular acti-
vation of MMP-2 by MT1-MMP [128].
In both cardiac myocytes and pulmonary endothelial
cells, as well as in brain neuronal cells, nuclear gelatin-
olytic activity is correlated with the processing of two

whereas the prodomain and the hinge-region alone
had no effect on the activation. Compared to the wild-
type MMP-3, lower promoter activation occurred in
the presence of catalytically dead MMP-3 mutants.
This suggested that MMP-3 can regulate the
CCN2 ⁄ CTGF promoter activity by two completely dif-
ferent mechanisms. One involves proteolytic processing
of one or several nuclear proteins, whereas the other is
independent of the processing capacity of the protein-
ase and involves the HPX domain. A DNA-binding
domain was found in the HPX domain, as an anti-
MMP-3 HPX antibody blocked the protein-DNA
interactions. The hinge region contains proline-rich
sequences found in some transcription factors. The
properties of MMP-3 as a transcription factor was
evaluated by analyzing nuclear MMP-3 associated pro-
teins (NuMAPs). Several NuMAPs were detected, such
as heterochromatin proteins, transcription co-activators ⁄
corepressors, RNA polymerase II and nucleosome ⁄
chromatin assembly protein. One of the NuMAPs,
HP1c, was demonstrated to interact with MMP-3 and to
co-activate the CCN2 ⁄ CTGF promoter with MMP-3.
Another identified NuMAP was the transcription
repressor NCoR1, suggesting that MMP-3 might
degrade NCoR1 to prevent transcription repression of
the CCN2 ⁄ CTGF promoter [130].
Cytosolic and vesicle localization
A study on dopaminergic neurons suggested a pro-
apoptotic role of active intracellular MMP-3. During
apoptosis, the proform of MMP-3 was cleaved to a

contractile function during the post-ischemic reperfu-
sion phase [11]. In vitro and in vivo studies suggested
that this was a result of MMP cleavage of the contrac-
tile protein regulatory element, TnI, and the cytoskele-
tal protein a-actinin [11,135,136]. Other possible
MMP-2 substrates in cardiac myocytes are desmin and
myosin light chain-1 [136,137]. The most probable
mode of MMP-2 activation inside cardiac myocytes
undergoing ischemia-reperfusion injuries is via peroxy-
nitrite.
Unlike the other members of the MMP family, and
despite the presence of the N-terminal signal peptide,
most of the MMP-26 (matrilysin-2 ⁄ endometase) pro-
duced is reported to be retained inside the cell
[138,139]. The conserved PRCGXPD motif in the
prodomain involved in the latency of other MMPs is
replaced by the unique PH
81
CGVPD motif in MMP-
26. This motif, along with other atypical structures, is
assumed to facilitate autocatalytical activation of the
enzyme inside the cell [140]. Furthermore, it has been
reported that MMP-26 has one high-affinity and one
low-affinity calcium-binding site [141]. Normal intra-
cellular calcium-levels probably maintain MMP-26 in
an inactive state and the active enzyme may only be
seen during transient intracellular calcium influx. An
increased level of MMP-26 in breast cancer has been
found to correlate with longer patient survival [142].
This positive effect of intracellular MMP-26 is

chondrocytes from resting zones, indicating that theses
enzymes are involved in ECM remodeling at the
hypertrophic cell zone in the growth plates of long
bones [144,145].
Ovarian carcinoma ascites-derived membrane vesi-
cles have been shown to contain both pro- and active
forms of MMP-2 and MMP-9, active urokinase-like
plasminogen activator, MT1-MMP and urokinase-like
plasminogen activator receptor. Ascites from patients
with late stage cancers had higher vesicle content and
contained more active enzymes than ascites from
patients with non-malignant lesions or early stage can-
cer. Purified ascites vesicles were found to stimulate
the invasion of cultured ovarian cancer cells through
matrigel, and this invasion was markedly inhibited by
the addition of either MMPI or serine proteinase
inhibitors [146]. Furthermore, fibrosarcoma cells are
also shown to shed membrane vesicles containing both
pro- and active forms of MMP-2 and MMP-9, as well
as urokinase plasminogen activator [147].
In oral carcinoma cells, both pro- and processed
forms of MMP-9 have been found in cytoplasmic
vesicular structures often co-compartmentalized with
trypsin-2, an activator of proMMP-9. In addition, the
same carcinoma cells expressed enterokinase, which is
an activator of trypsinogen, the zymogen form of tryp-
sin-2. This suggests the existence of an intracellular
cascade where enterokinase can activate trypsin-2,
which may further activate proMMP-9. The intracellu-
larly activated MMP-9 had a slightly higher molecular

interactions may concentrate the enzymes close to or
on target substrates, and can also affect the activation
of inactive proenzymes. The binding of MMPs to
other macromolecules may also regulate the activity
of the enzymes either through stabilization or through
induction of autodegradation. Enzyme activity may be
regulated through binding partner interactions that
includes noncatalytic or exosites, and thereby inhibit
or prevent the processing of a specific substrate.
MMPs are involved in a large number of physiologi-
cal and pathological processes. An MMP may prevent
disease by processing one substrate, whereas the
same enzyme may promote disease by process-
ing another substrate. More research is needed to
increase the understanding of the localization of
MMPs, in vivo binding partners, substrate processing,
involvement of exosites in substrate processing and
the regulation of enzyme activity by binding partners.
On the basis of such information, new specific MMPI
targets for novel drugs may be discovered that are
both MMP- and substrate-specific. Hopefully, such
specific MMPIs can be used in therapy against rele-
vant diseases and result in less side effects compared
to the present MMPIs.
Acknowledgements
The Northern Norwegian Regional Health Authorities
are acknowledged for their support of Bodil Fadnes.
We apologize to those authors whose work could not
be cited as a result of space limitations.
References

try 48, 10830–10845.
8 Sela-Passwell N, Rosenblum G, Shoham T & Sagi I
(2010) Structural and functional bases for allosteric
control of MMP activities: can it pave the path for
selective inhibition? Biochim Biophys Acta 1803, 29–38.
9 Overall CM (2002) Molecular determinants of metallo-
proteinase substrate specificity: matrix metalloprotein-
ase substrate binding domains, modules, and exosites.
Mol Biotechnol 22, 51–86.
10 Overall CM, McQuibban GA & Clark-Lewis I (2002)
Discovery of chemokine substrates for matrix metallo-
proteinases by exosite scanning: a new tool for
degradomics. Biol Chem 383, 1059–1066.
11 Wang W, Schulze CJ, Suarez-Pinzon WL, Dyck JR,
Sawicki G & Schulz R (2002) Intracellular action of
matrix metalloproteinase-2 accounts for acute myocar-
dial ischemia and reperfusion injury. Circulation 106,
1543–1549.
12 Starckx S, Van den Steen PE, Verbeek R, van Noort
JM & Opdenakker G (2003) A novel rationale for
inhibition of gelatinase B in multiple sclerosis: MMP-9
destroys alpha B-crystallin and generates a promiscu-
ous T cell epitope. J Neuroimmunol 141, 47–57.
13 Descamps FJ, Martens E, Proost P, Starckx S, Van
den Steen PE, Van Damme J & Opdenakker G (2005)
Gelatinase B ⁄ matrix metalloproteinase-9 provokes cat-
aract by cleaving lens betaB 1 crystallin. FASEB J 19,
29–35.
14 Murphy G & Nagase H (2010) Localising MMP
activities in the pericellular environment. FEBS J 278 ,

20 Das S, Mandal M, Mandal A, Chakraborti T &
Chakraborti S (2004) Identification, purification and
characterization of matrix metalloproteinase-2 in
bovine pulmonary artery smooth muscle plasma
membrane. Mol Cell Biochem 258, 73–89.
21 Lefebvre V, Peeters-Joris C & Vaes G (1991) Produc-
tion of gelatin-degrading matrix metalloproteinases
(‘type IV collagenases’) and inhibitors by articular
chondrocytes during their dedifferentiation by serial
subcultures and under stimulation by interleukin-1 and
tumor necrosis factor alpha. Biochim Biophys Acta
1094, 8–18.
22 Lindstad RI, Sylte I, Mikalsen SO, Seglen PO, Berg E
& Winberg JO (2005) Pancreatic trypsin activates
human promatrix metalloproteinase-2. J Mol Biol 350,
682–698.
23 Sorsa T, Salo T, Koivunen E, Tyynela J, Konttinen
YT, Bergmann U, Tuuttila A, Niemi E, Teronen O,
Heikkila P et al. (1997) Activation of type IV
procollagenases by human tumor-associated trypsin-2.
J Biol Chem 272, 21067–21074.
24 Bannikov GA, Karelina TV, Collier IE, Marmer BL &
Goldberg GI (2002) Substrate binding of gelatinase B
induces its enzymatic activity in the presence of intact
propeptide. J Biol Chem 277, 16022–16027.
25 Freise C, Erben U, Muche M, Farndale R, Zeitz M,
Somasundaram R & Ruehl M (2009) The alpha 2
chain of collagen type VI sequesters latent proforms of
matrix-metalloproteinases and modulates their activa-
tion and activity. Matrix Biol 28, 480–489.

34 Okamoto T, Akaike T, Sawa T, Miyamoto Y, van
derVliet A & Maeda H (2001) Activation of matrix
metalloproteinases by peroxynitrite-induced protein
S-glutathiolation via disulfide S-oxide formation.
J Biol Chem 276, 29596–29602.
35 Viappiani S, Nicolescu AC, Holt A, Sawicki G,
Crawford BD, Leon H, van Mulligen T & Schulz R
(2009) Activation and modulation of 72kDa matrix
metalloproteinase-2 by peroxynitrite and glutathione.
Biochem Pharmacol 77, 826–834.
36 Migita K, Maeda Y, Abiru S, Komori A,
Yokoyama T, Takii Y, Nakamura M, Yatsuhashi H,
Eguchi K & Ishibashi H (2005) Peroxynitrite-mediated
matrix metalloproteinase-2 activation in human hepa-
tic stellate cells. FEBS Lett 579, 3119–3125.
37 Rajagopalan S, Meng XP, Ramasamy S, Harrison DG
& Galis ZS (1996) Reactive oxygen species produced
by macrophage-derived foam cells regulate the activity
of vascular matrix metalloproteinases in vitro. Implica-
tions for atherosclerotic plaque stability. J Clin Invest
98, 2572–2579.
38 Frears ER, Zhang Z, Blake DR, O’Connell JP &
Winyard PG (1996) Inactivation of tissue inhibitor of
Regulation of MMP activity E. Hadler-Olsen et al.
40 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
metalloproteinase-1 by peroxynitrite. FEBS Lett 381,
21–24.
39 McCarthy SM, Bove PF, Matthews DE, Akaike T &
van derVliet A (2008) Nitric oxide regulation of
MMP-9 activation and its relationship to modifications

surface chondroitin sulfate glycosaminoglycan in
melanoma: role in the activation of pro-MMP-2
(pro-gelatinase A). Biochem J 403, 553–563.
47 Morrison CJ, Butler GS, Bigg HF, Roberts CR,
Soloway PD & Overall CM (2001) Cellular activation
of MMP-2 (gelatinase A) by MT2-MMP occurs via a
TIMP-2-independent pathway. J Biol Chem 276,
47402–47410.
48 Miyamori H, Takino T, Kobayashi Y, Tokai H, Itoh
Y, Seiki M & Sato H (2001) Claudin promotes activa-
tion of pro-matrix metalloproteinase-2 mediated by
membrane-type matrix metalloproteinases. J Biol
Chem 276, 28204–28211.
49 Brooks PC, Silletti S, von Schalscha TL, Friedlander
M & Cheresh DA (1998) Disruption of angiogenesis
by PEX, a noncatalytic metalloproteinase fragment
with integrin binding activity. Cell 92, 391–400.
50 Brooks PC, Stromblad S, Sanders LC, von Schalscha
TL, Aimes RT, Stetler-Stevenson WG, Quigley JP &
Cheresh DA (1996) Localization of matrix metallopro-
teinase MMP-2 to the surface of invasive cells by inter-
action with integrin alpha v beta 3. Cell 85, 683–693.
51 Deryugina EI, Bourdon MA, Jungwirth K, Smith JW
& Strongin AY (2000) Functional activation of inte-
grin alpha V beta 3 in tumor cells expressing mem-
brane-type 1 matrix metalloproteinase. Int J Cancer
86, 15–23.
52 Deryugina EI, Ratnikov B, Monosov E, Postnova TI,
DiScipio R, Smith JW & Strongin AY (2001) MT1-
MMP initiates activation of pro-MMP-2 and integrin

Hashimoto G, Nemori R & Okada Y (2005) Pericellu-
lar activation of proMMP-7 (promatrilysin-1) through
interaction with CD151. Lab Invest 85, 1489–1506.
59 Goldberg GI, Strongin A, Collier IE, Genrich LT &
Marmer BL (1992) Interaction of 92-kDa type IV
collagenase with the tissue inhibitor of metalloprotein-
ases prevents dimerization, complex formation with
interstitial collagenase, and activation of the proen-
zyme with stromelysin. J Biol Chem 267, 4583–4591.
60 Hibbs MS, Hasty KA, Seyer JM, Kang AH &
Mainardi CL (1985) Biochemical and immunological
characterization of the secreted forms of human
neutrophil gelatinase. J Biol Chem 260, 2493–2500.
61 Olson MW, Bernardo MM, Pietila M, Gervasi DC,
Toth M, Kotra LP, Massova I, Mobashery S &
E. Hadler-Olsen et al. Regulation of MMP activity
FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS 41
Fridman R (2000) Characterization of the Monomeric
and Dimeric Forms of Latent and Active Matrix Me-
talloproteinase-9. Differential rates for activation by
stromelysin 1. J Biol Chem 275, 2661–2668.
62 Triebel S, Blaser J, Reinke H, Knauper V & Tschesche
H (1992) Mercurial activation of human PMN leuco-
cyte type IV procollagenase (gelatinase). FEBS Lett
298, 280–284.
63 Winberg JO, Kolset SO, Berg E & Uhlin-Hansen L
(2000) Macrophages secrete matrix metalloproteinase 9
covalently linked to the core protein of chondroitin
sulphate proteoglycans. J Mol Biol 304, 669–680.
64 Winberg JO, Berg E, Kolset SO & Uhlin-Hansen L

type IV collagenase: functional analysis of fibronectin
and carboxyl-end domains. Kidney Int 43, 158–162.
71 Cha H, Kopetzki E, Huber R, Lanzendorfer M &
Brandstetter H (2002) Structural basis of the adaptive
molecular recognition by MMP9. J Mol Biol 320,
1065–1079.
72 Kolkenbrock H, Hecker-Kia A, Orgel D, Kinawi A &
Ulbrich N (1996) Progelatinase B forms from human
neutrophils. complex formation of monomer ⁄ lipocalin
with TIMP-1. Biol Chem 377, 529–533.
73 Yan L, Borregaard N, Kjeldsen L & Moses MA
(2001) The high molecular weight urinary matrix
metalloproteinase (MMP) activity is a complex of gela-
tinase B ⁄ MMP-9 and neutrophil gelatinase-associated
lipocalin (NGAL). Modulation of MMP-9 activity by
NGAL. J Biol Chem 276, 37258–37265.
74 Ray S, Lukyanov P & Ochieng J (2003) Members of
the cystatin superfamily interact with MMP-9 and pro-
tect it from autolytic degradation without affecting its
gelatinolytic activities. Biochim Biophys Acta 1652, 91–
102.
75 Allan JA, Docherty AJ, Barker PJ, Huskisson NS,
Reynolds JJ & Murphy G (1995) Binding of gelatinas-
es A and B to type-I collagen and other matrix compo-
nents. Biochem J 309, 299–306.
76 Collier IE, Krasnov PA, Strongin AY, Birkedal-Han-
sen H & Goldberg GI (1992) Alanine scanning muta-
genesis and functional analysis of the fibronectin-like
collagen-binding domain from human 92-kDa type IV
collagenase. J Biol Chem 267, 6776–6781.

teinases-9 and -2. Biochem J 392, 127–134.
84 Xu X, Wang Y, Lauer-Fields JL, Fields GB & Steffen-
sen B (2004) Contributions of the MMP-2 collagen
binding domain to gelatin cleavage. Substrate binding
via the collagen binding domain is required for hydro-
lysis of gelatin but not short peptides. Matrix Biol 23,
171–181.
Regulation of MMP activity E. Hadler-Olsen et al.
42 FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS
85 Malla N, Berg E, Uhlin-Hansen L & Winberg JO
(2008) Interaction of pro-matrix metalloproteinase-
9 ⁄ proteoglycan heteromer with gelatin and collagen.
J Biol Chem 283, 13652–13665.
86 Rosenblum G, Van denSteen PE, Cohen SR, Gross-
mann JG, Frenkel J, Sertchook R, Slack N, Strange
RW, Opdenakker G & Sagi I (2007) Insights into the
structure and domain flexibility of full-length pro-
matrix metalloproteinase-9 ⁄ gelatinase B. Structure 15,
1227–1236.
87 Polticelli F, Bocedi A, Minervini G & Ascenzi P
(2008) Human haptoglobin structure and function – a
molecular modelling study. FEBS J 275, 5648–5656.
88 Bannikov GA, Mattoon JS, Abrahamsen EJ, Preman-
andan C, Green-Church KB, Marsh AE & Lakritz J
(2007) Biochemical and enzymatic characterization of
purified covalent complexes of matrix metalloprotein-
ase-9 and haptoglobin released by bovine granulocytes
in vitro. Am J Vet Res 68, 995–1004.
89 Williams RJ (2003) Restriction endonucleases: classifi-
cation, properties, and applications. Mol Biotechnol

peptide bond hydrolysis. EMBO J 23, 3020–3030.
97 Overall CM & Kleifeld O (2006) Towards third
generation matrix metalloproteinase inhibitors for
cancer therapy. Br J Cancer 94, 941–946.
98 Overall CM & Kleifeld O (2006) Tumour microenvi-
ronment - opinion: validating matrix metalloproteinas-
es as drug targets and anti-targets for cancer therapy.
Nat Rev Cancer 6, 227–239.
99 Hu J, Van denSteen PE, Sang QX & Opdenakker G
(2007) Matrix metalloproteinase inhibitors as therapy
for inflammatory and vascular diseases. Nat Rev Drug
Discov 6, 480–498.
100 Minond D, Lauer-Fields JL, Cudic M, Overall CM,
Pei D, Brew K, Visse R, Nagase H & Fields GB
(2006) The roles of substrate thermal stability and P2
and P1¢ subsite identity on matrix metalloproteinase
triple-helical peptidase activity and collagen specificity.
J Biol Chem 281, 38302–38313.
101 Lauer-Fields JL, Minond D, Chase PS, Baillargeon
PE, Saldanha SA, Stawikowska R, Hodder P &
Fields GB (2009) High throughput screening of
potentially selective MMP-13 exosite inhibitors
utilizing a triple-helical FRET substrate. Bioorg Med
Chem
17, 990–1005.
102 Lauer-Fields J, Brew K, Whitehead JK, Li S,
Hammer RP & Fields GB (2007) Triple-helical
transition state analogues: a new class of selective
matrix metalloproteinase inhibitors. J Am Chem Soc
129, 10408–10417.

E. Hadler-Olsen et al. Regulation of MMP activity
FEBS Journal 278 (2011) 28–45 ª 2010 The Authors Journal compilation ª 2010 FEBS 43
109 Heim-Riether A, Taylor SJ, Liang S, Gao DA, Xiong
Z, Michael August E, Collins BK, Farmer BT II,
Haverty K, Hill-Drzewi M et al. (2009) Improving
potency and selectivity of a new class of non-Zn-che-
lating MMP-13 inhibitors. Bioorg Med Chem Lett 19,
5321–5324.
110 Butler GS, Butler MJ, Atkinson SJ, Will H, Tamura
T, van Westrum SS, Crabbe T, Clements J, d’Ortho
MP & Murphy G (1998) The TIMP2 membrane type
1 metalloproteinase ‘‘receptor’’ regulates the concentra-
tion and efficient activation of progelatinase A. A
kinetic study. J Biol Chem 273, 871–880.
111 Cha J, Pedersen MV & Auld DS (1996) Metal and pH
dependence of heptapeptide catalysis by human
matrilysin. Biochemistry 35, 15831–15838.
112 Crabbe T, O’Connell JP, Smith BJ & Docherty AJ
(1994) Reciprocated matrix metalloproteinase activa-
tion: a process performed by interstitial collagenase
and progelatinase A. Biochemistry 33, 14419–14425.
113 Gadher SJ, Eyre DR, Wotton SF, Schmid TM &
Woolley DE (1990) Degradation of cartilage collagens
type II, IX, X and XI by enzymes derived from human
articular chondrocytes. Matrix 10, 154–163.
114 Kao WW, Ebert J, Kao CW, Covington H & Cintron
C (1986) Development of monoclonal antibodies rec-
ognizing collagenase from rabbit PMN; the presence
of this enzyme in ulcerating corneas. Curr Eye Res 5,
801–815.

proteinase-9 associated with heparan sulphate chains
of GPI-anchored cell surface proteoglycans mediates
motility of murine colon adenocarcinoma cells. J Bio-
chem 143, 581–592.
121 Lambert E, Bridoux L, Devy J, Dasse E, Sowa ML,
Duca L, Hornebeck W, Martiny L & Petitfrere-Char-
pentier E (2009) TIMP-1 binding to proMMP-9 ⁄ CD44
complex localized at the cell surface promotes
erythroid cell survival. Int J Biochem Cell Biol 41,
1102–1115.
122 Mantuano E, Inoue G, Li X, Takahashi K, Gaultier A,
Gonias SL & Campana WM (2008) The hemopexin
domain of matrix metalloproteinase-9 activates cell
signaling and promotes migration of schwann cells by
binding to low-density lipoprotein receptor-related
protein. J Neurosci 28, 11571–11582.
123 Radjabi AR, Sawada K, Jagadeeswaran S, Eichbichler
A, Kenny HA, Montag A, Bruno K & Lengyel E
(2008) Thrombin induces tumor invasion through the
induction and association of matrix metalloproteinase-
9 and beta1-integrin on the cell surface. J Biol Chem
283, 2822–2834.
124 Kwan JA, Schulze CJ, Wang W, Leon H,
Sariahmetoglu M, Sung M, Sawicka J, Sims DE,
Sawicki G & Schulz R (2004) Matrix metalloprotein-
ase-2 (MMP-2) is present in the nucleus of cardiac
myocytes and is capable of cleaving poly (ADP-ribose)
polymerase (PARP) in vitro. FASEB J 18, 690–
692.
125 Si-Tayeb K, Monvoisin A, Mazzocco C, Lepreux S,

ing the CTGF ⁄ CCN2 gene. Mol Cell Biol 28, 2391–
2413.
131 Choi DH, Kim EM, Son HJ, Joh TH, Kim YS,
Kim D, Flint Beal M & Hwang O (2008) A novel
intracellular role of matrix metalloproteinase-3 during
apoptosis of dopaminergic cells. J Neurochem 106,
405–415.
132 Limb GA, Matter K, Murphy G, Cambrey AD,
Bishop PN, Morris GE & Khaw PT (2005) Matrix me-
talloproteinase-1 associates with intracellular organ-
elles and confers resistance to lamin A ⁄ C degradation
during apoptosis. Am J Pathol 166, 1555–1563.
133 Boire A, Covic L, Agarwal A, Jacques S, Sherifi S &
Kuliopulos A (2005) PAR1 is a matrix metallopro-
tease-1 receptor that promotes invasion and tumori-
genesis of breast cancer cells. Cell 120, 303–313.
134 Yang E, Boire A, Agarwal A, Nguyen N, O’Callaghan
K, Tu P, Kuliopulos A & Covic L (2009) Blockade of
PAR1 signaling with cell-penetrating pepducins inhib-
its Akt survival pathways in breast cancer cells and
suppresses tumor survival and metastasis. Cancer Res
69, 6223–6231.
135 Bolli R & Marban E (1999) Molecular and cellular
mechanisms of myocardial stunning. Physiol Rev 79,
609–634.
136 Sung MM, Schulz CG, Wang W, Sawicki G, Bautista-
Lopez NL & Schulz R (2007) Matrix metalloprotein-
ase-2 degrades the cytoskeletal protein alpha-actinin in
peroxynitrite mediated myocardial injury. J Mol Cell
Cardiol 43, 429–436.

Kennedy S, Duffy MJ, Rozanov DV, Krajewski S &
Strongin AY (2006) Matrix metalloproteinase 26 pro-
teolysis of the NH2-terminal domain of the estrogen
receptor beta correlates with the survival of breast can-
cer patients. Cancer Res 66, 2716–2724.
143 Taraboletti G, D’Ascenzo S, Borsotti P, Giavazzi R,
Pavan A & Dolo V (2002) Shedding of the matrix
metalloproteinases MMP-2, MMP-9, and MT1-MMP
as membrane vesicle-associated components by endo-
thelial cells. Am J Pathol 160, 673–680.
144 Dean DD, Schwartz Z, Schmitz J, Muniz OE, Lu Y,
Calderon F, Howell DS & Boyan BD (1996) Vitamin
D regulation of metalloproteinase activity in matrix
vesicles. Connect Tissue Res 35, 331–336.
145 Schmitz JP, Dean DD, Schwartz Z, Cochran DL,
Grant GM, Klebe RJ, Nakaya H & Boyan BD (1996)
Chondrocyte cultures express matrix metalloproteinase
mRNA and immunoreactive protein; stromelysin-1
and 72 kDa gelatinase are localized in extracellular
matrix vesicles. J Cell Biochem 61, 375–391.
146 Graves LE, Ariztia EV, Navari JR, Matzel HJ, Stack
MS & Fishman DA (2004) Proinvasive properties of
ovarian cancer ascites-derived membrane vesicles.
Cancer Res 64, 7045–7049.
147 Ginestra A, Monea S, Seghezzi G, Dolo V, Nagase H,
Mignatti P & Vittorelli ML (1997) Urokinase plasmin-
ogen activator and gelatinases are associated with
membrane vesicles shed by human HT1080 fibrosar-
coma cells. J Biol Chem 272, 17216–17222.
148 Vilen ST, Nyberg P, Hukkanen M, Sutinen M,