| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
Reviews |
MediCity Research Laboratory and Department of Medical Biochemistry, University of Turku; and
* Department of Dermatology, Turku University Central Hospital, FIN-20520 Turku, Finland
1Correspondence: University of Turku, MediCity Research Laboratory, Tykistökatu 6, FIN-20520 Turku, Finland. E-mail:jukwes{at}utu.fi
 |
ABSTRACT |
|---|
 TOP ABSTRACT MATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
Key Words: extracellular matrix • MMPs • interleukin • tissue inhibitors • SCC
|
MATRIX METALLOPROTEINASES |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
2
3)
. Tumor invasion, metastasis, and
angiogenesis require controlled degradation of ECM, and increased
expression of matrix metalloproteinases (MMPs) is associated with tumor
invasion and metastasis of malignant tumors with different histogenetic
origin (4,
5)
. MMPs are a family of at least 17 human zinc-dependent
endopeptidases, collectively capable of degrading essentially all ECM
components (see refs 1,
3,
5
). According to their substrate specificity
and structure, members of the MMP gene family can be classified into
subgroups of collagenases, stromelysins, gelatinases, membrane-type
MMPs, and other MMPs (see Fig. 1
). In general, MMPs contain a signal peptide, a propeptide, a catalytic
domain with the highly conserved zinc binding site, and a
hemopexin-like domain linked to the catalytic domain by a hinge region
(Fig. 1)
. In addition, gelatinase-A (MMP-2) and gelatinase-B (MMP-9)
contain fibronectin type II inserts within the catalytic domain and
MT-MMPs contain a transmembrane domain in the carboxyl-terminal end of
the hemopexin-like domain (Fig. 1)
. The hemopexin domain is absent in
the smallest MMP, matrilysin (MMP-7). The substrate specificity of
distinct MMPs has been determined by their ability to degrade different
components of ECM in vitro. However, direct evidence for the
proteolytic activity of MMPs in vivo is still limited.
|
Collagenase-1 (MMP-1), collagenase-2 (MMP-8), and collagenase-3
(MMP-13) are the principal secreted neutral proteinases capable of
initiating degradation of native fibrillar collagens of types I, II,
III, and V, and apparently play a crucial role in degradation of
collagenous ECM in various physiological and pathological situations
(see refs 1
2
3
). They all cleave fibrillar collagens at a specific
site, resulting in generation of amino-terminal 3/4 and
carboxyl-terminal 1/4 fragments, which then rapidly denature in body
temperature and are further degraded by other MMPs, e.g., gelatinases
(see refs 1
2
3
). Expression of collagenase-1 (MMP-1) in vivo
is seen in areas of rapid ECM remodeling both in physiological and
pathological situations. MMP-1 is expressed by several types of cells
including fibroblasts, keratinocytes, chondrocytes, monocytes and
macrophages, hepatocytes, and a variety of tumor cells. Substrates for
MMP-1 include collagens of types I, II, III, VII, VIII, X, aggrecan,
serpins, and 
2-macroglobulin (1
2
3)
. Collagenase-2 is synthesized by
polymorphonuclear leukocytes, stored in their secretory granules, and
released in response to the appropriate stimulus. In addition,
expression of MMP-8 has been observed in chondrocytes, synovial
fibroblasts, and endothelial cells (6)
. MMP-13 was originally cloned
from human breast carcinoma tissue; compared with other collagenases,
its substrate specificity is exceptionally wide, including fibrillar
collagens of types I, II, III, and XI, basement membrane and cartilage
collagen types IV and X, collagen type IX, gelatin, laminin, tenascin,
aggrecan, and fibronectin (7)
. Apparently due to its ability to degrade
a wide range of ECM components, the physiological expression of MMP-13
is limited to situations in which rapid and effective remodeling of
collagenous ECM is required, i.e., fetal bone development and postnatal
bone remodeling (8,
9)
. On the other hand, MMP-13 is expressed at sites
of excessive degradation of collagenous ECM in osteoarthritic cartilage
(9)
, rheumatoid synovium (9,
10)
, chronic cutaneous ulcers (11)
,
intestinal ulcerations (12)
, and periodontitis (13)
, as well as in
malignant tumors such as breast carcinomas (14
15
16)
, squamous cell
carcinomas (SCCs) of the head/neck and vulva (17
18
19)
, cutaneous basal
cell carcinomas (18)
, and chondrosarcomas (20)
.
The stromelysin subgroup contains stromelysin-1 (MMP-3), stromelysin-2
(MMP-10), and two other MMPs with similar substrate specificities that
are less closely related structurally: MMP-7 and macrophage
metalloelastase (MMP-12). MMP-3 and MMP-10 are expressed by
fibroblastic cells and by normal and transformed squamous epithelial
cells (18,
19)
. Stromelysins degrade basement membrane components, type
IV collagen, nidogen, and fibronectin; both matrilysin and macrophage
metalloelastase have the ability to degrade elastin (1
2
3)
.
MMP-2 (72 kDa gelatinase-A) is expressed by a variety of normal and
transformed cells. MMP-9 (92 kDa gelatinase-B) is produced by
keratinocytes, monocytes, alveolar macrophages, PMN leukocytes, and a
large variety of malignant cells. In addition to the capacity of MMP-2
and MMP-9 to degrade gelatin, laminin, and nidogen, MMP-2 has also been
reported to degrade native type I collagen and proteolytically activate
MMP-9 and MMP-13 (4,
21)
.
The first membrane-type MMP (MT1-MMP) (MMP-14) was cloned from invasive
lung cancer cells (22)
and revealed a typical five-domain modular
structure resembling collagenases and stromelysins; it also contains an
RXXR domain susceptible to intracellular proteolytic activation by
furin (23)
, an additional short carboxyl-terminal transmembrane domain,
and intracellular domain. Three additional MT-MMPs have been cloned:
MT2-MMP (MMP-15); MT3-MMP (MMP-16), and MT4-MMP (MMP-17) (1
2
3)
. Active
MT1-MMP serves as a cell membrane receptor for the complex formed of
latent MMP-2 (proMMP-2) and tissue inhibitor of metalloproteinases-2
(TIMP-2); at least MT1-MMP and MT2-MMP have been shown to
proteolytically activate proMMP-2 at the cell surface (21)
. In addition
to activation of MMPs, MT1-MMP and MT2-MMP degrade type I, II, and III
collagens, gelatin, fibronectin, laminin, vitronectin, and aggrecan.
In vivo, MT1-MMP is expressed both in stromal fibroblasts
adjacent to invasive tumor and in malignant epithelium (19,
24)
.
The cDNA of stromelysin-3 (MMP-11) was cloned from invasive breast
cancer tissue. The predicted structure of MMP-11 resembles that of
other stromelysins and collagenases. It has not been shown to degrade
any ECM component, but it degrades serine proteinase inhibitors
1-proteinase inhibitor and 1-antitrypsin (1
2
3)
. Cossins et al.
(25)
cloned cDNA of a novel MMP from a human mammary gland, which is
identical to MMP-19 cloned from liver (26)
. The predicted protein
structure of MMP-19 has the closest identity with MMP-1, -3, -10, and
-11. However, the presence of a carboxyl-terminal arginine rich domain
suggests that this MMP belongs to a new group of MMPs (19,
26)
. The
ability of MMP-19 to degrade native ECM components is not known.
Enamelysin (MMP-20), cloned from the odontoblastic cells, is the latest
member of the human MMP family (27)
. The expression of enamelysin has
been observed only in dental tissues, which together with its ability
to degrade dental amelogenin suggests a specific role for enamelysin in
the remodeling of dental enamel.
Most MMPs are secreted as latent precursors (zymogens) that are
proteolytically activated in the extracellular space, with the
exception of MMP-11 and MT1-MMP, which are activated prior to secretion
by Golgi-associated, furin-like proteases (see refs 1
2
3
, 21
). The
activity of MMPs in extracellular space is specifically inhibited by
tissue inhibitors of metalloproteinases (TIMPs), which bind to the
highly conserved zinc binding site of active MMPs at molar equivalence.
The TIMP gene family consists of four structurally related members,
TIMP-1, -2, -3, and -4, which show 30 to 40% identity at the amino
acid level and possess 12 conserved cysteine residues (28)
. TIMP-1, -2,
and -4 are secreted in soluble form whereas TIMP-3 is associated with
ECM. TIMPs have biological effects that extend beyond their role as
inhibitors of MMP activity. They induce changes in cell morphology,
stimulate growth of several cell types, and are involved in
steroidogenesis and germ-cell development of both sexes (see ref 28
).
TIMP-1 and -3 are antiangiogenic (28)
; TIMP-2 is also involved in
activation of MMP-2 (see ref 21
).
|
EXPRESSION OF MMPs IN CANCER |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
5
), but evidence for the activity of distinct
MMPs in tumor tissues in vivo is limited. However, there is
recent direct evidence that MMPs play an important role in tumor
invasion and progression. For example, increased expression of TIMPs by
either host or tumor cells results in reduced invasion and metastatic
capacity of transformed cells (29,
30)
. In addition, it has been shown
that MMP activity is required for increased motility of the epithelial
cells (31)
and for growth of metastasized tumor cells (32)
.
Furthermore, MMPs have been shown to play an essential role in
angiogenesis and tumor cell intravasation (33,
34)
, both of which are
required for tumor cell growth and metastasis. Additional evidence for
the role of MMPs in invasion and growth of tumors has recently been
provided by MMP knockout mice. Mice lacking MMP-7 showed reduction in
intestinal tumorigenesis (35)
, and MMP-2-deficient mice show reduced
angiogenesis and tumor progression (36)
. MMP-11 knockout mice show
reduced tumorigenesis in response to chemical mutagenesis (37)
.
The expression of MMPs in tumors is regulated in a paracrine manner by
growth factors and cytokines secreted by tumor infiltrating
inflammatory cells as well as by tumor or stromal cells; recent studies
have suggested continuous cross talk between tumor cells, stromal
cells, and inflammatory cells during the invasion process (4,
5,
16,
19)
. Accordingly, tumor cell-derived factors that increase the
expression of several MMPs in stromal cells have been purified. One of
these factors, EMMPRIN, was originally identified from the cell
membrane of LX-1 carcinoma cells and has been shown to induce the
expression of MMP-1, MMP-3, and MMP-2 by fibroblasts (38)
. Furthermore,
the ability of MMPs to degrade and inactivate interleukin-1ß
(IL-1ß) (39)
and cleave the tumor necrosis factor (TNF-)
precursor to a biologically active form (40)
, as well as the capacity
of TIMP-3 to inhibit activation of TNF- (41)
, indicate that besides
degrading ECM components, MMPs and TIMPs may also regulate the
availability and activity of inflammatory cytokines at the site of the
tumor invasion. The current view about the role of stromal component in
tumor cell invasion points out that stroma actually participates in
tumor progression and that stromal fibroblasts can cause tumorigenic
conversion of epithelial cells (42,
43)
. Evidence for the interplay
between tumor cells and the stroma has also been provided from findings
that in the early stages of epithelial malignancy, when the basement
membrane is still intact, angiogenesis is observed in stroma (44)
. In
many tumors, stromal cells also actively express ECM components, e.g.,
type I collagen, indicating that the surrounding stromal compartment of
malignant tumors is undergoing extensive tissue remodeling and that
many tumors induce formation of a specific tumor stroma present both at
the primary and metastatic sites (45,
46)
.
As mentioned above, tumor cell invasion is a multistep process that
involves tumor cells, fibroblasts, and inflammatory cells, and all
these cell types have been shown to produce ECM-degrading proteases (4,
5). In vivo expression of MMPs is localized to both tumor
and stromal cells at the invading margin of the tumor, providing a
mechanism for highly concerted degradation of ECM (4,
5,
15,
19)
.
Increased expression of MMP-1 has been observed in lung carcinomas (47,
48)
, squamous cell carcinomas of the head and neck (17)
, and colorectal
tumors, in which expression of MMP-1 correlates with poor prognosis
(49)
. In all these tumors the most abundant expression of MMP-1 was
observed in the stromal cells, but significant expression of MMP-1 has
also been observed in stromal or intratumoral endothelial cells.
Expression of MMP-1 was also observed in cancer cells located at the
invasive edge of head and neck squamous cell carcinoma (SCC) cells
adjacent to stromal cells, whereas the expression of MMP-13 is
predominantly confined to tumor cells at the invading margin; in a
subset of tumors, significant expression of MMP-13 was also observed in
stromal fibroblasts (17)
. This is in contrast to breast carcinomas, in
which MMP-13 is expressed primarily by stromal fibroblasts (15,
16)
.
Both gelatinases MMP-2 and MMP-9 are abundantly expressed in various
malignant tumors (4,
5,
15)
. MMP-9 is mainly expressed by malignant
cells, but also by inflammatory cells, including tissue macrophages and
eosinophils (15,
19,
50)
. Immunohistochemical analysis of the
expression of MMP-2 has shown increased abundance of the immunoreactive
enzyme at the neoplastic epithelium of breast, colon, and gastric
adenocarcinomas; however, in contrast to the immunohistochemical
results, in situ hybridizations show an increase in MMP-2
mRNA mainly in the stromal component of tumors. This discrepancy may be
explained by recent findings that latent MMP-2 binds to the cell
surface of malignant cells by interaction with cell membrane-associated
MT1-MMP (22,
23)
. Furthermore, activated MMP-2 binds to the cell
surface of invasive melanoma and endothelial cells via vß3
integrin (51)
. Recent findings that both MMP-2 and MT1-MMP mRNAs are
also expressed by stromal fibroblasts of human vulvar, breast, lung,
and head and neck carcinomas (15,
19,
52)
suggest that MT1-MMP binds to
and activates MMP-2 at the cell surface of fibroblasts. In SCCs, MMP-2
and MMP-9 may also be activated by MMP-13, which in turn may be
activated by MT1-MMP (19,
21)
. These kinds of activation cascades may
be required to ensure cooperation between tumor and stromal cells; they
greatly enhance cell surface localized proteolytic activity in
vivo and indicate that selective inhibition of one MMP may be
sufficient to block activation cascade and ECM degradation during tumor
invasion.
In summary, enhanced expression of different MMPs in cancer tissue and
experimental in vitro data indicate an essential role for
MMPs in tumor cell invasion. Although the expression of MMPs in
malignancies has been studied widely, the specific role of distinct
MMPs in the progression of cancer may be more complex than has been
assumed. For example, it has recently been shown that MMP-3, MMP-7,
MMP-9, and MMP-12 can generate angiostatin from plasminogen, indicating
that their expression in peritumoral area may in fact serve to limit
angiogenesis and thereby inhibit tumor growth and invasion in
vivo (53
54
55)
. This recent view about the role of stromal cells in
the progression of cancer cell growth and metastasis is particularly
interesting, and additional studies about the regulation of MMP gene
expression and activity in in vivo malignancies are clearly
needed to understand the role and regulation of MMPs in tumor cell
invasion.
Although TIMPs potently inactivate MMPs in vitro, the role
of TIMPs in the regulation of MMP activity and cancer progression
in vivo is unclear. Earlier studies have shown decreased
expression of TIMPs at the site of tumor invasion, but a positive
correlation between TIMP and MMP expression and poor prognosis of
malignant tumors has recently been shown (see ref 28
). The
contradictory roles of TIMPs may result from the bimodal function of
TIMPs as inhibitors of MMPs, but also as key players in the cell
surface-targeted MMP activation cascades (28)
. Another aspect to the
role of TIMPs and MMPs in tumor growth is the observation that
overexpression of TIMP-3 can induce apoptosis in normal and malignant
cells (30,
56,
57)
, providing evidence that MMP activity in general may
be important not only for invasion, but also for the survival of
malignant cells.
|
REGULATION OF MMP GENE EXPRESSION |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
42,
43
). The promoter regions of inducible MMP
genes (MMP-1, MMP-3, MMP-7, MMP-9, MMP-10, MMP-12, MMP-13) show
remarkable conservation of regulatory elements (Fig. 2
), and their expression is induced by growth factors, cytokines, and
other environmental factors such as contact to ECM (1
2
3,
58,
59)
.
Increased expression of other MMPs, the promoter regions of which do
not contain conserved cis elements (MMP-2, MMP-11), has also
been observed in malignancies, indicating overlapping mechanisms in the
regulation of the expression of these genes.
|
Role of AP-1 transcription factors in the regulation of MMP gene
expression
A single AP-1 element, which binds members of the AP-1
transcription factor family, is found at approximately -70 in the
promoter region of each inducible MMP gene (Fig. 2)
. AP-1 transcription
factors are leucine zipper proteins that bind to a consensus DNA
sequence (5'-TGAG/CTCA-3') as a dimeric complex (60,
61)
. Three
distinct members of the Jun family (c-Jun, JunB, and JunD) and four
members of the Fos family (cFos, Fra-1, Fra-2, and FosB) have been
characterized. Members of the Jun family can bind to DNA either as
Jun/Jun homodimers or Jun/Fos heterodimers, whereas Fos proteins do not
form homodimers and bind DNA in the absence of Jun (60,
61)
. Different
AP-1 dimers bind DNA with different affinities, which is thought to be
at least partly responsible for the diverse biological effects of
distinct AP-1 complexes.
The proximal AP-1 element located between -72 and -66 plays a major
role in the transcriptional regulation of MMP-1 gene expression, as
shown by the results that mutation of this element dramatically reduces
the basal activity and responsiveness of the MMP-1 promoter to external
stimuli (62
63
64)
. Although the importance of additional AP-1 elements
found in the promoters of MMP-1, -3, and -9 is not clear, it has been
shown that in the rabbit MMP-1 promoter, they bind Fos and Jun
containing AP-1 dimers and partially mediate the effect of phorbol
ester (59)
. Several cotransfection studies have shown that
overexpression of Jun and Fos proteins enhance MMP-1 promoter activity
and, in contrast, that expression of antisense mRNAs for
c-jun abrogates the induction of MMP-1 gene expression.
Furthermore, it has been shown that simultaneous induction of
c-jun and junB mRNAs precedes induction of MMP-1
gene expression by several types of stimuli (for references, see refs 1,
58,
59
). c-Jun is capable of inducing the minimal MMP-1 promoter
activity as a Jun/Jun homodimer, whereas induction of minimal (76 bp)
MMP-1 promoter by JunB requires the presence of several AP-1 elements
or simultaneous expression of c-Fos (65,
66)
. However, in HT-1080
fibrosarcoma cells, in which c-Jun expression is not induced by tumor
promoter okadaic acid, JunB containing AP-1 complexes mediate the
activation of the full-length (3.8 kb) MMP-1 promoter (67)
. We have
also observed that in NIH-3T3 fibroblasts, in which overexpression of
JunB alone does not stimulate minimal MMP-1 promoter, JunB is a potent
activator of full-length MMP-1 promoter (64)
. These findings indicate
that c-Jun is an independent activator of MMP-1 gene expression,
whereas trans-activation of MMP-1 promoter by JunB, and
possibly by other AP-1 transcription factors, is dependent on the
interaction with other transcription factors binding to additional
regulatory cis-elements in the 5'-flanking region of the
MMP-1 gene.
Recent studies with c-Fos knockout mice reveal that c-Fos is required
for malignant and invasive progression of skin papillomas and for
induction of mouse MMP-3 and MMP-13 gene expression by platelet-derived
growth factors and epidermal growth factor, but not by phorbol ester
(68,
69)
. In addition, overexpression of c-Fos in transgenic mice under
an interferon-inducible promoter was shown to induce expression of
mouse MMP-13 in thymus, spleen, and predominantly in bone, revealing
that the capacity of c-Fos to activate mouse MMP-13 gene expression
in vivo is cell type specific (70)
. The finding that neither
mouse MMP-9, MMP-3, nor MMP-10 expression was affected by c-Fos
overexpression suggests that c-Fos differently regulates expression of
distinct MMP genes in vivo and that the AP-1 element alone
does not determine the inducibility of the MMP promoter by c-Fos
in vivo (70)
.
Little is known about the regulation of AP-1 activity in malignant
tumors in vivo. Increased expression of AP-1 genes has been
reported during growth of malignant tumors, but there is no consistent
pattern of AP-1 complexes that would serve as a marker for increased
invasion or malignancy. Decreased expression of c-jun, junB,
and c-fos genes was observed in human lung carcinomas as
compared with normal tissue (71)
. As the expression of different AP-1
components at sites of MMP expression during tumor invasion is not
known, determination of the specific AP-1 complex pattern responsible
for induction of MMP gene expression in vivo would be
important in developing specific approaches to prevent tumor invasion
and metastasis.
ETS transcription factors in the regulation of MMP expression
Conserved PEA3 elements that bind members of ETS transcription
factors have also been found in all inducible MMP promoters; with the
exception of the MMP-12 promoter, they are located adjacent to at least
one AP-1 element (Fig. 2)
. ETS transcription factors are
helix-turn-helix proteins that share a modular domain structure
characterized by a highly conserved ETS domain, which recognizes the
purine rich PEA3 element A/CGGAA/T (72)
. Although ETS proteins have
been shown to trans-activate artificial promoter constructs
containing only the PEA3 element, ETS proteins do not usually dimerize
and bind to DNA alone, but prefer to form complexes with other
transcription factors, e.g., AP-1, for which they function as
coactivators (72,
73)
.
Expression of ETS-1 colocalizes with several MMPs to the stromal
fibroblasts adjacent to the invasive edge of several types of tumors
(47,
48,
74)
. ETS-1 expression has also been detected during
angiogenesis in human embryos, in granulation tissue, and during tumor
vascularization (75)
. Furthermore, overexpression of the ETS-related
transcription factor E1A-F, which activates MMP-1, MMP-3, and MMP-9
promoters, has been shown to induce an invasive phenotype in MCF-7
cells (76)
. In addition, overexpression of ETS-1, ETS-2, and PEA3
enhances the activity of full-length MMP-1, MMP-3 and MMP-9 promoters
(64,
77,
78)
. The functional interplay between AP-1 and ETS factors in
the regulation of MMP gene expression has been recently shown by us and
others. Our results indicate that expression of structurally distinct
ETS factors (ETS-1, ERGB/Fli-1, PU.1) differentially modulate
AP-1-dependent MMP-1 promoter activation (64)
. ETS-1 markedly enhanced
the stimulation of MMP-1 promoter by both c-Jun and JunB, whereas
ERGB/Fli-1 alone does not regulate MMP-1 gene expression and
potentiates only the effect of c-Jun. In contrast, PU.1 overexpression
potently blocked activation of MMP-1 promoter by overexpression of
c-Jun and JunB or by stimulation by phorbol ester and okadaic acid
(64)
. ETS factor Erg was recently shown to physically interact and
cooperate with Jun/Fos in the up-regulation of MMP-1 promoter activity
(61)
. These results indicate that in vivo interactions
between distinct transcription factors may modulate the response of MMP
promoters in situations where simultaneous induction of the expression
of both ETS and AP-1 factors occurs, such as tumor cell growth and
invasion.
|
MITOGEN-ACTIVATED PROTEIN KINASES |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
80
81
). Three MAPK pathways have
been characterized in detail (see Fig. 3
). The ERK1,2 pathway (Ras->Raf->MEK1,2->ERK1,2) is activated by a
large variety of mitogens and by phorbol esters, whereas the JNK/SAPK
(MEKK1–4->MKK4/MKK7->JNK/SAPKs) and p38 (TAK1->MKK3/MKK6->p38s)
pathways are mainly stimulated by environmental stress and inflammatory
cytokines (Fig. 4
). Other stress-activated MAPKs, including ERK5 and ERK6, have been
characterized recently (80,
81)
.
|
|
It is thought that the balance between distinct MAPK pathways regulates
cell growth, differentiation, survival, and death. Constitutively
active mutants of Raf-1 and MEK1 have been shown to transform
fibroblasts in vitro (82,
83)
, and in vivo
activation of ERK1/2 pathway has been observed in renal and breast
carcinomas (84,
85)
. Furthermore, the full transformation capacity of
oncogenic Ras has been shown to require activation of stress-activated
MAPKs, indicating functional interplay between mitogen and
stress-activated MAPK pathways during the transformation process (86)
.
Transcriptional activity and c-Jun protein stability is increased by
phosphorylation of serines 63 and 73 by JNK/SAPK, whereas
phosphorylation of threonines 91 and 93 leads to enhanced c-Jun
DNA binding (see ref 79
). Activation of ERK1,2 was recently shown to
induce c-Jun expression and phosphorylation, indicating cross talk
between ERK1,2 and JNK/SAPK pathways in the regulation of c-Jun
activity (87)
. Although p38 MAPK does not directly phosphorylate c-Jun,
it contributes to enhanced AP-1 activity by activation of transcription
factors (ATF-2, MEF2C, Elk-1, SAP-1) that up-regulate c-jun
and c-fos promoter activities (79)
. Phosphorylation of c-Fos
by p90rsk and ERK2 stabilizes the c-Fos protein
and results in increased trans-activation and transformation
capacity of c-Fos. Furthermore, the activity and DNA binding of ETS
transcription factors SAP-1 and Elk-1, which participate in
c-fos promoter activation through serum responsive element,
is regulated by ERK1,2, JNK/SAPK, and p38 pathways (80,
81,
88)
.
Activation of Ras has been shown to trans-activate ETS-1 and
ETS-2, and ERK1,2 and JNK/SAPK pathways have been shown to coordinately
mediate activation of ETS factor PEA3 by Ras (89)
.
Taken together, both AP-1 and ETS transcription factors are subject to
regulation by mitogen- and stress-activated MAPK signaling pathways.
Recent identification of specific adapter proteins that physically
connect kinases of MAPK cascades suggests the existence of highly
specific MAPK cascades that are activated in specific situations (90,
91)
. These signal transduction cascades allow strict control of
amplification, feedback, cross talk, and branching of the initial
signals triggered from the cell membrane resulting in a precise
regulation of gene expression in situations such as tumor cell
invasion.
Role of MAPKs in the regulation of MMP gene expression
As discussed, the role of phosphorylation in the regulation of the
AP-1 and ETS transcription factors is fairly well established. However,
the role of distinct phosphorylation-dependent signaling pathways in
the regulation of MMP gene expression has been determined only
recently. Increased serine/threonine phosphorylation as a result of
inhibition of serine/threonine phosphatase PP2A by tumor promoter
okadaic acid induces expression of MMP-1 and MMP-3 at transcriptional
level (67,
92)
. Evidence for the role of MAPKs in the transcriptional
regulation of MMP gene expression was shown by results indicating that
blocking the ERK pathway abrogated Ras, serum, and TPA-elicited
induction of a minimal promoter construct containing 72 bp of MMP-1
promoter, harboring the proximal AP-1 element (93)
. Furthermore,
overexpression of dominant negative MEK was shown to block
insulin-elicited induction of a reporter construct under the control of
an AP-1 element from the MMP-1 promoter (94)
. Recently, overexpression
of dominant negative Raf-1 was shown to block oncostatin M elicited
induction of a promoter construct containing copies of oligonucleotides
encompassing the AP-1 and an earlier unidentified STAT binding site of
the MMP-1 promoter in front of a heterologous promoter (95)
. Protein
kinase C (PKC) has been shown to activate ERK1,2 signaling pathway, and
it was recently shown that selective overexpression of PKC isoforms
PKC
, PKC
, and PKC
activate the full-length MMP-1 promoter
(96)
. Taken together, these recent results indicate that
mitogen-activated ERK1,2 pathway would be the major activator of MMP-1
gene expression.
In normal skin fibroblasts, enhancement of MMP-1 expression by TNF and
the IL-1-induced lipid second messenger ceramide is mediated by
coordinate activation of ERK1,2, JNK/SAPK, and p38 MAPK pathways (97)
.
Our recent results also show that activity of ERK1,2, JNK/SAPK, and
particularly p38, is required for induction of MMP-1 expression and
promoter activity by the tumor promoter okadaic acid (98)
. The
involvement of stress-activated MAPK pathways in the regulation of MMP
gene expression is also supported by findings that inhibition of p38
activity by specific chemical inhibitor SB 203580 blocks the IL-1
elicited induction of MMP-1 and MMP-3 expression in human fibroblasts
and vascular endothelial cells (99)
. Further evidence for the
importance of stress-activated pathways in migration and invasion of
fibroblasts is provided in our findings showing that induction of
MMP-13 expression by contact to 3-dimensional collagen is dependent on
the activity of p38 MAPK (100)
.
The role of MAPKs in regulation of MMP expression has also been
examined in malignant cells. In addition to MMP-1 and MMP-3, increased
transcriptional activity of the MMP-9 promoter in Ras-transformed
OVCAR-3 cells was shown to be mediated by MAP kinases, but blocking the
ERK pathway was not sufficient to abrogate the signal, suggesting a
role for stress-activated MAPK pathways as mediators of effect of Ras
(78)
. On the other hand, induction of MMP-9 promoter activity by
oncogenic Ras in SCC cells was abrogated by blocking either the ERK1,2
or JNK/SAPK pathways, indicating cell-specific differences in the role
of distinct MAPKs in the regulation of MMP promoter activity (101)
. The
first evidence for the role of MAPKs in the regulation of cancer cell
invasion was recently demonstrated by showing that inhibition of p38
activity by SB 203580 blocks phorbol ester-elicited induction of MMP-9
expression by SCC cells as well as their invasion through Matrigel
(102)
.
Taken together, these results indicate cooperation of multiple MAPK
pathways in the regulation of MMP transcription in response to
different signals. In addition to growth factors and cytokines,
invasive cancer cells and stromal cells at the invading edge of the
tumor are under physical stress, such as hypoxia and hyperthermia.
Based on the current view on the importance of stress-activated MAPKs
in the regulation of MMP expression, it could be assumed that
activation of AP-1 and ETS transcription factors by these MAPKs is
responsible for maximal activation of MMP expression in stress
conditions present in the tumor and that, in addition to its role in
the gene regulation, activation of ERK1,2 pathway may be important for
survival of cancer cells by protecting them from the programmed cell
death caused by stress-induced JNK/SAPK activation (103,
104)
. It will
be of great interest to determine whether invasion specific MAPK –>
AP-1/ETS routes exists in vivo and whether inhibition of
these pathways could serve as a therapeutic target to specifically
prevent tumor cell invasion and metastasis.
|
MMPs AS TARGETS TO INHIBIT TUMOR CELL INVASION |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
). In this respect, it is
important to understand the mechanisms of the regulation of MMP
expression in malignant tumors in vivo. For example, it was
recently shown that overexpression of transforming growth factor ß
(TGF-ß) under keratin 6 promoter inhibited formation of benign skin
tumors in transgenic mice but enhanced progression to invasive
carcinomas in mouse skin multistage carcinogenesis model (105)
,
indicating that different cell populations may alter their response to
factors such as TGF-ß. Furthermore, detailed characterization of
growth factor and cytokine receptors responsible for activation of
signaling pathways would help in designing receptor antagonists and
neutralizing antibodies to prevent the action of these factors. For
example, we have previously shown that TNF- stimulates MMP-1
expression in fibroblasts through 55 kDa TNF receptor (106)
, suggesting
targeting of this receptor pathway, as has been shown in organ cultures
(107)
Regarding inhibition of MMP expression at the level of kinase pathways,
it is possible that selective chemical inhibitors for distinct
signaling pathways (e.g., MAPK, PKC) will soon be available for initial
clinical trials. It is therefore important to identify growth and/or
invasion specific signaling cascades that could serve as targets for
chemical inhibitors or dominant negative kinase mutants and antisense
oligonucleotides (108)
. Overexpression of selective dual-specificity
MAPK phosphatases shown to prevent MMP promoter activation (97,
98,
101)
could also be used as a novel strategy to prevent activation of
AP-1 and ETS transcription factors and MMP promoters in
vivo.
Interactions between members of different transcription factors
increases capability of fine-tuning the transcriptional regulation of
MMP promoter activity. It is possible that constitutive expression of
specific transcription factors as a result of transformation may
modulate the response of MMP promoter to extracellular signals. For
example, overexpression of several ETS factors has been observed in
hematological malignancies as a result of proviral insertion into the
ETS gene or by chromosomal translocation, possibly resulting in
potentiation of AP-1 induced up-regulation of MMP expression, as shown
recently in vitro (64,
77)
. It was recently shown that
peritoneally injected, double-stranded AP-1 oligomers potently prevent
AP-1 dependent activation of MMP gene expression in mouse arthrosis
model (109)
, indicating that it is possible to block transcriptional
activation of MMP gene expression in vivo by eliminating
binding of the transcription factors. Also, novel retinoids that
potently inhibit cell proliferation have been shown to selectively
abrogate AP-1 dependent gene expression (110)
. Besides the treatment
strategies targeted to inhibit MMP promoter activation, degradation of
MMP mRNA by antisense RNA or ribozyme techniques may also provide
efficient and specific tool to prevent tumor cell invasion as shown
in vitro and in vivo (111,
112)
.
Inactivation of MMPs in pericellular space by either overexpression of
TIMPs or synthetic MMP inhibitors may prevent initiation of MMP
activation cascades and excessive ECM degradation. Several small
molecule MMP inhibitors are currently in phase I/II clinical trials,
and a hydroxamate MMP inhibitor, marimastat, is already entering phase
III trials in a number of solid tumors (113,
114)
. Although initially
aimed at inhibiting tumor invasion and metastasis and inhibiting tumor
growth by inhibiting angiogenesis, it is possible that MMP inhibitors
also directly inhibit tumor growth, as inhibition of MMP activity by
TIMP-3 and tetracyclines has been shown to induce apoptosis in
malignant cells (30,
56,
57,
115)
.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
|
REFERENCES |
|---|
|
TOPABSTRACTMATRIX METALLOPROTEINASESEXPRESSION OF MMPs IN...REGULATION OF MMP GENE...MITOGEN-ACTIVATED PROTEIN...MMPs AS TARGETS TO...REFERENCES |
|---|
and doxycycline. J. Biol. Chem. 272,31504-31509. Nature (London) 385,733-736[Medline]
converting enzyme (TACE) is inhibited by TIMP-3. FEBS Lett 435,39-44[Medline]
vß3. Cell 85,683-693[Medline]
B,
nuclear factor-
