Content uploaded by Kenneth M Yamada
Author content
All content in this area was uploaded by Kenneth M Yamada on Jul 18, 2015
Content may be subject to copyright.
INTRODUCTION
Cell-matrix interactions affect a diverse range of cellular
functions including cell differentiation, migration, proliferation
and survival. Information provided by the extracellular matrix
(ECM) can control processes of embryonic growth and
differentiation, tissue remodelling and repair. It follows that
localised proteolytic degradation of the ECM, and the release
of matrix fragments and matrix bound growth factors and
cytokines will have profound effect on signals reaching the cell
from the ECM (for review see Werb, 1997).
The matrix metalloproteinases (MMPs) have been
implicated in normal matrix remodelling events such as
mammary gland involution (Werb et al., 1996), and in
pathological conditions, including tumour invasion and
metastasis (Chambers and Matrisian, 1997). MMP-2
(gelatinase A) is associated with tumour tissues, and the
appearance of active MMP-2 is closely correlated with tumour
metastasis (Sato and Seiki, 1996). In common with all the
MMPs, MMP-2 is synthesised as a latent proenzyme, requiring
proteolytic removal of the propeptide for activation. The
physiological mechanism by which proMMP-2 is activated is
2789
Journal of Cell Science 111, 2789-2798 (1998)
Printed in Great Britain © The Company of Biologists Limited 1998
JCS4598
We have assessed the effect of fibronectin and laminin-1 on
the expression of molecules involved in the activation
pathway of MMP-2, a key proteinase in tissue remodelling.
HT1080 fibrosarcoma cells cultured on fibronectin were
shown to activate endogenous MMP-2, to a level
comparable with that elicited by treatment with phorbol
ester. In contrast, the MMP-2 expressed by HT1080 cells
cultured on laminin-1 was mainly in the pro- (inactive
form). Culture of the cells on peptide fragments of
fibronectin derived from the central cell binding domain
also promoted MMP-2 activation, indicating that signals
via fibronectin binding to integrin receptors may be
involved. HT1080 cells cultured on immobilised antibodies
to the α5 and β1 integrin subunits secreted levels of active
MMP-2 similar to those observed for full length
fibronectin, whereas cells cultured on an antibody to the α6
integrin subunit secreted mainly proMMP-2. The data
demonstrate that the activation of MMP-2 by HT1080 cells
is regulated by the nature of the extracellular matrix, and
that signals via the α5β1 integrin receptor may be involved
in the fibronectin induced up-regulation of MMP-2
activation.
We then assessed the effect of fibronectin on the
components of the putative MT1-MMP/TIMP-2 ‘receptor’
complex implicated in MMP-2 activation. Levels of TIMP-
2 protein expressed by HT1080 cells did not vary detectably
between cells cultured on fibronectin or laminin-1.
However, the expression of MT1-MMP protein was up-
regulated when the cells were cultured on fibronectin,
which could be attributed to an increase in levels of a
truncated 45 kDa form. Parallel studies using gelatin
zymography demonstrated that the up-regulation of the
production of the 45 kDa band was concomitant with
MMP-2 activation. Inhibitor studies revealed that the
truncation of MT1-MMP to a 45 kDa form is MMP
mediated, although not inhibited by TIMP-1. In vitro, the
45 kDa form could be generated by cleavage of membrane-
bound native MT1-MMP with several recombinant MMPs,
including both active MT1-MMP and MMP-2. The
implication that either MMP-2 or MT1-MMP can process
MT1-MMP to 45 kDa, raises the possibility that truncation
of MT1-MMP represents a self-regulatory end-point in the
activation pathway of MMP-2.
Key words: Fibronectin, MMP-2 activation, MT1-MMP, α5β1
integrin
SUMMARY
The activation of ProMMP-2 (gelatinase A) by HT1080 fibrosarcoma cells is
promoted by culture on a fibronectin substrate and is concomitant with an
increase in processing of MT1-MMP (MMP-14) to a 45 kDa form
Heather Stanton1, Jelena Gavrilovic1, Susan J. Atkinson1, Marie-Pia d’Ortho2, Kenneth M. Yamada3,
Luciano Zardi4and Gillian Murphy1
1School of Biological Sciences, University of East Anglia, Norwich NR4 7TJ, UK
2INSERM U492, Faculté de Médecine de Créteil, rue du Général Sarrail, 94010 Créteil, France
3Craniofacial Developmental Biology and Regeneration Branch, National Institute of Dental Research, National Institutes of
Health, Bethesda, Maryland 20892, USA
4Laboratory of Cell Biology, Instituto Nazionale per la Ricerca sul Cancro, Viale Benedetto, XV 10 16132 Genoa, Italy
Author for correspondence (e-mail: g.murphy@uea.ac.uk)
Accepted 14 July; published on WWW 27 August 1998
2790
under intense study, and recent evidence suggests that this
event is triggered by an interaction between proMMP-2 and
cell surface bound MT-MMPs (Sato et al., 1994; Strongin et
al., 1995; Butler et al., 1998). From this new subclass of
MMPs, MT1-MMP is the best characterised, and is
overexpressed in several tumour tissues in which activated
MMP-2 is found, including lung and colon (reviewed by Sato
and Seiki, 1996). Our laboratory has demonstrated that MT1-
MMP performs the first proteolytic ‘clip’ in the removal the
MMP-2 propeptide, leading to MMP-2 autolysis and the
production of fully active enzyme (Atkinson et al., 1995; Will
et al., 1996). As MT1-MMP is an ECM degrading enzyme in
its own right (Pei and Weiss, 1996; d’Ortho et al., 1997; Ohuchi
et al., 1997) active MT1-MMP and active MMP-2 at the cell
surface provide a powerful combination for the localised
remodelling of the ECM (d’Ortho et al., 1998).
Although MT1-MMP mediated activation of proMMP-2 can
be stimulated in vitro by concanavalin A or by phorbol esters,
little is known of the factors that influence this process in vivo.
ECM macromolecules influence cellular expression of MMPs,
a process which can be mediated via integrin receptors (Werb
et al., 1989; Larjava et al., 1993; Arner and Tortorella, 1995).
MMP-1 expression is elevated in response to collagen type I
culture substrates (Riikonen et al., 1995; Langholz et al., 1995).
Increases in MMP-2 activation were reported following culture
of several cell types on or within collagen type I gels (Azzam
and Thompson, 1992; Gilles et al., 1997; Haas et al., 1998).
Vitronectin receptor (αvβ3) ligation also stimulates MMP-2
expression, as demonstrated by Seftor et al. (1992) who
observed that treatment of a melanoma cell line with an anti-
αvβ3 integrin antibody increased MMP-2 expression and
stimulated the invasion of the cells through Matrigel.
The aim of this study was to investigate the regulation by
matrix macromolecules of both MMP-2 activation and MT1-
MMP expression, using HT1080 fibrosarcoma cells as a model.
We report here that fibronectin promotes proMMP-2 activation
by HT1080 cells. Culture of HT1080 cells on fibronectin, or
on fragments of fibronectin encompassing the RGD integrin
binding site, up-regulated processing of proMMP-2 to the
active form. In marked contrast, culture of these cells on
laminin-1 did not promote MMP-2 processing. The potential
for fibronectin integrin receptors to signal MMP-2 activation
was examined. We investigated the effects of the ECM
substrates fibronectin and laminin-1 on the expression of
TIMP-2 and MT1-MMP by HT1080 cells and found that the
truncation of MT1-MMP protein to a 45 kDa form increased
when the cells were cultured on fibronectin.
MATERIALS AND METHODS
Reagents
Human plasma fibronectin was purchased from Bioproducts,
Hertfordshire, UK, and murine laminin-1 was from Sigma Chemical
Company. Culture reagents were from Sigma or Gibco BRL with the
exception of fetal calf serum (FCS) which was obtained from
Globepharm, Surrey, UK. All chemicals were purchased from Sigma,
ICN or Pierce. Radiochemicals were purchased from Amersham Life
Sciences. The MMP inhibitor CT1746 (N1-(1-(S)-carbamoyl-2,2-
dimethylpropyl)-N4-hydroxy-2-(R)-[3-(4-chlorophenyl)-
propyl]succinamide) was a gift from Dr A. Docherty (Celltech
Research, Slough, UK). The 120 kDa and 110 kDa fibronectin
fragments were prepared by thermolysin digestion of human
fibronectin as described (Borsi et al., 1986). The fibronectin fragment
Fn III 6-10 was a gift from Dr S. Aota (National Institute of Dental
Research, National Institutes of Health, Bethesda, Maryland) and was
prepared as detailed by Danen et al. (1995). Recombinant human
proMMP-2 and proMMP-3 were purified from media conditioned by
the relevant transfected mouse myeloma cells: proMMP-2 (Murphy
et al., 1992b); proMMP-3 (Murphy et al., 1992a). Recombinant
proMMP-13 (Knäuper et al., 1996) was a kind gift from Dr V.
Knäuper (University of East Anglia, Norwich, UK). Recombinant
MT-MMP (∆502-559 ∆TM-MT1-MMP; d’Ortho et al., 1997) was
kindly provided by Dr H. Will, InVitek GmbH, Berlin-Buch,
Germany. Recombinant human tissue inhibitors of metalloproteinases
(TIMPs -1 and -2) were expressed in mouse myeloma cells (Murphy
et al., 1991; Willenbrock et al., 1993) and MMP inhibitory activity
was assayed as described by Murphy et al. (1981).
Preparation of extracellular matrix and antibody
substrates
Culture plates were coated with human plasma fibronectin, fragments
of fibronectin, or laminin-1 following the method of Tremble et al.
(1994). ECM was added to culture plates at 30 µg/ml in phosphate
buffered saline (PBS) and incubated overnight at 4°C. The solution was
then aspirated, the wells washed with PBS and blocked with 1% (w/v)
heat-denatured bovine serum albumin (BSA) in PBS for 1 hour at room
temperature. Wells were washed with PBS and used the same day. The
following monoclonal antibodies to integrin subunits were coated on
culture plastic at 100 µg/ml: anti-α5, mAbs 16 and 11 (Akiyama et al.,
1989; LaFlamme et al., 1992); anti-β1, mAb 13 (Akiyama et al., 1989)
and anti-α6, GoH3, (TCS Biologicals, Botolph Claydon,
Buckinghamshire, UK). The antibody-coated wells were washed with
PBS and blocked with BSA according to the protocol above.
Cell culture
The HT1080 human fibrosarcoma cell line was purchased from the
European Collection of Animal Cell Cultures, Wiltshire, UK. A
second HT1080 line that is N-ras transformed (Paterson et al., 1987)
was a gift from Dr C. Marshall, Institute of Cancer Research, London,
UK. MMP-2 activation by cells cultured on fibronectin was compared
in the two lines; both cell lines responded similarly. The N-ras
transformed cells were used routinely for this study as their
constitutive expression of the gelatinases was greater. HT1080 cells
were cultured in DMEM supplemented with 10% (v/v) fetal calf
serum (FCS), 2 mM glutamine, 100 units/ml penicillin and 100 µg/ml
streptomycin. A third HT1080 cell line, stably transfected with wild-
type MT1-MMP using the HCMV/gpt vector pGW1HG (Green et al.,
1994) was a gift from Dr J. Clements, British Biotech Pharmaceuticals
Ltd, Oxford, UK. These cells were used for the preparation of MT1-
MMP enriched cellular membranes and maintained in a selection
medium of DMEM supplemented with 10% FCS, 4 mM glutamine,
100 units/ml penicillin, 100 µg/ml streptomycin, HT supplement, 20
µM mycophenolic acid and 2 mM xanthine.
Preparation of conditioned media
Subconfluent cultures of HT1080 cells were trypsinised and
centrifuged, firstly in serum containing medium, followed by two
washes in serum free medium. Cells were seeded onto uncoated 24-
well tissue culture wells (Costar) or substrate-coated wells in DMEM
supplemented with 0.1% (w/v) BSA and cultured for 48 hours. In
some experiments, phorbol 12 myristate 13-acetate (PMA) was added
after the first 2 hours in culture, along with MMP inhibitors. At
harvest, cells were trypsinised and the number of cells per well
determined by counting the cells with a Neubauer haemocytometer.
TIMP-2 complexed with MMP-2 was purified from the conditioned
medium by adsorption to gelatin-Sepharose (Butler et al., 1998). Bound
material (about 10-100 ng) was eluted using Laemmli reducing sample
buffer and analysed for TIMP-2 by western blotting (see below).
H. Stanton and others
2791Fibronectin promotes processing of ProMMP-2 and MT1-MMP
Preparation of cell lysates
HT1080 cells were cultured on dishes coated with ECM or
monoclonal antibodies, or treated with PMA as described above. After
48 hours in culture, the conditioned medium was harvested for
analysis by zymography and the cell monolayers washed with cold
PBS before lysis. Cells were lysed as described by Lohi et al. (1996),
using a lysis buffer of 50 mM Tris-HCl, pH 8.0, containing 0.15 M
NaCl, 1% (v/v) Triton-X-100, 0.02% (w/v) azide, with the protease
inhibitors pepstatin A (1 µg/ml), phenylmethylsulfonyl fluoride
(PMSF; 100 µM), trans-epoxysuccinyl-l-leucylamido (4-guanidino)-
butane (E-64; 1 µg/ml), and EDTA (10 mM). The cell lysates were
cleared by centrifugation and total protein estimated using the
bicinchoninic acid assay (Sigma).
Zymography
Conditioned media from HT1080 cultures were analysed for gelatin
degrading activity by electrophoresis under non-reducing conditions
on SDS-polyacrylamide gels containing 0.5 mg/ml denatured type I
collagen (Heussen and Dowdle, 1980). The volume of conditioned
medium loaded per lane was standardised on the basis of the cell
counts obtained at harvest. Gels were incubated overnight at room
temperature in 100 mM Tris-HCl, pH 7.9, 30 mM CaCl2and 0.02%
(w/v) sodium azide. White zones of lysis indicating gelatin degrading
activity were revealed by staining with Coomassie brilliant blue.
Isolation of RNA and northern blot analysis
HT1080 cells were cultured in DMEM 0.1% (w/v) BSA on plastic or
ECM protein coated dishes. Cells were cultured for 24 hours prior to
lysis and extraction of total cellular RNA by the guanidinium
isothiocyanate method (Chomczynski and Sacchi, 1987). RNA samples
were separated on an agarose 2.2 M formaldehyde gel, transferred to a
Nylon membrane (Boehringer Mannheim) and probed with
digoxygenin-labeled riboprobes as described (Atkinson et al., 1995).
Western blot
Lysate proteins (25 µg protein per lane) or gelatin-Sepharose eluates
(10 µl) were separated by 10% SDS-polyacrylamide gel
electrophoresis. Proteins were transferred to nitrocellulose by
electroblotting and probed with either a polyclonal antibody to TIMP-
2 (Ward et al., 1991) or a polyclonal antibody to human MT1-MMP,
affinity purified as described (d’Ortho et al., 1998). Bound antibody
was detected using a peroxidase conjugated secondary antibody
followed by chemiluminescence detection.
Preparation of cell membranes
HT1080 cell membranes enriched in the 60 kDa form of MT1-MMP
were prepared from HT1080 cells stably transfected with wild-type
MT1-MMP (described above). Processing of MT1-MMP to the 45 kDa
form was prevented by culturing the cells in the presence of CT1746
inhibitor for 48 hours prior to harvest (Butler et al., 1998). To remove
excess CT1746 at harvest, the cells were washed with ice-cold DMEM
(containing inhibitors) with pH adjusted to 9.0. Cells were scraped,
centrifuged and the pellet washed twice with DMEM pH 9.0, using 25
ml solution/175cm2cell monolayer per wash. The membranes were
resuspended in buffer (20 mM Tris-HCl, pH 7.8, 10 mM CaCl2,
0.025% (v/v) Brij 35, 0.02% (w/v) sodium azide) containing the
protease inhibitors pepstatin A (1 µg/ml), E-64 (1 µg/ml) and PMSF
(100 µM). The amount of protein in the membrane preparation was
estimated by the bicinchoninic acid assay to be 2.6 mg/ml and the
amount of active MT1-MMP present in the membrane preparation was
estimated by a quenched fluorescent peptide assay, assuming a kcat/Km
value of 1.9×105M−1s−1(Butler et al., 1998).
MMP-2 cleavage of membrane bound MT1-MMP
Recombinant proMMPs-2 and -3 were activated essentially as
described by Murphy et al. (1991). ProMMP-13 was activated by
incubation with 2 mM APMA for 1 hour at 37°C and ∆TM-MT1-
MMP was activated as described (Butler et al., 1998). MT1-MMP
enriched cell membrane preparation (10 µl) was incubated alone or
with active MMPs for 4 hours at 37°C. The cleavage reaction was
stopped by the addition of Laemmli reducing sample buffer (Laemmli
and Favre, 1973). The molecular mass of the resulting MT1-MMP
fragments were estimated by SDS-polyacrylamide gel electrophoresis
followed by western blot analysis as described above.
RESULTS
A comparison of the effects of fibronectin and
laminin-1 substrates on HT1080 fibrosarcoma
processing of proMMP-2 to the active form
HT1080 cells cultured on plastic constitutively synthesised and
secreted MMP-2 and MMP-9 (gelatinase B) as detected by
gelatin zymography (Fig. 1A). The substrates plasma fibronectin
and laminin-1 both supported cell adhesion and spreading. Cells
cultured on laminin-1 expressed mainly latent, proMMP-2,
molecular mass 66 kDa (Fig. 1A, lane 2). In contrast, cells
cultured on fibronectin (Fig. 1A, lane 1) displayed an apparent
increase in processing of proMMP-2 from the 66 kDa latent
form, via an intermediate at 62 kDa, to the fully active 59 kDa
protein. This increase in MMP-2 processing induced by
fibronectin was similar to the levels of processing induced by
PMA (Fig. 1A, lane 3), a known stimulator of MMP-2 activation
for these cells (Lohi et al., 1996). Increased processing of
MMP-2 by cells cultured on fibronectin was also evident when
secreted proteins were labelled with [35S]methionine,
immunoprecipitated with an antibody to MMP-2 and analysed
by autoradiography (data not shown).
MMP-9 secretion was up-regulated by PMA treatment (Fig.
1A, lane 3); however, neither culture of the cells on fibronectin
nor on laminin-1 affected MMP-9 expression (Fig. 1A, lanes
1 and 2). Activated MMP-9 was not detected in this system;
activation of MMP-9 is known to proceed via different
mechanisms to MMP-2 (Murphy et al., 1992c).
Fig. 1. Increased processing of proMMP-2 to the active form by
HT1080 cells cultured on fibronectin in comparison with laminin-1.
HT1080 cells were cultured under serum free conditions on
substrates of fibronectin or laminin-1, or on culture plastic with or
without PMA for 48 hours. Samples of conditioned media were
analysed for gelatin degrading activity by zymography. Cells
cultured on fibronectin (lane 1), cells cultured on laminin-1 (lane 2);
cells cultured on plastic in the presence of PMA (lane 3) or in the
absence of PMA (lane 4). Arrows indicate the electrophoretic
mobility of recombinant proMMP-9 and recombinant pro- and active
MMP-2. Molecular mass markers are indicated on the right.
2792
As the activation of MMP-2 is a cell membrane-mediated
process, one way in which fibronectin might effect an up-
regulation of MMP-2 processing is by binding MMP-2 close
to the cell surface. Solid phase studies have demonstrated the
binding of a C-terminal domain fragment of MMP-2 to
fibronectin (Wallon and Overall, 1997); however, using the
ELISA method of Allan et al. (1995), we were unable to
demonstrate binding of the full length pro- and active MMP-2
to fibronectin (data not shown). It is unlikely, therefore, that
fibronectin acts to ‘trap’ MMP-2 close to the cell surface.
Ligands that promote cell attachment via the α5β1
integrin up-regulate proMMP-2 activation by HT1080
cells
To determine whether specific regions of fibronectin have the
potential to affect MMP-2 activation, defined peptide
fragments of fibronectin were coated to culture plastic. The 120
kDa chymotryptic fragment of fibronectin comprising the
fibronectin type III repeats 2-11 of the central cell binding
domain (CCBD; Fig. 2A) supported HT1080 adhesion and
spreading. MMP-2 activation on this fragment (Fig. 2Bi, lane
3)was equivalent to that observed for full length fibronectin
(Fig. 2Bi, lane 1). A second fragment of 110 kDa, which lacked
the alternatively spliced ED-B domain (Fig. 2A) also induced
MMP-2 activation (data not shown), indicating that regions in
the CCBD other than the ED-B domain effect changes in
MMP-2 activation.
The CCBD contains an RGD sequence in fibronectin type
III repeat 10, which is of key importance for cell attachment
via integrin receptors (for review, see Mohri, 1996). We
cultured HT1080 cells on fibronectin type III repeats 6-10 (Fn
III 6-10; Fig. 2A), which comprises the RGD sequence and
subregions in type III repeats 8 and 9 that act synergistically
with the RGD sequence for full adhesion activity (Danen et al.,
1995). Despite poor cell spreading on this fragment, the
activation of MMP-2 (Fig. 2Bii, lane 2) appeared greater than
that observed for cells cultured on a laminin-1 substrate, on
which the cells were well spread (Fig. 2Bii, lane 3). This
indicated to us that Fn III 6-10 may contain the information
capable of signalling MMP-2 activation.
We mediated direct interactions with the integrin receptors
that bind the CCBD, by using antibodies to integrin subunits
as substrates for cell adhesion. We confirmed that the HT1080
cell adhesion to fibronectin was completely inhibited by
mAb16, a monoclonal antibody to the α5 integrin subunit,
(results not shown; Yamada et al., 1990). Since mAb16
supports cell adhesion and spreading and mimics the action of
the fibronectin ligand (Akiyama et al., 1989) we chose to
culture HT1080 cells on this antibody. MMP-2 activation was
up-regulated when the cells were cultured on this antibody
(Fig. 2Biii, lane 2), with a processing profile comparable to that
when cells were cultured on fibronectin (Fig. 2Biii, lane 1). In
contrast, the monoclonal mAb 11 to the α5 integrin subunit,
which promotes receptor cross-linking but does not mimic
H. Stanton and others
B
A
Fig. 2. Processing of proMMP-2 to the
active form is up-regulated by fragments of
fibronectin comprising the α5β1 integrin
binding site and by a monoclonal antibody
to α5 integrin. (A) Schematic diagram of
intact plasma fibronectin and the fibronectin
fragments used in this study. (B) HT1080
cells were cultured under serum free
conditions for 48 hours on fibronectin or
fibronectin fragments, or on immobilised
antibodies to integrin subunits. Samples of
conditioned media were analysed for gelatin
degrading activity by zymography. (i) Cells
were cultured on fibronectin (lane 1), on
laminin-1 (lane 2) or on the 120 kDa
fragment (lane 3). (ii) Cells were cultured
on fibronectin (lane 1), on Fn III 6-10 (lane
2) or on laminin-1 (lane 3). (iii) Monoclonal
antibodies to integrin subunits were used as
substrates. Cells cultured on fibronectin
(lane 1), on the anti-α5 antibody mAb 16
(lane 2), on anti-β1 antibody mAb 13 (lane
3), on anti-α5 antibody mAb 11 (lane 4), or
on anti-α6 antibody GoH3 (lane 5). Cells
cultured on laminin-1 (lane 6). Molecular
mass markers are indicated on the right.
2793Fibronectin promotes processing of ProMMP-2 and MT1-MMP
ligand action (LaFlamme et al., 1992) did not up-regulate
MMP-2 processing when used as a substrate for HT1080
adhesion (Fig. 2Biii, lane 4). MAb 13 to the β1 integrin subunit
also promoted MMP-2 activation when it was used as a culture
substrate (Fig. 2Biii, lane 3).
HT1080 cell adhesion to laminin-1 is via the α6β1 integrin
and can be completely inhibited by the addition of a
monoclonal antibody to the α6 subunit, GoH3 (von der Mark
et al., 1991). As the processing of MMP-2 by HT1080 cells on
a laminin-1 substrate is poor (Fig. 1, lane 2), we speculated
that the α6 subunit is not involved in promoting MMP-2
processing. HT1080 cells were cultured on a substrate of GoH3
antibody, on which the cells spread well, but they failed to
process MMP-2 to the levels seen when fibronectin or mAb16
was used as a substrate (Fig. 2Biii, lane 5).
Culture of HT1080 cells on fibronectin or laminin-1
does not influence the expression of TIMP-2 protein
The activation of MMP-2 is exquisitely regulated by the levels
of TIMP-2 at the extracellular surface (Strongin et al., 1995;
Butler et al., 1998). We analysed the effect of culture substrate
on HT1080 TIMP-2 expression. TIMP-2 complexed to MMP-
2 in the conditioned medium was concentrated by gelatin-
Sepharose chromatography and analysed by western blot using
a polyclonal antibody to TIMP-2 (Ward et al., 1991). Although
the levels of TIMP-2 synthesised by HT1080 cells varied
between experiments, within experiments no difference was
noted in the levels of TIMP-2 secreted when cells were cultured
on fibronectin as compared with laminin-1 (Fig. 3A). Western
blot of cell lysates showed no obvious differences in the levels
of cell associated TIMP-2 between the substrates (Fig. 3B).
Messenger RNA levels for MT1-MMP expressed by
HT1080 cells are not altered by culture on
fibronectin or laminin-1 substrates
Fibronectin substrates were examined for their potential to
regulate MT1-MMP expression at the messenger RNA level.
RNA prepared from HT1080 cells was analysed by northern
blot and a single band of MT1-MMP was detected (Fig. 4),
consistent with the 4.5 kilobase transcript reported by Sato et
al. (1994). Culture of the cells on fibronectin or laminin-1 did
not alter the steady state levels of MT1-MMP mRNA (Fig. 4).
Culture of HT1080 cells on fibronectin up-regulates
the processing of MT1-MMP protein to a 45 kDa form
MT1-MMP protein in HT1080 cell lysates was analysed by
western blot using a polyclonal antibody to MT1-MMP
developed and affinity purified as described (d’Ortho et al.,
1998). This antibody detected major bands of MT1-MMP at
molecular mass 60 kDa and 45 kDa and a faint doublet at 63
kDa (Fig. 5A). Levels of the 60 kDa form of MT1-MMP were
unaffected by the different culture conditions used in this study.
However, culture of the cells on fibronectin increased the total
MT1-MMP protein level; more specifically the 45 kDa band was
up-regulated. By densitometric scanning the total MT1-MMP
protein produced by the cells cultured on fibronectin was 1.6
times that detectable from cells cultured on laminin.
Densitometric scanning also revealed that cells cultured on
fibronectin produced at least 3 times more 45 kDa MT1-MMP
than cells cultured on laminin. PMA also up-regulated the 45
kDa band. Strikingly, where increases in the 45 kDa band were
detected, there was a concomitant increase in the processing of
MMP-2 to the active form, as detected by zymography (Fig. 5B).
Processing of MT1-MMP to the 45 kDa form is
mediated by matrix metalloproteinase(s)
To investigate whether MMPs are involved in the processing
of MT1-MMP to 45 kDa, HT1080 cells were cultured on
fibronectin in the presence of inhibitors of metalloproteinase
Fig. 3. TIMP-2 protein expression by HT1080 cells cultured on
fibronectin or laminin-1. HT1080 cells were cultured under serum
free conditions on substrates of fibronectin or laminin-1 for 48 hours.
(A) Western blot analysis of secreted TIMP-2. TIMP-2/MMP-2
complexes were purified from the conditioned media by binding to
gelatin-Sepharose. Bound proteins were eluted with reducing sample
buffer and separated by 10% SDS-polyacrylamide gel electrophoresis.
Proteins were transferred to nitrocellulose by electroblotting and
probed with a polyclonal antibody to TIMP-2. Two separate
experiments are depicted. Experiment 1: cells cultured on fibronectin
(lane 1), cells cultured on laminin-1 (lane 2,). Experiment 2: cells
cultured on fibronectin (lane 3), cells cultured on laminin-1 (lane 4).
Recombinant TIMP-2 (lane 5). (B) Western blot analysis of HT1080
cell lysates. Cells were lysed in a buffer containing Triton X-100.
Lysate proteins were separated by 10% SDS-polyacrylamide gel
electrophoresis, transferred to nitrocellulose and probed with a
polyclonal antibody to TIMP-2. Two separate experiments are
depicted. Cells cultured on fibronectin (lanes 1, 3), cells cultured on
laminin-1 (lanes 2, 4). Recombinant TIMP-2 (lane 5).
Fig. 4. HT1080 fibrosarcoma mRNA levels for MT1-MMP are not
altered by culture on fibronectin or laminin-1. Northern blot of RNA
from HT1080 cells cultured on fibronectin, laminin-1 or on plastic.
Cells were cultured for 24 hours prior to extraction of total RNA.
Samples (5 µg/lane) were separated on agarose gels, transferred to a
Nylon membrane and MT1-MMP mRNA detected by hybridisation
with a digoxygenin-labelled riboprobe. The MT1-MMP probe
reveals the 28S ribosomal RNA (uppermost band of upper doublet).
The position of 18S ribosomal RNA is shown. A glyceraldehyde-3-
phosphate dehydrogenase (GAPDH) riboprobe was used as an
internal standard. Cells were cultured on plastic (lane 1) on laminin-1
(lane 2); or cultured on fibronectin (lane 3).
2794
activity for 46 hours prior to harvest for zymography and
western blot. The hydroxamate inhibitor CT1746 is a general
metalloproteinase inhibitor when used at the concentration
employed here. Inclusion of CT1746 in a culture of HT1080
cells on fibronectin completely abrogated the generation of the
45 kDa band (Fig. 6A, lane 2), indicating the requirement for
metalloproteinase activity in the processing of MT1-MMP to
45 kDa. Exogenously added TIMP-2 also inhibited this
processing of MT1-MMP, further defining the processing
activity as an MMP (Fig. 6A, lane 3). TIMP-1, even at high
concentrations (150 nM), was ineffective as an inhibitor of
processing of MT1-MMP to 45 kDa (Fig. 6A, lane 4). The
conditioned media from this experiment were examined by
zymography. Both CT1746 and TIMP-2 inhibited MMP-2
activation, but TIMP-1 had no effect (Fig. 6B).
To investigate which MMPs are capable of processing MT1-
MMP to the 45 kDa form, we used HT1080 cells stably
transfected with wild-type MT1-MMP to prepare cell
membranes. The membranes were enriched in the 60 kDa active
form of MT1-MMP by culturing the cells in the presence of
CT1746 prior to harvest (to inhibit processing to the 45 kDa
form); however, care was taken to wash away the CT1746 at
harvest (see Methods). The amount of active MT1-MMP
present in the membrane preparation was estimated at 23
pmol/mg of membrane protein by a quenched fluorescent
peptide assay. Based on this value, recombinant active MMPs
were added at a 1:1 molar ratio with native MT1-MMP, with
the exception of recombinant soluble ∆TM-MT1-MMP, which
was added at a 5:1 molar ratio to native MT1-MMP. Incubation
at 37°C for 4 hours resulted in the generation of the 45 kDa
form of MT1-MMP, with active preparations of MMP-2, and
∆TM-MT1-MMP capable of cleaving native MT1-MMP to the
45 kDa form (Fig. 7, lanes 3 and 4). In addition, active MMP-
3 and MMP-13 cleaved MT1-MMP to 45 kDa in vitro (Fig. 7,
lanes 5 and 6), indicating that MT1-MMP has a cleavage site
that is susceptible to several members of the MMP family. We
cannot detect MMP-3 or MMP-13 protein expression in
HT1080 cells (data not shown) and so it is unlikely that these
proteinases are involved in HT1080 processing of MT1-MMP
to 45 kDa. However, they may play a role in other cell systems,
particularly as MMP-13 is activated by membrane associated
MT1-MMP/MMP-2 (Knäuper et al., 1996). A second band of
processed MT1-MMP at 37 kDa was observed in all membrane
preparations incubated with active MMPs, but this band was not
detected in the lysates prepared directly from cells in culture.
Incubation of the membrane preparation alone also resulted in
the generation of the 45 kDa band, but to a lesser extent than
when active MMPs were added, demonstrating intrinsic MT1-
MMP processing activity in the membrane preparation.
DISCUSSION
Fibronectin matrices influence cellular functions including
adhesion, migration, and differentiation via interactions with
cell surface integrin receptors. These interactions are of
importance in early wound repair (Greiling and Clark, 1997)
and tumour development where several peptide and antibody
inhibitors of fibronectin/integrin function effectively inhibit
metastasis (Akiyama et al., 1995).
We have demonstrated in this report that fibronectin up-
regulates the activation of MMP-2 by HT1080 fibrosarcoma
cells, and that fibronectin induces a change in the levels of
H. Stanton and others
Fig. 5. The processing of MT1-MMP protein to a 45 kDa
form by HT1080 fibrosarcoma cells is increased by culture
on fibronectin. HT1080 cells were cultured under serum
free conditions on substrates of fibronectin or laminin-1, or
on culture plastic with or without PMA for 48 hours.
(A) Western blot analysis of HT1080 cell lysates. Cells
were lysed in a buffer containing Triton-X-100. Lysate
proteins were separated by 10% SDS-polyacrylamide gel
electrophoresis, transferred to nitrocellulose by
electroblotting and probed with a polyclonal antibody to
MT1-MMP. Cells cultured on fibronectin (lane 1), cells
cultured on laminin-1 (lane 2); cells cultured on plastic in
the presence of PMA (lane 3) or in the absence of PMA
(lane 4). (B) The conditioned media were analysed for gelatin degrading activity by zymography. Cells cultured on fibronectin (lane 1), cells
cultured on laminin-1 (lane 2); cells cultured on plastic in the presence of PMA (lane 3) or in the absence of PMA (lane 4). Molecular mass
markers are indicated on the right.
Fig. 6. Cellular processing of MT1-MMP to the 45 kDa form is
mediated by a matrix metalloproteinase. HT1080 cells were cultured
on fibronectin under serum free conditions, with or without protease
inhibitors for 48 hours. The conditioned medium was harvested for
zymography and the cells lysed for western blot. (A) Western blot
analysis. HT1080 cell lysates were prepared and subjected to 10%
SDS-polyacrylamide gel electrophoresis. Proteins were transferred to
nitrocellulose by electroblotting and probed with a polyclonal
antibody to MT1-MMP. Cells cultured on fibronectin (lanes 1-4),
alone (lane 1), or in the presence of metalloproteinase inhibitor
CT1746 at 1 µM (lane 2), or recombinant MMP inhibitors TIMP-2
and TIMP-1 at 150 nM (lanes 3 and 4, respectively). Cells cultured
on laminin-1 (lane 5). (B) The conditioned medium was analysed by
gelatin zymography. Lanes 1-5, see legend for A. Molecular mass
markers are indicated on the right.
B
A
2795Fibronectin promotes processing of ProMMP-2 and MT1-MMP
processed MMP-2 which is similar to that induced by phorbol
ester. We compared the up-regulation of MMP-2 activation in
response to fibronectin with activation on laminin-1.
Processing of MMP-2 by cells on laminin-1 appeared to be
quite low. Differences in MMP-2 activation were observed
without major changes in cell shape, because the cells spread
well on both fibronectin and laminin-1.
In our study, to induce an up-regulation of MMP-2
processing it was necessary to coat the fibronectin to culture
plastic; soluble fibronectin added to the culture medium did not
affect MMP-2 processing (data not shown). It has previously
been shown that cells do not respond to soluble fibronectin,
except at high concentrations, partly due to the need for a
conformational change induced by substrate adsorption that
opens up the molecule to expose the cell adhesive domain
(Yamada and Kennedy, 1984; Fukai et al., 1995). Reich et al.
(1995) reported that soluble laminin added to HT1080 cultures
for 6 hours increased MMP-2 mRNA and protein expression,
but no comment was made as to whether MMP-2 activation
was affected. In our study, soluble laminin-1 added to cultures
of HT1080 for 24 or 48 hours did not alter MMP-2 protein
expression (data not shown), or MMP-2 activation.
To determine whether specific domains of fibronectin signal
MMP-2 activation, we cultured HT1080 cells on polypeptide
fragments of fibronectin. We noted that the processing of
MMP-2 on the 120 kDa fragment of fibronectin was similar to
that observed for full length fibronectin. Werb et al. (1989)
reported that rabbit synovial fibroblasts respond to the 120 kDa
fragment, but not to full length fibronectin, up-regulating the
synthesis of MMP-1 -3 and -9. Our data differs from that study
on several points. Firstly, in our study, full length fibronectin
as well as the 120 kDa fragment signalled changes in MMP-2
activation; secondly, HT1080 synthesis of MMP-9 did not
appear to be affected by fibronectin substrates and thirdly,
although rabbit synovial fibroblasts express MMP-2, there was
no activation of MMP-2 apparent when these cells were
cultured on fibronectin (Werb et al., 1989). In a later study with
rabbit synovial fibroblasts, Huhtala et al. (1995) demonstrated
that the reason full-length fibronectin did not signal MMP
synthesis was due to opposing signals from α5β1 integrin
binding to the CCBD and α4β1 integrin binding to the
alternatively spliced CS-1 peptide region of fibronectin (Fig.
2A), as the latter down-regulates MMP synthesis. As HT1080
cell adhesion to fibronectin can be completely inhibited by an
antibody to the α5 integrin subunit (Yamada et al., 1990) it is
unlikely that the CS-1 region is involved in the regulation of
MMP-2 activation by HT1080 cells.
We decided to concentrate on the CCBD of fibronectin and
to culture HT1080 cells on smaller fragments comprising the
RGD site and synergy site. However, we encountered a
problem with poor cell spreading on fragment Fn III 6-10;
nevertheless, MMP-2 activation was observed. To circumvent
this problem of altered cell shape, cells were cultured directly
on a substrate of anti-integrin antibodies. HT1080 cells spread
well on antibodies to the α5, α6 and β1 integrin subunits.
Adhesion to the anti-α5 integrin subunit monoclonal antibody
mAb 16 promoted MMP-2 processing, but interestingly,
monoclonal mAb 11, also to the α5 integrin subunit failed to
support MMP-2 activation. Both antibodies are known to
promote integrin receptor clustering; however mAb 16, in
contrast to mAb 11, also mimics the action of the fibronectin
ligand (Miyamoto et al., 1995). Fibronectin receptor
occupancy would therefore appear to be important in the
signalling of MMP-2 activation. The lack of processing of
MMP-2 when cells were cultured on laminin-1 was also
observed when the cells were cultured on an antibody to the
laminin receptor α6 integrin subunit. However, mAb 13 against
the β1 subunit of α5β1 also promoted MMP-2 activation. This
result is consistent with the fact that mAb 13 can mimic
receptor occupancy. Collectively, our data with fibronectin
peptides and integrin antibodies indicate that the mechanism
by which fibronectin up-regulates MMP-2 processing by
HT1080 cells may involve signalling via the α5β1 integrin.
Several reports have described the regulation of MMP-2 and
MMP-9 expression in response to treatment with anti-integrin
antibodies (Larjava et al., 1993; Seftor et al., 1992). Kubota et
al. (1997) demonstrated that soluble anti-α2 and anti-α3
integrin antibodies induced proMMP-2 secretion and activation
by human rhabdomyosarcoma cells. Studies with two
glioblastoma cell lines showed that MMP-2 expression was
increased by treatment with anti-α3β1 or anti-α5β1 integrin
antibodies (Chintala et al., 1996). Clearly, increases in MMP-
2 expression and activation may be signalled by several of the
integrin receptors and the response to individual integrins is
cell type specific. Similarly, MMP induction by ECM
molecules is dependent upon the cell type. In our laboratory,
HT1080 cells respond to full length fibronectin by increasing
MMP-2 activation whereas MMP-2 activation by human
foreskin fibroblasts (which also express α5β1 integrin) is not
affected by culture on full-length fibronectin or the 120kDa
fragment (data not shown). This differential response may
Fig. 7. In vitro cleavage studies: incubation of membrane bound
native MT1-MMP with active MMPs. Cell membranes enriched in
the 60 kDa form of MT1-MMP were prepared from cultures of MT1-
MMP transfected HT1080 cells grown in the presence of CT1746
inhibitor. Membrane preparations were incubated alone, or in the
presence of active MMPs for 4 hours at 37°C, Laemmli sample
buffer added and the protein mixture separated by 10% SDS-
polacrylamide gel electrophoresis, electroblotted to nitrocellulose
and probed with the polyclonal antibody to MT1-MMP. Cell
membranes incubated alone (lane 2), or in the presence of active
MMP-2 (lane 3), active ∆TM-MT1-MMP (lane 4), active MMP-3
(lane 5), active MMP-13 (lane 6). Cell membranes, unincubated
(lane 1). Recombinant active ∆TM-MT1-MMP alone (lane 7).
Molecular mass markers are indicated on the right.
2796
indicate diverging intracellular signalling pathways following
integrin ligation in these two cell types, and future studies may
elucidate the mechanisms involved.
As the experiments described here were conducted over 24-
48 hours, we have not ruled out the possibility that fibronectin
induces the expression of an endogenous cytokine or growth
factor, that in turn signals changes to MMP-2 activation. It has
been demonstrated for early passage fibroblasts that the
induction of MMP-1 activity by PMA proceeds via an
interleukin-1 autocrine loop (West-Mays et al., 1995), and
similarly, the RGD peptide-induced expression of MMPs -1,
-3 and -9 by rabbit chondrocytes is augmented by an
interleukin-1 autocrine loop (Arner and Tortorella, 1995).
Evidence suggests that growth factors synergise with
extracellular matrix/integrin mediated signalling pathways
(Schwartz et al., 1995). Ligand-mediated integrin clustering
leads to the accumulation of growth factor receptors (Plopper
et al., 1995; Miyamoto et al., 1996). Furthermore, β1 integrin
receptor occupancy results in the enhancement of growth factor
receptor tyrosine phosphorylation and the transiently enhanced
activation of mitogen-activated protein kinases (Miyamoto et
al., 1996). The signalling mechanisms involved in fibronectin
up-regulation of MMP-2 activation and the possibility that an
endogenous growth factor or cytokine is required will be the
subject for future investigations.
In order to effect increases in proMMP-2 activation, we
speculated that the fibronectin matrix may be signalling
changes to other molecules involved in the activation cascade.
Interactions between the C-terminal domain of proMMP-2 and
a membrane-bound component are known to be important
(Strongin et al., 1995; Ward et al., 1994). MMP-2 has been
reported to bind αvβ3 integrin on the surface of β3 integrin
transfected cells and to purified αvβ3 integrin in solid phase
studies (Brooks et al., 1996, 1998), a process apparently
mediated by the C-terminal domain of MMP-2. It is unlikely,
however, that MMP-2/αvβ3 interactions play a role in MMP-
2 activation in the current study, for immunolocalisation
studies with several anti-αv or αvβ3 integrin antibodies have
indicated that the HT1080 cells used in these experiments
apparently do not express αvβ3 integrin (A. Messent and J.
Gavrilovic, unpublished observations). Membrane bound
MT1-MMP is known to initiate MMP-2 activation (Sato et al.,
1994; Atkinson et al., 1995; Will et al., 1996). It has been
hypothesised that MT1-MMP and TIMP-2 form a ‘receptor’
complex, that binds MMP-2 via its C terminus (Strongin et al.,
1995; Imai et al., 1996; Butler et al., 1998). Evidence for this
trimolecular complex has been provided by cross-linking
experiments (Strongin et al., 1995) and a model for MMP-2
activation has been described in which proteolysis of MMP-2
bound in the complex requires an adjacent MT1-MMP
molecule that is TIMP-2 free and therefore catalytically active
(Butler et al., 1998). This model predicts that the balance of
TIMP-2 and MT1-MMP is of critical importance in
determining the activation status of MMP-2. In agreement with
this model, addition of low concentrations of TIMP-2 to cell
membranes increases MMP-2 activation, presumably by
increasing the concentration of MT1-MMP/TIMP-2 complex
available for MMP-2 binding. Conversely, high concentrations
of TIMP-2 inhibit MMP-2 activation (Strongin et al., 1995;
Butler et al., 1998), probably because all MT1-MMP
molecules form MT1-MMP/TIMP-2 complexes, leaving no
free MT1-MMP molecules available to initiate proteolysis
(Butler et al., 1998). We postulated, therefore, that the
increased activation of MMP-2 by HT1080 cells cultured on
fibronectin in comparison with cells cultured on laminin-1
could reflect a variation in either TIMP-2 or MT1-MMP status.
The data indicated that the secretion of TIMP-2 was similar
when HT1080 cells were cultured on fibronectin or laminin-1
matrices. Furthermore, western blot analysis of cell lysates
indicated that the amount of TIMP-2 associated with the cells
was similar for both matrices. Using solid phase studies, we
also found that TIMP-2 does not bind fibronectin, thus ruling
out the possibility that a fibronectin matrix might act to trap
TIMP-2 (and therefore MMP-2) close to the cell (data not
shown). Collectively, these observations suggested that the
effect of fibronectin on the activation of MMP-2 by HT1080
cells was unlikely to be due to alterations in TIMP-2 status.
We also investigated the expression of MT1-MMP by
HT1080 cells. It has been shown that the culture of several cell
types on collagen type I gels or within collagen lattices induces
MT1-MMP mRNA expression and a corresponding increase in
MMP-2 activation (Gilles et al., 1997; Haas et al., 1998). In
the current study MT1-MMP mRNA levels were similar when
cells were cultured on fibronectin or laminin-1 substrates. To
determine whether post-translational regulation of MT1-MMP
is affected by fibronectin, MT1-MMP protein was analysed by
western blot using a newly described antibody to MT1-MMP.
Two major immunoreactive bands were detected at 60 kDa and
45 kDa, and a faint doublet at 63 kDa. This profile is very
similar to that reported by Lohi et al. (1996), who detected the
doublet at 63 kDa and major bands at 60 and 43 kDa in
stimulated HT1080 cells using two polyclonal antibodies to
MT1-MMP. The 60 kDa band in the current study corresponds
most likely to the active form of MT1-MMP (Pei and Weiss,
1996; Will et al., 1996; S. Atkinson, G. Butler, G. Murphy,
unpublished results). Although the level of the 60kDa form was
not affected by the various culture conditions employed, it was
clear that the overall MT1-MMP protein levels increased when
the cells were cultured on fibronectin, as the 45 kDa band was
up-regulated. Similarly, PMA treatment up-regulated the
expression of the 45 kDa band, as has been noted by Lohi et
al. (1996). Increases in the 45 kDa band were noted to be
concomitant with the activation of MMP-2, an observation that
agrees with the report by Lohi et al. (1996). It is yet to be
demonstrated whether there is a direct link between MMP-2
activation and MT1-MMP processing to 45 kDa, and what role
the 45 kDa band plays in MMP activation mechanisms.
According to the study by Lohi et al. (1996) the truncation of
MT1-MMP to the 43kDa form involves the loss of the N
terminus, since antibodies raised to peptides from the C-
terminal hemopexin domain and the intracellular domain of
MT1-MMP detect the 43kDa form in western blot. Allowing
for the loss of the N terminus, simple calculations would
predict that MT1-MMP is cleaved within the catalytic domain.
We postulate that this would render the molecule inactive,
which would represent an important regulatory step in MT1-
MMP activation cascades.
To unravel the sequence of events involved in the fibronectin
stimulation of MT1-MMP processing to 45 kDa more
information is required on the nature of MT1-MMP processing
and on the identity of the protease(s) involved. We added
protease inhibitors to HT1080 cells cultured on fibronectin.
H. Stanton and others
2797Fibronectin promotes processing of ProMMP-2 and MT1-MMP
Inhibition of MT1-MMP processing to 45 kDa by the
hydroxamate inhibitor CT1746 and by TIMP-2 indicated that
the protease was an MMP. Interestingly, TIMP-1, even at high
concentrations, failed to inhibit processing to 45 kDa. Our
laboratory has demonstrated that MT1-MMP activity is
efficiently inhibited by TIMP-2 and TIMP-3, but not by TIMP-
1 (Will et al., 1996). Our TIMP-1 data would therefore imply
that MT1-MMP processing to 45 kDa is autolytic. However
TIMP-1 also failed to inhibit MMP-2 autolysis from the
intermediate to the fully active form, a surprising result, as
TIMP-1 is an effective inhibitor of MMP-2 (Murphy et al.,
1992b). It is possible that MMP-2 bound in an MT1-
MMP/TIMP-2 complex is not susceptible to TIMP-1 inhibition
(suggested by the kinetic study of Willenbrock et al., 1993).
As the TIMP-1 inhibition data are difficult to interpret, both
MT1-MMP and MMP-2 must be considered as candidate
proteases for HT1080 MT1-MMP processing to 45 kDa.
MMP-2 is constitutively expressed in most cell types and, as
noted above, the presence of active MMP-2 correlates well
with the appearance of the 45 kDa band. We demonstrated by
an in vitro study that both MT1-MMP and MMP-2 cleave 60
kDa MT1-MMP to the 45 kDa form. Furthermore this cleavage
site proved to be susceptible to several members of the MMP
family in vitro. These results raise the intriguing possibility that
the end step in the activation pathway of MMP-2 involves the
cleavage of the activator MT1-MMP. To fully understand the
effects of fibronectin on MT1-MMP processing to 45 kDa,
further work is needed to define the exact enzymes involved.
In summary, we have shown that the culture of HT1080 cells
on fibronectin up-regulates the activation of MMP-2 and that
signals via α5β1 integrin are likely to be involved. We chose
to study MT1-MMP and TIMP-2, which are key components
of MMP-2 activation, to determine whether a fibronectin
substrate alters their expression or processing by HT1080 cells.
We observed that proteolytic processing of active 60 kDa MT1-
MMP to a 45 kDa product was concomitant with MMP-2
activation and that this processing was up-regulated by
fibronectin. We demonstrated that MT1-MMP processing is
MMP mediated. Work is underway to determine the N-terminal
sequence of the 45 kDa band and to assess whether it is
catalytically inactive. We speculate that proteolysis of active
MT1-MMP to a 45 kDa form may represent an end-point in
the activation pathway of MMP-2.
The up-regulation of MMP-2 activation by fibronectin may
be of particular importance in wound healing where fibronectin
is laid down in the provisional matrix forming a conduit for
inwardly migrating cells (Greiling and Clark, 1997). During
cancer progression fibronectin is also a prominent component
of the host stroma and may modulate MMP-2 activation,
essential for tumour invasion.
We thank Dr S. Cowell for the MT1-MMP riboprobe and for
assistance with northern blotting. This work was supported by the
Medical Research Council, UK, the Cancer Research Campaign, UK,
the Arthritis and Rheumatism Campaign, UK, INSERM, France and
Synthélabo, France.
REFERENCES
Akiyama, S. K., Yamada, S. S., Chen, W.-T. and Yamada, K. M. (1989).
Analysis of fibronectin receptor function with monoclonal antibodies: roles
in cell adhesion, migration, matrix assembly and cytoskeletal organization.
J. Cell Biol. 109, 863-875.
Akiyama, S. K., Olden, K. and Yamada, K. M. (1995). Fibronectin and
integrins in invasion and metastasis. Cancer Metast. Rev. 14, 173-189.
Allan, J. A., Docherty, A. J. P., Barker, P. J., Huskisson, N. S., Reynolds,
J. J. and Murphy, G. (1995). Binding of gelatinases A and B to type I
collagen and other matrix components. Biochem. J. 309, 299-306.
Arner, E. C. and Tortorella, M. D. (1995). Signal transduction through
chondrocyte integrin receptors induces matrix metalloproteinase synthesis
and synergizes with interleukin-1. Arthritis Rheum. 38, 1304-1314.
Atkinson, S. J., Crabbe, T., Cowell, S., Ward, R. V., Butler, M. J., Sato, H.,
Seiki, M., Reynolds, J. J. and Murphy, G. (1995). Intermolecular autolytic
cleavage can contribute to the activation of progelatinase A by cell
membranes. J. Biol. Chem. 270, 30479-30485.
Azzam, H. S. and Thompson, E. W. (1992). Collagen-induced activation of
the Mr72, 000 type IV collagenase in normal and malignant human
fibroblastoid cells. Cancer Res. 52, 4540-4544.
Borsi, L., Castellani, P., Balza, E., Siri, A., Pellecchia, C., De Scalzi, F. and
Zardi, L. (1986). Large-scale procedure for the purification of fibronectin
domains. Anal. Biochem. 155, 335-345.
Brooks, P. C., Strömblad, S., Sanders, L. C., von Schalscha, T. L., Aimes,
R., Stetler-Stevenson, W. G., Quigley, J. P. and Cheresh, D. A. (1996).
Localization of matrix metalloproteinase MMP-2 to the surface of cells by
interaction with integrin αvβ3. Cell 85, 683-693.
Brooks, P. C., Silletti, S., von Schalscha, T. L., Friedlander, M. and
Cheresh, D. A. (1998). Disruption of angiogenesis by PEX, a noncatalytic
metalloproteinase fragment with integrin binding activity. Cell 92, 391-400.
Butler, G. S., Butler, M. J., Atkinson, S. J., Will, H., Tamura, T., Schade
van Westrum, S., Crabbe, T., Clements, J., d’Ortho, M. P. and Murphy,
G. (1998). The TIMP2 membrane type 1 metalloproteinase ‘receptor’
regulates the concentration and efficient activation of progelatinase A. A
kinetic study. J. Biol. Chem. 273, 871-880.
Chambers, A. F. and Matrisian, L. M. (1997). Changing views of the role
of matrix metalloproteinases in metastasis. J. Nat. Cancer Inst. 89, 1260-
1270.
Chintala, S. K., Sawaya, R., Gokaslan, Z. L. and Rao, J. S. (1996).
Modulation of matrix metalloprotease-2 and invasion in human glioma cells
by α3β1 integrin. Cancer Lett. 103, 201-208.
Chomczynski, P. and Sacchi, N. (1987). Single-step method of RNA isolation
by acid guanidinium thiocyanate-phenol-chloroform extraction. Anal.
Biochem. 162, 156-159.
d’Ortho, M. P., Will, H., Atkinson, S. J., Butler, G., Messent, A. J.,
Gavrilovic, J., Smith, B., Timpl, R., Zardi, L. and Murphy, G. (1997).
Membrane-type matrix metalloproteinases 1 and 2 exhibit broad-spectrum
proteolytic capacities comparable to many matrix metalloproteinases. Eur.
J. Biochem. 250, 751-757.
d’Ortho, M. P., Stanton, H., Butler, M., Atkinson, S. J., Murphy, G. and
Hembry, R. M. (1998). MT1-MMP on the cell surface causes focal
degradation of gelatin films. FEBS Lett. 421, 159-164.
Danen, E. H. J., Aota, S., van Kraats, A. A., Yamada, K., Ruiter, D. J. and
van Muijen, G. N. P. (1995). Requirement for the synergy site for cell
adhesion to fibronectin depends on the activation state of the integrin α5β1.
J. Biol. Chem. 270, 21612-21618.
Fukai, F., Ohtaki, M., Fujii, N., Yajima, H., Ishii, T., Nishizawa, Y.,
Miyazaki, K. and Katayama, T. (1995). Release of biological activities
from quiescent fibronectin by a conformational change and limited
proteolysis by matrix metalloproteinases. Biochemistry 34, 11453-11459.
Gilles, C., Polette, M., Seiki, M., Birembaut, P. and Thompson, E. W.
(1997). Implication of collagen type-I induced membrane-type 1-matrix
metalloproteinase expression and matrix metalloproteinase-2 activation in
the metastatic progression of breast carcinoma. Lab. Invest. 76, 651-660.
Green, D. R., Banuls, M. P., Gearing, A. J. H., Needham, L. A., White, M.
R. H. and Clements, J. M. (1994). Generation of human umbilical vein
endothelial cell lines which maintain their differentiated phenotype.
Endothelium 2, 191-201.
Greiling, D. and Clark, R. A. F. (1997). Fibronectin provides a conduit for
fibroblast transmigration from collagenous stroma into fibrin clot
provisional matrix. J. Cell Sci. 110, 861-870.
Haas, T. L., Davis, S. and Madri, J. A. (1998). Three-dimensional type I
collagen lattices induce coordinate expression of matrix metalloproteinases
MT1-MMP and MMP-2 in microvascular endothelial cells. J. Biol. Chem.
273, 3604-3610.
Heussen, C. and Dowdle, E. B. (1980). Electrophoretic analysis of
2798
plasminogen activators in polyacrylamide gels containing sodium dodecyl
sulfate and copolymerized substrates. Anal. Biochem. 102, 196-202.
Huhtala, P., Humphries, M. J., McCarthy, J. B., Tremble, P. M., Werb, Z.
and Damsky, C. H. (1995). Cooperative signaling by α5β1 and α4β1
integrins regulates metalloproteinase gene expression in fibroblasts adhering
to fibronectin. J. Cell Biol. 129, 867-879.
Imai, K., Ohuchi, E., Aoki, T., Nomura, H., Fujii, Y., Sato, H., Seiki, M.
and Okada, Y. (1996). Membrane-type matrix metalloproteinase 1 is a
gelatinolytic enzyme and is secreted in a complex with tissue inhibitor of
metalloproteinases 2. Cancer Res. 56, 2707-2710.
Knäuper, V., Will, H., López-Otín, C., Smith, B., Atkinson, S. J., Stanton,
H., Hembry, R. M. and Murphy, G. (1996). Cellular mechanisms for
human procollagenase-3 (MMP-13) activation. J. Biol. Chem. 271, 17124-
17131.
Kubota, S., Ito, H., Ishibashi, Y. and Seyama, Y. (1997). Anti-alpha3
integrin antibody induces the activated form of matrix metalloproteinase-2
(MMP-2) with concomitant stimulation of invasion through matrigel by
human rhabdomyosarcoma cells. Int. J. Cancer 70, 106-111.
Laemmli, U. K. and Favre, M. (1973). Maturation of the head of
bacteriophage T4. I. DNA packaging events. J. Mol. Biol. 80, 575-599.
LaFlamme, S. E., Akiyama, S. K. and Yamada, K. M. (1992). Regulation
of fibronectin receptor distribution. J. Cell Biol. 117, 437-447.
Langholz, O., Röckel, D., Mauch, C., Kozlowska, E., Bank, I., Krieg, T.
and Eckes, B. (1995). Collagen and collagenase gene expression in three-
dimensional collagen lattices are differentially regulated by α1β1 and α2β1
integrins. J. Cell Biol. 131, 1903-1915.
Larjava, H., Lyons, J. G., Salo, T., Makela, M., Koivisto, L., Birkedal-
Hansen, H., Akiyama, S. K., Yamada, K. M. and Heino, J. (1993). Anti-
integrin antibodies induce type IV collagenase expression in keratinocytes.
J. Cell Physiol. 157, 190-200.
Lohi, J., Lehti, K., Westermarck, J., Kähäri, V. and Keski-Oja, J. (1996).
Regulation of membrane-type matrix metalloproteinase-1 expression by
growth factors and phorbol 12-myristate 13-acetate. Eur. J. Biochem. 239,
239-247.
Miyamoto, S., Teramoto, H., Coso, O. A., Gutkind, J. S., Burbelo, P. D.,
Akiyama, S. K. and Yamada, K. M. (1995). Integrin function:molecular
hierarchies of cytoskeletal and signaling molecules. J. Cell Biol. 131, 791-
805.
Miyamoto, S., Teramoto, H., Gutkind, J. S. and Yamada, K. M. (1996).
Integrins can collaborate with growth factors for phosphorylation of receptor
tyrosine kinases and MAP kinase activation: roles of integrin aggregation
and occupancy of receptors. J. Cell Biol. 135, 1633-1642.
Mohri, H. (1996). Fibronectin and integrins interactions. J. Invest. Medicine
44, 429-441.
Murphy, G., Cawston, T. E. and Reynolds, J. J. (1981). An inhibitor of
collagenase from human amniotic fluid. Purification, characterization and
action on metalloproteinases. Biochem. J. 195, 167-170.
Murphy, G., Houbrechts, A., Cockett, M. I., Williamson, R. A., O’Shea,
M. and Docherty, A. J. P. (1991). The N-terminal domain of tissue inhibitor
of metalloproteinases retains metalloproteinase inhibitory activity.
Biochemistry 30, 8097-8102.
Murphy, G., Allan, J. A., Willenbrock, F., Cockett, M. I., O’Connell, J. P.
and Docherty, A. J. P. (1992a). The role of the C-terminal domain in
collagenase and stromelysin specificity. J. Biol. Chem. 267, 9612-9618.
Murphy, G., Willenbrock, F., Ward, R. V., Cockett, M. I., Eaton, D. and
Docherty, A. J. P. (1992b). The C-terminal domain of 72 kDa gelatinase A
is not required for catalysis, but is essential for membrane activation and
modulates interactions with tissue inhibitors of metalloproteinases.
Biochem. J. 283, 637-641.
Murphy, G., Ward, R., Gavrilovic, J. and Atkinson, S. (1992c).
Physiological mechanisms for metalloproteinase activation. Matrix (suppl.)
1, 224-230.
Ohuchi, E., Imai, K., Fujii, Y., Sato, H., Seiki, M. and Okada, Y. (1997).
Membrane type 1 matrix metalloproteinase digests interstitial collagens and
other extracellular matrix macromolecules. J. Biol. Chem. 272, 2446-2451.
Paterson, H., Reeves, B., Brown, R., Hall, A., Furth, M., Bos, J., Jones, P.
and Marshall, C. (1987). Activated N-ras controls the transformed
phenotype of HT1080 human fibrosarcoma cells. Cell 51, 803-812.
Pei, D. and Weiss, S. J. (1996). Transmembrane-deletion mutants of the
membrane-type matrix metalloproteinase-1 process progelatinase A and
express intrinsic matrix-degrading activity. J. Biol. Chem. 271, 9135-9140.
Plopper, G. E., McNamee, H. P., Dike, L. E., Bojanowski, K., Ingber, D.
E. (1995). Convergence of integrin and growth factor receptor signaling
pathways within the focal adhesion complex. Mol. Biol. Cell 6, 1349-1365.
Reich, R., Blumenthal, M. and Liscovitch, M. (1995). Role of phospholipase
D in laminin-induced production of gelatinase A (MMP-2) in metastatic
cells. Clin. Exp. Metast. 13, 134-140.
Riikonen, T., Westermarck, J., Koivisto, L., Broberg, A., Kähäri, V. M.
and Heino, J. (1995). Integrin α2β1 is a positive regulator of collagenase
(MMP-1) and collagen α1(I) gene expression. J. Biol. Chem. 270, 13548-
13552.
Sato, H., Takino, T., Okada, Y., Cao, J., Shinagawa, A., Yamamoto, E. and
Seiki, M. (1994). A matrix metalloproteinase expressed on the surface of
invasive tumour cells. Nature 370, 61-65.
Sato, H. and Seiki, M. (1996). Membrane-type matrix metalloproteinases
(MT-MMPs) in tumor metastasis. J. Biochem. 119, 209-215.
Schwartz, M. A., Schaller, M. D. and Ginsberg, M. H. (1995). Integrins-
emerging paradigms of signal-transduction. Annu. Rev. Cell Dev. Biol. 11,
549-599
Seftor, R. E. B., Seftor, E. A., Gehlsen, K. R., Stetler-Stevenson, W. G.,
Brown, P. D., Ruoslahti, E. and Hendrix, M. J. C. (1992). Role of the
αvβ3 integrin in human melanoma cell invasion. Proc. Nat. Acad. Sci. USA
89, 1557-1561.
Strongin, A. Y., Collier, I., Bannikov, G., Marmer, B. L., Grant, G. A. and
Goldberg, G. I. (1995). Mechanism of cell surface activation of 72-kDa
type IV collagenase. J. Biol. Chem. 270, 5331-5338.
Tremble, P., Chiquet-Ehrismann, R. and Werb, Z. (1994). The extracellular
matrix ligands fibronectin and tenascin collaborate in regulating collagenase
gene expression in fibroblasts. Mol. Biol. Cell 5, 439-453.
von der Mark, H., Durr, J., Sonnenberg, A., von der Mark, K.,
Deutzmann, R. and Goodman, S. L. (1991). Skeletal myoblasts utilize a
novel beta 1-series integrin and not alpha 6 beta 1 for binding to the E8 and
T8 fragments of laminin. J. Biol. Chem. 266, 23593-23601.
Wallon, U. M. and Overall, C. M. (1997). The hemopexin-like domain (C
domain) of human gelatinase A (matrix metalloproteinase-2) requires Ca2+
for fibronectin and heparin binding. J. Biol. Chem. 272, 7473-7481.
Ward, R. V., Hembry, R. M., Reynolds, J. J. and Murphy, G. (1991). The
purification of tissue inhibitor of metalloproteinases-2 from its 72 kDa
progelatinase complex. Demonstration of the biochemical similarities of
tissue inhibitor of metalloproteinases-2 and tissue inhibitor of
metalloproteinases-1. Biochem. J. 278, 179-187.
Ward, R. V., Atkinson, S. J., Reynolds, J. J. and Murphy, G. (1994). Cell
surface-mediated activation of progelatinase A: demonstration of the
involvement of the C-terminal domain of progelatinase A in cell surface
binding and activation of progelatinase A by primary fibroblasts. Biochem.
J. 304, 263-269.
Werb, Z., Tremble, P. M., Behrendtsen, E., Crowley, E. and Damsky, C.
H. (1989). Signal transduction through the fibronectin receptor induces
collagenase and stromelysin gene expression. J. Cell Biol. 109, 877-889.
Werb, Z., Sympson, C. J., Alexander, C. M., Thomasset, N., Lund, L. R.,
MacAuley, A., Ashkenas, J. and Bissell, M. J. (1996). Extracellular matrix
remodeling and the regulation of epithelial-stromal interactions during
differentiation and involution. Kidney Int. 49 (suppl. 54), S68-S74.
Werb, Z. (1997). ECM and cell surface proteolysis: regulating cellular
ecology. Cell 91, 439-442.
West-Mays, J. A., Strissel, K. J., Sadow, P. M. and Fini, M. E. (1995).
Competence for collagenase gene expression by tissue fibroblasts requires
activation of an interleukin 1αautocrine loop. Proc. Nat. Acad. Sci. USA
92, 6768-6772.
Will, H., Atkinson, S. J., Butler, G. S., Smith, B. and Murphy, G. (1996). The
soluble catalytic domain of membrane type 1 matrix metalloproteinase cleaves
the propeptide of progelatinase A and intiates autoproteolytic activation.
Regulation by TIMP-2 and TIMP-3. J. Biol. Chem. 271, 17119-17123.
Willenbrock, F., Crabbe, T., Slocombe, P. M., Sutton, C. W., Docherty, A.
J. P., Cockett, M. I., O’Shea, M., Brocklehurst, K., Phillips, I. R. and
Murphy, G. (1993). The activity of the tissue inhibitors of
metalloproteinases is regulated by C-terminal domain interactions: a kinetic
analysis of the inhibition of gelatinase A. Biochemistry 32, 4330-4337.
Yamada, K. M. and Kennedy, D. W. (1984) Dualistic nature of adhesive
protein function. J. Cell Biol. 99, 29-36.
Yamada, K. M., Kennedy, D. W., Yamada, S. S., Gralnick, H., Chen, W.-
T. and Akiyama, S. K. (1990). Monoclonal antibody and synthetic peptide
inhibitors of human tumor cell migration. Cancer Res. 50, 4485-4496.
H. Stanton and others